Concepts of Sharedness: Essays on Collective Intentionality 9783110327175, 9783110326734

Citation preview

Hans Bernhard Schmid, Katinka Schulte-Ostermann, Nikos Psarros Concepts of Sharedness Essays on Collective Intentionality

Philosophische Analyse Philosophical Analysis Herausgegeben von / Edited by Herbert Hochberg • Rafael Hüntelmann • Christian Kanzian Richard Schantz • Erwin Tegtmeier Band 26 / Volume 26

Hans Bernhard Schmid Katinka Schulte-Ostermann Nikos Psarros

Concepts of Sharedness Essays on Collective Intentionality

ontos verlag Frankfurt I Paris I Ebikon I Lancaster I New Brunswick

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 http://dnb.d-nb.de.

North and South America by Transaction Books Rutgers University Piscataway, NJ 08854-8042 [email protected]

United Kingdom, Ireland, Iceland, Turkey, Malta, Portugal by Gazelle Books Services Limited White Cross Mills Hightown LANCASTER, LA1 4XS [email protected]

Livraison pour la France et la Belgique: Librairie Philosophique J.Vrin 6, place de la Sorbonne; F-75005 PARIS Tel. +33 (0)1 43 54 03 47; Fax +33 (0)1 43 54 48 18 www.vrin.fr

2008 ontos verlag P.O. Box 15 41, D-63133 Heusenstamm www.ontosverlag.com ISBN 978-3-938793-96-1

2008 No part of this book may be reproduced, stored in retrieval systems or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use of the purchaser of the work

Printed on acid-free paper FSC-certified (Forest Stewardship Council) This hardcover binding meets the International Library standard Printed in Germany by buch bücher dd ag

Contents

Concepts of Sharedness

Part I: Shared Attitudes

Raimo Tuomela, Helsinki Collective Intentionality and Group Reasons

3



Sondra Bacharach, Wellington; Deborah Tollefsen, Memphis Collaborative Arts and Collective Intention

21



Clotilde Calabi, Milano Winks, Sighs and Smiles? Joint Attention, Common Knowloedge and Ephemeral Groups

41



Hans Bernhard Schmid, Basel Shared Feelings. Towards a Phenomenology of Collective Affectice Intentionality

59

Part II: Analysing Sharedness

David P Schweikard, Köln Limiting Reductionism in the Theory of Collective Action

89



Frank Hindricks, Groningen The Status Account of Corporate Agents

119



Björn Petersson, Lund Collectivity and Circularity

145



Antti Saaristo, Helsinki On the Ontology of Collective Intentionality: A Constructivist Perspective

165



Katinka Schulte-Ostermann Agent Causation and Collective Agency

191

Part III: Sharedness and Normativity

Facundo Martin Alonso, Stanford Reliance and Intending the Joint Activity

211



Jennifer Hudin, Berkeley The Logic of External Reasons and Collective Intentionality

225



Monika Betzler, Bern Valuing Interpersonal Relationships and Acting Together

253



Nikos Psarros, Leipzig Mental Non Self-sufficiency, Sociality and Common Agency

273



Francesca Raimondi, Potsdam / Frankfurt Oder Joint Commitmenr and the Practice of Democracy

285

About the Authors – About the Editors

305

Concepts of Sharedness The present volume contains a selection of papers presented at the Fifth Conference on Collective Intentionality held at the University of Helsinki August 31 to September 2, 2006. The collection includes two additional contributions, namely the invited paper co-authored by Sondra Bacharach and Deborah Tollefsen, and a contribution by Katinka Schulte-Ostermann that completes this volume. This volume is a follow-up on earlier collections of essays on collective intentionality.1 The common aim of the papers gathered here is to explore the structure of shared intentional attitudes, and to explain how they underlie the social, cultural and institutional world. The volume is divided into three parts. The first part – entitled Shared Attitudes – contains papers devoted to the phenomenology of sharedness. Raimo Tuomela (Collective Intentionality and Group Reasons) opens the discussion by presenting his analysis of the We-mode, which is a strong form of collective intentionality that involves group reasons based on the group’s construction of it as a reason to act qua a group member. We-mode acting is thus essentially based on group reasons, i.e. on what the group decides, orders, or requires. In order to be acting as group members, the individuals therefore have to transfer some of their own authority to the group. This contrasts with the I-mode case where an individual is fully in charge of whatever she or he undertakes. In the second paper, Sondra Bacharach and Deborah Tollefsen (Collaborative Art and Collective Intention) bring the recent discussion on collective intentionality to bear on a new field, namely aesthetics. The background of their contribution is the debates in the philosophy of art concerning the role intentions play in interpretation. Intentionalism is the view that intentions are relevant to the interpretation of a work of art. Despite postmodern criticism, intentionalism of some form or another has maintained a healthy following. But those who advocate an intentionalist theory of interpretation have focused entirely on the intentions of an individual author. As Bacharach and Tollefsen convincingly show, the individualism presupposed in traditional intentionalist theories of interpretation poses a 1 Selections of papers presented at previous Conferences on Collective Intentionality appeared in G. Meggle (ed.), Social Facts and Collective Intentionality, Frankfurt: Hänsel-Hohenhausen (2001), in A. Meijers and F. Hindriks (eds.), in: Philosophical Explorations (2003) VI, in L. Tummolini and C. Castelfranchi (eds.), in: Cognitive Systems Research (2006) 7(2&3), special issue, and in L. May and R. Tuomela (eds.), in: Journal of Social Philosophy (2007) 38(3): iv-503.

II

serious problem for interpreting those artworks, which are collaboratively produced. It seems that in such cases an intentionalist interpretation has to go beyond an individualistic understanding of intentionality. In the next chapter, Clotilde Calabi (Winks, Sighs and Smiles) addresses a topic that has spurred considerable controversy in recent philosophical research. Developmental psychology has shown how basic joint attention is in the development of human cognition. By contrast, philosophers usually believed joint attention to involve a highly complex structure, which makes the received accounts of joint attention unfit for the purpose of explaining the base of human cognition. Calabi examines and assesses various new philosophical accounts of the structure of joint attention, and she addresses the core question of how infinite iterations in the explanation of joint attention can be avoided. Hans Bernhard Schmid asks why emotions have received only marginal notice in the analysis of collective intentionality (Shared Feelings – Towards a Phenomenology of Collective Affective Intentionality). The reason, he argues, is the predominance of cognitivism in recent philosophy of the emotions. According to cognitivism, emotions are intentional only insofar they involve cognitions and intentions. If cognitivism is correct, an account of shared cognitions and shared intentions is all that is needed to understand how emotions can be shared. However, Schmid argues that the analysis of shared emotions has to extend beyond the cognitive and practical components of emotions, and has to include the phenomenal aspects, i.e. the feelings. To share an emotion, Schmid claims, people have to share a feeling. Schmid addresses the question of how feelings can be shared, and he shows how the strong conception of sharedness, which he advocates, relates to the view that feelings are essentially individual. The second part of the volume is devoted to the analysis of the concept of sharedness. David Schweikard (Limiting Reductionism in the Theory of Collective action) addresses a concern that has been in the back of the mind of most philosophers of collective intentionality, namely that by reducing collective intentionality to individual intentionality, important phenomena are excluded from the analysis. Schweikard proposes a taxonomy of different types of collective action, and he examines which types of reductionism or non-reductionism can be applied to which kinds of phenomena. Schweikard concludes that there is no single reductive or nonreductive approach applicable to all types of collective action. Frank Hindriks (The Status Account of Corporate Agents) aims at developing an account of collective agents from the external perspective. According to Hindriks, corporate agents may come into existence not only by

III

means of the participants’ own taking themselves to be (and acting as) its members, but also by virtue of external assignment of status. The status account of corporate agents is refined and discussed with reference to the major competing accounts of corporate agency in the received literature. Björn Petersson (Collectivity and Circularity) addresses an objection raised against some of the most successful among the received accounts of collective intentionality, especially against Michael Bratman’s. In Bratman’s analysis, shared intentional activity is analyzed in terms of a notion of individual intentionality that has collective action in its content. Petersson aims at showing that this does not render Bratman’s account viciously circular. Petersson argues that at the basic level, the collective content can be accounted for in terms of causal structures and dispositions. Antti Saaristo (On the Ontology of Collective Intentionality – A Constructivist Perspective) lays down the outlines of a social constructivist view of the ontology of collective intentionality in general, and the we-perspective, including the collective mode of beliefs, desires, intentions and actions, in particular. Saaristo seeks to combine uncompromising naturalistic materialism about the mind with strong social constructivism about the origins of intentionality and intentional agency. Katinka Schulte-Ostermann (Agent Causation and Collective Agency) takes up some of the threads pursued in this second part of the volume by drawing attention to the controversy between individualistic and holistic explanations of collective agency. Her claim is that this controversy rests on the action-theoretical concept of agent causation. Only on the presupposition that the main difference between individual and collective action lies in the different nature of the subject of each action type does it become relevant to give an account of the relation between an action and a plural subject. There is an explanatory gap not only between collective actions and collective agents, but also between individual agents and their actions. Schulte-Ostermann presents two possible solutions to fill this gap: The first would be to develop a more convincing theory of agent causation, the second would be to endorse an anti causalist approach to action. The third and final part of the volume explores various aspects of the structure of collective intentionality in general, and of the intricate relations between sharedness and normativity in particular. Facundo Alonso (Intending the Joint Activity and Reliance) examines the cognitive conditions of an individual’s intention to participate in a joint activity. When A and B share an intention to act together, A’s intention that A und B act depends on his taking B to be similarly disposed. Facundo’s aim is to determine

IV

what precisely the cognitive attitudes towards the other participants are, and he proposes a concept of reliance for that purpose. Jennifer Hudin (The Logic of External Reasons and Collective Intentionality) argues that external reasons have different logical features from internal reasons and therefore require a different ontology. The features of external reasons involve first person plural indexicalizing rather than merely first person singular indexicalizing. Thus, external reasons are We-reasons rather than I-reasons. We-reasons do not motivate but impel action, therefore they do not require a sound deliberative route to produce intentional action. Recognition of truth conditions alone is sufficient to produce action. We-reasons impel action because of moral intuition and commitment to a social group. Moral intuition and social commitment are pre-theoretical notions an agent acquires by virtue of sharing a social background with other members of his social group. Monika Betzler (Valuing Interpersonal Relationships and Acting Together) draws our attention to a particular class of reasons that play an important role in explaining joint action, which should not be ignored – as it is often done in the received accounts. These are reasons generated by a particular kind of relationship between agents. Theories of joint action that do not take these relationship-dependent reasons into account are at least incomplete. The distinct explanatory role that can be assigned to interpersonal relationships, and to the reasons generated by them, is revealed by the fact that even when persons act jointly, we sometimes believe that we are justified in criticizing them for not acting jointly in the right way. An adequate analysis of a particular class of joint action has to refer to relationship-expressive reasons. Nikos Psarros (Mental Non Self-Sufficiency, Sociality and Common Agency) tries to give a theoretical account of the problems that accompany intentionalist accounts of collective activity by claiming that these problems stem from the idea of self-sufficiency underlying most classical accounts of personhood. This idea logically excludes the idea of cooperation that is essential for the concept of collective action. Building upon the concept of the needs of the human soul as analyzed by the French philosopher Simone Weil, Psarros outlines a concept of common agency that is based on the idea that human beings as persons and individual agents are mentally non-sufficient entities, i.e. that humans rely necessarily on other humans in order to develop their ability of conscious agency. Francesca Raimondi finally (Collective Intentionality and the Practice of Democracy) examines some aspects of Margaret Gilbert’s recent attempt to develop a theory of political obligation that is based on her plural sub-

V

ject account. Gilbert’s approach seems promising because it does not ground political obligations on metaphysical facts, or on individualistic values such as self-preservation. Rather, Gilbert ties political obligations to shared practices. Raimondi argues, however, that Gilbert’s concept of social group may be inadequate for the purpose of analyzing political collectives. Furthermore, the concept of group runs the risk of suggesting too much identity, and seem to be unfit to capture the plurality that is part of a liberal democratic ideal of political collectives. We hope that this volume shows how rich and lively the philosophical research focused on the analysis of collective intentionality has become, and will provide further inspiration for future work in this rapidly evolving field.2 Basel, Essen and Leipzig, April 2008 The Editors

2 We wish to thank the organizers of the fifth Conference on Collective Intentionality, especially Raimo Tuomela and Antti Saaristo, and the authors for their patience and cooperation. Special thanks go to Jan Hanisch who played a pivotal role in the preparation of the layout of this volume.

Part I: Shared Attitudes

Collective Intentionality and Group Reasons Raimo Tuomela, Helsinki

1. The we-mode and the I-mode Human beings are social beings living and adapted to living in groups, indeed, to functioning in several different groups during their lives. They can think and act as group members and as “private” persons. Functioning as a group member requires functioning on the basis of group reasons, or so I argue, and this again requires thinking based on collective intentionality (“aboutness”). Collective intentionality in its central sense is based on “we-thinking,” thinking in terms of a “we-perspective.” The following two basic kinds of collective intentionality can be discerned: a) Full-blown or “we-mode” collective intentionality, which involves (i) shared “for-groupness” based on collective acceptance, and (ii) collective commitment to the shared content. Here for-groupness satisfies a strong kind collectivity condition (see below) making precise the intuitive idea of the participants necessarily “standing or falling” together with respect to the satisfaction or truth or attributability (as the case may be) of the shared content. b) “I-mode” collective intentionality, which in the case of acting as a group member contingently involves shared “progroupness” but not the satisfaction of the (strong) collectivity condition and which the person still acting in a privately committed way, in the “progroup” I-mode, as I say.1 Functioning as a group member in the full sense accordingly is functioning in the we-mode (in sense (a)) and is of course to be distinguished from functioning as a private person, in the “I-mode,” even in cases of kind (b) where group reasons become involved. We-mode acting is essentially based on group reasons, viz. on what the group decides, orders, or requires (etc.) and where the group members qua group members give some of their ”natural” authority to the group. This contrasts with the Imode case where a person in principle is fully in charge of whatever she 1 Mere mutual belief need not involve “progroupness” but may still be regarded as Imode collective intentionality. N. Psarros, H. B. Schmid and K. Schulte-Ostermann, Concepts of Sharedness, Frankfurt 2008: 3-19; ontos verlag

4

Raimo Tuomela

undertakes. The we-mode level can be spoken of as the proper group level while the I-mode level gives the private level of social life. Let me elaborate on the above. As said, acting as a group member in the we-mode sense constitutively involves acting for a collectively constructed group reason – the group gives a group member reasons to think, “emote,” and act in certain ways. For instance, the group’s constitutive goals, values, and beliefs provide such group reasons. Viewed from the group members’ we-perspective we have examples of group attitudes such as “We will build a bridge together” (group intention), “We believe that stars determine our fate,” or “We believe that euros are our money” (group beliefs). Such we-mode group attitudes can serve as full-blown group reasons. Thus, that “we believe, etc, that thus and so” can serve as a group member’s reason to act in certain appropriate ways amounting to his part of the group member’s collective action, broadly understood. In contrast, the I-mode is concerned only with “private” personal and interpersonal reasons and relations as well as with groups involving such ingredients. Group reasons in a weak sense may contingently be involved. When they are, we are dealing with I-mode collective intentionality. More typically, shared attitudes had as a private person represent I-mode collective intentionality. A central case illustrating that the indispensability of group reasons at least in some contexts is provided by social institutions. They involve special fully collectively constructed institutional, “non-natural,” contents (e.g., “Euros are our money”), and this falls or at any rate seems to fall, beyond the conceptual resources of the I-mode. The important divide here is between a group thinking and acting as one agent versus some agents acting and interacting, perhaps in concert, in pursuit of their (possibly shared) private goals. Group reasons (we-mode reasons) and I-mode reasons (private reasons) for acting and having attitudes thus are clearly of a different kind, and accordingly the group level (containing group reasons) is not reducible to the private, I-mode level, although these levels as such are compatible. Accordingly, only the we-mode can properly account for the generality that the group level involves with respect to group members (change of membership, future members, etc.) and the kind of (partial) depersonalization that group life involves. We-mode thinking, feeling, and acting presuppose collective acceptance of the group’s constitutive goals, values, norms, standards, traditions, etc., briefly its “ethos,” or of some other, non-constitutive content as the object of the group’s “attitudes.” The collectively accepted contents must be taken to be for group use, viz., to be collectively available and in force for

Collective Intentionality and Group Reasons

5

the group members, and, when broadly conceived, for the benefit of the group’s goals and interests. It is often useful to view a group as an agent capable of acting as a unit. Thus it can be taken to accept views, form intentions, act, and be responsible. However, it is not an extra agent over and above the group members. When a group acts, its members must act as group members. In a sense, one can thus redescribe the group’s functioning and acting at the group member level, in terms of the group members’ functioning in appropriate ways as group members. This is basically we-mode activity. It follows from the idea of a group acting or functioning as one agent that the members ought to function appropriately. They can be said to be necessarily “in the same boat,” “stand or fall together,” or share a “common fate.” Here the strong Collectivity Condition is satisfied. Formulated for the special case of goal satisfaction it necessarily connects the members as follows: Necessarily (as based on group construction of a goal as the group’s goal), the goal is satisfied for a member if and only if it is satisfied for all (other) members. Thus, if you and I have the goal to go to Alfonzo’s for lunch, this goal as a we-mode goal is satisfied for me only if we both go to Alfonzo’s (and similarly for you). However, if the goal is had as an I-mode goal, it becomes satisfied for me if I go to Alfonzo’s – irrespective of what you do. Another element, over and above thinking and acting as a group member, conceptually involved in the we-mode is collective commitment. In order for a group member to act as a group member, it must thus be required that she be committed (bound) to performing actions that further the group’s ethos and other matters that the group is pursuing.2 Indeed, the members should be collectively committed, viz., committed as group members, to participating in group activities. Their collective commitment involves that they are also “socially” committed (viz., directly committed to other group members) to each other to act in the right group ways. Collective commitment has two basic, intertwined roles here. First, it “glues” the members together around an ethos. This gives the foundation for the unity and identity of the group. Secondly, collective commitment serves to give joint authority to the group members to pursue ethos-related action.3 They can and must in their own thinking and acting take into account that the group members are collectively committed to the group ethos and 2 In weakest cases the collective commitment need not be properly normative, e.g. morally normative. 3 This joint authority can be delegated to special members or representatives, below called “operative” members (see (GINT)).

6

Raimo Tuomela

to the group members and that they are jointly responsible for promoting the ethos. Every group member is accountable not only to himself for his participatory action but also to the other members. All this shows how group unity as formed by collective commitment to the ethos relates to action as a group member. In view of the above, it can be said that the we-mode is constituted by two elements, a content element and a “practical” (action-related) element, viz., collective commitment. To illustrate, we consider a two-person case in which a goal (or intention, belief, etc.) is collectively accepted (constructed) and held by two persons, you and me. The case involves two elements: (i) G is our goal, where “our goal” satisfies the aforementioned Collectivity Condition; (ii) we (you and I together) are collectively committed to goal G (as our goal). I claim that (i) and (ii) give the intuitive “rock bottom” of the we-mode. Actually, (ii) is part and parcel of (i) and can be regarded as entailed by it. The participants’ being collectively committed to goal G involves that they are committed to doing their parts of their joint action concerned with their achieving G. The joint goal that they here have constructed for their group serves as their reason for their performing their parts. The notion of a joint goal satisfies the Collectivity Condition. Due to its being satisfied, the notion of “we” is not reducible to the conjunction “you and I” although it entails it.4 The we-mode elements (i) and (ii) can intuitively be viewed as (partial) translations of group level descriptions of the following kinds: (i') (group g’s goal is G (where g has you and me as its sole members) (ii') group g is committed to goal G (as its goal). (i’) and (ii’) can be regarded as equivalent. Hence also (i) and (ii) are seen to be equivalent from our present group-level perspective. The present

4 Collective acceptance and construction relevant to our goal example are discussed in Tuomela (2007a, Chapter 8, in terms of the reflexivity-involving thesis (CAT) about collective acceptance that is shown to amount to a general form of the Collectivity Condition. While one can perhaps speak of construction of goals, etc., also in the “progroup” I-mode case there is no group construction but only aggregated private construction. A we-mode group is taken by its members to form a unit capable of action, whereas this kind of collective acceptance is missing in the case of I-mode groups, which only require interaction and dependence between the members.

Collective Intentionality and Group Reasons

7

point applies, mutatis mutandis, also to intentions, beliefs, and other voluntary attitudes. The concept of the we-mode in itself contains the notion of group (and expresses part of the group-perspective). It is a holistic institutional concept while ontologically it can be regarded as being about “groupishly” interrelated group members. As said, thinking and acting in the we-mode basically amounts to thinking and acting for a group reason, i.e., a group member’s taking the group’s views and commitments as his authoritative reasons for thinking and acting as the group “requires” or in accordance with what “favors” it (viz., its goals, etc.). A central notion that is needed is that of a social group, and in this paper I focus on the “we-mode” group, based on the one hand on a “we” concerned with togetherness and, on the other hand, on the group’s ethos. A we-mode group is one that in principle can act as one agent. If and when the members jointly intend to satisfy the ethos they are collectively bound (committed) to it, because intention can be taken to entail commitment in the relevant sense. Additionally, the full notion of the we-mode requires that the group members accept the group’s ethos and other goals and beliefs, etc. This acceptance concerns both their action and their acceptance (or at least disposition to accept) these group features in the reflective and reflexive sense that the specific constellation of goals, values, beliefs, etc., indeed is the group’s ethos and that those other elements indeed are the group’s nonconstitutive goals and beliefs. The acceptance referred to above is assumed to be publicly available knowledge in the group. Especially social psychologists tend to speak of group identification in the present kind of context. If this is taken to mean identifying (or partially identifying) oneself with a group it cannot be literally true at least if a group is regarded as a collective entity (e.g. agent). A member of a collective cannot literally be identically to what he is a member of. I will not here go deeper into the issue but wish to point out that in my account the elements of full-blown “groupness” discussed above help to make clear what group identification involves: In my approach it basically amounts to functioning (thinking and acting) in the we-mode – as a full-blown group member – relative to the group in question. As seen, the we-mode basically involves functioning for ethos-compatible group reasons with collective commitment and “being in the same boat” with the other group members (so that the strong Collectivity Condition is satisfied). To end this introductory section, let me present a brief characterization of we-mode and I-mode reasons in more precise terms as above. The later discussion will amplify the account of we-mode (or group) reasons. The

8

Raimo Tuomela

criteria below are given for a simple case relating to the context of a specific group, assuming that a reason is a fact or a fact-like entity expressible by a that-clause: (IMR) Reason R is a group member’s motivating I-mode (or private) reason for performing an action X if and only if R is the agent’s main motivating private reason for his performing X. Typically, R is a state that the agent wants or intends to be the case or a state that, according to his belief, obtains; and X is an action that is a means to R or an action that R requires for its obtaining such that the agent is privately committed to performing X for the reason R. (WMR) Reason R is a group member’s motivating we-mode (or full-blown group) reason for performing an action X if and only if R is the agent’s main motivating group reason for his performing X. Typically, R is a state that the group in question wants, intends, or requires to be the case or is a state that, according to the group’s belief, obtains; generally speaking R is a state that is “for the group.” X is an action that is the individual’s part of a collective action that is a means to R or a collective action that R requires for its obtaining, where the group members are collectively committed to performing the collective action for reason R and mutually believe so. In (WMR) X can be a collective (or group) action with multiple tokens (e.g., going to church on Sundays) or a joint action like cleaning up a park as a many-person action. As the group members are collectively committed to performing the collective action in question for reason R (a state expressible by a that-clause), they are also socially committed to the group members to performing their parts of the collective action for reason R. A full-blown, viz., a we-mode, group reason will have to satisfy the Collectivity Condition. Having a private commitment means in (IMR) that the person privately (rather than as a group member) has psychologically bound himself to a “content,” e.g., to performing an action for a reason. In general, private commitment is dependent on an intention, e.g., the intention to reach a goal. As said, functioning in the we-mode is necessarily connected to a thick group reason, to what one’s group has committed itself to in the situation at hand, where the group’s commitment serves as an authoritative reason for the participants. In contrast, functioning in the I-mode is at most only contingently connected to a group reason – that must be an I-mode one, if there is one involved.5 5 A group reason is a “desire-independent” reason in Searle’s (2001) sense (see Tuomela 2007a: ch. 6).

Collective Intentionality and Group Reasons

9

In this the rest of this paper I will discuss group reasons by focusing on the case of group intention and joint intention (Section II) and also discuss reasoning based on group reasons (Section III).6

2. Joint intention and we-intention My focus in this paper will be on the case of intention, especially so-called “we-intention” and joint intention in the we-mode sense. I will next discuss them in an analytical fashion – keeping in mind such examples as the joint intention jointly to lift a table, to sing a song, or to make a contract.7 We can say that the participant intends to perform her part of a joint action X at least in part for the (“internal”) reason that the group intends as a group to perform X (this may simply amount to the participants’ having formed the joint intention to perform X as a group).8 The group’s intending can be based on some collectively authorized, “operative” agents’ intending rather than all the members’ so intending. A person intends to perform her part of X primarily because the group intends to perform X and only derivatively because the others (distributively considered) intend to participate. The group reason here is a conceptual condition for a participant’s we-intention and a reason for her intending to perform her part (and for her performing it). That also the others or a sufficient number of them participate is both a conceptual and a rationality condition for one’s intention to participate. In the we-mode case the participants are functioning as group members and taking the group as their authority for their intention formation. This contrasts with the I-mode case, where they are functioning merely as private persons committed to a goal. A we-mode joint intention is expressible by “We as a group will perform X jointly.” Here the participants are assumed to take themselves to form a thick “we” (“we-together” or “we-as-a-group”), which is not a mere aggregative “we” that signifies “the two of us” (the basic I-mode sense of “we”). The “we-as-a-group” notion indicates the we-mode sense of “we.” 6 I have recently, in Tuomela (2007a), developed and presented a detailed theory of collective intentionality that considerably improves my earlier work in this field. This paper is largely based on the ideas and formulations presented in this book and on my paper on reasons in Tuomela (2007b). 7 The discussion in this section draws from Chapter 4 of Tuomela (2007a). In that chapter I also discuss I-mode joint intentions. 8 Here X can sometimes be an end to be collectively achieved, e.g. the end that the bridge shall be built collectively by the participants or, more generally, by their group. (Notice that in the latter case not all members need to participate in the full sense but some can be ”in reserve”, cf. the account (GINT) below.).

10

Raimo Tuomela

We can say that the intentional subject of a we-intention is “we” while the ontological subject of a we-intention is a single agent.9 Suppose now that the joint intention to perform X together is attributed to m real agents, say A1,...,Ai,...,Am assumed to understand what the joint action X involves. (The participants need not be identified by their standard names, but, e.g., as position holders in a group.) In this case, every participant Ai has, as his personal “slice” of the participants’ joint intention, the we-intention to perform X together with the others. The group reason that each Ai here has for participating in the performance of the joint action X is that their group so intends, which upon analysis amounts to the fact that the participants jointly intend X. They form a group here already because their having a joint intention means that they take themselves to be capable of acting as a group, and this will in effect make them a social group. My analysis can now be summarily formulated as follows:10 (WI) A member Ai of a collective g we-intends to do X if and only if (i) Ai intends to do his part of X (as his part of X); (ii) Ai has a belief to the effect that the joint action opportunities for an intentional performance of X will obtain (or at least probably will obtain), especially that a right number of the full-fledged and adequately informed members of g, as required for the performance of X, will (or at least probably will) perform their parts of X, which under normal conditions will result in an intentional joint performance of X by the participants; (iii) Ai believes that there is (or will be) a mutual belief among the participating members of g (or at least among those participants who perform their parts of X intentionally as their parts of X there is or will be a mutual belief) to the effect that the joint action opportunities for an intentional performance of X will obtain (or at least probably will obtain); (iv) (i) in part because of (ii) and (iii). I have assumed that the participants actually exist, but I allow that a participant might in some contexts be mistaken in his (rationally presupposed) beliefs (ii) and (iii). (On pain of not we-intending at all, he cannot 9 However, note that in my account ”we” can be represented by a predicate, WE, applying to m individuals for some m. Thus no ontologically significant plural subject is literally postulated here, seemingly contrary to, e.g., the view by Gilbert (1989). 10 This account is a slightly modified version of the one in Tuomela and Miller (1988).

Collective Intentionality and Group Reasons

11

be mistaken about the general situation at stake, viz., that there are some agents about to perform a joint action.) Thus, in such abnormal circumstances a single agent can in principle have a we-intention, and here a we-intention is not objectively a “slice” of a joint intention. Below I will, however, assume that the acceptance of “We will do X together” must be veridical and entail the existence of a joint intention. Thus it will be assumed below that all the agents in question really have the we-intention. As to (iv), it lays bare the fact that the presuppositional beliefs cannot be idle and serve to make the participation intention rational. As indicated, it is a mutually believed conceptual (presuppositional) condition for an agent’s we-intending to participate in performing X with the others and for his intention to perform his part of X that also the others (of sufficiently many of them for X getting performed) similarly participate. Furthermore, that they all (or even some of them) actually have formed the intention to participate in X is a contingent fact that a rational agent will take as a proximate reason for taking part. As said, the primary reason for a participant’s intention to perform his part of X in the we-mode case – on conceptually “internal” grounds – is the group’s intention (here in effect made up of the agents’ joint intention). The group’s intention conceptually – and often also causally – precedes an agent’s we-intention, and it is always a partial reason for the agent’s performing his part.11 The joint intention accordingly is the participant’s reason at least for her intention to perform her part and will in the case of an antecedently existing joint intention be a “newcomer’s” reason for her joining in, viz., for her forming her we-intention on the basis of the group’s previously formed intention. Recall that the participants are collectively committed to seeing to it that their intention to perform X together will be satisfied. This collective commitment is a conceptual feature involved in their intention.12 The presupposed beliefs (ii) and (iii), expressing the minimal rationality of the we-intender concerning what a we-intention conceptually involves, as well as condition (iv) will not be commented on in detail here.13 Let me 11 The joint intention can be based on a joint decision, agreement-making, or just be what a collective acceptance of a joint plan involves. As such, the formation of a joint intention (that amounts to the group’s intention) need not be based on a group reason. This point resembles that involved in saying that agreement making need not be based on agreement making. 12 In Tuomela (2002) I have characterized the we-mode in terms of three features: (1) the attitude or action must be collectively accepted as the group’s attitude or action, (2) it must be for the use and, typically, benefit of the group’s interests, and (3) the participants must be collectively committed to the attitude or action. 13 See especially Tuomela and Miller (1988) for justification.

12

Raimo Tuomela

only say that the joint action opportunity conditions include, besides the relevant mental and physical abilities of the participants, also that the others (or at least sufficiently many of the “right” kinds of them, as required for an intentional performance of X) will indeed participate.14 Supposing that joint intentions can be expressed by “We as a group will do X” or its variants, in order to cover “standing” intentions in addition to directly action-generating intentions we must also take into account dispositions to we-intend.15 The following can accordingly be regarded as a true claim: (JI) Agents A1,...,Am have the joint intention to perform a joint action X if and only if: (a) These agents have the we-intention (or are disposed to form the weintention) to perform X; and (b) there is a mutual belief among them to the effect that a). The entailment from left to right is obvious. As to the entailment of the joint intention by clauses (a) and (b), notice that there is a kind of holistic interaction effect here, because the agents’ we-intentions presuppose joint intention.16 This fact guarantees the truth of the entailment in question. The joint intention towards X is assumed to entail joint commitment towards X. Another point that still needs to be made here concerns broader cases in which a group can be said to intend to perform an action. The simple intuition here is the consensus intention: a group intends as a group when its members qua group members so intend. However, in the case of structured groups the matter can be viewed somewhat differently. Here one can make a distinction between group members who are somehow authorized (typically by the group members) and thus given the group-social power to form intentions for the group and possibly to realize these intentions. (There can be several hierarchical layers connected by the authority-relation but in order not to complicate issues unnecessarily only the twolayered case is considered). The authorized members are called operative members. The following summary analysis of group intentions seems ad-

14 See the account in Tuomela (1995: ch. 3. 15 See also Tuomela (1995). 16 Intuitively, the circularity problem arises because a we-intention basically amounts to one’s intention to perform one’s part of a fully intentional, viz. joint intentiongoverned joint action. I have discussed the problem elsewhere (e.g. in Tuomela, 2007a) and argued that the circularity is not vicious.

Collective Intentionality and Group Reasons

13

equate for groups that are externally and internally free (e.g., dictators are thus not allowed by the present formulation):17 (GINT) Group g intends to perform action X (or intends that a state X obtain) as a group if and only if there are operative members of g such that (1) these agents, when acting as group members (and accordingly performing their positional tasks due to their exercising the relevant decision making system of g), intentionally collectively form the joint intention towards X (e.g., partly by accepting the conative expression “We will do X” or one of its cognates for g) and are collectively committed to bringing about X; (2) there is mutual belief among the operative members to the effect that (1); (3) because of (1), the (full-fledged and adequately informed) non-operative members qua members of g tend to tacitly accept with collective commitment – or at least ought so to accept – that their group g intends to perform X (as specified in clause (1)); (4) there is mutual belief in g to the effect that (3). Clauses (1) and (3) express the most central ideas here. (2) and (4) express rationality conditions that might not be properly fulfilled while there still is a group intention on the basis of the fulfillment of (1) and (3). We can see from the present account, which mutatis mutandis applies to all voluntary group attitudes, that the non-operative members can in a central way take part in the group’s intention simply by functioning as group members and (tacitly) accepting the operatives’ joint intention, or at least being normatively obligated to such acceptance. This tacit acceptance can amount to the non-operatives’ endorsing the operatives’ joint intention qua members of g but they can be “in reserve,” e.g., concerning the execution of the intention. For instance, the sales personnel of a department store (an organization) can take part in the organization’s intentions and actions just by doing their work and without perhaps knowing very much about what the operatives for decision making are doing. They may but need not actually have the we-intention in question, but only take it to be the case that the operatives have accepted “We will do X” for the group. They still can be said to (weakly) participate in the group’s 17 See Tuomela (2007a: ch. 4 and 6). I would like to emphasize that the present account of group attitudes (such as intentions) applies to the case of voluntary attitudes. As to non voluntary one’s, e.g. “experiential” beliefs or emotions, a different approach, and aggregative I-mode one, in terms of shared we-attitudes in terms of, respectively, shared non voluntary we-beliefs and shared we-emotions is needed (see Tuomela, 2007a, chapter 3).

14

Raimo Tuomela

intention. It is even possible that the analysans of (GINT) is satisfied even when some non-operatives fail to tacitly accept “We will do X” for g and the participation intentions that this intention expression entails for them (together with other relevant information). Such persons may be dissidents, but also persons hired to help the group to achieve X may be a case in point. Let me end this section by commenting on the so-called “bootstrapping” problem. In the single agent case, an agent’s intention cannot at least always be her reason for the action that the intention is about. This is because one can rather arbitrarily form an intention and act on it. Such bootstrapping generally need not be involved in the case of a joint intention.18 A joint intention – as a jointness level counterpart of a group reason – can be a reason both for the members’ joint action and for their performing their parts of it. To show this, consider the example of some agents forming the joint intention to paint a house together. Then the previously formed joint intention may collectively serve as their reason for their painting the house together and it also serves as each participant’s reason for his performing his part of the joint action. In general, there may be a prior group intention or belief that is a group reason for each group member qua a group member to think and act according to it, and no bootstrapping is involved. But when the group member concurrently participates in the very formation of the group intention or belief – e.g., by agreeing with the others that they are to do something together or accept a view for the group – there is a kind of bootstrapping involved. However, we are here dealing with a (possibly) emergent jointness phenomenon centrally involved in the group’s acceptance and, on the group member level, collective acceptance. No single member can of course create this jointness, nor in general can mere aggregation. Thus, the jointness often is non-aggregative and emergent (“creative”) relative to the member’s Imode attitudes (or his “proposed” we-mode attitude input). Let me add that if the participants have an underlying joint want, then that want (or its content, rather) may sometimes serve as a deeper joint reason than does the joint intention formed on its basis. But think of cases involving negotiations and bargaining between the participants that finally lead to a joint intention that is a compromise and is against some participants’ underlying wants (at least first choices). Here the jointness of the intention indeed is creative and can serve as the meant kind of group reason (and the we-mode proattitude that it involves can serve as a distant motivational reason). Once the group attitude is in existence, the group members, 18 See Bratman (1987: ch. 2).

Collective Intentionality and Group Reasons

15

when functioning qua group members, can, and ought to, use it as their reason for thinking and acting. It follows that no harmful bootstrapping needs to be involved.

3. Joint intention and group reason An I-mode account of joint intention is concerned with putting together individuals’ private intentions when they for some reason want to act together or achieve something together.19 To exaggerate somewhat, the basic model in the two-person case, is: I regard you as a (cooperative) part of my environment rather than as a proper fellow group member and we face a task to be solved together by our actions.20 In contrast, in my wemode account you and I form a thick “we,” a group, which acts possibly in order to solve a jointly accepted task, or we just act together for purely social reasons (“each of us values and enjoys the company of the others, basically no matter what we are doing together”). What I will say below will apply not only to intentions but also to beliefs and actions, although I focus on the intention case. In my we-mode account the intention-dependencies go as follows in an unstructured group g that has formed an intention: (1) Group g, consisting of the agents A1,…,Am, intends to bring about X (or see to it that X), where X may in principle be any state or event (as long as no rationality constraints on g are imposed). As groups are not literally agents, we must speak of the group members’ relevant states and actions: (2) The members A1,…,Am of g jointly as a group intend to bring about X; put in more linguistic terms, these agents jointly accept that the intention expression “We will bring about X,” satisfying the Collectivity Condition, is true of them collectively (and also individually, see (3)) as group members. A more general way to arrive at group members’ joint intentions here is to speak of the group’s relevant authority over them. Note that in (2) the members’ (and in the normatively structured case the operative members’) joint intentions may have been formed for the reason that the group’s previous decisions and plans play an authoritative role for them. 19 Bratman’s (1999) account discussed in Chapter 3, is a case in point. 20 One can, however, cooperate in the I-mode and to achieve a shared goal and act for a group, and one can also value the other group members’ company – see Chapter 7 for I-mode cooperation.

16

Raimo Tuomela

After (2), the next step is the following “distribution” assumption: (3) Each one of the current members A1,…,Am of g in effect accepts, qua a member of g, in this situation that “We will bring about X” is true of himself. (4) Each member must reason or be disposed to reason as follows: (i) We will bring about X Therefore, taking (i) as my reason, (ii) I will participate in our (viz., here A1,…,Am’s) jointly bringing about X or jointly seeing to it that X As a paraphrase of (ii) we have (5) I will perform my part of our jointly seeing to it that X. It is inessential here what exactly a part of X involves – it can even be passive and conditional, accordingly to what the functionality of the group’s performance of X requires. In general, the participants must at least rationally hope that X will come about due to their joint effort. This serves to give a partial indirect characterization of what “a part” and “participation” involve. We can spell out (5) more fully to get (6) I intend to perform my part of our jointly seeing to it that X as my part of X when acting as a member of g. (6) presupposes that our jointly seeing to it that X is our jointly intentional action performed as jointly intended (that is, for the reason that we jointly intended, possibly in a rudimentary, preanalytic sense only, to jointly see to it that X). (6) gives the main individual-level content of what my notion of we-intention involves. Its group-level content is given by (1) and its jointnesslevel content is basically given by (2) while (3)–(6) concern the individual personal level of acting as a group member. Note that any collection of agents that satisfy (1)–(6) form a group capable of action, and so does a collection capable of forming a joint intention (and having mutual belief about this).21 We may leave it to psychologists to find out how people in ordinary life, as it were, simplify (6) when intending and acting jointly so that the instrumentally and functionally right things come about. There is some cir21 Notice still that the concept of joint intention can be taken to entail: If the members A1,…,Am of g jointly intend to bring about X, then they are jointly committed (bound) to (continue) intending and to bringing about X.

Collective Intentionality and Group Reasons

17

cularity here: a person’s intention to perform his part comes to depend on joint intention, although perhaps only on an unarticulated, preanalytic notion of joint intention. This does not matter much as long as people function satisfactorily in terms of the above account and produce X in about the right way. To end, I will present a simple and stylized example related to action reasons. I-mode reasons and we-mode reasons may lead to dramatically different results in the case of strategic social action e.g. in collective dilemma situations. To illustrate, consider a simple two-person two-choice case of a Prisoner’s Dilemma with the familiar choice alternatives C and D, where the row player’s preference ranking is DC, CC, DD, CD and the column player’s symmetric ranking is CD, CC, DD, DC. Here the agent may choose either C or D when he acts for an I-mode reason (involving I-mode intention). Thus, considering the single-shot case under uncertainty, if the agent – the row player, here “I” – thinks strategically and intends to maximize his value or utility he can reason thus: I prefer the joint outcome DC to all the other joint outcomes; however, I realize that if you are reasoning similarly and planning to go for CD, we will end up in DD. Still, wanting to avoid the worst outcome CD, I cannot rationally aim at the Pareto optimal CC. So I choose D and rationally expect DD to result in the single-shot case. In this case the agent’s reason for choosing D is to secure at least the third best alternative given his beliefs of the nature of the game and the other player’s rationality. Thus, as long as change of the game structure is not allowed, in a single-shot PD (under uncertainty) mutual defection is the rational outcome, and this result holds independently of how egoistic or altruistic the participants’ preferences and utilities are. In contrast, acting for a we-mode reason can lead to the group members’ rationally cooperating in collective action dilemmas (on the basis of their joint intention). Considering what is a Prisoner’s Dilemma situation from the point of view of the group members’ private preferences, we now adopt the group’s agent’s point of view and assume that the group (“our group,” “we”) accepts the dominance principle (“higher payoff dominates lower payoff”) and thus intends to choose C (over D). Thus, as the group can be taken to intend and act only if the members correspondingly jointly intend, as a group, to cooperate and accordingly jointly cooperate, it follows that the participants form the joint intention to realize the joint cooperation outcome and perform their parts of the group’s achieving it, viz. do C. As the group members are assumed to act as one agent, the joint outcomes CD and DC simply drop out as prohibited (unless there is a special division of tasks requiring them, here this is not considered).

18

Raimo Tuomela

The Prisoner’s Dilemma simply disappears in principle (and we are dealing with a simple “Hi-Lo” coordination situation). The agents act collectively rationally when they act for a we-reason as here, and indeed there is no room for private individual action and private rationality at all. The upshot then is that while strategically acting agents acting on the basis of the relevant I-mode reason involving maximization of private value (utility) they can rationally choose only D in the single-shot case (assuming full uncertainty), but in the case of agents acting for the relevant we-mode reasons (e.g. “our group will maximize the value for the group by acting appropriately”) the agents will rationally choose C. Of course, it must be added to all this that in actual practice groups may not act fully as units in the above committed way. There may be free-riders and there must be then be some sanctioning involved. Thus, from the backdoor, as it were, we may get the Prisoner’s Dilemma alternatives back. Furthermore, intergroup cooperation in a collective action dilemma is not of course resolved by member level, intra-group we-mode acting, but that will require the use of a group reasons involving the “we” formed out of the groups in question (cf. the EU states cooperating with each other).

4. Conclusion This paper has emphasized a strong form of collective intentionality called the we-mode that essentially involves a group reason, one based on the group’s construction of it as a reason to act qua a group member. Wemode acting thus is essentially based on group reasons, viz. on what the group decides, orders, or requires (etc.) and where the group members functioning qua group members give some of their ”natural” authority to the group. This contrasts with the I-mode case where a person in principle is fully in charge of whatever she undertakes. The above ideas were elaborated in the paper and discussed especially in the context where the group members jointly intend to do something together.

References Bratman (1999): Bratman, M. Faces of Intention, Cambridge: Cambridge University Press, 1999. Gilbert (1989): Gilbert, M. On Social Facts, London: Routledge: 1989.

Searle (2001): Searle, J. Rationality in Action, Cambridge/MA: The MIT Press, 2001.

Collective Intentionality and Group Reasons

19

Tuomela (1995): Tuomela, R. The Importance of Us: A Philosophical Study of Basic Social Notions, Stanford: Stanford University Press, 1995. Tuomela (2002): Tuomela, R. The Philosophy of Social Practices: A Collective Acceptance View, Cambridge: Cambridge University Press, 2002. Tuomela (2007a): Tuomela, R. The Philosophy of Sociality: The Shared Point of View, New York: Oxford University Press, 2007. Tuomela (2007b): Tuomela, R. Motivating Reasons for Action, in: Timmons, M., J. Greco, and A. Mele (eds.), Rationality and the Good: Critical Essays on the Epistemology and Ethics of Robert Audi, New York: Oxford University Press, 2007. Tuomela and Miller (1988): Tuomela, R. and Miller, K. We-Intentions, in: Philosophical Studies (1998) 53: 115-137.

Collaborative Art and Collective Intention1 Sondra Bacharach, Wellington/New Zealand and Deborah Tollefsen Memphis/TN

1. Introduction Until recently, the notion of collective or group intentionality was likely to be discussed only in conjunction with the work of Emile Durkheim and early 20th century debates over the proper methodology of the social sciences. We now find, however, philosophers of mind and action theory turning their attention to the issue. One of the reasons for the resurrection of collective intentionality has been to explain collective action. Individual intentions shape and inform individual actions. But we do not always act alone and it is collaboration with others that raises interesting issues regarding the possibility of collective intentions.2 Many prominent philosophers of action theory, including John Searle, Michael Bratman, Margaret Gilbert, and Raimo Tuomela have argued that collective action cannot be understood merely in terms of individual intentions to perform some action.3 Rather, collective action requires for its explanation the positing of collective intentions. In this paper we bring this recent discussion of collective action and collective intention to bear on the debate in the philosophy of art concerning the role intentions play in interpretation. Intentionalism is the view that intentions are relevant to the interpretation of a work of art. Despite post1

The research for this paper was supported, in part, by a Marsden Fast Start Grant from the Royal Society of New Zealand. The authors would like to thank James Hamilton (Kansas State University) and John Gibson (Temple University) for their helpful comments on earlier versions of this paper. We would also like to thank audiences of the 2005 Eastern division meeting of the American Society for Aesthetics and the 2006 Central States Philosophical Society meeting for their insightful questions and comments. 2 There are a variety of other phrases used to capture the notion of collective intention or something similar to it. A survey of the literature will turn up the phrase “shared intention,” “group intention,” “joint intention,” among others. For ease and clarity, we have chosen to use the phrase collective intention and when discussing Bratman’s account, shared intention. 3 Cf. Searle (1990, 1995), Bratman (1992; 1993), Gilbert (1989; 1996; 2003), and Tuomela (1995; 2006). N. Psarros, H. B. Schmid and K. Schulte-Ostermann, Concepts of Sharedness, Frankfurt 2008: 21-40; ontos verlag

22

Sondra Bacharach, Wellington/New Zealand and Deborah Tollefsen Memphis/TN

modern criticism, intentionalism of some form or another has maintained a healthy following. But those who advocate an intentionalist theory of interpretation have focused entirely on the intentions of an individual author. This is due, in part, to the fact that theories of interpretation have been developed with literary cases in mind and literature tends to be a solitary endeavor. We think, however, that the individualism presupposed in intentionalist theories of interpretation poses a serious problem for interpreting other forms of art. In contrast to literature, many artworks are collaboratively produced – consider for example, Hollywood movies, ballets, theater, musical performances, jewelry and furniture.4 In these cases it isn’t clear whose intentions are relevant to the interpretation of the work of art. The focus in the philosophy of art on individual intentions should come as no surprise. As is the case in action theory, individualism is the default position. The renaissance of interest in collective intentions, however, has encouraged us to investigate how collective intentions might be able to help philosophers of art accommodate a broader range of art, including collaboratively produced art. We shall argue that acknowledgment of collective intentions is essential to any intentionalist theory of interpretation that wishes to be applicable to collaboratively produced art. A consequence of this argument is that we will have shown that collective intentions play a role in explaining human action and the artifacts that result from human action. Although the current accounts of collective intention have done much to reveal the structure of joint action, the explanatory power of collective intention has been neglected in these discussions.5 So, our aim in this paper is to contribute both to debates in the philosophy art and aesthetics, as well as discussions concerning collective intentionality. In section I, we provide a brief discussion of the debate regarding the role intention plays in the interpretation of art. In section II, we present several examples of collectively produced artwork in order to reveal the deficiencies of any brand of intentionalism that focuses entirely on individual intentions and to motivate the need for an appeal to collective intentions. In section III, we suggest various strategies an intentionalist might use to accommodate our cases and find each wanting. In section IV, we draw on Michael Bratman’s (1993) account of collective intention, what he calls 4 Our discussion here will be restricted to co-authored artworks – that is, artworks created by more than one individual, as opposed to artworks that are created by a single individual, which are then performed or interpreted by a number of artists. 5 An exception to this is Tollefsen (2002a).

Collaborative Art and Collective Intention

23

shared intention, and show how it might be used as a basis for an intentionalist interpretation of collectively produced art.

2. Intentionalism In interpreting the behavior (both linguistic and non-linguistic) of individuals and the outcome of their behavior, we regularly appeal to intentions. If I want to explain why Sue opened the window, I will appeal to her desire for fresh air and her intention to fulfill this desire by opening the window. Since an artwork is the outcome of human behavior, appeal to intentions in interpreting art seems natural. The work has the features it has, in part, because the artist acted with some intentions in place. The recent debate regarding interpretation centers around the role of intentions in understanding and interpreting literature. While many agree that intentions matter, there has been much discussion about the nature and extent to which intentions are relevant to interpreting art. The two main forms of intentionalism on offer these days are modest actual intentionalism (MAI),6 advocated most notably by Noël Carroll,7 and hypothetical intentionalism (HI), advocated by Levinson, Tolhurst, and Nehamas.8 According to actual intentionalism, the actual intentions of the author determine the meaning of the work of art. This position has stronger and weaker versions, depending on the degree to which intentions fix the meaning of the artwork. The most popular version of actual intentionalism is moderate, claiming that the author’s or artist’s intentions constrain interpretations of the artwork: “With reference to literary texts, the modest actual intentionalist argues that the correct interpretation of a text is the meaning of the text that is compatible with the author’s actual intention”.9 The modest actual intentionalist takes seriously those intentions of the author that the linguistic/literary unit can support. When the intentions of the artist can support a variety of different interpretations, modest actual intentionalism appeals to social, historical, and other contextual factors to identify the correct interpretation. Hypothetical intentionalism (HI), on the other hand, claims that the correct interpretation or meaning of an artwork is constrained not by the actual intentions of the author or artist, but by the best hypothesis available 6 Intentionalism has many adherents, among the most prominent: Iseminger, (1992), Hirsch (1967; 1976), and Stecker (2003; 2005). 7 Cf. Carroll (1992; 1994; 2000). 8 Cf. Levinson (1996), Tolhurst (1979), and Nehamas (1981; 1986). 9 Carroll (2000: 76).

24

Sondra Bacharach, Wellington/New Zealand and Deborah Tollefsen Memphis/TN

about what the artist intended. On this view, the meaning of a text is the one that an ideal reader is justified in attributing to the author. Like modest actual intentionalism, when an ideal reader is justified in attributing to the author a variety of possible intentions regarding what the artwork meant, the correct interpretation will be the one that also takes into account various social, historical and other contextual factors. There are numerous arguments for and against these competing versions of intentionalism, many of which are well rehearsed elsewhere.10 We do not intend to enter the debate between the modest actual intentionalist and the hypothetical intentionalist, nor do we intend to defend intentionalism of either form here. Rather, we are interested in tackling a different problem: can intentionalism accommodate collaboratively produced art? When many artists are involved, whose intention is relevant to the interpretation of the work of art?

3. Collaboratively produced artworks In what follows, we present some cases of collaboratively produced art and use them as a way of revealing the difficulties current versions of intentionalism face with respect to the interpretation of art work that is collaboratively produced. These cases provide a representative sample of various types of collaboration but are no means meant to be exhaustive.11 3.1 Iri and Toshi Maruki Iri and Toshi Maruki are a Japanese couple who created The Hiroshima Panels, a group of large panels, painted over the course of three decades, that depict the devastation of the bombing of Hiroshima. Although the Marukis are both widely known as individual artists, they are also famous for their collaboration in this and other series. What is distinctive about Iri and Toshi is that they have radically distinct styles and methods of painting. Toshi was trained in Western style oil painting, focusing on representing humans in a relatively realistic manner, while Iri was trained in traditional Japanese brush painting, focusing on landscapes in a rather expressive and non-realistic manner.12 In spite of 10 See, for instance Carroll (2000) and Levinson (1996). 11 Artists working collaboratively is fairly common: pvi collective, Jochen Gerz, The Atlas Group, Abramovic and Ulay, Mel Ramsden and Ian Burn, Atelier van Lieshout, Bureau d’Etudes, The Yes Men, Christo and Jeanne-Claude, Papunya Tula Artists, Komar and Melamid, Group Irwin, Dumb Type, Tim Rollins and K.O.S., Seymour Likely, to name just a few. 12 Dower and Junkerman (1985: 11).

Collaborative Art and Collective Intention

25

this, however, they managed to negotiate a space that allowed them to act together to produce the murals. Given their very different backgrounds and skills, it would be difficult for them to agree on all of the aesthetic properties of the resulting murals. In fact, it seems that they often disagreed on how a painting should look: “He [Iri] and Toshi were really rivals rather than partners, he said – and Toshi agreed – and in innumerable ways their rivalry contributed to heightened creativity.”13 Instead, for some paintings, Toshi and Iri decided that each would be responsible for different parts of the paintings.14 In other cases, they worked on all aspects of the painting, correcting and fine-tuning each other’s work (though they did not always work on the same painting at the same time). One historian describes the process in the following way: “When actually painting the final version of a mural, the great sheets of paper laid flat on the studio floor, they tended to work in silence, each going over what the other had done, back and forth, in what in the end can only be described as a virtually unprecedented act of pure artistic collaboration”.15 Although the styles of each individual are present in the murals, the collaborative process produced aesthetic features of the works that are wholly unique. 3.2 Gilbert & George Gilbert Proesch and George Passmore, known simply as Gilbert & George, also work collaboratively. Unlike the Marukis, however, they have never exhibited individually; indeed, they have always and only worked together. Further, their collaboration differs from the Maruki’s in that their collaboration seems to be a focus of their work. Thus, their collaboration has a reflexive element in that they aim to represent the “collective” nature of their work in the work itself. The first body of works by Gilbert & George involves the artists themselves appearing as ‘living sculptures’ on display. As living sculptures, they actually constitute the work of art being exhibited. On display in these public appearances, they are dressed in highly stylized attire, performing banal activities and gestures in a very exaggerated and ritualized manner. Their actions draw attention to their selves and to the ordinariness of their lives, and by performing these simple rituals of life in the museum setting, they are declaring them to be art. The forming of the sculptures requires that 13 Dower and Junkerman (1985: 25). 14 “When the Marukis decided to paint Hiroshima together in 1948, they had no model for either what to paint or how to collaborate. It was clear only that Toshi, by virtue of her training, would be responsible for most of the figures” (Dower and Junkerman 1985: 25). 15 Dower and Junkerman (1985: 25-26).

26

Sondra Bacharach, Wellington/New Zealand and Deborah Tollefsen Memphis/TN

they cooperate physically. Further, the ideas for the sculptures are arrived at via a joint deliberative process. The decision to perform live had important implications for the artists’ lives as individuals. For one, such a decision required the artists to travel around in order for their work to be on display. As a result, exhibiting their work meant giving up their own individual private lives for the sake of their art.16 This has led many to interpret their work in a way that makes reference to this sacrifice and to their unique collaboration.17 By creating a new name “Gilbert & George,” they also make clear that responsibility for the work is shared. In sacrificing their own identities, the living sculptures themselves acquire an identity wholly distinct from the individuals who constitute it.18 The collective spirit of Gilbert and George’s art is evident throughout their entire oeuvre. It is a central theme in their art, and hence a central component of the interpretation of all of their works. For example, the very title of their self-portrait, They of 1986, highlights their cooperative spirit, and emphasizes the degree to which their work explores the idea of collectively produced art. 3.3 Third case: Raqs Media Collective A third kind of collectively produced art is art that has multiple authors, some of whom are either unknown or remain anonymous.19 Such art is more common among artist collectives – groups of artists working loosely 16 “Gilbert & George linked their living sculpture’s believability to their total selfabsorption, creating a meta-identity that encompassed both artists, relegating them to the status of automata or puppets” (Green 2001: 197). 17 “Since 1965, when Gilbert and George met and began working and living together, they have merged their identities so completely that we never think of one without the other; no surnames, individual biographies, or separate bodies of work hinder their unique twinship.” Gilbert & George Biography. Retrieved 31 October 2007, from www(dot)guggenheimcollection(dot)org/site/artist_ bio_52. html. 18 “Their strategy was to make themselves into sculpture, so sacrificing their separate identities to art and turning the notion of creativity on its head. To that end Gilbert and George became interchangeable ciphers and their surnames were dispensed with.” (Andrew Wilson, “Gilbert and George” Grove Art Online. Oxford University Press, Retrieved 30 October 2007, (www(dot)groveart(dot) com/ shared/views/article(dot)html?from=search&session_search_id=546237589&hi tnum=1§ion=art.032200). 19 There are many different degrees and types of collaboration in the art world that have not been touched upon here. We simply hope to motivate, using these three kinds of collaboration, the idea that theories of interpretation ought to take a more serious look at collectively produced art in informing those theories. A more

Collaborative Art and Collective Intention

27

together, and among artists who appropriate from others (either artists or non-artists). The Raqs Media Collective - a collective of artists working as part of the Sarai Media Lab, in Delhi, India – is one such example. Their work, “Network of N_odes” is a HTML-based audio and visual work containing videos, photographs, texts and words from texts. It investigates the way that information travels and is used and then appropriated by others in the Indian computer culture. Borrowing and appropriating many of the texts and images that appear in the “Network of N_odes”, this work explicitly acknowledges the difficulty of identifying the author of much computer-based art. How will an intentionalist interpret these cases? In what follows, we present different possible approaches that the intentionalist – actual or hypothetical - might take in order to handle them. In doing so, we shall highlight some of the unique challenges that arise when interpreting collaboratively produced art.

4. The limits of intentionalism 4.1 Modest Actual Intentionalism Recall that the aim of moderate actual intentionalism (MAI) is to construct an interpretation of an artwork that reflects and is constrained by the intentions of the artist. In our cases, however, there is no single artist that is responsible for the creation of the artwork; rather there is a group. The challenge for the MAI, then, is to determine whose intentions are relevant to the interpretation of the work. Call this the “Whose intentions?” problem. The most plausible strategy for MAI is to choose one individual in the group and interpret the artwork in light of his or her intentions. Our examples, however, suggest that such a strategy will fail. In all the above cases, no one individual had intentions with respect to the artwork as a whole. Rather, individuals have intentions regarding the aesthetic properties of the parts to which they will contribute and they may have general intentions to work with other artists but the final product seems to have aesthetic properties which no one individual intended. Consider again, the Marukis, whose collaborations of differing styles produce artworks with a strange mix of both. Given that these properties became apparent

serious and extensive theory of interpretation, however, would very likely need to consider far more types of collectively produced art than are raised here.

28

Sondra Bacharach, Wellington/New Zealand and Deborah Tollefsen Memphis/TN

only after their collaboration, it seems difficult to say that either of them had an intention to create a work with these properties. But even if one individual in the group did have intentions regarding the final product, which individual does one focus on in interpretation? Do we choose Iri’s or Toshi’s intentions? In cases where both artists contribute equally to the resulting work, focusing on one person’s intentions will always be arbitrary. The work of the Raqs Media collective poses an even more difficult problem: because they have not only appropriated, but appropriated from anonymous authors, it is impossible to discern or identify who has created which portions of the works. The MAI might respond to these worries by offering a summative account of interpreting collectively produced art, according to which the work as a whole is interpreted in light of the sum of intentions of the individual artists responsible for creating the work. In the case of the Marukis, for example, one might cite Iri’s intentions to paint the human figures, and Toshi’s intentions to design the landscapes, and so on. All of these intentions together are then used to construct the appropriate interpretation of the work as a whole. But how is this supposed to work, exactly? In some cases the individual intentions do not seem not to “add” up. We know, for instance, that Iri is skeptical of realism and directness, but Toshi’s painting is rational and literal; how do their individual intentions, which no doubt are informed by these styles, “add” up to explain the overall style of their paintings? Given that Iri and Toshi were sometimes in opposition regarding their style, their paintings often reflect neither Iri’s nor Toshi’s intentions about the painting20. And since we know that sometimes Iri and Toshi would let the other person completely overwrite everything they had painted, it’s not as if there are particular parts of the painting whose properties we can attribute to one or the other artist. Further, in cases where there is no conflict, it is still difficult to understand how the process of summation would produce an interpretation. Consider again the “Network of N_odes”: although there is no explicit conflict, there are at least five artists who are directly acknowledged as the ‘artists’, with three additional people cited as having provided ‘additional research’, and seven others who are ‘acknowledged’ (not to mention all those from whom they appropriated and borrowed without direct acknowledgement). Do we simply ‘add up’ all of these people’s intentions 20 Because their relationship is so intuitive, and they paint directly on the paper without prior outlines or minutely detailed sketches, the final form of the work is never certain.” Dower and Junkerman (1985: 26).

Collaborative Art and Collective Intention

29

to arrive at a suitable interpretation of the work? Surely the intentions of those who were simply acknowledged or those who provided additional research are not nearly as relevant to a correct interpretation of the work as those who actually created the work. It is not obvious how to reflect their lesser contribution in the summative approach. The MAI might attempt to point to one particular intention (or set of intentions) that the artists in the group have in common, rather than simply summing up different individual intentions. They might appeal, for instance, to the fact that Gilbert & George both have the intention to make a sculpture, while Toshi and Iri both have the intention to paint a picture depicting the bombing of Hiroshima, and the Raqs Media Collective members all share the intention to create an HTML-based audio and visual work. Thus, these artists share (in the sense of each having the same) executive or categorical intention with respect to the type of art they wish to create. Such a proposal goes a long way to accommodating the interpretation of collective produced art but falls short in that it seems to leave out the fact that the artwork was the product of collaboration. It is not just that the individual artists had the intention to make a certain type of art (sculpture, mural, and so on) but they intended to make the work of art together. Iri and Toshi could have had the intention to produce a work that depicted the bombing of Hiroshima but created two separate works independently of each other. Moreover, in all of the cases we have examined, the interpretation of these works by art historians and critics makes reference to their collaborative nature. The conflict of styles in Iri and Toshi’s work, for instance, is often noted in discussions of the Hiroshima Panels.21 Discussions of George and Gilbert highlight the collective nature of their work.22 It appears that any adequate interpretation of the work in question involves making reference to the fact that the art is the result of a collaboration. Simply pointing out that the artists all share a common intention, however, fails to acknowledge that they collaborated to produce a single work of art.

21 “the style [of the murals of Iri and Toshi Maruki] derives exceptional depth and energy from the fact that the paintings are a collaborative act by two strong and stubborn individuals whose temperaments and artistic backgrounds differ greatly: one trained in traditional Japanese brush painting and the other in Western-style oils; one devoted to landscapes, flora, and fauna, and the other to the human figure; one skeptical of “realism” and directness, and the other highly rational and literal (the dichotomy here being, in each case, Iri/Toshi, husband/wife).” Dower and Junkerman (1985: 11). 22 See footnotes 18, 19 and 20.

30

Sondra Bacharach, Wellington/New Zealand and Deborah Tollefsen Memphis/TN

4.2 Hypothetical Intentionalism What about hypothetical intentionalism (HI)? Recall that HI claims that the correct interpretation or meaning of an artwork is constrained not by the actual intentions of the author or artist, but by the best hypothesis available about what the artist intended. On this view, the meaning of a text is the one that an ideal reader is justified in attributing to the author (which is not necessarily what the author did, in fact, intend). In interpreting art, our goal is to explain why a given artwork has the particular properties it does, and HI proposes that we do so by generating the best hypothesis available about what intentions the artist must have had. In ordinary cases, this involves hypothesizing the intentions of the individual who created the work in question. Presumably, the hypothetical intentionalist proposes a similar explanation in the case of collectively produced artworks: we need to generate the best hypothesis available about the intentions of individual artists in the group that explains why the work has the properties that it does. Unfortunately HI falls prey to the same difficulties posed for MAI. Whose intentions do we hypothesize? Do we provide a hypothesis about each individual artist’s intentions? Or do we focus on one individual and hypothesize about his or her intentions? Which artist should we choose? If the artwork was genuinely collaborative, as in our cases, there seems no reason for choosing one individual over another. And if we hypothesize about all the individual intentions of the members of the group it is difficult to see how these individual intentions function together to produce an interpretation of the work of art. What if there is reason to hypothesize intentions that are in conflict? The real issue here is that HI requires that we hypothesize the intentions of the actual author. In the cases discussed above there is no single person who is the author; rather there is a couple or a group. Whose intentions are the subject of hypothesis? 4.3 Diagnosing the problem: We are now in a position to think more generally about the challenges posed by collaboratively produced art. In all of the above collaboratively produced works, a crucial social aspect of the work isn’t captured by an intentionalism that focuses on individual intentions, viz., the fact that the artists work together and that they respond to one another’s intentions are central to properly understanding and interpreting such works. Art historians and art critics do not interpret these works as if they were created by a single author. Multiple authorship is often the focus of discussion. Interpreting collectively produced art requires not just understanding the intentions of the individual artists, but

Collaborative Art and Collective Intention

31

also appreciating the way the artists work together, the way their art arises out of a complex interaction, the way the art is generated by shared visions and shared goals. Current forms of intentionalism lack the resources to handle this social aspect of art because they focus on a single individual author/artist. This individualism hinders intentionalism’s ability to provide adequate interpretations of collaboratively produced art work. In the next section, we turn to recent work on collective intention. Our focus is on Michael Bratman’s analysis of shared intention and its role in joint action. We suggest that the structure of collective intentions can help to provide a richer interpretation of collectively produced artwork and this can be done within an intentionalist framework.23

5. Collective intention We know that actions of individual agents are shaped and informed by, among other things, intentions. Intentions also play a significant role in the interpretation of human action. We often attribute intentions in the course of explaining (and predicting) the actions of others. But humans do not always act alone and it is capacity for joint or collective action which has led many to consider the intentional structure required for joint action and its explanation. Many philosophers including Michael Bratman, Margaret Gilbert, John Searle, and Raimo Tuomela, have argued that the underlying intentional structure of joint action cannot be simply a matter of summing up the individual intentions to do X. What is required are collective or shared intentions. Consider the following example Dave and Marianne are washing the dishes together. This is something they do. That is, it is an intentional action. And it is something that they do. That is, they are doing it together. Why can’t we explain this intentional action in terms of Dave’s intention to wash the dishes and Marianne’s intention to wash the dishes, perhaps adding the condition of common knowledge? Even if Dave and Marianne know of the other’s intentions, this does not seem to be enough to guarantee that Dave and Marianne are washing the dishes together. After all, these intentional states could be in place in a case in which Dave washes the dishes in the kitchen sink and Marianne washes them in the upstairs bathroom. In what sense would they be washing the dishes together? This suggests that what is necessary for joint action is that there be in place 23 See Livingston (2005) for a discussion of joint authorship and shared intention. Livingston’s focus is on providing an account of joint authorship rather than how to interpret jointly authored art. Our focus in this paper is on the latter issue.

32

Sondra Bacharach, Wellington/New Zealand and Deborah Tollefsen Memphis/TN

an intention that we do the dishes. But how do make sense of this weintention? There are a variety of interesting accounts of collective intention on offer. Some, like John Searle,24 see we-intentions as primitive forms of individual intentionality had by humans and other species. Others have argued that certain groups, themselves, are the bearers of intentions.25 In the cases we have presented here, we think the most interesting and helpful account of collective intention is Michael Bratman’s. For Bratman the phrase “collective intention” does not refer to the intention of some supra-agent but to the way in which, during joint action, individual intentions are shared. Thus, Bratman uses the phrase “shared intention” rather than “collective intention.” Bratman begins his account of shared intention by noting the role it plays in joint action. He identifies three interrelated jobs. First, shared intention helps to coordinate our intentional actions. For instance, our shared intention of washing the dishes will guide each of our intentional actions towards satisfying the goal of washing the dishes. Thus, someone will wash the dishes before rinsing them and someone will rinse them before drying them. Second, our shared intention will coordinate our actions by making sure that our own personal plans of action meld together. If I plan to do the washing then I will check with your plan and see if there is any conflict. Third, shared intentions act as a backdrop against which bargaining and negotiation occur. Conflicts about who does the washing and who does the drying will be resolved by considering the fact that we share the intention to do the dishes. Thus, shared intention unifies and coordinates individual intentional actions by tracking the goals accepted by each individual. The analogy between shared intentions and individual intentions, then, is quite strong. Having laid out the roles played by shared intentions, Bratman asks “what set of individual attitudes are interrelated in appropriate ways such that the complex consisting of such attitudes would, if functioning properly, do the jobs of shared intention?”26 Here is a somewhat simplified version of Bratman’s answer to this question. We intend to wash the dishes if and only if: (1) a. I intend that we wash the dishes. b. You intend that we wash the dishes. 24 Cf. Searle (1995). 25 Cf. Tollefsen (2002b). 26 Cf. Bratman (1993: 104).

Collaborative Art and Collective Intention

33

(2) I intend that we wash the dishes in accordance with and because of 1a and 1b; you intend likewise. (3) 1 and 2 are common knowledge between us. As a first approximation, this complex of intentional attitudes seems plausible. But consider a case in which we each intend to wash the dishes together and we each do so in part because of the other’s intention. However, I intend to wash the dishes with hot water and you intend to wash them with cold water. All of this is common knowledge and we will not compromise. Do we still have a shared intention? It seems not. In this case we do not have our subplans coordinated in the appropriate way. Recall that one of the jobs that shared intention has is to coordinate our individual plans and goals. In the example above our individual subplans are in conflict and this would prevent us from achieving our goal of getting the dishes washed. Bratman avoids this counterexample by adding a clause about participant’s subplans. It is not necessary that our subplans match but they must mesh. So, if my subplan is to wash the dishes with hot water and your subplan is to wash them with blue dish soap and I have no preference about the color of the soap, then our subplans mesh though they don’t match exactly. But if we have subplans to wash the dishes with completely different water temperatures then our subplans do not mesh. Bratman reformulates the account in the following way: We intend to J if and only if: (1) (a) I intend that we J and (b) you intend that we J; (2) I intend that we J in accordance with and because of 1a and 1b, and meshing subplans of 1a and 1b; you intend the same; (3) 1 and 2 are common knowledge.27 A shared intention is the complex of attitudes of individuals and their interrelations. An individual cannot have a shared intention.28 I cannot have a shared intention to wash the dishes with you, though I can intend that we wash the dishes. It is the interrelatedness of individual intentions which makes joint action possible. This complex of attitudes is what makes ascriptions such as “They intend to wash the dishes” true. Bratman acknowledges29 that when one appeals to shared intention in explanatory contexts we will attribute attitudes to 27 Cf. Bratman (1993: 106). 28 This is the sense, in which Bratman differs from Searle. Shared intention is not to be located “in the head.” 29 Bratman, personal communication.

34

Sondra Bacharach, Wellington/New Zealand and Deborah Tollefsen Memphis/TN

the group but this does not mean that there is a group mind but rather that the complex of attitudes is present in the group. Shared intentions are states of affairs realized by this complex of interrelated individual attitudes. How can this account of collective intention help us to interpret the kinds of cases of art considered here? Recall again the “whose intention?” problem. Intentionalism is committed to the view that intentions are relevant to the interpretation of art. But in many cases art is produced by many people. Which person’s intentions are relevant to the interpretation of the work? In many cases, this is not obvious. Recall Raqs Media Collective’s “Network of N_odes.” This work is made up of materials from not only by the Sarai Media Lab, but also by additional researchers, people whose ideas are acknowledged, as well as various other people whose work is appropriated unwittingly and unknowingly. Even if we could, strictly speaking, identify the intentions of all the contributors to “Network of N_odes”, this would not help us – intuitively, the intentions of those whose works have been unwittingly appropriated are not relevant at all to the interpretation of the work: those of the Raqs Media Collective are relevant, but surely not those whose work has been unwittingly appropriated. But, how do we rule out these contributors with a lesser role, while still including the actual artists of Raqs Media Collective? Bratman’s account can help us identify the group of artists who legitimately can be considered to have authorial command- those with the shared intention to create the work. The members of the Raqs Media Collective, but not the unwitting contributors whose images have been appropriated, share an intention to create the work where this is understood as the complex set of attitudes identified by Bratman’s analysis. As a result, it will be the intentions of those that participate in the shared intention that will be relevant to the interpretation of the art work. The intentions of these individuals may be about their particular part of the work but they will also involve the intention to create the work of art with others and to be responsive to the intentions of others in the group. The individuals in the group will have intentions relevant to their sub-plans and these will, of course, be relevant to interpreting the work of art. But these sub-plans themselves will be informed by the shared intention. The aesthetic properties found in the work will have been determined by these shared intentions even when they worked on different parts of the painting independently because each part will have to be responsive to the overall shared intention to make a work of art with the others. Consider, again, the Marukis. In times of conflict, when, for instance, their styles differ significantly, it is their shared intention to create the murals with

Collaborative Art and Collective Intention

35

certain properties that allows for negotiation and ultimately a resolution to such conflicts. It is plausible to think that the Marukis had a shared intention to depict the aftermath of Hiroshima where this doesn’t simply mean that Iri intended to depict it and Toshi intended to depict it. These individual intentions wouldn’t allow us to understand the complex interaction between the artists that produced the panels; their individual intentions to paint in certain styles and so on where informed by the shared intention that they paint the murals together. When they worked on different aspects of the painting, the shared intention guides and directs their individual intentions and actions. Indeed, it is difficult to see how they would have been able to collaborate for over 30 years without shared intentions in place. Notice that Bratman’s account of collective intention can be accommodated by MAI and most forms of HI. This can be accomplished by investigating the way the art arises out of a complex interaction among the different agents, and studying how the intentions of the individual are coordinated and where they diverge. And these requirements are all perfectly consistent with both MAI and HI. We can easily modify MAI to consider the interrelations among the individual intentions and to allow for the attribution of shared intentions to the group of artists. Doing so would enable MAI to take into account the features that are distinctive of collaboratively produced art while maintaining their emphasis on the role of intention. Of course MAI will be faced with the task of figuring out if the complex of attitudes identified by Bratman as shared intention are present during the collaboration but this is no more difficult a task than the one MAI faces in identifying the intentions of individuals in a case of art that is produced by a single artist. Interviews, journals, and so on, will provide the evidence that such a complex is present as these sources provide information about the intentions of individual artists. So, there is no, in principle, reason why MAI couldn’t be modified to account for collectively produced art. What about HI? Like MAI, we can easily modify HI to construct hypotheses about the intentions of the agents that take into account these interrelations and interactions among the individual artists of the group. Moreover, there is no reason why we cannot hypothesize shared intentions required in the group case in the exact same way as Levinson, for example, suggests we hypothesize individual intentions for individual cases: for any case where Bratman’s theory identifies the relevant interrelations among agents that support the existence of multiple authors, Levinson’s theory would tell us to construct the best hypothesis of that group’s inten-

36

Sondra Bacharach, Wellington/New Zealand and Deborah Tollefsen Memphis/TN

tions. So, in principle, there is no reason why HI couldn’t be modified to account for collectively produced art. There are forms of HI, however, that might not be able to be augmented by Bratmanian shared intentions. On at least one understanding of HI, we are not interested in hypothesizing the intentions of the actual artist, but rather in constructing a hypothesized author whose intentions can accommodate as many features of the work as possible. This is Nehamas’ version of HI (or NHI for short).30 One might be tempted to suggest that NHI can already account for the group cases. Why not simply postulate a single author and attribute intentions to this “fictional” author that provides the best interpretation of the work. In interpreting the work of Iri and Toshi, then, we would postulate an artist, perhaps “Iroshi”, and attribute intentions to this hypothesized agent. But postulating the intentions of a single author is a move that seems to miss something essential about collectively made artworks – viz., that they are made by many hands. The artwork is produced by individuals whose intentions make essential reference to the other participant(s). In many cases interpretation of the work requires that we appreciate that it is made by individuals working together. Indeed, treating the outcome of a collective process as if it were the result of a single artist runs the risk of excluding important and potentially insightful interpretations. Return to the Maruki case: if NHI postulates an imaginary author (perhaps a sort of melding of the two people together), and attributes an intention to this imaginary author, then this would rule out alternative interpretations which would depend for their plausibility on there being two distinct individuals who contributed to the work. For example, an interpretation according to which the authors were actually trying to capture something about martial strife in their disagreements about their work.31 Correctly interpreting the work of art as metaphorical for marital strife and reconciliation, for instance, would involve interpreting the work as a product of a back-and-forth between the two artists, in the same way that marital disagreements and reconciliations also involve a back-and-forth between the partners. Such an interpretation necessarily requires the existence of two artists – and hence would be unavailable to the NHI. Similarly, “Network of N_odes” also resists being properly interpreted by NHI. Recall that this work directly addresses problems of authorship raised 30 Cf. Nehamas (1981; 1987). 31 We do not know of anyone who raises this interpretation. However, it is a possible interpretation, consistent both with the works, and with the manner in which the Marukis produced their art.

