Action in Context 9783110898798, 9783110188936

The book illustrates the concept of action in three different contexts - the justification of actions, people's lif

170 88 1MB

English Pages 398 [404] Year 2007

Report DMCA / Copyright


Recommend Papers

Action in Context
 9783110898798, 9783110188936

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

Action in Context

Action in Context Edited by

Anton Leist

Walter de Gruyter · Berlin · New York

앝 Printed on acid-free paper which falls within the guidelines 앪 of the ANSI to ensure permanence and durability. Library of Congress Cataloging-in-Publication Data A CIP catalogue record for this book is available from the Library of Congress.

ISBN 978-3-11-018893-6 Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at 쑔 Copyright 2007 by Walter de Gruyter GmbH & Co. KG, 10785 Berlin, Germany. All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage or retrieval system, without permission in writing from the publisher. Printed in Germany Cover design: Martin Zech, Bremen. Printing and binding: Hubert & Co GmbH & Co KG, Göttingen.

Contents Anton Leist Introduction: Through Contexts to Actions . . . . . . . . . . . . . . . .


P. M. S. Hacker Thought and Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


I. Acting for Reasons Frederick Stoutland Reasons for Action and Psychological States . . . . . . . . . . . . . . . .


Thomas Schmidt: Reasons for Action without Metaphysics? . . .


Maria Alvarez The Causalist/Anti-causalist Debate in the Theory of Action: What It Is and Why It Matters . . . . . . . . . . . . . . . . . . . . . . . . .


Philipp Huebl: In Defence of Causalism . . . . . . . . . . . . . . . . . .


Rowland Stout Two Ways to Understand Causality in Agency . . . . . . . . . . . . . .


John Bishop: On the Importance of Reconciling Two Models of Causality in Agency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


Todd Lekan Actions, Habits, and Practices . . . . . . . . . . . . . . . . . . . . . . . . . . .


Norbert Anwander: The Reach of Habit . . . . . . . . . . . . . . . . . .




II. Action, Persons and Life J. David Velleman What Good is a Will? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


Neil Roughley: On the Ways and Uses of Intending: Lessons from Velleman’s Bratman Critique . . . . . . . . . . . . . . . . .


Stefaan E. Cuypers Personal Identity and Agency: Towards Analytical Personalism .


Marc Slors: Looking for the Real Enemy . . . . . . . . . . . . . . . . . .


Rdiger Bittner The Units of Living . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


Holger Baumann: Making Sense of Ourselves . . . . . . . . . . . . . .


III. Action and Epistemology Jennifer Hornsby Knowledge in Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


Katia Saporiti: A Seeming Solution to a Seeming Puzzle in Explaining Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


Anton Leist Cognition and Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


Joachim Schulte: Notes on Distinctions . . . . . . . . . . . . . . . . . . .


Christopher Hookway Action and Inquiry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


Stephen Hetherington: Knowledgeable Inquiry . . . . . . . . . . . . .


Name Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .


Introduction: Through Contexts to Actions Anton Leist 1. Action in Context: Why not Holism in the Theory of Action? In one of his last articles, D. Davidson pointed out the remarkable fact that in the history of (Western) philosophy there have been only two periods of pronounced interest in what actions are—as opposed to the usual, moral interest in how we ought to act—, namely in Aristotle’s time and the present day (Davidson 2005, 277). This observation could be complemented by another: it is perhaps no coincidence that during Aristotle’s period of thought, as in current philosophy, traditional enthusiasm for metaphysics and epistemology was and is counterbalanced by a critical distance towards these disciplines, so that in the resulting philosophical pluralism, in addition to the pervading interest in all aspects of knowledge and the mind, it was and is possible to take an extensive look at actions. In actual fact, a wealth of literature and discussion covering an independent discipline known as ‘action theory’ has been open to ‘analytic philosophers’ for decades, but the fact that this discipline is a speciality alongside many others already underlines the theoretically ambivalent and uncertain role of action theory within the context of philosophy taken as a whole. The traditional quibbles concerning language, mind and consciousness continue to loom more heavily than such small-scale quandaries surrounding why and how human beings act. If it is true that action theory has been pushed into a specific box by other disciplines, it is also true that it has not been overwhelmingly enthusiastic to enlarge upon its specific topic. For Davidson and the representative parts of analytic action theory he has influenced, “what actions are” is the central issue which upon second glance is typically not as harmless as it first appears. What actions are, or what constitutes an action, is usually viewed as the central question of action theory, usually phrased with the help of Wittgenstein’s formulation: “What is left over if I subtract the fact that my arm goes up from the fact that I raise my arm?” (PU, I. §621) Many authors have accepted this question as the central


Anton Leist

question1 and quite a lot have taken up Davidson’s agenda and tried to elaborate actions under his central idea that belief and desire simultaneously rationalise an action and causally produce or explain it.2 Davidson himself admitted that his proposal contained a number of problem spots and open angles, but he continued to support it right up until he ceased to write (see Davidson 2001). Davidson and others have connected the standard model of action theory with two insights in particular, both adopted from Wittgenstein and Anscombe: firstly, that actions have to be identified via descriptions and that descriptions of one and the same action form an open set; secondly, that the subjective states of the agent have to rationalise an action and that, correspondingly, the action can be explained as rationally comprehensible. Davidson’s specific achievement is usually regarded as having successfully reconciled these two requirements with a third which previously, in the intellectual climate of Ryle and Wittgenstein, had appeared to be incompatible with them: the hypothesis of a causal relationship, as opposed to a logical-conceptual one, existing between the motives behind an action and the action itself. Until today, the attractiveness of this project, started by Davidson in 1963, lies in its awe-inspiring claim to facilitating a connection between these three requirements: dependency on description, rationalisation and causality. Enthusiasm is being dulled, however, by the strange narrowness of this agenda, already apparent in the ‘central question’. Whereas, in his linguistic theory, Davidson supports an unrestricted holism, with individual sentences acquiring their meaning from the meanings of all the other sentences of a language (Davidson 1984, essays 9 – 12), actions are grasped and understood more narrowly. Despite fundamentally conceding a mutual dependency between meanings, desires and beliefs, and despite accepting the dependency of actions upon descriptions, he 1


For variations on the ‘central question’ of action theory, see M. Brand: “to adequately specify the nature of the proximate cause of action” (Brand 1984, 33). Or H. Frankfurt: “The problem of action is to explicate the contrast between what an agent does and what merely happens to him, or between the bodily movement that he makes and those that accur without his making them.” (Frankfurt 1978, 157). Also Velleman 2000, chap. 1. The belief-desire model of action plays a dominant role not only in action theory, but also in the explication of practical reasons (Anscombe, von Wright, M. Smith) and in metaethics (Hare), as well as in the social sciences. Its historical forebears are Plato (Phaidros, Politeia), Aristotle (De Anima 433a, NE Book 3), Hobbes (Leviathan, chap. 6) and Hume (Treatise II.3.iii).

Introduction: Through Contexts to Actions


remained in consistent pursuit of an action model which was more reductive than holistic. Instead of viewing both the description and the rationality of actions as dependent upon the description of many other things (the agent, his self-comprehension, his attitude, his knowledge, his history, his circumstances, his social position and role, his society, his tradition, etc.), comprehension of actions is reduced to a comprehension of descriptions of individual events. Not only is this ontologically emaciated version of action theory barely compatible with Davidson’s own increasingly detailed hermeneutic agenda (Malpas 1992), it is also no longer tenable in the light of numerous objections which have been put forward over the past few years. What happens when we act can certainly be described with the help of psychological states and events, but also equally well with the help of attitudes, memories, plans, phases in agents’ lives and their life-stories. It can be explained by referring to the immediate beliefs and desires of the agents, but also by referring to the beliefs and knowledge of the community, the argumentative standards of those communicating. It can be viewed both in the light of individual characteristics and in the light of social expectations and norms, under the individual claim of convictions, but also under the aspect of normative opinions and collective beliefs. Actions can be dealt with as individual actions, but also as parts of collective actions. The intentions which are connected to a certain action can be made clear enough by that individual action, but in some circumstances may not be decipherable until a comprehensive tale has been told. Why should we think that all actions are equal and, more to the point, that they are all equally simple? Like all human manifestations, acting can be simple, but it can also be very complicated, and the advantage of holism lies in its acceptance of this complexity, presupposed even for the most basic elements of acting. One could only speak of a ‘holistic’ theory of action if the dependencies hinted at above were to become visible, whereas the standard theory of action has done its best to suppress them. The ‘standard theory’ of action explains individual actions by using the rationalising relation between beliefs, desires and (in part) intentions with action-events. Current debates pursue various goals with regard to the standard theory, something I shall make more explicit in the next sections by giving an overview of diverse problems. Such critical discussions can, or so it appears, have the following three effects. Corrections intend to reconstruct in an alternative way the interior structure of acting with regard to its rationality, causality or intentionality


Anton Leist

assumptions, whilst retaining the standard model itself. A revision of the model is imminent when some of the taken views on acting—for example views concerning the relations between reasons and actions or the agent and his action—withstand elaboration. Expansions are possible when actions are studied in their relationships to various additional contexts which are relevant for the comprehension of those individual actions but unclear in their functional role so far. In my opinion, the debate within action theory is currently in the second phase. Perhaps the contributions to this book could elucidate this and possibly encourage a transition into the third phase. ‘Action in Context’ suggests that something which is already known— an action—is, as it were, experimentally pursued and illuminated in contexts and correlations independent of this action, partly due to interest in the action and partly due to interest in the contexts and correlations. Some devote themselves to study, for example, how normally valid assumptions within the standard action theory change when the environment changes (e.g. Bratman 1999a). Others see themselves in Dewey’s pragmatic (and perhaps Darwinian) tradition, observing actions basically as reactions to pre-given situations, and thus never being without context (e.g. Stoutland 1998; 2001; Bittner 2001, 68 – 71). These attempts are particularly fruitful when they are made in a conscious effort to revise the fundamental concepts of action themselves and to formulate them more openly.3 If Davidson’s holism for descriptions is correct and can be transferred equally to actions and the meaning of sentences in general, these attempts must lead to a continuing reformulation of action descriptions, individually and in general. Caught up in the hermeneutic circle, one can only speak of the ‘context’ of acting using a previously identified action—but at the same time the context also facilitates a new and extended description of the action. Even if actions were only materially detectable ‘atoms’, they could in each case be redescribed. Since they (unlike physical atoms) play differently complex functions in different personal and social contexts, their semantic contexts are even more 3

It goes without saying that there is quite some interest among empirical researchers in the social preconditions, both diachronic and synchronic, of single actions and the style of actions. Well-known present-day contributors to such a ‘frame analysis’ are the sociologists H. Esser and S. Lindenberg. Philosophy of action unduly neglects this research because of its still being bewitched by the ‘concept’ of action.

Introduction: Through Contexts to Actions


important. The line of argument followed in this introduction could, rather than using the action/context distinction, equally well be described as a plea for action holism, for externalism or for the social normativity of conditions for action. Rather than setting a fixed idea of action into a fixed idea of context and looking out for interrelations, this is an attempt to open up perspectives on actions, and then perhaps make visible the importance of action for other topics as well. 2. Standard Theory of Action: Basic Questions and Intriguing Answers On the condition of this inevitable circle, our talk of individual actions requires a wealth of background knowledge. One obvious possibility in the light of this dependency would be to describe, or at least to outline, the background necessary to individual actions. It would be reasonable to talk first of agents, rules, expectations, shared knowledge, the body, etc., and then about individual intentions, hand movements, spilled coffee, sunken ships, etc. The analytic theory of action married to Wittgenstein’s ‘basic question’ did not proceed like this, possibly out of fear that this procedure would involve a damaging circle, as well as, incidentally, an unnecessary digression. In actual fact, it is impossible to speak of agents without actions—but the reverse is also true: working with a simple notion of action can be a useful method to enter the circle at a certain point as well, albeit not the only method. What I call the ‘standard theory’ of action, in agreement with much of the literature, delivers just such a model of simplicity, which surely is helpful to make oneself acquainted with the multivarious aspects of actions and acting. The standard theory describes actions as a complex mixture of beliefs, desires and the execution of actions, and explains the latter as causing by rationalising through a ‘primary’ reason (primary in relation to possible other reasons), itself being a compound of belief and desire. This description and explanation of actions is not only outwardly impressively simple, it also excels by creating particularly intelligent answers to really difficult questions. And it is capable of linking different ontological, theoretical and practical dimensions.4 Just how good the 4

Because Davidson (1980) raised this synthetic view most strongly and put forward the most innovative contributions, in the following treatment I shall refer primarily to the debate he triggered. The terminology used thus far has already been taken from Davidson. Within the framework of his writing,


Anton Leist

answers are, especially in Davidson’s original version, can only be properly appreciated after taking a brief look at the situation before Davidson’s programmatic article appeared. The difficulties surrounding action theory prior to Davidson were primarily located in two areas: firstly a lack of clarity regarding the roles of desire and intention, and secondly scepticism regarding the function of causality within action (see Thalberg 1985, esp. 9 – 16). As the last in a long series of attempts to unite mind and body, intention and causality, desire and control, J. St. Mill advocated the notion of action as a combination of desire and physical effect, yet leaving open the mutual relationship between the two. In this situation, three solution strategies became dominant, all three of which encountered massive problems within a short space of time.5 Firstly, what could be called the ‘volition position’. In order to avoid the incoherence of radically different actions, mental as well as physical, Prichard declared only acts of will to be actions, with physical movements being their effects. This radical solution created more difficulties than it dispelled. Behind what we had so far taken to be ordinary actions were now, suddenly, countless previously unnoticed acts of will possessing a dubious status with regard to action. Ryle rendered this apparent in his well-known cutting and witty manner by combining them with attributes (“quickly”, “simply”, “enjoyably”) for normal actions (Ryle 1949, chap. 3; Anscombe 1957, 47, 49). This and later critiques of other versions of Prichard’s proposal—‘tryings’ rather than acts of will (Hornsby 1980; O’Shaughnessy 1980)—showed the reducibility of actions to acts of will to be an awkward solution. Secondly, the position of ‘agent’ or ‘perpetrator causality’. Rather than in a mental act, action is seen as being based on the agent and his


‘standard theory’ always means the basic model of 1963 and its immediate extensions (Davidson 1980, essays 1 – 3). The later revisions, especially in ‘Intending’ (1975), will be mentioned later on. There are controversial opinions concerning the extent in which Davidson grew to deviate from the standard theory, including Stoutland’s 1999 argument for a nearly total split. For other versions of the standard model see Dretske 1988; Seebaß 1993, 144 – 160; Smith 1987. M. Smith is the most important defender of desires as practical reasons. I shall address this in detail in section 5. For overviews and references regarding the historic and contemporary literature on these three positions, see Thalberg 1985 and Seebaß 1993, 49 – 68. For a critique of Hornsby’s and O’Shaugnessy’s elaborations on the will see Seebaß 1993, 256. Donagan 1981 outlines the classical history of action theory up to Davidson.

Introduction: Through Contexts to Actions


bringing forth of the action, a proposal which strangely enough is emerging again in the current debate under the label ‘agency’ (see section 6 below). Defended especially by Chisholm, this proposal encountered the problem of an independent and exotic concept of causality incompatible with event causality. Agent activity neglected in the standard theory but now rediscovered therefore has to be developed without the claim of being an ontologically independent form of causality. Thirdly, there was what came to be called the logical ‘reasons/causes dichotomy’. In brief this refers to the tendency of Wittgensteinians (Anscombe, Melden, von Wright, Stoutland) to focus on action rationality and the reasons justifying the agent, and from this practical perspective to prove that the causal explanation of actions is secondary or even impossible. The dichotomy began to look like a kind of dualism when Davidson was in turn able to introduce the expectation that practical reasons for an action would also have to generate it, or that actions would somehow have to be causally aroused in order to fit into the general causal fabric, a notion which was (and still is) alien to the followers of Wittgenstein. Whether dualism or not, whether causality relevant or not, the question of the relationship between reasons and causes was not satisfactorily answered from their side. Compared to these three problematic strategies and models, the (Davidsonian) standard theory is impressive for its simplicity, rendering it apparently better able to answer a series of central questions pertaining to individual actions. Here is a short outline of these questions and answers. Firstly, when is an action ‘basic’ in the sense that it cannot be broken down into other actions? The answer is: if one is doing something not done by doing something else. Normally, but not necessarily always these basic actions are bodily movements. Bodily movements are events and thus potential candidates for causal relationships. Bodily movements are caused by primary reasons (beliefs/desires) which in turn are not actions. Bodily movements, beliefs, desires, causes and effects thus seem to enter into a conceptually fitting and homogenous relationship—extravagances like acts of will as actions sui generis, intrinsic-causal agent activities or causally isolated reasons become superfluous. Secondly, how is the relationship between intentions and actions to be more precisely understood? The answer is: despite first appearances, intentions are not independent acts or events, but structural features of the relationship between primary reason and action. Intentions are thus indirectly linked to the causal events behind an action, find themselves in


Anton Leist

an indirect causal relationship to the action and not simply in a logical one, as postulated by the ‘logical connection argument’. Not all actions are intentional actions, but non-intentional (unsuccessful, unwanted) actions are dependent upon intentional ones, thus only exist by dint of an intentional action. Thirdly, how does the standard theory explain the possible diversity of action descriptions for one action? If actions are fundamentally bodily movements and events (both itself under a description), they can be described as such in different ways, so that different descriptions present different actions, without the bodily movement thereby causing different simultaneous actions. Bodily movement as a basic action provides, as it were, the material basis to which the most diverse—in principle an unlimited number of—action descriptions (and in this sense also actions) relate. Using a concept taken from J. Feinberg, a bodily movement (flicking a switch) gives rise to an ‘accordion effect’-production of partly intentional actions (turning on a light) and partly unintentional ones (alerting a prowler). Fourthly, can reasons for an action be simultaneously rational and motivating? Are these functions not mutually exclusive? And, if not, how can the rational reason motivate; how can the motivating reason be rational? The answer is: motives are rationalizing when they suitably adhere to beliefs, especially the belief that the action is suited to fulfilling a desire. When linked, beliefs and desires can therefore be both rationalising and motivating. This does not necessarily mean that there are no additional reasons; what is important is that every action arises from a primary reason of a combination of beliefs and desires. The somewhat narrow expression “desires” can be replaced by the term “proattitude”, encompassing all possible forms of a mental propensity to act. A pro-attitude, that is a consenting stance towards an action, seems to be essential to each and every action. And finally, fifthly, how can reasons be causes or be linked to them? Not necessarily through plain identity, but so that reasons are comprehended as linguistic descriptions of those causal and material events which cause another event. The description is the primary reason, the cause is an event of consciousness and the effect is the action as a material event. Primary reasons are thus indirectly both reasons and causes, and they are only causes because they are reasons, viz. can only be reasons for actions because they are causes.6 6

For a more detailed account, which especially demonstrates why action ex-

Introduction: Through Contexts to Actions


Beyond the valid reasons for an action, causality seems to be indispensable, confirmed by the observation that an agent can have several reasons for an action, but acting out of one special reason. Only if the ‘because’ of one or other reason is a causal ‘because’, can one be sufficiently certain as to why one is actually acting or has acted. But then reasons for action can not only be linguistic entities, can not only belong to the ‘logical realm of reasons’, but also actually have to bring about actions causally, within the ‘material realm of events’. Reasons must at the same time be able to justify and explain, and this includes their causal effectiveness. If one accepts my portrayal of the historical situation, in which action theory had to choose between different radical positions (acts of will, agent causality, acausal reasons), the strength of the standard theory becomes obvious, resulting from its reaction to these options. Paramount is its response to the reasons/causes problem and the coherence of causality and action (see Mele 2003). Nevertheless, these gains in presumed coherence come at the price of doubts as to whether the phenomenal wealth of the earlier debate has really been captured by the Gordian knot proposal of the standard theory and as to whether new, even more pressing problems have not been created in the process. These doubts lead, on the one hand, to the objection that this theory portrays actions too simply and, on the other hand, to criticism of the distorting role of causality in the view of actions held by the standard theory. I shall start by depicting these problems more explicitly (section 3), then in turn sketch three attempts at solving them (sections 4 – 6). 3. New Problems A representative overview of the problems and fields of discussion surrounding the standard theory may be thought to include at least the following central points from the standard theory and/or problems typically classified accordingly.7


ACHTUNGREplanations, because they carry on where descriptions leave off, are not causal in the sense of law-dependent explanations but merely empirical-inductive explanations, see Stoutland 1999; Davidson 1987. For illuminating and succinct critiques of the standard model, see Frankfurt 1978; Thalberg 1985; Stoutland 1985; 1998; 2001; Bratman 1985; Mele 1992; Hornsby 1993; Stout 1996; McKann 1998; Velleman 2000; Searle 2001; Stoecker 2002.


Anton Leist

Events: Are all actions really events? Can all actions be localised in space and time? Is it really possible to say: “How long was I acting this time?” or “Wow, that was a short action!” Are acts of omissions, although actions, really events—and not the precise opposite of events? Are mental acts, although actions, really events? When does a decision (clearly an action) begin and when does it end? These questions can be summed up as the ‘event problem’. Causality: Closely linked to the event problem is the unsolved ‘causality problem’, a collective name for all the doubts and objections regarding primary reasons causing actions. Beliefs and desires are attitudes, but attitudes are not events, which is why the standard theory has to postulate the springing into existence of primary reasons for each action, a similarly bold hypothesis to Prichard’s acts of will. Can such effective events really always be found ‘close to’ attitudes? Sometimes the desire for an action emerges at the same time as the action itself, sometimes not until afterwards. To what extent are primary reasons still causes of action if their descriptions only result from the overall action? Is accordingly the assumption of causality not restricted and the notion of a ‘logical connection’ between reasons and actions not reintroduced? Intentions: Is the reductive view of intentions with regard to the beliefs/desires couplet really accurate? Do intentions not presuppose beliefs rather than encompass them? Are intentions not particular, referring to real actions, whereas primary reasons are general, are proattitudes towards a potential action? Do intentions (plans) not themselves create reasons rather than be reasons of their own? These and numerous other objections are evidence of the fact that the reductive concept of intentions within the standard model distorts everyday understanding and does no justice to the relative autonomy of intentions in our descriptions of actions. The overall result is an ‘intentions problem’. Rationality: How is it possible for primary reasons to rationalise? Is this an internal characteristic which requires no further conditions? Desires can explain why somebody has done something; how is it that they are also capable of justifying? Does explaining then coincide with justifying? Is it possible, within our conventional interpretation, that irrational primary reasons possess a rational aspect? Where is the difference when the agent is not capable of seeing his action in any positive light at all? It seems that it is possible to act without reasons, whether rationally or irrationally. Is a playful, spontaneous action ‘without reason’ an irrational action then, or, because it is without reason, even not an action at all? The ‘rationality problem’ of the standard model

Introduction: Through Contexts to Actions


consists in eliminating the incoherence between explaining and justifying, and in finding evidence for a normative aspect within the primary reason.8 Normativity: Admittedly, the standard theory might have no intention of providing a comprehensive conception of practical reasons, instead merely proposing the general condition that primary reasons also have to possess a rational aspect. Even then, and bearing in mind the simultaneous proximity of explaining, it is still unclear how the primary reason can be linked to both explanations and justifications. Explanations require facts; justifications require more than facts. Reasons demand something, are commanding. How can beliefs and desires command? That is the ‘normativity problem’ of the standard theory. Agent activity: Actions are only the actions of an agent if an agent executes them. That is why it is permitted to say that an action has an agent if someone with a primary reason for that action exists. But this condition is not sufficient: the agent must act from this primary reason. Causality alone (in this case agent causality) is not sufficient since unconscious desires (for example) would not constitute actions. The standard theory has a problem with agent activity, that is the active bringing about of an action by an agent. It does not pay this active aspect of actions enough attention. How is this wealth of questions and problems surrounding action theory—partly new, partly old, partly aggravated—to be accounted for? One could possibly read it in the light of an interest in a gradually extended focus on actions, as especially being documented by the inclusion of the agent. Has the focus upon actions really shifted to acting agents and co-agents, compared with the earlier focus on single actions? One could also read the newer debate as documenting the breakdown of the standard theory and the turning back again, at least partly, to earlier positions. The intentionalist-rationalistic tradition prior to Davidson emphasised the subjective, deliberate and practical view of action, while neglecting causal explanations and running constant danger of itself creating inexplicable mental constructions. In contrast, the standard 8

These objections to the rationalising aspect of the standard theory become stronger if the theory is assumed to have an explicit conception of practical reasons. This is true for explicit normative variations such as M. Smith’s (1987), but not for Davidson. The debate about external reasons referred to in the following (see section 5) only demonstrates the limitations of the standard theory; it does not refute its concept of practical reasons because it does not have one.


Anton Leist

theory emphasises the objective, causal and descriptive view of action, while neglecting singular causes of action and running the danger of ignoring the active, practical and rational side of acting. Due to these sustained tendencies within the standard theory, its ontological-causalistic onesidedness, this dichotomy has never been fully overcome.9 Davidson did not really succeed in linking the two strands of his proposed synthesis, the ontological-causal strand and the subjectivepractical strand. It is obviously not the case that a primary reason causes an action by rationalising it. Primary reasons are psychological states, and thus not events, and it is unreasonable that they could be located ‘near to’ events, just as the notion that a mug of beer could be located ‘near to’ my thirst is unreasonable. Not only do causal and psychological talk miss each other; it is also unclear why they should have to link up in the manner aimed at by Davidson. The psychological (and by extension social) demands made of our grasp of actions are all fairly familiar to us if we think about it, and without them we could not manage. The causal requirements may be familiar to some brain researchers and, especially, to metaphysical philosophers. In our everyday lives we might need to know how a desire or a belief has caused an action, but only in an arACHTUNGREguACHTUNGREmentative-coherent manner and not in a physically causal one. We hardly ever say, “If only my primary reason of buying a cow at auction would cause me to drive to Haverford!”, yet far more often “If only I weren’t so unsure as to whether or not to drive to Haverford to buy a cow!”, or something similar. In our everyday lives we do not have much use for ‘purely’ causal relationships, and it is our everyday lives in which actions take place. If we really do not miss much by missing causality, the inevitable consequence is to describe the development of the standard theory as somewhat confused from the outset, insofar as it has been linked to a causal-ontological agenda. With reference to Davidson, it particularly becomes clear why he refused to apply his meaning holism to actions. If actions are also causal events, then they seem to resist generation through description: diseases are not easily healed, avalanches not easily stopped 9

Davidson was well aware of the deficit with regard to a ‘practical’ view of acting, as shown by the following comment in ‘Intending’: “The trouble is that the attitude of approval which the agent has … has been left out. It cannot be put back in by making the premise ‘The agent wants to improve the taste of the stew’: we do not want a description of his desire, but an expression of it in a form in which he might use it to arrive at an action.” (1980, 86)

Introduction: Through Contexts to Actions


by describing them afresh. If actions are not causal events, as I believe, then they are just as mutually networked as sentences and words are. Meaning holism is necessarily also action holism. For both meanings and actions, it is true that they must obey practical and not ontological requirements, and they do. For those who do not share this diagnosis, it may seem astonishing that recent debate, in the main part following on critically from the standard theory, has largely concentrated on the practical-normative aspects of actions, with the causalistic analysis of actions having shifted to the philosophy of the mind, an area which does less clearly answer practical interests. Three more recent sub-debates arise out of corresponding shady areas within the standard theory, namely the neglect of intentions, the one-sided concept of reasons and the omission of the agent. In the following I shall briefly outline the corresponding rehabilitation of three different classes of phenomena, concentrating in each case on one outstanding figure within these debates. Visible within these developments will be, as the minimum, the expansion of our view on actions by including the action-constitutive context. It should then, further, be easier to evaluate the extent to which a decidedly normative and non-causal concept of action is borne out by these developments. 4. Intentions and Commitments The intentions problem within the standard model is conspicuous and has been attacked in many comments. The problem partly springs up quite simply from the standard model’s reductive treatment of intentions and partly from the non-coincidence of intentions and causality, something becoming apparent in the ‘deviant causal relationships’.10 Interpreting intentions reductively means equating them to the existence of beliefs and desires; in so doing, various functions fulfilled by intentions are disregarded. Bratman’s ‘planning theory of intention’ is the most systematic and detailed attempt to develop these functions in a non10 On intentionality see esp. Thalberg 1985, 20 – 24; Stoutland 1985; Bratman 1984; 1985; 1987; Wilson 1980; 1985; McKann 1991; 1998. Also note Davidson: “At one time (about twenty-five years ago, when I wrote ‘Actions, Reasons and Causes’) I thought there were no such states as intending; there were just intentional actions. This was, I now believe, an error.” (Davidson 1987, 106) The problem of deviant causal relationships is one of the motors firing the agency debate. See section 6 below.


Anton Leist

reductive form. The reductive treatment of intentions seems to be descriptively inappropriate, it is true, but it does have, as we saw earlier on, good reasons by reacting to the difficulties surrounding the voluntaristic-rationalistic tradition. Is the planning theory capable of avoiding these difficulties? Davidson himself toned down his initial (1963) aversion to “mysterious acts of will” (1980, 83) by the time he wrote his article ‘Intending’ (1978), in which he proposed a less reductive analysis of intentions, giving intentions a certain autonomy on the basis of the elements inherent in the standard model. In a first step proceeding from acts of ‘pure’ intending, i.e. those not linked to subsequent action, he explains intentions along the lines of a practical syllogism; in a second step, he explains them as ‘all out’ judgments of what is desirable (1980, essay 5). One difficulty in linking a particular individual action not only to primary reasons, but also now to an evaluation as desirable, arises from the particular reference to actions. Desires and what is desirable frequently refer to actions in a special regard, not to special actions. With the conditions of an all out judgment, Davidson wants to guarantee that this judgment refers to a particular, real action. Assuming that were true, how is a similar analysis of intentions meant to refer to future actions, as all out judgments can hardly address particular actions taking place in the future? Reversing the analysis and interpreting (in my opinion more plausibly) the assignation of ‘desirable’ as general in every case, then the opposite question arises of how one can intend a particular, real and now executable action under this general term. Bratman’s planning theory of intention comes into play with regard to the differences between desires and intentions, differences which cannot really be bridged using judgments of what is desirable. For decision-makers facing a Buridan’s ass situation, in which competing alternatives are desired equally and are equally desirable, an intention to seize one alternative is obviously possible even though the desires cancel each other out (Bratman 1985, 219). For rational agents, intentions are guided by an agglomeration principle, whereas desires are not (220). The problems in Davidson’s theory surrounding a particular or general reference to actions mask an additional fundamental difficulty in how to cope with intentions directed at future actions. Davidson tries to comprehend intentions directed at future actions as extended versions of intentions directed at present actions, but he fails because some intentions to act in the future will not be desirable due to unenvisaged circumstances impeding those actions. Bratman wishes, on the one hand, to agree with

Introduction: Through Contexts to Actions


Davidson that beliefs/desires are essential components of rational action, but, on the other, wishes to grasp the role of intentions more decidedly by way of those subfunctions of acting which render an agent a rational agent. The planning theory is a normative reconstruction of intentions by means of their rational function within individual and cooperative plans of action. In a reversal of Davidson’s attempt to reduce future-directed intentions to present-directed ones, the planning theory assumes that present-directed intentions can only be explained against a background of future-directed intentions (Bratman 1987, 4). Unlike intentions directed at typical actions in the present, future-directed intentions raise considerable difficulties. How are intentions which are shaped today for an action taking place tomorrow linked with tomorrow’s action? Do they control that action with a ghostly hand? If they are irrevocable, does that not render them irrational due to their inflexibility regarding possible interim events? Such intentions would then indeed be irrational. These ontological and rational difficulties with future-related intentions miraculously dissolve if they are understood as commitments within plans of action, which have a normative character due to the mere fact of their integration in rational conditions, such as consistency or the coherence of means and ends. The Copernican point, as it were, from which intentions are comprehended as parts of plans, is marked by the term commitment, for only by means of this explicitly normative concept of binding oneself do the various normative functions of intentions within the open network of plans become accessible.11 Intentions involve commitments, which can be described in terms of their various practical functions. A commitment is thus expressed in a concomitant controlling of action, as well as in a sensitivity to reasons. The commitment involved in future-directed intentions ensures that the aim of an action remains firm into the future, as well as that it is rationally adapted to the actual requirements only revealed with time. Via these two elementary functions the actions of an 11 The introduction of this concept is Copernican in that it permits a radical renunciation of Davidson’s individualistic perspective, albeit without, similarly with Kant himself, making clear reference to the social community in a normative respect. See Brandom’s illuminating comments on the difference between Kant and Hegel on this point: Brandom 2002, chap. 7, esp. 216. Bratman does emphasise the social significance of commitment but ignores their normative-constitutive role.


Anton Leist

individual agent can be anticipated by others, which in turn facilitates cooperative action within the collective (Bratman 1987, chap. 2). It seems to be descriptively plausible to say that intentions fulfil such practical functions, but if they additionally become rationally effective, if the rational structure of acting is shifted to the formation and control of intentions, what significance do beliefs/desires then retain? The planning theory differentiates between internal and external perspectives with regard to an action. The internal perspective is the one which is characterised by the intentions already formed and the decisions taken by the agent. These commitments are not only necessary in order to continue pursuing plans of action in a controlled and rational manner into the future; they are also advantageously unburdening for boundedly rational agents, ‘boundedly’ in the sense of H. Simon’s concept of bounded rationality. The external perspective, on the other hand, is the perspective free from the agent commitments, partly stemming from the agent’s beliefs and desires, partly also revising them. The external perspective is accessible to the agent himself, but also of course to potential observers. Both perspectives additionally possess an ideal dimension and therefore contain internal and external norms (109). To summarise thus far, intentions are psychological states capable of forming commitments which are rationally connected to plans of action structured by internal and external norms. As innovative as Bratman’s normativistic reconstruction of intentions is in its various ways, some important points within it remain unexplained, partly as a result of its ongoing attachment to basic ideas of the standard theory. Firstly, the importance of commitment, as well as the connected plans and norms for intentions as mental acts remain unclear, and the same goes, secondly, for the relative weights of the internal and external perspectives. Both points are important for any evaluation of just how ‘Copernican’ Bratman’s proposal is. Intentions and plans are psychological states (29) which contain commitments. Commitments are reason-centered and, as we have seen, contain internal and external norms. Normativity of this kind cannot be fully dissolved into psychological states. According to Bratman, plans are hierarchically structured (29). Yet is a hierarchical structure itself a psychological state? Intentions contain commitments, but are not commitments themselves. But what, then, are intentions independently of commitment? Like Davidson, Bratman attempts to avoid the idea of a “ghostly influence on the future” (110), but leaves open the dependency of psychological states upon normative demands. Comments like “plans are … intentions writ

Introduction: Through Contexts to Actions


large” (29) may not be meant literally, but are still leaning heavily on the mental fundamentalism of the standard theory. Bratman wishes to hold on to the beliefs/desires rationality norm of the standard model, chiefly discussing the rational dependency of intentions by referring to a requirement that intentions be revised in the light of a conflict between the intentions and the beliefs/desires of one and the same agent. All principles of internal consistency, the coherence between means and ends, the external norms, are seen as expression of the rationality of desire (43, 109). The planning theory is thus subject to all the objections which have amassed to the rationality of beliefs/desires and which will be addressed below. The normative conflict between the internal and external perspectives requires an answer which cannot be provided using the standard view of practical rationality. The external perspective is not simply just a potentially better awareness of empirical facts, but a comprehensive social view of the norms governing perceptions, desires and knowledge, being reproduced by the agent’s ongoing work of shaping his assumptions in response to social challenges. “Reason-responsiveness” or “discursivity” would therefore pinpoint the function of the external standpoint better than the enlistment of corrective knowledge regarding desires/facts. If we interpret Bratman’s planning theory a little more radically than he intended, however, it provides an impressive argument for how constituent components of individual actions, like intentions, cannot be comprehended without an open horizon full of agents all acting for reasons and all coordinating their actions. Without reliable commitment cooperation is impossible (Bratman 1999, chap. 5 – 8), but without external norms reliable commitment is equally impossible. Various social contexts challenge the intentional and thus firmly future-directed actions of an individual agent. But social contexts do not only represent challenges (Bratman 1999, chap. 1), facilitating the rational revision of intentions and plans; they are changing scenes for individual actions and a changing stage at the same time.


Anton Leist

5. Practical Reasons: Internal or External? According to the standard theory, beliefs/desires are primary reasons which at least precede all other reasons.12 In this context “reason” can have more than one meaning: “cause”/“motive”/“motivating reason”/“explanatory reason”, or “justification”/“justifying reason”/“normative reason” * *

This bifurcation is meant to distinguish a list of (relatively) descriptive terms from a list of (relatively) normative terms. In addition, the differences between the individual terms are complex and cause for some controversy within the current debate. Due to the ambiguity of most of these terms, the type of action and rationality theory one subscribes to (including the context of these theories themselves) decides the mutual relationships in which one views these terms. At the centre of these differences reside divergent views on what it means to act ‘rationally’ or ‘morally’, what the ‘normative’ resources for these actions are, and in which normative framework actions therefore have to be perceived. Let us try to bring some kind of order to these various terms (see also Dancy 2000, chap. 1), starting with reasons as ‘causes’. In view of the thin relevance causality has within the explanation of ordinary actions, we can in this context disregard the idea of reasons for action being causes.13 Explanations of actions refer explicitly to motives (and thus at best to ‘causal terms’), not to causes and laws. Reasons as ‘motives’ and reasons as ‘motivating reasons’. The difference seems to be that talk of a mere motive has no normative claim, whereas talk of a motivating reason has at least a weak normative claim. “His motive was jealousy” and “his motivating reason was jealousy” differ in that the former is closer to “his motive was insomnia” and the latter closer to “his motive was the belief that he was being betrayed”. Talk of a reason thus oscillates between a reason which from the point of view of the agent cannot be rationalised and a rational reason, whereby the concept of ‘rational reason’ can in turn be interpreted more or less broadly. 12 Some authors, e.g. Smith 1987, believe them to be the only kind of reason. 13 See Thalberg 1985, 16 – 20; Stoutland 1986. Davidson might not have maintained that the primary reason in explicit form has to be the cause of an action (Davidson 1987), and yet he, too, advocates its causal role. Similarly, Hornsby finds it strange “that the rational explanation of action mentions causes but is not itself causal explanation” (Hornsby 1993, 135).

Introduction: Through Contexts to Actions


Reasons as ‘explanatory reasons’. The explanation of why an agent has done something can be, but does not have to be due to a motive, for example in an omission to act, still perceived as an action even though it results from negligence. Disregarding the vague category of omissions (are they actions or not?), explanation can result from motives or motivating reasons and is subject to the abovementioned differing degree of consciousness or rationality. A resum about the first group of terms would be: reasons in the sense of ‘motives’ or ‘motivating reasons’ are necessary elements in the description of actions. Actions cannot be without motive, and, depending on the meaning of the term, can hardly be without a motivating reason either. Reasons as ‘justifications’, ‘justifying reasons’ or ‘normative reasons’: The difference between ‘justifications’ and ‘justifying reasons’ is minimal. The former are usually given afterwards; the latter are usually consciously linked to current activity. More important is the difference between justifications and normative reasons. Justifications require an agent whose justification they are, whereas normative reasons do not. The two can accordingly diverge. A is justified in taking the book away because he believes it to be his. But this may not have been a good reason to take the book away if the book actually belonged to B. Justifications are reasons which are presumed to be good from the standpoint of the agent; normative reasons are reasons which are not good from anybody’s particular standpoint, but are the best reasons available. The exact connection between the two raises considerable problems, as we shall see in a moment. Talk of ‘normative reasons’ certainly leads us into problematic terrain because precisely this term is used and reconstructed differently in different theories of rationality. Whereas the Humean tradition traces back any normativity of reasons to desires under suitable conditions, the realistic and intuitionistic opponents of this tradition (Platonists, Reid, Kant, Moore, Korsgaard, Dancy) view reasons as not subjectively reducible.14 Whether one believes normative reasons to exist or not, and 14 Wittgensteinians, who deplore fits of dizziness confronted with any step towards abstraction, already struggle with talk of ‘normative’ and ‘normativity’. “Normative” and “normativity” are certainly predicates with more than one meaning and unite the characteristic of being prescriptive with that of being objective. The problem with normativity lies essentially in this combination which is not immediately reconcilable. How can something which is supposed to be objective motivate, and how is a subjective motive supposed to be objective?


Anton Leist

how one reconstructs them are also closely linked to how one sees the relation between the descriptive and the normative versions of reasons (the first and second types in the list above). In any case the relation has to be explained via an identity bridging the two types of reason. Motivating reasons are reasons which enable an action to appear positive and therefore suggest to an agent, command him in fact, to execute that action. Advocates of the standard theory choose to relate motivating and normative reasons as follows: Corresponding to the belief and attitude of a primary reason for an action, we can always construct (with a little ingenuity) the premises of a syllogism from which it follows that the action has some (as Anscombe calls it) ‘desirability characteristic’. Thus there is a certain irreducible—though somewhat anaemic—sense in which every rationalization justifies: from the agent’s point of view there was, when he acted, something to be said for the action. (Davidson 1980, 9) To serve as reasons for an action, beliefs and desires need not be reasonable, but a normative element nevertheless enters, since the action must be reasonable in the light of the beliefs and desires … . (Davidson 1980, 84) What is special about the (action) explanation is that Eve’s desire and belief explain her action in part because they provide Eve with a reason so to act. (Bratman 1987, 14)

To sum up these remarks on reasons so far: firstly, “reasons” in connection with actions is a term with many meanings and an especially descriptively and normatively ambiguous and dithering one. Secondly, “reasons” can refer to motives for action which are rational to differing degrees, i.e. deemed good and thus executed by agents to differing degrees. The assumption that an agent always presumes his actions to be good, otherwise he could never be asked why he does what he does, is not implausible (albeit highly controversial).15 Thirdly, motivating reasons are, by virtue of the use of the term “reasons”, to be comprehended as Disregarding these problems, talk of ‘normative’ reasons seems to be essential if one wishes to say why a reason prescribes something or is a prescriptive reason. To reject normative reasons is also a rejection of prescription or practical necessity. 15 It is possible to argue against this assumption, especially with reference to examples of arational acting (Hursthouse 1991). For other examples and the discussion of irrational action situations, see Anwander 2002. Opposing Raz 1999, Anwander believes that an action does not have to include being deemed good in order to be understandable. In my opinion, this debate should be read as centering on the framework or the system of understanding action, not on single actions.

Introduction: Through Contexts to Actions


justifying reasons. Fourthly, anyone employing justifying reasons positions himself, we can assume, within the ‘space’ of normative reasons. All actions are then indirectly tied to normative reasons, even if they are not necessarily correspondingly motivated. I believe that one should better not read Davidson and Bratman in the sense that they would oppose extended embedding and back-reference to normative reasons, thus erecting a barrier between action theory and rationality theory. Even though he did not observe the difference between justifying and normative reasons, Davidson has, with his demand that explanations for action should be “rationalisations” (1980, 3), embedded actions within the space of rational action and thus also that of normative-rational action. Davidson did not, surely, advocate an explicit normativity thesis, according to which actions are necessarily embedded within the space of normative reasons, and it is doubtful whether he wished to advocate one. According to the normativity thesis not all actions are necessarily justified as such, justified at all or motivated by reasons, but all actions are per se subject to the ideal of being justifiable or of being related to normative reasons. Without such an assumption, there would be no connection between the areas of motivated and right action. One consequence of the normativity thesis is that the shift of interest away from the issue of which reasons must be necessary components of an action towards what constitutes practical and especially moral reasons represents not a transition to another topic, but makes for a part of an extended analysis of actions, analoguous to the transition from intentions to plans in Bratman’s planning theory. If reasons for action depend on ‘good’ reasons, then one cannot understand actions without understanding the foundations underlying these good reasons; action theory cannot do without normative foundations, which are usually seen as being located firmly within metaethics and ethics. The same result springs from the two most important questions surrounding the rationality of practical reasons: in addition to the obvious question of what makes these reasons normative, the second crucial question of how normative reasons can motivate would be incomprehensible if it were not reasons for actions which were being discussed. Before I address these questions and the normativity thesis in more detail, I should like to draw attention to an additional integrative hunch definitively represented by the standard model and partly in competition with the normativity thesis. Advocates of the standard model, especially those who are interested in the causal relationship between reasons for action and action, frequently impute a naturalist claim, according to


Anton Leist

which a close (reductive, corresponding or other) relationship exists between brain states and propositional attitudes. Naturalistic assumptions or strategies can be articulated and pursued in different ways, but overall they provide strong motives for viewing reasons for action as subjective states, in contrast to normative-social relationships (see Stoutland 1998, 48 f.). Following a reductive conception of reasons, it is possible to reconstruct beliefs and desires as symbolic representations which are not actually characteristics of persons, but ontological states in their own right. Opinions differ as to whether the naturalist claim is necessarily linked to the standard model or not. Davidson’s pronounced intention of comprehending reasons as causes certainly suggests naturalistic tendencies, but this does not necessarily have to be linked to a full naturalism.16 Just how Davidson’s original aim of uniting reasons and causes is supposed to be achieved, remains an open question and cannot be pursued any further here. In line with probably the majority of those involved in the current debate on reasons (including Davidson himself ), I am of the opinion that normativity takes priority over reduction attempts. In the words of F. Stoutland: … nothing from the physical realm can determine what is a real reason for action or make it true that an agent acted for this reason rather than that. (1998, 53; more detailed Dancy 2000, 163)

Let us return to the normativity thesis and the embedding of motives in normative reasons. In Davidson’s works the difficulty surrounding the precise depicting of reasons for action—disregarding the controversial problem of the causality relation—is not all that ponderous because no explicit difference is made between justifying and normative reasons, and the paradoxical bifurcation of practical reasons being normative and motivating remains hidden. The gulf between motivation and objectivity becomes more visible if a clearer distinction is made between motivation and objectivity, as suggested by Williams’ famous article (1981). According to Williams’ terminology, a restaurant guest desiring alcohol has an ‘internal’ (subjective-justifying) reason to drink a glass allegedly containing gin. In addition, there can be an ‘external’ (objectivenormative) reason not to drink the alleged gin: the fact that the glass actually contains petrol. What is the relationship between subjective16 The opposite diagnosis referring to Davidson in his early work is defended by Stoutland 1998. In his opinion, Davidson later distanced himself from the standard model in this naturalistic sense. See Stoutland 1999.

Introduction: Through Contexts to Actions


justifying and objective-normative reasons, and what, in turn, is the connection to a motivating reason? If not to drink from the glass is a good reason, even though I am unaware of it, what significance does it then have for me that reasons are connected to my psychological attitudes, as postulated by the standard theory? The problem can similarly be formulated using “explaining” and “justifying”. If the primary reason were simultaneously to explain and to justify, the normativity thesis would be met. Objections? The standard theory was simply not articulated explicitly enough in order to render visible the extent to which it is capable of fulfilling the demands made of practical reasons. One criticism is therefore that the normative profile of the standard theory needs to be developed in more detail; another criticism addresses the resulting desire-based conception of reasons. Criticism of the first type arises from questions such as: Is a primary reason the entire normative reason or merely one of its constituent parts, the belief or the desire, possibly in each case in a certain dependency on the other part? Can a psychological state, such as a belief or a desire, be a normative reason at all, in contrast that is to the object of a belief or a desire? The standard model is usually vague about whether the reason is to be found in the belief and/or the desire.17 Firstly, it is vague regarding the way in which expansive talk of ‘primary reasons’ is certainly helpful in order to be able to reconstruct varying examples which include by name not desires, but all manner of emotional attitudes, in other words the way in which a pro-attitude can always be proven on the basis of these examples; it also becomes unclear, however, what else is claimed alongside the primary reason. J. Searle’s example, for instance, in which one would have no reason to pay for a beer one has ordered if one felt no desire to pay for it (2001, 27), is hardly an appropriate illustration of the model. But does the beer drinker really have a belief that sanctions are to be expected if he refuses to pay, and does he really have a desire to avoid those sanctions? How do we proceed if someone does not explicitly have the corresponding beliefs/desires? Secondly, there is a conflict between beliefs/desires as psychological states and as normative attitudes. One still has normative attitudes even when they are not psychological states, and wanting to reduce them to psychological states is problematic. In this point the standard model is 17 “The main idea is that the agent’s desires and beliefs at a certain time provide her with reasons for acting in various ways at that time.” (Bratman 1987, 15)


Anton Leist

generous, with its terminology covering both, thereby posing a challenge to the current debate on practical reasons, in which reasons as psychological states are certainly no longer primary reasons. Put another way, one could also say that the standard model is vague in two respects and has correspondingly given rise to two types of criticism: first, it stimulated reservation that desires or other psychological states could be the only or the dominant source of normativity; second, it gave way to doubts that subjective psychological states could be capable of carrying truth-claims. The first sort of criticism follows the tradition of hedonistic and related empiristic value theories; the second sort of criticism is of an epistemological and ontological nature, namely interested in the correlation between belief, reality and truth. The first concern corresponds to a traditional scepticism as to whether desires can be normative or bearers of truth-claims, whereas the second concern springs from various sources. In its constructive version it consists in an attempt to reconstruct the link between the different types of reason, an attempt to link the subjective and the objective. These constructive attempts also try to counter a doubt about whether the standard model can conceive of acting as real acting, acting within a real world. Criticism of the first kind could be called the naturalist-criticism 18, the sceptical second kind the idealist-criticism. In the following I shall skip the rather well-known naturalist problems within ethics and will concentrate instead on the idealism/realism puzzle.19 The idealist-criticism is targeted at desires, beliefs and possible combinations of these. It is not directed at beliefs and desires per se or at beliefs and desires in a certain mutual relationship, but at beliefs and desires as psychological states. The key point of criticism is: normative reasons for an action must prove this action to be good or correct, and 18 In the wake of G.E. Moore’s use of “naturalism” in ethics, which is not identical to the previous, much broader concept of naturalism. 19 Dancy criticises desires as the basis for reasons with the charge—familiar from the Kantian tradition—that desires are contingent and ill-founded. Because desires themselves require reasons, they can be jumped over in justifications (Dancy 2001, chap. 2). Rebukes like these are certainly exaggerated and would render theoretical and practical reasons mutually indistinguishable (see Darwall 2003). It would also be impossible to explain why there are (correctively effective) reasons for desires if desires themselves play no role at all. Scanlon’s ‘buck-passing account’ (1998) does far more justice to this problem of coherence and shows that flouting desires need not lead to a realistic conception of reasons as Dancy’s.

Introduction: Through Contexts to Actions


psychological states, with very few exceptions, are not capable of doing this (Dancy 2000, 102 – 107; Bittner 2001, chap. 7; Stoutland 2001).20 Since this objection refers to psychological states in general, it is equally valid for beliefs and desires, and indeed is meant to be so. A desire is not a normative reason because it is only as good as what it is directed towards; a belief is only as correct as that which is believed. Beliefs and desires are capable of subjectively justifying only if they can prove correct, and this can only be done by referring to the factive qualities of their objects. This criticism is partly based on examples from our everyday thinking, but overall it is just as plausible as the suggested factive conception of normative reasons is tenable. Such reasons are usually states of affairs which are independent of an agent, and very occasionally states of affairs identical with psychological states of the agent.21 In some cases an agent believes he has to react to his psychological states, usually dispositions, and if this assumption proves to be correct because not taking his disposition into account would mean advantages or disadvantages relevant to his situation, he then has a good reason. An agent prone to vertigo, for example, has a good reason not to approach the edge of a cliff if he is afraid of losing control and if he actually would lose control. Typically, however, an agent acts from assumptions about others and about the world surrounding him; the typical case is not formed by selfperception of one’s own dispositions, but by perception of the conditions governing the surrounding world. In contrast to this the standard theory promotes exceptional cases to typical cases by raising psychological states, at best suited to justifying, to the status of general normative reasons. Viewing normative reasons as states of affairs represents, in more than one way, the precise opposite of the standard model’s notion of normative reasons. This is also demonstrated by three consequences which Dancy, in particular, is prepared to face. First, the strongest possible answer to the question of how motivating and normative reasons are supposed to be linked is a claim to their being identical or nearly identical, reasons are motivating due to the qualities of their states of affairs (2000, esp. chap. 5.3; also Bittner 2001, §§ 236 – 242). From a diametrically 20 A number of authors, including these three, put forward the idealist-criticism and an alternative ‘factive’ concept of reasons at more or less the same time. Below I shall concentrate particularly on Dancy 2000 because his theory is both the most comprehensive and the most radical. The differences between Dancy, Bittner and Stoutland are slight, however, and will be mentioned in passing. 21 ‘States of affairs’ are understood here in the sense of actual states, not in the sense of logically possible facts, following the Tractatus convention.


Anton Leist

opposed standpoint, Dancy thus once again approaches the standard theoretical idea that reasons and motives coincide (see Dancy 1995). Second, due to the now ontologically reversed point of view, explanations for an action are not causal explanations for that action, stemming from psychological states of the agent, but normative explanations. The theory thus circumnavigates the problem of action causality (2000, 84 f., 163; Bittner 2001, chap. 5). Third, normative reasons are not (as in the standard model) opposed to empirical reasons; external states of affairs are normative states of affairs: “(R)eality is practical.” (2000, 137) (In contrast, Bittner is of the opinion that reasons are always only empirical reasons/states of affairs and normativity per se is something which is extremely dubious (2001, 69).) Can this normative realism of reasons really be upheld in the light of such amazing extensions? Once again, Dancy in particular has clearly identified and attempted to deal with two key difficulties. Reasons as states of affairs somehow need to be linked to the agent as his reasons for his actions: how else than through the psychological states of the agent? Do they not need to be taken into account once more as part of the reasons? And what about the erring agent? In the case of an erring agent motivating and explanatory reasons cannot be states of affairs as, due to the error, there is no such state. Is this not evidence of the secondary role of states of affairs and the primary role of psychological states? Dancy has provided inventive answers to both problems, their plausibility remaining controversial, including in parts of this book, and raising reservations as to whether these answers have not caused the difference between the factive reason-conception and the standard model to become considerably narrower. According to Dancy, the psychological states of the agent are “facilitating conditions” or “appositional” conditions of the realistic explanation, but not its component (2000, 126 – 130). An agent must be in the correct frame of mind for a state of affairs to be, for him, a reason, without the frame of mind itself being part of the reason—with the exception of rare or pathological self-related cases (124). In this way, states of affairs giving rise to reasons are integrated in the agent’s perspective of consciousness; reasons for actions are always accompanied by a knowledge of the reasons without them becoming psychological reasons (Bittner 2001, §§191 – 193). How to deal with the erring agent who thinks he has a reason but in reality does not? One possibility would be to resort to propositions or mental entities which can step in for what is maintained but is not really the case. In the light of the idealist-criticism this move would be a

Introduction: Through Contexts to Actions


regression and would more or less amount to a silent acknowledgment of the standard theory.22 Thus the only things remaining are references to parts of factive circumstances. This is the solution Dancy favours. He chooses throughout to adhere to the notion that a state of affairs explains normatively and motivationally why an agent finds an action good; in the case of an erring agent, this state of affairs is a non-given state of affairs. A state of affairs which is merely not given can be eliminated because it would beg the question in what way (in psychological states?) it is not given. Dancy therefore postulates, if not exactly in these words, nonexistent states of affairs as states of affairs. This seems clearly a reductio ad absurdum, which Dancy chooses to ignore with ambiguous, no longer realistic comments.23 Bittner opts for a different solution. If an agent thinks he is acting for a reason which, so it transpires, does not actually exist (a state of affairs which is not given), then he acts without reason. The agent acts from a psychological state of belief, which does not prove, however, that this state is his reason (2001, §§199 – 212). A nasty aftermath of this version of events seems to me to lie in the following. Either one chooses an artificial way of talking according to which this agent acts ‘for’ a reason, if not ‘out of ’ a reason, then one leaves open how both these kinds of acting for reasons are normatively linked. In not being explicit regarding the connection one refuses to try to achieve coherence between the subjective and the objective side. Alternatively one could say that the agent acts for no reason at all, thereby neglecting the fact that he thinks he has a reason. Bittner says that the erring agent acts for no reason at all, whereas we would normally say he acts for a mistaken reason. Someone acting for no reason at all is, with the exception of a specific class of spontaneous or extravagant actions, acting irrationally, whereas the erring agent is acting as rationally as one could possibly imagine. Neglecting this difference does not seem to be a good thing either, so the factive reason theorist still has to give an answer to the case of the erring agent. Several authors in the present volume elaborate on this puzzle. 22 Dancy discusses a theory which is slightly different from the standard theory with equal criticism (2000, 121 – 126). 23 “… there is no metaphysical problem about the nature of that which is believed when the belief is false, because of the intentionality of belief. A belief can have, as it were, a non-obtaining content without having no content at all.” (147) “Perhaps all that needs to be said … is that the content of a belief, what is believed, is something that either is the case or not.” (148)—In both cases the content is dependent on the belief, not vice versa!


Anton Leist

If Dancy’s and Bittner’s radical attempts tangle things up, some conventional differentiations once again need to be observed, for example the difference between explaining and justifying. J. Wallace has emphasised that the focus of explaining is on the actual state of the agent, whereas that of justifying is on the truth of a state of affairs, meaning that it cannot be correct to say that the two types of reason are almost identical (Wallace 2003). It seems to me that the appropriate relationship between explaining and justifying is to be found at the core of the conflict between internal and external reasons. Dancy connects the two because he sees human motivation as being essentially characterised by the rationality of believing something to be good, the latter in turn being founded in real states of affairs (Dancy 1995). Instead of directly identifying both by means of individual states of affairs, perhaps it would be better to bring them together at a more background level, for example through ‘rules’, ‘practices’ or ‘habits’. In this point Stoutland’s sketchy, Wittgenstein-inspired comments (1998; 2001) seem to be more promising than Dancy’s (and Bittner’s) adherence to immediate states of affairs. On the one hand, rules and regularities solve the task at hand, i.e. make it transparent why an agent acts, from his own point of view as well as that of others, and they are also normatively easier to understand than states of affairs beyond actions (“the sun goes down”). On the other hand, they have the background character to give particular justifications and explanations the necessary scope. This may be narrower or broader, depending on the type of action. Stoutland’s (2001) example of stopping at a stop sign, a practice within a strictly regulated system, is surely not representative of most actions. But indicating that there must be a common system of rules governing our reasons to act does point us in the right direction for tackling the difficulties in Dancy’s theory of reasons. If these observations are right, then the problem of connecting internal and external reasons, explanatory and normative reasons, the first-person and third-person view of actions, the different standpoints (Wallace 2003, 431) arising from the confrontation between more subjective and more objective action theories, should be transformed into more fruitful questions: Out of which common social system of rules can individual agents meet each other in different types of reason? How do the stipulated rules permit an explanation of individual actions even if they are based on false assumptions or a failure to adhere to those rules? The standard model and the factive reasons-position are theories which have been driven to opposite extremes, and this opposition can only be

Introduction: Through Contexts to Actions


overcome by using a level which is common and necessary to all the elements involved, in order to create the desired connection between explaining, justifying and truth. Similarly to the case surrounding Bratman’s planning theory, the debate about reasons therefore leads into an extended view of actions, in order to understand actions. 6. Actions and Agency Some, like Frankfurt, choose to formulate the central question of action theory by explicating the contrast between what an agent does and what merely happens to him; others simply concentrate on the difference between individual action and a bodily movement. Agents are already implied with the mere mention of actions, or so one could think. This assumption might have put the minds of standard theory advocates at rest about the fact that when they begin their analysis of action with individual actions they construct no actions, at least in principle, without agents—similarly to the way that linguists see no fundamental problem in directing their theory at sentences and not at speakers, silently assuming that sentences are always the products of speakers. Two key motives behind the standard theory were, however, first, to avoid the regress within Prichardian volitions and, second, to allow reasons and causes to become compatible. These two motives force the standard theory to exclude the agent of an action to a greater degree than is apparent at first sight. Unlike acts of will, agents can possess beliefs and desires independently of their special role as agents; dispositions of the agent can lead him to act without recognising these acts himself as his own. As a result of its programmatic motives, the standard model objectifies action in a way which renders it difficult to recognise the role of the agent within the actions analysed. To use a term which in the last few years has received increasing attention: the standard model has a problem with ‘agency’, a central characteristic of the agent and the action. The term “agency” is, like all general terms used in action theory, an artificial, philosophical expression and accordingly unclear. At the heart of the term is an intuition that the action cannot be separated from the agent, or that it shares something with the agent whose action it is: an element of the activity, the perspectivity of the agent or the execution from his point of view. If—taking the terms “agent” and “action” as given—“agency” is a relational term, then it comprises both the aspect of agency, an activity of the agent, and the ‘actorial’ characteristics of the


Anton Leist

action, characteristics which cannot be described without reference to the relevant agent. Because of its relational character, both the agent side and the action side of agency can be emphasised, both agents and actions can be ‘actorially illuminated’. Given this potential purely verbal emphasis becomes problematic when, as in Davidson’s (1970) discussion of action activity, the agent is not only no longer included, but is also no longer conceivable (e.g. Velleman 1992; Hornsby 2004a). The difficulty the standard theory displays is that it not only does not provide a satisfactory concept of action activity, but—ringing a little crazy—systematically excludes any active involvement of an agent. What deficits within the standard theory spark this accusation? Firstly, the inappropriate concept of a causal explanation of actions. By turning the everyday explanation of an action into a causal relationship between events due to beliefs/desires, it abstracts from the teleological sense of that action. One would not expect the explanation of an action to consist in information as to why a mental event brings forth a bodily movement. By abstracting from the activity, a causal relationship does not explain. Secondly, and somewhat more special, the standard theory cannot distinguish genuine actions from determined behaviour or from actions due to unconscious beliefs and desires. Explanations based solely on psychological states lack the actorial authority which typical actions have to possess. This objection is independent of causalistic problems and is orientated towards the everyday demands made upon explanations of action. Thirdly, it is hard to see how the standard theory wishes to solve the normativity problem if it does not permit any critical reflection regarding motives for action, i.e. the beliefs/desires pair. A complementing of the premises surrounding critical reflection methods to date is barely conceivable if the agent is not taken into account in his relationship to his beliefs, his desires and his intended action.24 These three difficulties can be supplemented with a fourth one which has been recognised for some time, the difficulty confronting the standard theory with regard to actions spinning out of control. More clearly than any of the other difficulties, ‘deviant causal relationships’ question the two basic elements—the belief/desire pair and the causal relationship. Since Davidson’s first indications (1980, 79) it has been 24 See Hornsby 2004a, 8, for the first deficit and Bratman 2000 for the remaining two. The first objection does not fully apply to Davidson, according to whom explanations of actions are not direct or explicitly causal explanations. See Davidson 1987; Stoutland 1999.

Introduction: Through Contexts to Actions


clear that a suitable psychological state of an agent can, in the course of an action running out of control, cause an action result which in content equals that which was desired, thus rendering the conditions of the standard theory fulfilled, even though it amounts to either no action or a failed action.25 What these four problems in the standard theory have in common is that they all arise from an activity-relevant, as well as actorially underdefined description of the premises of the action. However, they are all inadequate at different levels of describing and explaining actions, so that different deficits and types of action activity are addressed. Although various action theorists (Frankfurt, Velleman, Bratman) have begun to deliver more and more detailed analyses of the ‘comprehensive’ agent, it remains controversial why and to what extent they are forced to do so from the narrow interest of the analysis of individual actions. The explanation of actions compels the inclusion of action activity—but how comprehensive must the concept of action activity be, leading to a concept of agency and a comprehensive theory of the agent, including his rational, biographical and social characteristics? To what extent must a theory of action lead into, or be developed from, a theory of the agent? The discussion becomes slightly clearer if, in the question of how broadly the concept of action activity has to be construed, one differentiates between ‘minimalists’ and ‘maximalists’. Minimalists (like Hornsby) believe the inclusion of agents to be necessary in the analysis of individual actions, but without aspiring to a normative construction of agents. Maximalists (like Frankfurt, Velleman, Bratman) believe that explanations of actions require per se inclusion of the agent and his role within the ‘rational aetiology’ of the action. The maximalists have a plausible reason for this, it seems. If one demands of an action that it has to be rational in a certain way in order to be an action—even if “rational” is only meant in the sense of a coherence of action and belief/desire— then the concept of agent which is to be included, itself has to be a normative or rational concept. In addition, as Hornsby suggests (2004a), a further motive of the maximalists is presumably their adherence to event causality, so that their proposals of action activity and agency can be 25 Famous examples of this are the climber who unintentionally allows his partner to slip, the nephew who by chance runs over the uncle he is to inherit, the bank robber who signalises through his nervosity the begin of a break-in, etc. Discussions are included in: Frankfurt 1978; Peacocke 1979; Bishop 1989; McCann 1998, chap. 4; Velleman 1992; 2000, Introduction.


Anton Leist

meant as an extended version of the standard theory, in which the explanans events now consist in more complex (normatively mediated) psychological states of the agent.26 In this controversial situation it could be helpful to separate between the philosophical-ontological perspective and the everyday demands made upon explanations of actions. Two interesting questions are then, first, whether the adherence to event causality is reasonable within the extended concept of action activity, and second, why, independently of that, the concept of the agent actually has to be normative, and to what extent. For the first question the manner in which one assesses the problem of deviant causal relationships is enlightening. Even ‘causalists’, or so it appears to me, can no longer stand by the view that causal relationships, if they exist, are informative for the type and existence of an action, in the conventional sense of action, so that (comparable to determinism seen compatibilistically) causal relationships can be conceded, but are eliminated from the analysis of actions as sources of information. Davidson’s answer to deviant causal relationships was that it is crucial to make a distinction between a ‘correct’ causal relationship and irregular ones, in order thus to contain actions causally as well as otherwise (1980, 79). The wealth of counterexamples and more recently the fiction of an external intervener who is free to redirect the neural pathways of my actions against my will (see Bishop 1989, chap. 4 – 5; McKann 1998, chap. 6), make the extreme dimensions of this task very plain. Whether or not the aim behind an agent’s action is actually achieved always also depends upon the contingent circumstances (whether the golfing green is flat; whether the telephone number is reached), and because these causal influences are not predictable, they cannot be anticipated within an intention to act. The necessary guidance is therefore more interactive, the relationship between intention and result a recursive process, and not the same as a causal happening within the action (or, in the extended theory, within reflexive states of the agent). The correspondence between an intention and the result of an action alone is not sufficient; what is necessary is guidance by the agent during a sufficiently long part of the action, differentiated according to the type of action. A causal model for

26 Such proposals are included, amongst others, in Velleman 1992; Bratman 2000; Smith 2004.

Introduction: Through Contexts to Actions


this condition is (or so it appears) neither possible nor in all probability interesting.27 In line with the objection stemming from the everyday relevancy of causal explanations of action already noted, Hornsby also objects to shifting agency into complexes of higher-order intentions and action strategies (Hornsby 2004a, 14), as she thinks that most agents do not possess such higher-order states when acting. Quite possibly, but under which conditions is agency appropriate, aptly expressing action activity? Hornsby does not give a clear answer to this. If not from the basic premises of the standard theory, why else be a maximalist, extending an action to include the characteristics of the agent? It seems to me that the abovementioned motive of the maximalists is worth considering, even under non-causalist premises: in order to permit his purpose to become an action, the activity of an agent must be ‘rational’ to a certain extent, which is hardly possible without the agent himself being ‘rational’. “To a certain extent” is a vague precept, the vagueness of which might lead to slightly overstretched activity conditions in order to be met by many of our particular actions. But why should such conditions not be idealised in their formulation and then be mitigated in order to match our everyday understanding? A necessary connection thus postulated between actions and rational agency seems to be directly contradicted by the objection that even irrational actions are obviously still actions. (In a similar manner, one could also contradict other forms of normative action, like autonomous or authentic actions, by pointing out that compulsive or alienated actions do not lose their action character.) And yet the notion that all actions are first and foremost actions, which then have the cited normative qualities added on, is wrong. Far more, some actions are normatively thin actions (my arm going up), and other actions are only possible because of their thick normative qualities (greeting, heeding a stop sign). If the standard theory only captured normatively thin actions, it would be inappropriate to the vast majority of human actions. The maximalists therefore choose to adopt the correct strategy, albeit partly from problematic motives, by taking normatively more complex agent conditions into account. In 27 The condition of guidance stems from Frankfurt 1978. An informative discussion of the various proposals for a solution regarding deviant causal relationships is provided by H. McCann (1998, chap. 6). However, his own postulate of ‘intrinsically actional’ volitions (112) is just as ad hoc as his criticised attempts at solution put forward by the causalists.


Anton Leist

contrast to Hornsby’s objection of expecting too much, this really could be interpreted as just that: more complex actions being taken into account. With the obvious wealth of different interests in actions, the assumption that all actions are first and foremost actions per se, independently of particular interests in them, becomes obsolete. Far more, actions are dependent upon interests and context, and this makes them differ. Frankfurt’s classic differentiation between the wanton and the selfgoverned agent can both illustrate this and diffuse any suspicion that the normative conditions might be too complex. The wanton individual (who converges with the agent in the standard theory) acts in order to satisfy his desire, but has no control over that desire. In contrast, the agent with a second-order desire which lends authority to his desire to act, acts in a self-governed manner; he not only controls the course of his action, as called for by Frankfurt (1978), but also its motives.28 If human actions are not to be reduced to the level of behaviour, a certain degree of higher-order governing of actions is definitely necessary—a circumstance which is neglected in the standard theory, in particular due to a concentration on examples which are too simplistic. It is one thing to explain the difference between raising my arm and my arm going up (Davidson 2004, 101), and quite another to do justice to typical human actions with all their various potential for failure. If, from this point onwards, we refer to the ‘maximalists’ as ‘normative actoralists’, we can then divide them into two categories: action-orientated actorialists and person-orientated actorialists. This distinction roughly corresponds to the distinction between ethics of action and of virtue, an attempt to set moral standards primarily for individual actions or primarily for agents. The action-orientated actorialists would like to describe individual action as normatively appropriate via agent conditions, whereas the agent-orientated actorialists would only like the agent and actions to be characterised indirectly. The latter dedicate themselves to this extended task not from an abitrary interest in agents, but because they regard the first attempt as unsuccessful. With his proposals regarding second-order desires and volitions, Frankfurt was primarily an action-orientated actorialist. The criterion of being satisfied now no longer refers solely to current states of 28 Although Frankfurt (1978) only demands tighter guidance for the formation of an action, even attributing this ability to spiders, ‘guidance’ is obviously a typically human capability which even extends to the motives behind an action.

Introduction: Through Contexts to Actions


the agent (like desires), but also in the longer term and cross-situationally to operative dispositions of the agent.29 In wishing to explicate the active quality of the term “agency” by means of a directly given agent state, the action-orientated actorialists find themselves in the same difficulties as Prichard and other voluntarists: problems of circles and regresses, or ad hoc postulates of ‘intrinsically active volitions’ or ‘volitional necessity’ in order to circumnavigate these problems. In contrast, the agent-orientated actorialists view agency as anchored in dispositional characteristics of the agent, such as plans to act, ‘self-governing policies’ (Bratman), or a manner of action directed towards longer-term goals.30 In the face of the destructive circularity of the first position, the only option is clearly to follow Bratman, who shifts the internal problem of the insoluble difficulty of complete governing of action to more general, personal characteristics, and become an agentorientated actorialist. Of the three extension tendencies the standard theory has so far been described as producing—following on from a missing concept of intentions, insufficiently normative practical reasons and the blind spot of agency—this movement away from action activity towards the normatively reconstructed agent is the most important in several respects, albeit the one which is least clear at present. Firstly, the agent comprehended temporally and normatively is capable of incorporating both the temporalised concept of intentions and the concept of external reasons, thus providing a suitable framework for the transformation of the standard theory outlined here. Understanding the agent temporally, Bratman (1999; 2007) in particular has managed to bridge the intentions behind an individual action and the agent conscious of his temporal identity. Secondly, the developments described above have taken the conflict between normativity and naturalism/causalism, which has been 29 “Satisfaction is a state of the entire psychic system—a state constituted just by the absence of any tendency or inclination to alter its condition.” (Frankfurt 1999, 104) 30 “We try to solve the problem of subjective normative authority, and to say what constitutes the agent’s deliberation, by appeal to higher-order intentions, plans, and policies, about which desired ends to treat as reasons in motivationally effective deliberation. This seems a promising strategy in part because such higher-order intentions, plans, and policies have, as a matter of function, tight connections to the temporal extension of agency. That is why they are candidates for attitudes that, because their role in our agency, can speak for the agent.” (Bratman 2001, 101)


Anton Leist

present in the standard theory from the outset, and elevated it to a higher platform, but without solving it. Following a premise of this overview, however, no other reaction to the problem of analysis in the confrontation with elementary concepts is possible than that of expanding the field of vision and to observe local problems in the light of ever more extended conditions.31 This premise is confirmed by the observation that the described development opens up several key questions. First, concerning the extent of the temporality of the agent: what temporal abilities does an agent have to possess in order to be a ‘complete’ agent? What would minimalism and maximalism mean in this context? Second, concerning the connection between motives and (external) reasons: to what extent does the agent participate in the social ‘space of reasons’, and how motivating are these reasons? Third, concerning interpretations of agency by means of ‘selfgoverning policies’: to what extent does every action already embrace the conditions (even if not their realisation) of autonomous or authentic action? In other words, to what extent is being autonomous and authentic not peculiar to especially complex actions, but already contained within the basic model of action, merely overlooked by the standard theory to date? 7. The Extended Context: Basic Attitudes, Policies, Practices The action theorists referred to above as minimalists, and others besides, could view the described tendencies of the extended field of action less as a profound analysis of actions and more as a transition to other themes. Indeed: social plans for action, external reasons and actorial action strategies might still be necessary as conditions underlying individual actions, but do they not kick off a never-ending series of possible action/ context relation pairs, with the threat of losing the real constitutive conditions for individual action, and thus its ‘essential’ concept? Do the maximalists not really dissolve the concept of acting rather than explicating it further? 31 For this reason I have my reservations about Bratman’s (2001, 102 – 3; 2002) attempts to eliminate the dreaded circle by explaining the activity of actions through agency by means of conceptual distinctions. Why is this circle to be abolished if the concepts of action and the agent cannot be perceived in mutual exclusion in the first place?

Introduction: Through Contexts to Actions


This problem of the relationship between action and context, bearing in mind the different concepts of ‘context’, and of the distinguishability of actions and context, rears its head when the event-causalistic ontology of the standard theory is definitively discarded and all actions are comprehended as “orthonomous” (Smith 2004), with different dimensions of ‘orthonomy’ and different types of interaction between the agent and the action context. Making a distinction between actions and their context is both legitimate and necessary in order to study the type of interaction between the two. But if, correcting Davidson’s original abstraction,32 actions only appear as normatively qualified actions in the world, two things seem to be necessary. First, the different dimensions of action normativity need to be distinguished and their connection more closely observed. A normative theory of actions would relate these dimensions to each other systematically. Second, actions need to be distinguished from their normatively reconstructed social conditions of origin—their socio-normative context. If actions are not ontologically self-involved pieces, but typical, average, widespread actions, then philosophical action theory has, just as the sociological action theory based on M. Weber’s ideal types always has, to portray ‘ideal’ actions. Typical actions could be important to the extent that they are a prerequisite to all descriptions of particular actions. The particular actions of individual agents may not always meet the normative demands made of actions, and yet confirm precisely this normativity in the process. Unintentional, irrational, compulsive, conventional action can only fail in different dimensions if it is evaluable in these different dimensions of normativity. Be that as it may, more important than a collective category for the actions discussed specifically by philosophers are lists of the various practical functions within and beyond actions. Which psychological functions must an agent fulfil in order to have control over his actions? And which normative functions do agents and actions have to fulfil 32 “A reason rationalizes an action only of it leads us to see something the agent saw, or thought he saw, in his action—some feature, consequence, or aspect of the action the agent wanted, desired, prized, held dear, thought dutiful, beneficial, obligatory, or agreeable. …. The word ‘attitude’ does yeoman service here, for it must cover not only permanent character traits that show themselves in a lifetime of behaviour, like love of children or a taste for loud company, but also the most passing fancy that prompts a unique action, like a sudden desire to touch a woman’s elbow.” (Davidson 1980, 3 – 4; also 87)


Anton Leist

before it is possible even to speak of such psychological functions?33 Against which normative background are they possible? Which normative roles can an action play, which of these are the most important and how do they relate to one another? Which tasks do actions typically have to fulfil? These questions are almost impossible to address unless they are read such that the (partly intentional, partly unintentional) function of actions results within the framework of typical contexts, such as in interaction with other people, within actions plans or previous histories, within the framework of projects, etc. The internal and external functions of actions can, however, only be mutually related through the typical normative dimensions of actions.34 A classification of action-constitutive contexts can thus only be arrived at by distinguishing different normative functions of actions. Four such dimensions have already emerged more or less clearly in the course of these remarks, intentions, reasons, self-governing and authenticity. 35 Intentions include commitments and are thus embedded within a normative space of reasons; reasons are generally accessible states of affairs and thus elements within a social space of reasons, stabilised by a background of shared assumptions which needs to be accepted; selfgoverning is higher-level agency, which in turn requires external and lastingly potent proofs and responses because it cannot succeed on its own strength alone; authenticity could be rendered as loyalty to one’s own characteristics and personal history. Whereas the first two 33 For the first functional question see Bratman 2001, 92, for the second see Santiago 2005, 92. Santiago distances himself from purely psychological functions, as discussed by Bratman, but himself suffers from a normatively dubious term of “social roles” and “social contexts”. 34 Single actions as part of social or cooperative actions are therefore merely the visible manifestation of the normative dimensions which have always been given with human actions; they are, as it were, second-level social actions, preceded by non-cooperative actions as first-level social actions. For social action in this manifest sense of the second level see Tuomela 1995; Bratman 1999, essays 5 – 8; Stoutland 1997. In my opinion all three cleave to an individualistic intuition— partly in conflict with their intentions—in that they do not explain the normativity of individual actions sufficiently out of the very social relationships they try to catch. In contrast see Brandom: “Both selves and communities are normative structures instituted by reciprocal recognition.” (2002, 217) 35 Prudence and morality are further normative dimensions, yet the difference between the two is controversial. For a critical treatment of their relation to actions see Enoch 2006. Both dimensions are less general than those cited so far, with it being easier to imagine imprudent and immoral agents than aimless, irrational, compulsive and self-denying ones.

Introduction: Through Contexts to Actions


dimensions take precedence within action theory, the difficulty of distinguishing rational and self-governing autonomous action is in itself evidence of how one-sidedly traditional these associations are. Minimalists and maximalists concerning agency are currently having difficulty making a distinction between action per se and autonomous action. While some authors move without hesitation from one term to the other (Velleman 1993; Frankfurt 1999), present analyses largely reduce autonomous action to action per se because they tend to fall back only on those elements which, at best, promote the elementary characteristics of action up several levels.36 The consequence of this can only be either to back out of the concept of autonomy or to name external conditions for autonomous action which are not valid for intentional action. Within the framework of a socio-normative view of actions these conditions could consist in the way in which the agent behaves towards others: whether in typically autonomy-relevant situations, such as the assumption of responsibility, reaction to practical and theoretical problems, reaction to questions regarding his own motives and reasons for acting, he demonstrates the corresponding skills or not. The significance of different dimensions of the normativity of actions extends the focus upon actions, as does comprehension of actions as typical actions. Depicting actions as typical also presumes a role they play in connection with other typical human characteristics and happenings, human primary goods or practices. The roles played by actions in individual and collective knowledge, in personal and extrapersonal history, in a narrower and broader society are surely three types of action context, the significance of which may be obvious for our actions in general terms, but difficult to follow in detail. The mutual dependencies of cognition and action, the entanglement of doing and experiencing within the biographical framework, the dependencies between individual and social action—these aspects are surely discussed in many disciplines of the arts and social sciences, but fatally they are largely and systematically ignored in philosophical action theory so far. And so much the worse for that, because if the development of action theory pursued in this overview is correct, normatively differentiated action theory is impossible without a thorough study of these contexts. 36 See similarly Buss 2007, who especially criticises the explanation of autonomous action by means of rationality and reflexivity. In my opinion, Buss’ own externalistic analysis of autonomous action refers more to authentic than autonomous action.


Anton Leist

Holism within semantics means comprehending the meaning of individual sentences in connection with all the other sentences of a language. In terms of rational argumentation holism means comprehending the quality of an argument in connection with all the other arguments which are possible and relevant. The extension of ‘relevant’ contexts for arguments improves our understanding of these arguments, the extension of ‘relevant’ contexts for sentences our use of these sentences. What is true of sentences and arguments cannot be false for actions for the simple reason that the latter are only possible in close connection with the former. Up to his latest statements (2001), however, Davidson tried to balance the importance of the event-causal form of actions and their normative role, instead of conceding priority to the latter. By doing this he unfortunately distanced his holism of meaning and knowledge from the theory of action. All the more reason for us today to bring the long tale of the standard theory of action to a happy end by closing it forever and replacing it with a normative, contextualised and holistic theory of action. 8. Articles and Comments in the Present Volume Most of the articles and comments in this book stem from a conference entitled “Action in Context”, organised by the editor in Zurich in April 2005, but some have been added afterwards. To start, P. Hacker’s contribution serves to bring home the far-reaching spirit of action theory which dominated recent philosophy prior to its ‘Davidsonian revolution’. The relations between thinking and acting, perception and moving bodies, intentions and autonomy which are skilfully handled by Stuart Hampshire in Thought and Action could serve as a programmatic contrast to the certainly more detailed but sometimes also too self-sufficient later debates. On the other hand, Hampshire’s and Hacker’s picture of actions and agents somewhat falls behind the externalist analysis of actions depicted thus far. It does focus on non-causal intentionality, but presentation of the agent draws narrowly on purely cognitive and psychological concepts. The agent commands special intentional knowledge with regard to his actions, is active in a special way with regard to his convictions and has rationality and autonomy at his disposal purely by being himself. This slightly superhuman concept of the agent contrasts strangely to the equally Spinozistic concessions to body and causality. In his programmatic approach to thinking and acting Hampshire ultimately

Introduction: Through Contexts to Actions


refrained from attempting a synthesis of the two phenomena, but the sheer scope of his agenda is still worthy of note today. The following articles and comments included within this book address three discussions outlined in this introduction: the appropriateness of the desire-based concept of reasons and the ontology of single actions (I. Acting for Reasons), the significance of action within the extended context of practical intentions, personal self-perceptions and life-stories (II. Action, Persons and Life), and the relationship between action and cognition (III. Action and Epistemology). The first round is dominated by the controversy between representatives of the factive theory of reasons and advocates of the standard theory. F. Stoutland defends the external or factive concept of reasons with regard to the difficulties particularly arising from a differentiation between explanatory and normative reasons. In his opinion, causal explanations, like those attempted by Fodor and Audi, are not tenable because their causal claims are strongly implausible. Davidson’s subjective version of reasons is for Stoutland too weak to describe rational acting. According to him, Davidson’s response to the requirements of understanding and causally explaining an action comes at the expense of a complete isolation of actions from their normative environment and thus also from normative reasons. With regard to the problem already observed by Dancy and Bittner, concerning how to explain the erring agent from the factive concept of reasons without recourse to subjective reasons, Stoutland contributes an additional attempt at a solution which appears to unite Dancy’s and Bittner’s proposals. In his opinion, explanatory reasons are always intentional reasons without a factual claim to truth and fulfil the same explanatory function as the belief of an agent, without that belief explicitly needing to be mentioned. The starting point of T. Schmidt’s comment is in turn the weak point of factive reasons, i.e. such contraintuitive “explanations invoking a false explanans”. In an overview he weighs up the strengths and weaknesses of the proposals by Stoutland, Smith and Bittner and comes to the conclusion that a semantic analysis could be a substitute for the assumption of non-existent states of affairs in which, as with penalties pronounced by a referee, existing or non-existing states of affairs are replaced by normative decisions. Schmidt would like to retain Dancy’s and Stoutland’s explanations in principle, but also to replace their ontology of non-existing states of affairs with explanations using normative authority. Normatively authorising explanations would then render Stoutland’s recourse to non-factive states of affairs unnecessary.


Anton Leist

Schmidt, however, also brings a contrast into play which reduces the usability of his example: a player’s foul in the eyes of the referee and the ‘real’ foul. As long as ‘real’ fouls exist without a normative decision, then it seems non-factive reasons also exist. M. Alvarez points out the close connection between the claim of causal explanations of action and Davidson’s hypothesis that reasons for action are psychological states. In contrast, the claim to causality is not the pivotal point of the factive concept of reasons and may even be excluded by it. Against the standard theory Alvarez attempts to show that the interest in causality is hardly compatible with normative and motivating reasons, and is indeed suppressed by the latter. As a further advocate of factive reasons, she criticises the identification of desires with motivating reasons and, following on from this, the identification of motivating reasons with psychological states. Unlike Dancy, Stoutland and Schmidt, and similarly to Smith and Bittner, Alvarez tends to view normative and explanatory reasons in the light of erring agents as disjunctive, and she wants to accept psychological states (beliefs, desires) as explanations only. But even in these explanations she believes that conceptual connections between the elements of explanation (desires, beliefs, actions) dominate, in contrast to empirical generalisations of a typical Humean causal explanation. According to Alvarez, explanations of reasons are teleological explanations to which no action-relevant information can be added by causal relations. That the standard theory has not lost any of its attraction today is validated by P. Hbl’s critique of Alvarez. In his opinion, facts as reasons necessarily have to be connected to an agent’s actions—via his psychological states. Reasons become reasons of the agent (as opposed to reasons for the agent) when he is conscious of them. According to Hbl, Alvarez’ teleological explanations cannot escape the distinction that an agent pursues either actual or fictive goals, that he has actual or apparent reasons. This distinction cannot be formulated without reference to psychological states. One consequence of Alvarez’ relinquishment of causality is also illuminated by Hbl. If reasons are facts beyond psychological states, is a story connecting facts and agents then totally impossible? “It is hard to see how reasons could be related to our moral and juridical practice of blame and punishment, if no causality were involved.” (this volume, 128) While some advocates of factive reasons, like Stoutland and Alvarez, tend towards weakening the relevance of causality in actions and explanations of action with a new version of the logical connection

Introduction: Through Contexts to Actions


argument, others hold that an explicitly causal analysis of action is necessary but has to be, in contrast to the standard theory, a version of agent causality. R. Stout thinks that the chain-model of causality needs to be replaced by a process model orientated towards Aristoteles’ qualitative concept of causality, with causality and acting in some form of convergence. At the centre of Stout’s proposal is the notion that causality has in each case to be qualified by a causal mechanism because only through such specifications can deviant relationships in causal statements be eliminated. But how, in the case of actions, is this mechanism to be picked out if it is to be distinguished from event-causally dissolvable ‘mechanisms’—like the wind drying the washing? Stout’s proposal, not unlike Aristotle’s, draws upon the mechanism of practical rationality. An agent attempts something with intention when that something happens due to a mechanism of practical rationality embodied in the agent. Practical rationality is then simultaneously qualitative agent causality. J. Bishop’s reaction, perceiving in Stout’s rationality mechanism only a conceptual and not an ontological analysis of agency, shows how not even Stout’s proposal can evade the classical criticism of Aristotle’s notion of a ‘substance causality’. Bishop tentatively proposes the coexistence of a process causality of rational action control through the agent, and an event-causal view of that same control in the ‘real world’. On the one hand, this proposal is a reaction to the difficulty of using Stout’s formulation of the rational agent mechanism really to grasp causality and not just a rational structure, for example the Practical Syllogism. On the other hand, Bishop’s proposal of different levels involves new difficulties and would certainly not be compatible with the claim of mechanisms to be real. For now this discussion therefore has to end in a stalemate. Unlike the authors of the two preceding contributions, T. Lekan is of the opinion that an analysis of actions should not begin with the connection between reasons for and causes of individual actions, but with a look at socio-psychological contexts, orientated towards the concept of habits. In the tradition of Dewey’s behavioural theory, Lekan reminds us of a repertoire of concepts and descriptions which, in the present analytic theory of action, has largely been put to one side. Dewey’s and Lekan’s vocabulary is aimed explicitly against a ‘Cartesian agent’ and it endeavours to reverse the direction of dependencies. Given the breadth of this attempt, Lekan’s contribution to this book is more of a programmatic outline than systematic proof of the fruitfulness of this theoretical reversal.


