Natural and Artifactual Objects in Contemporary Metaphysics: Exercises in Analytic Ontology 9781350066328, 9781350066359, 9781350066335

Citation preview

Natural and Artifactual Objects in Contemporary Metaphysics

Also available from Bloomsbury Ontology and Metaontology, by Francesco Berto and Matteo Plabani God, Existence, and Fictional Objects, by John-Mark Miravalle Metaphysics, by Jonathan Tallant Indian Epistemology and Metaphysics, edited by Joerg Tuske Mereology, by Giorgio Lando

Natural and Artifactual Objects in Contemporary Metaphysics Exercises in Analytic Ontology

Edited by Richard Davies

BLOOMSBURY ACADEMIC Bloomsbur y Publishing Plc 50 Bedford Square, London, WC1B 3DP, UK 1385 Broadway, New York, NY 10018, USA BLOOMSBURY, BLOOMSBURY ACADEMIC and the Diana logo are trademarks of Bloomsbury Publishing Plc First published in Great Britain 2019 This paperback edition first published 2020 Copyright © Richard Davies, 2019 Richard Davies has asserted his right under the Copyright, Designs and Patents Act, 1988, to be identified as Editor of this work. Cover design: Irena Martinez Costa Cover image: Cabinet of curiosities (with glass doors) with various objects, by Andrea Dominico Remps (1621–1699). © De Agostini Picture Library / G. Nimatallah / Bridgeman Images All rights reser ved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage or retrieval system, without prior permission in writing from the publishers. Bloomsbur y Publishing Plc does not have any control over, or responsibility for, any third-party websites referred to or in this book. All internet addresses given in this book were correct at the time of going to press. The author and publisher regret any inconvenience caused if addresses have changed or sites have ceased to exist, but can accept no responsibility for any such changes. A catalogue record for this book is available from the British Librar y. A catalog record for this book is available from the Librar y of Congress. ISBN: HB: 978-1-3500-6632-8 P B: 978-1-3501-7543-3 ePDF: 978-1-3500-6633-5 eBook: 978-1-3500-6634-2 Typeset by Deanta Global Publishing Ser vices, Chennai, India To find out more about our authors and books visit www.bloomsbury.com and sign up for our newsletters.

Contents Carving Nature at the Joints Richard Davies

1

Part One Does Nature Carve Itself? 1 Mental Acts, Externalism and Fiat Objects: An Ockhamist Solution Riccardo Fedriga 2 Is the World Really a World of Objects? A Note on Quinean Ontology Antonio Rainone 3 Spatial Fictionalism. A Solution to the Grounding Problem Nicola Piras 4 Talking about Properties: A Couple of Doubts about Hofweber’s Internalist View Elisa Paganini

19 48 58 80

Part Two Where Do Limits Lie? 5 The Eye of the Needle: Seeing Holes Clotilde Calabi 6 Bona Fideness of Material Entities and Their Boundaries Lars Vogt 7 A Conceptualist View in the Metaphysics of Species Ciro De Florio and Aldo Frigerio

93 103 121

Part Three Where Do Tools Come From? 8 Artifacts and Fiat Objects: Two Families Apart? Massimiliano Carrara and Marzia Soavi 9 The Semantics of Artifactual Words Marco Santambrogio 10 Are Linguistic Objects Fiat or Bona Fide? An Ancient Proposal Maddalena Bonelli

141 156 181

Part Four What Does Mind-Dependency Depend On? 11 Leibniz’s Principle and Psycho-Neural Identity Andrea Bottani and Alfredo Paternoster

199

vi

Contents

12 Do We Exist? Mereological Nihilism, Collective Thinking and Dualism Alfredo Tomasetta Abstracts Index of Names and Principal Subjects

215 225 232

Carving Nature at the Joints Richard Davies

The theme on which the contributions to this volume are variations is one of the oldest and most intractable in the Western philosophical tradition. Not only can we make out strains of similar melodies in such different pre-Platonic philosophers as Parmenides and Democritus, but the improvisations to which the theme has given rise have often enough led to cacophonous outcomes (an evaluation that may be as fitting to atomism as it is to monism). A variety of more-or-less traditional distinctions are in play, from the ancient pairing of physis and nomos, of the natural as against the conventional, artificial, rulegoverned or law-abiding, to, more recently, the pairing of bona fide and fiat, but perhaps there is no entirely neutral way of getting at what is at issue, and I hope that my musical metaphor may be forgiven or at least rapidly forgotten. But one thing the metaphor of variations on a theme is meant to convey is a sense in which the chapters here presented come at a core of long-standing concerns in metaphysics from diverse angles and against various backdrops. My aim in this introduction is to illustrate some of the points of contact and of divergence among the approaches adopted and the theses advanced, and then to offer a few notes on what might appear to be the unexpected omission of the status of social objects from these discussions.

Finding the joints A useful point of departure is the oft-cited passage of Plato’s Phaedrus (265e) alluded to in my title and referred to by several of the present authors. Here, Socrates swiftly invokes an analogy between various ways of making discourses and various ways of dismembering a carcass. The account given of the various ways of making discourses raises very thorny questions in Plato exegesis and we

2

Natural and Artifactual Objects in Contemporary Metaphysics

shall not go into them here, but we may say that the point about cutting is fairly clear and depends on the difference between how a skilled carver will proceed and what a bungling butcher (kakos mageiros, 265e2–3) will let happen. Where the former will cut along the natural joints, the latter will end up smashing the parts. I have already cited what was already in Plato’s day the hoary antithesis between physis and nomos, but it may be worth pointing out that, in the passage in question, the key element in the phrase that is typically and, indeed, rightly rendered as ‘the natural joints’ (arthra he pephyken, 265e1–2), though it is certainly germane to physis, is most directly derived from the verb phyo, whose primary sense is to grow or to bring forth. In this sense, to carve skilfully is to cut at the points or interfaces around which the animal grows. There seem to be two aspects to the advantage of the skilled carver. One is that he expends less effort if he follows the articulations of the bones of the animal he is carving; it is easier to cut ligament or muscle than bone; this is an advantage of efficiency and it ensures that the knife will not be damaged.1 The other is that the resulting portions will correspond to the original segmentation of the living animal. The latter seems to be the advantage that Socrates is most interested in exploiting for the purposes of his analogy with making clear and consistent discourses: there ought to be a reciprocity between the process of dividing and that of unifying (diaeresis and synagoge, 266b). In some sense, the result of skilled carving will allow the body to be recomposed, while haphazard chopping will just leave a mash of pulp. Though we would be torturing Socrates’s analogy were we to ask what purpose there might be to recomposing a dead cow or pig, the distinction between the skilled carver and the bungling butcher depends on the notions not only (i) that a living body is internally differentiated by nature into bones, tissues, organs and so on but also (ii) that someone who acts on these differentiations has acquired knowledge that, in one way or another, goes beyond or perhaps behind mere acquaintance with the outward appearance of the animal. While Socrates employs the concrete case of cutting as an emblem for the making of discourses, we may exploit (i) and (ii) to raise some even more general questions. Corresponding to (i), there is a nexus of problems about how much sense it makes to talk about nature as a whole being articulated along some basic or primary contours such that all other ways of categorizing things will be either consequences of or at odds with that articulation and, hence, either superfluous or misleading. For instance, there is a standing temptation (again going back to Democritus) to say that the fundamental building blocks of what is are material minima, which we may choose to call ‘atoms’ or ‘elementary particles’ whatever

Carving Nature at the Joints

3

they may turn out to be. If we give in to this temptation, then we may be induced also to say that anything that can be explained in nature will be ultimately explicable in terms of the interactions of these minima. This way reductionism lies. If – again by way of merely gestural example – chemical interactions can be explained in terms of the interactions of physical minima, then the terminology of chemistry is, ultimately, dispensable in giving an account of how nature works, and its adoption is hardly more than a convenient shorthand for describing complexes of atomic processes. By contrast, if there are phenomena that are not reducible to the basic terms, then they may be treated either as merely subjective or as ultimately misleading. To take the classic case of colours, if there is no one-to-one correlation between, say, redness and any identifiable recurring configuration of atoms, then red things do not form a natural class in the way that, say, hydrogen atoms do; for this reason, our ascriptions of redness may say more about us than about the things to which it is ascribed. Conversely, though one can go quite a long way towards a plausible theory of fire and burning on the basis of the postulation of phlogiston, in the end, a way to carve nature up closer to the joints invokes oxygen. Just as not every way of generalizing (i) need commit itself to the starkly reductionist monism just outlined, so reasons for resisting (i) may come under many guises. Among the more interesting of these for our purposes are those that cast doubt on the universality or exclusivity of the knowledge that would be called for in generalizing (ii). This generalization would in some way or other relate to the claim that all genuine knowledge is knowledge of nature; but doubts about the justification of a universal or exclusive claim in this direction need not call on notions of the supernatural, the spiritual or the transcendent. For, even if an understanding of the underlying mechanics of physical processes provides a means for explaining many or even all natural phenomena, there remains room for manoeuvre in relating that understanding to other realms and aspects of our experience of and interaction with the world around us. To recur again to the case of colour, the reflectance of a given surface is a product of a certain atomic (and, so, chemical) structure, but that structure is not what we see when we see the redness of an apple in certain conditions of illumination. Even if, with Democritus, we say that the redness is a matter of ‘convention’ (nomos, for example, DK 68B9) or, with Locke, that it is a ‘secondary quality’ (Essay, II, viii), that is not to deny that it is something we see. Likewise in the case of artifacts: the fact that my corkscrew is of a certain shape and is made of metal is one among the material preconditions for its being able to do what it is for, but its having that function does not seem plausibly reducible to its physical configuration.

4

Natural and Artifactual Objects in Contemporary Metaphysics

As may have been emerging from what I have already said, the various ways of responding to the generalizations of (i) and (ii) are in some degree conditioned by how one conceives of the relations among pairings of terms, in each of which one term has to do with what holds in nature or is primary and fundamental. The expectation is that such terms will be one-placed, either because they are singular terms or because, when they stand for predicables, having a certain shape, mass, extension or whatever turns out to be the basic characteristics of subjects (call them ‘primary qualities’, if you like) they may be expected to be intrinsic properties of them. By contrast, the other term that figures in the pairings to which we have alluded will tend to be at least two-placed. Thus, for instance, being red may be understood as being such as to look red to certain creatures (non-Daltonic humans) in certain conditions of illumination, or being a corkscrew may be understood as being such as to enable the removal of corks in certain conditions. In an effort to articulate and coordinate the chapters that make up the volume, we propose that, at least ideally, they form groups that address the following four questions: ‘Does nature carve itself?’; ‘Where do limits lie?’; ‘Where do tools come from?’ and ‘What does mind dependence depend on?’ The first two questions are more germane to the generalization of (i), namely the issue of whether and to what extent it makes sense to think of the world as presenting differentiations that are, in one way or another, prior to our conceptions of it and that serve as guides to our perceptual and scientific elaborations of it. The second pair of questions elicits considerations of how the finalistic notion of ‘use’ interacts with that of the intentional and the mental. Proceeding, then, to consider whether nature carves itself, Riccardo Fedriga illustrates how, on William of Ockham’s mature account of mental acts, the formation of a judgement, for instance that this wall is white, involves the combination of two intuitions of singular objects, namely ‘this wall’ and ‘this white’, that are presented to us by way of a causal interaction. But talk of ‘causal interactions’ seems merely to name the problem of how we can ensure that the segments and features of the world that impinge on our senses correspond to the way things really divide up rather than merely to the ways that habit and convenience dictate. Though Fedriga is by trade a historian of medieval philosophy, his underlying aim is to bring out the relevance of the Ockhamist position for contemporary forms of anti-foundationalism, and specifically those that are diffident of there being such a thing as ‘the way things really divide up’. Yet, even if there is an irreducible plurality of ways in which we can divide things up, there remains some sense in which the causal efficacy of the segments and features of the world that figure in our judgements involve the meeting of the

Carving Nature at the Joints

5

organized matter of the wall and its reflectance with the organized matter of our perceptual apparatus. One approach that seeks to do justice both to the pluralism in question and to the need to recognize units of organized matter as fully fledged objects emerges from the account that Antonio Rainone gives of Quine’s ontological reflections. In the first instance, a certain caution dictates that we treat the terms used to refer both to the contents of everyday experience and to the outcomes of scientific enquiry as posits. That is to say, the items to which we are committed by the outlook of common sense and by more sophisticated and self-conscious procedures of investigation are ultimately relative to the theories in play and, hence, to the vocabularies adopted within those theories. The relativity of such schemes is not, however, entirely egalitarian, for it allows room for increase in explanatory power and generality with the adoption of evermore advanced scientific terminologies. And, as Rainone notes, Quine is also committed to a ‘robust realism’ according to which the posits of our currently best theories are the objects that, for the time being, we are justified in recognizing as such. Elaborating on related well-known Quinean themes, not least to do with Quine’s ‘taste for desert landscapes’, Nicola Piras presents and defends a strong version of the thought that ultimately we can do without the idea of a spatial object. The presentation of this thesis is couched not only in part in terms of the formalism of a standard mereology but also in good measure in a sustained and many-sided dialectic with the principal views that have been being developed in recent years regarding the interrelations between matter and the things of which it is the matter. On Piras’s view, the fundamental relation should be seen as one that stands between matter (or ‘stuff ’) and the regions that it occupies. Since he takes the notion of a region to be that of a bounded space, and boundaries to be arbitrary or, at best, fixed by some fiction, his defence of the doctrine he calls Spatial Fictionalism means that what drops out of the story is what we might think of as the naïve – or, some might say, commonsensical – view of physical continuants as preserving their integrity not only through change of place but also through change of parts and growth and shrinkage. By contrast, the position of Thomas Hofweber, examined in Elisa Paganini’s chapter, proposes that the notion of reference is quite sufficient to fix objects as extra-linguistic realities, perhaps conceived of as what an older terminology would have called ‘bare particulars’. But, because Hofweber finds that the attribution of properties does not appeal in an appropriate way to the notion of reference, he maintains that predicable terms do not stand for properties but rather should

6

Natural and Artifactual Objects in Contemporary Metaphysics

be regarded as dependent only on our linguistic competence, which is a matter of convention. Paganini raises several doubts about this brand of conceptual idealism, which is not exclusive to Hofweber though he is a leading and lucid exponent of it. For one thing, the notion of reference in play appears to be less than even-handed in its exclusion of properties as individuable independently of language, when, after all, language-users themselves seem to have the property of being such, that is, the property of being language-users, which does not seem language-dependent. Again, it is less than clear how the contours of the objects that Hofweber and other idealists invoke can be determined without attributing properties such as shape (or other ‘primary qualities’) to them. Piras is avowedly fictionalist about spatial boundaries and Hofweber is at least implicitly committed to an internalist and language-dependent account of the outlines of objects. In both cases, it might be noticed that there is a high level of abstraction in the determination to speak generically of ‘things’ and ‘properties’. And it may be no surprise that, at this level of generality, it is indeed hard to find criteria of individuation that derive from the way the world is or the way we are bound to encounter it. For this reason, the mitigations of the nominalist urge to which Ockham and Quine give voice might also be expressed by saying that, because they are not themselves count nouns but proxies for categories, the vocabulary of ‘thing’ and ‘property’ is destined to produce a sort of indeterminacy about what things and properties there ultimately are. To put it another way, while it makes sense to ask how many tables or how many chairs there are in this room, a question about how many things there are in this room gives no indication about how to proceed in giving an answer. Likewise, one might ask how many red things there are in the room, but not how many properties are instantiated. Perhaps it is in part for this sort of reason that, of late, there has been some resurgence in the literature of a readiness to look more closely at the identity and persistence conditions of various branches, groupings and sorts of things.2 The chapters that address our second question, of where limits lie, may be regarded as contributing to this more piecemeal examination of how lowerlevel categories should be characterized. Clotilde Calabi’s chapter examines a particular instance of the border between an object and its surroundings. A hole is a configuration of part of the surface of a material object. So far forth, holes are not themselves material objects; but from this it does not obviously follow that we do not see holes in much the way that we see material objects. Yet there is a widespread inclination to say that, insofar as the delimitation of a hole is highly relativizable, seeing one is rather like seeing an abstract entity like the line that is the equator, and this sort of

Carving Nature at the Joints

7

seeing should be thought of as indirect. Against this inclination, Calabi urges, in part in consideration of the factive nature of seeing, that seeing a hole or seeing that something has a hole in it may indeed be seeing an immaterial object but, what with all the visual cues that serve to locate and delimit the hole, the experience of such a seeing need not itself represent the hole’s immateriality. Thus, there seems to be a respect in which what we see in such cases really is very like seeing a natural unity (albeit of an identifiable absence), and hence may be assimilated to other sorts of direct perception. Given the analogy with anatomizing an organic body from which we started, it may come as something of a surprise that the idea of individuating natural unities is far from unproblematic in the life sciences. The Platonic–Socratic idea of cutting at the joints presupposes (as in our (i)) that there should be just one point or line at which the knife goes in between the parts to be set asunder. The following two chapters consider two ways in which the knowledge that we have been thinking of as characteristic of the skilled carver differs from that of the bungling butcher, one on the level of the single organism and the other on the level of taxonomic structures. In common with the chapters that precede it, Lars Vogt’s discussion sets out from the intuitive idea that the outer boundary of an object can be considered as a mathematical abstractum, which is not properly part of the object it bounds and will have one dimensionality less than it (the surface of a three-dimensional object is itself two dimensional). But Vogt’s expertise as a biologist leads him to elaborate two features of the ways that natural unities can be delineated in defiance of the purely geometrical notion of a boundary. In one direction, he emphasizes that, in defining the parts of living beings, the notion of a causal or functional unit, which may be internally heterogeneous in its physical make-up, takes precedence over that of a mere region, where such units may be traced to an evolutionary frame of reference. The other important observation he makes regards the way that, in considering such items as cells, we do well to think of the membrane that demarcates what is inside and what is outside it rather as a region than as a surface. Given that the cell could not survive without the membrane that covers and protects it, the membrane goes with the cell and has depth; so, natural borders can be proper parts of and of the same dimensionality as the units of which they are borders. One interesting corollary of Vogt’s attention to the granularity of how to individuate biological parts is that the priority attributed to causal unity means that, even though there may be osmosis across the membrane of a cell, neither the particular matter nor its particular location at a given moment determines whether a given molecule is inside or outside the cell.

8

Natural and Artifactual Objects in Contemporary Metaphysics

But this indeterminacy does not lead to any problem about vagueness and does not undermine the bona fide status of the cell under consideration. Questions of indeterminacy arise also in relation to the notion of speciesmembership. On the pre-Darwinian model of the fixity of species, the possession by an animal of a certain essence determined the species to which it belonged. Though there could be surprises with the discovery of new species (such as duck-billed platypuses) that called for adjustments to the higher orders of classification, or with the discovery of members of a given species with unexpected characteristics (such as Cygnus atratus, which is a very fine swan indeed, but not white), such ‘Australian’ phenomena did not introduce any vagueness about the sets in question. As Ciro De Florio and Aldo Frigerio illustrate, the development of evolutionary models led to adaptations of the notion of species in relational terms. Among the leading proposals in this direction are the interbreeding criterion and the phylogenic principle. But since both proposals apply only to organisms that reproduce sexually and, in any case, the interbreeding relation is not unrestrictedly transitive, these approaches are less than satisfactory. Indeed, these relational specifications are neither necessary nor sufficient to determine species-membership. Yet they point towards conceiving of a species as coming to be through evolutionary processes and then passing away with extinction. For this reason, many proponents of these sorts of criteria have seen species as scattered particulars, albeit of a stipulative or fiat sort. But, if the motivation for this position is weak, De Florio and Frigerio suggest that, without returning to any sort of essentialism about species-membership, it may be fruitful to re-examine the ranges of properly structural characteristics of the members that typify a given species, and allow that, though grounded in those characteristics, it may be indeterminate whether a given organism exemplifies to a sufficient degree a sufficient number of them to count as a member of it. In this direction, a species may be considered rather as a cluster than as a set. The element of decision-making involved in a biologist’s fixing species membership does not mean that the resulting grouping is a mere artifact, because such decisions will be based on the biologist’s knowledge of natural saliences and explanatory criteria. To speak of artifacts as ‘mere’ appears derogatory when applied to the objects of the natural sciences, but it is quite out of place when applied to corkscrews (my favoured case of artifactualness). But the standing of artifacts is by no means uncontroversial, as illustrated in the three chapters that discuss our third question, that of where tools come from. Marzia Soavi and Massimiliano Carrara consider the interrelations between the classical approaches to defining artifacts, either in terms of the intentions of

Carving Nature at the Joints

9

the artifex or in terms of the functions performed, and Barry Smith’s notion of a fiat object. For, in a dichotomy that is at least in part analogous, a fiat object may come into being either by explicit definition (intention) or as a collateral of activity that takes advantage of natural features of the artifact (function). Though it is pretty clear that the categories of fiat objects and of artifacts strongly overlap, Soavi and Carrara suggest that the ontological genus of which they are species may be identified as that of functional objects. The following two chapters concern the notion of an artifact in relation to the ways in which natural languages are clearly not natural but conventional. In her exploration of how Aristotle, at the outset of the On Interpretation in one of the earliest attempts to distinguish the ways that a sound can be the vehicle of a meaning, was understood by his great commentator Alexander of Aphrodisias, Maddalena Bonelli examines the evidence remaining in Ammonius and Boethius of Alexander’s attempt to deal with this crux. While a noun/name’s being the noun/name that it is is a matter of convention, Alexander asks why Aristotle says that words in general are primarily signs or symbols of affections of the soul (which would themselves be bona fide objects), and apparently only secondarily of the things they stand for. For all that Alexander buys into the naturalness of fit between the contents of the soul and the contents of the world that Aristotle takes for granted, the presence of an element that makes even the truth of propositions a matter of convention is a forerunner of more recent debates, as Marco Santambrogio makes clear in his chapter. Santambrogio’s first move is to argue that, even if they are not rigid designators of the sorts that have attracted attention in recent decades, such as proper names and natural-kind words, the semantics of terms for artifacts is nevertheless externalist. By this he means that the reference of ‘corkscrew’ does not depend on what I believe or anyone else believes about the characteristics of corkscrews, but rather on how a certain technique for their production has propagated over time. He then turns his attention to how words function as artifacts, and observes that this depends not so much on what individuals believe about their meanings as on how they are actually deployed in a linguistic community. For this reason, it may be said that, for all that the introduction of neologisms may be a matter of fiat, once they have been introduced into a language, their semantics become a matter that is causally derivative of the original stipulation. And this is almost entirely independent of what current users believe about their meanings. The last section addresses aspects of our fourth question: given that what is conventional or fiat is to some degree mind-dependent, it is of some concern to consider to what extent the mental is itself natural or bona fide. Andrea Bottani

10

Natural and Artifactual Objects in Contemporary Metaphysics

and Alfredo Paternoster investigate those most mind-dependent of things, namely mental states, and consider objections to the thesis that these things can be identical with the perfectly natural and biological entities that are neural states. Perhaps more radical still is the challenge that Alfredo Tomasetta explores to the claim that mental states require a bearer, and suggests that, in some sense reminiscent of certain Buddhist traditions, thinking things are themselves nothing but fiat objects.