Collaborative Art and Collective Intention

37

by the culture of appropriation in the Indian computer environment. It does so by itself appropriating and borrowing words, images and texts from the computer environment in the work itself. An essential quality of this work is not simply that it is produced by a number of artists working together, but also that others’ words, images and texts were incorporated into the work. “Network of N_odes”, then, would and could not be the work that it is if its authors were amalgamated into any single hypothetical or constructed author. Likewise, the inherently collaborative nature of George and Gilbert’s art also seems to be lost when one posits a single artist. The interpretation of their painting “They” would be radically different and, we would argue, inadequate if one failed to acknowledge that it was the result of a collaboration. This suggests that at least some forms of HI cannot accommodate collaboratively produced artworks. If we are correct, intentionalism needs to broaden its understanding of intention in order to acknowledge the role of collective intentions in interpreting collaboratively produced art. This doesn’t mean that the intentions of individuals will be ignored. But interpretations of collaborative art that make use of collective intentions will go beyond ascribing attitudes to individuals and will also make ascriptions of intention to groups, ascriptions whose truth conditions depend upon the complex of attitudes identified by Bratman being present within the group. Our attempt to show that collaborative art requires for its interpretation collective intentions has consequences for debates in action theory over the nature of joint or collective agency. This debate has generally focused on providing an analysis of joint action such that one reveals the conditions under which it would be appropriate to say “We did X.” Many have argued that collective attitudes of some form (either shared goals, collective intentions, joint commitments) are needed in order to capture the nature of joint agency and distinguish it from other forms of social interaction. The discussion, however, has focused on examples of joint agency that are simplistic and contrived. There have been no attempts to take a concrete case of joint agency and unpack it, so to speak, to get at its intentional and conceptual structure. Nor has there been any extensive discussion of the explanatory power of collective intentions. We think collaborative art provides a wealth of concrete cases of joint agency, which can be used as test cases by which to judge theories of collective intentionality and agency. Insofar as Bratman’s account of shared intention gives us a mechanism for identifying whose intentions are relevant to interpretation

38

Sondra Bacharach, Wellington/New Zealand and Deborah Tollefsen Memphis/TN

and a way of accommodating a variety of social aspects crucial to the interpretation of collaboratively produced art, his theory fairs well.

6. Conclusion We have seen how collectively produced art suggests that we need a more social and less individualistic account of interpretation. On our view, appealing to Bratman’s account of shared intention provides the basis for determining whose intentions are relevant to interpretation and further provides a way of capturing the ways in which intentions shape and inform the art work. We suggested that most current forms of intentionalism can accommodate collaboratively produced art work by appealing to shared intentions where this refers to the complex of attitudes identified by Bratman’s analysis. Of course, more work on how to interpret collectively made works is still needed. We have offered Bratman’s account of collective intention to accommodate the instances considered here, but there are other accounts on offer32 and it could be that others are better at capturing some aspects of collectively made artwork. For example, works of art produced in institutional settings, like those of the Bauhaus, may be better understood as resulting from something like corporate intentions, where this refers to the way in which groups can literally be the bearer of intentional states. This paper highlights that current forms of intentionalism need to broaden their understanding of intention and we hope that our proposal goes some measure towards understanding and interpreting collaboratively produced art. We also hope that collaborative art can serve as a concrete case by which to assess theories of joint action and collective intentionality. We believe that reflection on actual cases of collaborative art will result in a richer understanding of the phenomenon of joint activity.

References Bratman (1993): Bratman, M. Shared Intention, in: Ethics (1993) 104: 97113. Bratman (1995): Bratman, M. Shared Cooperative Activity, in: The Philosophical Review (1995) 101: 327-341. Carroll (1992): Carroll, N. Art, Intention and Conversation, in: Iseminger, G. (ed), Intention and Interpretation, Philadelphia: Temple University Press, 1992, pp. 97-131. 32 Cf. Tollefsen (2002b), Gilbert (1989), and Tuomela (1995).

Collaborative Art and Collective Intention

39

Carroll (1994): Carroll, N. Identifying Art, in: Yanal, R. (ed.), Institutions of Art, Pennsylvania State University Press: University Park, 1994, pp. 3-38. Carroll (2000): Carroll, N. Interpretation and Intention: The Debate Between Hypothetical and Actual Intentionalism, in: Metaphilosophy (2000) 31, 1/2: 75-95. Davies (1982): Davies, S. The Aesthetic Relevance of Authors’ and Painters’ Intentions,in: Journal of Aesthetics and Art Criticism (1982) 41: 6576. Dower and Junkerman (1985): Dower, J. and J. Junkerman. The Hiroshima Murals: The Art of Iri Maruki and Toshi Maruki, New York: Harper & Row Publishers, 1985. Gilbert (1989): Gilbert, M. On Social Facts, New York: Routledge, 1989. Gilbert (1996): Gilbert, M. Living Together, Lanham, Maryland: Rowman and Littlefield, 1996. Gilbert (2003): Gilbert, M. The Structure of the Social Atom: Joint Commitment as the Foundation of Human Social Behavior, in: Schmitt, F. (ed.) Socializing Metaphysics, Lanham, Maryland: Rowman & Littlefield, 2003, pp. 39-64. Green (2004): Green, Ch. (ed.). The Third Hand: Collaboration in Art from Conceptualism to Postmodernism, Minneapolis: University of Minnesota Press, 2001. Hirsch (1976): Hirsch, E. D., Jr. The Aims of Interpretation, Chicago, Illinois: University of Chicago Press, 1976. Hirsch (1967): Hirsch, E. D., Jr. Validity in Interpretation, New Haven, Connecticut: Yale University Press, 1967. Iseminger (1992): Iseminger, G. Intention and Interpretation, Philadelphia, Pennsylvania: Temple University Press, 1992. Levinson (1996): Levinson, J. The Pleasures of Aesthetics: Philosophical Essays. Ithaca, NY: Cornell University Press, 1996. Livingston (2005): Livingston, P. Art and Intention, Oxford, UK: Oxford University Press 2005. Nehamas (1981): Nehamas, A. The Postulated Author: Critical Monism as a Regulative Ideal, in: Critical Inquiry (1981) 8: 133-149. Nehamas (1986): Nehamas, A. What an Author Is, in: Journal of Philosophy (1986) 83: 685-691.

40

Sondra Bacharach, Wellington/New Zealand and Deborah Tollefsen Memphis/TN

Searle (1990): Searle, J. Collective Intentions and Actions, in: P. Cohen, J. Morgan and M.E. Pollack (eds.), Intentions in Communication, Cambridge, MA: Bradford Books, MIT Press, 1990. Searle (1995): Searle, J. The Construction of Social Reality, New York, NY: The Free Press, 1995. Stalnaker (1996): Stalnaker, N. Intention and Interpretation: Manet’s Luncheon in the Studio, in: Journal of Aesthetics and Art Criticism (1996) 54, 2: 121-134. Stecker (2005): Stecker, R. Aesthetics and the Philosophy of Art: An Introduction, New York, NY: Rowman & Littlefield, 2005. Stecker (2003): Stecker, R. Interpretation and Construction: Art, Speech, and the Law, Malden, MA: Blackwell Publishers, 2003. Tolhurst (1979): Tolhurst, W. On What a Text Is and How It Means, in: British Journal of Aesthetics (1979) 19: 3-14 Tollefsen (2002a): Tollefsen, D. Organizations as True Believers, in: Journal of Social Philosophy (2002) 33 (3): 395-411. Tollefsen (2002b): Tollefsen, D. Collective Intentionality and the Social Sciences, in: Philosophy of the Social Sciences (2002) 32 (1): 25-50. Tuomela (2006): Tuomela, R. Joint Intention, We-Mode and I-Mode, in: Midwest Studies in Philosophy (2006) XXX: 35-57. Tuomela (1995): Tuomela, R. The Importance of Us, Stanford, CA: Stanford University Press, 1995.

Winks, Sighs and Smiles? Joint Attention, Common Knowledge and Ephemeral Groups Clotilde Calabi, Milano As is well known, Searle explains collective intentional behaviour in terms of we-intentions plus some very general and pervasive Background skills.1 The two basic skills he has in mind are the capacity to recognize that other people are importantly like us in a way that waterfalls, trees and stones are not, and what he calls “a sense of the other”. Concerning this latter skill, he remarks that The collective behaviour certainly augments the sense of others as cooperative agents, but that sense can exist without any collective intentionality, and what is more interesting collective intentionality seems to presuppose some level of sense of community before it can ever function.2 Not all social groups are engaged in goal-directed behaviour all the time: people sitting around in the living room may form a social group and they do form such a group if they enjoy some kind of communal awareness: “Some of the time are just, for instance, sitting around in living rooms, hanging out in bars, or riding on the train. Now the form of collectivity that exists in such cases isn’t constituted by goal directed intentionality, because there isn’t any. … Nonetheless have the type of communal awareness that is the general precondition for collective intentionality”.3 In this paper I argue that Searle’s communal awareness has the structure of joint attention, and I discuss and assess different accounts of it. I start with three examples and two intuitions.

1 Many thanks to Hans Bernhard Schmid: his remarks have considerably helped me to shape the final version of this paper. Of course, I am the only responsible for what I have written. I am also grateful to Paolo Casalegno for comments and to Marco Santambrogio for comments and for the many discussions on joint attention we had. 2 (Searle 2002: 103). 3 (Searle 2002: 104). N. Psarros, H.-B. Schmid and K. Schulte-Ostermann, Concepts of Sharedness, Frankfurt 2008: 41-58; ontos verlag

42

Clotilde Calabi

1. Three examples and two intuitions 1) A is looking at the moon and soliloquising. Unnoticed by her, B sees that A is talking to the moon and looks at the moon too. The reference of A and B’s mental states is the same (they are paying visual attention to the same object and share the same demonstrative thought), but although B is aware that A is paying attention to the moon, A is not aware of B’s awareness. 2) A man and a woman are strolling and suddenly realize that it is raining. Their attention is captured by the rain and each of them notices that the other notices that it is raining. They share the same perceptual experience and are aware of their states of attention. That is, A is aware of B’s awareness and B is aware of A’s awareness. They do not exchange glances, they do not smile nor show their disappointment. Each of them just looks at the rain and is aware that the other is doing the same. 3) Two parents are looking at their daughter playing, they exchange glances and smile. Let me concentrate on the third case. Although my description of it is rather sparse, this is the most complex situation. Each of the two parents is aware that the other is looking at the girl and aware that the other is aware of his/her own state of awareness. Moreover, each of them is willing to make the other aware of his state of awareness and does so by a glance and a smile. Thus, if we zoom in onto the scene, we notice that: a) Each parent attends to the same object (they both simultaneously see that the girl is playing). b) They coordinate their visual attention. More precisely, each of them, while attending to the girl, is also monitoring the direction of the other’s visual attention. c) Each of them is aware of the state of awareness of the other. d) Each of them intentionally makes the other participant to his/her state of awareness and does so by a glance and a smile. In this example glance and smile are not an expression of delight only. They may be that, too, but they also indicate acknowledgement that one is aware of the other’s state of awareness. In fact, glance and smile are one way to manifest recognition (i.e. to acknowledge) that there is mutual awareness. Let me call condition (c) the mutual awareness condition and (d) the mutual acknowledgement condition.4 The three situations are ob4 In using the expression “mutual acknowledgement” I draw from Gilbert (2007), where a rich and interesting analysis of mutual acknowledgement and mutual

Winks, Sighs and Smiles? Joint Attention, Common Knowledge and Ephemeral Groups

43

viously different. In the moon situation no coordination of vision occurs and there is neither a mutual awareness nor a fortiori acknowledgement of mutual awareness. In the promenade example, there is mutual awareness only, and neither coordination of vision nor acknowledgement that there is mutual awareness. My question is which among the three cases is and which is not an example of joint attention. The first case is clearly not an example of joint attention while the third case is. What about the second case? Are the man and the woman jointly attending to the rain also a case of joint attention? What are the necessary and sufficient conditions a group of people should satisfy for being in a state of joint attention? This is the problem I want to address. In my analysis I concentrate on cases of joint perceptual attention in which two people are involved. Moreover, I consider only those cases in which attention to the scene is essentially visual. 5 The two intuitions concern the nature of joint attention (from hereon JA). The first is that when you are involved in a JA situation, you grasp this in the blink of an eye. This means that you realize immediately or instantaneously both that you and your companion are looking at the same object and that you are both in a state of mutual awareness. The second intuition is that in a JA situation everything is in the open: when two people enjoy an experience of joint attention, the fact that they are mutually aware that they are attending to the same object or state of affairs and they are immediately aware that it is so is “out in the open”. Philosophers interested in JA have tried to accommodate those two intuitions by taking two different routes. Some philosophers (most notably Lewis, Schiffer and, with some relevant differences, Margaret Gilbert, as well as researchers in game theory) have identified JA with a special type of common knowledge involving the possession and exercise of inferential capacities. Other philosophers (namely, Searle, Campbell and Peacocke) consider it a special kind of immediate perceptual knowledge not involving inferences. Following Campbell, I call the former analysis a reductive analysis of JA and the latter a non reductive analysis. According to the reductive analysis, it is possible to describe a JA experience by saying which individualistic states of its participants are involved in their experience, without already implying that there is JA among them. In contrast to this type of analysis, recognition are contained. 5 There is no limit to the number of people that may be involved in a joint-attentional situation. I have no objections to the idea that hundreds of people at a rock concert are paying joint attention to the music. Notice that in this case joint attention may not be visual attention (only), it might be auditory attention, as well.

44

Clotilde Calabi

in the non-reductive analysis, JA is a primitive. In fact, the description of the mental states of each participant to a JA situation already entails that there is someone else with whom that participant is jointly attending.6 I proceed as follows. In § 2, I give an intuitive characterization of JA and present the inferentialist account of it. In § 3, I present some objections raised by Peacocke against the inferentialist account and I reply to those objections. In § 4, I present a new version of the reductive inferentialist account of JA. Finally, in § 5, after criticising the new version, I conclude that JA may be a primitive after all.

2. The openness requirement and the inferentialist account Let me go back to the two intuitions of immediacy and openness. Any account of JA should make room for them. The requirement concerning openness is more subtle than the requirement concerning immediacy. In fact, it seems to me that three different ideas lie beneath the claim that a certain state of affairs S is out in the open. The first is the idea that S is out in the open for two subjects A and B only if it does not belong solely to their own private world. The second is the idea that S is out in the open if and only if there is nothing hidden in it, that is, if and only if S is right there, entirely present to the consciousness of A and B. The third is that S is out in the open if and only if it is right there and available for inspection by A and B. These are different concepts of openness and each of them may apply to certain things, but not to others (of course this does not exclude there being things or facts to which all three concepts apply).7 Consider the case of beliefs: different people can have the same beliefs and hence, in some sense, they share them. Thus, these beliefs do not belong to their own private world, and yet they are not right there, available for inspection. In fact, we can only find out what people’s beliefs are on the basis of what they say and how they behave. For this reason, beliefs are out in the open in the first, but not in the second sense. Moreover, if beliefs are dispositions towards making assertions, they are not entirely present to consciousness and hence do not fall under the second concept of openness. Pain has completely different features. It is right there, entirely present to consciousness, but there may be some eyebrow raising if we say that 6 Campbell (2005: 288). 7 For example the objects of demonstrative reference satisfy all three concepts: they do not belong to one’s private world, they are right there, and in some sense they are both entirely present to consciousness and available for inspection.

Winks, Sighs and Smiles? Joint Attention, Common Knowledge and Ephemeral Groups

45

it can be shared: some people will readily object that each person has his own pain. Finally, is the quale of pain available for inspection as something distinct from my feeling it? Difficult to say. Consider now JA. Which concept(s) of openness is/are involved here? Let me make my question more specific. What exactly is out in the open for the joint attenders? I shall start by presenting the inferentialist account of JA and the concept of openness it involves. The inferentialist account stresses that JA is awareness of a very complex state of affairs. To explain in what sense this is so, let me elaborate further on the promenade case. The man is not merely paying perceptual attention to the fact that it is raining and aware that the woman is paying perceptual attention to that fact, too (and the same holds for the woman): each of them is also aware that the other is aware that he/she is paying perceptual attention to the fact that it is raining. Yet, this is not the end of the story: if A and B jointly attend to the fact that it is raining and thus A is aware that B is attending to the fact that it is raining and aware of B’s awareness, and B is aware that A is attending to the fact that it is raining and aware of A’s awareness, then A is aware that B is aware that A is aware that B is aware that A is attending to the fact that it is raining. This can go on forever. In other words, a JA situation is such that if its participants have a first-order act of attention or are in a state of perceptual awareness directed to a state of affairs that p, then they have an n-order awareness that p (for all n).8 Such described, JA gives rise to the following puzzle. Two people jointly attending to some p are in a state that involves an infinite number of iterations of “is aware that” clauses. If the joint attenders are in a state with such infinite content, we cannot reconcile this feature of its content with the immediacy requirement, that is, with the fact that one is immediately aware of being in a JA situation. We cannot do so, because we all suffer from some limitations. How could the man and the woman be immediately aware of jointly attending to the fact that it is raining and, at the same time, each of them be aware of a state of affairs whose description is infinitely complex? Normal people simply cannot grasp infinitely complex content all at once. The inferentialist answers this question by considering JA a state that depends upon reasoning. In the promenade example, according to the standard inferential interpretation, the man and the woman will 8 A 1st order awareness of some state of affairs p is to pay attention to p or to be perceptually aware that p. An (n + 1) order awareness that p (for n > 1 or n = 1) is an awareness of someone else’s n-order awareness that p.

46

Clotilde Calabi

reason as follows. I see that it is raining and I see my companion. I am staring at the sky and have my eyes wide open. I know (in some way) that anyone in my condition with his eyes wide open, will see that it is raining, and that anyone in my condition knows that anyone in these conditions, with his eyes wide open, will see that it is raining. My companion has his eyes wide open and is in these conditions. Thus, he sees that it is raining and he sees me and knows (in some way) that anyone in these conditions and with his eyes wide open and staring at the sky sees that it is raining. And he knows that I know, and so on. In this manner, two people jointly attending to some state of affairs that p are such that they can make an inference from their first-order act of attending to p, to an n-order act of attending to p, for all n. Each step of the inference will be based upon the recognition that we have similar cognitive and perceptual capacities (such as the capacity to see well, identify what we are seeing, and the capacity to draw inferences). One may urge that the immediate knowledge that seemingly characterizes our participating in a JA situation, is by definition the contrary of inferential knowledge. The inferentialist can readily reply that “immediate” and “inferential” are not opposite terms, once “inferential” is dispositionally or counterfactually understood: the participants in a JA situation can become aware that p for all orders of awareness only if solicited. Thus, they have the logical capacity to reach all levels of awareness, without being obliged to engage in never-ending reasoning. They are immediately aware of a specific situation that would enable them (given their logical competence) to reach higher and higher levels of awareness. For example, in the parents’ case, if the mother is aware that the father is aware that the mother is paying attention to the girl playing, then the father, if asked would be ready to answer that he is aware that the mother is aware that he is aware that the mother is paying attention to the girl playing. If we analyse a JA experience along these lines, we are neither forced to impose on its participants any reasoning of infinite length, nor to attribute to them a mental state with content of an infinite nature. The suggestion is to consider the participants in a JA situation as immediately aware of the iterability of their awareness. This would accommodate the immediacy requirement. However, this solution to the puzzle does not come without a price, for it would limit JA to individuals who possess the concept of infinite iterability. Children who do not posses that concept, strictly speaking cannot participate in a JA situation. This is implausible.9

9 Thanks to Hans Bernhard Schmid for raising this objection.

Winks, Sighs and Smiles? Joint Attention, Common Knowledge and Ephemeral Groups

47

It should be emphasized, too, that the promenade example and more generally all JA situations have the following peculiarity: unlike many other common knowledge cases, they are situations of co-presence in which the participants have simultaneous perception of some state of affairs. Yet, for the inferentialist JA is exactly like standard common knowledge cases in other respects: the participants see themselves as enjoying some relevant property (here, it is the property of being normally alike) and knowing they have that property would allow them to go on forever, if only they were allotted a better memory and a longer life. For the man and the woman, knowing that they have that property allows them to see that a certain fact (that they are both seeing that it is raining and the other person) is out in the open. They can see that a certain fact is out in the open if they see themselves as similar under some relevant respects given that specific context. Thus, in the standard inferentialist analysis the belief that the other participant possesses the property of being normal (or normally alike), co-presence and simultaneous perception of S, are necessary and jointly sufficient conditions for JA experiences. In the end, the standard inferentialist account of JA identifies JA with common knowledge based on perception. The concept of openness involved here needs some more explication. It seems to me that in the inferentialist account, “openness” is understood in the first and the third sense, not in the second: “out in the open” means not belonging to one’s private world and right there, available for inspection. Peacocke has a nice way of capturing this type of openness. He calls it “mutual open-ended perceptual availability” and defines it as follows: A state of affairs is mutually, open-endedly available to two subjects “if the obtaining of the state of affairs and the operation of perceptual and attentional mechanisms in the two subjects bring it about that one of them perceives that it obtains, or perceives that the other perceives that it obtains, or perceives…, then the state of affairs (thus brought about) of his so perceiving is available to the other to perceive”.10 Now, Peacocke while accepting the idea that JA’s openness involves mutual open-ended perceptual availability raises some important objections the inferentialist account of JA. He thinks that it is flawed for three different reasons to which I turn now.

10 Peacocke (2005: 302).

48

Clotilde Calabi

3. Peacocke’s objections The first objection is not strictly connected with the issue of openness. It is a more general attack against the inferentialist strategy. The second and third one are more relevant to the matter at the centre of my concerns. The objection from lacking a concept of normality The inferentialist analysis requires that participants in a JA situation possess the concept of normality. Individuals, who lack the concept of normality, thus having no beliefs about the capacities of normal people, lack the kind of basic knowledge that is required for the inferentialist to pursue the inference. Yet they can still participate in a JA situation and perceive it as such. The inferentialist can readily reply: to possess the concept of normality is just to know how to draw a distinction between individuals who are like oneself under some relevant respects and individuals who are not (having one’s back turned to an object or having one’s eyes closed are not normal conditions for seeing it). Of course, it is one thing to have the ability to distinguish between those who are normally like ourselves in a certain situation and those who are not, and another to have a whole range of beliefs about what it means to be normal. One may have the capacity to recognize when someone else is similar to oneself under some relevant respects, without being able to make this explicit by pointing out some criteria for similarity. Yet, to have such a capacity for recognition would be enough to engage in a JA situation or have a JA experience.11. Of course, everything is pretty much similar to anything else in some respect. Yet context matters and it helps: our immediate grasp of the context helps us to find out whether there is some relevant similarity to go by in order to draw the inference and then things can proceed smoothly. The counterfactual objection In the inferentialist interpretation openness is dispositional. But the openness of JA is not something merely dispositional or counterfactual. JA is like perception and perception is not dispositional.12 I agree that there is an analogy between JA and perception, but I do not think that the analogy shows that the inferentialist interpretation is wrongheaded. In fact, it seems to me that the objection hinges on some ambi11 Interestingly enough, having this capacity would be compatible with lack of other concepts - with the exception, perhaps, of the concept of infinite iterability. 12 In fact, one may even make the stronger claim that JA is a special type of perception.

Winks, Sighs and Smiles? Joint Attention, Common Knowledge and Ephemeral Groups

49

guity concerning the second meaning of “out in the open”. Let me pause here. As I said, according to it, S is out in the open if nothing is hidden in it, that is, if S is right there, entirely present to the consciousness. Suppose that I am seeing a car in front of me. Is that car entirely present to my consciousness? If I perceive the car in front of me, I do perceive it in the blink of an eye. Thus, in a way, one answer to that question is: yes, if this means that I am not perceiving a detached portion of the car. Yet to say that I perceive the car entirely in this sense is perfectly compatible with my seeing parts of it only: I can see only its front but not its back. Thus, in another sense, the answer is: no, I cannot perceive the car entirely.13 Of course, to perceive an object is to be able to see whatever surfaces it has, if we move around it and this is perfectly compatible with its objects being entirely and instantaneously present to the consciousness of the perceivers. Let me turn now to JA and consider once again A and B noticing that it is raining. My former description of the situation can be further developed as follows: each of the participants is aware that the other is noticing that it is raining and aware of his/her state of awareness. That is, they both know that the fact that each of them is aware of the state of awareness of his companion is mutually and open-endedly available to them. The following conditions are satisfied: 1) They know that each of them is looking at the rain 2) They know that each of them is aware of the other’s awareness 3) They know that each of them knows that (1) - (3) are true. As I have already remarked, for the inferentialist to assert (1) - (3) is not to assert that A and B are involved in reasoning with an infinite number of steps or that they grasp a proposition with infinite content. To assert (1) (3) is to assert that given any level of awareness A and B have, and given their capacity for reasoning, they could ascend to the next level of awareness. There is a sense in which they can see immediately that it all is out in the open, that is, that (1) - (3) is true, and yet there is another sense in which, if they want to make the “sense of openness” they experience explicit, they have to go through the sort of reasoning I outlined above. Notice further that in my first description of this example I decided to leave it vague whether A knows or thinks that B is aware that A is aware of B’s awareness. Clause (3) is what allows the participants to ascend to higher order levels. Thus, the inferentialist has no reason to reject the analogy between JA and perception: for both JA and perception there is a sense in 13 Consider further that if there is a modal completion, the visual system brings it about precisely because parts of the object are hidden from view because occluded by other objects.

50

Clotilde Calabi

which the scene is entirely present to consciousness and a sense in which it is not. But this is no objection to her approach. I venture to say that this is the reason why the inferentialist concentrates on the first and the third meaning of “out in the open”. The concept of openness objection How would normal people come to know that normal people see things in front of them, that normal people know this? How do they know that they know? And so forth. According to Peacocke, “this knowledge is attained by generalization from experience with particular situations in which one sees that someone is seeing something and in which this is wholly open” (p. 311, my emphasis). Thus, experience of openness of certain situations and the knowledge it brings about is rationally prior to knowledge of generalizations about normal persons. The idea is the following: the criterion for establishing that a subject possesses the concept of openness is that he is capable of recognizing that something is out in the open when it is so. But if the subject should already possess and make use of that concept in order for something to appear out in the open to him, then to specify the condition for possessing the concept of openness would require possession of that very concept in the explanation and the explanation would be circular. If only those who have the concept of openness can have an experience of openness, an explanation of the conditions of possession of that concept stating that the subject should be able to use it when he has an experience of openness is circular. It presupposes possession of that very concept. According to Peacocke, for this reason the experience of openness is rationally prior to the possession of the concept of openness. The third objection has an important role in Peacocke’s overall strategy: since common perceptual belief presupposes the experience of openness and according to him joint attention is such an experience, he urges that common perceptual belief presupposes joint attention, rather than being a type of joint attention. The objection is strong indeed. Yet, Peacocke formulates it in a way that doesn’t defeat the inferentialist’s reasoning and he shouldn’t really feel worried. The inferentialist can readily acknowledge that one needs to presuppose an experience of openness without prior possession of the concept of openness, in order to specify the conditions of possession of that concept. Yet, he can counter that, in order to have the experience of openness, the subject needs to have at least the concept

Winks, Sighs and Smiles? Joint Attention, Common Knowledge and Ephemeral Groups

51

of normality.14 Thus, the concept of normality is logically prior to the concept of openness: its possession makes the experience of openness possible and hence makes it possible to explain what it means to possess the concept of openness. Notice that the lack of concept of normality objection cannot be raised with respect to the experience of normality, since the inferentialist is not claiming that experience of normality necessarily presupposes the possession of the concept of normality. One can acquire the concept of normality after exposure to the experience of normality and then apply that concept in order to have the experience of openness. At this point the inferentialist can still conclude that the kind of grasp of openness he has in mind is available to everybody. Margaret Gilbert, who endorses a particular version of the inferentialist account, expresses this idea very clearly: she says that this openness “is the kind I presume normal human beings to have, willy-nilly. If I am right, everyone possesses this concept, though most could not define it explicitly with any ease. I take such possession to go, generally, with the ability to apply the concept without hesitation when one is in a situation of openness”.15 My first conclusion is that none of the above objections against standard inferentialism is conclusive. Still I think that we cannot rest content with the idea that JA is just a case of common knowledge, based on a situation of co-presence and simultaneous perception: if this were a definition of JA, it would make it unavailable to children and this does not seem correct. Thus, either we provide an alternative definition of JA, or – following Searle and Campbell – we consider it a primitive. We shall see.

4. The phenomenological glue I shall now consider whether an alternative definition of JA can be provided. A slight modification of the example involving the parents sets the scene. Once again, the parents watch the girl playing, but this time there are no smiles. Suppose further that there is a glass panel between them and that each of them falsely believes that this glass is a one-way mirror, allowing him/her to see the other, but preventing the other from seeing him. They both perceive a state of affairs that is mutually and open-endedly available to them, without being aware of this availability. The par14 It should also be remembered that if someone has the concept of openness and p is open to him and to other people, then (other things being equal), he knows that it is. That is, he will realize, or perceive or notice the openness. One cannot possess the concept of openness, and not recognize an openness situation, if it occurs. 15 M. Gilbert (1989: 468).

52

Clotilde Calabi

ents lack mutual awareness and hence are not jointly attending to the girl. At some point, someone walks in, and tells each of them that the other is sharing his/her experience, too.16 Does this additional information make the man and the woman shift from an experience of convergent attention (attention directed to the same object) to an experience of joint attention? Let me make my question more specific. Their perception of the situation does not change, since they still perceive the state of affairs that the girl is playing. In other words, the representational content of their perceptual experience is exactly the same it was before acquiring the new information. What does the additional information change? Of course, their beliefs change, since they acquire a new belief about the state of affairs they simultaneously perceive. The inferentialist will say that when the informant walks in, the new piece of information he provides, supplemented with the recognitional belief already possessed by the parents that they are similar under relevant respects, allows them to have a JA experience: it obliges them to see that a certain fact is out in the open, namely the fact that each of them has common knowledge that each of them is personally attending to the girl. It could be objected that mere common knowledge of the new piece of information does not make the parents jointly attend to the girl (as I just said, it only gives them the common knowledge that each of them is personally attending to the girl). I think that the example deserves some consideration because upon closer inspection it reveals an aspect of JA experiences that is often overlooked. I call this aspect the “phenomenological glue”. Interestingly enough, both the inferentialist (Schiffer is a typical case) and Peacocke who criticises him, analyse “A and B jointly attend to p” as “(A is aware of the state of awareness of B) & (B is aware of the state of awareness of A)” In other words, they both analyse a joint attention state as an aggregation of individual awareness states, thus presenting an individualistic account of JA. If one follows the individualistic account, one overlooks the special tie that links persons in joint attention. Now, to be fair to the inferentialist, she is not insensitive to this aspect of JA experiences. In fact, she will come up with a new definition of JA that 16 I have taken the idea of the glass panel from Peacocke and I have added to it the detail that the informant tells the parents what is going on.

Winks, Sighs and Smiles? Joint Attention, Common Knowledge and Ephemeral Groups

53

might fit. Let me elaborate on the example to present her new definition. Imagine that after receiving the information, the two parents look at each other and nod, or make any gesture that unequivocally makes the other aware that he has received the information and that he is intentionally expressing this fact. A new fact has occurred and the situation has changed: each of them has volunteered to publicly express the acknowledgement that he has received that piece of information. The inferentialist remarks that at this point the parents have common knowledge of a new fact, namely that each of them has intentionally and openly expressed his willingness to accept that information. More precisely, the parents have openly expressed their willingness to be members of a group whose essence is to share in the belief that the information is in the open: they have individually acknowledged: “we share that piece of information”. She will claim the parents’ situation is a JA situation precisely because each of them publicly expresses acknowledgement to the other that he/she has acquired that information and thus they have both acquired common knowledge of their mutual acknowledgement. It is when this particular type of common knowledge arises that their state of awareness is no longer a state of I-awareness: it becomes a state of we-awareness. In doing so, the inferentialist has come up with a new definition of JA: it is a special type of common knowledge, namely common knowledge of a mutually and simultaneously expressed attitude of a specific kind based on co-presence.17 Does the more sophisticated version of the inferentialist analysis satisfy all our theoretical needs? If the correct answer is yes, then we have a theory of common knowledge that is an extension of Lewis and Schiffer’s theory and explains the phenomenon of JA. If the correct answer is no, then JA may be a primitive, after all. Again, we shall see.

5. Is this the correct theory of JA? I need to take stock. I have argued that to claim that JA is a case of common knowledge (plus co-presence and simultaneity) is to defend the idea that JA is an aggregate of individual states. This leaves out jointness. Then I have considered an alternative account of JA according to which a sufficient condition for its occurrence is that the participants must show each other their willingness to share in a belief, a perception, or simply the acknowledgement of the other’s presence. That is, each member of the group must express willingness to share in something with the other 17 This alternative definition of JA is in fact advanced by Margaret Gilbert and I was myself at some point convinced that it was the correct one.

54

Clotilde Calabi

members. Unless there is such acknowledgement of mutual recognition, noting similar properties, physical proximity and perceptual awareness of each other’s presence, although contributing to the emergence of a JA experience, would not be sufficient for its occurrence. According to this line of argument, any account of JA that reduces it to a mere example of common knowledge based on perception is mistaken, not because JA is an immediate experience of a state of affairs, as Peacocke contends, but rather because it provides only the necessary conditions for JA, but not sufficient conditions for it. In other words, although participants in a standard common knowledge situation take their awareness of the situation in some sense away from their private world and put it out in the open, it is still their own awareness. It becomes a collective affair only if it is based upon a special type of common knowledge, namely common knowledge of a mutually expressed attitude of a specific kind. Mere perceptual common knowledge is one thing, JA as a we-state is another, and the former is a necessary, but not sufficient condition for the latter.18 The philosophers who concentrate on immediacy and the philosophers who focus on mere perceptual common knowledge forget jointness. They make the same mistake. As I have already said, the inferentialist can make an even stronger claim. She can say that the special common knowledge I have described does not merely constitute the necessary and sufficient condition for the occurrence of JA: JA as a we-state can be defined in terms of this special type of common knowledge. The inferentialist will further contend that: (1) this special type of common knowledge is not in contrast with the immediacy requirement; (2) her new account sheds light not only on the case of the parents, but on many other common situations as well. Concerning (1), she will refer once again to her reasons for holding that immediate and inferential are not opposite terms. Let me consider now (2). Unless otherwise specified, the two parents looking at the girl are in a situation similar to one in which two people are casually strolling past each other and have common knowledge of this fact. They are not justified in asserting “We are walking together”. If their stroll were, for example, the result of a decision collectively taken or the result of a tacit agreement to conform to certain rules, they would be walking together. Yet an explicit decision or a tacit agreement is not required to perform joint actions. In fact, this type of action can come into existence on a more fleeting basis, for example, it may occur by virtue of a mere public and voluntary expression of the acknowledgement of each other’s presence. In the case of the two walk18 Gilbert (1989: 185).

Winks, Sighs and Smiles? Joint Attention, Common Knowledge and Ephemeral Groups

55

ers, they would form an ephemeral group if, in addition to the common knowledge that they are walking past each other, they were publicly expressing an acknowledgement of this very fact that they are walking along the same path and they know they are.19 A sigh, a cough, a shiver, a smile or just a glance would be enough. In this particular case, the joint action could be grounded on any of these expressions of mutual acknowledgement of the others’ presence.20 Similarly, the parents would not be allowed to say “we see that p”, just because they have common knowledge of their seeing the girl. The “we” in “we see” is not justified just by virtue of common knowledge. What is also required is a special type of common knowledge, namely common knowledge of the expression of mutual acknowledgement of the presence of the other. Going back to the informant case, when he walks in and provides the relevant information, no joint attention occurs, unless there is some expression of mutual recognition of that information. When the information is provided, it produces common knowledge and therefore some facts are put out in the open. But that type of openness is not enough for JA. If there were also some expression of mutual recognition that the information has been provided (for example by nodding or smiling), that expression would produce the special common knowledge that grounds the existence of an ephemeral group. Only if each individual willingly showed acknowledgement of the information to the other, would they be jointly attending to the girl. Is this the right theory of JA?

6. More problems Three questions loom. Firstly, how do I know that we have common knowledge of mutually expressed acknowledgment? Would I need another wink, another smile or another sigh? Would public acknowledgement of the mutual recognition require another acknowledgement? One might try to say that public recognition is entirely out in the open: when it oc19 My example is modelled on an example discussed by Gilbert in Gilbert (1989: 202). 20 On mutual recognition of the other’s presence see Gilbert (2007) and particularly her discussion of the Merton Street Library case, previously analysed in Gilbert (1989: 9). She is sitting at a table of Merton Street Library with another person and at some point each of them recognizes the other’s presence: “At a certain point I looked up, looked somewhat fixedly at the person in question, until he looked up. I caught his eye …; we looked at each other . I nodded and smiled briefly; he did also. We then returned to our respective concerns and had no further interaction”. She argues that she and her companion in that situation were jointly committed to recognize that they were co-present, as a group.

56

Clotilde Calabi

curs, it is right there, you get it in the blink of an eye and nothing else lies hidden. Although it is tempting to respond in this way, I am afraid this is hardly an answer to the infinite regress problem. Secondly, isn’t this theory too demanding? Recall the objection regarding children. If they have not yet acquired the capacity to participate in games of winking, sighing and smiling, let alone the capacity to use language to manifest their mutual acknowledgement of each other’s presence, they would be unable to enjoy experiences of joint attention. One could claim that children could enjoy only incipient joint attention, but not a full-blooded joint attention experience. Again, this seems implausible if we reflect on the fact that JA may be a condition for communication and hence cannot be explained in terms of communication (plus some common knowledge). Thirdly, given that error is always possible, the new theory allows for one type of mistake only: if I am mistaken in thinking I am having a JA experience, it can only be because I have a wrong belief. I thought I shared that special type of common knowledge with another person, but I did not: the world is not the way I thought it was. Now, as Searle points out, errors concerning collective actions are not only mistaken beliefs: if I am mistaken, my error concerns what I am doing, too. In the specific case of JA, I thought that I was jointly attending with you, but my heeding was not part of our heeding, that is, you were not co-attending with me at the scene. Thus my heeding was a different type of action than the one I thought it was. The three problems I have raised encourage us to think that JA may be a primitive, after all. This is Campbell’s view. He calls his own non-reductive account of JA “relational”. According to him, JA is an experience with two constituents that play different roles; one constituent is the object of attention, and the other constituent is the person one is jointly attending with. This latter is not another object of one’s attention along with the first object. Nevertheless, like the object of attention, it belongs to the conditions defining the experience content. The experience of JA would not be what it is (i.e. would not have the content it has) unless it contained reference within its content to a co-attending person. This is why any definition we attempt to give of JA that does not exclude children from participating in a JA situation turns out to be a circular definition, including the very notion of the co-attending person within itself. And this is why the notion of joint attention must be a primitive. I think that Campbell copes well with the three problems raised above: In his account infinite regress is avoided; children, who do not posses the concept of infinite iterability, can enjoy joint attention experiences; finally, he provides an answer to Searle’s question about error because if the other is necessarily a constituent of JA experiences and there is no

Winks, Sighs and Smiles? Joint Attention, Common Knowledge and Ephemeral Groups

57

other there, then the experience cannot be a JA experience: it must be a different experience.21 However, something more should be said concerning my story about the parents and the informant. The inferentialist says that no phenomenological glue ties the parents together in a JA state, after reception of the information, unless some nodding, winking or smiling occurs. Before the nodding and smiling, the parents have only acquired common knowledge of the fact that each of them is personally heeding the girl’s playing. She argues that in order to enjoy a JA experience, the persons involved should manifest either their intention to engage in a joint project (be it a joint investigation, moving toward the object, or a coordinated attack on it) or merely acknowledge each other’s presence with a sigh, a cough or a smile. This would allow them to gain special common knowledge of their intention to communicate their presence or their other intentions to each other. This is what JA amounts to. I am inclined to say instead that the winking and the smiling are an expression of JA, not a condition for its occurrence. More precisely, my hypothesis is that the winking and smiling coincide with the exercise of JA. JA is an action collectively taken and in this action the other actor enters as a constituent. Of course, we can carry on our actions in many different ways and we can jointly heed in many different ways: in some cases we smile, in other cases we cough or sigh. Finally, there are still other cases in which we simply form an ephemeral group in silence, as happens when we sit around in living rooms, hang out in bars, look through our books trying to study at the library, or when we ride on the train. In many cases there are no cough, no sighs, no smiles, nothing at all. We just attend to each other’s presence and do nothing else. The inferentialist says that in order to engage in JA it is necessary at least to acknowledge the other’s presence and manifest such acknowledgement as a basis for common knowledge. I think that we do not need even that. Of course, it may turn out that one thinks one is jointly attending to something with someone else but is wrong: it may turn out that he was not. Thus, when the informant walks in and gives each of the parents the same information, they may be jointly attending or they may not, depending on his reliability. If they are, they are performing a particular type of joint action; if they are not, they are performing a solitary act of heeding.

21 Campbell (2005: 289): “I took it that you were a constituent of my experience as a co-attender, when in fact you were not”.

58

Clotilde Calabi

References Campbell (2005): Campbell J. Joint Attention and Common Knowledge, in: Eilan N. and J. Roessler (eds.), Joint Attention, Communication and Other Minds, Oxford: Clarendon Press, 2005, pp. 287-297. Gilbert (1989): Gilbert M. On Social Facts, Princeton: Princeton University Press, 1989. Gilbert (2007): Gilbert M. Mutual Recognition, Common Knowledge and Joint Attention, in: Ronnow-Rassmussen, T., B. Petersson, J. Josefsson, and D. Egonsson (eds.), Hommage à Wlodeck. Philosophical Papers dedicated to Wlodek Rabinowicz, www.fil.lu.se/hommageawlodek (8.02.2008; 12:00 MET), 2007. Lewis (1969): Lewis D. Convention, Cambridge Mass.: Harvard University Press, 1969. Peacocke (2005): Peacocke C. Joint Attention: its Nature, Reflexivity and Relation to Common Knowledge, in: Eilan N. and J. Roessler (eds.), Joint Attention, Communication and Other Minds, Oxford: Clarendon Press, 2005, pp. 298-324. Schiffer (1972): Schiffer S. R. Meaning, Oxford: Clarendon Press, 1972. Searle (2002): Searle J. Collective Intentions and Actions, in: Searle J. Consciousness and Language, Cambridge: Cambridge University Press, Cambridge 2002: 90-105. Originally published in: Cohen P., Morgan J., Pollack M.E. (eds.) Intentions in communication, Cambridge Mass.: MIT Press, 1990. Tuomela (1995): Tuomela R. The Importance of Us: A Philosophical Study of Basic Social Notions, Stanford: Stanford University Press, 1995.

Shared Feelings Towards a Phenomenology of Collective Affective Intentionality Hans Bernhard Schmid, Basel Up until very recently, collective intentionality analysis has almost exclusively been concerned with shared intentions and – more recently – with shared beliefs. Next to no effort, however, has been devoted to the analysis of collective affective intentionality so far. This apparent research lacuna seems all the more surprising because the analysis of emotions has been among the key topics of international philosophical research for the last twenty years at least, with a plethora of conferences, monographs and collected volumes on the topic. Thus the question is: why is the emotional dimension so conspicuously absent from collective intentionality?

1. Affective intentionality: A matter of feelings One might first think that this lacuna simply reflects the general neglect of the affective in the wider history of the theory of intentionality. In the Anglo-Saxon context, intentionality has always primarily been regarded as a matter of conative (or practical) intentionality, whereas on the continental side, cognitive intentionality has usually served as the paradigmatic case. Of course, there are many exceptions to this rule on both sides of the “great divide” (among these Carl Stumpf, Martin Heidegger, Max Scheler, Gilbert Ryle). But these exceptions cannot distract from the rule: in neither of these two traditions of intentionality analysis, the affective has received an amount of attention that comes anywhere close to the one devoted to the cognitive and conative types of intentionality. Most philosophers of intentionality simply did not deem the emotional essential to the concept of intentionality. Thus it might indeed seem that the neglect of the affective dimension in the debate on collective intentionality is part of a more general anti-affective strand, which permeates most of the received philosophy of intentionality. Something of a mirror image of this can be seen in early conceptual analyses of the affective. Just as the affective has been neglected in the theory of intentionality, intentionality has been neglected in the theory of the afN. Psarros, H. B. Schmid and K. Schulte-Ostermann, Concepts of Sharedness, Frankfurt 2008: 59-86; ontos verlag

60

Hans Bernhard Schmid

fective in turn. According to what has come to be called “feeling theories” of the affective,1 the intentional is at best of minor conceptual importance to the affective. The essential feature of the affective is seen in the feelings qua states of bodily arousal, as in William James’ What is an Emotion.2 As a consequence, the question of the intentionality of emotional states is not seen as particularly important. In this view, a theory of the affective should be concerned with a taxonomy of feeling experiences, and with the analysis of the causal role of states of arousal, rather than with such ventures as intentional analysis. Thus the emerging picture is the following. There is a rift between an affectivity-free theory of intentionality, and an intentionality-free view of the affective. Let’s call this the intentionality/feeling dichotomy. At first sight, it might seem that much time must have passed since the days of affectivity-free intentionality and intentionality-free affectivity. In contrast to the days of the “old” image of the emotions, current philosophy of the emotions tends to be particularly interested in the very feature that has so shamefully been neglected in much of earlier philosophy. The following has come to be widely accepted: it is central to our understanding of the emotions that they be seen as a way of how our mind is “directed at” the world, and how our world is “given” to us. In short, emotions are now acknowledged as intentional states with specific content, rather than as mere states of bodily arousal. But this is only part of the story. Upon a closer look, it becomes obvious that with the recent turn to the intentionality of emotions, the days of the intentionality/feeling dichotomy are finally over. The view of the intentionality of the emotions that has been dominant over the most part of the recent debate tracks both the neglect of the emotional in much of earlier theory of intentionality and the fact that the intentionality of the emotions has been overlooked in earlier theories of the affective to one and the same source. Both shortcomings, it is claimed, are due to the fact that the emotions have been wrongly identified with the phenomenal aspects of the emotions, i.e. with the feelings. In many theories of the intentionality of the emotions, this is regarded as a mistake. Turning away from feeling theories, cognitivists claim that emotions are intentional in somewhat of the traditional sense of intentionality. Emotions are intentional insofar – and only insofar – as they imply cognitive and conative aspects, i.e. beliefs and action dispositions. And this does not seem implausible at all. Most certainly, the belief to be in some kind of danger does play a role in standard 1 Solomon (2003; 2006). 2 James (1884).

Shared Feelings Towards a Phenomenology of Collective Affective Intentionality

61

cases of fear, as does the tendency to flee and seek refuge, and similarly for all other emotions. The “directedness” of emotions – the question of what fear, anger, joy, pride are all about – is clearly a matter of those very beliefs and action tendencies, rather than of the feelings involved in standard cases of emotions. Perhaps the strongest arguments for this latter claim are the following. First, “feelings do not have ‘directions’”3 no fact about the twinge we feel, taken in itself, can tell us whether the twinge is indeed one of, say, remorse, rather than one of rheumatism. Thus it seems that whoever takes the twinge to be the core feature of emotional experience cannot at the same time focus on affective intentionality. Insofar as such feelings describe the phenomenological aspects of emotions, it seems that the phenomenological approach to the emotions misses the intentional character of our affective lives. Second, as opposed to emotions, feelings cannot be true or false4. Third, feelings are by definition conscious states; emotions, on the other hand, can be unconscious, as in the famous case in which after years of psychoanalysis, one finds out about the previously unconscious emotions one has had all along. Fourth, the emphasis on the “feeling” component tends to contribute to the “Cartesian” internalist image of the emotions, according to which emotions are “in the mind” rather than a form of engagement and entanglement with the world. We shall come back to these issues shortly. For the moment, let us simply take the cognitivist view of the intentionality of emotions for granted, and ask how this view answers our initial question. How does this contribute to explaining why so little work has been devoted to the analysis of collective affective intentionality? If the story I have told so far is basically correct, one does not have to assume some particular anti-affective bias in order to explain the apparent neglect of shared affective intentionality in current collective intentionality analysis.5 The explanation might actually be much simpler. If the cognitivist view of the intentionality of emotions is indeed right, the answer to the question concerning the reasons for the alleged research lacuna simply is that there is no such research lacuna. In this view, it simply makes no sense to weigh the amount of work done on collective conative and cognitive intentionality on the one hand against 3 Solomon (2003 [1975]: 4). 4 Solomon (2006: Lect. 3). 5 Yet, there may actually be such a tendency at work in some accounts of collective intentionality. Defending an anti-affective stance would be to claim that collective intentionality extends to shared beliefs and shared intentions, but not (or not to the same degree) to shared emotions. Among the received accounts of collective intentionality, Christopher Kutz (2000) seems to come closest to this view (cf. Kutz 2000: 196).

62

Hans Bernhard Schmid

the amount of work devoted to collective affective intentionality on the other. To claim that the philosophers of collective intentionality have neglected the analysis of collective affective intentionality by focusing on the cognitive and conative dimensions of intentionality would amount to a simple category mistake – because in the cognitivist view, shared intentions and shared beliefs are what the intentionality of shared emotions is all about. Thus the analysis of shared emotions appears to be nothing more than another realm to which the tools of the received analysis of shared conative and cognitive intentionality can be applied.6 The decisive question concerning the alleged research lacuna therefore is the following: is the cognitivist view of the intentionality of the emotions right? In the light of the most recent developments in the philosophy of the emotions, it appears that there is no straightforward answer to this question. Most certainly, the cognitivist view is right in emphasizing that emotions do indeed imply some sorts of beliefs, judgments, and cognitions. And many emotions are intrinsically related to action. At the same time, however, it has become increasingly clear over the last couple of years that the cognitivist view of the intentionality of emotions is distorted. In many respects, the intentionality of emotions is very different from the intentionality of beliefs. This is obvious from the following facts: first, we can have emotions independently of whether or not we have any corresponding beliefs, or even “in spite of” contrary beliefs. It is possible to be afraid of a snake without believing that it poses any real danger, as it is vividly illustrated by Charles Darwin’s experiment with the snake behind the glass. It is implausible to explain these cases with the assumption that they entail having an inconsistent set of beliefs, because these cases do

6 This seems to be the case in Margaret Gilbert’s very important contributions, which mark the most conspicuous exception to the general neglect of shared emotions in received collective intentionality analysis (cf. Gilbert 1997; Gilbert 2002). Gilbert discusses the case of collective guilt feelings. In standard cases, however, guilt presupposes action, as far as one tends to feel guilty for what one has done (or failed to do). This makes it particularly tempting to adopt a cognitivist view for the analysis of this particular shared emotion, and this is indeed what Gilbert does. Gilbert’s analysis of shared guilt feelings uses the very tools she has developed for the analysis of joint action, particularly the notion of collective commitment (Gilbert 2002). Thus shared emotions appear to figure as just another application for the theory. I think that in order to avoid any potential bias, cooperatively less marked examples should be chosen. In the following, the case of shared grief is used as a paradigm. In contrast to guilt, grief does not imply action.

Shared Feelings Towards a Phenomenology of Collective Affective Intentionality

63

not seem to make us irrational in the same way inconsistent beliefs do.7 Secondly, the respective relation of beliefs and emotions to the will is fundamentally different: We can neither believe nor have affective attitudes at will, but we can (and indeed very often do) cultivate our affective attitudes towards things in a way that has no direct analogy in the case of beliefs.8 Thus in this respect, too, the intentionality of emotions is very different from that of beliefs. If affective intentionality is different from the intentionality of beliefs (or cognitive intentionality), one might be tempted to think that this is because it is more of the kind of the intentionality of desires (or conative intentionality). This, again, might not seem implausible at first; being afraid implies the desire to flee and seek refuge, and similar action tendencies play an important role in many emotions. But this view does not hold up to closer scrutiny either. In fact, some emotions do not seem to involve any desire or action tendency at all, such as the pride one feels at the success of one’s daughter; these emotions rather seem to be of perception-like quality9. Thus it seems that the cognitivist account of emotions, instead of defending the importance of affective intentionality against ignorant feeling theorists, leads to a very distorted view of the intentionality of emotions. Cognitivism about emotions tends to model affective intentionality much too closely on the traditional model of intentionality of beliefs and desires, thereby neglecting the very peculiar kind in which emotions are (or can be) intentional. The intentionality of emotions does not seem to conform to the core feature of the belief/desire model of intentionality, which is the idea of a “direction of fit”. Put very bluntly: whereas beliefs are to fit the world, the world is to fit our desires. In a core sense, however, it seems that affective intentionality has neither of these two “directions of fit.” If emotions are indeed a form of “engagement” and “entanglement with the world”, as Robert Solomon emphasizes, this entanglement is not of the cognitivist kind. John Searle, to whom the conceptual elaboration of the “direction of fit” is owed, sometimes speaks of the “nil direction of fit” with regard to 7 In contrast to this, cognitivists tend to emphasize those cases where emotions vanish as a consequence to the formation of some new beliefs. To use Robert Solomon’s example, it seems natural to cease to be angry with James if one learns that it was not him who stole one’s car after all. This is a very mature reaction; I’m afraid, however, that in the normal real-life case, the anger will quickly resurge under a new mask. Although the “normative judgments” may change, the emotions are likely to persist. 8 For a detailed discussion cf. Goldie (2000). 9 Goldie (2000).

64

Hans Bernhard Schmid

the emotions – which is not a direction after all, as one might be tempted to add. If affective intentionality is not just a matter of cognitive and conative attitudes – what is it, then? In the light of the most recent development in the philosophy of the emotions, it seems that we have to overcome the intentionality/feeling dichotomy in order to understand the way in which emotions are intentional. It is the deep-seated preconception that feelings are not intentional – a view that feeling theorists and their cognitivist opponents held alike – which prevents us from seeing how emotions “disclose” the world.10 By playing the intentionality of emotions off against feelings, the cognitivists drive a wedge between the intentionality of emotions on the one hand, and the phenomenology on the other, leaving the feelings the role of mere contingent accompaniments of emotions.11 Suffice it to state for a fact that the tide in the current debate has turned: more often than not, feelings are now being seen as a core feature of emotions. Many authors claim that feelings can be (and indeed are) intentional, that the intentionality of feelings is different from cognitive and conative intentionality, and that feelings are at the heart of affective intentionality (which does, of course, not mean that all feelings are affective (or emotional). There are cognitive or epistemic feelings, too, such as the feeling of clarity). There are replies to all cognitivist points mentioned above.12 Feelings do have “directions”, feelings can be appropriate or inappropriate (if not true or false), they are not just contingent “accompaniments” of emotions; the very notion of “unconscious emotion” (i.e. emotion without feeling) is skewed, if not outright mistaken; and if the emphasis on the role of feelings for the emotions is taken to convey some illicit ‘Carte10 For an early critique of feeling theories cf. Bedford (2003 [1956/57]); the tendency to play the intentionality of emotions off against feeling theories is particularly conspicuous in Robert Solomon’s work. Cf. among many other examples Solomon 1993. 11 Margaret Gilbert also adopts this in her analysis of collective guilt feelings (Gilbert 2002). Gilbert made it her business to defend the claim that in a certain sense, collectives can “feel guilt”. Her analysis, however, makes it clear that the second part of her title - “collective guilt feelings” - is a misnomer. Gilbert does indeed claim that collectives can be said to “respond affectively” in terms of guilt, but she emphasizes that qua phenomenological ingredients, “feelings” are no necessary part of this emotion, but “only frequent concomitants” (Gilbert 2002: 119) or “accompaniment” (ibid.: 141) of emotions. Against this, Burleigh Wilkins has convincingly argued that a collective commitment without feeling might not be enough for there to be genuine collective guilt (Wilkins 2002). 12 For points 1, 2 and 4 cf. Goldie (2000), or, for an even more radical view, Ratcliffe (2005); for point 3 cf. e.g. Clore (1994); LeDoux (1994).

Shared Feelings Towards a Phenomenology of Collective Affective Intentionality

65

sian’ internalism, this is a question of how feelings are conceived of rather than of the concept of feeling itself. One might call this recent development the “phenomenological turn in the philosophy of emotions”. Here, feelings are conceived of as intentional – as “feelings towards”, to use Peter Goldie’s expression.13 For the purpose of this chapter, I take this development for granted, and I shall address the question: what are the consequences of this phenomenological turn for the purpose of an analysis of collective affective intentionality? For it seems clear that if affective intentionality is a matter of feelings rather than just of beliefs and desires, the question of how affective intentionality can be collective cannot simply be answered by pointing towards the received accounts of shared beliefs and joint intentions. If the phenomenological view of the intentionality of emotions is right, then there is a research lacuna. And it seems plausible to assume that on the basic level, it has to be filled with an analysis of how feelings can be shared.

2. ‘Sharing’: The metaphorical and the straightforward sense The expression of “sharing a feeling” does not seem to be altogether uncommon in everyday parlance, primarily in the sense of expressing one’s feelings to others. But there is more to the sharing of one’s feeling than the mere act of expression. The expression “Sharing a feeling with somebody” refers to the result as much as to the act of communication. And especially from a third person perspective, people might be said to share a feeling even independently of any prior act of communication. How are we to make sense of such expressions? I shall distinguish three loose or metaphorical senses of “sharing a feeling” from one strict or straightforward sense. In the loosest (and most innocent possible) sense, feelings are sometimes said to be shared insofar as the feelings of the respective individuals are qualitatively similar, and converge on one object or on one type of object. In the latter case, a certain feeling can be said to be shared in a popula13 In focusing on the phenomenological aspects of affective intentionality, however, one should be careful not to run feelings and emotions too closely together. Feelings extend to all sorts of things. One can “feel convinced”, for example, or “feel real”, or “feel stared at”. There are many feelings, which are not emotional in a narrow sense of the word/term. In the following, I shall use the term “affective intentionality” in a looser sense, extending it to those feelings, which are intentional, but not strictly emotional.

66

Hans Bernhard Schmid

tion even if the intentional objects of the individual feeling episodes are numerically different. In this sense, fear of dogs might be said to be widely shared even though people are afraid of different dogs, and might not even be aware of each other’s feelings towards dogs. This is “sharing” in a very loose sense of the term. In quite another sense, one might talk of “sharing somebody’s feelings” in the sense of sympathizing with another person. Here, the object of person A’s feeling is different from the object of person B’s feeling (the latter being person A, or person A’s feeling). Perhaps person B identifies with person A, and approves of A’s feeling to a degree in which she might be called to “share” her feeling. An even more demanding sense of “sharing” consists of a kind of combination of the two cases. People have feelings towards the same object or object type, and know about each other’s feelings, and the feelings of the individuals participating in a case of shared feeling might be causally connected in one or the other way. In this case, again, it is not an outside observer, but (some of) the participants who take themselves to share a feeling, either unilaterally or reciprocally.14 Knowledge of sharedness might deeply transform the way we feel. Particularly as far as the affective dimension of their minds is concerned, humans like to be “in tune” with each other. Thus Adam Smith has emphasized the role of the awareness of “emotional attunement” between people, claiming that it can be a source of great solace and indeed pleasure. As the saying goes – pain shared is pain halved and shared joy is a double – this type of sharing one’s feelings with others accounts for much of the structure of our social life, playing an important role in the emergence of social norms. In this account, sharing a feeling implies a “feeling of sharedness”, i.e. a higher-order feeling, which is “towards” the relation between lower-order feelings. Robert Sugden15 claims that Adam Smith’s theory of sympathy thus offers a convincing account of the emotional side of collective intentionality. The participants may have developed the lower-order feeling independently of each other, or some of the participants’ lower-order feelings may be caused by their knowledge or subliminal perception of the feelings of others (such as in more rudimentary forms of sympathy or emotional contagion). Whatever the underlying mechanisms, the core feature of fellow feeling is some kind 14 Note that the feelings of the participating individuals need not be qualitatively identical, or even just similar. In some cases of sharing a feeling, the participating individual’s feelings need to be suitably matching rather than qualitatively similar. I will discuss such cases below. 15 Sugden (2002).

Shared Feelings Towards a Phenomenology of Collective Affective Intentionality

67

of affective attunement between people that is sustained by the pleasure that flows from the awareness of this affective attunement. We might talk of sharing a feeling in all of these meanings in everyday parlance. But is this sharing in the straightforward sense? I submit that these cases do not conform to what we normally mean when we use the term “sharing something”. Let’s have a look at the simplest and most basic use of the word. Consider the case in which I propose to share a bottle of wine. Certainly, I do not thereby suggest that you and I each open a bottle, the two bottles being of the same vintage, or brand. Rather, I suggest that we enjoy one and the same (token) bottle. Similarly, the idea of sharing a car is not that of each one driving his or her own car, the cars being of the same brand. The point is to use one and the same (token) car together. The idea is this: one car, many users, one cake, many pieces, one apartment, many inhabitants, and so on, and so forth. This is what I will call the straightforward sense of “sharing something”. In the straightforward sense, sharing is not a matter of type, or of qualitative identity (i.e. of having different things that are somehow similar), but a matter of token, or numerical identity. That’s the straightforward sense of the concept of sharing, and it does not seem to apply to any of the three cases mentioned above. Note that in contrast to the straightforward sense, it is part and parcel of the very concept of fellow feeling that the feelings of the participating individuals are numerically different facts. Talk of “attunement” makes sense only if there are at least two relata that can be in (or out of) tune. There never was (and never will be) one token feeling with many parts and participants in these cases, but only individual feelings, perhaps with mutual cognitive and affective attitudes. This is what makes the talk of sharing a feeling merely metaphorical as applied to these cases. The question I wish to pursue in this chapter is this: Can feelings be shared in something of the straightforward sense of the word? Following is a description that suggests that we can: Father and mother are standing at the dead body of a beloved child. They feel ‘the same’ grief, ‘the same’ pain. This does not mean: A experiences this grief and B experiences it, too, and in addition to that they know that they feel it – rather, their feeling is a feeling-together (Miteinanderfühlen). A’s grief is not in any sense an object of B’s beliefs or feelings, as it is in the case of C, who approaches the parents with sympathy for them or for their pain. Rather, they feel their grief together in the sense of a feeling-together, an experiencing-together (Miteinander-erleben) not simply of the same value-state (Wertsachverhalt), but of one

68

Hans Bernhard Schmid

and the same emotional impulse (Gefühlsregung). [A’s grief and B’s grief are] not two different facts.16 In this passage, the German phenomenologist Max Scheler does indeed claim that in this case, the feeling of grief is shared in something of the bluntest straightforward sense of the word: one (token) feeling, many participants. The claim clearly is not that these parents simply experience feelings of the same type, or “matching” individual feelings, perhaps together with some form of (mutual) knowledge, (mutual) sympathy or fellow feeling. And neither is there any proper sympathy involved in this case. Rather, the claim is that while both individuals experience a feeling of grief, it is not the case that there are two different emotional facts involved here. According to Scheler, the parent’s feeling experience is not just qualitatively similar. Rather, it is numerically identical. It is needless to say that the idea that feelings can be shared in the straightforward sense of the word is quite unusual and provocative. Indeed it might even seem that Scheler’s case is simply too weak to be pursued any further. I take it, however, that in philosophy, unusual ideas should be highly valued and examined closely, even if their prospects of turning out to be true might appear to be rather dim. For even if they should turn out to be outright wrong at the end, the examination of such ideas might serve the purpose of helping us to clarify the reasons we have for rejecting them, and thus shed some light on why we conceive of matters in the way we normally do. This is what I shall be doing next: I shall use Scheler’s claim that feelings can be shared in the straightforward sense to examine the reasons we might come up with in order to defend our strong pretheoretic (or folk psychological) intuition that there is something wrong with this idea.

3. Individualism about feelings: Three versions I shall use the label “individualism about feelings” for the view that feelings cannot be shared in the straightforward sense of the word. I take it that individualism about feelings is a deep-seated conviction, that permeates not only most theories of affectivity, but large parts of our pre-theoretical views (i.e. folk psychology) as well.

16 Scheler (1974 [1913]: 23-4); my translation; my emphasis.

Shared Feelings Towards a Phenomenology of Collective Affective Intentionality

69

At least three different versions of individualism about feelings can be distinguished. First, feelings are ontologically individual. Feelings are conscious states. As such, they are ontologically subjective, i.e. somebody’s feelings, and there seem to be no conscious subjects other than individual beings. Second, feelings are epistemically individual. If it is true that only individuals can have feelings, it is also true that individuals can have only their own feelings. Third, feelings are, it seems, physically individual. If it is true that individuals can have only their own feelings, it also seems to be true that individuals experience their feelings as localized in their own bodies. This seems to force us to individualize feelings in the exact same sense as bodies; under normal circumstances, however, the only bodies there are seem to be the individuals’. Let’s have a closer look at each of these versions of individualism about feelings in turn, starting with the last version. Up until now, we have not discussed the nature of feelings beyond stating that they have a qualitative, phenomenal nature, i.e. that there is something “it is like” to have them. So Let’s have a closer look at what kind of phenomena might count as feelings. In his Concept of Mind, Gilbert Ryle gives a list that is often quoted: “thrills, twinges, pangs, throbs, wrenches, itches, prickings, chills, glows, loads, qualms, hankerings, curdlings, sinkings, tensions, gnawings, and shocks”.17 This list is particularly well suited to emphasize the individuality of feelings. Clearly, there has to be somebody who experiences these phenomena – a role for which individuals seem to be the only suitable candidates. And clearly, the itches, throbs etc. individuals feel are their own – they are epistemically exclusive, as it were. Moreover, individuals experience the feelings they have as localized in their own phenomenal bodies. Feelings are body-related. A twinge is felt as a twinge in the stomach, the “throbbing feeling” is in the breastbone, and so on. Feelings are felt as localized in the body. In that, feelings are spatiotemporal entities. This is also illustrated by the fact that feelings can be, to some extent at least, co-present. A twinge in the stomach is compatible with a throbbing feeling in the breastbone, and similarly for other cases. How does localization in the body amount to individuality? In the normal case, phenomenal bodies are individual in a twofold sense. They are individual in themselves, and they are individual in that they are related to our physical bodies, because normally, physical bodies are individual. In themselves, phenomenal bodies are individualized by their mode of access: nobody feels a twinge in somebody else’s stomach, let alone a throbbing feeling in some collective Leviathan’s breastbone. That is to say, phenomenal bodies are our own bodies, and they are individual, not collective. 17 Ryle (1949: 83-84).

70

Hans Bernhard Schmid

Phenomenal bodies are individual in their relation to physical bodies, because in the standard case (i.e. if we disregard for the moment phenomena such as phantom pains, which are parasitic in that their very possibility depends on the normal case), phenomenal bodies are co-extensive with physical bodies. And, if we disregard the extremely rare case of conjoined twins, and the possibilities of collective bodies such as Leviathans, physical bodies are the bodies of single individuals. Let’s call this form of individualism about feelings B-individualism. Tying feelings to our bodies, B-individualism seems to accommodate a deepseated intuition, which might explain some of the reluctance in recent collective intentionality analysis against affective intentionality. Cognitions and proattitudes might be collective; feelings, by contrast, seem to be individual, because they are localized in individual bodies. Yet there seems to be one exception. Not only has Margaret Gilbert devoted more attention to the affective than all other participants of the current debate, taken together, but she has also provided a standard formula for joint intentionality according to which the participants are “jointly committed as a body” (her italics). She thereby seems to suggest that intentionality has to be embodied in some way, and that shared intentionality therefore requires some kind of shared body. But in personal communication, she has made sufficiently clear that any such reading completely misses her point. Gilbert’s plural subjects do not have bodies of their own. The formula of “being jointly committed as a body” should be read as a metaphor, which uses the unity of the individual’s own body as an analogy for the collective unity that is generated through joint commitment. The collective “command center” of the plural subject does not have any arms and legs other than those of the participating individuals.18 For all of its apparent plausibility, there is something dubious about BIndividualism. The problem is not that there are collective phenomenal or physical bodies. Rather, the problem is that not all feelings are bodyrelated in the sense of being localized in the phenomenal body. There is an alternative to this individualizing reading of feelings. In the course of the current phenomenological turn in the philosophy of emotions, the focus of attention has shifted from those feelings, which are bodily sensations, and has turned to what is usually called psychic feelings.19 In contrast to bodily sensations, these feelings are much more a matter of the soul than of the body. Very much along these lines, Descartes says of some passions that they are felt “as if they were in the soul itself”. A rather useful distinction 18 Gilbert (2000: 5). 19 Stoecker (2007).

Shared Feelings Towards a Phenomenology of Collective Affective Intentionality

71

between bodily feelings and psychic feelings can be found in Scheler.20 It is telling that Scheler limits his concept of “immediate feeling-together” (unmittelbares Miteinanderfühlen) to psychic feelings. Following Scheler, psychic feelings are, in contrast to bodily feelings, intentional and not localized in the body, even though they are not without any relation to the body. The peculiar “heaviness” of the feeling of grief – the paradigmatic case of a psychic feeling in Scheler – does have a physical aspect. But grief is not localized in the body, as if one’s grief were felt in one’s left leg. According to Scheler, the peculiar nature of psychic feelings is such that it does not constrain the possibility of such feelings’ being shared in any way. Remember that Scheler’s example for sharedness of feelings in the straightforward sense is the parents’ grief for their deceased child. Scheler suggests that this feeling has to be individualized in a way different from the participating bodies: one (token) feeling, two bodies. As opposed to what some of Scheler’s interpreters claim,21 this identity is not just a matter of the content of the feeling, but a matter of the identity of the feeling as an impulse (Regung). Even if we grant that there are indeed such things as psychic feelings, and that the relation of this type of feelings to the body is not of a kind that leads to B-individualism, there are at least two further reasons for doubting that sharing of feelings in the straightforward sense could be possible. Let’s first turn to what one might call ontological individualism about feelings (or O-individualism). The claim is the following: as mentioned above, feelings are conceptually tied to consciousness. (This does, of course, not mean that they are either permanent, or in the focus of attention, or that one cannot be mistaken about one’s feelings. None of these claims is implied in the definitional consciousness of feelings.) Consciousness, however, is ontologically subjective, i.e. it is someone’s consciousness. There has to be someone who has it – the good old subject. It seems, however, that individuals are the only members of the class of subjects of consciousness we know of. Therefore it appears that only individuals – but not collectives – can have feelings. Ontological subjectivity does not only concern conscious states. It is a mark of the mental as such. Even those intentional mental states – such as beliefs and intentions – are ontologically subjective, which are not conscious states. Thus it is not surprising that the question of the subjectivity of collective intentions and shared beliefs has already stirred considerable debates in the received literature on collective intentionality. The point at 20 Scheler (2000 [1916]: 335-345). 21 van Hooft (1994).

72

Hans Bernhard Schmid

issue is the following. If it is claimed that there is such a thing as collective intentionality, and if it is further claimed that collective intentionality is in some way different from individual intentionality, we are hard-pressed to think that while individual intentionality is “had” by individuals, the subject of collective intentional states are the collectives. Yet there is considerable resistance against this apparently innocent assumption in current collective intentionality analysis. Most authors explicitly refuse to ascribe any mental states to collectives. Shying away from the group mind assumption, authors such as Michael E. Bratman and John R. Searle, resort to one or another version of a distributive reading of collective intentionality. Bratman claims that collective intentions are basically a web of individual intentions with collective content, while Searle, for his part, claims that collective intentions are irreducibly collective in mode, but “inside” the mind of individuals. In order to ban the group mind, some form of singularism about intentionality is adopted.22 The central insight behind this seems to be the worry that if any subject other than the participating individuals were the subject of the collective intention, the subjectivity and agency of the individuals involved in the process would be somehow impaired or compromised. It is true that individuals are and remain intentional agents when they join their forces for the purpose of collective actions, and they remain individual ‘cognizers’ when they share a belief. Individuals do not act and hold beliefs “on remote control”, as it were, when they share a belief and act jointly. Participants of collective intentional states are agents and cognizers rather than mere parts, organs, or instruments of some unified collective mind. Each individual participant can be ascribed beliefs and desires, which rationalize his or her behavior. In other words: any conception of collective intentionality should be compatible with the intentional autonomy of the participating individuals.23 22 Schmid (2003). 23 It is important, however, to distinguish the assumption of individual intentional autonomy from two further claims, with which it is usually mixed up. Individual intentional autonomy is the claim that in the normal case, each individuals’ behavior instantiates his or her own action (this excludes intentional heteronomy, i.e. behavior “on remote control”). The assumption of individual intentional autarky, by contrast, states that each individual ultimately acts only on his or her own desires, i.e. the intentional explanation of each individuals’ behavior should bottom out in that individuals’ own proattitudes (this excludes intentional heterarky, i.e. the possibility of acting on other people’s proattitudes without making them one’s own, or acting on another corresponding proattitude of one’s own). The third claim, intentional individualism, is weaker than individual intentional autarky. Intentional individualism is the claim that people can act only on individual inten-

Shared Feelings Towards a Phenomenology of Collective Affective Intentionality

73

Although individual intentional autonomy might explain why most philosophers of collective intentionality shy away from the group mind it is by no means a good reason for doing so. There are robust conceptions of the group mind, which are perfectly compatible with individual intentional autonomy. This is particularly obvious in Philip Pettit’s work. In his analysis of the discursive dilemma, Pettit has developed the view that (certain types of) collectives can be ascribed some sorts of mental states and do belong to the class of intentional subjects. He even ascribes them some sort of personhood.24 In this sense, groups can have “a mind of their own”.25 In his analyses, Pettit’s concern is with the rational unity of the groups’ perspective, which under some circumstances requires a peculiar discontinuity with the participating individuals’ own perspective. The phenomenon to which Pettit draws our attention is that the rational unity of a group perspective sometimes requires that this perspective be distinct from that of any of the participating individuals’. At the same time, Pettit argues that the group mind is solidly grounded in the volitions of the participating individuals. Groups have minds of their own based on the participating individuals’ insight into the problems of aggregating individual decisions to collective decisions, and on the participating individuals’ desire to act consistently and rationally, as a group. It is telling that while Pettit endorses the group mind, he stoutly rejects any notion of group consciousness.26 In his view, collectives might be ascribed some collective subjecthood in that they have mental states of their own; collectives might be treated as responsible for their actions, and thus be ascribed some sort of personhood. But there is no “what it is like” to have these states and to be these collective persons. Collective minds and persons do not involve any qualitative, phenomenal dimension, because collectives do not have any consciousness of their own. The consciousness of the participating individuals is the only consciousness there is. Thus even those philosophers who are liberal with regard to the collective mind exclude the possibility of collective feelings; this is the immediate consequence of their ruling out the possibility of collectives exhibiting some sort of consciousness in the first place. tions (this excludes intentional commonality, i.e. the existence of intentional states which are shared in the straightforward sense). In Schmid (2007), I claim that individual intentional autonomy does not imply individual intentional autarky and intentional individualism. Individual intentional heterarky and intentional commonality are compatible with individual intentional autonomy. 24 Pettit (2002); cf. Rovane (1998). 25 Pettit (2003). 26 Pettit (2002: 443).