Anton Leist

Regarding the breadth of the term “habit”, N. Anwander’s differentiation between social practices and individual dispositions is helpful. The dependency of our action on socially widespread manners of acting, on a pregiven logic, is far greater than we usually think. Nevertheless, the scope for individual behaviour with regard to given practices is in itself so great that Anwander does not view the predominant philosophical interest in actions per se, as opposed to particular individual actions, as in any danger. Social practices are not merely acted out, but interpreted individually. A similar thing is true of habits as individual dispositions. Dispositions, practices and reasons are mutually interdependent, the first two actually entrenching reasons more than suppressing them. According to Anwander, the different currents in action theory are on a convergence, not a collision course. Pragmatism would bring a particular social enrichment to the external conception of reasons, not suppress it. How do the premises of the standard theory stand up when confronted with the Kantian notion of moral will, with the ontological problems of personal identity or the question of what makes a life-story a reasonable unit? These are the questions which Velleman/Roughley, Cuypers/Slors and Bittner/Baumann take up in the second part of this book. D. Velleman raises the question of what an agent’s ability to will is good for. According to Velleman, there are some difficulties regarding this question within Bratman’s planning theory. Not all intentions are subordinate to the pragmatic function of being fitted into action plans. Many actions are not planned in advance, instead resulting spontaneously within a situation. Also, in Bratman’s model, intentions are too weakly connected with a self-knowledge of the agent, too closely linked to a narrowly comprehended pure volition, meaning that they cannot fulfil the pragmatic function of action coordination which Bratman alleges. From this Velleman concludes that a pragmatic-functional explanation of intentions and volitions is missing the point altogether, and that the ‘myth of intelligent design’ of the will should be replaced by a ‘myth of chance’. Just as Mediaeval architects invented the spandrel by chance, the will could also have its origins in chance and consequently not be completely subordinate in its functionality to the rational optimisation of desires—something necessarily to be assumed in the planning theory. Velleman, following on from Kant, views the will as an irreducible source of self-knowledge and conflicts, from which desires are systematically shaped, but can also be arbitrarily shed. In his comment, N. Roughley suggests that Bratman’s pragmatic function of planning optimisation should be taken as an abstract-

Introduction: Through Contexts to Actions


theoretical explanation, meant neither conceptually nor evolutionistically, and not in need to cohere directly with the intentions of particular agents. Like Velleman, Roughley takes proximate intentions to be inexplicable through distal intentions, and thus believes differing explanations to be appropriate. In contrast, Roughley remains reticent about Velleman’s attempt to explain intentions through cognitive attitudes and, further, through the ‘deep desire’ to know what one is doing postulated by Velleman. Bratman’s non-cognitive theory appears to him to be less speculative and therefore preferable. S. Cuypers reports what has been achieved in the debate on ‘personal identity’ by analytical philosophers to date as a stalemate between the empiristic and the Cartesian tradition. Whereas the Cartesian tradition could explain personal responsibility but is metaphysically problematic, the empiristic tradition, especially in the shape of Parfit’s theory, leads to unacceptable practical consequences, to the reduction of human beings to raindrops, as Cuypers puts it. In his attempt to develop a less paradoxical concept of personal identity, Cuypers draws upon Strawson’s famous view of persons as irreducible, ‘logically primitive’ units of body and mind, and especially upon his portrayal of persons as agents with intentions and reactive attitudes, embedded in a social world of co-agents. Taken further, Cuypers views these irreducible persons as agents within Davidson’s rationality holism, in which agents’ manners of acting prove them to be temporally identical. All paradoxes and practical absurdities of the metaphysics of persons could be avoided, according to Cuypers, if persons and their characteristics were grasped from the premise of their prevailing existence as agents. M. Slors wishes less to criticise Cuyper’s plea for a ‘logical personalism’ and more to convey support through the neurosciences. He tries to show that Strawson’s hypothesis of the irreducibility of persons was directed less at the metaphysical and more at a ‘descriptive dualism’, as anchored in everyday speech. Descriptive dualism has to be scientifically corrected, not philosophically. According to Slors, current research in the neurosciences provides stronger confirmation of personalism than the metaphysical debate. Cuypers believes narrative identity to be a more particular, existential form of agent identity, and he would have little difficulty connecting the role of acting with the biography of a person and his or her ‘life’. From the point of view of its actorial personalism, a human life would have to consist essentially in a number of actions. R. Bittner believes this view to be wrong, seeing its philosophical origins in the Sophists’ conviction of


Anton Leist

the teachability of virtues and in Aristotle’s synthesis of virtue and responsibility. Unlike the Sophists, he believes that virtues at best ‘grow’ inside us, not that we actively control them. What we achieve—more or less by chance—as ‘deeds’ or results is crucial, not our governing of actions. This scepticism towards the agent extends critically to the notion that an individual life could be given a direction or an aim. Rather, our lives lie beyond our activities and are best observed when we are dead. The same is principally true of what Bittner thinks to be the most suitable way of dividing up our lives, namely into units which he calls ‘stories’, but which he refuses to define. Stories are pieces of individual history, within which responsible action is possible at selected times, but without the ability to control their entirety. Bittner could agree to the idea of a ‘narrative identity’, but not to a personal identity as one of agents. Persons find themselves in the imponderabilities of the stories absorbing them. Constitutive of Bittner’s scepticism towards the relevance to life of actions is a distinction between the perspective of the agent and that of the spectator. In his comment H. Baumann criticises Bittner’s application as too sweepingly exclusive. In acting, agents are often aware of their role from the spectator perspective, our conventional understanding of action always permitting a certain degree of passivity when acting—in contrast to Bittner’s tendency to assume a strict opposition in acting between activity and passivity. Furthermore, a swimmer is aware of being carried by the waves and perceives his action as facilitated by this, not prevented by it. Only if the untenable reductive view of acting as pure activity is relinquished can certain forms of passivity acquire their actorial sense, namely in a perception as failures with regard to a particular action. Bittner could not give any plausibility to his concept of stories if he did not silently concede a logical priority of acting. In the third part of this book, Hornsby/Saporiti, Leist/Schulte and Hookway/Hetherington address the mutual dependency of acting, beliefs and knowledge. J. Hornsby takes up the problem, already touched upon several times in this introduction, of the compatibility of reasons as beliefs and as factive reasons. She finds the attempts at synthesis by Ryle, Dancy, Smith and Stout all unsatisfactory in different ways. Perhaps Smith’s attempt is the most reasonable, according to which normative and explanatory reasons are simply claimed to be mutually independent, but it is criticised by Hornsby for robbing explanatory reasons of their potential normative content. Without an internal connection between the two types of reason, explanations of actions are reduced to explanations of physical processes. Smith and Stout have attempted to eliminate this

Introduction: Through Contexts to Actions


problem (also relevant to Bittner), yet unsuccessfully in Hornsby’s view. Her own attempt at a solution draws upon an analysis of the distinction between belief and knowledge, according to which the difference between belief and knowledge consists in the adding-on of the given external state of affairs, without a difference in the subjective state of the believer. For Hornsby acting is always acting at least from a belief which can objectively become knowledge when the corresponding state of affairs applies. In her comment K. Sporiti illuminates the predicament which moves Hornsby to find identical meaning of the term “reasons” in the explanatory and normative uses of the term, and arrives at the conclusion that both uses can exist harmlessly side by side without, as Hornsby believes, a puzzle emerging. According to Saporiti the rationality of the erring agent is not called into question when no factive reason corresponds to his subjective-justifying use of “reason”. The two uses of “reason” are usually closely linked, meaning that, even in individual cases to the contrary, they can still be linked by presumption. Moreover, she believes Hornsby’s proposal to be unworkable. A belief is not alleged knowledge, but simply a belief. Belief and knowledge are separated by the concept of being-true, which there would be no point in discovering if one could content oneself with the belief alone. The two concepts of justifying and explanatory reasons (and according to Saporiti some more uses of reasons besides) can exist side by side very well without the need for semantic bridges. Like Hornsby, Leist/Schulte and Hookway/Hetherington are of the opinion that the theory of knowledge has to be connected with the theory of action, albeit in a somewhat different scope or context. A. Leist is interested in how cognition and action might be generally connected and criticises the traditional attempts to separate them, including in particular the ontological model of two ‘directions-of-fit’. Instead of separating cognition and action, he proposes grasping cognition as a special form of action, as action with its own sort of action aims. With regard to this antiontological strategy, J. Schulte points out in his comment that not even the simplest concept of action can survive without the concept of environment, and that ontological prerequisites can therefore not be evaded altogether. He perceives the relationship between cognition and action less as a pair of opposites and more as components within a list, from which no serious problems are to be expected. In contrast to Leist’s attempt to subsume all representative human characteristics within these


Anton Leist

two concepts, he sees in some passive behavioural forms a third possible category also belonging to a potentially open list. Ch. Hookway embeds knowledge in the action process of inquiry and proves it to be irreducibly practical. Following Rorty, he views truth not as a goal of cognition, so that the classical search for truth would have to be replaced by a more particular logic of epistemological rationality. He sees this proposal confirmed in the appropriateness of a practical rather than theoretical answer to scepticism. St. Hetherington is less critical of Hookway’s epistemic pragmatism, instead extending it by means of an analysis of the capabilities of knowledge linking propositional knowledge to non-propositional knowledge. Action theory could surely play a more dominant role within the philosophical disciplines than is the case in their current agenda, this at least is amply shown by the papers in the second and third part of this book in particular. Actions are not objects, but a background for many things which, without reference to action, would not be sufficiently comprehensible. Whether, as suggested by the title of this introduction, actions themselves then also have to be comprehended differently, is a question which this book will undoubtedly help to examine.37 Bibliography Anwander, Norbert (2002): “Intelligibilitt und Normativitt”, in: Analyse & Kritik 24, 231 – 248. Bishop, John (1989): Natural Agency, Cambridge: Cambridge University Press. Bittner, Rdiger (2001): Doing Things for Reasons, Cambridge: Cambridge University Press. Brand, Myles (1984): Intending and Acting, Cambridge/MA: MIT Press. Brandom, Robert (2002): Tales of the Mighty Dead, Cambridge/MA: Harvard University Press. Bratman, Michael (1985): “Davidson’s Theory of Intention”, in: Ernest LePore/ Brian McLaughlin (eds.): Actions and Events, Oxford: Blackwell Publishers (also in Bratman 1999). — (1984): “Two Faces of Intention”, in: Philosophical Review 93, 375 – 405 (also in Bratman 1987). — (1987): Intentions, Plans and Practical Reason, Cambridge/MA: Harvard University Press. 37 I would like to thank H. Baumann, R. Bittner, J. Schulte, F. Stoutland for critical comments and S. Kirkby for fundamental linguistic help. A shortened version of this introduction appeared in Deutsche Zeitschrift fr Philosophie 55.4, 2007. Also I am grateful to the Swiss National Science Foundation for financial support.

Introduction: Through Contexts to Actions


— (1999): Faces of Intention, Oxford: Oxford University Press. — (1999a): “Practical Reasoning and Acceptance in a Context”, in: Bratman (1999). — (2001): “Two Problems about Human Agency”, in: Proceedings of the Aristotelian Society 2000 – 2001, 309 – 326 (also in Bratman 2007). — (2002): “Hierarchy, Circularity, and Double Reduction”, in: Buss, Sarah/ Overton, Lee (eds.): Contours of Agency, Cambridge: Cambridge University Press, 65 – 85 (also in Bratman 2007). — (2007): Structures of Agency, Oxford: Oxford University Press. Buss, Sarah (2007): Autonomous Action: Self-Expression in the Passive Mode (Ms.). Dancy, Jonathan (1995): “Why there is Really No Such Thing as the Theory of Motivation”, in: Proceedings of the Aristotelian Society 95, 1 – 18. — (2000): Practical Reality, Oxford: Oxford University Press. — (2003): “Replies”, in: Philosophy and Phenomenological Research 67, 468 – 490. Darwall, Stephen (2003): “Desires, Reasons, and Causes”, in: Philosophy and Phenomenological Research 67, 436 – 443. Davidson, Donald (1970): “Agency”, in: Davidson 1980, chap. 3. — (1980): Essays in Action and Events, Oxford: Oxford University Press. — (1984): Inquiries into Truth and Interpretation, Oxford: Oxford University Press. — (1987): “Problems in the Explanation of Actions”, in: Davidson (2004). — (2001): “Aristotle’s Action”, in: Davidson (2005). — (2004): Problems of Rationality, Oxford: Oxford University Press. — (2005): Truth, Language and History, Oxford: Oxford University Press. Donagan, Alan (1981): “Philosophical Progress and the Theory of Action”, in: Proceedings and Addresses of the American Philosophical Association, 25 – 52. Dretske, Fred (1988): Explaining Behavior: Reasons in a World of Causes, Cambridge/MA: MIT-Press. Enoch, David (2006): “Agency, Shmagency: Why Normativity Won’t Come from What is Constitutive of Action”, in: Philosophical Review 115, 169 – 198. Frankfurt, Harry (1987): “The Problem of Action”, in: American Philosophical Quarterly 15, 157 – 162. — (1999): “The Faintest Passion”, in: Frankfurt, Harry (1999): Necessity, Volition, and Love, Cambridge: Cambridge University Press, 95 – 107. Hornsby, Jennifer (1980): Action, London: Acumen. — (1993): “Agency and Causal Explanation”, in: Heil, John/Mele, Alfred R. (eds.): Mental Causation, Oxford: Oxford University Press, 129 – 153. — (2004a): “Agency and Action”, in: Steward, Helen/Hyman, John (eds.): Agency and Action, Cambridge: Cambridge University Press, 1 – 23. — (2004b): “Agency and Alienation”, in: De Caro, Mario/Macarthur, David (eds.): Naturalism in Question, Cambridge: Cambridge University Press, 173 – 187. Hursthouse, Rosalind (1991): “Arational Action”, in: Journal of Philosophy 88, 57 – 68. Malpas, Jeff E. (1992): Donald Davidson and the Mirror of Meaning, Cambridge: Cambridge University Press.


Anton Leist

McCann, Hugh (1991): “Settled Objectives and Rational Constraints”, in: American Philosophical Quarterly 28, 25 – 36. — (1998): The Works of Agency, Ithaca-London: Cornell University Press. Mele, Alfred R. (1992): Springs of Action, New York: Oxford University Press. — (2003): “Philosophy of Action”, in: Ludwig, Kirk (ed.): Donald Davidson, Cambridge: Cambridge University Press, 64 – 84. O’Shaugnessy, Brian (1980): The Will: A Dual Aspect Theory, Cambridge: Cambridge University Press. Peacocke, Christopher (1979): “Deviant Causal Chains”, in: Midwest Studies, 123 – 155. Raz, Joseph (1999): “Agency, Reason and the Good”, in: Raz, Joseph (1999): Engaging Reason, Oxford: Oxford University Press, 22 – 45. Ryle, Gilbert (1949): The Concept of Mind, Oxford: Oxford University Press. Santiago, John (2005): “Personal Autonomy: What’s Content Got to Do With It?”, in: Social Theory and Practice, 31.1, 77 – 104. Scanlon, Thomas M. (1998): What we Owe to Each Other, Cambridge/MA: Harvard University Press. Searle, John (2001): Rationality in Action, Cambridge/MA: MIT Press. Seebaß, Gottfried (1993): Wollen, Frankfurt: Klostermann. Smith, Michael (1987): “The Humean Theory of Motivation”, in: Mind 96, 36 – 61. — (1994): The Moral Problem, Oxford: Oxford University Press. — (1998): “The Possibility of Philosophy of Action”, in: Bransen, Jan/Cuypers, Stefaan (eds.): Human Action, Deliberation and Causation, Dordrecht: Reidel, 17 – 41. — (2003): “Humeanism, Psychologism, and the Normative Story”, in: Philosophy and Phenomenological Research 67, 460 – 467. — (2004): “The Structure of Orthonomy”, in: Hyman, John/Stewart, Helen (eds.): Agency and Action, Cambridge: Cambridge University Press, 165 – 193. Stoecker, Ralf (2002): “In den Zeiten, wo das Wnschen noch geholfen hat”, in: Analyse & Kritik 24, 209 – 230. Stout, Rowland (1996): Things that Happen Because they Should: A Teleological Approach in Action Theory, Oxford: Oxford University Press. Stoutland, Frederick (1985): “Davidson on Intentional Behaviour”, in: LePore, Ernest/McLaughlin, Brian (eds.): Actions and Events, Oxford: Blackwell Publishers. — (1986): “Reasons, Causes, and Intentional Explanation”, in: Analyse & Kritik 8, 28 – 55. — (1997): “Why are Philosophers of Action so Anti-Social?” In: Alanen, Lilli/ Heinmaa, Sarah /Wallgren, Thomas (eds.): Commonality and Particularity in Ethics, London: Macmillan Press, 45 – 74. — (1998): “The Real Reasons”, in: Bransen, Jan/ Cuypers, Stefaan E. (eds.): Human Action, Deliberation and Causation, Dordrecht: Reidel, 43 – 66. — (1999): “Intentionalists and Davidson on Rational Explanation”, in: Meggle, Georg (ed.): Actions, Norms, Values, Berlin-New York: DeGruyter, 191 – 208.

Introduction: Through Contexts to Actions


— (2001): “Responsive Action and the Belief-Desire Model”, in: Grazer Philosophische Studien 61, 83 – 106. Thalberg, Irvin (1985): “Analytical Action Theory: Breakthroughs and Deadlocks”, in: Seebaß, Gottfried/Tuomela, Raimo (eds.): Social Action, Dordrecht: Reidel, 1 – 41. Tuomela, Raimo (1995): The Importance of Us: A Philosophical Study of Basic Social Notions, Dordrecht: Reidel. Velleman, David J. (1992): “What Happens, When Someone Acts?”, in: Mind 101, 461 – 481 (reprinted in: Velleman 2000). — (2000): The Possibility of Practical Reason, Oxford: Oxford University Press Wallace, R. J. (2003): “Explanation, Deliberation, and Reasons”, in: Philosophy and Phenomenological Research 67, 429 – 435. Williams, Bernard (1981): “Internal and External Reasons”, in: Williams, Bernard (1981): Moral Luck, Cambridge: Cambridge University Press, 101 – 113. Wilson, G. (1980): The Intentionality of Human Action, Amsterdam: North Holland. — (1985): “Davidson on Intentional Action”, in: LePore, Ernest/McLaughlin, Brian (eds.): Actions and Events, Oxford: Blackwell Publishers. — (2002): “Action”, in: Stanford Encyclopedia of Philosophy.

Thought and Action P. M. S. Hacker 1. Agency Human beings are agents, acting on things around them, and producing, preventing or sustaining change. They are not unique in their agency. Other living creatures are agents too, and so are inanimate things. An agent, in the most general sense of the term, is something that does something or acts. Inanimate agents are paradigmatically inanimate substances and partitions (lumps, nuggets, drops) or specific quantities (pints, pounds or yards) of stuffs. Inanimate agents do things. Among the things they do are things they do to something or other. We must distinguish, among things done to something, between those the description of which entails a change to the patient (causative verbs such as “break”, “bend”, “melt”) and those the description of which does not (non-causative verbs such as “hit”, “touch”, “push”). In these ways, inanimate agents act on patients. In some such cases, they bring about change. But they may also prevent or suppress a change that would otherwise have occurred. Both the causative and the non-causative verbs, applied to an inanimate agent, are used to describe things it does, but it performs no deeds. Inanimate agents neither act nor take action. They may have an activity (as enzymes do), but they do not engage in an activity. Artefacts may have an action, or be put out of action. The powers of inanimate agents are one-way powers. If the conditions for their actualization are satisfied, these powers are actualized. But inanimate agents do not exercise their powers—only beings that can take advantage of an opportunity to act can be said to do that. Animate agents may be sentient or insentient. Sentient agents of developed kinds possess perceptual and volitional powers. Their volitional powers are two-way powers to act or refrain from acting as opportunity is available. They act in pursuit of goals, persist in the face of obstacles, display a relatively wide range of adaptive behaviour in response to the circumstances of their goal-pursuing behaviour. They also have active and passive hedonic powers. What an animal that possesses such a


P. M. S. Hacker

rich repertoire of behaviour and response does in pursuit of its goals typically warrants a description in terms of its knowledge of the circumstances and of its wanting to attain, and aiming at, a certain goal. Its knowledge and its wanting its goal are not causes of its behaviour— they inform it. The link between perceptual, and hence cognitive, volitional and hedonic powers is not coincidental, but conceptual. So too is the connection between them and the power to act and to take action. A criterion for whether an animal wants a certain thing is that if it perceives that thing and apprehends an appropriate opportunity for obtaining it, then, other things being equal, it takes steps to get it. When it gets what it wants, it is typically pleased, and does with it what it wanted it for. Only volitional agents can be said to act or to take action. To take action is to act voluntarily in response to a circumstance (e.g. a threat or perceived danger) or in pursuit of a goal given perception of an opportunity. It involves purposive movement in pursuit of a goal or in avoidance of a perceived danger (we would not deem ‘freezing’ at the perception of a threat as taking action). But only a being that can take action can act. For a creature that can be said to take action is a creature that can do or refrain from doing something voluntarily, that can have and seize an opportunity (as well as forego one), that can opt for or choose one course of action over another. The boundary between mere animal behaviour—in the case of insects or fish—and the voluntary actions of more developed creatures is broad and blurred. We speak more readily of animal behaviour than of the acts animals perform. This may be a consequence of the fact that we invoke the formal concept of an act largely in contexts in which intentional and rational adjectival and adverbial modifications of what was done by an agent are in question. So, having been informed that someone V-ed, we may query whether his act was intentional or unintentional, deliberate, reasonable, spontaneous, benevolent or malicious, or negligent, and so forth— through a wide range of epithets that have little or no application to nonlanguage using agents. Human beings are rational agents. They have not only perceptual and volitional powers, as do other animals, but also rational powers—the powers of the intellect (the ability to reason that informs our lives) and of the will (the ability to act for reasons that is a guide to our lives). The faculty of reason is exhibited in reasoned discourse and in action and reaction, in giving reasons why and reasons for, and in reasoning from grounds to conclusions. Our rationality, our limited responsiveness to

Thought and Action


reasons, and our fallible ability to reason are corollaries of our being language-users. For only language-users can engage in reasoning, and only language-users can deliberate and give reasons for what they think and do. Only creatures that can deliberate (weigh reasons) and form their beliefs and resolve to act on the basis of reasons can be said to have reasons for thought and deed. The essential unity of thought and action, the multiple conceptual connections between cognitive and cogitative concepts, on the one hand, and volitional concepts and concepts of action, on the other, was, and continues to be, a sorely neglected or misunderstood theme in the empiricist tradition. No analytic philosopher in the twentieth century was more aware of this flaw in empiricist thought than Stuart Hampshire, and few made such illuminating efforts to delineate the web of connections between thought and action than he. His book Thought and Action1 is unjustifiably neglected now. But it contains a wealth of insights that should not be forgotten. For the rest of this paper I shall try to sketch the main contour lines of Hampshire’s work, both in Thought and Action and in his later book Freedom of the Individual (Hampshire 1965), in the belief that it contains much that provides part of the essential foundation for philosophical reflection on human beings as agents.2 2. Thought and Action “We are in the world”, Hampshire nicely remarks, “as bodies among bodies, not only as observers but as active experimenters. We could not ever be observers unless we were sometimes active experimenters, and we could not ever be experimenters unless we were sometimes observers.” (53) We could not be spectators of the passing scene unless we were also actors in the unfolding play, could not be perceivers unless we were also manipulators, could not be thinkers unless we were also doers. This is a truism to which philosophers have paid too little attention. For it has been part of the Cartesian heritage that has cast its long shadow over philosophy of the modern era that thought, sensation, sense-perception, emotion and will are given, incorrigibly and indubitably, in logical 1 2

Hampshire 1959; parenthetical references in the text are to this volume. The remainder of this paper is a somewhat altered version of my paper “A Tribute to Stuart Hampshire” (Hacker 2005).


P. M. S. Hacker

independence of knowledge of ourselves as active animals in a world of material things. It has been part of the empiricist heritage handed down to us by Hume, that our experience consists of a succession of impressions and ideas. And even Kant, who sensed more clearly than any other philosopher the fundamental flaws in the Cartesian and empiricist conceptions of ourselves and our experience of the world around us, entertained the idea, elaborated in his imaginary subject of transcendental psychology, that what he called ‘the unity of consciousness’ is the result of the mind’s transcendental synthesis of intuitions given it in the forms of space and time. Such synthesis is indeed an activity, but an activity of the transcendental self, not of an active agent who is a spaceoccupying particular moving within a world of objects. It was Hampshire’s contention that failure to apprehend the unity of thought and action, vitiated and continues to vitiate a great deal of modern philosophy. The task he set himself in Thought and Action was to unravel the conceptual commitments and involvements consequent upon recognition of the primacy of intentional action in our conceptual scheme. That we are, and that we unavoidably conceive of ourselves and are aware of ourselves, as active agents occupying a particular place at a particular time in a world of material things through which we move intentionally has implications that ramify throughout our conceptual scheme. Our conceptions of ourselves as space-occupying particulars and of other space-occupying particulars as distinct from ourselves, our concepts of our perceptions and their objects, of our intentions and their execution in action, of our beliefs and their truth or falsity form a seamless web, each element of which is necessarily connected, directly or indirectly, to all the others. The interconnectedness of these concepts, Hampshire argues, renders incoherent the philosophical vision embodied both in the Cartesian and in the empiricist world-picture. Hampshire announces his themes in the ‘Introduction’ as an exploration of … certain familiar contrasts: the contrast between that which is unavoidable in human thought and that which is contingent and changeable: between inner thought and its natural expression in speech and action: between that which confronts a man as the situation before him and that which is his own response to it: between knowledge and decision: between criticism and practice: between abstract philosophical opinions and the concrete varieties of experience. (9)

Thought and Action


His aim, he declares, is “not to find definite solutions to the problems connected with these contrasts” (9), but to trace connections between them, and thereby to bring moral argument closer to philosophy of mind. 3. The Subject We are in the world as bodies—spatio-temporal particulars—among bodies, not only as observers but as active experimenters. Every observer is aware of himself as one such space-occupying item in the world. The external world is a system of independently existing objects, acting and reacting upon one another. We perceive such objects, not as ‘outside the mind’, but as outside, or better, as independent of, us. They are independent of us not merely in the sense that their existence is independent of our perception and of our thought about them, they are spatially discontinuous with our body, and also, unlike our body, independent of our will—there being no such thing as psycho-kinesis. A cardinal error of empiricist philosophers was to represent us as passive observers receiving impressions from ‘outside’ the mind, where ‘outside’ includes our own body. But we find ourselves, from the very beginning able to act upon objects around us. And to act thus is to move at will and to bring about changes in other bodies, to move them by pushing or pulling them, to pick them up and handle them. I not only perceive my body, I also control it. I move, and move my limbs, at will. I do not always need to infer from my observations alone that I have made a movement. Rather, I know directly that I moved or tried to move, or that I did not try, but was rather moved by something else. No knowledge, Hampshire stresses, is more direct and underived than this knowledge of the fact of my own intention to move or to bring about change (47 f.). I find my power of movement limited by the resistances of objects around me. This felt resistance to my will, in conjunction with my perceptions, defines me and my own situation as an object among other objects. It is a mistake to try to delineate fundamental features of our conceptual scheme on the basis of the supposition that our only contact with objects and so with the world of physical things, is through perception, in which objects are presented to our passive mind. Both in perception and in the least of our movements, we are aware of ourselves as one object among others, one space-occupying particular among others. Being self-moving voluntary agents, we move among the objects around us. Being perceivers, our viewpoint changes as we move. The changing standpoint


P. M. S. Hacker

of the observer, Hampshire stresses, is to be understood as a change of his situation in the world, not in his relation to the world from the ‘outside’ as it were—as the empiricists represented it as being. One might maintain that we are aware of ourselves as bodies solely through the impressions of the senses. One might then further argue that when we move and act, we always verify this by the sense of touch and kinaesthetic sensation. But even if we say this, we cannot go further and say that it is only by touch and the other senses that we know that we possess a body and are not merely detached thinking things (50). 4. Perception So, again, we are not only observers, we are also actors—experimenters. To observe is, inter alia, to learn what obstructions there are in the environment. To experiment, in its original sense, is to act with a view to perceiving what then happens. Each language-using human being perceives objects around him, and in identifying them, also fixes in his mind his own situation in the world. All perception of which we can form any idea is the perception of a finite observer moving among the objects he observes from changing points of view in a common world. The conception of experience as a succession of impressions and ideas passively received and then compared by the mind, detached from embodiment in any one position, is incoherent. For then it would be a contingent matter that one perceives with one’s eyes, feels things with one’s hands and body; that one’s visual and auditory perceptions change as one moves; that one can switch one’s observations at will from one object to another. But these are not contingent matters of fact at all. Sense-datum philosophers inherited the empiricist conception. Their failure to translate categorical material object statements into hypothetical statements about sense-data was rooted in the fact that, at the very least, the antecedents of the envisaged hypotheticals must specify the objective standpoint of the observer—and that fact signifies the futility of their programme. The fundamental error is not to acknowledge that the standpoint of the observer is one physical fact among others—that the observer is always a self-moving body among other bodies which he observes and intentionally manipulates. A physical object is recognized as a potential obstruction or as something to be manipulated, occupying some definite position in relation to the observer at the moment of perception. Touch, not sight, is primitively the most authoritative of the

Thought and Action


senses, the natural criterion of physical reality, just because acting upon objects necessarily involves touching—the contact of one’s body with a resisting body that is not one’s own (48). And touching, handling and manipulating things are misrepresented as the passive reception of data— for here perception and action are united. It is essential to action that we should in general be able to discover by observation whether we have in fact achieved what we intended to achieve. Action and perception are thus complementary, and can neither be assimilated one to the other nor separated one from the other (52). We have here a battery of insightful remarks. But this bird’s eye view needs supplementation by detailed analysis of the conceptual connections indicated but not demonstrated. Let me sketch a few fragments of the kinds of arguments that might further the enterprise. It is indeed a non-contingent fact that we see with the organ of vision, denominated “the eye”. For an organ to be the organ of vision, it is necessary, but not sufficient, that it be light sensitive. In addition, it must be used in detecting from a distance the presence of visible objects and their visible properties, such as colour, shape, size, and movement. It must be useless in pitch dark; and it must be such that it cannot be used to detect objects that are occluded from the line of vision. A criterion for whether an animal sees something is whether it detects what it detects by the use of an organ satisfying such conditions, e.g. that it opens its eyes in order to detect and cannot thus detect visibilia with its eyes closed, that it moves closer in order to discern better, moves around objects to perceive what was occluded or partly occluded, puts its eye to an aperture in order to see through it, etc. Such characteristics are internal properties of what we call ‘possessing an organ of vision’. Similarly, it is not a contingent fact that an animal apprehends the presence of solid objects by touch—for it is a criterion for ascribing to an animal apprehension of the solidity of an object that it reacts appropriately to touching the object with its limbs and body and apprehending its relative impenetrability. These reactions are various, depending upon the size and nature of the object, the character of the perceiving animal and its purposes. That it does so apprehend the properties of the object is manifest in its behaviour, in whether the object is apprehended as an obstruction, which the animal then skirts around, or as something that can support the weight of the animal who then treads on it or climbs up or across it, or as something manipulable. And to manipulate an object, to stand on it, to lean against it, to push and pull it, to pick it up and throw it are not mere forms of perception, let alone of


P. M. S. Hacker

passive reception of ideas or impressions, but rather forms of interaction with things. Ideas and impressions may stand in regular relations of coexistence and succession, but, as Berkeley rightly noted, not of causation in its most primitive and basic sense, viz. making something happen by acting on it. And the hallmark of action and activity is substantiality. I can engage with my environment only because I am a living, self-moving substance. To manipulate an object is bring about changes in or to it by acting on it. To move an object is to bring about a change in its position by pushing, pulling or picking it up, etc., and to experience an object as an immovable obstacle is to try to move it by pushing or pulling it and to fail. These are actions, not impressions—actions of an intentional selfmoving substance among substances. There is no doubt that the empiricists mistakenly assimilated perception to passivity. But while there is a sense in which one is passive in seeing what one sees when one opens one’s eyes, it is obvious that in looking, inspecting, watching, scrutinizing one is active. It is not up to me to see or not to see what is visible before me when I open my eyes (although I may fail to notice things in view)—for our perceptual powers are not, in this respect, two-way powers to do or refrain at will, and their actualization is not a voluntary act. But it is up to me whether to look again or to look elsewhere, whether to follow a moving object with my eyes or to look at another object or just to shut my eyes, whether to move closer or to pick up the object and carry it to the light—and these are voluntary acts. In this sense, even in the case of vision, perception is intimately interwoven with action. And that too is built into our concepts by way of the criteria for ascribing to an animal visual detection—for it is a criterion for whether an animal has spotted a moving object that it follow the course of the object by moving its eyes and head to keep it in view, that it respond to the object’s visible change of position by changing its position. That my visual perceptions change as I move is indeed noncontingent. For if I move left or right, then A, which is occluded by B, becomes visible. If I move closer to B, then more of A becomes occluded. And if I move around B, then its back comes into view. It is part of the concept of an opaque material object that it has surfaces, not all of which are visible from one point of view. The occlusion properties of objects are their visible spatial relations.3 Given that the line of sight is straight, the mathematical theory of vision, first elaborated in Euclid’s Optics, is an a 3

For more detailed elaboration, see Hyman 1989, 80 f.

Thought and Action


priori theory of what is visible from a viewpoint—and there is no such thing as vision which is not from a viewpoint. The complementarity of action and perception is essential to the possibility of action in the execution of a plan. I would not be an agent unless I could act on items in my environment, and I would not be an intentional agent unless I could perceive what I intend to act on and form reasonable expectations accordingly. And, as Hampshire emphasizes, I could not foresee the future in my intentional actions unless I possessed such background inductive knowledge as enables me to form plans, nor could I execute my intentions unless I could follow the evolving trajectory of my actions. Such considerations, and a multitude of others pertaining to our concepts of space and time, of material object and of motion, of a perceiving agent and of perceptual faculties and organs, help to illuminate the manifold conceptual connections which Hampshire invokes and intimates, but often does not pause to elaborate. 5. Action We distinguish ourselves from other things, Hampshire notes (51), first, as being in a certain situation, as being here rather than there, and secondly, as being capable of planning to move from here to there. My situation is fixed for me as the place from which I move my arms and other limbs, which is also the place from which I see and move things around me. The thought of the possibility of changing my situation in the world is always present to me. As I move at will, my point of view correspondingly changes, and it is in this way that I explore the world arranged, as it were, around me as its centre (54). Even at the most rudimentary stages of consciousness, I always distinguish that which I myself do intentionally in changing my situation from that which happens to me (51). A human being is aware of himself as the moving locus from which he perceives all that he perceives. It is intentional movement that gives us our sense of ‘being in the world’, a sense which we would lack were the ‘given’ nothing but a succession of impressions and ideas. That empiricist thought would require us to think of our own movements as mere impressions too, and hence that the correlation between changes in one type of impression (bodily sensations) and changes in the other (the data of external sense) was merely a contingent and surprising matter of fact. “The most unavoidable feature of our


P. M. S. Hacker

consciousness”, Hampshire observes, “is the initiation of change at will, the changing of position and therefore of our relation to other things.” (68) That is why the Humean notion of experience is incoherent. For, not only should we be sceptical of the obtainability of the notions of impressions and ideas independently of notions of ourselves as selfmoving agents in a world of independent objects, but also we should be sceptical of the possibility of extracting from Hume’s meagre conceptual resources the essential distinctions between what we do and what happens to us, between what we wish and what we intend and do intentionally (70). An agent could not learn or coherently use a vocabulary, or make statements about himself and the world he encounters, unless he regarded himself as one self-moving observer among others, and one body among others, changing his own position at will. For it is a necessity in all but the most rudimentary uses of language that we should refer to persisting objects, employing some criteria of identity through change; and it is equally necessary that the speaker should have the means of indicating his own viewpoint. The first-person pronoun is not just one indexical among others. For the first-person singular form is the nucleus upon which all the other referential devices depend. It is, as it were, their point of origin. Any place referred to is at some definite distance from here, and any event or time referred to is traceable to now, to the speaker’s standpoint (87). And the final point of reference, Hampshire holds, the point at which all a person’s statements are ‘attached to reality’, is the speaker’s reference to himself as one thing, one person, among others. The Cartesian idea that the first-person pronoun could have an intelligible use to refer to a subject of experience independently of that subject’s existence as a being in the world of objects is not coherent. Some of the fundamental categories of thought can be traced to our elemental intentional acts of moving and manipulating and their conceptual involvements. It is, Hampshire contends, above all in intentional movement that temporal intervals of before and after impose themselves as necessary forms of all experience. For it is not only in perception, but also in making a movement or initiating a change, that we perceive the relation of before and after, as directly as we perceive the changing qualities of things. For any action, as the intended bringing about of an effect, has a trajectory, and so a relation of before and after within it (71 ff.). The future is that which is alterable in principle by action. And since we look upon the present situation as arising from the immediate past by some agency and as passing into some other situation

Thought and Action


by some force or agency that is operative now, the categories of causal explanation have their roots within our own experience of ourselves as agents (73). “To myself ”, Hampshire writes, “I am always a continuing thing producing changes, of which I know directly that I am the cause. The notion of a perceiving subject is also the notion of a continuing, embodied intentional agent, who displaces or is displaced by the things around him.” (85) And “I did it”, as opposed to “It just happened” or “Someone else did it”, is the primary and unquestionable indication of my own, utterly distinct existence as an object of reference (72). This requires elaboration. Let me indicate the direction in which thought must proceed. There would be no use for the indexical “here” other than as spoken by a space-occupying intentional agent located at a particular place, and no use for “now” save as indicative of the time of utterance. For their user, for me for example, ‘here’ is where I am and ‘now’ is when I utter it. Moreover, it is evident that there would be no use, either public or private, for the first-person pronoun, unless it were emitted from the mouth of the speaker located somewhere detectable or determinable at a given time. If the words of all speakers were always heard in the same tones from one and the same loudspeaker, no matter where we were located, the first-person pronoun would lose its point. The notion of my body is equally tightly woven into a web of internal relations. Although I can be said to be a body, i.e. a space-occupying particular, it would be wrong to say that I am my body. The use of the possessive here, which does not indicate possession, is singular. It is employed to describe a limited range of corporeal characteristics of a human being—surface features (A’s body is covered with dirt, sweat, mosquito bites), features of fitness and health (A’s body is fit or out of condition, healthy, ageing, feeble), aesthetic attributes (handsome, beautiful) and sensation (A’s body aches all over, itches, hurts). It is complementary to the phrase “my mind”; and I am no more identical with my mind than I am identical with my body. What has both a mind and a body is the living human being. A living human being is a body, i.e. a space-occupying particular of a specific kind, but cannot be said to be identical with his body. It is, I fancy, hyperbole to suggest that the thought of the possibility of changing my situation in the world is always present to me. No thought is ubiquitous. Paralysis and constraint apart, however, the possibility of moving is always with me, otherwise I would not be a voluntary agent, who can move at will. And it is also true that, paralysis and constraint


P. M. S. Hacker

apart, I always know that I can change my position and location at will, otherwise I would not be a self-conscious intentional agent, who can intentionally move from here to there for some reason. To be sure, I may become paralysed, lose my powers of voluntary movement, and still retain my powers to think and reason at will. But it is arguable, although it cannot be argued here, that the idea of an intentional agent who never had the power of voluntary movement (and speech), but who nevertheless had the power of thought is not coherent. There may be some sense in which it is true that “even at the most rudimentary stages of consciousness, I always distinguish that which I myself do … from that which happens to me.” But I am not sure what Hampshire has in mind here. To be sure, babies and animals alike react differently to what happens to them, e.g. when they are moved, from the way they behave when they themselves move. But a language user distinguishes in thought and speech between what he does voluntarily and intentionally and what happens to him. He possesses a multitude of concepts of activity and passivity, which he applies unreflectively to himself, and can say, without evidence, of any movement of his whether he moved or was moved. And this capacity is indeed integral to our concept of an intentional, language-using agent. Intentional agency necessarily involves not only the power to express one’s intentions without evidence, but also the knowledge of what one is doing when one executes one’s intentions. For if I V, and do not know that I am V-ing, then I am not V-ing intentionally, but unintentionally, accidentally or by mistake. So what I do intentionally, I do consciously and am conscious of doing, e.g. trying to persuade you that such and such. But although what I do unintentionally I do not do consciously, it is something I may become and be conscious of doing, e.g. boring you to tears. An intentional agent is the author of his intentions, and in determining my own intentions, I also know, ceteris paribus, what I shall be doing in the future when I execute them. For I foresee my future actions in my intentions. That knowledge is not inductive—I do not know that I shall be in London next week because I intend to go to London and know by induction that I normally execute my intentions. Rather, I can only be said to have formed the intention to V if I believe that I have both the ability to V and the opportunity to V, and have no reason to believe that I shall fail or be prevented from V-ing. So, having formed the intention to V tomorrow, I know, other things being equal, that I shall V tomorrow, in as much as my will is set, and I know that

Thought and Action


there are no obstacles to my project. Furthermore, my ability to say what I intend, and my knowledge of my own intentional actions are not derived from any impressions or ideas. To intend is not to have an impression, which I report in saying that I intend. It is neither an experience nor an inner act or activity. 6. Intentional Action It is one of the major strengths of Hampshire’s philosophical vision of ourselves and of our nature to give primacy to the concept of intentional agency. For he apprehended that the notion of intention is a node in our conceptual web at which a multitude of crucial conceptual strands meet: “the notion of the will, the relation of mind and body, the difference between observing a rule or convention and merely having a habit” (96), as well, we may add, as the notions of belief, of reasons for believing, acting and feeling, the various distinctions between what is passive and what is active in our lives, the notion of control and self-control, and that of freedom. An error running through empiricist philosophy is to represent human consciousness primarily as a state of passive awareness (94), of reception of experience in all its multicoloured variety, passing before the mind, of awareness of “a bundle or collection of different perceptions, which succeed each other with an inconceivable rapidity, and are in a perpetual flux and movement … a kind of theatre, where several perceptions successively make their appearance; pass, re-pass, and glide away, and mingle in an infinite variety of postures and situations.” (Hume 1739, T I.iv.4) But, Hampshire suggests, a much more decisive difference between waking and sleeping lies in the necessity of intended action in the one case and mere natural movement without intention in the other. In one’s waking life, there is necessarily always an answer to the question “What is he doing?”. To be a conscious human being, and therefore a thinking being, is to have intentions and plans, and to be able to answer the questions “What are you doing?”, and “What are you intending to do?” (119). Intention, Hampshire observes, is necessarily associated with the possibility of declaring one’s intention. It is characteristic of an intention that it may be formed long in advance of the action intended, and that it may have existed even though it was never executed. That, he suggests, is why intention, as opposed to purpose, is out of place in the context of animal behaviour. It is senseless to speak of what a dog intended to do


P. M. S. Hacker

before it was interrupted or prevented, unless “It intended to do so-andso next” is taken to mean “It would have done so-and-so next, but for being interrupted or prevented”. But that is not what it means in the human case. It is senseless to attribute intentions to an animal which has not the means to reflect upon, and to announce to itself or to others, its own future behaviour (97 f.). Only language-users can be said to have intentions. Expectation and decision are two complementary aspects of the notion of intentional action (111). One cannot say what a person is trying to do, unless one knows what he expects to happen. To try to V is necessarily to intend to V or to come as close as possible to V-ing. To decide to V is necessarily to form the intention to V. To intend to V is necessarily to believe that one will try to V. To intend something to happen as a result of one’s V-ing is minimally to believe that it may or could happen. For one cannot say “I intend E to happen, but I am sure it won’t” or “… but I believe it impossible”, although one may attempt the impossible to demonstrate its impossibility (134). A subject cannot be ignorant of his intentions, although he may make mistakes in stating them (134). A person unfailingly knows what he is trying to do, simply because it is his intention. For doing something intentionally entails consciously doing it; and intending to do something in the future entails knowing that one will do it or at least try to do it on the future occasion. At this point in his argument, Hampshire makes a striking move which is especially interesting in the light of one of Wittgenstein’s radical ideas. The foregoing considerations, he suggests, explain why there is no need for the reflexive knowing that would be implied by the phrase “knowing what one intends”. “I know now what I intend to do”, Hampshire notes, is a redundant way of saying “I know now what I shall do”, and “I know what my intentions in doing this are” is an impossibly redundant way of saying “I am doing this intentionally (or: with intention)” (102). There is no question either of a source of knowledge or of the justification of a claim to knowledge when someone announces that he knows what he is going to do. For he is not making a cognitive claim of any kind, but is rather declaring his intention, and in so doing revealing that his mind is made up (104). The claim that it is senseless to attribute intentions to animals is controversial. Apart from dogmatic behaviourists, we can all agree that animals are voluntary agents, they possess two-way powers of action, can do what they do because they want to do it and refrain from or resist doing something because they do not want to do it. Certainly they have

Thought and Action


purposes, and act in pursuit of their goals, try to fulfil their purposes and succeed or fail in their attempt. Some of the things they do are accidental, and surely to do something accidentally is to do it unintentionally? And a creature can do something unintentionally only if it can also do things intentionally. This is too quick. One can do something accidentally if what one does is not done on purpose, and we have granted that animals can act with a purpose. So if the dog, in rushing to the door to greet its master, knocks over a vase, it has done so accidentally—doing so was not part of its purpose. Should we deny that mere animals are intentional agents? It is interesting that two leading philosophers of psychology working today take partially opposing views on the matter. Anthony Kenny is in agreement with Hampshire. The grounds for his denial that animals can have intentions is that only language-using creatures can act for reasons, and to act intentionally is to act for a reason. But only language-users can act for a reason, in as much as a creature can only be said to act for a reason if it can give a reason to itself or to others. Unlike humans, animals cannot perform acts as answering to a description. So although they can obey commands, they cannot obey a command with the intention that their act satisfy the description specified by the command (Kenny 1989, 38 f.). By contrast, Bede Rundle holds that although it makes no sense to ascribe to an animal an intention formed in circumstances remote from those of its execution, and hence of ascribing to it an intention or having an intention which has been formed in advance of action, nevertheless “intentional” is interchangeable with “purposive”. Furthermore, when, for example, a lion manoeuvres itself into position for attack and waits for the appropriate moment to charge, we can say that it has the intention of charging, because we are in effect watching a pattern of behaviour that is already under way (Rundle 1997, 69 f.). I agree with Kenny that intentional action is action done for a reason. So if mere animals cannot have reasons for acting and therefore cannot act for a reason, then they cannot have intentions. And if only creatures who can give reasons (and, one might add, can reason), can act for a reason, then animals cannot act for reasons. I agree with Hampshire that if one can only have intentions if one can form intentions in advance of action, reflect on one’s intentions, and have intentions that are never translated even into incipient action, then animals cannot be said to have intentions. But I also agree with Rundle that “intentional” converges on “purposive”, and that we are inclined to say of an animal that exhibits a


P. M. S. Hacker

trajectory of action already underway that it has the intention of continuing the trajectory of the action. The question is whether one is willing to invoke the notion of reason in an attenuated sense, and to apply it to animals, even though they cannot give or reflect on reasons and cannot reason; and so too whether one is willing to apply the notion of an intention in an attenuated sense, and apply it to animals who exhibit a trajectory of behaviour, even though they cannot be said to form intentions in advance of action, cannot announce to themselves or to others their intentions, and cannot reflect on their intentions and abandon them on reflection. The matter seems to me more one for reflective decision than for discovery or for realisation of a familiar conceptual link. Hampshire’s suggestion that a subject cannot be ignorant of his intentions, that there is no need for the reflexive knowledge that would be implied by the phrase “knowing what one intends”, since there can neither be a source of such ‘knowledge’ nor a justification of a knowledge claim when someone announces his intention, and his contention that “I know now what I intend” or “I know what my intentions are in doing this” are hopelessly redundant ways of saying “I know now what I shall do” and “I am doing this intentionally” constitute what might be called a ‘redundancy theory of knowledge of first-person intention’. Accordingly, in saying “I know now what my intentions are”, one is not making a cognitive claim at all, but declaring one’s intention and, in so doing, revealing that one’s mind is made up. This is interestingly reminiscent of Wittgenstein’s non-cognitive account of avowals, not only of intentions, but also beliefs, thoughts and sensations. Wittgenstein’s suggestion, with which I think Hampshire would agree, is not that the operator “I know” is redundant here, but that it is not used to make a cognitive claim. Here, “I know” may serve the purpose of expression of decision, or of concession, or, in appropriate contexts, of making a grammatical statement—as in “Only you can know your intentions”, which amounts to the statement that the subject’s word carries special, non-inductive, weight, and that in the first-person present tense case, there is no such thing as doubt. What is important is that the epistemic operator here does not fulfil an epistemic role.4 Ignorance is excluded, both in the case 4

There is nothing odd, let alone contradictory, about this. In “While you were with me, I forgot all my troubles”, the epistemic verb “forget” plays no epistemic role. The assertion does not signify a failure of memory or loss of knowledge, but a distraction of attention.

Thought and Action


of intentions and of beliefs, simply because the intentions and beliefs are, in a non-trivial sense that will be clarified below, one’s own intentions and beliefs. One is, Hampshire emphasizes, the author of one’s intentions, and if not the author of one’s beliefs, then at any rate answerable for them not only to others but also and necessarily to oneself. And if ignorance is logically excluded, then so too is knowledge. For any proposition ascribing knowledge is itself a contingent proposition, which has sense only in so far as it excludes an intelligible alternative, namely ignorance or lack of knowledge. But if there is no such thing as ignorance in these kinds of case, then there is no such thing as knowledge either, since there is nothing for an ascription of knowledge to exclude. 7. Intention, Prediction and Autonomy A self-conscious agent has intentions and beliefs, and acts intentionally. He cannot non-redundantly be said to know his intentions or beliefs, since he cannot, in paradigmatic cases of intending and believing, be ignorant of them, although he may, of course, misstate them. One cannot intelligibly ask an agent “How do you know that you intend that?” or “How do you know that you believe that?”. If someone says “I intend to V”, one cannot ask “Why do you believe that you intend to V?”, as one can ask “Why do you believe that he intends to V?”, but only “Why do you intend to V?”. And that question is a request for his reasons for Ving, which are ipso facto his reasons for intending to V, not reasons for his thinking or knowing that he intends. Similarly, if someone says “I believe that p”, the question “Why do you believe that?” (unlike “Why do you believe that he believes that?”) is a question which asks for the addressee’s reasons for believing that p, not for his reasons for believing that he believes that p. Indeed, in saying “I intend …” and “I believe …”, as opposed to “I intended …” and “I believed …”, one is expressing one’s intentions and beliefs, rather than ascribing intentions and beliefs to oneself. An ascription of intention is compatible with denying that it will be executed, as an ascription of belief is compatible with denying its truth. But one cannot say “I intend to V, but I shall not”, any more than one can say “I believe that p, but not-p”.5 5

hat is why it is perhaps a little misleading to speak, as Strawson does, of selfascription of P-predicates “in the same logical breathe”, as it were, as one speaks of other-ascription (see Strawson 1959, chap. 3).


P. M. S. Hacker

Nevertheless, in intending to V, one does have special knowledge, not of one’s intention, but of one’s future action. One can assure another that one will V tomorrow, precisely because one intends to V tomorrow. That knowledge of the future is not inductive, although it presupposes inductive knowledge of possibilities of action open to one and of one’s abilities, and any such claim regarding one’s future can be rebutted on the basis of inductive knowledge. But my knowledge of my future action is based only on my having the firm intention that I have. In virtue of my resolution, I know without evidence that if I have the means, and nothing untoward happens, I shall V (FI, 60).6 I may be unsure or not know what another person intends to do, and if so, I may need to find out. But if a self-conscious agent is unsure or does not know what he intends to do, that does not mean that he intends something, but needs to find out what it is. It means that he needs to decide what to do, i.e. to form an intention. And if he decides, his reasons are not reasons for his prediction that he will V, but are reasons for him to V. His reason for the prediction that he will be there tomorrow is that he has decided to go. So “I don’t know what I intend to do”, unlike “I don’t know what he intends to do”, is not a confession of ignorance, but of indecision (FI, 68). And, as remarked above, “I know what I intend to do” is not a cognitive claim but equivalent to “I have made up my mind”. So, I am the author of my intentions, obviously so when they are formed by decision, but also, in a looser sense, when they are not. For here too, I have, or think I have, reasons for doing what I intend to do. If I intend to V, and am asked “Why?”, and can think of no reason, then my actions will be, as it were, opaque to me. I shall not understand why I am doing what I am doing. But if I further know that V-ing would be a mistake for me, I will change my mind. If I do not, but continue to insist that I shall V, then I imply that I can’t help V-ing. If I go on to V, despite the knowledge that I have weighty reasons against V-ing and none for it, then I am in the grip of a compulsion, and am not, in this respect, a free agent (FI, 70 f.). Again, Hampshire draws attention to the striking first/ third-person asymmetry: I can be certain that A will not V, if I know that it would be a mistake for him to V and I know that he will know or believe it to be so. But the agent does not need this reflexive knowledge about himself. If he comes to know or believe that it would be a mistake 6

All parenthetical references flagged “FI” are to Freedom of the Individual (Hampshire 1965).

Thought and Action


for him to V, he does not need to ‘ascertain’ that he has thus concluded (FI, 74). One who avows or reports his beliefs or intentions is not in the position of a third-person reporting another’s beliefs or intentions, i.e. of having to ensure that things are as he says they are. In his own case, he just has to be truthful.7 But, in addition, he must conceive of his beliefs and intentions as appropriate, or at any rate not as inappropriate (FI, 76 ff.) To avow that one believes that p is simultaneously to assert that p. It is to take a stand on whether it is the case that p, for one cannot say “I believe that p, but it is not the case that p”. So one cannot find oneself with a belief which one knows to be subjectively incredible, by contrast with finding oneself saddled with passions that are groundless. To recognize its subjective incredibility is to cease to hold it. If one is still haunted by such an absurd or groundless thought, one will not say “Nevertheless, I still believe that p”, but rather “I cannot help thinking (imagining) that p”. Here one is in the grip of a fantasy or delusion. Neither believing nor intending are passions. Rather, Hampshire argues in Spinoza’s wake, we are active in believing what we believe and in intending what we intend. Irrational, groundless emotions are passions, from which an agent can dissociate himself. They are not truly his—he is invaded by them. Learning their causes may enable him to control them by means of various instrumental techniques. But beliefs and intentions are not like that. I do not control my beliefs, but I cannot dissociate myself from them, as I can from my passive emotions—I am answerable for them before the tribunal of reason as I understand it. I cannot have a belief which I simultaneously hold to be misguided. Unlike phobias, I cannot view a belief as something I just happen, deplorably, to have. And this difference between beliefs and passions is exhibited in the fact that to admit to a phobia is not to endorse it, whereas to declare a belief is. I am not the author of my passions, Hampshire observes, but I am the responsible author of my intentions, and, to some degree, of my beliefs (FI, 80). Hence first-person authority with regard to my beliefs and intentions stems not from privileged, introspective ‘access’ to one’s mind, which yields, as the empiricists thought, indubitable and infallible knowledge, but stems rather from the fact that I am their author. I form some of my intentions and beliefs through deliberation and decision, and 7

As Wittgenstein remarked, in these kinds of cases, truthfulness guarantees truth. Matters are, however, more complicated when it comes to self-deception and unconscious beliefs and intentions.


P. M. S. Hacker

others, though not the upshot of deliberation, are endorsed by me in their expression. So they are fully mine, in a sense in which my passive attitudes and emotions may not be. They are, as Joseph Raz has put it, responsive to reason as I, rightly or wrongly, see it (see Raz 1999). Hampshire links his Spinozist conception of the activity and passivity of the mind with a correlative conception of human freedom. “The notion of freedom”, he writes, properly enters into the analysis of intentional action. The man who is comparatively free in his conduct of his life is active in the adoption of his own attitudes and his own way of life; his decisions and intentions are the best guide to his future actions; and just this is the significance of calling him ‘free’. He is a free man, in so far as he is the authority on his own future actions as issuing from his decisions; then his self-knowledge is predominantly of the kind that comes not from observation and induction, but from making up his mind what his attitudes and actions are to be. (FI, 112)

The conception Hampshire advances is, I think, a profound elucidation of rational agency. It seems to me to illuminate not so much the notion of human freedom as that of human autonomy and self-determination. If I am forced to live my life in a prison, I am not free, no matter how active I am in the formation and endorsement of my beliefs, intentions and active emotions. If I retain my rationality, I may be free from irrational passive attitudes and emotions that might hold me in their grip, even though I am not a free man. So, imprisoned as I am, I retain my autonomy—I am, as it were, master in my own house. Bibliography Hacker, Peter (2005): “A Tribute to Stuart Hampshire”, in: Philosophy 80, 175 – 197. Hampshire, Stuart (1959): Thought and Action, London: Chatto and Windus. — (1965): Freedom of the Individual, London: Chatto and Windus. Hyman, John (1989): The Imitation of Nature, Oxford: Blackwell. Hume (1739): A Treatise of Human Nature, ed. L.A. Selby Bigge, Oxford: Clarendon Press. Kenny, Anthony J. P. (1989): The Metaphysics of Mind, Oxford: Oxford University Press. Raz, Joseph (1999): “When We Are Ourselves: the Active and the Passive”, reprinted in: Raz, Joseph (1999): Engaging Reason, Oxford: Oxford University Press. Rundle, Bede (1997): Mind in Action, Oxford: Clarendon Press.

Thought and Action


Strawson, Peter F. (1959): Individuals. An Essay in Descriptive Metaphysics, London: Methuen.

I. Acting for Reasons

Reasons for Action and Psychological States Frederick Stoutland 1. Introduction The defining claim of the Belief-Desire Model of reasons for action (henceforth B-D model) is that rational explanations are (in the final analysis) explanations in terms of the agent’s own desires for ends and beliefs about the means to satisfy them. This can be divided into two: 1) all rational explanations are in terms of (or derive their force from) the agent’s ends and what he takes to be means to those ends—are, that is, instrumental; 2) all instrumental explanations of action are in terms of the agent’s own beliefs (about means) and desires (for ends). I shall argue that both claims should be rejected: 1) some rational explanations (in particular, explanations of responsive actions) are neither reducible to nor depend upon instrumental explanations; 2) the B-D model does not fit even instrumental explanations because it mistakenly psychologizes them in taking beliefs and desires to be the reasons for which we act. My own view is that reasons for action are not psychological states of the agent but (past, present, or future) states of affairs in the world that (from the normative point of view) favor her acting in a certain way and (from the explanatory point of view) are reasons for which she acted. I do not regard this as a metaphysical theory about reasons or psychological states, but as an elucidation of the constitutive human practice of describing and explaining action. 2. Two Points of View on Reasons for Action To consider an agent’s reasons from an explanatory point of view is to consider whether they do explain why an agent acted intentionally as he did, that is, whether they really were the reasons for which he acted. To consider an agent’s reasons from a normative point of view is to consider whether they were reasons for him to act; that is, whether they meet the relevant standards of evaluation. The latter may be normative in the


Frederick Stoutland

(strict) sense of requiring (or forbidding) a certain action, but reasons for action are not typically normative in that sense but rather considerations that favor my so acting: show that it would be good, manifest it as appropriate, and so on. Note that I treat a reason as a consideration that favors (or disfavors) an action even if it is outweighed by other reasons for (or against) the action. Reasons for action also have a normative dimension when considered from the explanatory point of view. That is not so with scientific explanation, for example: “The reason the car radiator broke was that it froze overnight and water expands when it freezes.” We use “reason” here since, as in the case of action, a reason functions to make phenomena intelligible: if the reason the radiator broke was that the water in it froze and expanded, then it is intelligible how it happened and why. But that sense of reason has no normative dimension since it would be mere confusion to ask whether the freezing temperature favored water expanding, whether it was appropriate that the radiator broke because the water in it froze, or whether it is rational of water to react that way. It is always in order, however, to ask an analogous question about the rational explanation of an intentional act: Was the reason that explained her act also a reason, normatively speaking, for her to act as she did? 3. Taking Something to be a Reason for Action Reasons for action are normative in a further sense: to explain an agent’s act in terms of the reason for which he acts presumes that he takes that reason also to be, normatively speaking, a reason for him to act. Taking a state of affairs to be a reason to act is what I shall call a practical stance, which has many forms, all of which involve evaluation—most often implicit—of a situation as a reason for one to act. If, for example, the reason for which Jim (who is an idiot about cars) pours anti-freeze into his car’s radiator is to clean the carburetor, he takes it that cleaning the carburetor is a (normative) reason for him to do that. In this case, taking has the form of belief, as it typically does when there is deliberation. But it often involves a non-reflective response to a situation as a reason for action. In stopping at the intersection, for example, I take the red light to be a reason for me to stop; but my taking is not believing the light is a reason but simply my responding to it as a reason for me to stop. Such response, unlike belief, is pre-reflective, but like belief it has a conceptual structure, and it may be in error. I stop at the intersection in

Reasons for Action and Psychological States


response to a red light, which I thus take to be a reason to stop, but I am mistaken if the light is only a warning light. What I took to favor my stopping did not favor it, and hence, although it was the reason for which I stopped, it was not a (normative) reason for me to do so. Taking is directed toward the practical significance of a state of affairs as a reason to act—toward those features that bear on the normative status of acting in a certain way. Jim took cleaning the carburetor to be a reason for him to put anti-freeze in his car, but he was wrong about that. He erred in believing that radiators are connected to carburetors and he erred in what he took to be a reason for him to put anti-freeze in his car, but these are different kinds of error. The former is theoretical, a matter of lacking scientific knowledge; the latter is practical since even if adding anti-freeze did clean carburetors, whether that was a reason for Bill’s action depends on his interests, his situation, and similar factors. The practical significance of states of affairs is manifest only to persons able to view the world from the point of view of the agents who take those states of affairs to be reasons for them to act. If there were such a thing as a ‘view from nowhere’, it would not be one from which the normative status of states of affairs is apparent. To understand what an agent takes to be a reason for her to act, we need not agree with her point of view but we must be able to understand how the world presents itself from that point of view. It is because rational explanations have this normative dimension that they are also agent-centered: to identify the state of affairs for which an agent acted, we must identify the point of view from which that state of affairs presented itself to her as such a reason. 4. Responsive Action and the B-D Model I turn now to critical discussion of the B-D model. My first objection is that its explanatory claims do not hold for responsive actions. We stop because someone is waving at us, we go to the door in response to a knock, we respond to a child crying. In such cases we encounter experientially a state of affairs1 to which we respond directly and thereby 1

I use the term “state of affairs” in an everyday sense to refer to any consideration, past, present, or future, that may or may not be the case, that can be taken to be a reason for one to act in a certain way. An end for which an agent acts can be characterized as a state of affairs that is not (yet) the case but that an agent acts in order to make the case. Other philosophers who think that normative reasons are


Frederick Stoutland

take it to be a reason for us to act in a certain way. By “encounter experientially” I mean through perception, feeling, emotion, using, reacting to, and so on—encounters that are typically pre-judgmental (though not pre-conceptual). By “respond directly”, I mean without deliberating on what to do. Even so, such action is intentional, uncoerced, usually appropriate, and done for a reason. A paradigm is Maria’s driving her car and stopping at the intersection for a red light. On my view, Maria perceived a metallic object as a stop light, and hence as a reason to stop, and because she responded directly to what she perceived, the light was the reason for which she stopped. She did not stop as a means to some end—and hence not because of her desire for an end or a belief about how to bring it about—but simply in response to a light that meant “Stop!”. Moreover, the light was a sufficient reason for her to stop, as it normally is for any driver. Defenders of the B-D model reject this account, arguing that a red light at the intersection cannot be the reason for which Maria stopped unless two things are available: an instrumental account of her action that refers to an end and a means toward that end, and a psychological account in terms of her desire for an end and her belief about the means. Maria must have, for example, the goal of conforming to legal signs, her desire being for that goal, and her belief being that stopping at the light will satisfy her desire. I do not deny that Maria’s action could be instrumental if, for instance, she has not acquired the capacity to perceive objects as stop lights and hence as reasons to stop. But this is not a typical case of a normal driver who stops just because the light is red—not in order to obey the law, avoid punishment or realize some other end. That such reasons depend on means and ends and the agent’s beliefs and desires about them is a dogma of the B-D model that should be rejected. One factor that may explain the attraction of this dogma is that intentional action always has an instrumental (or teleological) dimension: if an agent responds to a red light by stopping his car, then he intends his behavior, whatever it is (the diverse movements of his feet, hands, head, torso, etc.) to be his stopping. There is an end (an intention) in his act— some description he intends his behavior to satisfy—as there is in every intentional act. But that end is not part of some purpose the agent seeks not psychological states use “true propositions” or “facts”, which belong to the same general category as “states of affairs”. I prefer the latter, however, since facts (unlike states of affairs) must be the case and we do not respond to propositions.

Reasons for Action and Psychological States


to realize by stopping—it is his stopping—and hence it plays no role in any instrumental explanation of his action. That there is an end in an action does not entail that there is an end of that action: the instrumental dimension of intentional action does not entail its instrumental explanation. 5. Instrumental Action and the B-D Model Responsive actions are not the only kind of actions; there are also instrumental actions, which suggests that the B-D model might be saved by restricting it to actions taken as means to ends. However, the claims of the model refer not only to means and ends but to beliefs about means and desires for ends, and while it is definitive of instrumental action that means and ends are reasons for action, that is not the case for beliefs and desires. The question, therefore, is not whether actions being means to ends can be reasons for action (which they surely can be), but whether that entails that beliefs about the means and desires for the ends must also be reasons for action. Since reasons for action can be seen from either a normative or an explanatory point of view, that question is two-fold. On my view, a reason for an action, normatively speaking, never consists in an agent’s states of belief or desire. They may consist in what he believes and desires but those are not psychological states but the objects of such states (that I call “states of affairs”). The arguments for this are very persuasive and, moreover, they are widely accepted even by advocates of the B-D model, and hence I shall assume the view here without further argument. I also claim that what is, from the explanatory point of view, a reason for action does not consist in an agent’s states of belief and desire. This is, however, rejected by proponents of the B-D model, even those who agree that reasons, normatively speaking, are not psychological states. The latter hold a ‘two-category version’ of the model, which maintains, as Michael Smith puts it, that The term ‘reason’ is ambiguous, referring to true propositions in normative contexts but to psychological states in explanatory contexts. … By contrast with normative reasons, which seem to be truths, motivating reasons would seem to be psychological states that play a certain explanatory role in producing action [and hence] are of quite different categories. (Smith 1994, 95)


Frederick Stoutland

Normative reasons, writes Joseph Raz, are “facts which should affect what one should do. … Explanatory reasons are all states of mind of the person whose actions are to be explained.” (Raz 1978, 4) This view wrongly takes the distinction between an explanatory and a normative reason to be a category difference. As Bernard Williams wrote, It must be a mistake simply to separate explanatory and normative reasons. If it is true that A has a reason to v, then it must be possible that he should v for that reason; and if he does act for that reason, then the reason will be the explanation of his acting. … This is a basic connection. (Williams 1995, 39)

Two-category versions of the model reject this basic connection in holding that “reason” refers to facts or true propositions in a normative context, to psychological states in an explanatory one. This entails that a reason for S to A can never be one for which she A’s and that a reason for which she A’s can never be a reason for her to A, which amounts to a kind of reasons-for-action epiphenomenalism—the view that reasons for agents to act never are reasons for which they act. Just as an ordinary epiphenomenalist contends that although persons may feel pain, they never behave as they do because of the pain, so this view maintains that although there are (good) reasons for agents to act, they never act because of such reasons. If we assume that normative reasons are states of affairs, it follows that explanatory reasons must also be the states of affairs she takes to be reasons to act and not her psychological states. The difference between a normative and an explanatory reason is not in the kind of consideration regarded as a reason but whether it is viewed from a normative or an explanatory point of view. 6. Normative and Explanatory Reasons Are Not of Distinct Categories It may be objected that although normative reasons are states of affairs and explanatory reasons psychological states, there is a relation between the two that is close enough to be a surrogate for the identity our ordinary working conception assumes when we act for a (good) reason. This relation is the one that holds between the content of a psychological state and the state itself, the claim being that normative reasons are the contents of the psychological states that are explanatory reasons. On this view, what explains my buying stock is (in part) my believing that doing

Reasons for Action and Psychological States


so will increase my pension fund, whereas a (normative) reason to do that is the content of the believing, namely, that it will increase my pension fund. Back of this is the commonplace distinction between a psychological state and its content (what is believed or desired), which is perfectly acceptable. The distinction is conceptual, not real: it does not distinguish two components of a state, but two points of view from which we may describe a psychological state (or describe the agent as in that state). We may describe it in terms of the kind of state it is—belief, desire, hope, expectation, etc.—or in terms of its content—what is believed, desired, hoped for, expected, etc. This distinction is not, however, sufficient for a two-category version of the model, which thinks of the distinction not as between two points of view but as between two categories of reasons—states of affairs (facts, true propositions) and psychological states. If that distinction is to be identified with the content-state distinction, then the latter must also be real and not merely conceptual, which means replacing the commonplace distinction with a metaphysical one, thus adopting what Jennifer Hornsby calls a “two-task account” of psychological states (see Hornsby 1997, who rejects such an account). An extreme version of that account is held by functionalists like Jerry Fodor who regard psychological states as functional states that are defined by their causal powers, and contents as propositions that are causally impotent. On such a view, beliefs and desires have causal powers in virtue of their syntactical structure, not in virtue of their content (see Fodor 1987). As an account of reasons for action this view is hopeless, not least because it has no plausible account of why there is any systematic correlation between the causal powers of psychological states and their contents. Moreover, it does nothing to meet the objection that a twocategory version of the model makes it impossible for a reason for which an agent acts to be a reason for him to act (or vice versa) since it holds that explanatory power belongs to syntactically structured states that (as such) have no content, while normative reasons are identified with contents that are explanatorily impotent. This only exasperates the objection by holding that it is the very nature of a psychological state that what has explanatory power cannot have normative significance and that what is normatively relevant cannot explain. A less extreme version of a two-task account of psychological states is exemplified by Robert Audi’s view, which he presents as reconciling “two concepts of explanations by reasons”: the hermeneutic, which “empha-


Frederick Stoutland

sizes the way in which an explanatory appeal to reasons makes the action intelligible in the context of the explanation”, and the nomic-causalist, which posits nomic connections between psychological states and the actions they generate or sustain. Reasons of the first kind Audi calls “reasons proper”: they are the contents of beliefs or desires, that are normative reasons for action. Genuine explanatory reasons are believings or desirings, which Audi calls “reason states” and which “nomically explain (or even cause)” the action-token the agent performed. Reasons proper are abstract entities and since “abstract entities are not terms in causal relations, reasons proper are not causes”. Reason states “unlike abstract contents can be causal factors” (Audi 1994, 16). This view differs from Fodor’s in rejecting the “formality constraint” and holding that a real distinction between state and content does not rule out psychological states having causal powers (at least partly) in virtue of their content. “… Because wants and beliefs, as intentional attitudes, are individuated by their content, they would not be the attitudes they are apart from that content … . Content therefore plays at least an indirect role in giving desires whatever generative or causal power they have … .” (Audi 1994, 15) This avoids Fodor’s view that the explanatory force of psychological states does not depend on their content. Audi’s view, however, has no satisfactory account of how the content of a psychological state is supposed to determine its explanatory (causal) power. The problem arises because he holds that a psychological state causes an action-token only in virtue of a nomic connection between the state and the type of which the action is a token. That means that if a psychological state’s content determines (even in part) its causal power, then the rational connection between content and action type must determine (at least in part) the nomic connection between state and action type. On Audi’s view, the rational connection depends on the content being a (normative) reason to perform an action of a specific type, while the nomic connection depends on the causal type to which the psychological state belongs. Hence the rational connection determines the nomic connection only if the fact that a content is a reason to perform an act of a specific (rational) type determines that a psychological state with that content belongs to a causal type with a nomic connection to the action.2 2

I develop this in Stoutland 1996.

Reasons for Action and Psychological States


Audi’s view requires, therefore, the claim that if the content of a psychological state is of a certain rational type, then it is determined that the state will be of a certain causal type. It is precisely that claim that defines the covering law model of explanation: a reason for an agent to act is the reason for which she acts just in case there is a nomic generalization connecting the reason (as psychological state) with the type of action (provided the agent is rational). Hence Audi is committed to the covering law model. That model of explanation has, however, been subject to extensive criticism as it applies to explanation of intentional action (and to explanation generally). This criticism is, in my view, sufficient to show that the model, and hence Audi’s account of rational explanation, is indefensible. 7. Davidson’s Distinctive Version of the B-D Model Donald Davidson has been the most influential recent critic of the covering law model of action explanation, and careful readers of his work may object that my characterization of the B-D model does not fit his account of action. While I agree that his account is different in certain respects from the B-D model as I have formulated it, I think my fundamental objections also apply to his account. The most important difference concerns the content of the beliefs and desires that figure in a rational explanation (which Davidson calls a rationalization). Proponents of the B-D model standardly construe desires as directed to the agent’s ends, and beliefs as directed to means to those ends. Davidson, however, has a different view of the content of the belief-desire pair that he calls the “primary reason” for an action (which in a rationalization is both a normative and an explanatory reason): the desire is not directed to an end but to a type of action, and the belief is not about a means to an end but about a token action the agent believes to be of the desired type.3 Consider George’s walking to the library to attend the seminar being held there: on Davidson’s view, this is an explanation of George’s walking to the library only if George has a desire


Davidson 1980, 5: “R is a primary reason why an agent performed the action A under the description d only if R consists of a pro attitude of the agent towards actions with a certain property, and a belief of the agent that A, under the description d, has that property.”


Frederick Stoutland

to perform an action of the type ‘attending the seminar’ and a belief that this token action—his walking to the library—is of that type. The advantage of Davidson’s account is that it does not require forcing rational explanations into the instrumentalist mold. A rational explanation of Maria’s stopping at the intersection, for instance, does not require identifying a goal for her stopping and construing her action as a means to that goal. On Davidson’s account, we characterize Maria as having a desire to perform an action of the type ‘stopping for a red light’ and a belief that her action—stopping here—is of that type. Her stopping is not a means to some end: it is her doing an action of the type she desires to do. Indeed, on Davidson’s view, a primary reason is never instrumental since it explains an action not as a means to an end but as a token of an action type with a feature the agent desires (in the pro-attitude sense). George’s desire to attend the seminar, for instance, does not mean that he expects personal satisfaction or pleasure from so doing; it means that there is some feature of an action of that type to which he is attracted—its being required for his work, being beneficial, morally obligatory, or what have you. Davidson’s view is subject, however, to other difficulties that confront the B-D model because of the way it relates the normative and explanatory dimensions of reasons for action. He accepts the distinction between reasons to act and reasons for which we act, and he agrees with the model that the latter are beliefs and desires. But he makes a distinction between two senses in which reasons are normative, a distinction often expressed as between objective and subjective reasons. Objective reasons are features of our actions that objectively favor them— their being beneficial, appropriate, required by duty, and so on—and hence ones we may often be mistaken about. Subjective reasons are considerations that favor an action ‘from the agent’s point of view’, which Davidson identifies with the agent’s beliefs and desires: what favors an action from the agent’s point of view are his beliefs and desires with respect to it. This implies that when an agent acts because of a reason—when a belief-desire pair explains (‘rationalizes’ in Davidson’s sense) her action— her subjective (normative) reason to act is also that belief-desire pair, which means that the explanatory reason for which she acts is identical to the subjective (normative) reason for her to act. It would appear, therefore, that Davidson escapes the objection that the reason for which an agent acts can never be a reason for her to act: the reason for which she

Reasons for Action and Psychological States


acts always is a reason for her to act—a subjective reason, of course, but only that is relevant to a rationalization. There is, for Davidson, a sense in which agents act because of reasons that are not normative reasons for them to act, namely, when the subjective reasons are not objective. George walked to the library because he believed the seminar was being held there; if his belief was false, there was a subjective but not an objective reason for him to do that, and if his desire to go to the library was irrational, it was a subjective but not an objective reason. Objective reasons are defined in terms of subjective ones: an objective reason is a subjective reason whose constituent beliefs are true and whose constituent desires are rational. An objective (normative) reason may therefore be thought of as a state of affairs because if a belief is false it is not an objective reason, and if it is true, what is believed (a state of affairs) is the objective reason. An analogous point holds for desires: if a desire is irrational it is not an objective reason, and if it is rational, what is desired is the objective reason. 8. Objections to Davidson’s Version of the B-D Model This view is subject to two inter-related objections: its notion of a subjective reason is not coherent, and it rules out an adequate explanatory role for the notion of taking a consideration to be a reason to act. The straightforward way to construe the notion of a subjective reason is that it denotes what the agent takes to be a reason for him to act. Attending the seminar is, from George’s point of view, a reason for him to walk to the library just in case he takes that to be a reason for him to do so, and hence if there is a seminar in the library, then it is a reason for him to go. Construed in this way, a subjective reason is not a special kind of reason, indeed not a normative reason at all, but a consideration an agent takes to be a reason for him to act, which may not be such a reason. Davidson cannot, however, construe a subjective reason in this way for two reasons. First, because his notion of a rationalization requires that a subjective reason always is a normative reason to act, namely one that consists in the agent’s own beliefs and desires in the light of which her action is reasonable. This leaves no logical space for her taking a consideration to be a reason to act in a certain way since such beliefs and desires cannot fail to be reasons to act—subjective reasons to be sure, but that is all that is relevant. On Davidson’s view any role for taking, hence


Frederick Stoutland

for mistakenly taking, a consideration to be a reason to act is external to rationalization. Secondly, Davidson understands the role of beliefs and desires in rational explanation in essentially the same way as standard proponents of the model: a desire (pro-attitude) is a disposition to act in a certain way and a belief is a state that may activate that disposition. The belief-desire pair that is a primary reason explains (rationalizes) an agent’s action just in case the belief activates the disposition that is the desire to perform an act of a certain type, something that will take place whether the agent has any grasp of the normative status of the belief-desire pair as a reason for her to act as she does. On such a view, a reason explains why an agent acts without her taking it to be, from the normative point of view, a reason for her to act. This point is evident if we are clear that, for Davidson, a belief-desire pair is a reason for the action from the agent’s point of view just in case there is a certain logical relation between the contents of the pair and the action it rationalizes. Being a reason in this sense has nothing to do with the normative status of the belief-desire pair or the action, both of which may be irrational, but is solely a matter of the logical relation between them, which is essentially constitutive. Which action a belief-desire pair rationalizes is not a normative (or empirical) matter but a matter of which type of action the content of the pair constitutes as action of that type. Given the type of action the agent desires to perform and given her belief that a token action is of that type, it is logically determined that if she acts intentionally because of the belief-desire pair, she will perform the act she believes to be of the desired type. If she performs some other act, that shows either that she did not have that belief-desire pair or that she did not act intentionally because of it. It does not show that she had that belief-desire pair but was so irrational that it brought about some other action, for there is no ground for ascribing beliefs or desires of any kind to an agent who is irrational in that sense.4 Belief-desire pairs (that are primary reasons) are not reasons to act in a normative sense because they do not show that it would be good, 4

Davidson 1980, xii puts it this way: “The propositional contents of the explaining attitudes must bear a certain logical relation to the description under which the action is explained—a description that gives us an intention with which the action was performed.” He denies that the relation is deductive, which is compatible with my exposition of his view. My discussion is indebted to Gler 2003, which spells out Davidson’s view on this constitutive relation. Gler, however, thinks that what I criticize in Davidson is a virtue of his account.

Reasons for Action and Psychological States


appropriate, or required for an agent to act a certain way. Within a rationalization, therefore, there is no evaluation of a consideration as a reason for an agent to act; such evaluation takes place outside rationalization and requires determining whether the primary reasons are themselves rational. This is external to our acting and its explanation since it affects what we do only through its prior effect on the beliefs and desires that are our primary reasons. Which beliefs and desires these are may depend on our evaluation but whether they explain our action is independent of the latter, depending entirely on whether they rationalize the action (which is a constitutive matter) and whether they causally explain it (which is a matter of their causal strength), neither of which depends on our evaluating the normative status of a primary reason. 9. The B-D Model’s Mistaken View of Agent-Centered Explanations What lies behind Davidson’s claim that action explanations are rationalizations in his sense is the important insight that the kind of intelligibility only rational explanations can provide requires that they be given from the agent’s own point of view on her action and on the world in which she acts and to which she responds. They must, as I put it above, be agent-centered. According to the B-D model, a rational explanation is agent-centered in that it explains an agent’s intentional action in terms of her own beliefs and desires as reasons for her action. I regard this identification of the point of view of the agent with her states of belief and desire as a fundamental mistake because the identification is rooted in confounding an explanation given from the point of view of what the agent desires or believes about her action and world (which rational explanation should do) with explanation given in terms of that point of view (which rational explanation should not do). One way of finding out an agent’s point of view is finding out what she believes or desires, but it does not follow that we must explain her action in terms of those beliefs and desires, since what she believes and desires are not psychological states. To be agentcentered, rational explanations must be formulated in terms of the agent’s actions and world as apprehended by her, but that is not at all the same as their being formulated in terms of her apprehension. What rational explanations require is not the agent’s point of view but the world and her action as it is taken to be from that point of view.


Frederick Stoutland

10. Reasons for Action Do not Require Beliefs It might be argued that we have to live with the B-D model and its difficulties since there is no alternative. I think there is a plausible nonpsychological alternative, but I cannot develop it here and will instead respond to two objections that might be made to it. The first correctly notes that agents are often mistaken about the states of affairs they take to be reasons for them to act. Maria may stop in response to her son’s waving at her from the side of the road only to discover that it was not her son but a complete stranger. The reply characteristically made is, “Well, I believed it was my son waving and that’s why I stopped.” Defenders of the model contend that such a reply means that the reason for which she acted must include that belief. They contend, furthermore, that reasons must include beliefs in every case since, from the point of view of an agent, there is no difference between actions done in response to states of affairs that are what she takes them to be and those that are not, since an agent may not know whether things are as she takes them to be, and hence there can be no difference in how the actions are explained. Such a view embodies two errors. The first is that it takes for granted that explanatory reasons must be specifiable in true statements. That is the case for normative reasons: if it is not true that it was Maria’s son waving at her, then that was no reason for her to stop. But it is not the case for explanatory reasons. “The reason for which Maria stopped was that her son was waving” is an intensional statement: its truth value does not depend on the truth value of “her son was waving”. If it wasn’t her son waving, that shows, not that the reason for which she stopped was something else, but that she was mistaken about that as a normative reason for her to stop. She believed the person waving to be her son, but she did not respond to her believing. It is a confusion to think we must find something she was not (or could not be) mistaken about (her believing) in order to specify the real reason for which she stopped. The other error is to misconstrue the function of belief statements by thinking that “The reason Maria stopped was that she believed her son was waving at her” expresses a different explanatory claim than “The reason Maria stopped was that her son was waving at her”. In this context adding “she believed that” to a claim does not alter its explanatory content, but qualifies it so as to leave a space for the agent’s being mistaken about whether there was a (normative) reason for her to act. It might, therefore, be more perspicuous to express the claim as “The reason

Reasons for Action and Psychological States


Maria stopped was, as she believed, that her son was waving at her”, which plainly differs from “The reason Maria stopped was that her son was waving at her” only in being explicit about the possibility of Maria’s being mistaken about the reason. There is no more difference in the explanatory content of those claims than in the content of these two claims about the cause of multiple sclerosis: “It is a virus” and “We believe it is a virus.” The latter does not differ from the former in claiming that the disease is caused by a belief rather than a virus, but only in suggesting that we may be mistaken about what the cause is. So claiming that Maria stopped because she believed her son was waving does not claim her reason was a belief but specifies the respect in which she may have been mistaken in thinking there was a reason for her to stop. This error parallels one made by many accounts of perceptual illusion, for example, the bent stick in water illusion. Many accounts hold that to see a straight stick half under water as bent requires that there be something we are visually aware of which is bent—a sense impression, for instance—and since there may be, from the point of view of the perceiver, no difference between an illusory and a veridical perception, whenever we see anything we must be visually aware of a sense impression. This is to infer from the fact that we have an illusory perception of some object to the conclusion that there must be some other object of which we have veridical perception (or awareness), which I take to be fallacious in the same way as the inference from the fact that the reason for which I acted was no reason for me to act to the conclusion that there must be some other reason for which I acted (my belief ) that was a reason for me to act. 11. The Distinction Between Having A Reason and Acting Because of One The second objection is that a non-psychological account of explanatory reasons fails to meet Davidson’s claim that if explanatory reasons are not psychological states that cause actions, we have no account of the distinction between: 1) an agent who has a reason for an action but does not perform the action because of that reason and 2) an agent who performs the action because of that reason. Since the expression “has a


Frederick Stoutland

reason” is easily equivocated on,5 I will reformulate this distinction as follows: 1) S A-ed and took R to be a reason for her to A but did not A because of R and 2) S A-ed because of R. Davidson’s objection is that a non-psychological account fails to specify what must be the case when S A’s intentionally because of R that is not the case when S A’s intentionally but not because of R. Although I reject Davidson’s account, I accept the requirement that it specify a condition that obtains only when S A’s because of R. I would specify this condition as follows: If S A’s because of R, then her A-ing is (also) her intentionally R-ing or trying to, whereas if she A’s but not because of R (even if she takes R to be a reason for her to A), then her A-ing is not her intentionally R-ing or trying to.

To put the point in another way: if S A’s because of R, then an intention with which she A’s is to R, whereas if she A’s but not for R, then she does not A with an intention to R.6 Here are two examples: If the reason for which Maria stopped at the intersection was the red light, then she (also) intentionally stopped for a red light. If she did not stop because of that reason, then she did not intentionally stop for a red light; to stop for the light was not an intention with which she acted. If the reason the Captain threw the cargo overboard was to save the ship, his acting is (also) his saving the ship (or trying to). If he took saving the ship to be a reason for him to throw the cargo overboard but did not act because of that reason but in order to revenge himself on the exploitative ship owner, then he did not intentionally save the ship (or try to), but rather threw the cargo overboard with the intention of taking revenge on the owner.

These examples also illustrate what Kant meant by a maxim, which he defined as the subjective principle of an agent’s action. A maxim is a principle because it specifies the type of action the agent takes himself to be doing—the type he must instantiate for his action to be successful. A maxim is subjective because the type an action must instantiate to be successful is not binding on anyone else. Here is how the notion applies:

5 6

“S had a reason to A” may be taken to mean 1) “There was a reason for S to A”, 2) “S acted for a reason”, or 3) “S took a consideration to be a reason for her to A”. Only the last is relevant here. I learned to put the point this way from a lecture Jay Wallace gave at the University of California, Riverside in 2002. “Intention with which she acts” need not, of course, denote a prior intention.

Reasons for Action and Psychological States


The maxim of Maria’s stopping at the intersection is: “I am stopping in response to the red light.” Her action will be successful in the subjective sense of a maxim if she stops, not as a result of a blockade, a child running across the road, or the like, but as a driver responding to a traffic light. The maxim of the Captain’s action is: “I am throwing the cargo overboard in order to save the ship.” His throwing the cargo overboard is intended by him to be a saving-the-ship type of action, and it will be successful in fulfilling his maxim if he intentionally (and not by luck or accident) saves the ship.

A maxim, then, includes not only what the agent is intentionally doing in an immediate sense but also what he takes to be the reason for which he is doing it, and hence we can say that the condition necessary for S’s A-ing because of R is that the maxim of S’s action include his R-ing.7 This account accomplishes everything Davidson’s does and everything such an account should. Like his, mine is modest in simply specifying what must be the case when one acts because of R that is not the case when one merely takes R to be a reason for one to act. Most defenders of the B-D model will not be satisfied, however, for they think an adequate account of rational explanation should specify not only the success sense of acting because of a reason—what must be the case if one acts for a reason—but also the process sense—what must go on in order for one to act because of that reason, presumably the psychological states that explain why one acts because of it. The process sense of acting for a reason is not, however, in the scope of rational explanation, and hence it is not something an account of the latter should be expected to deliver. A rational explanation will not, for instance, explain why we acted because of one reason rather than another, as opposed to specifying the reason for which we did act. It can tell us what must be the case if the Captain threw the cargo overboard in order to save the ship but will not explain why he did it for that reason rather than, say, in order to take revenge on the ship owner. Davidson agrees on this point: a rational explanation, he writes, “provides no reason for saying that one suitable belief-desire pair rather than another (which may have also been present in the agent) did the causing” (Davidson 2004, 45). So on my account, a rational explanation does not explain why an agent acted because of one reason rather than another. 7

My account of a maxim is very rough. For a fuller account, see Rawls 2000, 168 f., which specifies the standard form of a Kantian maxim as follows: “I am to do X in circumstances C in order to bring about Y unless Z.” Kant, of course, held that to be morally permissible, a maxim had to be universalizable in a certain sense.


Frederick Stoutland

Proponents of the B-D model, however, generally take it to be a main virtue of their account that it does yield such an explanation, one they think explains what must go on in order for an agent to act for a reason, namely those believings and desirings that, on their account, produce her acting. Beliefs are said to guide us and desires to motivate us, the claim being that an action that does not have that causal history cannot be done because of a reason. Such a claim, however, confuses the success sense of acting for a reason with the process sense by maintaining that an adequate rational explanation specifies the psychological states that must occur in order to act because of a reason. But an agent who meets the condition I have given has acted because of a reason no matter what processes guided her or what events caused her action. As long as she does what someone who acts because of a reason does (does what someone would do whose maxim includes that reason), not by mistake, accident or luck but intentionally, then she is acting for that reason (acting on that maxim) regardless of the specific psychological states or processes in place when she acts. This does not mean there is nothing to be said about how agents succeed in acting for reasons, why they act for some considerations they take to be reasons for them to act but not for others, or why they act because of a reason on one occasion but not on others. It means that the connections between the various considerations agents take to be reasons for them to act and the reasons because of which they act differ enormously from case to case. They are sometimes simple, sometimes very complex, but always diverse, and no uniform account can possibly cover or illuminate them. It is a mistake to seek an account of the processes an agent must go through or the states he must be in that explain how or why a consideration he takes to be a reason for him to act becomes the reason for which he acts. 12. Conclusion If to accept the B-D model is simply to agree that psychological states like beliefs and desires play a significant, indeed, indispensable role in the way we understand, interpret, and explain human action, then, of course, I accept the model. But agreement at this very general level does not mean that there are no substantive differences between my account and the B-D model’s. I think there are fundamentally two.

Reasons for Action and Psychological States


The first is that the B-D model takes explanatory reasons to consist of beliefs and desires whereas I take the latter to be at most enabling conditions for reasons to be explanatory. Defenders of the model may think that distinction is ad hoc, introduced merely to save my view, but I think it is substantial and that recognition of the non-psychological character of rational explanations has wide implications for a range of issues. Unfortunately, however, I cannot pursue this matter in a paper of this length. The second is that on my view there are many actions for which the agent’s beliefs and desires do not play a role even as enabling conditions, the paradigm being responsive actions, where an agent responds directly to a state of affairs encountered experientially. Such action presumes that the agent and the state of affairs encountered belong to the same practical world, so that the agent has evaluatively significant concepts of the state of affairs to which she responds and of the action that is her response, but it does not presume that her beliefs and desires play a role in her acting. While it is an enabling condition for her response that she takes the state of affairs to be a (normative) reason for her to act, in typical responsive action her taking is not a matter of her beliefs and desires but is her direct response to a state of affairs as a reason for her acting. This means she is connected to the world in a more intimate and primordial way than through her beliefs about it or desires for it. The world presents itself directly as a domain of reasons to act—too many no doubt and often incompatible, which requires that she sort out and evaluate them—but they are there, nevertheless, no matter what she believes and desires, and she can respond to them directly. That defenders of the B-D model find this incredible is ultimately due to a certain picture of what it is to be an agent in the world, a picture whose grip on us helps account for the persistent tendency to psychologize reasons for action. On this picture, agents and world are related only externally, by world-to-mind relations of causality, on one hand, by mind-to-world beliefs and desires, on the other. Apart from the psychological attitudes of individual agents, the world is without normative or motivational significance, nothing being a reason for anything. Agents have various attitudes, which can themselves be reasons for action or perhaps constitute states of affairs as reasons, but apart from those attitudes there are no reasons for action. On this picture, the basic distinction between agents and world is between bodies and the physical environment. This is to reject John Dewey’s claim that “in every interaction that involves intelligent


Frederick Stoutland

direction, the physical environment is part of a more inclusive social or cultural environment” (Dewey 1938, 20). It is to think of us as not responding to our environment’s social-cultural meaning since, according to the picture, our environment has no such meaning except what our attitudes give it. The world is not thought of, to quote Dewey again, as a world whose meaning is “embedded in traditions, institutions, customs and the purposes and beliefs they both carry and inspire” (ibid., 43). To shake the grip of this picture would be to recognize that the reasons for which we act exist in the world—in everyday objects, tools, roads, buildings, public and private spaces, social organizations, and so on—and present themselves to us as normatively significant and explanatorily relevant even when our beliefs and desires play no role in our acting. Bibliography Audi, Robert (1994): Action, Intention, and Reason, Ithaca: Cornell University Press. Dancy, Jonathan (2000): Practical Reality, Oxford: Oxford University Press. Davidson, Donald (1980): Essays on Actions and Events, Oxford: Oxford University Press. — (2004): Problems of Rationality, Oxford: Oxford University Press. Dewey, John (1938): Logic: The Theory of Inquiry, sNew York: Henry Holt and Company. Fodor, Jerry (1987): Psychosemantics, Cambridge: MIT Press. Hornsby, Jennifer (1997): Simple Mindedness, Cambridge: Harvard University Press. Rawls, John (2000): Lectures on the History of Moral Philosophy, Cambridge: Harvard University Press. Raz, Joseph (1978) (ed.): Practical Reason, Oxford: Oxford University Press. Gler, Kathrin (2003), “Is There Such a Thing As Weakness of Will?”, in: Segerberg, Krister/Sliwinski, Rysiek (2003) (eds.): A Philosophical SmorACHTUNGREgasbord, Uppsala: Uppsala Philosophical Studies 52. Smith, Michael (1994): The Moral Problem, Oxford: Blackwell Publisher. Stoutland, Frederick (1996): “Review of Robert Audi, Action, Intention, and Reason”, in: Philosophical Quarterly 46, 538 – 541. Williams, Bernard (1995): Making Sense of Humanity, Cambridge: Cambridge University Press.

Reasons for Action without Metaphysics? Thomas Schmidt 1. Stoutland on Reasons for Action Suppose that Maria, while approaching an intersection in her car, recognizes that the traffic lights are red and stops. A perfectly acceptable explanation of her so acting seems to be this: the traffic lights were red, and this is why she stopped. In so explaining her action, we mention the reason for which Maria acted, which was that the traffic lights were red. Saying that this was Maria’s reason for stopping appears to amount to saying that Maria took the fact that the traffic lights were red to favour her stopping. Given ordinary circumstances, it is plausible to say that Maria was right in this since the fact that the traffic lights were red did indeed favour her stopping. The reason for which she acted was that the traffic lights were red, and this reason was a good reason. In his paper, Frederick Stoutland claims that the belief-desire theory of reasons for action has difficulties in accounting for actions of this kind, which he calls “responsive actions” (77 – 79). Our “constitutive human practice of describing and explaining action” (75) speaks against the view that reasons are psychological states.1 When explaining actions in situations of the above kind, we refer to states of affairs as reasons, and a philosophical account of reasons should not explain this away by maintaining that reasons actually are something internal to the agent. What is more, the belief-desire theory, says Stoutland, faces specific problems in making sense of the normative dimension of reasons for action.2 As an alternative to the belief-desire theory, Stoutland proposes an account of reasons for action the essential elements of which are as follows: reasons for action are “(past, present, or future) states of affairs in 1 2

Parenthetical references in the text are to Stoutland 2007 (in this volume). Stoutland’s conception of reasons for action is very similar to the one proposed in Dancy 2000. In chap. 2 – 3 of this book, Dancy provides a detailed argument to the effect that the belief-desire theory has difficulties in accounting for the normativity of reasons for action. For the view that the belief-desire theory is at odds with our ordinary practice of giving reason explanations see also Bittner 2001, 4 – 5. Bittner describes, in a similar way as does Stoutland, acting for reasons as responding to the way things are (see Bittner 2001, chap. 4).


Thomas Schmidt on Frederick Stoutland

the world” (75) which the agent takes to favour the respective action. In explaining an agent’s actions, we need to find out about those states which she took to favour her acting in the way she did. In so doing, “we need not agree with her point of view but we must be able to understand how the world presents itself from that point of view.” (77) This is what we are doing when we consider reasons “from an explanatory point of view” (75, Stoutland’s italics). Since it is possible that an agent takes something to be a reason for acting in a certain way which actually does not favour her acting in that way at all, explanatory reasons are not necessarily normative reasons. When enquiring whether the reason for which an agent acted did indeed favour her so acting, we consider her reason “from a normative point of view” (75, Stoutland’s italics), which involves finding out whether the agent’s reasons “meet the relevant standards of evaluation” (75). Note that, according to Stoutland, reasons do not come in two sorts, explanatory and normative. Rather, reasons of one and the same type can be considered from two perspectives, from an explanatory and a normative point of view. This account, says Stoutland, is not to be regarded “as a metaphysical theory about reasons or psychological states, but as an elucidation of the constitutive human practice of describing and explaining action.” (75) Much of the initial plausibility of Stoutland’s proposal derives from its way of dealing with situations in which a person takes something which is indeed the case to be a reason for her acting in a certain way. How about situations in which agents get things wrong? Given Stoutland’s proposal, agents can, with regard to reasons for action, get things wrong in two different ways. For one thing, an agent can correctly believe that something is the case and wrongly take what is the case to be a reason. In situations of this sort, the agent has, to put the matter in Stoutland’s terms, a reason considered from the explanatory point of view; this reason, however, is not a reason from the normative point of view. For another thing, agents sometimes believe something to be the case which actually is not the case and take what they (wrongly) believe to be the case to be a reason. Stoutland discusses the following example: “Maria may stop in response to her son’s waving at her from the side of the road only to discover that it was not her son but a complete stranger.” (88) He maintains that the statement

Reasons for Action without Metaphysics?