Social joints I hope in the preceding sketches to have given some sense of the unifying strands that connect the chapters gathered in this volume. But, as announced at the outset, it might seem that one set of debates in recent ontology has been left quite out of account, namely that to do with the standing of social objects, and that this might seem a culpable omission, given that these debates have enjoyed a relatively high profile of late. At the risk of offering a excusatio non petita, I permit myself a few words of explanation, the first of which is the purely contingent fact that the Italian National Research Project, from which the core of these chapters grew, took as its main focus the various distinctions between the natural and the artificial or mind-dependent, of which the distinction between bona fide and fiat is one prominent descendant. Conversely, the debates over how to approach social ontology have given rise to a literature so luxuriant that they deserve at least a volume to themselves if justice is to be done to them, so that it would have been arbitrary or partial to include just a few chapters devoted to these topics. A third point is a little more substantive. Relative to the distinctions that this volume thematizes, it looks likely at first glance that such social objects as one might be willing to contemplate will fall on the side of nomos, of convention, of the mind-dependent and of the end-related, where the ends in question are set by individual humans interacting with others. The question then arises of just what objects we are, in the final accounting, willing to contemplate as properly social. For, it does indeed seem that the interactions of individual humans give rise to structures that may be grounded in those interactions, but that are no more reducible to them than the being-a-corkscrew of my corkscrew is reducible to its being made of metal and being of a certain shape. But it is also apparent that the sorts of interactions that humans go in for are highly variable from time to time and place to place, so that it is not always a straightforward matter

Carving Nature at the Joints

11

whether the objects that we might be willing to recognize as emerging from the interactions in one time or place are the same sorts of objects that emerge in other times and places. Just as one might recognize corkscrews that are made of wood, plastic or ceramic or that have any of many shapes as corkscrews, so one might recognize any of many arrangements for, among other ends, the care of infants as families. There are temptations to say, as for instance Aristotle not unthinkingly did (Pol., I, ii) and many in his wake have done perhaps less reflectively, that some such arrangements are more natural than others. But it is well to resist such temptations by bearing in mind the variabilities in the ways that the objects that are families can be configured and the diverse ends they can subserve more or less successfully. The caution that is in order here may cashed out as a reminder to relativize to times and places. For instance, given that Aristotle thought, and by doing so represented the communis opinio of his time and place, that a family required a male head ruling over just one woman as well as over their offspring and such slaves as might accrue, we might say that that arrangement was a family-for-fourth-century-Athens (as well as for many other times and places). Yet, as any anthropologist will tell you, in some times and places, the sort of arrangement that is recognized as a family may differ with respect to the salient characteristics that Aristotle recognized. Once we have made these relativizations, it is worth noting that being a family is not just a matter of being recognized as such at the time and in the place in which it emerges. The point being that, for instance, even in fourthcentury Athens, an arrangement in which two (or more) women took equal roles in the care of their mothers would have been a family, despite its not being recognized as such by Aristotle, his contemporaries and followers. If this is correct, it seems that at least some social objects can subsist even in the absence of social recognition or sanction at the time and in the place where they emerge. Of course, families-for-fourth-century-Athens were recognized as families in fourth-century Athens; but that does not mean that only families-for-fourthcentury-Athens were families in fourth-century Athens: relativizing to times and places in order to resist the temptations to say that some social objects are more natural than others is consistent with saying that the objects in question are not mere functions of the more general arrangements (e.g. marriage laws) of the societies in which they emerge. A range of interesting political consequences flow from the sort of mitigated relativism just outlined. But I limit myself to just a few observations that arise out of it relative to some of the ongoing debates in social ontology.

12

Natural and Artifactual Objects in Contemporary Metaphysics

Perhaps not all theorists in this field subscribe to the view that, in one way or another, the social objects that are recognized and sanctioned in a given society can be identified by reference to the modes of recognition and sanctioning that that society accords them. Nevertheless, in their different ways, two of the leading schools of thought, that inspired by John R. Searle’s account of the construction of social reality in terms of collective intentionality and that which follows Maurizio Ferraris’s identification of social objects with inscribed acts,3 seem to be committed to at least some version of this view. If the assignment of social standing to some arrangement requires, as on Searle’s account, that there be some institution that recognizes and sanctions it, then the gynocentric filial collective alluded to above would not be a family in fourth-century Athens because there was no such institution in that time and place ready to confer that standing on such an arrangement. Likewise, if there was no recognized mode for such a family to register itself as such, Ferraris too would have to say that it was not a family. But I am ready to say that, despite its defiance of what were then and there the marriage laws, it was indeed a family. And one reason that Searle and Ferraris might be ready to concede is that it is hard to exclude the possibility of there being a society that does recognize and sanction such a union so that, by their lights, it would be a family in that society. To put it hypothetically, if some anthropologist discovered a society in which such units were recognized and sanctioned, then she would be right to say that what she was describing was one way in which families are formed in that society. Whatever word the members of such a society used for it in their language would be perfectly well translated as ‘family’. The aporia that arises here may be put rather paradoxically as the claim that a social ontology ought to be able to contemplate social objects that exist regardless of the actually existing institutions, as expressed in collective intentionally or by the archiving of documents, which are at the core of the theories that are currently most vigorously pursued. Another point on which Searle and Ferraris converge regards what they think of as the paradigm cases of social objects. Perhaps Searle is rather more guarded than Ferraris and more often speaks of social facts and social reality, but both take their explanations of the coming into being of social reality as exercises in ontology, in the theory of what things there are, and Ferraris’s recurrent formula (presented more than thirty times in Documentality) is ‘Social Object = Inscribed Act’. Among the instances on which Searle and Ferraris converge in thinking of as paradigmatic cases for a social ontology are marriages, money, university degrees and national borders. And in each of these cases, they offer

Carving Nature at the Joints

13

accounts of how these things come into being, and they take it that the fact that their theories account for the coming into being of such things is a motive in favour of accepting the theories. From what I have already proposed about families, it follows fairly naturally that one can be a member of a family only if the family exists in whatever way such things exist, as structures, arrangements or units that emerge from certain types of human interaction. But I am less sure that my being married or not depends on the existence or otherwise of an object that is a marriage. There are, of course, ceremonies that are weddings or marriages and that, by way of promising or swearing, codify certain rights and duties of the parties to it. Yet a ceremony is not an object but rather an event. The rights and duties that are codified in such a ceremony can hold good even in the absence of the ceremony. They can, for instance, be matured by the parties acting towards each other in certain ways over a certain period of time, thus generating reasonable expectations of the observance of them. Again, the parties to a marriage may acquire titles such as ‘husband’ or ‘wife’ and these titles do indeed denote objects, but the objects they denote are persons with the rights and duties that may or may not have been codified in a ceremony and that can hold good in any case. With specific reference to Ferraris’s theory, we may also observe that it would be very curious to suppose that a marriage is to be identified with a piece of paper issued by a registrar of births, deaths and marriages. While the rights and duties that are codified by the ceremony that leads to the issue of the piece of paper may be enforced by the production of the piece of paper (for instance in a divorce court), the base or explanatory category here seems to be that of the rights and duties themselves, rather than some registration of them. As regards money, it is not problematic that the €10 banknote that I have in my pocket is an object, but it would be curious to say that the money it embodies is an object and it would be downright misleading to say that the corresponding amount that I may be fortunate enough to have on my current account is an object. Here, Searle and Ferraris converge on the idea that the money in question should be regarded as a symbolic object whose powers or tendencies are actualized in moments of exchange. I think that this retreat from thinking that economic and financial transactions generate objects is perfectly reasonable. After all, the redness or weight of a tomato may be accounted for by the propensity to reflect light of a certain wavelength or by the tendency to turn the scales to a certain degree, and we feel no push to invoke ‘chromatic objects’ or ‘gravitational objects’. In this sense, it seems that, unlike families, money is not really a social object, but rather a function that holds among objects like pieces

14

Natural and Artifactual Objects in Contemporary Metaphysics

of paper (some, but not all, of which are banknotes), commodities of various sorts and economic agents (probably of more than one sort). The curiosity here is that a plausible exercise in explaining what money is will recognize social facts without having to invoke any social objects; so one might wonder in what sense it is ontology at all. Just as being a husband does not require there to be an object that is a marriage, so being a graduate does not require there to be an object that is a university degree. The qualification is a record of things done in lecture rooms, laboratories, libraries and examination halls, and it is not unreasonable to allow that the university is an object that in one way or another embraces these buildings, but is not reducible to them. One might even say that the university is a function of those activities and buildings, rather as being a corkscrew is a function of a certain matter and form. Like a marriage certificate, but unlike a €10 banknote, a degree certificate is non-transferable in exchange for goods or services. If someone buys a degree certificate that is not a record of the successful performance by the purchaser of certain academic tasks, then that inscription is hardly worth the paper it is written on. There is a certain variability from time to time and place to place (and from discipline to discipline) about just what tasks the performance of which warrants the issue of a degree certificate; and there have been times and places in which the tasks are so little onerous that the title may be in odour of fraud. Quite apart from the Mickey Mouse degrees sold by institutions that are not universities, it is a feature of both Searle’s and Ferraris’s accounts of social reality that they have difficulty explaining the parasitic status of what happens when a respectable university awards a degree, such as an honorary doctorate or an Oxbridge M.A., even in the absence of the performance of any academic task. On both the collective intentionality and the documental accounts, the overall legitimacy of the institution seems to be sufficient to guarantee that when, for instance, Europe’s most venerable university, Bologna, awarded an honorary degree to popular musician Lucio Dalla (1999), that made him a graduate. I do not deny that it did; but one might be forgiven for doubting that the social fact thus brought into being has the force of a degree awarded in recognition of examinations successfully passed. Yet neither Searle nor Ferraris can quite capture the causes for doubt in this case or the causes for indignation in cases in which a university awards a degree to a local magnate or potentate (which has been known to happen, after all). Rather, as with Aristotle’s account of what a family naturally is, both Searle and Ferraris seem to have to regard the ‘universitiness’ of certain institutions as a given from which flows their right to award degrees as they see fit. And, as with

Carving Nature at the Joints

15

the case of families, this fact raises questions of justice that it is not our present purpose to pursue. If there is a question about the legitimacy of this or that institution’s calling itself a university, the question can, in the happy cases, be answered in a noncircular way by looking at the activities of teaching, learning, researching and examining that it promotes as its focal ends. By contrast, national borders, which are also among the paradigm cases for Barry Smith’s delineation of fiat objects, do not lend themselves to specification in terms other than the acts that bring them into existence. But, once more, one might nourish the doubt that what is brought into existence is in any clear sense an object. For instance, if I have to present certain documents in order to pass relatively unhindered along the road that leads from Ventimiglia to Menton, that is not because there are such objects as Italy and France or the Italy–France border, but rather because the French border police are licensed by the French state to threaten to use, and indeed to use, force against me if I do not possess or refuse to present documents that they find convincing. But the Frenchness of the French state seems to amount to little more than the acquiescence for the time being of the Germans, Spaniards, Italians and Belgians in the French saying that their state can license the threatening behaviour of the border police (and many other functionaries). As with the anthropology of what counts as a family from time to time and place to place, so a glance at a historical atlas of Europe will quickly convince an impartial reader that the nations just mentioned are hardly objects but, if anything, processes or ensembles of rules and administrative systems. In these ways, it seems that the approaches to social ontology that appeal to the functioning of recognized assignments of social roles, whether by collective intentionality or by the inscribing of documents, not having recourse to anything so parochially normative as Aristotle’s notion of natural societies, run into problems in saying what makes such institutions valid or otherwise. Not unlike the aporia about the inter-societal recognizability of certain objects as tokens of the same type, here the problem is that the assigning institutions may themselves be subject to evaluation, for instance in terms of efficacy, efficiency and reasonableness as well as justice, in ways that the theories that appeal to them cannot countenance. In short, though recent debates in social ontology are of great interest and importance, not to say ethical and political urgency, the feeling is that many of the issues that arise in them may be more adequately handled when some of the distinctions on which they rely have been more fully explored. It is to the

16

Natural and Artifactual Objects in Contemporary Metaphysics

exploration of some of these distinctions that this volume is primarily dedicated, by way of what Locke might have called under-labouring (‘Epistle to the Reader’ of the Essay).

Notes 1 The point is made in another tradition that uses carving as a paradigm of skilled encounter with nature: Chuang Tzu: ‘Nourishing the Lord of Life’ §2. 2 Among the first into the field was Lynne Rudder Baker, whose ‘Why Constitution Is Not Identity’ appeared in the Journal of Philosophy in 1997 (vol. 94, No. 12, pp. 599–621); her Persons and Bodies (Cambridge University Press, Cambridge 2000) takes the particular case of persons, but elsewhere she examined other sorts of ‘everyday things’ (e.g. The Metaphysics of Everyday Life, Cambridge University Press, Cambridge 2007). In this vein, we may also cite the significant titles of the books by Amie L. Thomasson, Ordinary Objects (Oxford University Press, Oxford 2007) and Daniel Z. Korman, Objects: Nothing Out of the Ordinary (Oxford University Press, Oxford 2017), and David Wiggins’s perhaps more inscrutable Continuants (Oxford University Press, Oxford 2016). 3 Searle’s book-length outings on this topic are The Construction of Social Reality (Penguin, Harmondsworth 1995) and Making the Social World: The Structure of Human Civilization (Oxford University Press, Oxford 2010) which, with its greater emphasis on collective intentionality, is in continuity with work in the social sciences inspiring and inspired by Margaret Gilbert’s Joint Commitment: How We Make the Social World (Oxford University Press, Oxford 2014). Ferraris’s most systematic statement of his views is Documentality: Why It Is Necessary to Leave Traces ((2009) Fordham University Press, New York 2013). Though Searle’s and Ferraris’s views differ on many significant points, their convergence on the importance of the topics studied can be seen in their joint publication Il denaro e i suoi inganni (Einaudi, Turin 2018), collecting an Italian version of a paper published by Searle in the Cambridge Journal of Economics (‘Money: Ontology and Deception’, 2017, vol. 41 No. 5, pp. 1453–70) with a discussion ‘On the Colour of Money’ by Ferraris and a comment on symbolic objects by Angela Condello.

Part One

Does Nature Carve Itself?

1

Mental Acts, Externalism and Fiat Objects: An Ockhamist Solution Riccardo Fedriga

Any theory aiming to give an account of human cognitive faculties must reckon with a number of crucial issues, such as the imperfection of our epistemic equipment, the (more or less defined) boundaries between subject and object of knowledge and the nature of the tools that allow connecting and classifying entities of the known reality into classes or subcategories. This chapter inquires into Ockham’s treatment of human cognitive tools and their fallibility. We shall elucidate Ockham’s conceptual framework and examine whether his evidently naturalist epistemology may be described through the conceptual tools provided by the analysis of fiat objects, given that a causalist framework is arguably not suited to exhaustively account for it. The analysis of Ockham’s responses to these epistemological problems shall thus entwine with the analysis of the responses that, according to philosophers and historians, Ockham would have given to contemporary questions. Such an approach will allow us to test Ockham’s position against the background of the externalism/internalism debate, on the basis of Wilfrid Sellars’s criticism to the ‘myth of the given’.

Human knowledge The foundations of Ockham’s theory of human knowledge In the mature version of his theory, Ockham envisages a universe constituted exclusively of singulars, in which every thing is either a substance or a quality. Within this framework, mental acts are conceived of as belonging to the category of quality.1 They are divided into two general categories, in accordance to their referring to the sensitive part of the mind (listening, desiring, fearing)

20

Natural and Artifactual Objects in Contemporary Metaphysics

or to the intellective part (conceptualizing, reasoning). A further distinction is that between apprehensive and adjudicative acts: both sensitive and intellective faculties are capable of apprehension, but only intellect is capable of judgement. Ockham, however, leaves aside those acts that are merely sensory and focuses on intellective acts.2 Since we can apprehend not only incomplex objects (i.e. singular terms) but also propositions, demonstrations and impossible or necessary objects (i.e. complex, propositional objects), the intellect’s apprehensive acts must occur with respect to both complex and incomplex objects. Adjudicative acts, on the contrary, are about incomplex objects only, for an adjudicative act is the assent or dissent that our mind grants to a proposition. Propositions may be spoken, written or mental. Mental propositions constitute concepts and occupy the semantic space of an intentional act referring to a singular object that exists in reality. Such an act (which is the same concept in term of mental language) is also a linguistic sign that takes the object’s place within the mind and replaces the object in mental propositions. In other words, it is a term that stands for (i.e. supponit) individual objects in the world and substitutes them, directly and with any intermediary, in the mind of the thinker. If we consider the issue in the light of the fiat objects theory, the question arises about whether it is possible to draw a boundary between objects (res), mental acts, mental terms and linguistic acts of reference, and in what way this could be done.3 It is evident that such a space cannot be delimited by (only) a causal nexus, as if the knower were the cause of the known object, insofar as it is known. Rather, the relation between these two ‘objects’ appears to be more fleeting and contingent. In our view, this primarily depends on categorial differences. For, first of all, Ockham assumes only two of Aristotles’s categories, namely substance and quality; and secondly, it is a matter of examining to what extent Ockham’s theory may be defined as externalist or internalist. Let us examine apprehensive acts first. They can be divided into acts of simple apprehension (notitia intuitiva), which can never have a propositional content, and acts of abstractive apprehension, which can be both propositional and nonpropositional. This means that, when a white wall is in front of me, I intuit the wall, or this particular whiteness, or perhaps both these things at once, but I do not intuit that the wall is white. Abstractive propositions, on the other hand, may have a propositional content (e.g. my non-assertive thought that Socrates is white), or not (the occurrence in the mind of a simple concept such as ‘man’ or ‘whiteness’).

Mental Acts, Externalism and Fiat Objects

21

In the Ordinatio,4 Ockham deals with the two kinds of abstractive apprehension, that is, that of the universal and that of the singular,5 before advancing his novel and crucial definition of intuitive apprehension. On the one hand, there is abstractive knowledge that relates to something abstracted from many singulars and therefore coincides with knowledge of a universal that can be abstracted from a number of entities. On the other, there is abstractive knowledge that relates to a single thing and abstracts (i.e. is independent) from the thing’s actual existence or non-existence. We shall turn to the latter in due course. Let us now focus on intuitive knowledge: Intuitive cognition of a thing is cognition that enables us to know whether the thing exists or does not exist, in such a way that, if the thing exists, then the intellect immediately judges that it exists and evidently knows that it exists, unless the judgment happens to be impeded through the imperfection of this cognition. And in the same way, if the divine power were to conserve a perfect intuitive cognition of a thing no longer existent, in virtue of this non-complex knowledge the intellect would know evidently that this thing does not exist. Ockham, Ordinatio, Prologus, q. 1, art. 16

Summing up, intuitive knowledge is the knowledge by which one knows that a thing exists when it exists, and that a thing does not exist when it does not exist. The act of intuitive cognition allows intellect to access reality, to realize that this or that thing exists, to know it in the immediacy of its existence. Starting from such an apprehension, the intellect may then formulate a judgement about the existence of the intuitively known object (the pen exists, the table exists) or, more generally, a judgement of assent or dissent in relation to a contingent truth. To give an example: if Socrates is in front of me and is white, I shall have the intuitive knowledge of Socrates and of whiteness, and by virtue of this I shall know with evidence that Socrates is white. On the other hand, abstractive knowledge of the singular (which is what we shall be referring to from now on) does not allow me to know, of a thing that exists, that it exists, nor, of a thing that does not exist, that it does not exist. It is, so to speak, a sort of knowledge that is ‘indifferent to existence’. For instance, reading a text about Socrates might produce in me a nonpropositional representation of Socrates and his whiteness, but such an incomplex abstractive knowledge does not allow me to know whether Socrates exists or not, nor whether he is white or not. Intuitive apprehension allows me to know contingent truths, especially concerning the present; abstractive knowledge does not.

22

Natural and Artifactual Objects in Contemporary Metaphysics

Unlike Scotus, Ockham thinks that intuitive and abstractive knowledge refer to the same object, that is, a singular object, and moreover, that the latter always accompanies the former. There is a difference, however: the intuition that a pen is on my table enables me to produce a judgement of existence of the kind, ‘There is a pen on my table’, but once the pen is removed from the table, my intuitive knowledge of it ceases; I can, however, still think about it. Such a cognitive act is still concerned with a singular object, that is, the pen, but does not allow me to assert, ‘There is a pen on my table’: it is independent from the actual existence or non-existence of the object. These two kinds of knowledge, in other words, have different effects: while intuitive knowledge enables the knower to produce judgements of existence, abstractive knowledge does not. But, one may ask, how do intuitions come about? Claude Panaccio7 points out that an act of intuitive cognition occurs naturally when a real singular object is placed in an adequate location with respect to the knower (e.g. the object is not too distant, it receives enough light, the knower’s sense organs are correctly disposed, etc.). Such a natural cause is the object of intuition: if my act of intuitive cognition is adequately caused by Socrates, who is located in front of me, then this is an act of intuitive apprehension of Socrates.

Ockham’s theory of intuition Ockham defines intuition in a purely functional manner, on the basis of the role it has within the cognitive process’s causal chain8: intuition is caused by some singular object in the world, and, in turn, causes a judgement of assent to a true contingent proposition about that object. No reference is made to what one feels when one intuits something, nor are intuitions framed as intrinsically recognizable experiences: they seem to be entirely deprived of phenomenal characters. In Panaccio’s view, it follows from this that Ockham’s theory of intuition is undoubtedly externalist. In particular, Ockham’s externalism concerns mental content: namely, two knowing subjects might, in principle, have very similar intuitive states, yet one of the two might be thinking of one thing, the other of an entirely different one. Let us suppose, for instance, the following situation: there are two individuals; one of the two is observing an egg (call this ‘egg n.1’), while the other is observing another egg, which is extremely similar to the former (call this ‘egg n.2’). Following Ockham’s argument, we should then say that the first individual is intuiting egg n.1 rather than egg n.2, for his current intuitive state is caused by

Mental Acts, Externalism and Fiat Objects

23

egg n.1 and not by egg n.2. Panaccio grounds his externalist reading of Ockham on three pieces of textual evidence.9 The first piece of evidence is drawn from Reportatio II, q. 16, where Ockham proposes the angel’s thought experiment. Let us suppose that an angel were able to see what is in my mind when I apprehend an object that is present in my vicinity. Would the angel be able to say exactly which object I am intuiting at the moment? No; the angel would not be able to identify the object I am intuiting by simply examining my internal act of intuitive cognition and comparing it with the many things in my vicinity, especially if two of them are very similar. For the angel to correctly identify the object of my intuition, she should know which object caused in me that particular act. The second piece of evidence is taken from Quodlibeta Septem I, q. 13, where Ockham claims that intuitive knowledge is a knowledge of the singular that is more adequate than abstractive knowledge, for an intuition and its object are linked by a relationship of causality rather than mere similarity. Finally, the third piece of evidence appears in Reportatio II, qq. 12–13: ‘similarity is not the reason for which the mind knows one object rather than another’; rather, knowledge is due to the fact that ‘every naturally produced effect is produced by a single efficient cause and nothing else’, and in the case of intuition, such a single cause is constituted by the singular object of the act. At the same time, Panaccio admits that an act of intuitive cognition includes some descriptive content about its object: this is what Ockham seems to be saying when he claims that both intuitive and abstractive apprehensions are similar to their object. Such a descriptive component, however, is not part of the semantic content: it is only pragmatically required to ensure the correct re-application of the concept or the correct intuitive recognition – for example, when I recognize a person I meet today as the same person I met yesterday. Such a descriptive component does not fix the reference, which is fixed exclusively by causality. Thus, Panaccio may conclude what follows: Intuitive acts, admittedly, are said by Ockham to be likenesses of their objects in some sense. This likeness, I take it, has to do with the internal features of the intuitive act. Such internal features, however, cannot uniquely determine what the proper singular object of a given intuitive act is. Only causality can do that, since causality links the intuitive act with one single determinate cause, while similitude relates it in principle with several similar objects. (Panaccio 2010, pp. 243–4)

24

Natural and Artifactual Objects in Contemporary Metaphysics

Panaccio also emphasizes the atomistic bent of Ockham’s epistemology and the effects of intuitive apprehension.10 First, Panaccio stresses that the cause, and thus the object of an intellectual cognition, is always a singular thing and not a fact or state of affairs: the cause of my intuition of Socrates is Socrates himself, not the fact that Socrates exists. In Ockham’s account, intuition is not a propositional attitude, and this explains why an intuition, while being a cognitive state, does not by itself constitute a piece of knowledge. Ockham’s theory of knowledge, therefore, is not only externalist, but also atomistic. Secondly, acts of intuitive cognition produce equally non-propositional concepts. (Ockham’s mature theory will identify the latter with abstractive acts.) The process can be reconstructed as follows: when I meet a certain individual, the intuitive cognition that is caused by this causes in my mind the formation of a general concept corresponding to the narrowest natural genus to which the individual belongs – for instance, the concept of man if I meet a man. The fundamental feature of this process is that a single encounter is sufficient for the natural acquisition of such basic concepts. Moreover, intuitions also produce a number of adjudicative acts about contingent propositions, which are constituted of concepts that are either just formed or re-actualized.11 The assents granted to contingent propositions, coupled with the cognitive habitus they impress on the mind, constitute proper knowledge, for they are cases of evident knowledge12: evidence is not a feature of the act of intuitive cognition in itself, but of the propositional knowledge that causally follows from that. This reading is not meant to deny the fundamental role that, according to Ockham, intuitive apprehensions play in the cognitive process – this is, in fact, precisely what earned the Venerabilis Inceptor the nickname ‘philosopher of intuition’. Intuitions, which are not propositional, constitute the raw matter for a sophisticated mental mechanism that assembles non-propositional atoms into complexes (i.e. propositions) to which we may give our assent. Our cognitive process starts from intuitions: by causing the formation of basic concepts, intuitions trigger our assent to propositions. Having said that, however, it must be stressed that the evident knowledge of a proposition such as ‘Socrates exists’ is caused by the act of intuitive cognition, but it is not that act itself.