74

Hans Bernhard Schmid

Let’s now turn to the third and last type of individualism about feelings. It is epistemological individualism (or E-individualism). If it is true that only individuals can have feelings, it is also true that the feelings individuals have are their own. Feelings are not only subjective. They are epistemologically exclusive (or monadic), too. This seems to preclude the possibility of shared feelings in the straightforward sense. If E-individualism about feelings is true, shared feelings cannot consist in a numerically (or token) identical feeling, for if people were to share a feeling in the straightforward sense, it seems that they would have to have each other’s (token) feeling (and not just a similar or suitably related feeling of their own), which is not compatible with the epistemological exclusivity of feelings. Again: it is clear that individuals can have similar or perhaps even qualitatively identical feelings. And it is undisputed that other people, or other people’s feelings, can be in the intentional content of individual’s feelings. E-individualism is not in conflict with any of the metaphorical forms of the sharedness of feelings. For those forms of affective attunement, sympathy, and empathy do not compromise the epistemological exclusivity which E-individualism is all about. Quite to the contrary: those forms of sympathy and fellow feeling presuppose that people have only their own feelings and that one person’s feeling is numerically distinct from another person’s. This is particularly obvious in simulation theories of empathy and sympathy, and in Robert Sugden’s view of Adam Smith’s concept of fellow feeling.27 As mentioned above: if fellow feeling is a matter of people enjoying their mutual affective attunement, it is clear that the feelings have to be numerically distinct, because it takes at least two feelings for there to be any attunement (or indeed disharmony) of feelings. There is no “fusion” between the feelings of the participants in these metaphorical cases. The feelings of the participants are (and remain) “different facts”, which is not the case in the straightforward sense of sharing a feeling. E-individualism about feelings enunciates our deep-seated conviction that it is inaccessible to us how other people’s joys and pains really feel, i.e. what it is like for other people to feeble in pain, or to experience joy. Our notions of sympathy, compassion and empathy – just as our philosophical theories of empathy from Theodor Lipps up to current simulation theory – are marked by the insight that those feelings and beliefs never really reach what they aim to grasp. There is no better illustration for our convictions concerning the inaccessibility of each other’s feelings than our recognition of other people as the ultimate epistemic authority concerning their own feelings. But still, there are some divergent views to be found in everyday 27 Sugden (2002).

Shared Feelings Towards a Phenomenology of Collective Affective Intentionality

75

life, literature, and philosophy. There is a way of talking about the relation between individuals that seems to imply that the monads do indeed have windows. It is perhaps no coincidence that Scheler chooses the parents’ feelings towards their child as an example. These bonds might be so tight that at times, it becomes difficult to distinguish one’s own feelings from the others’. A literary illustration is provided in Nadine Gordimer’s The House Gun, where the parents of an accused murderer constantly swap their rapidly changing feelings, so that it does indeed seem that some of their feelings are, in quite a literal sense, not their own.28. But the intuition concerning the possibility of sharing a feeling in the straightforward sense goes far beyond the range of intimate relationships. Taken with a grain of salt, one might say that even a presidential campaign has been won based on this claim. Remember that Bill Clinton cast an important part of his public persona with his claim not only to know what others feel, but actually to feel it. One might suspect Clinton’s famous “I feel your pain” (which was so often repeated and endlessly quoted) of Frankfurtian bullshit. But even if this should be true, the question is: why and how does it work? It seems that Clinton’s claim is at odds with E-Individualism. For it seems clear from the context, that Clinton’s statement cannot be understood in some of the metaphorical senses of sharing a feeling. Clinton does not mean that he understands the other person’s pain (“She doesn’t feel your pain, she understands it” – according to the influential political commentator Joe Klein, that’s precisely the difference between Bill and Hillary). Does Bill Clinton mean that he knows about the other’s pain, and that he feels sorry about it, is saddened by it, or that he has some other sympathetic feeling of sorts? This is rather implausible because of the wider context in which Clinton made his statement. The place was a Night Club in Manhattan, the time was the evening of March 26, 1992. With his statement “I feel your pain”, Clinton addressed Bob Rafsky, an Aids activist who rather violently disrupted one of Clinton’s stomp speeches by expressing his concern and indeed anger at the government’s apparent inactivity in that domain. Surely, Clinton cannot mean that he knows about Rafsky’s pain (which would amount to no more than stating the obvious), and that he is saddened by the fact that Rafsky is in pain. Clinton’s concern is clearly not so much Rafsky as the disease. It is the suffering caused by the disease and the fact that the government neglected the issue that pains him. Does Clinton therefore simply mean that Rafsky has no monopole of being pained by the disease, and by the way the disease is dealt with, and that he, Clinton, feels similarly or even iden28 I am grateful to Margaret Gilbert for drawing my attention to Gordimer’s novel.

76

Hans Bernhard Schmid

tically when it comes to aids? It is true that according to the transcripts of the incident, Clinton mentions that close friends of his had died of aids, thereby suggesting that aids is the source of a great deal of personal pain to himself. But still, this reading seems implausible, because Rafsky had made it quite clear that he himself is terminally ill of aids. So how could Clinton ever compare his own pain to Rafsky’s? Bullshit or not: what could Clinton’s statement possibly mean? We should not dismiss the possibility of the plainest and most straightforward sense: Clinton means that he feels Rafsky’s pain; he does not mean that he sympathizes with Rafsky, or that he has a similar feeling of pain of his own, but simply what he says: that he feels Rafsky’s pain. This would be an asymmetric case of sharedness in the plain and straightforward sense: the pain is – and remains –primarily Rafsky’s; but Clinton takes part in Rafsky’s pain in something of the sense in which Clinton entered the room in which Rafsky was waiting when he came to deliver his stomp speech. This reading seems appealing enough –€only that it is incompatible with E-individualism about feelings. To sum up this discussion of the three versions of individualism about feelings, it seems that while there are some rather strong intuitions concerning the individuality of feelings that should not be dismissed lightheartedly, it might still be worthwhile to consider alternatives. Let’s therefore have a second look at the notion of straightforward sharedness as it seems to be implied in Scheler’s view. Scheler seems to reject both O- and E-individualism. As to E-individualism, Scheler advances his perception theory of the consciousness of others, in which he claims that there is no fundamental difference between our access to our own consciousness and to that of others. There is, Scheler claims, an “inner perception” of the feelings of others that leads us to experience other people’s feelings in just the way in which we experience our own feelings,29 which is what E-individualism denies. Scheler’s further comments on the matter lead him into conflict with Oindividualism, too. While he does not claim that there are any collective subjects of consciousness, he simply denies the very idea of the ontological subjectivity of consciousness as such, stating that in its original form, the stream of consciousness is “quasi anonymous” in character. The distinction between my and your consciousness is not fundamental, as Scheler suggests, but only the result of a process of differentiation. So at the basic level of consciousness, it is not anybody’s, and therefore cannot be ontologically subjective.

29 Scheler (1974: 283).

Shared Feelings Towards a Phenomenology of Collective Affective Intentionality

77

This, however, seems to be in blatant conflict with a widely shared intuition that is at least as deep-seated as any of the other intuitions mentioned above. It is the idea of the separateness of individual persons. Contrary to what some philosophers such as Arthur Schopenhauer thought, the fact that we are separate persons is not only a matter of appearances. Our pretheoretic intuitions, our theories, and much of literature are based on the assumption that we really are separate persons, that the otherness of other people is not just a superficial “construction”. This intuition, too, should not be dropped light-heartedly. Love lyrics – just to quote one example – is full of vivid descriptions of how the innermost of the feelings and experiences of other persons elude us, how some permanent fusion of feeling remains unattainable however hard we might try. Thus the question is: is there any way to reconcile these two intuitions, i.e. the straightforward concept of sharedness of feelings on the one hand, and the intuition concerning the separateness of persons on the other?

4. Shared feelings: Defending the straightforward sense In Scheler’s remarks, as well as in the short but lively debate that was ignited by Scheler’s proposal, a solution to the dilemma seems to emerge. The point of departure is a structure that does not seem to have received sufficient attention from the part of analytical philosophy as yet. It is the fact that conscious states are not only just ontologically subjective in the sense that they are to be attributed to someone who “has” it. To be in a conscious state also means to have some pre-reflective awareness of one’s being in a conscious state. In some, however low degree of clarity, to be in a state of consciousness involves some self-referential awareness. Two clarifications are in order. First, this does not mean that one cannot be wrong about one’s consciousness. It seems obvious that we are very prone to misinterpretations of our own conscious states indeed. Second, the claim that to have a conscious state implies some awareness of one’s being conscious does not mean that in order to be conscious, subjects have to reflect on themselves all the time. Self-reflection only serves to make explicit the peculiar pre-reflective awareness of any kind of consciousness, and is very different from that awareness. Conscious states are – pre-reflectively and un-thematically – conceived and interpreted by the subjects who have them. To be in a conscious state means to conceive of this state in some or another way. In the case of intentional states of consciousness, this concerns the intentional content as

78

Hans Bernhard Schmid

well as the mode and the subject of the intention. In the latter respect, to be in an intentional state of consciousness means to conceive of oneself pretheoretically and un-thematically as the subject of that intentional state. Consciousness conceptually involves some form of self-awareness, some conception of oneself as the one who “has” the consciousness in question. This brings in an important distinction. The subject of a conscious state can mean either of the following: a) the subject who has the conscious state in question; b) the subject as whom the subject takes himself or herself to have the state in question. If we allow for the possibility that a) differs from b), this seems to open up a perspective to make the two conflicting intuitions compatible. Of course, individuals can have only their own conscious states, especially feelings – but this does not answer the question as whom those individuals take themselves to have their conscious states. In other words: O- and E-individualism might be true with regard to subjectA, but be wrong with regard to subjectB of the consciousness in question. And our intuition concerning the connectedness of persons might be true with regard to subjectB, but not with regard to subject-A. Without doubt, the parents in Scheler’s example are two different persons each of whom has his or her own feelings. But this does not preclude the possibility that both of them experience their feelings as theirs (together) rather than as their separate personal feelings. The following phenomenological consideration seems to lend further support to this assumption. It seems that in everyday life, we experience only very few of our conscious states as our personal conscious states. In fact, it seems that we take our conscious states to be our own only where we have reason to think that our conscious states might be different from anyone’s. Where this is not the case, we simply think what one thinks or what is generally thought, in an a-personal or anonymous mode, as it were. We do not take our thoughts or feelings to be our own in any meaningful sense. In other cases, we take our conscious states to be that of other people’s, such as in cases of Lippsian “internal co-action”, in which we identify with other people. In these cases, the “metaphysical psychic substance”, as Scheler puts it – i.e. subjectA – is individual. But subjectB differs from subjectA. This is the direction in which Scheler’s remarks – together with the (admittedly very few) defenders of Scheler’s view –€seem to point. The question is: how much weight can be put on the structure “SA feels x as had by SB”? Does it really make good sense to allow for a systematic difference between SA and SB in order to make room for cases such as the one of strong sharedness? It has the disadvantage to hinge on a poorly analyzed structure, and it raises the following criticism: are cases in which the two levels of subjectivity diverge, not simply cases of self-misinterpretation? Why not

Shared Feelings Towards a Phenomenology of Collective Affective Intentionality

79

say that people are simply mistaken whenever they take their feelings to be anybody else’s rather than their own personal feelings? Isn’t it obvious that people who take their feelings to be some other individual’s, or some collective’s, or nobody’s in particular, are on the wrong track? And this is only the first of a whole series of critical questions. I think that a valid straightforward notion of sharedness should conform to the following four conditions of adequacy (the conditions are drawn from the debate of Scheler’s account in the German speaking world of the late 1910s and 1920s).30 1. The self-awareness condition. As any notion of sharedness it must be compatible with veridical self-awareness. It cannot require that one should mistake oneself for another person, or for no person at all, or for a collective. 2. The fallibility condition. It must be compatible with the insight that one can always be mistaken about other people’s feelings. No feeling, however strongly felt, and however intimately connected one believes it to be to another person’s life, provides infallible information about other people’s feelings. In other words: no feeling is in itself the criterion of its being shared. Even parents might be mistaken in their belief that their feelings for their child are shared. (A particularly touching example from the belles letters can be found in Thomas Hürlimann’s Das Gartenhaus. Here, again, we are introduced to a couple having lost their child. The child’s mother reacts with shock when she finds out that contrary to what she thought, her grief is not their shared grief after all, and that for years, the father’s sole reason to accompany her to the child’s grave has been to feed the graveyard cats.)31 3. The holism condition. Feelings, as any intentional conscious states, do not occur in isolation. Rather, they depend on each other within the total stream of consciousness (Edmund Husserl calls this the “horizontal structure” of intentionality). So an adequate account of sharedness has to be able to explain how in spite of the holistic structure of consciousness, it is possible to share some intentional states, without sharing the total stream of consciousness. 4. The difference condition. In many cases, sharing a feeling is a matter of the differences between the feelings of the participating individuals rather than any form of identity. If Clinton claims to feel Bob Rafsky’s feeling, he 30 References can be found in section ii. of the list. 31 I am grateful to Angelika Krebs for pointing me to Hürlimann’s novel.

80

Hans Bernhard Schmid

certainly does not mean that he, Clinton, a presidential candidate, feels Rafsky’s pain – a terminally ill aids-activist’s – in the same intensity, and indeed in the same quality as Rafsky. By way of another illustration, take the following example: imagine the shared feeling of joy at the success of the first performance of a symphony. If the man at the triangle, the composer, some member of the audience and the stage manager take themselves to share a single feeling of joy, this is because in their perception of the situation, their individual feelings “match” with that of the others rather than being qualitatively or even numerically identical. In order to be taken as “matching”, these feelings have to be taken to be different from each other according to the different roles the participants play in the joint activity. If the composer takes the man at the triangle and the member of the audience to share her joy, she will not, in her right mind, take them to experience her exuberant exaltation; rather, she will take the shared feeling to entail her own exuberant elation together with, for example, the audience member’s delight, and the man at the triangle’s silent satisfaction. Thus taking part in a shared affective state sometimes seems to entail some form of awareness of the difference between the feelings of the participating individuals rather than any awareness of identity. At least in the general thrust, all of these points seem convincing enough. It is, I believe, reasonable to say that any concept of sharedness of feelings, however straightforward, should be able to meet the following requirements. Firstly, it has to be compatible with basic forms of individual self-awareness. People do not have to mistake themselves for another, or completely feel dissolved in some group consciousness in order to share a feeling. Secondly, it has to be compatible with the knowledge that any feeling one takes to be shared might not actually be shared at all. Thirdly, it has to leave room for the experience of (partial) separateness of our conscious lives. Not all feelings are shared. And ultimately, it has to conform to the experience that very often (if not always), the sharedness of a feeling is a matter of the qualitative difference between the individual contributions. Can any straightforward conception of shared feelings possibly meet these conditions? I think it can be shown that it can. Let me start with the third point. The main problem the holism condition seems to be driving at is the following. If conscious states depend on each other within a horizontal structure, and if one role of the concept of the subject of consciousness is precisely to capture this holistic form, consciousness can be either collective or individual, but cannot be switched ad libitum. If one conscious state

Shared Feelings Towards a Phenomenology of Collective Affective Intentionality

81

were to be straightforwardly collective, all would have to be collective.32 I do not think, however, that this argument poses any serious threat to a straightforward notion of sharing. Even if we grant that the general thrust of the argument is right, it is overly restrictive since there are not just single conscious states on the one hand, and the total stream of conscious on the other. There are intermediate structures: episodes of our conscious lives. While it is true that the experience of one single feeling will not be taken as shared in complete isolation, this does not mean that one’s total stream of consciousness would have to be shared in order for one element of an episode to be shared. It is within shared intentional episodes that these phenomenological fusions of feelings occur, and presumably, such episodes require some form of common life and shared practices. One does not take oneself to share feelings with perfectly strange creatures. Strong sharedness requires a context, perhaps some form of intimacy between the participants, or at least some shared cognitive and conative attitudes. Episodes may differentiate shared conscious states from other conscious states that are had in separation of others. In this way, the straightforward sense of sharedness seems to be in tune with the fact that we do not share our entire conscious lives. The overlap – if there is any – is only partial. This brings us to the second point in the above list. I expect that an analysis of the structure and presuppositions of these episodes should also yield some insights into the independent truth conditions to which a feeling has to conform in order to count as shared in the straightforward sense. This is an issue not to be pursued further here. It is clear, however, that whatever we might feel when we share a feeling cannot itself be the criterion of the truth of the sharedness, and that we can always be wrong about our feelings actually being shared. This does not preclude, however, strong sharedness in cases where these conditions (whatever they might be) are met. Furthermore, it seems clear that one does not have to mistake oneself for another person (let alone some group consciousness) to experience her or his feelings, if one’s feelings are taken to be shared with that person. In other words, the individuals experiencing a feeling as shared can be aware of the difference between what we have called subjectA and subjectB above. Thus it seems that the self-awareness condition can be met, too. The most serious attack on the idea of a phenomenological “fusion of feelings”, however, comes from the last of the abovementioned arguments. Remember that the problem is this: if we grant that in most if not all 32 If I understand him correctly, this is the upshot of Alfred Schütz’s critique of Scheler’s views (Schütz 1991 [1931]: 147).

82

Hans Bernhard Schmid

cases of shared feelings, the experiences of the participating individuals are taken to be qualitatively different by the participants qua “matching” contributory feelings to a shared collective feeling, it seems plausible to see these feelings as numerically distinct qua suitably matching parts to a whole. It is essential to the shared feeling of joy at the success of a performance that the exuberant exaltation of the composer is not the delight of the member of the audience, or the silent contentment of the man at the triangle, etc. This, however, seems to make necessary to abandon any straightforward notion in two ways at least. Firstly, if the feelings of the participating individuals are conceived of as (matching) parts, they are different from each other, not numerically identical (just as a bicycle’s frame cannot be its wheels if it is to serve its function within the total structure of the bicycle). And secondly, if the shared feeling is taken to be a whole with parts, it seems to be different from the parts – in other words, a collective “we” beyond the individual “I” and the “thou” comes into play as the proper subject of the shared feeling.33 I think, however, that these metaphysical conclusions can be avoided (in part, my solution is inspired by composition-as-identity theory as developed by Donald Baxter, David Lewis and David Armstrong). Firstly, if selfawareness is indeed compatible with straightforward sharedness, it seems that there are two ways of counting the number of feelings involved. With regard to subjectB (which is a “we”), the number is one. With regard to subjectA, the number is two (in the dyadic case). There is no reason why one way of counting should be more legitimate than the other. At the same time, however, there is no legitimate way of counting that yields three. The feelings can be counted either by phenomenological subject, or by ontological subject, but to count the “we” as a third subject on the list seems about as futile as the attempt to make some extra money by trying to sell one’s coin collection as a whole after one has already sold all its parts, or to wait for the teams to appear on the field after all the players have entered. The shared feeling is nothing in addition to what the participating individuals feel. Rather, it is that feeling, and it is that feeling in a certain respect. The individuals’ feelings are the one shared feeling insofar as the conditions are met under which individuals are not mistaken in experiencing their feelings as being shared by the other participants.34 This leads to a robust 33 These are indeed the two main revisions demanded by Edith Stein (1917: 18) in her discussion of Scheler’s view. 34 It should be at least noted in passing that the main problem of this view (which conforms to the composition-as-identity theory as put forward by David Lewis, David Armstrong, and Donald Baxter; cf. Baxter 2005) is that whereas the shared feeling is one, the individual feelings are many: “what’s true of the many is not

Shared Feelings Towards a Phenomenology of Collective Affective Intentionality

83

notion of a collective consciousness that avoids the pitfalls of some forms of anti-individualism in that it does not force us to reject either O- or Eindividualism. The first of the challenges posed by the argument from difference, however, is still not met. It is that sharedness in the straightforward sense seems to be incompatible with the awareness of differences between the feelings of the participating individuals. It seems that in most cases at least, the participating individuals have to take their feelings to be different from the others’ feelings when they take themselves to share an emotional state. I do not think, however, that this disproves the claim that people can consistently take their feeling to be token-identical with the others when they share a feeling. Sharing in the straightforward sense leaves ample room for difference qua aspects of the whole: the shared feeling of joy after the successful first performance of a symphony, insofar as it is felt by the composer, is one of wild exuberance; insofar as it is the man at the triangle in the orchestra’s, it is rather one of silent contentment – still, it is one feeling. Thus the numerical identity of the feeling does not preclude difference, but the difference in question here is one between aspects of one feeling rather than one between numerically different feelings. In a paper from 1997 – five years before she put forth the views on the structure of collective emotions that I have mentioned above –€Margaret Gilbert has pointed out the following phenomenon. Individuals can feel guilt or pride for actions of their groups, even if they did not serve any part whatsoever in the respective group activity. Their feelings, Gilbert claims, are of different quality than the feelings of guilt or pride individuals may feel regarding their own participation in joint actions. Gilbert has labeled the former, non-personal type of feelings membership feelings. Membership feelings, Gilbert claims, are an important feature of collective emotional states. In her paper from 2002, however, Gilbert has moved away from this earlier, largely phenomenologically inspired account. She quotes two reasons for her dissatisfaction with her own earlier views. First, membership feelings lack the unity of genuinely collective emotional states; they have no collective subject in the sense of a subject that can be distinguished from the participating individuals. Membership feelings appear to be purely individualistic, i.e. not of the cloth of which plural subjects are made. exactly what’s true of the one” (Lewis). Composition-as-identity-theory might be controversial. But the mere fact that it is around should remind us not to give up a corresponding pre-theoretical phenomenological intuition too early just because it does not seem to fit easily with our metaphysical views.

84

Hans Bernhard Schmid

Secondly, and perhaps more seriously, Gilbert points out that membership feelings are not sufficient as conditions for ascribing emotional states to collectives. If there is no common knowledge of the respective feeling among the participants, it is possible that all individuals feel membership guilt without there being any shared feeling.35 In this case, the collective cannot be ascribed any guilt feeling, even though all members feel membership guilt. These seem to be the two reasons for which Gilbert takes a cognitivist line on the matter in her later thought. In the course of her development, she turns away from phenomenology. Gilbert now thinks that the phenomenal aspects are no more than contingent accompaniments of collective emotional states, and that an account of collective emotions should be based on an analysis of the shared cognitive and practical components of these emotions, rather than on a phenomenology of the feelings. I hope that over the course of this chapter, it has become clear that neither of the two reasons Gilbert quotes for her cognitivist turn holds water. The problem of her earlier account is not that it is based on feelings, but that the feelings are conceived of in an overly individualistic fashion. A more straightforward notion of feelings avoids the problems Gilbert takes to be the reasons for her turning away from phenomenology. First, if feelings are shared in the straightforward sense, they do have a collective subject, which is different from the subject of the participating individuals (i.e. subjectB). And secondly, some structure of mutual openness is an integral part of sharing in the straightforward sense – remember Scheler’s remarks on the parents’€not needing any additional mutual awareness in addition to their shared feelings, because the mutual openness is part and parcel of the very feeling itself (which does not mean, however, that the participants might not be mistaken). My proposal thus is to understand membership feeling as shared feelings in the straightforward sense. Gilbert’s turning away from phenomenology is based on a mistake that is due to an individualistic understanding of feeling. Feelings can indeed be shared in something of the simplest sense of the word. Our basic intuitions concerning the individuality of feelings leave ample space for a strong straightforward understanding of the sharedness of feelings. Thus there is no reason not to take the phenomenological turn in the philosophy of the emotions to the collective level. Collective affective intentionality is a matter of shared feelings.

35 Gilbert (2002: 135 ff.).

Shared Feelings Towards a Phenomenology of Collective Affective Intentionality

85

References i. General References Baxter (2005): Baxter, D. Altruism, Grief, and Identity, in: Philosophy and Phenomenological Research (2005) LXX/2: 371-383. Bedford (2003 [1956/57]): Bedford, E. Emotions, in: Solomon, R. (ed.), What Is an Emotion? Classic and Contemporary Readings, New York: Oxford University Press, 2003, pp. 207-216. Clore (1994): Clore, G.L.: Why Emotions are Never Unconscious, in: Ekman, P., and R.J. Davidson (eds.), The Nature of Emotion. Fundamental Questions, Oxford: Oxford University Press, 1994, pp. 285-290. Gilbert (1997): Gilbert, M. Group Wrongs and Guilt Feelings, in: The Journal of Ethics (1997) 1: 65-84. Gilbert (2002): Gilbert, M. Collective Guilt and Collective Guilt Feelings, in: The Journal of Ethics (2002) 6: 115-143. Goldie (2000): Goldie, P. The Emotions. A Philosophical Exploration. Oxford: Oxford University Press, 2000. James (1884): James, W. What is an Emotion?, in: Mind (1884) 9: 188-205. Kutz (2000): Kutz, C. Complicity. Ethics and Law for a Collective Age. Cambridge University Press: Cambridge UK, 2000. LeDoux (1994): LeDoux, J.E. Emotional Processing, but Not Emotions, Can Occur Unconsciously, in: Ekman, P., and R.J. Davidson (eds.), The Nature of Emotion. Fundamental Questions, Oxford: Oxford University Press, 1994, pp. 291-292. Ratcliffe (2005): Ratcliffe, M. The Feeling of Being, in: Journal of Consciousness Studies (2005) 12/8-10: 43-60. Ryle (1949): Ryle, G. The Concept of Mind, London: Hutchinson, 1949. Solomon (1993): Solomon, R. The Philosophy of Emotions, in: Lewis, M., and J.M. Haviland (eds.), Handbook of Emotions, New York: Guilford, 1993, pp. 3-15. Solomon (2003): Solomon, R. Emotions and Choice, in: Not Passion’s Slave. Emotions and Choice, Cambridge UK: Cambridge University Press, 2003, pp. 3-24. Solomon (2006): Solomon, R. The Passions: Philosophy and the Intelligence of Emotions, Chantilly/VA: The Teaching Company, 2006.

86

Hans Bernhard Schmid

Stocker (2003): Stocker, M. The Irreducibility of Affectivity, in: Solomon, R. (ed.), What Is an Emotion? Classic and Contemporary Readings, New York: Oxford University Press, 2003, pp. 258-264. Sugden (2002): Sugden, R. Beyond Sympathy and Empathy. Adam Smith’s Concept of Fellow-Feeling, in: Economics and Philosophy (2002) 18: 63-87. Wilkins (2002): Wilkins, B. Joint Commitments, The Journal of Ethics (2002) 6: 145-155. ii. References to the debate on Max Scheler’s theses Becher (1917): Becher, E. Die fremddienliche Zweckmäßigkeit der Pflanzengallen und die Hypothese eines überindividuellen Seelischen, Leipzig: Veit, 1917. Becher (1921): Becher, E. Geisteswissenschaften und Naturwissenschaften, Munich: Duncker & Humblot, 1921. Binswanger (1922): Binswanger, L. Einführung in die Probleme der allgemeinen Psychologie, Berlin: Springer, 1922. Bühler (2000 [1927]): Bühler, K. Die Krise der Psychologie, Weilerwist: Velbrück, 2000. Cohn (1919): Cohn, J. Geist der Erziehung. Pädagogik auf philosophischer Grundlage, Leipzig/Berlin: Teubner, 1919. van Hooft (1994): van Hooft, S. Scheler on Sharing Emotions, in: Philosophy Today (1994) 38/1: 18-28. Kronfeld (1920): Kronfeld, A. Das Wesen der psychiatrischen Erkenntnis. Beiträge zur allgemeinen Psychiatrie, Berlin: Springer, 1920. Roffenstein (1926): Roffenstein, G. Das Problem des psychologischen Verstehens. Stuttgart: Püttmann, 1926. Scheler (1954 [1912]): Scheler, M. The Nature of Sympathy, transl. by P. Heath, London: Routledge and Kegan, 1954. Scheler (1973 [1912-1916]): Scheler, M. Formalism in Ethics and Non-Formal Ethics of Values. A New Attempt toward the Foundation of an Ethical Personalism, Evanston/IL: Northwestern University Press, 1973. Schütz (1991 [1931]): Schütz, A. Der sinnhafte Aufbau der sozialen Welt. Eine Einleitung in die verstehende Soziologie, Frankfurt: Suhrkamp, 1991. Stein (1917): Stein, E. Zum Problem der Einfühlung, Halle, 1917. Volkelt (1920): Volkelt, J. Das ästhetische Bewusstsein, Munich: C.H. Beck, 1920.

Part II: Analysing Sharedness

Limiting Reductionism in the Theory of Collective Action David P. Schweikard, Köln Reductive explanations only distort the phenomena to be explained. (Joseph Raz)

1. Introduction One way of reading my title points to the diagnosis that gives rise to the main criticism concerning theories of collective action I intend to formulate. An alternative reading points to the therapy I suggest for future theorizing in this field. I think and diagnose that a universal reductionism limits the scope of a theory of collective action by excluding phenomena that such a theory aspires to account for. Instead, in terms of therapy I suggest classifying the phenomena of collective action and specifying to what kinds of phenomena reductive accounts are applicable, thereby limiting reductionism itself. This strategy does not, as should be clear ex ante, favour anti-reductionist accounts in general. Rather, the conviction that underlies my proposal is that there is no single reductive or non-reductive approach that could claim applicability to all types of collective action. Furthermore, searching for a differentiated overview over the phenomena of collective action helps locate and formulate with precision the core questions for further research. Throughout the discussion I will not take issue with the ways in which reduction can (or is suggested to) be carried out. Thus, I do not define reductionism or reductionist views with regard to their claims about how some X can be reduced to some Y. I will instead only look at the resulting claims according to which the Y serves to account for a certain kind of phenomena. By setting the focus on the plausibility and the efficiency of an account of collective action – in the following I refer to the accounts that imply reduction as “reductive accounts” – I set aside the strictly methodological issue concerning the structure of reduction itself.1 1 As I see it, there is still a lot more work to be done regarding this methodological issue. Most accounts of collective action (or collective intention) of which I am aware simply take reductive strategies to be successful and work with theses of explanatory replacement (e.g.: “What is explained in terms of X can/should be N. Psarros, H. B. Schmid and K. Schulte-Ostermann, Concepts of Sharedness, Frankfurt 2008: 89-117; ontos verlag

90

David P. Schweikard

Large parts of the paper consist in clarifying the overall claims by differentiating between kinds of reduction regarding shared or collective intention (section 3) and between types of phenomena within the extension of this theory (section 4). To prepare the ground for these differentiations, I introduce two crucial questions in the theory of collective action by way of an adaptation of “Wittgenstein’s question” known from general philosophy of action (section 2). With the specified versions of reductionism and a taxonomy of action types at hand, I lay out my suggestion for what kinds of collective action reductionist strategies are applicable to (section 5). My approach in this paper is mainly metatheoretical, since I shall restrict my discussion to clarifications and suggestions for how to theorize about certain phenomena. This results in a sketch of the landscape of accounts of collective intention, at the same time specifying the respective extensions, i.e. the ranges of applicability. To complete the overview, I will briefly reconstruct and comment on one of the most popular defences of the claim that groups of a certain kind are to count as agents in their own right (section 6).

2. Two core questions That people act together, cooperatively, jointly, collectively or as a group, is a ubiquitous phenomenon. Thus, philosophers of collective action do not need to prove that people act together. Assuming that people can, in a more or less straightforward way, have common goals, they do not usually pay much attention to why people do so. What they focus on is how people act together. This is the question that promises genuine and novel insights into both the nature and varieties of agency and the structure and subtleties of social life. In the case of individual action, it is common sense among philosophers of action to refer (at least partly) to the agent’s intention to describe and explain the distinctive features of her actions.2 Wittgenstein famously phrased this centrality of intention-to-act in the question that has become known as “Wittgenstein’s question”: “What remains if I subtract the fact explained in terms of Y.”). While explanatory reduction is not the only kind of reduction, I consider it to be a more interesting task to show why such replacement works or, rather, which other arguments make plausible replacement or reduction. 2 All references to “intention” and “collective intention” in the following are to intentions-to-act. This entails no comment on whether what is said also applies to other kinds of intentional states.

Limiting Reductionism in the Theory of Collective Action

91

that my arm went up from the fact that I raised my arm?”3 This question puts some weight on the relation between the plainly physical event (the arm’s going up) and the action (my raising my arm), thereby putting the issue of naturalism on the agenda. One of the tasks thus defined is to explain how the theoretically distinguishable aspects of action, the physical event and the mental event that causes or guides the bodily movement,4 are related to one another. The case of collective action is not studied along exactly the same lines. Here, too, we deal with movements of individuals that are somehow intricately related to their intending to move in certain ways. Yet, the problem is not how the agents involved each do what they do, but what is distinctive about their doing something together. Consider Margaret Gilbert’s well-known example of two walkers who start out going for their separate walks along the same path and eventually team up and walk together.5 For the present purpose, let us simplify the example such that the two walkers walk side by side, at the same speed and on the same route. Most cases of collective action in real life are less harmonious and, concerning the individual actions involved, less uniform, but this artificial scenario shall serve as a basic case. Whereas Gilbert analyzes what the walkers do in forming a group of walkers and what this group formation implies, I propose to focus on the difference between their walking separately and their walking together, in order to get a first idea of the problem of collective action. This difference can be illustrated by adapting and re-phrasing Wittgenstein’s question as follows: What remains if I subtract the fact that two people walked alongside each other from the fact that they walked together? Or, in other words, what’s specific about a number of people acting together in contrast to the same individuals” performing parallel individual actions? While Wittgenstein pointed to the role of intending in an event of someone’s raising his arm, this re-phrasing points to the importance of successful coordination of individual intending and doing. While the original form of Wittgenstein’s question serves to carve out the phenomenon of individual intention, this adaptation for a paradigmatic case of collective

3 Wittgenstein (PI: § 621). 4 Since I do not here argue for or against causal theories of agency or theories about the nature of mental states, I can leave this sketch of a central issue in the philosophy of action as rough as it stands. As I will point out in section 3, commitments regarding these questions can, when taken as fundamental for a comprehensive theory of action, indeed constrain accounts of collective intention. 5 See Gilbert (1990).

92

David P. Schweikard

action can serve to mark the locus of collective, joint or shared intention.6 The big questions about intentions of the first kind turn on how they are related to other mental and physical phenomena, on their semantic form and their conditions of satisfaction. The major interest with respect to intentions of the second kind is, phrased preliminarily, concentrated on whether and in how far they are reducible to individual intentions or to individual intentions and their interrelations. Consequently, the question of reduction lies at the core of reflections on collective intention right from the start. It can, moreover, be seen as an issue where deeper or more general philosophical convictions culminate and find expression. In addition to this question concerning the differentia specifica of collective intention, I want to formulate a second key question that every account of collective action, reductive or non-reductive, should be prepared to answer. It concerns the conditions that must be met for the collective action to be successful and it reads as follows: What ensures that people act together successfully instead of just performing unconnected individual actions parallel to one another? This success condition for collective action parallels the individual case by specifying the conditions of satisfaction for the underlying intentional attitudes. Now, these two questions are indeed crucial, but they do not indicate the only genuine problems in the theory of collective action. Furthermore, phenomena of collective action create problems for accounts of act individuation7 and yield intricate follow-up questions concerning the social ontological implications, especially regarding the ontological status of (should there be such) group agents.8 While I do think that these problems deserve more attention than they have received in recent debates, I will focus here on the problem of the intentionality of collective actions and leave the two core questions mentioned figure as conditions of adequacy. 6 I use the term “collective intention” as generic term for the intentional states involved in collective action. The alternatives mentioned, “joint intention” and ‘shared intention” are meant to allude to exemplary accounts; see Tuomela (2000) and Bratman (1999: ch. 5 - 8), respectively. 7 See Chant (2006). 8 See Pettit (2003a) and Pettit and Schweikard (2006). This conception of the core problems in the philosophy of collective action is similar, yet more fine-grained that the one offered by Kirk Ludwig (in Ludwig 2007). Ludwig identifies two main problems, one regarding the ontology, the other regarding the psychology of collective action. To put it very broadly, in this paper I first focus on psychological questions and then deal with certain ontological implications of answers to these questions.

Limiting Reductionism in the Theory of Collective Action

93

In the next section, I will specify what sorts of reductive strategy are available concerning collective intention before turning to the underlying commitments and saying more about which reductive strategies are (or might be) motivated by which commitments.9

3. Reductive strategies specified Preliminarily, I should point out that it is not my aim in this paper to deal explicitly with all accounts of collective action known from the literature. Rather, I distinguish between kinds of accounts in explaining the systematic options in the study of collective intention. In focussing on collective intention I will bracket concise discussion of the beliefs that accompany intentions to act. For the purpose of describing the structure of collective action it shall suffice to assume that beliefs play an important role in the process forming individual or collective intentions, but that in action (i.e. in the course of performing an action) this role is entirely integrated in the guiding intentions. Thereby, I also set aside consideration of what can be called the epistemic problems of collective action that concern possible discontinuities between individual and collective conceptions of what the collectively intended act is or entails. In line with large part of the literature I assume concordance between the content of collective intentions and the related individual beliefs. (i) Let us assume for the following that a collective intention is the intention that is referred to in describing and explaining collective action. Let us assume further that it plays an important role in bringing about the collective action by contributing to the coordination of individual intentions and actions. Given these assumptions, I can list three paramount questions that have triggered wide-ranging debate on the structure of collective intention among philosophers of collective action. These are: (1) Whose intention is the collective intention? (2) In which mode is the intention held? (3) What is the form of the content of such an intention? 9 An alternative way of reconstructing the problem of collective action and of criticizing reductive accounts is suggested by Frank Kannetzky (2005). His main claim is that intentionalism – i.e. a certain set of reductive (or individualist) views about mental states and action intentions he attributes to the most widely discussed theorists of collective action (namely Searle, Gilbert, Tuomela and Bratman) – is an insufficient basis for a theory of collective action, for it fails to account for the irreducible statues of shared forms of action and genuinely social practices. Although I agree with Kannetzky on a number of important issues, this is not the line I want to take in this paper.

94

David P. Schweikard

Every action intention, so the assumption that leads to question (1), has to have a subject. There has to be someone who has the intention that is involved in an intentional action. But this demand does not seem to indicate standard cases of collective action, in so far as it means that someone, i.e. one individual intentional agent, has to have the intention in question. Admittedly, there are cases such as the “mafia-case” described by Michael Bratman, in which only one member of a group has the intention that the group perform a certain action, e.g. that they go to New York together.10 There are good reasons for not calling such degenerate cases of doing something together collective actions and for formulating the further requirement that coercion be absent if a collective action is to take place. But still, excluding these cases from the phenomena that we want to account for does not answer question (1). The options are as follows: (a) the collective intention is an intention of individuals, taken singularly, (b) somehow interrelated individuals together have the collective intention or (c) the group11 itself is the subject of the collective intention. The addendum “taken singularly” and the qualification “interrelated” mark off (a) and (b): whereas (a) does not claim that any relation between the individuals involved is relevant to the existence of the collective intention, (b) states that some interrelation is indeed necessary for the collective intention to exist. The existence claim in (c) is even stronger, for it claims that groups themselves can be subjects of intention. Questions (2) and (3) refer primarily to the semantic form of intentions. Although this preliminary decision is not entirely innocent, the semantic form is taken to consist in two elements, the mode and the content of the intention. Concerning the mode referred to in question (2), the options are basically – adopting a terminology which Raimo Tuomela has transferred to the study of collective phenomena12 – a) the I-mode and b) the We-mode. Keeping separate the answers to questions (1) and (2), we should explain these options by saying that whoever or whatever is the subject of a collective intention, it can stand either in the form “I intend …” or in the form “We intend …”. Specifying the content of collective intention necessitates another distinction concerning the concept of intention. It has been suggested not to constrict the form of intentions to intentions to do something, but also to 10 See Bratman (1999: 100). 11 I use the term “group” here in its intuitive meaning. 12 For more recent usages of this terminology, see Tuomela (2002, 2003 and 2006).

95

Limiting Reductionism in the Theory of Collective Action

include intentions that something happen or be the case.13 Thus, not only one’s own actions can be the object of one’s intention, but also states of affairs that may but need not include one’s own contribution. Applied to the case of collective action, this means that the participating individuals” intentions need not be limited to their individual contributions, but it can appeal to the group’s collective action of which an individual contribution is only a part.14 This can be used to mark the distinction between a) intentions “that I do …” and b) intentions “that we do …”. Summing up, the basic options regarding the three elements of collective intention I have specified are the following: 1) subject of intention

2) mode of intention

3) content of intention

a) individual

a) I-mode

a) intention that I do …

b) interrelated individuals

b) We-mode

b) intention that we do …

c) group Table 1: The structure of collective intention

(ii) Now, what reductive strategies are available? Using the above table, we can call any theory of collective intention reductive that makes use of at least one element that is not given in the bottom line of the respective column. I will refer to the elements given in columns as “higher” and “lower” elements, such that e.g. 1a) is higher than 1b). This does not apply diagonally. More specifically, there are two ways, in which a reductive theory of collective intention can proceed: either it claims that collective intention can be analyzed on the basis of a theory that is reductive in the general sense just given, i.e. which makes use of higher elements of the table. Or it claims that it is possible to substitute an account using a higher element for an account that (ceteris paribus) uses a lower element in the same column. 13 See Vermazen (1993), Bratman (1999: 98) and Pettit/Schweikard (2006). 14 This conception of “appeal to the group’s action” has also been proposed by Michael Bratman in a paper entitled ‘shared Action, Shared Intention, and Shared Valuing”, presented at the Workshop on Collective Intentionality in Groningen, June 2006. – Appeal to the intentions or actions of others should be understood as involving the corresponding beliefs that, e.g., others also intend to perform the collective action. But I do not think it necessary to make this appeal explicit by saying that the individuals involved intend plus they believe that the others will cooperate.

96

David P. Schweikard

Hence, there is a variety of reductive accounts which are, as we can now say, reductive at least with respect to one aspect of collective intention. Reversely, this allows saying more precisely what can be meant when an account is called “non-reductive” or when a theory is called “anti-reductionist”. Instead of simply labelling an account “non-reductive” in case it makes use of at least one, but just any “higher” element, the suggestion is to specify the aspect with respect to which an account is non-reductive. This suggestion could also be extended to the classification of accounts in the debate on collective intentionality in general by insisting on a specification of exactly which aspect of the structure of intentionality is accounted for reductively. Let us now list very briefly some combinations of these elements that qualify as reductive accounts and of which most have actually been opted for, before we look at their underlying commitments. 1. The strictest reductive strategy combines the elements 1a), 2a) and 3a), thereby claiming that all aspects of collective intention can be accounted for in terms of individuals as intending subjects and the intentions standing in the I-mode and appealing only to single individual actions. 2. One of the accounts that have been at the centre of interest in the literature on collective action combines, according to this scheme, the elements 1b), 2a) and 3b). By calling the complex of interlocking individual intentions a ‘shared intention”, Michael Bratman has offered this type of reductive account for small scale act-types.15 3. Another group of comparatively popular accounts can be described as combining 1a), 2b) and 3b). Although they don”t agree on how this combination of elements is to be analyzed, Raimo Tuomela’s and John Searle’s accounts belong to this group.16 Here we have individuals intending in the we-mode (or “we-intending”) to do something together. According to both accounts, the individual intentions that guide the individual actions in pursuit of a collective action are derived from collective intentions (or we-intentions). 4. Finally I would like to mention one account that can be read as a combination of the three “lowest” elements from the scheme – 1c), 2b) and 3b). But, upon closer analysis, this account – Margaret Gilbert’s theory of plural subjects17 – turns out to refer to groups not in a non-reductive, 15 See Bratman (1999: Chap. 5-8). Since he never elaborates on any of those statements, I simply ignore the fact that Bratman sometimes refers to a ‘shared intention” as the “intention of a group”. 16 See the “classic” papers Tuomela and Miller (1988) and Searle (1990). 17 See especially Gilbert (1989; 1996; 2000 and 2003).

Limiting Reductionism in the Theory of Collective Action

97

but rather in a reductive fashion, for it takes the group as intending to perform some action X only in virtue of the group members” jointly committing themselves to the performance of X. This is not the place to reconstruct her approach and to develop this classification, but I conjecture at this point that Gilbert’s account is best described as combining 1b), 2b) and 3b). (iii) What are the deeper commitments that may lead to adopting one or the other reductive strategies? By offering a brief overview of the most important options, I intend to shed light on some interconnections between the theory of collective intention and issues in the philosophy of mind and in social philosophy more generally. Once again, my aim is not to provide a concise interpretation of the existing accounts of collective intention, but rather, to indicate at which points the particular shape of such accounts could be due to implicit commitments to other, closely related views. Regarding the first aspect, the prime question is whether only individuals can be subjects of intentions. This question is closely connected to the more general issue concerning the nature of intentional states. Claiming that intentional states are nothing but mental states of single individuals, i.e. individuals outside any relation to other individuals, comes down to opting for 1a). One fundamental commitment that would lead to this claim could be some kind of naturalist view concerning the nature of mental states according to which they can be reduced to physical states of the brain. Although philosophers of collective action do not need to be explicit on this point, putting forward an account that is reductive with respect to the first aspect in my scheme can be interpreted as part of a theory that is meant to be compatible with some version of naturalism about the mental. Another related issue that may result in a reductive account concerning the ‘subject of intention” in collective action regards the individuation of mental states. Here, the position usually labelled as “internalism” or “individualism” claims that the form and content of an intentional state can be specified or individuated by appealing only to facts inside the individual whose state it is. Put differently, it claims that an individual’s mental events supervene only on physical events inside (or internal to) this individual’s body. A commitment to this position would also result in a commitment to 1a) in the given scheme, for it explicitly excludes the possibility of an intentional state being relational or even a state of a group. The second position concerning this first aspect of collective intention is represented by the claim that not only individuals taken singularly,

98

David P. Schweikard

but also interrelated individuals can be ‘subjects” of intention. This view is based on an entirely different view of intentional states. It holds that intentional states need not be states of single individuals (in the above sense) but can also be states held by interrelated individuals. As will be obvious, the tenability of this view depends crucially on the nature of the relation between individuals that has to obtain if they are to count as interrelated subjects of intention. I basically see two ways of understanding these interrelations: one is semantic, the other ontological. The semantic conception limits the claim about interrelation to an interconnection of individual intentional states, i.e. to the claim that the interrelation between participants consists in their mutually referring to one another in their individual action intentions, thereby creating a web of interrelated intentions.18 The ontological conception, on the other hand, is based on the claim that the interpersonal relations between individuals participating in collective action are not just semantic, but ontological in the sense that standing in such relations is a constitutive feature of the sociality of individuals.19 Thus, a commitment shared by all views that endorse 1b) is that in intentionally acting together, the individuals involved are related to one another in such a way that a complex of interrelated individuals constitutes the subject of the relevant collective intention. While differing on the nature of the interrelation, these views are united in rejecting the reductive strategy that uses 1a). Regarding the connection to issues in the philosophy of mind that were hinted at above, the semantic version seems to be independent of at least the internalism/externalism-question,20 whereas the ontological version, by stressing the existence of (ontologically) constitutive relations, seems to be incompatible with internalism about the individuation of mental content. This surely holds for the content of the 18 Needless to say – given the allusions in the text –, I take Michael Bratman to be committed to a view very close to this characterization. In fact, in personal conversation (in June 2006) he referred to the relation implied by his conception of interlocking and interdependent intentions as ‘semantic interconnection”. 19 Very broadly, I take Hans Bernhard Schmid (2005a and 2005b) to be committed to this kind of view. 20 A semantic conception of interrelated individual intention that is to be compatible with internalism about mental states would have to adopt the minimal claim that each of the agents involved needs only to refer to the other(s) in the content of the individual intention. If it is a condition of satisfaction of such an intention that there are other(s) who intend likewise, the account is committed to an externalist view of mental states. On the question whether an internalist theory of mental states can serve as a basis for a theory of collective intention see especially Schmid’s (2003) critique of Searle (1990).

Limiting Reductionism in the Theory of Collective Action

99

intentional states of individuals acting together, which could not plausibly be individuated by reference to just one isolated participating individual. For some time in the debate on collective action, authors have referred to just any collective action as an “action of a group”, but recent reflections on collective intention have led to questioning the plausibility of attributing the relevant intentions to the group as such. It may be one thing to attribute intentions to interrelated individuals, but it is quite another to say that groups can be subjects of intentions in a non-metaphorical sense. The issue here is not so much the nature of intentional states, but rather, the social ontological question of whether groups have an ontological status of their own. One way of approaching these issues is to treat them as connected issues whose answers are interdependent. What is abandoned, then, is the search for a proof of the existence of groups or collective entities apart from their being capable of intending. And what is envisaged, instead, is an argument according to which the capability of intending is exactly what grounds speaking of groups as irreducible entities.21 This line of reasoning still does not imply that groups can or do have minds just as individuals have minds, at best, they have minds and are capable of intentions of a different kind. One way of making sense of the connection between intentions of the group and individual intentions is to show that such group intentions are supervenient on the intentions of the group members.22 Alternatively, one could argue that the group members” intentions constitute the group intention.23 In any case, an account that is committed to the view that groups can be subjects of intentions has to assume (i) that groups exist and it seems plausible, on such an account, to assume further (ii) that if a group has an intention, this intention is somehow accompanied by its members” intentions.24 In the following I 21 See Tollefsen (2002a and 2002b), Pettit (2003a) and Pettit and Schweikard (2006). There may be other contexts and other arguments that hint at or even prove the existence of groups. But since the discussion in this paper is focused on collective intention, I will not try to capture such alternatives. 22 Cf. Tuomela (1989) and List and Pettit (2006). 23 Pettit (2003a) points at this kind of view, but does not elaborate along this line. 24 I use the relatively neutral term “accompanied”, leaving room for different accounts of the relation between the group’s and its members” intentions. The second assumption, however, is not necessary, for one could imagine cases where the group members behave somewhat automatically, not showing any sign of intentional action, and still it is reasonable to attribute intentions and actions to the group. Just think of a particularly ecstatic music performance or the final phase of a soccer match where players are too exhausted to really coordinate their moves.

100

David P. Schweikard

will assume that these latter intentions stand in the We-mode and appeal directly to the action of the group. Regarding the other two aspects of collective intention distinguished above, the underlying commitments are not as controversial, although the plausibility of a comprehensive account obviously hinges on (the consistency of) its conception of all three aspects. Concerning the decision as to whether collective intentions are best conceived as being held in the I-mode or in the We-mode, the connection to the first aspect of collective intention is comparatively straightforward. This question only arises in connection with a commitment to 1a) or 1b), the views that only individuals or interrelated individuals can be subjects of intention. In fact, of the 2x2 matrix consisting of 1a), 1b), 2a) and 2b), all four possible combinations of elements are (given some qualifications) viable options in the theory of collective intention, so the choice between 2a) and 2b) is not determined by a prior decision in favour of 1a) or 1b). This is especially important, for putting forward a non-reductive account concerning the mode of intention does not exclude commitment to the view that only (single) individuals can be subjects of intentions.25 On the other hand, a commitment to a non-reductive view including 1b) does not preclude commitment to 2a) – in other words, even if interrelated individuals are taken to be the bearers of a collective intention, their appropriate intentions can still be conceived as having the form “I intend …”.26 Finally, regarding the third element of collective intention, the question is whose action is appealed to in the intention. Assuming that it is plausible to talk of “intentions that something happen”, this appeal can be either (1) to the action of an individual, (2) to a “We” as a collection of individuals or (3) to a “We” as a collective unit of agency. It should be obvious that the first option is only open to accounts that make use of other “higher” elements in the scheme, especially those opting for 2a). This view should not be confused with a view according to which there is a collective or we-intention from which the corresponding individual (or participants”) intention is then derived. Such a view does provide a reductive analysis of collective intention, but it does not suggest replacing “We-intentions” by “I-intentions”.

25 It is in this vein that John Searle (1990) classifies his account of “we intentions”, which is meant to be compatible with his commitment to “methodological solipsism”, as non-reductive or “anti-reductionist”. 26 These two commitments are combined in Michael Bratman’s (1999) account.

Limiting Reductionism in the Theory of Collective Action

101

The other two options are more controversial. Whether in combination with 2a) – “I intend …”27 – or in combination with 2b) – “We intend …” – accounts that opt for 3b) are not definitely committed to the non-reductive interpretation of the “We” appealed to in the content of collective intention. If, as for instance in Bratman’s account, the collective intention is taken to be an intention had and shared by interrelated individuals and to be of the form “I intend that we do X”, then it would be mistaken to interpret this “we” as necessarily denoting an irreducible collective entity. Whereas such a view is indeed compatible with (if not required by) the view that a group can (literally) be the subject of a collective intention, an alternative account would take the “we” to refer not to an irreducible unit but to a collection of individuals that has the properties attributed to it only distributively.28 Once again, the decisive point seems to be whether the “we” is taken to be a placeholder for reference to interrelated individuals, or whether it is taken to stand for an irreducible collective entity. The respective view depends on the underlying social ontological commitments of an account. This rather lengthy discussion and differentiation of various forms of reductive accounts of collective intention puts us in a position to launch the critical claim of this paper. The claim of limiting reductionism is not that all reductive accounts fail, but that no single approach, reductive or nonreductive, is sufficient to account for all phenomena of collective action. It is, in other words, not as simple as just selecting one of the available combinations of elements from the scheme provided and then applying it to the analysis of all kinds of collective action. In the next section I will develop a taxonomy of action types that will serve to specify the types of cases where I think reductive accounts of collective intention are applicable (and should be applied). This will also be useful for formulating the crucial disagreement that underlies the opposition of reductionist and non-reductionist approaches.

27 See the discussion between Baier (1997), Stoutland (1997), Velleman (1997) and Bratman (1999: chap. 8), where Bratman defends the plausibility of “I intend that we …”. 28 For the definition of “collections of individuals” to which I adhere see Pettit (2003b). – The concept of “distributive properties” does not need any further technical specification at this point. I use it as the opposite of “collective properties”, which are those that can be attributed to a collective entity as such, so that “distributive properties” are the properties a collective entity (in this case, a collection of individuals) has by virtue of its members having them.

102

David P. Schweikard

4. Sorting the phenomena: A taxonomy of action types Before I sketch the taxonomy of action types I want to use to specify the suitably limited extensions of reductive accounts of collective intention, I should first mention that by talking of action types I do not intend to open the discussion about the ontological status or the ontological structure of actions. I do not refer to action types (or act-types) as being exemplified by action tokens (or act-tokens29). Rather, I take them to be classes of actions, i.e. marked off by a defining characteristic. Furthermore, for the sake of brevity and clarity, I do not intend to elaborate on taxonomies of types of collective action put forth by other authors or to comment on differences in definition or terminology.30 The taxonomy I suggest here is designed to meet, first of all, the most salient basic intuitions about the phenomena of collective action. To avoid question-begging, the claims implicit in the taxonomy – e.g., that there are genuine kinds of joint action and that there are group agents – will have to be buttressed by further argument. The organising principle of the taxonomy I suggest is the question as to how many agents are involved in an action and how they are involved. There are, of course, actions that are neutral with respect to the number of agents involved and the kind of their involvement. I can, for instance, buy a house all by myself, but this is clearly something two or more people, groups, corporations and institutional agents can do also. So in working with this organising principle, I do not attempt to taxonomise all kinds of action.31 On the standard account, an action is marked off from a mere doing by being intentional in the sense that the agent who performs it either simply wills to perform it or takes its performance to be necessary for achieving some end or other. Thus, my adding basil leaves to the pasta sauce I am preparing is an action of mine, whereas my adding a whole packet of salt is not.32

29 This view is expounded, for example, in Alvin I. Goldman’s A Theory of Human Action (1970: chap. 1). 30 Nevertheless, my classification is (in some respects) similar to the ones proposed by Stoutland (1997) and Ware (1988). 31 Thanks to Ludger Jansen for suggestions concerning this question. 32 I am aware that the use of gerund nominalizations in the examples is not entirely innocent, but I am ignoring the complications at this point.

Limiting Reductionism in the Theory of Collective Action

103

The basic distinction is the one between individual and non-individual (= collective) actions. Individual actions can be solitary, which means that they can be performed independently of external circumstances, and relational or (in Weber’s sense) social, which means that they are performed by a single individual but conditioned by or directed at other individuals. An example for solitary individual actions is my brushing my teeth; a relational individual action would be my waving to my friend. Non-individual actions include joint actions and, more controversially, group actions. (I am simply bracketing so-called mass actions here.33) Joint actions can, as Michael Bratman emphasized early on in the debate, be such that they could be performed by single individuals, such as carrying a table, painting a house or going for a walk, or they can be such that they can only be performed by two or more individuals, such as playing tennis, singing a duet or dancing a tango. Call the first contingently joint actions and the latter necessarily joint actions. Both of these latter types can take cooperative or competitive form. Finally, group actions are actions that can only be performed by groups or collectives and that are attributed to them both by their members and by external observers. Examples for this type of action are a soccer team’s scoring a goal or a committee’s making a decision. Further distinctions can be based on the question of how group actions are physically performed or carried out. Especially cases, as the soccer example, allow for both individual actions (a solo by a single player) and joint actions (a pass play or a corner kick) to figure as representative actions that are then attributed to the group (or the team). This taxonomy can be illustrated as follows:

33 It is controversial whether mass actions should also be grouped with the other types of non-individual actions. If I define mass actions as actions in which a (usually relatively large) number of people together bring about a collective outcome without (collectively) intending to do so, I do not see how they exactly qualify as actions in the sense just given. Describing collective absence from national elections – given that it is not a coordinated absence of the members of a political party – as a collective action might be helpful for some other purpose, but the analytical framework chosen here requires that there be a specifiable and specifically attributable intention to act precisely this way.

104

David P. Schweikard

actions

individual actions

sol itary actions

relational actions

non-individual actions

joint actions

contingently joint actions

group actions

necessa rily joint actions

Table 2: Taxonomy of action types

Some clarifying and specifying remarks are in place. Firstly, the taxonomy as it stands allows those cases to be pinpointed to which the initially introduced rephrasing of Wittgenstein’s question applies. The question was: what is the difference between individuals acting parallel to one another in a seemingly coordinated but actually unconnected fashion and their acting together? This difference does not come into view in types of action that presuppose at least a minimal degree of coordination (as necessarily joint actions) or even organization (as group actions). Thus, it is precisely the class of contingently joint actions, the ones Bratman terms “cooperatively neutral”34, in which it is the distinct form of intentionality that makes the difference between just seemingly coordinated and actually coordinated actions. Necessarily joint actions and group actions still require a different kind of intentionality from the one operative in individual action, but unless a collective intention is formed these types of action cannot be performed at all. Secondly, the nomenclature introduced with this taxonomy serves to avoid two terminological confusions that are manifest in a large proportion of the debate on collective action. The first is the use of the Weberian term ‘social action” to refer to all kinds of collective action. As indicated above, the opposition between individual and social action35 is one that does not comply with Weber’s use of this term, and moreover, it blurs the differences between different kinds of non-individual action just explained. The second confusion concerns the term “group action”. 34 Bratman (1999: 96-97). 35 This opposition (or at least this distinction) is emphasized in the titles of two recent compendia on philosophy of action (Holmström-Hintikka and Tuomela 1997a; 1997b).

Limiting Reductionism in the Theory of Collective Action

105

It is tempting to use it for all kinds of non-individual action, since this reflects the broad range of cases in which the term “group” is applied in everyday discourse. But philosophically, simply adapting this use is unsatisfactory, for it turns an important sociological, social psychological and social ontological concept into an uninteresting generic term for all social associations, from spontaneous gatherings to complex social structures. These confusions might at least in part be traced back to one or the other of the reductive strategies as they were specified in the last section. Although the connection is not necessary, one might hold the view that since only individuals can have intentions and only individuals (as individuals) can act, any reference to “collective action” or “group action” is metaphorical anyway, so there is no need to be more precise about internal distinctions within that category. But even this argument is not necessarily conclusive, for even a reductionist about collective intention could argue that actions can be attributed to groups, even though groups do not count as agents in the full sense. Such considerations, however, miss the intent of the given taxonomy in that they take it to be optional whether we just talk about names or sortals for phenomena of collective action, or whether we talk about the respective (token-) events themselves. My aim in developing this taxonomy is to classify events and not just to sort out the ways of describing or redescribing individual actions. With the taxonomy at hand, I can proceed to the more constructive part of this paper and look at the classes of action to which reductive strategies are plausibly applicable. As I said at the outset, my claim is not that reductive accounts of collective intention and so of collective action fail outright, but that they should be limited in scope. Consequently, the next task is to say more precisely what kind of reductive account applies to what type of action, thereby mapping (at least the most salient of) the reductive strategies onto the taxonomy of action types.

5. When to be a reductionist Based on the first table given in section two (Table 1), we can distinguish twelve accounts of collective intention, although not all of these accounts are consistent. In this section I consider five of these accounts and explain briefly why I think each of them is applicable only to certain types of action. The condition of adequacy with which I will work in this section is whether the accounts under consideration can answer the core questions of the theory of collective action that I formulated in the first section. The re-phrasing of Wittgenstein’s question highlights that what is to be ac-

106

David P. Schweikard

counted for is the differentia specifica; the second question turns on the success condition of collective action. My aim is to specify the types of action, to which the accounts considered are applicable in the sense that they can spell out both the specific intentionality and the success condition of those types of action. i) The first to consider is the most consequently reductive account which combines the elements 1a), 2a) and 3a); on this account, a collective intention is reducible to intentions of individuals, taken singularly. It stands in the I-mode and it appeals only to the individual action of the respective individual. This account denies that there is anything specific about collective intentions by saying that they are reducible to individual intentions. These intentions, then, do not appeal to individuals others than those whose intentions they are, neither in mode, nor in content. Thus the claim is that there is nothing specific about the intentionality of collective action. However, what about the success condition? Can such an account explain what guarantees the successful performance of a collective action? – I do not think it can. Without appeal to other (possibly) involved individuals, this theory of intention provides no clue to how individuals could actually manage to act together, how they could coordinate at all, not to mention fine tuning during the performance. Therefore, I think an account of this kind is effectively applicable only to solitary and relational individual actions. The maximum of sociality that can be built into this kind of intention is in the type of individual action which can be such that it is directed at others (e.g. waving to someone). ii) The second account employs the elements 1a), 2b) and 3b), i.e. it claims that a collective intention is (reducible to) intentions of individuals, taken singularly, that it stands in the We-mode and that it appeals directly to the collective action. Here, the appeal to others, which may be spelt out by reference to beliefs about others” intentions or beliefs,36 is present in the entire form of the intention. This specific appeal makes for the specific difference of the intentions involved in collective action. But is it plausible to assume that individuals, taken singularly, intend in the We-mode? Is it convincing to say, as Searle does,37 that this mode of intention is irreducible if it is the mode of an intention that completely isolated individuals can have? – How we answer these questions depends on the stand we take concerning the underlying commitments that inform this account. Independent of that discussion, we can still ask whether such an 36 Just to note this systematic place: this is where considerations of such complex conceptions as “mutual belief” or “common knowledge” enter the scene. 37 See Searle (1990).

Limiting Reductionism in the Theory of Collective Action

107

account can explain the circumstances under which a collective action is performed successfully. And here I think the adherence to “methodological solipsism” does not explain but rather mystifies the workings of collective action. Consequently, it is only applicable to relational individual actions. iii) A third account, the one I take to be closest to Michael Bratman’s account of what he calls ‘shared intention”, makes use of the elements 1b), 2a) and 3b). Spelt out, it claims that a collective intention is an intention had by interrelated individuals who each have an intention of the form “I intend that We do …”, i.e. an intention in the I-mode appealing to the collective action. As for the question concerning the specific features, this account interprets collective intention as an intention shared by interrelated individuals, each of whom appeals to the joint endeavour in their respective individual (I-mode) intentions. Furthermore, it accounts for the success of collective action by treating the interrelations between individuals and the mutual reference between the individual intentions (of which individuals have to be aware) as a necessary and sufficient condition for the successful performance of collective action. In my view, this account is applicable to all kinds of joint action. It does not, however, grant special status to group actions, since it does not postulate or intend to show that groups can count as agents. iv) Only slightly differing from the third account, the fourth account I want to mention is the one that combines the elements 1b), 2b) and 3b). The difference is that on this account the intentions had by interrelated individuals stand in the We-mode. This difference is due to the way these accounts conceive the realization of a collective intention that is shared by interrelated individuals. It might not yield different results in the analyses of joint action, but neither is this account including 2b) subject to the same charge of circularity.38 Put briefly, the charge is that account iii) presupposes the we-sociality of which it says it is only appealed to in the content of individual intentions. Circularity lurks if it is claimed that this appeal to collective action does not presuppose something like a sense of collectivity. And this is exactly what account iv) grants. That interrelated individuals are taken to intend in the We-mode implies that they are taken to conceive the collective action as an action performed by “Us”. Moreover, the claim is that this holds not only for necessarily joint actions, but also for contingently joint actions.39 38 For this charge against Bratman’s account see Schmid (2005a). 39 There are two related issues that cannot be dealt with at length in this paper. The first is the difference already mentioned above between conceiving the interrela-

108

David P. Schweikard

v) Finally, I would like to mention the account that does not make use of any reductive elements but combines 1c), 2b) and 3b). Here, it is claimed that a group can be the subject of an intention and that in that case, individuals have intentions in the We-mode which appeal directly to the collective action. This account is tailored (precisely) to group actions. Obviously, postulating this class of action and referring to a group’s intention is not enough to prove that groups have to be considered as agents. Even if examples such as the soccer example given above are intuitively plausible and even if it seems as though we quite naturally refer to group actions in everyday talk, this does not prove the case for a non-reductive account of collective intention. Here the burden of proof seems to have been shifted to the social ontological question concerning the existence of groups. I will come back to this question in the next section. But it should be clear at this point that due to its inclusion of 1c) this account is not applicable to other types of non-individual action. To sum up, we can map these accounts on the table of action types: actions individual actions sol itary actions

i)

non-individual actions

relational actions

ii)

joint actions

contingently joint actions

group actions

nec essa rily joint actions

v)

iii) an d iv)

Table 3: When to be a reductionist

tions between individuals as being represented only by semantic interconnections between the individual intentions as ontologically constitutive relations. Secondly, a lot more could be said about the dimensions of the “We” alluded to in this passage. One interesting question on this count is whether the “We” as representing a sense of community is fundamental and pre-reflexive or generated by reflexive processes; here Hans Bernhard Schmid (2005b), for instance, argues for the first option, whereas a Bratmanian account as the one in iii) would support the latter option.

Limiting Reductionism in the Theory of Collective Action

109

This is the scheme I want to suggest for limiting reductionism in the theory of collective intention. It consists in limiting the scope of reductive accounts of collective intention to those types of action with respect to which they can answer the core questions defined above. By outlining the reasons for these limitations, I have argued for the claim that none of these reductive accounts is applicable to all types of action, especially not to all types of non-individual (or collective) actions. What follows from this is, up to this point, that the theory of collective action is well equipped to tailor the description or reconstruction of collective intention to the type of action it is meant to guide.

6. Implications for Social Ontology Any theory of collective action is concerned with social ontology right from the start, for it cannot but face the question as to whether collectives can act.40 One way of answering this question about the ontological commitment implied in the ascriptions of actions is to say that all reference to acting collectives is nothing but a makeshift or metaphorical reference for individuals (somehow) acting together.41 With regard to collective intention, an analogous reaction can be observed in the attempt to deny the possibility that groups have intentions. My contention in the preceding sections was that this answer is unsatisfactory since it falls short of an important part of the range of phenomena to be considered. It can be shown – and this is my positive answer – that there are types of action only groups can perform and there are ways in which groups can be organized that ground speaking of such groups as agents in their own right. While the first part of this answer rests on the appeal to its intuitive plausibility, the second part needs further argument and illustration. Before I turn to this I shall briefly comment on the reductive intuitions that lead to the answer I reject. On what grounds is the idea that collectives can be agents or subjects of intention repudiated? What motivates or even requires reductionism about collective entities? – As indicated above, I take these to be typical questions on whose answers implicit commitments in other fields of philosophy make a decisive impact. In view of the discussion about collective intention, the crucial issue is probably the nature of intentional states. 40 In this section, I will use the terms “group”, “collective” and “collective entity” interchangeably. 41 An example for such an eliminativist account is the view famously expressed by Anthony Quinton (1975).

110

David P. Schweikard

Here, strict internalism grounds methodological solipsism, which in turn is smoothly compatible with social ontological individualism.42 But the question is whether the overall consistency of an account that is developed starting from issues in the philosophy of mind should be valued higher than its plausibility with respect to social phenomena. It may be that these issues, the nature of intentional states and the ontological status of groups, are independent of one another, but internalism about mental content cannot then serve as an argument against the possibility that groups have intentions. Nevertheless, this seems to be a line John Searle is inclined to take when he warns against postulating some (at best) mysterious supraindividual mind floating over our heads.43 An important step in the debate on collective action as well as in the more general discussion about collective entities would be to lay open such underlying commitments and argue for the plausibility of their extensive impact. Alternative lines one can take on this issue consider the vernacular reference to groups as agents and the processes of their formation, the structure of standard ways of interpreting certain groups as agents or group-internal procedures that guide collective decision-making. While it is certain that these approaches make use of some notion of group agent, it is uncertain whether defenders of these approaches would agree to be committed to the irreducible existence of groups. Due to lack of space I cannot not go into the details of all the accounts alluded to, i.e. Margaret Gilbert’s plural subject theory44 and Deborah Tollefsen’s interpretationist account,45 but will limit my remarks to the social ontological implications of the widely debated account of judgement aggregation that Philip Pettit has adapted for the study of social groups.46 The account departs from the well-known “discursive dilemma” which emerges for a group when the connection of majority votes on the premises of an argument contradicts the majority vote on the conclusion.47 Pettit 42 I do not insist on the order or the exact terminology used in this phrase, as long as the idea about close interconnections of these issues gets across. 43 Cf. Searle (1990: 404; 2000: 118). 44 Cf. Gilbert (1989; 1996; 1997; 2000). 45 See Tollefsen (2002a; 2002b). 46 See Pettit (2001; 2002; 2003a; 2003b). 47 More perspicuously, a dilemmatic case would be one in which in a group of three, A, B, and C, each casts her vote on an issue of the form P∧Q→R (where P, Q, and R are separate propositions). If only A and C say that P holds, where B says is does not, and if only B and C say that Q holds, where A says it does not, then majority voting on the conjunction of P and Q, R, would yield a negative result, whereas taking the majority votes on the separate propositions P and Q would logically necessitate the affirmation of the conjunction. Even though every member of the

Limiting Reductionism in the Theory of Collective Action

111

claims that “[t]here is a type of organization found in certain collectivities that makes them into subjects in their own right, giving them a way of being minded that is starkly discontinuous with the mentality of their members.”48 This certain “type of organization”, so the argument runs, solves the discursive dilemma by ensuring that the decisions made collectively by the members of such groups fulfil the conditions of coherence and consistence over time that make for rationality on the group level. If a group is structured in respect of this ideal of rationality, it is accepted that a group decision may diverge from the decision which a majority of its members individually would have favoured. The discontinuity between the members” and the group’s decisions or judgements is spelt out as a supervenience relation: the group decision supervenes on the individual decisions, meaning that “if we replicate how things are with and between individuals in a collectivity […] then we will replicate all the collective judgments and intentions that the group makes.”49 This supervenience relation is claimed to hold with respect to group decisions, judgements and intentions, but what about the ontological status of groups? Here, Pettit claims programmatically that “groups that collectivize reason deserve ontological recognition as intentional and personal subjects.”50 Such ontological recognition is taken to be justified by the fact that these groups “have a rational unity that constrains their performance over time and that makes them distinct from their own members.”51 It is the relation between the idea of supervenience and the idea of distinctness that I want to draw attention to.52 Although they differ with respect to

48 49

50 51 52

group is perfectly consistent, the two procedures of aggregating the votes generate contradictory results. Pettit (2002: 443). Pettit (2002: 460). It is important to note here that Raimo Tuomela invested a lot of analytical effort much earlier into spelling out such relations between the individual level and the group level. At this stage it would lead too far off to reconstruct the framework of this complex account, but see Tuomela (1989). Pettit (2002: 451). Pettit (2002: 463). Pettit adopts a conception of “rational unity” that is close to the one developed by Carol Rovane (1998). Compare this other place where the specific distinctness is appealed to: “The integrated collectivity will not be distinct from its individual members, in the sense that it will not be capable of existing in the absence of such members. But it will be distinct in the sense of being a centre for the formation of attitudes that are capable of being quite discontinuous from the attitudes of the members. This is one of the lessons of the discursive dilemma.” Pettit (2003a: 184)

112

David P. Schweikard

their ontological implications,53 both ideas seem to express anti-reductive conceptions of the ontological status of groups. The upshot of Pettit’s account of groups for the context of the present paper is that there are certain types of groups that can be counted as subjects of intention. On condition that their internal organization ensures the coherence of their decisions, i.e. rationality on the group level, these groups should be recognized as intentional agents. What can be disputed, however, is the range of applicability of these procedures for collectivizing reason that ensures the “distinct” status of groups. The argument based on the discursive dilemma is based on assumptions that seem to limit its scope. It is assumed that: i. the individual decisions, which are then aggregated to form the group decision, are made in complete isolation from one another. Absent of all social hierarchies, this blind voting entails that the individuals who, for instance, vote on a set of interconnected issues, are assumed not to pay attention to the voting behaviour of others. ii. among the members of the group there is complete agreement on the way the entirely distinct issues are connected and how the conclusion is drawn. In other words, they agree exactly as to what the overall issue is and into which sub-issues it can be divided. These assumptions go through quite smoothly in the case of a collegiate court where two independent facts of a case, e.g. cause of harm and duty of care,54 have to be decided upon and the way of drawing a conclusion is prescribed by law (or a practice of deciding cases of a certain type). But generalizing this case to all sorts of social groups seems difficult precisely because neither deciding in isolation nor total agreement on the connection of issues seems to be something natural to assume. Thus, to conclude this brief excursus, there are some indications that further elements are needed to ground a more general account of group action and of groups as irreducible collective entities. On my account, there are four elements that need further consideration: i) Instead of assuming a kind of ideal rationality at the basis of collective decision-making, I would suggest spelling out a conception of pragmatic rationality. Coined with regard to what was said about the assumptions of the argument from the discursive dilemma, this kind of rationality does not categorically require total agreement about the connection of issues, 53 For the notion of “ontological innocence” and its relation to supervenience see Bennett and MgLaughlin (2005). 54 Cf. Pettit (2003a).