(1) The reason for which Maria stopped was that her son was waving can serve as an explanation of Maria’s action even though (2) Maria’s son was waving is false. Consequently, Stoutland’s account is committed to the possibility of non-factive explanations, i. e. to explanations invoking a false explanans. 2. Alternative Views For many, including myself, it is not easy to accept that an explanation of something with reference to some statement ‘p’ can be successful while ‘p’ is false.3 The commitment to non-factive explanations does not, however, render Stoutland’s conception outright unacceptable. Yet, given the uneasiness just articulated it seems clear that Stoutland’s theory is not to be had without philosophical costs. Whether or not the costs are too high depends on how alternative options fare. Here is, then, a brief glance on three alternatives to Stoutland’s account which seem available once one wants to say that not all reasons for action are psychological states. One option to go for is Michael Smith’s proposal which has it that explanatory (or, in Smith’s terminology, motivating) reasons and normative reasons are of a different kind: explanatory reasons are psychological states, while normative reasons are not. This proposal faces the problem that it seems to be the reason for which an agent has acted and which we refer to when explaining the agent’s action, of which we ask whether it favours the action. Reasons, the thought is, are those sorts of things which reasonable agents act for since they take those very reasons to favour the action and of which, consequently, it is both meaningful and sensible to ask whether they do indeed favour the action in question.4 Doubts can also be articulated with regard to the ‘disjunctive view’ saying that reasons are facts if they obtain and beliefs if what the person 3


Jay Wallace is one of the many who find it “strange and misleading” to explain actions by saying things such as: “His reason for doing it is that it would increase his pension, though as it happened it failed to do so” (both quotes from Wallace 2003, 430). For Smith’s distinction between motivating and normative reasons see Smith 1994, chap. 4 – 5, esp. 94 – 98. Stoutland discusses and rejects Smith’s account on


Thomas Schmidt on Frederick Stoutland

takes to be a reason is not the case. To be sure, this account fits in nicely with the fact that in cases in which the agent gets the situation wrong, we are often inclined to say that the reason the agent acted in the way she did was that she believed that p, while in cases in which p is true we would say that she acted because of p. The account under discussion seems, however, to fail to satisfy the apparently plausible requirement that the form of action explanations in terms of reasons should be invariant with regard to whether what the agent takes to be the case actually is the case or not.5 A third option to go for if one sets out to reject the identification of reasons for action with psychological states has been worked out by Rdiger Bittner. According to Bittner, actions for reasons are responses to the way things are, and, consequently, reasons are states of affairs which actually obtain. This implies that an agent who takes something both to be the case and to favour an action of hers which she consequently performs cannot be said to act for a reason, if things are not the way she takes them to be (see Bittner 2001, esp. 115 f.). Although this account avoids the problems pointed out above, one may find it undesirable to be forced to say that agents in situations of the type just described have no reason at all for acting in the way they do. The upshot of this very brief survey can be put in the following way: if one wants do deny that reasons for action are psychological states and if one wants to have a uniform notion of reasons for action which does not allow for two fundamentally different kinds of reasons—explanatory and normative ones—, and if action explanations in terms of reasons are to be invariant with regard to the truth or falsity of what the agent believes to be the case, then one seems to be forced to accept either Bittner’s view or a


p. 79 f. of his paper. The view that explanatory and normative reasons should not be assumed to be of two different kinds has already been expressed in Williams 1989, 38 – 39. This is, to be sure, only a sketch of an argument which would need to be worked out in greater detail in order to be conclusive. Especially, the plausibility of the requirement that the truth value of what the agent takes to be the case should not affect the form of the relevant explanation of her action (for this constraint on action explanations see Williams 1980, 102) needs to be assessed, since the argument against the disjunctive view just outlined crucially depends on the relatively strong claim that explanatory and normative reasons are identical (for a more detailed discussion of the disjunctive view see Garrard/McNaughton 1998, 55 – 57; Dancy 2000, 138 – 145).—Stoutland himself held the disjunctive view in an earlier paper (see Stoutland 1998, esp. 60 – 61).

Reasons for Action without Metaphysics?


position of the sort Stoutland opts for—which is committed to the possibility of non-factive explanations. The rough estimation of philosophical costs and benefits of alternatives to Stoutland’s conception just given was meant both to locate this conception within the space of theoretical options which seem available and, more importantly, to give credit to the idea of non-factive explanations since alternative accounts seem also not to be without cost. 3. A Diagnosis Let us, then, come back to the issue of non-factive explanations. If Stoutland is right, a statement such as (1) The reason for which Maria stopped was that her son was waving could figure in an explanation of Maria’s waving even though (2) Maria’s son was waving is false. The uneasiness about subscribing to both (1) and the negation of (2) can, it seems to me, most appropriately be put in a question: if this were so—what are reasons, generally speaking? This question does not ask for Maria’s reason but for the type of things reasons are. Reasons are, according to Stoutland, not supposed to be beliefs or other psychological states, and they cannot be facts either. This is why Stoutland has it that reasons are states of affairs. The term “states of affairs” here is meant “to refer to any consideration past, present, or future, that may or may not be the case [….].” (77, footnote) He prefers this term to “facts” since “facts (unlike states of affairs) must be the case” (78, footnote). Since Stoutland explicitly introduces the term “states of affairs” as referring to the general category of things to which reasons for action belong, we are invited to think of states of affairs as philosophically accomplishing something. It is, however, at least in the present context, difficult to see what this can be. If the notion of states of affairs is to do philosophical work with regard to explaining actions in terms of reasons, states of affairs (including non-obtaining ones!) need to be placed within an explanatory framework such that non-obtaining states of affairs stand in explanatory relations to actions. There is, it seems to me, reason to be


Thomas Schmidt on Frederick Stoutland

sceptical about whether the details of the underlying metaphysical picture can be satisfactorily filled in. It is not surprising that Stoutland has not provided an elaborated metaphysical theory on how reference to non-obtaining states of affairs can do explanatory work, since, as I have indicated above, metaphysics is not what he means himself to be doing in the first place. He seems, however, to be doing it nevertheless. 4. Perspectives Can the idea of non-factive reason explanations get off the ground without a commitment to a metaphysical project which is relatively unappealing? The only way of doing so, it seems to me, is to resist the temptation to give a theory about the type of things reasons are and, consequently, to avoid even asking the general question what reasons are. One should, in other words, not interpret (1) The reason for which Maria stopped was that her son was waving to imply an existence statement about reasons of which one needs to make metaphysical sense (much in the same way as one should not interpret a synonymy statement as implying the existence of meanings which need to be given a proper metaphysical place). What one would have to do instead of metaphysics is to embark on the project of arguing that the conjunction of statements such as (1) and the negation of (2) Maria’s son was waving is semantically respectable. Jonathan Dancy, in his defence of a theory of reasons which is very similar to Stoutland’s account, is well aware of the problem of non-factive explanations. What he does in order to argue for their acceptability can be understood along the lines of the philosophical task just envisaged. Dancy asks us to look for contexts which have a “factive pull” (Dancy’s phrase)— much in the same way as do action explanations in terms of reasons— which they, however, on closer inspection don’t have. Here is an example which is not Dancy’s but structurally analogous to one discussed by him (see Dancy 2000, 136). It is acceptable to hold that

Reasons for Action without Metaphysics?


(3) The soccer player was sent to the bench for having committed a foul is perfectly respectable even when the player sent to the bench has not committed a foul. The truth of (3) is, it seems, invariant with regard to the truth or falsity of (4) The player has committed a foul. To be sure, the conjunction of (3) and non-(4) sounds odd, but this oddity might not be due to the fact that in the face of the falsity of (4), (3) also is false. Rather, in the face of (4)’s falsity, the oddity of (3) could be explained by pointing out that (5) It was inappropriate to send the player to the bench. The analogy between this case and the case of non-factive reason explanations is obvious. It might be the case that reason explanations function in much the same way and are respectable for the very same reasons as is (3) in the face of the falsity of (4). This is, at any rate, a line along which one needs to argue in order to make the Dancy/Stoutland view plausible. The theoretical option just pointed out involves an attempt to defend a view about reasons for action without attempting to find an appropriate metaphysical category in which reasons can be placed.6 What is required instead is, as Stoutland is right to point out, “an elucidation of the constitutive human practice of describing and explaining action.” (75) I take the brief assessment of philosophical costs and benefits of different theoretical options given in this commentary to lend support to the claim that it still is an open question whether or not this very practice allows for a uniform conception of reasons which does justice to the different constraints one would wish such a conception to meet.7



The latter seems to be a fair description of what Dancy is aiming at. In his “Prcis of Practical Reality”, he states that the arguments put forward in chap. 5 of his Practical Reality (2000) “concern the need for normative and motivating reasons to be creatures of broadly the same metaphysical type” (Dancy 2003, 425). For helpful discussions on topics dealt with in this commentary I am indebted to Holger Baumann, Jonathan Dancy and Jan Gertken.


Thomas Schmidt on Frederick Stoutland

Bibliography Bittner, Rdiger (2001): Doing Things for Reasons, Oxford: Oxford University Press. Dancy, Jonathan (2000): Practical Reality, Oxford: Oxford University Press. Dancy, Jonathan (2003): “Prcis of Practical Reality”, in: Philosophy and Phenomenological Research 67 (2), 423 – 428. Garrard, Eve/McNaughton, David (1998): “Mapping Moral Motivation”, in: Ethical Theory and Moral Practice 1, 45 – 59. Smith, Michael (1994): The Moral Problem, Oxford/Cambridge (Mass.): Blackwell. Stoutland, Frederick (1998): “The Real Reasons”, in: Bransen, Jan/Cuypers, Stefaan E. (eds.): Human Action, Deliberation and Causation, Dordrecht: Kluwer, 43 – 66. Stoutland, Frederick (2007): “Reasons for Action and Psychological States” (this volume). Wallace, R. Jay (2003): “Explanation, Deliberation, and Reasons”, in: Philosophy and Phenomenological Research 67 (2), 429 – 435. Williams, Bernard (1980): “Internal and External Reasons”, in: Williams, Bernard (1980): Moral Luck, Cambridge: Cambridge University Press 1981, 101 – 113. Williams, Bernard (1989): “Internal Reasons and the Obscurity of Blame”, reprinted in: Williams, Bernard (1995): Making Sense of Humanity, Cambridge: Cambridge University Press, 35 – 45.

The Causalist/Anti-causalist Debate in the Theory of Action: What It Is and Why It Matters Maria Alvarez 1. Introduction One of the tasks with which philosophers have been occupied in the past fifty years has been that of giving an account of intentional actions, in particular of actions that are performed intentionally and for a reason. Anscombe’s 1957 book Intention was one of the first sustained attempts to provide such an account. The account she offers in that book, much influenced by Wittgenstein in content and, especially, in approach, has the character of what Strawson called a “connective analysis” of a concept (Strawson 1992, 19). This makes it difficult to summarise it in a few lines. Anscombe takes the reader on a—not always easy to follow—tour of the conceptual territory, drawing our attention to the connections and differences between some of the central concepts such as intention, prediction, voluntary and involuntary action, wanting, ends and goals, practical knowledge and practical reasoning, etc. One strain in Anscombe’s account, inspired by Wittgenstein and running contrary to much of the philosophical tradition, is the view that we ought to distinguish between mental causes and reasons, and that, in general, the concept of causation is unlikely to throw much light on our understanding of intentional action and of the connection between the reasons we act for and our actions. It was rather ironic, then, that despite Davidson’s avowed admiration for Anscombe’s book,1 the headline news of his 1963 paper “Actions, Reasons, and Causes” was, precisely, that the reasons for which we act are the causes of our actions; that reason explanations are a species of causal explanation; and that this was not only Aristotle’s but is also the common-sense, pre-philosophical view. That paper proved to be extremely influential, so much so, that causalism 1

Which he described as “the most important treatment of action since Aristotle” (frontcover of the 2000 Harvard reprint of Intention).


Maria Alvarez

about reasons soon became the orthodoxy and, despite the efforts of a resolute minority, it has remained so for the past forty years.2 Partly because causalism has been such a dominant orthodoxy for so long, and partly because the label covers a variety of positions, it is perhaps not clear what exactly the nature and significance of the disagreement between causalists and anti-causalists is. In this paper I shall try to make the disagreement more explicit so as to clarify what its significance might be. 2. Causalism in the Theory of Action In order to help bring out the nature of the disagreement, it might be useful to eliminate some candidates. The disagreement is not, for example, about the claim that “mind makes a difference” (Heil/Mele 1993, iii), if this means simply that what people believe and want makes a difference to the way they act, and hence to what happens in the world around them. Nor is the dispute about the idea that there is something that could be called ‘mental causation’, if this means simply that a psychological occurrence, that is, an episode in someone’s mental life, might be said to cause some other episode (psychological or not) in that person’s life; in the way, for instance, in which seeing a mouse in the kitchen, that is, perceiving a mouse, might cause one to scream or to jump onto a chair; or in which suddenly remembering that you must call the butcher might be said to cause you to do so right now. Nor does the disagreement concern the fact that all sorts of psychological elements such as motives, desires, thoughts, beliefs, etc. are often said to cause people to act. Anti-causalists do not deny the reality of this usage, nor do they claim that the relevant expressions enshrine a philosophical mistake. The mistake, they think, is to construe these expressions as causalists typically do, and to derive from them a certain philosophical picture about what kind of thing reasons are and how they are connected to actions. At a first approximation, a causalist is someone who claims that any elucidation of the relation between an action done for a reason and the reason for which it is done must invoke ineliminably the notion of 2

These are some recent publications from among that minority: Bittner 2001; Hutto 1999; Schueler 2003; Sehon 1994; 1997; 2000; Stoutland 1998; 2001; Wilson 1989.

The Causalist/Anti-causalist Debate in the Theory of Action


causation. But even this is too vague, as it is consistent with at least two claims. The stronger claim (which I shall call ‘the causal claim’) is that the reason for which someone performs an action is the cause of the action. I take this stronger claim to involve the view that reasons are mental things of a kind (events, states, combinations of these) that makes them suitable candidates to be the relata of causal relations, and hence to be called the causes of actions. The weaker claim (‘the causal-explanation claim’) is that, even though the reasons for which we act may not be the causes of our actions, an explanation of an action whose explanans is the reason for which the agent acted is a causal explanation.3 Thus, anti-causalists object to both the idea that reasons are mental entities that cause actions (sections 3 – 6, below); and to the idea that reason explanations are causal explanations (sections 7 – 9, below). 3. Are Motivating Reasons Mental States? The causal claim goes hand in hand with a widely accepted conception of reasons to which Davidson gave particularly sharp articulation. According to Davidson, a ‘primary reason’ for an action, i.e. a reason that shows immediately that the action was rational from the agent’s perspective, is a 3

See Hornsby 1997, 129 – 153. Although Davidson can be credited with having established causalism as the dominant view, it is hard to decide whether he held the stronger or only the weaker claim. On the one hand, he unequivocally states that a “primary reason for an action is its cause” (Davidson 1980, 12) and his replies to several of the objections he canvasses in that paper (e.g. the logical connection argument and the objection based on ‘the nomological character of causation’) depend on taking reasons to be particulars that can be variously described. Both of these suggest that he endorsed the stronger claim. On the other hand, he says that “primary reasons consist of attitudes and beliefs, which are states or dispositions, not events” (ibid.). And since Davidson thought that causation is a relation that obtains only between events, it seems that he did not think that reasons, which he took to be combinations of mental states, could be, strictly speaking, causes. Perhaps he thought that reasons constituted antecedent causal conditions of actions and that they could, on account of this, be called ‘causes’ of those actions (see Davidson 1980, 12)—which would suggest that he held the weaker claim. But my concern here is not so much to decide which is the right construal of Davidson’s views as to examine criticisms of the two causalist positions outlined above. And it is worth noting that other causalists see no objection to thinking of states and of reasons as, strictly speaking, causes. See Evnine 1991, 30.


Maria Alvarez

pair of mental states of the agent’s, typically a state of belief, and a state of desire or “pro-attitude”, to use his terminology (see Davidson 1980, 4). This conception of reasons has special appeal for defenders of the causal claim because it fits in nicely with a picture of how reasons cause, and hence explain, actions. According to this picture, reasons are mental states that are identical to, or somehow ‘realized’ in, states of the brain. Thus, to say that reasons cause actions is, ultimately, to say that brain states cause actions. This conception of reasons, then, seems to offer the most promising framework for what is sometimes called ‘the project of naturalising intentional action’, that is, of showing that explaining actions by reference to reasons does not require us to postulate entities that cannot be described in the vocabulary of sciences such as physics, neurophysiology, etc., but which have the power to cause changes on material entities; for, on this picture, talk of reasons as causes of actions is, ultimately, talk of brain states, events and processes as causes of actions. In the past few years, however, several philosophers have challenged the view that the reasons for which we act are our mental states.4 These philosophers have provided a variety of arguments against it, and have also endorsed somewhat different conceptions of what our reasons for acting are.5 I shall outline the objection emphasising its negative aspect, viz. that reasons are not mental states. I cannot claim that my articulation of the objection captures any common thread running through all of those arguments, though I hope it captures something of the spirit that inspires at least many of them.



The challenge I have in mind here is not that reasons are not mental states because, even if the reasons for which we act are that we believe and want certain things, our believing and wanting something is not rightly characterised as a mental state. (For various arguments to this effect see: Collins 1987; Hacker 1990, 28 ff.; Steward 1997). Rather, I have in mind the view that our reasons are not our believing and wanting things. See, e.g, Bittner 2001; Dancy 2000; Stout 1996; Stoutland 1998; 2001; Schueler 2003; Alvarez 2008 (forthcoming).

The Causalist/Anti-causalist Debate in the Theory of Action


4. An Alternative Conception of Motivating Reasons The view that the reasons for which I act are mental states is the corollary of the view that the reasons for which I act are my believing and wanting certain things.6 But, the objection goes, this is simply wrong: the reason for which I act is typically something that appears to me to make my action right or appropriate, and that is normally what I believe and not my believing it (for reasons that shall become clear, I shall concentrate on beliefs for the moment). The reasons for which we act are not states of believing; rather, as we might say, they are what those states are about, whether we conceive of the latter as particular things, or their features, as states of affairs, as facts about the world, etc. Accordingly, reasons are things such as that it’s late, that she’s in danger, that this will increase my profits, that I made a promise, that it’s very beautiful, that it’d be selfish, etc. And such things are clearly not mental states. Now, although the objection seems prima facie quite decisive in establishing the view that my reason is not my believing something but rather what I believe, there are considerations that may appear to show that the objection fails and that my reason must, after all, be my believing something, and hence a mental state. The first consideration is that, in order for what is believed, ‘that p’, to be someone’s reason for acting, the agent must believe (or know or suspect) that p, for unless he does, that p cannot be his reason: if I don’t believe that she has left, that she has left cannot be my reason for doing anything. This is true but, in fact, the very way the objection is articulated undermines the idea that reasons must be mental states. For the objection says that ‘that p’ cannot be a reason unless the agent believes that p. And this suggests that if the agent does believe it, then his reason is that p, what he believes, and not his believing.7 Consider an analogy. Unless I have money, I cannot pay for anything with it. But this does not mean that if I do have money, when I pay for something, I pay with ‘my having money’, rather than with the money that I have. The second, more powerful, consideration concerns cases where what an agent believes is false, i.e. it is something that is not the case. The 6


The expressions “that he believed that p” and “his believing that p” are treated as interchangeable ways of designating the mental state that believing something is said to be. Although these expressions are importantly different, I shall ignore the difference here and follow that practice. See Dancy 2000, 127 ff. for further discussion of this point.


Maria Alvarez

objection goes as follows. If my reason for acting is what I believe, and yet I believe something that is not the case, then my reason cannot be what I believe, because I believe something that is not the case, and what is not the case cannot be a reason (though it can appear to be one). Thus, the argument goes, in such a case my reason must be my believing what I believed, which is to say, my reason must be a mental state. It is tempting to go on to argue that if this is right, and in error cases my reason is my believing it, then that must be my reason also when I am right; and to conclude that the reasons for which we act are always mental states. However, the reasoning that since error cases suggest that reasons must be mental states, then reasons must be mental states also in veridical cases is a form of the ‘argument from error’ and, as has often been pointed out, we have good reason to reject that argument. But even so, the problem does appear to suggest that, at least when people act on the grounds of false beliefs, i.e. of what is not the case, their reason must be their believing those things. And this appears to suggest that, at least in error cases, reasons must be mental states. But does it? I think that there is an alternative way of characterising error cases that does not lead to that conclusion. The minority view now under attack is that the reason for which I act is what I believe, and this seems to lead to problems in error cases. But perhaps the minority view needs to be qualified: what I believe is my reason for acting when I believe what is the case. What I believe, however, is not a reason when it is not the case, even if I acted in the light of that false belief. In other words, if I act in the light of something I believe which is the case, then I act for a reason (though that reason may fail to justify my action). If, on the other hand, I act in the light of something that is not the case I do not act for a reason—though I took what I believed to be a reason to act, since I took it to be the case. We might say that in such cases, I acted for an apparent reason. But an apparent reason is not an actual reason, though it can be mistaken for one, just as a fake Vermeer is not a Vermeer, though it can be mistaken for one. Thus, if I act in the light of a false belief, the reason for which I acted was not that I believed what I believed for, in such a case, I did not act for a real reason. (It is important to note that this does not mean that in a case where I do not act for a real reason, I act for no purpose. If, for instance, I take my umbrella because, as I falsely believe, it is raining, then I did not act for a—real—reason but I acted for a purpose, say, to keep dry.)

The Causalist/Anti-causalist Debate in the Theory of Action


Admittedly, since the expression “Her reason was that p” implies that p, this form of words cannot be used, without an air of paradox, by those who know or believe that p is false.8 And thus, in such cases, we must resort to expressions such as “She did it because/Her reason was that/she believed that p”. This has appeared to many to suggest that in these cases reasons are mental states. However, the fact is that these forms of words do not imply that the reason in the light of which she acted was that she believed that p. Rather, they convey the idea that her acting as she did can be explained by her believing that p. But it does not follow from this that her believing that p was the reason that motivated her; what follows is that her believing that p can explain her action. The reason why we do something is a reason that explains—makes it intelligible—why we did something. But not every reason that explains is a motivating reason. The reason why Tom sits at the back of the class is that he is shy but that is not the reason that motivates him to sit at the back of the class—his reason is probably that the teacher is less likely to see him and ask him questions if he sits there. Hence, although in error cases the agent acted for no reason, this does not imply that there was no reason why the agent acted, that is, that there is no reason that explains her action; there was, namely that she believed what she believed.9 Thus, the minority view can account for error cases in a satisfactory way without having to concede that in such cases the agent’s reason was that she believed that p and hence that it was a mental state. 5. Desires and Motivating Reasons So far, I have examined the argument about the reasons for which we act exclusively in terms of beliefs. However, the widely prevalent DavidACHTUNGREsoniACHTUNGRE an conception of reasons says, as I mentioned a moment ago, that reasons are constituted by states of belief and of desire. And it might seem possible to accept that the belief part of a reason is not a mental state but maintain that the desire part of a reason is a mental state. As a matter of fact, many who reject the view that reasons are mental states also reject the idea that reasons must include desires. The thought 8 9

An air of paradox which is analogous to the Moorean paradox: “I believe that p but not p.” I discuss these issues further in Alvarez 2008 (forthcoming).


Maria Alvarez

that they must finds support in a view, attributed to Hume, that beliefs (even understood as “what is believed”) alone cannot motivate. But these philosophers plausibly hold that what is believed alone can motivate: that I am in need can motivate you to help me, so long as it is something you believe. If that is right, then consideration of desires does not further the view that reasons are mental states. The Humean response to this claim, however, is that, unless you are in a state of desire, you will not act. So, they conclude, states of wanting are at least part of the reasons for which we act. But even if it is true that unless you are in a state of desire, you will not act, it does not follow that states of wanting are part of the reasons for which we act, that is, that my wanting to do A must be part of my reason for doing A. If acting requires an agent to be in a state of desire, all that follows from this is that his being in a state of desire will contribute to explain why he acted, not that it was part of his reason for acting. But, it might be objected, it may be true that in some cases what an agent believes may be all that motivates him to act. But this cannot be true in cases where we do something simply because we want to, because in such cases the fact that one wants to do A seems the only candidate for the reason for which the agent does A. Consider this example (adapted from Mele 2003, 82). Suppose I’m feeling cheerful and feel (as it were, welling up in me) a desire to sing, and consequently break into song. I don’t sing in order to impress anyone, nor to exercise my voice, nor yet to show how happy I am, or because I love singing. I simply sing because I have a desire to do it: I feel the urge to do it. Then surely, in this kind of case, as Mele says, my reason for singing just is my desire to sing. He says the following to support his claim: (S)ingings motivated by such desires are intentional actions; it is standardly held that intentional actions are done for reasons; this wholly intrinsic desire is the best available candidate for a relevant reason; the desire plays a central role in explaining the action (Mele 2003, 82).

Whether we accept the claim that this singing was done for a reason depends on whether we accept both that this singing was intentional, and that all intentional actions are done for reasons. Neither of those claims, however, seems indisputable. First, it is plausible to claim that this kind of singing is not intentional. This is not as odd as it might seem, because to say that the singing was not intentional does not imply that it was unintentional, or accidental, or coerced. We might say that it was not intentional but voluntary, in the way in which many animal actions are

The Causalist/Anti-causalist Debate in the Theory of Action


surely voluntary (those animals do what they do because they want to) but not intentional (if intentional implies ‘done for a reason’, since animals do not act for reasons—at least most don’t). And this seems the right thing to say of things we do somewhat distractedly, such as rapping on a surface or doodling. Singing and whistling may also be like that. These actions are on what we could call the periphery of the intentional, but they are not done for a reason. This is all true but it is also true that sometimes we do sing for a reason. And the question here is whether the singing described in the example above is done for a reason. I don’t think it is and consequently I do not think that my reason was that I wanted to sing. In general, if my reason for ¦-ing was that I wanted to x, then it follows that I ¦-ed in order to x. So, if my reason for singing had been that I wanted to sing, then it should make sense to say that I sang in order to sing (compare: “My reason for driving was that I wanted to get there in time”—“I drove in order to get there in time”). But I don’t think that that makes much sense; or if it does, it is surely only as a roundabout way of saying that I sang for no reason, although I sang because I wanted to sing. I think that behind the view that in the example above I sang for a reason might be the thought that, since my singing can be explained by my wanting to sing, then my wanting to sing was my reason for singing. But that doesn’t follow. What follows is that my wanting to sing was the reason why I sang. But as we saw above, not every reason that explains why someone did something is a reason for which they did that thing. Thus, although when we explain why someone acted their wanting certain things may be mentioned in an explanation of why they did it, it does not follow that their wanting that thing was part of their reason for acting. I must emphasize that I am not arguing that people never do things just because they want to—the example above is such a case—my point is that when they do, and there is no more to say about their reasons for wanting to do that, their action may not be an action done for a reason, even though there is a reason why they did it.10 10 I have said that the fact that you want something is not itself normally a reason for you to act, because it can be, though that is rare (the same is true of my believing something). For example, someone might reason that the fact that they want x (e.g. want to talk to their boyfriend) very badly is a reason to satisfy that want because if they don’t, they won’t be able to concentrate on other things, or whatever. But even in these cases we might say that, properly speaking, the agent’s reason is some feature of the action, say, that calling their boyfriend will enable them to get on with important work. (Of course, it’s much more plausible that


Maria Alvarez

So it seems that there is no compelling argument to accept the Humean view that states of desire must be part of motivating reasons; the Humean arguments at most show that explanations of action must include references to the agent’s wants. 6. Motivating Reasons and the Causal Claim If the above is right, then it seems that motivating reasons, the reasons for which we act, are not mental states of agents’; they are, rather, what those mental states are about, or their contents. And these might be things such as propositions, or facts, or states of affairs, or things in the world, or features and aspects of things, etc. This conception of reasons, if right, seems to have important consequences for the causal claim. As we saw above, the picture that defenders of the causal claim tend to endorse presents motivating reasons as mental states that are identical to, or somehow ‘realized’ in, states of the brain. According to this picture, to say that the reasons for which we act cause actions is, ultimately, to say that brain states cause actions. However, if motivating reasons are not mental states, but are the sorts of things mentioned in the previous paragraph, then two things seem to follow. One is that it seems much less plausible to say that motivating reasons are causes of actions.11 The other is that the prospects of ‘naturalising’ intentional action in the manner described in section 3 seem much less bright. For, if motivating reasons are not mental states, then they are not states that can be identified, or said to be realised in, the brain states that cause motions of the body. This does not mean that someone who is interested in human action should not be interested in brain events, states and processes and their relation to motions of the body. However, what this suggests is that interest in the neuro-physiological knowledge to be discovered concerning the causation of the bodily movements that typically happen when someone acts, is not interest in, or knowledge of how reasons motivate the reason why one actually calls a boyfriend is that it’ll be nice to talk to him but that is not to the point here.) 11 At least not in any straightforward sense. One might think that perhaps reasons cause actions in some roundabout way. For instance, Crane and Mellor talk about mental entities that may be ‘causal surrogates’ for our reasons; see Crane/ Mellor 1990, 9.

The Causalist/Anti-causalist Debate in the Theory of Action


actions, or of how our reasons explain our actions, rather, it is knowledge of what happens in someone’s body when they act for a reason. Now, if motivating reasons are not causes of action, what remains of causalism? As I said above, some people who describe themselves as ‘causalists’ reject the idea that the reasons for which we act are causes of our actions but nonetheless maintain that an explanation of an action whose explanans is the reason for which the agent acted is a causal explanation. This is what I termed the “causal-explanation claim”, to which I now turn. 7. The Teleological Character of Reason Explanations of Action In this section I shall examine reason explanations of action and argue that they are teleological explanations. Of course, the view that reason explanations are teleological is probably not very controversial. What is controversial is whether these explanations are also causal—or reducible to, or explanatorily-dependent on, causal explanations. That issue I shall examine in the following section. A teleological explanation is an explanation whose explanatory power depends on reference to an end (or goal, or purpose) ‘for the sake of which’ something is done. Aristotle was the first philosopher to emACHTUNGREphasAiCHTUNGRE se that the explanation of human action by reference to the agent’s reason is teleological. And he thought of teleology as one of the four ‘kinds of cause’, namely the final cause, the telos or end, the ‘for the sake of which’ something is done (see Physics,198a18 ff.). Of course, the fact that Aristotle regarded an end as a kind of cause does not mean he thought that teleological explanations are causal in our sense because his concept of a cause was much wider than ours. Our concept of causation corresponds, very roughly, to Aristotle’s “efficient cause”. This is a rough correspondence because what characterised efficient causes for Aristotle is simply that they are initiators of change: he often describes an efficient cause as the mover, that which brings about the change, an agent. Moreover, he certainly thought that substances could be, literally, genuine efficient causes, whereas on the view that dominates nowadays, only events, and perhaps states, are thought to be efficient cases.12 12 Davidson and other causalists have claimed Aristotle for their camp on the grounds that he says that “the origin of action—its efficient, not its final cause— is choice, and that of choice is desire and reasoning with a view to an end”


Maria Alvarez

Be that as it may, our question is whether an explanation whose explanans is the agent’s reason for acting is a causal explanation. Let us consider a reason explanation of an action: (1) The doctor prescribed him antibiotics because antibiotics are effective in fighting infections. This explains the doctor’s action, namely, prescribing antibiotics, by citing her reason for doing so, namely, that antibiotics are effective in fighting infections. (Although it is true that anyone who understands this explanation does so because they are aware of certain implicit facts, e.g. that the patient had an infection, etc. Because of this, perhaps we should say that the explanans includes also those implicit facts. I leave that issue aside as it does not affect my point here.) Now, in order to understand (1) as an explanation in terms of the doctor’s reason, one also needs to ascribe an aim or goal to the doctor, the telos of her action, that for the sake of which she acted. In our example, that is, presumably, to promote her patient’s health—which in the circumstances took the form of ridding him of an infection. It is in the light of that end that the doctor’s reason, that antibiotics fight infections, makes sense of her action of giving him antibiotics. In other words, it is in the light of that end that the fact that antibiotics are effective in fighting infections provided her with a reason for what she did. So, a reason explanation depends on implicit or explicit reference to the agent’s goal. The reverse is also true: an explanation in terms of purposes often depends on implicit or explicit reference to the agent’s reason. For example, (2) The doctor prescribed him antibiotics in order to rid him of the infection. depends on an implicit reference to a reason for prescribing antibiotics, namely, that antibiotics are effective in fighting infections, for the doctor’s goal or purpose makes sense of the action in light of that reason. We might say that the full explanation is: (Nicomachean Ethics 1139a 32 ff.). See also Aristotle Metaphysics 1048a3 ff.; De Anima, 433 a31 – 32. However, in that passage in the Nicomachean Ethics Aristotle goes on to say: “Hence choice is either desiderative reason or rationative desire, and such an origin of action is a man” (Nicomachean Ethics 1139 b2 – 3; my italics). What seems clear is that Aristotle thought that teleological explanations in terms of final ends were compatible with explanations in terms of efficient causes, i.e. explanations in terms of the origination of change.

The Causalist/Anti-causalist Debate in the Theory of Action


(3) The doctor prescribed him antibiotics in order to rid him of the infection, because antibiotics are effective in fighting infections. Thus, reason explanations, which are explanations by reference to a reason in the light of which the agent acted, are teleological in the sense that their explanatory power depends on relating the agent’s reason to a goal or purpose for the sake of which he performed the action. 8. Are Teleological Explanations Causal? The idea that reason explanations are teleological in this sense I take to be fairly uncontroversial. But causalists would likely say that this is only half the story about reason explanations, and that when the full story is told, it becomes clear that reason explanations are causal. According to causalists, the next thing to say about reason explanations is something that brings us back to a point I mentioned above in sections 3 and 4, a point that concerns the proper form of reason explanations. In section 3, I argued that reasons are not ‘believings’ but ‘beliefs’, i.e. what is believed. And in section 4, I argued that the reason for which someone acts does not normally include their wanting something. However, I also emphasised that we must distinguish between reasons that motivate and reasons that explain actions because a reason may explain an action without being the reason for which the agent acted. Thus, it is open to a defender of the causal-explanation claim to argue that, although motivating reasons are not mental states, explanatory reasons are. Or to put the point differently, to argue that all explanations of action explain by reference to the agent’s believing and wanting things; in short, all explanations of action conform to the ‘Humean model’: “She acted because she wanted … and believed that …”. And in support of that claim, the causalist can bring two arguments to bear. One is that for someone to act for the reason that p and with a purpose or goal, that person must believe, or know that p, and must have that purpose or goal. And although this may not be explicit in every reason explanation (as is the case with (3) above), such explanations are enthymematic and, if they were given in full, they would be shown to fit the Humean model. The second argument is that, as we saw in section 3, when someone’s belief is known or suspected to be false, the explanation of their action must take the Humean form. Accordingly, (3) above, if made fully explicit would be something like this:


Maria Alvarez

(4) The doctor prescribed him antibiotics because she wanted to rid him of the infection he had, and believed that antibiotics are the means to do so. Moreover, they argue, such Humean explanations, i.e. explanations whose explanans are ‘desirings and believings’ (that is, the having of reasons and goals rather than the reasons and goals had), are causal explanations. Now, I see no reason to accept that all reason explanations are elliptical for Humean-type explanations, but I shall not argue the point here because Humeans are right that many explanations of action take the Humean form, that in error cases an explanation of an action must take that form (or something quite similar), and that a Humean explanation can be constructed for most, if not all, actions performed for a reason. So these explanations are central to explaining actions and the question remains whether, as defenders of the causal-explanation claim argue, Humean explanations are causal explanations. 9. Are Humean Explanations Causal Explanations? One of the main difficulties in deciding whether Humean explanations are causal is that there is no straightforward, never mind widely accepted, account of what makes an explanation causal. And without such an account, it is hard to decide what hangs on agreeing that Humean explanations are indeed causal. One plausible test of whether an explanation is causal is whether its explanans and its explanandum mention items (events, substances, etc.) that are related to each other as cause and effect. However, on this criterion Humean explanations do not seem to be causal because, as we have seen, the reason for which someone acts (mentioned in the explanans) and their action (mentioned in the explanandum) are not related as cause and effect. Another possible suggestion here is to say that a causal explanation of an action is an explanation that cites the causal antecedents of the action: the explanans of a Humean explanation, then, mentions states of believing and desiring which are the causal antecedent of actions. And therefore Humean reason explanations might be said to be causal because they are particular instances of empirical generalisations about human behaviour,

The Causalist/Anti-causalist Debate in the Theory of Action


in particular of empirical generalisations about what people who have certain beliefs and desires tend to do.13 The problem with this suggestion, however, is that it is not clear that the generalisations that underwrite Humean explanations are empirical. In other words, it is not clear that the thought that people tend to do the things that they believe will bring them what they want, or any specification of this general principle, is an empirical generalisation. Consider the generalisation that people who want to improve their health will tend to do things that they believe will contribute to improving their health. First, it seems implausible to suggest that this is, as genuine empirical generalisations are, a generalisation that we have derived from observation or experience. Second, it is not clear that anything could count as evidence against that generalisation. That is, it is not clear that anybody’s behaviour could count as evidence that this generalisation is false, rather than as evidence that the person did not really want to improve their health; or that they didn’t really believe that the relevant activity was so healthy, or that they were being weak-willed. But if a generalisation is not derived from experience, and if nothing could count as evidence that the generalisation is false, then it is not at all clear that the generalisation in question really is an empirical one. In fact, it seems to me that these generalisations about behaviour are not empirical, and that Humean explanations do not cite mental states that are causal antecedent of actions; rather the relevant generalizations reflect the conceptual connections that, as some of the philosophers that Davidson was arguing against had claimed, exist between ascriptions of beliefs and wants, and descriptions of people’s behaviour as being this or that kind of action. So the argument goes—an argument I find compelling—the kind of conceptual connection that there is between wanting something and doing what will satisfy the want suggest that states of wanting in general are not causal conditions of the intentional actions they explain (mutatis mutandis the same is true of having a belief, and indeed of having an intention).14 Such actions are the manifestations, rather than the effects, of 13 This seems to be Hornsby’s suggestion: “[These] explanations rely on a network of empirical interdependencies, recorded in counterfactuals (‘If she had not wanted—, but had still believed that—, then …’)” (Hornsby 1990, 165fn). 14 This point is not the old ‘logical connection’ argument that Davidson attributes to Melden (see Davidson 1980, 13). There may be no logical implication from “I wanted to turn on the light” to “I turned on the light”—but there is a conceptual connection between my wanting something and what I do, for if I claim to want


Maria Alvarez

the agent’s wanting something.15 Or to put the point differently, whether someone with the appropriate beliefs does what will satisfy their want is one of the central criteria for attributing the want to them. So, what people do is part of the criteria for what beliefs and wants it is right to ascribe to them, just as how people behave is part of the criteria for whether they are, e.g. generous. A person is not generous if he never performs generous acts; performing generous acts just is part of what being generous is. Likewise, part of what it is for a person to want x is that, when faced with the opportunity and given the ability to get x, the person will try to do the things that he believes might bring it about that he has x.16 Because of this, it seems wrong to think of their having the want as the cause of the relevant action; their intentional behaviour is partly constitutive of what having that want is, in a way similar to that in which generous acts are constitutive of the virtue of generosity. That is, although to have the virtue is to be inclined or predisposed to perform generous acts, this does not mean that the virtue causes those acts; rather the performance of such acts when the occasion arises is what having the virtue, what being generous, consists it. The view that there is a conceptual connection between whether someone has a want and what actions they perform has the consequence that there is a certain indeterminacy of the mental; not so much that we cannot know whether someone has a particular goal until we see how they act; but rather that their having the particular goal, or how strong their desire for that goal is, is partly constituted by how they act. Thus, in advance of the relevant action, it is partly an open question whether they indeed have the goal, or the extent to which they have it—it is an open question not because we cannot yet know it but because the answer is still indeterminate. (I say partly because, if an agent says they do, that counts for something, given that that is another of the criteria for the attribution of wants.) However, this does not mean that ‘there is no fact of the matter’ about whether an agent wants something, or about whether he acts for a reason or not (or for one reason rather than another). The point is rather that such facts about the having of wants and beliefs are not wholly determined by how things are in advance of the action, i.e. they to do A but do not do it when the opportunity arises, even though I am not prevented, my claim to want to do it is, to that extent, undermined. 15 As Anscombe puts it: “The primitive sign of wanting is trying to get.” (Anscombe 1957, §36) 16 I am putting aside complications created by the incompatibility of the various things a person may want.

The Causalist/Anti-causalist Debate in the Theory of Action


are not fixed merely by antecedent conditions. And this is because the fact whether the agent had the relevant wants and beliefs is itself partly constituted by the way the agent acts. Some people think that this point trades on a conflation of epistemology and metaphysics. So, they argue, it is true that we cannot know whether people want x or believe that p until we see how they act but that does not mean that people don’t want or believe those things until we find out that they do. But this is to misunderstand the point being made. The idea is not that we cannot know whether someone wants x unless we also know what he believes and how he acts. The idea is, rather, that whether someone wants x is partly determined by whether, given certain beliefs he has, he behaves in a certain way. (And the same goes for whether A believes that p.) As I noted above, whether someone wants x is partly determined by how, given certain beliefs he has, he behaves, because there are other factors that contribute to determining that fact. For instance, what A says if asked what he wants, the kind of feelings and emotional reactions A has (whether displayed or concealed) to the possibility of getting or losing x, etc. are also relevant in determining facts about what he wants and believes. The same is true of the example of generosity. The generous man is the person who has certain feelings, emotions and reactions when faced with the possibility of helping others, etc. Nonetheless, a person who has the feelings and emotions but never or rarely performs generous acts even though not prevented from doing so is not generous: he may be just sentimental. But if Humean explanations cannot be said to be causal on account of their being particular instances of empirical generalisations about human behaviour, because these generalisations are not empirical, what does the claim that they are causal amount to? Perhaps all that the claim that Humean explanations are causal explanations is supposed to amount to is the claim that someone’s wanting and believing certain things makes a difference to what he does. If so the claim seems fairly unobjectionable, so long as one remembers that what someone wants and believes is also partly determined by, and therefore not wholly fixed in advanced of, the relevant actions. As we might put the point, though this way of putting it needs qualification and explanation, what someone wants and believes makes a difference to what he does, but what he does also makes a difference to what he wants and believes. Davidson argued that the concept of causation needs to be introduced in order to distinguish between having a reason and acting


Maria Alvarez

in a way that is consistent with that reason, and having a reason and acting because of that reason (for the full argument, see Davidson 1980, 9 ff.). He claimed that the cases might be distinguished by saying that in the second case, but not in the first, the reason caused the action. We have seen that the idea that the reasons for which we act are causes is fairly problematic. Moreover, since Davidson made that claim, a number of philosophers, including Davidson himself, have pointed out problems that show that introducing the concept of causation is certainly not sufficient to account for the relation between an agent’s reason and his action—I have in mind the problem of so-called ‘deviant causal chains’ (and the corresponding possibility of deviant causal explanations of action) and the charge that causalism about reasons has the consequence that the mental is, ultimately, explanatorily irrelevant, to name but two. On the other hand, a teleological account enables us to distinguish between the two cases, for according to the teleological account, the reason for which someone acts is uniquely related to the goal the agent pursued in acting. Suppose you have two reasons to visit your rich aunt: that it’ll alleviate her loneliness, and that it’ll make her more likely to name you her sole heir. And suppose you visit her only for the second reason. This reason gives your goal in visiting her, namely to inherit her money. What characterises the reason for which one acts is that it in fact indicates the goal, purpose, or end for the sake of which the action was done. We thus have identified the pattern that characterises explanations by reference to the agent’s reason: they are teleological in nature. To explain an action by reference to the agent’s reason is to introduce (perhaps implicitly) the notion of a goal, in particular the goal or end for the sake of which the action was done. The defender of the causal explanation claim says that the ‘because’ of Humean explanations is a causal ‘because’. The question is what this amounts to. If all that the claim amounts to is that what the person wanted and believed made a difference to what he did, it is not at all obvious that this elucidates the supposedly obscure ‘because’ of explanations of action—especially as we cannot construe that difference in terms of cause and effect. Davidson claimed that we understand the force of a causal ‘because’ in the context of the relation between reasons and actions. We might respond that, given the problems mentioned above, it is not clear that we do. Moreover, the fact is that we understand a teleological ‘because’ at least as well as we understand a causal one: agents act because of the goals they have, guided by the things they believe (see Schueler 2003, esp. 44 –

The Causalist/Anti-causalist Debate in the Theory of Action


69). The notion of a rational goal is at least as familiar as that of a rational cause. Indeed, given the discussion so far, we might say that we understand the force of a teleological ‘because’ in the context of reason explanations of action better than we understand the force of a causal one. 10. Conclusion I hope the foregoing discussion has served to throw some light on the character of the causalist/anti-causalist debate in the theory of action and its significance. On the one hand, as we saw in section 4, causalism has encouraged, and been encouraged by, a conception of what the reasons for which we act are, i.e. that they are mental states. The prevalence of this conception of reasons has had an odd consequence: that for many years, areas of philosophy where the concept of a reason is central, such as the theory of action and ethics, have operated as if the reason for which someone acts and the reason there may be for him to act are things of very different kinds. But, as Dancy has argued, this implies that we can never act for a reason there is for us to act (see Dancy 2000, chap. 5). And although it is undeniable that we do not always act for the reasons we ought to, there must be something wrong in our conception of reasons if it has the consequence that we can never, as a matter of conceptual necessity, do so. Sometimes causalism is presented in its weaker form simply as a doctrine that says that a person’s wanting and believing some things makes a difference to what she does. This form of causalism, if it is right to call it that, seems in itself unobjectionable (except, perhaps, to some eliminativists). However, although unobjectionable in itself, it is not wholly harmless, as the insistence that it is not possible to explain the relation between reasons and actions without invoking the notion of causation certainly contributed to establishing and perpetuating the stronger causalist claim, that is, the idea that reasons are mental states that cause actions. And the dominance of this causal claim in the theory of action encourages the idea that an account of the causation by brain events and processes of the motions of the body that typically happen when someone acts is an account, though at a different level of description, of the causation of actions by reasons. But it is not. And if what we are interested in is actions, reasons, intentions and the interconnections between them, it is misguided to think that the way


Maria Alvarez

to get answers is to assign priority to occurrences or processes in people’s heads, brains and bodies. It is a mistake because, among other things, reasons, intentions and actions can only be understood in the complex contexts in which human behaviour takes place. And, in suppressing that fact, causalism merely has the effect, in the end, of changing the subject.17 Bibliography Alvarez, Maria (forthcoming, 2008): “Motivating Reasons and the Explanation of Action”, in: Philosophical Explorations 11. Anscombe, Elizabeth (1957): Intention, Oxford: Blackwell. Aristotle (1984): The Complete Works of Aristotle. The Revised Oxford Translation, Barnes, Jonathan (ed.), Princeton: Princeton University Press. Bittner, Rdiger (2001): Doing Things for Reasons, Oxford: Oxford University Press. Collins, Arthur (1987): The Nature of Mental Things, Notre Dame: University of Notre Dame Press. Crane, Tim/Mellor, David H. (1990): “There is No Question of Physicalism”, in: Mind 99, 185 – 206. Dancy, Jonathan (2000): Practical Reality, Oxford: Clarendon Press. Davidson, Donald (1963): “Actions, Reasons, and Causes”, reprinted in: Davidson, Donald (1980): Essays on Actions and Events, Oxford: Oxford University Press. Evnine, Simon (1991): Donald Davidson, Stanford: Stanford University Press. Hacker, Peter (1990): Wittgenstein: Meaning and Mind. An Analytical Commentary on the Philosophical Investigations, Vol. 3, Oxford: Blackwell. Heil, John/Mele, Alfred (eds.) (1993): Mental Causation, Oxford: Clarendon Press. Hornsby, Jennifer (1997): Simple Mindedness, Cambridge (Mass.): Harvard University Press. Hutto, Dan (1999): “A Cause for Concern: Reasons, Causes and Explanations”, in: Philosophy and Phenomenological Research 59, 381 – 40. Mele, Alfred (2003): Motivation and Agency, Oxford: Oxford University Press. Schueler, George (2003): Reasons and Purposes: Human Rationality and the Teleological Explanation of Action, Oxford: Oxford University Press. Sehon, Scott (1994): “Teleology and the Nature of Mental States”, in: American Philosophical Quarterly 31, 63 – 72. — (1997): “Deviant Causal Chains and the Irreducibility of Teleological Explanation”, in: Pacific Philosophical Quarterly 78, 195 – 213.

17 I am grateful to Dawn Phillips, Aaron Ridley, and to participants in the “Action in Context” Conference for their comments on earlier drafts. This paper was written during my tenure of a Mind Association Research Fellowship.

The Causalist/Anti-causalist Debate in the Theory of Action


— (2000): “An Argument Against the Causal Theory of Action”, in: Philosophy and Phenomenological Research 60, 67 – 85. Steward, Helen (1997): The Ontology of Mind: Events Processes and States, Oxford: Oxford University Press. Stout, Rowland (1996): Things that Happen Because they Should, Oxford: Oxford University Press. Stoutland, Frederick (1998): “The Real Reasons”, in: Bransen, Jan/Cuypers, Stefaan (eds.): Human Action, Deliberation and Causation, Dordrecht: Kluwer Academic Publishers. — (2001): “Responsive Action and the Belief-Desire Model”, in: Grazer PhilACHTUNGRE oACHTUNGREsophische Studien 61, 83 – 106. Strawson, Peter (1992): Analysis and Metaphysics, Oxford: Oxford University Press. Wilson, George (1989): The Intentionality of Human Action, Stanford: Stanford University Press.

In Defence of Causalism Philipp Huebl 1. Introduction In her essay, Maria Alvarez attacks the ‘orthodox’ causalist approach in the theory of action. For simplicity’s sake, let us identify—as Alvarez does— the orthodoxy with Donald Davidson’s action theory. Alvarez claims that the causalist approach of taking reason explanations to be elliptical causal explanations is confused. For her, reasons are facts, and thus they cannot enter into causal relations. More generally, action explanations that refer to reasons cannot be given a causal interpretation under any reading. Besides, Humean explanations of actions, which reduce the causal antecedents of actions to pairs of beliefs and desires, are overrated. Alvarez puts forward what I take to be a version of the logical connection argument by claiming that actions are the “manifestations” of mental attitudes such as desires, and hence desires cannot causally explain actions. She argues that there is a normative core notion of reason. Accordingly, she does not distinguish between different kinds of reasons, like normative reasons and motivating reasons. For her, all reason explanations are teleological explanations that make sense only in the light of the purposes or goals of the agent. Most of Alvarez’ arguments are based on evidence from English usage. She discusses occurrences of the term “reason” in our everyday parlance and examples of how we typically talk about actions. My comment has five parts. First, I will put in perspective Davidson’s causal claim, namely the claim that reasons are causes. Second, I will argue that Alvarez fails to support her opposing view that reasons are facts. Third, I will show that her arguments against Humean action explanations go astray. Fourth, I will argue against her view that desires cannot be the causes of actions since actions are the ‘manifestations’ of desires. My final remark is on philosophical methodology based on the overall critique from the first four points. I believe that in making explicit our practice of explaining actions in everyday life and the sciences, we will rather make progress by conceptual analysis than by taking ordinary language at face value.

In Defence of Causalism


2. The Causal Claim What exactly is the ‘causal claim’ put forth by Davidson, that is, the claim that the reason for which someone performs an action is the cause of that action? For Alvarez, the causal claim involves “the view that reasons are mental things of a kind (events, states, combinations of these) that makes them suitable candidates to be the relata of causal relations, and hence to be called the causes of actions.” (105)1 That is not exactly Davidson’s view, but some careless formulations in his early essays invite such a reading. For Davidson, causality is a relation that holds between events only, no matter how they are described. And events are changes in the course of the world (see Davidson 1995, 272). Hence, at first sight it seems peculiar that Davidson holds that citing or indicating reasons as belief-desire pairs is citing or indicating the causes of an action, since beliefs and desires are mental states and thus they cannot literally be causes. However, Davidson was aware of this tension already in his essay “Action, Reasons, and Causes” (see Davidson 1963, 32 f.). It can be broken down as follows. His claim that mental states are causes must be read with a grain of salt. Literally only particular events, namely changes of those states, are causes of actions. [I]t is changes in the attitudes, which are events, which are the often unmentioned causes. […] And we can often turn a causal explanation which mentions beliefs or desires into an explanation which refers to an event or events by saying the cause of the action was the advent of one or both of the belief-desire pair. (Davidson 1993, 288)

However, there is a second, though unmentioned, spring for Alvarez’ quarrel with Davidson’s approach. In his later work, Davidson defines “action” by drawing a close parallel between the relation of rationalization and the relation of causation. Again, the two are of different ontological kinds: rationalization (or rational explanation) is an intensional relation between propositions, whereas causation is an extensional relation between events. Davidson never solved this problem, and there are strong arguments that it cannot be solved within his framework.2

1 2

Parenthetical references in the text are to Alvarez 2007 (in this volume). For criticism, see Keil 2007, Fn 17.


Philipp Huebl on Maria Alvarez

3. Reasons as Facts? Alvarez takes the opposite tack. She claims that reasons are what is believed or desired, not the believing or desiring of it. In other words, reasons are not mental attitudes such as states of believing or desiring, but the content of those states. The common term for the content of beliefs and desires is “proposition”. Since Alvarez also says that reasons are facts, I take her to equate true propositions with facts. At any rate, propositions and facts are abstract entities and thus reasons cannot be causes. In short, Alvarez is an objectivist about reasons. Her arguments are based on evidence from everyday parlance. According to Alvarez, a semantically congruent answer to the question “What was your reason for doing it?” is “she was in need” or “to help her”, but not “I believed that she was in need” or “I wanted to help her”. For Alvarez, the answers refer to facts, but not to mental states. Some authors distinguish between two kinds of reasons: normative reasons and motivating reasons.3 A normative reason says what I ought to do, given all relevant facts. A motivating reason says what is rational for me to do, given my beliefs and desires. If I know all the relevant facts and if I am rational, my normative reason becomes my motivating reason. But if I do not know all the relevant facts, I still have a motivating reason, though it may differ from my normative reason. To paraphrase an example from Derek Parfit, if I falsely believe that my hotel is on fire, I may have a motivating reason to jump out of the window, though I have no normative reason to jump. There is a controversy about what the exact relation between normative reasons and motivating reasons is. Alvarez assumes at least implicitly that the word “reason” has only one core meaning, which is displayed in everyday reason explanations. Evidence from ordinary usage is taken to answer the ontological question “what kind of thing reasons are” (104) and to describe their role in a larger framework of normativity. According to Alvarez, our verbal practice shows that reasons are facts and that they should be separated from motivation. In fleshing out her view, she does not distinguish between examples that are commonly taken to belong to either one of the two different kinds of reasons. Let me point out three problems in this line of argument. 3

See Parfit 1997, 99 f. Bernard Williams uses the terms “internal reason” and “external reason”, but in a slightly different way (see Williams 1980). This distinction is also drawn by Michael Smith (see Smith 1994, 95).

In Defence of Causalism


First, it is only a contingent fact about the context of an utterance, whether in explaining actions we refer to our mental states or merely to the abstract propositional content of those states, or the “facts” as Alvarez would have it. Consider a courtroom case. The defendant answers to the question “What was your reason for doing it?” by uttering “to help her”. The attorney goes on asking “Did you want to help her or did somebody ask you to help her?” In this situation, it can be perfectly normal for the defendant to answer “I wanted to help her and I believed she was in danger”, or in a different situation “I did not want to help her, but my friend asked me to help her and I wanted to do him a favour.” Note, that the first sentence need not involve a Gricean implicature that in fact there was no danger at all. The courtroom example shows that in our reason explanations we tacitly assume or presuppose that propositions cited as reasons are represented by pairs of beliefs and desires. This holds even if we only mention the content of one of those states, and even if our other non-linguistic assumptions are largely left unspecified, maybe even to us. What looks like a mere terminological issue originates from asking for the ontological category of reasons as such. I believe that this question already neglects the practice of explaining and rationalizing. What Davidson should have said in order to avoid confusion is something like this: only propositions can be reasons, but a proposition is not a reason unless it is actually used in reasoning. More specifically, the proposition only becomes a reason when the person has a mental attitude towards it. But this does not turn the proposition itself into a reason. Talk of reasons as such is derived from the relational use of reasons in action explanations. In other words, a fact is only a reason if it is represented as a proposition by a mental state. Second, consider error cases. Someone went to attend a lecture, but the lecture was cancelled. Alvarez claims that the agent acted for an “apparent” reason, but she had no “real” reason. Note that here she deviates from her methodology of ordinary language analysis by drawing a technical distinction that comes close to the distinction between motivating reasons and normative reasons. In common speech, we use the term “reason” for both cases. We even use the term “reason” in cases where with some philosophical care only the term “cause” is appropriate. Apparently, the term “reason” has at least two readings as reflected in the technical terms, namely a subjective reading in “she had a reason” (motivating reason) and an objective reading in “there is a reason for her” (normative reason). In any case, what we should be interested in is not a lexicographic analysis of the term “reason”, but how the concept of reason


Philipp Huebl on Maria Alvarez

can be analysed. Alvarez holds that one might have an apparent reason to go to the lecture, but no reason, since reasons are facts and—as a matter of fact—the lecture was cancelled. Apparently, this is only a redescription in different terms. Is there any difference between Alvarez’ wording of “The agent had no reason, but an apparent reason” and the traditional wording of “The agent had no external reason, but a motivating reason”? At least, I can see none. There is, however, a substantial problem about characterizing reasons without reference to mental states. I think that mental states, ushered through the front door, creep in through the back, namely counterfactually. Here is why: in order to explain what a real reason is, we need to say how it differs from an apparent reason. Why was my reason to go to the lecture only an apparent reason and not a real reason, given the lecture was cancelled? The answer is a counterfactual conditional: Had I known that the lecture was cancelled, I would not have gone. Knowing implies believing. So again, an explication of the concept of a real reason requires mental concepts. Third, when we are asked for our motivating reasons for action, we are asked to answer why we did something or what motivated us to do something. Again, for Alvarez all reasons are abstract entities, namely facts. Facts cannot motivate us or make any change in the course of the world. As a consequence of her assumptions, reason exACHTUNGREplaAn CHTUNGRE ations could never say or explain why something happened. Not even purpose explanations can, unless they are linked to causation. But this opposes our actual practice. We cite reasons in order to explain changes in the world that we have brought about ourselves, changes that would not have occurred, had we not acted. The concept of both, motivational and normative reason is relational. A practical reason is always a reason for an agent to do or want something, and a theoretical reason is always a reason for an agent to hold a certain belief. Doing something is changing the course of the world. It is hard to see how reasons could be related to our moral and juridical practice of blame and punishment, if no causality were involved. Alvarez separates two things in her worldview. On the one hand, there are psychological or physiological explanations of what is going on in the minds or bodies of agents when they act. On the other hand, there are reason explanations of actions in our common vernacular. But do we really talk about two different things when we explain someone’s action in the light of her reasons and when we explain it in terms of her beliefs and desires?

In Defence of Causalism


3. Motivation Without Desires? The Humean schema of action explanation takes the causal antecedents of actions to be pairs of beliefs and desires. Beliefs alone cannot motivate, they must be supplemented by some kind of positive attitude towards the action. Alvarez’ arguments against Humean action explanations are problematic in several respects. Let me discuss three of them. She considers the following example: “I am in need can motivate you to help me” (110), and takes this to be a case where no desire is involved. But apparently, if you are ignorant or sadistic, you would not help me, since you have no desire to do so. You might even have the opposite desire to see me suffering. Thus, the above example tacitly presupposes that you have a standing desire to help those in need. Alvarez equates shared preferences with facts. Suppose most of us human beings share the preference of helping people in need. From that it does not follow that the proposition expressed by this preference is a fact, namely that one should help the poor. Nor does it follow that your belief that I am in need alone can motivate you to help me. Take another example: “The doctor prescribed him antibiotics because antibiotics are effective in fighting infections.” (114) A Humean theorist might analyse this as follows: The doctor held the belief that antibiotics are effective and she wanted to help the patient. Two examples might illustrate that in action explanations we implicitly ascribe beliefs and desires to the agent. In the first case, the doctor is evil and has no inclination to help her patients. In the second case, the doctor is ignorant, for she does not believe that antibiotics are effective in fighting infections. Both examples show that the statement above makes sense only against a background of ascribing certain mental attitudes to the doctor. Alvarez believes that this view is mistaken. As a counter-analysis, she claims that in order to understand the doctor’s reason we have to ascribe a goal or purpose to the doctor. I do not see how the concept of goal or purpose can shed more light on reason explanations than the concepts of belief and desire. Goals are simply desires in sheep’s clothing. As desires, goals have content and they must be harboured or entertained by the agent. Wants or desires are only satisfied or fulfilled if the world changes according to their content. In John Searle’s terminology, they have a mind-to-world direction of causation and a world-to-mind direction of fit (see Searle 1983, 7 f.). The same holds for goals. Introducing goals into a set of explanatory concepts is only superficially helpful for anticausalists. At first sight it seems that goals can be picked out


Philipp Huebl on Maria Alvarez

independently of the mental states of persons, because the term “goal” has an extensional as well as an intensional reading like “reason”. Think of a football goal. It stands there on the green even if nobody is playing the game. Yet, this metaphysical independence vanishes on closer examination. A goal is only a goal if someone has intentionally chosen or constructed it to be a goal, or if someone wants to reach it at some point of time. Andrew Woodfield has put this point thus: I agree that if a goal is achieved, we say that it exists or has been actualised. But this teetering from intensional to extensional usage is a loose faÅon de parler. Intentional objects can never break free of their shackles, for they can never become real obACHTUNGREjects. What is actualised, strictly speaking, is always some action or state of affairs that matches the goal by satisfying a goaldescription. (Woodfield 1976, 211)

In criticising Alfred Mele, Alvarez equates two action explanations schemas, the subsumption-reading and the means-end-reading. She claims So, if my reason for singing were that I wanted to sing, then it should make sense to say that I sang in order to sing (compare: ‘My reason for driving was that I wanted to get there in time’—‘I drove in order to get there in time’). (111)

The example about driving has a means-end-reading. The agent wants to be somewhere in time, and believes that driving is the mean to get there in time. However, the example about singing is different. The agent was singing because she had a desire to sing. The particular action can be subsumed under the action type ‘singing’. In this case there is no meansend-reading. Some actions are basic in the sense that we do not have to perform another action in order to perform the basic action. For Davidson, whether an action is basic depends entirely on the description under which it is picked out. I think this assumption needs some qualification. In my view, whether an action is basic depends on the content of the intention it is performed with. I may intend to move my fingers one by one according to the musical score. But if I know how to play the sonata, I can intend to play it as a whole. In the first case, every single finger movement may be a basic action, and in the second case, playing the whole sonata may be a basic action. My reason for playing a sonata may be to entertain others, to make money, or to fulfil a commitment. But I may also play it just for the sake of it, and then my reason is my plain desire to play it. The means-end reading is not a necessary condition for attributing reasons to an agent. And consequently, a causalist is not forced to claim that all reason explanations fit the

In Defence of Causalism


instrumental means-end-schema. However, since sometimes the term “Humean explanations” is used exclusively for the instrumental kind, it is worth mentioning that at least Davidson is not a Humean in the narrow sense. 4. Actions as Manifestations of Desires? Alvarez admits that she finds an argument compelling that goes something like this: the conceptual connection between desires and actions is so close that there cannot be a causal connection. In her words “Such actions are the manifestations, rather then the effects, of the agent’s wanting something.” (117 f.). I take this to be a version of the logical connection argument, but since Alvarez explicitly denies this, I am happy to call it the ‘conceptual connection argument’. I am not convinced either way. From the fact, that we use behavioural evidence in attributing mental states to other persons, it does not follow that we cannot have desires before putting them to action. One can have the desire to climb Mount Everest without ever getting the chance to satisfy it, for instance if one serves a life sentence in prison. Besides, the same independence occurs in the reverse direction: Some actions satisfy or fulfil some of our desires, but are not related to them in any direct sense. It is difficult to tell what Alvarez means by her remark that “[…] in advance of the relevant action, it is partly an open question whether they [the agents] indeed have the goal, or the extent to which they have it—it is an open question not because we cannot yet know it but because the answer is still indeterminate.” (118) An open question to whom? To us, the observers of the actions of others? Or to us, the performers of our own actions? I believe Alvarez merges two kinds of uncertainties. On the one hand, there can be uncertainty in attributing beliefs and desires to other persons. We do this solely on behavioural evidence, since there is no other kind of evidence. Needless to say, this enterprise is not safe of error. On the other hand, we can be uncertain about the status of one of our own desires among the set of others. Alvarez seems to have the second kind of uncertainty in mind. But then it is not a point about desires in general and whether they are ‘constituted’ (whatever exactly that means) by actions, but rather a point about our biological species. Although we as human beings have largely first-person authority about our own mental states, we do not always have perfect access to them. And even when we have, sometimes we misrepresent the world. However, we could easily


Philipp Huebl on Maria Alvarez

imagine higher developed rational beings that always have this kind of access and that almost never misrepresent the world. Yet, even given our limited capacities, uncertainty does not render the relation of desires and actions conceptual. Let me flesh out the last remark: Sometimes we do not know what we prefer, until we have tried the options. Consider examples involving preferences for chocolate bars. I have two bars in front of me, a bitter sweet chocolate bar and a milk chocolate bar. What are my preferences? There are four possibilities: I prefer x to y, I prefer y to x, I like x and y to the same degree, or I lack preferences concerning them, for instance if I do not even know what chocolate is. A misrepresentation of my preferences can have at least two sources. On the one hand, in the first two cases of asymmetric preferences, my desire is specific enough to decide between the two bars. Yet, I might be mistaken about how one or each of them really tastes like. In other words, I might be mistaken about how I represent the world. For example, I believe that this bar has the taste I prefer most, but in fact it has a different taste. And by comparison, I find out that I like the other chocolate bar better. On the other hand, in the case of symmetric preferences my desire is not specific enough to decide between the two bars. I just do not know myself well enough in this very moment. Either, I do not have perfect access to my first order desires. Technically speaking, I may need to develop a higher-order mental state to represent my first-order preference. Or, I may need to acquire a preference hierarchy by actually tasting the bars. The same holds for cases, where I have no preferences at all. In the cases discussed, I have to try the chocolate bars in order to ‘manifest’ my desire, at least in one sense of the term “manifest”. And by tasting the bars I might find out something new about myself. In a loose sense of the term “open question”, before tasting the chocolate bars it was an open question for me what preferences I had. In any event, most cases involving desires and actions do not involve misrepresentation or underdetermination of preferences. I know what I want and I could tell you without first observing my own actions. And even in cases of misrepresentation: my desire is what makes me act in the first place. It is part of my reason for actually tasting the chocolate. There is another way of reading the passage, namely as a point about metaphysics and not epistemology. Desires are mental states that represent the world via their content. The same desire can be fulfilled in an indefinite number of ways, because there are blank positions in its content that can be occupied in many ways by the world without

In Defence of Causalism


affecting the fulfilment of the desire as a whole. The world itself is ontological denser, so to speak, than its representation.4 However, this metaphysical fact does not make having the desire or goal indeterminate, as Alvarez believes. It concerns something entirely different, namely the actual ways of fulfilling a desire, or of reaching a goal. These different permutations leave the content of the desire untouched. Here is another argument. Consider desires and intentions. Both are kinds of motivational or pro-attitudes in having the world-to-mind direction of fit: they are satisfied or fulfilled, if the world changes according to their content. Intentions, however, concern only changes that can be brought about by actions, whereas desires include changes beyond the agent’s control. At least if I am rational, I cannot have the intention that the sun may shine tomorrow or to live forever. But I can have a desire, or want, or wish with the same content, even if I do not believe that I could do anything to make it happen. Apparently, those desires ‘manifest’ themselves somehow in speech, but there are no specific actions that are ‘partly constitutive’ for entertaining those desires. 5. Conceptual Explication as Philosophical Methodology I take philosophy to be the science that deals with the most basic concepts and how they are related to one another.5 A philosophical action theory reconstructs explicit and implicit assumptions about actions explanation in everyday life and the sciences by relating the basic concepts to one another, which underlie this practice. Alvarez adheres to a narrower approach. She refers to overt linguistic evidence from ordinary usage of English. According to her, we typically say “the reason was that it was raining”, but not “the reason was that I believed that it was raining”. Alvarez draws substantial conclusions from that finding. For her, since we do not use the words “belief ” and “desire” in any direct paraphrase, we have no grounds to assume that the concepts of belief and desire play a role in reason explanations. Yet this misrepresents the whole point about conceptual analysis. The causalist’s claim is that reason explanations are implicit, or elliptical, or metonymic causal explanations. In order to make our largely tacit practice of explaining actions explicit, we need to introduce two basic mental attitudes, belief and desire, and take changes 4 5

See Keil 2007, 81 – 83. What Peter Strawson calls “connective analysis”; see Strawson 1992, chap. 2.


Philipp Huebl on Maria Alvarez

of pairs of them as causes. In other words, talk about actions rests on some of our most basic concepts such as “belief ”, “desire”, “cause”, and “event”. I do not see the point in taking ordinary language examples at face value, since when we talk, only few of our assumptions and only few ascriptions of mental attitudes have a reflex in our words. Hence, as long as Alvarez does not provide a theory of language comprehension, which specifies how overt linguistic data and implicit linguistic and other assumptions interplay, it is futile to regard examples from a single natural language as sacrosanct. Especially, as long as these generalisations are not shown to hold for all (possible) languages. In two cases, Alvarez even departs from her own course. First, when she allows technical terms to come into play, like “real reason” and “apparent reason”. Evidently, non-philosophical speakers often use the same word “reason” for both cases, but can be made aware of the difference. So shall we take their initial use of words seriously, or their use after they have learned the distinction? Second, Alvarez says “[t]o explain an action by reference to the agent’s reasons is to introduce (perhaps implicitly) the notion of a goal, in particular the goal or end for the sake of which the action was done.” (120) When Alvarez allows ‘implicit introduction’ of concepts, she assumes precisely what causalists already said about beliefs and desires. Does her analysis in terms of the concept of goal have any advantage over the orthodoxy? I do not think it has. To say someone has a goal, presupposes that she wants to reach this goal. In other words, the concept of goal presupposes the concepts of intention or desire. In the end, one has to assume that the goal is somehow mentally represented by the agent. Something similar applies to downplaying the role of Humean explanations. The defenders never claimed that they often occur in their genuine form. Humeans never made a point about relative frequency of schemas of explanations, but a point about the presupposed form and the most basic concepts of actions explanations. Bibliography Alvarez, Maria (2007): “The Causalist/Anti-causalist Debate in the Theory of Action: What it is and Why it Matters” (this volume). Davidson, Donald (1963): “Action, Reason, and Causes”, reprinted in: Davidson, Donald (1980): Essays on Actions and Events, Oxford: Oxford University Press.

In Defence of Causalism


— (1993): “Replies”, in: Stoecker, Ralf (ed.): Reflecting Davidson. Donald Davidson Responding to an International Forum of Philosophers, Berlin/New York: De Gruyter. — (1995): “Laws and Cause”, in: Dialectica 49, 263 – 279. Keil, Geert (2007): “What Do Deviant Causal Chains Deviate From?” in Lumer, Christoph/Nannini, Sandro (eds.): Intentionality, Deliberation and Autonomy: The Action-Theoretic Basis of Practical Philosophy, Aldershot: Ashgate (forthcoming). Parfit, Derek (1997): “Reasons and Motivation”, in: The Aristotelian Society, Supplementary Volume 77, 99 – 130. Searle, John (1983): Intentionality, Cambridge: Cambridge University Press. Smith, Michael (1994): The Moral Problem, Oxford: Blackwell Publishers. Strawson, Peter F. (1992): Analysis and Metaphysics: An Introduction to Philosophy, Oxford: Oxford University Press. Williams, Bernard (1980): “Internal and External Reasons”, reprinted in: Williams, Bernard (1981): Moral Luck, Cambridge: Cambridge University Press. Woodfield, Andrew (1976): Teleology, London: Cambridge University Press.

Two Ways to Understand Causality in Agency Rowland Stout 1. Introduction I want to argue for a causal approach to agency based on an Aristotelian model of causal processes rather than on the more familiar model of a network of causal links.1 Processes in this Aristotelian model are the realisations of potentialities. I will talk of mechanisms as the things that have potentialities; so processes can also be thought of as the workings of mechanisms. An influential philosophical conception of our mind’s place in the world is as a site for the states and events that causally mediate the world we perceive and the world we affect. According to this conception, states and events in the world cause mental states and events in us through the process of perception. These mental states and events then go on to produce new states and events in the world through the process of action. Our role is as hosts for these states and events that causally mediate the states and events on the input side and those on the output side. The picture can be made a bit more complicated by adding extra loops. So mental states and events interact with other mental states and events before they cause the body to move. And feedback loops should be added so that the effects of these bodily events themselves cause further mental states and events which then cause different bodily and worldly states and events. But however complicated the causal chain is—perhaps looking more like chain mail—the structure of states and events in the middle of the chain constitute the mind, and causally mediate between non-mental inputs and non-mental outputs. This picture can be described as Cartesian without yet committing it to any sort of immaterialism. It is Cartesian inasmuch as it takes the role of the person and their mind to be located in the middle of the chain of events passing from perceived inputs through the person’s body to 1

I have been arguing for this in various ways, some of which recur here, for some time now. See Stout 1996; 2002; 2005; 2006.


Rowland Stout

achieved outputs, just as Descartes in The Passions of the Soul placed mental events sandwiched in a causal chain between two sorts of movements of the pineal gland, the first affected by external stimuli and the second causing the body to move. Many philosophers, often influenced by Wittgenstein, have rejected such a picture even without the immaterialism, thinking that in some problematic way it still makes mental states and events internal and unworldly. They also find it difficult in this picture to give reasons for action their proper role in agency. How might reasons for doing things—teleological reasons in particular—be thought to have a place in action given the Cartesian model of our causal place in the world? First they might be taken to be facts about the world outside. The fact that the only way to get water is to dig a deep hole under this palm tree is a reason for digging the hole if you need or want to get water. But in the Cartesian picture this fact does not figure in action. It figures in perception by causing various mental states including the belief that the only way to get water is to dig the hole. But once it has caused this belief that fact has no further causal role in the story. The belief combines with other mental states to cause the desire or intention to dig a hole and this desire or intention causes the body to move. If we accept all this then we must deny that action essentially involves reasons. The problem simply put is that reasons figure on the input side of the Cartesian picture and action figures on the output side. The two sides exist independently of one another. So reasons turn out not to be essential to action. We might try to avoid this separation of reasons and action by taking reasons to belong to the output side of the picture. This would be done by describing the mental states and events (or facts about those mental state and events) as the reasons for action. So we could say that your reasons for digging the hole are your desire for water combined with your belief that digging the hole is the way to get water. But this strategy undermines the idea of reasons for action. Facts about your mental states are not facts about means and ends. So they certainly cannot count as teleological reasons. And indeed as many people, including Fred Stoutland in this volume, argue, such facts are not really reasons for action.2 At best they are reasons why you act. Someone’s knowing reasons why you act can make your action rationally intelligible to them, but such reasons do not favour or justify your acting that way. It 2

See also Dancy 2000; Stout 2004; Stoutland 2007 (this volume).

Two Ways to Understand Causality in Agency


is the fact that digging a hole is the way to get water that favours or justifies digging a hole; the fact that you believe that digging a hole is the way to get water does not. It may seem that this Cartesian picture is quite inescapable if we acknowledge the causal nature of perception and action. How else could our mental states and events stand in the world, causally receptive and causally active, except as links in the causal chain (albeit a chain with feedback loops) from the world we perceive to the world we act on? With this worry in mind philosophers who reject the Cartesian picture very often tend to reject causal approaches to perception and action altogether. But I think that the apparent inescapability of construing our causal role as the role of links in a causal chain derives from a prevalent but quite avoidable conception of causation itself. When we see our causal role in another way then we can reject the Cartesian picture without being anticausalists. This prevalent conception of causation takes a basic causal process to be constituted from two particular events or perhaps complexes of events and states—the cause and the effect—and a generic relation of causation linking them. The causal world can be described according to this conception by specifying which structures of events and states are linked to one another by this relation. The causal world will look like a huge chain or network of events and states linked together. A diagram of such a causal world will have names for the events and states and a complex structure of arrows linking these names. Let me call this the causal chain model of causation. The philosophy of causation is dominated by this model. The questions it asks concern how to understand the causal relation. For example, is it to be understood as a counterfactual relation linking two events? To what extent does causation involve necessitation? Does it only link events or also states and perhaps facts or other things? But this is certainly not the only model of causation. There is also an Aristotelian model that allows in addition to a cause and an effect not just a generic relation of causation linking them, but individual causal processes. In this model we can ask what the cause is and what the effect is, but in addition we can ask how the effect results from the cause—i.e. what the mechanism is. And this how-question is not to be answered just by introducing more links in a causal chain. It is answered by specifying the potentiality whose realisation is the process that the effect belongs to. This Aristotelian model of the causal world cannot be fully pictured as a network of arrows, since each arrow would have to represent a


Rowland Stout

different sort of causal process. It is a world of mechanisms having potentialities which are realised in the workings of the mechanisms and which realise each other. And applying this model to our minds’ place in the world we can still take minds to stand between the perceived world and the affected world—causal intermediaries in a sense. But our mental states are not taken to be causal intermediaries in the sense of being nodes in the middle of a chain of causal links—caused by the inputs and causing the outputs. Instead they are taken to be the states of causal mechanisms transforming the inputs into outputs. According to this conception, states and events in the world are inputs into causal mechanisms that constitute our nature as both subjects and agents. Essential to these mechanisms are the potentialities whose realisations constitute the workings of these mechanisms. The way our minds are is the way these mechanisms work—the way they transform perceived inputs into produced outputs. These states of mind are causally affected by the environment as well as by each other; but on this model a change in one’s state of mind is not an essential part of a causal chain between the environment and one’s behaviour. There has been a revival of interest in mechanisms recently in the philosophy of science. See Machamer et al. (2000), though it is possible that these philosophers mean something slightly different by mechanisms. An important earlier attempt to argue for an Aristotelian conception of causation is Harr and Madden (1975). And the task of trying to understand dispositions and potentialities has always been part of the remit of contemporary metaphysics (e.g. Mumford 1998, working on well-established debates) even when this has not been taken to be part of the task of understanding causation. Thinking of the mind in terms of mechanisms, potentialities and dispositions is not new either. While Ryle (1949) is the most famous proponent of a dispositional view, Putnam’s (1967) machine-table functionalism, taking mental states to be analogous to states of a computer programme, does something similar. In the philosophy of action too, some philosophers are either using or feeling their way towards the idea that agency needs to be understood in terms of the operation of certain sorts of mechanisms. I will refer to some of these later. But the debate here has not been set out clearly yet as a debate between two alternative conceptions of causation. When it is set out in this way I think that much of the debate that is supposed to concern causalism versus anti-causalism would be much better recast as concerning alternative conceptions of causality in agency.

Two Ways to Understand Causality in Agency


Gilbert Ryle’s apparent anti-causalism serves as a good example. He claims that appealing to dispositions or potentialities in explaining things is not to provide causal explanations: There are two quite different senses in which an occurrence is said to be ‘explained’. … The first sense is the causal sense. To ask why the glass broke is to ask what caused it to break, and we explain, in this sense, the fracture of the glass when we report that a stone hit it. The ‘because’ clause in the explanation reports an event, namely the event which stood to the fracture of the glass as cause to effect. But very equivalently we look for and get explanations of occurrences in another sense of ‘explanation’. We ask why the glass shivered when struck by the stone and we get the answer that it was because the glass was brittle. (Ryle 1949, 88)

And in the same way, according to Ryle, when explaining actions in terms of reasons we are appealing to dispositions and so not providing causal explanations. His anti-causalism is anti-causal-chain-causalism. Alfred MelACHTUNGREden is the same: Stating the motive is not offering a (Humean) causal explanation of the action. The explanation does not refer us to some other event—the motive— which explains causally how the action comes to be. (Melden 1961, 102)

He attributes the causal chain model of causation to Hume and argues that this model does not apply in the explanation of action. But this leaves open the possibility that an Aristotelian model of causation might apply to the explanation of action. Elizabeth Anscombe (1957) does the same thing. She rejects the idea that explaining an action is done by describing the ‘mental cause’, and in so doing she appears to reject causalism. And many of the crop of contemporary anti-causalists can be described in the same way. They reject a particular version of causalism— one which perhaps identifies reasons for action with mental causes— without considering other ways that reasons for action may be involved in causal processes. In this paper I want to present some considerations in favour of thinking of agency in terms of the Aristotelian causal process model rather than the Cartesian causal chain model. These considerations by themselves do not show the causal mechanism model to be correct. They just show what might be gained in the philosophy of action if we could show this model to be correct. In each case the traditional approach that looks at what causes what is seen as failing because what seems to be important is how the behaviour is caused, not just by what.


Rowland Stout

2. Making Events Happen and States Obtain My first task is to motivate a causal approach to agency. The standard causal theory of action, which identifies action with mental states or events causing physical behaviour, employs the causal chain conception of causation. My defence of causalism will not go through this route. So I want to start a bit further back with the claim that agency is best understood in terms of the intentional achieving of goals. Indeed we can go further back still and consider whether agency is best understood in terms of intentionally making events happen and states occur. This is something which is certainly not accepted by everyone, and not even by everyone who accepts some sort of causal approach to agency (see Thalberg 1967). I want to consider a two stage argument for the claim that agency is best understood in terms of making things happen. Firstly, agency should be understood in terms of intentionally doing things. Then, doing things should be understood in terms of making the things done happen. For example, raising one’s arm is making one’s arm rise. So agency should be understood in terms of intentionally making things happen. The first stage of this argument is not particularly controversial. Someone might object that failing to do something is just as much a manifestation of agency as doing something. But trying and failing to do something is still doing something. And in any case, it is likely that failed trying is best understood in terms of successful trying. So I think we are safe to start thinking about agency in terms of intentionally doing things. The second stage of the argument is much more tricky. It might be thought that action is a paradigmatic causal concept. Philosophers have often looked for accounts of natural causation in terms of our concept of human agency. But it would not follow that we should seek a causal analysis of doing, since there may be no causal notion more basic than that of doing by which doing should be understood. No one doubts that causation itself is a causal concept, but causation should not be understood in causal terms. If doing is a very basic causal concept then a causal analysis of doing might be as futile as a causal analysis of causation. The example of raising one’s arm is a very questionable choice as a paradigm of agency. Philosophers of action very often look at these very simple acts or achievements as the targets of their analysis. But it is not at all clear that what characterizes agency are acts like this rather than activities. Many of the things we do seem to be activities rather than achievements. Eating a banana, going for a walk, thinking about a

Two Ways to Understand Causality in Agency


problem, chatting with a neighbour are all doings that do not seem to amount to mere achievements. Eating a banana is not just achieving the goal of the banana being eaten. I can achieve that goal by getting my four year old daughter to eat a banana, without having to eat it myself. And it is not obvious that it should be understood in terms of making something happen or in causal terms at all. The difficulty here is finding something other than the activity of eating the banana that counts as the thing caused when you eat a banana. It is certainly not right to identify eating a banana with causing oneself to eat a banana. For in that case eating a banana would be causing oneself to cause oneself to cause oneself and so on ad infinitum to eat a banana. But is there anything other than the activity itself that might be the thing one causes when one acts? In the case of actions which are movements of one’s body there is a very obvious candidate. My raising my arm might be identified with my causing my arm to rise. My arm rising is not an action—it is not what I do. To employ the Aristotelian terminology made familiar by Jennifer Hornsby’s work, it is an intransitive movement, whereas moving my arm is a transitive movement (Hornsby 1980, 2). But even in this limiting sort of case of an action, it might be objected that I am not really causing my arm to rise; I am just raising it. The idea of causing my own arm to rise might sound rather odd—as if I have to employ some non-standard device like autohypnosis to make the arm rise. But I think this oddness may easily be attributed to the fact that in normal cases we have a simpler way to express what happened—namely that I raised my arm. The implication that there is some non-standard route to the arm rising when I make it rise can easily be cancelled. Suppose there was something else like a force field or an electric shock that might have made my arm rise, though it did not. Then a sensible question for someone to ask is: “Did you make that happen?” And the sensible reply is: “Yes I caused my arm to rise.” So I think it is right to identify moving one’s body with causing one’s body to move. But not everything one does is moving one’s body. Most of what one does involves moving one’s body in some ways, but it is not at all clear that eating a banana, going for a walk, or having a chat with one’s neighbour should be identified with moving one’s body, even in a very complex way. But even if these activities should not be identified with a structure of intransitive body movements perhaps they can be identified with a structure of other intransitive achievements. It is too simple to say that


Rowland Stout

my eating a banana is my making the banana get eaten. But perhaps it is not so far wrong to identify it with my making a structure of events/states happen/obtain that are characteristic of my banana eating—i.e. roughly speaking by making the banana become ingested through my mouth into my digestive system. Perhaps my going for a walk is my causing some characteristic structure of achievements associated with going for a walk. It is not obvious exactly how to describe such a characteristic structure of achievements; but it looks like a perfectly reasonable philosophical task to work it out. One consideration in favour of thinking of actions in terms of achievements is that with these activities it is always still possible to try but fail to do them. It would be strange, but not impossible, to try but fail to eat a banana, go for a walk or chat with one’s neighbour. This suggests that with these activities there is something that can be successfully achieved or not. And this in turn suggests that these activities involve the achievement of goals. If eating a banana is achieving a structured goal then it is also making that goal become achieved. So I propose that a subject, S, v-ing can be identified with S causing a structure of results characteristic of their v-ing. Certainly this identification seems pretty awkward in some cases. But the move I make in the next section may help to reduce this awkwardness. For there I will argue that we can give a causal account that does not merely state what is caused, but states how it is caused. My eating a banana may then be seen to be not merely causing a structure of results characteristic of my eating a banana, but causing that structure of results by means of the operation of my eating mechanism. The thing about these activities is not just that they realise the capacity of agency as such, but also that they realise particular further instinctive or learnt capacities. This idea can be found a place in the Aristotelian model. 3. How Results Are Caused A traditional schema for analysing a causal notion—F—is as follows: S F-s O if and only if some aspect of S causes O to have some quality characteristic of F.

Consider the wind drying the washing. Using the traditional schema this is to be understood as some aspect of the wind causing the washing to be dry. Which aspect of the wind should be specified here is up for

Two Ways to Understand Causality in Agency


discussion. Perhaps it is the wind itself that is doing the causing or perhaps it is some event or state associated with the wind. This traditional analysis explains the causal notion in terms of a causal link between two things—a characteristic cause and a characteristic effect. As such it uses what I have called the causal chain model of causation. The obvious inadequacy of this account of the wind drying the washing is that by only specifying the cause and the effect it misses out the way the effect is caused, and this way too has to be characteristic of wind-drying. This can easily be seen by constructing a deviant causal chain counterexample. Suppose the wind blew open the window, which banged into the button that turned on the tumble dryer in which the washing soon dried. The wind caused the washing to be dry but it did not dry the washing. This is because it caused the washing to be dry in the wrong sort of way. It did not result in the washing becoming dry through the realisation of a wind-drying mechanism. This explanation suggests an alternative way of analysing causal notions—what I have called the causal process model. S F-s O if and only if O having the F-quality belongs to a process that is the working of an F-mechanism (the realisation of its F-potentiality) embodied by S.

So the wind dries the washing if and only if the washing being dry belongs to the working of a drying mechanism embodied by the wind. In the deviant causal chain example the washing being dry does not belong to the working of such a mechanism. In general the causal process model is immune from deviant causal chain counterexamples precisely because it does not employ the idea of a causal chain. However it does employ ideas that many modern philosophers of causation find deeply suspicious. First of all is the idea of a mechanism. A mechanism is identified by its potentiality. Here I am using the Aristotelian notion of a potentiality. Aristotle said: “Motion is the fulfilment of what exists potentially, in so far as it exists potentially.” (Physics, 201a10 – 11) So a process is the realization of a capacity or disposition for certain results in certain circumstances. You have to characterize a structure of stages to specify the potentiality. But what is required for the process to be happening is not just that that structure of stages is in train, but that there is a potentiality for such a structure and that this potentiality is being realised. There are two sides to something having a potentiality. There is the set of conditional statements that describe what the potentiality is a


Rowland Stout

potentiality for. And there is that underlying nature or set of conditions that grounds the potentiality.3 These latter conditions are divided into two categories—the conditions that constitute the potentiality or mechanism itself and the conditions that constitute the realisation of the potentiality or the operation of the mechanism. Usually this distinction is made in a pragmatic way, and for simplicity we can lump them all together as the underlying conditions of the process. The potentiality is characterized by describing a set of general conditional statements, knowledge of which consists in entitlement to make a set of inferences. Such inferences are not universally valid. You are entitled to make the inferences only when their underlying conditions are fully in place. In order to know that a certain potentiality is being realized and so be entitled to make the respective inferences, you do not have to know exactly what these underlying conditions are; you just have to know that they are satisfied, whatever they are. In this way the entitlement to make such inferences is grounded in the nature of the situation. Given this conception of a process, in order to characterize a causal process it is neither necessary nor sufficient to specify all the stages in a causal chain. Instead, one must describe the potentiality whose realization that process is. And to do this one must specify a structure of conditional statements. The difference between this approach and David Lewis’s (1973) nonAristotelian approach to causation should be clear. In both cases counterfactual conditionals are crucial. The difference is that, for Lewis, conditionals are truths about the world of all possible worlds, whereas, according to the view being recommended here, the general conditional statements constitute descriptions of processes. The truth of the conditional claim is localized to the situation in which the underlying nature of the process is present. So my suggestion is that the analysis of a causal notion should not be of the form: S F-s O if and only if some aspect of S causes O to have some quality characteristic of F. (For example, the wind dries the washing if and only if some aspect of the wind causes the washing to be dry.)

Instead it should be of the form: 3

Philosophers opposed to this talk of mechanisms and potentialities are also suspicious of this idea of a grounding relation, and certainly the Aristotelian approach owes a treatment of it, even though I will not attempt one here.

Two Ways to Understand Causality in Agency


S F-s O if and only if O having the F-quality belongs to a process that is the working of an F-mechanism (the realisation of its F-potentiality) embodied by S. (For example, the wind dries the washing if and only if the washing being dry belongs to a process that is the working of a drying mechanism embodied by the wind.)

“F” appears on both the left and the right hand side of this biconditional, which means that this is not a reductive account. But there is no vicious circularity here. The characteristic sensitivity that goes with a certain kind of thing can be established at least partially independently, and then we can establish on any particular occasion whether something belongs to a process that is the working of a mechanism with this sensitivity. For example, the law describing the mechanism of the wind drying washing will be something like this: if the washing is in the line of the wind for such and such a time, then it will dry. If the drying of the washing belongs to a process that is the working of a mechanism with this sensitivity, then we can say that the washing was dried by the wind. A quite different law describes the mechanism responsible for the washing drying as a result of the tumble dryer being turned on by the window being banged into it by the wind. It is not sufficient that the conditional statements describing such sensitivity should merely hold in the situation. The drying of the washing must belong to the process which is the realization of a potentiality that is described by those conditional statements. Put more simply, the process described by those conditional statements must result in the washing drying. This leaves some important work to be done in the philosophy of causation in explaining what it is for an event to belong to a process. And until this work is done the suspicion that something like deviant causal chain counterexamples may be found cannot be completely eradicated. But process causation is certainly no more mysterious than event causation. To explain what it is for one event to result from another is at least as difficult as explaining what it is for an event to belong to a process. Let us go back to agency. In the previous section I argued that we should start with trying to understand what it is for an agent intentionally to make an event happen or a state obtain. More generally, we should start with understanding what it is for an agent intentionally to cause things to be G, where G stands for the intransitive characterisation of what they are achieving. A typical way for the causal chain model to explain this is the standard causal theory of action, which works in the following sort of way, though of course there are many variations of this:


Rowland Stout

An agent intentionally causes things to be G if and only if the agent’s intending to cause things to be G causes things to be G.

The Aristotelian causal process model gives a different approach as follows: An agent intentionally causes things to be G if and only if things being G belongs to a process that is the working of an intentionally-causing-thingsto-be-G mechanism embodied by the agent.

This awkward phrase, “an intentionally-causing-things-to-be-G mechanism”, needs a bit of explanation. The point is that the laws that describe the potentiality of this mechanism must be characteristic of intentionally causing things to be G. What are such laws? How should we specify the potentiality that is characteristic of intentionally causing things to be a certain way? This is a central question of the philosophy of agency, and I will simply provide what I take to be the answer without any argument. I think the answer is that this involves sensitivity to reason—in particular teleological reasons. On the one hand, in doing things it seems we are changing the world; we are making new states obtain and events occur. On the other hand we do things for reasons. And very often these reasons are teleological; we do things for the sake of other things. So my proposal is that an intentionally-causing-things-to-be-G mechACHTUNGREanism is simply a mechanism that results in what should happen for things to be G. It results in what should happen according to some particular, and possibly flawed, version of practical rationality. For of course an agent may fail to achieve some goal while intentionally working towards it.4 So we get the following causal process theory of action: An agent intentionally causes things to be G if and only if things being G belongs to a process that is the working of a mechanism embodied by the agent which results in what should happen according to some version of practical rationality for things to be G.

For example, suppose an agent intentionally posts a letter. I argued in section 2 that this means the agent intentionally causes things to be characteristic of the activity of their posting a letter. This will include, perhaps among other things, causing the state of the letter being in the postal system to obtain. In the Aristotelian causal process model what it is 4

This idea of doing what you should do according to some version of practical rationality needs a lot more explanation. I will not say more about it here, but have tried to do so elsewhere—for instance in Stout 1996; 2006.