Sellars and Ockham the externalist Unlike what happens in Wilfrid S. Sellars’s ‘myth of the given’, therefore, intuitive apprehension causally produces knowledge but does not constitute knowledge. According to Panaccio, this is precisely what allows Ockham to escape Sellars’s

Mental Acts, Externalism and Fiat Objects

25

criticism. Since Ockham’s approach might appear as the perfect example of the so-called myth of the given, however, it is worth going through this label once more. Sellars coins the phrase ‘myth of the given’ to name ‘the conviction that there are sense-contents that are immediately accessible to the epistemic subject, that is, pure objects of knowledge that are not mediated by the inferentialconceptual apparatus’; against such a myth, Sellars goes back to Kant’s thesis according to which experiences without concepts are blind.13 The myth of the given, in other words, claims that the mind is purely receptive when it comes to the information that might be used to find an epistemic justification. This seems to correspond to Ockham’s theory: intuition is but the mind’s immediate acquisition of particular objects, and by virtue of this, it provides the foundation for all the knowledge we have of the world. In normal conditions, a real singular object causes an intuition, and this in turn causes an evident knowledge of a true contingent proposition. According to Sellars, says Panaccio, such a theory is problematic because, either the intuitive connection to things in the world has a purely causal nature (in which case it would be difficult to see how it may justify a belief), or intuitive awareness is itself a piece of knowledge, and if so must presuppose categorization and concepts. The bottom line of Sellars’s attack on empiricism is that knowledge always requires categorization, and thus cannot derive from brute, non-conceptual data such as intuitions.’ Panaccio’s thesis is that Ockham’s theory of intuition constitutes a way out to Sellars’s dilemma. Let us see exactly how. Saying that divine action can separate two things amounts to say, in Ockham’s philosophical jargon, that the existence of one thing that does not logically entail the existence of the other, for God cannot do what implies contradiction. Therefore, the relationship between intuition and proper knowledge is, in Ockham’s view, a causal relation rather than an inferential, logical one.14 The assent to a true contingent proposition that is caused by acts of intuitive cognition constitutes knowledge because (a) the assent is correctly caused by such acts of intuitive cognition and (b) such a process, though in principle fallible, is highly reliable in normal and natural conditions. Ockham’s theory succeeds in warding off Sellars’s criticism precisely because the reliability of the cognitive process rests on a form of epistemological externalism. Sellars’s thesis, as a matter of fact, presupposes an internalist attitude: in his view, the knowing subject has the concept of green only if he possesses a whole battery of concepts of which the concept of green is one of the elements.

26

Natural and Artifactual Objects in Contemporary Metaphysics

This is not true in an externalist framework, and in particular, it is not true in Ockham’s theory. In order for a belief to count as knowledge, it is not necessary for it to be internally represented in the mind of the knowing subject as the result of an inference, and it is thus not necessary for the knowing subject to be aware of it. Knowledge that a certain thing is green is simply caused by an act of intuitive cognition, whether the knowing subject is aware of it or not, and there is no need for a rational justification of this belief. In Ockham’s account, for evident knowledge to occur, it is sufficient that the concept of a green thing and the judgement that such a thing is green are adequately caused by acts of intuitive cognition. Panaccio concludes therefore that such an externalist account of the cognitive process allows Ockham to escape both Sellars’s criticism and the implausible epistemic holism that grounds it.

Knowing an object that does not exist: Strength or weakness of Ockham’s theory? In opening this new section, it is worth repeating once more that intuitive knowledge is that by virtue of which one knows that a thing exists when it does exist, and that a thing does not exist when it does not exist. Katherine H. Tachau stresses how such a definition seems to broaden the intuition’s potential (which is traditionally understood only as non-discursive, direct, simple and immediate awareness) so as to include the awareness that an object does not exist.15 Conceiving of an intuition that enables us to apprehend that something exists is straightforward enough; grasping the notion of an intuition that enables us to know that something does not exist is an entirely different matter. Indeed, when we do not perceive an object, we may, at the most, infer that the object is not perceptible, or that it is not present, but we are certainly not in the position to say that the object does not exist. How can one have evident knowledge that a certain thing does not exist, then? And, at a more basic level, can we even conceive of acts of intuitive cognition that have non-existent things as their object? In order to solve this dilemma, we can turn to Elizabeth Karger’s article ‘Ockham’s Misunderstood Theory of Intuitive and Abstractive Cognition’. As Karger points out, Ockham admits of a theological principle according to which, though nothing ever causes any effect without God’s cooperation, God can act alone to cause any of the effects any given thing is capable of causing. If this principle is applied to acts of apprehension, it will have to be recognized that acts of intuitive cognition can exist, which are produced in an intellect directly by God.

Mental Acts, Externalism and Fiat Objects

27

However, if an act of intuitive cognition is caused by God alone, then the thing itself, that is, the act’s object, has no role in causing it: there is no need for the thing to be present or even to exist for the act to exist. It follows that intuitive cognition of non-existents may occur, but such acts are supernaturally caused by God acting alone to produce them in an intellect. A question, Karger points out, immediately arises: would an act of intuitive cognition of a non-existent thing be capable of causing an act of evident assent to a contingent proposition? Ockham answers that it would, adding that the proposition assented to would be one correctly asserting that the object does not exist.16 Since the case of intuition de re non existente is possible – though only outside of the ordinary course of events and by divine intervention, namely by the potentia absoluta Dei – we may comprehend Ockham’s definition of intuition as that kind of cognition by virtue of which one knows that a thing exists when it exists, and that it does not exist when it does not exist. To be sure, one may still wonder why Ockham insists so much on a definition of intuition that allows also for the (miraculous) case of knowledge of non-existents. According to Costantino Marmo, ‘the thought experiment of divine intervention allows Ockham to eliminate the connection between intuitive cognition on the one hand, and existence and presence of the object on the other hand, in order to identify what is truly essential to it’,17 namely, the fact that intuitive cognition is capable of producing evident assent to a true contingent proposition. Ockham’s position seems thus to differ from that of John Duns Scotus, for whom the only possible intuition during our earthly existence is the one that recognizes the existence and presence of a singular object that is located in the vicinity of the knowing subject. In Scotus’s view, moreover, abstraction is concerned with the essence of things, which is universal and thus indifferent to the actual existence or non-existence of that thing. ‘While recognising that the existence and presence of an external object are the necessary conditions for intuitive cognition in the ordinary course of things’,18 Ockham also admits of cases in which they are not. Ockham’s omnipotent God, in other words, shortens the process that leads to the evidence of a negative existential judgement, for, in the natural course of things, such evidence is not immediate.19 Indeed, the difference between intuitive and abstractive cognition does not lie in their object – since they both have a singular as their object; rather, it lies in their effect. An interesting reading is provided by John F. Boler, according to whom the cognitio de re non existente has a specific anti-Platonic function.20 According to Boler, Ockham’s concern here is to avoid Platonism, not scepticism. Platonism, in this context, must be understood as the philosophical position that (a) allows for

28

Natural and Artifactual Objects in Contemporary Metaphysics

the belief in a greater number of things than the world contains and (b) requires to hold on to this position in order for the world to contain everything that is in thought and language. Boler argues that Ockham insists on explaining assent to propositions through the apprehension of terms because his aim is to provide an adequate theory of language and thought, not because he is concerned with justifying assent. There are no extra-mental entities to which true propositions correspond; there is no intermediary between a true proposition and the world; and one must not muddle the level of thought with things that are independent from it. However the apprehension of terms is exactly to be understood, it is certainly an incomplete component of knowledge: it is employed to construe assent, but cannot be used to justify it. Thus Boler claims that Ockham rejects Scotus’s definition of intuitive cognition as an apprehension of present and existent things precisely because such a definition blurs the distinction between things and thought. Indeed, Ockham affirms that intuitive cognition must be defined with reference to the existence of objects, but intuition is a kind of apprehension that is required to form propositions; it precedes the judgement about a proposition. Ockham is thus concerned with excluding from the definition of intuition any element that may institute a logical connection between the terms and the existence of objects. The notion of an intuitive cognition of non-existent things is needed precisely in order to write off the idea that there is a necessary logical connection between intuition and existence, thought and things. Ockham, therefore, asserts that (1) God can produce in us an act of intuitive cognition of a non-existent thing, which, in turn, can cause an act of evident assent to a contingent proposition and (2) that the proposition assented to is one that correctly asserts that the object does not exist. Such a reconstruction would prima facie appear as a clear counter-example to the externalist causal reading of Ockham’s doctrine: for the object is not there, it does not exist, so it cannot cause anything at all. The possibility of knowing a non-existent object as non-existent, therefore, seems to pose serious problems to an externalist interpretation of Ockham’s thought purely grounded on causal processes. This is the position held by Susan Brower-Toland,21 who quotes two cases of supernaturally produced intuition – which Ockham himself admits of – that seriously undermine Panaccio’s externalist interpretation. The first case is represented by the intuition of a distant object that is caused by God: in BrowerToland’s view, this shows that, if Ockham’s theory is indeed externalist, it has to be a kind of externalism. For strong externalism claims that a given mental

Mental Acts, Externalism and Fiat Objects

29

state depends on the existence of a certain entity in the vicinity of the knowing subject (which is not the case in the example just given), while weak externalism has looser criteria, as it allows for the relevant entity to exist even if not in the vicinity of the knowing subject. Ockham’s theory, therefore, may, at most, be framed as a version of weak externalism. However, Brower-Toland mentions a further case of divinely caused intuition, namely that of a non-existent object, which seems to decisively exclude a relation of foundation between causal processes and externalism, be it weak or strong: for in this case, the object does not exist at all, neither close to nor far from the subject’s location. Panaccio strenuously defends his externalist interpretation in his article ‘Intuition and Causality: Ockham’s Externalism Revisited’, where he tries to counter Brower-Toland’s objections. For what concerns the first case, Panaccio offers a response that is in line with what Ockham thought at the time of the Ordinatio: there is undoubtedly an immediate causal connection between the distant object and the act of intuitive cognition. True, Ockham writes that the object would be too far away to alone be able to naturally cause a sensory cognition that then turns into an act of intuitive cognition; however, if God decides to skip the usual natural process, he can play the role that is normally assigned to sensibility and collaborate with the external object in order to produce an intellectual intuition or a judgement of existence. In this case, the external object would be both the object of intuition and its partial cause. As for the second case – which, according to Brower-Toland, completely rules out any externalist foundation of every causal process – Panaccio recalls that in Ockham’s view, intuitive cognition of a non-existent object cannot occur naturally, but can only result from divine intervention, in which case God is the efficient cause of intuition, but not its object. Here, the distinction between cause and object of intuitive cognition emerges: in such a situation, according to Ockham, the intuition’s cause is God, but its object is the (possible) object that would have caused that intuition, had that intuition been naturally caused. In the case of supernaturally produced intuitive cognition, its object is correctly identified by looking at the closest possible world in which such an intuitive cognition would be naturally produced. The problem, however, stands as follows: how can we get to know the object of such an intuition? Since the act of intuition lacks any internal feature that may univocally determine its object, as Panaccio insists, in the case of a supernaturally produced act it would be impossible, for both us and an angel, to know with

30

Natural and Artifactual Objects in Contemporary Metaphysics

certainty what its object is solely on the basis of the act’s internal features, for the act may be traced to many different things. According to Ockham, however, any act of intuitive cognition has one and only object, which is fixed by the unique way in which God ordained the world. God created the natural order by connecting singular causes to singular effects, in a one-to-one relationship. And since God ordained the world in such a way, it follows that for each naturally produced thing there is an individual thing – at least one possible individual thing – which is its only possible cause within this natural order: ‘The internal aspect of my intuitive grasping … is not in itself sufficient to uniquely determine the object of my particular intuition. Causality is required.’22 The object of intuition, in this case, is that one singular thing that would have caused this particular act of cognitive intuition, had this act been naturally caused. In Panaccio’s view, Ockham’s insistence on providing a causal account of all intuitive cognitions, even those that are directly produced by God, proves that the externalist reading of his theory of intuition can and must be safeguarded.

Ockham’s implicit externalism According to Panaccio, Ockham’s framework is not only unmistakably externalist: it even embodies a particularly strong form of externalism. As a consequence of this reading, Panaccio assigns to Ockham a full-blown theory, which he reconstructs by piecing together references scattered throughout Ockham’s philosophical works. Such a theory is grounded on a fully fledged linguistic externalism, from which an externalism about mental contents follows; the latter, in turn, grounds a form of epistemic externalism. Starting from the rejection of the Modist principle conveyed by the formula ‘sicut intelligitur, sic et significatur’, according to which a term receives signification only if it has first been comprehended, Ockham introduces the notion of supposition, which expresses the idea that, within a proposition, terms replace concepts rather than real things.23 The semantic link that connects linguistic terms to concepts, on the other hand, is dubbed subordination: concepts are natural signs of things, for they naturally signify them, whereas linguistic terms, while they do directly refer to things, they do so by virtue of the conventional act of impositio, which originally institutes the link between word and meaningful terms; terms thus hold a secondary form of signification (as opposed to the primary and direct signification of mental language), which was established by he who first introduced a certain word.

Mental Acts, Externalism and Fiat Objects

31

A certain spoken word primarily signifies several things equally because it has been imposed by a single imposition to everything which a determinate concept of the impositor is common to, so that the word and the concept are to each other like ordered signs … such is the spoken word man, and this is why it is simply univocal: the person who imposed this word man intended that it should signify every thing which a determinate concept of the mind is true of Ordinatio I, 2, q. 4.24

It is in this sense, Panaccio points out, that the terms of spoken language are conventionally subordinate to the terms of mental language (i.e. concepts), in such a way that the former, by virtue of the relation of subordinatio, share the concepts’ signification regardless of what the concepts signify. Thus there is a gap between the meaning of the words a speaker is using and the concepts that same speaker has in mind the moment he utters them. This means that it may be possible for a speaker to correctly employ terms without possessing the corresponding concept: ‘Suppose that you have baptized your dog Fred. I can be willing to use the name Fred and you have imposed it, even though I have never seen the dog in question. I might not even know that Fred is a dog.’25 The criterion for the validity of a term or a statement, therefore, is not constituted by the reference to different concepts, but by the reference to a reality that lies outside of the conceptual domain: it is, in other words, a purely externalist criterion. The semantic area of mental terms does not coincide with the pragmatic area of written and spoken language, and is thus not sufficient to establish the correctness of the cognitive act with which we approach the plane of things. However, if we turn back to Ockham’s distinction between sensitive and intellective intuition and ask about the determination of the former, we reach a second level of externalism, namely, that about mental contents. If objects are the cause of sensitive intuition, what happens when two similar objects present themselves to us? Panaccio makes reference to the case of angelic cognition of individual realities, a topos of medieval theology; but the same point can equally well be made through the example of two eggs: no one would be able to say which of two eggs caused a certain intuition, for the analysis of the corresponding mental content would not provide sufficient elements to decide. Two subjects, therefore, may obtain the same sensitive intuition from the perception of two similar though not identical objects: only the external reference can determine the difference between the cognitive act that grasps object x and that which grasps object y.

32

Natural and Artifactual Objects in Contemporary Metaphysics

General concepts fall under the same rule: general (also called simple or absolute) concepts possess a content that is entirely determined by their extension. For instance, the moment I see a horse and the concept of horse is formed in my mind, the latter immediately becomes the natural sign I apply to all the horses I shall see. Concepts, therefore, have a twofold natural signification: the particular horse, which is the cause of the concept itself, and each and every horse. Asking what determines the content of a general concept and what the latter is, amounts to asking about its extension, which is in turn determined by two external factors: the causal connection and the likeness between substances. Ultimately, the process involved seems to be fully externalist: The extension of the concept is determinate externalistically. But that in no way prevents the concept from representing its objects in some way. If the concept is to be helpful for categorizing objects, it must incorporate some perceptual schema. This is what conceptual likeness is all about. Ockham’s concepts thus offer, after all, a hook out to perception, to use Normore’s phrase.26 In addition to that, one must take into account the signification determined by linguistic conventions. For along with natural concepts, there are other concepts that are acquired on the basis of the words’ representation: our mind, that is, forms representations of words from which, at a second stage, our intellect abstracts common concepts and employs them. Intellect, in other words, subordinates the concept that was formed in this way to the true word that such a concept signifies by nature (naturaliter). Such a process entails that concepts have two different significations, one that is natural and one that is instituted, and that extra propositional factors are necessary to disambiguate between the two. In this respect, Ockham introduces the relationship of reverse subordination: concepts are subordinate to words, and this results into a semantic redefinition and a new employment of those concepts within propositions. One may thus affirm that, in certain cases, the signification of concepts takes place at the level of language pragmatics, since it depends on the conventional use the speaker makes of them. Although the subject might have only a vague understanding of the concept he is using, ‘the Ockhamistic process of reverse subordination allows me to have a very precise concept of carburetor with all and only carburetors in its extension’.27 The same is shown by Ockham’s syncategorematic terms of mental language, whose conventionality seems to violate the naturalness of the cognitive

Mental Acts, Externalism and Fiat Objects

33

process, according to which an act of apprehension grasps directly and with evidence (notitia intuitiva) existent individual objects. According to Panaccio’s interpretation of reverse subordination, syncategorematic terms, such as ‘not’, ‘and’, ‘if ’, do not pose problems to the purity of mental language; rather, they provide a precious insight into the correct understanding of Ockham’s theory as a pragmatic-contextual framework. In this view, mental language is not impure because it contains the conventionality of syncategorematic terms: rather, a language is what it is precisely because such terms function as a signpost within a linguistic context that allows men to freely and responsibly find their way in the world-order (ordinatio). We shall see this feature reappear in the analysis of sensory illusions, prophecies and future contingents, for in all these cases an explanation based on causal necessity seems inadequate to regulate the relationships and the boundaries between explanandum and explanatum. In the Prologue to his commentary on Aristotle’s Physics, Ockham distinguishes four meanings of the term scientia (which is better understood as knowledge rather than science)28: regardless of the differences he envisages, what is relevant to our inquiry is that all four kinds of knowledge share a single principle, that is, they are all founded on the notion of ‘evident cognition’ (notitia evidens), which Ockham describes as follows: ‘I say that an evident cognition is a cognition of some true propositional compound, which is such as to be sufficiently caused, either immediately or mediately, by the incomplex cognition of the terms.’29 No knowledge is evident if the proposition is not true, and evidence, in turn, depends on the way it is caused. Ockham, in short, is making a claim of the following sort: even if no one tells me that the wall is white, I know it is white, because I see the wall being white. On comparing Panaccio’s remark, ‘It’s clear from Ockham’s own explanations that the incomplex cognition of the terms in this context, can be taken to be the incomplex cognition of the referents of the terms.’30 In conclusion, it is possible to ascribe the following strong theses to Ockham: (1) the meaning of words does not in the first instance depend on what we have in mind when we use them; (2) the content of our thoughts depends on the causal chain that produced those mental states; (3) the reference to mental factors does not allow to distinguish knowledge from mere belief. Although Ockham did not set out these three theses as parts of a single unified theory, it is undeniable that they, even taken singularly, point to an externalist framework.