Limiting Reductionism in the Theory of Collective Action

113

but it sets priority on the goals for whose attainment the group is formed. In the light of their shared goals and in favour of the possibility of realizing them, it is rational of the individual members to take other opinions into account and to allow for partial internal disagreement. And in view of changing circumstances, it can be rational for the collective to accept inconsistency over time as long as the attainment of the group’s goals is not endangered.55 ii) This conception of rationality is closely connected to a revised understanding of important aspects of the sociality of individuals. Calling them “group members” does not do the trick, since this leaves unclear what is required for membership and what it entails for the individuals. Rather, what it means for individuals to commit themselves to and identify with the goals or the views of a certain collective needs to be spelt out.56 For this is what enables them to use the pronoun “We” in a robust sense, referring to the group, of which they are members. In the context of collective action this grounds the shift from intending in the I-mode to intending in the We-mode; it is an essential prerequisite of group action. iii) Given a suitably complex internal structure, a strong commitment to joint goals and a high level of identification with the group on part of its members, the basis is laid for treating collectives as agents both in the attribution of actions and in the interaction with them.57 With this we complement our inquiry into the nature of collective intention with reflections on the practice of acknowledging groups as agents. iv) This acknowledgement is, however, in practice not on a par with the ontological recognition of groups as irreducible entities. This is the point at which the other elements converge and are supplemented with an ontological account of collective entities. The basic idea is that the relation between the individual members of a group and the group is best described as a constitutive relation. This relation requires (1) that individuals be (horizontally) interrelated in certain ways so that they together constitute a collective entity, (2) that these individuals be (vertically) related to the group and (3) that the collective entity have certain (intentional) properties that are attributed to it as such and not in virtue of some properties of the individuals that constitute it. For this view, based as it is on constitutive relations to be an ontological understanding of collective en55 Obviously. this last point is contained in Pettit’s illustration of the “modus tollens generalization”. See Pettit (2003). 56 Cf. Gilbert (1989). 57 This is meant as an integration of important insights of Tollefsen’s interpretationism.

114

David P. Schweikard

tities, the interrelations cannot just be semantic interconnection as, for instance, between the content of mutually referring intentions. The relations must be such that they provide the relata with properties they could not acquire outside these relations. This much for a brief sketch of how I think one could argue for the irreducible existence of groups that can count as agents. While this account is not restricted to cases where conditions of ideal rationality can be (and are in fact) met, neither it is intended to serve as basis for a theory of social community. I do not claim to have accounted for all characteristics of groups or of group action. In particular, I have bracketed social dynamics internal to groups as they are studied in social psychology, but I do claim that the ones mentioned are the necessary features of groups which are to be counted as agents. The view established thereby, i.e. the view that the capability to act is not reserved for singular individuals, can be labelled non-singularism.

7. Conclusion Finally, I should come back to the issue of reductionism and merge the differentiations suggested in the course of this paper. The aim has been to show that reductive accounts of collective intention are applicable only to certain types of collective action. To specify and make plausible this claim of limiting reductionism I have offered taxonomies both of reductive accounts of collective intention and of action types. The result obtained is that in order to account for the success of collective action, it is necessary that the relevant intention be conceived as an intention that is either shared by interrelated individuals or had by a group agent. Understanding collective intention as an intention held by isolated, singular individuals distorts the phenomenon of collective action by leaving it to mere coincidence whether an action is successfully performed. What remained after defining and illustrating the limits of reductive accounts was to pick up the thread that linked issues of collective action and of social ontology. In fact, developing an understanding of the sociality of individual agents, of the interrelations they enter into explicitly for the purposes of achieving common goals and of the nature of collective entities, is central to the philosophy of collective action. Concerning the latter issue, a sketch of a constitution view of groups has been given. Of course, the details of this view still need to be elaborated and it needs just

Limiting Reductionism in the Theory of Collective Action

115

as much further defence as the action theoretic framework in which it was placed, but it is worthwhile to have an overview of this framework.58

References Baier (1997): Baier, A.C. Doing Things With Others: The Mental Commons, in: Alanen, L., S. Heinämma, and T. Wallgren (eds.), Commonality and Particularity in Ethics, London: MacMillan, 1997, pp. 15-44. Bennett and McLaughlin (2005): Bennett, K. and McLaughlin, B. Supervenience, in: Zalta, E.N. (ed.), Stanford Encyclopedia of Philosophy, URL = http://plato.stanford.edu/entries/supervenience/. Bratman (1999): Bratman, M. Faces of Intention: Selected Essays on Intention and Agency, Cambridge: Cambridge University Press, 1999. Chant (2006): Chant, S.R. The Special Composition Question in Action, in: Pacific Philosophical Quarterly (2006) 87/4: 422-441. Gilbert (1989): Gilbert, M. On Social Facts, Princeton: Princeton University Press, 1989. Gilbert (1990): Gilbert, M. Walking together, in: Gilbert, M. Living Together: Rationality, Sociality, and Obligation, New York: Rowman and LittlefieldGilbert , 1996, pp. 177-194. Gilbert (1996): Gilbert, M. Living Together: Rationality, Sociality, and Obligation, New York: Rowman and Littlefield, 1996. Gilbert (2000): Gilbert, M. Sociality and Responsibility: New Essays in Plural Subject Theory, New York: Rowman and Littlefield, 2000. Gilbert (2003): Gilbert, M. The Structure of the Social Atom: Joint Commitment as the Foundation of Human Social Behaviour, in: Schmitt, F. (ed.), Socializing Metaphysics – The Nature of Social Reality, Lanham: Rowman and Littlefield, 2003, pp. 39-64. Goldman (1970): Goldman, A.I. A Theory of Human Action, Princeton: Princeton University Press, 1970. Holmström-Hintikka and Tuomela (1997a): Holmström-Hintikka, G. and Tuomela, R. (eds.), Contemporary Action Theory, Vol I: Individual Action, Dordrecht and Boston: Kluwer Academic Publishers, 1997. 58 An ancestor of this paper was presented at the Conference on Collective Intentionality V, in Helsinki in August 2006. Thanks to Hans Bernhard Schmid and Michael Quante for discussions.

116

David P. Schweikard

Holmström-Hintikka and Tuomela (1997b): Holmström-Hintikka, and Tuomela, R. (eds.), Contemporary Action Theory, Vol II: Social Action, Dordrecht and Boston: Kluwer Academic Publishers, 1997. Kannetzky (2005): Kannetzky, F. Konsequenzen des Privatsprachenarguments für Theorien gemeinsamen Handelns, in: Kober, M. (ed.), Soziales Handeln – Beiträge zu einer Philosophie der 1. Person Plural, Humboldt-Studienzentrum, Universität Ulm, 2005, pp. 115-129. List and Pettit (2006): List, Ch. and Pettit, Ph. Group Agency and Supervenience, in: Southern Journal of Philosophy (2006) XLIV (Spindel Supplement): pp. 85-105. Ludwig (2007): Ludwig, K. Foundations of Social Reality in Collective Intentional Behavior, in: Tsohatzidis, S.L. (ed.), Intentional Acts and Institutional Facts – Essays on John Searle’s Social Ontology, Dordrecht: Springer, 2007, pp. 49-71. Pettit (2002): Pettit, Ph. Collective Persons and Powers, in: Legal Theory (2002) 8: pp. 443-470. Pettit (2003a): Pettit, Ph. Groups with Minds of their Own, in: F. Schmitt (ed.), Socializing Metaphysics – The Nature of Social Reality, Lanham: Rowman and Littlefield, 2003, pp. 167-193. Pettit (2003b): Pettit, Ph. Akrasia, Collective and Individual, in: S. Stroud, S. and C. Tappolet (eds.), Weakness of Will and Practical Irrationality, Oxford: Oxford University Press 2003, pp. 68-97. Pettit and Schweikard (2006): Pettit, Ph. and Schweikard, D.P. Joint Actions and Group Agents, in: Philosophy of the Social Sciences (2006) 36/1: 18-39. Quinton (1975): Quinton, A. Social Objects, in: Proceedings of the Aristotelian Society 75 (1975): 1-27. Rovane (1998): Rovane, C. The Bound of Agency – An Essay in Revisionary Metaphysics, Princeton: Princeton University Press, 1998. Schmid (2003): Schmid, Hans Bernhard. Can Brains in Vats Think as a Team? in: Philosophical Explorations (2003) 6/3: 201-218. Schmid (2005a): Schmid, H.B. Wir-Intentionalität - Kritik des ontologischen Individualismus und Rekonstruktion der Gemeinschaft, Freiburg: Alber 2005. Schmid (2005b): Schmid, Hans Bernhard. Wir-Identität: reflexiv und vorreflexiv, in: Deutsche Zeitschrift für Philosophie (2005) 53/3: 365-376.

Limiting Reductionism in the Theory of Collective Action

117

Searle (1990): Searle, J.R. Collective Intentions and Actions, in: Cohen, P., J. Morgan, and M.E. Pollack (eds.), Intentions in Communication. Cambridge/Mass.: Bradford Books, MIT Press 1990. pp. 401-416. Searle (1995): Searle, J.R. The Construction of Social Reality, New York: The Free Press, 1995. Searle (2000): Searle, J.R. Mind, Language and Society – Philosophy in the Real World, London: Phoenix, 2000. Stoutland (1997): Stoutland, F. Why are Philosophers of Action so AntiSocial? in: Alanen, L., S. Heinämma, and T. Wallgren (eds.), Commonality and Particularity in Ethics, London: MacMillan, 1997, pp. 45-74. Tollefsen (2002a): Tollefsen, D. Collective Intentionality and the Social Sciences, in: Philosophy of the Social Sciences (2002) 32/1: 25-50. Tollefsen (2002b): Tollefsen, D. Organizations as True Believers, in: Journal of Social Philosophy (2002) 33/3: 395-401. Tuomela (1989): Tuomela, R. Collective Action, Supervenience, and Constitution, in: Synthese (1989) 80: 243-266. Tuomela (2000): Tuomela, R. Collective and Joint Intention, in: Mind and Society (2000) 1/2: 39-69. Tuomela (2002): Tuomela, R. The Philosophy of Social Practices – A Collective Acceptance View. Cambridge: Cambridge University Press, 2002. Tuomela (2003): Tuomela, Raimo. The We-mode and the I-mode, in: Schmitt, F. (ed.), Socializing Metaphysics: The Nature of Social Reality, Lanham: Rowman and Littlefield 2003, pp. 93-127. Tuomela (2006): Tuomela, R. Joint Intention, We-Mode and I-Mode, Midwest Studies in Philosophy (2006) 30/1: 35-58. Tuomela and Miller (1988): Tuomela, R. and Miller, K. We-Intentions, in: Philosophical Studies (1988) 53: 367-389. Velleman (1997): Velleman, J. D. How to Share an Intention, in: Philosophy and Phenomenological Research (1997) 57: 29-51. Vermazen (1993): Vermazen, B. Objects of Intention, Philosophical Studies (1993) 70: pp. 85-128. Ware (1988): Ware, R. Group Action and Social Ontology, in: Analyse und Kritik (1988) 10: 48-70. Wittgenstein (PI): Wittgenstein, L. Philosophical Investigations, translated by G.E.M. Anscombe, Englewood Cliffs/N.J.: Prentice Hall, 1958.

The Status Account of Corporate Agents1 Frank Hindriks, Groningen In the literature on social ontology, two perspectives on collective agency have been developed. The first is the internal perspective, the second the external one. The internal perspective takes the point of view of the members as its point of departure and appeals, inter alia, to the joint intentions they form. The idea is that collective agents perform joint actions such as dancing the tango, organizing prayer meetings, or performing symphonies. Such actions are generated by joint intentions, a topic which has been investigated extensively in recent decades.2 The external perspective takes the point of view of outsiders as its point of departure. Here the idea is that outsiders ascribe certain properties to collections of individuals that thereby acquire a new status. An example that is discussed in some detail below is provided by universities that, at least initially, had to be recognized by the pope in order to perform the actions that are characteristic of universities, in particular the awarding of academic degrees. This perspective has received considerably less attention.3 In this paper I present and defend a new account of collective agents from the external perspective. I call it “the status account”. At the heart of this account lies the idea just mentioned. A collective agent can come into existence by outsiders granting a collection of individuals some kind of status. I call a collective agent for which this holds a “corporate agent”, because such an agent is or has been, in a sense to be explained, incorporated. I develop this idea by distinguishing between two kinds of rules: status rules and constitutive rules. Status rules explicate the normative attributes that are characteristic of a particular kind of corporate agent. 1 Discussions with Lynne Rudder Baker, Uskali Mäki, Barry Smith, and Raimo Tuomela have contributed greatly to the substance of this paper. I thank them very much for the opportunities to benefit from their insights. I also thank Siegfried Van Duffel, Francesco Guala, Heikki Ikäheimo, and the editors for helpful comments. 2 See Bratman (1999), Gilbert (1989; 1996), Searle (1990; 1995), Tuomela (1995; 2005), and Tuomela and Miller (1988) for the most prominent accounts of joint intentions. 3 As we shall see below, its most prominent proponent is Searle (1995, 2005). Tollefsen (2002) develops an account of the intentionality of organizations from an external perspective. N. Psarros, H. B. Schmid and K. Schulte-Ostermann, Concepts of Sharedness, Frankfurt 2008: 119-144; ontos verlag

120

Frank Hindriks

Constitutive rules specify the conditions that have to be met for a certain collection of individuals to have those attributes. The status account shares some important features with John Searle’s social ontology.4 The advantages that my account has over Searle’s are discussed in section 2.2, immediately after the account has been introduced in section 2.1. Before doing that, however, I shall discuss the internal perspective. My claim is that, in addition to joint intentions, a collection of individuals needs to have developed or accepted a collective decision mechanism in order to really be a collective agent. Following Philip Pettit, I call such agents ‘social integrates”.5 Against Pettit, however, I argue that this mechanism need not meet particular conditions of collective rationality. Insisting on this requirement rules out many organizations that are plausibly regarded as collective agents. The relation between the internal and the external perspectives, between social integrates and corporate agents, is discussed in section 3. By and large, the two accounts I defend should be regarded as complementary.

1. The internal organization perspective Consider Amy and Bob. Amy and Bob are walking together. Do they form a collective agent? Margaret Gilbert answers this question in the affirmative.6 Amy and Bob perform a joint action. This joint action is generated by a joint intention. During the process of forming a joint intention, Amy and Bob have formed a collective agent, or in Gilbert’s terms “a plural subject”. This process can roughly be explicated as follows. Individual agents can form a plural subject by mutually expressing their willingness to be jointly committed to the performance of a particular action in conditions of common knowledge.7 They are committed to it “as a body”. Because they are jointly committed to walk together, it is appropriate for Bob to criticize Amy if she happens to disregard him and walks too fast. As Gilbert puts it, he is entitled to rebuke her. According to Gilbert, forming a joint intention is sufficient for forming a collective agent.8 It is not obvious that this is correct, though. The one4 5 6 7 8

Searle (1995). Pettit (2003). Gilbert (1989; 1996). Gilbert (1996: 349). This is not a necessary condition, in Gilbert’s view, because plural subjects also come into existence when some other joint attitude is formed, such as a joint belief.

The Status Account of Corporate Agents

121

time performance of the action of going for a walk together seems to be too fleeting to ground the existence of a new kind of agent. Rather than forming a new entity, Amy and Bob just join forces temporarily. Michael Bratman uses the terms ‘shared cooperative activity” and ‘shared agency” in this connection.9 These terms seem apt to describe what is going on and do not require us to postulate a collective agent. I believe we should refrain from doing so, because in cases such as the walking example just discussed it does not seem right to say that the decision to perform the joint action at issue is made by a collective agent. Instead, the joint intention simply originates from the individuals involved.10 As collective agents perform joint actions, there is no reason to doubt that the formation of a joint intention is a necessary condition for forming a collective agent, at least as characterized from the internal perspective. What must be added in order to arrive at an adequate conception of collective agency? Philip Pettit has provided an answer to this question that focuses on the way in which a collection of individuals forms collective judgments. The motivation for his approach is that without taking special precautions, collections of individuals can easily end up forming inconsistent collective judgments. It is Pettit’s view, though, that being responsive to the demand of consistency in the formation of beliefs is a requirement for being an agent. Let me elaborate. Consider three individual workers A, B, and C who have to decide to forego a pay-raise in order to introduce some workplace safety measures. They make this decision by considering three questions. First, are the workers faced with a serious danger if the safety devices are not installed? Second, are the proposed measures effective? And third, is foregoing the pay-raise a bearable loss? They rely on majority voting. The safety measures will only be introduced if a majority is in favor of doing so. It appears that whether a majority is in favor of doing so can be determined in two ways. 9 Bratman (1999). 10 Pettit and Schweikard do not regard the formation of a joint intention as sufficient for the formation of a collective agent either, as is evidenced by their claim that “it is we severally who intend that we act together, not we in the sense in which we might constitute a subject proper” (2006: 29-30). However, they reject this option for a more theoretical reason: according to their view, a joint action is (or need) not be generated by a single intention. They make this point as follows: “We see no metaphysical reason why a joint intentional action has to be the product of a single agent or a single state of intending”. (Ibid., 30; emphasis added). I do not think this is correct. It appears to be more natural to say that it is the (single) joint intention that gives unity to the behaviors of the individuals involved and is what makes it proper to describe them as elements of a single joint action.

122

Frank Hindriks

The first way is by voting on the three issues and doing what is implied by the votes. Only if a majority expresses a favorable opinion for each of the three issues will the individuals involved forego the pay-raise and introduce the safety measures. The second way is by voting directly on the issue whether to forego the pay-raise and introduce the safety measures. As illustrated in table 1, the outcomes of these two procedures can differ from one another. This in turn means that if they were to vote both on the overall issue and on its components, an inconsistency would arise.11 Table 1 Effective measure? q Yes

Bearable loss? r Yes

Pay sacrifice?

A

Serious danger? p No

B

Yes

No

Yes

No

C

Yes

Yes

No

No

Majority Yes

Yes

Yes

No

No

Pettit calls this problem of inconsistency “the Discursive Dilemma”. The problem is called “discursive” because it pertains to a process of reasoning involving what can be regarded as premises and a conclusion (p, q, and r in table 1 are the premises and their conjunction is the conclusion). It is taken to be a dilemma because it is regarded as forcing a choice on us between one of two ways of avoiding the inconsistency. The first is to vote only on the premises and to accept whatever is implied by the majority votes on them as the decision of the collective. This is the premise-based procedure. The second is to vote only on the conclusion and arrive at a collective decision in this way. This is the conclusion-based procedure. Pettit holds that the formation of a collective agent requires opting for one of these solutions to the problem.

11 In Pettit’s view such inconsistencies only pose a threat to collective agents, and not to individuals who want nothing more than to perform a joint action: “No plausible analysis of joint action, and none in the literature, requires that just for the purposes of acting jointly people have to take precautions against the appearance of such inconsistencies.” (Pettit and Schweikard 2006: 31).

The Status Account of Corporate Agents

123

By adopting one of these procedures, Pettit argues, the collection of individuals at issue develops “a mind of its own”.12 The easiest way to see what Pettit means by this is to consider the premise-based procedure. It can result in the collective accepting a proposition and thereby making a decision that none of its individual members support. Note that the key to a group having a mind of its own is a lack of responsiveness to individual views on some matter. Such a lack of responsiveness is also present in the conclusion-based procedure. After all, that procedure ignores the fact that each premise is supported by a majority of the individuals, possibly by all of them. In early papers on the topic, Pettit singles out the premise-based procedure and argues that if a collection of individuals that forms joint intentions has adopted this procedure there are good reasons to regard it as a collective agent.13 After all, it has “collectivized reason”, as Pettit put it: it has imposed consistency at the collective level by accepting as its decision what is logically implied by the collective judgments on the premises. In a more recent paper co-authored with David Schweikard, this focus on one of the two methods has disappeared. Presumably, the idea is that either one will do.14 The conception of collective agency presented there is as follows: A group of individuals will constitute an agent, plausibly, if it meets conditions like the following. First, the members act jointly to set up certain common goals and to set up a procedure for identifying further goals on later occasions. Second, the members act jointly to set up a body of judgments for rationally guiding action in support of those goals, and a procedure for rationally developing those judgments further as occasion demands. And third, they act jointly to identify those who shall act on any occasion in pursuit of the goals, whether they be the group as a whole, the members of the group individually, certain designated members, or certain agents that the group hires.15 In addition to forming joint intentions, then, Pettit and Schweikard regard the formation of joint goals and rational judgments as necessary conditions for collective agency. Furthermore, the formation of rational judgments must be systematized by adopting the premise-based or the conclu-

12 13 14 15

Pettit (2003). See also Pettit and Schweikard (2006). Pettit (2001; 2003). See also Pettit (2007). Pettit and Schweikard (2006: 33).

124

Frank Hindriks

sion-based procedure. Finally, the individuals jointly determine who is to perform a particular action. This conception of collective agency is rather attractive. Collections of individuals who satisfy all these conditions form collectives that are certainly not fleeting (which is not to say that they need to be in existence for a long period of time). In this respect, they differ from Amy and Bob, who perform only a single joint action. So Pettit’s account solves the problem from which Gilbert’s account suffers. However, it appears to be too strong. The first problem is that it excludes organizations that are governed in an authoritarian way. Instead of relying on majority voting it might be that a single individual makes all the decisions by considering only her own views on the matter. In the literature on the Discursive Dilemma this decision procedure is called “dictatorship”. If the so-called dictator is sensitive to consistency, then so will the organization be, too. Some organizations that are run in this way may be more effective than some others run in a democratic way. Intuitively, there is no reason not to regard them as collective agents. After all, such organizations act and they consist of several individuals. Dictatorships do not display the strong discontinuity between collective and individual judgments that Pettit appeals to in order to argue that collections of individuals can have a mind of their own. This only suggests, however, that (the possibility of) such a discontinuity is not a necessary condition for collective agency. Another consideration provides further support for this hypothesis. By requiring consistency, Pettit and Schweikard in effect develop a conception of rational collective agency rather than of collective agency per se. Consider the following passage: Did individuals come together in the manner characterized, then they would be in a position as a group to mimic or simulate the performance of an individual agent. The group would have goals corresponding to individual desires, judgments corresponding to individual beliefs, and just as rational individuals act so as to satisfy their desires according to their beliefs, so this group would be able to act rationally so as to achieve its ends according to its judgments.16 This reveals that the target of analysis is rational action and rational agency rather than mere agency.17 In other contexts, Pettit invokes the same considerations in order to develop a conception of collective personhood, 16 Pettit and Schweikard (2006: 33-34; emphasis in original). 17 See Pettit (1993) for an analogous distinction between mere intentional agents and thinking agents.

The Status Account of Corporate Agents

125

or of collective agents that can be regarded as bearers of moral responsibility.18 The purpose of this paper is more modest than that. It is to investigate the conditions for ordinary collective agency, not for its more demanding relatives. Of course, this second argument can only succeed if there is an intermediate option, one that is stronger than Gilbert’s account and weaker than Pettit’s. I shall now argue that there is such an alternative. Instead of a procedure for forming rational judgments, mere collective agents should be required to have a procedure or a mechanism for making decisions or forming intentions, presumably against the background of some overarching goal the group wants to pursue. In order that a collection of individuals can have a collective decision mechanism, the individuals involved must have jointly accepted the procedure as their method for making joint decisions or for forming joint intentions. Arguably, they can only properly regard a decision mechanism as their own if they have successfully exercised it on several occasions. Thus, genuine collective agents will have developed a practice of decision making. Given the existence of such a practice, it would be inappropriate to say that the decisions originate from the individuals. As there is an established way of transforming the ideas individuals have about what to do into a joint intention, it makes sense to say that there is a collective agent that makes decisions of its own. This proposal bears a close similarity to Raimo Tuomela’s conception of social groups as authority systems.19 Authority systems are a kind of collective decision mechanism. In Tuomela’s terminology, they are groupcommitment systems, the exercise of which leads to the formation of a group will or intention.20 They are transformation functions, or sets of such functions, that represent or describe the process of forming a joint intention on the basis of their individual intentions.21 At the heart of Tuomela’s proposal lies the notion of authorization, which is a matter of transferring the power to act. He makes a concomitant distinction between authorized or operative members on the one hand, and non-authorized or nonoperative members on the other, and uses these notions to describe the core idea as follows: As a way of creating a group will, the members commission the authorized members to act for the group while still retaining the ultimate responsibility – although not necessarily “authorship” 18 19 20 21

Pettit (2001; 2007). Tuomela (1995; 2007). Tuomela (1995: 174). Tuomela (1995: 180).

126

Frank Hindriks

– for such authorized action. Thus, the relevant power to act by the former is transferred to the latter, and the end result of this transfer of power is that the operative members gain the power to act in relevant, commissioned ways.22 Although by and large I accept Tuomela’s proposal, I think that the term “authorization” is of limited use in getting at the core issue, which is better captured by the general notion of a collective decision mechanism. One can easily get the impression that when a process of authorization is involved, some members always authorize others. Whereas this can be the case, it need not be. It can be that, in a sense, all members authorize themselves. What is crucial, in my view, is that the individuals involved switch from their individual or private perspectives to the perspective of the group and that they act on the basis of the joint intention that is formed rather than giving priority to their individual or private intentions. The joint intention is prior in the sense that it provides the point of departure for the members to determine how to contribute to the joint action. A final feature I want to add to this conception of collective agents, or ‘social integrates” as I call them, concerns the scope of the collective decision mechanism.23 It must pertain not only to what the group will do, but also to who belongs to the group. A genuine collective agent, at least as characterized from the internal point of view, can change membership while retaining its identity. It can lose members and acquire new ones as, for instance, sports teams can and often do. It is this feature that provides the very basis for regarding social integrates as genuine entities. Nothing in Gilbert’s conception of plural subjects guarantees that they possess it, but it is crucial for the ontological significance of collective agents as conceptualized from the internal perspective. Because of it, social integrates have criteria of identity over time that differ from those of other entities, most notably from those of the sets of individuals that are their members at particular points in time. It is the presence of such criteria that grounds their reality.24 In sum, in my view, forming a joint intention is not sufficient for a group to become a collective agent. Doing something together does not require or involve the formation of a new distinct agent. Instead, establishing a way of making decisions together or of forming joint intentions is the mark 22 Tuomela (1995: 176). 23 As mentioned in the introduction, I take this term from Pettit (2003). 24 As discussed in section 3, my view differs from that of Tuomela in this regard. He does not ascribe ontological significance to social integrates, whereas I do.

The Status Account of Corporate Agents

127

of collective agency as characterized from the internal perspective. Thus, in order for a group to be a social integrate it must have a minimal level of internal organization. The relevant individuals must have integrated themselves into a new unity by jointly accepting a decision mechanism that serves to determine what to do and who belongs to the integrate. There is no need for this mechanism to meet particular requirements of rationality. While such requirements may be crucial to collective moral responsibility, they are of little relevance to ordinary collective agency.25

2. The external attribution perspective 2.1 The Status Account introduced Many types of collective agents have characteristics that they owe to the wider institutional context in which they operate. This holds, for instance, for universities and limited liability companies. In order to be a university or a limited liability company, an organization needs to be recognized as such by outsiders, agents other than the organization itself or its members. Consider universities. Since they came into existence in the twelfth and thirteenth centuries, the prime characteristic of universities has been that they have the right to award titles of higher education (see note 26 for a qualification). They owe this right to external authorities, initially ecclesiastical authorities, in particular the pope, and later civil authorities. As a consequence of being guaranteed by an external authority, these degrees are valid in a wide range of places. In the Middle Ages universities granted licenses called “licentiae ubique docendi” that gave those who received them the opportunity to teach throughout Christendom.26 25 In the locus classicus on collective moral responsibility, Collective and Corporate Responsibility, Peter French (1984) develops a conception of collective agency that bears some similarity to the one presented here. At the heart of both accounts lies the notion of a collective decision mechanism, in French’s terms a “Corporation’s Internal Decision (CID) Structure”. French does not require that particular conditions of rationality be met in order that a collective agent bears moral responsibility. My claim, then, is not that the conception of social integrates proposed in this paper is irrelevant to collective moral responsibility. Instead, the claim is that it does not count against the proposal if it turned out to be inadequate as an account of collective moral agency. French’s conception of collective agency is in fact somewhat more demanding than my conception of social integrates, more closely resembling Tuomela’s notion of a task-right system (see section 3 for a discussion). 26 The account of the history of the university presented here is due to Verger (1992).

128

Frank Hindriks

Since their inception, universities have been communities of teachers and students.27 These communities had a special standing in that they enjoyed some independence from the local authorities, be they ecclesiastical or civil. For instance, the pope granted students the right to enjoy the revenues from ecclesiastical benefices without actually having to take up residence in them. Furthermore, universities were corporate entities that could endow themselves with statutes and enforce obedience of such statutes on the part of their members. In addition to this, they could sue in civil actions in their own names. Again, universities and their members owed these privileges to the wider institutional setting, including the external authorities mentioned. I shall call these privileges, i.e. features such as that of having the right to award the licentia ubique docendi, “normative attributes”. Normative attributes abound in institutional reality. Consider as another example the limited liability company (LLC). Just like universities, they can sue in their own name. The core characteristic of LLCs, however, pertains to its debts, obligations, and other liabilities. By forming an LLC the individuals involved limit their liability to the amount of money they invest. The members of an LLC are not personally liable for any of the debts of their company other than for the value of their investment in it. Limited liability, then, is a normative attribute of LLCs. The collection of normative attributes that is typical of a certain type of organization is what I shall call “a status”. Just like the right of a university to award academic degrees, the limited liability of an LLC is grounded in an external authority. In order to form an LLC a certificate of organization has to be signed and delivered to the Secretary of State. The existence of agents such as universities and LLCs, then, depends on recognition or (collective) acceptance by people or institutions external to them. Rather than (merely) to their internal organization, they owe their existence to the external ascription or attribution of a status. As indicated in the introduction, I call agents of this kind “corporate agents” because a collection of individuals is incorporated into a new 27 In fact some universities were, initially communities of students only (those modeled after Bologna rather than Paris, the universitates scholiarum rather than the universitates magistrorum et scholarium). The students hired teachers on annual contracts. The teachers created their own organization, the college of doctors. This college was responsible for examinations and the conferring of degrees. The claim that universities have the right to confer academic degrees, then, must be qualified. It is more accurate to say that universities are the organizations students go to in order to meet the requirements necessary for receiving academic degrees. See note 41 for more on this issue.

The Status Account of Corporate Agents

129

entity by being ascribed a status.28 The account of these agents I develop in this paper is the status account of corporate agents. Why should we regard corporate agents as agents distinct from their members? Perhaps this is because many of them can change membership without losing their identity. A more compelling reason, however, is that they have properties or powers that the individuals involved do not have. A university can confer academic degrees, whereas individuals cannot, at least not without being authorized to do so. Similarly, the liability of an LLC is distinct from that of the individuals that are its members. Corporate agents are distinct from their members because they have properties that are distinct from the properties of their members (section 104a of the Revised Uniform Limited Liability Company Act drafted by the National Conference of Commissioners on Uniform State Laws in the United States in 2006 states in fact that an LLC “is an entity distinct from its members”). In order to explain how this can be I employ the notion of constitution, which I regard as a relation similar to but distinct from identity.29 The core idea is that a collection of individuals constitutes a particular type of corporate agent exactly if it meets the required conditions for having the status of that type in the context at issue. Filing what is called “a certificate of organization” is one of the conditions that has to be met in order for some individuals to constitute an LLC in the US. An LLC is not identical to its members, because its members can survive its dissolution and it can survive the death of its members. Just like identity, constitution is a non-causal notion. An LLC is not caused by its members. However, its members do enact the actions of an LLC.30 28 In contrast to integration, incorporation is an institutional matter. As I use the term, though, it need not be a legal matter. 29 See Baker (1997; 2000) for a defense of such a view. Wasserman (2004) provides a useful but critical discussion of various conceptions of constitution. 30 Wilson (2005) formulates an analogous condition that collective agents and their members are supposed to satisfy. This is the condition of agency coincidence, which is a matter of two agents undertaking precisely the same actions at a particular point in time. Although it is defined for agents, Wilson maintains it is satisfied by collective agents and the (collection of) individual agents of which they are made up. Moreover, although it is put forward in the context of a discussion of constitution, Wilson appears to require identity. What I call “the Enactment Condition” is a more felicitous formulation of basically the same idea: any action of a collective agent is enacted by some (possibly all) of its members, or by one or more agents who have been authorized by the collective agent. This condition can play the same role Wilson ascribes to his condition of agency coincidence, in

130

Frank Hindriks

Both the notion of a status and the conditions that have to be met in order for something to have a particular status can be explicated in terms of rules: in terms of status rules and constitutive rules, respectively. A status rule specifies the normative attributes of the status at issue. It is a rule or a set of rules because normative attributes can be formulated as rules. Consider the status term “university” and suppose, for simplicity, that its only normative attribute is the right to award master’s and doctorate degrees. Given this stipulation, its status rule is straightforward: a university has the right to award master’s and doctorate degrees.31 The first universities developed out of pre-existing schools. It appears that nothing else was required in order for a school to be a university than that it be recognized as having the status of a university, a status that could at some point only be imposed by the pope. In contemporary societies the conditions that have to be met are much more complex and systematic. For such cases, they can be explicated in terms of constitutive rules.32 Such rules specify the necessary and sufficient conditions for a status to be instantiated in a particular context. In the case of corporate agents, it explicates the properties that a collection of individuals or some (other) entity must have in the context under consideration in order for it to constitute a corporate agent that has the status at issue. Status rules and constitutive rules characterize types of institutional entities. They can – but need not – be formal or codified. Features that are particular to a token of a type can be specified in what I shall call “a statthat it can replace the requirement of spatial coincidence that is usually invoked in accounts of the relation of constitution. 31 Status rules are closely related to regulative rules, which have as their structure “Do X”, or “If Y, do X” (Searle 1969). Only some regulative rules employ status terms, while all status rules do so. Status rules are (possibly stipulative) definitions, whereas regulative rules are not. See Hindriks (2007a) for a more detailed discussion of the relation between status rules, regulative rules, and constitutive rules. 32 This may well be what Searle is after in the following passage: “[W]here the imposition of status function according to the formula becomes a matter of general policy, the formula acquires a normative status. It becomes a constitutive rule.” (1995: 48) As I argue in Hindriks (2003), Searle relates this closely to codification and the possibility of there being counterfeit instances of a status. I believe the link is not so tight. Constitutive rules need not be codified. Furthermore, counterfeit instances can only exist when they are, and once an institutional authority is involved in status attribution: in the end, the crucial difference between a genuine and a fake instance of a certain status is often whether the status has been conferred by the relevant authority (consider a Central Bank issuing money). Constitutive rules do not necessarily involve authorities.

The Status Account of Corporate Agents

131

ute”. Both the status rule and the constitutive rule of the LLC are specified in the Act mentioned above. Its key articles pertain to the conditions for the formation of an LLC on the one hand (article 2), and to the relations of members and managers to persons dealing with an LLC on the other (article 3). These articles present the core of the constitutive rule and the status rule of American LLCs, respectively. The Act also discusses the operating agreement, which can basically be regarded as a statute. An operating agreement will be unique to a particular LLC. The kinds of provisions that the members of an LLC can make in it are specified and limited by the Act. The operating agreement serves as the foundational contract among the entity’s owners. In sum, a corporate agent as defined in this section owes its existence to external recognition of its status. As evidenced by the role that outsiders play, the status of a corporate agent is always embedded in a wider institutional context. The status can be explicated in terms of a status rule that specifies the normative attributes that are characteristic of the type of corporate agent at issue. If generally accepted conditions have to be met in order for a status to be instantiated in a particular context, a constitutive rule for the relevant status is operative in that context. It will by now be clear that corporate agents are institutional agents. As institutional reality is often very complex, the picture presented here must be qualified and extended in order to do justice to those complications. A start will be made in the next section, in which the status account is compared to some of its rivals.33 2.2 The Status Account defended and developed The status account of corporate agents presented in the previous section shares some important features with the social ontology that John Searle has developed in particular in his book The Construction of Social Reality.34 Nevertheless, my view differs from Searle’s in significant respects, two of which I shall discuss here.35 First of all, my notion of a status, as it is developed below, is more general, transparent, and precise than Searle’s notions of status function and deontic power. At the heart of Searle’s conception of deontic power lies the notion of permission. In his view, all

33 The ontology of many other institutional entities can be captured in terms of status rules and constitutive rules as well. The status account of corporate agents, then, can be generalized into a status account of institutions (see Hindriks 2007a). 34 Searle (1995; 1969). 35 See Hindriks (2007a) for some others.

132

Frank Hindriks

such powers can be defined in terms of permission and negation.36 This is a natural suggestion to make, given that prohibitions and obligations can indeed be defined in these terms. It is too restrictive, however, because it leaves out important kinds of rights. Consider the four basic Hohfeldian incidents: the privilege, the claim, the power, and the immunity. The first two of these can be defined as the absence or presence of a duty or obligation. Hence, they can be reduced in the way Searle envisages. The last two, however, pertain to ability or a lack thereof to effect a change in Hohfeldian incidents.37 These cannot be reduced to the notion of permission. Searle has little to say about status functions. The key claims he makes about them are, first, that they are functions entities cannot perform merely in virtue of their physical features, and second, that they can perform those functions only because they are collectively accepted to have those functions.38 It appears that he no longer thinks the notion of a status function adds much to that of deontic power: “[T]he forms of the status function in question are almost invariably matters of deontic powers.”39 What I want to suggest is that we can indeed do without the notion of a status function, but only if we accept a broader conception of deontic power that also includes powers and immunities. Searle uses the term ‘status function” primarily in relation to money. The function of money, one might say, is to serve as a means of exchange. This idea, however, can also be captured in terms of power. In economic terms, money provides one with purchasing power. This power is a matter of being able to change the distribution of property rights, which is, after all, the effect of purchasing something. According to the Hohfeldian conception, a power consists of the ability within a set of rules to alter one’s own or someone else’s rights.40 The notion of ability lies at the heart both of powers and immunities. This implies that, although we can do without the notion of a status function, it is not possible to build all deontic powers out of the notion of permission.41 36 Searle (1995: 106-07) claims that it is tempting to accept this idea. He puts it forward without qualification in Searle (2005: 16-17). 37 Wenar (2005). 38 Searle (1995: 46 and passim). 39 Searle (2005: 13). 40 Wenar (2005). 41 Of course, institutions often serve important functions. However, these are often not explicitly and deliberately attributed to them. When they are, this is done by specifying certain kinds of actions or activities that are, for instance, to be promoted. As will become clear below, this can be accommodated in terms of status

The Status Account of Corporate Agents

133

As I conceive of them, deontic powers can be explicated in terms of actiontypes and modes or modalities. The modalities include those of permission, prohibition, and obligation. In order to do justice to the complexity of institutional reality, we should, however, allow for other modalities. These include power or ability, which is more or less the possibility of performing certain actions. Perhaps yet other modalities should be included as well, so as to accommodate actions that are deemed appropriate, or activities that are to be promoted. In order to allow for this I use the more general term “normative attribute” rather than deontic power. The upshot is that the notion of a status and that of a status rule are more general and provide a better understanding of the underlying institutional phenomena than Searle’s notions of status function and deontic power.42 The second difference between my view and Searle’s that I discuss here pertains to the notion of a constitutive rule. According to Searle, the form or structure of a constitutive rule is given by the counts-as locution “X counts as Y in C”. In contrast to this, the status account uses the following schema for constitutive rules: In C, X is Y. X, Y, and C should be regarded as schematic letters for predicates. Y refers to a status. X predicates the property or set of properties that an entity has to have in order to have the relevant status. C stands for a context or set of

rules. I do not want to deny that the notion of a function can play a useful role in relation to institutional entities in a way that is not captured by the status account. The claim I want to defend instead is that the work that is supposed to be done by the Searlean notion of a status function can be done by my notion of a status. As I have argued in my 2003, there is some evidence for regarding the status function as the function of representation in relation to entities that do not exist independently of being represented. Given that interpretation, the notion of a status function could be taken to convey nothing but the idea that, insofar as institutional reality is concerned, deontic powers exist only because they are represented as existing. This is in fact compatible with the line I defend in the main text. 42 Some statuses are truncated. Consider a university modeled after Bologna that consists only of students (recall note 26). Such a university does not itself have the right to award degrees, but serves for students to meet the requirements for being eligible to receive a degree, what I have called “the X-conditions”. Searle also allows for statuses without “real powers” (Searle 2005: 10n2). He only mentions cases in which the recipient has the (dis-) honor of the status, as in the case of an honorary degree, for which he uses the term “honorific status functions”.

134

Frank Hindriks

circumstances. On the status account of constitutive rules they are to be regarded as universal generalizations.43 A notable difference between this status schema and the counts-as locution is that the phrase “counts as” has disappeared. The reason for this is that I take the phrase to be indicative of the role played by collective acceptance. On the view defended here, a constitutive rule is in force exactly if it is collectively accepted.44 Accepting a constitutive rule amounts to accepting that any entity of a certain kind has the status that appears in the rule, at least if the appropriate contextual conditions have been met. If in the appropriate context an entity meets the relevant X-conditions it actually has the relevant Y-status.45 More precisely, if in C an entity is X then that entity constitutes an entity that is Y. The upshot is that, once the role of collective acceptance has been made explicit, the phrase “counts as” becomes redundant. It is not fully clear to me how important this difference is. Surely, Searle had something like this in mind. However, I ascribe more ontological sig43 In order to make explicit that an instantiation of the schema concerns types of institutional entities rather than tokens, the schema of constitutive rules can be expanded as follows: [CR] ∀α (Cα→ (Xα ↔ Yα)), with α as a schematic letter for a variable. Using Z as a schematic letter for a predicate pertaining to the normative attributes belonging to a particular status, the schema for status rules can be formulated as follows: [SR] ∀α (Yα ↔ Zα). The schemas describe extensional equivalences. They do not make explicit that the relation between X and Y on the one hand and that between Y and Z on the other hand are in fact very different from one another. The relation at issue in constitutive rules is that of constitution. The extensional equivalence made explicit in the schema for status rules is due to the semantic equivalence between Y-terms and Z-terms: being Y just means having property Z. 44 This can be made formally explicit in terms of what I call the Collective Acceptance Principle (see Hindriks 2007a for further discussion): [CAP] (In G, p) ↔ CAG(p), with G for a group, “CAG” for collective acceptance by group G, and “p” for an institutional proposition that expresses a constitutive rule. This Principle is based on Tuomela’s Collective Acceptance Thesis, which he presents in chapter 5 of Tuomela (2002). 45 We often introduce labels for statuses in particular contexts. Currencies of particular denominations are a case in point in relation to the universal status of money. In such cases, the features that occur in the constitutive rule belong to the relevant status by (stipulative) definition. For instance, 5-euro bills are gray and 10-euro bills are red, and their sizes in millimeters are 120 x 62 and 127x 67 respectively.

The Status Account of Corporate Agents

135

nificance to institutional entities than Searle does. This is mainly due to the account of constitution on which I rely. It appears that Searle’s view is that X-terms and Y-terms are just different descriptions of the same entities.46 In my view, however, entities that are X constitute other entities that are Y (see section 2.1 and note 57). An important advantage of the status account of constitutive rules as compared to Searle’s account is that it can solve a problem that, it turns out, afflicts the latter, which I shall call the problem of free standing Y-terms.47 Barry Smith has criticized Searle, arguing that for some institutional entities there is no suitable candidate for an object X on which the status at issue is imposed.48 When discussing some examples to which Searle’s account of constitutive rules applies successfully, he observes: In all such courses we have certain distinctively patterned parts of physical reality, which in certain specific kinds of contexts fall under certain specific kinds of descriptions. The corresponding objects and events, correspondingly, come to be ascribed certain properties or powers of non-physical sorts. As falling under such descriptions the X term counts as a Y term of a certain sort.49 The idea, then, is that on Searle’s account there has to be a particular object X on which a status is imposed in order for that status to be instantiated. Having made this point, Smith turns to his critique: Unfortunately, however, there are entities in social reality – debts, rights, obligations, bond derivatives (and games of blind chess), which do not fit well with Searle’s formula. For here there is no physical X term to which the corresponding properties or powers could be ascribed.50

46 See Smith (2007: 12). 47 Another advantage, which I discuss in Hindriks (2007a), is that the status account can be used to elucidate the relation between regulative rules, and constitutive rules. 48 I have also discussed Smith’s criticisms in Hindriks (2007a; 2007b). Here I focus on Smith’s most recent formulation and also discuss the alternative he presents there. Another advantage of the presentation here is that the differences between the status account and Searle’s account of constitutive rules are made fully explicit here, because here I develop the formal details of the former. I thank Barry Smith for a very helpful exchange on this topic. 49 Smith (2007: 7). 50 Smith (2007: 7).

136

Frank Hindriks

Smith calls statuses for which no physical object can be identified on which it is imposed “free standing Y-terms”, because they are not grounded in physical reality. In response Searle has accepted that there are cases to which his counts-as formula does not apply. Furthermore, he has argued that corporations are a case in point: The laws of incorporation in a state such as California enable a status function to be constructed, so to speak, out of thin air. Thus, by a kind of performative declaration, the corporation comes into existence, but there need be no physical object which is the corporation. The corporation needs to have a mailing address and a list of officers and stock holders and so on, but it does not have to be a physical object. … There is indeed a corporation as Y, but there is no person or physical object X that counts as Y.51 Searle claims here that there is no X, no object on which the status of a corporation is imposed. Furthermore, he suggests that this is not needed. Thus, he admits that there are free standing Y-terms to which his countsas locution does not apply, but he does not take this to be a significant problem for his approach. In fact, he now maintains that it is a mistake to treat the object as the unit of analysis, and he goes on to suggest that we should focus our attention on processes, and facts.52 Representations play an enabling role in this connection: “The real picture is this: we have a set of processes and we have a set of representations which enable these processes to function”.53 In addition to this, Searle claims that “in many cases the representation is all the reality we need to make the entity function”.54 Smith objects to this and maintains that “we need to take seriously the fact that such processes involve objects”.55 He does so by introducing the notion of a quasi-abstract object. Free standing Y-terms are not ascribed to physical objects in his view. Instead, they are quasi-abstract objects anchored “in the realm of records and representations”.56 That they are quasi-abstract means that they are abstract or non-physical on the one 51 Searle (2005: 14-15; emphasis added). See also Smith and Searle (2003: 305-07). 52 Smith (2007: 19; the last section of this paper consists of a discussion between Searle and Smith). 53 Smith (2007: 21). 54 Smith (2007: 20). 55 Smith (2007: 21). 56 Smith (2007: 13).

The Status Account of Corporate Agents

137

hand but historical on the other. Collective acceptance or human agreement is relevant through the role documents play, which grounds the historical aspect of these institutional entities. He criticizes the idea that free standing Y-terms might just be concepts pointing out that concepts can exist without corresponding objects. We can have the concept of a debt at a point in time at which there happen to be no debts. Both Searle and Smith maintain that the universe is partly populated by a large number of free standing Y-terms, and that there need not be particular physical objects to which those terms apply. I regard both their views as unattractive. The main point of Searle’s social ontology used to be to show how institutional reality is embedded in physical reality. Searle is correct in pointing out that he has always used fact talk, but this need not be at odds with object talk. After all, it is a fact that certain pieces of paper are euro bills. Moreover, without an account of how debts and organizations involve objects, or at least entities such as actions or persons, Searle’s ontology adds nothing to existing accounts in the social sciences that appeal to norms or rules that specify the functions or powers, or as I would put it the normative attributes involved. Smith’s more daring proposal is attractive in that it is a genuine attempt to provide an ontology of the entities at issue. However, Smith underestimates the extent to which institutional entities can be grounded in physical entities. As I shall argue, we can understand the ontology of entities such as debts and organizations by abandoning the idea that statuses are nothing other than new descriptions of objects that are familiar under other guises. The first thing to realize is that institutional entities need not be objects. They can also be properties of objects. A debt, for instance, is to be regarded as a (relational) property rather than as a self-standing object. Even though George Bush can be the president of the United States, he cannot be a debt (not even if he has one). When the X and Y-terms that appear in the constitutive rule are interpreted as predicates, this can be accommodated in a rather natural way. The bound variable ranges over persons in this case. The Y-term stands for having a debt. If Bush has debts, he must meet the X-conditions for having a (particular kind of) debt and be in the relevant context C. Given all this, Bush’s being X constitutes his being Y (which is a matter of standing in a particular relation to some other person). Secondly, rather than regarding a status as a different description of the same object, it can be regarded as a predicate that first and foremost applies to another object, the one that is constituted by the object to which the

138

Frank Hindriks

status is ascribed. This alternative appears to work for corporate agents, and it can also be accommodated by the status account of constitutive rules. As we saw above, Searle maintains that in the case of a corporation “there is no person or physical object X that counts as Y”.57 In my view, a (collection of) person(s) can count as a corporation. This means that it is a corporation in that the relevant person(s) constitute(s) it. So we have two distinct objects here: a collection of persons and a corporation. The former constitutes the latter, because it meets the X-conditions and occurs in the appropriate context C (and because it is collectively accepted that in C entities that are X are corporations). The two entities are distinct from one another in the sense that the corporation has the normative attributes involved in its status essentially, whereas the collection of individuals has them only in a derivative sense.58 The upshot is that, pace Searle and Smith, corporate agents are not free standing Y-terms. Note that there is a sense in which any plausible view of institutional reality will have to allow for free standing Y-terms. Positions within organizations can be vacant. Especially if it is a statutory requirement for there to be a certain position, it is implausible to say that it does not exist because it is not filled. Consider the position of a treasurer and suppose it is in fact vacant in a particular society or association. In such a situation we should say that the position exists within this association, but that at the moment there is no one who is the treasurer; there is no one who has the properties that constitute the property of being a treasurer.

57 Searle (2005: 15). 58 The status account involves both the notion of property constitution and that of object constitution. In relation to constitutive rules, the idea of property constitution is that, when an entity has a property X, a property that figures in a particular constitutive rule that is collectively accepted, and it occurs in the context C that appears in the same rule, that property thereby constitutes the relevant property Y. If an object has property Y essentially, then this object is constituted by the object that is X. It is a contingent fact about a person that she has a debt. Limited liability is an essential property of LLCs (and a contingent property of collections of individuals). In line with this, debts are to be explicated in terms of property constitution, whereas corporate agents are to be understood in terms of object constitution. An indication of the fact that a relation of constitution is at issue is that an entity can be X without being Y, which is the case when being Y depends on being in the appropriate circumstances C. This view of constitution is very similar to that proposed by Baker (1997; 2000; 2008) on whose work I also rely for the distinction between having a property essentially and having a property derivatively (see note 29 for a difference between Baker’s view and mine).

The Status Account of Corporate Agents

139

Given the criticisms voiced regarding the notion of a free standing Y-term above, the question arises of what the existence of an unfulfilled position amounts to. It amounts to nothing else than the collective acceptance of a statute according to which the corporate agent has the position of a treasurer. This phenomenon is in fact quite common and uncontroversial. Few will, for instance, claim that the relevant royal or ecclesiastical institutions have disappeared during periods in which there is no king or no pope. In fact, the Catholic Church has a doctrine of perpetual papal succession, which is left unperturbed by periods in which there is no pope (known as papal interregna). In all such cases, the institution exists, but it is not instantiated.

3. The internal and the external perspective combined In the preceding sections, collective agency has been investigated both from the internal and from the external perspective. This has resulted in a conception of social integrates on the one hand and of corporate agents on the other. The question arises how social integrates and corporate agents relate to one another. By and large, the conceptions defended here should be regarded as complementary. It is quite common, for instance, for a social integrate to develop into a corporate agent. A social integrate can also come into existence at the same time as the concomitant corporate agent.59 This need not be the case, however. Corporate agents can in fact be constituted by single individuals. Sometimes this even holds by definition, as in the case of sole proprietorships. The limiting case of a corporate agent, then, is not a collective agent. In this respect, the claim that section 2 concerns the investigation of the external perspective on collective agents, which was made in the introduction, must be qualified. In fact, it pertains to a kind of agent that is distinct from individual human beings, which can but need not be a collective agent. So the status account also covers agents that are not social integrates.

59 It may be that a collection of individuals decides on a decision procedure when going through the process of establishing a corporate agent (for instance, by formulating an operating agreement for an LLC). It appears that in such cases a social integrate comes into existence at the same time at which the corporate agent comes into existence, even if the decision mechanism has not been exercised several times. This condition mentioned in section 1, then, is only plausible in relation to social integrates that are not corporate agents.

140

Frank Hindriks

It is important to note that statuses can exist within collective agents. Being a professor is a status within the university, just as being a manager is a status within LLCs. These statuses depend for their instantiation on collective acceptance both from insiders and outsiders. Statuses that are internal to a collective agent do not exist in all social integrates, but only within social integrates with structure. Tuomela has introduced the notion of a task-right system for such collective agents.60 Such a system is a division of tasks and rights governed by formal or informal norms.61 Such tasks and rights are normative attributes. As such, they can be accommodated by the status account of corporate agents. According to that account, the norms that are definitive of the relevant tasks, rights, or roles are status rules. So, even though it was developed above in relation to statuses of corporate agents attributed by outsiders above, the status account can be extended to normative attributes within those agents (see note 32).62 Collective acceptance from insiders is also important for statuses of a corporate agent as a whole. It is hard to see how, for instance, rights involved in a status could be exercised if insiders did not recognize them. Furthermore, there might be statuses that corporate agents can have in virtue of acceptance by its members only. Consider, for instance, the power of the Catholic Church to excommunicate some of its members. No external recognition appears to be required for it to possess this power. As another example, consider a national minority that declares its independence from the state to which it officially belongs. It is conceivable that it has some powers, even if it is not recognize as a state by outsiders. The thing to note, however, is that in such a situation its powers pertain to its members only. The accounts also differ in relation to the ontological criteria they employ concerning the existence of agents, at least apparently. The internal organization account of social integrates appeals to criteria of identity over time. The status account of corporate agents invokes new attributes or properties (i.e. statuses). Many philosophers deny the existence of collective agents. Tuomela, for instance, argues in relation to what I have called ‘social integrates” that they are “not ontically real entities, even if our for60 Tuomela (1995). 61 Using these notions, Tuomela (1995: 438) provides a definition of organizations that is very similar to the conception of collective agents defended by French (1984), and, as Tuomela points out, to that of Bach and Harnish (1979). 62 By virtue of their having the power to impose internal statuses, it seems plausible to say that task-right systems or organizations are themselves statuses.

The Status Account of Corporate Agents

141

mulations and phrasing – used for convenience – might suggest this”.63 Tuomela appears to regard the notion of a social group or integrate here as a useful tool, perhaps a useful fiction. On his view a more felicitous way of talking would be to say that certain individuals form a social group or integrate (rather than that they are one): “Thus, “groupness” is a relation between persons … . [A] group is not an entity”.64 Apparently, Tuomela rejects the availability of independent criteria of identity as a criterion for existence, and only accepts the criterion of new attributes or properties.65 The view that corporate agents are nothing but useful fictions, known as fictionalism, is in fact very popular in law and the social sciences. In law corporate agents are often regarded as legal fictions. According to a prominent view in economics, the firm is taken to be a nexus of contracts, and this view is usually regarded as a kind of fictionalism. I cannot do justice to these views in this paper. Let me note just two things. First, the status account of corporate agents focuses on normative attributes. These are exactly the kind of things that are central to contracts and to law. Second, such attributes can be used for developing a robust kind of realism, as I have tried to show in this paper. Perhaps the way in which I have employed the notion of constitution here in order to defend realism about corporate agents will convince some fictionalists that a genuine alternative is in fact available.

4. Conclusion It has been argued that the most plausible account of collective agents viewed from the internal perspectives is the internal organization account of collective agents as social integrates. From an external perspective, collective agents are best seen as corporate agents, a notion that has been explicated in terms of the status account. Social integrates basically consist of collections of individuals that have jointly accepted a decision mechanism. This internal organization account of collective agents was defended against weaker and stronger accounts. The former was rejected because people can share a joint intention without forming a collective

63 Tuomela (1995: 194). 64 Tuomela (1995: 194). 65 Tuomela (1995: 200-201 and 374-75) presents the ontologically noncommittal view as his official view, leaving open the prospect that future science will prove this view to be mistaken.

142

Frank Hindriks

agent. The latter was rejected because people can form a collective agent without satisfying particular criteria of collective rationality. The external perspective resulted in an account of institutional agents, the status account of corporate agents. Corporate agents are statuses that are usually imposed on collections of individuals, and that consist of normative attributes such as rights and obligations. This view was developed in terms of status rules that specify the nature of particular kinds of corporate agents, such as universities or limited liability companies, and in terms of constitutive rules that specify the conditions a collection of individuals has to meet in order to constitute an entity such as a university of a limited liability company. It is to be preferred to the Searlean account of constitutive rules, inter alia because the former succeeds in positioning corporate agents in a broader ontology where the latter fails.

References Bach and Harnish (1979): Bach, K. and Harnish, R.M. Linguistic Communication and Speech Acts, Cambridge (MA): MIT Press, 1979. Baker (1997): Baker, L.R. Why Constitution is Not Identity, in: Journal of Philosophy, (1997) 94: 599-621. Baker (2000): Baker, L.R. Persons and Bodies: A Constitution View, Cambridge: Cambridge University Press, 2000. Baker (2008, forthcoming): Baker, L.R. Non-Reductive Materialism, in: McLaughlin, B., and A. Beckermann (eds.), The Oxford Handbook for the Philosophy of Mind, Oxford: Oxford University Press, 2008. Bratman (1999): Bratman, M.E. Faces of Intention, Cambridge: Cambridge University Press, 1999. French (1984): French, P.A. Collective and Corporate Responsibility, New York: Columbia University Press, 1984. Gilbert (1989): Gilbert, M. On Social Facts, London: Routledge, 1989. Gilbert (1996): Gilbert, M. Living Together. Rationality, Sociality, and Obligation, New York: Rowman and Littlefield, 1996. Hindriks (2003): Hindriks, F. The New Role of the Constitutive Rule, in: Koepsell, D. and L.S. Moss (eds.), John Searle’s Ideas about Social Reality: Extensions, Criticisms, and Reconstructions, Oxford: Blackwell, 2003, pp. 185-208. Hindriks (2007a): Hindriks, F. But Where Is the University?, mimeo.

The Status Account of Corporate Agents

143

Hindriks (2007b): Hindriks, F. Facets of Sociality. Nikos Psarros and Katinka Schulte-Ostermann (eds.) [book review], in: Notre Dame Philosophical Reviews, ndpr.nd.edu/review.cfm?id=10284, (2007-07-05). Pettit (1993): Pettit, P. The Common Mind. An Essay on Psychology, Society, and Politics, Oxford: Oxford University Press, 1993. Pettit (2001): Pettit, P. A Theory of Freedom, Cambridge: Polity Press, 2001. Pettit (2003): Pettit, P. Groups with Minds of Their Own, in: Schmitt, F.F. (ed.), Socializing Metaphysics. The Nature of Social Reality, Lanham: Rowman and Littlefield, 2003, pp. 167-94. Pettit (2007): Pettit, P. Responsibility Incorporated, in: Ethics, (2007) 117: 171-201. Pettit and Schweikard (2006): Pettit, P., and Schweikard, D. Joint Actions and Group Agents, in: Philosophy of the Social Sciences, (2006) 36: 18-39. Searle (1969): Searle, J.R. Speech Acts: An Essay in the Philosophy of Language, Cambridge: Cambridge University Press, 1969. Searle (1990): Searle, J.R. Collective Intentions and Actions, in: Cohen, P.R., Morgan, J., and M.E. Pollack (eds.), Intentions and Communication, Cambridge/MA: MIT Press, 1990, 401-415. Searle (1995): Searle, J.R. The Construction of Social Reality, New York: The Free Press, 1995. Searle (2005): Searle, J.R. What Is an Institution?, in: Journal of Institutional Economics, (2005) 1: 1-22. Smith (2001): Smith, B. Fiat Objects, in: Topoi (2001) 20: 131-48. Smith (2007): Smith, B. The Foundations of Social Coordination: John Searle and Hernando de Soto, in: Psarros, N., and K. Schulte-Ostermann (eds.), Facets of Sociality, Frankfurt: Ontos Verlag 2007, pp. 3-22. Smith and Searle (2003): Smith, B., and Searle, J.R. The Construction of Social Reality: An Exchange, in American Journal of Economics and Sociology, (2003) 62: 283-310. Tollefsen (2002): Tollefsen, D. Organizations as True Believers, in: Journal of Social Philosophy, (2002) 33: 395-410. Tuomela (1995): Tuomela, R. The Importance of Us. A Philosophical Study of Basic Social Notions, Stanford: Stanford University Press, 1995.

144

Frank Hindriks

Tuomela (2002): Tuomela, R. The Philosophy of Social Practices: A Collective Acceptance View, Cambridge: Cambridge University Press, 2002. Tuomela (2005): Tuomela, R. We-Intentions Revisited, in: Philosophical Studies, (2005) 125: 327-69. Tuomela (2007): Tuomela, R. The Philosophy of Sociality: The Shared Point of View, Oxford: Oxford University Press, 2007. Tuomela and Miller (1988): Tuomela, R., and Miller, K. We-Intentions, in: Philosophical Studies, (1988) 53: 115–137. Verger (1992): Verger, J. Patterns, in: De Ridder-Symoens, H. (ed.): A History of the University in Western Europe: Volume I, Universities in the Middle Ages, Cambridge: Cambridge University Press, 1992, pp. 35-74. Wasserman (2004): Wasserman, R. The Constitution Question, in: Nous, (2004) 38: 693-710. Wenar (2005): Wenar, L. Rights, in: Stanford Encyclopedia of Philosophy, E.N. Zalta (ed.), http://plato.stanford.edu. Wilson (2005): Wilson, R.A. Persons, Social Agency, and Constitution, in: Social Philosophy and Policy, (2005) 22: 49-69.

Collectivity and Circularity Björn Petersson, Lund

1. Introduction1* According to a common claim, a necessary condition for a collective action (as opposed to a mere set of intertwined or parallel actions) to take place is that the notion of collective action figures in the content of each participant’s attitudes. Insofar as this claim is part of a conceptual analysis, it gives rise to a circularity challenge that has been explicitly addressed by Michael Bratman and Christopher Kutz.2 I will briefly show how the problem arises within Bratman’s and Kutz’s analyses, and then proceed to criticize some possible responses, including the ones proposed by Bratman and Kutz. My conclusion is that in order to avoid circularity and retain the features that are supposed to make this sort of account attractive, we need a notion of collectivity that does not presuppose intention. I suggest that we should make a distinction between collective and non-collective activity merely in terms of dispositions and causal agency. There are independent reasons to think that we actually possess such a distinct causal conception of collectivity. It is not necessary for the participants in a jointly intentional collective action to possess a stronger notion of their intended collective activity than this. In particular, they do not need to possess the concept of a jointly intentional collective action.

2. The circularity challenge Michael Bratman discusses the circularity challenge to his theory of collective action in “Shared Cooperative Activity” (1992) and “I Intend That 1* Versions of this article have been presented at the Higher Seminar in Practical Philosophy in Lund, at ECAP V in Lisbon 2005, and at Collective Intentionality V in Helsinki 2006. I am grateful to participants in these seminars and conferences for valuable criticism and suggestions. I am also indebted to Michael Bratman for helpful comments on an earlier version of the paper, and to Margaret Gilbert for useful suggestions in conversation. I have also benefited from comments by anonymous referees. This article was first published in The Journal of Philosophy, vol CIV, No 3, March 2007. 2 Bratman (1992: 97, 1997: 148), Kutz (2000: 86). N. Psarros, H.B. Schmid and K. Schulte-Ostermann, Concepts of Sharedness, Frankfurt 2008: 145-164; ontos verlag

146

Björn Petersson

We J” (1997). In his account, collective actions spring from shared intentions, and ’our shared intention’ should be understood in terms of each of us having an intention of the form ‘I intend that we J’. What is it, then, that I intend when I intend that we J? How does this intention of mine differ from, say, an intention of the form ‘I intend to contribute to J-ing’? It is tempting to think that ‘we J’ as it appears in the content of my intention, must refer to a collective action, presupposing shared intention. In that case, though, the account will not help us to a better understanding of what it means for an intention to be shared, and for an action to be collective. It will not provide the meaning of ‘sharedness’. In his book Complicity from 2000, Christpher Kutz contrasts his own account of collective action with a reductionist view, according to which such acts can be seen as sums of individual choices “on the basis of expectations about the choices to be made by those with whom they wish to coordinate.”3 Kutz argues that this would be insufficient, and claims that “the core, minimalist, notion of collective action requires /…/ that individuals act on overlapping participatory intentions”.4 Implicit in the idea of a participatory intention is “a conception that one is doing one’s part in a collective project”.5 Participatory intentions are a special class of intentions, “differentiated by their group-oriented content.”6 The content of this type of intention is “irreducibly collective”.7 So, the question is whether we can understand what kind of content these participatory intentions have without understanding the concept of a collective action in the first place. 8

3 4 5 6 7 8

Kutz (2000: 75). Kutz (2000: 75). Kutz (2000: 76). Kutz (2000: 86). Kutz (2000: 86). Margaret Gilbert’s theory of collective action is a third example of an analysis that is open to a circularity challenge of the kind discussed here. “What is key is the incorporation of an holistic notion of joint commitment. The formulation [in On Social Facts, p.198] correctly implies that the parties to any joint commitment must possess this notion” (Gilbert 2003: 53.) So, the individual states in which the notion of joint commitment must occur are also necessary conditions for a joint commitment to obtain. Since her approach differs in important aspects from Bratman’s and Kutz’s (she regards the enterprise as essentially normative, and it not evident that her account of how joint commitments are created should be read as a conceptual analysis), it would reach too far to give a fair treatment of her view here.

Collectivity and Circularity

147

3. Responses Bratman’s first response to the circularity challenge was to assume that we should appeal to act descriptions “characterized in cooperatively neutral ways”.9 My individual intention must refer to an act of the sort that may be co-operative but need not be. The act description must satisfy the “behavioral conditions” for joint activities without presupposing “the attitudes essential to cooperative activity”. “There is, for instance, a clear sense in which we can go to New York together or paint the house together without our activity being cooperative.”10 This initial response to the circularity challenge, Bratman claims, is similar to a common way of escaping circularity in the analysis of ‘intentional action’. If we want to say that an action is performed intentionally when it is explained by the right kind of attitudes, we must “eschew appeal to attitudes that include in their content the very idea of intentional action. Instead, we limit ourselves to contents that, at most, involve a concept of action that leaves it open whether the action is intentional.”11 Margaret Gilbert objects to Bratman that it is unclear how he construes ‘that we J’ in ‘I intend that we J’. She suspects, though, that “he has an individualistic construal in mind. Consider his central case of painting the house together. I shall take it that he construes my intending that we paint the house together as something like this: my intending that the house is painted by virtue of my actions along with complimentary actions of yours.”12 In that reading, though, the analysis will fail to capture intuitions about a stronger sense of ‘sharedness’ or ‘collectivity’ in action, which prompted the need for revising standard theory of action to begin with. Gilbert’s objection might seem unfair in the light of Bratman’s further restrictions on shared intention, which require that interdependence and interlocking subplans also figure in the participant’s conception of ‘our J-ing’. In terms of the full necessary and sufficient conditions of Bratman’s analysis, we have a shared intention to smash the window if and only if (1) (a) I intend that we smash the window and (b) you intend that we smash the window, and (2) I intend that we smash the window in accordance with and because of 1a, 1b, and meshing subplans of 1a and 1b, and you intend similarly, and (3) 1 and 2 are common knowledge between us. 9 10 11 12

Bratman (1992: 97). Bratman (1992: 97). Bratman (1997: 147). Gilbert (2000: 156).

148

Björn Petersson

However, I think that Bratman’s initial strategy against circularity nevertheless makes Gilbert’s doubts reasonable. According to that strategy, it must be left open whether the ‘we smash the window,’ as it occurs in the content of your intention and my intention, is a case of shared intention. Suppose I want the window smashed. When I note your presence on the street, I think that if you act in a certain way, the window can be smashed as a result of both our acts, and I form an intention accordingly. What I intend in that case is merely to get the window smashed, while predicting that your actions will be components in the process leading to that result. This prediction may rest upon my knowledge that your intentions are similar to mine, and that our subplans are likely to mesh in a way that enables me to reach my goal. There is mutuality and interdependence, in line with Bratman’s requirements. Still, I would say, nothing in this picture captures ‘sharedness’ or ‘collectivity’ in any sense distinct from what we can construe in terms of standard individualistic theory of action. What we have is a complex set of individual intentions, beliefs about other people’s intentions, and means-ends reasoning. In other words, if the characterization of a shared intention allows the individual intentions that make it up to be directed towards a mere set of actions that happen to be grouped together in a purely behavioral sense, the distinction between a shared intention and a set of individual nonco-operative intentions will be undermined. The additional conditions do not suffice to ground collectivity. (It is even more apparent that this strategy would not work in Kutz’s case, since he explicitly states that additional conditions such as mutual knowledge, interdependent subplans, etc., are superfluous.) Bratman’s analogy between avoidance of circularity about ‘intentional action’ and about ‘collective action’ is somewhat misleading. Collective and individual intentional actions both belong to the species of ‘intentional action’. The present way of drawing the distinction between them presupposes that the intentions that explain an act (in a way that makes the act intentional) contain a reference to the acting unit itself – the individual or the collective. Without presupposing any references to intention it seems unproblematic to conceive of an individual acting unit, and to imagine that sort of conception placed in the content of an intention to act individually. The trouble here is that if shared intentions are supposed to be what ties members together in a collective unit, it seems unavoidable that we refer to those very intentions when we characterize the conception of the acting unit that is supposed to figure in their distinctive content. The natural solution working for intentional action – to avoid reference

Collectivity and Circularity

149

to attitudes involving the concept to be defined – has a higher price when applied to ‘collective action’. So, Bratman’s first solution seems to make the theory more reductionist than it was supposed to be. With explicit reference to an earlier work by Christopher Kutz, he also appears to acknowledge this in the later paper.13 This brings us to the second response. In Bratman’s terms, that alternative strategy is to “seek basic cases in which each participant intends a joint activity understood in a way that is neutral with respect to shared intentionality. […] We then try to understand those cases in which the agent’s conception of the shared activity explicitly supposes it is a shared activity, in a way that builds on our understanding of these basic cases.”14 Kutz notes that there are cases where it seems proper to say that an intention is directed towards a jointly intentional activity. “[T]he intentions to perform many kinds of jointly intentional acts cannot be made sense of except in collective terms: noncooperative chess is not chess but something else – mutual chess solitaire.”15 “[T]here is no way to characterize the intention of an individual planning to dance the tango except as an intention to engage in jointly intentional tango dancing.”16 The proper response to the circularity challenge, according to Kutz, is therefore not “an analysis that tries to show how each instance of collective action is built out of non-collective materials, but rather a genealogical account that shows generally how the capacity to engage in collective action emerges out of capacities explicable without reference to collective concepts”.17 This article discusses circularity in the definition of ‘collective action’. This second response by Kutz and Bratman indicates that what they have in mind could be another issue – about circularity in the explanation of collective actions. What we need, in Kutz’s words, is “a genealogical account that shows generally how the capacity to engage in collective action emerges out of capacities explicable without reference to collective concepts”.18 Roughly, it says that the content of an executive intention (i.e. an intention regarding an outcome as a whole, whose function is to generate and command other subsidiary intentions) can come to refer to a jointly intentional act through a kind of bootstrapping interaction with

13 14 15 16 17 18

Bratman (1997: 147-148). Bratman (1997: 147-148). Kutz (2000: 86). Kutz (2000: 86). Kutz (2000: 86). Kutz (2000: 86).

150

Björn Petersson

simpler non-collective intentions, which then gradually become subsidiary to it. 19 In learning to play chess, I first conceive of the elements of the game in nonjoint terms: the knight moves so and so; it is best to open with pawns; and if you threaten my queen, I should check your king. As I learn to play, these constituent elements of chess come to be represented and internalized as “my playing chess”, or, alternatively, “my doing my part of our playing chess together.” By such a bootstrapping process, the collective joint activity thus becomes the object of an agent’s executive intention, having been built out of non-collective elements.20 This response meets one sort of circularity worry, by describing how an intention directed towards a jointly intentional activity can emerge from other sorts of intentions. One could imagine a similar response to Hume’s circularity insinuation against the idea that a desire to act justly is what explains a just act.21 The reply to Hume would then perhaps describe how the desire to perform a just action emerges out of purely self-interested desires. That would show how one might explain the desire in question, and thereby the performance of a just act, in a non-circular way. Explanans does not have to contain explanandum just because the explanation refers to an attitude in the content of which a conception of explanandum figures. Assume, though, that Hume’s argument should instead be understood as a claim about circularity in definition, directed towards an analysis of the following sort: A ‘just act’ should be understood as ‘an act performed out of the right kind of motive’, and ‘the right kind of motive’ means ‘a desire to act justly’. That would be an analytically circular definition, and the genealogical response would not remove its circularity. If the definition is supposed to provide the meaning of ‘just act’, it fails. In order to understand what it is that a person comes to desire when a desire to act justly emerges, we would have to know what a just act is to begin with. The challenge to be met here is that in order to make clear what it means for an intention to have an “irreducibly collective content”, we must avoid references to collective action. Otherwise, we cannot avoid circularity in the definition of the latter concept. Kutz’s description could be regarded as a response to the accusation that explanans contains explanandum. It might 19 A similar genealogical response to the circularity challenge, along with the claim that the ensuing conception of jointness should be regarded as primitive, is suggested by Frederic Schmitt (on behalf of Gilbert) in the introduction to Schmitt (2003: 11). 20 Kutz (2000: 87-88). 21 Hume (1739: Book III, part II, section I).