Two Ways to Understand Causality in Agency


for an agent intentionally to cause things to be characteristic of their posting a letter is for things being characteristic of their posting a letter to belong to a process that is the working of a mechanism that they embody that results in what should happen for such characteristic letter posting events/states to happen/obtain. When the way to get the letter posted is for the agent to walk to the post box the mechanism will result in the agent walking to the post box. If a mechanism with this sort of teleological sensitivity is what results in the letter getting into the postal system then we can say that the agent has intentionally posted the letter. An account of trying to achieve a goal, with or without actually achieving it, falls out naturally from the account: An agent is trying to cause things to be G if and only if a process is happening that is the working of a mechanism embodied by the agent which results in what should happen according to some version of practical rationality for things to be G.

Trying involves the same teleological mechanism operating as acting does; it just does not require that the achievement of the result belong to it. So if the agent walks to the post box as a result of the working of the mechanism that results in what should happen for the letter to get into the postal system then they are trying to post the letter. If the agent tries and fails then they are not ultimately doing what they should do in order to post the letter. But they are doing what they should do according to some flawed version of practical rationality—a version that characterises the mechanism that is working in this situation. The causal process theory of action can be ramified to handle the possibility raised in section 2 that certain sorts of activities like eating a banana or going for a walk, or indeed posting a letter, may not be very satisfactorily understood just in terms of making things happen. If indeed there is something more that is essential to these activities than simply causing a characteristic structure of results it is that these results are achieved in a characteristic way—by the operation of a special mechACHTUNGREanism. So let us assume that eating a banana is causing things to be characteristic of one’s eating a banana by means of the working of one’s eating mechanism. Then we can apply the general causal process model to the case in the following way:


Rowland Stout

An agent intentionally eats a banana if and only if things being characteristic of that agent’s eating a banana belongs to a process that is the working of an eating mechanism embodied by the agent which results in what should happen according to some version of practical rationality for things to be characteristic of their eating a banana.

4. Deviance and Responsibility For some time philosophers of action have been pushing their causal theories in this sort of direction. In 1964 Charles Taylor gave the following account of the teleological explanation of action: To offer a teleological explanation of some event or class of events, e.g., the behaviour of some being, is, then, to account for it by laws in terms of which an event’s occurring is held to be dependent on that event’s being required for some end. To say that the behaviour of a given system should be explained in terms of purpose, then, is, in part, to make an assertion about the form of laws, or the type of laws which hold of the system. (Taylor 1964, 9)

Adam Morton said: “intentional action is action that is guided by information to which it is responsive” (Morton 1975, 14). For many philosophers the problem of deviant causal chains has been what has motivated this sort of approach. Christopher Peacocke, responding to the problem of deviant causal chains, introduced the notion of differential explanation to attempt to make more precise the vague suggestion that “intentional behaviour is in some way characteristically sensitive to certain facts” (Peacocke 1979, 57). And David Lewis (1980) provided a similar resolution to the analogous problem of veridical hallucination that faces causal accounts of perception.5 Converging on what I take to be the same target, philosophers of action have also introduced the idea of proximate or sustained causation (see Brand 1984; Thalberg 1984). The rough idea here is that if there is any causal gap between cause and effect, then there is room for a deviant causal chain to be interpolated. Process causation is sustained causation. The results of a process happen while that process is still happening. It is much less easy to see how a chain of causal links may sustain its results, since the earlier links are finished by the time the later ones come along. 5

Other philosophers who have introduced guided control into their accounts of action include Davis 1970, 23; Thalberg 1984; Audi 1993; Bishop 1989.

Two Ways to Understand Causality in Agency


Harry Frankfurt argued most explicitly for the claim I am making here that action must be regarded as a sensitive guided process, and thus that the causal chain theory of action is mistaken. [T]he state of affairs while the movements [of a person’s body] are occurring is far more pertinent [than the causes from which they originated]. What is not merely pertinent but decisive, indeed, is to consider whether or not the movements as they occur are under the person’s guidance. It is this that determines whether he is performing an action. Moreover, the question of whether or not movements occur under a person’s guidance is not a matter of their antecedents. Events are caused to occur by preceding states of affairs, but an event cannot be guided through the course of its occurrence at a temporal distance. (Frankfurt 1978, 158)

For Frankfurt, what distinguishes guided action is the causal mechanism not the causal antecedents. Deviant causal chains are not a problem for this sort of account since action is not characterized in terms of a causal chain of any sort. There is no space in the process of a mechanism being realised in which a deviant causal chain can be interpolated. With the deviant causal chain in place the result belongs to the operation of a different mechanism—not a mechanism which works by making happen what should happen in order for some result to be achieved. The situation gets a bit complicated if the deviant causal chain is itself a manifestation of some other agency. Robert Audi (1993) considers deviant causal chains due to alien intermediaries. The example is of Tom, who intends to look at his watch to shorten a conversation. Ann, the alien intermediary, “likes to think she is making people do things that they would do anyway” (Audi 1993, 164). She presses her buttons and Tom, as a result, looks at his watch when and in the way he was intending to. He does not notice a thing. He is not intentionally looking at his watch; yet his behaviour does belong to the operation of a teleological mechanism—Ann’s mechanism. The problem here is not whether to attribute agency to the behaviour, but to whom the agency should be attributed. In the account as I have stated it the agency belongs to whoever embodies the teleological potentiality the realisation of which results in the behaviour. And I have left this phrase “embodies the potentiality” unexplained. Although Tom’s body is part of the mechanism in this case, it is not his mechanism, but Ann’s. Certainly this needs to be worked out, but it is a challenge rather than an objection to the causal process approach to agency.


Rowland Stout

In just the same way, Fischer and Ravizza (1998) have developed a causal process theory of responsibility with two elements. Firstly for an agent to be responsible for some behaviour that behaviour must result from the operation of a ‘reasons-responsive’ mechanism. Second, that mechanism must belong to the agent. Their motivation is not so much the problem of deviant causal chains but the related problem of alternative possibilities, something also dealt with by Harry Frankfurt, though some years earlier (Frankfurt 1969). Frankfurt argued that responsibility for action does not depend on the truth of some counterfactual conditional but rather on the nature of the actual process constituting the action. We can adapt Audi’s example above so that Ann only intervenes when Tom is not about to do what she wants him to do. Suppose that she has decided Tom will look at his watch whether he is inclined to do so or not. If he is not about to decide to look at his watch, she will press the levers in his brain and he will as a result decide to look at his watch. But if he decides to do this for himself she will do nothing. In this latter case, according to Frankfurt, we attribute responsibility to Tom, even though in some sense he could not have done otherwise. Fischer and Ravizza explain this by saying that in this latter case Tom’s behaviour issues from a reasons-responsive mechanism that is Tom’s own. What matters is the nature of the actual process resulting in Tom’s behaviour, not the existence or otherwise of alternative possible processes resulting in different behaviour. So Fischer and Ravizza are proposing something like a causal process model. But this proposal is treated as a fix within the philosophy of action to the particular issue of responsibility. They presume that we might have a notion of agency that does not employ this model which can then be beefed up by introducing the model to provide an account of really responsible agency. I suggest instead that once we start thinking of agency in causal terms at all the causal process model should be in place. Bibliography Anscombe, G. E. M. (1957): Intention, Oxford: Blackwell. Audi, Robert (1993): Action, Intention and Reason, Ithaca NY: Cornell University Press. Bishop, John (1989): Natural Agency, Cambridge: Cambridge University Press. Brand, Myles (1984): Intending and Acting, Cambridge MA: MIT Press.

Two Ways to Understand Causality in Agency


Davis, Lawrence (1970): Theory of Action, Englewoood Cliffs: Prentice-Hall. Fischer, John/Ravizza, Mark (1998): Responsibility and Control, Cambridge: Cambridge University Press. Frankfurt, Harry (1969): “Alternate Possibilities and Moral Responsibility”, in: Journal of Philosophy 66, 829 – 839. — (1978): “The Problem of Action”, in: American Philosophical Quarterly 15, 157 – 162. Harr, Romano/Madden, Edward H. (1975): Causal Powers, Oxford: Blackwell. Hornsby, Jennifer (1980): Actions, London: Routledge. Lewis, David (1973): “Causation”, in: Journal of Philosophy 70, 556 – 567. — (1980): “Veridical Hallucination and Prosthetic Vision”, in: Australasian Journal of Philosophy 58, 239 – 249. Machamer, Peter et al. (2000): “Thinking About Mechanism”, in: Philosophy of Science 67, 1 – 25. Melden, A. I. (1961): Free Action, London: Routledge. Morton, Adam (1975): “Because He Thought He Had Insulted Him”, in: Journal of Philosophy 72, 5 – 15. Mumford, Stephen (1998): Dispositions, Oxford: Oxford University Press. Putnam, Hilary (1967): “Psychological Predicates”, in: Capitan, William H./ Merrill, Daniel D. (eds.): Art, Mind, and Religion, Pittsburgh PA: Pittsburgh University Press, 37 – 48. Ryle, Gilbert (1949): The Concept of Mind, London: Hutchinson. Stout, Rowland (1996): Things That Happen Because They Should, Oxford: Oxford University Press. — (2002): “The Right Structure For a Causal Theory of Action”, in: Facta Philosophica 4 (1), 11 – 24. — (2004): “Internalising Practical Reasons”, in: Proceedings of the Aristotelian Society 44, 229 – 243. — (2005): Action, Teddington: Acumen. — (2006): The Inner Life of a Rational Agent, Edinburgh: Edinburgh University Press. Stoutland, Frederick (2007), “Reasons for Action and Psychological States” (this volume). Taylor, Charles (1964): The Explanation of Behaviour, London: Routledge. Thalberg, Irvin (1967): “Do We Cause Our Own Actions?”, in: Analysis 27, 196 – 201. — (1984): “Do Our Intentions Cause Our Intentional Actions?”, in: American Philosophical Quarterly 21, 249 – 260.

On the Importance of Reconciling Two Models of Causality in Agency John Bishop 1. Introduction: A Causal Process Model of Agency The explanation of intentional action in terms of an agent’s reasons is clearly not a case of causal explanation according to any Humean or covering-law model of causal explanation. Nevertheless, an intentional explanation (as we may call it) fails unless it claims, and claims truly, that the behaviour constituting the agent’s intentional action occurred because the agent had the reasons cited in the explanation. Without such a— broadly speaking, causal—clause, the attempted explanation can do no more than show the agent’s behaviour reasonable in the light of his or her reasons for action. With these Davidsonian insights taken as given, the question then arises how we are to understand the broadly causal claim implied in any intentional explanation.1 Rowland Stout advocates a causal approach to agency that implicitly takes on board these Davidsonian insights. Stout maintains, however, that the right way to understand causality in agency is by means of “an Aristotelian model of causal processes”. On this model, causes yield effects through the operation of a causal mechanism with the potentiality to produce the kind of effect concerned—and the causal process through which the effect results from the cause is a particular realisation of that potentiality. We humans possess causal mechanisms “that constitute our nature as both subjects and agents” (140).2 Intentional agency may thus be understood, Stout proposes, by means of the following schema: An agent intentionally causes things to be G if and only if things being G belongs to a process that is the working of an intentionally-causing-thingsto-be-G mechanism embodied by the agent. (148)

This is to say no more than that an agent’s intentionally bringing G about is the actualisation of a causal mechanism that constitutes the agent’s power as an agent to bring about G. And, of course, the issue now is to answer the question: what kind of a causal mechanism is that? Amongst 1 2

These insights were first set out and defended in Davidson 1963. Parenthetical references in the text are to Stout 2007 (in this volume).

Reconciling Two Models of Causality in Agency


all the causal mechanisms embodied in agents, how are we to distinguish that mechanism whose operation realises the capacity of agency with respect to whether or not G results? Stout does elaborate his initial schema, by incorporating the requirement that a mechanism that realises intentional agency must involve “sensitivity to reason—in particular teleological reasons” (148). This, he suggests, yields the following: An agent intentionally causes things to be G if and only if things being G belongs to a process that is the working of a mechanism embodied by the agent which results in what should happen according to some version of practical rationality for things to be G. (148)

The point of Stout’s reference to “some version of ” practical rationality is not entirely clear: does it mean that the operation of the mechanism is a process of practical reasoning involving the particular intentions and beliefs the agent has in the circumstances concerned? “What should happen for things to be G” will indeed vary depending on the agent’s prior intentions and salient beliefs, and the reference to “some version of practical rationality” may be intended to capture this point. In any case, however, the question how to distinguish a mechanism of the kind that realises genuine intentional agency still remains—as Stout’s discussion of deviance in section 4 of his paper makes clear. 2. The Problem of Causal Deviance In that section, Stout recognises the difficulty that there can be ‘deviant’ causal mechanisms embodied in an agent (at least, for a short time) that do yield behaviour for which the agent has good reason yet do not constitute a power of intentional agency. To return to Stout’s schema: an embodied mechanism could “result in what should happen for things to be G” given the norms of practical rationality as applied to the agent’s desires and beliefs and yet clearly not constitute the agent’s power of intentional agency over whether G does or does not obtain. (For example, Davidson’s alpinist embodies a causal mechanism that results in just what needs to happen for him to let go of the rope given that he desires to be free of the weight and danger of holding his colleague and believes that all he has to do to satisfy this desire is let go. Yet, since that causal mechanism involves the triggering of nervousness sufficient to make the


John Bishop on Rowland Stout

alpinist lose his grip, it is a mechanism that undermines his power of intentional agency rather than realises it (see Davidson 1973, 79).) Stout reasonably concludes that, to constitute a genuine power of intentional agency, a causal mechanism needs to satisfy some condition that meets Adam Morton’s requirement that intentional action be “guided by information to which it is responsive” (150). The condition Stout favours is the requirement that a causal mechanism that constitutes genuine intentional agency should realise a kind of ‘sustained causation’ in which the agent’s guiding intention remains in control of the outcome throughout the process.3 In addition, Stout notes (151) that it will be necessary to exclude causal mechanisms where one agent’s sustained causation of his own behavioural outcome is achieved only through the intervening agency of another agent—although finding the right excluding condition here will require some subtlety, since it is imaginable that a second agent might sometimes function as a prosthetic aid to preserve the first agent’s power of sustained agent-control.4 3. Causal Process versus Causal Chain Models of Agency Stout presents this Aristotelian causal process view of agency as preferable to what he takes to be the prevailing ‘causal chain’ account to be found in causal theories of action. These accounts depict an agent’s intentional action as (in some sense in need of further specification) reducible to, or constituted by, a causal relation linking certain of the agent’s mental states, processes and events to the behaviour intrinsic to the kind of intentional action concerned. Such accounts of fully fledged, deliberate, intentional action typically involve a linked set of states and events: for example, an initial intention to achieve a certain goal that initiates practical reasoning yielding a judgment, made in the light of salient beliefs, desires and further intentions, that such-and-such a ‘basic’ action is, all things considered, the best action to perform in the given circumstances—a judgment that itself produces a ‘proximate’ intention to perform such an 3


The term “sustained causation” is due to Irving Thalberg: “A full-blown causal theory [should] prescribe a tighter hookup [between intention and behaviour]— what I call ‘ongoing’, ‘continuous’ or ‘sustained’ causation. If someone’s behaviour is to count as intentional Xing, I believe her or his intention must continuously regulate the Xing.” (Thalberg 1984, 257) I have discussed this possibility at some length in Bishop 1989, 155 – 64.

Reconciling Two Models of Causality in Agency


action which then, finally, causes the agent to behave in the way required for the action concerned.5 To meet the deviance problem, of course, these accounts will need to specify that each step in the causal chain is the right kind of, non-deviant, causation: and there are well known strategies for achieving this (such as those of Peacocke and Lewis referred to by Stout (150)), although it remains controversial whether these strategies are fully successful.6 The claim that the ‘causal process’ understanding of causality in agency is to be preferred to this ‘causal chain’ account ought not, I think, to be at all controversial, provided it is construed as a claim about the meaning of causality in agency, or (in other words) about how our concept of agent causality should be defined. Is it not obvious that what it means to say that an agent intentionally makes happen a certain behavioural event or structured set of events is not reducible to saying that certain of the agent’s mental events, states and processes cause that event or set of events to come about?7 To act is more than to be caused to behave, even by one’s own ‘rationalising’ mental states. Each link of the causal chain in the account sketched above needs to be understood as something the agent does, and not merely as something that is caused to happen to the agent. If I intend to get some water, then I must engage in the reasoning that (given my beliefs) yields my judgment that, all things considered, the thing to do is start digging under this palm tree, and I must then intend to start digging, and, finally, I must carry out that intention. A mere chain of events in which a prior intention causes an allthings-considered judgment which then causes a proximate intention that finally causes my digging behaviour lacks something conceptually essential to the intentional action of digging.8 5 6 7


As I use the term here ‘basic’ actions are those the agent can perform directly, without the need to work out the means to their achievement: no ‘foundationalist’ theory of action need be implied by this notion. For a recent critical discussion of attempts to solve the problem of causal deviance faced by causal theories of action see EnÅ 2003, chap. 4. I describe intentional agency in terms of an agent’s making happen a certain behavioural event or structured set of events in order to acknowledge Stout’s useful remarks (142 – 144) about the need to expand the usual paradigm of intentional agency from raising one’s arm (making one’s arm go up) to activities such as eating a banana or going for a walk. Compare David Velleman: “(R)easons cause an intention, and an intention causes bodily movements, but [on this account] nobody—that is no person— does anything.” (Velleman 1992, 461)


John Bishop on Rowland Stout

4. Agent-Causation and the Causal Process Model Understood in this way—as a claim about the content of the concept of agent causality—Stout’s claim is hardly contentious. The Aristotelian causal process view preserves, and, indeed, emphasises, the notion that the agent brings about the event intrinsic to the intentional action, whereas the causal chain view leaves that notion out altogether. One typical way to put this point is by insisting that our concept of intentional agency involves a notion of agent-causation that is conceptually irreducible to the notion of some specific type of chain of event-causation. Conceptually, agent-causation is a relation between a person or agent (a personal substance) and an event or structured set of events. Agent-causation is thus categorially distinct from event-causation, which relates events or states with other events or states. Surprisingly, Stout makes no mention of agent-causation in his paper. Yet surely he ought to identify those who affirm the conceptual irreducibility of agent-causation as very much on his side? For, the claim that agent-causation is conceptually irreducible is effectively equivalent to the rejection of the event-causal chain model as adequate to our concept of intentional agency. And Stout is concerned to reject that model, while also insisting that rejecting that model is by no means to abandon the view that the notion of agency essentially involves at its core a certain concept of causality, nor the view that intentional explanation holds only when a certain kind of causal claim is true. (Here I note that Stout makes the valuable point that some philosophers who present themselves as anticausalists about action are in fact rejecting only the causal chain model (140 f.).9 The view that one ought to be some kind of causalist about intentional action is, of course, implied by the Davidsonian insights with which I began.) 5. Causal Chain Models and the Ontic Constitution of Agency Since it is readily apparent that an event-causal model cannot adequately define our concept of what it is for an agent to perform an intentional action, why have some philosophers advocated such a model? Stout does not ask this question. And so he does not consider the important possibility that event-causal models of intentional agency may be 9

The philosophers Stout mentions are Ryle, Melden and Anscombe.

Reconciling Two Models of Causality in Agency


advocated, not as conceptual analysis, but rather as providing an account of what it is for intentional actions to be ontologically realised in the natural world. The preferability of a causal process model as a conceptual analysis of intentional agency thus at least leaves it open that an eventcausal chain model might be preferable as an ontological account. In a footnote, Stout notes that the Aristotelian approach “owes a treatment of ” what he calls “the grounding relation” (see 146, footnote 3). The point here is that, for some causal mechanism to have a certain potentiality, there must be some “underlying nature or set of conditions that grounds the potentiality” (146). This is to say, I think, that a mechanism can have its characteristic potentiality only in virtue of some actual properties—and some explanation must at least in principle be possible of how it is that a mechanism whose actual properties are thus and so should realise this characteristic potentiality.10 This is the closest Stout gets, I think, to recognising that there is an ontological issue about how intentional agency can be realised in the actual world. There is an issue about what properties a mechanism needs to have in order to possess an intentionally-causing-things-to-be-G potentiality, and there is a (closely related) issue about what goes on in the world when such a potentiality is realised in a particular case. My suggestion, then, is that event-causal, ‘causal chain’, accounts of intentional agency may be aimed at dealing with this ontological issue, and so may not compete with the Aristotelian causal process model urged by Stout, nor the agentcausationist analysis of the concept of intentional agency which (I have suggested) is effectively equivalent to the Aristotelian account. A central aim of Stout’s paper, then, may be misguided: the debate about the nature of intentional agency, it may transpire, is not to be “set out clearly (…) as a debate between two alternative accounts of causation” (140). There are, indeed, two distinct accounts, but they may both apply—one at the level of defining our notion of intentional agency, and the other at the level of explaining what it is for intentional action to be ontically realised. 10 Compare the account of explanatory appeals in science given by Machamer, Darden and Craver in an article to which Stout refers in passing (140). This account is ‘dualist’ in the (special, technical, sense) that it maintains that an adequate description of a mechanism must be in terms of both entities and the activities in which those entities are involved. On this account, mechanical explanation (explanation in terms of mechanisms) achieves an intelligibility that “consists in the mechanisms being portrayed in terms of a field’s bottom out entities and activities” (Machamer et al. 2000, 21).


John Bishop on Rowland Stout

It might be argued, of course, that once we concede that the causal process model is conceptually preferable, we should accept—at least prima facie—that it is also ontologically preferable. If, however, the causal process model involves conceptually irreducible agent-causation (as I have suggested), then construing it ontologically obviously brings commitment to ontologically irreducible agent-causation. Such a commitment, however, is inconsistent with what is widely thought to be the most plausible naturalist metaphysics—the metaphysics, that is, that best fits the implicit presuppositions of natural scientific theories. Though the coutenancing of irreducible substance-causation might theoretically be consistent with naturalism, naturalist metaphysics has in fact developed in quite the opposite direction. Irreducible substance-causation and primitive substance-causal mechanisms have been rejected.11 And so it is salient to consider whether, if naturalist metaphysics does indeed have no place for irreducible agent-causation, it may nevertheless be reasonable to accept that intentional agency can be ontically realised in the natural world, even though it cannot conceptually be understood for what it is without appeal to irreducible agent-causation. Otherwise, there will be a serious clash between the metaphysics we need for naturalism and the metaphysics we need to ground intentional—and hence morally responsible—agency. To be reassured that there is no such clash, and no grounds for the skepticism about responsible agency that would threaten if there were, we will need to appreciate how an event-causal ontological reduction of conceptually irreducible agent-causation can indeed be a possibility. And it is the vindication of that possibility, I suggest, at which eventcausal theories of intentional action aim. What they seek to show is that, whenever there occurs an intentional action, that intentional action is wholly realised by a causally related set of states and events. Arguably, this 11 In “Thinking about Mechanisms”, Machamer, Darden and Craver make it clear that any given mechanical explanation in the sciences often (perhaps always?) simply presupposes a “bottoming out” level of entities and activities (Machamer et al. 2000, 13). Higher-level mechanisms, in other words, are rendered intelligible by assuming the intelligibility of the operation of lower-level mechanisms out of which they are constituted. This does not entail, however, that a contemporary mechanical scientific worldview is committed to a metaphysics of irreducible mechanisms: indeed Machamer, Darden and Craver themselves take the ‘reductionist’ view that “objects simpliciter … may be said to be causes only in a derivative sense. It is not the penicillin that causes the pneumonia to disappear, but what the penicillin does.” (ibid., 6)

Reconciling Two Models of Causality in Agency


can be shown only by defending the claim that certain conditions stateable purely in event-causal terms are necessary and sufficient for the occurrence of an intentional action—or, in other words, that we rightly attribute an event or activity to an agent when and only when that event or activity is caused in the right sort of way by states, events and processes that belong to the agent.12 6. Conclusion: Reconciling the Two Models Stout presents two ways of understanding causality in agency as exclusive alternatives. I have argued, however, that they need not be so regarded— except when they are put forward as competing definitions of what it is for agents intentionally to bring about events (and then a causal process or agent-causationist view seems clearly preferable). The event-causal, or causal chain, understanding of agency comes into its own in the context of attempting to meet the sceptical challenge posed by agent-causationists who insist that agent-causation is an ontological, as well as a conceptual, primitive. Such insistence leads to scepticism about the very possibility of intentional agency in the natural world only on the assumption, of course, that a naturalist metaphysics has no place for ontologically irreducible causing by agents (or, indeed, by substances generally). But that assumption is sufficiently well founded for it to be important to attempt to defend an event-causal theory of intentional action, in the service of a reconciliatory monist naturalism that would preserve the fundamental commitments of our ethical and common sense understanding of ourselves as responsible intentional agents. Bibliography Bishop, John (1989): Natural Agency: an Essay on the Causal Theory of Action, Cambridge: Cambridge University Press. Davidson, Donald (1963): “Actions, Reasons and Causes”, reprinted in: Davidson (1980). — (1973): “Freedom to Act”, reprinted in: Davidson (1980). — (1980): Essays on Actions and Events, Oxford: Oxford University Press. 12 I attempt a defence of an event-causal theory of action along these lines in Bishop 1989. It remains controversial, of course, whether any such attempt succeeds.


John Bishop on Rowland Stout

EnÅ, Berent (2003): How We Act: Causes, Reasons and Intentions, Oxford: Clarendon Press. Machamer, Peter/Darden, Lindley/Craver, Carl F. (2000): “Thinking about Mechanisms”, in: Philosophy of Science 67, 1 – 25. Stout, Rowland (2007): “Two Ways to Understand Causality in Agency” (this volume). Thalberg, Irving (1984): “Do Our Intentions Cause Our Intentional Actions?”, in: American Philosophical Quarterly 21, 249 – 260. Velleman, J. David (1992): “What Happens When Someone Acts?”, in: Mind 101, 461 – 481.

Actions, Habits, and Practices Todd Lekan 1. Introduction For all their differences, classical pragmatists such as Peirce, Dewey, and James, seem to agree that much previous philosophy has erred by isolating its methodology and its results from action or practice. Thus, at least initially, it is strange to ask the question “What is the pragmatist theory of action?” since action is such a ubiquitous notion in pragmatism. Pragmatism is foremost a meta-philosophical thesis about the nature of philosophical theorizing, enjoining us to view even the most recondite and abstract activity of philosophers as modes of human practice. Philosophical reflection is not different in kind from the sort of intelligent problem-solving that we find in practical domains such as medicine, scientific inquiry, music, engineering, and cooking. Bodies of practical knowledge, and procedures for the appropriate application of such knowledge, arise from these domains as participants work to resolve novel problems and creatively search out new possibilities. Thus, practice has a logical and temporal priority to any theorizing. Theorizing is a mode of practice—practice made intelligent. The purpose of theorizing is to improve practice. Practices may be improved in a variety of ways: morally, aesthetically, scientifically, religiously (among others). A comprehensive philosophical account of ‘action’ must be sensitive to the differences among human practices. However, notwithstanding a reputation for being ‘anti-theoretical’ or ‘anti-intellectual’, pragmatist philosophy does not abandon the goal of offering general accounts of notions such as ‘action’ which apply across various domains of human life. What distinguishes pragmatism is its insistence that such philosophical generalizations are modes of practice to be judged by how well they improve the very practices from which they emerge. I cannot defend this controversial meta-philosophical position here, but it is important to keep it in mind as we explore the main outlines of a pragmatist theory of action.


Todd Lekan

Many philosophical accounts of action begin with an analysis of the concept of a reason for action. The debates focus on questions such as whether reasons for action must be analyzed into beliefs and desires or whether belief alone is motivationally adequate; whether reasons can be causes of action; or whether acting for a reason involves an implicit normative commitment to endorsing the reason as good for anyone similarly situated. As important as these debates may be, we should pause before we just assume that the best way to begin to study action philosophically is by inquiring into the nature of reasons for action. For pragmatists, the practice of reason-giving is situated inside other human activities and practices which are themselves made possible by practical knowledge embodied in socially shared habits. To begin asking “What is a reason?” or “What makes some reasons better than others?” is to run the risk of forgetting that reason-giving occurs in the context of human practices that have become questionable in some specific respect, and that the people who raise questions about reasons are only able to do so because they are equipped with practical knowledge pertaining to such practices. It is better, claims the pragmatist, to begin by inquiring into the conditions that make possible the practical knowledge necessary for intentional action. If our philosophical aims had little or nothing to do with the improvement of human practices, it might make sense to inquire into the nature of reasons or justification, consigning any discussion about the concrete socio-psychological conditions of human action to the ‘descriptive sciences’. Yet, adoption of a pragmatic account of human action will necessarily begin its inquiry by exploring the social and psychological contexts that make action possible precisely because it is in these contexts that justificatory questions about good and bad reasons are raised. In what follows, I argue that intentional action depends on a practical knowledge best understood in terms of socially shared habits. The account of habits is based on ideas that John Dewey develops in his Human Nature and Conduct. I begin by sketching a brief account of skilled action, focusing on basic features of skilled action including knowhow, novelty, and complexity. I then offer a brief outline of a Deweyan account of habits that helps explain the nature of practical knowledge. I conclude with some thoughts about the normative implications of a habit-based account of practical knowledge and intentional action.1 1

My account in this paper is based largely on the pragmatism theory offered in Lekan 2003, esp. chap. 1 and 4.

Actions, Habits, and Practices


2. Practical Knowledge and Intentional Action Actions might be divided into discrete units of behavior by reflective analysis. But the experience of action is one of a seamless flow of activity that involves spontaneous responses to a variety of social and personal contexts. When something goes wrong, or when we wish to learn about ourselves or others, we might individuate actions into discrete bits of behavior interpreted by reference to the goals or purposes that provide reasons for these actions. But when the wheels of practices and activities run smoothly, appropriate actions simply happen. Flipping light switches upon entry to rooms, the circular motions performed while brushing teeth, genuflecting upon entrance to churches, riffing the right blues scales while playing jazz, may appear to be bits of disconnected behavior but in fact they are manifestations of a conflux of sensibilities, attitudes, abilities—ways of responding that have been honed over time. Such actions are the fruit of a practical knowledge that is not best thought of primarily in terms of beliefs about rules or reasons. This knowledge is embodied in habits. Although skilled expertise is only one species of intentional action, it will be useful to begin with a brief phenomenology of it.2 Doing so reveals important dimensions of practical knowledge that might otherwise be ignored. I want to focus, in particular, on three features of practical knowledge well illustrated in skilled expertise: know-how, novelty, and complexity. Know-How and Seeing-As People acquire skilled expertise by participating in structured practices and activities. Through repeated participation in such practices, people internalize norms of appropriate conduct. Cooking a gourmet meal well requires years of training in specific practical contexts. The display of skilled expertise requires interaction with practical contexts that consist of tools, materials, machines, and institutional arrangements. In these contexts the expert cook sees ingredients as potential constituents of dishes. She sees tools as implements for various tasks. The novice, when confronted with the same kitchen as the expert, does not see the situation in the same way because she lacks the relevant skills and practical knowledge. In short, skilled-expertise involves skilled know-how that is 2

For a similar account see Dreyfus/Dreyfus 1990.


Todd Lekan

not primarily represented as beliefs or judgments about rules of appropriate conduct. Novelty The practice of cooking has a history that precedes the particular projects of particular cooks. Over time, people modify practices and activities as they struggle to solve novel problems. Practices are continually modified through trial and experimentation. Experts are capable of modifying practices in significant ways unavailable to novices, except by accident. Such intentional modification of practices is also impacted indirectly by other changes in the conditions under which the practice is pursued. For example, a practice like cooking is shaped by expert innovation as well as changes in conditions including the invention of electric ovens and food processors, the importation of spices, fruits and vegetables. Complexity Consider the complex circumstances that enter into the skilled pursuit of cooking: What ingredient may be substituted for this dish? What sorts of people frequent this restaurant? What kinds of cooking equipment are available? The endless permutation of circumstances relevant to the expert’s responses does not imply that these responses are arbitrary and without reason. An expert’s responses to such situations may involve explicit reflection on rules or reasons for actions when things go wrong, or when she is involved in training novices. But it is highly implausible to maintain that experts perform actions through explicit consultation of rules, principles, or reasons when performances are flowing well. The Novice Three features of skilled expertise have been identified: know-how, novelty, and complexity. My claims about the centrality of ‘know-how’, ‘novelty’ and ‘complexity’ are based on a very brief phenomenological description of what it is like to perform actions with skill. Now consider the novice. His or her actions are much less fluid than the expert’s because the novice lacks the requisite know-how. Novices need to engage in much more reflective deliberation and analysis about actions because they lack the skills to see contexts as requiring the appropriate responses. Novices typically do not solve novel problems in ways that significantly modify practices. Practices themselves are novel situations for novices. With respect to the complexity of practices, the basic skills, tasks, tools, and materials are experienced by the novice as bewildering array of foreign

Actions, Habits, and Practices


material. Beliefs about rules are usually indispensable for the beginner. Novices use rules as guides for learning about important elements of practices. Assimilation of rules is often a necessary condition for the development of skill, but not always. Language is a good case. One learns the rules of grammar long after gaining a solid grasp of the language. A good many actions, to be sure, are performed spontaneously without explicit reflection on rules or reasons, but which hardly deserve to be called ‘skilled expertise’. Many simple actions like flipping light switches, waving hello, opening a window, or tossing a yo-yo do not appear to involve complex abilities or creative potentials. But the appearances here are misleading. Simple actions seem different from skilled expertise conducted in structured practices because we take for granted the background of even the simplest actions. Turning off a light switch may be further described in any number of ways, depending on the context. It may be “the first step in preparation for love-making”, “closing the act of a play”, “getting ready for bed”, “saving electricity”, and so forth. The description may not go much beyond “this is the sort of thing a person does when entering an exiting room in her house”. (Even this description will be part of a network of assumptions about what she does in her house). Usually, the background of actions may be left unarticulated. When something goes wrong, we bring this background into view. We may be interested in why a light switch was flipped when, for example, a stagehand misses a cue during a performance of a play. It is then appropriate to ask who flipped the switch and what the person doing so thought he was doing. This type of analysis helps us to become aware of the complex background assumptions needed to comprehend the meanings of bare intentional actions and needed to determine whether an action counts as intentional in the first place. Thus, it seems that the practical knowledge of even simple intentional actions outside structured practices displays some degree of complexity. 3. Habits What is the best explanation of the practical knowledge that figures in intentional actions? Beliefs, desires, and reasons are notions that are important for understanding intentional action. However, the pragmatist claims that these notions only make sense over and against a background of practical knowledge that could be described as a know-how—an ability


Todd Lekan

to react appropriately. We need an interpretation of the nature of this practical knowledge that sheds light on its function. The pragmatist hypothesis that I offer, following Dewey, claims that practical knowledge is embodied in socially shared habits that function together in a holistic context. Habits have diachronic and synchronic properties. I distinguish three dimensions of habits. The first, interpenetration, displays diachronic and synchronic properties. The second, functional interaction is synchronic. The third, historical transmission is diachronic. I do not regard these three dimensions of habits as ontologically separable. However, for purposes of analysis, it is helpful to divide habits in these ways. Interpenetration of Habits Dewey’s Human Nature and Conduct offers his most worked out account of habits. Not to be thought of as mechanical, rote responses, habits for Dewey are dynamic response patterns that help to focus behavior and attention. Habits are concrete ways of thinking, feeling, and doing. A habit can reside in many activities and practices. Its character and meaning can only be determined in light of its role in a practice or activity and in its connection to other habits. Consider the range of qualities that Dewey would characterize as habits: impatience as a personality trait; a tendency to hard drinking after hours; skill in manipulating surgical instruments; an exemplary singing voice; class prejudice; loyalty to the Roman Catholic church; commitment to truthtelling; the ability to speak several languages fluently. The term “habit” may easily mislead us because we associate it with specific, fixed patterns of behavior isolated from each other. However, a habit does not function in isolation from other habits. When we make a list of habits, we run the risk of distorting this holist dimension of habits. Practical knowledge is not primarily the application of isolated rules to instances but the simultaneous mobilization of concrete ways of thinking, feeling, and doing. Although he uses the term “practices” Frederick Will’s observations readily apply here if we substitute the term “habit” for “practice”. Will says, One who learns a practice as if he were a soldier being imprinted with elements of close-order drill learns to behave in a way so odd and eccentric that Bergson could make it an important element in his theory of laughter. One essential thing missing from such learning is an understanding of how the more obvious, immediate responses that exemplify practices are determined in character by features of the conditions of performance,

Actions, Habits, and Practices


including therewith other dispositions to act, other forms of procedure, in short, other practices. This means that in following practices one is performing in a way that is governed in a great degree and in a highly complex way … . Practices do, so to speak confront circumstances, but not as individuals. Rather, in the metaphor employed by W. V. Quine in speaking of the coordination of statements with sense experience, they do so as corporate bodies. (Will 1997, 69 f.)

Will points to a certain holism of practices or habits at work as people learn appropriate responses. Some philosophers of science adopt a similar holism in understanding how scientific hypotheses are properly tested. The idea, roughly, is that we cannot assess the acceptability of a hypothesis simply in terms of whatever tests would falsify the hypothesis. A hypothesis might be confirmed by an experimental test but could still be acceptable in light of sufficient revisions of other auxiliary assumptions associated with the theory in question. Similarly, one might say that the appropriateness of an action can be adequately determined only through the application of a host of interlocking habits. Let’s call this feature of habits ‘interpenetration’. Interpenetration is both a synchronic and diachronic property of habits. It refers to the fact that a habit always functions with other habits in a holistic context. A singing habit, for instance may synchronically express itself with other habits such as participation in religious services, social relations in a choir, and so forth. Such habits develop over time, diachronically, from childhood. Habits and Transactions Transaction with environments is a synchronic feature of habits. A person’s habits are in continuous transaction with multilayered environments. Dewy compares this transactional aspect of habits to physiological functions: Habits may profitably be compared to physiological functions, like breathing, digesting. The latter are, to be sure, involuntary, while habits are acquired. But important as this difference is for many purposes it should not conceal the fact that habits are functions in many respects, and especially in requiring the cooperation of organism and environment. (Dewey 1983, 15)

The physiological function of breathing requires both lung and air. But air and lung are not externally connected; rather, they are functions that mutually modify each other. Socially learned habits like house building depend on both physical and cultural environments. The kinds of houses we build depend on the available materials, the type of physical


Todd Lekan

environment, the level of technology of a culture and so on. As houses are built, the physical and cultural environment is transformed, thereby transforming socially shared habits connected to domestic life, work, and travel. Habits as Historical Transmissions Historical transmission is a diachronic property of habits. A habit has a particular history in the life of a person as well as in the lives of generations of people. Both the practices of the Roman Catholic Church qua institution, and the vocational pursuit of a priest qua person result from historically transmitted habits. Most habits are acquired early in life and develop over time. Many habits, like pitching baseball, are prone to degeneration well before our lives are spent. Although habits are historically transmitted in the lives of individuals, they reside in evolving practices and institutions whether these are formal such as a church or informal such as family traditions. Dewey writes, “Individuals flourish and wither away like the grass of the fields. But the fruits of their work endure and make possible the development of further activities having fuller significance.” (Dewey 1983, 19) The fact that habits are historically transmitted does not imply that we are passive recipients of the transmission process. As we adapt habits to new circumstances we are capable of transforming them to some degree. One need only explore the history of an art such as jazz improvisation in order to appreciate the ways in which transmitted habits are modified during the course of their use. Understanding the fact that habits are socially shared historical transmissions helps to see that they are not simply physical aspects of behavior. Many habits embody deep-seated commitments, ideals, and aspirations. Patriotism, religious zeal, and fidelity to a culture represent embedded patterns of evaluation and belief passed on from generation to generation. 4. Habits and Motivation My claim, so far, is that the practical knowledge that figures in intentional action is embodied in habits with three dimensions: interpenetration, transaction with environments, and historical transmission. Consider the implications of this account of practical knowledge for understanding ‘motivation’. Habits are relatively organized systems of impulsive energy. Part of the meaning of the claim that habits interpenetrate is that actions

Actions, Habits, and Practices


are manifestations of a calibration of multiple elements in a person’s motivational economy. This balance can occur in the simple ability to carry on a conversation while eating dinner with appropriate table manners, or it can manifest itself in more general ways such as person’s ability to play multiple social roles with friends, family members, work colleagues, and so forth. It is misleading to think of these many calibrated actions as caused by desires and beliefs. If my conversation partner makes what appears to be an insulting joke at the dinner table, I may get angry. At that point, I might reflect on my beliefs and desires. Which belief is correct? Did he mean to insult me or was this a harmless ironic jest? I become aware of many desires: a desire to restore wounded pride by a quick verbal riposte, an aversion to creating conflict, and a desire to just get through with dinner so that I can do other things this evening. The desires and beliefs that are salient for me in this situation are the results of acquired habits. I form just these beliefs and desires because I am a certain kind of person constituted by a unique calibration of habits keyed to respond to situations in just this sort of way. Dewey writes, The essence of habit is an acquired predisposition to ways or modes of response, not to particular acts except as, under special conditions, these express a way of behaving. Habit means special sensitiveness or accessibility to certain classes of stimuli, standing predilections and aversions, rather than bare recurrence of specific acts. It means will. (Dewey 1983, 32)

Habits are propulsive—they are always operative even when they are not overtly expressed in specific actions. Even my ability to evaluate my beliefs and desires in order to control my actions is itself the product of propulsive habits. If I suppress my desire to verbally swipe back when you make a poor joke, because I believe it is not worth the potential conflict, it is because I have acquired second-order evaluative habits. The Deweyan account does not imply that we are prisoners of our habits. They are not external forces that govern mind and will, rather they constitute mind and will. However, this account does complicate questions pertaining to responsibility and change of habits. Dewey remarks that even though many have surrendered magical beliefs that wishes can control nature, they still believe that simply wishing for an end of action can bring the latter about. Dewey writes, … Formation of ideas as well as their execution depends upon habit. If we could form a correct idea without a correct habit, then we possibly could carry it out irrespective of habit. But a wish gets definite form only in connection with an idea, and idea gets shape and consistency only when it has a habit back of it … . The act must come before the thought, and a habit


Todd Lekan

before an ability to evoke the thought at will. Ordinary psychology reverses the actual state of affairs. (Dewey 1983, 25)

Consider the alcoholic struggling to give up drinking. Habits transact with specific circumstances. The alcoholic needs to do more than just wish for a future in which he no longer drinks. One important step is to enter new circumstances—to abandon, at least for a while, those friends, family members, and social situations that transact with his drinking habits. Since habits interpenetrate with each other, he may return to those old friends after he has established other ways of relating to people acquired, perhaps, in a half-way house. Consider a more complex example. Some years after World War II various participants in the Manhattan Project admitted to a short-sightedness with respect to their rush to build an atomic bomb in the summer of 1945. Some reported that before the surrender of Germany, people reasonably perceived the urgent need for the bomb, given intelligence had revealed that German scientists were well on their way to building the weapon. Once Germany surrendered, a few scientists involved in the atomic weapons program headed by Oppenheimer in New Mexico reported a vague sense that the need for a bomb was not as urgent. Japan had not surrendered, but that seemed imminent. Nonetheless, there was no real discussion about whether the project ought to be continued. Afterwards, members reported feeling a sense of inevitable momentum, fueled by very concrete conditions. The military had offered unlimited resources to support the project. The team had grown close, developing bonds of affection and camaraderie—bonds made possible by the special isolated, ‘small-town’ conditions under which the secret research was conducted. These scientists had, of course, dedicated themselves to lives of research. The habits of team members made it possible to ignore important questions about the consequences of their research as conditions changed. The example illustrates the complexity of retrospective responsibility judgments as well as proposals about how best to handle this sort of situation in the future. The Deweyan account of habits helps to locate the failures in attention and deliberation. It is not quite right to say that team members were just outright ignorant of morally salient principles and the facts of the case. Nor is it right to say that they simply flaunted moral principles or that they were overruled by passions. The collective blindness was the result of the acquisition of socially shared habits involving professional vocation, camaraderie and the like. These habits were shaped by transactions with military and research institutions whose

Actions, Habits, and Practices


collective focus was geared towards addressing the threat of war. Once we get clear about this context, we may be better able to address questions about how to best arrange the conditions in military-cum-research institutions so as to prevent such moral short-sightedness now, and in the future. The pragmatic account helps us guard against ‘magical thinking’ about motivation by situating our judgments in the context of socially shared habits whose meaning and import are determined by historical transmission, interpenetration, and transaction with circumstances. The claim so far is that the practical knowledge that makes possible intentional action is embodied in habits. Beliefs and desires are relevant to explanations but they are not, claims the pragmatist, explanatorily basic. To be sure, much more needs to be said to prove that the pragmatist hypothesis does the best job in explaining the phenomenon of intentional action. Rival theories would have to be shown inadequate, and we would need to say more about the causal role of habits. Nevertheless, I hope to have shown that the pragmatist theory is a credible account of intentional action, and that it does an especially good job in helping us think about the concrete contexts of action. As was already mentioned, pragmatists maintain that our philosophical models and theories ought to usefully inform efforts to improve practices. We have some prima-facie reason to think that the account of habits meets this normative criterion insofar as it usefully illuminates motivation, responsibility, and the prospects for social and personal change. 5. The Justification of Action The habit-based account just given may be granted by many nonpragmatist philosophers who might reply: “Yes, something like this might be true for understanding practical knowledge and intentional action, but when it comes to questions about good and bad reasons for action, we need to turn to a theory of practical reason that answers basic questions about normativity. The pragmatist account may offer corrective insights with regard to understanding intentional action. These insights may have normative implications on particular evaluations of agents and acts. However, questions about the authority of norms, values, and the like need to be addressed. To say that we learn how to act by acquiring habits in practices is all well and good. But what makes habits good or bad? Does pragmatism has anything to contribute to this question?”


Todd Lekan

The pragmatist account that I defend is decidedly contextualist and pluralist about normative questions. It is pluralist in that it maintains that people develop habits in a variety of contexts, for a variety of purposes. What counts as a good habit in art may or may not be a good habit in morality. Moreover, what counts as a good habit in art varies depending on the aesthetic practices and institutions that support the habit. Thus, pragmatism tells us that we should be very cautious about issuing general statements about principles, norms, or values in advance of detailed attention to particular circumstances. Thus, the account of habits just sketched will be friendly to a certain kind of contextualism or particularism about values. Particularist accounts of moral judgment have enjoyed attention in recent ethical theory. Some of these moral theorists have been inspired by (one way of reading) Aristotle’s ethics (see McDowell 1979; Dancy 1993). Such accounts maintain that we do not determine whether actions are morally correct by appeal to general rules or moral theories. Rather, right and wrong are determined by perceiving particular features of a situation. Possession of certain virtuous character traits makes possible such responsiveness to the particulars. If the goal of theory is to furnish us with a general account of those principles or rules that must be consulted in order to determine what counts as a good or bad reason for action, then this version of virtue ethics is decidedly anti-theoretical. Substitute “habits” for “virtues” and it appears as if a pragmatic account might be a variant on this approach to virtue ethics. Is pragmatism anti-theoretical in just this way? Briefly, the answer depends on what is meant by “theory”. Pragmatism is anti-criterial in the sense that it denies that normative theory can provide principles in advance of our collective efforts to resolve problems. However, pragmatism does attempt to explain the function of general considerations in answering normative questions. As such, it does offer an account of the justificatory structure of practical reason. This structural account is itself a generalization of habits of inquiry that appear to have worked well in human experience. The claim is that humans have created various second-order habits—habits to transform first-order habits that have become problematic in some respect. Such second-order habits involve intellectual and imaginative capacities to generalize beyond particular situations and envision alternative possibilities. In the remaining part of the paper I distinguish three kinds of generalization in practical justification. The distinctions are functional—they track different roles that general considerations play depending on the kind of problem. The

Actions, Habits, and Practices


first type of generality consists of summaries of successful particular judgments or responses. The second type involves using rules, principles, or ideals as regulative considerations for construing salient features of problematic situations. The third type of generality is a more detailed representation of a problem-solving process—what I call a representation of a ‘deliberative event’. Let’s briefly consider these three types. Generality 1: Summaries of Prior Judgments Consider the first type of generality: summaries of successful judgments or responses. After careful study, we may glean appropriate response patterns and encapsulate those into rules. We might see, for example, that “when the cook runs out of paprika, several substitutes may be appropriate including cayenne, Tabasco etc.; or it is better to use a right handed pitcher against a right handed batter even if the pitcher’s record is not as good as the left-handed alternative”. Even though, as has already been stated, much skilled action involves nuanced sensitivities that could never be fully encapsulated in rules, it does not follow that rules have no use. When training novices, it is frequently helpful to decompose complex activities into discrete actions that map to specific circumstances. Such rules take the form of “when in circumstances C, one ought to do x”. Such rules may be used to help train novices by focusing their attention on appropriate aspects of the situation. Generality 2: The Regulative Use of Principles Nevertheless, it would be wrong to conclude that the only function of general considerations is to summarize response patterns after particular judgments or responses have been made. When action is blocked because of some novel, unforeseen problem, general considerations function to regulate our judgments about salient features of the situation. To see this, consider a difficult life-changing decision. Imagine Jane Addams confronted with the difficult choice about whether or not to devote her life to social service, especially when such a vocation may conflict with responsibilities to family. We might describe her reasoning justifying a decision to devote herself to social service in the following way: (1) If one accepts social service as a value, then one should plan to make time to help others. (2) Jane Addams accepts social service as a value.


Todd Lekan

(3) Founding Hull House—a settlement house to address problems in immigrant communities—is an appropriate way for Addams to use her time to help others. (4) Therefore, Addams ought to found Hull House. The reasoning here appears to be deductive. Given the truth of the premises, the conclusion must follow as true. However, as Tyles points out, the appearances are misleading. One reason is that in practical reasoning, “the inference to a means merely picks out one of several alternative ways to realize an end” (Tyles 1998, 287). The first premise need not be true because other means may be available. Adams might spend her time working in a homeless shelter, for example. Even if (1) and (2) are true, other values might conflict with these, such as Addams’ responsibility to take care of a sick family member. While in the midst of a practical problem, a person’s efforts are devoted towards discovering the appropriate response. Addams is trying to learn about her situation. Her practical reasoning involves, at least initially, efforts to define the problem. Is this situation—prior to the decision—one best described as ‘wasting one’s talents that could be used for others’ or is it better described as a ‘guilt complex based on early childhood memories’? Specific features of a situation are selected as salient—given special emphasis—in an effort to better describe the nature of the problem. Such selective emphasis is guided by general considerations that typically represent values or ideals such as ‘social service’, ‘care for family’ and the ‘imperative to avoid personal self-delusion’. While Addams is in the process of deliberating, these considerations are tools of analysis—helping her to construe what is important in the situation. Generality 3: Deliberative Events In many problematic situations, we move quickly from formulation of the problem by way of general considerations to an action plan that resolves the difficulty. We do not, as a rule, formulate our reasoning in much detail. But sometimes it is useful to lay out the details of our reasoning in a step by step fashion. This is especially true when the problem is complicated, or when others demand an account of our reasoning. Dewey maintains that careful analysis of problem-solving reveals a structure that he calls the “pattern of inquiry”. Dewey argues that we can find the same basic pattern across domains of inquiry, be they moral, scientific, or practical. This claim is meant to be a description; however, Dewey (following Peirce) also maintains that this pattern

Actions, Habits, and Practices


represents a method of inquiry that is superior to other methods of ‘fixing belief ’. Part of the reason this method is superior is that it is selfcorrecting and experimental. It enables us to continuously improve our knowledge claims. Many philosophers would object that even if there is a pattern that represents the best deliberative practices for discovering what is true or right, it does not follow that such a method will answer questions about whether such claims are justified. The ‘order of discovery’ is a psychological process that should be conceptually distinguished from the ‘order of justification’ which is normative. The issues here are deep, and require more discussion. Briefly put, pragmatism maintains that if genuinely problematic situations exist, then it is not plausible to divide out the process of discovering correct judgments, from the process of determining whether judgments are justified. After we have deliberated about a novel problem, we might formulate our justification in ways that leave out the details of how people arrived at the judgment. We might, for example, simply state that ‘When in such and such circumstances, one ought to do x. She was in such circumstance. Therefore she do x.’ (As was stated earlier, such generalizations may be useful to make explicit salient rules for those novices new to such situations.) But during the thick of deliberation about a novel problem, there are no such ready-made policies available. In such cases it is very often necessary to make explicit the phases of problem-solving, including the details of the particular problem. ‘Case-based’ approaches to applied ethics acknowledge this very point. Such ethicists claim that more progress is made in resolving difficult moral dilemmas by formulating the circumstances of the case, relating these to paradigmatic cases that are not in doubt (see Jonsen/ Toulmin 1988). Such detailed formulations are more helpful for establishing consensus about how to address difficult moral problems, than appeals to contentious principles from which judgments are supposedly derived. Detailed formulations of reasoning about problem cases are still generalizations that may be applied to new instances. Such generalizations display the general structure of the pattern of inquiry. I will briefly lay out this structure, and demonstrate its plausibility by showing how it nicely squares with the features of habits identified earlier. A deliberative event can be analyzed in terms of five stages. These stages represent the temporal ordering of deliberative events. That deliberative events display these phases does not imply that people consciously think of their deliberations according to this structure.


Todd Lekan

Nevertheless, the pragmatist claims that explicit formulation of this structure can provide some theoretical insight into practical justification. The phases are as follows (see Dewey 1991, 105 – 122): There is an indeterminate situation, including a felt sense of trouble due to a failure of habits. There is a preliminary interpretation of the problem, including a view of what important goods and evils are at issue. Action plans are created which attempt to take account of the important goods and evils at issue. These action plans are tested in imaginative trials.3 The most promising plan is tried and tested in actions bringing about change in the situation. *



* *

These five stages are a schema of a deliberative event. It is important to distinguish this abstract schema of a deliberative event from a particular representation of a deliberative event, and the deliberative event itself. A deliberative event is a relation between a person, a representation, and a problematic situation. The representation is a sign for, or an interpretive object of, a problematic situation. Dewey writes that “ … it is a unique doubtfulness which makes that situation to be just and only the situation it is” (Dewey 1991, 109). Indeterminate situations are the result of a specific transaction of habits and circumstances. Addams has just this situation of doubt because of the unique conflux of acquired habits and social circumstances that make possible just this very indeterminacy. It would be wrong to locate the indeterminacy that gives rise to deliberative events simply in the psychology of a person. Addams is searching for a justified action that will reconcile her habits and circumstances. Habits interpenetrate in both personal and social aspects. Addams’ indeterminate situation arises from the fact that she has acquired personal habits of care for family and social service. The possibility of joining the Hull House project brings out the tension between these habits. Since habits are socially shared, Addams’ deliberative event can be explained in 3

The first and fourth phases of the deliberative event call attention to the role of feeling and imagination in deliberation. I do not develop these important dimensions of deliberation in this paper. However, the fact that the pragmatic meta-ethical model puts these features at the center of deliberation promises an additional argument in its favor to the extent that it can be shown that other extant models inadequately treat imagination in ethics. For the best extended defense of the centrality of imagination in pragmatist moral deliberation see Fesmire 2003, 55 – 68.

Actions, Habits, and Practices


terms of larger social habits. Her dilemma is not completely idiosyncratic because her habits are shaped by social practices and institutions. Educated middle-class women in the late nineteenth-century were confronted by similar dilemmas. Women who had not married were expected to care for siblings or other family members. But at this time, more women forestalled marriage because of increased educational and occupational opportunities. While such social dimensions need not be part of Addams’ own explicit self-understanding of her problem, we may easily imagine circumstances in which they could be. This would especially be the case if, for example, she and other women shared their deliberations. (In fact, Hull House served as a social site for these kinds of shared deliberations.) Now, consider historical transmission as a property of socially shared habits. We who come later may participate in deliberative events about the very same socially habits of those who came before. Those who take up social service today inherit habits of thinking, feeling, and doing that were shaped by Addams and her cohorts. These habits may be problematic in some respects, given our current circumstances. Some have argued, for example, that Addams was not critical enough of the traditional association of women with care-giving. What may have been an adequate action-plan given the circumstances in the late nineteenth century may be inadequate today. It seems correct to say that pragmatism offers a ‘theory’ of practical justification if by that we mean a general account of the structure of justification, including an interpretation of the various ways that general considerations figure in practical judgment. But, if by “theory” we mean something like a philosophical adumbration of ends, principles, or rules that could be worked out independently of deliberative-events, pragmatism is decidedly anti-theoretical. The deliberative event is meant to model the structure of practical justification. It is certainly not meant to provide a substantive criterion of justification of action. If genuinely problematic situations exist, then the most we might do in advance of analysis of the particulars of such problem situations is to adumbrate those procedural processes that foster good deliberation. We might also sketch an account of those habits that make for good problem-solving. (We might think of these as the virtues of the good inquirer.) It is true that the formal end of deliberative events is the resolution of indeterminate situations by way of satisfactory action plans. We might say that action plans are justified if they resolve the problematic situation. But this is vacuous. Consider the analogy with good explanations. It is


Todd Lekan

true, but uninformative, to say that an explanation is good if it explains the phenomena. The question is how the explanation accounts for the phenomena—what specific observations does it predict, how does it fit in with other hypothesis and so forth. Similarly, an action plan is good depending on the way that it resolves conflicting habits and the various values that those habits may express. Much more can be said about a pragmatist account of practical justification. Perhaps the most important point to drive home here is that any pragmatist account of justification will emphasize the fallibility, sociality, and historicity of human action. Actions occur in the context of habits and practices. The primary normative question, for pragmatism, is how we might continually learn about these habits and practices so that we can more intelligently reform these in response to new problems. Norms, values, and ideals arise out of our deliberative activities. Their authority is conferred by the tribunal of experience. Bibliography Dancy, Jonathon (1993): Moral Reasons, Cambridge: Oxford University Press. Dreyfus, Hubert I./Dreyfus, Stuart E. (1990): “What is Morality?”, in: Rasmussen, David (ed.): Universalism versus Communitarianism, Cambridge: Cambridge University Press, 237 – 264. Dewey, John (1983): Human Nature and Conduct, in: Boydston, Jo Ann (ed.): The Middle Works Volume 14, Carbondale: Southern Illinois University Press. — (1991): Logic: The Theory of Inquiry, in: Boydston, Jo Ann (ed.): The Late Works Volume 12, Carbondale: Southern Illinois University Press. Fesmire, Steve (2003): John Dewey and Moral Imagination, Bloomington: Indiana University Press. Jonsen, Albert R./Toulmin, Stephen (1988): The Abuse of Casuistry: A History of Moral Reasoning, California: University of California Press. Lekan, Todd (2003): Making Morality: Pragmatist Reconstruction in Ethical Theory, Nashville: Vanderbilt University Press. McDowell, John (1979): “Virtue and Reason”, in The Monist 62, 331 – 350. Tyles, James (1998): “Rationality Beyond Deduction: A Guide for the Perplexed and Disappointed”, in: Westphal, Kenneth (ed.): Pragmatism, Reasons, and Norms, New York: Fordham University Press, 265 – 292. Will, Frederick L. (1997): “The Rational Governance of Practice”, in: Westphal, Kenneth (ed.): Pragmatism and Realism, Lanham, Md: Rowman and Littlefield, 63 – 83.

The Reach of Habit Norbert Anwander 1. Introduction While there is no lack of dispute about the proper characterisation of reasons for action, most philosophers would readily agree that reasons are the via regis to understanding action.1 In his exposition of a pragmatist theory of action, Todd Lekan presents a refreshing alternative to this standard view. Rather than starting off with an analysis of the concept of a reason for action, he suggests that we begin by thinking about habits. Lekan motivates this methodological shift by reminding us that “actions occur in the context of habits and practices” (180).2 This is not just the modest point that it would be mistaken to think of actions as isolated phenomena. True, the amount of literature dealing with behaviour such as blinking or raising one’s arm might suggest that this is indeed what action theory has largely been doing. In any case, the more ambitious thesis put forward by Lekan is that intentional action is importantly dependent on habits. As he argues, intentional action requires practical knowledge, and the latter is in turn “embodied in socially shared habits” (164). Moreover, as Lekan sees it, it is always in the context of such habits that justificatory questions about reasons for action are raised. The pragmatist urges us therefore to begin the philosophical study of human action with “the social and psychological contexts that make action possible” in the first place (164). My comment will follow suit and focus on habits. While I welcome Lekan’s attempt to broaden our perspective on human action, I am sceptical about the Deweyan concept of a habit, on which his account is explicitly based. For this concept seems to run together two quite different ideas. On the one hand, there is the idea of habits as social practices, and on the other hand, there is the idea of habits as individual dispositions. The aim of my subsequent discussion of these two notions is to get a clearer view of the reach of habit. I will argue that although 1 2

See e.g. Anscombe 1957; Davidson 1980; Audi 1997; Mele 1997; Dancy 2000; Bittner 2001; see also Alvarez 2007; Hornsby 2007 and Stoutland 2007 (all in this volume). Parenthetical references in the text are to Lekan 2007 (in this volume).


Norbert Anwander on Todd Lekan

Deweyan habits complete the picture neither of the two notions provides a genuine alternative to reasons when it comes to understanding action. In the final section, I explore the possibility of a more thorough way of explaining people’s actions by referring to their habits. 2. Habits as Social Practices When Lekan offers the examples of “house building” (169), “the practices of the Roman Catholic church” (170), and “[the art of ] jazz improvisation” (ibid.), what he seems to have in mind are habits in the sense of social practices. I take such habits to be more or less regulated ways of doing things. An essential feature of these specific types of activity is that they are socially shared. The relevant group may be very large as with the members of a world religion, or it may be rather small, consisting only of the aficionados of some particular type of music. The practices are constituted by some set of rules, which are, however, in turn subject to change and even explicit modification. The development of literary genres is a particularly striking example, but such transformation is also to be found in scientific standards, table manners, and so on. Such practices can be essential for human action in various ways. For a start, some kinds of activity are not even possible without some specific established social practice. In the absence of the respective institution, you simply cannot sign a contract, get married, or take part in a competition. Yet this uncontroversial point cannot be the one Lekan wants to make, for there is no specific way in which this conditionality on practices is a matter of the practical knowledge that is embodied in them. The relevant way in which practices inform human action is by providing us with a repertoire of activities. For most people all of the time and for all people most of the time, the things they do are not novel, but have been done by others many times before and are being done by many others simultaneously. Now, all these activities follow a certain pattern, or a script, as we might say. It is obvious that there are certain rules that you have to follow if you are to write a sonnet (the relative rigidity of such rules is itself a feature distinctive of different practices: the rules for classical poetry are highly rigid, those for slam poetry are rather lax), but even such mundane activities as mailing a letter, going to work, and watching a film follow a certain pattern. As long as you are engaged in these activities, it is to some extent settled what you are to do. If you are writing a philosophical comment, you cannot insert a poem about your

The Reach of Habit


cat even if that is what you feel like doing. Needless to say, without such established patterns of doing things, life would be even more difficult and disorienting than it already is as everything would constantly be up for decision. Social practices are extremely helpful sets of decisions you no longer have to make yourself. Social practices also make actions intelligible. Knowledge of the kind of activity in which someone is engaged helps us to understand why he is behaving in this rather than in that way. Unless you know that somebody is writing a sonnet, it must seem utterly bizarre why he would bother about rhyme and rhythm instead of simply conveying his message in the most convenient way. If tomorrow’s newspaper contained poetry rather than information, you would be quite puzzled. As numerous novels and movies prove, describing people’s behaviour while leaving out and pretending ignorance of the conventions against which they make sense is a guaranteed recipe for comedy. Alasdair MacIntyre (1986) has argued that the intelligibility of all actions is dependent on social practices. At first sight, this strikes one as an exaggeration and putative counterexamples easily come to mind: Is not my thirst sufficient to explain why I am drinking this glass of water? It just does not seem as if we needed to appeal to any social practice in order to make sense of this piece of behaviour. But, as MacIntyre points out, there are situations where even appeal to such basic needs as hunger and thirst will not make an action intelligible. His example is that of a scientist who is shown the only specimen of a new hybrid fruit which is easy to cultivate and, thanks to its high nutritional value, could help to overcome starvation. We would be very puzzled if the scientist reacted by grabbing the fruit and gobbling it down. His explanation that he was hungry and liked fruit would not make his behaviour more intelligible but rather less. For, MacIntyre argues, “in the context of this kind of scientific practice eating something just because one felt hungry is not an intelligible way of behaving” (MacIntyre 1986, 73). His point is that it is itself a matter of social practice whether or not in a particular situation reference to basic needs will make an action intelligible: “There are no contextless good reasons.” (ibid., 74) This then is the priority claimed for habits over reasons for action: For any particular situation it is dependent on social practices which considerations count as reasons. In the absence of a background of shared practices, no human action, not even the seemingly most natural one of satisfying one’s needs, can be understood. Let us for the sake of argument


Norbert Anwander on Todd Lekan

accept this priority thesis.3 The question now is what the priority thesis means for standard theories of action, in which the concept of a reason for action takes pride of place. Very little, I think. The thesis that reasons for action (less ambiguously: considerations that count as reasons) are dependent on social practices just does not engage the analysis and understanding of action itself in terms of such reasons. Granted, the repertoire of actions available to a particular agent in a particular situation is contingent on social practices, and reference to them may be required to make any chosen course of action intelligible. No doubt, if you want to make full sense of someone’s specific behaviour, more is needed than standard action theory has to offer. However, this theory’s primary focus is a more abstract one, viz. the issue of what distinguishes action proper from mere behaviour. The aim is not so much to make sense of particular actions or even to explain how human beings come to act, but rather to understand what action per se is. For this, there is indeed no need to inquire into any specific reason and why it could be seen as such. It appears then that the question of where the study of action should begin is basically a matter of what you are interested in. Very roughly speaking, standard action theory, on the one hand, seems most at home in the criminal court, where one major issue is whether something was done intentionally, deliberately, and so on. Social practices, on the other hand, are particularly of interest to anthropologists as well as to those interested in changing our ways of doing things. Either context is perfectly legitimate and important. Moreover, even if appeal to social practices is necessary in order to make some piece of behaviour intelligible, it is clearly not sufficient. Reference to a habit in the sense at issue may help you to understand what somebody is doing—he is writing a sonnet, he is playing football, he 3

For the record, I have my doubts about this priority thesis. For it may be that basic needs always provide some reason for action, albeit not a conclusive one. Moreover, even the view that in some contexts satisfying a basic need is no consideration in favour at all, does not commit one to the priority thesis. Any particularist, even of the most realist bent, is happy to hold such views, see Dancy 2004.—There is also the issue of how intelligibility relates to justification. As I see it, some consideration may help to make sense of an action without being seen as providing any justification for it, not even from the agent’s point of view, see Anwander 2002. Because (of my view that) normativity in the strong justificatory sense relevant for prudential and moral reasoning is beyond the province of action theory proper, I ignore the second half of Lekan’s paper, in which he turns to questions about good and bad reasons for action and presents a pragmatist account of the justificatory structure of practical reason.

The Reach of Habit


is writing a comment. If you further want to understand why some person is engaging in this or that activity at all, social practices provide you with no answer.4 They define the repertoire available, but this still leaves us with the question of why some individual person here and now chooses to play it in this rather than in that way. 3. Habits as Individual Dispositions The alternative notion of a habit as an individual disposition seems better suited when it comes to answering the question of why some individual acted as he did. Instances of this notion can be found in the list of qualities that Lekan offers as Deweyan habits: Impatience as a personality trait, a tendency to hard drinking after hours, skill in manipulating surgical instruments, an exemplary singing voice, class prejudice, loyalty to the Roman Catholic church, commitment to truth-telling, the ability to speak several languages fluently (168; Lekan 2003, 27). The items on this list strike me as rather diverse, but there is at least one thing they have in common: They are qualities of an individual agent. While the question of where social practices are located is potentially an interesting one, it is just obvious that particular people display impatience, are prone to hard drinking, have surgical skills, exhibit an exemplary singing voice, and so on. Of course, other people can have these habits too, but they are still the properties of individual agents. There is no interesting sense in which, for instance, the ability to speak several languages fluently may be socially shared. Habits as social practices, I agreed, are relevant for understanding what somebody is doing. But I also noted that they provide no room for addressing the question of why some agent is engaged in some practice in the first place. Habits in the sense of individual dispositions, however, seem suited to this task exactly because they refer to a particular agent’s qualities. Why is this teacher snapping at his students? Well, he is just a rather impatient person and hates explaining things twice. Why does that


Admittedly, the truth is more complex. For the explanation why some individual is engaging in a piece of poetry or philosophical comment may in turn refer to him being a poet or an academic philosopher respectively; and writing such things is what such people do. While this deferment is entirely legitimate, I doubt that it is satisfactory in the end.


Norbert Anwander on Todd Lekan

woman refuse to lie? Well, she is a Kantian and very committed to truthtelling. Again, my interest is in how this explanatory scheme relates to those theories for which the concept of a reason for action is pivotal. In his discussion of skilled expertise, Lekan plausibly argues that, when performances are flowing well, experts do not “perform actions through explicit consultation of rules, principles, or reasons” (166). Indeed, we would not get much done and we would not live for long if everything we do had to result from careful deliberation or the explicit following of rules. Lekan is perfectly right to draw our attention to the practical knowledge necessary for intentional action. I also agree that this practical knowledge is not propositional, but should be understood as embodied know-how. This is most conspicuous in the following two kinds of behaviour: (i) skilled action: We are able to walk without having to think about the next step, some of us have the ability to play the piano in their fingers as it were, and others can at least type that way. (ii) immediate response: Wellmannered people immediately pass the salt when requested to do so, and experienced drivers immediately push the brake pedal when a child runs into the street. Their responses to the world are direct, without need for mediating thought. Undoubtedly, such automatic behaviour based on embodied know-how is a real part of our lives, and a very important one too. But is automatic behaviour alien to the standard reasons theory of action? I do not think so. First of all, we should not confuse acting for a reason with action performed on the basis of practical reasoning (see Audi 1997, 104). No reasons theorist holds that the agent permanently has to weigh reasons—what are the pros and cons of not hitting the child?; or to engage in a piece of instrumental reasoning—what is the best means to bring the car to a stop? It is certainly true that most actions are not the result of a conscious decision, yet no reasons theorist ever said they were. A more delicate issue is whether an agent needs to be presently aware of his reasons as reasons. Robert Audi, for example, presents a dispositional account, according to which an agent v-s for some reason if she is noninferentially disposed to attribute her v-ing to that reason (ibid., 83 f.). The opposite view is, among others, held by Jonathan Dancy, who specifies reasons as “the features in the light of which the agent acted” (Dancy 2000, 129). On that account, reasons have to be conceived by the agent as favouring the action. It looks like there may be a problem if Dancy is right: Since as we act reasons are often not present to us, it

The Reach of Habit


would seem that there are not all that many actions done for reasons after all. We must not be rash here, though. For much of what we do is best described in terms of subsidiary actions (see Searle 1983, 84; Mele/Moser 1997, 231 f.). For example, my writing this comment on Lekan’s paper involves a number of subsidiary actions, including typing. My typing may be considered intentional, not on account of a present intention to press this or that particular key, but because it is part of a larger activity for which I have the relevant intention. Analogously, I may be said to push this or that particular key for a reason because this piece of behaviour is suitably related to my writing this comment, which undoubtedly I do for a reason. Some people may want to dismiss this as a mere verbal manoeuvre for the reasons theorist to save his face. Even so, there is an important lesson here. Which role reasons for action play is itself a matter of context: Yes, I can type automatically, driven (or guided, if you like) by the force of habit, but as to writing this comment, there is, of course, a reason for which I am doing it. The same holds for Lekan’s example of the expert cook: While the cook is cutting the carrots, boiling the rice, and stirring the meat, he may not think much in terms of reasons, but I am sure there is something in the light of which he is preparing the meal. It turns out, then, that skilled action appears to be independent of reasons only as long as our fascination with the how-question—how does he do it?—keeps us from asking the why-question—why is he doing it? What about immediate response? As you ask for the salt, the wellmannered person passes it without thinking. On the account that I favour, the right thing to say here is that she passes the salt for a reason, the reason being your request.5 In characterising this person as wellmannered, I have referred to the fact that she sees legitimate requests as reasons for action and acts accordingly. This opens up a way not only for reconciling reasons for action and habits, but for having them work in tandem. For the well-mannered person’s automatic and reliable pattern of seeing requests as reasons seems to be just the sort of thing that Lekan would call a habit. The conciliatory picture that suggests itself is this: If you want to understand intentional action, you always have to look at the reasons for which the agent acted. Habits, however, often explain why some specific agent took certain features to be reasons for him to act, in 5

This is reasons externalism; see Dancy 2000; Bittner 2001; see also Alvarez 2007; Hornsby 2007; Stoutland 2007 (all this volume).


Norbert Anwander on Todd Lekan

particular when both his recognition and his response are automatic and reliable. 4. Habits in lieu of reasons for action? On any plausible understanding of what it means to act for a reason, it is perfectly possible for an activity both to be done for a reason and to flow automatically. Moreover, there is no conflict between an agent’s conceiving of and acting on some feature as a reason and that recognition being habitual. The conciliatory picture towards which I gestured in the previous section makes it look as if all to which the argument of the habits’ advocate boils down is that we should pay more attention to distal factors—habits, rather than proximal ones—reasons for action. Given Lekan’s pragmatist aim to improve practice, this is sensible advice: For if change in behaviour is what ultimately you are aiming at, it is very often a good idea to address those factors that erroneously, but stubbornly make you see certain factors as reasons for action. With regard to the theory of action, however, a mere debate about the comparative importance of distal and proximal factors in explaining action has little impact, if any at all. While it is different aspects that are being highlighted, the basic explanatory pattern remains the same. In the philosophy of action, the real issue, I suggest, is this: What figures in the best complete answer if you want to know why someone acted as he did? For all I know, it has never been denied that actions are often best explained via the reasons for which they were done. But friends of reasons usually make the more ambitious claim that the explanation of an action always involves reference to the agent’s reasons. More precisely, they cherish rationalisations in terms of reasons for action as a privileged kind of explanation. Dancy, for instance, writes: “The explanation of an action succeeds to the extent that it enables us to see how the agent might have taken certain features of the action as good reasons to do it.” (Dancy 2000, 95) And Audi, focussing on agents, states: “If we do not know for what reason a person acts, we do not fully understand that person.” (Audi 1997, 75) It is this pretension to an explanatory monopoly that the friends of habits (in the individualist sense here at issue) should challenge. The claim would be that for some actions, the best explanation mentions the agent’s habits rather than his reasons. Again, this is not just the weak thesis that for some purposes it is more illuminative to mention distal factors—habits, rather than proximal ones—reasons for action. That

The Reach of Habit


would leave action theory exactly as it is. What I have in mind is the strong thesis that habits may explain actions in lieu of reasons. In the remainder of this comment, I will step beyond Lekan’s own account and tentatively explore how such a claim might be corroborated. As a preliminary task we have to make sure that it is not any old piece of behaviour that we are explaining but respectable intentional action. Intentionality is generally seen as the hallmark of our agency (as opposed to the mere doings of things) and a requirement for accountability (see Guttenplan 1994, 63 ff.; Mele 1997, 16 f.). On the face of it, habituality does not preclude intentionality: People’s habitually brushing their teeth before going to bed is fully compatible with the brushing being suitably related to their agency. If it came to it, we would hold them fully responsible for having brushed their teeth. The problem, however, is that in most theories of action there is a very close connection between intentionality and acting for a reason.6 The coughing that is caused by a bad cold is not intentional, but when I politely cough with a view to signalling my presence, we are talking about a fully-fledged action. If the advocate of habits accepts this connection, he must either limit his ambitions to non-intentional behaviour, or accept that whenever he offers his explanation of an intentional action, there is a reason for action around and a standard explanation in its terms available. But the right thing to do, I think, is to reject the view that intentionality is tied to reasons for action. Here are some factors that may instead secure a behaviour’s connection to agency: (i) intervention control (at any time, I can stop brushing my teeth); (ii) modification and adaptation (I can apply more or less force, use dental floss if necessary); (iii) internal goal (I can tell whether or not brushing my teeth was successful).7 We do not need reasons for action to distinguish behaviour closely connected to agency from mere behaviour such as coughing, snoring, and absent-mindedly scratching one’s head. Note that the above factors accounting for intentionality have no explanatory force: None of them gives an answer to the question of why I brushed my teeth. So there is actually room for other things to come into play. Are habits suited to do the explanation? The answer depends on how we shape our concept of a habit. It is easy to see why the conciliatory 6 7

See e.g. Anscombe 1957, 9; Davidson 1980, 84 f.; Guttenplan 1994, 65; Audi 1997, 102 ff.; Mele 1997, 20. For intentionality without reasons, see Brett 1981; Anwander 2002, 232 – 237; Pollard 2006, 60 – 63.


Norbert Anwander on Todd Lekan

picture that I sketched at the end of the previous section will not do: Its definition of habits makes them dependent on the concept of a reason for action. Among other things, this rules out commitments and loyalties, as mentioned in Lekan’s list that I cited above. For to be committed to truth-telling crucially involves seeing oneself as having strong reasons to tell the truth. Other items on that list will be of no help either, albeit for different reasons: For instance, the ability to speak several languages fluently, just as skills generally, can never explain an action. They are at best enabling conditions, necessary for another explanation to go through, but not making a positive contribution (see Dancy 2000, 127 ff.). For a positive account, I can do no better than quote Bill Pollard’s characterisation of a habit as “a pattern of a particular kind of behavior which is regularly performed in characteristic circumstances, and has become automatic for that agent due to his repetition” (Pollard 2006, 57). One crucial aspect of Pollard’s definition is that it does not at all mention mental states of the agent, such as desires, beliefs, or intentions. Furthermore, it highlights the repetition involved in a habit. It is not simply because of a tendency or a personality trait that we attribute a habit to a person, but because some person has actually repeatedly acted in a specific way. Partly, this is a metaphysical point: While there can be dispositions without manifestations, it is impossible for you to have a habit of v-ing without having repeatedly v-ed. (Incidentally, this is a reason for thinking that individual actions enjoy conceptual priority over habits.) Finally, this concept is unambiguously behaviouristic, in that a habit just is a regular pattern of behaviour. Explaining an action in terms of such habits would be highly contextual. The context in question is the agent’s own history. We would make sense of an agent’s v-ing now by showing how it relates to his previous v-ing in similar circumstances and fitting it into a pattern of similar behaviour. Are there plausible instances of actions that are best explained in this way? To make a convincing case, detailed stories might be needed, but here are two brief examples: My parents used to say grace, and I used to finish meals by smoking a cigarette. The crucial thing here is that my parents continued their habit long after religion had lost its grip on them, and I certainly was never committed to having a post-lunch fag. Neither was anyone responding to a feature of the situation conceived of as a reason for action. It seems to me that in both cases, a perfectly intelligible and probably the best answer to the question as to why we acted as we did, would have been: “Oh, they always do that!”

The Reach of Habit


It remains to be seen whether such an alternative pattern of explanation by reference to habits can really be established. It may well be that the reach of reasons for action remains unlimited, i.e. that at some stage or other, the explanation of actions must mention reasons. Nevertheless, both aspects of Deweyan habits, social practices and individual dispositions, provide an illuminating way of looking at action in context: the often neglected context in which human beings come to take particular considerations as reasons for action.

Bibliography Alvarez, Maria (2007): “The Causalist/Anti-causalist Debate in the Theory of Action: What It Is and Why It Matters” (this volume). Anscombe, G. E. M. (1957): Intention, Oxford: Blackwell. Anwander, Norbert (2002): “Intelligibilitt und Normativitt”, in: Analyse und Kritik. Zeitschrift fr Sozialtheorie 24, 231 – 248. Audi, Robert (1997): “Acting for Reasons”, in: Mele, Alfred R. (ed.): The Philosophy of Action, Oxford: Oxford University Press, 75 – 105. Bittner, Rdiger (2001): Doing Things for Reasons, Oxford: Oxford University Press. Brett, Nathan (1981): “Human Habits”, in: Canadian Journal of Philosophy 11, 357 – 376. Dancy, Jonathan (2000): Practical Reality, Oxford: Oxford University Press. — (2004): Ethics Without Principles, Oxford: Oxford University Press. Davis, Lawrence H. (1994): “Action (1)”, in: Guttenplan, Samuel (ed.): A Companion to the Philosophy of Mind, Oxford: Blackwell, 111 – 117. Davidson, Donald (1980): Essays on Actions and Events, Oxford: Clarendon University Press. Guttenplan, Samuel (1994): “An Essay on Mind”, in: Guttenplan, Samuel (ed.): A Companion to the Philosophy of Mind, Oxford: Blackwell, 1 – 107. Hornsby, Jennifer (2007): “Knowledge in Action” (this volume). Lekan, Todd (2003): Making Morality: Pragmatist Reconstruction in Ethical Theory, Nashville: Vanderbilt University Press. — (2007): “Actions, Habits, and Practices” (this volume). MacIntyre, Alasdair (1986): “The Intelligibility of Action”, in: Margolis, Joseph/ Krausz, Michael/Burian, Richard M. (eds.): Rationality, Relativism and the Human Sciences, Dordrecht: Nijhoff, 63 – 80. Mele, Alfred R. (1997): “Introduction”, in: Mele, Alfred R. (ed.): The Philosophy of Action, Oxford: Oxford University Press, 1 – 26. — /Moser, Paul K. (1997): “Intentional Action”, in: Mele, Alfred (ed.): The Philosophy of Action, Oxford: Oxford University Press, 223 – 255. Pollard, Bill (2006): “Explaining Actions with Habits”, in: American Philosophical Quarterly 43 (1), 57 – 69.


Norbert Anwander on Todd Lekan

Searle, John (1983): Intentionality, Cambridge: Cambridge University Press. Stoutland, Frederick (2007): “Reasons for Action and Psychological States” (this volume).

II. Action, Persons and Life

What Good is a Will? J. David Velleman 1. Introduction As a philosopher of action, I might be expected to believe that the will is a good thing. Actually, I believe that the will is a great thing—awesome, in fact. But I’m not thereby committed to its being something good. When I say that the will is awesome, I mean literally that it is a proper object of awe, a response that restrains us from abusing the will and moves us rather to use it respectfully, in a way that does it justice. To say that the will is a good thing, however, would imply that having a will is better than not having one, or that using it is better than not using it— neither of which I am prepared to assert as a general rule. Speaking metaphorically, I would say that the will is like a magic wand. In fairy tales, the character who looks upon a magic wand as an unalloyed good is destined to be sadder but wiser in the end. Being a magician isn’t better than being an ordinary human, just different; and a magician must value his powers by respecting them and therefore using them appropriately, even sparingly, not by using them as much as possible. Kant expresses something like this view in the first section of the Groundwork: In the natural constitution of an organic being—that is, of one contrived for the purpose of life—let us take it as a principle that in it no organ is to be found for any end unless it is also the most appropriate to the end and the best fitted to it. Suppose now that for a being possessed of reason and a will the real purpose of nature were his preservation, his welfare, or in a word his happiness. In that case nature would have hit on a very bad arrangement by choosing reason in the creature to carry out this purpose. For all the actions he has to perform with this end in view, and the whole rule of his behaviour, would have been mapped out for him far more accuArCHTUNGRE ately by instinct; and the end in question could have been maintained far more surely by instinct that it can be by reason. (Kant 1956, 62 – 63, Acad. 395)


J. David Velleman

Later in the Groundwork, Kant is going to declare that “the will is nothing but practical reason” (80; Acad. 412). So when he belittles the utility of reason in “a being possessed of reason and a will”, Kant is actually belittling the utility of the will itself. If Nature had merely been looking out for our well-being, according to Kant, she would have placed our behavior under the rule of instinct, without bothering to endow us with a will at all. An opposing view of the will can be found in Michael Bratman’s classic book Intention, Plans, and Practical Reason. Bratman doesn’t speak of the will as such, but his conception of intention, as an effective commitment to act, can be interpreted as a conception of that which issues from the will and mediates its influence on behavior. The will can therefore be conceived as the faculty of intention, in Bratman’s sense of the term. Bratman argues convincingly that intention is distinct from desire and belief, and hence that the capacity for intentions is supplementary to the capacity for desire-belief motivation. In Bratman’s view, this supplementary capacity “has a pragmatic rationale, one grounded in its long-run contribution to our getting what we (rationally) want”.1 Bratman thus appears to think, in opposition to Kant, that the will is designed for our preservation, welfare, and happiness. In particular, Bratman believes that the capacity for intentions enhances the effectiveness of desire-belief motivation, by enabling it to secure greater desire-satisfaction in the long run. Many non-human animals have the capacity for desire-belief motivation, which enables them to navigate their environment intelligently in the pursuit of survival and comfort; but they cannot frame commitments of the sort that would store the results of advance problem-solving and coordinate their actions over time. The latter capacity belongs exclusively to humans, and in Bratman’s view it makes them even more intelligent than other animals in the pursuit of desirable ends. I don’t want to deny that the will often provides the instrumental advantages that Bratman claims. What I deny is that the will is best understood as designed to provide those advantages. The way to figure out how the will works is not to figure out how it serves the purpose of making us more effective in pursuing desirable ends. Indeed, the way to figure out how the will works may not be to assume that it has any 1

See Bratman 1987, 35. See also p. 34: “[T]his role of intentions in providing a background framework for the weighing of desire-belief reasons is itself grounded in pragmatic considerations concerning the satisfaction of (rational) desire.”

What Good is a Will?


purpose at all. For in my view, the will may well be an accident, of the sort that Richard Lewontin and Stephen J. Gould have called a spandrel—that is, a feature formed not by design but by the accidental confluence of other designed features (see Gould/Lewontin 1979; Gould 1997). 2. Bratman’s Theory Bratman calls his theory of intention a “planning theory” (Bratman 1987, 9), because it takes future-directed plans as the paradigm for all intentions, including intentions directed at present or immediately forthcoming actions (4). Future-directed plans serve two purposes, according to Bratman (2 f.). First, they facilitate scheduling of deliberative effort, by enabling us to deliberate in advance, when time and materials for deliberation are plentiful, and then to store the results until the occasion for action arrives, when resources for deliberation may be scarce. Second, plans facilitate coordination of actions, by committing us to future actions whose performance we and others can count on. Plans can serve these purposes because they embody commitments that are, in Bratman’s terminology, both volitional and reasoningcentered. In their volitional aspect, plans determine our behavior: unless we change our minds, we will do what we plan, when the time arrives. In their reasoning-centered aspect, plans set the agenda for, and constrain the scope of, further deliberation (15 – 20). Bratman analyzes the latter, reasoning-centered aspect of plans in terms of norms and corresponding dispositions of practical reason (29 – 49). First, a plan is rationally required to resist reconsideration, so as to constitute a stable making-up of our minds. Second, a plan is rationally required to be means-end coherent—that is, to be filled in with instrumental details in time for its execution. Third, our plans are rationally required to be consistent with one another and with our beliefs, so that their joint fulfillment is possible given the facts as we see them. In accordance with the first requirement, a plan to spend one’s vacation in Maine, for example, tends to close the question where to spend one’s vacation, and it tends to keep the lid on that question unless circumstances change significantly. In accordance with the second requirement, the same plan tends to prompt deliberation about means of getting to Maine, issuing in a plan about how to get there. In accordance with the third requirement, the plan excludes from consid-


J. David Velleman

eration any plan of spending one’s vacation elsewhere, or of doing anything else that would be incompatible with being in Maine at that time. In their volitional and reasoning-centered aspects, plans occupy a different functional role from desires. A desire to spend one’s vacation in Maine might influence one’s behavior but wouldn’t determine it, given the possibility of countervailing desires that may be stronger; and so the desire would lack the volitional nature of a plan. A desire to spend one’s vacation in Maine would also lack the reasoning-centered features of a plan, since it could rationally be abandoned at any time, could rationally be retained in the presence of a conflicting desire—say, to spend the vacation in Florida—and wouldn’t oblige one to plan a means of getting to Maine. Bratman sums up these differences by saying that a desire fails to settle our course of action (18 f.). Because a plan would settle things, as a desire does not, it is entitled to be regarded as a distinct mental state, which Bratman calls intention. So characterized, intentions differ even from those desires which set goals for our deliberations and actions. As Bratman points out, we can rationally adopt goals that are not jointly attainable, as a way to improve our chances of at least partial success, but we cannot rationally adopt incompatible intentions (chap. 8). For example, we can rationally aim at being awarded each of two fellowships, even if the award of either would automatically cancel our application for the other, but we cannot rationally plan or intend to win each of these fellowships, given their mutual incompatibility. Intentions have the function of settling the outcomes that they represent, and we cannot settle outcomes that are incompatible; we can aim at incompatible outcomes because aiming doesn’t settle anything. Bratman believes that every future-directed intention requires a present-directed intention to convey its motivational force: Future-directed intention involves volitional commitment because of its relation to present-directed intention: if my future-directed intention manages to survive until the time of action, and I see that this time has arrived and nothing interferes, it will control my action then. Presentdirected intentions have a special relation to action, and future-directed intentions are the sort of state that will have this special relation if they survive until the time of action is seen to have arrived. (108)

Bratman identifies no other useful role for present-directed intention, and he might therefore be expected to conclude that it exists only to play this role and hence only when some formerly future-directed intention

What Good is a Will?


has matured to the point of requiring present action. Yet Bratman believes that whenever an agent does something intentionally, he has an intention to do something (though not necessarily the same thing; see 119 ff.).2 Bratman is therefore committed to the existence of presentdirected intentions that didn’t develop out of future-directed intentions, motivating actions that are intentional but were not antecedently planned. Bratman’s view on the prevalence of present-directed intentions can be extrapolated from remarks that he makes about the association between intentional action and intention. In considering potential counterexamples to this association, Bratman mentions only “automatic and unreflective” actions, as illustrated by this example: “Suppose you unexpectedly throw a ball to me and I spontaneously reach up and catch it.” (126) Bratman suggests that such spontaneous actions, which do not spring from an intention, fail to disprove the association because they are not quite intentional, either. What’s revealing about these remarks is that they show how narrow is the class of actions that do not spring from an intention, in Bratman’s view. Even if actions are unplanned, they spring from an intention so long as they are not “automatic and unreflective”. 3. A Problem for Bratman’s Theory Now, I agree with Bratman that even in circumstances for which we have no antecedent plans, we often make up our minds before we act: we settle on a course of action before taking it, rather than taking it automatically and unreflectively. And as we have seen, settling what we are to do is characteristic of intention. Thus, Bratman is right in thinking that many unplanned actions are informed by a mental state of the sort that he calls intention. Yet the frequency with which we form immediate intentions de novo raises a potentially troubling question for Bratman’s theory. The question is why we bother to form intentions when there is no longer any opportunity for them to serve the functions for which, according to Bratman, the mental state of intention is designed. When the plate of cookies is held out to us, why do we make up our minds to take one? Why doesn’t our hand just shoot out and grab, as it does when we spontaneously and automatically react to a sudden throw? 2

Bratman concedes the possibility of minor exceptions to this rule.


J. David Velleman

We might have planned to take a cookie if we had known in advance that we would be offered one. In that case, an intention would have placed our action under the control of deliberations for which there would be no time when the plate of cookies was already under our nose; and it might have given us grounds for skipping dessert earlier in the day, so as to coordinate our diet. But if we didn’t know in advance, and consequently didn’t plan, then the offer of cookies signals that we have already missed the opportunity to form a well-considered intention and to benefit from its coordinating role. We have to make up our minds on the spot, to take a cookie or not to take one. The question, to repeat, is why we bother to make up our minds. The intention to take a cookie doesn’t incorporate the results of any prior deliberation, or set the stage for any further planning, or provide a basis for any coordination. There appears to be nothing for the intention to do.3 The same question can be raised not only about the intentions involved in unplanned actions but also about various cases of planning. Why, for example, don’t I wait until I’m standing in the voting booth to settle how I am going to vote? Of course, I have good reason for starting to think about my vote in advance: I want to be sure of making sufficient progress in my deliberations before time runs out. But having started my electoral deliberations in good time, why don’t I leave them open until the moment arrives? Why conclude them prematurely? No prior preparations are needed for marking my ballot in one way rather than another, and there is no other behavior, either mine or anyone else’s, that needs to be coordinated with my vote. The only foreseeable effect of planning my vote is that I will become resistant to reconsidering it, and yet there seems to be no harm, and potentially some good, in leaving it open to reconsideration, or in continuing actively to consider it in my spare time. Making up my mind in advance therefore seems to carry potential costs but no conceivable benefits. So why don’t I—indeed, why doesn’t everyone—show up in the ‘undecided’ column of pre-election polls?


One might think that forming an intention at least serves the purpose of terminating whatever brief deliberation we have time to undertake on such an occasion. But a spontaneous action would serve this purpose equally well; and we sometimes make up our minds without deliberating at all. What does intention add to action?

What Good is a Will?