34

Natural and Artifactual Objects in Contemporary Metaphysics

If this reading of Ockham is true, it is also true that we can admit a cognitive theory purely rounded on causal processes that doesn’t need the metaphysical assumption of temporal necessitas per accidens, nor an externalist foundation, because the causal chain simply describes the operative mode of the cognitive processes, also in absence of an external reference: Ockham’s appeal to the counterfactual dependence of an intuitive state on its natural cause is motivated by his views about the nature of causal dependence in general. And, as we have now seen, Ockham holds that such dependence is grounded in certain essential dispositional features of the effect-features that would remain even when its natural or de facto cause does not exist, or at least does not produce it. It is precisely these dispositional features of the effect that the supernatural cases highlight. Even when a given effect – an intuitive cognition, say – is produced by God acting alone, it will by nature be such that it is counterfactually causally dependent on its natural cause.31

False judgements: Deception by God or by the senses? If cognitive acts are founded on a form of intentionality whose nature is causal, because they are externalistically grounded on the direct reference to things, how are we to understand those cases in which objects deceive our senses, as in the much-quoted example of the stick immersed in water that appears to be bent? This is the case of sensory illusions, which, along with prophecies and future contingents, constitute the case studies that allow reinterpreting Ockham on a not-metaphysical ground and establishing whether he is to be considered as an internalist or an externalist, and of what precise sort.32 In relation to this, Karger33 warns us against the flawed reading that Philotheus Boehner34 gave to Ockham’s response to Chatton in his ‘historical’ 1943 article. According to Boehner, Ockham conceded to Chatton that it is possible for God to make us judge a non-existent object as existent, causing in us cognition of that thing, which would then cause the false judgement. On this assumption, Boehner argued that Ockham’s riposte to Chatton consisted in denying the intuitive character of cognition and affirming its abstractive character. On the basis of this bad interpretation, Boehner and subsequent scholars inferred that, in Ockham’s view, intuitive cognitions are infallible while abstractive cognitions are deceitful.35 Under this reading, the infallibility of intuitive cognitions rests on the fact that not even God can produce an act of intuitive cognition such that the latter produces a false judgement. However, Karger points out, no textual evidence may

Mental Acts, Externalism and Fiat Objects

35

be invoked to support the thesis that, in Ockham’s view, intuitive apprehensions are infallible while abstractive apprehensions are not. Such a thesis, moreover, would be in contradiction with a number of other of Ockham’s theses: first of all, abstractive apprehensions, like intuitive ones, are capable of causing evident judgements, not about contingent truths but about necessary truths. Secondly, claiming that an act of apprehension is by nature incapable of causing false judgements does not imply that it is absolutely impossible for it to do so: intuitions are in fact not infallible, because they can, in particular circumstances, cause false judgements. Before turning to Ockham’s theory of sensory illusions, Karger focuses on Ockham’s doctrine of sensory knowledge. Man has two souls: one rational and one sensitive; both souls are within one body, though the former is nonextended, the second is extended. The body is endowed with sensory organs, which can be external (i.e. those located on the body’s surface and thus coinciding with the senses) or internal (i.e. imagination, which is the internal sensory organ). All acts of apprehensions, whether through internal or external senses, are acts by virtue of which a material thing is apprehended. Acts of external senses are acts through which a thing is perceived, while acts of internal senses are acts through which a thing is imagined. But the sensitive soul, Ockham holds, is not capable of adjudicative acts; therefore, no act of sensory apprehension can directly cause an adjudicative act. The process must thus be as follows: sensory apprehension of a thing causes an intellective apprehension of that thing in the intellect, and such an intellective apprehension causes in turn an adjudicative act about that object. As we said at the start, acts of intuitive and abstractive cognition can be both intellective and sensory.36 In accordance with the definitions of acts of intuitive and abstractive cognition, we may say that a sensory act is intuitive if it is capable of being the indirect cause of an evident assent to a present-tensed contingent proposition, while it is abstractive if it is not capable of doing so. All acts of external senses therefore are acts of intuitive cognition, while all acts of internal senses are proper to imagination and thus abstractive. Now, a sensory illusion is a false judgement that attributes a real thing some contingent property the thing in fact does not have; it is, in other words, a judgement indirectly caused in the intellect by a sensory apprehension of that thing. For instance, the visual illusion whereby a stick immersed in water appears to be bent is but the false judgement about the stick being bent, and such a judgement is indirectly caused by a wrong visual apprehension of the stick.

36

Natural and Artifactual Objects in Contemporary Metaphysics

Ockham thus admits of cases in which intuitive cognition causes false judgement; however, it must be stressed that the cause of a false judgement, via intuitive cognition, always depends on the specific circumstances in which our senses come into contact with the object. In the case just mentioned, visual apprehension leads intellect to judge the stick as bent because half of the stick is seen in one medium (air), and the other half is seen in a different medium (water). An intuitive cognition that, in specific circumstances, produces a false judgement about the apprehended object would still by nature be capable of causing a true and evident judgement.37

Prophecies, suppositions and future contingents The deception of senses showed us that it is possible erroneously to attribute to an object a quality it does not possess, and, moreover, that it is possible to avoid this kind of errors of judgement by carrying out a logical and linguistic analysis of the incorrect inference between statements reporting a deceptive state of things.38 The case of prophecies and of the apprehension of future contingents, which are founded on and verified by the assumptions of logica fidei (which in the Tractatus are called suppositiones), presents us instead with a different situation, in which a correct judgement attributes a quality to an object that is not ‘yet’ real. The inquiry thus shifts onto a different ground, one that is not tied to the causal nexus proper to logical inference or to necessitas per accidens.39 The mechanism that takes place in this particular type of cognitive acts makes them particularly relevant to the issue of Ockham’s externalism. Ockham tackles the problem of prophecy in the Tractatus de praedestinatione et de prescientia dei respectu futurorum contingentium.40 His analysis works at the level of the logic of propositional rules: a prophecy is understood as a statement about future contingent events, whose truth-value is determined by revelation. Ockham asks whether prophetic revelations ‘necessarily happen in the way they were revealed, or not’. The key issue concerns the fatalistic implications that derive from propositions whose truth-value is guaranteed by their being objects of divine revelation but which, as far as their content is concerned, are still about indeterminate and ‘open’ states of affairs. The issue of prophecy is addressed through a distinction between two planes. First of all, there is the ontological plane, in which Jonah the prophet, in Ockham’s example, prophesies at t1 that event p will occur at t2. In order for this statement to be recognized as a prophecy, it is necessary to wait for the occurrence of event p at t2, but the need

Mental Acts, Externalism and Fiat Objects

37

for such a verification – which will obtain only a posteriori – does not detract from the fact that even at t1 the statement ‘p will occur at t2’ was to be considered a prophecy, that is, a true statement. Determinism, however, is ruled out, for the statement will be recognized as true only on a different, epistemological plane – that is, when it is possible to verify that the predicted state of affairs has actually taken place. Prophecies, therefore, seem to belong with those statements to which one may apply Ockham’s distinction between past-tensed propositions secundum rem (i.e. propositions about events that actually took place in the past and are thus fixed by necessity, such as ‘Caesar crossed the Rubicon yesterday’) and secundum vocem (i.e. propositions that are only verbally about the past, such as ‘Peter is predestined to salvation’). The prophetic sentence uttered at t1, for example, ‘Petrus praedestinatur’, requires that the predicted event take place at t2; thus, while the sentence may be said to be true already today, insofar as it is guaranteed by the divine auctoritas that revealed it, it still needs to be completed by a series of actions that determine the actual occurrence, at t2, of the state of affairs described at t1. It must therefore be understood as a proposition that is only verbally about the present, and as such escapes both Aristotle’s necessity of the present and the necessitas per accidens of the past.41 A prophetic statement, therefore, does not describe at t1 a fact that is completely determined at t2, but defines the truth-conditions of a future-tensed statement that always presents itself as an ‘implicit conditional proposition’. In case the event that functions as a premise takes place, then the conditional proposition shall be true, and necessarily so. The prophecy predicts the actual occurrence of an event in the following manner: the event will (necessarily) take place in case a (contingent) combination of situations occurs, which causes the event to become real. Thus, in Ockham’s view, the truth-value of prophecies is that of necessitas consequentiae, that is, the necessity that follows from the logical implication between two propositions, and not the ontological truthvalue expressed by the necessitas consequentis, which presupposes that the future event will deterministically occur. A prophecy institutes a relationship between the propositional complex and the res, in the absence of the res themselves. Prophetic statements, therefore, refer by virtue of suppositio to states of affairs that are real to divine cognition, but are not yet so in actuality – at least as far as human intellect is concerned. The implicitly assumed premise is that the divine subject that uttered the prophecy provides sufficient guarantee for the inferences that one may draw from prophetic statements according to the logic of belief.

38

Natural and Artifactual Objects in Contemporary Metaphysics

Theologians from the second half of the thirteenth century (and later Ockham) found suitable elements to legitimize the scientific status of theological knowledge and to guarantee the truth of prophetic sayings in the definition of science that Aristotle offers in the Posterior Analytics.42 In the Prologue to the Expositio in libros Physicorum, indeed, Ockham lists four possible definitions of science, which go back to Aristotle: among them, one in particular claims that science may be also identified with knowledge grounded on assent given to propositions that are considered true. Divine auctoritas, expressed in the form of fide digni statements, functions in this case as the necessary premise (regula)43 that leads the argumentative chain to true conclusions. Such an auctoritas constitutes a normative and prescriptive ground, which can indicate the correct way of performing cognitive acts and free will’s choices (in the non-causal manner expressed by the notion of ‘bringing about’) without thereby causally determining the contents and truth-values of our pieces of knowledge. In Ockham’s view, therefore, prophecies are to be understood as ‘signs’. Prophecies supply a paradigmatic model for the way the res in the ordination must connect to one another through cognitive acts; at the same time, they function as precepts whose normative force applies to human conduct on the path to salvation. Prophecies are thus linguistic indications that outline what might be defined as the ‘pragmatics of theological language’.

Grounding and logica fidei In the context of prophecies, the role of ground is performed by the certainty of the revelation and thus by the divine world-order; in the case of future contingents, it is played by some assumptions that pertain to the logica fidei. Indeed, in the Tractatus, Ockham’s solution to future contingents relies on what he calls suppositiones. It is important to note that the term suppositio acquires here a different meaning from the one it usually has in Ockham’s works. In the Summa logicae, for instance, the term suppositio indicates the referential property that allows words to replace things within propositions, directly anchoring the logical plane to the ontological one. In the Tractatus, on the other hand, Ockham uses suppositiones to indicate the premises of an argumentative procedure, in the modern sense of ‘necessary hypotheses’ or ‘postulates’, which is the meaning attributed to the term suppositio in many thirteenth-century commentaries on the Posterior Analytics (see, for instance, Thomas Aquinas and Robert Kilwardby).44 In that work, indeed, Aristotle defined the term ὑπόϑεσις as follows: ‘Everything which, being capable of proof, is assumed without being

Mental Acts, Externalism and Fiat Objects

39

proved, if admitted by the learner is a Hypothesis, which hypothesis is not an absolute hypothesis but only one with reference to the person who accepts it.’45 In scientific reasoning, therefore, a hypothesis is a statement that is assumed without being proved and functions as the necessary premise to the reasoning. Aristotle stresses that such premises are ‘capable of proof ’, but are ‘assumed without being proved’; in Ockham’s account, on the other hand, they are incapable of proof but they are equally necessary. In the Tractatus, the role of legitimizing and demonstrating the truth of hypotheses/suppositiones is played by the auctoritates and by the epistemic function of ‘suppositional’ statements, within the framework of what might be defined as ‘logic of belief ’. The suppositiones are thus both necessary and indemonstrable, and in this respect they are similar to the postulates of Euclidean geometry: they are propositions whose truth-value is epistemologically indemonstrable and indeterminable, and yet are certain, for they rest on an act of belief that is fundamental to logical-demonstrative reasoning. The suppositiones, that is, have an axiomatic character, which is not grounded on the self-evidence of their truth or on their inductive verification, as in the case of scientific propositions, but on the certainty of belief, which is indemonstrable and, in Ockham’s view, equally self-evident. The epistemological foundation of the logic of belief does not detract from the fact that, once assumed as premises, the suppositiones provide Ockham with a normative and pragmatic rule that allows moving within the framework of propositional logic, while maintaining a reference ground that guarantees the stability of the divine world-order and the possibility to determine the truth-value of propositions with certainty. Such a normative function seems analogous to that performed by syncategorematic terms within a proposition. The former has a stronger force, as it anchors the whole propositional context to a state of the things and not only two terms their external individual references (that is the singular res the terms are referring to) but both have a conventional origin and perform the same normative function to different degrees. Sensory illusions show, and prophecies and future contingents confirm, that there are states of affairs and events that can be verified through causal inference, but within which it is not possible to identify and prima facie distinguish the explanandum from the explanatum. These cases point to the existence of a normative ground. Such an ordinatio does not function according to causal mechanisms only; rather, it seems articulated by (1) causal relations, (2) purely logical relations and,

40

Natural and Artifactual Objects in Contemporary Metaphysics

for what concerns the logic of belief, also (3) relations characterized by a more flexible bond (i.e. that of ‘bringing about’). Such flexibility might be understood in the terms of Barry Smith’s distinction between fiat and bona fide objects – though Ockham would not be comfortable with the ontological rigidity of this distinction.46 The conventionalist and pragmatist reading of Ockham advanced in this chapter highlights the difficulty in univocally establishing boundaries and definitions of the cognitive act’s objects. A useful approach may thus be that of Achille Varzi and, partially, Umberto Eco – whose solution fit the context of a sort of normative ground.47 With regard to the distinctions and the boundaries we can trace through our cognitive acts, Umberto Eco quotes the example of a carver that wants to cut up a cow or a calf: different cultures have produced different cuts of meat, but no culture has ever conceived of a cut that joins the tail and the snout, for instance. This is because reality in itself, before the categories and the language we try to grasp it with, shows certain ‘natural joints’, certain ‘natural lines of cleavage’, that do not impose (according to a rigid causality) but propose (according to the softer notion of ‘bringing about’) certain perspectives and directions to the explorer, while forbidding or making others impossible: ‘We cannot say everything we want. Being may not make sense, but it has many senses: perhaps no sense is compulsory, but certainly some senses are forbidden.’48 By applying Smith’s distinction and the notion of normative ground, the solution proposed here consistently develops the reading of a pragmatist Ockham, who is capable of admitting of grounded belief contexts and yet avoiding Sellar’s critiques to Ockham’s internal conceptualism. The ground thus appears as a bona fide object, which functions as Eco’s natural joints and allows to anchor the linguistic plane to a realist foundation in such a way that propositions can correctly latch on to things in the form of prophetic statements or future contingents, which instead have the nature of fiat objects. In this sense, the normative ground and the relation of ‘bringing about’ that links those statements to the states of affairs or events they describe show a kind of causality that is different from the rigid one that features in natural sciences. The latter is made possible because it is founded on the normative nature of the former. A prophecy, therefore, must not be understood as a statement secundum vocem that anticipates itself as a statement secundum rem; rather, it must be interpreted as a normative proposition, which indicates the rules of the ordinatio along which one must move, and within which, nonetheless, humans

Mental Acts, Externalism and Fiat Objects

41

maintain the possibility to act one way or another. Such a capacity to act, finally, can also be found in that space of freedom and autonomy that is made possible, in Ockham’s view, by the flexibility produced by the lack of boundary between intentional acts, individual concepts and mental terms.

Notes 1 The Ordinatio is the first part of the Commentary on the Sentences of Peter the Lombard, that is, it is the commentary on the Sentences’ first book. While this part of the commentary was compiled by Ockham himself with a view to publication, the remaining part is constituted by notes derived from his lectures, and it is thus named Reportatio. For the exposition of Ockham’s theory of knowledge, we refer to the works of Brower-Toland, Marmo, Mc Cord Adams, Panaccio and Tachau (see bibliography). 2 Why does Ockham limit his analysis to intellective acts, one may wonder? A possible answer is offered by Tachau 1988, p. 116: ‘Ockham reveals that the restriction of his initial discussion to the intellect is far from incidental. Rather, on his view, only this faculty is capable of the adjudicative acts requisite to scientific knowledge (scientia), the immediate objects of which are propositional.’ See also Read 2015, p. 15. 3 Read 2015, p. 19. 4 Compare Ockham, Ordinatio, Prologus, q. 1, art. 1. 5 This is one of the striking novelties of Ockham’s theory: while in his predecessors’ view abstractive knowledge was only of universals, he admits of abstractive knowledge of the singular. 6 Translation by Ockham 1990, p. 23. 7 Compare Panaccio 2012 (2014), pp. 7–12. 8 Ibid. 9 Panaccio 2010, pp. 242–4. 10 Panaccio 2012 (2014), pp. 9–12. 11 Klima 2015, p. 5: ‘One particularly important result of Panaccio’s investigations is that Ockham’s externalism critically hinges on his account of the semantic concept of concept as being causally determined, quite independently from what Panaccio’s calls “the perceptual schema” enclosed in concept.’ 12 Compare Tachau 1988, p. 122. Although scholars before Ockham had argued whether scientific knowledge is characteristically evident, their use of that term had remained relatively untechnical by comparison to Ockham’s. As he defines evident cognition (notitia evidens), however, it is precisely ‘the cognition of some

42

13

14 15 16 17 18 19 20 21 22 23 24 25 26

Natural and Artifactual Objects in Contemporary Metaphysics true proposition (complexum), the sufficient mediate or immediate cause of which is a non-complex knowledge of its terms’ (Ord. Prol, q.1: [OTh: I: p. 5, lin 18-21]): ‘Notitia evidens est cognitio alicuius veri complexi, ex notitia terminorum immediate vel mediate nata sufficienter causari’; for, writes Tachau, the building of scientific knowledge (as evident cognition of necessary propositions) upon premises known evidently. That is because, just as scientific knowledge (scientia), understanding (intellectus) and wisdom (sapientia), evident knowledge concerns complex objects, that is, propositions, which are the intellect’s specific objects. Adam Wodeham, for instance, commented Ockham’s technical definition of notitia evidens as infallible. On Wodeham’s internalist and externalist point of view, see Karger 2015. Compare Sellars 1974, p. 44 ff; in part. pp. 324–7, where Sellars refers to Ockham as ‘the first major breakthrough in the theory of categories’. The new strategy of Ockham, according to Sellars, ‘is illustrated by Ockham’s explanation of such statements as A) Man is a species – Roughly, he construes it to have the sense of B) “Man” is a sortal mortal term – Where, since mental terms are to be conceived as analogous to linguistic expressions in overt speech, the quotation marks are designated to make it clear that in a statement (A) we are mentioning concept rather than using it, as we would be if we were to judge that Tom is a man. To this we must add, I believe, that whereas in (B) the expression “Man” [presents] itself as a name, it need not, and should not, be so construed. Its “depth grammar” places it in a quite different box’. … ‘To apply Ockham’s strategy to the theory of categories is to construe’ – concludes Sellars (p. 327) – ‘categories as classification of conceptual items.’ But this not true, as we will show: if we correctly consider Ockham in its context, also its theory of construe categories is nowadays much more pragmatic oriented of the over-interpretation that Sellars calls, improperly, ‘Ockham’s strategy’. Ockham in fact appears more linked to the use of concepts than Sellars itself could have been thought. Compare Ordinatio, Prologus, q. 1. See Pasnau 1997, p. 62. For a full definition of intuitive knowledge, see Tachau 1988, pp. 123–9. Compare Ockham Sent. I Prol. I (35); Karger 1999, p. 210. Marmo 2001, p. 8. Ivi, p. 7. See Ockham 1991, n. 2, p. 95. Compare Boler 1973. Brower-Toland 2007. See also Panaccio 2015, pp. 173, 183–5. See also Karger 2015. Marmo 2010. Translation in Panaccio 2015, p. 169. Panaccio 2014, p. 172. See also Karger 2014. Panaccio 2014, p. 177. See also Klima 2010.

Mental Acts, Externalism and Fiat Objects

43

27 Ivi, p. 180. For a broader analysis of the conventional use of terms in Ockham, See also Conti 2012. 28 Compare, Expos. in Physic., Prologue, Opera Philosophica IV, pp. 5–6. According to the first definition, science is secure knowledge of a true proposition; according to the second, it is evident knowledge of a true proposition; according to the third, it is evident knowledge of a proposition that is always true and thus necessary; the fourth definition, finally, defines science as evident knowledge of a necessary proposition that derives from necessary premises by way of valid syllogisms. 29 Ordinatio I, Prologue (cited in Panaccio 2015, p. 182). 30 Panaccio 2015, p. 182. 31 Brower-Toland 2017, p. 76. 32 We talk of ontological or metaphysical grounding when a rigid and causal necessity bond links a state of affairs to a subsequent one, which thus gets determined by it; of ‘explanatory ground’, when an explanatory foundation acts as truth-maker; of a ground as a ‘propositional operator’, when it connects the propositions ‘This sheet is white’ and ‘This sheet is rectangular’ to form the proposition ‘This sheet is white and rectangular; of a normative ground, when it provides guidance to pragmatically determine actions, as in the case of prophecies and their performative value in Ockham’s Tractatus de praedestinatione. Further distinctions have been made in relation to the degrees a ground may have, but they are not relevant to our inquiry. Setting such complex issues aside, the notion of ground is understood here as an explanatory criterion that is not exclusively epistemic but is structured along certain natural joints and trends, and which can be reconciled with some realist assumptions and employed to determine lines of action. Such a ground allows establishing relational bridges that connect propositions about intra-worldly states, not between possible worlds or statements located across different worlds. Such a position seems quite similar to Ockham’s one, which is centred on a ground that is certainly rigid, but still avoids determinism by virtue of the linguistic plane onto which propositions of divine foreknowledge and human intellect move. Such a plane involves temporality and the use of term suppositio in the Tractatus, understood as normative principles that direct actions on the basis of the logic of religious belief. On the reading of intra-temporal grounding of hard facts in Ockham and the critiques to Plantinga, see Bottani-Fedriga 2018. On the definition of metaphysical, analytical and normative grounding and the many connected issues, see also Schnieder 2006, Schaffer 2009 and Fine 2010. 33 Karger 1999, pp. 211–15. 34 Compare Boehner 1943. 35 On the role of Chatton in the construction of Ockham’s direct reference theory (and consequently elimination of any sort of species or other intermediate entity), see Mc Cord Adams 1987, p. 82, Pasnau 1997, p. 277 ff. More recently, BrowerToland 2015 has showed the interplay between Chatton and Ockham in the

44

36 37 38 39

40 41 42 43

44 45 46 47

Natural and Artifactual Objects in Contemporary Metaphysics development of their closed theories of object of judgement (and not only the role played by Chatton’s criticism in changing Ockham’s solution). Pasnau 1997, pp. 76, 85. Pasnau 1997, p. 181 ff. See also Karger 2004 and Amerini 2009. Brower-Toland 2015 Since William of Sherwood (thirteenth century), medieval philosophers and theologians call necessitas per accidens the principle according to which, if a certain event or state of affairs is in the past, then there is nothing that one could do about it now. For in Ockham’s contingent framework, necessitas per accidens means that statements of divine foreknowledge, though expressed at the past tense (e.g. Petrus fuit praedestinatus), are not accidentally necessary when the fact the proposition is talking about is really about the future. So it is within our power to act in such a way that God would not have believed what in fact he does. See de Ockham 1978. On prophecies in Ockham, see Fedriga and Limonta 2015. Compare Quodlibeta Septem, IV, q. 4, p. 316, ll. 43–6. Compare An. Post. A, 2, 71b20-23. In the Summa Logicae Ockham employed the term regula to name the guarantee of correctness on the basis of which an argumentative chain can proceed to reach true conclusions (Summa logicae, II, c. 27, ed. cit., p. 334, ll. 10–14). The regula, in other words, functions as a norm to establish the truth-conditions of a proposition. For the discussion about the necessitas per accidens, see Marenbon 2005 and Fedriga and Limonta 2016. An. Post., I 10, 76b29-31. Compare Smith 2001. Compare also Smith 2003, pp. 164–6. The example goes back to the Phaedrus, in which Plato recommends to cut the animal ‘along its natural joints’, avoiding to break any part after the manner of a ‘bad carver’ (quoted also in Varzi 2010, pp. 69–73). Varzi is not so distant from Eco’s realist perspective, though he rejects a fundamental feature of this position, that is, the distinction between fiat and bona fide objects, and maintains that all objects are fiat and thus conventional. In Varzi’s view, the distinction is redundant if its aim is to avoid nihilism or a drift into interpretations, since both kinds of objects rest on the existence of a concrete substrate and thus on a foundation of solid realism onto which the activity of conventional definition of boundaries is performed. Varzi, unlike Eco, does not attribute natural joints and guidelines to the bona fide ground onto which fiat objects are carved, for he understands the ground as a purely neutral substrate: it is true that, without it, definitions could not be; yet, it does not seem decisive in characterizing them as Eco’s ‘natural joints’. What is relevant to our argument, however, is not what the two theories share, but the idea that that the perspective and epistemic nature of the objects of our analysis does

Mental Acts, Externalism and Fiat Objects

45

not undermine their ‘objectivity’ and their necessarily being rooted in something substantial. On Varzi’s notion of boundaries, see Varzi 2007. On Varzi’s and Eco’s riposte, see Varzi 2010, pp. 73–80. On the relationship between these arguments and the tradition of material/formal ontology, see also Bottani and Davies 2007, Introduzione. 48 Eco 1997, p. 39.