Collectivity and Circularity

151

also be read as a defense of the individualistic presupposition that the capacity to act collectively cannot be just as basic as the capacity to act individually. The account above describes how intentions directed towards jointly intentional activities can emerge, on that presupposition. But the initial circularity challenge is neutral with respect to this individualistic assumption, as well as to the question about circularity in explanation. The challenge is not that intentions with a non-collective content must come before intentions with a collective content, but that we need a characterization of the content of the components referred to in the definition of a collective action – a characterization that does not rely on that very concept. This requirement is reasonable independently of which genesis we assume that these components have. As far as I can see, Kutz’s strategy fails to meet that challenge. Even if his story above is correct, we still want to know what makes an action collective. What kind of entity is it that becomes the object of the agent’s intention through the bootstrapping process? His answer, like Bratman’s, seems to be that the action turns collective when the executive intentions of the individuals contributing to it become directed towards a jointly intentional achievement. That answer brings us back to the original circularity challenge, independently of how we think that these executive intentions come into existence. A third response to the challenge would be to argue that the analysis is sufficiently informative in spite of its being circular. Kutz hints at that strategy when he claims that the circularity problem is “more methodological than substantive”. He also notes approvingly that Bratman has “relaxed his strictures” in this matter.22 However, neither Bratman nor Kutz makes clear in which sense and to which extent the strictures concerning circularity should be relaxed in this specific context. Circular analyses may be informative and explanatory. Philosophers have offered non-reductive explications of (e.g.) normative and modal concepts, while presupposing some primitive notion of modality or normativity. That sort of analysis might be the best we can do in some cases. Network definitions that place a concept within a larger scheme of interrelated concepts can be informative even if each concept is defined entirely in terms of others within the scheme. And so on. One thing that may make some inclined to accept circularity in the present case is that the analysis merely suffers from what I. L. Humberstone calls analytical circularity. It is not inferentially circular.23 Inferential circularity is generally vicious, while analytical circularity “might not always be a bad 22 Kutz. (2000: 86). 23 Humberstone (1997).

152

Björn Petersson

thing” according to Humberstone.24 Suppose we claim that a necessary condition for ‘I smash the window intentionally’ to be true is that I know that I smash the window intentionally (and that knowledge presupposes truth). This would be a case of inferential circularity. If you know that this specific component of the analysis applies to a case, you already know that the concept analyzed applies. You cannot know that I know that I smash the window intentionally unless you know that I smash it intentionally. So, the analysis would neither enhance our understanding of the meaning of the term ‘intentional action’, nor improve our ability to apply the term to the right kind of objects. But suppose we want to analyze the notion of acting collectively, in the strong sense of ‘acting with a shared intention’. We assume that a necessary condition for ‘we smashed the window collectively’ to be true is that our smashing the window was explained by my intending that we smash the window and your intending that we smash the window. (The two latter occurrences of ‘we smash the window’ should be taken as presupposing shared intention.) In that case, we still have analytical circularity – the concept analyzed appears (implicitly) in a component of the analysis. However, this would not suffice to make the analysis inferentially circular, since we could know that the conception of a collective action, involving the notion of shared intention, appeared in the contents of your intention, without knowing that ‘we smashed the window collectively’ applies to the case. As Humberstone puts it, the concept analyzed has a suitably protective embedding in this specific condition. Analytically circular analyses often occur in contexts where we have a common pre-theoretical understanding of the concept analyzed, but a controversy or uncertainty concerning the applicability of the term to some given type of item. In such cases, we may regard the list of application conditions “not so much as a way of getting the concept across to someone not familiar with it, but as a recipe for telling us when it applies”.25 This possible excuse for analytical circularity does not apply here. ‘Collective action’ is not alleged to be like ‘art’ – a concept of which we have a common pre-theoretical understanding while we sometimes need to be informed about boarders for application. On the contrary, the starting point for analyses of ‘collective action’ is an assumption about what kind of entities this concept applies to. Proponents of the concept (like Bratman, Gilbert, and Kutz) typically focus salient cases of co-operative, joint action, as opposed to mere parallel behavior. The analytical task they 24 Humberstone (1997: 254). 25 Humberstone (1997: 251).

Collectivity and Circularity

153

set themselves is to make clear how the description of these phenomena should be fitted into some overall action-theoretical framework. As Humberstone and Roseanna Keefe both stress, the purpose of the analysis determines whether we should accept its circularity.26 Circularity should at least not be accepted when “it obstructs the transfer of understanding the application conditions of a concept may be designed to effect.”27 One sort of understanding that a definition of ‘collective action’ should provide is practical. Although the theories discussed here are descriptive, they will have relevance in moral and legal contexts. (Kutz explicitly builds his theory about legal and moral accountability on his analysis of ‘collective action’.) To assess normative claims about group actions and participation, we need to know what it is that the individuals are supposed to intend or be committed to when a collective action is said to emerge from their attitudes. Since the circularity discussed here makes it unclear what kind of objects these attitudes have, it obstructs the transfer of one important kind of understanding that these analyses are designed to effect. I suppose that the main theoretical purpose of the analyses at hand is to establish a substantial distinction between collective actions and noncollective sets of individual actions. Circularity clearly threatens that aim. So, in this case, analytical circularity is a methodological problem, and a substantive one.

4. Collectivity without intention Bratman suggested initially that circularity could be avoided by appealing to act descriptions that satisfy the behavioral conditions for joint activities but do not presuppose shared intention. On that assumption, there is no reference to shared intention in the notion of joint activity that is supposed to figure in the content of the individual ‘I intend that we J’intention. My complaint was that if we skip that implicit reference, Bratman’s theory does not uphold the distinction between a collective action, and a set of individual interdependent actions. The additional conditions about mutual interdependence etc. are insufficient for that. However, I believe that Bratman’s suggestion could be elaborated. An act description might be neutral with respect to shared intention without be26 Keefe (2002). 27 Humberstone (1997: 251).

154

Björn Petersson

ing neutral about collectivity. The strategy is to employ a broad concept of collective activity that does not presuppose shared intention but is nevertheless substantively distinct from the notion of a mere set of intertwined acts. That might require the description to go beyond mere behavior, though not very far. Susan Hurley employs a useful notion of a “unit of activity”.28 She does not offer any precise explicit definition, but develops the conception by utilizing it in a series of arguments directed towards an individualistic presupposition about instrumental rationality. An important point is that the root of problems of co-ordination and co-operation (exemplified e.g. by prisoner’s dilemmas) is not “the nature of the individuals’ goals, or the instrumental character of rationality. Rather it is individualism about rationality, which holds the unit of activity exogenously fixed at the individual”.29 Like Pettit and others, Hurley thinks that considerations from game theory and social choice theory suggest that we must allow collectives to be regarded as agents or units of activity. She denies, though, that this must involve regarding the unit of activity as the source of the evaluation of outcomes. “Thus, collective activity does not require collective goals or preferences. /…/ That is, collective action does not require that individual preferences be shared or even aggregated into some collective preference”.30 No joint goals, preferences or intentions are required. If a definition of ‘unit of activity’ could be extracted from Hurley’s account, it would be purely operational. Roughly, a unit of activity is an entity for which options can be evaluated instrumentally, at least from a spectator’s point of view. “When I speak of a unit of activity, I mean that activity within certain boundaries (which may be defined in spatial or other terms) has consequences, calculated against a background of what happens and is done outside those boundaries; the consequences of this activity can be evaluated instrumentally, in light of the goals of the unit or parts of it”.31Now, if a collective activity does not require shared preferences, goals or intentions, where then is the glue that keeps the participants together in a unit?32 To frame the question differently, are there any essential intrinsic features that distinguish collective units of activity from mere sets of agents? How do we distinguish a member from a non-member? 28 29 30 31 32

Hurley (2004). Hurley (2004: 201). Hurley (2004: 202). Hurley (2004: 201). This was Michael Bratman’s rhetorical question at Hurley’s presentation of the paper I mainly refer to here (CollInt IV, Siena 2004).

Collectivity and Circularity

155

Hurley offers several suggestions about how collective units of activity can be formed, and about how their boundaries become determined. One approach focuses “the capacity to identify with others as part of a collective unit of activity”.33 The ability to attribute mental states to others plays an important role for the development of that capacity, Hurley argues. However, ‘each participant’s identification with the others as part of a collective unit of activity’ cannot be taken as conceptually essential to a unit of activity. To begin with, this would actualize a circularity challenge. Furthermore, “unit of activity,” as Hurley employs the term, applies to entities apparently incapable of identification, at least of the sort requiring the possession of intentionalistic concepts. Slime mould can form units of activity. “An ant colony can be regarded as a unit of activity governed by various local chemical signals that determine the behavior of individual ants”.34 It seems compatible with Hurley’s use of the term to regard any entity as a unit of activity as long as it produces consequences such that it makes sense for a spectator to evaluate these effects in terms of goals assigned to the entity. Hurley stresses that different units of activity have different causal powers. Attributing such powers to an entity appears to be a condition for regarding the entity as having options to begin with. My suggestion is that that the problem of what binds a unit of activity together could be solved in terms of causal powers or dispositional properties. This solution is based on the assumption that assignments of dispositional properties carry with them defining limits of the object to which these properties are assigned. To regard an entity as having causal powers or dispositions is to view that entity as a causal agent.35 This notion of an agent is even weaker than Hurley’s ‘unit of activity’. Assignments of purely causal agency do not presuppose that any goals can be assigned to the agent in question. All causal agents are not units of activity in Hurley’s sense. This use of the terms “agent” and “agency” might seem linguistically inappropriate to some. Perhaps they would say, with Davidson, that agency by definition is intentional under some description: “a man is the agent of an act if what he does can be described under an aspect that makes it intentional.”36 “I am the agent if I spill the coffee meaning to spill the tea,

33 Hurley (2004: 208). 34 Hurley (2004: 206). 35 Tim Crane employs this term when he exposes his view that dispositions are independent causes. Crane (1998: 220). 36 Davidson (1971: 46).

156

Björn Petersson

but not if you jiggle my hand.”37 My spilling the contents of the cup is intentional, and that aspect is what makes me an agent of (unintentionally) spilling coffee. However, we sometimes categorize as acts reflexes, twitches, manifestations of common clumsiness and other types of behavior, which are not intended under any description. They fall under the notion of agency as different types of unintentional activities. Under the adjective “reflex,” OED lists “reflex action (independent of the will, caused as automatic response to nerve-stimulation)” and the noun “reflex” is said to be synonymous with “reflex action.” When I unknowingly talk loud to myself, or when I sneeze, you may think of me as doing something. How, then, do these sorts of acts differ from things that merely happen to me, like when my hair grows or the earth beneath me moves me away from the sun? We cannot appeal to the empirical fact that I cannot choose to influence the latter sorts of events.38 That inability is part of the story about reflexes as well. These are acts I cannot help doing. Verbs such as “grow” or “move” can be used to refer to processes or tendencies, or to activities. A man’s hair grows, and he may also grow a beard. The ferry moves me away from the island, and I may actively move away from things. Growing a beard and moving away from things can be actions that are intentional under some description. However, that active sense of “grow” or “move” does not entail the presence of any intention. A plant grows, but it is also capable of growing a new bud, as well as of moving its leaves towards the sun. Without resorting to anthropomorphism, we might even explicitly say about a chemical, that it acts upon a certain substance by making it corrode or dissolve, and we may point to explanatory conditions as being responsible for their effects. These everyday expressions indicate that ordinary speakers are capable of distinguishing a weak notion of agency, which is not intentional under any description. This way of speaking simply places the object in a certain causal role, and refers to an effect for which internal features of the object is a condition. The notion of agency we employ when we talk about reflexes and other types of human behavior without intentional aspects does not only bear a metaphorical resemblance to the causal notion characterized above. If you point out that your jiggling made me spill coffee, my causal role for that effect is stressed – although no intentions are in play. My spilling cof37 Davidson (1971: 45-46). 38 This is Irving Thalberg‘s suggestion about how to distinguish ”action verbs” from ”bodily process and reaction verbs”. Thalberg (1972: 62). For more detailed criticism of this suggestion, see Persson (1981: 15).

Collectivity and Circularity

157

fee is in this case something that I do, albeit unintentionally, so it is an act, but not in a stronger sense than the one outlined. The weak causal assumptions made about me in this case differ widely from the conditional statements that would have to be true in order for me to intentionally spill coffee. Although the causal notion of acting is weaker than the concept of ‘acting that is intentional under some description’, there is a substantial difference between descriptions in terms of causal agency and purely behavioral judgments. The latter are even less committing. The coffee incident could be de described in purely behavioral terms – your hand comes in contact with my wrist, my hand trembles, this causes coffee to be spilt, etc. – and that sort of description would be neutral with respect to causal agency. Similarly, we could describe the series of events that occur when the liquid and the metal meet, while leaving the question of causal agency open. A description of our going to New York, or our painting the house, which merely satisfies the behavioral conditions for these activities, would in itself contain no commitments concerning causal agency. Connected with the use of a broad merely causal notion of agency is a claim about dispositional properties. To say that the chemical acts upon a substance is to assume that there is something about the constitution of the chemical, which explains what happens to the substance under the actual conditions. “The plant moves its leaves” (as compared to “the plant’s leaves move”) contains a similar allegation about the constitution of the plant. If you stress that I was the one who spilled coffee, albeit unintentionally, you are implicitly claiming that there is something about my features, which explain why I spilled coffee under those conditions. This sort of claim is closely associated with the assignment of a dispositional property to the object in question – in the last case, perhaps clumsiness or feebleness. When we explain what happens with the metal in terms of what the acid does to it, and thereby place the liquid in the role of the causal agent, we presuppose a specific level of explanation. If we were chemists, we could have chosen to describe the process in terms of what the molecules do. We would then assign dispositional properties on the molecular level. Presumably, molecular dispositions form the base of the acid’s disposition of being corrosive. That does not imply conceptual reduction of descriptions of the latter to descriptions of the former. Even a chemist in the habit of thinking about chemical processes in molecular terms switches easily to macro level descriptions in contexts where it is linguistically in-

158

Björn Petersson

appropriate to go into molecular details when talking or thinking about what a certain chemical can do. Our choice of a certain object as the causal agent implicitly involves defining the unit to which causal powers are assigned (in the sense of drawing a boundary, not in the semantic sense of ‘define’.) The object’s components belong to it in virtue of their roles for the object’s intrinsic dispositional properties or causal powers. To regard a set of objects as one causal agent is to imply that something glues its components together, without being committed to the attribution of some specific glue. To illustrate my claim that disposition-assignments implicitly define the unit to which causal powers are attributed, let me use an example from the philosophical debate on the nature of dispositions. George Molnar presents the existence of “intrinsic maskers” of dispositions as a case against the idea that disposition-assignments can be reduced to conditional causal statements of a certain kind. According to Molnar, an object x may have a disposition D, and at the same time there may be another disposition intrinsic to x, D*, such that its being triggered prevents the manifestation of D. Intrinsically masked dispositions, like the “power of ingested poison masked by the power of ingested antidote”, are “everywhere once you start looking for them”. (So, Molnar claims, we cannot reduce assignments of dispositions to talk about conditional predictions of their manifestations.)39 It is essential to Molnar’s argument that the masking and the masked disposition are intrinsic to one and the same object. The assumption that they are intrinsic presupposes some boundary around whatever they are intrinsic to. Suppose substance y makes us tired. In the drug x this is outweighed by the effect of an additional substance z. We could say with Molnar that x has two dispositional properties. It is fatiguing, and it has the property of blocking manifestations of the first type of disposition. The drug then has two causal powers that happen to cancel each other out. A causal conditional analysis of dispositions would not admit that x possesses both these dispositions. On the other hand, we might deny that x has any dispositional properties of the kind mentioned. We could instead assign the property of being fatiguing to substance y, which is an element of x. That would be compatible with a (sufficiently subtle) conditional causal understanding of the disposition in question.40 39 Molnar (1999: 5). 40 The „intrinsic maskers“ example is meant to undermine even a subtle version of the conditional analysis, of the type proposed by David Lewis. (Lewis 1997) Lewis‘ analysis allows for the possibility of extrinsic maskers. If we assign the

Collectivity and Circularity

159

My point is not that any of these two ways of regarding the matter would be more plausible than the other, but that they exploit different underlying assumptions about the individuation of the bearer of the dispositional property in question, and that any assignment of causal powers or dispositional properties comes with such implicit assumptions. Our responses to the “intrinsic maskers” problem will not only depend on our intuitions about the conditional analysis of dispositions, but also on our ways of determining the boundaries of their bearers.41 How much structure do we require from a set of entities before we start placing them within the boundaries of one causal agent? Vice versa, under what circumstances will we switch from viewing an object as one causal agent to regarding it as a set of distinct but perhaps interdependent causal agents? These are questions of pragmatics. Their answers depend on which level of assignment of causal agency that is most convenient for explanatory or predictive purposes in the context. Someone familiar with ecology may find it natural to think about what one species does to another, or to the ecosystem, while others would tend to regard a species of plants or animals as nothing but a loosely connected set of distinct units of agency. On the other hand, when it comes to saliently demarcated macroscopic objects, like a spot of acid on a metal plate, few of us find it useful to regard that as a mere conglomerate of causal agents. Only certain spectators with specific background knowledge and training do that, and only in special circumstances. There are also cases such that most of us switch easily between the two perspectives. Ant colonies or swarms of bees would be paradigmatic exproperty of being fatiguing to substance y, and the property of blocking the manifestations of that disposition to substance z, the latter disposition would be an extrinsic masker of the first. 41 Dispositions are the subject of an old but still ongoing metaphysical debate and few if any claims about their nature are uncontroversial. Two claims relating to the concept of dispositions are essential to my argument here. Firstly, that regarding an entity as a causal agent – like when we think of the acid as acting upon the metal – is closely associated with assigning dispositional properties to the entity. Secondly, that assignments of dispositional properties carry with them defining limits of the entities to which they are assigned. These are claims about commitments inherent in ordinary people’s thinking and talking. My purpose here is to explicate a certain conception of causal agency that may figure in the content of ordinary people‘s attitudes. Although I think that these two assumptions are compatible with quite differing accounts of the nature of dispositions, they may not square with any such account. To explore the exact metaphysical implications of these proposals would reach beyond the ambitions of this paper, though.

160

Björn Petersson

amples. Evidently we are capable of thinking about a swarm of bees as a set of interdependent but distinct causal agents. We may also watch the swarm as one causal agent and think about what it might do, wonder what makes it fly this way rather than that, think that some of its acts seem unexpected, etc. That way of thinking does not require much knowledge about what keeps the swarm together. You may think that it has something to do with chemical signals while the small child beside you has a completely different theory. What makes both of you regard the swarm as the unit of causal agency need not be any specific knowledge about its internal structure, but just that its behavior gives you reasons to think that there is some such structure. In a similar manner, we are capable of changing perspective when we regard the behavior of people. Without invoking any concept of intention, we may distinguish between the behavioral disposition of a group of agents, and the set of behavioral dispositions of its members. A group of people may be disposed to smash a shop window under certain conditions. The base of this disposition could be a set of behavioral dispositions of the group’s members. This does not imply, though, that the behavioral disposition of the group is conceptually reducible to that set of individual dispositions. In other words, we can regard a group of individuals as the unit of causal agency, in a sense clearly distinct from that in which we view it as a set of causal agents. My contention is that to think of a set of individual actions as a collective activity is to regard the set as the causal agent. In regarding the group as the causal agent, we imply that there is some glue – there is something about the intrinsic features of the group and about the participants’ role in the base of the group’s causal powers, which distinguishes members from non-members – although we refrain from specifying this glue. This way of implying that there are internal features of an object responsible for the object’s behavior, without specifying those internal features, is characteristic of any kind of assignment of dispositional properties. Typically, assignments of dispositions are informative and explanatorily valuable in cases where it is non-trivial that the explanation of an object’s behavior is to be found among its internal features, although it would be impractical to go into the details of those features (because they are unknown, or it would take too long time, or be too costly for some other reason). The weak notion of collective activity, in which the group is conceived of as the causal agent, is not an ad hoc construction, designed to save a certain type of account of collective action from circularity. There are, I think, independent reasons to regard this way of thinking about groups as part of common sense. Apparently, we have the capacity to lump things

Collectivity and Circularity

161

like bees together in a way that enables us to regard sets of them as distinct units of causal agency. I see no reason to think that this conceptual ability must make an exception when it comes to sets of people. One sort of indication of our possession of a concept of collectivity without intention is that there is some empirical evidence for ascribing a conception of collectivity to two-year-old children.42 There is also well-known evidence strongly suggesting that the capacity to assign intentional states (beliefs, desires, intentions) to others as well as to ourselves is developed later in life – usually around four years.43 If both findings are correct, they prove that two-year-old children have a conception of collectivity that does not presuppose intention. However, experimenters’ descriptions are unavoidably theory–laden and empirical reports on issues like these are usually open to interpretation in accordance with more than one conceptual scheme. When it comes to the claim that the weak notion of collectivity is an element in ordinary people’s way of thinking and talking I will in the end have to rely on appeal to linguistic intuitions. Suppose you are engaged in a discussion while walking with a small group of people. After a while, you all realize that you have walked past the place you were supposed to stop at. Each of you just automatically followed the others without thinking about where you were heading. None of you intended to walk that far, let alone that the group should do so. Still, it seems plausible to regard your walking behavior as collective, in a sense distinct from a set of individual acts of walking that far. I would regard this as a typical example of a collective activity where no intentions concerning the collective activity (in this case, the activity of walking that far) bind the participants together. The analysis suggested here provides a way of understanding how we can think of such activities as being collective in a distinct sense.

5. Conclusion When someone is watching us smashing the shop window, he can think of what we are doing as a collective activity – a result of the group’s disposition being triggered. Then he simply thinks of the window’s being smashed as explained by features of the group as a whole, and the group as the unit of causal agency. In other words, he can regard our behavior as collective, 42 This empirical research was presented by Hannes Rakoczy, Max-Planck-Institute for Evolutionary Anthropology, at CollInt IV Siena 2004 (“Pretend Play and the Development of Collective Intentionality”, work in progress). 43 Cf. Gopnik (1993).

162

Björn Petersson

in a sense that does not presuppose shared intention. If each of us intends that the group smashes the window, and conceives of the intended activity as collective, i.e. of the set of agents as the unit of causal agency in what is intended, then we collectively intend to smash the window. (Here I assume that conditions of mutual knowledge and other requirements of the preferred analysis are fulfilled. I leave it open which additional conditions, if any, that are needed.) To use a handy hypostatization, we might say that this set of I-intend-that-we intentions is a collective intention. Insofar as this collective intention is what explains our smashing the window, it is a collective action. We do not need to place the notion of a jointly intentional achievement in the content of the intentions of the parties to a collective action. The weaker notion of collective activity is sufficient for that role. Take a paradigm case of collective action. A good soccer player thinks partly in terms of how his or her team should act, not merely in terms of individual achievement. Suppose each player at a certain moment in a game, perhaps at a signal from the coach, starts thinking that the best thing for the team to do would be to settle with the present score and go for mere defense. They should not attempt to score more goals at the price of risking the points they have at present. Each player then comes to intend that the team becomes more defensive. That requires each of them to think about the team in a way that concerns the team’s behavioral dispositions. This is the most convenient way of thinking about this change of strategy. They would think of the team as the unit of agency, and evaluate the options that lie within that unit’s causal powers. I guess most of us can recognize this simple switch between an individual and a collective perspective on the behavior of a group to which we belong. It seems farfetched to require that the players conceive of their intended change of strategy as being like tango dancing, i.e. an activity for which it is essential that each participant conceives of what they are doing as a jointly intentional activity. Occasionally thinking in that way about each other’s engagement might be important for getting a greater enjoyment out of playing soccer, compared to the pleasure of playing computer games or solitaire, for instance. Simple group actions within the game do not require that sort of self-reflective awareness, though. My claim is that a weak notion of collective activity must figure in the content of the intentions of the parties to a collective action. That claim is nevertheless stronger than Bratman’s initial proposal that it is sufficient that the act description, as it figures in the content of the agent’s intention, merely satisfies the behavioral conditions for a joint activity. Such a description would be neutral with respect to causal agency. (A behavioral description like ‘their arms move’ makes no allegations about causal agen-

Collectivity and Circularity

163

cy.) My additional requirement is that the notion of the team as the causal agent must enter that content. It might even be misleading to regard this requirement as appealing to intentions “characterized in cooperatively neutral ways,” in line with Bratman’s initial proposal.44 If we allow for a weak sense of co-operation, corresponding to the notion of causal agency, ants, slime mould and molecules can be said to co-operate in relation to the effects of the unit of causal agency to which they belong. The implicit assumption about there being something that binds the elements together could, as far as I can see, be expressed in terms of “causal co-operation.” Finally, I should make it clear that I do not deny that participants in some groups actually think of their intended common activity as being jointly intentional, i.e. that they intend their collective activity to be a collective action. Each participant may intend that each participant intends the activity to be collective. It is essential to some individual action-types, like voting, that they are carried out intentionally. When I intend to vote, I conceive of this act as intentional. In a similar manner, it may be essential to some collective action-types, like tango dancing, that they are jointly intentional. When we intend to dance a tango together, each of us conceives of our intended act as being jointly intentional. There is no reason to exclude the possibility of self-reflexive capacities of that sort. It is not implausible to think that such abilities may develop in the bootstrapping manner Bratman and Kutz suggest. The important thing is that this higher order capacity is inessential to the ability of performing a jointly intentional collective act – a collective action. All collective actions do not have to be that sophisticated.

References Bratman (1992a): Bratman, M. Shared Cooperative Activity, in: Bratman, M. Faces of Intention, New York: Cambridge University Press 1999, pp. 93-109. Bratman (1997b): Bratman, M. I Intend That We J, in: Bratman Faces of Intention New York: Cambridge University Press 1999, pp. 142-165. Crane (1998): Crane, T. The Efficacy of Content, in: Bransen, J., and Cuypers, S.E. (eds.) Human Action, Deliberation and Causation, Dordrecht and Boston: Kluwer 1998, pp. 199-225. Davidson (1971): Davidson, D. Agency, in: Davidson, D. Essays on Actions and Events, New York: Oxford U P 1980, pp. 43-61. 44 Bratman (1992: 97).

164

Björn Petersson

Gilbert (2003): Gilbert, M. The Structure of the Social Atom: Joint Commitment as the Foundation of Human Social Behavior, in: Socializing Metaphysics, Schmitt, F. (ed.), Lanham: Rowman and Littlefield 2003, pp. 39-65. Gilbert (2000): Gilbert, M. Sociality and Responsibility, Lanham: Rowman and Littlefield, 2000. Gopnik (1993): Gopnik, A. How We Know Our Minds: The Illusion of First-person Knowledge of Intentionality, in: Behavioral and Brain Sciences (1993) 16: 1-15 and 90-101. Humberstone (1997): Humberstone, I L. Two Types of Circularity, in: Philosophy and Phenomenological Research (1997) LVII/2: 249-280. Hume (1739): Hume, D.A. Treatise of Human Nature London: Fontana 1982, Book III. Hurley (2004): Hurley, S. Rational Agency, Cooperation, and Mind-reading, in Gold, N. (ed.) Teamwork, Multi-Disciplinary Perspectives, New York: Palgrave, 2004, pp. 200-15. Keefe (2002): Keefe, R. When Does Circularity Matter?, in: Proceedings from the Aristotelian Society (2002) CII: 253-270. Kutz (2000): Kutz, C. Complicity, New York: Cambridge University Press, 2000. Lewis (1997): Lewis, D. Finkish Dispositions, in: The Philosophical Quarterly, (1997) 47: 143-158. Molnar (1999): Molnar, G. Are Dispositions Reducible?, in: The Philosophical Quarterly, (1999) 49: pp. 1-17. Persson (1981): Persson, I. Reasons and Reason-governed Actions, Lund: Studentliteratur, 1981. Schmitt (2003): Schmitt, F. (ed.) Socializing Metaphysics, Lanham, MD: Rowman and Littlefield, 2003. Thalberg (1972): Thalberg, I. Enigmas of Agency, London: Allen and Unwin, 1972.

On the Ontology of Collective Intentionality: A Constructivist Perspective1 Antti Saaristo, Helsinki

1. Introduction In this paper I sketch a (social) constructivist view concerning the ontology of collective intentionality in general and the we-perspective, including the collective mode of beliefs, desires, intentions and actions, in particular.2 The argument I wish to put forward seeks to combine uncompromising naturalistic materialism in philosophy of mind with strong social constructivism concerning the origins of intentionality and intentional agency. Naturalism and constructivism are, admittedly, odd bedfellows, but my suggestion is that it is possible and indeed important to aim for a view that incorporates the best of both worlds. To keep the paper within an acceptable length the argument will have to remain somewhat tentative, but I do believe that the paper nonetheless sketches an important perspective to the ontology of collective intentionality.3 The somewhat platitudinous starting point of my argument is Max Weber’s assertion that when studying human affairs from a social scientific perspective we are interested in actions, i.e. meaningful behaviours, as opposed to mere bodily movements.4 Actions have conceptual content, the rationality of which can be assessed in terms of the reasons the agent has for her actions. When we want to account for meaning in this sense, we face the problem of rule following. In this paper I want to revisit the rule 1 *I would like to thank the audience at the 5th International Conference on Collective Intentionality (Helsinki, August - September 2006) for helpful comments and criticisms on an earlier version of this paper. 2 The animating idea behind the present argument features prominently also in Saaristo (2006b). The philosophical background of the view is discussed at length in Saaristo (2007). 3 In particular, the view I suggest is in direct contrast with John Searle’s (1990 and 1995) and Margaret Gilbert’s (1989) famous accounts of collective intentionality, for Searle thinks that collective intentionality is a biologically primitive feature of the (individual) human brain, while Gilbert opposes explicitly the Wittgensteinian view of intentionality that plays a major role in the present argument. 4 Weber (1947: 88). N. Psarros, H. B. Schmid and K. Schulte-Ostermann, Concepts of Sharedness, Frankfurt 2008: 165-189; ontos verlag

166

Antti Saaristo

following problem as presented by Saul A. Kripke in his interpretation of Wittgenstein’s Philosophical Investigations.5 More precisely, I shall examine the so-called social solution to the problem as a way to defend a materialistic and dispositional (and in that sense naturalistic) but nonetheless essentially social (and in that sense constructivist) view of the ontology of (collective) intentionality.6 Thus, although the social solution is far from being widely accepted, in this paper I am not defending the social solution as such but rather using it to create a novel view of the ontology of collective intentionality. Hence the main argument of my paper is conditional: if the social solution to the rule following problem is sound, then the kind of naturalistic constructivism I sketch is a strong candidate for a central component of the theory of collective intentionality.

2. The problem of rule following The rule following problem is a general problem that all accounts of meaningful activities, including accounts of meaningful behaviours (i.e. actions), must ultimately face. Kripke formulates the problem in the form of a sceptical challenge that is meant to question the basic assumption that some performances, such as claims, actions, forming of beliefs etc., have a determinate meaning. Let us begin by recalling the sceptical challenge. The starting point is the idea that we seem to learn the meaning of an item (a sentence, an action, etc.) on the basis of a finite sequence of examples. However, a finite sequence of examples satisfies infinitely many rules that determine meaning – how, then, is a determinate meaning possible? In the rule following literature it is customary to distinguish two aspects of this challenge: (1) The Infinity Problem: A finite sequence of examples instantiates infinitely many rules. Thus, whatever one does to continue the rule instantiated in the sequence is in accordance with the rule on some interpretation. On what basis does one decide how to continue the sequence, since whatever one does is compatible with the sequence on some interpretation? (2) The Normativity Problem: It must be possible to follow a rule, e.g. to use a term, correctly and incorrectly. Thus, there must be conceptual room for the distinction between how one in fact continues a sequence and how one ought to continue it. What determines the correct way to continue a sequence of examples? 5 Kripke (1982). 6 An important inspiration and precursor here is Esfeld (1999).

On the Ontology of Collective Intentionality: A Constructivist Perspective

167

The sceptical challenge is, in short, that it is difficult to see how the two problems could be resolved simultaneously, for they seem to pulling in opposite directions. Since the Infinity Problem is based on the logical fact that a finite sequence instantiates infinitely may rules, maybe there is no rational solution to the problem.7 However, there might nonetheless be a causal solution: Although any given finite sequence logically speaking instantiates infinitely many rules, perhaps for us it exemplifies only one rule.8 Perhaps one way of continuing the sequence simply strikes us as the natural way to continue the sequence. In other words, although the sequence does not by itself determine logically how it is to be continued, perhaps our shared biological nature as well as (social) upbringing make us also to share the sense of what counts as the obvious way to continue the sequence. Thus, perhaps the Infinity Problem forces us to hold that at the fundamental level of rule following we, as Wittgenstein puts this, do not consciously choose but obey the rule blindly in the manner we are causally disposed to do.9 This dispositional solution to the Infinity Problem seems to resolve the problem satisfactorily. Indeed, as Wittgenstein and Kripke argue, the Infinity Problem guides us away from theories of meaning that build on interpretations and other mental acts. Understanding a meaning is, according to this view, a purely biological event in which our causal dispositions activate, and not a moment of contentful illumination. The twist here, as we know, is that this move, essential as it is for resolving the Infinity Problem, serves only to make the Normativity Problem even more difficult. As Kripke points out, the dispositional solution tends to equate performance and correctness, and this appears to make the Normativity Problem practically irresolvable.10 The obvious difficulty here is that if in order to solve the Infinity Problem we need to see our applications of rules (how we continue sequences of examples) as based not on reasons and rational justifications but on blind causal dispositions, then there appears to be no room for assessing the correctness of our applications, for that would reintroduce rational deliberation and contentful interpretation into the picture, and thus fall prey to the Infinity Problem. But solving the Infinity Problem is not enough; we need to resolve also the Normativity Problem, 7 Note that it does not help to hold that the rule is a Platonist meaning, an abstract object, for then whatever one does to follow the rule would again be compatible with the rule on some interpretation. Interpretations are always open to re-interpretations – cf. Wittgenstein (PI: §198). 8 Pettit (1990: 36). 9 Wittgenstein (PI: §219). 10 Kripke (1982: 24).

168

Antti Saaristo

but unfortunately the two problems appear to require completely different and mutually exclusive solutions.11

3. The social solution to the problem of rule following The apparent dilemma between the requirements of the Infinity and Normativity Problems motivates the so-called social solution to the rule following problem. In short, the idea is that to resolve the Infinity Problem we must hang on to the core insight of the causal, dispositional solution. However, to cope with the Normativity Problem we must embed the individual disposition to a social setting, where the interplay of several individuals creates the background against which the correctness of each individual application of a rule can be assessed. In this paper I do not pretend to defend the social solution against all possible objections. Rather, my aim is to examine the picture of collective intentionality in general and we-mode action in particular that is implied by a well-developed version of the social solution. Thus, the view of collective intentionality and we-mode action I sketch in this paper stands or falls with the social solution to the rule following problem. In my view the core of the social solution can be summarised as follows.12 The Infinity Problem shows that to avoid circularity rule following must be analysed in strictly non-intentional, causal terms that do not already presuppose meaningful (or mental) content. But to resolve the Normativity Problem we must build a social element into our analysis so that for each individual participating in a social practice there is a difference between performance and correctness. However, also the social element must be built out of strictly non-intentional, causal building blocks. This is usually conceptualised so that the rule following problems requires a dispositionalist solution that builds on two sets of dispositions: the first-order dispositions to continue sequences of examples in one of the infinitely many possible ways, and the second-order dispositions to monitor the functioning of the first-order dispositions such that one non-intentionally adapts one’s first-order dispositions and behaviour to the dispositions and behaviour of others. In particular, one must be disposed to sanction those whose behaviour strikes one as inappropriate, and to be receptive 11 McDowell (1998: 242). 12 Cf. especially Esfeld (1999), Haugeland (1990), Haukioja (2000), Kusch (1999) and Pettit (1993 and 2002).

On the Ontology of Collective Intentionality: A Constructivist Perspective

169

to the sanctioning by one’s fellows. If in a group of individuals everyone is thus disposed, the practice of the group ought to converge towards the equilibrium where the first-order dispositions of different individuals are identical. To avoid falling back to the Infinity Problem this social harmonisation of first-order dispositions must be assumed to be non-intentional, practical agreement. This has two interesting implications. First, as Robert Brandom emphasises, this means that we should not think of rule following in terms of individuals relating to an external rule (for this would simply re-introduce the Infinity Problem).13 Rather, the rule is implicit in the blind, social practice of a group of individuals harmonising their firstorder dispositions. One could say that rule-governed behaviour is in this sense conceptually prior to and constitutive of rules, for explicating a rule is but an attempt to describe in words the equilibrium, or the coherent collective practice, towards which the first-order dispositions converge and which, of course, may change if individuals with radically different first-order dispositions invade the group.14 Second, the necessity to postulate co-operative, social dispositions already at the pre-intentional level shows that we should not think of individual agents fundamentally as self-sufficient (a)social atoms who simply follow their individual gameplans in an environment containing both living and non-living factors one needs to take into account, but as highly social members of communities such that their pre-intentional inclinations toward co-operation and practical agreement are constitutive of their individuality.15 This, of course, is methodological holism par excellence.16 So how, exactly, could we theoretically capture the intuitive idea that we have first-order dispositions to continue sequences of examples in a certain way without rational justifications, in addition to which we have secondorder social dispositions to harmonise the first-order dispositions with our fellows? Although I find Brandom’s, Pettit’s and, especially, Kusch’ attempts of explicating the social-dispositional view very promising, in my mind the strongest way of accounting for the idea that individuals are essentially and irreducibly social is the theory of collective intentionality in general and the we-mode/I-mode distinction in particular. This theory is explicitly meant to expound the idea that individuals are capa13 Brandom (1994). 14 This is the core of the much-criticised meaning finitism of the so-called Edinburgh school (e.g. Barnes et al. (1996)) – i.e. the claim that meanings are open-ended in the sense of being dependent on the development of our practices. 15 Cf. Barnes (2000: 56). 16 Saaristo (2003).

170

Antti Saaristo

ble of acting and thinking essentially and irreducibly qua group-members such that the individual applications are, as Searle puts this, derived from collective considerations.17 The notion of derivation of individual roles from a collective-level plan ought to capture the idea that the collective practice is normatively guiding individual applications. The problem is, however, that for example in Searle’s version of the collective intentionality theory the basis of sociality is intentional all the way down; in his view it is a primitive fact that biological brain states simply are intrinsically intentional and meaningful, which is of course precisely the view of mind and agency that the rule following argument is meant to challenge. Thus, Searle’s version of the collective intentionality theory is of no help in the present context. The same, unfortunately, must be said of Gilbert’s version of the theory, for her theory is explicitly anti-Wittgensteinian in this very aspect.18 However, the third major collective intentionality theorist Tuomela adopts a different line when he argues that the contents of intentional attitudes are social constructions.19 Accordingly, in his theory of collective intentionality Tuomela leaves room for a strong notion of togetherness and co-operation already at the purely causal pre-intentional (or sub-personal) level, and this in my mind is precisely where the rule following considerations and the theory of collective intentionality can meet fruitfully. However, since also Tuomela is mainly interested in analysing social actions in the context where we already have full-blown intentional agents, it is somewhat difficult to find an account of group action in his writings that would suit the present purposes, where, to avoid circularity, we need an account that employs exclusively non-intentional notions. Let us, for 17 Searle (1995). 18 Gilbert (1989). 19 Tuomela (2002: 79). Tuomela’s argument for social constructivism proceeds in the context Sellarsian semantics and thus it is a contribution to the rule following debate only indirectly (for Sellars builds on rule following considerations). Indeed, Tuomela ends up supporting something like the social solution sketched above only as a contingent claim in the sense that a social element is not in his view a conceptually necessary element of all meaningful activities, but simply the way in which we as a social species happen to solve the rule following problem. In his view there are possible worlds with non-social rule following. This is also the view of, for example, Haukioja (2000) and Pettit (1993). I tend to side with e.g. Brandom (1994), Kusch (1999) and Winch (1958) in thinking that a social element is necessary here, but for the rather modest purpose of the present paper – i.e. to show what we should think of collective intentionality if the social solution is accepted – I need not take a stand in this matter (see, however, Saaristo 2004b).

On the Ontology of Collective Intentionality: A Constructivist Perspective

171

example, take a look at Tuomela’s analysis of a reason-based we-intention of X to participate in a social practice P (where P is a social action type such as following a rule):20 X has a reason-based we-intention to participate in P iff (i) X intends P (ii) X believes that everybody in the relevant collection of agents intends P (iii) X believes that there is a mutual belief in the collection that everybody intends P (i) at least partly because of (ii) and (iii). As said, for the present purposes we need to turn the above reason-based analysis, which operates with intentional states such as intentions and beliefs, into a non-intentional, purely causal dispositional analysis. Consequently, the connection expressed by (iv) must be seen not as a reasonrelation but as a purely causal connection stating how the second-order social dispositions (ii) and (iii) monitoring the behaviour of others regulate the first-order disposition (i) to continue a sequence of examples. This reading of Tuomela’s analysis of collective we-mode intentionality results in the following social process where an individual’s dispositions (solving the Infinity Problem) are susceptible to the pre-intentional harmonisation with the dispositions and behaviour of others (solving the Normativity Problem): (RF) For any individual X in a community C, X’s performance p (an eventtoken) in a situation s (a unique situation) can be said to be rule-governed (by the rule “do P in S”) iff (i) X is disposed to treat s as an instance of a situation-type S. (ii) X is disposed to perform P (an event-type) in S. (iii) X performs p as an instance of P. (iv) The members of C share the dispositions (i) and (ii) and go along with (iii) (the members treat – tacit agreement in practice – p as a token of P). (v) (i) – (iii) hold at least partly because of (iv) (other causes may include X’s individual, biological first-order dispositions etc). The causal mechanism in (v) is to be understood in terms of second-order dispositions towards the practical harmonisation of the first-order dispositions, i.e. Barnes’ notion of “non-rational inclination toward agreement and co-ordination” which creates “agreement in practice”.21 Moreover, the 20 Tuomela (2002: 26). 21 Barnes (2000: 56 and 54).

172

Antti Saaristo

types S and P are constituted by the dispositions and thus do not commit the analysis to the existence of Platonist universals or anything similar. In sum, (RF) defines an interlocking system of first- and second-order dispositions that forms a social practice such that the totality of the practice implicitly constitutes the rule that determines the correct performance of P. Since the analysis operates with causal dispositions instead of rational interpretations, the Infinity Problem does not arise. However, since the individual applications are embedded in a communal practice within which the applications are (tacitly) assessed, from the point of view of the members of the community C there is a difference between individual performance and socially sanctioned correctness, and thus the social element in (RF) takes care of the Normativity Problem despite the fact that the sociality in question boils down to interlocking dispositions that as such are non-normative causal mechanisms. Or at least this is the bold claim of the social solution to the rule following problem. Thus, according to the social solution, meanings – and intentionality in general – reside, strictly speaking, in social practices instantiating intersubjective normativity, for only within practices such as (RF) can a biological state count as meaningful. Thus, an individual can be seen as an agent capable of intentional actions, contentful mental states and meaningful talk only to the extent she participates in such practices. In short, the psychological is constitutively dependent on the social, as one could formulate this claim that expresses also the core thesis of methodological holism.22 This means that the social solution is profoundly committed to the view that the evolutionary psychologists John Tooby and Leda Cosmides have famously labelled the Standard Social Science Model, or the SSSM for short, and which they wish to replace with the Integral Causal Model, or the ICM.23 Tooby and Cosmides characterise the philosophical core of the SSSM as locating meanings to the public sphere of culture such that individuals are seen to be capable for meaningful action only to the extent they participate in cultural practices.24 This is also the view the social solution is committed to. The ICM, in contrast, operates with biological, causal dispositions that are assumed to be intrinsically intentional and contentful. The main argument Tooby and Cosmides offer against the SSSM’s view of meaning and content is that in their view the SSSM is essentially committed to treating the human mind as a tabula rasa or, at most, as some 22 Bhargava (1992: 12). 23 Tooby and Cosmides (1992). 24 Tooby and Cosmides (1992: 26).

On the Ontology of Collective Intentionality: A Constructivist Perspective

173

kind of general-purpose information processing machine.25 Contrary to this model, Tooby and Cosmides think that the empirical neurosciences as well as evolutionary studies have established beyond all doubt that the human brain includes a number of different context-specific mechanisms (dispositions).26 In other words, the brain is neither a blank slate waiting to be written on nor a general-purpose computer. Instead of that the brain instantiates a system of innate dispositions with very specific scopes. An advocate of the social solution, however, may very well agree with this line of thought. A crucial step in the social solution to the rule following problem was the assumption that the brain is not merely a passive receiver of inputs or a general-purpose computer capable only of purely logical operations. Indeed, the resolution of the Infinity Problem assumed explicitly that the human brain contains context-specific, intrinsic (and evolved) dispositions, for example, to classify objects in a certain ultimately unjustifiable way and to harmonise the first-order dispositions to match the behaviour of others. Thus, the SSSM in the sense of the social solution is not only compatible with but actually presupposes the view that the human mind (qua biological) consists largely of functionally specialised mechanisms (dispositions) to produce behaviours that solve particular adaptive problems, such as preconditions for language, co-operation and so on. Moreover, one has to assume that these dispositions are sufficiently richly structured for constituting a complex architecture of first and second-order dispositions, including the social element. In short, I have argued that far from rejecting the picture of the biological mind as a collection of functionally specialised causal dispositions, philosophically mature versions of the SSSM (i.e. those building on the social solution)27 actually presuppose the view Tooby and Cosmides say is implied by the brain sciences. Thus, Tooby and Cosmides may be right in pointing out that the traditional empiricists’ picture of the biological mind as either a blank slate or a general-purpose computer is based on very primitive understanding of evolutionary biology and neuroscience. However, Tooby and Cosmides themselves are rather keen to cut corners when it comes to philosophical semantics and the metaphysics of intentionality. They appear to think that the rejection of the tabula rasa and the computer pictures of the biological 25 Tooby and Cosmides (1992: 28-29). 26 Tooby and Cosmides (1992: 24). 27 Needless to say, my short defence of the much-ridiculed SSSM applies only to the mentioned philosophical core of the SSSM. In particular, I am not committed to defending the specific views of any of the social theorists Tooby and Cosmides identify as paradigmatic representatives of the SSSM.

174

Antti Saaristo

mind implies a view where the context-specific, innate causal dispositions are intrinsically meaningful, i.e., the ICM. According to the social solution, in contrast, the mechanisms (dispositions) Tooby and Cosmides talk about may well be necessary for meaning, content and intentionality, but they cannot be sufficient (because of the Normativity Problem). The social solution holds that these phenomena include – exactly as the SSSM insists – an irreducibly social element: When it comes to the social sciences, the proper task for the natural scientific approaches is to study the biological preconditions of the SSSM, not, pace Tooby and Cosmides, replace the SSSM with the ICM or indeed with any other naïve oversimplification. Thus, the animating idea in the social solution is to drive a wedge between the causal dispositions of an individual on the one hand and the continuing intersubjective practice that carries meaning and normativity on the other. This idea sits remarkably well with the empirical work on human language acquisition and the evolution of human cognition carried out by Michael Tomasello and his co-workers.28 Note, however that although I argue that the actual empirical findings of Tomasello are compatible and supportive of the constructivist view of (collective) intentionality as sketched in this paper, Tomasello’s own pronounced understanding and interpretation of his findings is not: Tomasello uncritically accepts the standard view that intentionality is, as Searle puts it, biologically primitive feature of the brain and not a social construction. In particular, Tomasello is by no means committed to defending the SSSM against the ICM or advocating a constructivist theory of (collective) intentionality. Thus, by using Tomasello’s findings to defend a constructivist picture I side with, for example, Jones and Zimmerman, who build both on their own empirical research and on Tomasello’s work in order to argue “that intentionality (and other aspects of ‘mindedness’), usually treated as dwelling in a private mental domain, parades around in full public view as integral, facilitating practices of interaction”.29 In particular, they emphasise that intentionality in Searle’s and Tomasello’s sense, i.e. as an innate, biologically primitive feature of individuals, is not something empirical analysts must invoke to understand what is going on in their experiments involving interacting individuals: Rather, it [i.e. intentionality] is something that is consequential for the participants within the interaction (and not just inside their heads). Thus, we cannot only see “intentionality” as a cog-

28 See, in particular, Tomasello (1999) and Tomasello (2003). 29 Jones and Zimmerman (2003: 157).

On the Ontology of Collective Intentionality: A Constructivist Perspective

175

nitive and psychological process, but also, and perhaps in the first place, as a social and interactional phenomenon.30 Tomasello’s argument begins with the following puzzle: The time that separates humans from other great apes is evolutionary speaking a too short time to allow us to account for the obvious differences between human and non-human cognition exclusively in terms of biological evolution. After all, modern humans and chimpanzees share something like 99 percent of their genetic material. According to Tomasello, the solution to the puzzle is that humans evolved only one biological adaptation, the capacity for a species-unique mode of cultural transmission of cognitive skills such that meaningful practices may, so to speak, be coded in culture rather than in biology. Many non-human primate species are capable of producing new behavioural innovations, including tool-uses, but they seem to lack the capacity to adopt these innovations from others and improve them. In other words, it is not the biologically based creative component that the non-human primates are lacking, but the capacity to uphold the process of cumulative cultural evolution.31 Moreover, in Tomasello’s view the biological mechanism required (on empirical grounds) for this is the capacity for adapting to the behaviour of others, i.e. the non-instrumental susceptibility to the behaviours of others – or indeed the second-order harmonisation view the social solution argues to be required (on conceptual grounds) for rule following and, thereby, meaning and intentionality. As Tomasello and Hannes Rakoczy put this, precisely this kind of essentially social “derived normativity” of individual actions that the collective intentionality theory emphasises is “a key characteristic” in the architecture of the human cognition.32 The dual (biological and cultural) inheritance of human cognition can be seen clearly also in developmental psychology: Human infants have a very long period of immaturity when they are completely dependent on one or more parents. The compensating advantage of this extremely risky lifehistory strategy is that “it enables ontogenetic pathways that incorporate significant amounts of individual learning and cognition” and thus allows mature human cognitive skills to be largely cultural in nature.33 In other words, this allows human cognitive skills to be prototypical examples of the Dual Inheritance Theory, i.e. the view that “the mature phenotypes” of humans “are seen to depend on what they inherit from their forbears both 30 31 32 33

Jones and Zimmerman (2003: 179). Tomasello (1999: 5). Tomasello and Rakoczy (2003: 127). Tomasello (1999: 13).

176

Antti Saaristo

biologically and culturally”.34 In sum, “individual human beings possess a biologically inherited capacity for living culturally”,35 i.e. living a meaningful life as an intentional agent. Recall the analysis (RF) above that I argued to be the best explication of the causal, innate dispositions that, according to the social solution are required for learning a meaningful language or indeed for performing meaningful performances in general. In essence the analysis says that such individuals must be able to (i) carve up the world around them into objects and events in a way that reveals rather stable patterns (resolving the Infinity Problem), and that (ii) all this must be subject to constant social monitoring and harmonising (resolving the Normativity Problem). Tomasello reviews of number of empirical studies of infants and nonhuman primates and concludes that it is a well-established fact that both humans and non-human primates have very impressive skills for recognising visual and auditory patterns in very complex sensory stimuli.36 Tomasello argues that this skill is crucial for the development of language, but that it clearly is not sufficient, for also non-linguistic primates possess the skill. What humans have more than non-human primates is, according to Tomasello, the further ability to harmonise the recognised patterns with the ways in which others use similar patterns in practice. These two sets of abilities are precisely what the analysis (RF) requires intentional agents to be capable of at the pre-intentional level.37 The reality and importance of these two sets of abilities – that could be called pattern-finding skills and social harmonisation skills – are relatively wellestablished by empirical psychology and, as said, feature prominently also in (RF). Tomasello describes the pattern-finding skills as follows: These skills […] begin to emerge early in human development (some prelinguistically) and include such things as: • the ability to form perceptual and conceptual categories of “similar” objects and events;

34 35 36 37

Tomasello (1999: 14). Tomasello (1999: 53). Tomasello (2003: see, in particular, the references on p. 50). I have here followed Tomasello’s terminology of skills and abilities. However, the crucial lesson of Tomasello is that humans do not only possess these skills but also put these skills and abilities in actual, practical use even when it does not serve any obvious further end. Thus, the terminology of dispositions would both be substantially preferable and make the connection to the social solution even more obvious.

On the Ontology of Collective Intentionality: A Constructivist Perspective

177

the ability to form sensory-motor schemas from recurrent patterns of perception and action; • the ability to perform statistically based distributional analyses on various kinds of perceptual and behavioural sequences; • the ability to create analogies (structure mappings) across two or more complex wholes, based on the similar functional roles of some elements in these different wholes. These skills are necessary for children to find patterns in the way adults use linguistic symbols across different utterances, and so to construct the grammatical (abstract) dimensions of human linguistic competence. They are skills that are evolutionarily fairly old, probably possessed in some form by all primates at the very least.38 •

Such pattern-finding skills are necessary for resolving the Infinity Problem and shared by many non-linguistic animals. The social aspect is brought into the picture by a second set of skills, those of social harmonisation. These skills are “very likely unique to human beings, and they probably emerged relatively recently in human evolution”.39 Tomasello describes them as follows: These skills first emerge in human ontogeny at around 9-12 months of age and include such things as: • the ability to share attention with other persons to objects and events of mutual interest; • the ability to follow the attention and gesturing of other persons to distal objects and events outside the immediate interaction; • the ability to actively direct the attention of others to distal objects by pointing, showing, and using of other nonlinguistic gestures; • the ability to culturally (imitatively) learn the intentional actions of others, including their communicative acts underlaid by communicative intentions.

38 Tomasello (2003: 4 – references to a number of empirical studies omitted). 39 Tomasello (2003: 4).

178

Antti Saaristo

These skills are necessary for children to acquire the appropriate use of any and all linguistic symbols, including complex linguistic expressions and constructions.40 When the two sets of skills are combined, the result is the picture implied also by the social solution and (RF), namely the second-order, social dispositionalist view of rule following. Tomasello claims to have empirically shown that this is what the biological and innate basis of human cognition and language acquisition indeed looks like: the meaningful realm is a cultural realm, and our evolved, highly social dispositions make it possible for us to enter and reproduce this realm.41 Philosophers are sometimes hostile to social constructivism for the reason that they suspect that constructivism, including the social solution, is somehow essentially anti-naturalistic in nature. I trust that this section has removed these worries. First, I explained how the philosophical line of thought supporting the social solution is almost overly naturalistic. All the fundamental building blocks of a linguistic practice must be purely causal, blind dispositions; a theoretician cannot help herself to intrinsically intentional items. Second, I showed how the view is thoroughly naturalistic also in the more specific sense of being highly compatible with the empirical special sciences. General naturalism simply does not exclude the constructivist view of intentionality. Before I can proceed to explain what a constructivist view of collective intentionality could look like, there is one more clarification I need to make. This will prepare the ground for the way the issue is discussed in the following section. As we know, philosophers have since Frege’s days emphasised that there are two aspects to meaning, Sinn and Bedeutung, or sense and reference (or intension and extension). According to the social theory of meaning sketched in the present paper, both these aspects are dependent on rules. The referential aspect of meaning is based on concrete, rule-governed practices that include physical objects – recall how Wittgenstein opens his Philosophical Investigations by introducing a rudimentary language game that includes slabs, pillars and rocks as well as terms denoting these objects.42 Similarly, actions are meaningful in vir40 Tomasello (2003: 3 – references to a number of empirical studies omitted). 41 Interestingly, according to Tomasello (2003: 88-89), cases where chimpanzees have come very close to mastering something like a human symbolic language are those in which the chimpanzee is not taught the symbols, but rather included in structured human, social practices where symbols are used as part of concrete everyday practices. 42 Similarly, Tomasello emphasizes that the basic practices on which language is built “are not dyadic – between child and adult (or child and object) – but rather

On the Ontology of Collective Intentionality: A Constructivist Perspective

179

tue of normative rules governing practical reasoning. The sense-aspect of meaning is, then, based on rules that govern the relations between terms and assertions, i.e. theoretical reasoning. In other words, the social solution, as I see it, subscribes to a social version of inferential role semantics vis-à-vis both these aspects of meaning.43 For the purposes of the present paper there is no need to dig deeper into these issues; it suffices to understand that the social solution is deeply committed to the view that all meaningful activities are rule-governed and that this presupposes a social practice of blindly assessing the acceptability of individual performances that typically have the form of theoretical and practical inferences. In particular, actions as meaningful behaviours are such that their appropriateness must be normatively assessable in terms of the rules of rationality governing practical inferences and, thereby, constituting the space of reasons.

4. The social solution and intentional action Behaviour counts as action to the extent that it embodies a meaning, i.e. to the extent it can be rationalised. The social solution to the rule following problem holds that that the relevant rules of rational action etc. cannot be external to our communal practices of blindly treating certain actions as rational actions. This is important, since an influential tradition in philosophical action theory holds that Weber’s assertion – that action is meaningful behaviour – boils down to the view that an intentional explanation of action that rationalises a piece of behaviour by explicating the reasons behind the performance of the behaviour must be seen as explicating the normative relations that constitute the piece in question as meaningful, intentional action.44 This broadly speaking Davidsonian conviction is the next element in the present argument. Thus, let us look at the standard form of an intentional explanation of action:45 1. X desires D they are triadic in the sense that they involve infants coordinating their interactions with both objects and people, resulting in a referential triangle of child, adult and the object or event to which they share attention” (Tomasello 2003: 21). 43 See Brandom (1994) for a related defence of inferential role semantics and Saaristo (2007) for a discussion of the social-inferentialist take on reference. 44 Davidson (1980). For my reading of Davidson, see Saaristo (2006b). 45 Adapted from Rosenberg (1985).

180

Antti Saaristo

2. X believes that A is the best means to attain D under the circumstances (L) If any agent, X, desires D, and believes that doing A is the best means to attain D under the circumstances, then X does A Ergo, X does A. In Davidsonian action theory physical entities, such as brain states and bodily behaviours, can count as intentional entities, such as mental states and meaningful actions, precisely to the extent that they can be embedded, via intentional redescriptions, into a rational pattern such as the intentional explanation of action above. This is how Davidson puts the point: “My point is that if we are intelligibly to attribute attitudes and beliefs, or usefully to describe motions as [meaningful] behaviour, then we are committed to finding, in the pattern of behaviour, belief, and desire, a large degree of rationality and consistency”.46 Moreover, “the content of a propositional attitude [or meaningful action] derives from its place in the pattern”47 and “the satisfaction conditions of consistency and rational coherence may be viewed as constitutive of the range of applications of such concepts as those of belief, desire, intention and action”.48 Or, as Dagfinn Føllesdal sums up this view, “to be a desire or a belief [or a meaningful action] is just to be a factor that can figure appropriately in reason explanation”.49 And what counts as appropriate is defined by the rules of rationality such as (L). A crucial question that is often left open in the Davidsonian tradition concerns the ontological nature of the rules constitutive of intentional states and meaningful actions. This, of course, is precisely the question the social solution to the rule following problems aims to answer. The crucial aspect of the social solution here is that it strongly resists all forms of Platonism when it comes to such rules. The very point of the solution is to relativize the rules to social practices. In particular, they cannot be self-sufficient abstract object in need of interpretation, for that would simply bring in the Infinity Problem. For example, according to the social solution, the intentional explanation above has the particular form it does because this is how rational agents are collectively required to arrange their desires, beliefs and actions in order to count as rational agents capable of desires, beliefs and actions in the first place. In this view (L) is an attempt to make explicit a norm that resides implicitly in our practices 46 47 48 49

Davidson (1980: 237). Davidson (1980: 221). Davidson (1980: 237). Føllesdal (1994: 306).

On the Ontology of Collective Intentionality: A Constructivist Perspective

181

of giving and asking for reasons for our actions. In particular, the social solution holds that the form or explicit model of an acceptable inference (such as the standard belief-desire model) and indeed (L) cannot be conceptually prior to accepting certain inferences as valid in practice. According to the social solution, the rule following considerations show that we cannot justify our rule-governed practices by appealing to explicit rules. Rather, formulations of explicit rules such as (L) or models of practical inference are attempts to make explicit the essential components of an implicit practice. In short, the social practice view I have sketched in this paper treats as a valid form of practical inference any form that is (tacitly) accepted as valid in concrete social practices. Crucially, there are familiar cases of practical reasoning that do not operate in terms of the standard I-mode belief-desire pair. Brandom uses the following example:50 (a) It is raining. (b) I shall open my umbrella. According to the traditional orthodoxy in action theory we ought to insist that such an inference is crucially incomplete – we must assume that at least the following premises are implicitly present: (c) I desire to stay dry. (d) I believe that it is raining. (e) I believe that if it is raining I must open my umbrella in order to stay dry. However, when we understand that the form of intentional explanation is determined by our own (collective) practice, it becomes clear that the insistence of adding the extra premises (c)-(e) is justified only if it is our practice to require them to be added. But, typically, we do not in fact (indeed in practice) require such additions. In our practices answering, “it is raining” is generally accepted as a full answer to the question concerning one’s reasons for opening one’s umbrella. Thus the crucial lesson here is that according to the social solution, the form of intentional explanation (such as the belief-desire model) is not forced upon us by any universal rules of rationality. For example, the inference from “it is raining” to “I shall open my umbrella” is acceptable because it sits well with our practices, which implicitly include a norm that could be made explicit as “if it is raining, one is rationally entitled 50 Brandom (2000: 87). I have discussed the same example in a related manner in Saaristo (2006b) and Saaristo (2007).

182

Antti Saaristo

(or even required) to open one’s umbrella”. Thus, no further justifications are required.51 Bearing this in mind, let us next consider a situation where an agent faces a situation she recognises to be a social dilemma situation where mutual co-operation is the collectively optimal (best for us) thing to do. Note that this is an issue that is generally seen as a paradigmatic example of the difference between I-mode rational action and we-mode rational action.52 I-mode agents conceptualise a social dilemma situation in terms of their individual preferences (that can of course be benevolent) and aim to realise the best outcome from the point of view of their preferences by acting strategically, i.e. by taking into account the expected behaviours of others. We-mode agents, in contrast, address the situation from a collective perspective of seeing the situation as a collective task where the players work together to realise whatever outcome is collectively best for them, dissolving thus the whole setting of a game and the need for strategic action in the sense of game theory. In this context the question about the nature of we-mode collective intentionality turns into a question concerning the fundamental difference between I-mode co-operation and we-mode co-operation in social dilemma situations. In light what I have said above, I suggest that this further question could be asked by querying what we ought to say concerning the following practical inference: (a) In this (social dilemma) situation co-operation is collectively optimal. (b) I shall co-operate. A champion of the theory of collective we-mode intentionality will hold that the inference from (a) to (b) captures accurately and completely the psychological processes underlying we-mode social action.53 The standard I-mode individualism, in contrast, insists that the inference has to be completed by adding the following premise: 51 Note that the seemingly universal principle (L) in the standard form of intentional explanation is, in fact, far from being universal: (L) includes an indispensable ceteris paribus clause – we can never know that all relevant issues have been taken into consideration (this is also why I said above that in the standard view at least (c)-(e) must be added to the inference). Although this is a very serious problem for non-constructivism, from the perspective of the social solution the ceteris paribus clause is indispensable precisely because there is no other fact of the matter as to when further qualifications are no longer needed than the social practice of (tacitly) accepting certain reasons as sufficient. 52 Cf. e.g. Bacharach (2006), Tuomela (2000) and Tuomela (2007). 53 Cf. Searle’s (1995) discussion on how we-mode agents derive their individual contributor-intentions from collective-level considerations.

On the Ontology of Collective Intentionality: A Constructivist Perspective

183

(c) I desire to realise collectively optimal outcomes. Or, alternatively (or indeed something similar), (c’) I desire to do what is right. (d’) I believe that in a social dilemma situation the right thing to do is to realise the collectively optimal outcome. However, according to the social solution, the individualistic obsession to add further, explicitly I-mode premises to the inference cannot be justified by appealing to any universal rules of rationality. According to the social solution, whether the inference from (a) to (b) is incomplete or not depends solely on our practices of treating it either as an acceptable form or as an unacceptable form of intentional explanation in the game of giving and asking for reasons. Thus, the question about the ontological status of collective we-mode intentionality, according to the view sketched in this paper, is ultimately a question about our inferential practices. If we in practice treat the inference from (a) to (b) as acceptable, then collective we-mode intentionality is indeed irreducible to I-mode intentionality. However, if it is our practice to require the inference to be completed by adding premises (c), (c’) and (d’), or something of that sort, then we must hold that we-mode intentionality does in fact reduce to more fundamental I-mode intentionality. This twist in the story concerning the ontological status of collective intentionality has some interesting implications, two of which I wish to emphasise here. First, to find out whether collective we-mode intentionality is a real phenomenon requires one to study concrete inferential practices rather than what is happening in the heads of individuals. This is a task for empirical social scientists.54 However, also philosophers and other theoreticians can play a role here. I am thinking of the work of, for example, Brian Skyrms, who argues forcefully that human populations must be assumed to evolve to include an implicit norm of practical inferences that could be made explicit precisely as the rule that the inference from (a) to (b) is to be treated as a valid inference.55 Second, since the reality of collective we-mode intentionality depends on what kind of inferences we treat as acceptable in our everyday practices it is in a sense up to us to make the collective we-mode intentionality either 54 See, for example, Kerr and Park (2001), whose work appears to suggest that the inference from (a) to (b), and thus collective we-mode intentionality, is indeed rendered acceptable in actual practices. 55 Skyrms (1996).

184

Antti Saaristo

a sui generis phenomenon or a feature reducible to individualistic I-mode intentionality. This means that the question concerning the ontological status of collective intentionality is not only a descriptive question. It is also a profoundly normative or indeed political issue: which form of agency we ought to make true in our practices of treating certain inferences as valid and complete and certain other inferences as invalid or at least incomplete?56 Do we want to live in societies that consist of individuals who reason in collectivistic terms (in the we-mode), or do we insist that people should be more autonomous, disparate individualists who always ask whether a collective task really is in their personal interest, be that interest egoistic, altruistic or what have you and, accordingly, reason in individualistic terms (in the I-mode)?

5. Conclusion In this paper I have argued that the so-called social solution to the problem of rule following can be developed into a forceful argument in favour of rather strong constructivism regarding the ontology of collective intentionality. According to this view, collective we-mode intentionality is – as is all intentionality – always constructed in social practices. However, this kind of constructivism is not anti-naturalistic in any suspicious sense, for the building blocks of the practice in question are – and must be – assumed to be purely naturalistic, biological dispositions. Indeed, from the point of the social solution the standard assumption that biological brain states are somehow intrinsically intentional or even contentful is a very dubious departure from causal naturalism.57 I think this line of thought can offer at least the following three important proposals to the current debates on the theory of collective intentionality. (1) An important but somewhat neglected58 matter in the theory of collective intentionality is the underlying view of the general philosophy of mind. Of the major collective intentionality theorists Gilbert and Searle are explicitly against the Wittgensteinian social view and Searle in particular emphasises that in his view intentionality, be it individual or collective, is a biologically primitive feature of the brain. Tuomela, on the other hand, seems to build a picture compatible – at least to a certain degree – with the constructivist view sketched in this paper, for although he 56 This is the main thesis in Saaristo (2006b). 57 Saaristo (2004a). 58 For example, Deborah Tollefsen’s (2004) influential review does not consider these issues at all.

On the Ontology of Collective Intentionality: A Constructivist Perspective

185

often toys with the idea that collective we-mode intentionality is a result of evolution and as such a seemingly purely biological phenomenon, he nonetheless emphasises strongly that the evolution in question must be seen as bio-cultural co-evolution.59 I would like to suggest that the social solution could be used to pinpoint what, exactly, is the scope of the biological (and, correspondingly, the social) in the emergence of full-blown (collective) intentionality. This is the motivation behind the strictly nonintentional, dispositionalist version (RF) of Tuomela’s analysis of we-intentionality in Section 3 above. Actually, this applies largely to Searle’s views on intentionality as well. Searle explains that “[i]ntentional states function only given a set of Background capacities that do not themselves consist in intentional phenomena […] By capacities I mean abilities, dispositions, tendencies, and causal structures generally”.60 Indeed, Searle even states explicitly “that much of Wittgenstein’s later work is about what I call the Background”.61 The social solution argues that rule following and, thereby, intentionality requires that a number of individuals possess a system of first- and second-order causal dispositions such that the blindness of the dispositions resolves the Infinity Problem and the social structuredness of the dispositions resolves the Normativity Problem by allowing an individual’s dispositions to be guided and shaped by the behaviour of others. Similarly, Searle explains that Instead of saying, the person behaves the way he does because he is following the rules of the institution [such as language], we should just say, First (the causal level [i.e. the dispositional solution resolving the Infinity Problem]), the person behaves the way he does, because he has a structure that disposes him to behave that way; and second (the functional level [i.e. the social solution to the Normativity Problem along the lines of (RF)]), he has come to be disposed to behave that way, because that’s the way that conforms to the rules of the institution.62 Moreover, Searle explicitly admits that human institutions include “a socially created normative component”.63 Now, all this sounds very compatible with the social solution and the constructivist view of (collective) intentionality. The main difference is that Searle appears to think that 59 60 61 62 63

Tuomela (2007). Searle (1995: 129). Searle (1995: 132; see also – or perhaps primarily – pp. 140 and 143). Searle (1995: 144). Searle (1995: 146).

186

Antti Saaristo

the brain states made possible by the non-intentional Background dispositions, including the ones producing the normativity-creating social element of human institutions, are nonetheless intrinsically intentional states – and the social solution denies this. Building on Wittgenstein’s rule following arguments, the social solution holds that the notion of intrinsically intentional physical states is not only ontologically very dubious, but also question-begging or powerless when facing the Infinity Problem. Instead, the social solution admits that certain physical states and systems are intentional, but not intrinsically. They are intentional in virtue of being assigned a normative status in social practices. The intentionality of our cognitive capacities is not intrinsic but dependent on a social system of interlocking first- and second-order dispositions. This is compatible with Searle’s discussion of intentional states being dependent on a complex system of non-intentional, causal dispositions that he calls the Background, but not with Searle’s portrayal of intentionality of a biologically primitive feature of the brain. (2) A problem closely related to Point (1) above is that the attempts to naturalise collective intentionality – be it in terms of neurology or evolutionary biology – tend to assume that certain rather primitive biological features (say, mirror neurons or co-operative dispositions) are somehow intrinsically intentional and contentful as such. The marriage of the theory of collective intentionality and the social theory of rule following that I put forward in this paper would re-interpret these attempts in metaphysically more modest light: We would no longer say that the studies of certain neural processes or evolved tendencies are studies of intentional capacities (or their realisers) as such. Rather, such studies would be seen as studies of the biological preconditions of the social practices (such as (RF) above) that constitute the possibility of all intentionality, for the ontological home of intentionality would be in social practices, not in biological processes. For many this would be a metaphysically more acceptable view than the idea that certain biological features are intrinsically intentional. Thus, the kind of social theory of (collective) intentionality I suggest in this paper would be both constructivist and highly naturalistic without implying eliminativism. (3) Finally, the view proposed in my paper would imply precisely the kind of constructivism required for nailing down the problematic role of the wemode in the theory of collective intentionality. Recall that the paradigmatic instantiation of we-mode collective intentionality is collectively rational behaviour in collective action dilemmas, which (in)famously is not individually rationalizable. By portraying intentionality essentially in terms of socially sanctioned inferential practices, the constructivist view defended

On the Ontology of Collective Intentionality: A Constructivist Perspective

187

in my paper allows us to approach collective we-mode intentionality in terms of social practices that support collectively rational co-operation in collective action dilemmas – practices that, for example, Skyrms has forcefully argued to evolve in most populations.64

References Bacharach (2006): Bacharach, M. Beyond Individual Choice: Teams and Frames in Game Theory, Princeton: Princeton University Press, 2006. Barnes (2000): Barnes, B. Understanding Agency: Social Theory and Responsible Action, London: Sage Publications, 2000. Barnes et al. (1996): Barnes, B., Bloor, D., and Henry, J. (1996). Scientific Knowledge: A Sociological Analysis, London: Athlone, 1996. Bhargava (1992): Bhargava, R. Individualism in Social Science: Forms and Limits of a Methodology, Oxford: Clarendon Press, 1992. Brandom (1994): Brandom, R. B. Making it Explicit: Reasoning, Representing, and Discursive Commitments, Cambridge Mass.: Harvard University Press, 1994. Brandom (2000): Brandom, R. B. Articulating Reasons: An Introduction to Inferentialism, Cambridge, Mass.: Harvard University Press, 2000. Davidson (1980): Davidson, D. Essays on Actions and Events, Oxford: Clarendon Press, 1980. Esfeld (1999): Esfeld, M. Rule following and the Ontology of the Mind, in: Meixner, U., and P.M. Simons (eds.), Metaphysics in the Post-Metaphysical Age. Papers of the 22nd International Wittgenstein Symposium, Kirchberg: Austrian Ludwig Wittgenstein Society, 1999, pp. 191-196. Føllesdal (1994): Føllesdal, D. The Status of Rationality Assumptions in Interpretation and the Explanation Action, in: Martin, M. and L.C. McIntyre (eds.), Readings in the Philosophy of Social Science, Cambridge Mass.: The MIT Press, 1994, pp. 299-310. Gilbert (1989): Gilbert, M. On Social Facts, London: Routledge, 1989. Haugeland (1990): Haugeland, J. The Intentionality All-Stars, in: Philosophical Perspectives, (1990) 4: 383-427. Haukioja (2000): Haukioja, J. Rule following, Response-Dependence and Realism, Turku: Turku University Press, 2000. 64 Skyrms (1996).

188

Antti Saaristo

Jones and Zimmerman (2003): Jones, S.E. and Zimmerman, D.H. A Child’s Point and the Achievement of Intentionality, in: Gesture, (2003) 3: 155-185. Kerr and Park (2001): Kerr, N.L. and Park, E.S. Group Performance in Collaborative and Social Dilemma Tasks: Progress and Prospects, in: Hogg, M.A., and R.S. Tindale (eds.), Blackwell Handbook of Social Psychology: Group Processes, Oxford: Blackwell, 2001, pp. 107-138. Kripke (1982): Kripke, S.A. Wittgenstein on Rules and Private Language, Oxford: Basil Blackwell, 1982. Kusch (1999): Kusch, M. Psychological Knowledge: A Social History and Philosophy, London: Routledge, 1999. McDowell (1998): McDowell, J. Mind, Value, and Reality, Cambridge, Mass.: Harvard University Press, 1998. Pettit (1993): Pettit, P. The Common Mind: An Essay on Psychology, Society, and Politics, Oxford: Oxford University Press, 1993. Pettit (2002): Pettit, P. Rules, Reasons, and Norms: Selected Essays, Oxford: Clarendon Press, 2002. Rosenberg (1995): Rosenberg, A. Philosophy of Social Science (2nd ed.), Boulder: Westview Press, 1995. Saaristo (2003): Saaristo, A. On the Objectivity of Social Facts, in: ProtoSociology, (2003) 18-19: 291-316. Saaristo (2004a): Saaristo, A. Personhood as a Social Status, in: Ikäheimo, H., J. Kotkavirta, A. Laitinen, and P. Lyyra (eds.), Personhood, Jyväskylä: University of Jyväskylä, 2004, pp. 192-198. Saaristo (2004b): Saaristo, A. Merkitys sosiaalisena käytäntönä: kohti ihmistieteiden filosofisia perusteita, in: Ajatus, (2004) 61: 81-114. Saaristo (2006a): Saaristo, A. Intentional Action and the Limits of Scientific Naturalism: Davidson’s Alleged Refutation of the Logical Connection Argument, in: Pihlström, S., H.J. Koskinen, and R. Vilkko (eds.), Science – A Challenge to Philosophy?, Frankfurt am Main: Peter Lang, 2006, pp. 247-257. Saaristo (2006b): Saaristo, A. There Is No Escape from Philosophy: Collective Intentionality and Empirical Social Science, in: Philosophy of the Social Sciences, (2006) 36: 40-66. Saaristo (2007): Saaristo, A. Social Ontology and Agency: Methodological Holism Naturalised, unpublished Ph.D. thesis, London School of Economics and Political Science (University of London), 2007.