4. A Solution: Anscombe’s Theory I think that this case points to a function of intention that Bratman has omitted from his account. So long as I am still deliberating, I don’t know how I am going to vote, and such uncertainty about my own future behavior is an uncomfortable state of mind, especially when the behavior in question is significant for my conception of myself. I am quite happy to leave my dinner selection undecided until the last moment, when the waiter’s pencil will be poised to record it. But my vote in an election is not just a momentary matter of taste: it defines my stance on questions of importance. Not to know whom I favor in an election is to suffer an undesirable kind of ignorance about myself. The idea that intention provides self-knowledge is a major theme of an alternative theory of intention—namely, that presented in Elizabeth Anscombe’s book Intention (1963). Anscombe’s primary claim on this topic is that an intention on which one acts embodies “knowledge without observation” of what one is doing. This claim has generally met with puzzlement and skepticism, and so I’d like to spend a moment defending it.4 Anscombe compares knowledge of one’s intentional action with knowledge of the position of one’s limbs, describing both as nonobservational knowledge. She expands on the latter case as follows: [A] man usually knows the position of his limbs without observation. It is without observation, because nothing shews him the position of his limbs; it is not as if he were going by a tingle in his knee, which is the sign that it is bent and not straight. Where we can speak of separately describable sensations, having which is in some sense our criterion for saying something, then we can speak of observing that thing; but that is not generally so when we know the position of our limbs. Yet, without prompting, we can say it. I say however that we know it and not merely can say it, because there is a possibility of being right or wrong: there is a point in speaking of knowledge only where a contrast exists between ‘he knows’ and ‘he (merely) thinks he knows’. (13 f.) It is not ordinarily possible to find anything that shows one that one’s leg is bent. It may indeed be that it is because one has sensations that one knows this; but that does not mean that one knows it by identifying the sensations one has. (49)


For recent defenses of Anscombe’s claim, see Hursthouse 2000; Falvey 2000. My interpretation of Anscombe has benefited especially from Falvey’s paper.


J. David Velleman

If a man says that his leg is bent when it is lying straight out, it would be incorrect to say that he had misjudged an inner kinaesthetic appearance as an appearance of his leg bent, when in fact what was appearing to him was his leg stretched out. (50)

These passages are best understood, in my view, as imprecise descriptions of the fact that, when it comes to the position of one’s limbs, one is subject to perceptual appearances that involve no sensory qualities. That one’s leg is bent is an appearance that doesn’t feel like anything—like anything other than one’s leg’s being bent, that is. The appearance is empty of any sensations, such as pressure, tingling, and the like. Appearances without sensations are simply perceptual beliefs—proprioceptive beliefs, in the present case.5 Anscombe’s claim is not that one is infallible about the position of one’s limbs: she not only acknowledges but emphasizes the possibility of being wrong. Her claim is rather that when one’s proprioceptive beliefs are true, they usually constitute knowledge, because they are connected in the right way with the facts that make them true. Anscombe is thus a reliabilist about proprioceptive knowledge. Her view is that one needs no sensory evidence on which to judge the position of one’s limbs because the position of one’s limbs generates proprioceptive beliefs via a mechanism that is generally reliable: a leg’s being bent reliably causes one to believe that it is bent. I think that recognizing Anscombe’s reliabilism is the key to understanding her view on non-observational knowledge of intentional action. What corresponds to a proprioceptive belief, in the case of intentional action, is the intention itself. Descriptions of one’s intentional actions are, in Anscombe’s phrase, “known by being the content of intention” (53). The notion that knowledge can be embodied in an intention requires explanation and defense, which Anscombe provides as follows. We need to distinguish between two kinds of indicative statements about the future: expressions of belief, such as “I am going to be sick”, and expressions of intention, such as “I am going to take a walk” (1). If someone responds to the statement “I am going to be sick” by asking 5

This interpretation implies that the second passage quoted above contains a slight error. Anscombe should not have said “it is because one has sensations [pl.] that one knows this”. What she should have said is “it is because one has sensation [sing.] that one knows this”. Proprioceptive appearances are received via sensory pathways, blockage of which, as in anesthesia, leaves one unable to judge the position of one’s limbs. But the sensations conveyed by these same pathways are not intrinsic to proprioception.

What Good is a Will?


“Why would you do a thing like that?”, he has misinterpreted the speech act, by failing to recognize it as an expression of belief rather than intention. Conversely, if someone responds to “I am going to take a walk” with “How can you tell?”, he has failed to recognize it as an expression of intention rather than belief. The difference between these statements cannot lie in the former’s being informative and hence potentially knowledge-conveying, since the latter is informative and hence potentially knowledge-conveying as well. As Anscombe puts it, “the indicative (descriptive, informatory) character is not the distinctive mark of ‘predictions’ as opposed to ‘expressions of intention’, as we might at first sight have been tempted to think” (3). The “indicative (descriptive, informatory) character” of a statement expressing an intention indicates that the statement simultaneously expresses knowledge on the part of the speaker. When one says “I am going to take a walk”, one lets the hearer know that one is going to take a walk. One’s assertion is meant to provide the justification in virtue of which the hearer then knows that one is going to take a walk, and it is meant to provide that justification by virtue of expressing one’s own knowledge to the same effect. Hence an expression of intention must at the same time be an expression of knowledge—of something known, in other words, by being the content of intention. In Anscombe’s view, the underlying difference between “I am going to take a walk” and “I am going to be sick”, given that both express knowledge on the part of the speaker, is that the latter expresses speculative knowledge, whereas the former expresses knowledge that is practical, in the sense that it causes the facts that make it true (87). “I am going to be sick” expresses a belief that is caused by evidence of the speaker’s becoming sick, whereas “I am going to take a walk” expresses an intention that causes the speaker to take a walk. Both the belief and the intention may amount to knowledge on the part of the speaker, if they are true and appropriately connected to the facts that make them true. The difference lies in the causal order of the connection. A belief amounts to knowledge if it is appropriately caused by (evidence of the) facts that make it true; an intention amounts to knowledge if it appropriately causes facts that make it true.6 6

Here I am using the purposely vague term “appropriately” to encompass whatever additional conditions are necessary to rule out various things that might defeat an attribution of knowledge. I am not trying to develop a precise epistemology for the knowledge embodied in intention, since the details of such an epistemology


J. David Velleman

5. A Problem for Anscombe and Its Solution The most puzzling part of Anscombe’s view is her claim that what is known by being the content of intention can include, not just how one is moving one’s limbs, or what one is thereby attempting to accomplish, but the fact that one is actually accomplishing it: I think that it is the difficulty of this question that has led some people to say that what one knows as intentional action is only the intention, or possibly also the bodily movement; and that the rest is known by observation to be the result, which was also willed in the intention. But that is a mad account …. Another false avenue of escape is to say that I really ‘do’ in the intentional sense whatever I think I am doing. E.g. if I think I am moving my toe, but it is not actually moving, then I am ‘moving my toe’ in a certain sense, and as for what happens, of course I haven’t any control over that except in an accidental sense. The essential thing is just what has gone on in me, and if what happens coincides with what I ‘do’ in the sphere of intentions, that is just a grace of fate. … But this is nonsense too. (51 f.)

Thus, Anscombe believes that if one is trying to shoot a bull’s-eye, intends to shoot one, and will consequently end up having shot one intentionally, then one already knows without observation that one is shooting a bull’seye, not just that one is intending or trying to do so, or moving one’s limbs with that aim. But how can the content of one’s intention embody knowledge of whether the bull’s-eye is going to be hit? An answer to this question is suggested by the following passage: ‘Why are you pumping?’—‘To replenish the water supply.’ If this was the answer, then we can say ‘He is replenishing the water supply’; unless indeed, he is not. This will appear a tautologous pronouncement; but there is more to it. For if after his saying ‘To replenish the water-supply’ we can say ‘He is replenishing the water-supply’, then this would, in ordinary circumstances, of itself be enough to characterise that as an intentional action. (The qualification is necessary because an intended effect just occasionally comes about by accident.) (38 f.)

The parenthetical remark at the end of this quotation indicates that, in Anscombe’s view, one can bring about an intended result without doing so intentionally if the result comes about accidentally. Hitting the bull’seye by luck doesn’t count as hitting it intentionally. are not relevant to my purposes. All that’s relevant is that the order of causation between facts and knowledge of those facts is reversed from that which is characteristic of speculative knowledge.

What Good is a Will?


Now compare the two passages just quoted, and notice that the concept of accident figures in the first passage as well. In that passage, Anscombe ridicules the view that what I do, and know that I’m doing, when I intentionally move my toe does not include my toe’s actually moving. What’s ridiculous about that view, according to Anscombe, is the implication that my toe’s moving is accidental, ‘a grace of fate’. Anscombe suggests that my toe’s moving is part of my intentional action, and hence part of what I know without observation, because it is a reliable result of my intending to move my toe. The lesson to be drawn from these passages is that the reliability of the connection between intention and movement ought to be necessary both to the movement’s being intentional and to the intention’s being knowledge, according to Anscombe’s reliabilist epistemology. Speculative knowledge, which is caused by the facts, must be caused by them via some reliable mechanism, or it doesn’t qualify as knowledge, after all. If knowledge can also be connected to the facts by virtue of causing them, this connection must be reliable as well. Unless an intention with the content “I’m going to move my toe” reliably causes my toe to move, it won’t amount to practical knowledge. If my toe moves accidentally, I will neither have moved it intentionally nor known that I was moving it. Anscombe’s view, then, is that what is done intentionally and what is known by being the content of intention will tend to coincide, because they are constituted by the same reliable connection and undermined by the same failures of reliability. If your intention to hit the bull’s-eye doesn’t amount to knowledge of the fact that you’re hitting it, the reason is probably that you can’t reliably hit the bull’s-eye; and in that case, your hitting it won’t qualify as intentional, either. If your hitting the bull’s-eye does qualify as intentional, then you must have a reliable way of hitting the bull’s-eye, in which case your intention to hit it is reliably connected to that result and will probably amount to knowledge.7 What you do 7

What is done intentionally is not perfectly coextensive with what is known without observation. Whether something is done intentAiCHTUNGRE onally depends on the reasons for and against doing it—especially against. An unreliable but lucky marksman may not be credited with hitting the bull’s-eye intentionally, but he may be blamed for hitting the President intentionally, simply by virtue of trying to hit the President and despite his lack of a reliable method for doing so. (This example is due to Gilbert Harman. Thanks to Gideon Rosen for reminding me of it.) In neither case does he know without observation that he is hitting his target. If his target is the President, then, what he does intentionally exceeds what he knows without observation. As I understand Anscombe, she believes that


J. David Velleman

intentionally, you tend to do knowingly, because intentional action and knowledge are generally two ends of the same chain of reliable causation. 6. Intention and Belief What I want to preserve in Anscombe’s theory is the claim that intention can embody knowledge of the resulting action, knowledge that is practical in the sense of causing what it represents.8 Anscombe’s view that intention embodies knowledge doesn’t lead her to assert that intention requires belief; rather, she seems to think that knowledge can subsist either in a true belief or in a fulfilled intention. But in viewing intention as potentially embodying knowledge, she clearly conceives of it as similar to belief in being a cognitive commitment to the truth of its propositional object. Anscombe is therefore at odds with philosophers who deny that intention entails belief, because they do so on grounds that would rule out its being any sort of cognitive commitment that could embody knowledge. Consider, for example, Bratman’s argument against the thesis that intention entails belief: First, there seem to be cases in which there is intention in the face of agnosticism about whether one will even try when the time comes. I might intend now to stop at the bookstore on the way home while knowing of my tendency toward absentmindedness—especially once I get on my bike and go into ‘automatic pilot’. If I were to reflect on the matter I would be agnostic about my stopping there, for I know I may well forget. It is not that I believe I will not stop; I just do not believe I will. Second, there seem to be cases in which there is intention in the face of agnosticism about whether one will succeed when one tries. Perhaps I intend to carry out a rescue operation, one that requires a series of difficult steps. I am confident that at each stage I will try my best. But if I were to reflect on the matter, I would have my doubts about success. (Bratman 1987, 37 f.)

Here Bratman argues that intending to stop at the bookstore, or to carry out the rescue operation, is compatible with remaining agnostic about the


intentional action is invariably known without observation. On this point, among others, I believe that Anscombe is mistaken. I do not want to preserve either Anscombe’s reliabilism or her view that intentional action can be analyzed in terms of intention. On the problem of analyzing intentional action, see the preceding note. For an evidentialist rather than reliabilist conception of knowledge without observation, see Velleman 1989, Part I.

What Good is a Will?


truth of the corresponding descriptions, “I am going to stop at the bookstore” or “I am going to carry out the rescue”. He seems to regard the intentions as conative commitments to making these future-tense descriptions true but not as cognitive commitments to their being true. But then if the agent does stop at the bookstore or carry out the rescue on the basis of his intention, he cannot claim to have known that he was going to, since he can hardly claim to have known things with respect to which he was cognitively uncommitted. Bratman must therefore deny Anscombe’s claim that intention embodies knowledge of the resulting action. This denial leaves much of Bratman’s functional account unmotivated. Why, for example, should an agent be rationally obliged to arrange means of carrying out an intention, if he is agnostic about whether he will in fact carry it out? Suppose that I form an intention to fly to Chicago next Tuesday, well knowing that I often forget to take trips that I have planned. (I am even more forgetful than Bratman.) Buying a ticket for my flight to Chicago will turn out to have been a waste of money if I forget to take the trip. If going to Chicago were merely a goal, in competition with my other goals, then the rationality of buying a ticket would depend on the expected benefits of attaining this goal, by buying the ticket, and the expected costs of foregoing other goals, by investing in the ticket and then possibly failing through forgetfulness to attain even this one. Part of the reasoning-centered function of intention, in Bratman’s view, is to cut through such cost-benefit calculations by generating a categorical requirement to identify means of doing what is intended. But why should I be categorically required to invest in means whose benefits I am not yet prepared to believe in? If I am still entertaining the possibility that a ticket will go to waste, why shouldn’t I weigh its expected benefits against those of alternative investments? Bratman also believes that a plan of flying to Chicago on Tuesday rationally constrains my subsequent practical reasoning by ruling out options inconsistent with my being in Chicago—for example, the option of accepting your invitation to dinner on the same day. But if, for all I know, I might forget my flight and still be here on Tuesday evening, then it would certainly be advantageous to have a dinner appointment to replace whatever activities I would then be missing in Chicago. If I form two intentions—to fly to Chicago and to have dinner with you here— then I will have a better chance of doing something that I intend, and a better chance of enjoying myself on Tuesday, just as I can raise my


J. David Velleman

chances of winning a fellowship by applying for two of them, despite the known impossibility of winning both. As I have explained, Bratman argues that rationality forbids such inconsistencies between our intentions, though not between our goals. But why should my intentions be subject to a requirement of consistency if I can remain cognitively uncommitted to their truth? If I am agnostic as to whether I’ll be in Chicago on Tuesday evening, why should my plans for Tuesday evening have to be consistent with my being there? Of course, the requirement of consistency among intentions is essential to their role in coordinating behavior, since behavior cannot be coordinated with inconsistent states of affairs. Yet the coordinating role of intentions would itself come into doubt if intentions did not involve a cognitive commitment. When an intention coordinates behavior, the agent and his associates proceed on the assumption of its being executed—which would be an odd way to proceed if the agent himself were agnostic on the question. If I am agnostic as to whether I will be in Chicago on Tuesday, why should anyone plan or act on the assumption of my being there? And why should anyone hesitate to plan or act in ways inconsistent with that eventuality? Bratman might say that intentions must be consistent because they are agglomerative, in the sense that rationality favors combining an intention to A and an intention to B into a single intention to A and B. Surely, I cannot rationally intend to be in Chicago and have dinner with you here at the same time. This argument would only shift the problem, however, by raising questions about the requirement of agglomerativity. Why are attitudes agglomerative? Beliefs are agglomerative because they aim to fit the world, of which there is just one, in whose complete characterization the contents of all true beliefs are conjoined. The rational pressure to conjoin beliefs is a pressure to fuse them into a single characterization of the single world that all of them aim to fit. An agglomerativity requirement would be equally appropriate for intentions if they had to be jointly executable in a single world. As before, however, the question is why my intentions must be jointly executable, if I can be agnostic as to whether they will be executed. If I can have a plan without believing that it will be carried out in this world, why should I confine myself to planning for a world in which it is carried out? Note finally that Bratman must deny, what is axiomatic for Anscombe, that the natural expression of an intention to do something is the assertion that one will do it. For if Anscombe were right about the natural expression of intention, then Bratman’s view would imply that

What Good is a Will?


someone could sincerely say “I am going to act” without believing it, since the assertion might still be sincere as an expression of intention. What’s more, if Anscombe were right, then the statement “I am going to act but I don’t believe it” would express an intention while reporting the lack of a corresponding belief—a combination of mental states that should not be paradoxical, according to Bratman. Yet such a statement is clearly an instance of Moore’s paradox, because someone who says “I am going to act” without believing it is guilty of insincerity by ordinary conversational standards. Unless Bratman wants to revise those standards, he will have to say that “I am going to act” is not the natural expression of an intention. Perhaps, then, he will say that the natural expression of an intention is “I intend to act”. Bratman might also point out that “I intend to” can be a way of avoiding an outright affirmative answer to the question “Are you going to act?”. But avoiding an outright affirmative answer does not necessarily indicate a lack of belief. After all, one can also avoid an outright affirmative answer by saying “I believe so”, which positively requires that one believe that which one is trying to avoid asserting. Hence expressions used to avoid an outright assertion cannot in general be interpreted as indicating a lack of belief. Now, if one answers the question “Are you going to act?” with “I believe so”, then failing to act will not make one guilty of having told a lie, but it will entail that one has misled a listener who was inclined to adopt or defer to one’s point-of-view. Surely, the same will be true if one answers the question “Are you going to act?” with “I intend to”. If one then fails to act, one can retrospectively defend this reply by saying “Well, I only said that I intended to”; but this defense will be just as narrowly legalistic as “I only said that I believed so”. In either case, one will have raised an expectation of action in a listener who was inclined to adopt or defer to one’s point of view. If intention didn’t entail belief, however, then the statement “I intend to act” wouldn’t license even a credulous listener to expect action. “I intend to” would be similar to “I hope so”—that is, the report of an attitude that warrants no expectation on the part of the listener because it requires none on the part of the speaker.


J. David Velleman

7. What is Intention For? In my view, these incongruities in Bratman’s account can ultimately be traced to the fact that for many of the purposes that Bratman assigns to intention, knowledge of one’s forthcoming actions is what actually does the work. Knowledge of one’s forthcoming actions is what provides the basis for coordination, and it is what constrains the options available to further practical reasoning. Intention would be ill-suited to serve these purposes if it were compatible with agnosticism about one’s forthcoming actions, as Bratman believes. Hence intention would better serve the purposes for which it is designed, according to Bratman, if it were capable of embodying self-knowledge in the manner envisioned by Anscombe. I have now argued that the epistemic role of intention supports the pragmatic functions elucidated by Bratman. Yet I previously considered cases in which no pragmatic purpose is served by the self-knowledge embodied in an intention. Making up one’s mind to take a cookie that has been offered, or making up one’s mind how to vote in the next election, puts one in the position of knowing what one is going to do, but not of being able to coordinate or plan any related activities. These cases therefore raise the question whether the epistemic function of intention is entirely subservient to its pragmatic functions. On the one hand, these cases might exemplify a re-purposing of intention—a use of the attitude for a purpose that, in the normal case, is ancillary to its ultimate purpose but, in the present cases, is the only purpose at stake. That is, intention might have been designed to embody self-knowledge for the sake of facilitating coordination, but it might then be used, on occasion, for the sake of self-knowledge alone, when coordination isn’t necessary. On the other hand, the pragmatic uses of intention might be the ones that exemplify re-purposing. That is, having been designed to embody self-knowledge for its own sake, intention might turn out to yield coordination as a fortuitous byproduct. If these alternative hypotheses seem too fanciful in attributing the human will to intelligent design, never mind: I am not going to argue for either of them. I suspect that the human will is an accident—an absurd and wonderful accident. But as I shall argue, it is an accident that more closely approximates the second rather than the first of the envisioned designs. Attributing the will to accident may seem just as fanciful as attributing it to intelligent design. Any hypotheses about the origins of the will must be closer to creation myths than to scientific theories. Why entertain creation myths at all?

What Good is a Will?


My justification for entertaining creation myths about the will is that such myths are implicit in the functionalist moral psychology that Bratman practices. Bratman seeks to characterize the attitude of intention in terms of the function that the attitude typically serves and the rational standards for serving that function. In this effort to identify the function of intention, he is guided by the assumption that it is a function, not merely in the value-neutral, philosophical sense that it is a causal role, but in the evolutionary sense that it confers some benefit on intentionforming creatures. He seeks to characterize intention, not just by observing what an intention does, but by figuring out what an intention might do that needs doing. He thereby adopts what Daniel Dennett has called the design stance, assuming for methodological purposes that the attitude of intention will turn out to be as an intelligent designer would have fashioned it. My own creation myth about the will, which follows shortly, is meant as an antidote to this methodological assumption.9 My purpose in telling this myth is not so much to gain acceptance for it as to show that the standard myth is optional. Our theorizing about the will should not necessarily be guided by the assumption that this practical faculty is designed to confer practical benefits. It often does confer practical benefits, of course, but those benefits may be accidental and should therefore not be assumed to have governed its design. 8. The Will as a Spandrel On many other occasions I have defended a particular explanation of the causal relation by which an intention is reliably connected to the facts that constitute its fulfillment. Let me take a moment to rephrase that explanation in the terms that I have adopted for the present occasion.10 An agent has a standing desire to understand what he does, and the best way for him to satisfy this desire is not to do anything until he is prepared to understand it, and then to do that which he is prepared to 9 For direct arguments against the assumption, see Fodor 2000, chap. 5. 10 An important difference in terminology is that, on the present occasion, I have avoided claiming that intention entails belief in what is intended. Rather, I have said that intention entails a cognitive commitment to the truth of what is intended—a formulation that leaves open whether that cognitive commitment should be called a belief.


J. David Velleman

understand.11 The act-description embedded in an intention, such as “I am going to help myself to a cookie”, conveys information about what the agent is up to, information that implicitly explains why he is behaving as he is—or, more precisely, would explain why he was behaving as he would be if he now proceeded to reach for something on the platter before him. This act-description thus prepares the agent to understand the behavior of reaching toward the platter and, conversely, to be puzzled by alternative behaviors. (“If I’m helping myself to a cookie, then why am I scratching my shin instead of reaching toward the platter?”) The agent’s desire to understand what he’s doing inhibits him from acting until he has committed himself to the truth of some such description, whereupon it reinforces his antecedent motives for acting in accordance with that description, with the result that he does what he’s prepared to understand and consequently understands what he’s doing. The agent can commit himself to the truth of any one of the actdescriptions that he has some antecedent motives for satisfying, because that commitment will reliably enlist reinforcement for the motives favoring satisfaction of the description. The agent’s commitment to the truth of the description, being a reliable cause of the facts that will make it true, then constitutes knowledge of those facts—practical knowledge, in Anscombe’s terminology. The commitment therefore plays the functional role of an intention. This conception of intention has the consequence that to have a will—that is, a capacity for intentions, so conceived—is to have freedom, in the form of an open future. Because my cognitive commitments to the truth of propositions about my forthcoming behavior reliably cause those propositions to come true, I can commit myself either to the proposition that I will help myself to a cookie or to the proposition that I won’t, and my commitment will amount to knowledge in either case. With respect to my own intentional actions, then, I can invent my knowledge of the future rather than discover it, and my future is therefore open in a sense in which it is closed with respect to other events. My will therefore gives me genuine options, in the form of alternative truths of which I could invent either one.12

11 The explanation offered in this paragraph involves empirical assumptions that, in my experience, strike psychologists as obvious while striking philosophers as bizarre. I summarize some of the psychological evidence in Velleman 2000a. 12 For further elaboration on this theme, see Velleman 2000b.

What Good is a Will?


My present point about this account of intention is that it posits nothing more than the predictable consequences of two motivational states whose utility in the design of a creature is far more general than that of the human will. The first of these states is the creature’s desire or drive to understand what’s going on in its environment. The second of these states is the realization on the part of the creature that it is one of the inhabitants of that environment, a realization that revolutionizes the creature’s conceptual scheme, by giving it an objective conception of itself. If a creature is driven to understand what goes on in its world, and if it also realizes that its own behavior is among those goings-on, then it will be driven to understand its own behavior, and it will have acquired the capacity for intentions. Of course, the creature may not immediately realize that it can understand its behavior by doing what it is prepared to understand, but that realization must naturally come, as the creature frames ideas of things to do, does them, and sees that it then understands what it is doing. Discovering its capacity for intentions is simply a matter of discovering how the egocentrically conceived world of doing things is connected to the objectively conceived world of things understood— which is rather like a dog’s discovering that what it wags is the tail visible to its rear. Once the creature finds its own tail, in the form of its behavioral contributions to the intelligible world, it is ready to start forming intentions. In short, the will emerges as a byproduct of curiosity plus selfawareness, which are fundamental endowments of human nature, designed for purposes far more general than scheduling deliberation and facilitating coordination. Once Nature had made self-aware inquirers, she didn’t have to give them a faculty to serve these specific purposes, because they already had one. I have already suggested one moral of this creation myth. The moral is that if we adopt the design stance in theorizing about the will, we should remember how medieval architects designed spandrels: by accident. We shouldn’t assume that the will is perfectly suited to the purposes that it serves, because it may not have been designed for those purposes, or indeed designed for any purposes at all. A further moral that one might draw from this creation myth is that the will and its freedom are less central to human nature than we previously thought. I decline to draw this moral for two reasons. First, I think that a feature acquired by accident can become central to the nature of a creature by shaping its subsequent evolution, as I imagine that the


J. David Velleman

will may have shaped ours. Second, I think that centrality is partly an evaluative matter, and that we are entitled to value our accidental endowments more than our designed features, if we like. I realize, however, that this question is one on which we are likely to bear out William James’s remark that philosophical disagreements are at bottom differences of temperament.13 The reason why I am willing to view our free will as an accident is that its being an accident wouldn’t undermine my sense of its value, mine being a temperament that welcomes absurdity. Those whose temperament favors order in the human condition may find that the idea of the will’s being an accident dampens their enthusiasm for it. My creation myth does have one additional moral, owing to the fact that human life goes better when informed by an understanding of human nature. We do better at managing our lives if we understand the kind of creature whose life we are managing. In the case of the will, I think, we stand to benefit in the management of our lives from an understanding of how it is related to our motivational psychology, over which it exerts a kind of supervision and control. Such an understanding can indicate how smooth or stormy we should expect the relations between these faculties to be. On this question, Bratman’s rationalist conception of the will counsels optimism. Our capacity for intentions is an enhancement of our capacity for desire-belief motivation, according to Bratman, designed to make us more effective in getting what we want. Of course, the will must sometimes restrain desire-belief motivation, but it does so for the sake of greater desire-satisfaction in the long run. In the long run, then, we can expect harmony between willing and wanting. My absurdist conception of the will is less optimistic. I agree with Bratman that our capacity for intention enables us to realize gains in desire-satisfaction, but I see no reason to assume that it is perfectly or even moderately well suited to this role. Our nature as motivated animals is sometimes enhanced by our nature as self-aware inquirers, but these two sides of our nature can also be at odds—fundamentally at odds, in a way that isn’t resolved in the long run. Managing a human life therefore entails striking an essentially uncomfortable balance between wanting and willing, or between motivation and intention.

13 James makes this remark in the first lecture of Pragmatism (James 1995, 2).

What Good is a Will?


9. Conclusion In closing, I’ll offer two illustrations of this balance and its discomforts. The first is an instance that favors willing, the second an instance that favors wanting. My conception of the will suggests that we should not be surprised by the conflict between desire-satisfaction and the requirements of morality. If the latter requirements inhere, as Kant believed, in the very structure of the will, then we should not expect them to be desire-syntonic, because the will is not just an enhancement to desire-belief motivation. The will makes its own demands, which can be fortuitously instrumental to those of desire-belief motivation, but only sometimes and only imperfectly. To seek desire-based reasons for being moral is therefore a mistake. What we should seek instead are ways of harmonizing these two sides of our nature, by cultivating desires that are at least minimally syntonic with the will. My conception of the will also suggests that, although it can enhance desire-belief motivation, we should also expect it to get in the way, and not just when it malfunctions. Simply by being what it is, the will disrupts many activities that go best when left to the guidance of our motivational nature. I am not saying, with Kant, that instinct would do better than practical reason at securing our very preservation, welfare, and happiness; but I am saying that instinct does better in guiding many activities. The activities that I have in mind have been characterized by the psychologist Mihaly Csikszentmihaly as providing an experience that he calls “flow” (Csikszentmihaly 1990). He conducted research in which subjects were prompted to record their activities, and their feelings about them, at regular intervals during the day. He then identified a category of “optimal experiences” that occur in the course of highly challenging activities in which the subject exercises appropriate skills. Czikszentmihaly writes: When all of a person’s relevant skills are needed to cope with the challenges of a situation, that person’s attention is completely absorbed by the activity. There is no excess psychic energy left over to process any information but what the activity offers. All the attention is concentrated on the relevant stimuli. As a result, one of the most universal and distinctive features of optimal experience takes place: people become so involved in what they are doing


J. David Velleman

that the activity becomes spontaneous, almost automatic; they stop being aware of themselves as separate from the actions they are perACHTUNGREforming. (53)

As Czikszentmihaly goes on to explain, this loss of self-consciousness “does not involve a loss of self, and certainly not a loss of consciousness, but rather, only a loss of consciousness of the self ” (64). Not surprisingly, it also involves a suspension of deliberation: In normal life, we keep interrupting what we do with doubts and questions. ‘Why am I doing this? Should I perhaps be doing something else?’ Repeatedly we question the necessity of our actions, and evaluate critically the reasons for carrying them out. But in flow there is no need to reflect, because the action carries us forward as if by magic. (54)

Activities that provide this experience can of course be initiated deliberately, but then they follow a course determined by sequences of stimulus and skilled response. They require intelligence, but they require it to be expressed directly in behavior rather than in prior decisions; they require us to think with our bodies, without pausing to make up our minds. For this reason, the will cannot enhance these activities; it can only disrupt them. I think that the phenomenon observed by Czikszentmihaly provides indirect support for my conception of the will. To begin with, it illustrates the intimate connection between willing and self-awareness, both of which must be suspended if we are to ‘find flow’ (see Csikszentmihaly 1997). What’s more, the difficulty of finding flow illustrates the self-assertive nature of the will. In order to enter the state of optimal experience that Czikszentmihaly describes, we need to find an engrossing activity that will draw us out of ourselves and thereby silence our deliberations and planning. The elusiveness of such experiences should remind us that will is not just a capacity for control but a drive toward control, which must sometimes be beguiled into relinquishing its hold on our behavior. Having a will entails having a tendency to be willful—a tendency that cannot be explained by conceptions of the will as a passive instrument of practical reasoning. Unlike that passive instrument, the will is a mixed blessing. Bibliography Anscombe, G. E. M. (1963): Intention, Ithaca: Cornell University Press. Bratman, Michael (1987): Intention, Plans, and Practical Reason, Cambridge, MA: Harvard University Press.

What Good is a Will?


Csikszentmihaly, Mihaly (1990): Flow: The Psychology of Optimal Experience, New York: Harper and Row. — (1997): Finding Flow: the Psychology of Engagement with Everyday Life, New York: Basic Books. Falvey, Kevin (2000): “Knowledge in Intention”, in: Philosophical Studies 99, 21 – 44. Fodor, Jerry (2000): The Mind Doesn’t Work That Way: The Scope and Limits of Computational Psychology, Cambridge, MA: MIT Press. Gould, Stephen J./Lewontin, Richard C. (1979): “The Spandrels of San Marco and the Panglossian Paradigm: A Critique of the Adaptationist Programme”, in: Proceedings of the Royal Society of London 20, 581 – 598. Gould, Stephen J. (1997): “The Exaptive Excellence of Spandrels as a Term and Prototype”, in: Proceedings of the National Academy of Science 94, 10750 – 10755. Hursthouse, Rosalind (2000): “Intention”, in: Teichmann, Roger (ed.): Logic, Cause and Action, Cambridge: Cambridge University Press, 83 – 105. James, William (1995): Pragmatism, New York: Dover Publications. Velleman, David J. (1989): Practical Reflection, Princeton: Princeton University Press. — (2000a): “From Self-Psychology to Moral Philosophy”, in: Philosophical Perspectives 14, 349 – 377. — (2000b): “Epistemic Freedom”, in: Velleman (2000), 32 – 55. — (2000c): The Possibility of Practical Reason, Oxford: Oxford University Press. Kant, Immanuel (1956): Groundwork of the Metaphysics of Morals, translated by H. J. Paton, New York: Harper & Row.

On the Ways and Uses of Intending: Lessons from Velleman’s Bratman Critique Neil Roughley 1. Introduction The human will, David Velleman plausibly suggests, is best understood as the faculty of intention. If at least some non-human animals are equipped with a rudimentary belief-desire psychology, then one specific feature of the human life-form lies in our character as intenders. If that is correct, then one might wonder whether evolution has done us a favour kitting us out with this faculty. Michael Bratman has argued that it has, as intentions are attitudes that facilitate intra- and interpersonal coordination of action, as a result of which intenders are more likely to get what they (rationally) want. Velleman sees this as an overly harmonious picture of the relationship between wanting and willing. An adequate characterisation of their relationship should, rather, show why they frequently pull in different directions, giving rise to what one might picturesquely label a cleft in human agency. The reason for this constitutive tension lies in the fact that the will is not, as is frequently assumed—amongst others by Bratman—a conative faculty like desire, but is rather an essentially cognitive matter. According to Velleman, the will is a ‘cognitive commitment’ to what one is going to do, under favourable circumstances a form of knowledge that causes its own truth. Velleman argues that this conception is, paradoxically, supported by Bratman’s delineation of the characteristic functional and normative roles realised by intentions: if intention didn’t involve a cognitive framing of one’s future action, there could be no rational requirements on intenders to form subordinate intentions or to refrain from forming incompatible further intentions. Moreover, the cognitive conception fills a gaping hole in Bratman’s planning theory: because Bratman analyses intention as whatever fulfils certain functional or normative roles within an agent’s future-directed planning practices, he is unable to make adequate sense of present-directed intentions, which appear to play none of the roles—and thus to bring none of the advantages that are characteristic of distal intentions. In contrast, Velleman’s conception of intentions as understandings of one’s behaviour-to-come offers both a unitary analysis of

On the Ways and Uses of Intending


proximal and distal cases and a suggestion as to the general point of intending—that of satisfying the standing desire to know what one is doing. If the desire that motivates intention formation is only one amongst our many motivations, that may explain why our intentions need by no means dovetail unproblematically with the rest of our desires. There are three points in Velleman’s discussion which seem to me to raise genuine problems for Bratman. Firstly, the planning theory does indeed leave us pretty much in the dark as to the character of proximal intentions. Secondly, if Bratman is right that intentions characteristically fulfil certain causal and normative roles, the question is pertinent as to whether this is not so because of some feature of intentions in virtue of which they can generally play these roles. Thirdly, we require an explanation of why willing can not only enable, but also impede agents in the pursuit of their goals. Velleman argues that the solution to each of these problems involves understanding intention as essentially a matter of ‘cognitive commitment’. Here, I disagree. 2. Plan-Independent Intentions Velleman’s first challenge is summed up in his question as to why we bother forming intentions to do something here and now (197 f.).14 This question clothes an important objection to the planning theory in somewhat surprising garb. The objection is straightforward: a theory that analyses intentions as whatever it is that makes certain functional and normative contributions to an agent’s planning processes has difficulties with those attitudes we take to be intentions that don’t fulfil these roles, in particular with proximal intentions. However, Velleman doesn’t phrase his objection in conceptual terms, but rather as a concern about the point of intending. Why, one wonders, does he put things this way? Velleman offers a justification for doing so, namely that asking this question enables him to criticise Bratman’s theory by its own lights. I shall argue in a moment that this strategy is unsuccessful. After arguing this, I will go on to explain why Velleman’s way of raising the issue nevertheless puts us onto an important feature of the concept of intention that is lost from view in Bratman’s approach. 14 Parenthetical references in the text are to Velleman 2007 (in this volume).


Neil Roughley

Velleman sees his question as to the point of intending—or the good of willing—as appropriate within the variant of conceptual functionalism advanced by Bratman. Bratman’s methodology, Velleman claims, is based on the assumption that the analysis of intention should pick out functions “in the evolutionary sense that it confers some benefit on intention-forming creatures” (209). Velleman’s strategy, which aims at hoisting Bratman by his own methodological petard, is however unconvincing as a form of immanent criticism. This is because it conflates two different levels within Bratman’s theory that we ought to be at pains to keep separate. One is the level at which ‘functions’ means the causal and normative roles played by whatever we conceive as an intention. At this level, the notion of “function” is the tool of conceptual analysis. Bratman thinks of intentions as being defined by their characteristically fulfilling functions such as making rational, and typically leading to, means-ends reasoning on the part of their bearers. The second level concerns what Bratman calls intention’s “pragmatic rationale”, namely to facilitate the process of intenders getting what they (rationally) want, a point he also puts in terms of intentions themselves being “universal means” (see Bratman 1987, 28, 53; Bratman 1999, 5 – 6). As far as I can see, the attributions at this second level aren’t meant to have conceptual status, but merely to clarify why we should, in general, be glad to be the bearers of those planning structures that are supposed to be definitive of intending.15 Whether the individual agent sees any point to forming an intention in a particular instance is therefore irrelevant for the question of whether Bratman’s conceptual functionalism can account for why the attitude thus formed counts as an intention. Moreover, the truth of Bratman’s plausible claim that intentions generally help us to realise our rational desires by means of inter- and intrapersonal coordination would not entail that agents either form or should form intentions with the explicit purpose of achieving this aim. In other words, if one accepts Bratman’s methodological assumptions, one need not be worried by the lack of an answer to Velleman’s ‘why bother’ question. The question that should worry a Bratmanian is, rather, 15 This is moreover, a different question to that as to the evolutionary function of intention. Within evolutionary theory, the relevant benefit conferred on a function’s bearers concerns the contribution of that function to their survival or to the distribution of their genes. This idea is independent of whether whatever fulfils the function is of use to its bearer or has a point from his or her perspective.

On the Ways and Uses of Intending


why certain attitudes concerning ones actions here and now should count as intentions when they don’t fulfil the roles that are supposed to be definitive of the attitude.16 However, perhaps the Bratmanian need not be worried. He may point out that Bratman has in fact specified a function that is charACHTUNGREacteristically fulfilled by proximal intentions, namely that of conduct control (see Bratman 1987, 16). This is true, but it raises the question as to whether an attitude that fulfils this particular function should be seen as being of the same type as an attitude that exerts both causal and normative guidance over an agent’s planning. This point is particularly pertinent in view of the fact that the former function is widely thought to be fulfilled by motivationally predominant desires. Compare a case of spontaneous intention formation: someone is walking along the street and, on seeing an acquaintance, forms the intention to greet him and as a result does so. It seems perfectly plausible that this action is a result of the agent’s being more strongly motivated to do so than to perform any other action or omission that is incompatible with doing so. There is thus a good reason for seeing the spontaneously formed intention to greet as simply the conscious desire to utter the greeting, a desire that proves to be motivationally sufficient to cause the agent to greet.17 The possibility that such spontaneously formed intentions might be nothing more than motivationally predominant, consciously occurrent wants is given little consideration by the majority of the protagonists in contemporary debates about the nature of intention. The main reason for this, as far as I can see, is the desire for a unitary theory. Deliberatively formed intentions can clearly not be understood in this way, as they are compatible with stronger countervailing motivation. The question is whether a unitary theory is adequate to the whole range of phenomena that we refer to as intentions, a question to which I would reply in the 16 Although Bratman argues against what a he calls the “strategy of extension” from proximal to distal intentions at work in Anscombe, Davidson and Goldman (Bratman 1987, 7), his planning theory embodies a strategy of extension in the opposite direction. 17 The addition of the requirement that the desire be conscious is necessary if we assume that there are such things as unconscious desires that might take control of our action if they have sufficient motivational strength. There seem to be cases in which the greeting may be voiced unintentionally—say, in the face of a previous decision to ignore the particular person. The unintentional character of the greeting is plausibly explicable by the lack of a consciously occurrent want to take the step.


Neil Roughley

negative. As it happens, this is a point on which Velleman and I are in agreement. In his earlier Practical Reflection, Velleman argues that “intention” is ambiguous, referring on the one hand to what an agent is “decided upon”, on the other hand to “an agent’s ultimate motivating desire” (Velleman 1989, 112). He adds that it is only the former he is interested in, a focus which carries over to “What Good is a Will?”. His example of a proximal intention there, to take a cookie from a plate, is indeed presented as a case of an intention formed by ‘making up ones mind’ or deciding. Note that this need not be the case. Only an agent who is uncertain as to whether to take a cookie need make up his mind. People presumably sometimes form spontaneous, non-decisional intentions to do so and then act in the same way that they frequently greet on the street. Note further that it is only in the case as described by Velleman that the question as to why the agent bothers to form his intention is intelligible. The spontaneous cookie taker, like the spontaneous greeter on the street, doesn’t bother. Neither of them performs any preparatory mental action at all. Only when the agent has gone through the process of decision does the ‘why bother’ question have application. This seems to me to be important not merely because it indicates substantial self-imposed restrictions on the subject matter of Velleman’s article. It also points to what one might see as a proto-analysis of the notion of intention in focus: intentions of the non-spontaneous type are at core the products of decisions about what to do. Moreover, if I have explained correctly how to distinguish between the two cookie takers, we also have an answer to the question as to why the non-spontaneous cookie taker forms an intention: prior to doing so, he is uncertain as to what to do and, as long as that is the case, he is unable to act. If he doesn’t want to just let the opportunity pass, he needs to decide one way or another. That is exactly what distinguishes him from the spontaneous cookie taker, whose activated wants immediately control his action.18 I agree, then, with Velleman that in the case of the decisional cookie taker the answer to the ‘why bother’ question is rooted in the concept of (decisional) intention. The controversial feature of Velleman’s position at this stage is his equation of uncertainty about what to do with uncertainty about what one is going to do, and his corresponding conception of 18 As I have been suggesting we use the terms, Velleman’s claim that a spontaneous action could terminate deliberation (Fn 3, 198), is false. Deliberation, I think we should say, requires uncertainty as to what to do; spontaneous actions exclude such optative uncertainty.

On the Ways and Uses of Intending


decision as cognitive commitment. He sees a disposition to commit ourselves cognitively prior to action as the effect of a standing desire to understand what we are doing. This has the consequence that the formation of (decisional) intentions is uniformly motivated, something he attempts to show by means of a further example of plan-independent intending, that of the voter who forms his intention to vote well in advance (198). According to Velleman, the voter has a special reason to satisfy his standing desire for self-understanding, namely that the matter at hand is of particular importance for who he is. This, Velleman says, should lead voters to want to get clear on the important question as early as possible. In the light of the fact that there are otherwise only costs, but no benefits of advance commitment in such cases, our tendency to decide in advance speaks, he claims, for a cognitive interpretation of intention backed by the desire for self-understanding. Certainly, one reason why voters spend time in advance thinking through who to vote for is that they take political matters to be important. We generally think of our political evaluations as sensibly developed over longer periods of engagement with the relevant issues. Where this is so, when we are called on to give institutional expression to those value judgements, the contents of the intentions to do so are usually fixed by the contents of such stable evaluations. That is one good reason why “everyone [doesn’t] show up in the ‘undecided’ column of preelections polls” (198). Another is that, even when the relevant evaluations aren’t stable, there obviously are benefits to deciding in advance. One such benefit is that one thereby unburdens ones mental capacities for other matters. There is thus little plausibility to the idea that the answer to the ‘why bother’ question here can be derived from conceptual considerations. The electoral case thus does nothing to support a cognitive conception of decision. Moreover, if we need a special reason to activate our standing desire to understand what we are going to do in the electoral case, one wonders why that desire should be thought to be activated before our performance of such trivial actions as cookie taking. According to Velleman, our desire to understand what we are doing restrains us from behaving in ways that would “baffle” us (Velleman 1989, 35). But neither the spontaneous cookie taker nor the spontaneous greeter on the street are in the least bit baffled by what they do. Their tokening a conscious want prior to action guarantees under normal conditions that they would be able to answer the question as to what they were doing, were they to be interrupted in the course of their action and asked. This is, I think, the sense in which we


Neil Roughley

generally ‘know what we are doing’, one of the everyday ideas out of which philosophers attempt to distil a technical concept of intentional action.19 If this is correct, then the everyday expression is not to be taken literally, marking rather the capacity of agents to immediately acquire such knowledge should it be required. But if the spontaneous cookie taker is in no need of a substantial cognitive stand, why should his hesitant counterpart require one? I can see no good reason. Moreover, the phenomenology of such cases is surely that the agent brings his uncertainty to an end by simply opting to take the cookie, where “opting” has the sense explicated by Anthony Kenny: the agent takes an optative stand of the form “Let me take the cookie” (Kenny 1994, 218 – 221). 3. Plan-Facilitating Intentions Velleman, I have been arguing, is right that Bratman’s planning theory has problems with proximal intentions. One way Bratman could respond would be to go disjunctive, reserving the planning theory for distal intentions and equating at least spontaneous proximal intentions with the predominantly motivating conscious wants that characteristically control immediate intentional conduct.20 Another response would be to rethink what functionalism claims is the conceptual status of intention’s roles in planning. The connection of paradigmatic intentions to decision, which looks like an excellent candidate for a conceptual relation, is sidelined by the planning theory,21 although the connection suggests a form of unity among paradigmatic cases of proximal and distal intending. For Velleman, the unity among non-spontaneous intentions is provided by their cognitive core. This, he argues, is revealed not only by focussing on 19 Another is “doing something on purpose”. The fact that the two phrases are clearly not equivalent is one reason why there is such divergence between the proposed explications of “intentional action”. 20 There is a disjunctive element in Bratman’s theory, namely the distinction between intentions and guiding desires (Bratman 1987, 137 – 8). But this distinction ignores the fact that we understand our spontaneous proximal guiding desires as intentions. 21 In his application of the planning theory to cases of so-called double effect, Bratman explicitly rejects ‘the intention-choice principle’, according to which one intends to do whatever one chooses to do (Bratman 1987, 145 and 152 – 155). His grounds are that choosing a package deal that includes side effects does not commit the agent to intending those side effects. I explain why this reasoning is unconvincing in Roughley (in press).

On the Ways and Uses of Intending


the plan-independent character of proximal intentions, but also by clarifying what it is about plan-facilitating intentions that enables them to play their characteristic roles. Velleman’s claim that intention’s conceptually central feature is a matter of cognitive commitment clashes directly with Bratman’s suggestion that intending has no positive doxastic implications (see Bratman 1987, 35 – 41). A large part of “What Good is a Will?” consists in the attempt to show that, by denying intention a cognitive essence, Bratman deprives the planning conception of necessary support: if someone intending to v can be agnostic as to whether she is going to v or not, then, according to Velleman, she cannot be rationally required to settle on means to v-ing or to refrain from intending to do things she takes to be incompatible with v-ing. If I invest in means to some end I am not sure I will be going for, I may be wasting my time, energy or money and if I avoid planning alternatives to a course of action I don’t know if I will be taking, I may end up doing nothing that I have planned and thus I risk dissatisfaction with the way things go (205 – 207). Where there is room for doubt, there is rationally room for weighing up pros and cons. It is only where the agent has no doubt as to what he is going to do, Velleman argues, that he can be under a ‘categorical requirement’ of rationality concerning his further intention formation and practical reasoning. Certainly, Velleman is right that Bratman offers no answer to the question as to what it is about intending that explains such norms as the means-ends principle. The key idea of conceptual functionalism is precisely that mental concepts are realised by whatever fulfils specific ‘functional’ roles. Bratman’s introduction of the idea that ‘function’ should also be understood in normative terms entails the denial that an account of what explains those roles can be given. The question for Velleman is why the feature in virtue of which intentions fulfil these roles must be cognitive.22 He mentions the idea of an optative (‘conative’) commitment to truth-making, parallel to that of 22 This assumption is shared, for instance, by Jay Wallace, for whom the rationality of instrumental reasoning depends not on the putatively cognitive character of intending, but on two connections that he claims hold between intention and belief: firstly, that intending to v entails believing one can v and secondly, that ‘minimally self-aware’ agents will generally, and should rationally develop true, and only true beliefs about their intentions. Given these connections, Wallace claims, the core principle of instrumental reasoning is merely a special case of the requirement of belief consistency. See Wallace 2001, 20 – 23.


Neil Roughley

commitment to truth-taking (205).23 Why he thinks no such attitude can do the job is not clear to me. One way of showing that it can would involve seeing the requirement as derived from the unique, personconstitutive role of the optative stands with which agents terminate episodes of practical deliberation. According to such a view, the agent’s termination of optative uncertainty as to what to do is criterial for where he or she stands on that question. It is, I think, because we take it that such episodes of opting belong essentially to their bearers that the metaphysics of agent causality, where it focuses on the power of decision (see O’Connor 1995, 200 Fn 36), has a certain intuitive plausibility. Obviously, a suggestion along these lines would need to be argued for in detail. But if something of this kind is plausible, then Velleman’s cognitivist suggestion is not the only candidate for an explanation of the intention-centred norms of practical rationality. It is only if one assumes that the commitment component of decision must be cognitive that one might think, as Velleman does, that someone who is not sure if she will v must be “still entertaining the possibility” of v-ing or not v-ing (205). Of course, for someone still in such a state of deliberation, opting for ways to make v-ing possible would indeed be premature. But on an optative understanding of intention, not believing that one will v is not equivalent to being unsure whether to v or not. The former, but not the latter is perfectly compatible with being settled on v-ing. In fact, the claim that intending is essentially a cognitive matter itself raises a ‘why bother’ question: if intending to v really did involve a cognitive commitment to the proposition that one will v, it would be unclear why an intender should concern herself with making sure that she v-s. Sometimes realising an intention’s content can turn out to be a strenuous affair. Where this is the case, an agent has to work at making that content true. But if she believes all along that the proposition that she will v is true anyway, why should she bother to make the effort? Velleman’s answer is that such effort is itself the effect of the belief in conjunction with the agent’s desire to know what she is doing (Velleman 1989, 44 – 51). This basic desire will be satisfied if the belief that she will v turns out to be true. But if the requirement that one realise ones intentions is entirely dependent on the desire to know what one is doing, it looks to be more rational for an agent to revise her ‘cognitive

23 This idea is prominent, for instance, in Broome 1999, 406 – 410.

On the Ways and Uses of Intending


commitments’24 as soon as they look endangered than to stick to them at the risk of getting things wrong. However, we generally believe that rationality recommends us to stick to our intentions unless we take it that we have a good reason to re-enter deliberation (see Bratman 1989, chap. 5 – 6). Velleman supports his position with a further argument that begins not with the requirements of instrumental reasoning, but with a piece of linguistic analysis. Following Anscombe, he argues that the sentence “I am going to v”, which is clearly one of our main ways of expressing an intention, has assertoric force (206). For the non-cognitivist about intentions, the assertoric form simply indicates the parallel between optative and cognitive commitment without speaking for their identity. This parallelism between non-identicals is illustrated by the sentence “I am going to see X tonight, but I don’t intend to”, which is Mooreparadoxical if the first conjoin is read as the expression of an intention, but is not if it is read assertorically. What, then, of Velleman’s own Moore-paradoxical sentence “I’m going to v, but I don’t believe it”? Assuming that the first conjoin is indeed the expression of an intention and not a prediction, we need to pay attention to a grammatical peculiarity of the English language that is apt to mess up our conceptual intuitions. “I don’t believe it”, like “I don’t want to” and “I don’t think so”, involves a surface grammatical transfer of the negative particle from the subordinate clause, where it semantically belongs, to the main clause, where it looks as though it governs the attitudinal verb. This is what grammarians call “transferred negation”. Where we say “I don’t believe I will v”, we normally mean “I believe I will not v”. That, however, is obviously not an expression of agnosticism, but of a positive belief concerning ones not v-ing. To intend to do something whilst believing one will not do it is indeed only possible for an agent characterised by a degree of irrationality that suggests a dual personality syndrome. To set yourself to make true some proposition whilst believing you won’t (or can’t) make it true would be an eminently pointless exercise. The noncognitivist about intention does not need to deny this. Note, further, that Velleman’s switch to the second-person perspective, claiming that a “credulous listener” is licensed to expect action on 24 Velleman’s terminological switch from talk of expectations in 1989 to talk of cognitive commitments (see his footnote 10, 209) leaves room for the claim that the norms associated with intending are explained not by intention’s cognitive component, but by its commitment component.


Neil Roughley

hearing the expression of an intention (207), does not provide him with the support he believes it does. It is simply not true that intention’s facilitation of interpersonal coordination depends on there being a conceptual connection between intending to v and believing one will v. To pick up Velleman’s formulation: the doxastic attitude ‘warranted’ on the part of a listener is not fixed by the doxastic attitude putatively ‘required’ on the part of the speaker. People are, of course, generally justified in expecting that other agents will realise their intentions. But only generally. Whether I should expect a certain person to do what she intends to do will depend on what I know about the person, as well as about the circumstances under which she is planning to act. Agents who are realistic in forming their expectations and strong-willed in carrying out their plans should rationally be relied on more than people who are not. Moreover, in this respect, the difference between intending and hoping to v (207) is far less drastic than is often claimed. Someone who intends to v may well believe that she will v and someone who hopes to v may well believe that her chances of v-ing are fairly minimal. But someone hoping to v cannot possibly believe that she has no chance of ving: hoping is not mere wishful thinking. Indeed, she might well think that she has extremely good chances of v-ing, and someone who knows her, her abilities and motivational capacities, might be more justified in believing that she will v than in believing that another, motivationally unstable person who intends to v will end up v-ing. The evidence that Velleman marshals against Bratman’s planning theory should thus not persuade us to adopt the cognitivist conception of intention. Nevertheless, Velleman has pinpointed a second problem in the planning theory of intention that is rooted in Bratman’s conceptual functionalism. The functionalist assumes that there is no informative answer to the question as to what it is about intentions that confers on them their characteristic causal and normative roles. Velleman’s argument that only cognitive commitment could make sense of intending’s characteristic normative roles doesn’t only presuppose that the functionalist story is incomplete, it also indicates that such an account is pitched at the wrong level: specifying characteristic roles is not conceptually decisive, but sets an explanatory task for conceptual analysis. Velleman broaches that task, but offers what I have argued is an unconvincing answer.

On the Ways and Uses of Intending


4. The Will in the Way Velleman concludes his article by arguing that two forms of tension between wanting and intending are best understood as symptoms of the divide that separates an agent’s conative attitudes from her cognitive commitments to her behaviour-to-come. It is the second of these, the phenomenon of ‘flow’, which I want to conclude by discussing. According to the psychologist Csikszentmihaly, the ‘optimal’ character of ‘flow’, experienced during successful performances involving bodily skills, derives in part from the absence of occurrent self-awareness on the part of the agent. Velleman presents this datum as supporting his claim that willing is essentially a cognitive matter (213 – 214). It is certainly plausible that experience of ‘being one with’ ones activity and ones environment in such cases positively requires the absence of perpetually intervening acts of will. However, it is anything but clear that the concept of the will that gives this claim its plausibility is cognitive in character. Velleman quotes Csikszentmihaly as saying that the flow experience involves “loss of consciousness of self ”, but omits to add that for Csikszentmihaly ‘the self ’ is “the information we use to represent to ourselves who we are” (Csikszentmihaly 1990, 64). In other words, according to Csikszentmihaly, the cognitive features that go missing in flow are general features of the agent’s self-understanding, what he also calls “the image we have of ourselves” (ibid., 63). Moreover, Csikszentmihaly is explicit that the kind of self-awareness absent is not a cognitive relation to ones behaviour-to-come. Quite to the contrary: “To enter flow, goals should be set clearly in advance, so that the athlete knows what he is to do. As the activity progresses, the athlete then knows moment-bymoment what to do next.” (Jackson/Csikszentmihaly 1999, 21) The athlete’s complete involvement with the activity excludes any thoughts about who he is or how others might see him; what it positively requires is “knowing in advance what he is going to do”, both “at the immediate level” and “in long-range terms” (ibid., 22). Far from supporting the claim that the will is essentially a cognitive matter, flow is in fact an excellent counter-example: in flow, the will is absent, but a doubt-free cognitive relation to ones behaviour-to-come is present. Indeed, it is precisely the clarity for the agent of what he is going to do next that enables him to get along without the intervention of the will. What is decisive for flow is the lack of uncertainty and the corresponding lack of the need for deliberation as to what to do next. In other words, the claim for which the phenomenon of flow provides


Neil Roughley

support ex negativo is the claim that willing requires a genetic component, a decision. Whether decisions are essentially cognitive, as Velleman believes, or essentially optative, as I think, is a matter into which the phenomenon of flow provides no insight. Note however that, in spite of the absence of willing, flow phenomena need by no means involve the complete suspension of intention. The professional cyclist in flow who responds to attempts of others to pass him (see Jackson/Csikszentmihaly 1999, 21) surely intends to do so, forming a spontaneous or routine intention to respond appropriately.25 What he doesn’t need to do is to make a decision, forming an intention that is even minimally deliberative. An understanding of the conditions of flow thus supports the claim on which Velleman and I agree: that a unitary theory of intending may be a hindrance to clarifying the phenomena we see in everyday life as appropriately referred to by the term and its cognates. It also points to the fact that we should not equate the will with a faculty for producing intentions of all kinds. On the evidence offered by flow phenomena, it seems that we will where we have formed decisional intentions, whilst intentions with no such decisional pedigree involve no willing. 5. Three Lessons I have been arguing that there are three lessons to be learnt from Velleman’s Bratman critique. Firstly, conceptual functionalism, including its normative variant, should be rejected as a comprehensive analytic methodology in philosophical psychology. The platitudinous truths about intending with which it works themselves set an explanatory task on the psychological, not just on the neurological level. Although Velleman doesn’t claim this, his arguments are, I think, best understood as presupposing it. Secondly, intention should be seen as a disjunctive concept, picking out either the attitudinal product of a decision or an agent’s motivationally strongest conscious want. The intentions at work within flow phenomena are of the latter kind, as are a great many spontaneously formed proximal intentions, whose genesis is not a response to uncertainty as to what to do. Much of what Velleman presents as supporting cognitivism about intentions actually speaks for 25 The authors talk explicitly of “clarity of intention” (Jackson/Csikszentmihaly 1999, 21).

On the Ways and Uses of Intending


the claim that paradigmatic intentions derive from conscious mental steps to dissolve uncertainty about what to do. What is constitutive of this second type of intention is not primarily its characteristic consequences, but its genesis. If these first two lessons are by and large consistent with what Velleman explicitly says in “What Good is a Will?”, the third isn’t. The third lesson to be learnt is that there is nowhere near sufficient evidence to convince the person on the street (or, presumably, Michael Bratman) to abandon the natural understanding of decisional intention formation as an optative—or conative—mode of committing oneself. Certainly, neither proximal intentions nor the phenomenon of flow provide the cognitivist with strong arguments. However, as Velleman’s paper shows, the cognitivist can offer a coherent answer to Kant’s question as to how the means-ends principle comes to exert normative force. Nevertheless, that answer—in short: that we want to have been right about what we thought we were going to do—is highly implausible as an explanation of why we generally do, and should, adopt the means to our ends. Rather, what we seem to strive for, and tend to believe we ought to strive for, is optative or motivational consistency. A consequence of the third lesson from Velleman’s critique is that the non-cognitivist needs to make clearer what the optative commitment consists in that anchors the striving for such consistency. The non-cognitivist therefore still has work to do here. Perhaps this is something on which David Velleman might agree.26 Bibliography Bratman, Michael E. (1987): Intentions, Plans and Practical Reason, Cambridge, MA: Harvard University Press. — (1999) Faces of Intention, Cambridge: Cambridge University Press. Broome, John (1999): “Normative Requirements”, in: Ratio (new series) XII, 398 – 419. Csikszentmihaly, Mihaly (1990): Flow. The Psychology of Optimal Experience, New York: Harper Collins. Jackson, Susan A./Csikszentmihaly, Mihaly (1999): Flow in Sports. The Keys to Optimal Experiences and Performances, Champain, IL: Human Kinetics. Kenny, Anthony (1994): Action, Emotion and Will, Bristol: Thoemmes Press. 26 My thanks to David Velleman for his response to an earlier version of this paper. I am also grateful to the Swiss National Science Foundation for funding the research of which this paper is a product.


Neil Roughley

O’Connor, Timothy (1995): “Agent Causation”, in: O’Connor, Timothy (1995): Agents, Causes, Events. Essays on Indeterminism and Free Will, Oxford: Oxford University Press. Roughley, Neil (in press): “The Double Failure of Double Effect”, in: Lumer, Christoph/Nannini, Sandro (eds.), Intentionality, Deliberation and Autonomy, Aldershot: Ashgate. Velleman, David J. (1989): Practical Reflection, Princeton, NJ: Princeton University Press. — (2007): “What Good is a Will?” (this volume). Wallace, R. Jay (2001): “Normativity, Commitment and Instrumental Reason”, in: Philosophers’ Imprint 1 (3), (last accessed 17. 03. 2006).

Personal Identity and Agency: Towards Analytical Personalism Stefaan E. Cuypers 1. Introduction I claim that there is something deeply wrong with the standard debate between materialist empiricists—defending the bundle theory—, and dualist metaphysicians—defending the ego theory—about personal identity through time. After introducing the problem of personal identity as I see it and diagnosing the influential impersonal solution of Parfit and Perry, I outline a third, innovative theoretical approach— which I call ‘analytical personalism’—to the problem, pointing towards a satisfactory theory both of the nature and the importance of personal identity. 2. The Aporetic Standard Debate If we can compare philosophical debates to the game of chess, then we could say that the standard debate on personal identity leads to a stalemate.1 The match between empiricists (notably Anthony Quinton and Sydney Shoemaker) and metaphysicians (notably Roderick Chisholm and Richard Swinburne) in analytical philosophy ends in a draw since 1

I restrict myself here to the standard debate between the adherents of the two most prominent theories. This restriction means that various non-standard theories will not be discussed. The underlying reason for this restriction is my belief that the non-standard theories either do not provide a clear alternative to the bundle theory or the ego theory, or else they can in some sense be reduced to one of these two. For an excellent introduction to the state of the art of the standard debate, see Shoemaker and Swinburne 1984. Also very helpful are Perry 1975; Oksenberg Rorty 1976; Noonan 2004. Among the non-standard theories of personal identity in contemporary analytical philosophy, one could mention the body theory put forward by Williams 1973, 1 – 81 and the brain theory put forward by Mackie 1976, 173 – 203 and by Nagel 1986, 43 – 45.


Stefaan E. Cuypers

neither of the two views gives a satisfactory account of the nature as well as the importance of personal identity. In light of generally accepted beliefs and values—so-called ‘common sense’—neither the empiricist nor the metaphysical approach provides an adequate view of both the ontology of personal identity and the relevance of personal identity in existential contexts such as responsibility, love and self-interest.2 Proponents of the bundle theory—the mental continuity theory— adhere to the common sense idea of knowable facts and to the scientific idea of respectable entities, but they cannot rationally justify the significant concept of perfect personal identity in existential practices. A close investigation of the empiricist bundle theory leads to the outcome that personal identity is only identity in a loose or conventional sense. However, the belief in perfect or real identity plays a prominent role in all sorts of contexts where personal identity is clearly of some importance. As a result, the empiricist declaration that personal identity is ultimately conventional inevitably clashes with the awareness of the importance of strict personal identity in existential matters such as responsibility, love and self-interest. So, although the empiricist view corresponds on the ontological level to the mechanistic and normal, common sense worldview of contemporary Western civilization, this conventionalistic outlook undermines our ordinary existential self-image. If one accepts this conventionalism, then one takes responsibility, love and self-interest to be practical fictions. That is to say, if this conventionalist empiricism is received as the explanation of how personal identity is constituted, then our common practices and ordinary attitudes, which are based on a conception of personal identity totally at variance with the conventionalist view, are eliminated or at best reduced to pragmatic illusions. Clearly, the bundle theory fails on the side of ‘what matters’. Conversely, adherents of the ego theory—the Cartesian soul-subACHTUNGREstance theory—can indeed provide a rational justification of the significant view of human nature that forms the background of existential practices, but they are far removed from the mechanistic and normal, common sense picture of the world’s make-up because of the appeal to further facts and unknowable entities. According to this view, the deep self is not a conventional fiction but an ontological reality. This leads to the result that personal identity is always a matter of real essence. ConACHTUNGREsequently, the metaphysical point of view provides a philosophical 2

For the full argumentation to substantiate this claim, see Cuypers 2001, 18 – 33; Cuypers 2005.

Personal Identity and Agency: Towards Analytical Personalism


foundation for the significance of strict personal identity in rational, emotional and moral practices. But while the metaphysical view supports our ordinary existential self-awareness, on the ontological level, this essentialistic viewpoint creates a miraculous awareness of transcendent spirituality. If one accepts this essentialism, then one postulates an ontological fiction. That is to say, if this essentialist metaphysics is received as an explanation of how personal identity is constituted, then our personal identity becomes clouded with ontological obscurity as it vanishes into an ontological darkness beyond the reach of experience. Clearly, the ego theory fails on the side of ‘what there is’. Herein lies the aporia that the standard debate on personal identity in analytical philosophy runs up against: In the light of our common sense as well as scientific beliefs about the world and our ordinary practical values, neither the empiricist approach nor the metaphysical approach can adequately account for both the ontology of personal identity and the importance of personal identity in existential contexts which involve responsibility, love, self-interest and the like. Again, the bundle theory gives rise to a destructive conventionalism, whereas the ego theory implies an obscure essentialism. ‘Every man of common sense’ is either stranded on the empiricist reef Scylla or sinks into the Charybdis of the metaphysical whirlpool.3 Are analytical philosophers forcing common sense to live under illusions? Must common sense make a choice between a practical or an ontological fiction? Until the present day, this aporia has not yet been given an adequate solution within the framework of analytical philosophy. The analytical problem of personal identity is, then, a real problem. 3. Reductionism and Instinctive Personalism In an attempt to find a way out of the dramatic situation into which the standard debate has fallen, some radical empiricists—specifically Parfit and Perry—have proposed, within the context of the mental continuity theory, a reductionistic or impersonal solution to the problem of personal identity.4 This solution consists of an extreme radicalization of the bundle theory of personal identity. What the reductionist emphatically denies is that persons are separately existing entities, distinct from the existence of 3 4

The expression “every man of common sense” is taken from Reid 1785, 108. See Parfit 1971; Parfit 1984, 199 – 347; Perry 2002, 84 – 166.


Stefaan E. Cuypers

experiences. Persons do not exist, in the sense that they do not exist independently, separated from their experiences. There are no further facts apart from the facts about experiences. The reductionist radically reduces the existence of persons to the existence of experiences. Persons are not fundamental. Unlike experiences, persons do not belong to the makeup of reality. Persons exist impersonally: bundles of experiences exist, but separate and underlying substrates of experiences do not exist. In the latter (metaphysical) sense, persons do not exist. Accordingly, personal identity is not so much constituted by as reduced to mental continuity. Persons are mental bundles without qualification; persons are just series of experiences; persons are streams of consciousness and nothing more. Personal identity is absorbed completely by mental continuity; in itself, apart from mental continuity, personal identity does not exist. What common sense calls ‘personal identity’ is really just mental continuity. Consequently, nothing more than mental continuity (and connectedness), according to the reductionist, is important for the continued existence— the survival—of the person. From the perspective of the impersonal theory, personal identity is not a real and important matter, but a nominal and trivial one. The radical empiricist—the reductionist—interprets the aporia of the standard debate about personal identity in analytical philosophy as a dilemma. Faced with the dilemma between conventionalism or essentialism, he chooses firmly for an empiricist conventionalism that he develops into an impersonal theory of personal identity. Metaphysical essentialism is out of fashion and has become totally implausible in (post)modern society. The reductionist sticks with his decision, and tries to live not only with its theoretical consequences, but also with its practical consequences. Although the common sense view cannot be rationally justified, our fictional belief in personal identity can be explained in terms of reductive naturalism. Responsibility, love and selfinterest are not mere practical fictions, but our rational, emotional and moral practices must nevertheless be radically reconceived and revised in light of the impersonal conception of the self.5 But someone who thought through the reductionist idea that a person is nothing more than a series of experiences to its extreme conclusions would end up being bewildered, and even dismayed. For, on closer inspection, the impersonal solution to the analytical problem of 5

For a further elaboration of the impersonal solution and attendant reductive revision of our existential practices, see Cuypers 2001, 38 – 55.

Personal Identity and Agency: Towards Analytical Personalism


personal identity ultimately leads to a serious distortion and even complete destruction of the personalist descriptive metaphysics and the personalist moral and emotional reactive attitudes of ‘every man of common sense’.6 The reductionist idea that persons exist as impersonally as raindrops is directly at odds with our intuitive, common sense conception of the nature of persons. And the revisionist idea that personal identity in existential contexts such as responsibility, love and self-interest is a mere matter of words sharply conflicts with our normal sense of value and sensibility. Of course, the fact that the impersonal theory does not square with the common sense view does not in itself constitute a philosophical argument against this theory. But our ‘instinctive’ and generally accepted beliefs and values seem to form the only basis upon which an adequate view of both the nature and the importance of personal identity can be philosophically constructed. The reductionist, however, rejects the instinctive personalism of common sense as an illusion. To opt for empiricist conventionalism ultimately amounts to acknowledging the fictionalism of the common sense view. Is the ordinary person, when he sticks to his guns, then forced to live with the fiction of personal identity? To salvage instinctive personalism by revitalizing the Cartesian ego cogito, or another variant of this metaphysical idea, seems an overreaction. To postulate a substantial (or transcendental) I is simply to replace the practical fiction of personal identity with the ontological mystery of personal identity. Is the ordinary person at the end of the day faced with the choice between either the fiction or the mystery of personal identity? The analytical problem of personal identity, as formulated at the end of the previous section, is not really resolved by the impersonal theory. On the contrary, the empiricist position in the standard debate simply becomes more extreme and hardened; the radical impersonal solution only adds fuel to the fire. In other words, the aporia of the standard debate is not a dilemma, but an impasse. There are different possible reactions to this deadlock. One reaction is to accuse analytical philosophy of boundless navet, to abandon it altogether and to look for inspiration in German idealism, existential phenomenology, process philosophy, French deconstructionism or psychoanalysis for another—perhaps better?— conception of personal identity. My reaction is less desperate, because I 6

The label “descriptive metaphysics” stems from Strawson 1959, the expression “reactive attitudes” from Strawson 1962; see further in the text for their meaning.


Stefaan E. Cuypers

am convinced that a solution to the ‘identity crisis’ in analytical anthropology, or at least the inception of one, can be found inside the framework of analytical thought itself. The instinctive personalism of common sense need not be rejected as a myth nor exalted as a miracle. The person is not a conventional fiction nor an essentialistic illusion, but rather—as I will try to show in the rest of this chapter—a tangible reality in the intersubjective world. 4. Philosophical Atomism and Descriptive Metaphysiscs There is something seriously amiss with the standard debate on personal identity in contemporary analytical philosophy. As set out in the two previous sections, in view of the commonsensical (and scientific) beliefs about the world and ordinary practical values, neither the (radical) empiricist nor the metaphysical approach can adequately account for both the ontology of personal identity and its importance in existential contexts which involve responsibility, love, prudential self-interest, and the like. To my mind, therefore, the analytical problem of personal identity—as articulated at the end of section 2—still remains unresolved. If the ego theory were the sole alternative to the bundle theory, or the impersonal variant of this empiricist approach, then it would follow that the impasse against which the standard debate runs up leads to the catastrophic conclusion that the analytical framework of thought is fundamentally unable to deal with the problem of personal identity. For that reason, I think that we cannot refuse to accept Coburn’s conclusion: (T)hat our [standard] ways of thinking about the problem of personal identity are radically defective and that what we need is an approach that is orthogonal to the ones that have heretofore filled the literature; not more epicycles, but a conceptual analogue of the Copernican Revolution. (Coburn 1985, 402 – 403)

I firmly believe that such a conceptual ‘Copernican Revolution’ in the debate on personal identity is possible within the framework of analytical thought itself. Fortunately there is another analytical alternative to the bundle theory, or its impersonal attendant, as well as to the ego theory of personal identity. The bits and pieces of this alternative—or at least the beginnings of such an alternative—lie scattered around in contemporary analytical anthropology. To escape from the disastrous situation into

Personal Identity and Agency: Towards Analytical Personalism


which the standard debate has fallen, one cannot but embark upon the attempt to construct an alternative solution to the problem of personal identity. In this chapter, however, I have no ambition at all to directly solve the problem of personal identity in an alternative way. Instead, I will limit myself to a necessary preparatory task, namely the sketch of an overall view in terms of which the standard debate can be diagnosed and offered a therapy.7 My diagnosis of what is wrong with the standard debate on personal identity suggests that this debate is severely infected with philosophical atomism. My central claim is that both the bundle theory and the ego theory are at bottom atomistic doctrines. These two traditionally opposed theories have much more in common than their proponents on both sides realize or are willing to admit. The empiricist approach and the metaphysical approach share, in particular, the selfsame epistemological (and semantical) presuppositions of philosophical atomism. Moreover, in a certain (structural) sense they even have the selfsame atomistic ontological presuppositions in common. Both the bundle theory—exemplifying ‘logical atomism’—and the ego theory—exemplifying ‘Cartesian atomism’—are grounded in the perceptual model of self-knowledge.8 Knowing oneself has to be explained by something like perceiving or observing one’s own mind and its contents. One of the most salient consequences of the perceptual model of self-knowledge is that the self is formally construed as a private object. If the soul-substance or the bundle of experiences is an object of introspective knowledge by acquaintance, then the self can be known directly and immediately to one and only one person and, consequently, must be private to each separate person. Consequently, the atomistic problem of personal identity is the problem of the self-identity of the first person. Nothing but the self—the mind—of the first person can be of relevance for the constitution of his identity. Both the bundle theory and the ego theory are built not only upon the selfsame epistemological foundations, but in a structural sense also upon the selfsame ontological foundations. There are many particulars—indivisible ‘atoms’—separated from each other which stand to one another in external relations. This 7 8

For a more extensive diagnosis and therapy of the standard debate, see Cuypers 2001, 58 – 73. I borrow the term “logical atomism” from Russell 1918 and the term “Cartesian atomism” from Zemach 1970. I shall allow myself to use these terms in a loose and broad sense.


Stefaan E. Cuypers

fundamental maxim in the ontology of philosophical atomism applies to the constitution of both worlds, the physical as well as the psychological. In the light of this atomistic doctrine the self is either a spiritual atom (Cartesian ego) or a collection of mental atoms (bundle of experiences) objectively given in introspection. Both kinds of atoms structurally fall under the same category of psychological substance and because of their self-subsistent or static character, both the self and the (bundle of ) experiences can exist separately from other selves and experiences. In sum, the portrait of the self which emerges from philosophical atomism pictures the self as a non-bodily, private, and static object with which the first person is intimately acquainted. The single most important source of any atomistic theory of personal identity is the perceptual model of self-knowledge. If it can be shown that this model is radically inadequate and ultimately untenable, then the pivotal motivation for upholding such a theory crumbles. Of course, the denial that there is any introspective perception of the self dates back to David Hume and has remained ever since a vital tenet of empiricist, and thus logical atomistic, theories of personal identity. Without giving an argument, Hume asserted that the experiences (sensations, feelings, images and the like) are much more accessible to the inner sense than the obscure subject of the experiences (Hume 1739, 161 f.). However, beyond Hume and more comprehensive than Hume’s denial, arguments for what Hume merely asserts can be given, I think, not only to underpin his denial of the self but also to support the perhaps surprising denial that there is any introspective perception of the experiences themselves. 9 The first person does not inwardly perceive anything whatever: neither the self nor the experiences. Let it be clear that this is not a denial of the existence of self-knowledge as such, but only of a certain model of self-knowledge, namely the perceptual model which strongly invites one or other atomistic theory of personal identity. The upshot of the complete dethronement of introspective acquaintance with the self and its experiences is that the main road to the idea of the self as private object will be fully blocked and, consequently, that both atomistic theories of personal identity—the bundle theory and the ego theory—will to a great extent be undermined. If the perceptual model of self-knowledge is bankrupt, then it follows that the atomistic idea of the self as a private object of introspective knowledge is untenable. The dethronement of introspective acquaintance 9

For some of these arguments, see Shoemaker 1986.

Personal Identity and Agency: Towards Analytical Personalism


with the self and its experiences makes it impossible still to interpret the problem of personal identity as the problem of the self-identity of the first person. In addition, such an epistemological criticism casts doubt on the atomistic ontological doctrine of the separateness of selves and experiences. Thus, if such an atomism is accepted, then the problem of personal identity reposes from the start on an intellectual illusion. Admittedly, further arguments—semantical and ontological ones—are required to show that interpreting the problem of personal identity in atomistic terms is a false start; but the initial and most important step in this ‘case against atomism’ has been taken. What, then, is the therapy for what is wrong with the contemporary debate on personal identity? The vacant starting position, to my mind, should be allocated to Peter Strawson’s descriptive metaphysics of the person. Such a metaphysics describes the actual structure of our thought about the subject of experiences (the person) and the experiences themselves. A descriptive metaphysics has to be contrasted with both a revisionary and a validatory metaphysics.10 Instead of describing our actual conceptual scheme, a revisionary metaphysics tries to change it and produce a better conceptual structure, while a validatory one tries to give a deeper ‘metaphysical’ justification for what we believe on instinct. In the context of my interpretation of the standard debate on personal identity in terms of philosophical atomism, the first of these oppositional metaphysics corresponds to what I have called ‘logical atomism’, while the second corresponds to ‘Cartesian atomism’. Looked upon in these terms, the bundle theory proposes a conventionalist, empiricist revision of our actual concept of personal identity, whereas the ego theory offers an essentialist, dualist validation of our common sense concept of personal identity. However, the bankruptcy of these two theories of the same atomistic manufacturing makes philosophically attractive a preliminary investment in a descriptive metaphysics of the person on the road to an alternative solution to the problem of personal identity. Strawson’s view of the concept of a person contains, as I try to show below, the necessary scaffolding for such a nonatomistic architecture of personal identity and, consequently, suggests a way out of the impasse in which the contemporary debate seems to be stuck. Logical atomism and Cartesian atomism have an important common denominator, namely, the belief that the concept of a person is analysable 10 Strawson 1985, 23; see also Strawson 1959, 9 – 11.


Stefaan E. Cuypers

or reducible. Both the bundle theorist and the ego theorist analyze the concept of a person in terms of the concept of a body and that of a mind (a self ). On both accounts a person essentially is a mind which is only contingently linked with a particular body.11 The bundle theorist reduces the concept of a mind still further to the concept of (a bundle of ) experiences, whereas the ego theorist insists on the unanalysability of the concept of a mind (a soul-substance). Thus Cartesian atomism is based on a single reduction of the concept of a person and logical atomism on a double one. However, if there is some truth in the criticisms of philosophical atomism mounted above, then this reductionism seems in principle to be unwarranted. It seems, therefore, that the following conclusion is compelling:12 (W)hat we have to acknowledge, in order to begin to free ourselves from these difficulties, is the primitiveness of the concept of a person. What I mean by the concept of a person is the concept of a type of entity such that both predicates ascribing states of consciousness and predicates ascribing corporeal characteristics, a physical situation andc. are equally applicable to a single individual of that single type. (Strawson 1959, 101 f.) 11 In this sense even materialistic empiricists adhere to soft dualism in that they require only that a bundle of experiences be supervenient upon an appropriate kind of body or brain (or reliable cause whatsoever) and not upon one particular body. Although a particular bundle of experiences always has to supervene on one or another particular material substratum of the appropriate kind, during its lifetime it can very well change from one particular substratum to another without losing its unity and identity, at least as long as there exists mental continuity between the different substrata. For this soft dualism, see, for example, Parfit 1984, 209, 282 – 287. 12 This conclusion is ultimately founded on two logico-ontological principles of Strawson’s descriptive metaphysics, both of which impose constraints on our actual use of the concept of experiences (mental states and events) and the concept of that to which experiences are ascribed (the subject). The first principle states that “identifying references to ‘private particulars’ depend on identifying references to particulars of another type altogether, namely persons.” (Strawson 1959, 41) In other words, individual (token) mental states—for example, this pain here and now—cannot even in principle be individuated and identified without reference to the subject of which they are states. The second principle states that “… it is a necessary condition of one’s ascribing states of consciousness, experiences, to oneself, in the way one does, that one should also ascribe them, or be prepared to ascribe them, to others who are not oneself.” (Strawson 1959, 99) In other words, kinds (types) of mental states—for example, pain—cannot even in principle be ascribed to oneself without the possibility of ascribing them to others. For further elaboration of, and argumentation for these fundamental principles, see Strawson 1958, 15 – 58, 99 – 101; Cuypers 2001, 68 – 70.

Personal Identity and Agency: Towards Analytical Personalism


Strawson calls predicates of the first kind P-predicates and predicates of the second kind M-predicates; the former are literally applicable only to persons, while the latter are also applicable to other material particulars which are not persons. P-predicates ascribing sensations, thoughts and actions should apply to the very same thing—the person—to which Mpredicates ascribing corporeal characteristics, a physical situation, etc. apply. This is so because the applicability of P-predicates (experiences) requires the possibility of the identification of that to which experiences are ascribed. This requirement implies that the subject of experiences has to be or to possess an entity, the identification of which should cause no problem within our spatio-temporal framework. In Strawson’s descriptive metaphysics material bodies occupy the basic positions in this framework: material bodies and bodily persons are basic particulars (see Strawson 1959, 23 – 30 and 38 – 40). Therefore, if persons (subjects of experiences) are or possess entities to which M-predicates apply, then there is no problem of identification at all because their material bodies can readily be individuated and (re)identified by ordinary physical criteria. In this sense persons are quite literally public objects of perception: their bodies can be seen, heard and touched. However, this does not mean that persons are identical with material bodies. For one thing, P-predicates which are applicable to persons cannot be applied to material bodies. In a descriptive metaphysics the concept of a person is neither reducible to the concept of a mind nor to the concept of a body. The concept of a person is logically prior to both that of a mind and that of a body.13 The concepts of a mind and a body are secondary, non-primitive concepts which are to be analyzed in terms of the primary, primitive concept of a person. Thus a living human body always presents itself as a personal body and a conscious mind always as a personal mind. Since a person has both a bodily aspect and a mental aspect, he can appear either as a bodily person or as a ‘minded’ person, according to which of these aspects is emphasized. To put it somewhat differently, a person is not a compound of parts but an indivisible unity to which both P-predicates and M-predicates are equally applicable. The admission of the primitiveness of the concept of a person requires, in other words, the acceptance of a sui generis type of entity. Among the things that exist there are persons. 13 For Strawson’s own statement of the logical derivativeness of the concept of a body, see Strawson 1980, 272 – 273.


Stefaan E. Cuypers

A reductive analysis of the concept of a person is generally thought to be clarifying and informative (X = Y), whereas the acknowledgement of the primitiveness of that concept seems to be obscure and uninstructive (X = X). What, then, makes the primitive concept of a person intelligible? How is such a concept possible at all? I concisely mention two very important ideas of Strawson which point to an answer. First, the primitive concept of a person is primarily the concept of an agent: “What I am suggesting is that it is easier to understand how we can see each other, and ourselves, as persons, if we think first of the fact that we act, and act on each other, and act in accordance with a common human nature.” (Strawson 1959, 112) The special ontological status of a person can best be appreciated if one pays heed to the fact that a person is not so much a passive or static being but rather an active or dynamic one. For the concept of an action equally seems to have a logically primitive character. An action—for example, writing a letter—involves both an intention to act and a bodily movement, without the action being reducible to either aspect or to a combination of both. When it is realized that the concept of action is analysable neither in terms of the concept of (pure) intention nor in terms of the concept of bodily movement, it becomes intelligible why the concept of the centre of agency—the person—should be unanalysable as well. Second, the primitive concept of a person is first and foremost the concept of an object of reactive attitudes and feelings: “(W)e should think … of the kind of importance we attach to the attitudes and intentions towards us of those who stand in these relationships to us, and of the kinds of reactive attitudes and feelings to which we ourselves are prone.”14 We not only act, and act on each other, but also react to each other in accordance with a common human nature. Now, these moral and emotional reactions are primitive in the sense of not being founded on a belief about the (inner) constitution of persons. We do not morally and emotionally react to other people because they have bodies of a certain kind or because they have minds of a certain type; we simply and immediately react to other people as other persons. So the primitiveness of the concept of the object of these reactions—the person—corresponds with the primitiveness of the reactive attitudes and feelings themselves. In 14 Strawson 1962, 6. Strawson’s idea of reactive attitudes and feelings is very similar to Wittgenstein’s idea of an attitude towards a soul. For an exciting exploration of these two closely related ideas, see Cockburn 1990, 3 – 52.

Personal Identity and Agency: Towards Analytical Personalism


other words, it is because people quasi-instinctively react to one another as persons that the primitive concept of a person is possible at all. In sum, the portrait of the person which emerges from Strawson’s descriptive metaphysics pictures the person as a bodily, public, and dynamic agent who engages with other persons and the world. 5. Bodily, Agential, and Narrative Identity Starting from the building blocks made available by Strawson’s descriptive metaphysics of the person, it is possible, I think, to construct a more satisfactory theory of both the nature and the importance of personal identity. For, if a person initially is defined as a bodily, public and dynamic agent, then we do not have to go beyond the facts as we know them to account for the significant role which the person plays in our ordinary rational, moral and emotional practices. When this portrait of the person replaces that of the atomistic self, the moral agent enters on the scene as a commonsensical reality. Let me elaborate on this fundamental Strawsonian insight. Strawson’s non-atomistic view conceptually describes the place of persons as moral agents in our commonsensical scheme of reality. Such a descriptive metaphysics stays in line with our ‘instinctive’ and generally accepted beliefs and values. In view of the fact that a Strawsonian view acknowledges the primitiveness of the concept of a person—persons as basic particulars have a sui generis existence—it is neither revisionary nor validatory. That is to say, such a view is neither conventionalist in the manner of the empiricist bundle theory nor essentialist in the same way as the dualist ego theory. As a consequence, Strawson’s non-atomistic view of the person is in full harmony with the instinctive personalism of common sense: every man of common sense cannot but subscribe to Strawson’s personalist descriptive metaphysics and his personalist moral and emotional reactive attitudes. So, given the Strawsonian view, the instinctive personalism of common sense need not be discarded as a myth nor exalted as a miracle. As a result, a Strawsonian view of the person is, I submit, the only viable basis upon which an adequate view of both the nature and the importance of personal identity can be philosophically constructed. Since the person as a moral agent is a tangible reality in the intersubjective world of common sense and science, one does not have to postulate a substantial (or transcendental) I in order to account for the importance of personalism in rational, emotional and


Stefaan E. Cuypers

moral contexts. Because the person as a commensensical reality is the subject of (moral) reactive attitudes and feelings, a Strawsonian view can provide a rational justification for the common sense concept of the person contained in existential practices which involve responsibility, love, prudential self-interest, and the like. So, given the Strawsonian view, the ordinary person is not faced with the choice between either a practical fiction or an ontological mystery. In sum, Strawson’s view of the person promisingly points towards an escape route out of the impasse into which the standard debate on personal identity has become stuck. I have reached this promising consequence and the conclusion stated above—that the person is a bodily, public and dynamic agent who engages with other persons and the world—within the bounds of analytical philosophy. To carry out my diagnosis of, and to suggest a therapy for what is wrong with the standard debate on personal identity, I have not invoked any other philosophical tradition, such as German idealism, process philosophy, existential phenomenology or postmodern deconstructionism. Rather, my refutation of philosophical atomism and my subsequent employment of descriptive metaphysics have taken place within the framework of mainstream 20th-century English-language philosophy. In conclusion, I sketch the way in which a non-atomistic theory of personal identity could be developed out of Strawson’s preliminary descriptive metaphysics of the person. Since such a theory builds upon a personalist metaphysics which honours the instinctive personalism of common sense within an analytical framework, I call this position analytical personalism. 15 Strawson’s descriptive metaphysics is an excellent tool to melt down the two erroneous positions in the standard debate on personal identity and, subsequently, to forge the scaffolding of a ‘Copernican Revolution’ in that debate. Although I have no ambition to completely solve the problem of personal identity, analytical personalism constitutes, to my mind, a serious analytical alternative to the bundle theory, or its impersonal attendant, as well as to the ego theory. According to analytical personalism, the phenomenon of personal identity is complex and multi-layered as it comprises bodily identity, agential identity and narrative identity. Unlike the bundle theory and ego theory, my alternative view is thus a hybrid view of personal identity. My only 15 Analytical personalism should be distinguished from both the Catholic personalism of, for example, Emmanuel Mounier (1950) and the continental personalism of, for example, Max Scheler (1913 – 1916), although close connections and significant parallels might exist between these varieties of personalism.

Personal Identity and Agency: Towards Analytical Personalism


limited goal here is to delineate the beginnings of such an alternative in order to escape from the disastrous situation into which the standard debate has fallen. I will briefly highlight the essential features of this alternative view. According to Strawson’s descriptive metaphysics, the concept of a person is neither reducible to that of a mind nor to that of a body. Furthermore, this primitive concept of a person is that of a dynamic agent related to the public world. In addition, it is the concept of an object of (moral) reactive attitudes and feelings. Hence, on a Strawsonian view, a person is a moral agent who has a body and a mind. Setting aside the aspect of moral agency for the moment, saying that the primitive concept of a person is that of an indivisible unity to which both P-predicates and M-predicates are equally applicable is the same as saying that such a concept conceives of a person as a substantial psychophysical unity. Accordingly, personal identity consists in the identity of a substantial psychophysical unity. Now the best interpretation of this psychophysical personalism in contemporary analytical anthropology has, in my opinion, been given by David Wiggins (1980). Wiggins’s interpretation is neo-Aristotelian in the sense that it conceives of a person as a living human organism with the powers of intellect, will and memory—or, for short, as a rational animal: (A) persisting material entity essentially endowed with the biological potentiality for the exercise of all the faculties and capacities conceptually constitutive of personhood—sentience, desire, belief, motion, memory, and the various other elements which are involved in the particular mode of activity that marks the extension of the concept of person. (Wiggins 1980, 160)

Correspondingly, personal identity consists in the identity of a rational animal. Let me spell out a bit this first essential ingredient of analytical personalism. An essential component of personal identity is animal identity which, of course, comprises bodily identity. Because a person is a psychophysical unity (and a basic particular), personal identity necessarily involves ‘incarnational’ identity. The view that bodily identity is always a necessary condition of personal identity goes back to Bernard Williams’s nonstandard corporeal theory of personal identity (Williams 1973, 1 – 81). On this view, person X today is one and the same person as person Y at some past or future time only if X is bodily continuous with Y; that is to say, only if X’s body is causally continuous with Y’s body in the unitary spatio-temporal framework.


Stefaan E. Cuypers

However, what is needed to analyse this incarnational identity is not the atomistic concept of res extensa but the non-Cartesian, Aristotelian concept of a living human organism, as revitalized by Wiggins. Correspondingly, the essential bodily aspect of personal identity does not so much depend upon the identity of ‘a lump or other quantity of matter’ but rather on the spatio-temporal (and causal) continuity of a personal body: A person is material in the sense of being essentially constituted [realized, composed] by matter; but in some strict and different sense of ‘material’, viz. being definable or properly describable in terms of the concepts of the sciences of matter (physics, chemistry, and biology even) person is not a material concept. … For the continuity principle defines a material entity in the ‘matter-constituted’ [‘matter-realized’, ‘matter-composed’] sense of ‘material’, while leaving it possible for the concept of person to be primitive relative to the concepts that pull their weight in the sciences of matter and primitive relative to the concept human body. If we understand what a living person or an animal is, then we may define the body of one as that which realizes or constitutes it while it is alive and will be left over when, succumbing to entropy, it dies. (Wiggins 1980, 164)

Analytical, psychophysical personalism emphatically is not mere animalism. On the latter view, I just am (identical with) my body; I am just a living human organism—a member of the primate species Homo sapiens—which began to exist as a single cell (a zygote) and which will go on existing until its biological death.16 The ontological necessary and sufficient condition for personal identity is simply that of a living human organism: person X today is one and the same person as person Y at some past or future time if and only if X’s body is identical with Y’s body; that is, if and only if X is bodily continuous with Y. The difference between analytical personalism and mere animalism can be explained by making use of the distinction between the ‘is’ of identity and the ‘is’ of constitution (realization, composition). The explanation draws on an analogy with the relation between a statue—for example, Rodin’s The Thinker—and the lump of bronze which constitutes it. The relation between a person and his body is like that between The Thinker and the lump of bronze of which it is composed. Because there are changes which The Thinker can survive but which the lump of bronze cannot (for example, losing its left hand), and vice versa (for example, being melted and solidified again as one of The Citizens of Calais), the statue and the lump of bronze are 16 For variations on this animal view, see Snowdon 1990; Snowdon 1995; Thomson 1987; Thomson 1997; Olson 1997; Olson 2003.