References Adams, M. M. (1987), William Ockham (II voll.), Notre Dame University Press, Notre Dame, IN. Amerini, F. (2009), ‘Realism and Intentionality: Hervaeus Natalis, Peter Aureoli, and William Ockham in Discussion’, in S. F. Brown, T. Dewender and T. Kobusch, (eds), Philosophical Debates at Paris in the Early Fourteenth Century, Brill, Leiden, 247–60. Boehner, Fr. P. (1943), ‘The Notitia Intuitiva of Non-existents According to William Ockham’ in Traditio, I, 223–75. Boler, J. F. ‘Ockham on Intuitive Cognition’, Journal of the History of Philosophy 11 (1973): 95–106. Bottani, A. C., and R. Davies (eds), 2007, Ontologie regionali, Mimesis, Milano. Bottani, A. C., and R. Fedriga (2018) ‘Ockham, Plantinga and the Row of Ants’, in M. Szatowski, (ed.), God, Time and Infinity, De Gruyter, Boston/Berlin, 1–17. Brower-Toland, S. (2007), ‘Intuition, Externalism and Direct Reference in Ockham’, History of Philosophy Quarterly, 24: 317–36. Brower-Toland, S. (2015), How Chatton Changed Ockham’s Mind: William Ockham and Walter Chatton Objects and Acts of Judgment, Klima, (ed.), 204–34. Brower-Toland, S. (2017), Causation and Metal Content: Against the Externalist Reading of Ockham, in Pelletier and Roques, (eds), 59–80. Conti, A. D. (2012), ‘Alcune note su logica, linguaggio vocale e linguaggio mentale in Ockham’, in F. Amerini and R. Messori, (eds), Sulle origini del linguaggio. Immaginazione, espressione, simbolo, ETS, Pisa, 133–58. Eco, U. (1997): Kant e l’ornitorinco, Bompiani, Milano. Fedriga, R., and R. Limonta 2015 ‘Prophetae non dixerunt falsum. Spazio percettivo e spazio semantico nelle teorie della profezia di Pietro Aureolo e Guglielmo di Ockham’ Documenti e Studi sulla tradizione filosofica medievale, XXVI, 399–432. Fedriga, R., and R. Limonta (2016), Mettere le brache al mondo. Compatibilismo, conoscenza e libertà, Milano, Jaca Book. Fine, K. (2010), ‘Some Puzzles Concerning Ground’, Notre Dame Journal of Formal Logic, 51 (1): 97–118. Floridi, L. (ed.) (2003), Blackwell Guide to the Philosophy of Computing and Information, Oxford, Blackwell.

46

Natural and Artifactual Objects in Contemporary Metaphysics

Karger, E. (1999), ‘Ockham’s Misunderstood Theory of Intuitive and Abstractive Cognition’, in P. V. Spade (ed.), The Cambridge Companion to Ockham, Cambridge, Cambridge University Press, 204–26. Karger, E. (2004), ‘Ockham and Wodeham on Divine Deception as a Skeptical Hypothesis’, Vivarium, 2: 225–36. Karger, E. (2015), Was Adam Wodeham an Internalist or an Externalist?, in G. Klima (ed.), 186–203. Klima, G. (2010), ‘Indifference vs. Universality of Mental Representation in Ockham, Buridan and Aquinas’, in F. Amerini, F. Marrone and P. Porro, (eds), Later Medieval Perspectives on Intentionality (Quaestio 10/2010), Brepols Publishers/Pagina soc. Coop., Turnhout/Bari, 99–110. Klima, G. (ed.), (2015), Intentionality, Cognition and Mental Representation in Medieval Philosophy: Text and Studies, Fordham University Press, New York (NY). Marenbon, J. (2005), Le temps, la prescience et l’éternité de Boèce à Thomas d’Aquin, Vrin, Paris. Marmo, C. (2001), Guglielmo di Ockham, in Filosofie nel tempo, a cura di P. Salandini e R. Lolli, opera diretta da G. Penzo, volume I: Dalle origini al XIV secolo, tomo II, Spaziotre, Roma 2001, 1157–88. Marmo, C. (2010), La semiotica del XIII secolo. Tra arti liberali e teologia, Bompiani, Milano. Ockham, G. (1978a) Tractatus de praedestinatione et de praescientia Dei respectu futurorum contingentium, rec. Stephanus Brown, St. Bonaventure University, New York. Ockham, G. (1991), Scritti filosofici, a cura di Alessandro Ghisalberti, Nardini, Firenze. Ockham, W. (1990), Philosophical Writings, edited and translated by Philotheus Boehner and revised by Stephen Brown, Hackett, Indianapolis 1990. Panaccio, C. (2010), ‘Intuition and Causality: Ockham’s Externalism Revisited’, Quaestio, 1: 241–43. Panaccio, C. (2012 (2014)), ‘Ockham: Intuition and Knowledge’, in L. Osbeck e B. S. Held (eds), Rational Intuition. Philosophical Roots, Scientific Investigations, Cambridge University Press, Cambridge, 55–74. Panaccio, C. (2015), Ockham’s Externalism, in G. Klima (eds.), 166–85 Pasnau, R. (1997), Theories of Cognition in the Later Middle Ages, Cambridge University Press, Cambridge. Jenny, Pelletier and Magali, Roques (ed.), The Language of Thought in Late Medieval Philosophy. Essays in Honour of Claude Panaccio, Cham: Springer, 2017. Read, S. (2015), Concepts and Meaning in Medieval Philosophy, in Klima, 9–28. Schaffer, J. (2009), ‘On What Grounds What’, in Chalmers, Manley and Wasserman (eds), Metametaphysics, Oxford University Press, Oxford, 347–83. Schnieder, B. (2006), ‘Troubles with Truth-making: Necessitation and Projection’, Erkenntnis, 64, 61–74. Sellars, W. (1974), Essays in Philosophy and Its History, Reidel, Dordrecht/Boston.

Mental Acts, Externalism and Fiat Objects Smith, B. (2001), ‘Fiat Objects’, Topoi, 20 (2): 131–48. Smith, B. (2003), Ontology, in L. Floridi (ed.), 155–66. Tachau, K. H. (1988), Vision and Certitude in the Age of Ockham, Brill, Leiden. Varzi, A. (2007), Confini, in Bottani – Davies (a cura di). Varzi, A. C. (2010), Il mondo messo a fuoco. Storie di allucinazioni e miopie filosofiche, Laterza, Roma-Bari.

47

2

Is the World Really a World of Objects? A Note on Quinean Ontology Antonio Rainone

Premise In recent years Barry Smith (1994) has developed some ontological reflections relative to the distinction between bona fide and fiat objects, in particular in the area of mereology (which will not be discussed here). The former are all those objects to which we recognize an authentic existence ‒ an existence that is independent from our linguistic and cognitive resources; the latter, instead, are entities dependent upon our conventions, and so they are somewhat arbitrary – our mere stipulations. (Think, for example, to a nation with its boundaries or a collection of objects such as the class of all men with red hair: this class is obviously just something we construct by our cognitive resources.) If this distinction is correct, we might then ask ourselves what are in general the real objects in contrast with the artificial ones. We could answer that objects of common sense (trees, animals, tables etc.) are real, while abstract mathematical objects such as classes and scientific objects such as molecules, atoms and electrons are solely artifacts – objects whose existence is only fictional. Perhaps some physicist ‒ or some physicalist philosopher ‒ would say that objects of common sense are real only in a prima facie sense, these objects being nothing but collections of physical particles. It could be so. Nevertheless, we have a great confidence in the bona fide existence of objects and bodies, with well-defined boundaries, that surround us. What is, then, the solution of this question? A solution has been proposed by W. V. Quine. It would seem that for Quine all objects are fiat, mere posits of our epistemic and linguistic resources. This contention may seem paradoxical, because Quine declares himself a realist. However, Quine also argues that the contrast between mere posits and real

Is the World Really a World of Objects?

49

objects is not very plausible since we are confined to our conceptual scheme, and we can know the world only through this scheme: real (or bona fide) for us are all and only the entities that fall within our epistemic, linguistic and scientific conceptual scheme, the roots of which are psychological and evolutionary.

Quine’s cognitive-linguistic theses: Objects as Posits There is in Quine’s approach to language and knowledge a curious assumption regarding objects and, more generally, entities that common sense conceives as genuine objects and entities really existent and provided of the features we experience, for, from the Quinean perspective, the objects of common sense are nothing more than posits. In his celebrated Two Dogmas, for example, Quine said that macroscopic physical objects are, for all we know, ‘irreducible posits comparable, epistemologically, to the gods of Homer’ (Quine 1951, p. 44). But, he also added, we have to consider ‘the myth of physical objects’ superior to other myths because of its efficacy ‘for working a manageable structure into the flux of experience’ (Quine 1951). We can find similar arguments also in Quine (1960b), which is explicitly dedicated to the notion of a posit.1 Are natural objects just mere posits? Do we really invent objects by decree, by mere fiat? This sounds like an idealistic or instrumentalist conception of our knowledge. But Quine refuses any instrumentalist (and obviously idealistic) view of our (common and scientific) language, and he declares himself a realist – indeed a strong advocate of a ‘robust realism’ (Quine 1981b, p. 21). Is there a problem in this connection? Maybe there is. But the story does not end there. For Quine considers that a fortiori posits also the entities and objects of the natural sciences, such as atoms, molecules, electrons and quarks. And much the same holds also for abstract mathematical objects (numbers, classes, functions). (In the last case, the opposite view would be a form of Platonic realism, which Quine rejects, although only within certain limits.) Thus, the objects of common sense, scientific and mathematical objects are all posits of our language for understanding the world. But then ‒ we might wonder ‒ what is truly real in a bona fide sense, independently of our language? Maybe we could accept without problems the thesis that the scientific world and the mathematical realm constitute two domains of postulated entities. But what about common sense physical objects? Physical objects are our ubi consistam: we are talking about physical objects (i.e. the common sense objects) since time immemorial and we believe in their independent existence. How could we regard as posits ‒ that is,

50

Natural and Artifactual Objects in Contemporary Metaphysics

as postulated, invented ‒ entities to which we make constant reference, on which we rely and in an intersubjective way? We do not know that physical objects are posits, Quine would perhaps answer, but they are. Why? The solution to this question is somewhat complex; it has to do with our relations with the external world, on the one hand, and with our linguistic resources, on the other hand.

How posits arise in language We all are human animals equipped with a linguistic capacity, and, just like non-human animals ‒ for example, birds, dogs, apes and particularly howler monkeys ‒ we are able to make sounds for communication in the presence of changes in the external surroundings. But it is not the objects and events that cause our linguistic responses and the animal sounds: these are caused instead by sensory stimulations of our nerve endings, that is, by the excitement of our exteroceptors. Animal responses to sensory stimulations consist in calls and cries, human responses in observation sentences. As Quine put the question in From Stimulus to Science (1995), ‘What I call observation sentences are, at their most primitive, the human counterparts of bird-calls and apes’ cries. Examples are “It’s raining”, “It’s cold”, “Dog!”’ (Quine 1995, p. 22). At such a primitive stage, as Quine says, our utterances are not yet committed to ontological beliefs in the existence of objects out there: utterances are only instinctive responses ‒ at the phylogenetic level ‒ to changes or events in the environment, whereas responses are learned by emulation of adults at the ontogenetic level. Ontological commitment to objects begins to occur only at a very advanced stage of phylogenetic evolution or of ontogenetic language-learning. Practically, there is not much difference between the acquisition of language by the species and by children (and certainly even our ancestors, at some stage of their evolution, should have taken advantage by emulation of their conspecifics). Articles, pronouns, endings for plural are the linguistic devices for reification and referring to objects. In Quine’s terms, we come to the reification of and objective reference to objects out there only when we begin to master sentences, even the simplest, that include at least a pronoun, such as ‘it’, for example. These sentences are what he calls ‘focal observation categoricals’, that present the form ‘Whenever … , it … ’, for example, ‘Whenever there is a Raven, it is Black’, where, notes Quine, ‘the pronoun is essential’ (Quine 1995, p. 28). Sentences of this type are the seminal forms of universal quantification ‘(∀x) (Rx ⟶ Bx)’, which can easily be converted into existentially quantified sentences such

Is the World Really a World of Objects?

51

as ‘¬(∃x) (Rx Λ ¬Bx)’. At this level of language-learning we can individuate, according to Quine, the roots of reference ‒ that is to say, our ability to refer to external objects in our environment. Once we reach mastery (both as a species and as children) of our language and its devices ‒ such as anaphoric reference by pronouns, universal and existential quantification, direct reference by singular terms or descriptions by definite articles ‒ we will have achieved a sophisticated way to talk of objects, which means that we have an ontology. But objects are posits, as Quine stated in Two Dogmas and in Posits and Reality. The question we might ask ourselves is then the following: is our ontology just a product of our language and its devices ‒ a set of posits ‒ or is it a domain independent of our language or ‘conceptual scheme’? In other words, are there really things such as mountains, lakes, trees, tables or are they mere postulations and our linguistic ways of splitting up the world? (Here we can perhaps recall the famous theses of Nelson Goodman on ‘ways of worldmaking’: cf. Goodman 1978.)

Posits as psychological and biological devices It is not easy to answer these questions. To begin with, according to Quine, we cannot do without speaking in terms of objects. Objects are part of our way of cutting up the world as the undifferentiated source of our sensory stimulations; we could not communicate without talking about objects. And objects are so deeply rooted in our speech that we would not know how to do without them. Here’s the famous argument which opens Speaking of Objects (1969): We are prone to talk and think of objects. Physical objects are the obvious illustration when the illustrative mood is on us, but there are also all the abstract objects, or so there purport to be: the states and qualities, numbers, attributes, classes. We persist in breaking reality down somehow into a multiplicity of identifiable and discriminable objects, to be referred to by singular and general terms. We talk so inveterately of objects that to say we do so seems almost to say nothing at all; for how else is there to talk? (Quine 1969, p. 1)

However, this argument is not conclusive, since Quine continues by noting that ‘it is hard to say how else there is to talk, not because our objectifying pattern is an invariable trait of human nature, …’ (ibid.; emphasis mine). And then he continues recalling the problems of radical translation. The point emphasized here is particularly important: our objectifying pattern is not an invariable trait of

52

Natural and Artifactual Objects in Contemporary Metaphysics

human nature, says Quine. Why should it not be an invariable trait? Of course, this is the central tenet of the thesis of the indeterminacy of radical translation. But what evidence do we have for thinking that our speaking of objects is not an invariable trait of human nature? This is a mere hypothesis that seems to go in a different direction from the one where Quine intends to go; for he is persuaded that we could hardly conceive, even if only hypothetically and on the basis of strong evidence, which however is never forthcoming (recall the famous there is not fact of matter), of a language where perverse sentences such as ‘lo! rabbithood’ or ‘it is rabbiting’ or ‘here undetached rabbit parts’ are just as plausible as ‘here a rabbit’.2 In this connection we can think of Gestalt theory and its way of considering perception of objects as whole bodies (see below). It is true that ‘we are bound to adapt any alien pattern to our own in the very process of understanding or translating the alien sentences’ (ibid.); but the ascription of what are, from our standpoint, ontologically odd beliefs is a last resort. A form of ‘charitable’ attitude ‒ in Davidsonian style ‒ should perhaps grant also to a radically different linguistic community our own ‘objectifying pattern’ and the belief in objects out there, due to the similar human nature. Hereinafter I will try to explain why on Quinean premises. Human animals posit objects already at the level of primitive experiences and mentality ‒ and this seems, as Quine notes, a natural trait of the human species, without which no speech or communication would have ever been possible. As Quine clearly argues on the first page of Word and Object, even though our knowledge comes from sensory stimulations, during evolution and thanks to language, there arises an entification/reification process relative to the external world from which our stimulations derive. Here’s his argument: This familiar desk manifests its presence by resisting my pressures and by deflecting light to my eyes. Physical things generally, however remote, become known to us only through the effects which they help to induce at our sensory surfaces. Yet our common-sense talk of physical things goes forward without benefit of explanations in more intimately sensory terms. Entification begins at arm’s length; the points of condensation in the primordial conceptual scheme are things glimpsed, not glimpses. (Quine 1960a, p. 1; emphasis mine)

This is the pivotal argument that undergirds Quine’s naturalistic-oriented epistemology. Like all other animals, we acquire knowledge from external forces that primarily cause our sensory stimulations. But ‒ and this is the point that Quine emphasizes ‒ stimulations of our body surfaces are only a part of the story,

Is the World Really a World of Objects?

53

for already at a prelinguistic level, in the primordial conceptual scheme, our ancestors were oriented to perceive bodies and objects out there. True, entification is not yet reification, but only a prior stage of reification by language; in any case, it entails beliefs about existing things in the external world. Then it would be a mistake to think, as Donald Davidson (1974) does, that Quine would claim a dualism between empirical content and conceptual scheme. Indeed, for Quine such a dualism does not hold good: if it is true that surface irritations are the causes of our knowledge, it is equally true, in Quine’s view, that the knowing subjects perceive bodies, not sense-data or stimulations in terms of which bodies would need to be accounted for. When the time comes for the development of language, human beings are talking about objects, not about stimulations or sense-data. Here’s how Quine answers Davidson: Our typical sentences are about bodies and substances, assumed or known in varying degrees, out in the world. Typically they are not about sense data or experiences or, certainly, surface irritations. But some of them are elicited by surface irritations, and others are related to surface irritations in less direct and more tenuous ways. … Nobody could suppose that I supposed that people are on the whole thinking or talking about the triggering of their nerve endings; few people, statistically speaking, know about their nerve endings. (Quine 1981c, p. 40)

In sum, according to Quine, we are naturally talking and thinking about bodies and objects ‒ and not at a linguistic community level, but as human beings. This point was also stated by Quine in The Roots of Reference (1974), where he deals with this problem from a psychological and biological perspective. In psychological terms he refers in particular to Gestalt psychology: unlike the classical empiricism of Hume and Berkeley, Gestalt experiments have taught us that we are naturally prone to perceive ‘structured wholes’ without stopping at smells, noises, feels, flashes, patches of colour ‒ in other words, we perceive bodies (Quine 1974, pp. 1–2). Moving between Gestalt and behaviourism, Quine explains the way in which we perceive things in the world, even if our perceptions derive from stimulations of our bodily surfaces. But there’s more. Our propensity to perceive bodies (and objects as soon as we learn to master the language) is not only based on our psychological mechanisms; it derives also from our biology. As Quine claims, It is no wonder that bodies, bodily identity and bodily persistence, are the mainstay of ontology. Bodies, for the common man, are basically what there are; and even for esoteric ontologist bodies are the point of departure. Man is a

54

Natural and Artifactual Objects in Contemporary Metaphysics body-minded animal, among body-minded animals. Man and other animals are body-minded by natural selection; for body-mindedness has evident survival value in town and jungle. (Quine 1974, p. 54; emphasis mine)

Survival value is then another trait of our positing bodies, and later reification of physical objects ‒ of our ontology. This point was emphasized by Quine also in From Stimulus to Science (1995), where he notes that ‘natural selection had heightened man’s responsiveness and that of other animals to the sight and smell of bodies, for there is survival value in prompt awareness of prey and predator survival’ (Quine 1995, p. 35). (For example, think of the difference between perceiving the appearance of a tiger and feeling the presence of a tiger.) We unconsciously posit bodies and then objects for psychological and biological reasons. How then could linguistic communities that did not develop the natural attitude toward the partition of the world into discrete bodies and objects ever exist? To think the contrary would be a form of unmotivated relativism (and indeed Quine often received accusations of relativism: cf., for example, Alston 1979) or, perhaps, something like a theory of the pre-objectual mentality. With this last italicized expression I am referring to the pre-logical mentality of Lucien Lévy-Bruhl, a concept that Quine does not accept (cf. Quine 1960a, p. 58). As Quine notes in Word and Object, if someone asserts a sentence that in our language would be translatable as ‘p and not p’, then this would mean that the linguist has committed an error in translation, not that the speaker is illogical (Quine 1960a, p. 58). Recall Quine’s famous argument, according to which ‘one’s interlocutor’s silliness, beyond a certain point, is less likely than bad translation’ (ibid., p. 59) ‒ which is a formulation of the well-known principle of charity. In this connection we might raise the following question: why should we not allow the possibility of contradictions while accepting – albeit as a last resort – the possibility of ontologies or conceptual schemes devoid of objects? Quine himself says that we, as human animals, can’t do without talking or thinking of bodies and objects. Objects appear to be necessary and indispensable in the same way as logically correct assertions, and perhaps even more so, given that they are part of our natural responses to sensory stimulations. Our ‘immemorial doctrine of ordinary enduring middle-size physical objects’ is for Quine a theory ‘second nature’, in contrast with ‘deliberate’ theories such as physics and chemistry (Quine 1960a, p. 11). This means, I believe, that objects really exist out there, at least for us: they are an essential part of our world ‒ the world of the human race in general, and therefore also of people unknown to us. Physical objects are

Is the World Really a World of Objects?

55

posits in an epistemological sense: they are the way we acquire knowledge of the external world. True, bodies and objects ‘are not given in our sensations’, as says Quine, but they are only ‘inferred from them’ (Quine 1974, p. 1); nevertheless they are all we have, for psychological, biological and linguistic reasons, and do not exist a ‘cosmic exile’, independent of our beliefs, theories and conceptual scheme from which we may look at the world in itself (Quine 1960a, pp. 275–6). The only world we know is our epistemic world ‒ and this is a Quinean thesis of Kantian type or, in other words, a form of what Hilary Putnam later called ‘internal realism’.

Ontology, realism and indispensability According to Quine there are also theories that are ‘deliberate’, rather than ‘second nature’ in the way that the common sense conceptual scheme is. In general, scientific theories such as those of physics and chemistry are of this sort. In these theories, we find so-called theoretical entities (molecules, atoms, electrons, quarks, etc.). And entities of this type arise from and are an extension of the ontology of bodies and objects of common sense (recall the famous maxim of Two Dogmas that ‘science is a continuation of common sense, and it continues the common-sense expedient of swelling ontology to simplify theory’: Quine 1951, p. 45). In From Stimulus to Science Quine put his argument in this way: Reification has continued beyond bodies and substances, adding new sorts of objects and new quantified conditionals to tighten the logical structure of science. Atoms tentatively entered in the ontology already in antiquity, in easy analogy to the primordial bodies. Abstract objects were reified even earlier, it would seem, though there was nothing so suggestive in the way of analogy. (Quine 1995, p. 39)

But are there really theoretical entities or are they only a way to organize our experimental procedures and results, as instrumentalists claim? Quine’s answer to this question is sufficiently clear. Our knowledge is confined to our theories about the world, in the same way in which our beliefs in external objects are dependent on our conceptual schemes. Our best theories are the best way of describing reality; if they are confirmed by successful predictions, then they are probably true, even though only until contrary evidence appears, and so we can accord, even if only provisionally, a real existence to entities that such theories are hypothetically positing. Scientific method ‒ according to Quinean naturalism ‒ is

56

Natural and Artifactual Objects in Contemporary Metaphysics

the unique and final arbiter of what exists out there; as Quine writes in Word and Objects, we can’t consider the standpoint of scientific theories as makebelieve, for we can never do better than occupy the standpoint of some theory or other, the best we can muster at the time. What reality is like is the business of scientists, painstakingly, to surmise, and what there is, what is real, is part of that question. The question how we know what there is simply part of the question … of the evidence for truth about the world. The last arbiter is so-called scientific method, however amorphous. (Quine 1960a, pp. 22–3)

A final point. Quine often spoke of benefits and convenience as regards the adoption of an ontology. And this also applies to the ontology of logic and mathematics. We could then consider his views as an elaborated interpretation of epistemic rationality in terms of practical rationality about decision-making (even if it is not always conscious) where cost–benefit analysis is of fundamental importance. As objects are indispensable and convenient for our communication at the common-sense level, so logic and mathematics are indispensable and convenient for our knowledge – and for this reason we have to accept their ontologies – and so too are the particles of science indispensable and convenient for our knowledge of the world.