On the Ontology of Collective Intentionality: A Constructivist Perspective

189

Searle (1990): Searle, J.R. Collective Intentions and Actions, in: Cohen, P. R., Morgan, J. and Pollack, M. E. (eds.), Intentions in Communication, Cambridge, Mass.: the MIT Press, 1990, pp. 401-415. Searle (1995): Searle, J.R. The Construction of Social Reality, London: Penguin, 1995. Skyrms (1996): Skyrms, B. Evolution of the Social Contract, Cambridge: Cambridge University Press, 1996. Tollefsen (2004): Tollefsen, D.P. Collective Intentionality, in: Internet Encyclopedia of Philosophy, http://www.iep.utm.edu/, 2007. Tomasello (1999): Tomasello, M. The Cultural Origins of Human Cognition, Cambridge, Mass.: Harvard University Press, 1999. Tomasello (2003): Tomasello, M. Constructing a Language: A Usage-Based Theory of Language Acquisition, Cambridge, Mass.: Harvard University Press, 2003. Tomasello and Racokzy (2003): Tomasello, M., and H. Rakoczy, H. What Makes Human Cognition Unique? From Individual to Shared Collective Intentionality, in: Mind and Language, (2003) 18: 121-147. Tooby and Cosmides (1992): Tooby, J. and Cosmides, L. The Psychological Foundations of Culture, in: Barkow, J. H., L. Cosmides, and J. Tooby (eds.), The Adapted Mind: Evolutionary Psychology and the Generation of Culture, Oxford: Oxford University Press, 1992, pp. 19-136. Tuomela (2000): Tuomela, R. Cooperation: A Philosophical Study, Dordrecht: Kluwer, 2000. Tuomela (2002): Tuomela, R. The Philosophy of Social Practices: A Collective Acceptance View, Cambridge: Cambridge University Press, 2002. Tuomela (2007): Tuomela, R. The Philosophy of Sociality: A Shared Point of View. New York: Oxford University Press, 2007. Weber (1947): Weber, M. The Theory of Social and Economic Organization, New York: The Free Press, 1947. Winch (1958): Winch, P. The Idea of a Social Science and its Relation to Philosophy, London:Routledge and Kegan Paul, 1958. Wittgenstein (1953): Wittgenstein, L. Philosophical Investigations, Oxford: Blackwell, 1953.

Agent Causation and Collective Agency Katinka Schulte-Ostermann, Essen

1. Introduction For the so called “agent causation theory” of action a given human behaviour counts as an action, if and only if an agent having the intention to perform the action in question has caused it.1 The theory of collective action and intentionality implicitly or explicitly rests on this assumption, since otherwise the distinction between individual and collective actions would not be possible. This is so because speaking about collective actions always presupposes that there is a collective subject consisting of a plurality of different individuals, each one of which has it own individual intentions concerning this action, and not that the collective action in question is caused merely by a plurality of distinctive intentions. In contrast, a single individual action can be caused by a single individual having different though interrelated intentions. The performance of an actor, for example, is not only a consequence of her intention to perform a certain role but as well of her intention to earn money and of her intention to become a famous person. She might not perform this role if it were not a means for her to earn money and to become a celebrity, but she might want at the same time to earn money and become famous only by performing this certain role, namely Antigone. All these intentions are necessary conditions for her action, which is the performance of the role of Antigone. Therefore an action caused merely by a set of interrelated intentions needs not be the action of a collective subject, although for a collective subject to act there have to be some interrelated intentions among its members concerning the collective action itself. One main problem of theories of collective actions is therefore the constitution or the “structure” of such a collective subject. Common examples of collective actions are cases of two or more persons walking together, playing soccer, or giving a violin concert. In all of these cases, the individual participants must have at least some intentions, which are the cause of the collective action. In case of individual actions, the intentions have to be the intentions of one acting person. This is not possible in case of a col1 Chisholm (1976). N. Psarros, H. B. Schmid and K. Schulte-Ostermann, Concepts of Sharedness, Frankfurt 2008: 191-208; ontos verlag

192

Katinka Schulte-Ostermann

lective action: In order to partake in a collective action everyone involved in it has to have an intention concerning the collective action itself. This means that there are at least as many intentions as agents.2 On the other hand, having an intention concerning a collective action does not entail that whatever one does is part of this collective action or that the totality of one’s life is identical with the participation in this action. What is needed is a kind of intentions that are directed towards a certain collective action and that are separate from all other intentions a person may form – this kind of intentions is called in the discourse of collective intentionality “we-intentions”. To be a participant in a collective action implies that one has the we-intention to perform the action in question together with the other participants. In recent discussions on the notion of collective intentionality two main positions have evolved: The first – taken by Raimo Tuomela – can be called individualistic or atomistic and claims that the notion of collective intentionality can and should be reduced to the notion of we-intentions and beliefs of individuals, since otherwise it seems necessary to postulate something like a “group soul”, which produces the relevant intentions and actions.3 The second position – taken by Margaret Gilbert – claims that it is impossible for a collective action to come into existence if only we-intentions of individuals and individual beliefs are presupposed, so that there must be some collectively shared commitments prior to the collective action, which are irreducible to the beliefs of the individual members of the collective.4 The main thesis of this paper is that the controversy between individualistic and holistic explanations of collective agency rests on the action-theoretical concept of agent causation. This is because only if one presupposes that the main difference between individual and collective action lies in the difference of the nature of the subject of each action type then it becomes relevant to give an account of the relationship between the action and a plural subject, a relationship, which is at least not easily reducible to the relationship between an individual person and her actions. Thus there exists not only an explanatory gap between a collective action and a collective agent (first gap), but also between an individual agent and her actions (second gap). In order to defend this claim I aim at showing that: a) there is up to date no convincing account of how to bridge the second gap, and that 2 Even if some of the individuals may only have the intention not to do certain actions, and thereby enabling the others to fulfil their task. 3 Tuomela (2001: 102 – 135). 4 Gilbert (2000: 14 – 36).

Agent Causation and Collective Agency

193

b) bridging the first gap presupposes bridging the second gap. In the first step of my demonstration I will show that all explanations within the framework of the theory of individual of agent causation are either reducible to the classical approach of event causation as introduced by Donald Davidson, or not convincing for various reasons. In the second step I will discuss the consequences of the failure to bridge the gap between agent and action by the notion of agent causation for the relationship between the collective and the collective action. Finally, I will argue that there are two possible solutions to this problem: The first consists in developing a more convincing theory of agent causation than the ones developed so far. The second is the suggestion to argue in favour of an agent theoretical account, which does not need to bridge the second gap. Thus the second solution is to undertake an explicitly strict anti causalist approach to the phenomenon of action.

2. Agent causation Defenders of agent causation take their starting point from the common observation that persons, and not some event, are the subjects of actions. A motivation for this account lies in the intuition that any meaningful concept of freedom of action entails that it is the agent who is responsible for an event and not just some other event. And because freedom of action is only granted if the agent is not himself caused by an event, the concept of agent causality is introduced. Its main characteristic is that an agent can be a cause without being himself caused while event causality always implies that every non-agentive event is caused by a non-agentive event and that every cause is caused by another cause.5 Within this conception the question arises how events caused by a person and events caused by other events are related. It is important to notice here that proponents of agent causation normally are not defending the position that the agent directly causes his bodily movements, but that his acquiring-an-intention is his “true” action and that this intention causes the bodily movement.6 Generally speaking, for the proponent of agent causation an action consists in acts of “bringing something about” or in acts of “intending to bring something about”. Bodily movements only 5 Usually it is assumed that this position has evolved from Kant’s distinction between the “causality from nature” and the “causality from freedom”. Cf. Kant (KdrV: Third Antinomy B472 – B479). 6 Chisholm (1966: 20), O’Connor (2000: 72), and O’Connor (2002: 352).

194

Katinka Schulte-Ostermann

count as actions because they are caused by such acts of bringing about or intending. The person counts as a kind of immovable mover and it is this ability, which is the conditio sine qua non of freedom of action. However, cognitive and neuronal science has apparently provided some evidence against this conception of agent causation and freedom, the experiments of Libet being one well-known example. This and other experiments led to the conclusion that bodily movements are directly connected with the neuronal states of the brain and are caused by them. The postulation of an agent causing this connection seems thus redundant. The argument is supported by the discovery that the brain has apparently no centre of consciousness or even of activity.7 Therefore the question of the bodily locus of personality is open again. Chisholm’s answer that a person’s having an intention causes the change of her neuronal states is problematic, because it does not clarify what kind of being a person is. The relation between personality and the human body remains thus a mystery. We will try to unveil this mystery by examining Kant’s, Bishop’s and O’Connor’s accounts of agent causation. Chisholm’s theory will not be pursued any further since it is confined to the description of the difference between agent and event causation and does not provide any explanation for agent causation.8 2.1 Kant’s “Causality from Freedom” This thesis is closely connected to Kant’s concept of “Causality from Freedom” as explicated in the Critique of Pure Reason. Kant discusses there two apparently contradicting claims, namely that: i) it is necessary to postulate the causality from freedom in order to give an complete account of everything in the world,9 and ii) that freedom does not exist, since everything in the (material) world is determined by the laws of nature.10 Kant’s solution of this dilemma gave reason to many critical reactions.11 However, this is not here our concern. Our attention will be focused instead to the question whether his account of “Causality from Freedom” can be seen as a form of agent causation and whether this form of causation is really something irreducible to event causation. It is therefore necessary to take a closer look at both claims i) and ii). 7 8 9 10 11

Dennett (1991: Ch. 5). Chisholm (1966). Kant (KdrV: A444/B472). Kant (KdrV: A445/B473). Cf. Smith (1923: pp. 492 – 495 and 512 – 519), Wilkerson (1976: 130 – 139), Kreimendahl (1998), and Allison (1998).

Agent Causation and Collective Agency

195

Claim i) is antireductionist in the sense that it states that there are events, which are not explainable by the laws of nature. Therefore it is necessary to postulate the sort of causality that he calls “Causality from Freedom”. i) can be taken as an epistemic claim insofar, as it tells us something about the possibility of explanations. In contrast claim ii) is an ontological one. It states that freedom is something that does not exist. Because claim i) is epistemic and claim ii) an ontological they do not necessarily contradict each other. Postulating freedom because of epistemic reasons does not imply that freedom (in general) really exists. It could just be that the sciences have not discovered yet the laws of nature that explain the phenomena in question so that we use out of pragmatic reasons the notion of freedom. The real contradiction between claim i) and ii) becomes apparent only if one takes Kant’s proofs for both of them into account. In these proofs Kant utilizes his notion of causality developed in the Transcendental Analytics. It asserts that whenever something happens we presuppose that what happened had always a cause and that a cause and its effect are always connected by a rule. This assertion does not suffice for excluding the possibility of “freedom from causality”. The impossibility of freedom from causality results from the additional claim that each single cause must be seen as an effect of some other cause or causes. Accepting this rules out the spontaneous coming into existence of an effect without any cause. If it where the case that a cause exists, which is not an event caused by another event,12 then its effect should be realised during the whole duration of the existence of this cause. Thus if for example stones were cases of such unconditional causes, then it could not be explained why the throwing of a stone by a person X caused a windowpane to break at 12:31 h on March 3rd, 1998 and not at a point of time before or after this particular date.13 Because if the stone were the unconditional cause of the breaking of the windowpane, then there must be a strict rule stating that stones always break windowpanes. However, if it is accepted that not the stone caused the breaking of the windowpane, but some event, in which the stone was involved then it is unproblematic that the stone is at a certain point of time involved in the breaking of the windowpane and not at every other point of time. This is so because the rule holds between two kinds of events and not between a substance (i.e. an unconditional cause) and an event.

12 Kant calls such an unconditional cause a “substance”. 13 Cf. the critical comments in Broad (1953: 156).

196

Katinka Schulte-Ostermann

However, Kant sees here another problem lurking: The circumstance that every cause has to be caused by another one results in a conflict with another aspect of his notion of causality, namely the condition that every cause must be sufficient for its effect. But if every cause is necessarily caused then it is impossible for a cause to be sufficient for its effect since every cause depends on an infinite chain of causes, which are necessary for its coming into being. Kant tries to circumnavigate this obstacle by introducing the notion of spontaneity. So the infinite chain comes to an end and the causes can be sufficient for their effects. Kant calls this spontaneity the Transcendental Freedom, or the capability of the Will to initiate an event solely through itself.14 Therefore Kant’s notion of “Causality from Freedom” cannot strictly be called agent causality, because the Will and not an agent or person is the cause of actions. Furthermore Kant’s notion of “Causality from Freedom” is confronted with the problem that it is impossible to deduce any effects from acts of spontaneity. This difficulty exists for every naïve theory of agent causality, which claims that nothing determines the future acts of will or of the intentions of the agent except the agent himself. Reasons like desires, wishes or beliefs cannot therefore be parts of causal explanations of actions. But even if the will is the spontaneous cause of actions the problem of how the agent himself is connected to that action persists. As long as there is no theories about how the person determines her will, Kant’s theory of “Causality from Freedom” cannot be regarded as an approach to agent causality. This is so because replacing the notion of Kantian “will” with the notion of “the Person who causes the action” does not solve the problem that the kind of spontaneity Kant has in mind is not suitable for being used in action explanation because there cannot be any meaningful relation established between the agent, his mental states and the action. Kant’s introduction of the notion of “Causality from Freedom” is one of the most opaque parts in his critical philosophy and a highly controversial issue. However, the aim of this short discussion was not to provide a contribution to the exegesis of the Kantian work, but it rather aimed at showing that Kant’s notion of “Causality from Freedom” cannot be used for the purpose of defending an agent causality approach. 2.2 The “Theory-Theory” of agent causation John Bishop chooses a completely different approach than Kant. In contrast to many other proponents of agent causality he is a compatibilist. This means that he is convinced that freedom and determinism are not mutually excluding notions, but that both can be true. This attitude be14 Cf. Kant (KdrV: B475/A447).

Agent Causation and Collective Agency

197

comes apparent in his discussion of the question of how event and agent causality are related. While he is in general defending a naturalist monistic ontology, he is also defending the irreducibility of the notion of agent causation.15 Fundamental for his account is the adoption of Sellars’ semantic functionalism. It consists mainly in the thesis that mental concepts are on the one hand irreducible to scientific concepts, but that, on the other hand, they should not be taken as concepts referring to a separate kind of entities. The function of mental concepts therefore can be only explanative in the context of a specific theory.16 The irreducibility of agent causation means therefore only that in the realm of a “Folk Action Theory” the function of agent causation cannot be reduced to other more fundamental concepts of the same theory: For each system treated as an agent, the theory postulates a certain ‚active nature‘, which consists in its repertoire of inferrings and basic actions and the ways this repertoire may vary over time. No further explication of what an agent is can be provided within the theory. [...] But this does not entail that questions about how agency is realized, physically or otherwise, are somehow illegitimate. They are just external to the theory […].17 Bishop defends thus only the irreducibility of agent causation in the realm of a “Folk Action Theory”, but this does not imply that elements of “folk theories” cannot be reduced to notions or concepts of other theories like neuroscientific or chemical ones. The difference between agent and event causation lies therefore for Bishop mainly in the fact that both concepts belong to different theoretical systems. Each of them, however, fulfils a similar function for its specific system in explaining how an event can occur. In “folk theory” agent causation explains the occurrence of events, which are actions, in scientific theory event causation explains the occurrence of events in general. This does not imply that event causalistic explanations play no role in “folk theory”, but that the notion of event causation in “folk theory” should not be identified carelessly with the notion of event causality in scientific theory. The task for the proponent of agent causality is to show that the notion of event causality cannot be used to analyse the notion of agent causality, because every attempt to explain the notion of action by the folk theoretical notion of event causality has at some point to introduce the notion of agent causation. Bishop claims instead that the notion of event causation cannot be found in “folk theory” 15 Bishop (1983: 66). 16 Cf. Sellars (1963: 1 – 40 and 127 - 196). 17 Bishop (1983: 74).

198

Katinka Schulte-Ostermann

but only in scientific theory, and it is this aspect, which marks the difference to the notion of agent causation.18 Therefore the principle of difference is violated if one accepts that the notion of event causality belongs as well as the notion of agent causation to the realm of “Folk Psychology”.19 The question therefore arises, whether the difference between agent and event causation does not lie in the nature of the relationship, but only in the nature of the relata. As mentioned above, the main motive for agent causation lies in the intuition that something can only be the action of a person, if the person itself is seen as the direct cause of the event in question. The process of causation cannot therefore be interpreted as a mere principle of regularity, but must be taken as real in an ontological sense. Otherwise it could be the case that agent and action stand in the relation of a mere contingent correlation or coincidence. The “theory-theory” theoretical approach is not open for this kind of realism. The notions of event and agent causality are theoretical notions, which have an epistemic function and they are only justified insofar as they fulfil this function. As theoretical notions they are constructs developed by persons and cannot therefore be interpreted as real. Even a notion of causality explaining the relation between cause and effect via a causal power could in the theoretical frame of the theory-theory not be taken as realistic because the notion of power itself is a postulate within the theory that fulfils the function to describe the relation between cause and effect. Whether there is in reality such a power, or not is of no interest for the functionality and therefore for the justification of the theory in question. The only criterion for the justification of theoretical concepts is their explanative and predictive power. Bishop is not undertaking any attempt to resolve the tension between the basic intuition that agents are the direct causes of their actions (which is one of the major reasons for the defence of agent causation) and his version of the “theory-theory” of agent causation. 20 There is, however, the possibility to take the notion of agent causality just as a theoretical postulate, which is necessary for the explanation of certain events without implying that there must be any real connection between the agent and his action. This would be in a sense circular and would not 18 Cf. Bishop (1983). 19 Actually there are remarks from Bishop, which at least give reason that he indirectly agrees with this point (Bishop 1983: 77). 20 In his book Natural Agency Bishop not only gives up the thesis of agent causation, but as well his adoption of theory-theory (Bishop 1989).

Agent Causation and Collective Agency

199

allow any epistemic approach to the phenomenon of action. The main aim of the “theory-theory” approach, namely to use the notion of agent causation for the purpose of explaining a certain kind of events cannot be therefore reached, because the notion of action is only via the notion of agent causality explainable, while the notion of agent causation is defined as the causal relation between persons and actions. From the above follows that Bishop cannot give any plausible argument for the difference between event and agent causality, despite his claim that in the case of agent causality a person and in event causality an event is the cause of the effect. But this is no more than repeating the thesis. His idea to use the “theory-theory” proves to be not useful because it does not give an informative explanation of the difference between agent and event causation and is of no help in defining the notion of action because of its inherent circularity. 2.3 The Emergence Theory of agent causation Timothy O’Connor does not, in contrast to John Bishop, try to interpret the concept of agent causality as a mere theoretical, but as an ontological notion. Despite this realistic claim he is like Bishop nevertheless defending a version of physicalist monism. Therefore on the one hand he defends the thesis that there are indeed two kinds of causality and claiming at the same instance that is just one kind of substance, from which everything else emerges. With the help of the notion of emergence he introduces agent causation, as a kind of “top-down” causality, while the causality of nature is always a “bottom-up” causality. In case of top-down causality it is not possible to reduce the causal power of a macro system to its microscopic elements. The macroscopic properties cannot therefore be reduced to the microscopic properties of a system with emergent causal power. In contrast in the case of the causality of nature it is always possible to explain the macroscopic causal properties through corresponding microscopic causal properties.21 This differentiation gives raise to the suspicion that emergence requires a kind of substance dualism. To avoid this suspicion it is necessary to give an account of the relation between the macro and the micro level properties. A first step of reconciling both levels is to state, that every macro level property is completely determined by its micro level properties, without being possible, however, to infer from the micro level properties which macro level properties are emerging. 21 Cf. O’Connor (2000: 108).

200

Katinka Schulte-Ostermann

In general it is necessary to distinguish between two kinds of emergence, diachronic and synchronic emergence. Diachronic emergence claims that it is possible that complex systems with emergent properties develop over the time from more basal systems. Synchronic emergence on the other hand, tells us that there exist complex systems with emergent properties that cannot be deduced from the properties of basal systems.22 O’Connor uses the notion of synchronic emergence for his account of agent causality. According to him agent causality is an emergent phenomenon connected to the neuronal states of the brain. However, O’Connor does not claim that the persons themselves are emergent. Emergent is only the capability of human beings respectively of human brains to initiate actions. He must defend this claim in order to maintain also the thesis that a person can be completely free to do what she wants: Nonetheless, there is no direct cause of my causing an intention to A, if that is what I do. No factor, internal or external, deterministic or indeterministic, brings about my acting.23 The capability to cause an action must therefore be a synchronic emergent property of the brain.24 At this point the following question arises: Is it justified to speak of agent causality if the capability to initiate actions is an emergent property of the neuronal system, which itself is completely determined by event causality? O’Connor gives no answer to this and the following really important questions: a) What is a person with respect to the neuronal system of her brain? b) How does a person make use of the emergent properties of her brain? There is only one very short passage where O’Connor tries to define the notion of a human person: [I defend] a picture, in which human persons are biological entities that have irreducible mental properties and capacities.25 However, he does not give any hint about what he means with the notion of “biological entity”. The term could be understood either as referring to something, which is of interest for biology – like cells, organs and living 22 Stephan (1999: 66-72). 23 O’Connor (2000: 97). 24 O’Connor (2000: 121-123). O’Connor himself does not speak directly of the brain but of “physical systems”, “basic biological functions” or “neural structures”. His reference to neural structures indicates, however, that he is mainly talking about the brain as the system from which the power to initiate something emerges. 25 O’Connor (2000: 73).

Agent Causation and Collective Agency

201

beings –, or as referring to the whole living being, that is the biological entity, which is capable of development and reproduction. Only in the latter case is it meaningful to say – using the O’Connor’s notion of person – that human beings as entities are the subjects of emergent properties. In contrary, is the notion used in the first meaning then it could be that every biological part of a human being displays the emergent properties in question.26 O’Connor announces that he has more to say on this issue, but unfortunately he fails to fulfil this announcement. The only thing he adds to the matter is that emergent properties are properties of the brain. It seems therefore that O’Connor reduces the notion of person to the notion of a biological entity, which has irreducible mental capabilities and properties. If this is right, it is impossible for him to give an account of agent causality that fulfils the function of introducing a new kind of causality, which is responsible for actions. This is so because according to his theory actions are not caused by a biological entity and its biological properties, but by mental properties, which are emergent to the biological properties of the biological entity. Therefore it is not the agent, but the emergent properties that are the causes of actions. Even if he speaks of emergent and therefore top-down causation is it not justified to speak of agent causation because the notion of agent has in this account no function. Additionally he has to accept that not only mental properties, but chemical biological or geological properties also can be emergent properties. Therefore any theory, which claims to give an account of agent causality by the notion of emergent properties, has to include why some of the possible emergent properties can be regarded as causes of actions, while others are excluded. Not every case of top-down causation is necessarily the cause of an action.27 O’Connor tries to overcome these difficulties by stating that mental properties are only the structuring causes of actions, i.e. that they determine which effects are possible, while the agent is the triggering cause.28 It is the

26 In this case one should understand the expression “human person” not as referring to a species. This is justified insofar as in the defining part of the sentence does not include the expression “human” or “human being”. 27 This problem will be solved if one is able to show that only mental properties are emergent properties, while all other seemingly emergent properties will show to be reducible in the long run. Cf. Broad (1925: 80); Levine (1983: 357); Stephan (1994) and Beckermann (2002). 28 The distinction between structuring and triggering causes has been introduced by Dretske (1988: 42-44). For the application of this distinction in action theory see Audi (1993) and Detel (2007).

202

Katinka Schulte-Ostermann

agent who determines the time at which an action occurs.29 With this differentiation, however, O’Connor just returns to the starting point of his investigation. This is because the reference to a structuring cause is not telling us anything about agent causality. Additionally, the difference between structuring and triggering cause is not identical with the difference between the causality of mental properties and that of agents. Both kinds of causes (structuring and triggering) can be found in contexts, which have nothing to do with agency: A soft ground, age and disease may, for example, be the structuring cause, why a tree falls, while a mighty windblast is the triggering cause of its fall. To say that mental causes are only structuring causes does therefore not allow the conclusion that the triggering cause is in any case an agent. There is no reason to believe that the emergent property of agent causality, as introduced by O’Connor, is the sole possible triggering cause for emergent mental properties, which have to be regarded as structuring causes. In order to defend the thesis that the agent is a triggering cause, O’Connor has to show firstly that the notion of the agent cannot be reduced to events or properties and secondly that it is possible at all that a substance in any state can be a cause and not only substances in certain particular states. Concerning the first task it has been argued that O’Connor reduces the notion of person to the notion of biological entity having mental capabilities and properties. Only the mental properties and capabilities are irreducible including their power to cause events. But these properties are only the emergent properties and capabilities of a biological system. Therefore it can be concluded that O’Connor’s attempt to give an account of agent causality by means of emergence theory does not work out. This diagnosis, however, does not imply that theories of emergence are in general not applicable in action theoretical discourses.

3. The first and the second gap The aim of our analysis so far has been to show that the three most discussed accounts of agent causation are unable to close the explanatory gap between the agent and her actions. None of them tells us something about the causal powers of the agent. A closer look at these theories reveals instead that at least in an ontological sense the agent cannot be seen as the cause of her actions.

29 O’Connor (2000: 99).

Agent Causation and Collective Agency

203

But my initial claim has been that the failure to close the gap between agent and action in case of the agent causation accounts has also consequences for the gap between individual actions and collective actions. In the following I will explore the impact of these consequences. In his contribution to this volume David Schweikard claims that a complete reduction of the notion of collective action to the notion used in explaining individual actions has the consequence that there is no real difference between individual and collective actions. What I want to do is to defend on the one hand, that there is actually no real qualitative interesting difference between these two cases, but that on the other hand this is not explained by reducing collective to individual actions. I will argue instead that what has been seen as the distinctive but uncontroversial trait of individual actions, namely that there is an unequivocal relation between an action and one singular mind of one individual, is not at all unproblematic and that therefore if this relation is taken as not granted, the absolute difference between individual and collective actions is undermined. Thus, as claimed at the beginning, bridging the gap between collective agent and collective action rests on bridging the gap between the individual agent and individual action. Before I proceed with my defence it is necessary to make the conditions explicit that are necessary for supporting the claim that an explanatory gap is bridged. Generally speaking, an explanatory gap is bridged or even closed if it is possible to deduce the explanandum from the explanans either by means of logical necessity or by means of natural laws. The explanatory gap between agent and action is thus bridged if it is possible to deduce from the agent – either logically or nomologically – her actions and the explanatory gap between a collective (consisting of at least two individuals) and its actions is closed if one can infer logically or nomologically the latter from the former. Now if it is the case that it is impossible to justify any statement about the relation between an individual agent and some events, which are regarded as her actions because there is no way to infer the action logically or nomologically from the agent itself, then it holds also for a collective that the behaviours of its members aren’t actions either. In order to sustain the view that there are nevertheless collective actions one can opt between three approaches: a) It is possible to regard the behaviour of a collective as an action, even if it is not possible to show that at least some of the behaviour of its individuals can be regarded as actions.

204

Katinka Schulte-Ostermann

b) It is not possible to regard the behaviour of a collective as an action as long as there is no way to prove that at least some behaviour of the individuals of the collectives can be regarded as actions. c) One can (in a Davidsonian manner) define the notion of action via the notion of their special cause, namely mental events. This strategy implies that the difference between individual and collective action cannot be directly found in the fact that individual actions are caused by just one individual and collective ones by a plurality of individuals because the Davidsonian notion does not imply that the causes of an action must be the mental states and events of just one individual human being. Within this framework if one wants to say that in the case of individual actions the mental states belong only to one and in the case of collective actions to different persons, one has not only to give an account of the notion of person, but to give as well a reason why this move is necessary without falling back to agent causalist approaches with their metaphysically mysterious notion of the agent as the “real” cause of actions. This last strategy does not have to bridge the gap between the individual person and her action, but the gap between mental events and bodily movements. If one wants to prove that some bodily movement of an individual counts as an action then one has to show that this movement was caused by a mental and not just by a physical or chemical event inside a human brain. That is, one is only justified to speak of a collective action if one is able to justify the claim that the bodily behaviour of the members of this collective are caused by mental events within this collective.30 And because bodily movements are always movements of the body of an individual human being, one cannot speak of a collective action of a given collective if one is not able to close the gap between mental events and the individual behaviour of the human beings belonging to this collective.31 Thus option c) is excluded as a candidate for supporting the claim that there are collective actions, leaving us with options a) and b). 30 Stoutland gave already very interesting and convincing arguments against the possibility of coming to a notion of collective agency via the notion of action of event-causation. It is his main thesis that because an action theory based on event causation can only be individualistic she necessarily fails to give an account of what it means that a collective as collective is acting. 31 I have argued elsewhere extensively against mental causation (Schulte-Ostermann, MS). Therefore it is sufficient for the purposes of this paper to mention that all prominent mental causation approaches (anomalous monism, supervenience theory, emergentism) are highly disputed and none of them can be salvaged by just minor reformulations. Cf. as well Walter (2006).

Agent Causation and Collective Agency

205

By opting for a) one is actually claiming that collective action is independent of individual action. A theory of collective action would therefore be independent of any theory of individual action. However, I think that a) is untenable and that there is a dependency between individual action and collective action, but not by means of reduction. If one opts for b) instead, then it follows that one either has to give a better account of agent causation than the ones discussed so far, or to admit, that it is impossible to justify the claim that some behaviour of a collective can justifiably be regarded as its actions. This is because there is no way to show that some behaviour of an individual was caused by an agent and not by some event. However, option b) does not commit one to any causal theory about actions. Therefore an approach to action, which justifies the claim that certain events are actions because the acting subjects themselves consider them as actions, seems very promising. Von Wright introduces a third kind of explanation in contrast to purely logical and to purely causal explanations, namely what he calls, teleological explanation of events. This kind of explanation describes these events as purposeful from the viewpoint of the acting person without implying that this person or some mental event caused the action. The question is therefore whether the explanation of an event by the purposes of the collective is in some interesting sense different form the explanation of an event by the purposes of a single person. What I wanted to show is that there is no interesting difference, at least not in regard to the nature of the agent. But this does not imply that collective action is reducible to the actions of its members. It only means that there is no special problem about collective actions, which is not encountered in the analysis of individual actions and vice versa. Thus the explanatory gap between the individual actions and the collective actions is confronted with the problem that somehow the notion of agent (collective or individual) does not have an object of reference, which can be taken as being the cause of action. It is not known, whether the behaviour of some beings is just mystically correlated to their mental states, or if it is really an action of these beings because they are the causes of it. Many theoreticians of collective agency, however, seem to presuppose that there are individual actions and that the discussion mainly concerns how the emergence of collective action has to be accounted for (via some kind of collective entity or just by the intervening mental states of individual agents). As mentioned above there are two possibilities to cope with this problem:

206

Katinka Schulte-Ostermann

i. By finding a more convincing account of agent causation ii. By developing an action theory, which does not depend on the concept of the agent as cause. I pledge for the second option. Pettit and others already have introduced an account of collective subjects, which depends on the coherence of intervening beliefs and intentions without presupposing that these beliefs and intentions need to be the intentions and beliefs of one individual subject. In contrast he argues that the subject is constituted by a coherent network of beliefs and intentions, which are in some kind shared by different individuals without implying that each individual is aware of each intention and belief concerning the collective action, but that only the main features of the collective action in question have to be mutually known. Two considerations speak in favour of a non-causal theory of individual and collective action in addition to the approach on collective and individual subjects: i) There is the still unsolved problem of mental causation, and therefore the epistemic problem, of how we are entitled to claim that something is an action, if we do have a causal theory, but no possibility to determine that some mental event really caused this action, and is not merely correlated to it. ii) The everyday experience that we seldom have problems to decide whether someone did something or whether something was only happening on him. Therefore it seems that we have quite a clear notion of individual and collective action, which does not depend on any causal theory of the relation between body and mind. Even if the debate on action theory started once from an analysis of ordinary language, it seems that the current state of affairs can be best described as a Lakatossian research programme that got stuck in a “degenerative problem shift”. In such cases launching a completely new research programme may help finding a way out of the impasse. Thus we should start again with the original question: How do we know that something is an action and not a mere happening?

References Allison (1998): Allison, H. The Antinomy of Pure Reason, Section 9. (A515/B543-A567/B595), in: Mohr, G. and M. Willaschek (eds.),

Agent Causation and Collective Agency



207

Immanuel Kant, Kritik der reinen Vernunft, Berlin: Akademie-Verlag, 1998, pp. 465-490. Audi (1993): Audi, R. Mental Causation: Sustaining and Dynamic, in: Heil, J. and A. Mele (eds.), Mental Causation, Oxford: Clarendon Press, 1993, pp. 53-74. Beckermann (2002): Beckermann, A. Die reduktive Erklärbarkeit des phänomenalen Bewusstseins, C.D. Broad zur Erklärungslücke. in: Pauen, M. and A. Stephan (eds.), Phänomenales Bewusstsein, Paderborn: Mentis, 2002, pp. 122-147. Bishop (1983): Bishop, J. Agent Causation, in: Mind (1983) 92/365: 61-79. Bishop (1989): Bishop, J. Natural Agency, an Essay on the Causal Theory of Action, Cambridge: Cambridge University Press, 1989. Boogerd et al. (2005): Boogerd, F.C., Bruggeman, F.J., Richardson, R.C., Stephan, A., and Westerhoff, H.V. Emergence and its Place in Nature: A Case Study of Biochemical Networks, in: Synthese (2005) 145: 131-164. Broad (1925): Broad, C.D., The Mind and Its Place in Nature, London: Harcourt, 1925. Broad (1953): Broad, C.D. Determinism, Indeterminism, and Libertarianism, in: Berofsky, B. (ed.), Free Will and Determinism, New York: Harper Colins, 1966, 135-159. Chisholm (1966): Chisholm, R. Freedom and Action, in: Lehrer, K. (ed.), Freedom and Determinism, New York: Random House, 1966, pp. 11-44. Chisholm (1976): Chisholm, R. The Agent as Cause, in: Brand, M. and D. Walton (eds.), Action Theory, Berlin: Springer, 1976, 199-211. Dennett (1991): Dennett, D. Consciousness Explained, Boston a.o.: Little, Brown and Company, 1991. Detel (2007): Detel, W. Mental Causation and the Notion of Action, in: Psarros, N. and K. Schulte-Ostermann (eds.), Facets of Sociality, Frankfurt: otos verlag, 2007, pp. 51-83. Giblert (2000): Gilbert, M. What it is for us to intend?, in: Gilbert, M. Sociality and Responsibility, Lanham: Rowman & Littlefield, 2000, pp. 14-36. Kant (KdrV): Kant, I. Kritik der reinen Vernunft, Hamburg: Meiner, 1998.

208

Katinka Schulte-Ostermann

Kreimendahl (1998): Kreimendahl, L. Die Antinomie der reinen Vernunft 1. and 2. Abschnitt (A405/B432 – A461/B489), in: Mohr, G. and M. Willaschek (eds.), Immanuel Kant, Kritik der reinen Vernunft, Berlin: Akademie-Verlag, 1998, pp. 431-446. O’Connor (2000): O’Connor, T. Persons and Causes, Oxford: Oxford University Press, 2000. O’Connor (2002): O’Connor, T. Libertarian Views: Dualist and AgentCausal Theories, in: Kane, R. (ed.), The Oxford Handbook of Free Will, Oxford: Oxford University Press, 2002, pp. 337-355. Schulte-Ostermann (MS): Schulte-Ostermann, K. Das Problem der Handlungsverursachung, Ph.D. Thesis, forthcoming. Sellars (1963): Sellars, W. Science, Perception and Reality, London: Routledge & Kegan Paul, 1963. Smith (1923): Smith, N.K. A Commentary to Kant’s ‘Critique of Pure Reason’, New York: Humanities Press, 1923. Stephan (1994): Stephan, A. Theorien der Emergenz – Metaphysik oder?, in: Grazer Philosophische Studien, (199) 48: 105-115. Stephan (1999): Stephan, A. Emergenz, Von der Unvorhersagbarkeit zur Selbstorganisation, Dresden: Dresden Univ. Press, 1999. Tuomela (2001): Tuomela, R. Collective Acceptance and Social Reality, in: Lagerspetz, E., H. Ikäheimo, and J. Kotkavirta (eds.), On the Nature of Social and Institutional Reality, Jyväskylä: Univ. of Jyväskylä, 2001. pp. 102-135. Walter (2006): Walter, S. Mentale Verursachung, eine Einführung, Paderborn: Mentis, 2006. Weber and Deacon (2000): Weber, B. and Deacon, T. Thermodynamic Cycles, Developmental Systems, and Emergence, in: Cybernetics and Human Knowing (2000) 7: 21-43. Wilkerson (1976): Wilkerson, T.E. Kant’s Critique of Pure Reason, A Commentary for Students, Oxford: Clarendon Press, 1976.

Part III: Sharedness and Normativity

Reliance and Intending the Joint Activity1 Facundo Martin Alonso, Stanford Philosophers have maintained that collective action is not simply a matter of a collection of agents acting individually. People can be acting in a coordinated way – e.g., Weber’s case of people simultaneously opening their umbrellas when it starts to rain – but still not be acting collectively in a proper sense. In addition, philosophers have claimed that every bodily movement, which is part of a collective action, is a movement of a member of the group. From these two claims they have concluded that what distinguishes collective action from other kinds of aggregated phenomena cannot lie in a feature of our bodily movements. On the contrary, they have suggested, the key property of collective action lies in its internal or mental component. It lies in the intentionality of the participating agents, that is, in their “collective” or “shared” intentionality. But how, we must ask, should we understand shared intention? According to the individualist approach, shared intention should be understood in terms of the intentions of the individual participants and the complex inter-relations that hold between them.2 The elaboration of an appropriate conception of the intentions of individuals in shared intention is thus one of the central concerns of this approach. With this concern in mind, Michael Bratman has proposed that we understand the intentions constitutive of shared intention as intentions of individuals whose content is the joint activity itself. In particular, Bratman’s proposal is to “... exploit the fact that we can speak not only of intentions to, but also of intentions that [...] Accordingly, we can speak of my intention that we J,” where “J” is the joint activity in question.3 Alternatively, 1 Previous versions of this paper were presented at the Fifth International Conference on Collective Intentionality in Helsinki and at the 2007 Pacific Meeting of the American Philosophical Association. I would like to thank the discussants for valuable comments and feedback. I owe special thanks to Michael Bratman for his helpful comments and suggestions. 2 The holistic approach, on the contrary, maintains that shared intention and action cannot be properly understood in terms of the intentions of individuals, their inter-relations and so on. Margaret Gilbert has developed an elaborate holistic account of shared intention and collective phenomena in general. See, for instance, Gilbert (1992; 1996; 2000a). 3 Bratman (1999a: 115. Emphasis in the original). N. Psarros, H. B. Schmid and K. Schulte-Ostermann, Concepts of Sharedness, Frankfurt 2008: 211-223; ontos verlag

212

Facundo Martin Alonso

Raimo Tuomela and Kaarlo Miller have claimed that in order for a group of agents to perform a joint intentional action X, they must each have a “we-intention” to do X: “… they must share a common intended goal, normally (but not necessarily) the goal to perform that total action … X.”4 Thus, a we-intention is in Tuomela and Miller’s view an intention of the individual whose content is “the full social action.”5 These two views of the intentions of individuals in shared intention differ in important respects. However, they share an important idea: namely, they both conceive of the intentions constitutive of shared intention as intentions about, or in favor of, the joint activity.6 The idea that I may intend our activity has been challenged in different ways.7 My aim in this paper, however, is not to dispute this conception of the intentions of individuals in shared intention. Rather, my aim is to contribute to its understanding. In particular, I want to explore the cognitive conditions of an individual’s intention in favor of the joint activity when he shares an intention with others to so act. More precisely, when you and I share an intention to act together, my intention that we so act depends upon certain cognitive attitudes I have about your attitude towards the joint activity and about your eventual actions. The same is true, of course, 4 Tuomela and Miller (1988: 370). 5 Tuomela and Miller (1988: 375). That the notion of we-intention has as its content, not the agent’s part of the joint activity but rather, the joint activity itself is purported to reflect, in Tuomela and Miller’s view, the fact that a participant in a joint activity will accept the locution “We will do X,” rather than that of “I will do my part of X,” as adequately describing his attitude (1988: 375-376). Tuomela has elaborated on the notion of we-intention in subsequent works. See, for instance, Tuomela (1995; 2005). For reasons of simplicity of presentation, however, in this paper I will for the most part focus on his earlier work. 6 Of course, Tuomela and Miller, as well as Bratman, think that more is required than these basic intentions of individuals in order for there to be a shared intention. Roughly, Tuomela and Miller require that there be mutual belief among the participants about the existence of each of their we-intentions. As Tuomela has recently put it, shared intention “… consists of the participants’ we-intentions about the existence of which the participants have mutual belief” (2005: 330). Alternatively, Bratman thinks that shared intention involves, in addition to these basic intentions in favor of the joint activity, intentions in favor of “meshing subplans”, “interlockingness” between the relevant intentions, and common knowledge of all this (1999a). I provide a critical discussion of these views in Chapters Four and Five of my Ph.D. Dissertation (MS). 7 It has been objected, for instance, that there is no clear sense in which I may intend our activity. Different formulations of this objection have been advanced by Annette Baier (1997), Frederick Stoutland (1997; 2002) and David Velleman (1997).

Reliance and Intending the Joint Activity

213

of your intention that we act. My aim in this paper is then to determine what precisely these cognitive attitudes about the other participants are. In other words, I want to explore what cognitive attitudes the individual is basically required to have in order for him to intend the joint activity in the way characteristic of shared intention.8 According to Tuomela and Miller, an individual’s we-intention to perform a joint activity X entails an intention on his part to perform his part or share of X. In addition, Tuomela and Miller think that in order for an individual to we-intend to do X, and therefore, to intend to do his part of X, he must believe, among other things, that the other participants will (or at least will try to) perform their parts of X.9 In their own words, “[u]nless a person, A, believes that another person, B, will do (or at least try to do) his share of X, A will not do his part, nor can he even unconditionally intend

8 There is an important clarification to make here. My concern in this paper is with the cognitive requirements of only those cases of intending the joint activity which are candidates for shared intention. These cases should be distinguished from other cases in which I intend the joint activity but in a way that rules out our sharing an intention in the proper sense. A clear example of intending the joint activity in the latter way is provided by Bratman’s “Mafia” case (1999a: 117-118). In this case, I intend that we go to New York together by way of kidnapping you and throwing you in the trunk of my car. But my intention that we go to New York together here involves an intention on my part to coerce you in a way that bypasses your intentional agency, and this is, as Bratman rightly suggests, irreconcilable with our sharing an intention in the proper sense. Notice, however, that this case of intending that we go to New York together may be framed by appropriate cognitive attitudes: I may have thoughts about how to kidnap you, about how to get there, etc. Still, my cognitive attitudes will probably include a reference neither to your attitude towards “our” going to New York together nor to how that attitude of yours interacts with mine, since I may not even care about that. Finally, it should also be noticed that intending the joint activity in the way involved in shared intention is not incompatible with all cases of coercion. It is compatible with certain cases of coercion, which fall short of the Mafia case – i.e., with certain cases of coercion in which the coercion in question does not bypass the intentional agency of the other agent. Here I have in mind certain cases involving threats, asymmetries in bargaining power, and the like. Henceforth, I will use the expression “to intend the joint activity” and its cognates to refer to the kind of intention in favor of the joint activity that is a candidate for shared intention. 9 Among the things that an individual must believe in order to we-intend to do X, Tuomela and Miller also mention the belief that there be (or will be) a mutual belief among the participants that they will do their parts of the joint activity (1988: 375).

214

Facundo Martin Alonso

to do it.”10 The previous cognitive requirement constitutes in Tuomela and Miller’s view a “precondition” on we-intentions.11 Alternatively, Bratman suggests that when you and I share an intention to act together, my intention that we act depends on my knowledge or belief that you intend that we act or that you will come so to intend; and vice versa.12 Thus, for instance, Bratman writes: “In the case of our painting the house together … each of our intentions that we paint depends on our knowledge of the other person’s intention that we paint: [my intention that we paint] depends on my knowledge that [you intend that we paint]; and [your intention that we paint] depends on your knowledge that [I intend that we paint].”13 Therefore, both Bratman and Tuomela and Miller think that when a group of individuals share an intention to act together, each individual’s intention in favor of the joint activity depends on certain beliefs (or knowledge) the individual has with respect to the other participants’ relevant attitudes or actions.14 Thus, these authors agree on the kind of cognitive attitude about the other participants’ attitudes or actions which is required for intending the joint activity: namely, belief. However, these authors disagree on the content of those beliefs. While Tuomela and Miller regard the other participants’ actions – i.e., their parts in the joint activity – as the main contents of those beliefs, Bratman focuses on the other participants’ relevant intentions as their fundamental targets. 10 Tuomela and Miller (1988: 374). Similarly, Tuomela and Miller write: “It is … an essential presupposition of a we-intention that at least a sufficient number of agents, as required for the performance of the joint action in question, will participate” (1988: 374). 11 It should be mentioned that Tuomela and Miller conceive of the belief that there be (or will be) a mutual belief among the participants to the effect that they will do their parts of the joint activity as another precondition on we-intending. Tuomela has recently claimed that this belief, together with the belief that the participants will perform their parts of the joint activity, express “the minimal rationality of the we-intender” (2005: 341). 12 Bratman introduces this suggestion in Bratman (1999b: 153-157). 13 Bratman (1999b: 154). Bratman’s original formulation of his notion of shared intention involved the idea that the intentions of individuals constitutive of shared intention depend on the individuals’ knowledge of each other’s intentions (1999a: 121). Later on, Bratman relaxed this condition to claim that these intentions may depend on the individuals’ beliefs, rather than on their knowledge, about each other’s intentions (1999b: 143, n 4). 14 In order to make matters simple, I will take Bratman’s notion of my intention that we act (or similarly, of my intention in favor of the joint activity) as canonical for this way of capturing the intentions of individuals in shared intention.

Reliance and Intending the Joint Activity

215

I think that both Bratman’s view and Tuomela and Miller’s view are right in directing our attention to the contents that each of them emphasizes.15 In contrast to these views, however, I think that both the appeal to the other participants’ relevant intentions and the appeal to the other participants’ relevant actions are central to intending the joint activity.16 In my view, both targets should be included as contents of the cognitive attitudes which are preconditions of intending the joint activity. There is another point on which I disagree with these views, a point which is central to the conception of the cognitive requirements of intending the joint activity I will be defending in this paper. I think that even though Bratman as well as Tuomela and Miller are right in singling out proper contents of the cognitive attitudes which are preconditions of intending the joint activity, they are wrong in assuming that the kind of cognitive attitude which embodies those contents needs to be belief. Consider the following example. Suppose that the high school I used to attend is organizing a fund-raising tango party for next Saturday. I very 15 Tuomela and Miller are right in emphasizing the dependence of an individual’s intention in favor of the joint activity on his cognitive attitudes with respect to the other participants’ actions. Suppose that I am considering whether to engage in a joint activity with you, an activity which by its very nature requires both my and your participation. Were I to decide and therefore form the intention that we act in these circumstances, I would be required to have certain cognitive attitudes towards your playing your part. Were I to think that you would not do your part of our joint activity, it would not be possible for me to intend that we act. For I cannot intend that we perform an activity which I think we will not be able to perform. Likewise, I think that Bratman is on the right track when he requires that an individual’s cognitive attitudes in shared intention include in their contents the other participants’ intentions in favor of the joint activity. Following David Velleman’s (1997) suggestion, we may regard intentions as the kind of attitudes that “settle” the deliberative question of whether to act or not . Accordingly, Bratman claims, we may say that by intending the joint activity I settle the deliberative question of whether we will act, provided that certain conditions are present. These conditions include, among other things, that the other participants also intend the joint activity. Therefore, if I think that the others do not or will not intend that we act, I cannot reasonably intend that we act. For I cannot intend an activity I do not take myself to settle. For an elaborate explanation, see Bratman (1999b). 16 It might be objected that this is not in general true of all cases of shared intention, especially, of those cases of shared intention which involve a large number of individuals. I grant that cognition of the other participants’ intentions and actions may not be necessary for large-scale cases of shared intention. However, note that the object of study here are small-scale cases of shared intention – i.e., cases in which an intention is shared by a small group of individuals, typically (but not exclusively) by two individuals.

216

Facundo Martin Alonso

much want to go to this party, and I am especially interested in going with Laura, an old acquaintance of mine and an excellent tango dancer. So I call her in advance and invite her to this party. Excited at the prospects of dancing the tango with an old acquaintance, Laura accepts the invitation. She forms the intention that we go to the tango party together next Saturday and makes this known to me. Since I believe (or know) that Laura has the relevant intention, it may seem that I can now proceed, in accordance with Bratman’s view, to form my own intention that we so act. However, this is insufficient. Laura has a reputation of being quite an unstable planner: she often commits herself to certain activities but later on changes her mind. Therefore, it seems that if I am to intend that we go to this party together, I should have certain cognitive attitudes not only about Laura’s relevant intention at a given time but also about her (future) relevant actions. In Tuomela and Miller’s view, I should believe that Laura will perform her part of our joint activity when the time comes. That is, I should believe that she will follow through with her intention. Yet, on careful reflection – that is, given what I know about Laura’s unstable or “flaky” behavior – I do not believe that she will follow through with her intention. I lack evidence backing up the expectation that she will perform her part. But even though I do not believe that Laura will perform her part, I may still rely on her to perform her part. Furthermore, such reliance may well be rational, even in the absence of belief. I will say more about this below. Now suppose that the acquaintance of mine with whom I am especially interested in going to the tango party is not Laura, but Maria. So I call Maria in advance and invite her to go to this party. Maria says that she will think about it and that she will let me know. I am not surprised by her response. Maria is known to be quite an indecisive person. She usually has trouble making up her mind. However, once she settles on a particular course of action, she follows through with it no matter what. Unlike Laura’s, Maria’s problem is not one of psychological instability or “flakiness.” Her problem is, rather, her indecisiveness. One of the questions we may want to raise with respect to this case is whether I have sufficient evidence about Maria’s attitudes to form the intention that we go to the tango party together. If we follow Bratman in this regard, we should respond in the negative, since in this case I lack the belief that Maria intends (or will come to intend) that we go. However, I think that Bratman is mistaken here. Granted, in this case I lack the evidence that would support the belief that Maria has (or will have) the

Reliance and Intending the Joint Activity

217

relevant intention. But in order for me to intend that we go to the tango party together, I do not need to believe that Maria intends (or will intend) likewise. It suffices that I rely on her to so intend, and that seems possible even in the absence of belief. Furthermore, once I rely on her to form the relevant intention, I will be in a position, given what I know about her character, to rely on her to follow through with her intention. So in the previous two cases I lack a belief either about the other person’s intention or about the other person’s action. Still, it seems that in both cases I can have an appropriate cognitive attitude about my partner’s intention and actions to intend that we act together. This suggests that the cognitive attitude about the other participants’ relevant intentions and actions central for intending the joint activity is not belief, but reliance. The question that immediately follows here is: what is reliance?17 As some philosophers have suggested, we may initially conceive of reliance as a three-place relation: namely, individual A relies on individual B to do action X.18 Thus, we may say, for example, that Peter relies on Mary to water his plants.19 It is worth noticing that an individual may rely on another not only to perform a particular action but also to have a particular attitude – for instance, to have a certain intention, belief or desire.20 It should also be noticed that the relation of reliance does not always hold between two individuals. Quite often we rely on some group, organization 17 In what follows I provide only a rough characterization of the phenomenon of reliance. For a detailed study of this phenomenon, see Chapter Six of my Ph.D. Dissertation (Alonso MS). 18 The idea of regarding reliance as, basically, a three-place relation has been advanced, in different ways, by Annette Baier (1986: 236), Simon Blackburn (1998: 30) and Richard Holton (1994: 67). It should be mentioned, however, that the main object of study in these investigations is the phenomenon of trust rather than that of reliance. Still, these authors think that both reliance and trust can be understood, initially, as a tripartite relationship. This is because trust involves, in these authors’ view, the attitude of reliance plus a moral feeling of some kind (on which they disagree). I will not consider the differences between reliance and trust in this paper. 19 Of course, the relation of reliance may be extended to a four-place relation to include some context or circumstance in which it takes place – i.e., A relies on B to do action X in circumstances C. Continuing with our example in the text, we may say that Peter relies on Mary to water his plants while he is away on vacation. The relation of reliance can naturally be extended in other ways. 20 Along these lines, Annette Baier has suggested that we can rely on another person to be motivated in a particular way (1986: 234). Similarly, Simon Blackburn has claimed that an individual can rely on another to be in a certain state – for instance, to be concentrated while reading a book, to be worried about a job interview, etc. (1998: 30).

218

Facundo Martin Alonso

or institution. And we even sometimes rely on inanimate objects, as when we rely on a rope to hold our weight when climbing down a rock.21 We can get a deeper understanding of the phenomenon of reliance if we focus on the central role that this attitude plays in the individual’s practical reasoning and action. In this regard, one of the characteristic features of reliance is that relying on an individual B to do X involves a disposition on one’s part to incorporate the assumption that B will do X into one’s practical reasoning and action. Thus, when I rely on Laura to show up at the tango party I incorporate the relevant assumption in my practical reasoning and action. For instance, I plan to buy two tickets, I deliberate about what time to pick her up, I look for directions about how to get to her place, etc. Still, the fact that reliance on an individual to do X involves a disposition to incorporate the assumption that he will do X into one’s planning and action does not distinguish, as Richard Holton has suggested, reliance from belief.22 For believing that an individual will do X also entails a disposition to incorporate the assumption that he will do X into one’s practical reasoning and action. However, these attitudes differ in an important respect: while belief is based or grounded on “convincing evidence,” reliance is frequently a matter of “pragmatic necessity.”23 When I rely on a rope to hold my weight while climbing down a rock I incorporate the relevant assumption into my planning and action.24 For instance, I plan to make certain moves that depend on the assumption that the rope will hold my weight. However, my reliance on the rope need not be grounded on convincing evidence that the rope will hold my weight.25 The reason 21 Similarly, it may be suggested, on some occasions we speak not just of an individual, but of a group, an organization or an institution as relying on somebody (or something) to do X. We may then say, for instance, that the government relies on Peter to pay his taxes. I find this suggestion to be initially plausible. It remains to be explained, however, what it is for a social group, rather than for an individual, to rely on somebody (or something) to perform a particular action. 22 Holton (1994: 68). 23 Holton regards the present difference between belief and reliance as one about how these states are causally originated rather than as one about how these states are justified or grounded. Thus, Holton suggests that while belief is caused by “convincing evidence,” reliance is sometimes caused by “pragmatic necessity” (1994: 68). 24 This example is widely discussed in the literature. See, among others, Blackburn (1998) and Holton (1994). 25 Sometimes our reliance on a person or object to do X is based on some evidence justifying our expectation that he or it will do X. In some cases, as Blackburn has suggested, we rely on somebody to do something because that is what he has

Reliance and Intending the Joint Activity

219

I rely on the rope may just be that I have no other alternative. In the circumstances, this may be the only way I have of climbing down the rock.26 As Holton maintains, “I have reason for simple reliance on an object if I need something done and reliance on it is my best bet for getting it done; likewise for simple reliance on a person.”27 For related reasons, reliance and belief stand in different relations to the will. It is widely held that since belief aims at truth and is shaped by appropriate evidence, it cannot be under the control of the will. For if I could acquire a belief at will, I could acquire it independently of its being true or false, irrespective of the reality it is purported to represent. But then I could not think of this attitude of mine as a belief.28 Things are different with reliance, however. Reliance can be, in contrast to belief, under the control of the will. In fact, we sometimes decide to rely on people to do certain things. That is, we sometimes decide to deliberate, plan or act based on the assumption that some people will do one thing or another. Of course, the idea that reliance can be under the control of the will does not mean that we can rely on anybody to do anything we may want. As Holton has suggested, we can decide to rely on certain occasions, but within certain limits.29 The following rational requirement on reliance illustrates the limits in question. While in relying on you to do X I need not believe that you will do it, it is a rational requirement that if I rely on you to do X I lack

26

27 28 29

always done (1998: 32). Thus, for instance, we rely on a responsible student to turn in his homework by the appropriate deadlines, on the traffic in the United States to drive on the right, etc. However, and this is the point that I am trying to make in the text, there are some other cases in which we rely on somebody or something to do X without having convincing evidence grounding the expectation that he or it will do it. Notice that my view is not that reliance can be justified by pragmatic considerations, which are independent of the evidence. Rather, I conceive of the pragmatic considerations that justify reliance as appropriately constrained by the evidence. In particular, pragmatic considerations for reliance fail to justify reliance in the presence of convincing evidence to the contrary. For instance, if I have convincing evidence that the rope will not hold my weight, the consideration that relying on the rope is the only means for climbing down the rock will not justify my reliance on it. Later on in the paper I express this idea in a slightly different way by suggesting that it is a rational requirement of reliance that it be consistent with belief. Holton (1994: 68). Bernard Williams has forcefully argued for this point. See, for instance, Williams (1973: 148-149). Holton (1994: 69).

220

Facundo Martin Alonso

the belief that you will fail to do it. In other words, what I rely on, if I am rational, has to be consistent with my beliefs. Holton has accounted for the plausibility of this requirement in terms of the important roles that reliance and belief play in our planning about the future. When I rely on you to do X I incorporate the assumption that you will do X into my plans. However, if I believe that you will not do X, I should also incorporate that assumption into my plans. But then I am led to an inconsistency, for I have to plan both on the assumption that you will do X and on the assumption that you will not do X. 30 Despite these differences with belief, reliance is similar to belief in the following respect. Reliance has a mind-to-world “direction of fit.”31 Thus, if I rely on you to do X, and you fail to do it, I will be mistaken in what I relied on. Reliance thus seems to have two faces. On the one hand, it is affected by pragmatic considerations –that is, the reasons for reliance are, in some cases, pragmatic. On the other hand, it presents features proper of a cognitive attitude.32 Let’s now come back to our previous example. Even though I have doubts about whether Laura will actually end up attending the tango party on Saturday, doubts which are grounded on her reputation as quite an unstable planner, I rely on her to be there. My reliance on her is based on pragmatic reasons. These reasons include the fact that her participation is necessary in order for our joint activity of tango dancing to succeed and that the success of our joint activity is something I very much want. Moreover, notice that my reliance on her to do her part in our joint activity may well be rational. After all, there are strong pragmatic reasons for relying on her and this reliance is consistent with what I believe. A similar argument applies, I suggest, to my relying on Maria to form her intention that we go to the tango party together. Given these considerations, we can conclude that reliance on the other participants’ relevant intentions and actions satisfies the cognitive requirements for intending the joint activity. Thus, Bratman and Tuomela and Miller are right about the contents of the cognitive attitudes which are preconditions of intending the joint activity. But they are wrong in

30 Holton (1994: 71-72). 31 The notion of direction of fit derives from Anscombe (1963). 32 There are interesting parallels between reliance and what some philosophers have called “acceptance” or “acceptance in a context.” On the notion of acceptance see among others Kaplan (1981), Cohen (1991). The notion of acceptance in a context is Bratman’s (1999c).

Reliance and Intending the Joint Activity

221

thinking that those cognitive attitudes are necessarily beliefs. Reliance, I have argued, can do the cognitive work in intending the joint activity.33 This may shed light on the vexed question of whether shared intention generates interpersonal obligations. In persisting in one’s intention in favor of the joint activity, one reinforces the other participants’ reliance on one’s performing one’s part, and this normally grounds relevant obligations to them. Since in shared intention each participant relies on the others in the aforementioned respects, and since these forms of mutual reliance are mutually reinforced, there will normally be interpersonal obligations. Whether this provides a general solution to the question of the relation between shared intention and interpersonal obligation is a matter for future research.34

References Alonso (MS): Alonso, F. Shared Intention, Reliance, and Interpersonal Obligations, Ph. D. Dissertation, Stanford University, forthc. Anscombe (1963): Anscombe, E. Intention, Ithaca: Cornell University Press, 1963. Baier (1986): Baier, A. Trust and Antitrust, in: Ethics (1986) 96: 231-260. Baier (1997): Baier, A. Doing Things with Others: The Mental Commons, in: Alanen, L., S. Heinämma, and T. Wallgren (eds.), Commonality and Particularity in Ethics, London: MacMillan, 1997, pp. 15-44. Blackburn (1998): Blackburn, S. Trust, Cooperation and Human Psychology, in: Braithwaite, V. and M. Levi (eds.), Trust and Governance, New York: Russell Sage Foundation, 1998, pp. 28-45. Bratman (1999a): Bratman, M. Shared Intention, in: Bratman, M. Faces of Intention. Selected Essays on Intention and Agency, Cambridge: Cambridge University Press, 1999, pp. 109-129. Bratman (1999b): Bratman, M. I intend that we J, in: Bratman, M. Faces of Intention. Selected Essays on Intention and Agency, Cambridge: Cambridge University Press, 1999, pp. 142-161.

33 Notice that this does not yet entail that reliance can do all the cognitive work in shared intention. 34 Margaret Gilbert and Michael Bratman have held an interesting dispute on the relation between shared intention and interpersonal obligations. See, for instance, Bratman (1999d) and Gilbert (2000a; 2000b).

222

Facundo Martin Alonso

Bratman (1999c): Bratman, M. Practical Reasoning and Acceptance in a Context, in: Bratman, M. Faces of Intention. Selected Essays on Intention and Agency, Cambridge: Cambridge University Press, 1999, pp. 15-34. Bratman (1999d): Bratman, M. Shared Intention and Mutual Obligation, in: Bratman, M. Faces of Intention. Selected Essays on Intention and Agency, Cambridge: Cambridge University Press, 1999, pp. 130-141. Cohen (1989): Cohen, J. Belief and Acceptance, in: Mind (1989) 98: 367-389. Gilbert (1992): Gilbert, M. On Social Facts. Princeton: Princeton University Press, 1992. Gilbert (1996): Gilbert, M. Living Together: Rationality, Sociality and Obligation. Lanham: Rowman & Littlefield Publishers, 1996. Gilbert (2000a): Gilbert, M. What is it for us to intend?, in: Gilbert, M. Sociality and Responsibility. New Essays in Plural Subject Theory, Lanham: Rowman & Littlefield Publishers, 2000, pp. 14-36. Gilbert (2000b): Gilbert, M. Obligation and Joint Commitment, in: Gilbert, M. Sociality and Responsibility. New Essays in Plural Subject Theory, Lanham: Rowman & Littlefield Publishers, 2000, pp. 50-70. Holton (1994): Holton, R. Deciding to Trust, Coming to Believe, in: Australasian Journal of Philosophy (1994) 72: 63-76. Kaplan (1981): Kaplan, M. Rational Acceptance, Philosophical Studies (1981) 40: 129-145. Stoutland (1997): Stoutland, F. Why are Philosophers of Action so AntiSocial?, in: Alanen, L., S. Heinämma, and T. Wallgren (eds.), Commonality and Particularity in Ethics, London: MacMillan, 1997, pp. 45-74. Stoutland (2002): Stoutland, F. Critical Notice of Michael Bratman’s Faces of Intention. Selected Essays on Intention and Agency, in: Philosophy and Phenomenological Research (2002) 65: 238-241. Tuomela (1995): Tuomela, R. The Importance of Us. A Philosophical Study of Basic Social Notions, Stanford: Stanford University Press, 1995. Tuomela 2005: Tuomela, R. We-Intentions Revisited, in: Philosophical Studies (2005) 125: 327-369. Tuomela and Miller (1988): Tuomela, R. and K. Miller. We-intentions, in: Philosophical Studies (1988) 63: 367-389.

Reliance and Intending the Joint Activity

223

Velleman (1997): Velleman, J. D. How to Share an Intention, in: Philosophy and Phenomenological Research (1997) 57: 29-50. Williams (1973): Williams, B. Deciding to Believe, in: Williams, B. Problems of the Self, Cambridge: Cambridge University Press, 1973, pp. 136-151.

The Logic of External Reasons and Collective Intentionality Jennifer Hudin, Berkeley

1. Moral sense Social and moral behavior are tightly interconnected notions and an analysis of one inevitably calls into consideration an analysis of the other. With regard to these notions, there are several fundamental questions: 1) What is their ontology? That is, what sort of facts are social and moral facts? What provides the truth conditions of their statements? and 2) What is the nature of a practical reason given social or moral concerns? Though these questions are interdependent – a clear answer to one would help in answering the others – they can be treated separately, and indeed have been. Considering moral behavior to be a kind of social behavior, which is enabled by cognitive capacities for social organization, we can make some progress understanding the ontology of both notions, social and moral, if we consider them from a socio-biological perspective. From such a perspective, social and moral behavior are interesting phenomena because while social behavior is not unique to humans – other species of animals exhibit coordinated movement capacities – moral behavior is only observed among linguistic creatures such as ourselves. This is particularly notable because other social creatures who are said to have some kind of language in the form of minimal linguistic representation, bees, for example, are not ascribed a system of morality.1 It is an interesting observation that socially unacceptable behavior for nonhuman species is not described in terms of breaking codes. A code describes a social norm, which has been codified in some way. Codification entails the possibility of transgression and transgression in turn involves another sophisticated notion, that of penalty. All three notions, codification, transgression and penalty are concepts that require language, implicitly if not explicitly. Not only are social codes and penalties linguistic, i.e., they are constructed by language and require language in order to be understood, their ontology and conceptualization also appears to require an additional cognitive capacity over and beyond that of knowing how to

1 Cf. von Frisch (1973). N. Psarros, H. B. Schmid and K. Schulte-Ostermann, Concepts of Sharedness, Frankfurt 2008: 225-251; ontos verlag

226

Jennifer Hudin

behave socially and to speak a language. Social codes require the capacity of moral sense. Moral sense is not just a social skill that arises from knowing how to coordinate behavior with other conspecifics, nor is it a primitive sense of survival, which allows appropriate social behavior. Rather, moral sense is a separate capacity that allows for the perception of states of affairs as socially binding, i.e., obligatory, and how socially binding states of affairs are equitably distributed among a given set of members of a social group. As stated, moral sense is the perception of social obligation and its measure – social equitability. The claim that moral sense is a logically separate cognitive capacity that is neither linguistic nor merely an outgrowth of social behavior can be argued from two directions: 1) It is possible to imagine, and indeed we can find, a linguistic social group which has socially errant behavior but lacks the notion of codification, transgression and penalty. And, 2) in the opposite direction, it is possible to imagine, and indeed we can find, a nonlinguistic social group which has punitive responses to social errancy but lacks the notions of codification, transgression and penalty. In both cases we can say that the group in question lacks the notion of social obligation and social equitability. Let’s consider each case in turn. As stated earlier, bees seem to have some sort of minimal language in which they can indicate indexical relations between the hive and the sources of honey. 2 Bee social behavior consists in gathering nectar and bringing it back to the hive. The method by which they gather nectar is by deciphering the dance of given bees, which tells them where to get the nectar. The dance indicates the direction and distance the nectar is from the hive. Socially errant behavior in bee groups could be defined by some behavior, which does not merely fail to get nectar – accidents can always happen – but rather, behavior in which the dancing bees simply disregard their duty. Now notice that social errancy cannot be described without a minimal sense of “ought” or duty. In order to build a story in which 2 In conversation, both Beatrice Kobow and John Searle have presented compelling arguments as to why the bee dancing may not be linguistic. Furthermore, Searle maintains that it is not possible to have any sort of linguistic representation without commitment (Searle 2007). With respect to the point I am making in the present argument, it is unimportant whether bee dancing is actually linguistic. The example is intended as a thought experiment only. We could easily imagine a human tribe that has language but lacks the notion of obligation. In this case, the human tribe would lack commissives (i.e., promising), but be able to perform all the other types of speech acts. The point is that language does not guarantee a sense of social obligation.

The Logic of External Reasons and Collective Intentionality

227

codification can play a role in bee life, we have to describe their work behavior as duty and duty implies obligation. But it is not the case that the working bees perceive their behavior as obligatory, nor as a duty. And the evidence for this is that in fact, failure of nectar gathering behavior does occur among bees in the sense of duty neglect in cases where they do not understand or “believe” the information that the dancing bee has provided. In such a case, they simply disregard her directions. Failure to follow the dancing bee’s directions is a beelike form of socially errant behavior but as far as we can tell, there is no penalty or punishment involved for disregarding the directions of the dancing bee. Now this is a case in which many of the operative features necessary for moral behavior are present: the bees do have a form of linguistic behavior and they do have socially errant behavior because the working bees neglect to do their job. But even given this amount of social structure, there is no codification in the bee society because the bees lack notions that express moral sense. Specifically, at the social level there is no notion of penalty. At the individual level, there is no sense of obligation or social equitability. We can safely infer that bees do not have a moral sense. On the other hand, wolves are a nonlinguistic3 social species with social rankings that include alpha males. Alpha males mark territory thereby “claiming” it as belonging to them. At this point, we already have several markers of a structured society – the notions of hierarchy rankings and territory are social notions. When another wolf wanders onto an alpha wolf’s territory, the alpha wolf inflicts punishment by attack. Now we have many features in place that look like they would allow at least wolf society to satisfy the criteria for having a notion of codification and therefore a cognitive capacity for moral sense: social ranking, territory, transgression, penalty. Don’t these features all add up to social coding by means of behavior alone? The wolf territory marking is not a case of social coding. Rather, nonlinguistic social groups such as wolves have social norms, i.e., conditions of appropriate behavior that oscillate depending on who is in charge, and the penalties for violating these norms are in brute, physical terms, generally violent. The wolf that wanders onto the alpha’s territory has not violated a law – i.e., a social code – but rather a social norm established by one male who is alpha at that moment.4 This social ranking is subject 3 I am using the term linguistic to refer to systems of representation that have syntactical structure. Wolves may have methods of representation, but there is no evidence to date that they have syntax. 4 The question arises as to whether social norms are indeed social practices, and whether non-human species can actually have social practices. According to

228

Jennifer Hudin

to change by any challenger and at any time. Furthermore, it is a metaphor to describe the physical attacks prompted by errant social behavior as penalties imposed. Rather, they are at best a method of discipline used to indicate who is temporarily in charge and to maintain the hierarchy status of the moment. Therefore at the social level, the brute physical attacks are not penalties but disciplinary measures.5 And with the lack of penalties, we can say that it is also not the case that the wolf’s wanderings count as transgressing a social code that marks a boundary. The wolf no more conceives of his wandering in terms of transgressing a social code than the alpha wolf conceives of his attacking him in terms of imposing a penalty because a law was violated. The physical attack prompted by territory transgression does not satisfy the criteria for social coding. Wolf social behavior lacks all the concepts required for social coding and we can safely conclude that they do not have the notions of obligation and social equitability. Wolves do not have a moral sense. As a marked contrast to both of these examples, let’s consider an analogous human case. A man walking in a park happens to wander off the path onto an area of freshly planted grass, which has a sign that says: No Trespassing. The man has indeed violated the park code, and he is now susceptible to a penalty. This penalty most certainly would not involve brute physical violence, but at most a mild fine or an official admonition. The man feels an immediate obligation to hop off the grass. The obligation arises from understanding the social code, his accidental transgression and the potential penalty. If we want to explain the cause of the man’s particular behavior in the park, the reasons we appeal to are not brute physical facts, but his comprehension of linguistic facts – the sign that Haugeland (1982), animals cannot have social practices because they do not have the capacity for conformism. Conformism, Haugeland maintains, is the strictly human capacity to censor one’s behavior in order to fit in with one’s social group. Although Haugeland argues this pro-human view of conformism in a Heideggerian spirit, this is also a Kantian argument as to why animals must lack practical reason. I am in the company of ethologists such as Frans de Waal (2006), which maintains that animals can deliberately censor their behavior. Deliberate censorship of behavior in the absence of immediate punishment or reward is evident among primates as well as canines, felines, and other non-human species. On my account, it is logically possible to have non-human social practices and social norms. I am grateful to the editor for this question. 5 If there is any doubt that the wolf disciplinary action is the same as a penalty, consider the eventual scenario in which a transgressor wins the fight with the alpha wolf and thereby deposes him. In cases of penalty impositions, there is no such scenario in which the transgressor challenges the penalty (a possibility) and thereby replaces the authority figure(s) who impose penalties.

The Logic of External Reasons and Collective Intentionality

229

says “No trespassing.” But this is only a partial explanation. A complete explanation needs to include what motivated the man to hop off the grass, and this requires making reference to the man’s capacity to understand what codification entails in this and every case: a code, transgression and penalty. But again, mere understanding of the park code is insufficient to motivate the man to hop off the grass. It is perfectly possible to understand utterances such as imperatives and ignore their directive. Something over and beyond understanding meaning is required to motivate the man to move and that something is a moral sense of social obligation, a sense of “ought” or “should”. In this human case, unlike the bee and wolf case, we can truly say that the man has a sense of “ought”, i.e., a moral sense, not merely because of his capacity to understand language, which he does, but along with his capacity to understand language and behave socially. There are three points in particular that these examples illustrate. Some of these points are to be expected and are not at all surprising given the preceding discussion. For example, the first point is that social behavior is more basic than moral behavior and moral behavior supervenes, so to speak, on the capacity for social behavior. This is not an unexpected result because moral behavior, which is grounded by a moral sense, requires the capacity for social perception. Moral sense as obligation and social equitability are relational notions, which always require more than one person. Second, and more interesting, behavior alone is capable of defining social normativity for all social groups, but social normativity that is codified is neither solely a case of behavior nor linguistic representation, nor merely a combination of behavior and language. Social codification represents our obligations and though it requires language, can only evolve in virtue of another capacity that allows for knowledge of being obligated. This kind of knowledge does not come for free with linguistic representation, but as I am claiming, is a biological capacity that works together with but functions separately from the capacity for language. And this leads to the third point and that is linguistic representation, having language capacity in and of itself, is not sufficient to derive a moral system. Again, it is possible that other species have some sort of linguistic representation yet lack a moral system. Let us consider the case of the bees again. Bees apparently represent indexicals such as distance and direction of sources of honey, but they cannot represent reasons for getting the honey or going to honey sources. So, what in addition to representation is required to make the move from merely representing what is the case, to representing what ought to be the case?