Personal Identity and Agency: Towards Analytical Personalism


distinct objects—that is, are not identical with one another—although they are not separate objects, as would be The Thinker and one of The Citizens of Calais. Statues are not identical with lumps of bronze, even though they are constituted by them. Likewise, persons are not identical with their bodies, even though they are composed by them.17 So in claiming that a person is his body, mere animalism uses the ‘is’ of identity, whereas analytical personalism uses the ‘is’ of constitution. By using this constitutive ‘is’ analytical personalism creates the necessary conceptual space to hold that persons transcend their bodies. Although bodily identity essentially composes personal identity, the latter is not reducible to the former. Like Rodin’s The Thinker is constituted by a particular lump of bronze without being identical to it, a person is constituted by a particular living human organism without being identical to it. In other words, bodily identity is a necessary but not a sufficient condition of personal identity. Since identity of body is not a sufficient condition of personal identity, according to analytical personalism, other considerations, of personal characteristics, must be invoked. Wherein then does personal identity consist over and above bodily identity? To answer this question, the second essential ingredient of analytical personalism must come into play. The conceptual development of this element starts from the Strawsonian thought that the concept of a person, as primitive relative to that of a body, is the concept of a dynamic agent related to the public world. As constituted by a living human organism, a person exhibits a particular mode of activity that marks him off as an agent. In comparison with other animals, the rational animal lives in a specific active mode, namely the mode of intentional agency. Correspondingly, personal identity necessarily comprises agential identity besides bodily identity. So both the ‘minimal’ structure of bodily identity and the ‘maximal’ structure of agential identity determine the nature of personal identity. Moreover, the importance of personal identity also stems from this identity of the person as an agent. Before expanding on the significance of personal identity in terms of narrativity, let me comment a bit on the second core feature of analytical personalism. The best account of intentional agency in contemporary analytical anthropology has, to my mind, been given by Donald Davidson (1980). According to his causal theory of action, a bodily movement is an action if and only if the movement is appropriately caused by beliefs and desires (P-predicates with propositional content or propositional attitudes). 17 For further elaboration of this constitution view, see Baker 2000.


Stefaan E. Cuypers

More specifically, Davidson defines acting with an intention as acting for a reason: “(A)n action is performed with a certain intention if it is caused in the right way by attitudes and beliefs that rationalize it.” (Davidson 1980, 87) The beliefs and other propositional attitudes constitute the reasons why the action was performed. Now Davidson’s causal theory of action in psychology forms a part of his wider theory of rational (and radical) interpretation in semantics (see Davidson 1984). In view of that, actions are rationally interpreted in terms of propositional attitudes from the ‘third person’ standpoint—a standpoint which is also the hallmark of Strawson’s descriptive metaphysics of the person. In light of the rationalization principle, those attitudes should be ascribed to the agent as to provide good reasons for his actions. The relation between the nature of intentional agency and agential identity becomes apparent when one important tenet of Davidson’s interpretation theory is put to the fore, namely the holism of the mental. What does the holistic character of propositional attitudes and intentional action involve? Attitudes and actions cannot be identified atomistically, i.e. on their own, isolated and cut off from other attitudes and actions. In order to demarcate what they are, attitudes and actions have to be related to other attitudes and actions. Rational relations amongst attitudes and actions make them what they are. Such relations are constitutive of the propositional attitudes and intentional actions. So the attitudes and actions form a logically coherent whole or network. Davidson makes this holism of the mental clear as follows: We identify thoughts [beliefs and other propositional attitudes], distinguish between them, describe them for what they are, only as they can be located within a dense network of beliefs. … To have a single propositional attitude is to have a largely correct logic, in the sense of having a pattern of beliefs that logically cohere. … The point extends to intentional action. Intentional action is action that can be explained in terms of beliefs and desires whose propositional content rationalize the action. (Davidson 1982, 320 f.)

Now the important thing about the holism of the mental with regard to personal identity is that in virtue of this holism the attitudes and actions are unified in rational patterns at a time and also over time. If I form the intention to do something in the (near) future, then my intention is only rationally intelligible if it is a part of a diachronic network of other intentions, beliefs, desires and other propositional attitudes. In executing my plan of action all the different actions performed to reach my goal only make sense if they belong to a diachronic sequence of actions that logically cohere. Attitudes and actions are thus intrinsically unified in

Personal Identity and Agency: Towards Analytical Personalism


virtue of the holistic rationalization principle. Such patterns of unified propositional attitudes and, in particular, sequences of unified intentional actions are constitutive of agential identity. So agential identity through time is just the playing out over time of the holistic nature of the attitudes and actions. According to analytical personalism, agential identity must be grafted on bodily identity for the full determination of personal identity. Some will object, however, that this additional maximal structure of agential identity cannot possibly capture our internal sense of identity. In reply, let me adumbrate how analytical personalism handles the internal aspects of personal identity. Of course, the capacity for agency depends upon the powers of intellect, will and memory. And these powers of the mind are in the case of persons self-consciously exercised. The self-conscious or reflexive employment of intellect, will and memory in the production of actions is then responsible for the subjective sense of agential identity and, moreover, the person’s ownership of a particular body. That is to say, the involvement of reflexivity in agential and bodily identity constitutes the fact that I produce these bodily movements and the fact that this moving body is my body.18 So whereas bodily identity determines personal identity from ‘the outside’ or third-person perspective, agential identity determines it from ‘the inside’ or first-person perspective. What is more, from this first-person perspective—in self-consciousness and memory—a person is continuously and immediately present to himself. This immediate and continuous self-presence makes self-identification superfluous. However, although a person does not have to identify or reidentify himself from the inside, self-presence in self-consciousness and memory is intelligible only as a subjective angle on something that has more to it than the subjective angle reveals, namely the career of an objective continuant—a personal body—in the unitary spatio-temporal framework.19 Agential identity (together, of course, with basic bodily identity) not only determines the nature of personal identity, it also establishes its importance. In a way, however, agential identity is too ‘thin’ to accomplish this; a more ‘thick’ maximal structure of personal identity is needed to 18 For this thought, see Madell’s (1976) discussion of Ayer 1963. 19 For interesting proposals—which are compatible with analytical personalism—of how the subjective (self-conscious) and the objective (bodily) standpoints can be combined in the constitution of personal identity and self-knowledge, see Campbell 1994; Cassam 1997; McDowell 1997.


Stefaan E. Cuypers

account for its importance. As the third, and last, crucial feature of analytical personalism I, therefore, bring in narrative identity. Yet thick narrative identity is in a way not a separate maximal structure on top of thin agential identity. Narrative and agential identity belong to the same level. Narrative identity is just the more existential and more literary corollary of agential identity. Narrative identity only introduces additional and richer methods for the unification of a person’s life. According to the narrativity theory of personal identity, actions and other events in a person’s life are unified into a single life by means of a coherent and intelligible life story or biographical narrative. Alisdair MacIntyre is a well-known proponent of this theory: It does follow … that all attempts to elucidate the notion of personal identity independently of and in isolation from the notions of narrative, intelligibility and accountability are bound to fail. … In what does the unity of an individual life consist? The answer is that its unity is the unity of a narrative embodied in a single life.20

Narrativity internally unifies the different actions of a person into a single coherent story and it externally relates them to the wider community in which he lives. Telling the story of one’s life involves rendering one’s deeds and omissions intelligible not only to oneself but also to others with whom one entertains multifarious relations in a public common world. Furthermore, besides just telling one’s life story, one also takes responsibility for it. Now taking responsibility for one’s life story involves acknowledging its importance. It is then in this way that one’s biographical narration plays an important role in rational, moral and emotional practices. So the concept of personal identity contained in existential practices which comprise responsibility, love, prudential self-interest, and the like is the concept of narrative identity. One’s biographical narration is the proper object of the (moral) reactive attitudes and feeling of one’s own and other people. These reactions testify to the significance with which persons invest their life stories. In this manner narrativity gives substance to the Strawsonian thought that the concept of a person is the concept of an object of (moral) reactive attitudes and feelings. Hence the importance of personal identity stems from the identity of the person as a rational and moral agent telling his life story. 20 MacIntyre 1985, 218. For other versions of the narrativity theory, see Schechtman 1996; Christman 2004. The idea of narrative identity is, of course, also prominent in continental philosophy: see, for example, Ricoeur 1990.

Personal Identity and Agency: Towards Analytical Personalism


In sum, according to analytical personalism, personal identity as agential, and by implication, bodily identity essentially consists in the narrative unity of the actions of a rational and moral agent in a social setting and within a historical tradition. 6. Conclusion Admittedly, my sketch of an analytical alternative to the bundle theory, or its impersonal attendant, as well as to the ego theory of personal identity stands in need of much more elaboration and it does not conclusively show that a non-atomistic view in accord with Strawsonian principles is in the last analysis a fruitful and stable position to take in the debate on the nature and importance of personal identity. But whatever the final details of such an alternative view—analytical personalism—may be, something along these lines must be true, unless we are willing to acquiesce in the dualistic dissolution of our psychophysical unified nature or, even worse, in the empiricist elimination of the significant role of our identity in common practices. Acknowledgment This chapter is based on a part of my book Self-Identity and Personal Autonomy. An Analytical Anthropology, Stefaan E. Cuypers, 2001,  AshACHTUNGREgate. Bibliography Ayer, Alfred J. (1963): “The Concept of a Person”, in: Ayer, Alfred (1963): The Concept of a Person and Other Essays, London: Macmillan, 82 – 128. Baker, Lynn R. (2000): Persons and Bodies. A Constitution View, Cambridge: Cambridge University Press. Campbell, John (1994): Past, Space, and Self, Cambridge, Massachusetts: The MIT Press. Cassam, Quassim (1997): Self and World, Oxford: Clarendon Press. Christman, John (2004): “Narrative Unity as a Condition of Personhood”, in: Metaphilosophy 35, 695 – 713. Coburn, Robert C. (1985): “Personal Identity Revisited”, in: Canadian Journal of Philosophy 15, 379 – 403. Cockburn, David (1990): Other Human Beings, London: MacMillan.


Stefaan E. Cuypers

Cuypers, Stefaan E. (2001): Self-Identity and Personal Autonomy. An Analytical Anthropology, Aldershot: Ashgate. — (2005): “Das Problem der personalen Identitt in der analytischen Philosophie”, in: e-Journal Philosophie der Psychologie 3, Davidson, Donald (1980): Essays on Actions and Events, Oxford: Clarendon Press. — (1982): “Rational Animals”, in: Dialectica 36, 317 – 327. — (1984): Inquiries into Truth and Interpretation, Oxford: Clarendon Press. Hume, David (1739): “Of Personal Identity”, in: Perry (1975), 161 – 172. Mackie, John L. (1976): Problems from Locke, Oxford: Clarendon Press. MacIntyre, Alasdair (1985): After Virtue. A Study in Moral Theory, London: Duckworth. Madell, Geoffrey (1976): “Ayer on Personal Identity”, in: Philosophy 51, 47 – 55. McDowell, John (1997): “Reductionism and the First Person”, in: Dancy, Jonathan (ed.): Reading Parfit, Oxford: Blackwell, 230 – 250. Mounier, Emmanuel (1950): Personalism, Mairet, Philip (transl.) (1952), Notre Dame: University of Notre Dame Press. Nagel, Thomas (1986): The View from Nowhere, Oxford: Oxford University Press. Noonan, Harold W. (2004): Personal Identity (2nd edition), London: Routledge. Oksenberg Rorty, Amelie (ed.) (1976): The Identities of Persons, Berkeley: University of California Press. Olson, Eric T. (1997): The Human Animal. Personal Identity without Psychology, New York: Oxford University Press. — (2003): “An Argument for Animalism”, in: Martin, Raymond/Barresi, John (eds.): Personal Identity, Oxford: Blackwell, 318 – 334. Parfit, D. (1971): “Personal Identity”, in: Perry (1975), 199 – 223. — (1984): Reasons and Persons, Oxford: Clarendon Press. Perry, John (1975) (ed.): Personal Identity, Berkeley: University of California Press. — (2002): Identity, Personal Identity, and the Self, Indianapolis: Hackett. Reid, Thomas (1785): “Of Identity”, in: Perry (1975), 107 – 112. Ricoeur, Paul (1990): Oneself as Another, Blamey, Kathleen (transl.) (1992), Chicago: Chicago University Press. Russell, Bertrand (1918): “The Philosophy of Logical Atomism”, in: Russel, Bertrand (1956): Logic and Knowledge. Essays 1901 – 1950, Marsh, Robert C. (ed.), London: Unwin Hyman, 175 – 281. Schechtman, Marya (1996): The Constitution of Selves, Ithaca: Cornell University Press. Scheler, Max (1913 – 1916): Formalism in Ethics and Non-Formal Ethics of Values. A New Attempt toward the Foundation of an Ethical Personalism, Evanston: Northwestern University Press, 1973. Shoemaker, Sidney/Swinburne, Richard (1984): Personal Identity, Oxford: Basil Blackwell.

Personal Identity and Agency: Towards Analytical Personalism


Shoemaker, Sidney (1986): “Introspection and the Self”, in: Shoemaker, Sidney (1996): The First-Person Perspective and Other Essays, Cambridge: Cambridge University Press, 3 – 24. Snowdon, Paul F. (1990): “Persons, Animals, and Ourselves”, in: Gill, Christopher (ed.): The Person and the Human Mind. Issues in Ancient and Modern Philosophy, Oxford: Clarendon Press, 83 – 107. — (1995): “Persons, Animals, and Bodies”, in: Bermffldez, Jos L./Marcel, Anthony/Eilan, Naomi (eds.): The Body and the Self, Cambridge, Massachusetts: The MIT Press, 71 – 86. Strawson, Peter F. (1959): Individuals. An Essay in Descriptive Metaphysics, London: Methuen. — (1962): “Freedom and Resentment”, in: Strawson, Peter F. (1974): Freedom and Resentment and Other Essays, London: Methuen, 1 – 25. — (1980): “P. F. Strawson Replies”, in: Van Straaten, Zak (ed.): Philosophical Subjects. Essays Presented to P. F. Strawson, Oxford: Clarendon Press, 260 – 296. —(1985): Skepticism and Naturalism: Some Varieties. The Woodbridge Lectures 1983, New York: Columbia University Press. Thomson, Judith J. (1987): “Ruminations on an Account of Personal Identity”, in: Thomson, Judith J. (ed.): On Being and Saying. Essays for Richard Cartwright, Cambridge, Massachusetts: The MIT Press, 215 – 240. — (1997): “People and their Bodies”, in: Dancy, Jonathan (ed.): Reading Parfit, Oxford: Blackwell, 202 – 229. Wiggins, David (1980): Sameness and Substance, Oxford: Basil Blackwell. Williams, Bernard (1973): Problems of the Self, Cambridge: Cambridge University Press. Zemach, Eddy M. (1970): “The Unity and Indivisibility of the Self ”, in: International Philosophical Quarterly X, 542 – 555.

Looking for the Real Enemy Marc Slors 1. Introduction In this brief comment on Stefaan Cuypers’ paper I’d like to argue two things: (1) However right I believe Cuypers to be on analytical personalism, I shall claim that he is wrong about ‘the enemy’. By “the enemy” I do not mean the bundle theory and/or the ego theory—the positions against which Cuypers argues. What I mean is that what these theories have in common on the basis of which they are both rejected. According to Cuypers both theories presuppose philosophical atomism and a quasi-perceptual theory of self-knowledge I shall refer to as introspectionism. And since both presuppositions are untenable, Cuypers argues, the ego- and bundle theories of personal identity over time must be replaced by an alternative—analytical personalism—that presupposes neither. Against this I shall argue that it is doubtful that atomism and introspectionism are necessarily presupposed by ego- and/or bundle theories. Moreover, neither atomism nor a perceptual theory of selfknowledge are necessarily incompatible with Sir Peter Strawson’s view on the primitive nature of the concept of “a person” on which Cuypers grounds his analytical personalism. With this criticism I do not mean to argue against analytical personalism itself. I believe that theory to be largely correct. Moreover, I believe that there is a real opposition between analytical personalism and the bundle- and ego-theory that does go back to Strawson’s view on the primitiveness of the concept of a person. Strawson, however, rejects first and foremost what I will label a ‘descriptive dualism’ with regard to persons (roughly: the idea that it makes sense to distinguish between the mental and physical properties of persons and to conceive of persons either as mental entities, as bodies, or as hybrids of both). The first and main thesis I’d like to defend is that Cuypers’ real enemy is descriptive dualism, not atomism and introspectionism. (2) The second point I’d like to make is that once it is recognized that descriptive dualism is the real enemy of analytical personalism, it should also be acknowledged that analytical personalism is strongly supported by recent developments that take place outside the core of analytical philosophy. Developments in neurophenomenology, cognitive neuro-

Looking for the Real Enemy


science and developmental psychology all count very much in favour of analytical personalism. But let me first give a very brief outline of what I take Cuypers to be arguing. 2. An Outline of Cuypers’ Argument The following represents what I take to be Cuypers’ main points and the structure of his argument: (1) There are rougly two main contenders for a feasible theory of transtemporal personal identity: the (empiricist) bundle theory and the (more metaphysical minded) ego theory. (2) Both theories hinge on philosophical atomism and a quasi-perceptual theory of self-knowlegde. (3) Humean and other considerations (here I take Cuypers to be referring to the transparancy of experience) lead to a rejection of the perceptual model of perception as well as to a rejection of atomism. (4) Impersonalism, which is proposed as a way out of the impasse by e.g. Parfit, is of no help. In fact impersonalist theories are just radical versions of the bundle theory that make the impasse even deeper. (5) The position proposed to dissolve the impasse is labelled “analytical personalism”. It takes its lead from Strawson’s idea that “person” is a primitive term. This idea is interpreted as a rejection of atomism. A person, on this account, is primarily an agent that is prone to reactive attitudes. (6) This analytical personalism is further elaborated on by showing that diachronic identity of the person involves three axes: bodily identity, agential identity and narrative identity. 3. Two Questions about Atomism and Self-Knowledge The motor behind the dialectic in Cuypers’ paper is the claim that atomism and introspectionism are implied by ego- and bundle theories while they are wrong-headed. I shall not discuss the question whether atomism and introspectionism are indeed wrong-headed. Instead I shall ask (1) whether atomism and introspectionism are indeed necessarily connected with bundle and ego theories. And secondly, (2) I shall ask


Marc Slors on Stefaan E. Cuypers

whether atomism and introspectionism are indeed avoided by the analytical personalism Cuypers wants to substitue ego- and bundle theories with. (1) Let me begin with the bundle theory and atomism. I agree that most of the available bundle theories conceive of diachronic personal identity in terms of a bundle of (mental) states, atomistically conceived, that stretches out over time. But is a non-atomist conception of a bundle view a conceptual impossibility? I think not. It is not impossible to conceive of psychological continuity—a psychological bundle—as constituting personal identity over time while rejecting a (Parfitian) atomistic definition of psychological continuity (see Parfit 1984, 204 – 9). A (diachronically) holistic version of the psychological continuity theory does not seem to be an a priori impossibility.1 I do agree that on a holistic conception the ‘units’ of which the bundle consists are less well delineated than on an atomist conception. But they are units nevertheless. For if not, psychological holism would imply that we cannot distinguish between individual mental states. It doesn’t imply that. Therefore the title “bundle theory” is still applicable on a holistic conception of psychological continuity. Now to the connection between the bundle theory and introspectionism. Again, I agree that most bundle theories implicitly presuppose such a form of self-knowledge. But I would deny that the connection is a necessary one. Non-introspectionist theories of self-knowledge such as theories that make use of what Robert Gordon calls “ascent routines” (Gordon 1996) or more elaborate ones such as Jos Lus Bermffldez’ developmental theory (Bermffldez 1998), for instance, are perfectly compatible with a bundle conception of personal identity as psychological continuity. Nothing in the bundle theory itself commits us to a homuncular observer when it comes to self-knowledge. What about the ego theory and a quasi-perceptual theory of selfknowledge? Here I believe the connection is much stronger than in the previous two cases. For an ego theory does postulate something similar to an inner, observing homunculus. Still there are difficulties. For when we are talking of self-knowledge, in the case of an ego theory we are talking about knowledge about the ego itself had by the ego itself. This kind of epistemic relation is notoriously tricky. Of course, relations such as direct acquaintance have been proposed to account for this type of selfknowledge. But these solutions remain very much under discussion. In 1

In fact I have defended such a theory; see Slors 2001a; 2001b.

Looking for the Real Enemy


the case of the ego theory, the idea of direct acquaintance may be especially difficult. For if the ego is the object of acquaintance, how can it also be the subject of it while at the same time being indivisible and unanalysable? Finally, what about the connection between the ego theory and atomism? Here again the connection seems to be very strong. Insofar as an ego is conceived of as an unanalysable entity that is the self, the ego is a philosophical atom. A theory of personal identity over time in terms of diachronic self-identity (as ego-identity), then, can justifiably be called an atomist position. Still I have misgivings. For the unanalysability of the ego itself can be construed as a rejection of atomism. The ego itself is not conceived of as being constituted by philosophical atoms. Of course it can be replied here that this does nothing against the charge that the ego is an atom itself. But the problem is that the same goes for Strawson’s primitive concept of a person which is in fact the starting point of the position Cuypers himself defends. The Strawsonian primitive concept of a person cannot be analysed in terms of its mental and physical constituents; it is a primitive entity that has mental and physical properties but is not constituted by them. (2) This brings us to the second question: to what extent does analytical personalism, grounded on the Strawsonian primitive concept of a person, rule out atomism and perceptual theories of self-knowledge? With regard to atomism, I can be brief: if personalism avoids atomism because the concept of a person is an unanalysable primitive concept, then neither can the ego theory be conceived of as a form of atomism. If the ego theory is an atomist position, then so is personalism insofar as it is based on the Strawsonian primitive. Does analytical personalism rule out introspectionism? Cuypers theory certainly does. But the relevant question here is whether this is due to the fact that a perceptual theory of self-knowledge is conceptually incompatible with personalism. I think not. Neither the idea of the primitiveness of the concept of a person, nor the idea that personal identity involves bodily identity, nor the concepts of agential and narrative identity seem conceptually incompatible with e.g. a Block-like version of ‘access consciousness’ (see Block 1978) or an Amstrongian ‘self scanning mechanism’ (see Armstrong 1968), both of which can be construed as versions of introspectionism. These are very sketchy considerations, of course. Yet I believe they warrant the conclusion that the implication of atomism and introspec-


Marc Slors on Stefaan E. Cuypers

tionism is not what is really wrong with the ego- and bundle theories from the viewpoint of analytical personalism. 4. What Is Wrong with the Ego and Bundle Theory? I’d like to reverse the question: let’s not ask why the ego and bundle theory ought to be rejected in favour of analytical personalism, let’s suppose that analytical personalism is to be preferred over the ego and bundle theories (which I believe is the case) and ask what is wrong with the ego and bundle theories from that point of view. In order to answer this rephrased question, we should look at the Strawsonian basis for analytical personalism. Strawson (1959, chap. 3) did indeed claim that “person” is a primitive term and that it cannot be analysed in terms of its being constituted by (atomistically conceived) mental and physical properties. But his discussion makes it clear that the main cause of his concern was not so much the idea that a person is made up of elements. Rather, his concern was with the idea that a person is a conjunct of mental elements on the one hand and bodily elements on the other. The focus of Strawson’s argument is the claim that a person is not a hybrid of mind and matter, body and soul. So, Strawson’s concern was to argue that a person is a single entity, not an aggregate. This can be understood as opposition to philosophical atomism. But it is first and foremost a rejection of what can be labelled descriptive dualism. Descriptive dualism is not an ontological thesis. Rather, it is the conjunction of (i) the thesis that the mental and the physical are at least conceptually separable realms and (ii) the thesis that when describing the person as an explanandum, it has separate mental and physical aspects (some of which may well be non-essential, so that purely psychological and purely bodily theories of personal identity are possible). From the perspective of a theory according to which descriptive dualism is mistaken, it is very clear what is wrong with the ego and bundle theories. Both kinds of theory are mistaken because they are psychological theories of personal identity as opposed to physical/bodily theories. An ego is a psychological entity. And though bundle theories can in principle just as well represent the person as a bundle of mental and physical items, in practice bundle theories are psychological continuity theories. The thesis that the real enemy of analytical personalism is descriptive dualism seems to be corroborated by the fact that Cuypers appears to be

Looking for the Real Enemy


just as much set against animalism as he is against ego and bundle theories. Animalism, however, is neither an atomist theory, nor is it committed to a perceptual theory of self-knowledge. But it is a theory according to which the person is constituted by her body rather than by her psychological make-up. In that respect, animalism is a descriptively dualist theory. 5. Support for Analytical Personalism The real enemy of analytical personalism, I contend, is descriptive dualism. But by robbing personalism of its supposed atomist and introspectionist enemies, I have also robbed it of supporting arguments. Rather than by rejections of atomism and introspectionism, analytical functionalism should be supported by reasons to reject descriptive dualism. Here it may be thought that analytical philosophy is littered with rejections of dualism, so there really is no problem. Unfortunately, the rejections of dualism to be found in the literature are almost always rejections of dualism as an ontological thesis, not rejections of descriptive dualism. Thankfully, though, there is ample support for a rejection of descriptive dualism from areas that border on but are distinct from analytical philosophy. Let me mention two examples of support for a rejection of descriptive dualism from outside analytical philosophy. First, part of the recent neurophenomenological movement is concerned with sketching a picture of the human mind as essentially embodied (see e.g. Gallagher 2005). For this purpose, insights from phenomenology are combined with recent finds in cognitive neuroscience and developmental psychology. For, as it turns out, the phenomenologists’ picture of the mind-body unity (e.g. Merleau Ponty’s Corps Sujet) that does not allow us to separate the two, is, to a substantial level of detail, backed up by contemporary science. Secondly I’d like to mention the discovery of mirror neurons in the early 1990’s (Rizzolatti et al. 1996). In my opinion, what this discovery shows, is that bodily expressions of primitive intentional states are not contingent correlates of these states, but rather integral aspects of them. A facial expression is part of the emotion we take it to indicate (see also Damasio 1994; Goldman/Sripada 2005). A gesture is part of an intention.


Marc Slors on Stefaan E. Cuypers

There is no reason to ignore such support for analytical personalism. And given the need for arguments for that position, I think there is enough reason for analytical philosophers that are sympathetic to Cuypers position to draw on the recent insights of cognitive neuroscience, phenomenology and developmental psychology. Bibliography Armstrong, David (1968): A Materialist Theory of Mind, New York: Humanities Press. Bermffldez, Jose L. (1998): The Paradox of Self-Consciousness, Cambridge MA: MIT Press. Block, Ned (1978): “Troubles with Functionalism”, in: Savage, Wade (ed.): Perception and Cognition: Issues in the Foundations of Psychology, Minnesota Studies in the Philosophy of Science IX, 261 – 326. Cuypers, Stefaan (2007): “Personal Identity and Agency: Towards Analytical Personalism” (this volume). Damasio, Antonio R. (1994): Descartes’ Error: Emotion, Reason and the Human Brain, New York: Putnam’s Sons. Gallagher, Shaun (2005): How the Body Shapes the Mind, Oxford: Oxford University Press. Goldman, Alvin/Sripada, Sekhar S. (2005): “Simulationist Models of Face-Based Emotion Recognition”, in: Cognition 94 (3), 193 – 213. Gordon, Robert M. (1996): “‘Radical’ Simulationism”, in Carruthers, Peter/ Smith, Peter K. (eds.): Theories of Theories of Mind, Cambridge: Cambridge University Press, 11 – 21. Parfit, Derek (1984): Reasons and Persons, New York: Oxford University Press. Rizzolatti, Giacomo/Fadiga, Luciano/Gallese, Vittorio/Fogassi, Leonardo (1996): “Premotor Cortex and the Recognition of Motor Actions”, in: Cognitive Brain Research 3, 131 – 141. Slors, Marc V.P. (2001a): The Diachronic Mind, Dordrecht: Kluwer Academic Publishers. — (2001b): “Personal Identity, Memory, and Circularity: an Alternative for QMemory”, in: Journal of Philosophy 98 (4), 186 – 214. Strawson, Peter F. (1959): Individuals, London: Methuen.

The Units of Living Rdiger Bittner 1. Introduction There is world history: everything that did or will happen. There is human history: that part of world history that involves humans. My question is how we conceive of human history. Normally we do not conceive of it as one big whole, but in smaller packages. My question is what these are. Bread comes in loaves, wine comes in bottles: in what form do things happening to and done by humans come? In all sorts of forms, you will say. Yes, but within this indefinite number of forms of things happening a limited range of types can be recognized. I shall try to illuminate some of these types, investigating especially when and why we think of what happens in terms of one rather than the other. 2. Actions The idea of an action has dominated the understanding of what happens in human lives for a long time. “Theory of action” is called, therefore, the reflection on what it is that we judge good or bad, or right or wrong, the assumption being that it is in the form of actions that human history comes in for moral appraisal. Actions are generally understood to be human doings: somebody does, or several people together do, something, and the event of his or their doing it is the action. Being events, actions are discrete and datable. They happen at a certain time, perhaps also at a certain place; and they are distinct entities, like the people producing them. They need not be individuals, though: one action may consist of several others. This is a common sense understanding of action. Philosophical explanation has run into difficulties with all of these characteristics. I shall only mention a few of these difficulties. With respect to datability, people have wondered when somebody’s doing something happens, and


Rdiger Bittner

in particular, when it is over: is a killing for instance completed once the killer has done his part, once the arrow, poetically speaking, has left the bow, or is it only completed once the victim is dead? (Stoecker 1993, sect. 4) With respect to discreteness, it is just not clear how actions are individuated: what does it take for a stretch of human history to form one action? Many of us would be hard put, therefore, to count their actions from morning to midday. And with respect to actions being somebody’s doings, it is a disputed question, first, whether omissions should not also, sometimes at least, count as actions (Thomson 1977, chap. 15; Birnbacher 1995), and second, what it is for a human being to be the doer of some action. Various authors have noted that the term “action” does not appear frequently in ordinary talk. This is not because we do not conceive of things human in terms of actions. It is because we normally move on right away to specify the action, i.e. to say who did what; and to do that we do not need the word “action”. In fact, using the word “action” in reporting what people do or did often leads theorists to speak, strictly, nonsense. “Jimmy closed the door” reports an action of Jimmy’s without using the word “action”. “Jimmy performed the action of closing the door” is nonsense, unless Jimmy be an actor. An action is already one’s doing something, and there is no doing of one’s doings, and no performing them either (Thomson 1977, 35 f.). The idea of somebody’s action, understood as a datable, discrete event, seems to have been with us not for as long as we can think. It seems to have arisen at a certain point in history only. Here is Aristotle in the course of explaining the voluntary and the involuntary: Something similar holds for the cases of throwing goods overboard in storms. For nobody simply throws goods away voluntarily, but to save oneself and the others everybody does who is in his right mind. Such actions, then, are mixed, but they look more like voluntary ones; for when they are done they are chosen, and it is relative to a certain time that something is the aim of the acting. (NE 1110 a 8 – 14)

The reasoning in the last sentence seems to be this. Somebody’s doing something is aimed at something on one occasion, and may not be aimed at it on another. So when on some occasion somebody throws goods overboard to save his life, he chooses to do so, and what he does must therefore be considered voluntary. The last word in this passage, “acting”, which translates “praxis”, is unspecific, it means activity in general; and the idea is that people’s activity is tied to a certain aim on one occasion, and not tied to it on another. The word “actions” in the middle of this

The Units of Living


passage, however, again translating “praxeis”, is specific and refers to instances of people’s being active, namely to this and that throwing goods overboard in a storm. So here we have a duality of meanings of “praxis”. There is on the one hand “activity”, in many other passages also “kind of activity”1, a usage that can also be found in other writers, for example in Plato. There is on the other hand here a very early occurrence of that concept of an action on which current theory of action is still founded. For these actions are datable: your throwing goods overboard happened then and there, they are discrete: your throwing goods overboard and mine were different actions, and they are somebody’s doings: for any action of throwing goods overboard there is at least one thrower whose working in some specific way constitutes the action. The word “praxis” is not especially rare in Plato. Yet nowhere in Plato does it have the second of the two meanings just distinguished in the Aristotelian sentences, except for one passage in the (very late) Laws, where the Athenian talks about judges who would be capable of setting for every transgression the punishment “fitting the suffering and the action” (876 d). This could be taken to mean that punishment ought to be meted out not merely in accordance with the amount of suffering endured by the victim, but also in accordance with the particular character of the action, with its exceptional meanness, say, or cruelty or the like. If this reading is correct, then we have here again the idea of an action as a distinct occurrence with particular characteristics. Everywhere else in Plato, though, “praxis” is used rather in the sense of “activity” or even “business”.2 It appears, then, that the concept of “an action”, i.e. an actionindividual, was introduced fairly late. How, then, did speakers without this concept refer to people’s doings, as they must have done? And if that concept had not been necessary before, what was the reason for introducing it? The answer to the first question was hinted at earlier. We do not frequently use the word “action”, because we normally move on right away to specify the action in question. Thus when the concept of ‘an action’ was not available at all, people as well moved on right away to specify the action in question. There is a characteristic turn of phrase in Plato’s early Charmides, when Socrates asks his interlocutor Kritias whether on his view “the doing, or making, or however you want to call 1 2

This must be the meaning, for instance, in NE 1094 a 7. Menon 99 b provides a good example of this usage.


Rdiger Bittner

it, of good things” is what moderation amounts to (163 e). Socrates is clearly speaking of a number of actions here, all those that would constitute, on the thesis under examination, somebody’s moderation. However, the actions are individuated here, not as individual doings, but as accomplishings of things done. One’s active life does not, on this view, come divided into action-individuals, but into completings of things, not into stretches of the agent’s working in some way, but into works done; and it is the works, not the workings which are then said to have a certain character, for example they are said to be good. The same conception is at work in Aristotle’s famous statement that we become just by doing just things, moderate by doing moderate things, courageous by doing courageous things (NE 1103 a34 – b2): what one does is divided here, not into complete action-individuals, but into complete achievements bearing a certain character. Just to have a short term ready and without presuming to capture the proper meaning of the English word, let me call a thing achieved in one’s activity a deed. Better even would be ‘a done’, but that may no longer be acceptable English. The battle was won, the house was built, the sick man was cured—these statements report somebody or other’s deeds. And I am suggesting that, before the concept of an individual action was introduced, it was in terms of deeds that people’s active life was understood. With regard to the second question, why actions succeeded deeds as the dominant conception of our active life, I can offer no more than a guess: a shift in moral sensibility may have brought about a corresponding shift in the conception of agency. This shift in moral sensibility is witnessed by chapter II 3 of the Nicomachean Ethics. Here Aristotle lets his opponent raise an objection to his statement that we become just or moderate by doing just or moderate things. The objector suggests that he who does just and moderate things is already just and moderate and so does not need to become just and moderate by appropriate practice. Aristotle replies by denying that he who does just and moderate things is already just and moderate: What is done under the requirements of virtue, is not yet done justly or moderately, if it is of a certain kind. The agent in acting needs to be of a certain kind too. He needs to know that what he is doing is just or moderate, he needs to choose doing something just or moderate and choose it for its own sake, and he needs to do it from a firm and unchanging attitude. (NE 1105 a28 – 33)

The Units of Living


This is a strange doctrine. One would have thought that for something to be done justly or moderately it is indeed sufficient that it be of a certain kind, and that the agent’s being of a certain kind, i.e. his knowing the just or moderate and choosing it for its own sake from a firm disposition, is irrelevant. Dirlmeier, commenting on a similar passage a few lines down, defends Aristotle as follows: Example: a young Athenian has done something brave. Viewed from the outside, it looks like the deed of a man who acts this way from a firm disposition. Still, do we know whether the young man carries this comprehensive attitude in himself ? […] Do we know then whether he followed right planning, whether he acted in accordance with ‘the right logos’ or his action was only a lucky hit [ein Zufallstreffer]? (Aristotle 1991, 307 f.)

We do not know that, but the difficulty is to see why it should be relevant. Clearly our view of the agent will be different depending on whether we think that his deed was a lucky hit or that it was in character. However, what Aristotle claims in the quoted passage is that our judgment of the deed should be different depending on whether the agent has one attitude or another, and it is difficult to see a reason for this claim. The deed, one would think, remains brave or just, whatever the agent’s attitude. After all, a lucky hit is no less a hit for being lucky. Strange or not, the doctrine makes intelligible the shift from deeds to actions in understanding our active life. On this doctrine, deeds do not contain the truth about the moral status of what we do, they are morally ambiguous; and if you suppose that it is of the essence of our active life that it be subject to moral assessment, then the moral ambiguity of deeds gives you reason to doubt that it is appropriate to understand our activity in terms of them. Furthermore, if the moral status of our doings is a matter not merely of what we do, but also of our knowing and deciding what we do, then, again supposing that our active life essentially is subject to moral assessment, it is natural to understand human activity in terms of these workings of agents, their knowing, deciding, and only therefore doing, what they do. It is natural, that is, to shift the focus of understanding human history from things done by agents to agents’ doings. For doings and their origins within the agent, that is now where, morally speaking, the kitchen is. Therefore actions, owed to agents’ knowing and deciding what to do, are the sort of thing for which agents can be held responsible. The natural setting for holding people responsible is a trial. So it may not be accidental that the early use of “praxeis” in the sense of “actions” in Plato’s


Rdiger Bittner

Laws comes in a discussion of legal proceedings, and the other passage I mentioned, Aristotle’s treatment of the voluntary in NE III 1, is also dealing, according to Dirlmeier, with problems that had come up in court (Aristotle 1991, 323). Perhaps, then, there are not first actions, and then agents who are taken to bear responsibility for them. Perhaps conversely, actions were invented as the sort of thing which agents could be asked to respond for in court. However, what prompted that shift in moral sensibility that brought in its wake such a change in understanding our active life? Why did people come to worry about the difference between lucky hits and reliable virtue? Here is, again, a guess. Lucky hits no longer sufficed when in the sophistic movement the moral and political competence of a man came to be seen as the result of a professional shaping of his character. If virtue can be taught3, only good action produced on a regular basis, not the odd splendid deed, will be evidence that a man has learnt his lesson; and a man’s being good at something or other is, according to the sophists’ educational claims, a matter of having learnt a lesson. Aristotle’s protestantism of NE II 3, i.e. his refusal to call just a mere deed unsupported by a principled commitment to justice, rests on the conviction that the individual thing we do depends on our general attitude, an attitude, however, that is not born with us, but acquired in training, hence changeable in principle. Actions succeed deeds as the conception of our doings, because actions are the product and the manifestation of the dispositions an agent is taught that make up his character. The sophists were wrong on this point, however, and virtues, while growing in people, are not taught. No doubt a virtue-friendly environment helps, but that is evidence for their plant-like growth rather than for their teachability. You do not acquire a virtue the way you acquire proficiency in a new language or in a special kind of dance. However, if virtue is not taught, there is no longer reason to despise lucky hits, for whether somebody has a virtue and how firmly it is anchored must itself be counted a matter of luck, not of planning. Thus one’s doings effectively become lucky, or unlucky, hits throughout, since one does not control their springs. With that the focus of moral assessment returns from actions to deeds. Rather than praising or disparaging the inner principle and the action supposedly flowing from it, we will praise or disparage the thing done. This perspective on human history is familiar 3

See the striking passage Euripides, Supplicantes, v. 913 f.

The Units of Living


from writings in the heroic tradition, a late remembrance of which appears in the passage from Pericles’ memorial speech that Nietzsche quotes in Genealogy of Morals: “everywhere on land and sea our daring opened up its path, erecting imperishable monuments for itself in good and ill.”4 Hero is he who achieves what is worth remembering. With human history viewed under this perspective, the concept of an action is dispensable. We can instead describe our lives in terms of works done. 3. Lives Here, however, is an alternative to both actions and deeds: we might describe our lives precisely in terms of lives. This is the oldest of the concepts I am going to consider here. It appears in Plato (Republic 581 c, e) and in Aristotle (NE I 3), but traces can already be found in earlier poets, in Bacchylides (9, 39 – 45) and Pindar (Ol. 8, 12 – 14; Pyth. 1, 41 f.).5 On the other hand, it was the origin of an important tradition. Taken up by Cleanthes in his hymn to Zeus (v. 27 – 29, SVF I, 537), it was incorporated into Christian doctrine by Augustine (de civ. dei XIX, 2, 3, 19) and became a tenet of medieval moral theology with Thomas Aquinas (S. th. II – II, qu. 179 – 182). Hannah Arendt still, with her Vita activa of 1960, took support from that tradition. A life in the sense of this tradition is not understood as the stretch of a human creature’s being alive, what Heidegger calls “die Erstreckung des Daseins zwischen Geburt und Tod” (Heidegger 1963, § 72; italics omitted). It is understood, rather, as a qualitative whole the unity of which lies in what basically the human being in question is striving for. Not that human beings are supposed to be striving for one individual thing which having reached they could rest content. In this respect Hobbes’ polemics against the idea of a satisfaction in this life (Hobbes 1651, chap. 11, 1st paragraph) is off the mark. Human beings are here supposed to be basically striving, not for one thing, but for one kind of thing, and that leaves open a restless pursuit of one individual thing after another. However, different human beings are pursuing different kinds of thing, and it is the kind of thing a human individual is after that determines the kind of life he is living. 4 5

Nietzsche, I 11. The quotation comes from Thucydides II 41, 4. von Wilamowitz-Mçllendorff 1913, 185 – 191.


Rdiger Bittner

Using this criterion, Plato distinguished three kinds of life or, as it is commonly phrased, three lives, one aimed at understanding, one aimed at honour, and one aimed at pleasure and at money. Adding money under the third heading is an awkward attempt to find room for the life of devoted business people, awkward because money presumably is not pleasurable in the way of good wine, say. If however it is argued that money above all gains you access to pleasurable things, as Plato does argue (Rep. 580e – 581a), this is rather reason to drop money from the list of life-forming aims, for then it is not pursued for its own sake, as Aristotle notes (NE 1096 a 6 – 7). Aristotle therefore settles for a list without business people: you pursue either pleasure or honour or understanding. Thomas Aquinas in turn cuts the list down to two: your life is either active or contemplative, the former a broadened version of Aristotle’s political life in pursuit of honour, the latter a narrower version of Aristotle’s theoretical life in pursuit of understanding. Gone is the lusty fellow, Thomas’ reason for dropping him from the list allegedly ( II – II, qu. 179, a.2, ad 1) being that his is a beastly life on Aristotle’s own admission (NE 1095 b 20). That is a bad reason, though, since beastly or not, it is still humans who are living that way. In fact the reason for discarding the life devoted to pleasure is probably that cutting the list of possible lives down to two makes it fit the biblical story of Maria and Martha (Luk 10, 38 – 42). What is perhaps most remarkable about the traditional idea of different lives is the lack of jealousy between them. To be sure, the authors, and especially the philosophers, hold determinate views on what life is best, but they do not therefore think that those taking other paths are not as they ought to be.6 The pleasure-seekers are not so much blamed as despised by Aristotle, since they choose a life fit for cattle, whereas those who pursue honour in the political or military field do not go wrong at all, they just miss out on what is best. Thomas similarly, having discarded the life of pleasure, does not denigrate the active life: it is just that the life of contemplation is a higher life and more meritorious, too. One reason for this liberal attitude is certainly the fact that it would be self-defeating for Aristotle, set to praise virtues whose field is an active political life, to demean that life in comparison to the life of the philosopher. Indeed, even with that liberal attitude the consistency between his treatment of the ethical virtues and the elevation of a 6

John Broome’s phrase, from his paper “Normative Requirements” (Broome 1999).

The Units of Living


theoretical life is dubious. Somewhat similar considerations apply to the monk. The official church, for all its praise of spiritual life, has never been tempted to tell people that to live in the world is to be on the wrong track entirely. A more important reason, however, for the generosity of the lives tradition may be the fact that a life is not really chosen but rather found congenial by a human being of this or that character. It is true, Aristotle uses the expression that the many “choose” a life of cattle (NE 1095 b 20), but right there he also describes them as “the coarsest”7, opposing them to “the refined” who seek to be active. Never mind the invidious distinction, invidious because there do seem to be refined pleasures and coarse politicians. The important thing is that people, already being coarse or refined, do not so much decide upon their lives as find in one life or another a suitable expression of what they are. Plato, accordingly, has people, not choose to be, but become one kind of human being rather than another, depending on what element happens to be dominant in their souls (Rep. 581 b – c). So there is no blaming their lives on people: lives just grow from how they are. This raises a problem. For our actions we are responsible. Indeed, on the argument above the point of understanding our histories in terms of actions is precisely that actions are the sort of thing for which we can be responsible. For our lives, however, we are not responsible: they grow from what we are. How can this be if lives, to a large and important extent, consist of actions? They don’t. A life is not a big chunk of that of which an action is a much smaller chunk. Lives and actions are alternatives for understanding one’s history. We can take your course through the world as a series of events, some consisting in something happening to you, others consisting in actions of yours; and since of the latter you are master, we may ask you to respond for them and, if appropriate, blame you for them. On this view, what you are like will appear in tendencies you have of producing such kinds of actions under such kinds of circumstances. On the other hand, we can take your course through the world as one thing, your life, in which will be expressed what you are. For that we cannot take you to task. Your life is not something you produced, it is what you turned out to be, it is you unfolding. To be sure, we can pay special attention to this or that event in your life, but when we do, it will still be your life we are considering, if from a particular angle. So we can either recognize you in 7

Plato, interestingly, uses the same word in his discussion of the three lives, Rep. 581 d.


Rdiger Bittner

what you became or we can judge you for what you did. We cannot do both. If, then, it is either life or actions, we should go for life. For one thing, the idea of an action one is responsible for rests on the assumption that over the doings in question we are master. Yet there is reason to doubt that we are ever master over things we do. The reason is not that, as neuroscientists urged in recent debates, our actions and reactions are conditioned by the state of our brain and therefore could not be considered free. As long as this “therefore” is merely asserted, as it constantly is in current discussions, this objection has little force. The reason is, rather, that the very idea of being master over what we do seems to require a duality within ourselves of something that decides and something that executes, as in Aristotle’s image of the kings announcing to the people what is to be done (NE 1113 a 8 – 9); and it may be simpler and more plausible to understand ourselves as in this way undivided.8 A second reason to go for life rather than actions is this. Assume, in contrast to the doubts just mentioned, that there is a range of doings over which we are in control. Why should the distinction between cases where we are and cases where we are not in control be deemed so important that we should fashion our conceptions of events involving us accordingly, actions on one side, things happening to us on the other? Actions are related in numerous ways to what happens to us and conversely. Moreover, they do not strikingly stand out in the sea of what goes on with us. It would thus seem artificial to give the distinction between actions and other things such a privileged conceptual position. In your life there does not appear a salient demarcation line between your actions and what happens to you. Such a line only becomes important when you are called upon to give account for what you did. That setting, however, the setting of being asked to justify what you did, has no particular importance, contrary to what Habermas and his followers keep telling us. It is true, when you appear in court because of that accident you were involved in, you have reason to be very precise about the line between what you did and what merely happened, but when you tell us about your vacations in Crete last year, you have no such reason, indeed you have reason not to be precise about that line, as it would cut right through the stretches of life that are of interest to you and to us. That swim on the first morning, are you going to divide it into the series of your actions and the series of things happening to you? It would not be natural to do so. Actions, then, 8

The point is argued more elaborately in Bittner 2001, chap. 12.

The Units of Living


contrary to what we may have come to suppose, are not the born candidates for a conception of our histories. It may be objected that lives are no good, either. For one thing, the kinds of life distinguished in the lives tradition are awkwardly schematic. People’s lives will not easily and will not illuminatingly be ordered into such types as life of pleasure, life of honour, life of understanding. Most people pursue lots of different things, one after the other and one beside the other, and presenting their lives as aimed simply at pleasure or understanding would not do justice to these lives. However, this is not an objection to conceiving of our histories in terms of lives. It is only an objection to using too narrow a range of lives for the purpose. The objection has force against the lives tradition with its scanty menu. If instead we are prepared to characterize somebody’s life by the detailed profile of what this human being is or was up to, the objection will be satisfied. And turning to the practice of describing human histories in terms of lives, i.e. to biography, it is indeed part of what we expect from a biography that we are told, not only what happened in this life and not only what got accomplished, but also where he or she was heading or, to borrow Heidegger’s language9, what the projects were to which someone was projecting him- or herself; and projects are plural. Aristotle supposed that one’s various projects would form a hierarchical order under one overarching aim (NE 1094 a 18 – 22). Yet in fact we are not heading ‘To the lighthouse’, as it were. What we are heading for is an entire landscape. As long as a life is understood in terms of such a range of things pursued, it need not present an absurdly schematic conception of somebody’s ways. It may be objected, secondly, that lives are too big to serve as a conception of what we do. We do not lead, for we do not understand, our lives in terms of lives, but in terms of smaller pieces. This is so not for the simple reason that, while leading and understanding our life, we are not yet done with it. While writing a book and understanding what we do, we are also not yet done with it, but we can certainly understand what we do in terms of the book we anticipate completing. The point is, rather, that a life, comprehending whatever project somebody may pursue, can no longer appear within the outlook of the one who is living it. Lives are a matter of what grammarians call futurum exactum or future perfect: there will have been my life. The living, however, are turned to the future, 9

For instance Sein und Zeit § 31, 145.


Rdiger Bittner

not to the future perfect. So lives, lying beyond our ken, do not provide a useful conception of our histories. This objection is valid. Lives with their characteristic closure do not do justice to the fact that what we did and what we start out to do are always mere parts of a whole that we, while living, never capture. In what we do and start out to do we are never dealing with our life. We are involved in this or that story. 4. Stories By a story I do not mean a report, but what may be reported. A story is a part of one’s history the limits of which are determined by what it is a story of. There is the story of my doing philosophy, which turned out to be a rather big piece of my history. There is the story of my going to the market later this morning, a fairly small affair. Big or small, stories are limited: some things belong to my doing philosophy and some things belong to my going to the market, and other things belong to neither. Certainly it may be debatable whether some particular incident belongs to some story or not, but that does not show that there is no fact of the matter as to what the limits of a story are, on the contrary. And stories are nested: the story of my doing philosophy comprises a lot of smaller stories, like the story of my reading Spinoza.—“How far down are you willing to go with these stories? Is there a story of your reading the second letter in line 12 on the fifth page of your edition of the Ethics?”—Yes, there is. Not that this story seems important. It is only that I may go down as far as proves helpful for shedding light on what I am doing and on what I care about in what I am doing. Lives, I argued, do not consist of actions. Neither do stories. They are chunks of one’s being involved with things. So they are not something over which one is master. For the stories in one’s life one is not responsible, only for what one did at some point in this or that story. You are responsible, say, for leaving Jane in the lurch at that point, but not for there having been a story with Jane in the first place. True, you could have cut this story very short, but you could not have erased it, as little as you could have created it. Stories comprise things that come to you: stories therefore you do not make, but only shape.10 10 This is why it is an absurd task that Alexander Nehamas reads Nietzsche as having completed: “Nietzsche created a character out of himself.” Nehamas 1985, 233.

The Units of Living


And stories are the unit of our concerns. This is not to say that you care about what your story with Jane will be like in the end. We are not eager, ordinarily, to leave behind only stories of this or that sort, romantic stories, say, or heroic stories. Nor would we succeed: “il n’y a pas de hros pour le valet de chambre” does not hold for heroes only, but for any one of our figures. The point is, rather, that we care about what happens next in one story or another in which we are involved. We are not concerned with our life being such and such.11 As I said, our life is too large to appear on our agenda. Nor are we concerned with our actions in contrast to what happens to us. There is really nothing particularly interesting about that subset of our history, except perhaps when appearing in court. What we are concerned with is how to get on with this or that story: passing this test, dealing with this client, finishing this job. And if it is stories rather than life or actions which form the frame of reference of our concerns, it should be in terms of stories that we can best understand human history. For thus we would be adopting the view on their history of those whom we try to understand. In the end the idea of stories as units of human history is not new. Aristotle tells us in the Poetics that tragedy represents “one action which is whole and complete and has beginning, middle and end” (1459 a 19 – 20). This at least is how the passage, and similar ones like 1449 b 24 – 25 and 1451 a 18 – 19, is usually understood. The trouble with this understanding is that it does not fit the tragedies we know: they do not contain just one action. Earlier I noted that we do not have settled criteria for the individuation of actions. So we have to be content with our usual ways of distinguishing one action from another. Judging on this unreliable basis, the tragedies clearly represent each a large number of actions. These actions are connected in various ways, and traditional drama theory made such connectedness sufficient for unity of action. This, however, is an interpretative coup de force: connected actions do not therefore form one action. There seems to be just one way of aligning Aristotle’s theory with the tragedies we know: Aristotle must be saying in the passages in question that tragedy represents, not one action, but one story. This has the drawback of making Aristotle use the word “praxis” in different senses, but as this word already proved ambiguous in another

11 Here I correct myself: Bittner 1974, 490.


Rdiger Bittner

way, namely, between “activity” and “action-individual”, that is not a serious loss.12 Materially, this reading makes excellent sense: what is represented in Oidipus Tyrannos and Antigone and Ajas is in each case one story, with beginning, middle and end. It is for instance the story of Antigone’s decision to bury her brother, of the ensuing conflict with Creon and finally of her and Haimon’s death. The best way to understand human history would thus appear to be to think of it in terms of stories like those told in tragedies. Not of course that we constantly, or indeed ever, are tragic heroes. Rather, we are constantly engaged in stretches of things going on that have a unity of the kind we are familiar with from tragedy. Bibliography Arendt, Hannah (1960): Vita activa oder vom ttigen Leben, Stuttgart: Kohlhammer. Aristotle (1991): Nikomachische Ethik, Franz Dirlmeier (ed.), Darmstadt: Wiss. Buchgesellschaft. Birnbacher, Dieter (1995): Tun und Unterlassen, Stuttgart: Reclam. Bittner, Rdiger (1974): “Maximen”, in: Akten des 4. Internationalen KantKongresses Mainz 1974, Berlin: de Gruyter, 485 – 498. — (1992): “One Action”, in: Oksenberg Rorty, Amlie (ed.): Essays on Aristotle’s Poetics, Princeton: Princeton University Press, 97 – 110. — (2001): Doing Things for Reasons, Oxford: Oxford UP. Broome, John (1999): “Normative Requirements”, in: Ratio 12, 398 – 419. Heidegger, Martin (1963): Sein und Zeit, 10th edition, Tbingen: Niemeyer. Hobbes, Thomas (1651): Leviathan. Nehamas, Alexander (1985): Nietzsche: Life as Literature, Cambridge Mass.: Harvard. Nietzsche, Friedrich (1887): Zur Genealogie der Moral I. Stoecker, Ralf (1993): “Reasons, Actions, and their Relationship”, in: Stoecker, Ralf (ed.): Reflecting Davidson, Berlin: de Gruyter, 265 – 286. Thomson, Judith Jarvis (1977): Acts and Other Events, Ithaca: Cornell. von Wilamowitz-Mçllendorff, Ulrich (1913): Sappho und Simonides. Untersuchungen ber griechische Lyriker, Berlin: Weidmann.

12 Here I revoke something I said in Bittner 1992, sect. 4.

Making Sense of Ourselves Holger Baumann 1. Introduction Rdiger Bittner raises two questions at the beginning of his paper: (i) In what forms do things happening to, and things done by human beings come? (ii) When and why do we think of “human history”1 in terms of one form rather than the other? (261)2 As to the first question, Bittner distinguishes—with ample reference to the history of philosophy—four ways of describing human history: in terms of actions, in terms of deeds (“completing of things”), in terms of lives, and in terms of stories.3 With regard to the second question, summarizing the upshot of Bittner’s arguments is more intricate since he seems ultimately to answer a somewhat different question. This becomes apparent when looking at the conclusions he draws from his discussion: stories are said to be primary in describing human history because they are the “unit of our concerns” (273). By conceiving of human history in terms of stories, we adopt “the view on their history of those whom we try to understand.” (ibid.) In contrast, actions are not “the born candidates for a conception of our histories” (271), nor do lives “provide a useful conception of our histories” (272). Obviously these claims do not easily fit the question Bittner takes as his starting-point. While he sets out to illuminate when and why we think of human history in terms of one form rather than the other, Bittner (at least in crucial passages of his paper) is quick to compare and evaluate the respective forms from a certain perspective. Unfortunately, however, he is not very explicit about this perspective, which makes it difficult to discuss his claims or to apprehend the supposedly critical implications his arguments have, e.g., for action theory. 1

2 3

“Human history” is a (slightly misleading) term Bittner uses to refer to “that part of history that involves humans”. More specifically, by speaking of “human history” he refers to things done by and happening to persons throughout their lives. Parenthetical references in the text are to Bittner 2007 (in this volume). Bittner does not claim this list to be exhaustive, and one might well think of other ways of understanding human history, e.g. in terms of plans or projects.


Holger Baumann on Rdiger Bittner

In the first part of my comment, I will try to shed some light on the perspective from which Bittner is speaking. In order to do so, I will briefly consider his notion of stories in contrast to actions and lives, and I will review the reasons why actions and lives are said to fail as “born candidates of our history”.4 It emerges that Bittner purports to speak from the everyday perspective of agents. Conceiving of human history in terms of stories, or so I understand his claim, is just the default mode of human agents who try to get along in their lives and to make sense of themselves and of others; in this sense, stories are primary. In the following sections I will criticize these claims at least in part. I will argue that stories are not primary in this way and that we should not confine ourselves to understanding human history in terms of stories. If we do, something important will be left out. In order to make sense of ourselves, we can and in fact do conceive of ourselves in terms of actions and in terms of lives. Furthermore, there is no reason to settle on one single form of understanding ourselves or to view the different forms as mutually exclusive. In this context it proves helpful to inquire into the reasons why Bittner emphasizes the importance of stories. I will indicate that his central hypothesis is strongly influenced by a broader picture of human agency, according to which we lack control over our lives and are devoid of any ‘overall task’ or ‘substance’. 2. How Do we Conceive of Human History? A story, Bittner asserts, is “a part of one’s history, the limits of which are determined by what it is a story of.” (272) The story of someone’s being the father of his daughter, for example, might consist of all the things done by and happening to the father within the relationship to his daughter. Also, “stories are nested” (ibid.)—they might comprise smaller stories: that the father bought his daughter a puppet when she fell off the shute at the age of seven, that he forbade her to attend that party because of a blizzard, and so on. This is a rather vague account of stories. One might ask, for example, whether there are certain formal constraints on stories, whether a story must conform to specific dramatic patterns, or how to deal with the fact that stories must be told by someone and do not ‘lie out there in the 4

I do not examine the notion of “deeds” in detail because it is not crucial to Bittner’s overall argument.

Making Sense of Ourselves


world’. Bittner does not address these questions in detail, but instead characterizes stories with a view to contrasting them with lives and actions: While the latter are discrete and datable events of our own doing, stories comprise things that happen to us. And while a life is understood as “a qualitative whole the unity of which lies in what basically the human being in question is striving for” (267; my emphasis), stories are limited in scope. Now, it is these two aspects that play a crucial role in Bittner’s case against the view that human history should be understood in terms of actions or lives: Firstly, he questions the importance of distinguishing between what we do and what happens to us when thinking about ourselves. His reasoning is as follows: conceiving of human history in terms of actions becomes relevant only if questions of responsibility arise. However, such contexts do not pervade our lives, and beyond those rather exceptional situations in which we are called to account for what we have done, we do not naturally draw a sharp line between what happens to us and what we do. Thus, actions are not natural candidates for describing our history. Secondly, understanding ourselves in terms of lives is said to be of no use either because lives evade the agent’s perspective. A life is “a matter of what grammarians call futurum exactum or future perfect: there will have been my life.” (271) This way of describing human history is the perspective of biographers; the “living, however, are turned to the future, not to the future perfect. So lives, lying beyond our ken, do not provide a useful conception of our histories.” (271 f.) What do these arguments tell us about the perspective from which Bittner is speaking? We learn that he is concerned with how we normally or ‘naturally’ conceive of human history, and we come to know that this everyday perspective is not third-person, but the perspective of agents living their lives. The question “How do we conceive of human history?” is thus posed from the everyday perspective of agents trying to get along in their lives and to make sense of themselves and of others. From this perspective, Bittner regards actions and lives as unsuitable for understanding human history. We should do better to describe human history in terms of stories, since we always understand ourselves in terms of stories, i.e. limited parts of our history having a kind of unity and comprising both things happening to and things done by us. Against this background I shall now critically examine the two arguments outlined above, starting with the case of lives; throughout my


Holger Baumann on Rdiger Bittner

discussion I also hope to illuminate further what it means to conceive of human history in terms of stories. 3. Lives and Stories At first sight, Bittner’s reason for dismissing the view that agents do or should conceive of human history in terms of lives seems to be quite conclusive: agents cannot understand their history in terms of lives since it is impossible for them to take up that perspective as long as they are living. In fact, the argument even appears to be too conclusive. If we cannot—for conceptual reasons—conceive of ourselves in terms of lives, one might wonder why anybody should ever have argued the converse. Is Bittner fighting a straw man then? One might be inclined to answer this question in the affirmative. Those who undertake to describe human history in terms of lives, one might say, just speak from a different perspective than Bittner. They want to throw light on somebody’s life, not from that person’s own perspective (see sect. 2), but from a third-person perspective after her death. If this were true, Bittner would be comparing apples with oranges. Both stories and lives would turn out to be useful in describing human history, though in different contexts and from different perspectives, and there would be no reason to force a decision between these ways of understanding human history. Suppose, however, that both Bittner and his opponent pose the question “How to conceive of human history?” from the perspective of agents. Is there a way in which the opponent can refute Bittner’s objection? I can think of the following argument: “You are misconstruing what it means to consider our history in terms of lives. Of course, we can never grasp the whole of our life—‘comprehending whatever project we may pursue’ (271)—as long as we are living it. In this sense, we cannot understand ourselves in terms of lives. But consider these examples: ‘If I set out to date this man, I will have been someone who betrayed her husband.’ Or: ‘If I had not accepted this job, I would never have become a self-absorbed and lonely person.’ As soon as we engage in these modes of thinking about ourselves, we consider our history in terms of lives. And it is an important way to make sense of ourselves.” The opponent concedes the point that lives are a “matter of … future perfect: there will have been my life”. But he emphasizes that—although we cannot grasp the life that will have been ours in a quantitative sense—

Making Sense of Ourselves


we can think of ourselves in terms of lives in a qualitative sense. When we settle on significant and wide-ranging issues, we often reflect on these from the standpoint of the life we will have been living. Equally, we sometimes consider our past in that way.5 At this point Bittner could respond that conceiving of life in the ways proposed by his opponent in fact amounts to conceiving of ourselves in terms of stories. There is this story of the woman: “I met this beautiful man at the grocery. He asked me for advice concerning his choice of wine. Now I want to see him again, but I am afraid of cheating on my husband and I know that I will risk my marriage.” Rephrasing the woman’s experience in this way, however, leaves out the crucial element: that this woman does not want to lead the life of someone who will have been cheating on her husband. Of course, we could complete the story by adding this element. But in this case, the woman’s story is nested in the ‘story of her life’ and the distinction between stories and lives is likely to collapse. The claim that stories form the frame of reference of our concerns would be in danger of becoming vacuous. Another way to respond to the opponent’s argument would be to argue that, although possible, we do not think about ourselves in these ways. Bittner sometimes seems to have in mind an argument along these lines. He claims, for example, that “we care about what happens next in one story or another in which we are involved. We are not concerned with our life being such and such.” (273) But as I have indicated, we are sometimes concerned with our life being such and such. Why, then, does Bittner disavow this way of thinking about ourselves? As far as I can see, his resistance is due to a broader picture of human agency he has laid out in other work. By introducing the idea of understanding ourselves in terms of lives, we also introduce the idea of a ‘self ’ that considers its life from a unified perspective and regards itself as being responsible. But Bittner does not believe in such a unified self and he thinks that the idea of being responsible for one’s life is misguided; in a somewhat similar fashion, he believes it to be irrational that we regret things we have done.6 Rather than showing that we cannot or in fact do not conceive of ourselves 5


Of course, we are not always in a position to judge or to foresee the importance a certain event may have for our lives. What I want to emphasize, however, is that we can lead our lives with a view to the life we will have been living. Strikingly, thinking about ourselves in this way entails that we regard ourselves as being in control of/responsible for our lives; by contrast, Bittner denies that we are ever in control or responsible for our lives (see below). These ideas can be found, for example, in Bittner 1992; 2000; 2001c.


Holger Baumann on Rdiger Bittner

in terms of lives, as he purports to do in this paper, Bittner can thus be understood as making the very different claim that we should not conceive of ourselves in terms of lives. I will return to this point in the final section. For now I shall turn to actions and stories. 4. Actions and Stories Two hypotheses play a crucial role in Bittner’s case against conceiving of human history in terms of actions: i) thinking about ourselves in terms of actions becomes relevant only if questions of responsibility arise; ii) that context is of no particular importance to us. I will question both hypotheses, especially the first one. To begin with, let us take a closer look at the relationship between the concept of an action and the concept of responsibility. Bittner relates the following story: as a matter of historical fact, people used to describe human history in terms of deeds first, i.e. in terms of things accomplished in human activity.7 The concept of an action was introduced later, due to a “shift in moral sensibility” (264). Referring to deeds when praising or blaming agents was deemed to be inadequate since it might just have been a lucky coincidence that someone happened to bring about good or bad. It thus seemed appropriate to take into account the “agent’s knowing, deciding, and only therefore doing” what he did (265). The lesson to be learnt from this story is that “perhaps … there are not first actions, and then agents who are taken to bear responsibility for them. Perhaps, conversely, actions were invented as the sort of thing which agents could be asked to respond for in court.” (266) The concept of an action is thus considered to be derivative from the concept of responsibility. If no one had an interest in holding people responsible, we would not have this concept at our disposal, or at least we could easily dispense with it. In my view, this conclusion is unfounded. Bittner draws too heavily on the genesis of the concept of an action and unwarrantedly restricts to the context of responsibility the role that actions play in our active lives. Even if people began to think about human history in terms of actions 7

For example, that my paper is finished or that the kitchen is cleaned are deeds of mine. What is at stake, then, in thinking about human history in terms of deeds are the products/outcomes rather than the processes. In contrast, writing the paper or cleaning the kitchen are actions of mine. They are discrete and datable events, springing from my knowing and deciding what to do.

Making Sense of Ourselves


because of a shift in moral sensibility, the concept of an action proved to be important within our active life and for our self-understanding in other respects, too. Most notably, it is linked to the idea of individual agents, who are active rather than passive and who can shape their lives in an autonomous or self-directed way. Once this point is granted, we might be more sceptical concerning Bittner’s second hypothesis that those contexts in which we conceive of ourselves in terms of actions are of no particular importance—it is based upon a too narrow understanding of the role that the concept of an action plays in our life. But what about Bittner’s general argument that we do not normally draw a sharp line between what we do and what happens to us because it is not natural to do so? The general idea seems to be that in everyday life we do not give actions a privileged position because something important will be left out if we focus exclusively on things we do: “That swim on the first morning, are you going to divide it into the series of your actions and the series of things happening to you?” (270) That would certainly be odd. But again (as in the case of lives), Bittner seems to misconstrue or to construe too literally what it means to conceive of human history in terms of actions. He seems to suggest that if we do so, we will exclusively focus on what we are doing and completely ignore everything else. I am not sure whether this reading of Bittner is correct, but if it is, this strikes me to be a rather strange interpretation (although ‘philosophers of action’ might sometimes provoke this understanding). If we think about ourselves in terms of actions we must obviously take into account what happens to us. We do so, however, in order to shed light on what we do or what we did. Take the following example: “After he had several times accused me of cheating on him and of being selfish, I finally left my husband. I thought it over again and again, because I really wanted to save our marriage, but in the end I decided to break up. I could stand it no longer.” This is the story of a woman who leaves her husband. It comprises things that happen to her—that her husband accused her of cheating on him and so on. But crucially, it is told in order to throw light on the woman’s decision and her subsequent action; the action has a privileged position. In my view, describing human history in terms of actions can be understood along these lines. And conceiving of ourselves in this way seems (at least sometimes) to be indispensable in making sense of ourselves and of others. Why does Bittner argue the converse? And why does he play down the importance of understanding ourselves in terms of actions in such a


Holger Baumann on Rdiger Bittner

tendentious way? Here is a guess: ultimately, Bittner does not believe that we are autonomous agents or that we are ever responsible for anything. As he has argued elsewhere in more detail, there is no way to make sense of the idea that we are masters of our actions (Bittner 2001a; 2001b; 2002). It requires a duality within ourselves—something that decides and something that executes—“and we should better understand ourselves as in this way undivided.” (270; see Bittner 2001, chap. 12) In this picture, “agents do not hold a position above their life from where to direct it. They are nothing beyond what in their life they grew to be.” (Bittner 2001b, 157) And although Bittner concedes (in his earlier work) that we are not used to viewing ourselves in this way, we should and can do so “without a loss”. As in the case of lives, then, Bittner does not really show that we cannot or do not conceive of ourselves in terms of actions in everyday life, but rather believes that we should not. 5. Conclusion To sum up: I have first tried to show that Bittner purports to speak from the everyday perspective of agents trying to get along in their lives and to make sense of themselves. From this perspective, Bittner contends, we are neither concerned with actions nor with lives—we are concerned with stories, i.e. parts of our history having a certain unity and comprising both things that happen to us and things we do. Throughout my discussion, however, it turned out that both understanding human history in terms of lives and thinking about human history in terms of actions can be and in fact are important ways to make sense of ourselves in everyday life. Bittner only dismisses these ways of describing human history because he misconstrues or construes too literally what it means to conceive of oneself in these ways. The admittedly cursory remarks at the end of the last section (as well as the end of the preceding section) suggest a reason for this: Bittner already presupposes a certain picture of human agency, and this picture is, on the one hand, congenial to the view that we do or should conceive of ourselves in terms of stories, while, on the other hand, it is incompatible with the view that we conceive of ourselves in terms of actions or in terms of lives.8 I am not in a position to discuss this original 8

As I have suggested, these two perspectives are internally connected: if we conceive of ourselves in terms of lives, we must also think of ourselves that, to some extent, we have control over our lives or are responsible for them.

Making Sense of Ourselves


and controversial view of human agency here, but for my purposes it suffices to indicate that Bittner’s overall hypothesis is strongly influenced by it. Put more provocatively, one might say that Bittner does not really answer the question “How do we conceive of human history?” from the everyday perspective of agents trying to get along in their lives. He rather answers it from the background of his own theory, telling us how we should understand ourselves.9 Bibliography Bittner, Rdiger (2007): “The Units of Living” (this volume). — (2002): “Autonomy, and Then”, in: Philosophical Explorations 5, 217 – 228. — (2001a): Doing Things for Reasons, Oxford: Oxford University Press. — (2001b): “Agents as Rulers”, in: Grazer Philosophische Studien 61, 143 – 158. — (2001c): “Masters without Substance”, in: Schacht, Richard (ed.): Nietzsches Postmoralism, 34 – 46. — (2000): “Ich kann nicht anders”, in: Betzler, Monika/Guckes, Barbara (eds.): Autonomes Handeln. Berlin: Akademie Verlag, 179 – 191. — (1992): “Is It Reasonable to Regret Things One Did?”, in: The Journal of Philosophy 89, 262 – 273.


For helpful comments on an earlier version of this paper I am indebted to Anton Leist and Christian Budnik.

III. Action and Epistemology

Knowledge in Action Jennifer Hornsby 1. Introduction When you learn what someone’s reasons were for acting as they did, you know why they acted: you have an explanation of their acting. In many philosophical accounts of reason-explanation, beliefs are supposed to be the only cognitive states of mind which need to be brought into the explanantia; what people know is made to seem irrelevant to what they do. Here I shall suggest that we have to allow a role for knowledge in action if we are to have a correct understanding of what it is to act for a reason. I take inspiration from Timothy Williamson who wrote “If we try to leave epistemology out of the philosophy of mind, we arrive at a radically impoverished conception of the nature of mind” (Williamson 2004, 41). I think that epistemology has to be brought into philosophy of action unless our conception of agency is to be radically impoverished. The prevalence of the assumption that beliefs alone (among cognitive states of mind) belong in an account of acting for reasons is owed to the widespread acceptance, on the part of philosophers of mind, of a certain account of actions and their provenance. According to this account, actions are “caused and rationalized by a desire and belief of the agent”, and a reason for acting is sometimes said to be a ‘belief-desire pair’.1 I think that the attractions of the account have to be reckoned with: we ought to be able to understand why belief should have been accorded such a central place in the understanding of human action. But I shall suggest that the story of action told by philosophers who adopt this account is put into doubt when knowledge is introduced. (There is an aspect of the belief-desire account that I shall not be concerned with at all here: I say nothing about non-cognitive states of mind such as desire.) One of my objectives is to bring together two different conceptions of reasons for acting—that held by the philosophers of mind of whom I have just spoken, and that of philosophers who work in ethics. The latter 1

The source of the view in the contemporary literature is Davidson 1963.


Jennifer Hornsby

usually take a reason to be a consideration which favours acting, and not a belief or desire or ‘belief-desire pair’. It can be puzzling how the two different conceptions of reasons are related. I present a puzzle about this in §1. In §2, I look at the accounts of four philosophers each of whom might seem to have a way of solving the puzzle. In §3 I suggest that a solution relies upon recognizing that the things that are agents’ reasons are things which agents know (§3). 2. A Puzzle About Reasons for Acting I start from something which I take to be obviously true, to which nearly all philosophers who write about these matters are committed, but which can seem puzzling. I call it (Conj): it records the fact that two different claims about an agent’s reasons for acting are often both true. (Conj) There are plenty of instances of the following: (B) X had a reason to v : they believed that p, and (F) A reason for x to v was that p.

An example of (B) is “A reason John had for going to the Plaza was that he believed that he would meet Fred there”; and an example of (F), “A reason for John to go to the Plaza was that he would meet Fred there”. One can think of “B” for “belief ”, and “F” for “fact”: we can equally say that “A reason for John to go to the Plaza was the fact that he would meet Fred there”. (B) and (F) record the stuff of reason-explanations. When x’s believing that p tells us of a reason that x had to v, we may be able to explain x’s having v-ed by saying that they did so because they believed that p. When a reason for x to v was that p, we may be able to explain x’s having v-ed by saying that they did so because p. Thus we have “John went to the Plaza because he believed that he would meet Fred there” and “John went to the Plaza because he would meet Fred there”. When both of these explanations can be given, John believed that he would meet Fred at the Plaza, and the fact that he would meet Fred there counted in favour of his going there. Thus John got something right, and acted for a reason there was for him to act. We must endorse (Conj) if we allow that, in practice, often enough, people do get things right and act for reasons.

Knowledge in Action


(Conj) starts to seem puzzling when we notice that we can have an instance of (B) without p’s being true, whereas p has to be true in an instance of (F). It can then seem as if an agent’s acting for a reason both did not require and did require the truth of their beliefs. Let me elaborate on the two points that produce this puzzle. The first point is that a reason a person has can be given by saying what they believed even where their belief was false. The point shows up in this little piece of dialogue in which an explanation is given. X: Why did Kim (a skater) keep to the edge of the pond? What reason did she have for doing so? Y: She thought that the ice in the middle of the pond was dangerously thin. X: Was it dangerously thin? Y: I don’t know; but certainly Kim thought that it was. Here Y knows Kim’s reason in knowing what Kim believed, but takes no view about whether her belief was true. That couldn’t be so unless a person might have a reason for doing something in virtue of believing something false. The other point was that p can only be a reason for someone to act if indeed p (if it is a fact that p). If the ice was not dangerously thin, then even if there had been a reason to keep to the edge of the pond, it was certainly not that the ice was dangerously thin. Unless the ice was thin, the ice’s being thin couldn’t be a reason for anyone to do anything, and thus not a reason for Kim to keep to the edge of the pond, whatever she thought. Thus one can give a reason why someone did something by saying what they believed even when their belief was false, but to provide a reason that there was for someone to act, one has to adduce a truth. It could seem then as though when corresponding instances of (B) and (F) are true, an agent has two different reasons for acting. There seems to be a reason which they would have even if (contrary to fact) their belief was false; and another reason which they couldn’t have—which would not be a reason—unless their belief was true. To have a reason, they are not required to grasp any truth; but in order for something they have grasped to be a reason, it has to be a truth that they’ve grasped. Of course we can allow that there may be more than one reason for someone to do something. But we surely don’t want to say that an agent has two reasons just in virtue of (correctly) believing some single thing—that p.


Jennifer Hornsby

Perhaps all of this is sufficiently familiar that we do not find ourselves puzzled by it. But even if we are not puzzled, we need an account of reason-explanations which allows us to endorse (Conj). Such an account would enable us to reconcile the different conceptions of agents’ reasons that I mentioned at the outset. Philosophers of mind focus more or less exclusively on schema (B): they take a reason for acting to be a state of belief or desire; and they set schema (F) aside. Those who work in ethics focus more or less exclusively on schema (F); and they set schema (B) aside. But given (Conj), we cannot set either of (B) or (F) aside: instances of both are often true simultaneously. We shall see that not everyone finds it easy to come by an account which makes room for instances of (B) and (F) in conjunction. In the next section, I look in turn at things said by Gilbert Ryle, Jonathan Dancy, Michael Smith and Rowland Stout. Each of these writers suggests a way of reacting to the fact that both (B) and (F) concern agents’ reasons. From their accounts, we can learn how not to try to solve the seeming puzzle, and we can discover some pointers towards the account of reasonexplanation that we need. ‘Unusual Cases’ Before proceeding, I should note that there can be cases in which x’s believing something is recorded in an instance of (F). Here is an example. Jack says “Having had nightmares about skating, I can’t help finding myself recurrently thinking that the ice in the middle is thin. I know that it’s dangerous for me to skate believing the ice to be thin: the nervousness induced by my believing this is likely to make me fall. So I’d better keep to the edge of the pond.” In this case, the fact that Jack believes what he does is itself a reason for him to keep to the edge. This is an unusual case, inasmuch as substituends for “p” in (F) usually concern such things as the thinness of the ice, the whereabouts of Fred, rather than the agent’s beliefs about such things. Usually when a philosopher says that someone’s believing that p was their reason for v-ing, they are not thinking of a case like Jack’s. The point of mentioning what I am calling unusual cases2 is to ensure that they are not confused with the usual ones. Such a confusion is easily made. For schema (B) might be summarized by saying that beliefs are reasons for action; and that could make it seem as though a person’s believing that p was ordinarily a reason for acting, in the sense of schema 2

Dancy calls them special and unusual cases (Dancy 2001, 124).

Knowledge in Action


(F). Making this confusion can lead to thinking that the facts that favour agents’ acting are always facts concerning their beliefs. But one need only reflect on instances of (F) in cases that aren’t unusual in order for it to become plain that reasons for acting are usually not facts concerning beliefs. Of course an account of reason-explanation must cover unusual cases. But they are accommodated simply by recognizing that among the reasons for agents to act—among the things mentioned at the place of “that p” in (F)—can be facts concerning states of mind.3 So having mentioned these cases and warned against confusion, I can leave them aside. 2. Four Philosophers on Reason-Explanations Ryle Ryle wrote: To say that he keeps to the edge [of the pond] because he knows that the ice is thin is to employ quite a different sense of ‘because’, or to give quite a different sort of “explanation”, from that conveyed by saying that he keeps to the edge because he believes that the ice is thin.4

Here Ryle reminds us that we can say what agents know, not only what they believe, in explaining what they do. Given my announced agenda, this is evidently a welcome reminder. But I don’t think that what Ryle says here can be right. How might Ryle’s insistence on the difference between knowledgeattributing explanations and belief-attributing explanations be thought to help in understanding (Conj)? Well, noticing that “x knows that p”, like “A reason for x to v was that p”, requires the truth of p, we can link Ryle’s “X v-ed because they knew that p” with our (F). But then, if we accept 3


Beliefs are not the only states of mind that can provide agents with reasons. Jill says: “I know that the ice in the middle is somewhat thin, and I find myself constantly aware of the fact. I fully appreciate that it’s perfectly safe: that the ice is not dangerously thin. Nevertheless it is dangerous for me to skate when I’m aware that the ice is somewhat thin … .” Jill then continues much as Jack did. Jill’s reason now is a fact about what she knows, not what she believes. Ryle 1949, 129 – 30. Hyman 1999, from which I have learnt a lot, quotes a version of this passage from Ryle with approval. Hyman’s emphasis was on how we should understand knowledge, whereas mine is on how we should understand action.


Jennifer Hornsby

what Ryle says, it will be natural to think that quite different sorts of reasons are alluded to in (B)’s instances from those alluded to in (F)’s. And we should not then find it at all perplexing that a person should have two reasons when (B) and (F) are both true. Following Ryle, we might say: “They are indeed different reasons: they belong in quite different explanations, employing different senses of ‘because’.” The trouble, though, is that this could only help us if we shared Ryle’s view that knowledge-attributing explanations are of a quite different sort from belief-attributing ones. This is difficult to swallow, as we can see by bringing someone else into the dialogue that I used in introducing the puzzle: X: Was the ice in the middle of the pond dangerously thin? Y: I don’t know; but certainly Kim thought that it was. Z: It was thin: that was something all of us skaters knew. The previous point was that Y, who knew a reason Kim had for keeping to the edge of the pond, had no need to take a view as to whether Kim’s belief was true. The present point is that Y does not need to take a view as to whether Kim knew the thing she believed. It isn’t as if Y should now say to Z “Sorry: the explanation of his keeping to the edge was of quite a different sort from what I took it to be. I hadn’t realized that Kim knew that the ice was thin.” If, as Ryle says, there were really two quite different explanations, then we should not be able to understand how Z’s remark chimes perfectly well with Y’s. Presumably what ensures that Y and Z aren’t at odds with one another is that Kim very likely did believe that the ice was thin if she knew that it was. Y and Z, so far from having quite different explanations to give of why Kim kept to the edge, can be in total agreement. Ryle for his part appears to think that knowledge excludes belief. If that were right, then we wouldn’t have our puzzle. For if knowledge excluded belief, then cases in which (B) and (F) both obtain would not be plentiful—as (Conj) says that they are. If knowledge excluded belief, then we should find instances of (B) and (F) conjoined only in cases where x believes that p, and p is true, but x does not know that p. ((F) isn’t true if p is false; and on the assumption that knowledge excludes belief, (B) isn’t true if x knows that p.)

Knowledge in Action


Dancy Dancy has sought an account of acting for reasons which satisfies two constraints. Dancy’s normative constraint says that “that in the light of which one acts, must be the sort of thing that is capable of being among the reasons in favour of so acting”; his explanatory constraint says that “normative reasons should be the right sort of thing to contribute to motivation” (Dancy 2000, 105). Inasmuch as reasons in (F) are thought of as favouring acting, and reasons in (B) are thought of as contributing to agents’ motivation, it seems that finding the account that meets Dancy’s two constraints would provide us with a way past our puzzle. I shan’t rehearse the various twists and turns of Dancy’s own arguments. Here I want only to consider a claim he makes about explanation which might be thought to help with the puzzle. Dancy speaks of there being a normative explanation whenever we have an explanation which gives the agent’s reason. We have the ingredients of a normative explanation when we can say: His reason for doing it was that p.

Dancy’s own example is: “His reason for doing it was that it would increase his pension.” Now Dancy tells us that statements such as this belong in non-factive explanations. He says that we see this by considering that it is perfectly acceptable to say: “His reason for doing it was that it would increase his pension, but he was sadly mistaken about that.” But then (so the idea runs), there is an explanation of his doing it which is compatible with the falsehood of his belief and goes as follows: “He did it because, as he supposed, it would increase his pension.” Schematically: He v-ed because, as he supposed, p.

If explanations on this pattern were really non-factive, then there would be a sort of reason-explanation which encompassed both truths and falsehoods as reasons for doing things. We might then say that the real psychological explanation is given by an instance of this schema; then we should no longer have to think of an agent as having had two reasons where instances of both (B) and (F) are true. It is not clear that this can really solve our puzzle, however. One of our data is that when a reason for doing something is that p, it is true that p. This datum is not brought into the reckoning now. Suppose that Jane, unlike the sadly mistaken man, is someone whose pension actually would increase if she did the thing—if she transferred certain funds, let us say.


Jennifer Hornsby

On Dancy’s account Jane is brought into line with the mistaken man when the behaviour of both is explained by saying that they had a certain normative reason. But what we want to acknowledge is not only that people may have reasons for acting whether they believe truly or falsely, but also that people sometimes act for reasons there really are. If Jane will increase her pension by transferring the funds, then there really is a reason for her to make the transfer, and her transferring the funds can be explained by mentioning this reason: she makes the transfer because it will increase her pension. It seems that we should not be content to assimilate all normative explanations to a non-factive sort.5 We know that there is a perfectly good factive explanation available for the case where what the agent supposed was false: “He v-ed because he believed that p”.6 And it is a very strange idea that explanations are ever non-factive. To many ears, “He v-ed because, as he supposed, p” is true only if it is true that p. (One plausible account of “as f supposes” used parenthetically within a sentence s will treat it [as a sentence adverb such as “luckily” should arguably be treated] as conveying something about what is said in s without affecting its truth-conditions. If so, then, given that “p because q” requires the truth of p and of q, introducing a parenthetic “as f supposes” within it will not produce anything nonfactive.) Smith Smith distinguishes between two sorts of reasons—“normative” and “motivating”. Normative reasons are considerations that justify; motivating reasons are psychological states that teleologically explain. Smith’s distinction might be employed in accounting for (Conj). The idea would be that (B)’s instances tell us of motivating reasons, (F)’s of normative ones. The puzzle about (Conj) might then be supposed to be created by a 5


I take it that Dancy himself would not be content. And I wonder whether Dancy’s normative constraint is too weak for his own purposes. Do we not want to say that, when all is well, that in the light of which one acts must be among the reasons in favour of so acting? Dancy allows that causal explanations are factive: he thinks they are to be distinguished from the normative explanations to which his claim of nonfactivity pertains. He describes the target of his 2004 as “the view that in addition to the normative explanation of action there could be a distinct causal explanation that appeals to the agent’s beliefs as causes.” Perhaps one can avoid this view by holding that a normative explanation may coincide with a causal one.

Knowledge in Action


failure to appreciate that there are these two quite different sorts of reason, belonging in different explanations. (Smith’s solution to the puzzle would then be much like the one we envisaged Ryle offering.) Well, Smith’s distinction appears to introduce an objectionable disconnection between the two uses of the word “reason” in (Conj). If a motivating reason is a quite different sort of reason from the normative sort, then perhaps it is to be treated as the sort of reason that explanations of any kind provides. Consider “The window broke because it was in the path of a cricket ball”, which we can paraphrase with “The reason why the window broke was that it was in the path of a cricket ball”. Since windows don’t have reasons for acting, the explanation here is evidently not a reason-explanation. But if a person’s believing that p is to be treated simply as an explanatory state, then it is hard to see why we should think of a person’s believing that p very differently from a window’s being in the path of a cricket ball. And then we should have to abandon the idea that explanations of what a person does which mention her psychological states sometimes reveal the person as under the influence of reason. Thus it has been an objection to Smith that, by separating motivating reasons from normative reasons as he does, he encourages us to see reasonexplanations as ‘purely causal’. Smith has responded to this objection. He acknowledges that in order that the explanations in which motivating reasons figure should genuinely be reason-explanations, a connexion must be forged between normative reasons and motivating ones. In a recent paper, he makes a connexion in the following way. (1) [There is] a rational requirement of response upon recognition of a normative reason. .. When agents recognize that they have a [normative] reason, they are rationally required to respond appropriately by acquiring corresponding motivations. … In other words, (2) [R]eason requires those who judge that they have a normative reason to v to be motivated to v.7

But, as we shall see now, this places undue demands upon agents. If, as Smith says at (1), recognition of a normative reason introduces a rational requirement, then a person does not count as under the influence of reason if she does something in consequence of believing something false. Just as p cannot be a reason for someone to do something unless it is true that p, so no-one can recognize that p is a reason to do something unless it is true that p. Claim (1), then, about response upon recognition 7

Smith 2004, 173. My lettering added.


Jennifer Hornsby

of normative reasons, touches so-called motivating reasons only where the agent’s motivating reason is a true belief.8 We would not have a puzzle, however, unless a person’s reason could be given by saying what they believe even if what they believe is false. Smith’s claim (2) might seem to fare better. Here what is required of a reasonable agent is not that they recognize a (normative) reason, but that they judge themselves to have one. False belief is accommodated, since even if p is false, one may judge that p is a reason to v. The problem now, however, is that (2) offers an overblown account of what reason requires. In order to be responsive to a reason, an agent does not need to judge that they have a reason. Perhaps this becomes clear by analogy with the case of theoretical reason. Where q is a reason for believing that p, a person might be under the influence of reason in believing that p as a result of believing that q, without judging, concerning q, that it is a reason to believe that p. All being well, and their being reasonable, the fact that q leads them to believe that p. Similarly, I suggest, someone’s believing that p may lead them to v, so that they v because they believe that p; and for this to be the case, there is no need for them to judge that they have a reason to v. A person can operate under the influence of reason on occasion without operating with the concept of reason on that occasion. Summing up on Smith, we may say that his account of motivating reasons seems in the first instance to divorce them from the reasons there are to act—from considerations that favour acting. And then his account of how reasons of his two sorts are connected seems to impose an unduly rationalistic view of the reasonable agent. On Smith’s account, so-called motivating reasons have the status of reasons only insofar as they are the products of agents’ judgements to the effect that they have reasons. Stout Stout has an objection to Smith which adds to the one that I have just made. Against Smith, Stout says that on Smith’s view “the only transformations that characterize the rationality of agency consist in the production of psychological attitudes” (Stout 2004, 202). That seems right. As Smith sees things, normative reasons do all their work in 8

Smith might say that he uses “normative reason” as Dancy does, so that “His reason for doing it was that p” does not require the truth of p. If so, then recognition of a normative reason will amount to recognition of something like “A reason for me to v is that p”, and then the objection I make to claim (2) comes into play.

Knowledge in Action


producing states of mind: “reason requires … that [agents] acquire motivational states”. In Smith’s account, there is no place for the influence of reason on what agents actually do. Although I agree with Stout in finding this objectionable, I am not convinced that Stout can solve our puzzle. Like Smith, Stout wants to keep statements of reason and reason-explanations at a distance from one another. But whereas Smith wants to achieve such a distance by insisting that reason-explanations speak only of motivating reasons (of psychological states), Stout wants to do so by “challenging the standard assumption … that a proper account of agency must appeal to the causal role of psychological states”. But Stout then makes it quite mysterious what it might be for a person to be under the influence of reason. The difficulty for Smith was to show how an explanation of someone’s acting which adduces their beliefs could show them to be under the influence of reason. On Stout’s account, the difficulty is to see why someone’s having a reason to v should ever explain why they v-ed. If psychological states play no causal role in agency, it is hard to understand why instances of (B) should ever be adduced in explanation. Well, Stout has an account of what we should ‘more properly say’ when actually we say “The reason why x v-ed was that she believed that p”. It goes as follows: X believed that p and that meant X had a reason to v, and that was why X v-ed.9

But then the pressing question for Stout is: Why, if psychological states play no causal role in agency, should x’s believing that p mean that x had a reason? Presumably Stout will tell us that a person is under the influence of reason if p is a reason for them to v and it is because p that they v. We find no mention of psychological states here. But nor do we find any idea of an agent’s having a reason. There is the fact which favours her acting on the one hand, and there is the agent’s acting on the other hand (and certainly, if the ‘because’ sentence is true, then the one explains the other). But if an agent was under the influence of reason, then must she not also have been cognitively involved, as it were, with whatever was her reason? 9

Stout 2004, 238. Actually Stout there gives a first-person version, and treats a particular case. I’ve generalized. I’ve preserved Stout’s punctuation, but note that the first “and” here must be taken to be outside the scope of the “believed”; otherwise agents would be required to have beliefs about their reasons, as they are in Smith’s account.


Jennifer Hornsby

3. Knowledge, Belief and Reasons The Role of Knowledge The answer, obviously, is Yes. One can only be influenced by that of which one is in some sense aware. Think once again about the skaters who keep to the edge of the pond because the ice in the middle is thin. Now consider George who is quite ignorant of the condition of the ice. Since the ice in the middle is thin, there may be a reason for George, as there is for other skaters, to keep to the edge. But the ice’s thinness cannot now be George’s reason for skating there. It might be that George is sociable, and skates at the edge because that is where the other skaters are; and it might then be true that he skates at the edge because the ice in the middle is thin (there is a two-step explanation of George’s skating there which adduces the thinness of the ice). Still, even if it is because of the ice’s thinness that George skates at the edge, this explanation of his skating there doesn’t give George’s reason. Contrast George with Fran, who knows that the ice in the middle is thin, and who keeps to the edge because it is. When we say about Fran that she kept to the edge because the ice in the middle was thin, we record her reason for keeping to the edge. The mystery created by Stout’s account goes away now. When an agent acts for the reason that p, we have not only the fact that p (on the one hand) and the agent’s acting (on the other hand), but also the agent’s knowing that p. It is important to see that Fran’s knowledge, not her belief, is what matters here. One sees this by contrasting Fran now not with George but with Edmund. Edmund does have an opinion about the ice in the middle of the pond: he believes that it is thin. He believes this because a normally reliable friend has told him so; but the friend was actually quite ignorant of the actual condition of the ice. (For some reason, the friend didn’t want Edmund to skate in the middle of the pond.) It so happens that the ice in the middle is thin—so thin indeed that there is a reason for Edmund to keep to the edge. But still we cannot record Edmund’s reason, as we could record Fran’s, by saying “He kept to the edge because the ice was thin”. It is not enough that an agent should have a true belief that p— not even a ‘justified’ true belief—for them to have acted for the reason that p. It might be thought that the real difference between someone who does and someone who does not v for the reason that p need not be a difference of knowledge. Could there not be a condition satisfied by Fran

Knowledge in Action


and not by Edmund which falls short of the condition that one knows that the ice is thin? Well, it is true that the argument from examples is not watertight. Nonetheless, I think that it has to be allowed that knowledge is what makes the difference here. This is because the difference that needs to be made is just the sort of difference that knowledge makes. The factive character of knowledge ensures that a person’s state of mind, if it is knowledge, relates them to actual reasons.10 If someone does not know that p, then there might be no actual reason for them to do that which their believing that p is for them a reason to do. But if someone knows that p, then the state of mind they are in guarantees the truth of p. We see now that the explanations from knowledge of which Ryle spoke are not merely linked with explanations deriving from (F): explanations of the two sorts go hand in hand. Where a reason which x had is given by “x v-ed because p”: (E) x v-ed because p


x v-ed because x knew that p.