Notes 1 This Quinean topic has not always received the attention it deserves. In recent years it has been discussed by Kemp (2006), pp. 110–15, and Hylton (2007), pp. 18–22. Here I have taken into account their comments, although my view is somewhat different. 2 On this problem see also the argument of Putnam (2015), section 10. If I understand his argument, he holds that in the psychology of perception there is no evidence for such perceptions as un-detached parts of rabbits, but, presumably, only of rabbits as bodies and objects out there.

References Alston W. P. (1979), ‘Yes, Virginia, There Is a Real World’, Proceedings and Addresses of the American Philosophical Association, 52: 779–808. Davidson D. (1974), ‘On the Very Idea of a Conceptual Scheme’, Proceedings and Addresses of American Philosophical Association, 47: 5–20. Repr. in D. Davidson, Inquiries into Truth and Interpretation, Clarendon Press, Oxford 1984, pp. 183–98.

Is the World Really a World of Objects?

57

Goodman, N. (1978), Ways of Worldmaking, Hackett, Indianapolis. Hylton, P. (2007), Quine, Routledge, New York. Kemp, G. (2006), Quine: A Guide for the Perplexed, Continuum, New York-London. Putnam, H. (2015), ‘Naturalism, Realism, and Normativity’, Journal of the American Philosophical Association, 2:312–28 Quine, W. V. (1951), ‘Two Dogmas of Empiricism’, The Philosophical Review, 60: 20–43; repr. in W.V. Quine, From a Logical Point of View, Harvard University Press, Cambridge (MA), 1953; Second Edition Revised, 1980, pp. 20–46. Quine, W. V. (1960a), Word and Object, The MIT Press, Cambridge (MA). Quine, W. V. (1960b), ‘Posits and Reality’, in S. Uyeda, (ed.), Basis of the Contemporary Philosophy, 391–400; repr. in W.V. Quine, The Ways of Paradox and Other Essays, Random House, New York, 1966, Revised and Enlarged Edition 1976, pp. 246–54. Quine, W. V. (1969), ‘Speaking of Objects’, Proceedings and Addresses of the American Philosophical Association, 31 (1957–8): 5–22; repr. in W. V. Quine, Ontological Relativity and Other Essays, Columbia University Press, New York and London, pp. 1–25. Quine, W. V. (1974), The Roots of Reference, Open Court, La Salle (IL). Quine, W. V. (1981a), Theories and Things, The Belknap Press of Harvard University Press, Cambridge (MA). Quine, W. V. (1981b), ‘Things and Their Places in Theories’, in Quine (1981a), pp. 1–23. Quine, W. V. (1981c), ‘On the Very Idea of a Third Dogma’, in Quine (1981a), pp. 38–42. Quine, W. V. (1995), From Stimulus to Science, Harvard University Press, Cambridge (MA). Smith, B. (1994), ‘Fiat Objects’, in N. Guarino, L. Vieu and S. Pribbenow (eds), Parts and Wholes: Conceptual Part-Whole Relations and Formal Mereology, 11th European Conference on Artificial Intelligence, Amsterdam, 8 August 1994: European Coordinating Committee for Artificial Intelligence, Amsterdam 1994, 15–23.

3

Spatial Fictionalism. A Solution to the Grounding Problem Nicola Piras

Introduction The world the spatio-temporal domain of quantification is full of objects. Each of these objects is the only occupant of the region of space it actually occupies; scarcely any human denies this. One might think that there is a sort of metaphysical law that bans co-occupation. I call it the No Co-occupation Law (NCL), though it is also known as the No-coincidence Law and the No Co-location Law: the abbreviation NCL works as well in each case. Despite its wide endorsement among humans, almost all philosophers deny the NCL. Indeed, its negation is called the ‘standard account’ (Burke 1994; also, Wasserman 2015). But here some problems arise: how is it possible for two objects to be coincident? What grounds their difference given that they share all the same material properties and parts? Questions of this sort feed what is called the ‘grounding problem’ (GP) (Bennet 2009b). There are two kinds of solution to the GP: either accept co-occupation and try to explain the difference or reject co-occupation and try to respond to counterexamples. There are several ways to explain co-occupation, but I do not find any of them completely compelling: they make too generous ontological commitments and have very extravagant metaphysical consequences. I also think that giving up the NCL is too extreme a revision of our intuitions. My aim here is to defend the NCL against counterexamples. The solution I shall seek to defend may be called ‘Spatial Fictionalism’, namely the claim that, because objects depend upon their boundaries and every boundary is of the fiat sort, that is, human-dependent, objects inherit the same ontological status.

Spatial Fictionalism

59

Thus, whereas objects are just human-dependent demarcations of qualified variation in space, and hence of the fiat sort, the stuff from which they are picked out is of the bona fide sort. Since fiat objects do not occupy space but are merely located in space, all of the co-occupation puzzles disappear. Despite appearances, Spatial Fictionalism revises our intuitions about the world to an extent less than giving up the NCL does. My strategy is as follows. I begin by stating the NCL. I then outline the counterexample to the NCL and how it gives rise to the GP. I sketch some prominent ways of defending the NCL, showing why they fail, or why they are inadequate on independent grounds. Finally, I set out Spatial Fictionalism and how it solves the problem.

No Co-occupation Law The NCL states that two different objects cannot occupy the same region of space at the same time. In case of violation of that law it is said that two objects coincide. In order to appreciate the force of the NCL, let me state some very basic assumptions. 1. Assume spatial dualism. There are only two kinds of entities: (i) entities that occupy other entities – ordinary things, lumps of stuff, events; call them objects; and (ii) entities that can be occupied by other entities; call them regions. Objects and regions are metaphysically irreducible, that is, they are neither identical to each other nor related by a reduction relation, such as a grounding relation, ontological dependence or supervenience relation. 2. Stipulate also that the occupation relation links at a given time t an object to a region with the same size, shape and topological properties1; for example, an object x shaped as a sphere at t occupies a region shaped as a sphere at t. We may represent this intuition with a time-indexed two-place relation that may be read as the following: at the very same time in which x occupies y, no part of x occupies a subregion not belonging to y and no subregion of y is free from a part of x. 3. Assume the lexical core of standard mereology. Parthood is a two-place relation, axiomatized as a partial ordering: reflexive (everything is part of itself), transitive (any part of any part of an object is part of that object) and antisymmetrical (distinct objects are not parts of each other).

60

Natural and Artifactual Objects in Contemporary Metaphysics

With parthood as our primitive, we can define other important mereological predicates, such as proper parthood, so that if an object is a proper part of another, then that object is not part of the first, overlap, when two objects have a part in common, and discreteness when two objects have no parts in common. Let me finally return to the NCL. The principle formalizes a very intuitive concept: if an object occupies a region at a given time, whatever there is in such region at that time is that object. For instance, suppose I lose my phone. In order to find it I just need to discover the spatial region it currently occupies. And, obviously, also the converse holds: if I find my phone, I encounter its region as well. How many other objects can I find in that region before I move my phone to my pocket? None, for the NCL maintains that identity of location is sufficient for identity in general. Clearly, you can find, as well as the phone, all of its parts. But since the NCL entails extensionality, that is, the thesis according to which an object is just its parts taken as a whole, this is not a problem at all. Yet, sharing location is not necessary in order to be the same object. Consider a Max Black world (1952), namely a world in which two otherwise indiscernible spheres are located in two different regions. These regions have the same size and shape. The two spheres also have all the other properties in common: colour, mass and so forth. Are they the same object? The NCL does not supply an answer, insofar as it neither bans nor affirms multi-location, that is, the thesis according to which two objects may be identical even if they do not share the same region. According just to the NCL two objects that share the same region are the same. Indeed, the NCL is a tensed instantiation of the Principle of Identity of Indiscernibles (PII), referred to regions and their occupants: PII: For every x, for every y, if x is P iff y is P, then x is identical to y. The PII claims that sameness of properties entails sameness in general, but it does not claim the converse, that is, the Principle of Indiscernibility of Identicals (PII*), which conversely states, PII*: For every x, for every y, if x is identical to y, then x is P iff y is P. Its locative tensed instantiation would claim that two objects are the same if and only if they share the same location in space at a given time. Hence, it would ban multi-location. Thus, a good formalization of the NCL might be as follows: NCL: For every x, for every y, for every region r at t, if x occupies r at t and y occupies r at t, then x and y are identical.

Spatial Fictionalism

61

According to the NCL, being co-located in a certain region is sufficient for identity. Since being located in r at t is the only property that matters for sufficiency of identity among objects, in the if-clause of the NCL there is a conjunction instead of a biconditional as in the PII. Although the NCL seems to capture our intuitions in following the PII, there is a serious challenge for it.

Against the NCL The NCL does not have a wide acceptance among philosophers. Critics of the NCL can be divided into two main categories: (i) those who believe that the NCL does not hold in the actual world and (ii) those who believe that it may but that it is not necessary. Those who believe the NCL is not valid even in our world usually argue that every region, when occupied, hosts at least two objects: a thing and the stuff of which it consists (Wiggins 1968; Fine 2003). Some of them argue that every region can host an object and a large number of events, for example, a ball, its rotating, its getting warm and so forth (Casati and Varzi 1999; Davidson 1969). Those who believe that the NCL may hold in our world although it is not necessary argue that the matter of which an object consists may be penetrable. If so, two objects can interpenetrate. In other words, two objects can coincide. Here I defend only the possibility of the NCL, even though I mention something about the question of necessity. The usual argument against the NCL runs as follows (Wiggins 1968): imagine a tree and the aggregate of cellulose of which it consists and suppose that one cellulose cell is destroyed. Clearly, the tree can survive such an event, whereas the aggregate of cellulose cannot. Mutatis mutandis, consider the case in which the tree is cut down and eventually transformed into a table, without losing any cells. Contrary to the former example, the tree ceases to exist, while the aggregate of cellulose survives.2 Since the tree and the cellulose have different properties, they share the same region at the same time without being identical. Thus, the NCL is violated. In order to avoid such counterexamples, one can either narrow the scopes of quantifiers to sortals or provide a mereological version of the NCL: Sortal relative-NCL (SR-NCL): For every x, for every y, belonging both to the same sortal s, for every region r at t, if x occupies r at t and y occupies r at t, then x is identical to y.

62

Natural and Artifactual Objects in Contemporary Metaphysics

Mereological version-NCL (MV-NCL): For every x, for every y, for every region r at t, if x occupies r at t and y occupies r at t, then x overlaps y. That is, the SR-NCL states that two different objects can share the same region as long as they do not belong to the same sortal. Since two different trees belong to the same sortal, they cannot be co-located, whereas a tree and the aggregate of cellulose of which it consists can be so, insofar as cellulose belongs to the a different sortal. The MV-NCL states that x and y can be co-located just in case they have all their parts in common in r at t. If so, since the tree and the cellulose of which it consists share all their parts, they do not contravene the principle. Both SR-NCL and MV-NCL appear to be very intuitive principles. Nevertheless, the fact that two different objects that do not overlap or belong to the same sortal cannot be co-located may be a contingent fact due to the physical laws of our world (Zimmerman 1996; Sider 2001). If so, the modal strength of the NCL has to be weaker: Weak-NCL: Possibly, for every x, for every y, for every region r at t, if x occupies r at t and y occupies r at t, then x is identical to y. Weak-NCL may also be challenged in various ways by assuming one of the following principles: (a) Necessarily, every thing is made of stuff. (b) Necessarily, every amount of stuff constitutes an object. (c) Necessarily, there are scattered objects. According to (a), immaterial things as ghosts and abstract objects such as numbers are necessarily outside space and time, because only objects made of stuff can exist in space. Thus, every inhabited region hosts a thing and its stuff as a matter of necessity. According to (b), there is no fixed minimum amount of stuff needed to be a thing. If so, every lump of stuff counts as a thing, no matter how large it is. According to (c), things have to be composed of topologically disconnected parts, regardless of whether such parts are close to each other. Thus, a little lump of stuff also counts as a thing insofar as it is a part of a bigger one. Therefore, also Weak-NCL may be sortally or mereologically relativized, in order to avoid such counterexamples. Sortal relative-Weak-NCL: Possibly, for every x, for every y belonging both to the same sortal, for every region r at t, if x occupies r at t and y occupies r at t, then x is identical to y.

Spatial Fictionalism

63

Mereological relative-Weak NCL: Possibly, for every x, for every y, for every region r at t, if x occupies r at t and y occupies r at t, then x overlaps y. Such relativizations may seem in substantial agreement with our modal intuitions, either way. Nonetheless, giving up the original, stronger formulation of the NCL turns out to be a source of modal problems.

The grounding problem Consider the GP (see, inter alia, Olson, 2001; Fine 2003; Bennet 2009a, b; DeRosset 2011; Einheuser 2011; Sutton 2012; Sidelle 2016): GP: how is it possible for two co-located objects to share all the same parts, shape, size, mass, causal history and so forth, and nevertheless be different? Consider the putative example due to Wiggins about the tree and the cellulose. Since the two co-occupants share all the same actual material parts they seem to be the same object, perhaps differently described. Unfortunately, they differ in some of their modal properties and in some of their persistence conditions. Thus, assuming the PII, the tree and the cellulose are different. So, there are two different objects with all the same parts, shape and location that are nevertheless different. What grounds such difference? It seems natural to think that modal properties and persistence conditions are grounded on or depend upon the actual and the material ones. In spite of the intuition, tree and cellulose share all the actual and material properties and nevertheless are different. How is it possible? The right way to avoid such a conundrum seems to be to resist the counterexamples to the NCL. Indeed, the NCL would prevent the GP for it denies that a region of space can host more than one object, whereas all the other narrower or weaker formulations obviously do not. Nonetheless, even if one were to reject all the counterexamples, the GP can be refined in order to challenge the NCL also. GP*: how is it possible for the very same portion of reality to be the source – or to be the owner – of two inconsistent sets of de re modal properties? The problem is deeper than it seems, since when two properties are mutually inconsistent, there are two objects, as a consequence of the PII. Consider again the example due to Wiggins. In this case we have two sets of properties: the first set collects all the modal properties we attribute to trees: ‘possibly, it survives annihilation of one part’, ‘necessarily, its parts are

64

Natural and Artifactual Objects in Contemporary Metaphysics

topologically connected’ and so forth. The second set collects all the modal properties we attribute to aggregates of cellulose: ‘necessarily, it does not survive the annihilation of one part’, ‘possibly, its parts are not topologically connected’ and so forth. The problem is deeper than it seems, again. The objects do not differ just in some properties; they differ in the properties that define what such objects are. As Bennet (2009a) and Sidelle (2016) point out, the intersection of modal properties and persistence conditions is what is referred to as sortal properties, namely the properties that establish what an object is: under what conditions it survives and what changes it can undergo without losing its identity across time and worlds. This means that if a certain portion of reality has a consistent conjunction of modal properties and persistence conditions, it belongs to a certain sortal. In other words, when you encounter such a conjunction of properties that is internally consistent, then you find a certain object. More abstractly, to every consistent set of modal properties and persistence conditions instantiated by the material content of a certain region there corresponds a sortal, whose essential properties are those collected in the set. Thus, in the case of the tree and cellulose, we have two internally consistent sets of properties, each of which corresponds to a sortal. Why is it that only modal properties and persistence conditions are sortal? Assume that modal properties are those properties that ensure the survival of an object, or of its counterpart, across possible worlds. Persistence conditions ensure the survival of an object across time. Consider a tree. It seems that also other properties define its membership in the set of tree: consistence, texture, atomic structure and so forth. But only the modal and persistence formulations of such properties are sortal, for they define what must be identical across worlds and time in order to preserve the identity conditions of such an object. Consider, say, the non-modal property P of an x of sortal S. If x loses P, since P is not necessary, x does not lose its membership of sortal S and nothing significant for the surviving of x happens. Suppose now that x loses the property of being necessarily P for belonging to the sortal S. Since P is necessary for x to be what it is, namely a member of S, x is no longer what it was before losing P. Therefore, since membership of a sortal is a function of modal properties, only the modal instantiation of a property may be a sortal one. Once the definition of sortals is fixed, it is necessary to understand how many sortals there are. In order to avoid metaphysical arbitrariness, I assume

Spatial Fictionalism

65

that there is a sortal for every consistent set of modal properties and persistence conditions.3 Unfortunately, the world is full of inconsistent sets of modal properties located in the same region, as in the case of tree and cellulose. What is to be done about this inconsistency? Apart from a desperate appeal to paraconsistency, there are three approaches or solutions: (1) we accept the inconsistency and reject the NCL, thus providing a reason for the modal difference instantiated by the same portion of reality; (2) we reject the inconsistency and try to defend the NCL against its critics in order to motivate how the content of one region may instantiate two different sets of properties without contradiction or (3) we deny one or both sets of properties and thus defend the NCL.

Prominent solutions to the grounding problem Many philosophers have tried to defend the NCL against its counterexamples. However, I do not find any of the already known solutions really compelling. In what follows I review and then briefly reject some of them.4 The following list of arguments can be divided into four subsets on the basis of the conclusions they reach. Some of them may belong to more than one subset: 1. There is no fact of the matter about which one of the coincident objects exists and which does not; 2. One of the coincident objects does not exist; 3. Exactly one of the coincident objects does not exist; 4. There is more than one object but that fact does not contravene the NCL. The arguments in the first set (b, c) endorse the thesis according to which the identity across worlds and time differs from world to world and from time to time. Identity crucially depends on which world or on which time we are considering. The arguments in the second set (a, c, g, h) may be employed to deny the existence of whatever object we want: either things, or lumps of stuff, or simples arranged thing-wise. It just depends on what ontology one supports. Nonetheless, such arguments have been employed in the literature to deny the existence of certain kinds of objects. Nevertheless, the arguments in the second set can be

66

Natural and Artifactual Objects in Contemporary Metaphysics

reformulated in order to deny another kind of objects just replacing the subject of the premises. I guess such variability of the conclusion is an intrinsic weakness of such kinds of arguments. The arguments in the third set (e, f) may be employed just to deny a particular kind of objects. In that case it is impossible to replace the subject and still have a valid argument. The arguments in the fourth set (b, c, d, g) reach the conclusion that even if there are two or more objects in the same region, that is not a proper violation of the NCL. a. Causal over-determination (Merricks 2001, 2017; Sider 2003; Horgan and Potrč 2008): 1. 2. 3. 4.

Object and lump of stuff do the same causal work; Just one is needed to preserve the worldly causal chain; Therefore, one of them does not exist; The NCL is not violated.

Those who support the causal over-determination argument against co-occupation, usually implicitly, or sometimes explicitly, assume that to be is to be causally active. In fact, the argument can be undermined if causal inert objects exist. If so, one can say, well, either x or y causes z; but such arguments do not prove that there exists only one between x and y. It is possible that both exist but that only one of them is causally efficacious. That’s only half of the story. Suppose one is really sure that existent objects are just causally efficacious objects. The problem now is to understand in a non-arbitrary manner which one between x and y causes z. I think there is no fact of the matter that permits us to decide which one between tree and cellulose causes my pain when I walk into them. Consider a counterfactual analysis of causation: ‘x causes z; iff had x not occurred, z would not have occurred.’ Which one between tree and cellulose has to replace x? Usually, it is argued that smaller objects are more fundamental than bigger ones. Cellulose, arguably, is just a plural name for a collection of cells; therefore cellulose exists and the tree does not. But, this is just an assumption, not a conceptual truth or a conclusion. Hence, to hold that smaller objects are causally efficacious because they are more fundamental than bigger ones is to beg the question since we are implicitly assuming two conclusions: (1) smaller objects are more fundamental

Spatial Fictionalism

67

than bigger objects; (2) fundamental objects are causally efficacious, whereas derivatives are not. b. Counterpart theory (Lewis 1986, 2003): 1. The object x has counterparts x1 and x2: (a) x1 is capable of surviving the annihilation of one of its parts, as a tree but not as an aggregate of cellulose; (b) x2 is capable of crumbling away, as an aggregate of cellulose but not as a tree; 2. Tree and cellulose are not two objects but the labels of two different counterpart relations; 3. Therefore, there is just one object with two different counterpart relations; 4. The NCL is not violated. According to Lewis (2003, p. 28), tree and cellulose are just two ways of describing the same portion of reality. Both the names label a sortal, each with some essential properties. Since essential or sortal properties are modal properties and since according to Lewis modal properties can be analysed through counterpart relation theory, there is no problem at all. We have a certain portion of reality and two labels for it: according to one label it has certain counterparts in certain worlds that behave in such and such ways and according to the other label there are other counterparts that behave in very different ways. Fair enough. My problem with that solution is that it works only under a certain assumption: the two objects diverge only in modal properties. As I noted, the two entities diverge even in persistence conditions: are those conditions modal properties or are they reducible to modal properties? If the answer is ‘yes’, then the solution by Lewis works well. Otherwise it does not. It is beyond the aim of the chapter to try to capture the essence of persistence conditions. It is enough to say that the solution by Lewis works only in case persistence conditions are reduced to modal ones. c. Conventionalism about modal properties (Heller 2005; Einheuser 2011; Sutton 2012): 1. Modal properties are just conventions; 2. There are no two sets of different modal properties, each of which defines the existence of a new object; 3. There is just one object without modal characterization; 4. The NCL is not violated.