230

Jennifer Hudin

A good place to start an investigation is by examining the facts at hand and in this case, it happens to be our own language and its logical form. Moral notions such as “ought” and “should” can be formally described by means of deontic operators such as Obligation [O] and Permissibility [P]. These operators operate over statements of facts about what is the case and turn them into facts about what ought to be the case. That there are these particular formal operators in human language and that there is even a need to have such operators points to the fact that part of human cognition is the capacity to think about general and particular states of affairs in terms of obligations and permissibility. This is not only a sophisticated linguistic operation, it is a highly sophisticated cognitive capacity. Language tells us that we in fact do perceive what is the case in terms of what ought to be the case, but it does not tell us how we perceive this way. While moral sense of obligation and equitability is indeed reflected by human language, indeed as I have argued, codification cannot exist without language, the origin of human language itself is a mystery. A perfectly sound route of investigation of moral sense and moral notions is to try to give an account the origin of human language and then try to use the same method by which we discovered the origins of human language for the origin of moral notions and moral sense. Now, there are two obvious ways to do this, either by an evolutionary-anthropological account, or by a logical-ontological, nativist, Chomsky-style innateness account. I think thus far we have indirectly argued that no account of the logical-ontological structure of language will give us moral sense. Even if we could discover that syntax was a part of our genetic neural makeup, we could not infer that similarly there is a genetic structure for moral sense. As I have tried to argue, moral sense is not linguistic. On the other hand, the anthropological-evolutionary account of language might prove more fruitful as a method for discovering how moral notions and moral sense evolve. If an evolutionary-anthropological account of language could be successful – an account which naturalizes language – perhaps we could similarly give an evolutionary-anthropological account of moral sense and also naturalize morality. Again, the most obvious way to begin such an account is to look at the facts at hand – the capacity to behave collectively and the capacity to speak a language. In this case, an evolutionary account of collective behavior coupled with an evolutionary account of language might yield an account of moral behavior.6 Such evolutionary accounts of moral behavior are not new at all. Generally, the picture painted of how social groups evolve moral systems is in 6 Tomasello (2000: ch.2)

The Logic of External Reasons and Collective Intentionality

231

terms of social efficiency, however efficiency is understood. But such an explanation will always fall short of adequacy because collective behavior and linguistic representation are logically insufficient for moral behavior, as we have argued: it is logically the case that you can have one without the other, language without moral sense or moral notions. While such an evolutionary account of moral systems has to have some truth – cultural systems do evolve and change in relation to the needs of particular social groups – there still remains the question of how a sense could evolve that perceives states of affairs as obligatory and equitable. The problem for any evolutionary account is that moral sense simply is not necessary for social behavior. And as we have seen, moral notions such as obligation and permissibility are not necessary logical features of nonlinguistic representation. There are several philosophical questions that an evolutionary-anthropological account misses when it provides accounts of moral behavior based solely on social behavior. For example, why is it that most social groups down the phylogenetic line do not have moral behavior when apparently they do have systems of representation, and some of it linguistic, plus the capacity for socially coordinated behavior? Why should only human societies have moral systems? On the evolutionary account, in principle, any social group with linguistic capacities and coordinated behavior could have developed moral systems. But it didn’t happen that way, and such an account should be able to tell us why. Another problem left unanswered by an evolutionary account is the following: codification appears to be the social feature that turns social systems into moral systems, and by moral systems, I mean those in which the members of social group not only understand obligations and permissibility as pertaining to themselves, but as they also pertain to others. And as I argued initially, codification requires language, but language capacity itself does not guarantee understanding obligation. How could an evolutionary account of a notion such as codification explain how obligation arises? Providing an evolutionary account of codification – a social notion at a very high level – requires having the notion of obligation – a cognitive notion – already in place. Thus, an evolutionary account must first explain how social groups evolve a sense of obligation before it can begin to make any sense of the notion of codification. But now we are back to the problem of explaining obligation and its related notion, codification, solely in virtue of collective behavior and linguistic representation. A third feature in addition to collective behavior and language is logically required to have moral behavior and moral notions. This third logical feature is a cognitive capacity for moral perception, i.e., moral sense.

232

Jennifer Hudin

The next question in our investigation then is what sort of cognitive capacity is moral sense? It is a weak and uncontentious claim that humans and other social animals have biological capacities for knowing how to behave in their social group, i.e., knowing the norms of their social group. But the question about innateness of moral sense is more interesting because it requires not merely knowledge of how to behave socially, but also knowing how to obligate oneself, and understanding what it is for something to be an obligatory act not just for one individual, but an obligatory act for any individual of the social set – i.e., understanding obligation as equitably distributed. And if such notions as obligation and permissibility are already too complex and too socially developed to be explained as something as simple as pre-theoretical moral senses, then what sort of pre-theoretical, pre-institutional senses would a species such as humans need in order to create the more complex notions such as obligation, permissibility, equitability? This is a much more difficult question than how social species coordinate their behavior through cooperation, or how normative behave arises. This is the foundational question for the possibility of creating moral systems. Analyzing the ontology and the enabling conditions for moral facts in terms of a pre-theoretical and pre-linguistic moral sense is the first step in understanding how a species can have moral notions that evolve into full blown moral systems. But in doing this, it is always important to keep one’s eye on the higher-level feature that we want to describe, in this case, how moral behavior is produced. Thus, we need to consider moral reasons, not merely reasons for holding ethical beliefs, but also reasons that motivate ethical behavior, i.e., reasons that direct actions taken in one’s life, from the perspective of rational behavior and moral behavior. An analysis of moral reason inevitably reduces to an analysis of human action and practical reason. The same questions about the nature of moral reasons can be asked about the nature of practical reason itself: Are moral reasons for action always motivated by personal desires and beliefs of a given agent? Or is it possible that individual action be motivated by reasons that have nothing to do with individual intentionality? That is, is it possible that some behavior, specifically moral behavior, be motivated from simply recognizing reasons that have nothing to do with satisfying personal desires? Clearly this is the traditional debate between the Humean and Kantian explanatory models of behavior and moral agency. In light of the biological considerations that enable humans to have deontic notions such as obligation, we can now ask what role these cognitive enabling conditions have in moral reasoning and motivation of behavior. Simply put, can moral behavior be produced by, and a product of, some

The Logic of External Reasons and Collective Intentionality

233

pre-theoretical innate moral sense that is not only common to all humans, but that functions independently of personal satisfaction? This is a major question indeed. Finally, tightly connected with the considerations of practical reason and ethical behavior are notions of rationality and morality. Specifically, the questions for an analysis of practical reason is whether a rational agent requires commitments to morality and whether it is possible to be rational and not be moral. Returning to the question of whether a species such as humans have pre-theoretical, biologically innate moral senses powerful enough to override self-interest, it is then important to consider how such moral sense operates. Does it operate solely within a system of rationality or can it operate independently of rational consideration? And then, can such nonreflective behavior really be considered moral? Kant most famously answered this question negatively.7 The discussion thus far about the cognitive primacy of moral sense has been developed with the ultimate goal of explaining a large portion of everyday social behavior that can be called moral, i.e., our behavior that is performed under the reason of obligation. In the first part of this paper, I have argued that moral sense does not rely solely on linguistic behavior nor adaptive, normative behavior, but is rather a type of perception that shapes linguistic meaning and adaptive normative behavior in its own particular way. It is a cognitive capacity on all fours with the capacities to act socially. With this claim in hand, in the second part of this paper, I will consider a model of explanation of moral facts and practical reason that fits between the Humean model of agency as consistently and always a matter of self-interest, and the Kantian model which not only allows for the possibility of moral action as disinterested agency, but demands a rational agent to be moral. In this third model, I will suggest a type of moral behavior that is produced by disinterested moral agency which functions independently of rationality: a moral agent need no more deliberate on his momentary moral behavior than he need deliberate on his momentary visual experiences. Moral behavior is a kind of perceptual experience – that of perceiving states of affairs as binding and socially equitable. A moral agent perceives what is as what ought to be and therefore acts without requiring any extra cognitive operation such as deliberation. Similarly, if the moral agent is not required to deliberate on moral acts, he need not invoke or consult any moral rules. Rather, a moral agent on this account is one who has the capacity for moral sense, a sense which impels moral action without any reflection whatsoever. 7 Kant (GMS).

234

Jennifer Hudin

Finally, with respect to the relation of moral sense to rationality, it follows from what I have said thus far that if moral sense does not require deliberation, it can operate separately from rationality yet in tandem with rationality also. As I will discuss shortly, moral sense that impels action without deliberation is the norm. In addition, rationality operates in tandem with non-deliberative moral actions. An individual endowed with such a moral sense and functioning rational capacities can always question obligations and equitability. In fact, the moral sense that allows an individual to understand obligation and equitability allows him to judge them also.

2. Evidence for an innate moral capacity As discussed above, several facts about human behavior and human moral reasoning need to figure into an account of moral cognition. Let’s review these features and introduce a few new ones: 2.1 Linguistic representation Linguistic representation is necessary but not sufficient for moral facts and moral practical reasoning. This point has already been discussed in detail. The fact that linguistic representation in itself and of itself is insufficient for moral facts leads to the second and third features of moral facts, namely the requirement of deonticity and understanding “ought” as a we-reason. 2.2 The requirement of deonticity The deontic operators such as Obligation and Possibility formally convert descriptions of states of affairs in the world into descriptions of deontic states of affairs in the world. Similarly, when sincerely uttered, they create deontic states of affairs out of states of affairs that are merely facts, i.e., they turn “is” into “ought”. 2.3 Understanding “ought” as a we-reason Human language contains not only formal deontic operators, but also the implicit illocutionary force of “ought” depending on the type of speech act: Thus the statement: “You owe federal taxes,” can have the illocutionary force of an assertion, as in “It is a fact that you owe federal taxes;” the force of a declaration, “We the government hereby state that you owe federal taxes,” and the illocutionary force of a directive: “You owe federal taxes (!).” Implicit in all three speech-acts is a reason for action from a third-person point of view: “This is a reason for you to act on.” But as Williams has noted, it is difficult to see how third-person reasons can be capable of motivating an agent to act until they are indexicalized to the

The Logic of External Reasons and Collective Intentionality

235

agent in question by the agent himself, that is, until he makes them into a first-person reason, a reason “for me”.8 This observation of Williams’ is analogous to the observation we made about the man walking in the park discussed earlier. As we noted then, it is not sufficient for the park walker to understand what the sign “No Trespassing,” means for him to move his body voluntarily. As Williams would agree, some factor in addition to understanding what the sign means is required for the park walker to adopt the reason to move his body, the “ought not to walk on the grass” as his reason. But contrary to Williams’ conclusion that the reason must appeal to something in his motivational set – roughly construed as the set of his desires – and thereby become indexicalized to him, our conclusion was that the man must have the particular capacity for moral sense and that moral sense is a social capacity which does not reference the individual, but rather the social group in question. Moral sense in the sense we have been discussing requires that the man recognize the reason stated by the sign that prohibits trespassing on the grass,9 is an obligation that is equitably distributed. Not walking on the particular patch of grass is not just a reason for him; it is a reason for all members of the set to which he momentarily belongs – the set of park walkers. Thus, the park walker recognizes the truth of the statement as applying to him and the entire set of park walkers, and the ought stated by the sign in the park is not recognized as a first-person reason indexicalized to the I, but rather as a first-person reason indexicalized to the We. More importantly, this recognition of the We-reason does not occur because of rational consideration, but only in virtue of a capacity that I have been calling moral sense. Contrary to Williams’ observation, first-person indexicalization of reasons cannot explain many reasons that we normally act on, reasons which flood our daily lives.10 Apart from the observation made by Searle that in a lot of cases, we do not always prefer the reasons we act on, 11 there is 8 Williams (1981: 102). 9 This is a strange sign that is posted in English-speaking countries. By definition, “trespassing” means law violation. It isn’t possible to have a sign that says “Trespassing permitted here.” Prohibiting trespassing is redundant, unless the explicit prohibition includes the notion of a penalty or a fine. 10 Williams would not disagree on this point. His sub-Humean argument could not account for several examples, e.g., needs and suicide. 11 In cases in which we do not strongly prefer or really even want to pursue a course of action, Searle claims that we form a desire that is secondary to the recognition of a reason for action. “In some very broad sense of “want” and “desire,” every intentional action is an expression or manifestation of a want or desire to perform

236

Jennifer Hudin

also the fact that the recognition of the truth of some reasons, specifically institutional reasons in particular, is not simply a function of understanding them as true of the agent in question. As we stated above, it requires understanding them as true of the agent as a member of a particular set. To repeat, reasons for action provided by institutional facts such as owing taxes are not merely I-reasons, true for me alone, but We-reasons, true for us as a group. As another example of this, let’s take the case above of owing taxes. An agent who owes taxes understands the speech acts above because he has a complex background understanding of himself as a member of a class of citizens, membership of which is in virtue of paying dues, in this case, taxes. The proposition, that X owes taxes is true of him as a member of this set, because at some time previous to the utterance of the speech act, X intentionally and with full responsibility made himself a member of this set.12 But just as important as what X did, is the fact that he did not or could not create this fact alone or in isolation – quite unlike the act of eating chocolate or drinking water which can be done alone. Rather, he accomplished this deed in the virtual presence of a community who not only made the creation of the fact possible, but also share the ownership of the fact with him. 13 To repeat, such facts as owing taxes and all other types of facts that serve as institutional reasons, if true of an agent, are necessarily also true of the members of his community. Ontologically then, we can see that We-reasons are different in kind from other kinds of practical reasons. They differ in two ways: 1) Truth conditionally: their truth conditions are true of a that action. I will claim that this is not so in the case of external reasons” (Searle 2001: 168). 12 Searle describes such states of affairs as “desire-independent reasons.” (Searle 2001: Ch.6). 13 As a counterexample, Asa Andersson suggested the possibility of a tax that was only imposed on one person. Such a case is possible for example, in high tax brackets in which only one very rich person meets the criteria for a particular tax rate. In such a case (unlikely as it is), the reason for that particular individual to pay taxes must be an I-reason since the reason for action is applicable to no other person. The set membership contains only one person. This story is not a counterexample since the rich taxpayer is not actually the sole taxpayer in the state. In the scenario stated above, the set of taxpayers is stratified according to level of income. The highest level of taxes is a subset of the larger set of taxes of which all taxpayers are members. Therefore, the single rich taxpayer is a member of the rest of the tax paying community. In the case of a state in which no citizen is taxed unless he has an income over a certain amount, this is not a situation of taxation but imposition of penalty for exceeding a set income level, since the general economy does not rely on taxation. I doubt such a situation could exist in a free market economy.

The Logic of External Reasons and Collective Intentionality

237

set of individuals rather than single individuals and 2) Mode of existence: We-reasons require collective recognition along with collective commitment (to be discussed shortly) to the state of affairs in question in order to exist. The mode of existence of We-reasons is really quite different from that of I-reasons. I-reasons, reasons that are true for me particularly, but not necessarily for any other person, do not require collective recognition to exist. In fact, it is possible for them not to require any observer at all as in the case of needs, e.g., “My body needs calcium, and therefore I should drink more milk.” No observer whatsoever is required for my body’s need for calcium serving as the reason I should drink more milk. And it follows that the reason still exists whether I never acknowledge it at all.14 Still, as Williams persuasively argues, a reason must motivate an individual to move his very own body, and he could argue that even granting the ontological distinction I made above with respect to We-reasons, an individual, not a set, moves his body. This is a core argument against any non Humean account of practical reason and we have to meet it head on. Simply put: how do We-reasons motivate? And just as importantly, if they motivate without meeting any desire, primary or secondary in the individual motivational set, how can they possibly be rational? One of the foundational arguments of rationality and practical reasons is the type distinction between beliefs and desires, with motivation falling in the set of desires. On a traditional Humean account of practical reason, mere recognition of a reason alone regardless of whether it is indexicalized to the first-person singular or first-person plural, is powerless to motivate action. Against this view, I will argue that We-reasons as practical reasons function differently from other types of practical reasons because they do not require rational deliberation in order to motivate therefore dispensing with any need for satisfaction of members in the motivational set, or any to appeal to desire (passion) in any form. And in answer to the rationality criterion, I will argue that We-reasons do not need to satisfy personal inclinations to meet demands of rationality. In fact, I will argue that the agent who needs to deliberate over a We-reason that genuinely exists for him in order to be motivated could very well be considered a candidate for irrationality. 14 This observation about needs leads to a further question about whether my ontological criteria for We-reasons is a subcategory of external reason, or whether it is the definition of the category of external reason itself. In the latter case, needs on this account would not count as an external reason. But for reasons beyond the scope of this paper, I am not committed to this point of view and am happy at this point to settle for the ontology of We-reasons as a subcategory of external reason. For a discussion about needs as external reasons see Searle (2001).

238

Jennifer Hudin

2.4 The requirement of collective intentionality and social commitment Let’s stipulate that intuitions are something closer to emotions because they do have a gut feel to them and are sometimes not fully expressible. Furthermore, let’s stipulate notions to be something closer to concepts and that they can be expressed propositionally. As we have already noted, human language has the capacity to represent states of affairs in moral terms, i.e., it can represent what is the case in terms of what ought to be the case. And as we have been arguing, the bridge between mere representation and moral representation is moral sense, a kind of perception of states of affairs, as they ought to be. The question for moral sense then is, where on the spectrum between intuition and notion, emotion and reason, does it lie? Is it closer to an intuition, or is it a notion, i.e., a concept that is gradually acquired? It would seem that moral sense, as we have been arguing, is something closer to a pre-theoretical, pre-intentional intuition that provides the enabling conditions for notions that are linguistically encoded and gradually acquired. The claim that I am making is that humans have a pre-intentional sense of deontic notions, which we call obligation and permissibility, but the pre-intentional notions themselves cannot be explained in terms of obligation and permissibility. States of affairs that are obligatory and permissible already presume a high level of social and institutional structure. So, the pre-intentional notions that enable these higher-level notions must consist in more primitive cognitive capacities. Our task is to try to determine what sort of cognitive capacities social creatures would require in order to allow them to create such notions as obligations. For moral sense, I suggest two cognitive capacities that work in tandem. These capacities are 1) the capacity for collective intentionality, which allows the perceiver to understand intentions of other minds,15 and 2) the capacity for social commitment. This commitment is defined as a kind of emotional bond that includes but is not limited to a particular individual and can extend to an entire social group. Social commitment is expressed by imitating the social behavior of a group to which there exists a commitment. Imitation of group behavior is the means by which social facts are perpetuated. These four points form the foundation of an argument as to how Wereasons, henceforth, external reasons, exist and how they can motivate 15 Tomasello claims that this capacity for collective intentionality in which the contents of other minds can be understood is exclusive to the human species. I am not necessarily committed to this view; Tomasello (2000: 53-54).

The Logic of External Reasons and Collective Intentionality

239

behavior without needing to invoke satisfaction of desires. That humans have a pre-linguistic capacity for deonticity, i.e., a moral sense, which is enabled by pre-intentional biological capacities for collective intentionality and commitment to social facts that the group collectively holds is not sufficient to argue for external reasons. On this account, wolves could possibly have external reasons for action if we discovered that they really could in some way represent to themselves what others in the group are thinking along with their capacity to imitate behavior. But the decisive break between humans and all other social animals is that humans not only have the pre-intentional capacities of collective intentionality and social commitment, they have a linguistic means to represent deontic notions. This is the beginning of an explanation as to how humans have moral facts and moral reasons for behavior. There still remains the problem of motivation. This leads to a fifth feature. 2.5 Moral intuitions Humans can and typically do make moral judgments that cannot be adjudicated by reasons. Such judgments are made quite effortlessly, quickly and non-reflectively. This phenomenon is called moral dumbfounding. Similarly, humans display the opposite phenomenon in which they have moral reasons for judgments but they do not have any of the associated moral intuitions in the emotional sense such as, for example, that of outrage, indignation, and fairness, all broadly construed. This phenomenon is called moral numbness. These phenomena have lead neuroscientists to suggest a “dual-process” theory of decision-making.16 The dual-process theory of decision-making claims that humans make judgments by means of two different functional areas in the frontal lobe, the ventral medial prefrontal cortex (VMPFC) and the dorsal lateral prefrontal cortex (DLPFC). There are two distinctive features that seem to separate these two functional areas, emotion and reason. It appears that decisions based on emotions associated with past experience are a function of the VMPFC while decisions requiring solely the logic of rationality are a function of the DLPFC. Time does not seem to be the decisive factor for either function. For example, snap decisions that are emotion based are produced by the VMPFC while decisions that require immediate logical action to new situations are produced by the DLPFC. On the other hand, reflective decisions can be a function of either the VMPFC or the DLPFC, again it is all a matter of the mode in which the decision is made: Decisions that are based on learning through past experience, specifically learning that resulted in reward or punishment, are produced by the VMPFC, while 16 Chaiken and Trope (1999), Sloman (1996)

240

Jennifer Hudin

reflective decisions based on the logic of the data alone (e.g., those of mathematics and scientific data) are produced by the DLPFC. The feature that determines which part of the prefrontal cortex to be used is whether the decision is made in virtue of emotion or in virtue of reason, i.e., logic. In light of this, moral dumbfounding, a decision-making process that is not rational, can be said to be an expression of the VMPFC while moral numbness, a decision-making process, which lacks emotion entirely but is rational, is an expression of the DLPFC.17 Moral dumbfounding is an interesting phenomenon in which humans make choices, many of them with serious consequences such as jury decisions, where choices made have no rational grounds. Rather, when asked why such choices are made, individuals refer to feelings about fairness, for example, rather than to the logic of their reasons. Such decisions are made as a matter of intuition and these types of decisions are far more common than we would like to think. Just as interesting, these non-rational intuitions appear to have a general and universal character – they can be quantified over, predicted in degrees of probability for a social group given the nature of framing for the problem to be solved. 18 In light of this fact, it is not a stretch to claim that decisions made by the VMPFC, both as a matter of dumbfounding and as a matter of rational decision making based on past experience of reward and punishment, are instances of Hume’s internal reasons. Such decisions always appeal to something in the personal motivational set. On the other hand, the polar phenomenon, moral numbness and the decisions made solely by the DLPFC could be considered the embodiment of Kantian morality in its ideal form. Ideally, in non-pathological cases, moral numbness deliberates, reflects and operates entirely on rational processes in order to make moral choices.19 Unfortunately, moral numbness can also be a product of brain pathology or socialized insensitivity caused by various factors, none of which are considered desirable.20 Human practical reason seems to operate best when it has the opportunity of employing both intuition and reason, i.e., when the VMPFC and DLPFC are functioning adequately. But now the point at hand for our purposes is the fact that the two phenomena, intuition (emotion) and reason (logic), 17 Kahneman and Sunstein (2005) 18 Kahneman and Sunstein (2005). 19 Science Fiction has idealized moral numbness in highly rational characters, which make the best decisions because they are not burdened by human emotions, which often negatively affect situations. Dr. Spock and Data in the Star Trek series are two such characters. 20 Phineas Gage is a case in point.

The Logic of External Reasons and Collective Intentionality

241

can operate isolated one from the other. We can establish that they are logically distinct systems. Now, with respect to our discussion about the ontology of moral sense, there are two conclusions we can draw from the dual-process theory: (Option 1) Moral intuitions are innate because they are fundamentally emotional responses that all humans share in the form of primal fear responses. That is, in light of the neurological data above, we can say that moral sense appears to be primarily a function of the VMPFC, or (Option 2) moral concepts, acquired from a social group, are the primary operators in moral decision making. Moral intuitions are a secondary phenomenon that gradually develop from the capacity to form moral concepts. Again, in light of the neurological data above, we could say that humans have the capacity for acquiring concepts and employing them in rational consideration – a function of the DLPFC. Social experience coupled with moral rationality leads to a social-moral sense, a function of the VMPFC, which is shaped by a given social group with a particular moral background. In light of the dual-process theory, and current neuroscientific data, these really are the only two options for explaining moral sense. In terms of an evolutionary account of moral behavior, I think the second option (Option 2) is probably the best choice, but I think it is not entirely correct. Option 2, which places all the responsibility for morality and rational consideration on the DLPFC can only be correct if moral behavior arises solely in virtue of behavior and language. But as we have argued, logically, moral sense is not entailed either by behavior, language nor their combination. But Option 1 is not a plausible choice either. Moral sense is not solely a function of the VMPFC, it is not entirely a matter of intuition based on the past experiences of reward and punishment of a single individual. A lot rides on this point because if there were neurological data to show that moral decisions are primarily a function of the VMPFC, then Hume’s model would certainly be vindicated: moral behavior is always a matter of acting on an internal reason, satisfying some member in the motivational set of an individual. Moral sense would be nothing more than an innate fear-based (and/or reward-based) response to outside stimuli. Science would prove that in the Humean spirit, ought reasons motivate behavior because they are in the best interest of the individual. But as I have been arguing from the beginning, moral sense is a more complex perceptual capacity than merely the capacity to learn normative behavior in order to survive, or the capacity to construct rational arguments. Moral sense supervenes on the capacity for social behavior, the skill of coordinating behavior with conspecifics, by adding two of its own

242

Jennifer Hudin

functions: the capacity for collective intentionality, and the capacity for social commitment. This additional capacity that moral sense requires is a two-pronged function, which combines the capacity for rational assessment, and the capacity for intuition and emotion filtered through the lens of collective consciousness. Where does moral sense situate itself then in the real neurobiological world of decision making? Without the capacity for collective intentionality, the VMPFC operates solely in terms of the individual, and the DLPFC operates impersonally but not with an eye to social or personal considerations, a crucial factor in moral judgment.21 Moral sense must be a product of biological function that works together with the VMPFC and DLPFC but allows for their perceptions to be collective rather than individual. I suggest a third functional area in the prefrontal cortex that works with the VMPFC and the DLPFC to produce moral considerations, damage to which results in pathological, amoral behavior. Rather than a dual-process theory, moral behavior requires a tri-process theory for an adequate explanation. In summary, there is good evidence to show that the VMPFC does operate in moral dumbfounding, and the DLPFC in moral numbness, and we can now suggest that it operates in virtue of an additional biological capacity for moral sense, a biological capacity separate from the VMPFC and DLPFC, but operates along with the VMPFC/DLPFC to allow for moral intuitions and moral notions. Moral behavior is produced when the capacity for moral sense operates – the perceptual capacity that views decision making and action by means of collective intentionality and collective commitment.

21 This claim is corroborated by neuroscience in the form of the pathologies of both the VMPFC and the DLPFC. Briefly, damage to the VMPFC results in “shortsighted” decision making in which long-term effects that benefit the individual’s own well being are neglected. In addition, such patients do not display emotions towards familiar faces or situations. It is thought that their decision-making does not employ the “gut feeling” that is necessary and common to non-pathological decision-making (Damasio 2005: 40). Damage to the DLPFC is characteristically associated with schizophrenia, and also higher order cognitive deficits such as rule learning (Spence 2002).

The Logic of External Reasons and Collective Intentionality

243

3. How moral sense impels action without desire How then could this moral sense motivate an individual to act on practical reasons that are moral? Motivation is a notion that when semantically fleshed out inevitably makes reference to desire. Even on Searle’s account of external reason, agents ultimately move their bodies because they will it, and they will it because they form secondary desires. But it seems that there is very little willing occurring during acts performed out of moral intuition. Let’s take two groups of examples of social acts that can be deemed moral but clearly are not produced by deliberation. The first set are exceptional acts: a soldier throws himself on a bomb to save his buddies does not will his own destruction; a starving man shares his last piece of bread with other starving neighbors has little desire to deprive himself of more food and no hope whatsoever that sharing the little food he has will save anyone;22 A man who jumps into the depths of a lake to save a drowning man, a complete stranger to him, does not will nor desire any ensuing danger to his own life; In February of 1847, Tamsen Donner of the ill-fated Donner party, was given the choice to be rescued from months of starvation in their snowbound encampment and go to Sacramento. Rescue though, meant leaving her dying husband at the encampment. Mrs. Donner chose to stay with her dying husband. Tamsen died the day after her husband died. As with the cases above, Tamsen Donner’s choice did not consist in any self-serving desire whatsoever.23 During WWII, Kamikaze pilots who plunged their planes into American aircraft carriers did not will nor desire their own deaths. In fact, many of them strongly disapproved of the war, and did not want to die.24 These are a few exceptional

22 Russians who lived through WWII in Leningrad told this story to me when I was living in the Soviet Union. 23 The Donner Party consisted in a group of families who joined the emigration to California, in the summer of 1846. In November 1846, after some poor choices as to which route to take, the Donner Party found themselves caught in an early blizzard at Alder Creek, high in the Sierra Nevada, California. The Donner’s did not build wooden structures, but lived in sheds made of brush, for four months. They died, one by one, of cold and starvation. It remains a controversial question as to whether cannibalism occurred, and how Tamsen Donner actually died. Cf. Goodyear (2006: p.146). 24 “I do not want to die! ...I want to live. No, I don’t want to die...I feel lonely. I don’t know why I feel so lonely.” from the diary of Hayashi Tadao, (a Kamikaze pilot) November 26, 1940 (Ohnuki-Tierney 2006: xiv).

244

Jennifer Hudin

social acts that took a moral sense to perform whether we agree about the value of such acts or not. Now let us take some less exceptional social acts, the kind that we perform everyday that also qualify as moral: a man unwraps a candy bar and walks to a garbage can to throw away the wrapper; a woman gets on a city bus and pays the fee; Two friends meet on Wednesday night at a restaurant because they promised to do so on Tuesday; A group of pedestrians cross the street when the light turns green; A wife prepares dinner every night for her husband. In the exceptional cases, we can say that all of the individuals are acting in virtue of moral sense – out of some exceptional moral sense of doing what they feel ought to be done rather than out of any normal sense of desire, no matter how broadly desire can be construed. And though in the purest Humean spirit, this moral sense that drives exceptional acts can always be redescribed as some extended form of desire, as one of Hume’s calm passions such as the love of children and family and friends, that drives one to virtuous acts. 25 But desire in the normal sense of the word is not at work in these cases and even the Humean redescription cannot make logical sense. Individual actions in which the individual’s own death is the logical outcome cannot be said to satisfy some internal, dispositional state that is virtue. For how can any virtuous dispositional state, supposedly a good thing, result in its own extinction by it’s own satisfaction? Such a notion as satisfaction by extinction defies Hume’s original idea that internal reasons are ultimately in the best interest of the individual and selfish for that reason. There is a striking inconsistency in claiming that the reasons for self-sacrifice for the benefit of another are internal reasons because we lose the sense of what an internal reason is, and what it is suppose to do. Clearly the reasons that acts of self-sacrifice require a moral sense, and moral sense perceives states of affairs as not pertaining to the individual, but to the set, i.e., they are not internal. In the less exceptional everyday cases though, the individuals are also acting out of moral sense – doing what ought to be done, not what they necessarily want to do, or what they are inclined to do given pragmatic reasons. Consider the cases of the candybar eater and the wife who prepares dinner every night. For the candybar eater, it really is more efficient and much easier to throw the wrapper to the ground. And the wife might enthusiastically cook every night, but just as probable, she gets bored, or tired, or would rather go out and eat on occasion. In either case, extra effort is required to perform these very mundane activities, activities, which 25 Hume (Treatise: 487), Rawls (2000: 59).

The Logic of External Reasons and Collective Intentionality

245

are performed out of sense of what ought to be done rather than what fulfills a primary or secondary desire of any sort. These acts require a moral sense, a sense of what must be done. And for both kinds of action, exceptional and unexceptional, acts are performed because of external rather than internal reasons. Thus said, are these actions really motivated? Does the soldier and does the wife really feel motivated to perform their selfless acts, do they really need motivation to act on their external reasons? Perhaps in explaining how external reasons become practical reasons, motivation is not the appropriate operator, nor the only operator in question. I suggest that all of the cases above are cases in which external reasons as oughts trigger moral intuition and it is moral intuition, which initiates action. Moral intuition as one part of moral sense is nonreflective, and requires a sense of the collective and that of social commitment. While collective intentionality allows for the perception of the individual as part of a whole, it is social commitment, the emotional bond to the social group – be that a set of two or two thousand members – that moves the body. Social commitment impels rather than motivates an agent to act and it impels as a force, operating differently from motivation. Thus, on this model, external reasons operate quite differently from internal reasons because they do not require rational deliberation on the part of individuals in order to operate. Recognition of the truth of external reasons – that they are reasons for us – is sufficient to move the body of an individual when that individual is also committed to the social group in question. The recognition of group membership coupled with the capacity for the emotional bond of social commitment moves the body in the manner appropriate for the context in question. For everyday social acts, the body moves as the group body moves for the given context – paying a fare, crossing when the light is green, throwing rubbish in garbage cans or attending to an evening meal as part of a social role. None of these actions requires reflection, deliberation, desire or anything resembling desire. It is commitment that impels social action in whatever appropriate form and not motivation driven by personal desire when external reasons are in effect. Equally, it is social commitment that impels such exceptional moral acts as saving a life, alleviating the hunger of another, or saving the honor of a nation. Finally, how does this model accommodate rationality? If it cannot, then the model allows for the possibility of obedient zombies. Williams’ best argument for his sub-Humean model and against the possibility of external reasons, especially as I have proposed them, is that practical reasons must meet a rationality constraint. That is, coming to recognize the truth of a practical reason must be a process in which one deliberates from what he knows, i.e., the truth of the reason must meet internal logical relations

246

Jennifer Hudin

among already existing states in the personal motivational set. In addition then, coming to indexicalize the reason must consist in a process of matching how the reason fits into the already existing set of internal states in the motivational set. Williams argues that rationality requires this route of deliberation and any other way of moving the body is not acting on a reason, i.e., it is not a matter for rationality. But surely we do move our bodies without requiring referencing our motivational set. And though the movements we make are not products of rational deliberation, they are rational and subject to assessment both from the first-person and third-person point of view. And how could this be possible? On the practical side, it is possible by allowing intuition to work in isolation from rationality, the VMPFC to work isolated from the DLPFC. On the philosophical side, rationality operates in virtue of veto power. Truth recognition is sufficient to impel movement, but movement can always be vetoed and there are two ways that veto power can operate: 1) The individual in question automatically and with no deliberation recognizes the truth of an external statement but has a strong desire not to be part of the social group in question. Therefore, his internal desire trumps the external reason thereby rendering it ineffective to impel action; or 2). The individual simply denies the truth of the external reason. In the first case, the individual freely opts out of the social group in question. Let’s consider the case of owing taxes again because tax owing is a sophisticated institutional fact and if the model can work to explain this kind of fact, it can work for simpler institutional facts. Suppose the taxes that the agent owes are in part, spent on political causes he does not like. After recognizing the truth of the external reason of owing federal taxes – that it is true for him – he can choose not to pay them, in which case, the choice is made in virtue of desire and the negative act of not paying taxes is a result of an internal desire, and/or may be a case for akrasia. In the second case though, if the individual in question denies the truth of an external reason, and if he has access to all the facts, he is irrational. External reasons are truth functional arguments. Thus, the statement: You owe federal taxes, is the consequent of an argument. The full argument of this particular external reason is something like the following: Because you are a citizen of the United States (and enjoy all rights and responsibilities thereof) you owe federal taxes. If an agent denies the consequent but accepts the antecedent, he presents a classic case of irrationality. Recognition of the truth of the consequent is automatic given the knowledge of the state of affairs that the implicit antecedent provides. The states of affairs stated by antecedents of all exter-

The Logic of External Reasons and Collective Intentionality

247

nal reasons are either (actively) intentionally created by a set of individuals (e.g., a soccer club), or they can be states of affairs that members are born into (e.g., citizenry of a given country). In either case, these states of affairs form the social background against which moral intuition and commitment operate by means of impelling action. In less complicated cases, our daily cases, external reasons impel without reflection. Searle buys a beer and pays his check without any reflection at all. We receive our state and federal taxes, we pay them automatically with not much thought beyond the usual complaints of how high they are this year, etc. Indeed, we do not need to depend routinely on a sound deliberative route to get through our institutional day. External facts impel action, they do not motivate in the same way that our desires drive us. And they can only operate within an individual who shares a collective belief and commitment as to what is socially fair and what is socially outrageous given the social distribution of ought facts.

4. The logic of external reasons The classic belief-desire model cannot capture external reasons because moral behavior has its own logic. There are several ways in which external and internal reasons logically diverge. First, the belief of internal reasons is an “I believe”, a first-person singular belief. The belief of external reasons is a first-person plural belief – a “we believe”. Second, the engine of internal reason is motivation, which is driven by desire. The engine of external reason is not motivation but commitment. Thus internal reasons are constructed by an “I believe that p” in tandem with “I desire that p”, whereas external reasons are constructed by a “we believe that p” along with “I commit to p”, where p stands for a social set, of which the I is a member. Third, because external reasons and internal reasons are constituted differently, the ways in which they move the body are different: internal reasons, constituted by desire, motivate by deliberation; external reasons, constituted by commitment, impel by recognition of truth conditions. Fourth, the content of both internal and external reasons is of a different nature. The content of the practical reason of internal reasons is identical with the content of the desire that motivates it. 26 I will use Searle’s example to illustrate, if I want to eat chocolate ice cream, the (internal) reason for my buying chocolate ice cream is the content of my desire: that I want chocolate ice cream. In contrast, the content of an external reason consists in its truth conditions. If I keep my promise to 26 Searle (2001: 170) makes this point.

248

Jennifer Hudin

come to your house on Wednesday, the (external) reason for my coming to your house on Wednesday is the recognition of the truth conditions of the external reason statement. Finally, desire, though it may form inconsistent relations with other desires in an agent’s motivational set, nonetheless always functions within an agent’s motivational set of beliefs, desires, etc. That is, it is quite possible to want to do two different things, to want to perform two different actions that are quite at odds with each other. For example, I can both want to take a nap and go swimming at the same time. But for any practical reason that I actually do act on, it must be consistent with my set of beliefs. Again, if I choose to go swimming, the desire to swim must be consistent with a set of beliefs such as the belief that an accessible pool of water exists, the belief that I know how to swim, etc. Thus, there is an inherent rationality of practical reasons that are internal, though the desires that motivate may form inconsistent relations within their own set, the set of desires. In contrast, external reasons that trigger moral intuitions, do not appear to be rationally constructed in the same way as internal reasons. Social commitment that impels individuals to move their bodies – to act on external reasons – is often inconsistent with the beliefs and the desires of an individual’s motivational set. And as we saw earlier in the case of dumbfounding, moral intuition often forms inconsistent relations with moral rational reasons.27 But can external reasons per se really be subject to rational assessment if as I have argued they do not require deliberation? We have already argued that external reasons can be assessed and judged against an individual’s internal motivational set, but this is an ex-post facto rational assessment intended to block William’s zombie charge. What about the rationality of movement that is impelled by nonreflective moral intuition, movement that is not thwarted by internal reasons? Do such intentional actions fit 27 For example, Kahneman and Sunstein (2005: 98) provide examples of moral dumbfounding in which subjects judged a single action as either fair (morally acceptable) or unfair (morally unacceptable) depending how the action is described (framed) as in the following example:

Unfair: An employer who is doing well cuts wages by 5%, because many unemployed workers would be willing to work at the lower rate. Fair: An employer who is doing poorly cuts wages by 5%, to avoid or diminish loes. Given how a single state of affairs is described, particularly with respect to social equitability, outcomes of moral decisions have a predictable probability. As the authors point out, lawyers make use of this fact in argumentation in jury trials.

The Logic of External Reasons and Collective Intentionality

249

within any rational schema? In fact, we can argue from data such as that of Kahneman and Sunstein’s that such nonreflective intentional movements do fit within a rational schema, though not a rational schema that is internal to the individual, but a rational schema that is internal to the collective. As I have argued, moral intuition operates through the lens of collective intentionality. The rationality of moral intuition is the rationality of a given social group, which forms the background against which intuition operates. Though moral sense is a biologically innate capacity that is held in common by all humans, the various modes in which this sense finds its expression is through the moral sensibility of the individual’s particular social background. Moral decision-making is a more complicated matter than rationalization, be this a matter of reflection or satisfaction of an individual’s motivational set. There are good epistemic and logical reasons to argue that external reasons are ontologically and logically distinct from internal reasons. External reasons are facts that are collectively believed as We-reasons, and impel rather than motivate action. They are a function of non-reflective moral intuition and impelled by social commitment and collective intentionality.

References Chaiken and Trope (1999): Chaiken, S., and Trope, Y. Dual-process theories in social psychology, New York: Guilford Press, 1999. Damasio (2005): Damasio, A. Disorders of Social Conduct Following Damage to Prefrontal Cortices, in: Changeux, J., A. Damasio, W. Singer, and Y. Christen (eds.), Neurobiology of Human Values, Heidelberg: Springer-Verlag, 2005, pp. 37-46. Deacon (1997): Deacon, T.W. The Symbolic Species, The co-evolution of language and the brain, New York: W.W. Norton and Company, 1997. de Waal (2006): de Waal, F. Primates and Philosophers, How Morality Evolved, Princeton: Princeton University Press, 2006. Gilbert (2000): Obligation and Joint Commitment, in: Gilbert, M. Sociality and Responsibility, Lanham: Rowman and Littlefield, 2000. Goodyear (2006): Goodyear, D. What happened at Alder Creek, The New Yorker, Apr. 26, 2006. Haugeland (1982): Haugeland, J. Heidegger on Being a Person, in: Nous, (1982) 16: ‘15-26.

250

Jennifer Hudin

Hudin (2006): Hudin, J. Motor Intentionality and its Primordiality, in: Inquiry, (2006) 49: 1-18. Hume Treatise: Hume, D. A Treatise of Human Nature, L.A., Oxford: Clarendon Press, 1968. Kahneman and Sunstein (2005): Kahneman, D. and Sunstein, C. Cognitive Psychology of Moral Intuitions, in: Changeux, J., A. Damasio, W. Singer, and Y. Christen (eds.), Neurobiology of Human Values, Heidelberg: Springer-Verlag, 2005, pp. 91-105. Kant GMS: Kant, I. Groundwork of the Metaphysic of Morals, transl. Patton, H.J., New York: Harper, 1964. Ohnuki-Tierney (2006): Ohnuki-Tierney, E. Kamikaze Diaries, Chicago: Chicago University Press, 2006. Rawls (2000): Rawls, J. Lectures on the History of Moral Philosophy, Cambridge: Harvard University Press, 2000. Searle (1990): Searle, J.R. Collective Intentions and Actions, in: Cohen, P. and M.E. Pollack (eds.), Intentions and Communications, Cambridge: MIT Press, 1990. Searle (2001): Searle, J.R. Rationality in Action, Cambridge, MA: MIT, 2001. Searle (2007): Searle, J.R. What is Language?, unpublished manuscript, 2007. Sloman (1996): Sloman, S.A. The empirical case for two systems of reasoning, in: Psychological Bulletin, (1996) 119: 3-22. Spence (2002): Spence, S. Neuroscience and the will, in: Current Opinions in Psychiatry, (2002) 15: 519-526. Tomasello (2000): Tomasello, M. The Cultural Origins of Human Cognition, Cambridge, MA: Harvard University Press, 2000. Tuomela (1988): Tuomela, R. We-intentions, in: Philosophical Studies, (1988) 53: 367-389. Tuomela (2005): Tuomela, R. We-intentions revisited, in: Philosophical Studies, (2005) 125: 327-369. Velleman (1997): Velleman, D. How to share an intention, in: Philosophical and Phenomenological Research, (1997) 57: 29-50. Von Frisch (1973): Von Frisch, K. Decoding the Language of the Bee, Nobel Lecture, Dec. 12, 1973.

The Logic of External Reasons and Collective Intentionality

251

Williams (1981): Williams, B. Internal and External Reasons, in: Williams, B. Moral Luck, Cambridge: Cambridge University Press, 1981, pp. 101-113. Williams (1981a): Williams, B. Ought and moral obligation, in: Williams, B. Moral Luck, Cambridge: Cambridge University Press, 1981, pp. 114-123.

Valuing Interpersonal Relationships and Acting Together1 Monika Betzler, Bern

1. Introduction Often, we do not act by ourselves but together with others. Whether father and son build a sand castle, whether three felons rob a bank together, or whether a team of doctors and nurses performs several bypasses on a patient with heart disease – in all of these different cases, two or more persons act together. The individual persons might be doing very different things – the son might be heaping up the sand, for example, while the father gives shape to the castle. But under a certain description, they are acting together, since neither father nor son is building a sand castle by himself. It seems that there are many things we simply could not do by ourselves. But what exactly is it about events like the building of a sand castle, a bank robbery, or an operation that turns them into instances of joint action? Contemporary theories of joint action usually answer this question

1 This article first appeared in German in Deutsche Zeitschrift für Philosophie (2007) 3: 441-445. I owe thanks to the Akademie Verlag for permission to publish this translation. Many thanks also to Ellen Coughlin, Nora Kreft, Nikos Psarros and Pranay Sanklecha for their help in translating the article. I am very grateful for extensive written comments by Holger Baumann, Michael Bratman, Christian Budnik, Nora Kreft, Hans Bernhard Schmid and David Schweikard that have saved me from many mistakes. For helpful oral comments and suggestions regarding earlier versions of this paper, I would like to thank Michael Bratman, Christoph Horn, Rahel Katzenstein, Julian Nida-Rümelin, Martine Nida-Rümelin, Martin Rechenauer, Jörg Schroth and Mark Textor. I also benefited from the discussions following several presentations at the Institute of Philosophy in Berne in December 2005 and in Bonn in January 2006, at the colloquium for moral and political philosophy at the Geschwister-Scholl-Institute in Munich in February 2006, at the Institute of Moral Theology in Vienna in June 2006, at the Collective Intentionality V conference in Helsinki in August 2006, and at the colloquium for political philosophy at the University of Fribourg in 2007. N. Psarros, H. B. Schmid and K. Schulte-Ostermann, Concepts of Sharedness, Frankfurt 2008: 253-272; ontos verlag

254

Monika Betzler

with reference to shared,2 participatory,3 collective,4 or “we”5 intentions. According to those intentions, a joint action can be made intelligible either by postulating a kind of interrelation between the intentions of two or more persons, or by pointing to the form or way in which certain intentions are entertained. One of the central questions of this dispute is whether joint actions can be accounted for by reducing them to individual intentions and their relations to each other, or whether we need a non-reductionist theory of plural or collective intentions. Not all of the different versions of intention-based accounts explicitly claim to present us with a comprehensive theory of joint action. But it seems reasonable to suppose that they try to account for joint actions exclusively by reference to intention, and some of them explicitly generalize their theory in that way. Despite the considerable differences that exist between the various accounts, they share the basic assumption that particular kinds of intentions are both necessary and sufficient to explain joint action.6 Intention-based explanations are marked by a twofold neutrality: on the one hand, they are neutral with respect to the individual reasons that motivate particular persons to form an intention to act jointly and to participate in the given joint action; on the other hand, they stay neutral with respect to the kind of relationship that exists (or does not exist) between the jointly acting persons and that influences exactly how persons act together.7 This twofold neutrality seems advantageous to a comprehensive explanation of different kinds of joint actions: in this way, the fact that particular persons act together for very distinct reasons does not present any difficulty for the explanation of joint action. Even if one of the felons robs the bank because he wants to live in the Caribbean for the rest of his life, the other because he has always wanted to do something extraordinary, and the third because he wants to own a mansion in the best part of town, their respective intentions to rob a bank seem to sufficiently explain their joint action in spite of their individually varying reasons. Equally, the fact that persons relate to each other very differently does not challenge their account of joint action. Whether doctors or nurses know 2 Cf. Bratman (1993; 1997); Gilbert (1997), however, speaks of “joint commitment”. 3 Cf. Kutz (2000). 4 Cf. Searle (1990). 5 Cf. Tuomela (2003) or Schmid (2005: ch. i and ii). 6 The so-called teleological accounts that were introduced as alternatives to the standard view are an exception to this. Cf. Miller (2001). 7 Here we can distinguish between the temporal extension, the frequency and the degree of participation and organisation of joint actions.

Valuing Interpersonal Relationships and Acting Together

255

each other, whether father and son are bound by close familial ties to each other, or whether there are any power asymmetries between the agents is irrelevant for explaining in what sense they are acting together. In this paper, I will discuss neither the considerable differences nor the problems that each of the various versions of intention-based accounts of joint action face. Also, I will not call into question the fundamental significance and relevance they have for explaining many cases of joint actions. However, the exclusive focus on intentions in contemporary theories of joint action misjudges the significance of a particular class of reasons that play an important role in explaining joint action, and should therefore not be ignored. These are reasons generated by a particular kind of relationship that exists between jointly acting persons. Any explanation of joint action that does not refer to these relationship-dependent reasons at least in some cases of joint action would be incomplete. Some joint actions do not occur - or, better, do not only occur - in order to achieve some joint aim. Rather, they occur in order to give expression to the kind of relationship that persists between the respective agents, a relationship that they value. The distinct explanatory role that can be assigned to interpersonal relationships, and to the reasons generated by them, is revealed by the fact that a certain type of rational criticism of joint actions can be made intelligible only by reference to them. Even when persons are acting together, we sometimes believe that we are justified in criticizing them for not acting together in the right way. Here, we are appealing to the distinct interpersonal relationship that persists between the respective persons and that gives them reasons to act together in a particular, normdependent way. For the explanation of some joint actions, it is necessary to refer to relationship-expressive reasons. In order to defend this thesis, I will first explore what sort these interpersonal relationships are and in what sense their participants value them. Afterwards, I am going to explain what reasons these interpersonal relationships generate for their participants: insofar as interpersonal relationships are valued, they generate reasons to realise them in a more or less normatively determined way. One of the conditions for realising an interpersonal relationship is that their respective participants act together and thereby realise the value of their particular interpersonal relationship. In order to defend these claims, I will close by considering some objections that could be brought forward against them. In this way I want to show that a solely intention-based explanation of joint action is at least in some cases incomplete.

256

Monika Betzler

2. The value of interpersonal relationships In order to call attention to the distinctiveness of interpersonal relationships and their potential value in a first step, I want to start by introducing the following two examples: (i) Max and Moritz are going on a journey together. They have been friends for more than ten years now, and they have decided to travel to Iceland this summer. After they arrive in Reykjavik and settle down in their hotel, they start discussing how they want to spend the rest of the day. Max would like to explore the city on foot by himself and have a few beers in some of the chic bars in town afterwards. Moritz is sulking because he would like to spend the rest of the day together with Max. Throughout the following days, Moritz gets increasingly annoyed with Max. Even when they are doing something together – for example, visiting the Laugadalur geothermal pool – Max does not seem to care much about his companion. Instead, it seems as if Max is only incidentally interested in the same activities as Moritz. Moritz feels neglected and issues a bitter complaint to Max. Max has difficulty understanding Moritz’ complaint. After all, they have planned the journey together, they share the same hotel room and go on some trips together. Moritz feels let down and comments angrily: “I thought we were friends.” (ii) Hans plays the piano and joins a jazz band that meets regularly for joint jamming sessions. He enjoys improvising at the piano and attends all the jamming sessions. In addition to that, he loves public performances and values the opportunity to play abroad with his band and get to know interesting places. But after only a few rehearsals, the other musicians take him aside and complain about his musical contribution. They say he doesn’t listen closely enough to the them and doesn’t react to their musical suggestions properly. Also, he does not show any willingness to participate in a dialogue with the other musicians. Hans had thought he was a master of improvisation at the piano. But the other musicians complain about his inability to create a composition together with them. They claim that Hans does not know what it means to jam together. In both cases, persons act together with others. However, a difficulty arises from at least one person not acting together with the others in the right way. Both Max and Hans have the intention to act together with the others. So the criticism they face is not grounded in a lack of commitment on

Valuing Interpersonal Relationships and Acting Together

257

their part to stick to their intentions and to adjust their actions to those of the others.8 The intentions themselves do not provide us with the relevant content and thus fail to have the necessary authority they need to justify and render intelligible the expressed criticism. After all, Max did travel to Iceland together with Moritz, he is sharing a room with him and joins him on some of his trips. We can thus assign to him a series of intentions to act jointly with Moritz that he shares with Moritz. Similarly, Hans could defend himself by pointing out that he acted upon his intention to improvise together with others. We can thus assign to him a relevant intention that he shares with the other musicians. Still, the criticism to which Max and Hans are subjected seems to be justified, because both of them also agreed to dedicate themselves to the realisation of a particular interpersonal relationship (or at least they appeared to agree to it). It is legitimate to assume that Max, by going on a journey with Moritz, also wants to realise their friendship. By taking part in the jam session, Hans can be expected to want to realise a certain musical relationship. We can assign the legitimate expectation to Moritz and the other jazz musicians in Hans’ band that the joint actions with Max or Hans take place at least partly because they each value their respective relationships. But since both Max and Hans make mistakes when it comes to the realisation of their respective interpersonal relationships, and since the other parties have good reasons to suppose that they (Max and Hans) intend to realise these relationships, their criticism appears rationally justified. After all, Moritz can assume that Max wants to go travelling with him, because he is his friend and because he believes it is valuable to do things together with his friend. Likewise, the other jazz musicians can assume that Hans wants to improvise together with them because he values jamming with others. The criticism cannot be grounded in the intentions assigned to Max and Hans, but in the valuation of their respective interpersonal relationships that is underlying their intentions. Only this valuation allows us to call these cases of joint action cases of friendship and of jazz improvisation. 8 In particular, M. Gilbert has drawn our attention to the duties that collective agents commit to when they voice their desire to perform a certain action together. The agents thus have a justified claim on each other to act accordingly (Gilbert 1990). Gilbert’s thesis is often criticized as too strong: Michael Bratman (1997) is more careful with regard to the normative consequences of intentions. According to him intentions at most generate prediction about how a person will act. He also mentions that further actions have to be coordinated according to the intentions.

258

Monika Betzler

The above examples thus show that particular types of interpersonal relationships – namely, those shaped by certain feelings of intimacy between friends, and those obtaining between musicians in a band qua musicians – are able to explain some cases of joint action more exhaustively than intentions. With this in mind, we must first clarify what characterises interpersonal relationships, understood in that sense, and what differentiates them as valuable and distinct. Put more precisely, friendship is a relationship between two persons, both of whom have certain feelings for the respective other and care about the friend. In contrast to other types of interpersonal relationships, friends entertain more intimate relationships to one another, partly shaped by shared interests and values, as well as by empathy, trust and the exchange of private information. The intimate relationship of friendship is valuable. Here, it is possible to distinguish the value of friendship generally from the value of a particular friendship, say the friendship between Max and Moritz. This distinction corresponds to the different values that such relationships have from the perspective of third parties and from the perspective of the first person. From the perspective of third parties, we can think of friendship as valuable without being friends with someone ourselves (for example, because we believe that friendships stabilise social structures). From the uninvolved perspective of third parties, what counts is to determine the value of such engagements with regard to their utility to society and to promote them accordingly. From the perspective of the first person, however, what matters is the value that relationships actually have for the persons participating in them. Being friends with someone is valuable for several reasons: among other things, friendships facilitate the improvement of one’s own life and the satisfaction of one’s own interests; this is because they allow for recurring experiences of intimacy with a concrete other person, who is perceived as valuable, in part at least for her own sake. When it comes to the value of jazz improvisation, we can again distinguish between two possible perspectives. A minister of arts and culture might think that free-jazz performances should be promoted more strongly (perhaps in order to increase the scope of publicly funded cultural events). But to actively take part in jazz improvisation is valuable because it allows for specific aesthetic experiences for the person participating in it of jointly producing music. In contrast to friendship, this does not require the musicians to relate to each other as individuals. All it requires is that they relate to each other as players of certain instruments.

Valuing Interpersonal Relationships and Acting Together

259

However, relationships of this kind are valuable from the perspective of the participating persons only if further conditions are fulfilled: they have to know, implicitly or explicitly, what it means to realise the value of such relationships. Only this will provide us with the relevant reasons to eventually act together in the right way. Otherwise they would only have reasons to make valuable experiences for themselves. But insofar as persons have reasons – from their first-person perspective – to act together, they have them because of the realisation-conditions of values such as jamming or particular friendships. We can therefore distinguish more precisely between the properties or aspects that render such relationships valuable to those participating in them, and the conditions that need to be fulfilled to achieve this. To think of jazz improvisation or friendship as in itself valuable presupposes not only a belief in the value of the friends, or of a certain sequence of notes as aesthetically interesting, but also being engaged in a relationship with such friends, or with the other musicians with whom a particular musical event is produced. A friendship or a band’s jazz improvisation can become a valuable engagement only if the participating persons relate to each other reciprocally. The specific way in which they relate to each other still needs to be analyzed more closely. This reciprocal reference to one another at least partially constitutes friendship and jazz improvisation, and bestows on them the value characteristic to them.9 Thus, the reason friendship is valuable for the respective friends is – among other things – that they relate to each other in friendly ways over time. Jazz improvisations are valuable for the respective musicians because interactively producing music comes along with particular aesthetically pleasing experiences. It is the friendly reference to as well as the musical communication with others that renders friendships and jazz improvisations valuable for their own sake. Friendship or jazz improvisation can develop only insofar as persons relate to each other reciprocally. What makes friendship and jazz improvisation valuable is – at least in part – that in both cases, two or more persons relate and refer to each other in distinct ways that are still to be determined more precisely.10 9 For a first approximation of the difference between intrinsic and extrinsic values, cf. Korsgaard (1983). For a more precise definition, cf. Rabinowicz and RönnowRasmussen (1999). 10 R. Long (2003) points out that friends value not only each other as persons, but also the relationship they have: “It’s only when something has gone wrong that we find ourselves valuing the friend but not the friendship”. N. Kolodny (2003)

260

Monika Betzler

In order to characterise more closely and in the relevant ways the value of this reciprocal reference, it is helpful to distinguish it from cases in which a person has certain experiences in the company of others. Joint experiences are not sufficient for realising the value of a relationship. When persons visit a rap concert, for example, more or less all of them have the same auditory experience of a particular musical genre. This experience might be intensified by the fact that other persons experience something similar at the same time and in the same space: their expression of fun and enjoyment might increase such emotions in others. However, this kind of joint experience is valuable only because each one of them engages in positive or enjoyable experiences by herself that are at best intensified by the expressions of others. The cause of these experiences is the rapper, and the joy of other listeners is nothing but a catalyst. In order to actually realise the value of reciprocal interpersonal reference, something must be added to the joint experience: the experience of others when we refer to them and when they refer to us is in itself valuable. In short, the mutuality that can be expressed in reciprocal relationships (such as friendships or jazz bands) has its own independent value. This is what makes friendship and jazz improvisation intrinsically valuable. To experience others as valuable in themselves is possible only if this very experience has been caused by the behaviour of others in reference to us. Our response to this behaviour can have an intensifying effect that sustains and supports the reciprocal reference. Companionship or mutuality can be thought of as distinctly valuable under these circumstances alone. The referring to others and the “being referred to” by others is what we evaluate as positive in a variety of different ways. And it is this reciprocal reference that, at least in part, renders friendship and jazz improvisation valuable for the persons involved. The following example shows that not every kind of reciprocal reference constitutes and thereby renders valuable such engagements.11 Two fellow workers compete in owning the fastest car. This is because, on the one hand, fast cars have the intrinsic property of particularly great engine power and, on the other hand, because they may have the extrinsic property of being faster than the cars of the competitor.

argues extensively that relationships have a distinct value and generate reasons of a special kind. 11 I am grateful to Hans Bernhard Schmid, who presented this example to me as an objection.

Valuing Interpersonal Relationships and Acting Together

261

In this example of so-called positional goods, however, we are not dealing with “interrelational value” in the sense in which I consider relationships as distinctly valuable. For a start, it is not constitutive of the value of fast cars to compare them with other peoples’ cars. Hence, the reciprocal relationship is not at all necessary. We could also reconstruct the example in the following way: it could be thought of as achieving the highest possible position in society. Ownership of fast cars could be regarded as useful in this respect. Comparing one’s own car with that of one’s colleague would thus be a means to promote this particular end. In this case the reference to the colleague is necessary, but it would not be valued for its own sake. Rather, it would serve the end of achieving a highly esteemed position in society. However, the reciprocal reference that is constitutive of friendship and jazz improvisation, and which also contributes to their value, is not a means to increase the value of friends or of aesthetic experiences. In order to realise such engagements at all, the act of referring to others has to be valued for its own sake. But what turns valuable engagements of the kind mentioned into interrelational values is the fact that, in order to be realised, the participating persons must refer to each other reciprocally and for the sake of their relationship’s own value. Nevertheless further elaboration is required on what exactly distinguishes different valuable engagements from each other and how the interpersonal reference characteristic of them is determined. That is to say, the reasons that persons have to realise such engagements have to be demonstrated. Interrelational values of this kind can only be realised if two or more persons accept the corresponding reasons to behave and to act jointly. They thus accept reasons to share interrelational values. I am going to turn to the question of what reasons persons really have to realise and to share interrelational values in this way.

3. Expressive reasons The particular reasons a person has to realise an interrelational value of the kind mentioned are determined by norms underlying this interrelational value. In other words, how exactly persons have to relate to each other in order to realise their friendship is determined by cultural and social norms and practices. In principle, each instance of friendship has to be defended with reference to such norms, and it thereby becomes subject to rational criticism too. It is part of the conditions that need to be fulfilled in order to

262

Monika Betzler

realise friendship at all that a person refers to another person with different emotional attitudes and acts with reference to this other person. The other person is also prone to such attitudes and actions. Thus, friendship might to varying degrees be determined by cultural and social norms. In order to realise a jazz improvisation, a different kind of reciprocal reference is needed. What distinguishes it from the reference among friends is that a musician does not have to refer to the other musicians as individual persons, but only as representatives of certain instruments. Thus, it usually does not matter much if musicians are exchanged, as long as for each instrument there is a musician with the relevant musical skills. The reciprocal reference must be such that the rules of the interplay are created while carrying out the reference itself. Paul Berliner points out: Over its course, players are perpetually occupied: they must take in the immediate inventions around them while leading their own performances toward emerging musical images, retaining, for the sake of continuity, the features of a quickly receding trail of sound. They constantly interpret one another’s ideas, anticipating them on the basis of the music’s predetermined harmonic events. Without warning, however, anyone in the group can suddenly take the music in a direction that defies expectation, requiring others to make instant decisions as to the development of their own parts. […] By the journey’s end, the group has fashioned a composition anew, an original product of their interaction.12 How exactly persons refer to each other to realise friendship or jazz improvisation is therefore to varying degrees determined by some set of given rules. In order to realise the values determined by these norms, the individual persons have to accept reasons to behave and act accordingly. These reasons differ in their concrete content (i.e., in what exactly they demand that we do), since they motivate different norm- and ruledetermined behaviour and actions, aimed at the realisation of different interrelational values. But they are alike in that they are always reasons to share certain interrelational values and to express this by behaving accordingly. They share common features insofar as they move agents – at least in part – to value the respective interrelational values for their own sake, i.e., non-instrumentally. The reasons are thus (i) norm-dependent and (ii) non-instrumental. In order to realise interrelational values, each partici-

12 Berliner (1994).

Valuing Interpersonal Relationships and Acting Together

263

pating person must think of herself as someone who shares these values. This means that she accepts reasons to behave and act accordingly.13 In order to realise the value of friendship, a person will typically accept reasons for certain emotional attitudes and actions. She has, for example, reason to feel happy when she sees her friend. She will long for him when she has not seen him for a long time, she is glad when he is doing well and feels compassionate when he is faring badly. A person who attends to her friendship will also accept reasons to act with reference to her friend. She will call him from time to time to find out how he is doing. She will give him presents occasionally, or meet him regularly to do something with him. She will converse with him and invite him to her house. The respective friend is prone to act similarly. A member of a jazz band that regularly meets for jazz improvisation has reasons to contribute to the adequate realisation of such improvisations. This implies that she refers to the other musicians as representatives of certain instruments and as producers of particular melodies. For example, she has to listen to them in order to properly time her cue. Further, she has reasons to blend into the group as a collective, rather than to try to place her own musical contribution at the centre of attention (unless, of course, this was agreed upon). Apart from determining the actions of each person, norm-dependent reasons also tell us how to act. In order to realise interrelational values, more is required than certain ways of behaving and acting: they have to be carried out in such a way that they are not simply viewed as means to some further end. That does not mean that the participants of friendships or jazz improvisation as friends and as jazz players do not also act out of instrumental reasons. Max, for example, may also value friendship because it allows him to go on more affordable holidays. Equally, Hans might value jazz improvisation because it allows him to be in the spotlight. However, neither of those two interrelational values can be realised if they are valued exclusively as a means to a further end. Firstly, the other mentioned ends (i.e., to go on more affordable holidays and to be in the spotlight) could not be achieved if friendship and jazz improvisation were exclusively valued as a means to them. Secondly, friendship and jazz improvisation might possibly be realised very insufficiently if the norm-determined 13 There are different ways to conceive of the connection between values and reasons. To give a complete account of this connection is beyond the scope of this paper, but I tend towards the following interpretation: the valuing of values implies that one accepts reasons to realise the respective value. For a discussion on the possible connection between values and reasons, see Heuer (2004).

264

Monika Betzler

reciprocal reference to the other persons is not aimed at and valued for its own sake. This means that whoever is not able to refer to other persons as her friends or as other band members, and who cannot think of this reference as valuable in itself, does not have the ability to realise friendships and jazz improvisation. The question still remaining to be answered is how each single person who wants to realise such valuable engagements can make sure that the other persons contribute their due. After all, friendship cannot be realised if there is only one person referring to others with the appropriate feelings and actions. Likewise, jazz improvisation cannot be accomplished if the other musicians are not contributing properly. How can they trust that the others also accept reasons to realise friendship or jazz improvisation? In the case of jazz improvisation, a person’s joining a jazz band suggests at the outset that she accepts such reasons. Following her joining, the other members seem to be justified in having normative expectations of her behaviour. Thus, they can legitimately assume that she became a member because she values the kind of reference to others that is typical for jazz improvisation, and because she wants to react accordingly to reasons to practise this reference. If the new member underwent an admission process, the other members would have even more reason to assume that she has the relevant knowledge and skills, and that she accepts her duty to contribute to the joint performance in an appropriate way. Should the new member not match the normative expectations, the old members seem justified in criticising her. In the case of friendship, we cannot identify a similarly clear line, the crossing of which allows us to assume that the other person will contribute her due to its realization. Rather, we are here dealing with a gradual process. The first indications of friendship develop over time. They can be perceived and experienced. Thus, a person may show signs of joy whenever she sees the other person again. She may signal her interest in the well-being of the other by asking the relevant questions and showing attentiveness. If she shows a number of such reactions, and they are intensified and deepened by the corresponding responses, we can expect a friendship to develop. However, insofar as a person can legitimately assume that she is the participant in a friendship, the thereby realised value of friendship has normative authority for the respective friends. In the case of the jazz band, the declared membership serves that function. The realisation of interrelational values commits the participants of this realisation to continue to share the value of norm-determined reciprocal refer-

Valuing Interpersonal Relationships and Acting Together

265

ence, ceteris paribus. The reasons generating interrelational values thus gain normative authority as soon as a person conforms to them.

4. Advantages: The normative framework of joint action and rational criticism In my view, the concept I have introduced of sharing interrelational values explains some cases of joint action better than the alternative accounts. Intention-based approaches not only remain consciously neutral with respect to the realisation conditions of joint aims as well as to the reasons a person might have to pursue them, but they also provide us with descriptions of single events of actions. As a result, Max and Moritz’ undertaking could be designated as an instance of joint travelling, and Hans’ participation in the jazz band could be counted as an instance of joint music production by reference to his intentions. Sometimes, however, joint actions are embedded in a normative infrastructure. They realise not only the intentions directly preceding them, but also the interrelational values that are shared by the participating persons and that ground these intentions. The case of the joint journey should be made intelligible not only with reference to the relevant intentions to travel together. Rather, the intentions causing the concrete sequence of actions would have to be explained by reference to the interrelational value of friendship. If a person cares for her friendship with another person, she accepts reasons to express this friendship in a number of emotional attitudes and corresponding joint actions. The intention to travel together with the friend would then represent the conclusion she has drawn from a number of reasons, which she accepts. However, a joint journey that can be traced back to such an intention is an expression of her care for her friendship only if she is willing to be motivated by further reasons for similar emotional attitudes and actions. But if Max only shares a hotel room with Moritz and goes on some trips together with him, without having the relevant emotional attitudes, that is, without acting together with him out of the appropriate friendly reference (i.e., without valuing this reference for its own sake), then his intention and the ensuing joint journey was not an expression of friendship. In this case, the joint journey does not take place to realise friendship, but to achieve other aims (like, for example, a more affordable holiday). Analogously, the intention to produce music together with others might be understood either as isolated or as an expression of the wish to realise the distinct reciprocal reference that is typical for jazz improvisation. If

266

Monika Betzler

the latter, the intention is at least partially an expression of valuing the above-mentioned reference for its own sake. Some joint actions take place because the participating persons realise interrelational values. In those cases, the corresponding actions are an expression of the realisation of these values. For this to happen, it is not sufficient for the participants to have the respective intention to perform some isolated joint action; rather, on top of that, their intention must be embedded in the acceptance of further reasons, according to which a particular joint action takes place, among other things, in order to value a reciprocal reference for its own sake. Furthermore, my suggestion can make more sense of the duties persons have to each other reciprocally, insofar as they act together. According to intention-based approaches, a commitment to act together stems from interdependently conditioned intentions.14 According to that, a person intends to contribute her part to a joint action as long as the others intend to do so, too. But conditional intentions of this kind do not generate the reciprocal commitments we are after. Just as my intention to go cycling whenever the weather is nice does not mean that I have any duties towards the weather, conditional intentions do not generally generate any commitments. And what is more, it is not consistent with the term “intention” to review it again and again, that is, to always think about how it is dependent on the intention of others.15 The conception of interrelational values avoids the problem of reciprocally conditioned intentions and accounts in a much more plausible way for the commitment persons have to one another, if they share interrelational values. Insofar as they commit to the realisation of an interrelational value, they agree to correspond to the norms that determine this respective value. These norms determine, among other things, how persons should behave and refer to one another. Insofar as they do not correspond to the norms, they can be rationally criticised.

14 Cf. Bratman (1997) and Gilbert (1993; 1999). Compare also Velleman (1997). 15 According to Abraham Roth (forthc.), this problem can be solved only if a person takes on another person’s intention directly, without re-considering it herself. But this is only possible in cases where there is some authoritative asymmetry between persons. However, I cannot respond to Roth’s suggestion more extensively at this point.

Valuing Interpersonal Relationships and Acting Together

267

5. Objections and responses At first sight, my thesis may seem vulnerable to several different objections. In order to defend it, I will discuss two central objections. Firstly, I would like to demonstrate that accounts of joint actions that refer to interrelational values are not circular in a vicious, but in an informative, sense. Secondly, I would like to clarify to what extent interrelational values and intentions are related. The relevance of intentions in explaining joint action does not have to be denied. 5.1 Circularity It could at first be objected that friendship and jazz improvisation are collectively “loaded” expressions.16 As interrelational values, they are grounded in the conception of collectivity or commonality. For this reason, they are not suited to accounting for joint actions – or so the objection might run. To explain joint actions by referring to interrelational values, therefore, seems to be circular. What was to be explained has already unreflectively been presupposed. In response to this, I simply want to deny that the premises include any term that has been presupposed unreflectively and used to explain joint actions. Rather, joint actions are interpreted as applications of the normative term under scrutiny. That is, they are being tested as to whether they correspond to the norms that determine interrelational values like friendship and jazz improvisation. To decide if such a correspondence is actually present or not, we are in need of some justification with reference to the interpretation of these norms. In that sense, the idea that constitutes interrelational values is not just uncritically presupposed, but is interpreted and defended by reference to the norms that determine each concrete interrelational value. Single instances of joint action can subsequently be examined as to whether they actually contribute to realising these conceptually thus defended interrelational values. Hence, we are not facing a viciously circular explanation, but an informative account that examines, interprets and justifies the application of a collectively “loaded” term by reference to norms that determine it. Our general understanding of reciprocal reference is thus interpreted as justified in particular, concrete instances. As a 16 Cf. Bratman (1992: 96 f.; 1993: 114). Bratman argues that for the analysis of joint action we should only consider those kinds of activities that do not conceptually presuppose acting together with others (like, say, dancing Tango or playing football).

268

Monika Betzler

result, Moritz’ taking offence becomes understandable: he can render his complaints plausible by referring to those norms that in his eyes justifiably determine friendship. 5.2 Interrelational values and intentions Further, it might be objected that I am wrongly presenting my conception of interrelational values as wholly distinct from an intention-based conception of joint action. After all, a person who joins a jazz band and participates in joint music production must also form intentions – or so the intention-based theoreticians might complain. She must, for example, intend to follow the rules of jazz improvisation and contribute her part to the joint endeavour of music production. The claim that interrelational values are special, however, does not necessarily deny the connection of these values with intentions. A member of a jazz band may actually have the mentioned intentions. But what explains the joint action equally well is the interrelational value of jazz improvisation – the fact that persons mutually refer to each other, which is regarded as valuable. To share interrelational values does not exclude that persons do and also must form intentions regarding the realisation of these values. In the end this makes plausible that the members of a band also have intentions. But they do not have to have intentions regarding the effectiveness of the intentions of the other musicians, nor do their intentions have to be such that they view their contribution as a means to a common end. But the distinctiveness of interrelational values can also show that some joint actions take place without any intentions being formed. This seems to be especially pertinent in the case of friendship. Sometimes we befriend a person without having intended to do so and without forming any further intentions as a result. Only a conception of interrelational value can explain joint actions that take place out of such an implicit understanding of friendship.17 According to a similar objection, interrelational values could not be regarded as anything other than shared intentions. Michael Bratman maintains that “shared valuings” are nothing else than so called “shared policies”, which are in turn described as intentions whose content is sufficiently general:18 For us to value X is, in a basic case, for us to have a shared policy to treat X as justifying in shared deliberation in which we intend 17 Brewer (2003: 554 f.); compare also Helm (forthc.). 18 Cf. Bratman (2004; 2007: 1-5).

Valuing Interpersonal Relationships and Acting Together

269

to engage and by which we intend to be guided. We say in what such a shared policy consists by extending the account of shared intention to the case in which the content of the relevant policylike intentions of the participants is our treating X in this way.19 In this vein we are dealing with a decision made by the participants regarding which considerations should be given special weight. But it does not seem very plausible that friendship or jazz improvisation is a special reason to act together, to which friends or jazz musicians agree after joint consideration. Friendship, for example, does not at all presuppose that friends explicitly and consciously agree to grant this value any normative privilege. Further, the talk of viewing x as a justifying reason after a process of joint consideration underdetermines the source of normativity just as much as what agents should subsequently do. In my opinion, an explication of interrelational values can explain far better why they should enjoy normative privileges. After all, an interpretation of the norms that determine them indicates what reasons agents have to share and realise them together.

6. Concluding remarks Action theories that traditionally deal with explanations of single persons’ actions generally have it that actions can only be accounted for sub specie boni. According to them, an action can be said to be an action if we can tell a plausible story about how the agent’s judgements about what is desirable correspond to what she does, and that she takes this to be the reason for which she acts. In theories of joint action the standard move is to explain joint actions by reference to some form of intentions. In my opinion this is partly because intentions have properties that seem to be particularly relevant for joint actions. Intentions are the conclusions of deliberation processes. They commit persons to actually carrying out what they intend to do, and they thus allow them to act stably and coherently over time. Attempts to explain joint actions face the challenge of showing why persons should engage in joint actions in the first place. The stability and the related normative privilege of intentions seem to be particularly suited to mastering this problem. The only thing left to explain is how the individual intentions relate to each other, and whether or not there are plural varieties of intentions. 19 Bratman (2007).