We should explore the significance of (E). It suggests a way of relating (B) and (F), and thus of solving the puzzle. Knowledge and Belief In responding to Ryle, we saw that an explanation which said that Kim knew that the ice was thin was perfectly compatible with one which said that Kim believed that the ice was thin. We need next to see how these two explanations are related. Ryle maintained that quite different relations are expressed by “x knows that p” and “x believes that p”; in his view, knowledge is a skill or capacity; belief is a motive, tendency or inclination (Ryle 1959, 133 – 4). Ryle’s holding this view might explain why he should have thought that knowledge excludes belief. But actually the view is no bar to allowing that a person might very well both know that p and believe that p. Indeed Ryle makes use of an example that helps one to see why belief that p should normally coexist with knowledge that p. 11 In Ryle’s example, a man who believes that the ice is thin is said to “behave … like [a] man who knows 10 See Williamson 2000, chap. 1 on the factive character of knowledge. 11 Presumably Ryle thought that knowledge and belief have different objects. I try to set aside the question whether that is right. Even if “that p” is ambiguous and behaves differently in “x knows –” and “x believes –”, it can still be a question how “x knows that p” is related to “x believes that p”. Even if facts are not the sort of thing that can be believed, it can still be a question whether if one knows the fact that p, one believes that p.


Jennifer Hornsby

until he is embarrassed by the question how he knows”. This makes us think of the man who believes as having (wrongly) treated p as something that he knew: his embarrassment then stems from his failure actually to know it. But in that case, the man, in believing that the ice was thin, was motivated to behave as someone who knew that it was thin would behave. So someone who did know that the ice was thin, though he can be thought of as possessing a skill or capacity, would thereby have towards action the very motive or tendency which Ryle attributes to the man who believes that the ice is thin. Unless possession of knowledge carried with it a disposition to act, knowledge would be a useless capacity in practice. And the practical disposition that knowledge engenders is to behave as if that which is known is true. But if one believes something, one behaves as if it were true. It is no wonder that so many philosophers nowadays endorse the claim that ‘knowledge entails belief ’. But the suggestion here need not be that it is conceptually required that someone in one state of mind (knowing) should be in the other (believing). The connection made between “knowing that p” and “believing that p” is informed by normative considerations. The suggestion is that someone’s knowing that p ought to secure their believing that p. If that is so, then when someone acts knowing that p and acts for the reason that p, their having a reason for acting can be recorded by saying either that they knew that p or that they believed that p. Whichever is said, we know that a reason they had was that p. There is no double-counting of reasons, then, when instances of (B) and (F) are conjoined. False Beliefs What about cases where the agent does not know that p? It is all very well to say that the agent doesn’t have two reasons when p is the case and they believe it. But this does nothing to explain why an agent may have a reason to act in virtue of believing something false (or indeed something true but which they do not know). The question we face now, then, is a version of the question that was pressing for Stout. We need to understand why someone’s believing something should mean that they had a reason. This is easier to understand once it is allowed that an agent who acts for a reason such as is given in an instance of (F) knows that p. We can make use of the connection between knowledge and belief once again. The idea was that in believing that p, a person is motivated to act in the ways that someone who knew that p should be motivated. Thus p, if it is

Knowledge in Action


believed by an agent, is treated by the agent as if it were true. But p, if it were known, would be true. So we can think of a person as having a reason to v of the sort that (B) records when there would be a reason for them to v—and a reason now of the sort that (F) might record—if what they believed were true. This ensures that a person can have a reason to v in virtue of believing something false. And although when they do believe something false, the reason they have to v is probably not a reason for them to v such as (F) records, we still understand what it is for them to have their reason by reference to such reasons as (F) records. (And notice that there is no suggestion [as there was in Smith’s account] that for someone to have a reason, they must have made a judgement to the effect that there is a reason: it could be that someone came to have a reason simply in coming to believe something.) Dancy was right then to say that we adduce an agent’s normative reason when we say “His reason for doing it was that p”, and right to allow that someone might be mistaken that p although p was his reason for doing the thing. But there is no need to resort to putative non-factive explanations in order to see how it can be that p may be the agent’s reason even when it is false that p. In such cases the agent, not the explanation of their acting, departs from the facts. The agent v-s because they believe what they do: they act as if p were an actual reason for v-ing although really it isn’t such a reason. As Dancy himself puts it, they act “for a reason that is no reason”.12 This may sound paradoxical. But we make sense of it by appreciating that the agent’s reason for v-ing is given in an instance of (B), and it is the falsehood of the corresponding instance of (F) which ensures that there was no reason for the agent to v. The Puzzle Revisited Our puzzle was why it should be that in order for a person to have a reason they are not required to grasp any truth, but in order for something a person has grasped to be a reason, it has to be a truth that they’ve grasped. We can see in a summary way how the introduction of knowledge provides a solution by considering these three schemas: (b) x v-ed because they believed that p. (k) x v-ed because they knew that p. (f ) x v-ed because p.

12 This is how Dancy puts it (Dancy 2000, 144).


Jennifer Hornsby

Where (b) gives a reason-explanation, we have an instance of (B); and where (f ) gives a reason-explanation, we have an instance of (F). By introducing (k), we make a connection between the notion of “reason” as it occurs in (B) and in (F), and we see why there should be a single reason at work when instances of both obtain. The connection between (B) and (F) may be made more vivid. For an agent who acts because she believes that p, treats p as if it were an actual reason for v-ing. But then they treat p as they would if they knew it. So (b) might be rewritten thus: x v-ed because they treated p as they would if they knew that p.

This enables us to assimilate an understanding of the motivation of someone who does something for which there is actually a reason to an understanding of the motivation of someone who believes that p. Someone who does something for which p is actually a reason, knows that p. But if p is a reason why they do something, then they know that p, and of course they treat p as they would if they knew it.13 4. Repercussions for the Standard Account According to the standard account in philosophy of mind, which I mentioned at the outset, beliefs alone (among cognitive states of mind) belong in an account of acting for reasons. It is easy to see how such an account is reached. When a person’s acting can be explained using an instance of (F), it can also be explained using an instance of (B); but there are cases—of false belief among others14—when no instance of (F) can be used, but an instance of (B) can be. Thus (B)’s instances cover the ground that (F)’s instances cover, and more ground besides. It seems then that if one wanted an account that covered all of the ground, one might confine

13 One might treat putative counterexamples to ‘knowledge entails belief ’ as examples in which someone, in spite of knowing that p, fails to treat p as one who knew it should. There is no need here to settle whether there are genuinely such counterexamples, but we can note that they might be accommodated with minor modifications. 14 Here and elsewhere, I’ve omitted to deal with cases where an agent knows that p and v-s because p but in which we don’t think that p is a reason for v-ing: it is at best a terrible reason for v-ing. Evidently such cases would need to be covered in a fuller account.

Knowledge in Action


oneself to (B). Belief, one might say, is a ubiquitous factor in acting for reasons. Those who do confine themselves to (B) usually tell a story about action according to which “the causal interactions [of ] mental states and events within an agent … constitute his being influenced by a reason.”15 But when internal states and events are supposed to do all of the work of a reason, the real influence of a reason is not in the picture. For consider that p cannot actually be an influence on anything or anyone when it is false that p. Certainly one knows why a person acted when one knows what they believed; and certainly a person may believe that p and treat it as a reason, even when it is false that p. But can we have an idea of what it is for someone to treat something as a reason, excepting as we understand the actual influence of a reason? An account of acting for reasons must have room for the idea of a person’s acting because p, where the fact that p favoured their acting as they did. Adherents of the standard view might respond to this by saying “We are concerned only with motivating reasons, such as are given in instances of (B); a separate account has to be given of reasons in the sense of (F).” They think that they can tell a story of acting for reasons without taking it into account that there are actual reasons for which people act. They assume that the role of belief in reason-explanation can be understood without introducing the idea that facts are reasons. But their assumption cannot be right if a person has reason to act by virtue of believing that p only insofar as they are then motivated to act in the ways that would be appropriate if they knew that p. In its concern with a ubiquitous factor in action explanation, the standard view loses sight of the most significant factor. An account of acting for reasons must contain a conception of a person as changing the world as a result of a reason.16 But then it cannot lose sight of the work of knowledge in practice. When the facts serve us in practice, they do so because we know them. The disposition to action which knowledge engenders is to behave as if the facts were as they actually are.

15 I’ve picked on a sentence from Velleman 1992, 130 as representative of a widespread view. 16 I make a related criticism of the standard story—that it leaves persons out—in Hornsby 2004.


Jennifer Hornsby

Bibliography Dancy, Jonathan (2000): Practical Reality, Oxford: Oxford University Press. — (2004): “Two Ways of Explaining Actions”, in: Hyman/Steward (2004), 25 – 42. Davidson, Donald (1963): “Actions, Reasons and Causes”, in: Davidson, Donald (1980), Essays on Actions and Events, Oxford: Oxford University Press. Hornsby, Jennifer (2004): “Agency and Actions”, in: Hyman/Steward (2004), 1 – 23. Hyman, John (1999): “How Knowledge Works”, in: Philosophical Quarterly 49, 433 – 451. Hyman, John/Steward, Helen (eds.) (2004): Agency and Action, Cambridge: Cambridge University Press. Ryle, Gilbert (1949): The Concept of Mind, London: Hutchinson. Smith, Michael (2004): “The Structure of Orthonomy”, in: Hyman/Steward (2004), 165 – 193. Stout, Rowland (2004): “Internalizing Practical Reason”, in: Proceedings of the Aristotelian Society 104, 229 – 243. Velleman, David J. (1992): “What Happens When Someone Acts?”, in: Mind 101, 461 – 481. Williamson, Timothy (2000): Knowledge and its Limits, Oxford: Oxford University Press.

A Seeming Solution to a Seeming Puzzle in Explaining Action Katia Saporiti 1. Hornsby’s Claim Jennifer Hornsby begins her paper “Knowledge in Action” by stating the following truism: “When you learn what someone’s reasons were for acting as they did, you know why they acted: you have an explanation of their acting.” (285)1 As is the case with most truisms, this does not tell us much. Think of all the different contexts in which we use the word “why” to raise a question and, again, all the different contexts in which we ask for an explanation. There surely are as many different kinds of things one might wish to know when asking “Why?” as there are different kinds of things one would be inclined to accept as explanations in some context or other. That a reason is something you can refer to in an explanation of something or in an answer to a “why”-question therefore does not say much. What counts as an explanation of an action and what as an answer to the question why a certain person acted the way she did and what counts as a reason for a particular course of action, are open and much debated questions. And in thinking about these questions we should keep in mind that there may be more than one answer to each of them, because there are indeed many different contexts in which we discuss our actions. Not only do we explain actions, we justify and judge them, we endorse them or warn against them, we rationalize them, etc. All these and various other practices we engage in may involve talk about reasons for actions. And what it is that we want to know if in a given context we want to know a reason for an action can always be tested by asking ourselves what kind of answer we would be likely to accept. Hornsby’s main thesis is that acting for a reason essentially involves knowledge on the part of the agent. There is an obvious sense in which this is true: We can attribute beliefs only to subjects whom we credit with knowledge. And only somebody who believes something can possibly act for a reason. But this is, as we will see, not the sense in which Hornsby’s thesis is to be understood. Hornsby claims that “we have to allow a role 1

Parenthetical references in the text are to Hornsby 2007 (in this volume).


Katia Saporiti on Jennifer Hornsby

for knowledge in action if we are to have a correct understanding of what it is to act for a reason” (285). And according to her, it is not knowledge in general that is required on the part of the agent. Rather, she believes that our concept of someone’s acting for a reason cannot be explained but with reference to some specific piece of knowledge attributed to the agent. According to Hornsby, the lack of acknowledgement of the role knowledge plays in explaining action is due to the widely accepted Davidsonian belief-desire-model of reason-explanation. But while philosophers of mind usually take reasons to be pairs of beliefs and desires, moral philosophers tend to take reasons to be facts that favour actions. Hornsby’s aim is to reconcile these two different conceptions of reasons. This, she claims, can be achieved only if we take into account the role knowledge plays in action. To her “the things that are agents’ reasons are things which agents know” (286). Hornsby tries to establish her thesis, first, by introducing a puzzle which she thinks is created by the difference in the conceptions of reasons prevalent in the philosophy of mind, on the one hand, and ethics, on the other, second, by presenting a solution to this puzzle. In what follows I will argue that the puzzle which Hornsby sometimes calls a seeming puzzle is not all that puzzling. Its construction rests on a confused appeal to the way we talk and (as does the presented solution) the unwarranted claim that different ways of talking about reasons for actions, which are of good use to us when we engage in various practices such as justifying, explaining, excusing or condemning actions and agents, should be unified by an account of what it is to act for a reason. While it is true that such an account must allow for the different ways we refer to reasons and while it certainly is a desideratum to shed light on how these different ways relate to one another, it is not clear why there should be one and only one answer to the question what it is that can be rightly called a reason for an action. (The question itself sometimes seems to be a result of an inappropriate hypostasis.) In the end we will also see that a person who considers the alleged puzzle a real puzzle, in the first place, should not be satisfied with the solution Hornsby presents. 2. The Seeming Puzzle and a First Reaction to It Hornsby begins to construct her puzzle by observing that “there are plenty of instances” where the following two things hold:

A Seeming Solution to a Seeming Puzzle


(B) X had a reason to v : they believed that p. (F) A reason for x to v was that p. (286)

Any account of what it means to act for a reason, Hornsby claims, has to endorse this point. And she goes on to point out that while (B) does not seem to imply p, (F) in fact does imply p. For (F) can only be true if p is, while the truth of p is not implied by (B)’s being true. This in Hornsby’s opinion gives rise to the following puzzle: “It can then seem as if an agent’s acting for a reason both did not require and did require the truth of their beliefs.” (287) Later she puts the puzzle somewhat differently: It could seem then as though when corresponding instances of (B) and (F) are true, an agent has two different reasons for acting […] a reason which they would have even if (contrary to fact) their belief was false; and another reason which they couldn’t have—which would not be a reason—unless their belief was true. (287)

A third formulation of the puzzle offered by Hornsby goes like this: “To have a reason, they [the agents] are not required to grasp any truth; but in order for something they have grasped to be a reason, it has to be a truth that they’ve grasped.” (287) A first reaction to this could be the feeling that there is nothing to be puzzled about here. (B) and (F) just are different claims. They can serve to describe different situations. We sometimes act on false beliefs. Those are the cases in which (B) is true, but (F) is not. Luckily, there are plenty of instances where we act on true beliefs. And in these cases both (B) and (F) obtain. In accordance with this first reaction one might try to reformulate (B) and (F) in the following way: (B’) – – (F’) – – –

X believed that, given her situation, it would be reasonable to v if p. X believed that p. X believed that, given her situation, it would be reasonable to v if p. X believed that p. p.

Given this reformulation (B) is implied by (F). But (F) is not a consequence of (B). And (F) implies the truth of p, while (B) does not. The Davidsonian dictum that is at stake here runs as follows: “Actions are events that are caused and rationalized by a desire and beliefs of the agent.” In the light of this first reaction to the puzzle one would have to say that what causes and explains the action is the agent’s belief that p rather than the fact that p. What might be felt to be awkward here is that in (F) what is called a reason seems to be the fact that p. For one might want to argue that if p does not obtain, somebody’s acting on the false


Katia Saporiti on Jennifer Hornsby

belief that p is acting without a reason. Such an agent would be mistaken about having a reason to act the way she does. X thinks she has a reason to v, while in fact she hasn’t. The reason in question (namely, that p) does not hold. After all, how can a fact that does not exist (a state of affairs that does not obtain) cause somebody to act in a certain way? But surely, even in the case where it is true that p it is in an obvious sense not the fact that p which causes our agent to act the way she does, but her belief that p— just as it is not the fact that p, but the fact that she believes that p, which explains her action to us (which makes her behaviour a reasonable one). For if it were a fact that p but the agent did not believe, let alone know, that p, we could not say that she v-ed because p. Does this mean that reasons aren’t causes any more? No, it does not mean that. To have a reason to act the way one does just means to act in a way that is caused by and can be explained with reference to one’s beliefs and desires. But often, and this is what seems to create the confusion here, in talking about our actions we do not refer to our beliefs and desires but to what we believe and desire as reasons we have for acting the way we do. There are different motives for doing this. Sometimes it is just an abbreviated and somehow simplified way to talk, suited to our needs. Often, if an agent acts on a true belief, we do not bother to mention his cognitive states, but only talk about the fact he believes to hold. (We would say “she v-ed because p” rather than “she v-ed because she believed that p”, using the latter kind of expression especially in cases where something went wrong, i.e. the agent acted on a false belief.) And if all goes well, these facts in one way or another are what caused the agent to have the relevant beliefs that made him act the way he did. So reasons are causes, all right, either in the sense that they prompt the action, if we talk about beliefs and desires, or in the sense that they are at the origin of a causal chain that led the agent to have the beliefs and desires that prompted the action. (Obviously, there are very different concepts of causation involved here.) In other contexts, though, we are not so much concerned with the agent’s beliefs (nor are we much concerned with what made him act the way he did) but with looking at an action in the light of the facts. This is often the case when we want to judge an action (or an agent for what he did or did not do). We often discuss different courses of action, talking about their pros and cons, thereby in a sense abstracting from the individual agent. Also, if somebody took a particular course of action despite there being good reasons not to do so, or if somebody didn’t do what there were good reasons to do, we will ask questions as to whether

A Seeming Solution to a Seeming Puzzle


he knew of these reasons or whether he could have known better etc. Our interests in these questions are manifold. Sometimes we want to judge the agent, at other times we want to find out how to improve a kind of behaviour in the light of certain goals etc. But what we refer to as a reason in this sort of discussion cannot straightforwardly be regarded as something which causes an action. Very often the actions whose reasons we are talking about have not even taken place. This does not mean, of course, that our talk of reasons here does not have anything to do with the reasons which explain and cause behaviour. Quite to the contrary, the different practices we engage in when we talk about actions and reasons for actions are involved and mutually dependent. Roughly speaking, we could not explain someone’s behaviour by referring to his belief that p unless we could regard p as favouring his action. And we could not regard a state of affairs as favouring an action, unless ceteris paribus we ourselves were inclined to take the respective action in case we believed the state of affairs to obtain. 3. Resolving the Puzzle Setting aside our first reaction to Hornsby’s puzzle, one may want to modify the reconstructions of (B) and (F) that I have given. (B), “X had a reason to v : they believed that p”, could again be spelled out as: (B’) – X believed that, given her situation, it would be reasonable to v if p. – X believed that p.

But (F), “A reason for x to v was that p”, would come out as something like (F*) – For anybody in the same situation as x it would be reasonable to v if p. – p.

(F*) would amount to the assertion that there was good reason for x to v, whether she knew it or not, and whether she in fact did v or not. (F*) does not entail that x v-ed. Nor does it entail that x believed that p. It does not even entail that, had x believed or known that p, she would have v-ed. (F*) also does not entail that x believes that it would be reasonable for anybody in her situation to v if p. In fact (F*) does not seem to tell us much about x at all. Instead, it tells us something about the situation we take x to be in and about what we would assume to be a reasonable reaction to p or what, given the situation of x, would count as rational


Katia Saporiti on Jennifer Hornsby

behaviour. But in asserting something along the lines of (F*) we do treat x as a rational agent; we assume that, had she believed that p, she ceteris paribus would have v-ed. We assume that, had she believed that p and indeed v-ed, we would have been able to explain her behaviour by saying “she v-ed because p”. We treat x as being one of us, as engaging in our practice of explaining, evaluating, judging and predicting actions. If (F*) were true, then, if x believed that p and didn’t v, we would be entitled to putting forward a hypothesis like “X didn’t, after all, believe p to be a reason to v (and there she was mistaken)” or “x had more powerful reasons not to v”. Note that (B’) and (F’), which resulted from the first reaction to the puzzle, stand in different relations from (B’) and (F*). Again, it could easily be the case that (B’) holds while (F*) does not. But also one can easily imagine cases where (F*) holds while (B’) does not. What is likely to be the most important point, according to Hornsby, is that there are cases, probably ideal cases, we think of when we give belief-desireexplanations of our actions, where both (B’) and (F*) hold. (F*), which I suggested as a reformulation of what could be meant by (F), certainly is not a reconstruction Hornsby had in mind, as she assumes throughout that from (F) we can derive an explanation of the form “x v-ed because p”. But this surely presupposes not only that p and that x did indeed v, but also, among other things, both that x believed that p and that x believed that, given her situation, it would be a good idea to v if p. Furthermore, it presupposes that x v-ed on these beliefs and not for some other reason. But we are not concerned here with giving a full, let alone a reductive, account of what it means to act for a reason. So we can set aside many of these complications. In concentrating on the explanation of actions that have taken place, Hornsby tends to overlook that the way we postulate the existence of reasons for actions nobody ever performs is instructive as regards understanding explanations along the lines of (F). Having started off with the widely shared assumption that “x has a reason to v” means that “x’s v-ing is caused by and can be explained by x’s beliefs and desires”, we can now see how we might want to adjust this. As (F*) implies neither that x v-s nor that she has the belief that p, we might want to say that “x’s having a reason to v” means that x’s v-ing, if indeed she did v in a given situation, could be explained and caused by her beliefs and desires. And we could say that “p’s being a reason for x to v” means that x’s v-ing, if indeed she did v in a given situation, could be explained and caused by her belief that p, if indeed she believed that p.

A Seeming Solution to a Seeming Puzzle


What about the puzzle Jennifer Hornsby has drawn our attention to? (F*) does imply the truth of p, while (B’) does not. Does this mean that having a reason to v at the same time requires having a true belief one acts upon and does not require having a true belief one acts upon? Or, to raise a slightly different, but clearly related, question: Do we use the word “reason” in different ways, does it have different meanings in (B) and (F)? And if so, does this undermine our understanding of reason-explanation? Hornsby insists that any account of what it means to act for a reason has to allow for the fact that there are plenty of cases where both (B) and (F) yield true sentences. This certainly is true. But throughout her paper she also seems to assume that this means that we have to come up with an account that not only allows (B) and (F) to yield true sentences but which in addition to that gives us one core-meaning of “reason” and a corresponding canonical use of the term that could be represented by another schema which would be entailed by both (B) and (F). This canonical use would lie at the bottom of all other uses of the term, which we would then have to regard as derived from it. If by using “reason” in (B) and (F) we refer to different things—e.g. beliefs and facts—‘reasonexplanation’ becomes unintelligible, or so Hornsby seems to think. But this worry is unfounded. While any account of reason-explanation should allow for a variety of well-established practices and, if possible, explain their relations, it certainly does not have to single out one sort of entity which we are justified in calling a reason. Neither is there a reason for assuming that from all the different things we do when we talk about actions and the reasons agents can have or should have for performing them, one can be singled out as underlying all the others. 4. Constraints on an Account of Reason-Explanation In presenting her puzzle and dismissing several solutions to it Hornsby tries to spell out some of the constraints an account of reason-explanation has to meet. A discussion of Ryle’s work leads her to maintain that (B) and (F) should not be treated as if they yielded different kinds of explanations and that an agent’s knowing that p should secure his believing that p. While the latter seems uncontroversial, the former claim is somewhat unspecific. If what is meant is that (B) and (F) should not be treated as excluding each other in the sense that they cannot be true at the same time, then it certainly is true. If what is meant is that there is no difference between (B) and (F), the claim is certainly false, as (F) has


Katia Saporiti on Jennifer Hornsby

implications which (B) does not have, namely the truth of p. This, after all, is how the puzzle is supposed to arise. Hornsby comments on Jonathan Dancy’s observation that it is perfectly all right to say things like “His reason for doing it was that it would increase his pension, but he was sadly mistaken about that”. From this Dancy, according to Hornsby, wants to draw the conclusion that reason-explanations as in (F) are non-factive. Hornsby objects that this does not take into account “that when a reason for doing something is that p, it is true that p” (291). But while Dancy may not succeed in showing that all reason-explanation is non-factive, his observation that we do say things along the following line, is certainly correct: “X’s reason to v was that p, but he was mistaken about p”. This reveals something Hornsby tends to overlook. When we refer to a certain state of affairs as a reason for an action, we sometimes imply that this state of affairs obtains and sometimes we don’t. And this is not, as Hornsby assumes, tied neatly to certain ways of expressing ourselves. Sometimes it is possible to say that p was the reason for x to v, or that x ved because of p etc., even though it was not the case that p. Therefore it is not a ‘datum’ “that when a reason for doing something is that p, it is true that p”, as Hornsby claims. Equally, we may in certain situations say that p is a reason for x to v and accordingly that x has a reason to v, even if x does not believe or know that p. Speech is not as tidy as Hornsby supposes when she confines herself to using the expression “to have a reason” in talking of beliefs that may be true or false, and to using the expression “to be a reason” in talking of facts that favour actions. Hornsby accuses Dancy of inappropriately assimilating an agent who acts on a true belief to an agent who acts on a false belief. She claims that “what we want to acknowledge is not only that people may have reasons for acting whether they believe truly or falsely, but also that people sometimes act for reasons there really are.” (292) But Hornsby does not succeed in explaining why this cannot be done by simply adding that the belief we refer to in explaining somebody’s action is true. Also it remains unclear why we should not want to assimilate somebody’s acting on a true belief to somebody’s acting on a false belief. Besides, talking of “reasons there really are”—presumably distinguishing them from reasons which are not really there—is an illegitimate hypostasis if it leads us into thinking of entities which are justly called reasons and which exist independently of a practice of invoking them. Hornsby puts her point somewhat differently when she asks: “Do we not want to say that, when all is well, that in the light of which one acts must be among the reasons

A Seeming Solution to a Seeming Puzzle


in favour of so acting?” (292, Fn 5) But here again she supposes that only a fact can favour an action. Why should we not allow other things to favour our actions (states of affairs which do not obtain or beliefs, desires, wishes, moods, habits, emotions, aims, goals, norms, laws, rules, challenges, provocations, etc.)? One reason for thinking this may be brought to light by looking at Hornsby’s discussion of Michael Smith’s distinction between normative and motivating reasons and Stout’s claim that reason-explanations are not causal explanations. Hornsby worries that if we separate whatever is referred to as a reason in (B) from what is referred to as a reason in (F) we render reason-explanation unintelligible in that what we get in (B) is the agent’s motivation while what we get in (F) is the rationale of the action. So it seems that only (B) and (F) together can reveal an agent as being rational. If we disconnect reasons in (B) from reasons in (F), we would either have to give up on the motivational aspect or on the rational aspect of reasons, or so Hornsby thinks. This way we would not be able to ever regard a person as being under the influence of reason. From her discussion of Smith and Stout it becomes clear that Hornsby assumes that to be under the influence of reason an agent has to be influenced by a reason, where this means that for an agent to act rationally there has to be a fact that favours and causes his action in a straightforward way. But surely, a person who acts on a false belief can be perfectly rational. And by assuming that we can state diverse things by (B) and (F) and that there are many different things that can be called a reason, we do not commit ourselves to the view that (B) and (F) are completely disparate and that there is nothing that connects the various things we can count as reasons. To summarize: We have to allow that there are many cases where (B) and (F) hold without giving rise to Hornsby’s puzzle. It seems that we can avoid doing so only if we assume that (B) and (F) yield different explanations and that what is called a reason in (B) is not the same as what is called a reason in (F). At the same time, though, we cannot allow explanations along the lines of (B) and (F) to exclude each other. And although we want to make a difference between somebody who acts on a false belief and somebody who acts on a true belief, we cannot allow (B) and (F) to drift apart in a way that renders it unintelligible that people who act for reasons are rational in what they do. Instead, we have to account for the fact that (B) and (F) are indeed closely related. Furthermore, and this is something I do not have room to discuss but which seems to me a convincing claim Hornsby makes in discussing


Katia Saporiti on Jennifer Hornsby

Stout: we must not assume that for an agent to act for a reason he needs to have mastered the concept of a reason. 5. The Seeming Solution Let us now look at how Hornsby tries to accommodate her findings by presenting a solution to her puzzle. She draws our attention to the connection she believes to hold between p’s being an agent’s reason to act the way he does and the agent’s knowing that p. She maintains that, unless x knows that p, p cannot be his reason for v-ing. (On the other hand, she claims that p could be a reason for him to v (296).) Somebody who has grasped a reason has knowledge, or so she maintains. This, according to Hornsby, is one of the things we have to bear in mind if we want to solve the puzzle. She construes a Gettier-like example in which an agent’s true and justified belief that p does not amount to knowledge, and she claims that in this case where p is indeed a reason for x to v (i.e. p is a fact that favours his v-ing) and in which x indeed v-s because he believes that p, we still cannot record x’s reason to v by saying “x v-ed because p” (296). Hornsby does not explain, though, why exactly we should not allow this last formulation to be a correct description of the situation. She just compares the agent to somebody who acts on a true and justified belief which undoubtedly amounts to knowledge as everything is normal and not Gettier-like. Hornsby claims that “knowledge is what makes the difference here. This is because the difference that needs to be made is just the sort of difference that knowledge makes. The factive character of knowledge ensures that a person’s state of mind, if it is knowledge, relates them to actual reasons.” (297) There are a number of obvious responses to this. 1) Gettier-like situations are exceptions and their bearing on a general practice is doubtful. 2) One may doubt that Gettier-examples prove that knowledge involves more than justified true belief, as they describe situations where it may be doubted that the relevant belief is truly justified. 3) It remains unclear what it could be that knowledge adds so that we are allowed to say “x v-ed because p” only if x knows that p. Surely, a true belief is as factive as knowledge! The second point we have to bear in mind to solve the puzzle is, according to Hornsby, that somebody who believes that p treats p as if he knew that p. Considering the following schemata of explanations

A Seeming Solution to a Seeming Puzzle


(b) x v-ed because they believed that p. (k) x v-ed because they knew that p. (f ) x v-ed because p. (299)

Hornsby suggests to rewrite (b) as “X v-ed because they treated p as they would if they knew that p.” By doing this we are supposed to arrive at a better understanding of the way somebody’s acting on a belief is related to somebody’s doing something for which what he believes is a reason. “By introducing (k), we make a connection between the notion of ‘reason’ as it occurs in (B) and in (F), and we see why there should be a single reason at work when instances of both obtain.” (300) What we still do not see, though, is what is accomplished by introducing knowledge that cannot be accomplished by acknowledging that (F) and (f ) can only be true if the agent’s belief that p is true. Hornsby solves her puzzle by ensuring that the truth of (F) implies the truth of (B) and correspondingly the truth of (f ) implies the truth of (b). But this, as we have seen earlier, can easily be done by plausible if not complete explanations along the lines of (B’) and (F’). And anybody who believed in ‘double-counting reasons’—that is, believed that in instances where (B) and (F) both yield true sentences we are dealing with two different reasons the agent had for acting—will still believe so. He would try and argue just as Hornsby has done that x has one reason (namely his believing that p or his treating p as if he knew that it were true) which he would have even if it were false that p, and another one which he would not have if it were false that p (namely p—the fact which favours his action). On the face of it, the assumption that the difference (if there is one) between p’s being a reason for x to v and p’s being x’s reason to v lies in x’s knowing that p does not seem plausible at all. If for p to be a reason for x to v x does not have to know that p, then neither does x need to know that p for p to be x’s reason to v. X just has to believe that p and p has to be true. But a true belief does not amount to knowledge. Furthermore, believing that p does not amount to treating p as something one knows; nor is believing that p to the believer in some sense indistinguishable from knowing that p. If this were the case, why should anybody ever try to find out which of her beliefs are true? What is true, though, is that believing that p just is taking it to be true that p. This corroborates the suspicion that regarding the alleged puzzle there is nothing to be gained from invoking knowledge.

Cognition and Action Anton Leist 1. Ontological and Rational Relationships Between Knowing and Acting We know and we act. We usually realise this upon not knowing something or upon failing to act in a situation demanding action. Errors and ignorance, or a weak will and lethargy, are only conspicuous, however, against a background of knowing and acting working as usual. Do we humans possess any other general attributes? Of course, we are made up of matter, are living creatures, are able to reproduce, are mortal, require oxygen. However significant these features might be, they do not characterise our perspective as human beings: they also hold true for other animals, which unlike us have no or at least no comparable perspective. “Perspective” means a first person standpoint from which to know and act. Human animals are animals whose standpoint is one of knowing and acting, and whose knowing and acting again is linked to a first person standpoint. Closely connected with this perspective of the world are general human characteristics like speaking, communicating, responding, thinking, leading a life, being in a mood, being concerned about something. These characteristics partly denote our rational and controlling abilities, partly our emotional and social ones. They are undoubtedly among the most elementary human characteristics with the consequence that a person lacking these abilities will not be able to lead a normal human life. And yet they are characteristics which rather differentiate and unite knowing and acting, not ones which themselves are independent of knowing and acting. We realize something by speaking, thinking, feeling, responding; and we act out of mental states, moods, concerns or thoughts. In other words, knowing and acting are personal human attributes picked out abstractly and by this providing a conceptual space for all other attributes to be classified within. At least this is the assumption forming the basis for the following. It is an assumption which is relevant to the scope of the question I would like to discuss: How do


Anton Leist

knowing and acting behave in relation to one another and, in particular, can one of these be given priority over the other? The difficulties obviously surrounding this question stem from its being extremely wide-ranging, as well as from the vagueness in the concepts of “knowing” and “acting” in the first place. In general terms, “knowing” covers both immediate perceptions and cognitions acquired through the use of language and theory, and in the following it is used this broadly.1 In both cases (and those in the grey area in between them), knowing is the recognition ‘of ’ something, something being known. Similarly acting, to our conventional understanding, is purposeful as well, with the most generally named aim being to change a situation in the world at large. Knowing and acting are thus both purposeful, even if in different ways. The usual, everyday differentiation is that knowing refers to something existing independently of knowledge, whereas acting produces something dependent on action. It is known that a candle is burning, yet the candle is burning independently of the fact that this is recognized. By contrast, the candle is only lit because it has been lit; its being lit is dependent on the fact of somebody lighting it. The difference between knowing and acting, as shown by these examples, is an ontological difference, for which some philosophers have recently developed the model of ‘opposing directions of fit’. I shall return to this model in more detail later, when I shall also attempt to show that it has to be rejected, along with the ontological difference between knowing and acting itself. Let us, however, assume for a moment that the abovementioned ontological difference does exist. Knowing and acting would then be so different that they could not fall into the same category. How could we then describe their mutual relationship? One possibility would be to say that they have no systematic relationship at all. Let us call this the ‘neutrality thesis’ of their relationship. The potential seriousness of this assumption would depend on what exactly a ‘systematic relationship’ is taken to be. Below I shall name various possibilities for how such a relationship could be formulated, all of which would be contradicted by the neutrality thesis. Independently of this, the following test demon1

The same goes for its latin correspondent “cognition” which is understood to be equivalent in the present text. To ‘cognize’ in English seems more akin to “erkennen” in German, but a more strenuous attempt than to ‘know’. Nevertheless “knowledge” and “cognitions” in the following most of the times will be used interchangedly.

Cognition and Action


strates how improbable the total neutrality of knowing and acting to each other is. Let us imagine some creatures who are able only to know or only to act, without having the respective other ability at their disposal. Since all creatures who act somehow knowingly control their actions, acting without knowing seems hardly conceivable. On the other hand, a knowing totally independent of acting would be a minimalist kind of knowing—if one could even speak of knowing given the reduction to reactions to environmental stimuli. Some non-human creatures react intelligently to the environment without being able to act; and their lack of knowing comes with a lack of ability to act. These cursory anthropological examples point towards a mutual dependence between human knowing and acting, a mutual relationship of constitutivity rather than a lack of relationship. How to describe the relationship between knowing and acting more explicitly? The two possibilities which immediately suggest themselves are to think of knowing and acting as one being embedded within the other, or one including the other. The corresponding hypotheses are thus: Inclusion of Knowing Knowing is a requisite concomitant phenomenon of acting. Inclusion of Acting Acting is a requisite concomitant phenomenon of knowing.

Both theses presume that neither knowing nor acting can be reduced to acting and knowing, but that one still has a precedence over the other. According to the inclusion of knowing thesis, acting takes precedence over knowing; knowing is somehow dependent on acting in a way which is not true in reverse. According to the inclusion of acting thesis, knowing takes precedence over acting; acting is somehow dependent on knowing in a way which is not true in reverse. Both theses presume an asymmetry in the relationship of dependence between knowing and acting; in turn it is suggested that these asymmetries constitute the origins of, and thus the explanation for knowing and acting. A better formulation for these theses would therefore be a dependence of ‘constitution’ or ‘explicability’. Action Constitution (AC) Knowing is ‘constituted’ or rendered explicable through action. Knowledge Constitution (KC) Acting is ‘constituted’ or rendered explicable through knowledge.

Knowing being embedded in action is now expressed as action ‘constituting’ knowing (AC) and, vice versa, action being embedded in


Anton Leist

knowledge is now expressed as knowing ‘constituting’ acting (KC). Now a more precise analysis of ‘constitution’ in this context is required, and an analysis in terms of necessary and sufficient conditions may be helpful. Weak and strong versions of both relationships can then be distinguished: Weak AC: Action is a necessary condition for knowing. Strong AC: Action is a sufficient condition for knowing. Weak KC: Knowing is a necessary condition for action. Strong KC: Knowing is a sufficient condition for action.

These hypotheses are still slightly vague in the sense that ‘action’ or ‘acting’ and ‘knowing’ may be interpreted as abilities and the regular use of these abilities, but can also be seen as individual actions and singular cognitions or pieces of knowledge. Adhering to these differentiations, the AC and KC theses break down further into type and token variants, where, alongside the just mentioned type-type versions, type-token and token-token versions will play a particular role. A type-token version of the weak AC would then be: Acting (of the type A) is a necessary condition for knowing Ki.

And a token-token version would be: Action Ai is a necessary condition for knowing Ki.

Many weak token-token examples are true in a trivial sense: singular actions are necessary for the realisation of singular pieces of knowledge, and vice versa many singular actions are impossible without singular pieces of knowledge. By contrast, the strong token-token examples are problematic because they call into question the very ontological difference which has just been assumed to exist between actions and pieces of knowledge. Following a strong token-token AC, the lighting of a candle, for example, would be sufficient for the recognition of something, e.g. that the candle has been lit, that a candle can burn, etc. If actions and cognitions are ontologically different, however, one cannot be sufficient for the other. And we can at least imagine that somebody lights a candle and does not know that it is burning, will burn, has burnt or can burn. Likewise, and maybe more obviously, somebody can know (in line with the strong token-token KC) that this is a candle, that it can burn, that it can be lit, etc., without actually lighting the candle or doing anything else directly resulting from this knowledge. Knowing things about an object does not per se commit the possessor of that knowledge to any particular action.

Cognition and Action


If we take a look at the relationships between singular cases of acting and knowing, it is true in a trivial sense that one is a prerequisite for the other, even if mutually, meaning that nothing can be deduced about the ranking of AC and KC. Such cases merely reflect everyday contingent relationships. In reverse, the examples containing sufficient conditions seem to be evidence that a strong dependency fails in the light of the ontological difference between acting and knowing and therefore cannot exist. Having presumed this difference, the relationships cannot be sufficient. Nevertheless, regarding the relationship between knowing and acting, cases can be found in which the strong hypotheses are fulfilled without the ontological difference being called into question; and these cases show that the differentiations made do not satisfactorily grasp the relationships. Completing a doctorate is a sufficient condition for gaining knowledge in a particular subject if we assume (maybe counterfactually) that it is impossible to be awarded a doctorate without gaining new knowledge on a subject. And inventing a vaccine is a sufficient condition for producing it if inventing a vaccine involves knowledge about that vaccine and that knowledge is impossible to come by without producing the vaccine. In the first case, the action of completing a doctorate is sufficient for a piece of knowledge; in the second case, knowing about the vaccine is sufficient for the action of producing it. What prevents these cases from calling the ontological difference into question is their characteristic as complex occurrences. Completing a doctorate is a complex action; discovering a vaccine is a complex piece of knowledge. “Complex” means here that the action or piece of knowledge includes other actions and pieces of knowledge and is only possible through them. In turn, this seems to be mutually possible, so that even with these complex cases there are no insights to be gained regarding an asymmetrical relationship between knowing and acting. For this, ‘simple’ actions and pieces of knowledge require closer study. Does the above picture change if we take a look at the relationship between types of knowledge or actions and singular pieces of knowledge or actions? The alternatives would then be as follows: (1) Weak type-token AC: Action of the type A is a necessary condition for knowledge Ki. (2) Strong type-token AC: Action of the type A is a sufficient condition for knowledge Ki. (3) Weak type-token KC: Knowing of the type K is a necessary condition for action Ai.


Anton Leist

(4) Strong type-token KC: Knowing of the type K is a sufficient condition for action Ai.

(1) and (3) seem clearly to apply to a large number of cases, regarding simple actions and cognitions as well as complex ones. This tells us nothing, however, about the relationship between knowing and acting in general, for the dependency of the two abilities is only outlined using individual examples. The strong hypotheses (2) and (4) contravene the ontological difference in the same way as the token-token relationships within the same examples. This comes as no surprise since the type of acting or knowing covers a class of singular actions or cognitions and it would not be plausible for the token to violate the ontological difference, but the type not to. As an interim conclusion we therefore ascertain that the presumed mutual relationship of constitutivity between knowing and acting does exist (as proven by the weak theses), but that clues regarding an asymmetry within one or other constitution thesis, AC or KC, are not found easily. If actions and cognitions are not, as we had excluded, identified one in terms of the other, then they can only be mutually included, one within the other, in complex actions and cognitions. The fact that cases for both types of inclusion can be found renders the fundamental precedence of one over the other improbable. Either we have overlooked an important characteristic of knowing and acting, or there really is nothing more to say about their relationship to one another. In actual fact we have thus far ignored one important aspect of knowing and acting, and it could well be this characteristic which gives rise to an asymmetry, either in connection with the ontological difference or in the form of its dissolution. This characteristic is that of their rationality, or the way in which cognitions and actions are justified. Even if simple cognitions and actions (unlike complex ones) cannot be reduced to the respective other, it could well be that, seen rationally, either cognitions or actions play a priority role. If we assume, relatively uncontroversially, that being rational is an essential feature both of cognitions and of actions,2 we then obtain not simply alternative, but more precise versions of the hypotheses AC and KC: (ACR) Knowing is justified and rationally explained by acting. 2

‘Uncontroversially’ in the sense that all cognitions and all actions have to be related to one or more ideals, not in the sense that all cognitions and all actions necessarily have to be rational.

Cognition and Action


(KCR) Acting is justified and rationally explained by knowing.

Knowing could ideally be founded in acting, or vice versa acting could ideally be founded in knowing, in each case presuming, of course, that the opposite is not the case. ACR would have as a consequence that a part of acting is rationally contingent, i.e. that it guides knowing, but it is not in turn guided by knowing. KCR would have as a consequence that a part of knowing is purely contemplative, i.e. that it is not connected to actions or expressed in terms of them. A precedence of acting would lead to a lack of self-transparency in knowing; a precedence of knowing would lead to a lack of self-realisation through acting. Because of these contrasting limitations and restrictions, it is also of practical relevance which of the two asymmetries applies. The majority of Western philosophers have attempted to circumnavigate the obviously unpleasant consequences of either alternative by viewing knowing and acting, in rational terms, as possessing equal status. Kant, in particular, attempted this with his dichotomy of ‘theoretical reason’ versus ‘practical reason’, just like the empiristic tradition in terms of different kinds of ‘impression’ and ‘passion’ before him.3 In an extension of Kantian and empiricist traditions, current philosophers, in particular those working within analytical philosophy, are likewise keen to bisect reason and so to harmonise the validity of knowing with that of acting.4 In the rest of this article, by contrast, I would like to demonstrate, firstly, that these attempts are doomed to fail, and, secondly, that many things point towards the action constitution thesis (AC) or, to put it in another way, towards knowing being ideally embedded in acting and not vice versa acting in knowing. In order to refute the dichotomy of reason, I shall start by focusing more closely on the problem of compatibility (2.), then criticise an attempt by the dichotomic (empiricist) tradition to harmonise knowing and acting (3.), in order then to demonstrate that an alternative, teleological idea of action is more appropriate (4.). Then follows a positive attempt to show how empirical cognition could be understood as 3


Because the empiricists presupposed a single sensory source of knowledge and because therefore a corresponding distinction was not possible until a later stage in the model, among some of them it remains unclear whether they are rational dualists or not, as Kant is. In many thematic complexes, in particular action theory, the analytical tradition is more strongly oriented towards the empiristic tradition than towards the Kantian one.


Anton Leist

a component of action (5.). In conclusion some practical consequences arising from the priority of action will be observed (6.). 2. The Problem of Compatibility: Factuality and Activity As the terms “knowing” and “acting” are so comprehensive, attempts to make them compatible with each other depend on interpretations and thus transformations. Viewing them as of equal status would mean their being responsible for different ‘phenomenal areas’, each separate from the other. Humeans favour differentiating between ‘empirical knowledge’ and ‘morality’, Kantians differentiate between ‘theoretical’ and ‘practical reason’, and both parties mean, broadly speaking the same thing in terms of everyday objects, even if not, surely, in terms of rational reconstruction. According to their pictures such different types of reason and justification are at work in theoretical knowing and practical acting that one cannot lead back to the other, nor need it do so. A general doubt about these dichotomous images arises, however, from the problem of how one and the same protagonist is supposed to be able to move within such strictly separated and neatly closed-off areas. How is one and the same actor and knower supposed to possess abilities from both areas, that is knowing and acting, if these abilities do not have a common core?5 Having no common core means that knowing and acting in their ideal or rational forms are mutually exclusive. Would it be problematic, really, if the ideal forms of knowing and acting are so different that imagining their fusion within one individual seems impossible? A first response to this question will necessarily have to deal with the mentioned ontological difference, since this difference will also dictate the structure of the different rationalities involved. The difference can be deduced from the different ways in which knowing and acting are related to ‘the world’ or engage with ‘the world’. Knowing encapsulates ‘something’ in the world which has to be given in order to be grasped; acting generates ‘something’ in the world which needs not to be given in order to be grasped. Knowing requires a given, the object of the knowing; acting originates from itself, or rather from potentially ideal (and often manifest) prerequisites for acting, such as orders, desires, 5

I shall neglect the various attempts by Humeans and Kantians to find a common core through Hume’s calm passions or Kant’s Categorical Imperative and faculty of judgment and take them as not convincing.

Cognition and Action


values, intentions. Another (equally figurative) way of expressing the difference between these directions of fit is as an internal/external difference: knowing requires an external object; acting by contrast requires an internal order from which it can originate. Just now I expressed the difference with the help of an ontological image, and it seems to me that this is the minimal version of its expression, linked to a realistic idea of the world. Other, somewhat modified versions are possible. Instead of saying that knowing requires a given, it would also be possible to say that the given is a measure for knowing or that knowing pictures the given. In contrast, talk of reference seems to be weaker and not presupposing a realistic concept of knowing. Within the recent direction-of-fit-discussion many participants also speak of related opposites, such as ‘statements/orders’ or ‘beliefs/desires’. To my ear talk of ‘directions of fit’ has per se a stronger claim to realism than talk of reference, as it is said that statements or convictions (following the truth requirement) have to ‘fit’ the world, whereas actions (following the correctness requirement) have to ‘fit’ orders or desires.6 In one case the direction of fit is from the world to the statements, in the other case from the orders to the world. Among these variations I shall focus in the next section on the (Neohumean) differentiation ‘beliefs/desires’ because it paradigmatically illustrates how the problem of compatibility regarding knowing and acting comprises several subproblems, including that of internally reconstructing an action and that of externally putting action into context within its action-based environment (in turn the object of knowing). How the ontological difference between knowing and acting is comprehended finds expression not only in the different ontological relationships knowing and acting have to the world, but also in the internal understanding of how an action is constructed. If it is right that the traditionally assumed two types of reason are based on ontological differences, then these types of reason can only be understood if these ontological differences are understood. Very generally speaking, they emerge from the question of how an action (the actor) 6

Picture and terminology originate from J. Austin and G. Anscombe and have been popularised by J. Searle and M. Smith (for a history of the distinction, see Humberstone 1992). See Anscombe 1957, §§ 2, 32; Searle 1983; 2001, 36. All later advocates of the distinction expound the problems of the ‘direction of fit’ metaphor: see Schueler 1991; 2003; Humberstone 1992; Velleman 1992b; Zangwill 1998.


Anton Leist

‘fits’ into the world or how coherent representations can be found for the connection between what is given and human activity.7 What is given and activity cannot be reduced one to the other because activity generates givens and, for this reason alone, cannot itself be merely given. In the face of this difficulty regarding the compatibility of givens and activity, attempts at coherence could conceivably come down more strongly on the side of what is given, or more strongly on the side of activity. Accordingly, knowing and acting would be comprehended more as givens or, alternatively, more as activity. The resulting impression that givens and activity really cannot be reduced one to the other does bestow considerable plausibility on the dichotomous tradition, trying as it does to do justice to both aspects. And yet, let us take a look at some of the problems encountered along the way. 3. Compatibility 1: Neohumeanism In the following I shall present a rough sketch of the Neohumean idea of the connection between knowing and acting, representative of a dichotomic attempt to solve the problem of their compatibility. According to this idea, the distinction between knowing and acting is such that knowing is essentially the knowing of something given, whereas acting is characterised by activity. If acting is characterised by activity, then that would seem to be incompatible with it in itself being something given. This in turn cannot be correct, since actions are the objects of knowledge. Acting is thus both a given and activity; examined more closely, the same can be said of knowing. Which begs the question: how are the two to be distinguished at all? One possibility presenting itself at this point would be not to associate knowing and acting with different areas, but to understand both as parts of a single, knowledge-linked activity, one capable of pursuing various aims and one which can be differentiated (in numerous ways) depending on these aims. The dichotomous interpretation of knowing and acting, by contrast, only makes sense if knowing is not viewed as activity, but instead as the picturing of givens. 7

In my opinion, “activity” or “being active” is the most general term characterising human acting. Because it includes mental actions, it extends somewhat further than the usual concept of “action”, which in this point is more uncertain.

Cognition and Action


In the Neohumean model this assumption leads to quite different difficulties within the realms of the empirical and practical. Since the idea of knowledge as one of something given is widely accepted for the realm of the empirical, the model seems unproblematic regarding this area.8 For the realm of actions we either have to assume that actions are ‘known’ through a specific given relating especially to actions, or that there is no ‘practical knowledge’ at all. Such a specific given for actions could be ‘values’, but since values would represent a given so far conceptually removed from empirical facts and things, it seems more reasonable to give up talk of practical knowledge altogether.9 The idea of practical givens with regard to knowledge ‘in the world’, such as values or imperatives, has to be rejected because it does not harmonise with empirical cognition, and the latter definitely cannot be dismissed. Regardless of the terminology chosen, cognition linked to the practical has to take a back seat to empirical knowledge. The consequences of restricting knowing to the realm of the empirical were considered by many to be so profound that, in recent ethics and starting with Kant, various attempts have been made to rehabilitate the concept of practical knowledge and to identify a given within the practical, either in the form of a law of reason (Kant) or in terms of individual moral duties (Ross). Such a reaction and its underlying need are completely understandable. Within the framework of the knowing/acting dichotomy, however, and without discounting empirical knowing, these attempts are doomed to fail from the outset. Under these conditions each instance of ‘practical knowing’ has to cohere with empirical knowing, yet ultimately fails due to their lacking compatibility. Practical knowing detached from acting would have to be a somehow paler version of empirical knowing, and would then no longer be practical knowing. The customary Neohumean reaction to this lack of practical knowledge is substitute talk of a ‘practical rationality’, in terms of ‘cognitively purified’, ‘informed’ or ‘enlightened’ desires, which only masks, however, the true dimensions of the ensuing deficit. Hume was the first to express in more drastic terms what it means to be at the mercy 8 9

I shall return to a critique in the section after next. This corresponds to the Humean tradition. Hume’s doubt as to whether an ought can be deduced from an is is equivalent to there being no mode of cognition for the practical field equal in status to empirical knowledge. For a more recent discussion see Mackie 1977.


Anton Leist

of unalterable desires, in truth to be a “slave” to them (T II.iii.3). Whether or not we call these weakly optimising procedures ‘rationality’, in their restrictedness they lend particular expression to the fact that the ability of knowing, when attributed to the same actor, is a largely contemplative ability of minimal practical use. The protagonists remain cognitively opaque in their practical existence, and that alone is proof of the desirability of an alternative model. The way in which practical abilities threaten to disappear in the Neohumean model due to the constraints of coherence can be seen even more clearly if we take a look at another, more famous problem, namely the compatibility of causality and action, or of explanation and justification.10 The Neohumean has to show how actors as material beings are able to move among material things and are able to change them. Accordingly, actions are not only activity, but at the same time also something given, just like things. What is the relationship, however, between given actions and active ones? If the dividing line runs between these two types of action, how is the actor able to be involved on both sides at once? Is the actor, when he brings forth an action actively, excluded from its causality? And if not, how then is he able to be the originator of his own action? These questions, which are typical within the framework of dichotomic thinking, now surface once again following on from D. Davidson’s classic attempt to find a simple solution to the problem of compatibility regarding causes of actions and the reasons for them.11 Davidson’s proposal, according to which actions both become explicable as events by resulting from the ‘primary reason’ of the actor and, at the same time, rationalise the assigned action (Davidson 1963, 13), would achieve both to embed individual actions harmoniously into the sequences of events outside the actor and make a connection between explaining and justifying inside the action. The empirical and practical areas would then necessarily be made to tally or to be stringently linked to one another within the action, due to the primary reason of the actor in each action. The fact that Davidson’s proposal fails, as well as the way in 10 I purposely avoid here adding the relationship ‘causality and freedom’ because that would load me with a string of further remarks, which could however closely follow the event/agency causality-alternative, to be touched upon shortly. 11 Davidson 1963. Davidson’s proposal is not strictly Neohumean because he neither limits himself to desires as pro-attitudes nor advocates a simple compatibilism. See Schueler 2003, chap. 3. If successful Davidson’s theory would surely be helpful for solving several central coherence problems for Neohumeans.

Cognition and Action


which it fails, elucidates the inherent tendency of the Neohumean model to categorise actions only as givens and to ignore their active part. To what extent does Davidson’s proposal fail? He himself noticed the problem of deviant causal chains, but was never successful in providing an answer and rather ignored them.12 Deviant causal chains point to the conflict between event and agent causality as well as to the insight of only the second one being appropriate to actions. To date opinions differ as to whether agent causality is possible in a non-metaphysical way, but neither is its possibility disproven; I shall come back to this in the next section. It remains to say that unavoidably deviant causal chains are strong evidence for the impossibility of fitting the actor into an event-causally structured world without destroying the fact of his being an actor.13 This first problem corresponds with a second one, according to which Davidson’s connection between explanation and justification either becomes contradictory or leaves open exactly what the connection is. An action is to be explained event-causally from the primary reason containing beliefs and desires. In order to have an event-causal effect, the primary reason must be an attitude which is actually given (or its event correlation). If ‘justification’, in contrast to explanation, includes a normative quality being attributed to what is justified, then it remains unclear how exactly this is supposed to be possible from an actually given attitude. An actual desire, for example, cannot justify an action, but instead only explain why it was carried out.14 In order not to exclude justifications in toto from the causal explanation, we would therefore have to assume that the primary reason has an ideal content, which itself, however, is not given in the same way as the attitude. At the very least we are left here with the problem of the compatibility of explaining and justifying. A more complex (normative) version of the primary reason will hardly manage to be event-causal, however, rendering explanation and justification no longer as easy to coincide. 12 See Davidson 1980, 78 f.; Frankfurt 1978; Bishop 1989, chap. 4 – 5; Moya 1990, chap. 11; Stout 1996, 86 – 95; Velleman 2000, chap. 6 13 For this reason all attempts at a solution reduce to being elaborations of alternative causality proposals: see Stout 1996, 89; Velleman 1990, 125. 14 That is not exactly true, since desires themselves are more appropriately comprehended as idealising representations: see Quinn 1993. Such a concept of desire is not usually found in the Humean tradition, however, and for good coherence reasons. In the Humean tradition desires are in their core mere bodily urges.


Anton Leist

In summary from both observations, in the Neohumean dichotomisation of knowing and acting how to understand actions faces at least two serious problems. First, any claim towards knowledge prioritizing actions, which would do justice to the active character of actions, is dismissed because knowing has to be limited to empirical knowing. The substitute concepts of practical rationality are just a pale version of practical knowledge. Second, even empiricist practical rationality is in conflict with the idea of actions as purely givens. Primary reasons (or their corresponding manifesting events) cannot simultaneously and in their entirety be causes and reasons. Whereas the first problem merely involves a practical scepticism, the second problem reveals the incompatibility of givens and activity in the interpretation of action. Neither observation, and particularly the second, excludes further attempts at harmonisation of course, but, should either succeed, this in itself would surely radically alter the Neohumean model itself.15 What conclusions are we to draw from these difficulties surrounding the Neohumean version of the knowing/acting dichotomy? As hinted at just now, it seems reasonable—against the background of distinguishing between the given and activity—to view knowing and acting as connected ‘in one’ since both knowing and acting have actual, as well as active characteristics. Doubting the distinction between the given and activity seems to make little sense because this is a distinction deeply rooted in our everyday experience of actions: that we are capable of completing an action and thus mastering activity within a context that is given. Neither reducing activity to the given nor dissolving all the given into activity would do justice to this experience. In the Neohumean model, the conception of knowing as something given leads to a conceptualization of acting as also largely given, in other words to creating coherence whereby the given takes precedence. The alternative is to attempt to create coherence between knowing and acting with activity taking precedence. With regard to the internal reconstruction of action, that leads—unlike the belief/desire model—to a teleological model of action which I shall have a look at in the next section. This altered concept of action likewise has a different external embedding of the action into the given as a consequence. However, teleological action 15 In my opinion this tendency can best be seen by looking at Velleman’s development from his Neohumean position in Velleman 1989 to his approaches in Velleman 2000.

Cognition and Action


harmonises only with a teleological concept of knowledge, something I will try to outline in the section after next. 4. Compatibility 2: The Teleological Model of Action The analytical model of action is dominated by the Neohumean model, according to which beliefs and desires are the causal and normative reasons for action, and actions can be understood on the basis of these elements. The tradition of defending this model of action stems from the certainty of already possessing a sufficient understanding of beliefs and desires, so that actions can be sufficiently explained from them. As we saw just now however, this approach destroys actions and activity, so instead it seems reasonable to take as a starting point the reverse certainty of being an actor and to deduce the elements of the action from this alternative certainty. This dichotomy is none other than again that of factuality and activity: the belief-desire model encapsulates a certainty of the factuality; the teleological model elaborates a certainty of the activity. If we dismiss the first as failed, we should attempt the second. Unlike the usual view that an action is understood as ‘teleological’ as soon as it has an aim, as soon as it follows an intention or a reason, in fact the teleological model—as an alternative to the belief-desire model—is made up of various elements which all result from the activity idea, according to which it is part of the self-perception of the actor to bring about results in an intentionally controlled manner and to achieve this successfully in the majority of cases. The actor has at his disposal the ability to recognise when he brings about which results for which reasons, without the necessity of an event causality independent of his judgment. In order to explain an action, no causal law is necessary (nor can one possibly be enlisted), only the correct self-perception of the actor. Without a connection to his own understanding of how and why he acts as he does, his actions are impossible to explain.16 In outlining the teleological model it is necessary to name three of its most important elements in more detail. These three are: ‘teleological reasons for action’, ‘agent causation’ and the priority of understanding over explanation. Let me start with the first, a teleological concept of 16 This is compatible with the fact that in certain cases he might be unsure of his reasons, and that in certain others a third party might be able to evaluate his reasons better than he himself can.


Anton Leist

practical reasons. In the Humean tradition practical reasons are seen as desires, informed desires or pro-attitudes. This concept of reasons is, again due to the priority of causal explaining over the justification of actions, inapt to link intended aims and actions in a subjectively justified way. Desires usually contain target information and thus indicate the direction an action can take, but only in one class of pathological cases are personal desires also reasons to carry out an action.17 In reverse, the identification of reasons with facts which is so popular with critics of desire-reasons is too narrow, in the sense that only facts relative to aims and needs are reasons for a correspondingly oriented action. The aims and needs of the protagonist must be more or less somehow given, only in atypical cases should reasons (facts) without an already given teleological background instil new aims in the actor. A gathering storm is, for example, a reason to seek shelter, but only for an actor who has the aim or need not to get wet or not to endanger himself. In addition to conscious aims, the function of a teleological background can also be fulfilled by needs, motives or feelings, against which, for the actor, facts are reasons. Calling these additional elements likewise teleological seems justified because, in connection with facts, they require a similar action orientation as conscious aims, and because they can at any time be translated into conscious aims or be consciously neutralised or corrected through alternative aims. Needs, motives and feelings are all teleologically structured, are needs for, motives for, feelings towards or for. Facts as practical reasons are therefore ‘teleological’ reasons to the extent that they only become reasons in connection with a teleological background, implicit or explicit. It seems to be unimportant whether this background is terminologically viewed as part of the reasons or merely as their precondition. R. Bittner (2001, 153), a recent defender of the fact-version of practical reasons criticised teleological interpretation of reasons by pointing out that reasons do not refer to aims of actions, but to obstacles or requirements obstructing the aims of actions or leading to them. The reason for opening a window is not that the window is open but so that the stale air can escape from the room. Bittner thinks that the one has nothing to do with the other because the aims of the action ‘opening the 17 One reason for avoiding a visit to the kitchen can be the desire to eat something there. Desires as reasons are usually reasons to avoid something. On the deviating character of desires as reasons as opposed to causes, see Bittner 2001, §196; Schueler 2003, 54, 58.

Cognition and Action


window’ can only be linked to ‘airing’ via an intention on the part of the actor, and intentions, unlike facts, cannot justify actions (ibid., 155). But of course the link between the aim and the connected intention is not only that of an intention, but also and beyond that a potentially real one: if the aim in actual fact brings about or includes the intended result. Therefore aim-based reasons are not simply a variation of desire-based reasons, as Bittner suggests, but fact-based reasons.18 Agent causation is a controversial notion, which does need to be acknowledged in some shape or form, however, if the endemic consequence of deviant causal chains within the causal model is to be avoided.19 Only if the actor (causally) brings forth the action solely ‘in the presence of ’ or ‘in connection with’ his reason, and only if the bringing forth of the action is at the same time causally controlled by the reason with no loopholes is the actor the actor of the action. The familiar counterexamples necessitate identification of being an actor and of causally bringing forth, in other words a form of agent causality. It then remains to be shown how event causality and agent causation can be compatible. D. Velleman’s proposal of presuming a rational desire to be part of both the actor and the commencing acting (Velleman 2000, 141 – 2) seems promising in this context. Velleman’s idea that actors ‘are’ then more or less motives (143) could also be teleologically expressed as the idea that actors are more or less their aims. Actors then identify themselves as actors through the aims of their action.20 18 Not only is Bittner’s critique of teleological reasons for action inaccurate, his own concept of reasons requires an objective in order to distinguish random from relevant reasons. “… anything might be a reason for which someone does something, but only some things actually are such reasons at a certain time, and which are depends on the person’s particular practical stance.” (Bittner 2001, 107) “Practical stance” is explained by Bittner as the manner in which “a person thinks and feels about what happened and might happen” (ibid.). I believe that if one does not comprehend “what might happen” in the sense of an intended aim, this practical stance is too far-fetched. An aim can of course be vaguely intended, resulting in correspondingly vague or weak reasons. If someone has a weak fear that his share prices might fall, bad news from the stockmarket will be a (weak) reason to sell his shares, but only because his fear contains a clear aim, namely not to make a financial loss. Bittner’s own concept of reasons is therefore indirectly teleological as well. 19 See Stout 1996, chap. 3 ; Stout 2005, chap. 6; Velleman 2000, 7, 125; Schueler 2003, 50. 20 On the close connection between explanations of action and agency, see also Hornsby 1993; Searle 2001, chap. 3. Searle’s theory of intentionality and his


Anton Leist

The priority of understanding over explanation means, firstly, that the actor can usually judge not only which reasons he has, but also which of his reasons is the one leading him to act effectively. And, secondly, that there can be no mental or submental event found within the actor from which the acts of the actor could alternatively be explained. A more detailed defence of this thesis would have to show, on the one hand, that ‘unconscious’ reasons are also ones held by the actor and that therefore not all reasons have to be conscious ones; on the other hand it would have to provide evidence that mere submental events lack the authority required to be viewed as valid reasons for action (see Searle 1992, chap. 7; Searle 2001, chap. 4). In a slightly modified way this is a repetition of the problem of why an event should be an event relevant to the actor. The priority of understanding is therefore an inherent component of being an actor and of agent causation. The arguably more or less important problem with the teleological model is the external aspect: how does an actor fit into a world which does not act? The severity of this problem depends on the extent to which an actor and his environment ‘fitting’ each other is problematic in the beginning. The probably most reasonable version of this question is only a special version of the more general question of whether and in what sense ‘realism’ is a problem. My proposal, which is only intended to be brief, is that the relationship between the actor and the environment may be called ‘explicative’ in a sense of practicability, and that it is to be separated from a classical ‘reductive’ relationship of the sort typical of idealistic and realistic accounts. At the moment I should like to postpone a more detailed answer and come back to it within the framework of the following discussion on the notion of realism. What is the output of these deliberations so far? A completely plausible understanding of action can surely only result by embedding the empirical-cognitive elements of acting within the completion of the action—and thus within the view of the actor and the aims behind the action striven for by the actor—and not vice versa, by embedding actions within an action-independent causal-event-construction of the world. The teleological model responds to the problem of compatibility more coherently than the causal model; it corresponds more closely with our everyday self-perception, or better, takes as relevant what is neglected in newer remarks on action theory are mutually incompatible, however. His isolation of the mental and his adherence to realism contradict his interpretation of what it means to be an actor.

Cognition and Action


the Humean conception. In a further step, the problem of compatibility now requires expansion beyond the convictions located within the framework of average actions to knowing itself. Can knowing and cognition itself be comprehended as acting, without qualification? 5. Compatibility 3: Knowing as Acting Are the conditions for successful knowledge principally different from those for average acting? Using the ‘directions of fit’ image, we could say that with knowing it depends on the ‘world’ whether the knowledge is true, whereas with acting it depends on the actor whether he manages to complete an action. And yet is this difference actually appropriate? The success of both knowing and acting generally depends on both the world and the skill of the actor. Discovering how prions are transferred between sheep depends on how prions are transferred between sheep, but also on which transfer paths are investigated in detail by the scientists. Finding out how to infect sheep with scrapies depends on the use of prions, but also on the fact that prions transfer scrapies. Skilful acting and the world’s constitution are the two necessary factors for knowing the world and for changing it by acting. The aims behind an action are different for knowing the world and for changing it, and, as more complex examples will demonstrate, pursuing these aims in the one case necessitates pursuing them in the other. This dependency alone seems to refute the idea that they could be mutually incompatible. Two objections to any reduction of knowing to acting immediately come to mind: first, that knowing is not acting at all, but reception; and second, that knowing may be acting, but is also reception. The complete equation of knowing and reception, as implied in the first objection, is a trademark of empiricism and may be considered refuted since Kant pointed to the non-receptive role of our cognitive concepts. Only the second objection remains troubling, therefore, according to which a particular class of actions can be separated as being also receptive, whereas the rest are not. One problem with this objection results from its terminology: either “receptive” is linked to the sense organs, in which case the objection remains reductionist; or “receptive” is understood nonsensually, in which case average world-changing actions are also partly receptive. A force resulting from the resistance of the wood is applied to the hammer and the arm holding it when driving a nail into the wood. A


Anton Leist

force from the water is applied to a swimmer, enabling him to stay on the surface. Both hammering and swimming, as well as actions like examining or noting, clearly show that knowing and acting cannot be separated to the extent of arriving at pure actions or pure cognitions. The person hammering or swimming requires simultaneous cognitive control of her actions. The examiner or note-taker requires simultaneous activity of her perceptions and ascertainments. Hammering and examining are actions and perceptions rolled into one. The only thing distinguishing them is their different aims, so that once again we have to question what separates cases such as hammering or examining as far as their aims are concerned. At first sight the answer seems simple: the former has to do with a changing of the world, the latter with ascertaining whether something is the case. Hitting a nail into some wood seems to be one thing, and ascertaining whether or not one has managed it quite another. According to this line of thinking, the aim of knowledge-aimed activities is to ascertain whether or not something is the case, whether or not something is given, whether or not a statement is true. The problem with this phrasing of the aim of activities, however, as emphasised by the pragmatists, is that it remains unclear whether anything at all, and more exactly what, is being referred to with such an aim (see Rorty 1991; 1998). Viewed intuitively, we seem to be using such an aim to capture a certain class of actions, but this intuition once again depends on the different ‘directions of fit’, and it is doubtful whether we should be using this model. What fundamentally opposes it within the present context is simply the circumstance that it is used to separate and at the same time identify a class of actions. Actions such as the hitting of a nail are characterised by the fact that the intended outcome, assuming it is realised, is a result of the action. The result is attributed to the action; the action brings it forth. If ‘knowledge-aimed actions’ are to have a special ‘cognitive aim’, then this aim is either an aim behind the action in the same ‘bringing forth’ sense, or—and this is suggested by the different ‘directions of fit’—it is not an aim at all. If the ontological difference between cognitions and actions is true, cognitive aims cannot at the same time be aims behind actions. The advocates of the ontological difference normally concede that actions cannot have knowledge as their aim in the same way actions aim at the results of acting, and yet they stick to the analogy. In their eyes, perceiving that a candle is burning is not an action in the same way that lighting the candle is. And yet it remains unclear why perceiving that a

Cognition and Action


candle is burning should be an action in any way if it is also to be described as a given, i.e. if somebody can only perceive the burning if it is true that the candle is burning. The lighting of a candle is an action if the actor really does bring about the burning: if the candle is burning because he made it burn. Perceiving the candle cannot be an action in this sense because the burning of the candle does not depend on whether the perceiving brought about the burning. The perceiving brings about a state of being perceived; yet the content of the perception is not the state of being perceived, but the burning of the candle. The cognitive and active parts of the action thus fall apart, and the concepts “cognitive action” or “cognitive aim” are contradictory if taken explicitly. The best answer to this observation seems to me to be twofold. First, knowledge-aimed actions are, like all actions, marked by the distinction between something given and activity, a distinction which we have no need to deny. The nail is not yet in the wood, I am hitting the nail into the wood. The candle is burning, we perceive that the candle is burning. I shall return shortly to whether this distinction is problematic or whether it is in need of justification. If, as is the case so far, it is comprehended as a distinction which we fundamentally experience in connection with our acting, then cognitions and actions are to be seen as completely equal in this regard, indeed in many cases the result of an action is only possible in connection with a cognitively registered result. An amount is paid, we perceive that the amount is paid. If nobody perceived that the amount was paid, the amount would be deemed unpaid. Second, the difference between knowing and acting can only be a distinction ‘within’ the acting because knowing is not simply given. The option that aims of knowledge could be separated from other aims of action by the help of an ontological difference is ruled out if, as we have just seen, the difference between the given and activity is true for all actions, including usual, world-changing actions such as the lighting of a candle. Aims of knowledge and cognition are ones, we could say, which can be referred to using cognitive terms such as “recognising”, “perceiving”, etc. This, however, provokes the return question of what distinguishes these words from ones like “lighting”, “hammering”, etc. Once again this will have explained nothing. My proposed explanation consists in a qualitatively different, but altogether only gradual difference in terms of action control existing between these types of action. All actions are linked to an intention to change something in the world, which per se can be viewed as a an attempt on the part of the actor to control part of


Anton Leist

the world.21 The difference between knowledge-aimed actions and other world-changing actions is to be found only in a more or less direct intention to control. In order to make talk easier I will call knowledgeaimed actions ‘epistemic actions’ in the following. The lighting of a candle with the aim of making the candle burn is a direct control attempt with regard to the state of the candle, part of the room (the part illuminated) and additional resulting potential actions. Perceiving a non-burning or a burning candle with the aim of establishing whether the candle is burning is a potential control attempt with regard to additional potential actions involving the candle. Now, it is undoubtedly true that knowing whether or not a candle is burning opens up a potential control over dealings with that candle—for example, control over whether the candle represents danger, whether it will be used up and thus not be available later, etc.; what is doubtful, however, is that knowing the state of the candle means no more than contributing to this potential to control. Is it not so that the additional potential actions result from the knowing and not vice versa? And are the ensuing possibilities not too diverse in any case to circumscribe a knowing of the candle in this way? Both objections are worth considering and show that the aims behind epistemic acting are graded and that potential action control merely represents the overriding aim and thus the one linking it with world-changing acting (direct action control). Epistemic actions fall more precisely into two categories, namely routine ones and experimental ones. Routine epistemic actions are the product of experimental ones and thus only possible under their premise. I shall return to them shortly. Experimental epistemic actions are ones in which examining actions are performed with the purpose of expanding action potential. ‘Experimenting actions’ are experimental manipulations or operations on objects intended to reveal their behaviour and characteristics. In addition to experimental epistemic actions, nonepistemic experimental actions are also possible: the latter are directed straight towards an intended action success, whereas the success of epistemic actions consists only in expanding action potential. In turn, 21 A spy does not wish to change anything about the position of some papers, but he does wish to change something about his information concerning the content of those papers. A swimmer, putting himself at the mercy of the waves, retains in control for as long as he is able to change the situation. Parachuting is an action as long as an opening of the parachute follows the free fall. If the chute does not open, the action becomes a mere happening.

Cognition and Action


experimenting actions are the same in both areas and always directed straight towards success. If one wishes to establish whether an egg is boiled, it is necessary to crack it open and examine its consistency. If one wishes only to establish whether an egg is boiled, one examines it with the purpose of expanding the potential for actions with the egg (including creating it in the first hand). The potential nature of examining renders the handling of the egg an epistemic type of acting which has a fundamentally experimental character. This feature of acting is not eliminated by the fact that a wish to eat the egg may ensue. And vice versa, the potential nature of examining remains even if the original intention was to eat the egg and examining was purely for the purpose of establishing its edibility. In turn, this description will probably be conceded easily, with the single objection that the expanded potential for action results from the aim of establishing the condition of the egg, and not vice versa that the aim of establishing the condition of the egg results from the aim of expanding the potential for action. Whether or not the potential for action is expanded, does, after all, depend on the actual constitution of the egg. This objection overlooks the role of examining actions, however. They have the direct function of establishing the characteristics of things; their link to merely potential dealings with those things renders them experimental actions. Establishing whether or not the egg runs out when the shell is cracked open is, accordingly, an action resulting in the established state of the egg. According to this reconstruction, whether or not something is the case always depends on whether an action succeeds or fails, whether the egg is made to run out or not. Several objections to this action-oriented reconstruction of knowing seem obviously to present themselves. Firstly, not all cognitive processes in the conventional sense are linked to examining actions. To this extent, experimental acting only seems to cover part of knowing. Secondly, examining actions (like probing the state of the egg) would be better described with the aim of ‘establishing a fact’ rather than a direct changing of the world. If this is true, then the representation ultimately fails to dissolve knowing within acting because this aim is precisely what is to be elucidated. Thirdly, the difference between the given and activity remains unexplained. If this difference can simply be assumed, why not others too? Why does it deserve special treatment? The answer to the first objection is to be found in the abovementioned routinised epistemic actions. They are ‘routinised’ to the extent that they are linked with statements not backed by actual


Anton Leist

experimental actions. Many everyday perceptions, as well as many backgrounds to experimental actions, are of this kind. We live in a highly artificial world and our perception of it is based on previous experimental and technical, world-changing actions. We only perceive that an egg is ‘done’ because we have access to the practice of boiling eggs and thus to a wealth of information regarding eggs, water, temperature, etc. Even apparently simple, fundamental perceptions like colour perceptions require a background of conceptual differences, of changing and controlling light, spatial position, etc. Not every instance of cognition is experimental, therefore, but every one is dependent on experiment. Second, do examining actions not in turn have to leave open what is to be examined with them? Do they not differ from technical actions in this respect? The merely inquisitive person wants to see whether the egg runs out, and that is all. The hungry person wants to know whether the egg is done and whether she can already eat it. This contrast is correct, and it is also correct that the term “examine” is a cognitive and not openly a world-changing concept. Examining actions extend to direct changing actions, such as the cracking of an eggshell, which can also be a direct success or failure, but cannot be reduced to them. It is also questionable whether the aim behind and success of examining actions does not inevitably have to be an irreducibly cognitive aim, so that the cognitive content expressed in terms such as “proving”, “perceiving” or “truth” could not be reduced to a successful action. The answer is that the success consists in the extended potential for action; because the ‘success’ is difficult to measure, this answer may appear less than satisfactory. The systematic connection between what is actually known and some additional potential for action seems uncontroversial to me. If the actor has no knowledge about the constitution of the egg, then he can purposefully do less with it than if he knows something. If he realises an egg as being boiled he will have more success with that egg than if he mistakes it for being boiled. These observations are indisputable. But the objection was directed at the elucidatory relationship: because (and insofar as) the knowing person correctly knows the state of the egg, his actions will be more successful, and not vice versa. Moreover, or so the objection could continue, if the protagonist can know that his acting has been/will be successful, then why not also know that a state independent of action, such as that of the egg, is given? Does not an explanation of knowing through acting already presume a certain knowing, namely that of the action and its effects on the world?

Cognition and Action


In these questions and answers, an everyday view of the relationship between knowing and acting and a special, comprehensive view of the same are becoming confused, however, rendering the entire preceding attempt somewhat peculiar. The comprehensive view endeavours to provide an answer to the question of how knowing and acting are generally connected; the answer being that knowing is always within the framework of and dependent upon acting. Therefore, establishing the success of an action cannot be reiterated as establishing an isolated state (egg is runny, egg is hard), but only as establishing the result of an action. Anyone regarding the consequences of actions in the manner of actionindependent states ignores the comprehensive view; he attempts in no way whatsoever to elucidate a general connection between knowing and acting. That is why the circularity objection raised just now is based on a misunderstanding of what exactly is requiring elucidation. Naive or philosophical realists can only raise such an objection because they assume other preconditions. If they confront the connection problem, naive realism falls away and philosophical realism proves to be unsuitable. But did I not also, according to the third objection, assume without justification a difference between the given and activity? If nothing may be assumed, nothing can be justified. It thus has to be a matter of showing what can be assumed and why. The alternative would be to reject all justifications which extend beyond the everyday as impossible and superfluous, a position which R. Rorty incidentally adopts in many of his texts.22 Even if it were viable without contradiction, a reduction to naivety surely is not possible for people asking questions. In addition, we really do need orientational aids exceeding the local validity of everyday routines and certainties. It is not unreasonable to enquire into the connection between knowing and acting. On the other hand, an extended attempt at comprehension need not be organized as foundational, in the Cartesian style of a deductive argument; instead, it can be a selection of attempts to reveal connections between parts of our world, the most important ‘parts’ being knowing and acting. What right do we have simply to assume a difference? 22 By this I mean a partly justified tendency of Rorty’s to dissolve philosophical ‘problems’ rather than solve them. But how far can dissolution be taken? So far that no more problems are left? By wishing to understand general concepts better than is usually the case in our everyday and scientific contacts, we already begin to acknowledge some new and nasty problems, which cannot at the same time be dissolved. So we surely are not to fall back simply into common sense.


Anton Leist

So much can be said, perhaps: our self-perception as protagonists would resolve itself without this difference. We act against a backdrop of the given and only act because some things are not given which we would like to see. We refrain from an action or are content with the outcome of an action because we are content with what is given. The motive for, and necessity of, acting would be ungraspable without this distinction, as would a reverse-reference to the effect of an action and/or how we acted. We do not have to comprehend ourselves solely as acting within a world in order to explain many things in that world; but in comparison to a self-perception as solely knowing, a self-perception based on acting does have considerable advantages, at least if the preceding considerations are correct. And all our other characteristics, at least in general, are already covered if we comprehend ourselves as acting. Describing knowing completely as a type of acting does have the advantage of not having to assume realistically that the world makes our knowledge come true, but only that the success of a class of our actions (epistemic actions) is of a more indirect kind than the success of others. It also has the advantage that the locally maybe enticing image of opposing directions of fit need not be blown up into a comprehensive picture. And it has the additional advantage of rendering further explanation superfluous regarding the unity in different parts of the person resulting from various abilities.23 Truth proves to be a form of the good: slightly different but still comparable. “‘The true’, to put it very briefly, is only the expedient in the way of our thinking, just as ‘the right’ is only the expedient in the way of our behaving.” (James 1975, 106) 23 Velleman’s discussion of the ‘directions of fit’ model suffers from both of these problems. Despite his ontological abstinence (2000c, 105 Fn14, 250 Fn12), he remains stuck to the realistic picture when he talks about “getting the truth value right” or pursue the “really true” (251 – 2). Throughout it remains unclear how truth as a state is supposed to become an aim for action (252). Also, it remains unclear how the two aims truth and realisability (2002b) are to be made mutually compatible. Humberstone (1992) elevates the distinction into a distinction between thetic and telic system intentions and, and in so doing introduces a new type of ontology. I like Zangwill’s proposal better which, by contrast, is an explicitly normative attempt: “The normative functionalist idea is that beliefs and desires have distinctive normative liaisons.” (1998, 195) According to this proposal, the ‘fit’ criterion is to be reconstructed for convictions and desires as a different normative criterion for success, even if in details Zangwill’s proposal remains dependent on fundamentalist intuitions, according to which perceptions and desires dichotomously rationalise convictions and intentions (196).

Cognition and Action


6. Practical Consequences? If acting and knowing cannot exist side by side as equals, then not only must one of them be included within the other, but one must also, sadly, be limited by the other. At the beginning of these remarks I formulated this as: either the actor cannot be completely lucid in his actions, or cognition must remain purely contemplative, i.e. impractical. Whether or not these presumptions and any ensuing verdict are informative, however, depends on what yardstick is used to measure complete transparency or to measure the extent to which cognition translates into complete practicality. Viewing one or the other model as limited is only informative against the backdrop of a possible alternative. If one model dispels the other, however, such an alternative cannot exist. A practical reconstruction of knowing such as the one suggested here has not, therefore, in contrast to perhaps sceptical fears, as a consequence any nasty restrictions of self-knowledge. Acting is what knowing primarily consists in. The distinction between a more contemplative knowing on the one hand, and a more technical knowing on the other, results far more from competition between more potential and more direct, instrumentally aimed action, as well as from their mutual practical suppression in everyday life. It is very tempting to act directly rather than to pursue uncertain, potential aims. And it often seems more satisfactory to achieve direct results than to develop indirect, potential ones. The lives of mechanics, salesmen or politicians outwardly seem more fulfilled and more successful than those of scientists, philosophers, or poets. The inherent danger of their actions, however, consists in the fact that they cut themselves off from other potential actions. Purely technical or commercial acting would lead to endless repetition, which is why it requires a permanent renewal from the framework of epistemic acting. Purely epistemic action would lead to a continuing expansion of possibilities without the latter ever being realised. What has to be substituted for the misleading idea of self-transparency is nothing less than the overall practical success of balancing these different ways of acting.24

24 I would like to thank H. Baumann and K. Saporiti for comments and S. Kirkby for linguistic help.


Anton Leist

Bibliography Bishop, John (1989): Natural Agency: An Essay on the Causal Theory of Action, Cambridge: Cambridge University Press. Bittner, Rdiger (2001): Doing Things for Reasons, Oxford: Oxford University Press. Davidson, Donald (1963): “Actions, Reasons and Causes”, in: Davidson (1980), 3 – 19. — (1971): “Agency”, in: Davidson (1980), 43 – 61. — (1980): Essays on Actions and Events, Oxford: Oxford University Press. Frankfurt, Harry (1978): “The Problem of Action”, in: American Philosophical Quarterly 15, 157 – 162. Hornsby, Jennifer (1993): “Agency and Causal Explanation”, in: Heil, John/ Mele, Alfred (eds.): Mental Causation, Oxford: Oxford University Press. Humberstone, I. Lloyd (1992): “Direction of Fit”, in: Mind 101, 59 – 83. James, William (1975): Pragmatism, Collected Works Vol. 1, Cambridge/MA: Harvard University Press. Lenman, James (1996): “Belief, Desire and Motivation. An Essay in QuasiHydraulics”, in: American Philosophical Quarterly 33, 291 – 301. Mackie, John L. (1977): Ethics. Inventing Right and Wrong, Harmondsworth: Penguin Books. Moya, Carlos J. (1990): The Philosophy of Action, Cambridge: Polity Press. Quinn, Warren (1993): “Putting Rationality in Its Place”, in: Quinn, Warren (1993): Morality and Action, Cambridge: Cambridge University Press, chap. 12. Rorty, Richard (1991): “Pragmatism, Davidson and Truth”, in: Rorty, Richard (1991): Objectivity, Relativism and Truth, Collected Papers 1, Cambridge: Cambridge University Press. — (1998): “Is Truth a Goal of Inquiry?”, in: Rorty, Richard (1998): Truth and Progress, Collected Papers 3, Cambridge: Cambridge University Press. Schueler, George F. (1991): “Pro-Attitudes and Direction of Fit”, in: Mind 100, 277 – 281. — (2003): Reasons and Purposes. Human Rationality and the Teleological Explanation of Action, Oxford: Oxford University Press. Searle, John (1983): Intentionality, Cambridge: Cambridge University Press. — (1992): The Rediscovery of the Mind, Cambridge/MA: MIT Press. — (1995): The Construction of Social Reality, New York: Free Press. — (2001): Rationality in Action, Cambridge/MA: MIT Press. Smith, Michael (1994): The Moral Problem, Oxford: Blackwell. Stout, Rowland (1996): Things That Happen Because They Should. A Teleological Approach to Action, Oxford: Oxford University Press. — (2005): Action, Chesham: Acumen. Velleman, J. David (1989): Practical Reflection, Princeton: Princeton University Press. — (2000): The Possibility of Practical Reason, Oxford: Oxford University Press. — (2000a): “The Possiblity of Practical Reason”, in: Velleman (2000), 170 – 199.

Cognition and Action


— (2000b): “The Guise of the Good”, in: Velleman (2000), 99 – 122. — (2000c): “The Aim of Belief ”, in: Velleman (2000), 244 – 281. Zangwill, Nick (1998): “Direction of Fit and Normative Functionalism”, in: Philosophical Studies 91, 173 – 203.

Notes on Distinctions Joachim Schulte 1. One of the most striking features of Leist’s paper is the generality of its approach. Presumably, this is meant to stimulate readers to reflect on, and perhaps rethink, their basic ideas in the area covered by Leist’s treatment. I, for one, felt greatly stimulated by reading this paper. In particular, I felt encouraged to respond in kind; that is, I regarded myself as entitled, perhaps even requested, to make a number of general observations more or less directly inspired by Leist’s remarks. There is a good deal in Leist’s paper that I find myself in agreement with. His staunch defence of the view that cognition involves acting is admirable. His dissatisfaction with certain widespread but all too simple models of human action should be applauded. His inclination to question the usefulness or aptness of received distinctions is something I find myself in sympathy with. At the same time, it is his way of dealing with, employing, and occasionally helping himself to, certain distinctions that has given rise to most of the following observations. As I have said, most of my remarks are of a very general kind, and are hence to be taken with a pinch of salt. After these fairly sweeping comments on the subject of distinctions and Leist’s discussion of certain distinctions I will mention two brief points connected with specific passages in his paper. 2. Before getting down to the first distinction (or ‘distinction’), I want to say something very commonplace which may serve as a foil for subsequent remarks. First, the success of a distinction, I take it, depends on its corresponding to intelligible or perceivable differences. If, for instance, I am told about different categories of angels, I do not have to be acquainted with angels to understand the distinctions involved; but I do have to grasp the intended differences between them in terms that would enable me to recognize them as members of this or that class of angels if I encountered them in a story, on stage, or in reality for that matter. In short, a distinction which does not involve some sort of criteria for (re)identifying members of the classes to be distinguished is not really comprehensible and fails to serve a clear purpose. Second, to be successful a distinction has to distinguish between things or terms belonging to the same (higher) category or figuring on the

Notes on Distinctions


same level of discourse. If, for example, you and I stand in my kitchen and I explain to you that this is a lentil and that is a bean, I may be said to have (ostensively) explained, and distinguished between, certain kinds of seed. If I go on to point to an artichoke, I may still be said to have distinguished between sorts of vegetable. But if my next step consists in pointing to a mortar or the dishwasher or a picture on the kitchen-wall, the best you will be able to come up with is a category like “things in his kitchen”, which does not seem particularly helpful. The point is: if two things can be shown to be distinct, this is by no means tantamount to showing that they are instances of two sides of a (helpful) distinction. This point is connected with the fact that we expect a distinction to be such that it will allow us to enumerate its subordinate categories or to form a relevant list, e. g. seraph, cherub, guardian angel …; beans, lentils, chickpeas, … . 3. Now, let’s try to apply these general considerations to the title distinction of Leist’s paper: cognition vs. action. The first difficulty with this distinction is that these terms are extremely general and vague and therefore hard to clarify or even define. Dictionary definitions like “cognition: the mental act1 or process by which knowledge is acquired, including perception, intuition, and reasoning” or “action: the state or process of doing something or being active” (Collins English Dictionary) are such that while one does not actually want to quarrel with these attempts at elucidation one does feel that the question “In contrast with what?” seems strangely pointless. And if that question has no point, it will be difficult to see why one would want to look for, or try to think up, criteria for telling when or where one is dealing with a case of cognition. What about membership in the same category or class? Again, it seems that we are dealing with completely disparate items. There is nothing to be compared or contrasted with here. And there is surely no appealing category that offers itself to accommodate both sorts of thing. But perhaps one could bring in the idea of a level of discourse mentioned above as well as the notion of a list of items. In this vein, one might say that pairs of contrastive terms of this kind (theory and practice, reflecting and doing, etc.) have played a great role in the history of ideas in general and in the history of philosophy in particular. To this, however, I would reply that in most intelligible cases these sorts of terms were not 1

NB: the dictionary does not seem to find it easy to talk about cognition without mentioning some kind of act(ion).


Joachim Schulte on Anton Leist

self-explanatory, they were clarified by their contexts and more or less illuminating examples, sometimes by the views they were meant to express opposition to. The real topics of discussion were more specific views and examples, and it was in their light that the general contrast was understood. Whether the remaining cases can be regarded as intelligible is at best an open question. And in part a possible answer will depend on whether we can somehow apply the idea of a list to the supposed distinction between cognition and action. This idea of a list is important as it hints at a claim which clearly stands behind most traditional versions of the title contrast: people wanted to say that these were the two sides characteristic of human beings. But this is a bit like saying that a bicycle consists of two wheels and the frame (comprising saddle, handlebar, pedals and a few other items). Saying this sort of thing does not involve any kind of distinction; it is just an enumeration of different, disparate, items forming a whole if arranged in a certain way. The right question in such cases is “What does a bicycle [a human being] consist of ?” and not “What are the distinguishing (contrastive) properties instantiated by bicycles [human beings]?”. Cognition and action are not items that would figure on a list prepared by Martian anthropologists trying to answer a question like “What’s special about human beings?” or “What is it that distinguishes human beings from those chaps from Alpha Centauri?”. 4. That cognition and action do figure on such a list is something one might be tempted to say for the reason that one believes that cognition and action are bound to be the sort of thing those (rational) chaps have in common with us as well as with these Martians. This kind of statement would surely be welcome to many, but to my mind it would be a far too easy way of fending off an important worry about certain presuppositions forming the background of the title contrast and informing much of Leist’s discussion. The worry is this: that suggesting that cognition and action are somehow exhaustive categories when it comes to characterizing human nature pushes aside the question whether these admittedly central notions are, if put together, adequate to yield a sufficiently accurate picture of what it means to count as a human being. This worry operates on two levels. On the one hand, it amounts to the question whether the two opposed categories—cognition and action—cover everything they are meant to cover; e. g. whether “cognition” suffices to give a satisfactory picture of that domain of

Notes on Distinctions


human capacities and achievements which it is meant to capture. This question is discussed by Leist when he deals with the view that knowing is essentially, or at least in part, a form of receptivity (§ 5). I think that his specific way of putting the matter must to some extent rest on a preconception which I will discuss in a moment. But if it were merely meant to express a biased reading of the notion of cognition as amounting to the claim that coming to know something is essentially a matter of sense perception, we would be dealing with a super-empiricist dummy that cannot really be worth discussing. If, on the other hand, this worry is meant to highlight the question whether the pair of terms “cognition” and “action”, understood as forming a vaguely exhaustive pair, may be misleading if one wants to arrive at an approximately satisfactory picture of human nature, we are getting on to something important. For this particular way of representing the manifold contemplative and active sides of human existence tends to give the impression that man is basically an active being. Leist’s specific account, with its emphasis on cognition as a kind of acting, gives the impression that someone is a human being only if he is actually doing, or at least striving to do, something. This picture, I think, leaves out something important, something I would want to regard as an essential part of human nature. Here I can of course only hint at what I mean. But I think it is quite obvious that many forms of utterly passive behaviour, of idleness, even laziness, and receptivity (in a sense which is not reducible to the super-empiricist form of sense perception alluded to above) are indispensable dimensions of our existence. And what is most important: they are dimensions of our existence that are not covered by terms like “cognition” or “action”. When we say that something dawns on us, that it pops into our mind, or simply occurs to us, we use these and similar idioms to express that we arrive at certain insights, not by some cognitive process or other forms of toil and sweat, but by devoting or abandoning ourselves to an issue which is given full rein in such a way that it may take possession of us and our faculties and will, by virtue of its dominion over us, sometimes (sometimes!) grant us knowledge (knowledge!) of a kind that is not captured by Leist’s term “cognition”, nor arrived at by anything that he would be willing to acknowledge as a kind of action.2


It is clear that a more detailed discussion of these issues would have to go into the role played by practices and techniques that have become so entrenched as not to


Joachim Schulte on Anton Leist

This point is important. And its importance is partly due to its connection with what we want to regard as the (Utopian) ideal involved in our favourite notion of human nature. I am afraid that if we are satisfied with the cognition-action kind of ideal which I sense in the background of Leist’s discussion, we will end up with a figure of classical proportions which at the same time is just as thin as the paper it is drawn on. I think we need something much thicker, much more complicated, much more messy; and that would, on the one hand, mean that the cognition-action pair is not sufficient to describe the figure we mean when we speak of “a human being”, and on the other hand it would mean that we have no clear list on which “cognition” and “acting” can be supposed to feature as items falling under some intelligible category or class of attributes. There is one additional matter to be mentioned here. If you try to extend the realm of activities or processes covered by the terms “cognition” and “action” to such an extent that they can be hoped to capture most of what it means to be a human being, it will be detrimental to your strategy to connect these notions (as Leist occasionally does) closely with ideas like ‘rational explanation’, ‘reason-giving’ or ‘intention’ and their ilk. For if you do so connect them, they leave too many glaring gaps between their streamlined ‘cognizing agents’ and the familiar figures of human beings we know and, sometimes, hate or love. 5. Here there is one last point on distinctions. While Leist seems willing enough to put most traditional distinctions or ‘dichotomies’ in question, he regards what he calls “the distinction between the given and activity” or between “what is given and human activity” as irreducible—a distinction that it would not make sense to cast doubt on, especially as it mirrors an important aspect of our everyday experiences of acting: we feel that we perform actions that take place in contexts that are somehow given. Well, this last statement is no doubt true. But what does it amount to? It amounts to saying that we can describe some parts of our experience as things that are given (things that simply happen, simply happen to us, etc.) and other parts as things that take place because they are done by us. But that is as far as it goes. Beyond that, I can only signal incomprehension: as far as I can tell, there is nothing that is simply given play a noticeable role in what we may want to call ‘knowledge-acquisition’. That is, we would have to discuss the whole area of our ‘second’ nature.