68

Natural and Artifactual Objects in Contemporary Metaphysics

The conventionalist approach is very similar to counterpart theory. Both theories hold that modal properties are reducible to something else: in one case conventions and in the other case counterpart relations.5 Here the problem is again the one mentioned above: are persistence conditions modal properties? Moreover, conventionalists have to explain why and how the very same portion of reality admits two mutually inconsistent modal descriptions, which is a grounding problem for conventionalists. It seems to me that there is still much work to be done. d. Four dimensionalism (Sider 2001): 1. Objects have four dimensions: three spatial and one temporal. That is, beyond spatial parts they also have temporal parts; 2. Two co-occupants are not two objects in the same region; rather they are two objects with a part in common in that region; 3. There are no coincident objects, but just overlapping objects; 4. The NCL is not violated. Four dimensionalism is often presented as a good way of solving the problem. It gives us the resources to easily explain the problem: when it seems that two objects coincide, it is just because they overlap. Thus, the coincidence is just a shared stage within the life of such objects. Nothing mysterious. There are two problems with four dimensionalism. (1) Not every philosopher agrees with such an ontology and therefore it is controversial to adopt a controversial thesis to solve a controversial problem. (2) Consider a changeless world, namely a world where time passes but temporal objects do not change. Consider a tree and the cellulose. Even though they do not change across time, they nevertheless may be different across possible worlds: the tree would have been made of different stuff, the cellulose would have constituted more than one tree and so forth. Hence, the problem reappears: what grounds such a difference? e. Hylomorphism (Fine 2008; Koslicki 2008): 1. A lump of stuff is a proper part of a thing along with its formal proper parts; 2. There are no two or more coincident particulars, rather just one composed of all of its parts; 3. The NCL is not violated. My worry with this position is twofold. (1) Since an object may be constituted by more than one part, it can be constituted by more than one formal part

Spatial Fictionalism

69

or manner of composition, however we take that phrase. Since formal parts dictate the structure, it is at least possible that one object owns more than one structure. There is nothing within the theories of Fine and Koslicki that can block such an argument. Therefore we have more than one object in the same region at the same time. (2) The ontology of a proponent of hylomorphism resembles a Meinongian picture: a jungle instead of the Quinean desert. According to a hylomorphist, such as Kit Fine, there are more kinds of parts, formal and material, and more kinds of objects, material, formal and mixed entities, than in a mainstream Quinean ontology. Not only does the ontology turn out to be richer, so too does the ideology: following Kit Fine (1999) there is more than one mereological system and hence a very rich ideological commitment. f. Mereological nihilism (Unger 1979; Van Inwagen 1990): 1. 2. 3. 4.

Parthood (almost6) never occurs; There is just one simple for every inhabited region; Every region may host at most one simple; The NCL is not violated.

The main problem with nihilists is that they throw away parthood because of its controversial consequences and its uncertain status, but they accept simples, that is, proper partless entities. As far as I am concerned, simples are as controversial as composed entities. I also have another worry: simples are constituted of a lump of stuff, thus employing the same old argument that there are two particulars in the same region at the same time. Hence, they have no serious way to avoid the challenge. g. Modal supervenience thesis (Sider 1999; Olson 2001) 1. An object’s properties supervene on the properties of its microphysical structure; 2. Coincident objects have the same microstructure; 3. Coincident objects are identical. There are many kinds of supervenience; given constraints of space, I cannot give reasons for rejecting each of them here. I would just say here that, as I noted above, there is no fact of the matter on the basis of which smaller parts are more fundamental than the whole itself or, at least, there is no general agreement about it.7 Not even the converse holds for sure.

70

Natural and Artifactual Objects in Contemporary Metaphysics

h. Sortal dominance (Burke 1994; Rea 2000): 1. When a lump of stuff composes a thing, just one sortal dominates, the one with the most properties instantiated; 2. There is just one object, the one whose sortal dominates the other. According to Burke, the dominant sortal is the one with ‘the widest range of properties’. Thus, since a thing has every property of its constituted stuff and then some other properties, then the sortal of the thing is the dominant one. Again, here my worry is twofold. (1) The fact that the sortal that entails the owning of more properties is the dominant one is either arbitrary or trivial: arbitrary, since having more properties does not seem a reason for being dominant, trivial, for one of the objects may be part of the other, and then it is obvious that the whole has more properties than one of its parts. (2) Even if the sortal with more properties is dominant, which one has more properties? What if the two sortals entail the same number of properties? It is perfectly admissible that there may be a case in which a lump of stuff and a thing are indistinguishable: for instance, in a changeless world without humans, an artwork and its stuff have the same number of properties, even though the properties are different. Another problem also arises: how many properties does a lump of stuff have? How many properties does a thing have? Questions like these are still outstanding.

Spatial Fictionalism Spatial Fictionalism8 is the conjunction of the following three doctrines: (i) Dependence on boundaries: objects derive their individuality from their boundaries. (ii) Fiat boundaries: every boundary is of the fiat sort. (iii) Stuffism: stuff is of the bona fide sort. The doctrines jointly taken entail that a world with only fiat boundaries is a world inhabited by only fiat objects, whereas the stuff which they are ‘made of ’ is of the bona fide sort. In a world like this, the NCL is clearly a conceptual truth since every region of space may host only stuff and every other object made of this is just of the fiat sort. My first task is to prove that what each doctrine claims is not so controversial as might appear, but rather less controversial than the other philosophical strategies for supporting the NCL.

Spatial Fictionalism

71

Dependence on Boundaries Before starting let me state a very intuitive notion of boundary. According to Aristotle (1984, p. V, 17, 1022a4–5), a boundary of an object is its last part, beyond which it is not possible to find this object anymore. In a pseudoformal way B1=df. x is a boundary of y iff x is the outermost part of y.

It is not enough, as Sorensen (1998, p. 275) argues. Suppose the universe, that is, the object of which every other object is a proper part, is finite. It has surely an outermost part but nevertheless it has no boundaries since we usually assume that boundaries have to demarcate an object from its complement, that is, the object with which it does not have parts in common. And the universe, by definition, has no complement. Then we have to supplement B1 with one more clause. B2=df. x is a boundary of y iff x demarcates y from its complement.

Then we obtain the following comprehensive definition of boundary: ABxy=df. x is a boundary of y iff x is the outermost part of y and x demarcates y from its surrounding.9

Assume also that a boundary is a lower-dimensional object, namely it has at least one dimension fewer than the object it bounds. For instance, the boundary of a three-dimensional cube is its two-dimensional surface, the boundary of a surface is its one-dimensional perimeter and so forth. This assumption, according to Smid (2015), does justice to the intuition that boundaries are less real than the entities they bound. Indeed, following Chisholm (1983) and Smid (2015), the mainstream theory of boundaries assumes that boundaries depend upon the bounded object.10 On the face of it, whereas an object can survive even lacking boundaries, a boundary cannot survive without being part of an object. I think this is a false start. An object is, by definition, something discrete from its surrounding and its neighbours. In other terms, an object is an individual: something that can be counted as one and that can be distinguished from its surrounding. I claim that boundaries are the bearers of that individuality11 and hence objects depend on boundaries in order to be individual. Consider the well-known example by Wiggins: tree and cellulose are different insofar as they belong to two different sortals. And they belong to their sortals because they have the needed properties. If those properties are rightly aggregated, then there arises an object of a certain sort. The space where such

72

Natural and Artifactual Objects in Contemporary Metaphysics

properties are instantiated is full of objects, maybe not completely full, but full enough. Consider the content of the region of tree/stuff and consider the content of the region next to it. Both have their content and their content instantiates some properties. Consider what it would be for both to be boundaryless: how would it be possible to distinguish which properties belong to the first entity and which to the second entity? Answers might be as follows. (1) They instantiate different sortals whose properties are already fixed; you need just to find where the properties of the considered sortal are and there is the object you are looking for. (2) Where there is a contradiction among properties, there is a new entity. In neither case do you need boundaries. Unfortunately, both answers are unsatisfactory. (1) There is a sortal for every consistent conjunction of modal properties and persistence conditions; then without boundaries it is impossible to stop the conjunction; it is thus impossible to discern the two objects. (2) The topic of the chapter is precisely the inconsistency of sortals instantiated in the same region and my aim is to avoid the possibility that two different objects occupy the same region; hence, that cannot be an answer. If every attempt to individuate boundaryless objects is either inconsistent or incoherent, then in order to individuate an object, boundaries are necessary. Thus, objects depend on boundaries.

Fiat Boundaries Let me begin with a brief introductory remark. According to Smith (2001), Smith and Varzi (2001) and Varzi (2011), there are two kinds of boundaries: fiat and bona fide. Bona fide boundaries are natural discontinuities which occupy regions of space. They are parts of the bounded object, made of a certain stuff, with certain atomic structures and so forth. Since bona fide boundaries belong to the furniture of the world, they are causal agents and ontologically independent of human beings. Fiat boundaries are fictional articulations of reality. They do not belong to the furniture of the world; rather, they are projections of human beings onto the world. We speak, for example, of national borders only by convention; our talk of the horizon is merely the result of our perceptions; galaxies are merely the products of our scientific schemes. Fiat boundaries are thus ephemeral elements,

Spatial Fictionalism

73

located in space but not taking up space. They are just concepts which are used to pursue human objectives. According to Varzi (2001, 2011) since every boundary is vague, that is, there are borderline cases, and since it is impossible for a bona fide object to be vague, then every boundary is of the fiat sort. His argument runs as follows (my reconstruction, based on Varzi 2001, pp. 49–50): 1. x is part of y, iff x lies within the boundaries of y; 2. if x is part of y, then the nearest entity x+1 to x is part of y; 3. Therefore, every object xn is part of y. The argument says if an object is part of a whole, then the nearest object is also part of the same whole. By a recursive application of the modus ponens, we reach the sorites-like paradox according to which every object is part of that whole. The argument proves that since it is vague where a boundary is located, it is also vague what belongs to an object. But since, following a venerable tradition, vagueness is a semantic problem, boundaries are semantic objects as well.12 According to Varzi (2011), by generalization every boundary of every object is vague; therefore every boundary is fiat. Such a generalization seems highly controversial. First, it seems that not every boundary is vague and secondly, vagueness may be a feature of bona fide reality, and not just of fiat one. Perhaps not every boundary is vague. For instance, my skin, that is, the boundary of my body, is a sharp cut-off surface. However, it is also true that on a closer look the surface of my body is a set of dense swarms of little particles, each of which seems to be a good candidate to be the surface of my body.13 Which of these becomes the surface is fiat act, since it seems absolutely arbitrary to choose one of them as the bona fide surface: which criterion may be employed to choose the surface? What differences are there among surfaces? Nevertheless, the boundary itself may be vague. Thus, it is true that there are many equally good candidates for being the surface of my body, for the surface itself is vague: the supposed many candidates are the very same surface whose position in space is not definite. That is, there are vague bona fide boundaries. Assume that there are indeed vague bona fide boundaries. Let K be an object, and let b be its bona fide boundary. Since b is vague there is at least one object where it is indeterminate whether or not it belongs to K. Furthermore, there are certain portions of that world where it is vague whether or not they

74

Natural and Artifactual Objects in Contemporary Metaphysics

belong to the indeterminate portion between K and its complement, and so forth. For all of that portion it is vague whether they are parts of either of K or of its complement. Therefore, it is vague how many objects there are. It is certainly possible to accept such a solution but it seems to me at least highly controversial. The fiat nature of boundaries can also be argued for in a different yet very similar way.14 Consider an entity C and its bona fide boundary b. Consider the boundaries of a cube, that is, its surface. Suppose that b is removed. Hence, the layer immediately behind b, namely b+1, becomes the new surface of the cube. Since b was two dimensional the mass of C does not change after the removal of b. Likewise since b+1 is two dimensional as well, if it were removed the mass of C would not change. Suppose that any time that a new surface of C becomes its boundary it is removed after a few seconds. At the very least, C may lose every one of its surfaces and yet its mass does not change. Is there a point in time after which C starts to decrease in mass? Certainly not, since layers are all two dimensional, and, therefore, without mass by definition. Therefore, we reach the highly controversial conclusion according to which the mass of an entity has nothing to do with the mass of its parts, unless such parts are of the fiat sort, namely they belong to representations of the world instead of the world itself. If we want to avoid the paradoxes two arguments arise. We have to accept that every object is of the fiat sort since every object has fiat boundaries. It does not mean that the stuff they are made of depends upon human beings. Rather, it means that since they own their individuality by means of their boundaries, and since their boundaries are fiat, then their individuality also depends upon a fiat act. They are there, where we think they are. Nevertheless, ‘there’ hosts just stuff, nothing more.

Stuffism Stuffism is a well-known thesis according to which what exists is not a collection of disparate entities, but just an undivided yet heterogeneous stuff (Heller 1990; Sidelle 1998; Dasgupta 2009). Stuff is usually understood as non-individuated reality, namely a cohesive pile of properties not divided into discrete entities. Instead, a world of stuff is a world with highly local variability, although to such variations there do not correspond independent entities. Stuffism is needed in order to avoid the conclusion that the conjunction of the first two doctrines might entail idealism or irrealism. There are many

Spatial Fictionalism

75

ways of arguing for Stuffism: conventionalism about modal properties, holism, arbitrariness. I set aside here such arguments and instead I just assume a soft form of ontological realism according to which stuff exists regardless of human beings.

Solution Recall the following problem: the same region of space hosts at least two different objects – a thing and a lump of stuff, contrary to the intuitiveness of the NCL. Moreover, it is mysterious how it is possible for two objects to be different despite sharing a region, parts and, more generally, every material actual property. It is the problem I called the GP. Clearly, the NCL would avoid the GP, since it precludes the condition that two different objects may share the same region. Spatial Fictionalism solves the problem of holding the NCL. Suppose that the object x that occupies r has the modal properties m1, … , mn and the set of persistence conditions c1, … , cn. Suppose that modal properties and persistence conditions are consistent. Thus, x is an object whose sortal properties are the ones resulting from the union of the two sets of modal properties and persistence conditions. Unfortunately, since there are no bona fide boundaries, x is virtually indistinguishable from its complement: there is no fact of the matter and no metaphysical way to pick up x, since there isn’t anything that differentiates x from its surrounding. Despite the fact that there is no bona fide difference between x and its surrounding, human beings single out x by means of a fiat act, namely through creating fiat boundaries. Such an act picks out a portion of stuff with its set of properties: the resulting object is just an arbitrary demarcation of space. Thus, there is no difference between thing and lump of stuff: it is the same quantity of stuff picked out through time and across possible worlds in two different ways: the boundaries of a thing – topologically connected, necessarily shaped in a certain manner and so forth – are the results of certain actions due to certain aims, and the boundaries of the lump of stuff – topologically disconnected, without any necessary shape and so forth – are the results of different actions due to different aims. They share their region. But since they are just fiat objects, they do not violate the NCL. Fiat objects, by definition, are just located in space but they do not occupy space. Thus, that region has just one host, namely a certain amount of stuff.

76

Natural and Artifactual Objects in Contemporary Metaphysics

Consider a thing: its boundaries have to be so and so in order to match the criteria fixed by our concept of thing; its boundaries, for instance, must preserve a certain, albeit vague, shape: certain imperceptible changes are admitted, whereas sudden changes are not and so forth. Consider a lump of stuff that such a thing is made of: it has very different persistence conditions. For instance, the preservation of the shape is not one of its persistence conditions. They are bona fide indistinguishable: if there are no boundaries, it is impossible to say where a thing has its surface and thus what its shape is. This does not entail that the shape is human dependent. It just means that we pick one of the many possible shapes. Such selection is based on both decisions and the structure of our way of seeing the world: how our perceptions and cognition work. They persist in different ways, for their persistence conditions are fixed by two different fiat criteria of drawing boundaries: those of the thing state that the thing persists if and only if its boundaries are continuous through time without sudden changes; those of the lump of stuff state that a lump of stuff persists if and only if its boundaries include every part of which it consists. In fact, a thing ceases to exist if and only if its shape is modified, and a lump of stuff ceases to exist if and only if one of its parts is annihilated. As I have tried to show, such conditions depend upon boundaries. And boundaries depend upon fiat acts.

Conclusion The aim of this chapter is to maintain the NCL in order to solve the puzzle of co-location, that is, the so-called grounding problem: how it is possible for two different entities to share the same region at the same time. By rejecting the condition that two entities may be coincident, the NCL avoids the problem. I hold the NCL as a consequence of a more general thesis which I called ‘Spatial Fictionalism’. That thesis is a development of the thesis of fiat boundaries defended by Smith and Varzi (2001) and Varzi (2011, 2013, 2014). Spatial Fictionalism is the conjunction of three doctrines: entities are endowed with individuality by means of boundaries; every boundary is of the fiat sort; reality is just an undivided yet heterogeneous pile of stuff. A world without boundaries is a world without discrete entities and therefore the difference between the hosts of a same region becomes just a conceptual consequence of our way of drawing boundaries, instead of being a bona fide fact.

Spatial Fictionalism

77

Notes 1 Such a characterization of occupation relation is equivalent to a tensed version of ‘exact location’ by Casati and Varzi (1999). 2 NCL also fails if one accepts as possible either of the two following cases: ‘The tree may have been constituted of a different stuff ’ or ‘The cellulose may have constituted a different object’. 3 This position was already argued by Bennet (2009a) and Hirsh (2011). 4 Let me briefly say why I have excluded from the catalogues two of the most famous solutions: deflationism (Hirsh 2005) and relative identity (Geach 1962, 1967). I rule out such theories, because I maintain that the problem we are discussing is really a metaphysical one and not just a verbal dispute, as a deflationist and a relative identity theorist would say. 5 According to Heller (2005) counterpart theory can be employed by conventionalists for their purposes. 6 van Inwagen allows that living organisms are composites; so he is a local nihilist. 7 There is also no empirical evidence that simples there exist. There is a big and controversial debate both in physics and metaphysics about it. See Hudson 2007. 8 Varzi (2011, 2013, 2014) has defended a very similar thesis with very similar consequences. However, the thesis is slightly different in few passages and in some of its consequences. 9 In a world without a universe B1 entails B2. The proof is trivial. 10 Assume that dependency is a non-symmetrical relation and x depends on y iff x exists and preserves its identity thanks to y. 11 See also Varzi 2011, p. 9. For an alternative view, see Clarke 1981. 12 See, inter alia, Russell 1923 and Lewis 1986. 13 This is the well-known argument of the many. See Unger 1979, Lewis 1993 and Sider 2001. 14 See Galton 2007.

References Aristotle, (1984), ‘Metaphysics’, in J. Barnes (ed.) The Complete Works of Aristotle, Princeton University Press, Princeton, vol. 2. Bennet, K. (2009a), ‘Spatio-Temporal Coincidence and the Grounding Problem’, Philosophical Studies, 118: 339–71. Bennet, K. (2009b), ‘Composition, Coincidence, and Metaontology’, in D. Chalmers, D. Manley and R. Wasserman (eds), Metametaphysics. New Essays on the Foundations of Ontology, Oxford University Press, Oxford, 38–76.

78

Natural and Artifactual Objects in Contemporary Metaphysics

Black, M. (1952), ‘The Identity of Indiscernibles’, Mind, 61: 153–64. Burke, E. (1994), ‘Preserving the Principle of One Object to a Place: A Novel Account of the Relations among Objects, Sorts, Sortals, and Persistence Conditions’, Philosophy and Phenomenological Research, 54: 591–624. Casati, R., and Varzi, A. C. (1999), Parts and Places. The Structure of Spatial Representation, MIT University Press, Cambridge MA. Chisholm, R. M. (1983), ‘Boundaries as Dependent Particulars’, Grazer Philosophische Studien, 10: 87–95. Clark, B. L. (1981), ‘A Calculus of Individuals Based on Connection’, Notre Dame Journal of Formal Logic, 22: 204–18. Dasgupta, S. (2009), ‘Individuals: An Essay in Revisionary Metaphysics’, Philosophical Studies, 145: 35–67. Davidson D. (1969), ‘The Individuation of Events’, in N. Rescher (ed.), Essays in Honor of Carl G. Hempel, Reidel, Dordrecht, 16–234. DeRosset, L. (2011), ‘What is the Grounding Problem?’, Philosophical Studies, 156: 173–97. Einheuser, I. (2011), ‘Toward a Conceptualist Solution of the Grounding Problem’, Nous, 45: 300–14. Fine, K. (1999), ‘Things and their Parts’, Midwest Studies in Philosophy, 23: 61–74. Fine, K. (2003), ‘The Non-Identity of a Material Thing and Its Matter’, Mind, 446: 195–234. Fine, K. (2008), ‘Coincidence and Form’, Proceedings of the Aristotelian Society, supplementary volume, 82: 101–18. Galton, A. (2007), ‘On the Paradoxical Nature of Surfaces: Ontology at the Physics/ Geometry Interface’, The Monist, 90: 379–90. Geach, P. (1962), Reference and Generality, Cornell University Press, Itaha. Geach, P. (1967), ‘Identity,’ Review of Metaphysics, 21: 3–12. Heller, M. (1990), The Ontology of Physical Objects. Four-Dimensional Hunks of Matter, Cambridge University Press, Cambridge. Heller, M. (2005), ‘Anti-Essentialism and Counterpart Theory’, The Monist, 88: 600–18. Hirsh, E. (2005), ‘Physical-Object Ontology, Verbal Disputes, and Common Sense’, Philosophy and Phenomenological Research, 70: 67–97. Hirsh, E. (2011), Quantifier Variance and Realism, Oxford University Press, Oxford. Hudson, H. (2007), ‘Simples and Gunk’, Philosophy Compass, 2: 291–302. Koslicki, K. (2008), The Structure of Objects, Oxford University Press, Oxford. Lewis, D. K. (1986), On the Plurality of Worlds, Blackwell, Oxford. Lewis, D. K. (1993), ‘Many, but Almost One’, in J. Bacon (ed.), Ontology, Causality and Mind: Essays in Honour of D. M. Armstrong, Cambridge University Press, New York, 23–38. Lewis, D. K. (2003), ‘Things qua Truthmakers’, in H. Lillehammer and G. RodriguezPereyra (eds), Real Metaphysics: Essays in Honor of D. H. Mellor, Routledge, London, 25–38. Merricks, T. (2001), Objects and Persons, Oxford University Press, Oxford. Merricks, T. (2017), ‘Do Ordinary Objects Exist? No’, in E. Barnes (ed.), Current Controversies in Metaphysics, Routledge, London, 135–48.

Spatial Fictionalism

79

Olson, T. E. (2001), ‘Material Coincidence and the Indiscernibility Problem’, The Philosophical Quarterly, 51: 337–55. Rea, M. C. (2000), ‘Constitution and Kind Membership’, Philosophical Studies, 97: 169–93. Russell, B. (1923), ‘Vagueness’, Australasian Journal of Psychology and Philosophy, 1: 84–92. Sidelle, A. (1998), ‘A Sweater Unraveled: Following One Thread of Thought for Avoiding Coincident Entities’, Nous, 32: 423–448. Sidelle, A. (2016), ‘Coincidence: The Grounding Problem, Object-Specifying Principles, and Some Consequences’, Philosophical Papers, 45: 497–528. Sider, T. (1999), ‘Global Supervenience and Identity across Times and Worlds’, Philosophy and Phenomenological Research, 59: 913–37. Sider, T. (2001), Four-Dimensionalism. An Ontology of Persistence and Time, Oxford University Press, Oxford. Sider, T. (2003), ‘What’s So Bad about Overdetermination?’, Philosophy and Phenomenological Research, 67: 719–26. Smid, J. (2015), ‘A Puzzle Concerning Boundaries, Dependence, and Parthood’, Analytic Philosophy, 56: 169–76. Smith, B. (2001) ‘Fiat Objects’, The Monist, 20: 131–48. Smith, B., and Varzi, A. C. (2001), ‘Fiat and Bona Fide Boundaries’, Philosophy and Phenomenological Research, 60: 401–20. Sorensen, R. (1998), ‘Sharp Boundaries for Blobs’, Philosophical Studies, 91: 275–95. Sutton, C. S. (2012), ‘Colocated Objects, Tally-Ho: A Solution to the Grounding Problem’, Mind, 121: 703–30. Unger, P. (1979), ‘There Are No Ordinary Things’, Synthese, 41: 117–54. van Inwagen, P. (1990), Material Beings, Cornell University Press, Ithaca. Varzi, A. C. (2001), ‘Vagueness in Geography’, Philosophy & Geography, 4: 49–65. Varzi, A. C. (2011), ‘Boundaries, Conventions, and Realism’, in J. K. Campbell, M. O’Rourke, and M. H. Slater (eds), Carving Nature at Its Joints: Natural Kinds in Metaphysics and Science, MIT Press, Cambridge (MA), 129–53. Varzi, A. C. (2013), ‘Fictionalism in Ontology’, in C. Barbero, M. Ferraris and A. Voltolini (eds), From Fictionalism to Realism, Cambridge Scholars, Newcastle (UK), 133–51. Varzi, A. C. (2014), ‘Realism in the Desert’, in F. Bacchini, S. Caputo and M. Dell’Utri (eds), Metaphysics and Ontology without Myths, Cambridge Scholars, Newcastle (UK), 16–31. Wasserman, R. (2015), ‘Material Constitution’, The Stanford Encyclopedia of Philosophy (Spring 2015 Edition), E. N. Zalta (ed.), URL = https://plato.stanford.edu/archives/ spr2015/entries/material-constitution/. Wiggins D. (1968), ‘On Being in the Same Place at the Same Time’, Philosophical Review, 77: 90–5. Zimmerman, D. W. (1996), ‘Could Extended Objects Be Made Out of Simple Parts? An Argument for “Atomless Gunk”’, Philosophy and Phenomenological Research, 56: 1–29.