270

Monika Betzler

In this article, I have tried to make fruitful a sub specie boni version of action explanation for the explanation of joint actions. However, my thesis is of modest scope. I have not argued that all types of joint actions should be explained by reference to values. All I have tried to show is that some joint actions can only be made intelligible if we refer to certain types of values, i.e., to what I called interrelational values. As an explanation of joint actions, interrelational values cannot be wholly substituted by intentions, no matter of what kind. This becomes evident if we consider that some forms of rational criticism participants of joint actions might be subjected to can only be justified by reference to such values. Also, we can show with reference to such values not only that persons might be motivated by different reasons to act collectively; but that, in order to realise these values, a fundamental acceptance is required to value a certain form of reciprocal reference for its own sake. I have talked a lot about friendship and jazz improvisation, and this focus might seem arbitrary. These are by no means the only interrelational values. I have merely chosen them as paradigms of interrelational reference. However, I hope to have shown that the concept of sharing interrelational values can more exhaustively account for some types of joint action, and that a certain form of rational criticism can be made intelligible only by reference to such values.

References Berliner (1994): Berliner, P. Thinking in Jazz. The Infinite Art of Improvisation, Chicago: Chicago University Press, 1994 Bratman (1992): Bratman, M. Shared Cooperative Activity, in: Bratman, M., Faces of Intention. Selected Essays on Intention and Agency, Cambridge: Cambridge University Press, 1999, pp. 93-108. Bratman (1993): Bratman, M. Shared Intention, in: Bratman, M., Faces of Intention. Selected Essays on Intention and Agency, Cambridge: Cambridge University Press, 1999, pp. 109-129. Bratman (1997a): Bratman, M. Shared Intention and Mutual Obligation, in: Bratman, M., Faces of Intention. Selected Essays on Intention and Agency, Cambridge: Cambridge University Press, 1999, pp. 130-141. Bratman (1997b): Bratman, M. I Intend that We J, in: Bratman, M., Faces of Intention. Selected Essays on Intention and Agency, Cambridge: Cambridge University Press, 1999, pp. 142-161.

Valuing Interpersonal Relationships and Acting Together

271

Bratman (1999): Bratman, M. Faces of Intention. Selected Essays on Intention and Agency, Cambridge: Cambridge University Press, 1999. Bratman (2004): Bratman, M. Shared Valuing and the Framework of Practical Reasoning, in: R. J. Wallace et al. (eds.), Reason and Value: Themes from the Moral Philosophy of Joseph Raz, Oxford: Oxford University Press, 2004, pp. 1-27. Bratman (2007): Bratman, M. Dynamics of Sociality, in: Midwest Studies in Philosophy (2007) 30: 1-15. Brewer, Talbott (2003): Two Kinds of Commitment (And Two Kinds of Social Groups), in: Philosophy and Phenomenological Research 66: 554-583. Gilbert (1990): Gilbert, M. Walking Together: A Paradigmatic Social Phenomenon, in: Gilbert, M. Living Together. Rationality, Sociality, and Obligation, Lanham: Rowman & Littlefield, 1996, pp. 281-311. Gilbert (1993): Gilbert, M. Agreements, Coercion, and Obligation, in: Gilbert, M. Living Together. Rationality, Sociality, and Obligation, Lanham: Rowman & Littlefield, 1996, pp. 281-311. Gilbert (1996): Gilbert, M. Living Together. Rationality, Sociality, and Obligation, Lanham: Rowman & Littlefield, 1996. Gilbert (1997): Gilbert, M. What Is It For Us To Intend?, in: Gilbert, M. Sociality and Responsibility. New Essays in Plural Subject Theory, Lanham et al.: Rowman & Littlefield, 2000, pp. 14-36. Gilbert (1999): Gilbert, M. Obligation and Commitment, in: Gilbert, M. Sociality and Responsibility. New Essays in Plural Subject Theory, Lanham et al.: Rowman & Littlefield, 2000, pp. 50-70. Gilbert (2000): Gilbert, M. Sociality and Responsibility. New Essays in Plural Subject Theory, Lanham et al.: Rowman & Littlefield, 2000. Helm (forthc.): Helm, B. Plural Agents, in: Nous, forthc. Heuer (2004): Heuer, U. Raz on Values and Reasons, in: R. J. Wallace et. al. (eds.), Reason and Value: Themes from the Moral Philosophy of Joseph Raz, Oxford: Oxford University Press, 2004, pp. 129-152. Kolodny (2003): Kolodny, N. Love as Valuing a Relationship, in: Philosophical Review (2003) 112: 135-189. Korsgaard (1983): Korsgaard, C. Two Distinctions in Goodness, in: Philosophical Review (1983) 92: 169-195.

272

Monika Betzler

Kutz (2000): Kutz, C. Acting Together, in: Philosophy and Phenomenological Research (2000) 61: 1-31. Long (2003): Long, R.T. The Value in Friendship, in: Philosophical Investigations (2003) 26: 73-77. Miller (2001): Miller, S. Social Action. A Teleological Account, Cambridge: Cambridge University Press, 2001. Rabinowicz and Rönnow-Rasmussen (1999): Rabinowicz, W., and Rönnow-Rasmussen, T. A Distinction in Value: Intrinsic and For Its Own Sake, in: Proceedings of the Aristotelian Society (1999) 100: 33-52. Roth (forthc.): Roth, A.S. Shared Agency and Contralateral Commitments, in: Philosophical Review, forthc. Schmid (2006): Schmid, H.-B. Wir-Intentionalität. Kritik des ontologischen Individualismus und Rekonstruktion der Gemeinschaft, Freiburg und München: Alber, 2006. Searle (1990): Searle, J. Collective Intentions and Actions, in: P. Cohen et al. (eds.), Intentions in Communication, Cambridge, MA: MIT Press, 1990, pp. 401-415. Tuomela (2003): Tuomela, R. We-Intentions Revisited, in: Philosophical Studies (2003) 53: 115-137. Velleman (1997): Velleman, D. How to Share an Intention, in: Philosophy and Phenomenological Research (1997) 62: 29-50.

Mental Non Self-sufficiency, Sociality and Common Agency Nikos Psarros, Leipzig

1. The incompatibility between collective agency and mental self-sufficiency In classical theories of collective intentionality that explain collective agency by means of we-intentions1 or joint commitments,2 it is tacitly accepted as a conditio sine qua non that the participants in a collective action3 are supposed to have a full-fledged cognitive and emotive “structure” – a personality –, which once established can sustain itself without any interaction with other similar entities. In other words, persons are taken to be self-sufficient mental entities. This assumption is commonly expressed in form of an axiom stating that persons participating in a collective action do this driven by individual and also self-sufficiently existing preference resp. intentional structures, which somehow coincide with the intentions that are ascribed to an entity that is postulated as the – real or theoretical – subject of the collective action, e.g. a group or a community. There are several accounts of how we-intentions or joint commitments come into being on an individualistic basis. However, they face at least three common difficulties: a) They cannot explain resp. give an account of source, from which joint commitments or obligations draw their coercive power that, on the one hand, forces the participants in a collective action to observe their commitments resp. obligations and to rebuke defectors, and on the other hand forces defectors to try to justify their defection by stating the reasons that necessarily render any further participation in the given collective action impossible. Among modern philosophers only Max Stirner has attacked this subjection under the “external obligation” as 1 Searle (1995), Tuomela (2003). 2 Gilbert (2000). 3 For the purposes of this presentation the term collective action will be used as a label embracing all types of cooperative agency including joint actions. N. Psarros, H. B. Schmid and K. Schulte-Ostermann, Concepts of Sharedness, Frankfurt 2008: 273-284; ontos verlag

274

Nikos Psarros

a kind of addiction to an utterly wrong idée fixe, proposing instead that persons should establish only free associations (Vereine4) retaining the right to quit membership at any moment without further explanation. All other authors, however, whether they are contractualists or not, seem to agree at least on the unlimited validity of the old legal principle pacta sunt servanda, i.e. the idea that accepting an obligation means in any case that one waives the right to cancel this obligation arbitrarily. b) They cannot explain why individuals can be urged, or even forced to participate in collective actions that they did not initiate or in actions that are not in accord with their current preferences or intentions, for example, why individuals can be conscripted into military or other compulsory services, or why citizens must obey legitimate laws5 with which they are not in accord. Individualistic accounts cannot also explain why it is not irrational when people remain voluntarily on their posts for the sake of a duty or an ideal even if this means that they will sacrifice pivotal individual goods like freedom or life. Especially in case of the sacrifice of the own life without any expectable benefit, if not for one itself then at least for one’s closest relatives,6 individualistic preference-centred accounts must fail completely, since there cannot be any ego-centred preference or commitment involving the own death when there is still the possibility of coming along by staying alive. c) They cannot explain why a person can still claim to participate in a collective action although she is actually not doing anything that could be identified as a participating act – i.e. why sentences like “we are dancing, but I’m taking a break” are meaningful.7 The problem is that the idea of self-sufficiency underlying the classical concept of personhood excludes logically the idea of cooperation that is essential for the concept of collective action – the term “cooperation” is used here in a strong sense, meaning that cooperating persons join a col4 Stirner (1982: 306; Verein is translated as union). 5 Cf. Raimondi, this volume. 6 Consider for example the case of a young soldier who has no family that eventually would be rewarded for his death on duty, but nevertheless stays at his post until he falls, or the case of the stokers on board of the Titanic who remained in the engine room ensuring the lighting of the ship during its abandoning, being aware of the fact that they so deprived themselves from any chance of survival and that because of the legal situation of seamen at those days no-one of their closest relatives would be compensated for loosing them. 7 Schmid (2007).

Mental Non Self-sufficiency, Sociality and Common Agency

275

lective action for the sake of a “common good” taking into account that their own interests may remain unsatisfied. This is so because self-sufficiently existing entities do not depend on other entities for their self-definition as entities of a given kind.8 The idea of mental self-sufficiency allows self-sufficient mental entities to “assemble” – and in this sense to do something “together” – if they decide individually and independently from each other to do so in a Stirnerian way, always retaining their right and their ability to leave the assembly without giving any reason. From the point of view of a self-sufficiently existing person there is no path leading to collective action except that of external force, either in form of practical necessity – e.g. the impossibility of hunting a large animal alone – or in form of a social environment that acts as an obstacle between the individual and its goals. Our human intuition is, however, that people often engage in collective actions not just because they are forced to do so by the circumstances, but because they feel the need to cook, walk, discuss, or make sport with others. People seek this form of cooperativity in order to establish an environment in which they can unfold their personal individuality as members of a community.9 The actualisation of a collective action entails the idea that the participants not only depend, but also rely on each other for their definition as participants of the given collective action. One cannot, for example, start a race without competitors and without taking the behaviour of her co-competitors during the race into account for ones own behaviour. In other words, the participants of a collective action act together for the sake of the collective action itself, and they do so because the collective action in question is part of their own self-definition. The defection of a participant results under these circumstances in the (more or less severe) impairment of the self-definition of the others, giving them the right to rebuke the defectors of the cooperation, unless they present good reasons for their behaviour, but also in the impairment of the selfdefinition of the defector itself, who seeks to repair this impairment by accepting the punishment imposed upon him. Animals, on the other hand, are typical mentally self-sufficient entities. They do not rely on conspecifics in order for their species-typical lives to unfold, including their cognitive and emotive capacities: a dog for example displays the typical behaviour of a dog even when it is raised in a dog8 Psarros (2005). 9 This does not exclude the possibility of the existence of a private sphere that is inaccessible to others without permission, nor does it exclude that individuals cannot pursue private aims without the aid or the knowledge of others.

276

Nikos Psarros

free environment and a sheep can live its sheep life without its herd. This holds also for the defective states of an animal, i.e. if an animal displays some stress symptoms because it lives isolated from its herd, these symptoms are species-typical for an isolated animal of its kind. Obviously there are lots of animal species, in which the lives of individuals depend more or less from the existence of conspecifics, with which they form internally structured groups, like bees, ants, lions, sheep, or higher primates. The existence of these groups is not maintained by cooperation, but by a mere “mechanical” behaviour that is based on the superposition of instinctive behavioural routines. This holds also for higher primates, who seem to have individual preference structures and are capable of pursuing individual aims. However, even among the mentally most developed primates there is no cooperation in a human sense, but only a kind of mutual “exploitation”.10 The fact that obviously some animals can learn new forms of behaviour by observing the behaviour of their conspecifics or even by observing their environment is compatible with the thesis of animal mental self-sufficiency because what animals can learn is confined by what they are. An animal does not learn its kind-specific behaviour. It can learn only to adjust a given kind-specific behaviour within the margins that are specified by the nature of its kind. With humans the situation is radically different: Babies raised in a human free environment adopt – in the very rare cases they survive at all – the features of this environment failing to develop the characteristic traits of human behaviour.11 It cannot be denied, however, that some animals living in the vicinity of humans display traces of cooperativity. A sheepdog for example, has to realize that his task is to supervise the herd in cooperation and on behalf of the shepherd in order to be consigned by the latter with this job. This means that the animal understands in a way that the sheep are neither a toy nor a prey, but a good by its own right that has to be taken care of, in spite of the fact that this care bears no immediate profit for the dog. However, it also seems in such cases that the dependence of the animal on the cooperation with its human master is more or less superficial and does not affect the underlying mental self-sufficiency of the involved animals, since unattended dogs lose these traits of cooperativity and turn wild.

10 Tomasello (2001). 11 Such cases are known in the literature as “wolf children”. Despite the fact that many of them remain unconfirmed or have been proven to be fictitious there are several documented reports of children that had been raised by animals. All of them displayed a similar behaviour including the inability to walk upright and to smile, and also had great difficulties in learning to talk.

Mental Non Self-sufficiency, Sociality and Common Agency

277

2. Humans as mentally non self-sufficient entities and the “Needs of the Human Soul” Accounts of we-intentions and joint commitments aim at explaining the inherent cooperative sociality of humans. However, they do not explain, but just reformulate it, because the underlying theories either try to derive we-intentions and joint commitments from individual attitudes and preferences, or they introduce joint commitments in an ad hoc manner as a counterweight to the individualistic self-sufficient persons. We-intentions and joint commitments are similar in their function to Hobbes’ “natural law of reason”12 that forces his bellicose individuals to accept that a truce under the supervision of Leviathan is a better means for achieving their individual aims than eternal war. If there is such a law of nature then it is a mystery why it does not apply also to other similarly developed, mentally self-sufficient creatures like primates. If it is a law that applies only to the human nature then it is tautological with the insight that humans are not animals. But, if humans are the only living beings to whom Hobbes’ “natural law of reason” applies, or who – in more modern words – can form we-intentions resp. become subject to joint commitments, and if this law, we-intentions and joint commitments are at odds with the proposed selfsufficiency of the self then perhaps we should give up the tacit assumption that participants in collective actions are mentally self-sufficient entities. The idea that humans are not mentally self-sufficient entities can be traced back to Plato and Aristotle. It underlies also the Christian concept of God as the ultimate transcendental judge, above every person and every concrete community. In modern philosophy this idea has been expressed and pursued mainly by Hegel, Wittgenstein and also by Simone Weil, whose work remained mostly unnoticed because of her early death during World War II and of her tendency towards a certain kind of religious mysticism. Simone Weil grounds her reflections on sociality13 on the idea that beings participating in collective actions are necessarily mentally non selfsufficient entities. In order to be constituted as persons and also in order to retain their personality, such beings have to enter an already socially ordered world that contains already formed persons and to stay therein. Weil calls this moment of primeval double dependency both from other 12 Hobbes (Leviathan, Book I, Ch. 14 and 15). 13 Weil (2001).

278

Nikos Psarros

persons and from the socially ordered world the Obligation. The Obligation affects both individuals and communities and is directed towards other “obliged” individuals and communities embracing also past and future generations. In other words, in order to become a person one has to be embedded not only in a local socially structured environment, but also in an environment that is itself embedded in a historical process: it has to have an intelligible past and an open future. Personhood, and all the features and the rights associated with it, arises from the interaction of obliged individuals and communities within this historical frame. The success of this interaction relies according to Weil on the fulfilment of a set of conditions that she calls the Needs of the Human Soul. Despite of this characterization the Needs of the Human Soul are not conceived as individual mental or psychological states. Weil presents them rather as elements of a social ontology leaving, however, their exact ontological status unclear. The fundamental premises of S. Weil’s social ontology are that 1) a good personal life can only be established in a good living community and that 2) this can be the case only when all Needs of the Human Soul are fulfilled in a proper manner. The list of the Needs of the Human Soul encompasses fourteen elements that are ordered in seven pairs of opposites: duties

rights

order obedience hierarchy punishment risk common property

freedom responsibility equity honour certitude private property

truth

freedom of opinion

aspect

moral poietic communicative

Their status as needs means that they are not only abstract and formal conditions for the successful unfolding of human life, but that humans experience them as their own desires and pursue their satisfaction. This always includes the possibility of exaggeration, which can be manifested as addiction or voracity. Being addicted to order, for example, means that one accepts any imposed system of social order no matter whether this particular order is an obstacle to the fulfilment of other Needs of the Soul or not. Voracity with respect to order means, that one declares one par-

Mental Non Self-sufficiency, Sociality and Common Agency

279

ticular system of order as absolutely valid or true, focussing all efforts on the extinction of any other existing order. In this context certain cases of cowardice have to be regarded as forms of addiction to order and political or religious fundamentalism a form of order voracity. The fact that achieving a state of addiction or voracity is at miss with the satisfaction of the Needs of the Soul entails the fact that a good life requires their continuous and balanced satisfaction, both by the acting individuals and their communities. This is possible only if Weil’s Needs of the Human Soul are regarded as describing necessary and sufficient conditions for the successful unfolding of a fundamental form of collective agency that has the form of freely organized common actions. Common actions depend on the coordinated cooperation of mutually mentally interdependent individuals, i.e. the cooperation of beings that are mentally non self-sufficient. Common actions are as such actions on their own behalf, i.e. their nature is independent from the nature of the actions of the participating individuals. Common actions are not the sum or the result of the iteration of individual actions, but rather the result of the interference of individual actions of various kinds, which are oriented towards the common goal that characterizes the common action and which are ordered according to the nature of the common action. Thus success and failure of a common action are not linear functions of the success and failure of the individual actions on which it relies. A good example demonstrating the complex relationship between a common action and the individual actions that realise it by their interference is an election: It can succeed even if a number of the voters did not vote correctly and it can fail in spite of the fact that every voter has successfully realised her own act of voting, e.g. in case that a participation quorum was not achieved – successful individual voting is just one of the conditions of a successful election.14 Human life is a form of common action that is characterized by the fact that the participating individual agents are free to choose among a number of options regarding their actual contribution to the common goal, which in this case is the continuation of human personal life itself. This kind of freely organized common actions has a general formal structure, the constitutive aspects of which can be identified with the Weilian Needs of the Human Soul listed above. The Needs subsumed in the first column of the table under the title duties15 are necessary for the constitution of the subject of the common action. Since such a subject, however, can be formed 14 For a detailed account of the concept of common action see Psarros (2006). 15 Simone Weil does not use the terms rights and duties as introduced here.

280

Nikos Psarros

only by the interaction of individual agents they figure as their duties towards the common goal. The Needs in the second column describe the conditions that are necessary for the constitution of the participants of a freely organized common action and are therefore referred to as the rights of the participating individuals. The first four pairs make up the moral aspect of a freely organized common action, i.e. they determine its structural parameters. The pairs risk/ certitude and common/private property comprise its poietic aspect, i.e. the parameters that determine how a common action can achieve a certain goal. Finally the pair truth/freedom of opinion addresses the communicative aspect of a freely organized common action, namely the circumstance that in order for such an action to succeed, the participants have to establish a form of communication that is based on those two conditions. Simone Weil does not examine the question whether the particular list of the Needs of the Human Soul is sufficient for establishing a good personal life in a good living community or not. She also does not make any reference to the philosophical tradition, to which she is indebted. However, although a final answer cannot be given easily, there is a striking similarity between the list of the Needs of the Human Soul and Plato’s and Aristotle’s reflections on the same topic. 2.1 Plato: The Needs of the Soul as constitutive factors of the Polis In the 9th book of Plato’s Republic, Socrates gives a sort of historical explanation of the fact that the perfect state described by him does not exist in reality. His claim is that the “current” situation is the result of the historical decay of a formerly existent perfect state. There was a time, so Socrates’ tale, when the perfect state existed according to the principles elaborated in the discussion with his friends. The political system of this state Socrates calls “axiocracy” because only the best reigned. However, with the passage of time the principles for the appointment of the rulers weakened somehow so that a cast of “honoured” people entered the stage taking over the government. The members of this cast were still the best from the axiological point of view, but they didn’t rule any more because they were the best, but because they were respected as being the best. So the political system changed to “timocracy”, i.e the rule of the honoured. Further down in time the principles of appointing the rulers were more weakened, so that the offspring of those honoured men took up the government in a hereditary manner, the political system mutating into an “oligarchy”, i.e. the rule of a closed group. At the beginning the oligarchs still ruled wisely defending bravely the state. However, since the criteria

Mental Non Self-sufficiency, Sociality and Common Agency

281

for appointing the best citizens as rulers were no more applied properly it turned out that the rulers began increasingly to pursue their own benefit instead of taking care of the common welfare – a “plutocracy” emerged, the rule of the rich. This behaviour evoked insurgency in the rows of the common people who seeked participation in the power, the wealth and the government. In cases they succeeded the political system of “democracy” was established, the rule of the many. The rule of the many was however unstable, since every one tried to put through his own interests so that a fight of everyone against everyone commenced. At the end a single person managed to overpower all others and to establish himself as a sovereign ruler, introducing the political system of “tyranny”. According to Socrates, both the polis and the individual soul have a similar constitution. Thus the shift in the political constitution affects also the mental constitution of the citizens, so that to each political state there is a corresponding mental configuration – the timocratic, the oligarchic, the plutocratic, the democratic and finally the tyrannic man. However, the 4th century BC Athenian political system cannot be identified with any particular political structure in Socrates’ tale. It was actually an amalgam of axiocratic, timocratic, oligarchic, plutocratic, democratic and tyrannical elements.16 Therefore, Socrates’ tale should not be read as the narration of a factual historical process, but as a sort of parable that tries to show something that was neglected in the debate of the Politeia so far. The process of transformation that Socrates describes resembles strikingly the decay of a corpse. At the beginning we still have the dead looking like the living person. With commencing decomposition the outermost layer of the body, the skin, dissolves, giving insight to the inner structure of the body, to the muscles, the soft tissues and finally, after all organic material has decayed, to the bones – the last principle of the body, which in its purest form has the least elements of life, nevertheless being the fundamental bearing of the living person. Similarly the decay process of the perfect state unveils the essential factors that are necessary for its maintenance, showing at the same instance that each of them alone is not sufficient. According to the story these factors are honour, bravery, property, freedom, and individual determination. Comparing them with S. Weil’s Needs of the Soul we find a striking correspondence despite the 16 Theoretically the best should rule. In fact the political power was distributed among few influential rich aristocratic families who by means of their wealth could control large parts of the people’s assembly. On the other hand, the plebiscitary nature both of this institution and of the courts, which were more or less tribunals, made for ambitious men possible to gain influence that was sometimes equal to a tyrant’s.

282

Nikos Psarros

fact that apparently Plato did not analyze the structure of the necessary factors for a properly functioning state as thoroughly as S. Weil did. 2.2 Aristotle: The Virtues as evaluative manifestations of the Needs of the Soul In the Nicomachean Ethics Aristotle introduces the concept of virtue in order to describe a fundamental characteristic of human beings, which is neither a property nor a capacity, but a kind of attitude towards both one himself and a common good. The concept of Aristotelian virtue presupposes an understanding of man as a mentally not self-sufficient being that has to rely on other non self-sufficient beings in order become what it is. Man is by nature a communitarian being, a ζῷον πολιτικόν, and the virtues that are needed for a good individual life in a good living community are at the end expressions of the Needs of the Soul in the mode of the evaluation of one’s own personality. The vicinity of the Aristotelian concept of virtue to S. Weil’s Needs of the Human Soul is emphasized by pointing out their similarity with respect to four fundamental traits: 1. They are necessary for the unfolding of a good life in a good living community, 2. they can be missed in the directions of exaggeration or deficiency,17 3. the right form of fulfilling them is achieving the balanced middle between those two extremes, 4. the fulfilment both of Aristotelian virtues and Weilian Needs of the Human Soul is an eternal process realized as active participation in freely organized common actions. The difference between the Needs of the Human Soul and the Platonic decay steps of the ideal state on the one hand, and the Aristotelian virtues on the other, consists in the fact that Plato and Aristotle examine the success conditions of respectively one aspect of human life, Plato being interested in the structure of the common action itself and Aristotle focussing on the part of the individual participant. S. Weil on the other hand, tries to give a complete account of the common action called human life. Thus the Needs of the Human Soul display this double structure of pairwise arranged duties and rights.

17 The virtue of bravery, for example, can be exaggerated becoming foolhardiness, or be deficient becoming cowardice.

Mental Non Self-sufficiency, Sociality and Common Agency

283

3. Common agency and collective responsibility The concept of Common Agency together with the concept of Obligation and the concept of the Needs of the Human Soul can shed new light on many problems connected with collective agency as for example on the problem of collective responsibility. According to Simone Weil, responsibility is one of the Needs of the Human Soul, i.e. a necessary condition for the constitution of freely organized common actions, which are pivotal for the constitution of personhood. Thus responsibility is fundamental for the constitution of personhood itself. Since from a praxeological point of view there is no difference in principle between an individual and a collective agent because both are the subjects of their respective actions, collective agents are likewise not self-sufficient regarding their constitution – they rely on the existence of other collective agents at the same level of structural complexity. This means that collective agents also have to participate in common actions, for which responsibility is a structural condition. Thus both individual and collective agents can be held responsible for the consequences of their actions at the corresponding level and both are subject to subsequent sanctions (or praise). This does not mean, however, that the participants in a common action can or should be treated in the same manner as the responsible collective agent, or that all of them can or should be ascribed to the same degree of responsibility. Since most of the common actions that characterize human life are freely organized actions that depend only on the individual contribution and do not prescribe the exact way it has to be delivered, the individual responsibility can be judged according to the individual contribution to the given common action. In this context the concept of Obligation provides a criterion for the exercising of one’s need for responsibility when confronted with a common action that violates moral standards: Because not only individuals but also collectives (as agents) are subject to the Obligation towards other individuals and collectives in historical perspective, a person who sees that her collective is about to commit a reprehensible common action is obliged to refuse cooperation, or to contribute as least as possible to it according to her abilities: ultra posse nemo obligatur. Thus the descendants cannot be held responsible for the deeds of their ancestors, notwithstanding the fact that they are eventually responsible for their rectification.

284

Nikos Psarros

References Gilbert (2000): Gilbert, M. Sociality and Responsibility, Lanham: Rowman and Littlefield, 2000. Hobbes Leviathan: Hobbes, Th. Leviathan, Oxford: Oxford Paperbacks, 1998. Psarros (2005): N. Psarros. Autonomie und Autarkie, in: M. Kober (ed.), Soziales Handeln – Beiträge zu einer Philosophie der 1. Person Plural, Ulm: Universität Ulm Humboldt Studienzentrum, 2005, pp. 85-95. Schmid (2007): H.-B. Schmid. On not Doing One’s part, in: Psarros, N. and K. Schulte-Ostermann (eds.), Facets of Sociality, Frankfurt/ Heusenstamm: ontos verlag, 2007. Searle (1995): Searle, J.R. The Construction of Social Reality, New York: Allen Lane, 1995. Stirner (1982): Stirner, M. The Ego and Its Own – The Case of the Individual Against Authority, transl. by St. Byington, London: Rebel Press, 1982. Tomasello (2001): M. Tomasello. The Cultural Origins of Human Cognition, Harvard: Harvard University Press, 2001. Tuomela (2003): Tuomela, R. ‘The We-mode and the I-mode’, in F. Schmitt (ed.), Socializing Metaphysics: The Nature of Social Reality, Lanham: Rowman and Littlefield, 2003, pp. 93-127.

Joint Commitment and the Practice of Democracy Francesca Raimondi, Potsdam/Frankfurt Oder For a certain strand in the analytical theory of action, “collective intentionality” is the concept that has to be elucidated in order to make sense of a peculiar type of action in which more than one person is involved. For those cases in which a person is not only acting ‘on her own’ in the presence of others but doing something together with them show a ‘jointness’ of the individual behaviours that supposedly calls for special explanation. The basic idea is that, in order to explain the ‘jointness’ of actions like taking a walk in a park together with a friend, the appeal to a specific sort of intentions is required, namely “collective” ones.1 The task of providing a suitable account of joint action is not merely supposed to be a problem for the theory of action. Margaret Gilbert, in particular, has stressed upon the link between collective intentionality and questions of social as well as political philosophy. Her account of joint action is in fact supposed to provide a model for the explanation of sociality as such and its normative structure. In her recent book A Theory of Political Obligation. Membership, Commitment, and the Bonds of Society, which sums up the lines of her political thinking deployed in a large number of articles, Gilbert tries to explain the structure and normativity of political collectives on the basis of her concept of joint activity. Put very briefly, the main thesis of the book is that membership in a political collective has to be described as a specific case of taking part in a social group, i.e. in a joint activity. Since for Gilbert membership in a social group grounds obligation towards the norms governing the practice of the group and towards its other members, political obligations in her view should be explained as resulting from engagement in a collective practice. 1 However, the various approaches to collective intentionality understand this notion in quite different ways. The scope reaches from reductionist positions (e.g. Bratman), which analyze collective intentions strictly in terms of individual intentions and their interrelations, to non-reductionist positions (e.g. Gilbert), which ascribe collective intentions to some sort of collective subject (a “we”) that cannot be reduced to the sum of individual intentions. Apparently, only nonreductionist positions take “collective intentionality” to be a genuine category of the theory of action, whereas reductionists still operate with the customary setting of an individualistic terminology. Thus non-reductionists hold that a new vocabulary is required in order to explain joint actions. N. Psarros, H. B. Schmid and K. Schulte-Ostermann, Concepts of Sharedness, Frankfurt 2008: 285-304; ontos verlag

286

Francesca Raimondi

In this paper I would like to investigate how Gilbert tries to shape her account of joint actions into a basis for political theory. On the one hand, her approach seems to be promising, for it grounds political obligations neither in metaphysical facts, like the will of god or a supposedly natural system of rights and duties, nor in individualistic values like self-preservation. Rather it correlates them directly with the fact of sharing a practice with others. On the other hand, the concept of a social group and of joint activity can appear to be inadequate for characterizing political collectives, since the way we become a member of such a ‘group’, as well as our relation to its norms, is quite peculiar. Furthermore, the concept of group runs the risk of being too narrow, or of suggesting too much identity, and so of failing to capture the plurality that is constitutive of political collectives. I will begin my investigation by analyzing Gilbert’s concept of small groups, for it is a central facet of her methodology to start precisely with small social entities in the assumption that their features also apply, mutatis mutandis, to larger entities like political collectives. I will argue that her concept of a (small) social group suggests a type of account of political obligation, which does not seem to match with Gilbert’s political theory. This reflects the fact that the traits that are supposed to found obligation within small groups do not really apply to political collectives. In particular, as I will argue, the genesis and kind of political membership are such that Gilbert’s analogy must fall short in some relevant aspects. This will lead me to the question how the notion of membership in political collectives should instead be understood and what the consequences of this shift are for the understanding of the collectivity of political societies. I will discuss this topic with special reference to the case of modern democratic societies, since in my view Gilbert’s theory, despite her explicit claim to account for political societies whatsoever, displays a distinctively modern understanding of political collectives.

1. Gilbert’s conception of social groups and the Contractualist tradition Just as joint actions are distinguished from merely spatio-temporally coexisting actions, so social groups differ from mere “aggregates”2 of per2 Gilbert (2006: 15 f.) Gilbert borrows the term “aggregate” from Rousseau, who opposes it to “associations,” which are populations characterized by “unity” (Rousseau 1983: 23).

Joint Commitment and the Practice of Democracy

287

sons. Gilbert sees the determining features of both joint actions and social groups in an underlying “intentionality of membership, unity, and consciousness of unity”.3 So, on her view, neither shared features (like ‘being born in X’) nor shared beliefs (like the belief that the others will also cooperate) are sufficient criteria for membership in a group. In both cases, the ‘sharedness’ of the relevant feature is acknowledged only from a thirdperson perspective. In social groups, by contrast, the members know each other as taking part in a shared practice and they therefore feel obligated to shape their behaviour in accordance with this practice. Gilbert calls such a mutual obligation to a shared practice, which she regards as the distinctive feature of social groups, a “joint commitment”. Only insofar as the members themselves acknowledge that they are committed to do something do they also have reason to act in the relevant way. And only when their commitment is a joint commitment are we entitled to speak of a joint action and not simply of a mere coexistence of actions. It is consequently the joint commitment of the members that “gives a social group a substantial kind of unity, a unity perceived by its members, without whose appropriate understanding it cannot be”.4 For Gilbert to enter a joint commitment with other people means to become a “plural subject” and to pursue a more or less definite goal “as a body”. A plural subject, however, is not supposed to designate an obscure ontological entity but rather express the specific normativity of joint actions in contrast to other kinds of action. With the terms ‘joint commitment’ and ‘plural subject’, Gilbert tries to make room for the fact that obligations in cases of joint action are not located only within the standpoint of the single subject. The single person is not only obligated by virtue of her own individual rationality but by reason of a social norm.5 3 Gilbert (2006: 96). 4 Gilbert (2006: 91) 5 With the notion of joint commitment, Gilbert signals her rejection of “singularist accounts” of collective actions as not being able to elucidate the sense of social obligations. If joint actions were to be rationalized in exactly the same way as individual ones (with the only difference that in the practical premises of joint actions figure also in assumptions regarding the conduct of other persons), then the point of reference of the obligation remains the same, namely the individual ‘I’ and his or her interests. According to singularist accounts, acting against my promise to make a trip with S means acting against my own intention, and is therefore a case of individual inconsistency but not the violation of a commitment to others. Since they locate the source of the obligation in the rationality of the single person, singularist accounts are not able to explain why I should keep a promise even if I have personally good reasons to break it (cf. Gilbert 2006: 123).

288

Francesca Raimondi

Furthermore, these terms support a specific view on how to conceive the structure of collective actions, namely as binding the individuals as parts of a single body.6 In order for a joint commitment and a plural subject to come into being, however, there must be some mutual acknowledgement of the fact that the individual members of the group want to engage in a joint activity: “each must manifest his willingness to constitute with the others a plural subject of the goal in question. When each has done so, the foundation is laid for each person to act in his capacity as the constituent of a plural subject of the relevant goal.”7 Undertaking a joint activity presupposes therefore a “willingness” or “readiness” to act with other persons, and it is the mutual demonstration of readiness that brings a plural subject into being.8 Gilbert leaves the meaning of “willingness” or “readiness” somewhat indeterminate in order to apply these terms to a multiplicity of cases. So there is readiness when two persons decide to go for a walk together, but also when one of them starts a dispute, which the other simply “falls into”. In both cases, though, something like a previous act of the will consenting to the joint action is supposed to be involved. In Gilbert’s view, then, acting together and being jointly committed comes about not simply through action, but rather requires each individual person to express his or her intention of doing so. Membership in a group and joint commitment are thus based upon intentional acts, which give rise to the collective obligation. In this sense, Gilbert’s formula “A joint commitment is a commitment of the will”9 can be read as expressing an internal relationship between individual intention and collective obligation. According to Gilbert, what binds individual persons to some shared rules is precisely membership in a group. She argues that this is shown by the fact that members of social groups have “a special standing to rebuke and make demands”10 on each other that outsiders supposedly lack. She therefore concludes that “the standing of the participants is a function of their joint activity. Thus it is special not only in the sense of not being shared by people generally, but also having a specific source, namely, the joint activity.”11 Since a distinctive feature of mutual acknowledgement of 6 Precisely this feature belongs in my view to a rather specific form of collective actions and makes, as I will argue in the last section, the application of Gilbert’s account of collective practice to other contexts of action precarious. 7 Gilbert (1989: 204). 8 On this point cf. also Gilbert (1994: 184 f.). 9 Gilbert (2006: 134). 10 Gilbert (2006: 103). 11 Gilbert (2006: 104).

Joint Commitment and the Practice of Democracy

289

willingness to act together is required in order to speak of a joint activity as “not being shared by people generally”, it is in the end precisely this distinctive feature that grounds entitlements and obligations. So far, then, one could say that Gilbert connects an individualistic element of consent with a non-individualistic understanding of norms and obligations.12 In showing his or her willingness to act together with others, the individual submits him- or herself intentionally to some shared norms that henceforth will govern the behaviour of each member of the group. It is therefore the individual person and her private intentions, which give rise to her obligation towards social rules.13 If this were in fact the central idea of Gilbert’s view of joint commitment, then it would be rather uncontroversial to place a political theory based on it in the tradition of contractualism. For it is a characteristic trait of contractualism to trace the validity of social norms back to the consent of individuals committing themselves mutually as if they were entering into a contract to conform to a social order. The contractualist theorists typically did not assume, however, that the scenario that underlies their argument also describes the historical genesis of political collectives. They did not understand the social contract as an historical event, which brought the social order into being, but rather as a theoretical construct. The point of the hypothetical scene of contract making, for the contractualist, is to show that the public order, which in fact is always already enforced, could in principle be justified to its members as preferable to the state of nature.

12 As Esposito argues, to conceive of a social entity in terms of subjectivity – as for instance a plural or general subject – relies of a still persisting link to the singularist paradigm (Esposito (2004: 7 f.). In Gilbert’s case, this seems to hold in the sense that – so far we have seen – from a genealogical point of view her concept of a group is still conceived as the correlate of the singular wills of its members, and the group as such functions therefore as a ‘larger person’ acting out this will. To put the point in another way, such accounts remain in the singularist paradigm, since they confer on collective phenomena the ‘monological’ structure of individual actions, and therefore fail to capture the constitutive plurality of actions in which different people do things together. In the last section of this paper I will come back to this point. 13 Taking note of the “weak analytic individualism” of On Social Facts, Baier (1997: 17) points out an “individualist bias” in Gilbert’s conception of collectivity, since, as Gilbert herself argues, “the concept of an individual person does not require for its analysis a concept of collectivity itself unanalysable in terms of persons and their non-collectivity-involving properties” (Gilbert 198: 435-436).

290

Francesca Raimondi

The aim of constructing a hypothetical scene of contract making is therefore not a descriptive, but a practical one. It does not explain how social norms arise, but is supposed to influence the attitude of subjects towards the political order. In its first formulation (in Hobbes), the social contract theory was indeed meant to sustain the subjection of the ‘real’ members of his society. Hobbes’ somewhat paradoxical claim to justify the demand for the actual obedience of real dissenters by way of the fictive consent of a fictive mankind was later on rejected by Rousseau. He was the first to explicitly work out that the function of the contractualist argument is in fact a normative one.14 In Rousseau, the counterfactual scene of the social contract does not serve the purpose of legitimating the actual society, but rather helps us to figure out what a truly legitimate order would consist in. In this case, imagining an original contract supports the articulation of a political order that free and equal persons (in the state of nature) would consent to. In the history of political philosophy, the contractualist position (as a position based on the legitimating power of consent) has consequently become more and more a way to determine the boundaries of a just society and to provide the guiding principles for evaluating and reforming actual societies.15 This normative function is therefore to be seen as a second characteristic trait of contractualism.16 In her account of joint commitment, Gilbert, like the contractualist, relies on the supposedly analytic relationship between a declaration of intent and an obligation. Since on her account the manifestation of willingness by the members of a social group is an actual one, however, her account lacks the normative features of contractualist theories operating on merely hypothetical agreements. Because the members of the group actually consent to engage in a joint action, there is in some sense (although this could be questioned) no need for a further specification of how social groups should be structured in order to fulfil some additional normative criteria (e.g. to be good, just or decent) and therefore legitimately to call 14 To be accurate, Hobbes’ theory already has normative traits, though in a very rudimentary and limited manner. For a detailed reconstruction of the contractualist tradition cf. Kersting (1994). 15 This constructive function of the contractualist argument is made explicit by Kant and systematically exploited by Rawls. 16 That this undertaking is also not free from paradoxical traits is quite evident. For the consent that is supposed to display the principles as just is not an agreement among full-blooded individuals but among idealized persons who are in some sense indistinguishable from one another. In gaining the principles of society by abstracting from the conditions of actual sociality, like the plurality of standpoints, the theory cannot fully grasp the traits of the political practice it is in fact concerned with.

Joint Commitment and the Practice of Democracy

291

for recognition. So far then, Gilbert’s theory of social groups can with some restriction be said to stay in a contractualist tradition. How about her political theory?

2. Gilbert’s theory of “Political Obligation” At first glance, Gilbert’s political theory seems to echo quite closely her theory of social groups. To begin with, the lack of the normative determinations just mentioned is characteristic also for Gilbert’s theory of political obligation. The opening question of Gilbert’s book looks prima facie like the classical problem of contractualism: “Does membership in a political society obligate one to uphold the political institutions of that society? For short: Are there political obligations?”17 As becomes clearer in the course of the subsequent investigation, however, Gilbert does not tie the obligating character of political institutions to their fulfilling certain specific functions, such as their guaranteeing justice or freedom or simply benefit and safety. According to her, the obligating character of political institutions results from the fact of political membership itself. It is thus the first part of the question quoted above that displays Gilbert’s real starting point. The crucial question of the book is: “must one obey the commands of one’s country simply because it is one’s country?”18 And Gilbert’s answer to this question is unequivocal: ‘Yes!’ Gilbert’s affirmative answer reflects her acceptance of a version of the so called “analytic membership argument”:19 “To say that the members of a political society are obligated to obey its laws is to say something that is true by virtue of the meanings of the terms involved. […] Given the meanings of its terms, it cannot be refuted by an appeal to experience. So it makes no sense to doubt it.”20 Gilbert maintains that this “analytic membership argument” does not entail a naturalistic fallacy. It does not derive the “ought” of the political obligation, from the “is” of ‘natural’ belonging to a political collective.21 So she seems to argue, contrary to a crude legal positivism, that obligation does not simply flow out of the fact of community membership but, on the contrary, has to do with specific attitudes of the members of society. 17 18 19 20 21

Gilbert (2006: 1). Gilbert (2006: 12). Gilbert (2006: 7). Gilbert (2006: 7). Gilbert (2006: 9).

292

Francesca Raimondi

Given her account of social groups, one is therefore inclined to think that this specific attitude must be a form of intentionality or consent. In fact, Gilbert answers the question of what exactly political membership means, and why it is essentially related to obligations, by adapting her concept of social groups, modelled on a small cluster of persons, to the case of political collectives.22 Thus the normativity of political societies is explained as a case group-normativity resulting from the fact of doing things together. In the same way, the structure of political obligation is supposed to be based on a joint commitment of the members of the political society. Political societies are thus to be seen as “plural subjects” and are supposed to act “as a body”. In order to show that her theory of joint commitment is capable of accounting for political collectives, Gilbert adds a couple of further specifications and distinctions to her original version. First, she stresses that a joint commitment does not necessitate that the members of a group ‘intimately’ identify with the common purpose. The reasons for entering a joint commitment do not need to be the same for all the members, and they may also be mere prudential reasons, which do not express their ‘private’ beliefs. The fulfillment of one’s duties constitutes acceptance of a commitment, no matter what reasons one may have for doing so: “It is worth emphasizing that a joint commitment to believe that such-andsuch as a body does not – as I understand it – require the parties to the commitment personally to believe anything. The commitment is, after all, together to constitute, so far as possible, a single body that believes that such-and-such. None of the individuals is that body. It is reasonable, then, to deny that their personal beliefs are in question.”23 The commitment to a democratic society for example does not imply that the members of that society privately must hold democracies to be the best form of governance. Furthermore, a joint commitment leaves open whether, and to what extent, a direct involvement of the members is needed to pursue the common purpose. An action is still a joint action even if some members delegate the relevant decisions to other members of the group. The group 22 According to Gilbert, countries are the primary political collectives. As a specific form of “political society”, countries are characterized as possessing a “relatively definite, relative permanent location”, as “including within [them] other, smaller societies” and as being such that, within them, “[their] members can live ‘whole lives’” (Gilbert 2006: 16-17). “Political societies,” in turn, are defined as societies with “political institutions”. Finally, societies are for Gilbert a sort of “social group”. 23 Gilbert (2006: 137).

Joint Commitment and the Practice of Democracy

293

members may play different roles and stand in hierarchical relationships; they can also stand in perfectly impersonal or anonymous relationship to one another and still count as acting together. Finally, a joint commitment does not necessarily imply fixed common goals. This is especially the case when the group members have agreed on some procedures of decisionmaking and thus have empowered these procedures to determine what the group’s goals will be. Although Gilbert’s concept of joint commitment is flexible enough to cover a wide range of cases, the adaptation of the model of social groups to political collectives becomes overtly problematic when we consider the formation of a joint commitment itself. For Gilbert also insists on some sort of readiness as the condition for membership in political collectives: “To enter a joint commitment, the parties must express their readiness to be jointly committed with certain others. I take this to imply that entry into a joint commitment is at some level intentional. That is not to say that it must be a matter of deliberation or forethought, and, though its implications must be understood, they need not be consciously noted or dwelt on at the time.”24 But what exactly does “readiness” mean in this context? At first, Gilbert gives a negative characterisation of this notion by way of rejecting “actual contract theories” which found political collectives on an actual “agreement to uphold the relevant set of political institutions”.25 Gilbert argues plausibly that such an agreement between the members of a society never actually takes place. The project of grounding societies in an agreement furthermore proves to be a circular undertaking: “It is plausible to suppose that making an agreement with another person is itself a case of acting together.”26 Gilbert’s “readiness” must therefore be less sophisticated and ‘social’ than such agreements. This holds insofar as the expression of readiness requires neither mutuality nor an accord about the object of agreement. Furthermore, it can also be exhibited in the case of practices that are already instituted and which the individual simply “falls into”. But whereas Gilbert’s critique of “actual contract theories” is plausible, her notion of readiness still remains unclear. When and how do the members of a political community express their readiness? And if the moment of mutuality is in some sense missing, how could it be justifiable to conceive of political collectives as analogous to social groups? The analogy seems to hold only insofar as we are entitled to assume that everyone who 24 Gilbert (2006: 168). 25 Gilbert (2006: 56) 26 Gilbert (2006: 118).

294

Francesca Raimondi

enters a political community (by birth or by immigration) also wants or is “ready” to do so. But how can such a claim be theoretically justified? If readiness is supposed to express a decision of the members (which according to contractualist thinking grounds the obligation of their members), then it cannot be taken for granted that everyone will make such a decision. If, by contrast, “readiness” does not stand for a decision of the members, it remains unclear what this term means and how (at least outside of a contractualist perspective) it could be a source of obligation. As it turns out, Gilbert does not seem to conceive of readiness as a decision, for she distinguishes readiness not only from agreement but also from voluntariness. This becomes apparent when she claims that even the coercion to readiness can still give rise to obligations.27 Moreover, she takes it for granted that political obligations obtain independently from their being just or unjust, and this suggests that for Gilbert political membership does not rest on an act of the will but is rather a sort of getting involved in a social practice, which can impose obligations one was not ‘ready’ to have. With this move Gilbert denies the moment of individual willingness, which is, however, central to her notion of a social group and a joint commitment. Although Gilbert explains the meaning of membership in terms of a theory of action, i.e. in terms of getting involved in a shared practice, the commitment qua membership seems to hold independently of the member’s own intentional attitudes. In the case of political collectives, the source of obligation seems to be participation in the practice as such, which is (or is not, as e.g. for refugees) already there whether or not one chooses it. This shift one can detect within Gilbert’s political theory is puzzling, for it has significant consequences for the whole account that Gilbert does not seem to recognize. If the membership in a political community does not rest on a decision, the joint commitment underlying it cannot be a commitment of the will – at least not in the sense that it is an act of the will, which brings about such a commitment. Instead one must say that the norms and institutions of a political society apply before the individual is able to perform acts of the will, so that individual intentions cannot be seen as the source of obligation, at least in the first instance. If these arguments are sound, then Gilbert’s attempt to explain political obligation by way of her concept of joint action becomes circular in so far as joint action in the political sphere presupposes political institutions and norms. Ac27 Gilbert (1993; 2006: 75 f.) Gilbert thereby takes a rather controversial position on the much-debated question of the binding force of coercive agreements, a question discussed by J. Simmons and M. Walzer, among others.

Joint Commitment and the Practice of Democracy

295

cordingly, joint action on the level of political obligation does not follow from a readiness or willingness of already constituted persons. To acknowledge this, however, ultimately results in rejecting the “weak individualism” that I have emphasized in Gilbert’s theory of social groups – although Gilbert does not explicitly acknowledge this consequence.28 For in the case of basic social norms – in which category I would also place political obligations – one cannot rely on intentional beings that are already capable of deciding about the norms they want to embrace; rather the converse is the case. The concept of person is not understandable independently of a collective and social sphere into which, as Heidegger puts it, the single individuals are “thrown”. So Gilbert’s picture of individuals binding themselves by some sort of willingness or readiness to a joint commitment cannot account for political collectives and the way persons are situated in them. The puzzle about Gilbert’s theory then, results from her rejection of intentionality and decision as the condition of membership on the one hand, and her attempt to continue to use joint action as an explanation for political obligation on the other.29 The rejection of willingness as a condition on membership thus changes the sense of obligation, since political (or social) obligation is not something one first has an external standpoint on and then (in some way or other) decides to embrace. This does not mean that the members of a political society cannot challenge their norms or institutions. On the contrary: as I will argue in what follows, it is precisely because there is no original agreement on such norms that the subjects living under them can begin to perceive (some of) them as unacceptable. Just as obligation has another sense for political collectives than it does in the case of going for a walk together, so also do “unity” and the “consciousness of unity” – which, besides “intentionality”, are the characteristic features of social groups. Since there is no such thing as an individual decision and its mutual acknowledgement at the origin of the political society, the assumption that a shared attitude towards norms can function as a source of unity fails to apply.30 As a consequence, conceptual tensions 28 I am therefore not at all sure if Gilbert would agree to the Heideggerian (or also Wittgensteinian) conclusions I draw in the following. 29 Schmid (2001: 89) calls this “an ambivalence” within Gilbert’s theory. In fact, Gilbert’s Theory of Obligation apparently bears traces of her earlier “weak analytic individualism”, while at the same time providing an account of membership that goes beyond the scope of her earlier works. 30 One can argue that, in the light of these remarks, a simple intentionalist account cannot hold for small social groups either, since they constitute themselves also on the basis of norms that are to be explained in a non-intentionalist manner. But

296

Francesca Raimondi

arise with respect to the notion of collectivity that Gilbert introduced at the outset of her theory.31 Since I assume that her underlying assumptions regarding collectivity and social unity are crucial for the question of how to conceive of political obligation, I will focus my considerations in the rest of the paper on this topic. So, let me go a bit more into details.

3. Two concepts of social unity If we reject an intentionalist picture of some sort of a pre-social agreement constituting a political collective, we are left needing some other elucidation of collectivity. In Gilbert’s account one can detect traces of two classical positions in political philosophy concerning the issue of social unity, namely those of Hobbes’ Leviathan and of Rousseau’s Social Contract. In Hobbes, the unity of the political collective is the product of the sovereign power. The unity of the commonwealth is guaranteed by means of laws and sanctions, which as such cannot be said to constitute a “substantial unity” among the subjects but only an institutional one. In order to maintain the unity of the commonwealth, the sovereign power must be absolute, i.e. there can be no sources of authority other than the particular person or group holding the place of the sovereign. Rousseau follows Hobbes in holding that the sovereign requires unconditional obedience in order for a political and social order to be possible. But in contrast to him, Rousseau combines the subjection of the “bourgeois” with his partaking in sovereignty as a “citoyen”. In Rousseau’s view only such a sovereign power is able to provide a social unity whose acts and institutions are conform to the “general will” of the members. Though this does not lead Rousseau to endorse democracy as political form, it nevertheless requires him to advocate democratic features such as “assemblies”32 of the soverI cannot further pursue the topic in this paper. For a challenge of the intentionalist assumptions of theories of collective intentionality, including that of Gilbert, cf. Baier (1997). 31 It would be worth investigating whether “collective intentionality” as a framework for the explanation of social contexts of action does not as such imply a rather specific and therefore not generalizable picture of collectivity. In the case of Gilbert’s account of joint commitment this seems to hold, but assessing to what extent this flows out of the general approach of collective intentionality would require a detailed and wide-ranging analysis that I cannot provide in this paper. Schmid (2006), however, emphasizes that the examples one can find in theories of collective intentionality quite across the board present only harmonious forms of joint actions. 32 Rousseau (1983: Book III, Ch. 12).

Joint Commitment and the Practice of Democracy

297

eign people as a means to enforce the “general will”. So whereas for both the unity of the political body originates from a sovereign act and is therefore not the result of individual intentions, Rousseau nevertheless tries to link the norms of society back to the will of its members. Gilbert seems to argue in accordance with Hobbes when she maintains that the forms of political collectives (democracies as well as monarchies, just as well as unjust ones) are all obligatory in the same way. According to her, political institutions are binding independently from their being just or unjust,33 since their validity flows out of the simple fact of membership. Following Gilbert and Hobbes, one should rather say that the question of the legitimacy of political obligations is not a relevant question, at least from the standpoint of political theory, since political obligations hold for anyone who is a member of society, no matter how repressive or discriminating that society is.34 Beneath this Hobbesian line of argument in Gilbert’s account, there is nevertheless also a manifest link to Rousseau’s political thought. For instance when Gilbert (suddenly) claims that “one interesting aspect of this conception [of joint commitment] is that it helps to explain the sense one might have that the source of political authority in every case lies with ‘the people’. Here a joint commitment of the whole population in question – the people – is taken to underlie whichever kind of rule is in place, whether governing rules, personal rule, or rules of governance.”35 This passage is astonishing, because if it is evident that the rules of a given political society, be they just or unjust, need to be practically undertaken by its members in order for the society to function then this by itself can hardly imply that political authority lies with ‘the people’. For this to be so, ‘the people’ must have some possibility of shaping the institutions of 33 Gilbert takes into account, though, that in some cases it may be necessary not to conform to unjust institutions. The source of such an obligation to resistance, however, is not commitment to the political collective, but moral duty. 34 The question whether there is a naturalistic fallacy in Gilbert’s theory is thus not as easy to dismiss as she suggests. For Hobbes it is clear that the subjects are obligated towards the sovereign and the law only as long as he guarantees their survival, and that they cease to be obligated when the sovereign fails to do so. Gilbert, like Hobbes, ties obligation to membership, but does not give an account of how obligation is motivated, and hence of how, in cases of failure, it might count as unmotivated or invalidated. Since it is not clear why this feature is missing from Gilbert’s account, the question about the fallacy must to some extent remain an open one. If Gilbert were actually arguing for the validity of political obligations no matter how unjust, then the threat of a naturalistic fallacy would indeed suggest itself. 35 Gilbert (2006: 213).

298

Francesca Raimondi

the political community. But this – although it can perhaps also apply in monarchical systems where the king tries to do justice to the will of his subjects – is certainly not a feature of all political systems whatsoever. Tyrannical power does not rest on the consent of the subjects. Given the Hobbesian traits of her account, it is difficult to see how Gilbert can at the same time speak in Rousseauian terms about political collectives. Yet she also maintains that her concept of joint commitment explains why the members of a society can call the institutions their “own”, as if they were the result of a collective action. But again, not every institutional order fits this description. To ground such rhetoric, Gilbert ought to understand her account as a normative one, specifying under which conditions membership in a political community can count as a case of joint activity. This would however contrast with Gilbert’s “analytical membership argument”, since not all actual political institutions would then count as unconditionally binding. That Gilbert does not explicitly shape her theory in this direction is somewhat surprising, for many features of the theory suggest exactly this. The general line of argument of Gilbert’s Theory of Obligation, for instance, can also be read as an attempt to account for an internal relationship between the concept of interpersonal actions, on the one hand, and the concept of obligation (and with it presumably also the concept of law), on the other. Personal rights and obligations are neither based on a given natural law, nor are they simply the outcome of the actual values in a society. According to Gilbert’s approach, they are rather institutions essentially tied to various forms of being and acting together, and they ascribe specific statuses to persons as participants in social practices. To connect the concepts of rights and obligations with contexts of actions is an interesting idea, since it suggests that these statuses might share the transformability proper to practices. Furthermore, this idea situates them in a space “among” people – which means that it rejects the view that rights are the exclusive property of specific members of society. This basically ‘inclusive’ sense of membership36 together with the focus on practical transformation can be read as attesting to a modern view of politics and can be turned against Gilbert’s recurring static Hobbesian understanding of obligation. It is furthermore also important to notice that her account of social practice can be read as having an underlying egalitarian basis, for being a member of a 36 In Gilbert’s conception of practice anybody who belongs to a society as part of its collective practice can be counted as a member of it. This contrasts with an institutional sense of membership, which is structurally exclusive since it binds the fact of membership to some specific features of persons – e.g. income, descent, nationality, and gender.

Joint Commitment and the Practice of Democracy

299

social group not only constitutes obligations according to social rules, but also to some extent the right of being a ruler in the practice.37 Seen from this angle, however, Gilbert’s separation of collective (political) obligation and private (moral) freedom can no longer be maintained in her original terms. This does not mean that morality should enter into politics. The ground for rejecting Gilbert’s division is rather that the act of criticizing or rejecting some current rules of society as unjust is in a sense a genuinely political act. So, in my view, Gilbert’s account of joint commitment is anchored in a modern and egalitarian view of politics, although – and this seems to me precarious – her attempt is to provide a generic account of political obligation independently from the specific form a political society might in fact have. I’m not convinced that such an attempt can succeed, since there is a connection between the institutional forms of a society (beginning with its governmental and legal structure) and the forms of collective actions possible in a society, as well as the (implicit) understanding of the value that acting together has.38 Provided that these arguments are sound, and that Gilbert’s account based on joint activity refers to democratic features of societies, one can ask if a Rousseauian conception of social unity, based on the notion of a “general will,” would provide the suitable frame for a democratic understanding of political collectivity. In view of Gilbert’s rejection of the contractualist picture and the correlative idea of a grounding consent, I think it won’t. In the last section of the paper I will sketch a motivation for this claim.

37 Actually this is acknowledged by the two ‘criteria’ for social groups, namely the “obligation criterion” and the “concurrence criterion” (cf. Gilbert 2006: 103 f.). Yet Gilbert takes them to apply also e.g. in monarchical systems. But I doubt that one can really speak of ‘concurrence’ there and that on the background of a genuine monarchical understanding of politics (say that of Henry VII.) political societies would have been described in these terms. 38 This is one of the basic insights of Claude Lefort’s political theory. He argues that the political order of a society is not only a question of how the overtly political institutions are shaped – in short, what form of government a society has. Rather, the development of democratic societies implies a change in the form of society as such – in the way the social practice is carried out (cf. Lefort 1988).

300

Francesca Raimondi

4. Democracy and the indeterminacy of collectivity Rousseau conceives of the political collective as a “body politic” whose unity (under the condition of legitimacy) is formed by a “general will”. Unlike Hobbes, for whom the unity of the body politic is the effect of a sovereign act imposing sanctions upon the subjects,39 Rousseau conceives of the political collective as being (potentially) held together by a common intent of the members. Being part of a social practice means for Rousseau to have a trans-individual perspective and therefore access to the domain of generality. He conceives of this generality as an intentional generality, i.e. as a fundamental accord over the public goods that are to be collectively pursued, an accord, which anybody can in principle acknowledge provided that he sets aside the egoistic perspective of individual advantage. Yet for Rousseau there are neither criteria nor guarantees that the “general will” is in fact achieved by the practice of ruling. But despite these complications, Rousseau holds onto the notions of a general will and of a body politic as the proper concepts to account for a (good) political society that deserves the title of an “association” and is not a mere “aggregate”. It is therefore this unity of intent that provides the unity of the “body politic” and grounds the obligating power of its institutions as the expressions of the “general will”. Gilbert’s notion of joint commitment and of a plural subject recalls, at least in formal respects, a Rousseauian approach to political collectivity. Also Gilbert’s “plural subject” is conceived as a unity of will, so that her account, like Rousseau’s, rests on the assumption of a consensus of intent as the glue of the social and political order – though both reject the idea that entrance in the social order is based on some sort of agreement. On these grounds Gilbert, like Rousseau, depicts the political practice as belonging to an irreducible “we” (that refers to the population of a particular nation), and operates with the body-metaphor in order to elucidate the relation between this “we” and the individual “I’s” constituting it. However, unlike either Rousseau, who operates with a ‘thick’ concept of reason, or Communitarians who hold culture as providing such a consensus of the will, Gilbert does not give any argument for the assumption of a unity of wills underlying the joint activity of a political society. 39 For Hobbes social unity can therefore apply only within the domain of sanctions, namely that of overt behaviour: it does not reach the ‘private beliefs’ of the subjects.

Joint Commitment and the Practice of Democracy

301

Due to the impossibility of an original act of agreement that would settle a relation of consonance and transparency between the wills of the members and the political institutions, however, a certain scepticism towards a political collectivity which conceives of itself as a substantial “we” and calls the political institutions its “own” seems appropriate. As we saw, Gilbert does not necessarily assume – as Rousseau does – that the consensus lying at the core of her notion of plural subject is a consensus about the member’s fundamental beliefs. The members of a political community can feel themselves committed to follow their rules and at the same time undertake this only out of prudential, pragmatic or other otherwise detached stance. But, one can add, just as members can be not fully satisfied about the shape of the political institutions, they can also find them obviously bad, unjust and in need of change. Entering a collective practice cannot simply be equated with becoming an agent who acts according to social rules, for such an entry at the same time puts the agent in a position to take up a critical stance toward these rules. To be a member of a collective governed by certain rules means to be capable of criticizing their legitimacy. The impossibility of an original consensus, then, implies also the impossibility of a consensus flowing out of the institutions or practices of a given society. This holds – and this might prima facie seem paradoxical – in particular from a democratic point of view. Paradoxical, because in making politics to a certain extent a collective practice, it might seem that democracies have a specially good claim to represent their institutions as ‘one’s own’. At least, it is a peculiar trait of democracies to strive for a political order that meets the approval of its members and is not merely enforced by authority. So the democratic “we” is one that is not only constituted by a given political order, but should in some sense also be constitutive for it. Seen from another angle, however, to represent politics as a collective practice displays the assumption that there are neither individuals nor groups that are entitled to achieve the “general will” and decide for all. Or to put it more directly, democracies go along with the assumption that there is no such thing as a “general will” that has a reality beyond or independent of the political process. Since it is impossible that the wills of all members enter the institution of a society,40 however, also democratic transformation cannot be thought as providing such a substantial collective unity. 40 This obtains also in cases where a new political collective gets constituted. As Derrida has shown in his reading of declarations of independence (cf. Derrida 1986), such a constituting act is in itself paradoxical, since the subject who signs it and therefore is supposed to provide its legitimacy – the people – is in fact first constituted by the declaration itself.

302

Francesca Raimondi

If ‘democracy’ nevertheless is the name for the ongoing claim of a political order coming from ‘the people’ (and that is to say from each member), then the assumption of an already constituted “we” actually works against the possibility of democracy. Such a substantive “we” can in fact only refer to a (hegemonic) part of the society, excluding other parts of it from the collective quest for the common good. And along the same lines, its intentions or goals can only be the ones belonging to a group or to a specific understanding of what the common good consists in. If the actuality of a democratic “we” and its intentions are already assumed, then politics turns out to be the execution of a predetermined script and loses the trait of being a transformative practice. Precisely this shows that for democracy ‘the people’ does not at all refer to a collective subject with one will and a substantial unity, but has rather the character of a claim for an inclusive order that at the same time cannot be fully achieved and has to remain controversial. Claude Lefort has therefore argued that democracy is only at work where the locus of power is instituted as “an empty place” that “cannot be occupied”,41 not even by ‘the people’, if by this one understands the image of an achieved totality (which would always be something particular like ‘the reasonable men’, ‘the Italians’, or whatever). The “invention démocratique” is more than the decapitation of the king and the establishment of an acephalic, autonomous body politic: “democratic society is instituted as a society without a body, as a society which undermines the representation of an organic totality”.42 In maintaining the claim for an inclusive political order and in making equality its leading principle, democracy breaks with an organic conception of society, where every part finds its teleological place in the whole. It institutes a form of generality that no longer has the positive traits of a “general will” or of an overall consent of individual persons. As an empty place, it is rather a negative mark, displaying the impossibility of thinking of an achievable form of generality. Such an understanding of democracy leads to a different picture of collective practice than the one underlying Gilbert’s account. In virtue of the constitutive plurality of democratic societies and their ‘devoided’ generality, it becomes clear that collective practice cannot be simply understood as a “pooling of wills”. Against the background of a democratic practice, acting together also means to criticize, to transform, or even not to fulfil those requirements of a given political order that appear to be unsustaina-

41 Lefort (1988: 17). 42 Lefort (1988: 18).

Joint Commitment and the Practice of Democracy

303

ble.43 Such actions, however, do not (at least in some cases) simply reform the given political order, in the sense that they provide a better application of the existing rules. Insofar as they enforce modes of being and acting that hitherto were not considered in the social practice, and thereby generate new collective significances, they rely precisely on the ‘void’ generality of democracy. The claims pertaining to this form of critique are thus controversial in a peculiar sense, since they are not settled on an established common ground, but challenge a given consensus and display its unfoundedness.44 Since they do not act in the light of the instituted order, but present new understandings, they can neither be ascribed to a “plural subject”, nor are they “joint actions” in Gilbert’s sense. In discarding an established ground for jointness, they paradoxically claim collective acceptance in disrupting its unity and in highlighting its contestedness. Understanding the democratic practice as a collective one implies acknowledging the irreducibility of conflicts and contingency in the institution and definition of the common, in contrast to the picture of a harmonious concurrence of wills. Commonality and contestedness both belong to democracy as inseparable features of its mode of sociality. If one can understand democracy, then one has to understand the coexistence of these two sides. This requires dismissing a Hobbesian concept of obligation as well as breaking with the assumption that collective actions are always to be understood as the realization of a “we” somehow constituted by a common project. To conclude, my attempt has been to show how Gilbert’s account based on the phenomenon of acting together is in fact an interesting approach for political theory. In order to make it suitable for a democratic concept of political practice and to deploy the features I have pointed to in the end of section 3, I have however challenged some aspects of her notion of joint action, such as the concept of a “plural subject” and the formula of acting “as a body” as implying a problematic understanding of social unity. In arguing that the concept of an acting person is not graspable outside the social dimension of human existence, I have implicitly shifted the challenge posed by joint actions from the question of how to account for individual persons joining one another in action to the question of how to 43 Schmid (2006) argues in a similar way in alluding to the collective dimension of dissidence. 44 This applies already to cases that apparently only claim to extend some rights to previously excluded groups. To extend the right to vote to women or to non-nationals changes the reasons for conferring such a right and therefore the concept of the right to vote itself.

304

Francesca Raimondi

inscribe plurality in the notion of collective acting, and therefore in some way to ‘politicize’ the concept of joint action.

References Baier (1997): Baier, A. Doing Things With Others: The Mental Commons, in: Alanen, L. et. al. (eds.), Commonality and Particularity in Ethics, Basingstoke: Macmillian Press, 1997, pp.15-44. Derrida (1986): Derrida, J. Declarations of Independence, in: New Political Science (1986) 15: 7-15. Esposito (2004): Esposito, R. Communitas, diaphanes: Berlin, 2004. Gilbert (1989): Gilbert, M. On Social Facts, London: Routledge, 1989. Gilbert (1993): Gilbert, M. Agreement, Coercion, and Obligation, in: Ethics (1993) 103: 679-706. Gilbert (1994): Gilbert, M. Walking Together: A Paradigmatic Social Phenomenon, in: Gilbert, M. Living Together: Rationality, Sociality, and Obligation, Lanham: Rowman and Littlefield, 1996, pp. 177-194. Gilbert (2006): Gilbert, M. A Theory of Political Obligation. Membership, Commitment, and the Bonds of Society, Oxford: Clarendon Press, 2006. Hobbes (1996): Hobbes, T. Leviathan, Cambridge: Cambridge University Press, 1996. Kersting (1994): Kersting, W. Die politische Philosophie des Gesellschaftsvertrags, Darmstadt: Wissenschaftliche Buchgesellschaft, 1994. Lefort (1988): Lefort, C. Democracy and Political Theory, Minneapolis: University of Minnesota Press, 1988. Rousseau (1983): Rousseau, J.-J. On the Social Contract and Discourses, Indianapolis: Hackett Publishing Company, 1983. Schmid (2001): Schmid, H. B. Wir-Intentionalität – Jenseits von Individualismus und Kollektivismus, in: Dialektik (2001) 1: 71-92. Searle (1990): Searle. J. R. Collective Intentions and Actions, in: Cohen, P. R. et. al. (eds.), Intentions in Communication, Cambridge/MA: The MIT Press, 1990, pp. 401-415.

305

About the Authors Facundo Alonso, M.A., is Ph.D. candidate of philosophy at Stanford University, California. Sondra Bacharach, Ph.D., is senior lecturer at Victoria University, Wellington, New Zealand. Monika Betzler, Prof. Dr., is professor of philosophy at the University of Berne, Switzerland. Clotilde Calabi, Ph.D., is associate professor of philosophy of language and philosophy of mind at the Department of Philosophy of the University of Milano, Italy. Frank Hindriks, Ph.D., is assistant professor at the Department of Ethics of the University of Groningen, The Netherlands. Jennifer Hudin, Ph.D., is lecturer in the Department of Cognitive Science at U.C. Berkeley, California, currently pursuing research on the Orbital Frontal Cortex at the Aging and Memory Center, at the University of California, San Francisco. Björn Petersson, Ph.D., is associate professor at the Department of Philosophy, Lund University, Sweden. Francesca Raimondi, M.A. phil., is doctoral fellow in the graduate program “Lebensformen und Lebenswissen” at the University of Potsdam and at the European University Viadrina of Frankfurt/Oder. Antti Saaristo, Ph.D., is development manager of the HSE research network at the Helsinki School of Economics, Finland. David Schweikard, M.A. phil., is researcher at the Department of Philosophy and at the Institute for Ethics in the Life Sciences at the University of Cologne, Germany. Deborah Tollefsen, Prof. Dr., is assistant professor at the Department of Philosophy of the University of Memphis, Tennessee. Raimo Tuomela, Prof. Dr., is professor of philosophy at the Department of Social and Moral Philosophy of the University of Helsinki, Finland.

306

About the Editors Nikos Psarros, Prof. Dr., is extraordinary professor of philosophy at the University of Leipzig, Germany. Hans Bernhard Schmid, Prof. Dr., is SNF-professor of philosophy at the University of Basle, Switzerland. Katinka Schulte-Ostermann, M.A. phil., is PhD-researcher at the Department of Philosophy of the University of Essen, Germany.

PhilosophischeAnalyse PhilosophicalAnalysis 1 Herbert Hochberg Russell, Moore and Wittgenstein The Revival of Realism

8 Rafael Hüntelmann Existenz und Modalität Eine Studie zur Analytischen Modalontologie

2

9 Andreas Bächli / Klaus Petrus Monism

ISBN 3-937202-00-5 334 pp., Hardcover € 94,00

Heinrich Ganthaler Das Recht auf Leben in der Medizin Eine moralphilosophische Untersuchung

ISBN 3-937202-01-3 167 pp., Hardcover € 58,00

3 Ludger Jansen Tun und Können Ein systematischer Kommentar zu Aristoteles’ Theorie der Vermögen im neunten Buch der „Metaphysik“ ISBN 3-937202-02-1 302 pp., Hardcover € 70,00

4 Manuel Bremer Der Sinn des Lebens Ein Beitrag zur Analytischen Religionsphilosophie ISBN 3-937202-03-X 134 pp., Hardcover € 58,00

5 Georg Peter Analytische Ästhetik Eine Untersuchung zu Nelson Goodman und zur literarischen Parodie ISBN 3-937202-04-8, 332 pp. Hardcover € 94,00

6 Wolfram Hinzen / Hans Rott Belief and Meaning Essays at the Interface ISBN 3-937202-05-6 250 pp., Hardcover € 58,00

7 Hans Günther Ruß Empirisches Wissen und Moralkonstruktion Eine Untersuchung zur Möglichkeit von Brückenprinzipien in der Natur- und Bioethik ISBN 3-937202-06-4 208 pp., Hardcover € 58,00

ISBN 3-937202-07-2 189 pp., Hardcover € 58,00

ISBN 3-937202-19-6 340 pp., Hardcover € 70,00

10 Maria Elisabeth Reicher Referenz, Quantifikation und ontologische Festlegung ISBN 3-937202-39-0 ca. 300 pp., Hardcover € 89,00

11 Herbert Hochberg / Kevin Mulligan Relations and Predicates ISBN 3-937202-51-X 250 pp., Hardcover € 74,00

12 L. Nathan Oaklander C. D. Broad's Ontology of Mind ISBN 3-937202-97-8 105 pp., Hardcover € 39,00

13 Uwe Meixner The Theory of Ontic Modalities ISBN 3-938793-11-2 374 pages, Hardcover,€ 79,00

14 Donald W. Mertz Realist Instance Ontology and its Logic ISBN 3-938793-33-3 252 pp., Hardcover, EUR 79,00

15 N. Psarros / K. Schulte-Ostermann (Eds.) Facets of Sociality ISBN 3-938793-39-2 370 pp., Hardcover, EUR 98,00

16 Markus Schrenk The Metaphysics of Ceteris Paribus Laws ISBN 13: 978-3-938793-42-8 192pp, Hardcover, EUR 79,00

EditedBy • HerbertHochberg • RafaelHüntelmann ChristianKanzian • RichardSchantz • ErwinTegtmeier

PhilosophischeAnalyse PhilosophicalAnalysis 17 Nicholas Rescher Interpreting Philosophy The Elements of Philosophical Hermeneutics ISBN 978-3-938793-44-2 190pp., Hardcover € 89,00

18 Jean-Maurice Monnoyer(Ed.) Metaphysics and Truthmakers ISBN 978-3-938793-32-9 337 pp., Hardcover € 98,00

19 Fred Wilson Acquaintance, Ontology, and Knowledge Collected Essays in Ontology

25 Laird Addis Mind: Ontology and Explanation Collected Papers 1981-2005 ISBN 978-3-938793-86-2 289pp., Hardcover, EUR 79,00

26 Hans Bernhard Schmid, Katinka Schulte-Ostermann, Nikos Psarros Concepts of Sharedness Essays on Collective Intentionality ISBN 978-3-938793-96-1 306pp., Hardcover, EUR 89,00

ISBN 978-3-938793-58-9 XX, 726., Hardcover, EUR 159,00

20 Laird Addis, Greg Jesson, and Erwin Tegtmeier (Eds.) Ontology and Analysis Essays and Recollections about Gustav Bergmann ISBN 978-3-938793-69-5 312 pp., Hardcover, EUR 98,00

21 Christian Kanzian (Ed.) Persistence ISBN 978-3-938793-74-9 198pp., Hardcover, EUR 79,00

22 Fred Wilson Body, Mind and Self in Hume’s Critical Realism ISBN 978-3-938793-79-4 512pp., Hardcover, EUR 139,00

23 Paul Weingartner Omniscience From a Logical Point of View ISBN 978-3-938793-81-7 188pp., Hardcover, EUR 79,00

24 Simone Gozzano, Francesco Orilia Tropes, Universals and the Philosophy of Mind Essays at the Boundary of Ontology and Philosophical Psychology ISBN 978-3-938793-83-1 196pp., Hardcover, EUR 69,00

EditedBy • HerbertHochberg • RafaelHüntelmann ChristianKanzian • RichardSchantz • ErwinTegtmeier