Notes on Distinctions


and nothing that is simply and straightforwardly a product of my or anybody else’s activity. To be given and being due to a certain activity are paradigm examples of context- or description-dependent predicates. Any attempt at transforming them into absolute category terms is bound to go against the grain of Leist’s own inclinations. He would be well advised to jettison them without regret. 6. Now I come to the first of my specific, and perhaps minor, points. On his opening page Leist uses a way of talking that seems to become increasingly common (which does not render it more intelligible). He speaks of human beings as having a specific, or particular, perspective: they have or occupy “a first-person standpoint from which to know and act”. Granted, I can make, or try to make, sense of the idea that I look at my actions from another person’s perspective or point of view or standpoint—which means that I can attempt to look at what I do or have done in the light of other people’s interests, prejudices, sentiments, opinions, cultural preconceptions, etc. To the same extent, I can imagine other people’s endeavours to look at matters from my point of view. I think these are exercises we have learnt to perform more or less well when we were brought up by our mothers or grandmothers, and the quality of our judgements will largely depend on the success of this sort of upbringing. But what would it mean to say that, in living and acting, I occupy my own perspective, or even a (or my?) first-person perspective. In contrast with what? Well, I may change my perspective or my point of view, but that does not amount to occupying a perspective different from my own. And where I have no choice in the matter, it simply makes no sense to speak of my perspective or my point of view. By the same token, in this kind of context it makes no sense to speak of perspectives or points of view at all. So, the sooner we drop this meaningless kind of talk, the better for our discussions of this sort of topic. 7. My last small point is not so much in the nature of a worry or an objection but more in the nature of a question or puzzle. Leist, like a number of other writers, uses the Peircean terminology of types and tokens as applied to actions. I am not clear about the consequences of this way of employing these notions. If I am not mistaken, this terminology was invented for the purpose of dealing with and capturing certain peculiarities of signs, conventional signs. I am aware of the fact that


Joachim Schulte on Anton Leist

the word “sign” has been used in all kinds of inflationary senses, but I won’t let this bother me here. For present purposes, the term “signs” may cover items like letters, musical notation, or words—but neither events nor scratches. So, what does it mean to prefix the words “type” or “token” to words like “action”? I must confess that I do not know the answer. My suspicion is that these words are simply used to indicate the difference between particular actions and actions of a certain type. But “action of a certain type” seems to indicate a very different sort of category from “typeaction”. To mention just one thing: if there were ‘type-actions’ they, as a kind, would have to be similar to ‘type-letters’ or ‘type-words’, which is an idea I cannot make sense of. To arrive at a category that would deserve the name of ‘actions of type T’ would involve a lot of hard work, and the result would probably be controversial, to say the least. This suggests that people who are prone to use the type-token terminology in this context wish to spare themselves some of the trouble that goes with finding out about or constructing types of actions.—But I am really unsure about this; and, as I said, these remarks of mine are very much in the nature of questions. There are several further topics that I would have liked to touch on. In particular, I would have liked to ask questions about what Leist has to say about the idea of directions of fit and the connected notion of “a world” or “the world”. But sensible questions along these lines would presuppose knowing Leist’s responses to some of the remarks I have made above. So, we may as well leave the next round of this game to another occasion.

Action and Inquiry Christopher Hookway 1. Epistemology and Action Why might we suppose that there are important connections between epistemology and the theory of action? We can start with a simpler question: what connections are there between knowledge and action? If we can show that there are philosophically important connections beACHTUNGREtween knowledge and action, then we can turn to the second question: given the connections between knowledge and action, are there any connections between the theory of knowledge and the theory of action. Do we need to take account of our philosophical understanding of action when we theorize about knowledge? And do we need to take account of our philosophical understanding of knowledge when we theorize about action? In this paper, I shall mostly be concerned with the simpler question, but I shall offer some speculations about ways in which the two areas of philosophical theory are not independent. There are many ways in which such connections might be identified and explored, but I shall start by identifying two, and then, in the remainder of my paper, I shall begin to develop just one of these. The two strategies exploit connections that go in opposite directions. One possibility is that a proper understanding of action—or perhaps of rational action—will need to appeal to (and place constraints upon) our understanding of epistemological concepts such as knowledge. We might argue that knowledge only matters because of its role in guiding action. The second possibility is that we can only understand how we can obtain knowledge when we take account of the facts that knowers are typically agents and that obtaining knowledge will often depend upon this fact: we can often obtain knowledge only if we are prepared to do things. One way in which we might carry out the first strategy is to begin from the observation that knowledge and justified belief (the phenomena that are normally taken to be central to epistemology) are important to us because they provide us with resources which can be employed in deciding what to do. Broadly speaking, states of knowledge and justified


Christopher Hookway

beliefs are valuable because actions based upon knowledge and justified beliefs are more likely to be successful than actions which are not so based. One way to defend this point is to follow Tim Williamson in arguing that our knowledge can always serve as a premise for practical reasoning (see Williamson 2000). If this is accepted, then we can focus on either (or both) of two consequences that it appears to have. First, theorists of action who are interested in practical reasoning need to use the concept of knowledge in setting out their views. And we might then suppose that epistemologists’ views about what knowledge is could provide information relevant to a better understanding of rational agency: if knowledge requires justification then so do the premises of our practical reasoning; if knowledge is true belief which tracks the truth, then so do the premises of our practical reasoning. The features that we appeal to in order to distinguish knowledge from mere true belief should be features that are also relevant to the rationality of our actions. And second, if it is accepted that the value of knowledge lies in its providing premises for successful practical deliberation, this may have implications for how we think about knowledge: we try to understand knowledge by asking what properties a state must have if it is to have this sort of role in practical deliberation. For example, we might try to defend a form of contextualism about knowledge or justification by arguing that how much justification is required for something to be a premise for practical reasoning depends on the risks of disaster we incur when we act on the basis of it. If the acceptability of premises for practical reasoning is context sensitive and such premises are typically items of knowledge, then it seems, knowledge, too, must be context sensitive. Epistemology would then need to appeal to our theoretical understanding of practical deliberation because of the role that knowledge fills within such deliberation. These connections are interesting and controversial, but I shall not explore them further here. The strategy I want to consider exploits a different but equally intimate connection between knowledge and agency since it holds that the subject matter of epistemology is itself a form of action or activity. There can be no doubt that much of our knowledge results from cognitive activities. I find out when the meeting I need to attend will take place by finding my diary, opening it at the appropriate page, and looking at the appropriate place on the page: I carry out a sequence of actions designed to take me to the correct belief about the time and location of the meeting. I find out when I need to leave home to attend the meeting by reflecting on how long it took me to do so last week, and considering

Action and Inquiry


whether I recall any information about traffic conditions that would lead the trip this time to take more or less time. In each case, I carry out a goal-directed activity that, if all goes well, provides me with knowledge or true belief. In evaluating my belief, we may have to reflect on how I carried out the inquiry (what structure the sequence of actions I executed had) and also evaluate those actions as a means to achieving my cognitive goal. Sometimes we will criticize my belief because my inquiry was poorly carried out or inappropriate; sometimes we will criticize it because I ought to have carried out some inquiry when, in fact, I failed to reflect upon my opinion at all. So we find a second kind of connection between knowledge and action: the rationality of my belief can depend upon the activities I carry out or the activities I neglect to carry out. Here, too, we find a possible connection between theory of knowACHTUNGREledge and theory of action. If we sometimes evaluate our beliefs by evaluating the actions and activities that we carry out (or neglect to carry out) when we form those beliefs or reflect upon their standing, then the rationality of our beliefs can depend upon the practical rationality which is manifested in our inquiries. In that case, it would be unsurprising if information from the theory of action were not relevant to the study of epistemic evaluation. But this does not take us very far; it does not show that the impact could be at all substantial. I now want to distinguish two different ways in which this relevance can be spelled out, one of which is more radical than the other. Just how does the activity of inquiry or deliberation relate to the objects of epistemic evaluation? If we look at familiar textbooks of epistemology, we find that the objects of epistemic evaluation are states. We are concerned with states such as beliefs, and the evaluative questions concern whether they are true, whether they are justified, and whether they count as knowledge. These are the main terms of epistemic evaluation; and they apply to our cognitive states. The task for epistemology is to explain what these different evaluations of these states mean, and to identify whether we actually possess any knowledge or justified belief. Such evaluations can be used as part of a survey of someone’s overall epistemic health. So far, we have seen that inquiries are important as means to ends. An inquiry is a good one if it will give rise to justified beliefs or knowledge; but it is the state of knowledge or justified belief that is of primary interest. So actions are important as means to our cognitive goals; but the primary objects of epistemic evaluations are not actions or activities, but are states. We defend methods of inquiry by showing that they give rise to knowledge or justified belief. As well as


Christopher Hookway

making states the primary objects of epistemic evaluation, this has consequences for the primary vocabulary of epistemic evaluation: it is made up of words like “knows”, “is justified in believing” and so on. Even if this is all that is to be said about the connections between epistemology and the evaluation of action, it is easy to see how results from the theory of action might have epistemological importance. The picture we have just offered employs some assumptions about what the goals of inquiries are: our goal is to obtain knowledge or justified belief, and it is reasonable to suppose that we seek true belief. Both Richard Rorty and Donald Davidson have recently argued that Truth—and therefore presumably Knowledge—cannot serve as the goals of our inquiries. In each case, this rests upon the combination of two claims. The first is a principle about action: we can adopt something as a goal only if we are able to recognize when that goal has been achieved. This is grounded in some views about how goals serve a normative role in the regulation of action. The second is a general acceptance of fallibilism: however well we inquire, we cannot eliminate the possibility that we have gone wrong. We cannot be “absolutely certain” that we have not made a mistake (Davidson 2000, 67). If they are right about this, then action theory provides the material for showing that our cognitive goals cannot be true belief, knowledge, or belief that is objectively justified. I shall return to this argument in section 2. However, there are epistemic evaluations which are internal to the activity of inquiry and which are not primarily concerned with the evaluations of states of belief. One way we can see this is by considering the role of reasons in theoretical deliberation. As well as looking for (defeasible) reasons that support a particular conclusion, we have to be alert for reasons for doubting our assumptions, reasons for suspecting that there may be defeaters for these defeasible reasons, or reasons for collecting more relevant information before we continue, and so on. Identifying such reasons is internal to the process of inquiry, and it is a distinctive kind of epistemic evaluation. Inquiry is a kind of practical activity; and identifying reasons is a kind of epistemic evaluation that is internal to such activities. Hence we would expect that the sorts of questions that action theorists raise about other kinds of action should arise too in connection with activities of inquiry. This is the possibility that I want to begin to explore. Section 2 of the paper explores in more detail what this approach to epistemology might involve. We then identify two ways in which this has an impact upon how we think about central epistemological problems.

Action and Inquiry


First (section 3) it can alter the ways in which we formulate the problem of scepticism, seeing it as a challenge to our ability to exercise our agency, responsibly and effectively, in carrying out inquiries. Then (section 4) we consider how it enables us to address issues about the nature and importance of knowledge. The final section attempts to draw some general conclusions suggesting that, in each case, we can formulate the question as an instance of a general question about practical rationality and action. 2. Epistemology as Theory of Inquiry In this section we consider how a theory of epistemic evaluation might look if we accept the strategy I shall call ‘epistemology as theory of inquiry’. The fundamental idea involved in this approach is that epistemic norms are primarily concerned with regulating an activity, the activity of inquiry. In that case, some, at least, of the norms that govern our epistemic activities should be special cases of norms involved in the conduct of actions and activities more generally. First, some issues about goals. Inquiry is typically thought of as a goal-directed activity, and we would expect that questions about how an inquiry should be conducted would be sensitive to our understanding of its goal. So what sorts of things can serve as goals of inquiry? We are often told that our aim in inquiry is true belief, but this is, at best, a partial answer to the question: not just any true belief will serve as an answer to a given inquiry, and a fuller specification of the goal should determine which true beliefs would count as achieving our goal. The most natural view is that an inquiry is an attempt to solve some problem concerning what is the case or, perhaps, to find the answer to a question. If I inquire into whether seventy-one is a prime number, than I seek the correct answer to the question “Is seventy one prime?”. If I inquire into the causes of global warming, I seek an answer to the question “Why does global warming occur?”. So we might identify the aim of inquiry as finding the correct answer to the question that prompts the inquiry: inquiry is always an attempt to find the correct answer to some question. In that case, the slogan “Truth is the aim of inquiry” is, presumably, a simplified way of saying that the aim of inquiry is to arrive at belief in a proposition which provides an answer to the question which is true. An answer to a question is correct so long as it is appropriately relevant to the question and is true.


Christopher Hookway

The attempt to explain the relevant notion of correctness in terms of truth can be questioned from two different directions (at least). Some may take it to be too weak: our aim is to arrive at knowledge of a correct answer to the question, not just possession of a correct answer to the question. Or our aim is to accept some correct answer because it is the correct answer. Others, such as Richard Rorty, claim that it is too strong. Rorty endorses a general principle about what is a reasonable goal for action. In “Universality and Truth”, he announces that his “grounding premise” is that “you can only work for what you can recognize” (Rorty 2000, 4). In slightly different terms, you can only adopt something as your aim if you are able to recognize when this aim has been achieved. It is a consequence of his fallibilism, that, although we can recognize when a belief is justified, we have no effective way of determining whether it is true. Given the principle about action, it follows that we can aim for an answer to our question that we are justified in accepting, but we should not aim to arrive at an answer that is true or known to be true. Indeed, since we may be fallible in the standards of justification that we accept, this too may be too strong: all we can reasonably aim for is an answer that meets the standards of justification that we currently accept. In a similar spirit, Donald Davidson insists that “we can’t consistently take truth to be both objective and something to be pursued” (Davidson 2000, 67). In consequence, he proposes that we should give up the idea that “truth is a norm, something for which to strive”. Truth is objective “but pointless as a goal” (ibid.). It is clear that, properly understood, these issues turn on some quite general issues of practical reason, issues about when something is available to us as a goal, and about the ways in which our understanding of our goals regulates the activities which aim at those goals. Indeed these important epistemological issues can best be addressed by considering examples from other kinds of activities. Suppose I need to buy an aeroplane ticket, and I tell you that my aim is to obtain the cheapest possible seat to my destination. The Rorty argument would suggest that this couldn’t be my goal because, however assiduously I search the Internet for cheap flights, I can never reach a position when I can be absolutely certain that there is no cheaper ticket to be had. When I turn my computer on, I may hope that I shall obtain the cheapest ticket, but all I can aim for is the ticket which I can believe is the cheapest. Or perhaps I could aim to reach a position where no one could blame me for missing any cheaper tickets that are available. But I cannot actually aim to get the cheapest. If I subsequently find that a cheaper ticket was available,

Action and Inquiry


I may well be disappointed, but would I (should I) have a sense of failure? Such examples bring out an interesting kind of complexity in how we think about goals. We might say that I aspire to buy the cheapest ticket. I then adopt a strategy that monitors my researches so that I end up buying the cheapest ticket I can find in a certain period of time searching for information that is available in certain ways—in this case, through the Internet. The strategy is designed as a means of achieving what I aspire to; but there is no guarantee that it will enable me to achieve this. Perhaps, then, the rationality of using this strategy depends upon it being reasonable for me to hope that I shall achieve what I aspire to and, in this case, a readiness to accept second best, a cheap ticket which is not the cheapest. But there are different kinds of cases here. If, after much experience, I believe that the strategy I adopt is likely to lead me to the cheapest ticket, then it is hard to see why the fact that there is no guarantee that it will do so should stop me identifying the cheapest ticket as my goal. I cannot decisively establish whether I have succeeded in obtaining my goal; but I can adopt a means because it has a very good chance of taking me to my goal. There is another case to consider too. Suppose someone has the goal of securing an excellent education for their children. This is a long-term project, and at no stage are they in a position to determine whether they have succeeded. What they can then do is adopt a variety of shorter-term goals whose achievement they believe, or hope, will contribute to the achievement of the long-term aim. They want their children to receive a good education, and they anticipate that the pursuit of all these shortterm aims will contribute to this, but it would be misleading to say that this was the goal of each of this goal. Such examples have an analogue in the epistemological case. A scientific inquirer may have the remote aim of understanding some range of puzzling phenomena. She may not believe, or even hope, that the results she reaches in the short term will turn out to be correct. She expects them to turn out to be false, and thus she does not expect that they will yield knowledge. But she anticipates that defending these views and allowing them to be criticized will be a valuable means towards eventual progress towards a better understanding. Current inquiries are intended to contribute to the achievement of the long-term aim, but they do so in an indirect way. Reflecting on the variety of kinds of goals our actions can have will thus yield an understanding of the variety of epistemic goods that can be sought through our inquiries. If inquiry is understood as a goal-directed activity, then much of the reflection involved in monitoring our inquiries will involve assessing the


Christopher Hookway

suitability of means to the achievement of our ends. Some examples are unproblematic: if you want to find out the time of the next train, you should consult the timetable; if you want to identify the bird you can see on the hilltop, you should look at it through binoculars; if you want to make progress in your inquiries, you should propose theories that are very risky; if you want to propose theories that are very risky, you should avoid ad hoc hypotheses; and so on (see Laudan 1990). These are all methodological norms, and they make proposals about the courses of action that are most likely to secure important epistemic goals. But it may not be clear that a similar pattern is involved in the use of all epistemic norms. In fact, the suggestion that inquiry is a goal-directed activity enables us to illuminate a lot of what goes by the name of reflective monitoring and evaluation in epistemic matters. Such monitoring and reflection always seems to involve engaging in further subordinate inquiries, raising and addressing questions whose answers will contribute to success in the overall inquiries. For example, when beginning an inquiry, we can ask what strategy we should adopt for completing it. We may subsequently find reason to ask whether that strategy was a good one in the circumstances, or reflect upon how it should be executed, or how well it has been executed. Carrying out the inquiry can require us to complete other activities: we can carry out experiments, consult book from our shelves, and make observations of our surrounding. Some of these involve physical actions, and all involve the management of attention and the consideration of how we should understand what we attend to in the particular contexts in which we find ourselves. The management of attention can be employed in other ways too: we examine patterns of belief on the alert for inconsistency; we examine the credentials of our secure beliefs on the lookout for signs of prejudice; exploiting our knowledge of how our opinions might be defeated, we examine the beliefs we hold in order to see whether any of them can defeat our current assurances, even if only provisionally. All of these kinds of reflection involve activities: we ask questions, we make observations, we pay attention to things, and so on. If we inquire well, this is because we have conducted these activities well. And if we are confident that we have inquired well, we shall be (provisionally) satisfied with the outcome of the inquiry. The pattern I have described is common to a range of different sorts of cases. Most importantly here, there is no epistemological distinction between public collaborative inquiries such as are encountered in

Action and Inquiry


laboratories and processes of reasoning that occur in foro interno. When we engage in solitary reasoning—perhaps trying to solve a practical dilemma or an arithmetical problem—it is a contingent matter whether we use public props, such as pen or paper or calculator. Solitary observation can involve just as much manipulation of the environment— in order to see better—and careful exercise of attention as scientific experiment does. We may reserve words like “reasoning” and “deliberation” for describing the solitary exercise of our epistemic capacities, but this is to use these words to activities that fall under a common genus of activity with scientific investigations. The word “inquiry” is used here to capture that common genus. It is in this spirit that Paul Grice criticized what he took to be a common philosophical picture of the nature of reasoning, that it simply consisted in a sequence of inferential steps each conforming to some pattern of (hopefully) valid argument. Instead, he urged that “reasoning is an activity, something with goals and purposes” (Grice 2001, 16) and emphasised the connection of reasoning with the will. I have argued that since inquiry is an activity, epistemic evaluation is concerned with a kind of epistemic rationality. Logical principles may give us information about which patterns of beliefs are consistent and inconsistent, and, given our epistemic goals, this information cannot fail to be an indispensable tool for managing our inquiries. But the logical rules rarely tell us what to believe and which inferences to draw (see Harman 1986). We often have to exercise judgment in deciding how to avoid the logical and probabilistic inconsistencies that threaten our opinions. The thesis of this section is that once we recognize that inquiry is a goal-directed activity, a host of epistemological problems arise concerning the character of these goals. We have identified some differences in the sorts of goals that actions can have, and seen some resulting differences in the ways in which our actions are related to those goals. And we have shown how this can enable us to identify some different kinds of epistemic goals. In the process, we have identified some different ways in which our fallibility can be reflected in the content of our goals and the ways in which we pursue them. The rational assessment of inquiry has a similar structure to the rational assessment of other activities. It is distinctive in two respects. First, in inquiry, pursue epistemic goals rather than other more practical kinds of goals: we want to solve problems about what is the case rather than problems about what to do. And second, there is often a role in the ways we carry out the activity for


Christopher Hookway

information of distinctive kinds—we use information about logical consistency, the force of evidence, about the goodness of explanation, and other distinctively epistemic presumptions. The common structure is manifested in how we use resources in securing our goals; and what is distinctive about epistemology concerns the nature of the specific goals we have and the precise character of the resources we make use of. Of course there are also issues about the rationality of our epistemic goals: how do we assess them. Recall that our epistemic goals are expressed in problems we face or questions we have to answer. The issue about the rationality of ends here concerns how we assess which situations we will find problematic, which questions we should find pressing. Often, a question will become pressing because its answer is pertinent to achieving some practical goal or carrying out a further inquiry. Some questions are important, some are trivial; some are interesting, others are dull. In no case is this an intrinsic feature of the question. The interest of the question will be a function of our other interests and goals, and our other beliefs and expectation. This seems to parallel the ways in which we assess our non-cognitive goals (see Dewey 1938). 3. Scepticism as an Issue in the Theory of Action When we examine the history of epistemology, it is inseparable from the history of philosophy’s engagement with scepticism. In introducing the subject to our students, it is common to begin with sceptical arguments, perhaps those of Descartes’s first Meditation which suggest that our supposed cognitive achievements are put at risk by our inability to respond to these challenges. Unless I can show that I am not dreaming, and I am not the victim of an all powerful deceitful demon, then I am not entitled to any of the everyday and scientific beliefs on which I rely. The sceptical arguments thus suggest that our apparent epistemic achievements cannot be taken at face value. For Descartes, this involved suggesting that we cannot be certain of any of our opinions and that, if this so, we are in no position to make a contribution to the sciences that will last. The Pyrrhonians had urged that whenever we take a belief to be true, we can be forced, by being presented with counter-arguments which are similar to the considerations we would appeal to in support of our opinion, to suspend judgment, content only to say how things seem to us. And one sceptical theme in Hume’s philosophy suggests that when we try to understand the processes by which we form our beliefs, constructing a

Action and Inquiry


naturalistic theory of cognition, the resulting theory makes it difficult to explain that we have formed beliefs in ways that ensure that they are likely to be true. All three sceptical strategies suggest that our confidence that we can achieve our cognitive goals by careful reflective inquiry is illegitimate. There is a kind of activity that we either cannot carry out or cannot carry out while having a legitimate kind of confidence in our ability to complete our tasks successfully. If we are to reject scepticism about our ability to inquire, we must either show what is wrong with the challenges that threaten our confidence that we can do so, or revise our cognitive goals so that the challenges do not engage with our practice. For example, the moderate sceptics of the seventeenth century drew the moral that we should qualify our cognitive goals: rather than trying to master the nature of reality, we should aim for bodies of theory that, so far we could now tell, made sense of how things appeared to us. When we look at how scepticism is explained in contemporary textbooks of epistemology, there is a marked difference. Sceptical arguments are usually supposed to show that we cannot actually know anything at all, or that we are never in a position legitimately to claim knowledge of anything. Alternatively, they may be taken to show that none of our beliefs are justified. The focus on knowledge is probably a twentieth century development in the subject, originating mainly in the work of Roderick Chisholm. We can usually adapt the earlier arguments to this vocabulary—indeed, many students are probably under the impression that Descartes’s concern, in the Meditations, was with whether we possess knowledge of the external world. It is not at all obvious that this way of thinking about scepticism is required by what we find in the earlier sceptical tradition. Let us ask how we should think about the threat of scepticism from the perspective of the evaluation of inquiry. One effect of such challenges might be to challenge the legitimacy of our confidence that we have the ability to carry out inquiries effectively and responsibly. This result could be obtained in two different ways (at least). First, the sceptical challenges may lead us to question whether our cognitive goals can be achieved: if our aim is certainty about some matter and we can be shown that legitimate certainty cannot be obtained, then we have no reason to carry out the inquiry. Secondly, the sceptical challenges may challenge our assurance that we possess particular resources that are required for the effective conduct of our inquiry. Consider, here, Descartes’s analogy of the rotten apples: once error enters our corpus of beliefs, there is no reason to think that it can be identified or eliminated. The rot will just spread through


Christopher Hookway

the corpus. So, perhaps, unless the beliefs we use as premises in our reasoning are certain, there is a risk that they are untrue, and if we rely upon untrue premises, we may be led to further false beliefs whose presence will prevent our ever uncovering the original error. So the sceptical challenges question whether our beliefs possess characteristics that they must have if they are to be used responsibly as resources for further inquiry. If we don’t possess certainties, we cannot inquire effectively. Challenges to our ability to possess certainty have a sceptical force only if a lack of certainty prevents our participating in inquiry effectively. If, like the pragmatists, we believe that our strategies of inquiry are selfcorrecting, that one false belief does not threaten to undermine the rest of the corpus, then challenges to certainty have no sceptical force. For example, Peirce recognizes that induction is fallible, but he also offers reasons for believing that persistence with induction will enable us to identify our errors and correct them. When we form a belief through perception, we stand ready to revise that belief when further experience or reflection shows that it is unlikely and can explain why things should have appeared to us, in context, as the belief represented them as being. Errors that enter our system of beliefs can be ejected through rational inquiry; there is no reason to suppose that, once acquired, the false belief inevitably lingers and poisons all of our other opinions. We can’t be absolutely certain of these beliefs, perhaps, but we can still learn from them. Similar remarks can be made about knowledge. Unless it can be shown that our cognitive aims have to be described used “know” and its cognates, or that information can only be used as (for example) evidence if it is known, then challenges to the possibility of knowledge may not threaten our confidence that we can carry out inquiries effectively. The best way to characterize scepticism, as a threat to our practice of epistemic evaluation, is that it suggests that we cannot carry out inquiries, with a suitable degree of monitoring and rational control, without relying excessively upon luck for our success. We cannot be confident that our errors will be eliminated as time goes by, so long as we continue to inquire well enough. And we cannot take responsibility for how well we do. What counts as success here is, of course, controversial. It may be required that each inquiry issues in knowledge; or it may be required that we can make progress, replacing old theories with better ones that retain the merits of the predecessor while overcoming at least some of their shortcomings. Perhaps we need to be certain that such progress can be made; perhaps all that is required is a fully reasonable hope that it will

Action and Inquiry


occur. Perhaps we require that there be no real reason for doubt that we will fail, that, given our background beliefs, this is not recognized as a real possibility (see Hookway forthcoming) However the details are worked out, scepticism threatens our participation in a range of activities: we lack the resources for pursuing our goals effectively. In an earlier book (Hookway 1990), I argued that the problem of scepticism was analogous to the free will problem. The supposed parallel is this: the challenges suggest that we cannot take responsibility for the success of our inquiries by participating in such activities through exercising our practical reasoning skills conscientiously. The more conscientious we are in carrying out our inquiries, and the better we understand the sceptical challenges, the more conscious we are that our success in achieving our goals requires a large measure of luck. Once again, a satisfactory response to an epistemological issue requires us to examine what sorts of cognitive goals are legitimate and what capacities we require in order to engage in a goal-directed activity in a responsible, self-controlled manner. 4. What Do We Do with Knowledge Leaving aside the sceptical questions mentioned above, two problems about knowledge have dominated recent epistemology. The first, which emerged in the 1960’s, arose from the failure of various attempts to provide reductive analyses of the concept of knowledge. Gettier’s famous counter-examples led to a variety of alternative analyses, designed to avoid the troubling counterexamples. And most of these were in turn challenged by ingenious counterexamples in turn. Many are now sceptical that an adequate reductive analysis of knowledge can be constructed: perhaps it is a simple primitive concept; perhaps there is a family of related conceptions of knowledge that serve different purposes. What cannot be denied is that, however pessimistic we may be that the goal of the analysis will be attained, this series of debates has introduced a set of concepts and considerations which has enriched epistemological debate. From Goldman’s emphasis upon the causation of our beliefs and the reliability of our methods of belief formation, through Dretske’s ideas about conclusive reasons and Nozickian discussions of tracking, to discussions of relevant alternatives and a number of other ideas, features are identified that clearly have an important role in our practice of epistemic evaluation.


Christopher Hookway

In recent years, these debates have spawned what is referred to as the ‘value problem’. This derives from Plato’s demand, in the Meno, that we explain why it is better to have knowledge than mere true belief. All that is required for a belief to be a successful basis for action is that it be true: what further value is there in possessing beliefs which are justified, which track the truth, etc.? The two problems are related. An account of knowledge will be correct only if a satisfactory account can be given of why, on the analysis being offered, knowledge was valuable. And even if we think that the search for a reductive analysis of knowledge should be abandoned, there is still scope for asking why we should value true beliefs that possess the different features that have been identified: why should we value beliefs which track the truth, which are reliably produced, and so on. The debate has enabled us to identify a range of epistemic properties that, in at least some contexts, we value. And it is important to understand why it is correct to value these properties. One promising way of approaching such issues is to ask why it may be important to identify beliefs and believers that have these different properties while carrying out inquiries and investigations. One possibility is that knowledge is the aim of inquiry: somehow, it identifies a fundamental epistemic good such that inquiries are only successful if their outcomes are good in this fashion. It is a necessary condition for a belief to provide the correct answer to the target question of the inquiry that it possess the characteristics that are required for it to count as knowledge. It is hard to see how this could be shown without taking account of what we might want this knowledge for, what we expect to do with the knowledge that we obtain. Perhaps the outcome of one inquiry provides resources to be used in further investigations, in which case the value of knowledge will not be understood without taking account of these further applications of it. Another approach might be to deny that knowledge is itself the aim of inquiry, but to point out that, if we are regulating our inquiries, we must be aware that the inquiry has succeeded and that the correct answer has obtained. We need knowledge because, without it, we will have no reason to be content with our answer, no reason to bring the investigation to an end. In that case, knowledge is not the primary aim of inquiry: all the inquiry is aiming for is the right answer to the question. But one of the subordinate inquiries we have to carry out in order to settle on the right answer concerns whether we know it to be true: I find myself drawn to the answer that p, I ascertain that I am in a position to know the answer, so I can then safely accept the answer that p. Somehow the

Action and Inquiry


possession of knowledge tells me when to stop, even if it isn’t the primary aim of inquiry. However, whether knowledge is required for this role is unclear: it may be enough that I take myself to be justified in believing that the answer I have arrived at is the correct one. A third approach is that we can understand what the value of knowledge is by examining the uses we make of knowledge within inquiry. Several different considerations may be relevant here, each pointing to ways in which the possession of knowledge—or, more often, the belief that knowledge is possessed—can affect the ways in which the activity of inquiry is carried out. One view, defended by Timothy Williamson, is that knowledge is what can be used as evidence: if we don’t know something, we shouldn’t use it as evidence for or against other propositions and theories. Hence we might be interested in whether we possess knowledge of some matter because we want to know whether we can use it as evidence in assessing other matters. A second view is that when we ascribe knowledge of some subject matter to someone, we certify them as reliable sources of information on the matter. If we put the two views together, we might conclude that when we identify someone else as possessing knowledge, then we hold that the information that we acquire from them can itself serve as evidence. Knowers are people to whom we can address our requests for information; and knowledge is material that we can use and trust when assessing the support for propositions and theories. The connections with our epistemic ‘actions’ suggested here become clearer when we take note of an important grammatical features of knowledge sentences. Unlike sentences about beliefs, which typically take a propositional complement, knowledge sentences can also (and in practice most often do) take an indirect question complement. I can be aware that someone knows when the next train leaves, they know why water expands on freezing, that they know how the internal combustion engine works, or where the nearest post office is, and so on. Indeed, any question that can be used to formulate the target of an inquiry can also be expressed as the object of a knowledge sentence (see Hookway 1990, chap. 11; Schaffer forthcoming). Indeed, one might even think of knowledge as a relation between a person and a question. An important feature of these indirect question knowledge claims is that when we assert that someone possess knowledge relative to a question, we do not need to know what her answer to that question actually is. All we need to be sure of is that her answer, whatever it is, is the correct one. In order to master the concept of knowledge, we need to


Christopher Hookway

recognize when we can ascribe such knowledge to someone while we are ignorant about just what the answer is. Thus we conclude that someone knows when the train leaves because we have seen them studying the timetable and thus expect that they can answer our questions about the train times correctly. In asking about the value of knowledge, we need to pay attention to this feature of our talk of knowledge. This reinforces two of the observations we have made about the value of knowledge. First, we use the concept of knowledge when identifying informants. We can only use the concept in this way if we can know that someone possesses knowledge independently of knowing what her answer is to the question at hand. Moreover, the knowledge ascription guides us in how to obtain the information that the knower has to offer: we ask them the question that is represented in the knowledge sentence uttered. We can inquire into this question by performing the action of asking this question to our informant. And we complete this stage of the investigation by taking note of the reply that we receive. The value of knowledge, and the importance of identifying knowers, becomes clear when we recognize a pattern of actions that is internal to our practice of inquiry. Something similar can occur when we are concerned with whether we possess knowledge of the outcome of our own inquiries. It might be expected that our fundamental goal is knowledge; and we ascertain whether we know by identifying whether our answer to the question is true and whether it meets the other conditions for knowledge. But this distorts what often happens. I find that I have reached a particular answer. Reflection on how the inquiry has gone assures me—independently of what my answer is—that I know the answer, that whatever answer I have is the right one. Hence the fact that I possess knowledge provides me with reason for thinking that this answer is the correct one. The aim of the inquiry is simply correctness, and the information that I possess knowledge is needed in order to satisfy myself that the answer is correct. We understand the value of knowledge by noticing the role of knowledge ascriptions in the activities of asking and answering questions that are important for our practice of inquiry. This perspective on the issue may help us with a problem that has used as a tool for addressing the value problem in recent discussions of the topic. The problem originated with Stewart Cohen, but is most often formulated using an example from Ernest Sosa. It begins by accepting the fallibilist point that knowledge can be obtained through inductive reasoning. The evidence we have for an inductive conclusion is always

Action and Inquiry


fallible: we may have false beliefs whose inductive support is just as good as this item of knowledge. Sosa’s example concerns someone who drops a bag of garbage down the garbage chute of her apartment building. Sound inductive reasoning leads her to conclude that the bag has arrived in the garbage room in the basement. We are ready to say she possesses knowledge in spite of the fact that we can imagine circumstances, compatible with all her evidence, under which the belief is false. Sosa’s example illustrates this by supposing that there might have been a nail sticking out of the chute on which the bag could have snagged (Sosa 2000; Greco 2003, 112). This hypothetical and small possibility of the belief being false does not prevent our ascribing knowledge to the woman. If someone wanted to know where the garbage bag was, we could say: Ask S, she knows. And what she knows is: where the garbage bag is. If it turns out that the bag did catch on a nail, then we would have to reconsider the position—she didn’t know where the bag was, after all. But as it is, S’s knowledge can be used in the sorts of questioning activities that we have described. Others can rationally rely upon her testimony; and, in reflecting on her own belief about the location of the bag, she can reinforce her rational confidence by noting that she knows where it is. If she knows where the bag is, and her inclination is to believe that it is in the garbage room, then the bag is indeed in the garbage room. (We should note here that if we are in a situation where the possibility that the bag has caught on the nail is being taken seriously, then we may question whether she really knows whether the bag is in the garbage room. Our confidence that she does possess this knowledge can then be questioned. We certainly could not appeal to her knowledge in order to dismiss the possibility that it was caught on the nail. That possibility has, perhaps, become salient; and this is reflected in how we understand the question “Is the garbage bag in the garbage room?”. In normal cases, the possibility that the bag is caught on the nail is not sufficiently salient to influence how we understand the question with respect to which she is said to possess knowledge.) The problem raised by Cohen and Sosa can be pressed by comparing this case with another. Suppose that the same person buys a ticket in a lottery where the chances of winning are one million to one. On excellent inductive grounds, she forms the true belief that her ticket will not win. As in the other case, our subject has an inductively grounded belief with a very high probability of truth. Moreover, in each case she would be rational to use her belief as a guide to conduct. So far, they seem to be on a par, epistemologically speaking. The problem arises because many


Christopher Hookway

theorists find it natural to say that while she has knowledge of the whereabouts of the garbage, she does not know that her ticket will not win. The intuitions that underlie these judgments are controversial: there are those who argue that she has knowledge in both cases, and others who deny that she has it in either. But since there seems to be a difference between the two cases, it is interesting to see how far the views defended here help us to explain why we might be reluctant to ascribe knowledge in the lottery case. First, let us consider using the subject as a source of testimony about whether this ticket will win. It is clear that she does not know which ticket will win. Nor is her knowledge even relevant to that question. Suppose she is asked “Which ticket will win?” and answers “Well I know that it won’t be mine”. If this is taken as ruling out that possibility, then someone might infer that it increases the chance of each of the other tickets winning. And of course, the information she possesses does not do that. There are things she does know: she knows what chance of winning attaches to her ticket, for example, and she knows that it is very unlikely to win. But learning that the chance of her ticket winning is so low does not increase one’s estimate of the chance of some other ticket winning. And this is because its chance of being the winning ticket is no lower than that of any other ticket. The claim to knowledge is usually salient in a context where an inquiry, or reflection upon he progress of an inquiry, makes a particular question salient: Where is the garbage bag? Which ticket will win? In claiming knowledge, one offers something that is presumed to be relevant to that question. It may be an answer to the question. It may suggest an instruction about how to answer the question: “But Jones knows where the garbage bag is.” It may just be part of an answer: “I know that the garbage bag is not still in the apartment.” This narrows down the range of possibility. So both in conversation and in reflection, we can expect that the knowledge claim made is the one that is most relevant to the salient question. The claim that the garbage bag is in the garbage room is evidently relevant to the question “Where is the garbage bag?”. The claim that S’s ticket won’t win is most relevant to the question “Which ticket will win?” if it reports some special knowledge of that ticket. If all it rests upon is the general observation that, like all the tickets, it has a very small chance of winning, then it is not relevant to the question. In that case, if S claims to know that her ticket won’t win in a context where the salient question is “Which ticket will win?” her utterance carries the Gricean implicature that her knowledge claim is relevant to that question, and

Action and Inquiry


hence carries the Gricean implicature that it rests on some knowledge that attaches specifically to that ticket (see Grice 1989, 24 ff.) Similar remarks could be made about the use of knowledge claims in identifying correct answers to the questions that prompt inquiries. However I shall not pursue that here. What I hope to have shown is that if we take seriously the fact that epistemology is concerned with the activity of inquiry, and if we recognize that inquiry is structured by acts of asking, and trying to answer, questions, we can understand some of the uses we make of knowledge claims. Moreover once we attend to these, we can see some differences between the lottery example and the garbage bag example. These explain why, in the cases described, we attach different epistemic values to the two items of information. However, for at least two reasons the conclusions that can be drawn should be qualified. First, even in the case as described, what I have said is compatible with two different views of the lottery case. One possibility is that the claim to knowledge that the lottery ticket will lose is false. The other possibility is that the claim to knowledge is true but misleading—it carries misleading implicatures. In this paper, I take no stand on which of these is correct. Second, in examining these phenomena, I made some assumptions about what the salient questions were: I took them to be “Where is the garbage bag?” and “Which ticket will win?”. These were natural choices, but we cannot rule out the possibility that scenarios could be constructed in which different questions were salient and in which our evaluation of S’s information might be different. 5. Conclusion I have suggested that the problems of epistemology are best understood as arising out of our participation in some distinctive activities, which I have called inquiries. Like other activities, inquiries are goal-directed: they are usually attempts to solve problems or answer questions. Again, like other activities, our ability to carry them out successfully depends upon our being able to employ effective means of achieving these goals. We often employ reflective self-control in carrying out these activities and their success depends upon our having the ability to do this. And in carrying out these activities, we need to be able to recognize and make use of resources of different kinds. Such resources may include things like instruments, things whose deployment can contribute to our success in performing the activities. So far, inquiry does not differ from many other


Christopher Hookway

sorts of goal-directed activities that can be carried out reflectively and responsibly. The paper has identified a number of areas where familiar epistemological problems can be understood as special cases of problems that can arise for other kinds of activities. If that is correct, we may anticipate that a better understanding of the issues about actions and activities will suggest fruitful ways of thinking about the epistemological issues. In section 2, for example, we considered some issues about the goals of inquiry: do we seek true answers to our questions? Or knowledge of those answers? Or objectively justified belief ? Or beliefs that we will take to be justified? We observed that arguments used to criticize particular positions here can exploit some general theses about the circumstances under which we can adopt something as a goal. This led us to note some differences in the kinds of goals that can be adopted and to raise questions about how these different kinds of goals can be manifested in the case of inquiry. In section 3 we argued that the epistemological challenges of scepticism are best understood as a special case of challenges to our ability to carry out activities legitimately confident of our ability to do so in a responsible and effective manner. And in section 4 we suggested that we can approach issues about the value of knowledge by viewing it as a resource or instrument that we can make use of in carrying out these distinctive activities. Bibliography Brandom, Robert (ed.) (2000): Rorty and his Critics, Oxford: Blackwell. Cohen, Stewart (1988): “How to Be a Fallibilist”, in: Philosophical Perspectives 2, 91 – 123. Davidson, Donald (2000): “Truth Rehabilitated”, in: Brandom 2000, 65 – 73. Dewey, John (1938): Logic: The Theory of Inquiry, New York: Holt, Rinehart and Winston. Greco, John (2003): “Knowledge as Credit for True Belief ”, in: DePaul, Michael/Zagzebski, Linda (eds.): Intellectual Virtue: Perspectives from Ethics and Epistemology, Oxford: Oxford University Press, 111 – 134. Greco, John (ed.) (forthcoming): Oxford Companion to Scepticism, Oxford: Oxford University Press. Grice, Paul (1989): Studies in the Way of Words, Cambridge MA: Harvard University Press. Grice, Paul (2001): Aspects of Reason, Oxford: Clarendon Press. Harman, Gilbert (1986): Change of View, Cambridge MA: MIT Press. Hookway, Christopher (1990): Scepticism, London: Routledge.

Action and Inquiry


— (2006) “Reasons for Belief, Reasoning, Virtues”, in: Philosophical Studies 130, 47 – 70. —(forthcoming): “Peirce and Scepticism”, in: Greco (forthcoming). Laudan, Larry (1990): “Normative Naturalism”, in: Philosophy of Science 57, 44 – 59. Rorty, Richard (2000): “Universality and Truth”, in: Brandom 2000, 1 – 30. Schaffer, Jonathan (forthcoming): “Knowing the Answer”, in: Philosophy and Phenomenological Research 73. Sosa, Ernest (2000): “Skepticism and Contextualism”, in: Philosophical Issues 10, 1 – 18. Williamson, Timothy (2000): Knowledge and its Limits, Oxford: Oxford University Press. Zagzebski, Linda (2003): “The Search for the Source of Epistemic Good”, in: Metaphilosophy 34, 13 – 28.

Knowledgeable Inquiry Stephen Hetherington 1. Introduction: Hookway’s Proposal In “Action and Inquiry”, Christopher Hookway proposes a pragmatist reinterpretation of epistemology’s primary focus. Inquiry is the epistemic phenomenon to which he directs our attention. Epistemology already attends to a reasonably canonical list of epistemic phenomena; is Hookway proposing just that we add one more item—inquiry—to that list? No. His idea is that inquiry is a central (indeed conceptually prior) epistemic phenomenon, in terms of which epistemologists should proceed to understand some other epistemic phenomena, including ones that have traditionally received independent discussion. For example, rather than asking simply, “What is knowledge, and why is it valuable?”, we should (if Hookway is right) reflect upon knowledge’s nature and value insofar as knowledge plays roles within inquiry. We should not think of knowing merely as a state, entire unto itself, which we attain in only some circumstances, and which we may choose to seek—or, alternatively, simply be grateful for possessing—for its independently describable value. Hookway’s proposal is pragmatist about knowledge because he emphasizes knowledge as being, first and foremost, a contributor to, and product of, something we do; and that ‘something’ is inquiry. Knowledge’s nature and value are relational— constituted somehow in terms that relate them to aspects of inquiry. Knowledge is whatever it is, with whatever value it has, because of its roles within and arising from inquiry. And inquiry is our primary manifestation as epistemic functionaries. In that sense, we are epistemic agents before we are ever epistemic subjects. We use knowledge in inquiry; we gain knowledge through inquiry; and these are our primary relationships to our knowledge. Undoubtedly, we inquire; and knowledge has whatever nature and value it has, because of this fact about us as agents. For short: Knowledge is as knowledge does, all in the service, and as a result, of inquiry.

Knowledgeable Inquiry


2. A Question About Hookway’s Proposal But might some instances of knowledge in fact ‘do’ nothing, serving the interests of no actual inquiry, not being used within any real inquiry, not even resulting from genuine inquiry? Maybe such knowledge never functions as evidence upon which subsequent inquiry is based. Perhaps it comes into existence without actual antecedent inquiry aiding its passage into the world. If this is possible, not all knowledge need play some role within real inquiry. What should Hookway say about such knowledge? He could seek to retain the conceptual link between knowledge and inquiry by broadening, as follows, the range of modalities being countenanced: Even a piece of knowledge which happens never to play a role within, or to result from, actual inquiry could be used within, or result from,1 real inquiry. Even this knowledge has the nature and value it does, due to its actual or possible roles or genesis within inquiry. Knowledge is whatever it is, because of how it helps us to inquire and because of how it can arise through inquiry. For short: Knowledge is as knowledge does or could do, in the service, or as a result, of inquiry.

Yet such an approach might be thought to weaken the pragmatist spirit of Hookway’s sort of proposal; for the suggested approach is compatible with no knowledge having its nature or value understood in terms of actual contributions to inquiry. The previous paragraph’s suggestion could be satisfied simply by adding to some traditional analytic conception of knowledge the fact that, for each instance K of knowledge, there is a possible world W—accessible yet distinct from this world— where K either plays an apt part in, or arises suitably from, some process of inquiry actually occurring within W. However, this would remain compatible with no such K’s being used or gained in that way within this world. Hence, there need not be anything clearly pragmatist in how such an analysis would be characterizing the knowledge actually present in this world. Inquiry would not be directly constituting K qua knowledge within this world; inquiry would be bearing only upon a modal aspect of K qua knowledge within this world. At most, this world’s knowledge would 1

That is (in the latter case), we could have reasoned our way to an acceptance of it, even if in fact we did not. This is part of why philosophers still regard the availability of a rational reconstruction—a description of how a person could have justified a particular belief—as pertinent, perhaps necessary, to an assessment of the particular belief ’s epistemic status, even for that person who did not engage in the described justificatory process.


Stephen Hetherington on Christopher Hookway

possibly be related to inquiry; it would never actually be constituted by inquiry. Some such danger of abstractness could persist, too, even if—in an actualist spirit—we try ‘locating’ the modality within the inquiry as such. If we wish not to talk of possible worlds, we might talk of possible inquiries instead. Even a piece of knowledge which never contributes to actual inquiries, and which has not arisen through inquiry, may have value as knowledge due to roles within possible inquiries, and due to the fact that there are possible inquiries of which it would be a conclusion. What is a possible inquiry, though? Is it a merely abstract object, such as an ordered set of propositions, perhaps with indices indicating times and (possible) users? That way, mere formalism beckons; again, pragmatism wanes. 3. An Ability Analysis of Knowing How may a pragmatist evade those potential worries, those critical lines of thought (developed late in §2)? How is he to retain—as constituted by contributions to inquiry—the actual nature and value of knowledge, while allowing that a piece of knowledge can exist, with value, even apart from either contributing to, or arising from, any actual inquiry? Here is one idea. On a pragmatist’s behalf, we might seek to understand the broadened modality (mentioned in §2) in terms of a potential within the knowledge itself, rather than in terms that reflect only a more extrinsic sense of possibility. For example, we might look for an aspect of a person herself which would constitute some particular knowledge’s being as knowledge does or could do, within or arising from inquiry. First, we would want to say that knowing is, within itself, whatever it does or could do in contributing to, or arising from, inquiry. Second, we would want to say that all of this is an actual part of the knower when the knowledge is present. When combined, these two points would enable us to close any otherwise looming conceptual gap between (i) the claim that something is knowledge and (ii) the pragmatist idea of that particular something’s being knowledge by being constitutively linked to inquiry. Closing that potential gap is the pragmatist’s goal, at any rate. How may we begin to capture this approach? Before I try doing so, a motivational word is needed, regarding my use just now of the phrase, “conceptual gap”. Hookway denies that knowledge need be the aim of inquiry or that it is generally required for

Knowledgeable Inquiry


assessing the success of an inquiry. Rather (he argues), knowledge’s value is a function of something internal to the practice of inquiry. That is a suggestive thought. However (apart from offering this thought), Hookway’s discussion leaves open what knowledge is. We are told only that knowledge’s nature and its value should intertwine, conceptually and explanatorily. Roughly speaking: if we understand the one, we will have understood at least much of the other. That is the aim of an approach such as Hookway’s. Again, though, can an avowed pragmatist really effect that desired conceptual linking of knowledge’s value with its nature? There is a danger of his not doing so, insofar as (i) pragmatists offer us an explication only in terms of inquiry, say, and insofar as (ii) they do not make this explication an identity—so that the nature of any piece of knowledge is somehow exhausted by its constitutive relationships to inquiry. In an admittedly programmatic and experimental spirit, then, I will propose an hypothesis that does postulate some such identity. On Hookway’s behalf, let us consider something like the following, which I will call an ability (proto-)analysis of knowing:2 One’s knowing, at time t, that p is, at t, either (i) one’s possessing a suitably constrained potential for engaging, or for having engaged, accurately in prelated inquiry, or (ii) one’s actually engaging suitably in such an inquiry.

Modality would thereby be present in the analysis of knowledge, courtesy of clause (i). But we may readily understand (i) in a pragmatist way, so that the potential of which it speaks would be constituted by nothing beyond actual aspects of the person. This will accord with how pragmatists should think of this feature of the person’s knowing. The proposed account will allow knowing that p to be either (i) an apt ability to contribute, or to have contributed, to inquiry or (ii) a manifestation of that ability. And, of course, we may then offer various possible instantiations of that generic account, as we consider different specific abilities with which one might contribute to inquiry. For instance, knowing that p may be the ability to respond accurately to questioning about p, or to ask questions about p, or to pose such questions—or any combination of those aspects of how one may inquire regarding p. Or one’s knowing that p would be one’s actually doing any of that, manifesting such an ability. In any case, the general point remains that insofar as an ability is a real-but-modal 2

For a fuller (albeit slightly different) such account, see Hetherington 2006.


Stephen Hetherington on Christopher Hookway

feature of a person, able to be present even if never manifested, knowing that p could be present even if the associated ability is never manifested; yet the ‘knowledge ability’ might involve nothing beyond features of the person’s ability to inquire or to have done so.3 Still, how might we conceive of the relevant constitutive modality within knowing, without looking beyond actual aspects of the person? Some of Nicholas Rescher’s (1979) remarks on possibility-as-conceptual seem apposite. Confronted by the challenge of understanding possibility in terms of actual acts of mental conceiving (rather than, circularly, in terms of possible actual acts of mental conceiving), Rescher advocates our looking to general categorial features of mental conceiving: What we are saying is that the ‘reality’ of certain possible states of affairs and things (that is, nonexistent possibilities) resides in the reality of possibilityinvolving processes [the construction of verbal descriptions, and the hypothesizing (assuming, postulating) of their existence]. We are saying that, when thepossibility-of-the-thing is its only ‘reality,’ this ‘reality’ inheres in a possibilistic intellectual process. Here actuality is indeed prior to possibility—the actuality of one category of things, namely, minds with their characteristic modes of functioning, underwrites the construction of the totality of nonexistent possibles that can be contemplated. (Rescher 1979, 174)


This proposed analysis enables a pragmatist about knowledge to avoid having to accept that the presence of some knowledge that p can be understood as an isolable state of a person at a time, entire unto itself at that time—able to be present independently of all other beliefs or hypotheses, say, that the person may have or entertain. On the contrary; according to this section’s proposal, the presence of some knowledge that p cannot be understood independently of understanding the actual or potential presence of a network of p-related inquiries. An inquiry would lead to the particular knowledge that p, perchance, without thereby leading to something that is nonetheless constituted as knowledge that p independently of all inquiry. This particular instance of knowledge that p would be knowledge only by having arisen from appropriate inquiry. (Note Hookway’s insightful observation that we often accept a particular belief as knowledge simply because we accept that appropriate inquiry has occurred.) Of course, as Hookway also observes, inquiry involves knowledge: often, our preferred premises within an inquiry are taken to be known, somehow independently of their role within the specific inquiry. Must we therefore understand this status of theirs as being present independently of all inquiries? Not at all: these instances of knowledge, even if they are constituted as knowledge independently of the particular inquiry within which they are serving as premises, are constituted as knowledge via other actual or potential inquiry. At any rate, that is what this section’s proposal would imply. The proposal is thus pragmatist and holist.

Knowledgeable Inquiry


Can something similar help us to understand, in pragmatist terms, knowledge’s modal link to inquiry? We might think of the ability to use an instance of knowledge that p within inquiry as being some relevantly sophisticated function of actual inquiries, by that same person, involving other pieces of knowledge. Presumably, these would be aptly similar to the knowledge that p. This picture allows you to know that p, in spite of never embedding this knowledge within actual inquiries, by having an ability to embed it within inquiries or to have done so; and this ability is constituted by your record, in the past and/or future, of embedding similar instances of knowledge within actual inquiries.4 That picture is programmatic, but not increasedly so. It merely shifts our focus, from modality to similarity. The challenge of understanding the modality in your ability to use, or to have used, the knowledge that p within inquiry becomes the challenge of understanding whatever network of actual inquiries by you has involved, and/or will involve, sufficiently similar instances of knowledge. One philosophical challenge gives way to another. At least the new one is recognizably pragmatist.5 A further detail: we should notice that although this section’s ability analysis of knowing does not rule out knowledge’s being a state of a person, it does not mandate that interpretation. Maybe a person could satisfy the proposal purely by how she acts, regardless of whether she is in any settled state of conscious confidence, say. Traditional analytic epistemology would have us locate a belief that p within the person; and this belief would be deemed to be the ‘stuff ’ or ‘matter’ of the knowledge that p (other things being equal). But this section’s ability analysis treats the belief that p as epistemically epiphenomenal. A belief that p has an epistemic character, if at all, only in terms of what it does or could do within inquiry, perhaps as part of what the knowledge that p does or could do within inquiry. My key suggestion, still following Hookway’s lead, is that one’s knowing that p is just some ability (manifested or not) within one, regardless of how it is present and regardless of what else is present. A belief that p could—but need not—be part of either this ability or the ability’s manifestation. 4 5

And if you never use any aptly similar instances of knowledge in inquiry, pragmatists might well regard this as good evidence of your not knowing that p after all. This analysis can accommodate Hookway’s suggestion that inquiry may be pursued either collectively or individually. The ‘knowledge ability’ could be possessed either collectively or individually. On an instance of a collective, but not individual, ability to inquire accurately, see Hardwig 1985.


Stephen Hetherington on Christopher Hookway

4. Two Applications of the Ability Analysis Hookway describes how to recast some sceptical challenges. Gesturally, I will add another challenge to that list. Consider a traditional epistemic regress argument, resulting in sceptiACHTUNGRE cism about there ever possibly existing any inferential knowledge. The heart of that argument takes this form: Having even a single piece of inferential knowledge, k1, requires any supporting evidence e1 to be knowledge itself: call this new knowledge k2. But if k2 is inferential, it rests upon evidence e2 which also needs to be knowledge: call this new knowledge k3. And so on, ad infinitum—unless some knowledge is non-inferential, say.

Suppose we adopt something like §3’s ability analysis, thinking of knowing that p as one’s either contributing, or being able to contribute, to p-related inquiry. This could be either a manifestation, or an ability to manifest, a p-related skill. It would be either a manifestation of, or the ability to manifest, some p-related knowledge-how. In partially Rylean terms (see Ryle 1949, chap. 2; Ryle 1971), it would be either an intelligent p-related action, or it would be the accompanying knowledgehow. Famously, Ryle asked whether intelligent actions owed their character as intelligent to their being guided by prior knowledge-that. He was wondering whether, in practice, knowledge-how amounted to nothing more than knowledge-that. The affirmative answer to that query is an intellectualism about the nature of intelligent action. But Ryle’s answer rejected that intellectualist thesis. Accepting an intellectualist interpretation of intelligent action, he argued, commits one to a vicious infinite regress. Accordingly, he concluded, one could never combine such an interpretation with an understanding of what makes an action intelligent. How does that Rylean result (as I take it to be)6 combine with the ability analysis of knowledge? It assists in deflecting the traditional epistemic regress argument, as follows. Consider any manifested instance of knowing that p, such as one’s contributing to a p-related inquiry. This is an intelligent action: one is exercising knowledge-how. But it need not have that epistemic nature by being derivable from prior and guiding evidence. First, it could rest upon prior, similar, actions (rather than evidence). These may end at some time 6

In Hetherington 2006, I defend Ryle against Jason Stanley’s and Timothy Williamson’s criticisms (see Stanley/Williamson 2001).

Knowledgeable Inquiry


without the sequence thereby not comprising intelligent actions. Second, the action’s knowledgeable contribution could (even if it need not) be its helping one to construct subsequent steps in the inquiry. The Rylean denial of intellectualism coheres well with these possibilities, once we conceive of manifested knowing as an exercise of intelligent activity. Yet—to adopt a rhetorical pose for a moment—how else should we conceive of acts of knowing? Even so, might that approach license a worrying dogmatism, with a correlative absence of epistemic attainment? In order to answer this, I will revisit Wittgenstein’s (1958) discussion of rule-following. We may interpret his discussion as itself including an argument against intellectualism about intelligently performed actions. His argument differs from Ryle’s in style; and it contains more besides. Nevertheless, part of it is plausibly viewed as a rejection, intriguingly akin in spirit to Ryle’s, of that sort of intellectualism. Consider Wittgenstein’s key examples. He asks (beginning at §151; resuming at §179) what is needed for a person to know how to continue a numerical series. And (starting at §156) he ponders how a person knows how to read, albeit in a restricted sense of “read”. In those cases, Wittgenstein’s main question is that of whether the person applies a rule that is already present to her consciousness or understanding. On the picture being probed, the rule would be meant to supply a direction or interpretation, providing explicitly articulated instructions for the person as to how next to act. In other words, it is meant to be some relevantly influential knowledge-that,7 already possessed by the person; and an associated question is that of whether, in acting intelligently (such as in the situations examined by Wittgenstein), the person must be using some such knowledge-that. Is knowledge-how, at its heart, a kind of knowledge-that? Does any exercise of knowledge-how involve an application of knowledge-that? Although Ryle posed those questions more crisply, they were pivotal to Wittgenstein’s multi-faceted discussion of rule-following. 7

It is at least partly knowledge-that; it might also include something imperatival (perhaps linking the knowledge-that to actions in potential circumstances). For this blend, see Wittgenstein’s §86. (There, he offers a precursor of his ‘official’ rule-following ruminations. And his remarks on “further rules to explain this one” presage the Rylean regress argument about prior intellectualist uses of knowledge-that.) His explication in that section of the use of a rule involves a formulated ‘table’, one which would be known by the intellect, as part of then contemplating how to apply this knowledge in acting.


Stephen Hetherington on Christopher Hookway

His answer to those questions, too, was similar to Ryle’s. Reflecting on various possibilities (especially experientially or subjectively described ones), Wittgenstein tried to ascertain what it would be like to proceed either definitively on the basis of prior knowledge-that—or not definitively at all. And he could find no such clear “what it would be like”.8 For instance (§§187 – 8), does one know in advance all the correct answers? No. Are we consciously guided (§170) by the appropriate rule (the grasp of which I have taken to involve some knowledge-that)? Not obviously by this-one-rather-than-another-one. Moreover, even when a particular rule is one by which you think you are being guided, you need to interpret it and its influence (§§86, 190). You need to know how to apply it; and so (on the intellectualist picture being tested) you will depend upon another rule’s prior presence and intervention. On and on this will need to continue—if, of course, it ever had to begin. So, should we infer that it never did need to begin? In a seemingly anti-intellectualist spirit, Wittgenstein tells us (§201) that “[t]here is a way of grasping a rule which is not an interpretation, but which is exhibited in what we call ‘obeying the rule’ and ‘going against it’ in actual cases.” It is not true, then, that all intelligently performed actions are controlled by knowledge-that. At least some intelligent actions need not be generated, even partly, via a knowing grasp of a rule. Instead (in a Humean vein), Wittgenstein says that, at least sometimes, acting intelligently is a custom (§199) or a practice (§202). In short: How can he know how he is to continue a pattern by himself—whatever instruction you give him?—Well, how do I know?—If that means ‘Have I reasons?’ the answer is: my reasons will soon give out. And then I shall act, without reasons. (§211; my emphasis of “how”)

Again § 219: “When I obey a rule, I do not choose. I obey the rule blindly.” Here, Wittgenstein is not necessarily talking of every occasion of following a rule. His focus (as shown in §217) is on occasions “when I have exhausted the justifications [and thereby] I have reached bedrock … . Then I am inclined to say: ‘This is simply what I do’.” And on these occasions an intellectualist analysis, even of intelligent action, is mistaken: an action can be manifesting knowledge-how without having required determinative or constitutive guidance from prior knowledgethat. As Ryle would likewise say, we need not expect knowledge-how only 8

Perhaps this is because, for Wittgenstein, there is no distinct phenomenon of knowing-that. Might all knowledge be knowledge-how? Below, I return briefly to this pragmatist idea. For more on it, see Hetherington 2006.

Knowledgeable Inquiry


ever to be present and manifested in accord with the dictates of some prior, guiding, knowledge-that. There is at least that much of a metaphysical gap between knowledge-how and knowledge-that. Are they therefore categorially distinct kinds of knowledge? 5. Conclusion That question returns us to this paper’s ability (proto-)analysis of knowledge. If my suggestion is categorially correct, it shows us how to take a first step towards unifying the phenomena of knowledge-that and knowledge-how. The unification would be unorthodox, explaining knowledge-that in terms of knowledge-how (rather than, as epistemologists more typically expect to be feasible, trying to reduce knowledgehow to knowledge-that). This would be a potentially significant step towards enriching the pragmatist program to which Hookway draws our attention; for the knowledge-how in question would be constituted by our pertinent skills as inquirers.9 Bibliography Hardwig, Jon (1985): “Epistemic Dependence”, in: Journal of Philosophy 82, 335 – 349. Hetherington, Stephen (2006): “How to Know (That Knowledge-That is Knowledge-How)”, in: Hetherington, Stephen (ed.): Epistemology Futures, Oxford: Clarendon Press, 71 – 94. Hookway, Christopher (2007): “Action and Inquiry” (this volume). Rescher, Nicholas (1979): “The Ontology of the Possible”, in: Loux, Michael J. (ed.): The Possible and the Actual, Ithaca, NY: Cornell University Press, 166 – 181. Ryle, Gilbert (1949): The Concept of Mind, London: Hutchinson. — (1971): “Knowing How and Knowing That”, in: Ryle, Gilbert (1971): Collected Papers, Vol. II, London: Hutchinson, 212 – 225. Stanley, Jason/Williamson, Timothy (2001): “Knowing How”, in: Journal of Philosophy 98, 411 – 44. Wittgenstein, Ludwig (1958): Philosophical Investigations, Oxford: Blackwell.


I am grateful to Paul Snowdon for his comments on a draft of §4’s discussion of Wittgenstein, and to Anton Leist for his remarks on the whole paper.

Name Index Alvarez, Maria 42, 103, 106, 109, 124–131, 133f., 181, 187 Anscombe, Elizabeth 2, 6f., 20, 103, 118, 141, 158, 181, 189, 199–208, 210, 219, 225, 323 Anwander, Norbert 20, 44, 181, 184, 189 Aquinas, Thomas 267f. Arendt, Hannah 267 Aristotle 1f., 43, 46, 103, 113f., 145, 174, 262, 264–271, 273 Armstrong, David 257 Audi, Robert 41, 81–83, 150–152, 181, 186, 188f. Augustine 267 Ayer, Alfred J. 249 Bacchylides 267 Baker, Lynn R. 247 Baumann, Holger 44, 46, 48, 101, 275, 341 Berkeley, George 60 Bermffldez, Jos L. 256 Birnbacher, Dieter 262 Bishop, John 31f., 43, 150, 154, 156, 161, 327 Bittner, Rdiger 4, 25–28, 41f., 44–48, 95, 98f., 104, 106, 181, 187, 261, 270, 273–283, 330f. Block, Ned 243, 257 Brand, Myles 2, 150 Brandom, Robert 15, 38 Bratman, Michael 4, 9, 13–17, 20f., 23, 29–32, 35f., 38, 44f., 194–197, 199, 204–209, 212, 216–219, 222f., 225f., 228f. Brett, Nathan 189 Broome, John 224, 268 Buss, Sarah 39 Campbell, John 249 Cassam, Quassim 249

Chisholm, Roderick 7, 231, 361 Christman, John 250 Cleanthes 267 Coburn, Robert C. 236 Cockburn, David 242 Cohen, Stewart 366f. Collins, Arthur 106, 345 Crane, Tim 112 Csikszentmihaly, Mihaly 213f., 227f. Cuypers, Stefaan 44f., 231f., 234, 237, 240, 251, 254–258, 260 Dancy, Jonathan 18f., 22, 24–28, 41f., 46, 95, 98, 100f., 106f., 121, 138, 174, 181, 184, 186–188, 190, 288, 291f., 294, 299, 310 Darwall, Stephen 24 Davidson, Donald 1–7, 9, 11–16, 18, 20–22, 30, 32, 34, 37, 40–42, 45, 83–87, 89–91, 103, 105f., 113, 117, 119f., 124f., 127, 130f., 154–156, 181, 189, 219, 247f., 285, 326f., 354, 356 Davis, Lawrence H. 150 Dennett, Daniel 209 Descartes 138, 360f. Dewey, John 4, 43, 93f., 163f., 168–172, 176, 178, 360 Dirlmeier, Franz 265f. Donagan, Alan 6 Dretske, Fred 6, 363 Dreyfus, Hubert I. 165 Dreyfus, Stuart E. 165 Enc, Berent 157 Enoch, David 38 Esser, Hartmut 4 Euclid 60 Euripides 266 Evnine, Simon 105


Name Index

Falvey, Kevin 199 Feinberg, Joel 8 Fesmire, Steve 178 Fischer, John M. 152 Fodor, Jerry 41, 81f., 209 Frankfurt, Harry 2, 9, 29, 31, 33–35, 39, 151f., 327 Gallagher, Shaun 259 Garrard, Eve 98 Gettier, Edmund L. 312, 363 Gler, Kathrin 86 Goldman, Alvin 219, 259, 363 Gordon, Robert M. 256 Gould, Stephen 195 Greco, John 367 Grice, Paul 359, 369 Guttenplan, Samuel 189 Hacker, Peter M. S. 40, 53, 55, 106 Hampshire, Stuart 40, 55–58, 61–72 Hardwig, John 377 Hare, Richard M. 2 Harman, Gilbert 203, 359 Harr , Romano 140 Hegel, Georg W. F. 15 Heidegger, Martin 267, 271 Heil, John 104 Hetherington, Stephen 46–48, 372, 375, 378, 380 Hobbes, Thomas 2, 267 Hookway, Christopher 46–48, 351, 363, 365, 372–378, 381 Hornsby, Jennifer 6, 9, 18, 30f., 33f., 46f., 81, 105, 117, 143, 181, 187, 285, 301, 303–305, 307–313, 331 Huebl, Philipp 124 Humberstone, I. Lloyd 323, 340 Hume, David 2, 56, 62, 65, 110, 141, 238, 322, 325, 360 Hursthouse, Rosalind 20, 199 Hutto, Dan 104 Hyman, John 60, 289 Jackson, Susan E. 227f. James, William 163, 212, 340

Jonsen, Albert R. 177 Kant, Immanuel 15, 19, 44, 56, 90f., 193f., 213, 229, 321f., 325, 333 Keil, Geert 125, 133 Kenny, Anthony 67, 222 Korsgaard, Christine 19 Laudan, Larry 358 Leist, Anton 1, 46f., 283, 315, 344–350, 381 Lekan, Todd 43, 163f., 181f., 184–190 Lewis, David 146, 150, 157 Lewontin, Richard C. 195 Lindenberg, Siegwart 4 Machamer, Peter 140, 159f. MacIntyre, Alasdair 183, 250 Mackie, John L. 231, 325 Madden, Edward H. 140 Madell, Geoffrey 249 Malpas, Jeff E. 3 McCann, Hugh 31, 33 McDowell, John 174, 249 McNaughton, David 98 Melden, Abraham I. 7, 117, 141, 158 Mele, Alfred R. 9, 104, 110, 130, 181, 187, 189 Mellor, David H. 112 Merleau-Ponty, Marurice 259 Mill, John S. 6 Moore, George E. 19, 24, 207, 225 Morton, Adam 150, 156 Moser, Paul K. 187 Mounier, Emmanuel 244 Moya, Carlos J. 327 Mumford, Stephen 140 Nagel, Thomas 231 Nehamas, Alexander 272f. Nietzsche, Friedrich 267, 272f. Noonan, Harold W. 231 O’Connor, Timothy 224 Oksenberg Rorty, Amelie 231

Name Index

Olson, Eric T. 146 O’Shaugnessy 6 Parfit, Derek 45, 126, 231, 233, 240, 255f. Peacocke, Christopher 31, 150, 157 Peirce, Charles S. 163, 176, 362 Perry, John 231, 233 Pindar 267 Plato 2, 263, 265, 267–269, 364 Pollard, Bill 189f. Prichard, Harold A. 6, 10, 35 Putnam, Hilary 140 Quinn, Warren 327 Quinton, Anthony 231 Ravizza, Mark 152 Rawls, John 91 Raz, Joseph 20, 72, 80 Reid, Thomas 19, 233 Rescher, Nicolas 376 Ricoeur, Paul 250 Rizzolatti, Giacamo 259 Rodin, Auguste 246f. Rorty, Richard 48, 334, 339, 354, 356 Rosen, Gideon 203 Ross, W. David 325 Roughley, Neil 44f., 216, 222 Rundle, Bede 67 Russell, Bertrand 237 Ryle, Gilbert 2, 6, 46, 140f., 158, 288–290, 293, 297f., 309, 378–380 Santiago, John 38 Saporiti, Katia 46f., 303, 341 Scanlon, Thomas M. 24 Schaffer, Jonathan 365 Schechtman, Marya 250 Scheler, Max 244 Schmidt, Thomas 41f., 95 Schueler, George F. 104, 106, 120, 323, 326, 330f. Schulte, Joachim 46–48, 344 Searle, John 9, 23, 129, 187, 323, 331f.


Seebaß, Gottfried 6 Sehon, Scott 104 Shoemaker, Sidney 231, 238 Simon, Herbert 16 Slors, Marc 44f., 254, 256 Smith, Michael 2, 6, 11, 18, 32, 37, 41f., 46, 79, 97, 126, 288, 292–295, 299, 311, 323 Snowdon, Paul F. 246, 381 Socrates 263f. Sosa, Ernest 366f. Spinoza 71, 272 Stanley, Jason 378 Steward, Helen 106 Stoecker, Ralph 9, 262 Stout, Rowland 9, 43, 46, 106, 137f., 148, 154–159, 161, 288, 294–296, 298, 311f., 327, 331 Stoutland, Frederick 4, 6f., 9, 13, 18, 22, 25, 28, 30, 38, 41f., 48, 75, 82, 95–101, 104, 106, 138, 181, 187 Strawson, Peter F. 45, 69, 103, 133, 235, 239–245, 248, 254f., 257f. Swinburne, Richard 231 Taylor, Charles 150 Thalberg, Irvin 6, 9, 13, 18, 142, 150, 156 Thomson, Judith J. 246, 262 Toulmin, Stephen 177 Tuomela, Raimo 38 Tyles, James 176 Velleman, J. David 2, 9, 30–32, 39, 44f., 157, 193, 204, 210, 216–218, 220–229, 301, 323, 327f., 331, 340 von Wilamowitz-Mçllendorf, Ulrich 267 von Wright, George H. 2, 7 Wallace, R. Jay 28, 90, 97, 223 Wiggins, David 245f. Will, Frederick L. 167–169 Williams, Bernard 22, 80, 98, 126, 231, 245


Name Index

Williamson, Timothy 285, 297, 352, 365, 378 Wilson, George M. 13, 104 Wittgenstein, Ludwig 1f., 5, 7, 28, 66, 68, 71, 103, 138, 242, 379–381

Woodfield, Andrew


Zangwill, Nick 323, 340 Zemach, Eddy M. 237

Subject Index ability 44, 54f., 167f., 171f., 317, 326, 329, 361f., 369f., 374–378, 381 action – ~ in context 1, 4f., 13, 17, 34, 36–38, 40, 54, 122, 164-167, 173, 180, 181, 183, 190f., 277, 280f., 303, 323 – explanation 8f., 9, 18, 20, 30–32, 42, 46, 75, 83, 87, 98, 100, 105, 112f., 115f., 120, 124, 127, 129f., 133f., 141, 188, 191, 301, 308, 331, see also explanation – intentional 8, 13, 17, 39, 56, 61, 65–67, 72, 78f., 83, 87, 103, 106, 110, 112, 150, 154, 156–161, 164f., 167, 170, 173, 181, 186f., 189, 197, 199f., 202–204, 222, 248f. – responsive 75, 77, 79, 93, 95 action theory 1–4, 6, 9, 11, 21, 29, 37, 39f., 44, 48, 124, 133, 181, 184, 189, 275, 321, 332, 354, see also theory of action activity (active) 7, 11, 12, 30-33, 35, 40, 46, 53, 55–57, 60, 64f., 72, 142–144, 159f., 161, 164–166, 168, 170, 175, 180, 213f., 242, 247, 262-269, 274, 280f., 322, 324–329, 334f., 337, 339, 345, 247–249, 352–363, 369f. agency 7, 29–31, 33, 35f., 38f., 43, 53, 72, 137f., 140–142, 144, 147f., 151f., 154–161, 189, 216, 231, 242, 245, 247–249, 264, 276, 279, 282f., 285, 294f., 326, 331, 352, 355 agent – activity 7, 11, 29 – causation 9, 11, 43, 157f., 224, 327, 329, 331f. – erring ~ 26f., 41f., 47


236–240, 244, 254–259

belief 2f., 5, 7., 10–13, 15–18, 20, 22–25, 27, 29–31, 41f., 46f., 55f., 65, 68f., 71f., 75f., 78f., 81–89, 91–95, 98f., 104–107, 109f., 115, 117–119, 124–129, 131, 133f., 138, 155–157, 164–167, 170f., 173, 177, 190, 194f., 200f., 204, 206f., 209, 212f., 216, 223–225, 232–236, 243, 245, 247f., 285–292, 294–298, 300f., 303–313, 323, 327–329, 340, 351–356, 358–365, 367, 370, 373, 376f. – false 108, 294, 298, 300, 305f., 310f., 362, 367 belief-desire model 2, 75, 77–79, 83–85, 87f., 91–93, 328f., see also theory of action, standard belief-desire pair 30f., 83f., 86, 91, 95, 123, 125, 216, 285f., 304, 308, 328 body 5f., 40, 45, 57–59, 62f., 65, 112f., 121, 137f., 143, 151, 231, 240f., 245–247, 249, 258f. bodily 137, 227, 238, 241–247, 258f., 327, 245, 247, 249, 251, 255, 257 – movement 7f., 29f., 112, 157, 202, 242, 247, 249 Cartesian atomism 237, 239f. causal 2, 7–9, 12f., 18, 30, 32, 40–43, 63, 82f., 87, 92, 103, 105f., 112–117, 119–121, 124f., 129, 131, 137–146, 148–152, 154–161, 173, 201, 209, 217–219, 226, 246, 292f., 295, 301, 327, 329–332 – chain (model) 137–142, 145–147, 150f., 156–159, 161, 306


Subject Index

– deviance 154f., 157, see also deviant causal chains – explanation 7, 11, 18, 26, 30, 33, 41f., 63, 103, 105, 113f., 115f., 119f., 124f., 133, 141, 154, 292, 311, 327 – mechanisms 43, 140f., 151, 154–156, 159f. – powers 81f. – relationship 2, 7f., 12f., 21, 30, 32f. causalism 35, 103–105, 113, 120–122, 124, 140–142 causality 2f., 6f., 9–13, 18, 22, 26, 31f., 40, 42f., 93, 125, 128, 137, 140, 154, 157f., 161, 326f., 329, 331 cause 2, 7–10, 12f., 18, 21, 29, 43, 54, 63, 71, 82, 89, 103–106, 112–114, 116, 118, 120f., 124–127, 134, 137–139, 141, 143–151, 154f., 157, 160, 164, 292, 305–307, 311, 326, 328, 330 cognition 39, 41, 47f., 315f., 318, 320f., 325, 333–335, 338, 341, 344–348, 361 cognitive aim 334f., 338, 362 commitment 13, 15–17, 38, 56, 130, 164, 168, 170, 185, 190, 194–196, 204–206, 209f., 216f., 221, 223–227, 229, 266 community 3, 15, 250 constitutive 13, 15, 36, 38, 86f., 118, 133, 224, 245, 247–249, 375f., 380 contexts 1, 4, 17, 38–40, 43, 122, 164–166, 173f., 181, 184, 232f., 235f., 244, 277f., 281, 303, 306, 348, 358, 364 control 6, 15f., 34, 37, 43, 65, 150, 156, 289, 335, 362, 369 coordination 44, 169, 195, 198, 208, 211, 216, 218, 226 decision 10, 14, 16, 41f., 56, 66, 68, 70–72, 175f., 183, 186, 214, 219–222, 224, 228, 281

deed 46, 53, 55, 250, 264–267, 275, 280 deliberation 35, 71f., 76, 166, 172, 177–179, 186, 195f., 198, 211, 214, 220, 224f., 227, 352–354, 359 descriptions (of actions) 2–5, 8–10, 12, 19, 37, 43, 117, 121, 130, 146, 167, 200, 205, 210 descriptive dualism 45, 254, 258f. descriptive metaphysics 235, 239–241, 243–245, 248 desires 2f., 5–8, 10f., 13–20, 22–25, 29f., 34f., 42, 44, 75, 78f., 81–87, 92–94, 104, 109f., 117, 124–129, 131–134, 155f., 164, 167, 171, 173, 190, 196, 213, 217–219, 222, 247f., 304, 306, 308, 311, 322f., 325–327, 329f., 340 deviant causal chains 13, 30, 32f., 43, 120, 145, 147, 150–152, 155, 327, 331 directions of fit 47, 129, 133, 316, 323, 333f., 340, 350 disjunctive view 42, 98 disposition 25, 29, 35, 44, 86, 105, 140f., 145, 169, 181, 185, 190f., 195, 221, 265f., 298, 301 distinction 344–350 enabling condition 93, 190 epistemology 119, 132, 201, 203, 285, 351–355, 360f., 363, 369, 372, 377 error cases 26, 108f., 116, 127 ethics 21, 24, 34, 121, 264, 272, 285, 288, 304, 325 events 3, 7–10, 12f., 15, 30, 32, 92, 105f., 112f., 116, 121, 125, 137–140, 142, 144, 148–151, 156–158, 160f., 176-179, 240, 250, 261, 269f., 277, 280, 301, 305, 326, 328, 332, 350 evolution 45, 209, 211, 216, 218 expectation 3, 5, 61, 66, 81, 207, 225f.

Subject Index

explanation – Humean ~ 116f., 119f., 124, 131, 134 – intentional 154, 158 – non-factive 97, 99–101, 291, 299, 310 – normative 26, 291f. – teleological 42, 113–115, 124, 150 – rational 18, 75–77, 83f., 86f., 91–93, – reason ~ 95, 100f., 103, 105, 113–116, 121, 124, 126–130, 133 first-person 28, 62f., 68, 71, 131, 237–239, 249, 295, 315, 349 flow 213f., 227–229 freedom 65, 72, 210f., 326 functionalism 140, 218, 222f., 226, 228, 259 future 14–17, 61f., 64, 66, 70, 72, 172f., 195–197, 199f., 205, 210, 216, 245f., 248, 271f., 277f., 377 goals 35, 42, 48, 53f., 67, 78, 84, 103, 113–116, 118, 120f., 124, 129–131, 133f., 142–144, 148f., 165, 189, 196, 205f., 217, 227, 248, 307, 311 – ~ of inquiry 353–361, 363, 366, 369f., 374 habit 28, 43f., 163–165, 167–174, 177–185, 187–191, 311 history 3, 38f., 46, 92, 166, 170, 190, 261f., 265–267, 269, 272–278, 280–283 holism 1–5, 12f., 40, 45, 169, 248, 256 human being 1, 45, 53–55, 58, 61, 63, 65, 129, 131, 184, 191, 262, 267, 269, 271, 275, 277, 315, 346–349 Humean model 115, 123, 325-329 idealism

24, 235, 244


identity 232, 236, 240, 245ff., 249ff. – agential ~ 244, 247–250, 255, 257 – bodily 244f., 247f. 249, 251, 255 – narrative 45f., 243f., 250, 255, 257 – self- ~ 237, 239 – temporal 237, 239 inquiry 48, 163f., 174, 176f., 351, 353–355, 357–362, 364–366, 368–370, 372–379 intending 6, 12–14, 65f., 69–71, 148, 151, 202–204, 216–218, 221–228 intensional 88, 125, 130 intention 3, 5–8, 10, 13–17, 21, 32f., 35, 38f., 40f., 43–45, 56f., 61, 64–72, 78, 86, 90, 103, 117, 121f., 130, 133f., 138, 155–157, 187, 190, 194–212, 216–226, 228f., 242, 248, 259, 323, 329, 331, 335–337, 340, 348 intentionality 3, 13, 27, 40, 189, 331 justification 11, 18f., 24, 28, 66, 68, 164, 173, 177, 179f., 184, 201, 322, 326f., 330, 339, 352, 356, 380 know-how 164–167, 186 knowing 66, 68f., 128, 204, 208, 227, 237, 265, 280, 287, 296, 298, 300, 309, 312f., 315–326, 328, 333–341, 347, 350, 366, 372, 374–380 knowledge 3, 5, 17, 26, 39f., 46–48, 54, 56f., 61, 64–66, 68–71, 77, 112f., 146, 163, 165, 177, 183, 199–205, 208, 210, 216, 222, 237f., 255f., 285, 289f., 296–301, 303f., 312f., 316–319, 321f., 324f., 328f., 333–335, 338, 340, 345, 347f., 351–358, 361–370, 372–381 – ~ how 378–381


Subject Index

– practical 103, 163–165, 167f., 170, 173, 181f., 186, 203, 210, 325, 328 – ~ that 378–381 – value of ~ 123, 352, 364–366, 370, 374 life

3, 41, 44–46, 65, 72, 104, 124, 170, 175, 212, 214, 250, 262, 264–273, 277–279, 281f., 315, 341 logical atomism 237, 239f. logical connection argument 8, 42f., 105, 117, 124, 131 maxim 90–92 mechanism 43, 137, 139–141, 144–152, 154–156, 159f., 200, 203, 257 moral 1, 18, 21, 34, 42, 44, 57, 84, 91, 128, 160, 163, 172–174, 176–178, 184, 213, 233–235, 242–245, 250f., 261, 264–267, 280f., 304, 325 morality 38, 174, 213, 322 motivation 22, 27f., 35, 93, 126, 128f., 133, 152, 164, 170f., 173, 194, 196, 211f., 217, 219, 226, 228f., 238, 291, 293, 295, 300, 311 motive 2, 8, 18–20, 22, 26, 29–31, 33f., 36, 39, 104, 141, 210, 297f., 330f., 340 narrativity 247, 250 naturalism 22, 24, 35, 160f., 234 normative 3, 11, 13, 15–24, 26, 31–35, 37–42, 46f., 75–77, 79–88, 93–98, 101, 124, 126f., 164, 173f., 177, 180, 216–219, 223, 226, 228f., 268, 291–294, 298, 311, 327, 340, 354 normativity 5, 11, 16, 19, 21–24, 26, 30, 35, 37–39, 95, 126, 173, 184 optative

220, 222–225, 228f.

passivity (passive) 46, 53, 57–60, 64f., 71f., 170, 214, 242, 281, 347 perceiving 58f., 61, 63, 104, 174, 237, 334–336, 338 perception 17, 25, 40, 46, 54–62, 65, 78, 89, 137–139, 150, 238, 241, 255, 316, 334f., 338, 340, 345, 347, 362 – self- ~ 25, 41, 329, 332, 340 personal identity 44–46, 231–239, 243–251, 254–258 – bundle theory of ~ 231–233, 236–239, 243f., 251, 254–256, 258 – ego theory of ~ 231–233, 236–239, 243f., 251, 254–257 – impersonal theory of ~ 234f. personalism 45, 231, 233, 235f., 243–247, 249–251, 254–260 planning theory 13–17, 21, 29, 44, 195, 216f., 219, 222, 226 point of view 18, 20, 26, 28f., 45, 58, 60f., 75, 77, 79, 81, 84–89, 96, 184, 207, 232, 258, 349 – explanatory 75f., 79f., 96 – normative 75, 86, 96 potentiality 139, 145–148, 151, 154, 159, 245 practical 5, 7, 11–17, 20f., 26, 37, 39, 41, 43, 45, 48, 76f., 93, 101, 148–150, 155, 163, 165, 176, 179, 201, 204, 209, 220, 224, 232–236, 244, 298, 321f., 325f., 328, 331, 341, 352–355, 359f. – justification 174, 178–180 – rationality 17, 43, 148-150, 155, 224, 325, 328, 353, 355 – reasoning 103, 155f., 176, 186, 205, 208, 214, 223, 352, 363 practice 28, 36, 39, 42, 44, 56, 75, 95f., 101, 107, 124, 126–128, 133, 163–170, 173f., 177, 180–183, 185, 188, 216, 232–234, 243f., 250f., 258, 264, 271, 298, 301, 303f., 307–310, 312, 338, 345, 347, 361–363, 365f., 375, 380

Subject Index

– emotional 233f., 243, 250, – existential 232, 234, 244, 250, – moral 233f., 242 – social 44, 179, 181–185, 191 pragmatic 4, 44, 146, 164, 173f., 178, 194, 208, 218, 232 pragmatism 44, 48, 163f., 173f., 177, 179f., 212, 374 prediction 69f., 103, 201, 225 priority of understanding over explanation 329, 332 process 32, 37, 43, 46, 48, 91f., 106, 112, 121f., 137, 139–141, 145–152, 154–161, 170, 175–177, 179, 217f., 220, 235, 244, 280, 337, 345, 347f., 354, 359f., 373, 376 prudence 38, 184, 236, 244, 250 purpose 33, 59, 65, 67f., 78, 94, 108, 113–115, 120, 124, 128f., 150, 163, 165, 174, 193–195, 208f, 211, 218, 222, 336, 359 rationalisation (rationalise) 2, 10, 18, 20f., 37, 83–87, 125, 188, 248f., 285, 303, 305, 326, 340, rationality 3, 7, 10, 16–19, 21, 28, 39f., 43, 45, 47f., 54, 72, 148–150, 155, 205f., 223–225, 294, 320, 325f., 328, 352f., 355, 357, 359f. reactive attitudes 45, 235, 242–245, 250, 255 realism 24, 26, 323, 332, 339 reason 54, 67f., 71f., 194f., 213, 356, reasons – explanatory 18f., 26, 28, 41f., 46f., 80, 82–84, 88f., 93, 96-98, 101, 115 – external 11, 22, 28, 35f., 126, 128, 187 – factive concept of ~ 24–27, 41f., 46f. – ~ for action 8f., 18, 21f., 26, 42, 75f., 79–82, 84, 88, 93, 95–99, 101, 128, 138, 141, 154, 164,


166, 173f., 181, 183f., 186–191, 288, 303f., 307f., 329, 331f. – having a ~ 89, 119f., 295, 298f., 306, 308f. – internal 18, 22, 28, 126, – justifying 18f., 21–23., 47, 2123, – motivating 8, 18–20, 23, 25f., 42, 79, 97, 101, 105, 107, 109, 112f., 115, 124, 126–128, 292–295, 301, 311 – normative 18–26, 28, 35, 41f., 46, 75–77, 79–85, 88, 93, 97f., 101, – objective 23, 84f. – primary 5, 7f., 10–12, 14, 18, 20, 23f., 26, 83f., 86f., 105, 326–328 – practical 2, 6f., 11, 18–24, 35, 128, 173f., 184, 321f., 330, – subjective 41, 84f. – ~ teleological 138, 148, 155, 329, 331 – two different conceptions of ~ 285f., 288, 304 reasoning 54f., 108, 113, 127, 157, 175–177, 184, 186, 195f., 205, 218, 222f., 225, 262, 277, 345, 359, 362, 366f. reduction 22, 45, 160, 240, 317, 333, 339 responsibility 39, 45f., 71, 147, 150, 152, 160f., 171–173, 176, 189, 232–236, 244, 249f., 265f., 269f., 272, 277, 279f., 282f., 322, 362f., 370 rule-following 379 rules 5, 28, 85, 165–168, 174f., 177, 179, 182, 186, 190, 311, 359, 379 scepticism 48, 355, 360-363 self-governing policies 35f. self-knowledge 44, 72, 199, 208, 237f., 249, 254–257, 259, 341 self-transparency 321, 341 speech 45, 56, 64, 127, 133, 201, 267, 310


Subject Index

state of affairs 25–28, 38, 41, 47, 75–81, 85, 88, 93, 95, 98–100, 107, 112, 130, 151, 172, 206, 306f., 310f., 376 story 42, 44, 250, 261, 268, 272–274, 276, 279–281, 285 sustained causation 150, 156

truth 48, 287, 289, 291f., 299, 305, 309, 313 338, 340, 352, 354–356, 364, 367 type vs. token 82–84, 86, 90f., 116, 130, 240–242, 318–320, 349f.

teleological model 328f., 332 theory of action 1, 3, 5, 31, 37, 40, 43, 47, 103f., 121, 124, 157, 188, 261, 263, 352-354, 360, see also action theory – causal ~ 142, 147f., 151, 161, 247f. – pragmatist ~ 163, 181 – standard ~ 3, 5–13, 16–18, 20, 23, 25, 27, 29–37, 40–44 thought 40, 55–57, 62–64, 171f. tradition 3, 94, 251

value 24, 124, 173–176, 180, 221, 232f., 235f., 243, 323, 325 virtue 34, 46, 92, 118, 174, 179, 264, 266, 268 vision 56, 59–61

uncertainty 131f., 199, 220, 222, 224, 227–229

want(s) 42, 82, 110–112, 117–119, 128–130, 132–134, 212, 216, 219–222, 227f. will 6f., 10, 14, 29, 44, 54, 65, 193–195, 208f., 211-214, 216, 227f.