4

Talking about Properties: A Couple of Doubts about Hofweber’s Internalist View Elisa Paganini

The distinction between natural objects and artifactual ones has a long tradition in metaphysics. It presupposes that objects may be divided into two kinds: (1) natural kinds, that is, objects marking true joints in nature and (2) artifactual kinds, that is, objects whose classification reflects the actions and interests of human beings. This distinction takes some notions of metaphysical priority as basic: natural kinds are generally considered more fundamental than artifactual ones. But what does natural mean? What does fundamental mean? According to Thomas Hofweber (2009 and 2016a), a certain way of doing metaphysics considers notions like natural, fundamental or ultimate as distinctively metaphysical and primitive, where this means that they are different from the notions we use in our everyday conversations and cannot be defined in simpler terms. The specificity and indefinability of these notions raises the concern that metaphysical distinctions are made using terms taken for granted without a correct and precise understanding of them. It is due to this concern that Hofweber denounces what he calls ‘esoteric metaphysics’ – a metaphysics that aims to answer questions involving primitively metaphysical terms – and proposes instead ‘egalitarian metaphysics’ – a form of metaphysics that tries to answer questions accessible to all and without special terms for metaphysical insiders.1 Within egalitarian metaphysics, being natural and being artifactual are on a par; they are predicates of natural language, used within science and common conversations. When we use these predicates, we may want to talk about properties, we may want to consider for example whether the property of being natural is more (or less) frequently instantiated than the property of being artifactual. And, according to Hofweber, when we talk about properties,

Talking about Properties

81

whatever we say depends on us and our minds and not on how the objects in the world are in themselves. In order to understand this claim, it is useful first of all to consider how Hofweber thinks we should answer the question ‘Are there properties?’ and then consider how he argues for the internalist view of our talk about properties. My reconstruction of his argument is not neutral; my aim is to raise a couple of doubts regarding his account of our talk about properties. The present work is organized as follows: first, I consider how Hofweber evaluates questions of existence, and after reporting his analysis of existence claims concerning properties, I present my first doubt; secondly, I take into consideration the idealistic consequences of Hofweber’s internalist view of our talk about properties and propositions, and raise my second doubt. I conclude by summing up my doubts concerning the internalist view of talk about properties.

Internal and external questions of existence In order to understand Hofweber’s view of our talk about properties, it is useful to consider how he interprets the assertion ‘There are properties’ or any assertion claiming existence. He actually proposes that any assertion of existence has two readings: an internal one and an external one. In so doing, he rehabilitates the distinction between internal and external readings of existence first proposed by Carnap (1956). As is well known, Carnap first observed that some important metaphysical questions like ‘Are there objects?’, ‘Are there properties?’ or ‘Are there numbers?’ have two readings: an internal reading and an external one. The distinction was intended as a reply to Quine’s (1948) contention that any existence claim requires an ontological commitment; for example, if we accept ‘There are numbers’ or ‘∃x (x is a number)’, according to Quine, we are committed to there being numbers in the world. According to Carnap, instead, if we adopt an internal reading, we may correctly accept ‘There are numbers’ without being committed to the supposition that there are numbers in the world, the idea being that, internally speaking, the question of existence may be answered only relative to a system of linguistic expressions and semantic rules for accepting or rejecting those expressions. Carnap’s proposal was notably criticized by Quine in 1951 (Quine 1951a, b) and there has been general agreement among philosophers that Quine was right and that the distinction between internal and external questions proposed by Carnap is inadequate. Hofweber rehabilitates the distinction between two different readings of existence claims, giving them an interpretation quite different from the one

82

Natural and Artifactual Objects in Contemporary Metaphysics

proposed by Carnap.2 According to Hofweber, our common usage of the existential quantifier is polysemous or ambiguous. We may want to adopt an external reading, or the domain condition reading, according to which the truth of what we assert depends on the objects of the external world in the domain of quantification. Suppose, for example, that I report what happened at the university and say, ‘Someone stole John’s wallet.’ The sentence is true if a person in the external world stole John’s wallet, false otherwise: it therefore has an external reading. But, let us now suppose that I am reading a novel reporting that Jane stole John’s wallet; supposing that fictional characters do not exist in the external world, we may infer that the sentence ‘Someone stole John’s wallet’ has an internal or inferential reading: the sentence is true but we do not expect the quantifier ‘someone’ to range over objects (or persons) in the world; its application depends merely on our use of language in fiction. Let us concentrate on the internal or inferential reading of the existential quantifier for a moment. Suppose that according to the novel ‘Jane stole John’s wallet’ is correct and true; suppose moreover that the name ‘Jane’ and the name ‘John’ do not refer to objects (or people) in the actual world; if so, we may have reason to infer that according to the novel ‘Someone stole John’s wallet’ (substituting the quantifier ‘someone’ for the name ‘Jane’) without being committed to the supposition that there is a domain of quantification in the actual world over which the quantifier ‘someone’ ranges. Now, in such a case, the sentence ‘Someone stole John’s wallet’ is true in the internal or inferential reading, its truth being guaranteed by there being a linguistic term (i.e. ‘Jane’) which the quantifier substitutes. It is quite important to note that the sentence ‘Someone stole John’s wallet’ (like any existentially quantified sentence) can have two opposite evaluations in a specific situation. And in particular, it has two opposite evaluations when the existentially quantified sentence is obtained by substituting a non-referring term, like ‘Jane’ in the novel. In such a case, the quantified sentence ‘Someone stole John’s wallet’ is true in the internal reading, but false in the external one; and, in general, according to Hofweber’s proposal, whenever we have an existentially quantified sentence obtained by substituting a non-referring term in a true sentence, the quantified sentence is true in the internal reading and false in the external one. Having presented Hofweber’s two readings of existential claims and his contention that whenever the existential quantifier substitutes a non-referring term in a true sentence we obtain a true sentence in the internal reading and a

Talking about Properties

83

false one in the external reading, we are ready to consider how, according to this approach, the sentence ‘There are properties’ is to be viewed.

The internalist view of talk about properties: The first doubt Let us now consider the existential claim “‘There are properties’. How can we obtain it? Well, we can obtain it through simple inferences like the following. We start from a true sentence like the following: (1) Fido is a dog. From it we infer the following: (2) Fido has the property of being a dog. And then we introduce an existential quantifier in the following way: (3) There is a property Fido has. If we accept that premise (1) is true, that (2) is truth-conditionally equivalent to (1), and that the rule of the introduction of the existential quantifier is always acceptable, conclusion (3) is difficult to deny. And for any true sentence claiming that an object has or does not have a property, we may conclude that there is a property this object has or does not have. It is therefore through simple inferences like the one reported above that we may conclude that there are properties. But the question we should consider is the following: does the quantifier in conclusion (3) have an external or an internal reading? This question may be answered only after establishing whether ‘the property of being a dog’ has a reference or at least a denotation in the external world. If the description has a denotation, the quantified sentence is allowed to be true in an external reading; if the description does not have a denotation in the external world, the quantified sentence is forced to be true in an internal reading (the sentence being false under the external reading). Hofweber claims that the definite description ‘the property of being a dog’ does not have a reference or denotation in the external world, and therefore that the sentence ‘There is a property Fido has’ is true in an internal reading and false in an external one. This observation, Hofweber says, may be generalized to any claim of property existence, so that the assertion ‘There are properties’ is true in an internal reading and false in an external one. And this observation grants his internalist view of our talk about properties.

84

Natural and Artifactual Objects in Contemporary Metaphysics

The affirmation that the definite description ‘the property of being X’ (for any X) does not have reference or denotation is crucial for Hofweber’s claim that the existence of properties can be claimed only within an internalist view. It is therefore useful to consider his argument in support of his contention that ‘the property of being X’ is non-referring. According to Hofweber, in order to establish whether the definite description ‘the property of being X’ is referential or not, we should consider whether it behaves as a singular term, the assumption behind this being that only singular terms may refer to entities in the external world, while predicates and modifiers do not.3 And in order to establish whether a certain linguistic expression behaves as a singular term, we should accept that any singular term obeys the following substitution rule: semantically singular terms should in general be substitutable for other terms that stand for the same entity.4 Now, according to Hofweber, supposing that ‘the property of being X’ stands for an entity, it should stand for the same entity that ‘being X’ stands for. And as long as ‘the property of being X’ cannot be substituted for the term ‘being X’, it follows – Hofweber claims – that it is not a singular term and that it does not denote an object. It is quite difficult, however, to understand why he thinks that ‘the property of being X’ and ‘being X’, if referential at all, should refer to the same entity. I suspect that the reason behind his claim depends on a specific interpretation of sentences like the following (which is a way of specifying conclusion (3) of the above argument): (4) There is a property Fido has, namely being a dog. As long as (4) may be inferred from (2) substituting a definite description (i.e. ‘the property of being a dog’) with a quantifier, he probably thinks that ‘the property Fido has’ and ‘being a dog’, if referential at all, should refer to the same entity, that is, a property. But, in my opinion, the inference is not correct: the property Fido has is the property of being a dog. We should therefore read (4) as follows: (5) There is a property Fido has, namely [the property of] being a dog. There is a reason behind this reading of sentence (4), and this has to do with the distinction between properties and the instantiation of properties; ‘the property Fido has’ – if it refers to anything at all – refers to a property and ‘being a dog’ does not refer to a property; it expresses a property instantiation. ‘Being a dog’ means ‘having the property of being a dog’, just as ‘being a philosopher’

Talking about Properties

85

means ‘having the property of being a philosopher’. And ‘having the property of being X’ is not at all equivalent to ‘the property of being X’, as ‘the property of being X’ is the complement of the verb having. This observation becomes quite evident if we reconsider sentences (1) and (2) above. We obtained (2) by substituting ‘is a dog’ in (1) with ‘has the property of being a dog’. ‘Being a dog’ is therefore equivalent to ‘having the property of being a dog’ and not to ‘the property of being a dog’. I therefore agree with Hofweber that, for example, ‘the property of being a philosopher’ cannot be substituted for ‘being a philosopher’, just as – for example – we may acknowledge as true ‘Being a philosopher is fun’, but we may consider as false ‘The property of being a philosopher is fun’,5 the reason being that instantiating the property is fun, but the property in itself is not. However, I am not ready to agree with Hofweber that the substitution failure shows that ‘the property of being a philosopher’ is not a singular term, as the description might possibly be substituted for any co-referring expression. The point is simply that ‘the property of being a philosopher’ is not co-referential with ‘being a philosopher’. It seems to me therefore that we do not have reason to maintain with Hofweber that ‘the property of being X’ (for any X) is not a singular term and it cannot have a denotation. From this consideration, my first doubt is as follows: we do not have reason to conclude that our talk about properties has only a true internal reading, but we may still believe that a true external reading is allowed by our use of language.

Nominalism about properties and conceptual idealism: The second doubt Hofweber is a realist about objects: he maintains that objects exist in the world quite independently of us. But he is a nominalist about properties, propositions and numbers because he thinks that our talk of properties, propositions and numbers may be true only in an internal reading and not in an external one.6 In the previous section, I explained why I think that his argument for property nominalism is not conclusive. I am now interested in the relation between what Hofweber calls ‘conceptual idealism’ – a consequence of proposition and property nominalism – and his nominalism regarding properties. Let me first present how I understand Hofweber’s conceptual idealism. According to conceptual idealism, whenever we attribute a property to an object,

86

Natural and Artifactual Objects in Contemporary Metaphysics

for example, we say that an object is a natural one or an artifactual one, what we say depends on us and our minds. It may be interesting to consider how the internalist view of our talk of properties and propositions leads to conceptual idealism as a consequence. First of all, it is important to acknowledge that the internalist view of talk about properties and propositions is incompatible with there being inexpressible properties and propositions.7 The reason is that if there were inexpressible properties or propositions, it would not be possible to express these properties or propositions in language and therefore there would not be linguistic expressions fit for talking about any of them; the existential quantifier on inexpressible properties and propositions would therefore range only over language-independent entities in the world and could not substitute expressions of the language, allowing only for an external and not for an internalist view of our talk of inexpressible properties. As long as, according to Hofweber, we may adopt only an internal reading of our talk about properties and propositions, it follows that an internalist view of talk about properties and propositions is incompatible with any true claim that there are inexpressible properties or propositions. How does Hofweber account for the fact that any property or proposition may be expressed in language? He claims that any talk of property or proposition rests on linguistic expressions and linguistic expressions rely on our representational abilities; therefore our talk about properties and propositions depends on us and our minds and not on any juncture in reality, and that amounts to maintaining a form of idealism, as he readily acknowledges.8 In order to better understand this form of idealism, it is useful to distinguish between ontological idealism, that is, ‘the view that what there is, and what it is like, are mind-dependent’ and conceptual idealism, that is, ‘the view that what can be said or thought about reality depends on us’ (Hofweber 2016a, p. 257). Hofweber denies ontological idealism and accepts conceptual idealism. The reason is not difficult to grasp: he is a realist about objects, and therefore maintains that what there is is independent of us and our minds, contrary to ontological idealism; he is instead a nominalist about properties and propositions: properties and propositions do not exist independently of us and our minds, and this allows for conceptual idealism. So, to run briefly over my argument once again, Hofweber’s conceptual idealism is the view that what can be said or thought about reality depends on us and our minds. As I understand it, it is the view that whatever we say or think about the objects in the world around us does not depend on how the

Talking about Properties

87

objects are in themselves, relying instead on us and on how we represent the objects in the world. I want now to concentrate on how this relation of dependence in our talk of reality depends on us and our minds. How may this relation of dependence be understood? One way of seeing the issue that I find very natural is that we have the property of representing reality in a certain way, and it is because we have this property that we may talk and think as we do. But my grasp of Hofweber’s idea cannot be correct, because, according to him, there are no properties in the external world, and therefore we cannot objectively have the property of representing reality in a certain way. My first interpretation of conceptual idealism cannot therefore be assumed. Let us try to make sense of conceptual idealism avoiding any reference to objective properties we may have. Suppose that, when we talk about reality, we do not objectively have the property of representing the objects around us in a certain way, because, according to Hofweber, there are no properties in an external reading of properties and we cannot objectively have them. What does this mean? Let us try to understand it in the following way: when we state conceptual idealism, that is, when we say that our talk of reality depends on us and our minds, we are describing ourselves in a certain way. The idea is that whenever we express conceptual idealism, that is, whenever we say that what can be said or thought is dependent on us, we are saying something that is itself dependent on us and our minds. But even this second way of characterizing conceptual idealism cannot be correct. The reason is that we may want to understand how our account of conceptual idealism depends on us. And the only reasonable account of this relation of dependence seems to require an objective and external property subjects have, that is, the property of representing the objects in the world (themselves included) in a certain way. Let me try therefore to formulate my second doubt. In order to account for conceptual idealism, it seems to me that we should acknowledge both an internal and an external reading of our talk about properties. In particular, we should accept that subjects have at least one property that may be thought or said only in an external reading of properties, that is, the property of representing objects and themselves in a certain way. But if my interpretation is correct, then the internalist view of talk about properties (according to which we may adopt only an internal interpretation of our talk about properties) is refuted by conceptual idealism (which requires the existence of at least one property that can be accounted for only in an external reading of talk about properties).

88

Natural and Artifactual Objects in Contemporary Metaphysics

Concluding remarks Let me sum up what I have been doing. I advanced a couple of doubts concerning Hofweber’s internalist view of our talk about properties, as, for example, the property of ‘being natural’ and the property of ‘being artifactual’. The first doubt relates to one of the premises of the argument in support of the internalist view of our talk about properties, namely the alleged co-reference of ‘the property of being X’ and ‘being X’. I suspect that the premise is not established and therefore, in my opinion, the conclusion remains dubious. My second doubt concerns a consequence of the internalist view of properties and propositions, that is, conceptual idealism. It seems to me that conceptual idealism is incompatible with the internalist view of talk about properties. In particular, in order to account for conceptual idealism, a property that is itself a condition of conceptual idealism is presupposed (i.e. the property had by subjects of representing objects and themselves in the world), and the existence of this objective and external property is incompatible with the internalist view of talk about properties.

Acknowledgements I am indebted to Alfredo Tomasetta for helpful and detailed advice. I had the opportunity to be the discussant of Hofweber’s ‘The internalist conception of (our talk about) properties’ at the Conference ‘The Metaphysics of Properties and Relations’ held in Bergamo in July 2015, organized by Andrea Bottani, Thomas Sattig and Alfredo Tomasetta; this contribution to the volume edited by Richard Davies has been an occasion to reflect more deeply on Hofweber’s proposal and to clarify my doubts.

Notes 1 The distinction between esoteric and egalitarian metaphysics is central to Hofweber (2009) and Hofweber (2016a, ch. 13). 2 Hofweber (2005, 2016a, ch. 3 and 2016b) acknowledges his debt to Carnap and his different interpretation of the internal and external reading of the existential questions. 3 See, for example, Hofweber (2016a), p. 208: ‘Predicates and modifiers are not referential, but singular terms are.’

Talking about Properties 4 5 6 7 8

89

See Hofweber (2016a, p. 29). The examples are in Hofweber (2016a, p. 29). See Hofweber (2016a, §11.7) See Hofweber (2016a, ch. 9). See Hofweber (2016a, p. 256): ‘What there is to represent about reality is somehow due to us and our minds. This is a version of idealism, and it is in a sense the natural companion of the effability thesis.’

References Carnap, R. (1956), ‘Empiricism, Semantics and Ontology’, in R. Carnap, Meaning and Necessity, Chicago, University of Chicago Press, 205–21 Hofweber, T. (2005), ‘A Puzzle about Ontology’, Noûs 39(2): 256–83 Hofweber, T. (2009), ‘Ambitious, Yet Modest, Metaphysics’, in D. J. Chalmers, D. Manley and R. Wasserman (eds), Metametaphysics, Oxford, Clarendon Press, 260–89 Hofweber, T. (2016a), Ontology and the Ambitions of Metaphysics, Oxford, Oxford University Press Hofweber, T. (2016b), ‘Carnap’s Big Idea’, in S. Blatti and S. Lapointe (eds), Ontology after Quine, Oxford, Oxford University Press, 13–30 Quine, W. V. O. (1948), ‘On What There Is’, Review of Metaphysics 2: 21–38 Quine, W. V. O. (1951a), ‘Two Dogmas of Empiricism’, Philosophical Review 60: 20–43 Quine, W. V. O. (1951b), ‘On Carnap’s Views on Ontology’, Philosophical Studies 2: 65–72

Part Two

Where Do Limits Lie?

5

The Eye of the Needle: Seeing Holes Clotilde Calabi

Immaterialism and indirect seeing Sam cannot thread a needle – he asks for help, saying that he doesn’t see the eye of the needle well enough. Mary is at the beach playing with sand. After making a hole, she says that she is going to fill it with water. Tom is approaching a tunnel and believes that it looks 500 meters away. Ben looks at the net and wonders whether a fox enlarged a small hole in it and killed three chicks. There are all sorts of holes that we claim we see (or not see well enough) – splits, clefts, orifices, apertures, tunnels in mountains and eyes of needles. In justifying our claims, we make allegations about our visual impressions: it looks like there is a hole of which we are visually aware, which can be filled, plugged, enlarged, searched, threaded and so forth. My question is this: what exactly do we see when our claim that we see a hole is true? Do we see holes as we see bona fide objects such as houses, chairs and cats? In our frequent statements about holes, we presuppose their extra-mental existence. Yet they ‘appear to be absences, thus non-existing items’ (Bertamini and Casati 2015). Do holes exist? Materialists and immaterialists have contrary opinions. Materialists hold that only concrete particulars exist and thus any statement about holes should be reduced to a statement about properties of material objects and particularly to statements about the linings of holes (Lewis and Lewis 1970). Immaterialists hold that holes exist, but they are not bona fide objects. Rather, they are abstract particulars (Casati and Varzi 1994). I start with three not-negotiable claims, namely 1. Holes exist. 2. We see holes. 3. Seeing a hole is a visual experience of some kind.

94

Natural and Artifactual Objects in Contemporary Metaphysics

The problem I shall address is the following. Immaterialists hold that, given that holes are not bona fide objects, we cannot see them directly. Nevertheless we see them indirectly. This is interesting because it shows that ontology impinges on visual phenomenology. However, it faces the following problem. A compelling answer to the question ‘How do we see ordinary objects, such as houses, chairs and cats?’ is that we see them directly. Moreover, we have the intuition that we see holes exactly in the same way. In other words, the intuition is that splits, clefts, orifices, apertures, tunnels in mountains and eyes of needles are something that we see directly. Therefore, immaterialism is somewhat in conflict with our pretheoretical intuitions. Casati and Varzi have a argument to the effect that, pace our intuition, we do indeed see holes indirectly. Call this claim ‘indirect seeing’. It is easy to show why immaterialism about holes entails indirect seeing. Indeed, it’s plausible to say that one can see directly material entities only. It follows that if holes are immaterial, then one cannot directly see them. Can holes be seen at all? Yes, because holes depend for their existence on material entities hosting them. Hence, it is plausible to say that we can see them, by seeing directly their material hosts. For example, we see the hole in the cheese by seeing directly the cheese surrounding the hole, and we see the eye of the needle by seeing the needle. Casati and Varzi give us two examples of indirect seeing that they consider paradigmatic, thus suggesting that when we see holes we are in a state sufficiently similar to the states involved in the examples. Their first example is the equator. The equator is an abstract entity and we can see it only by seeing its physical realization (e.g. we can see a line that has been drawn on a terrestrial globe). If we take this example at face value, we have the claim that to see indirectly an F is to see (directly) some G and believe that it is the physical realization of F. If seeing a hole is like seeing an abstract entity, we have the conclusion that seeing it amounts to the perception-based belief that a certain region in front of us is a hole. Surprisingly, Casati and Varzi admit that we recognize holes on the basis of light information and shadows distribution. Information concerns the surface texture of the objects hosting the holes: our visual impression of a hole is generated by some kind of irregularity of texture that we detect. For example, our impression of a hole is generated by the irregularity of a drawing, the void in a weft, the empty space between the books aligned on a bookshelf, a clearing in a dense forest. Of course, our visual impression can be the basis of the belief that there is a hole there, but to have an impression of a hole is not to have that belief, nor does it involve having it. In other words, they admit that seeing

The Eye of the Needle

95

holes is a visual achievement. Unfortunately, this makes unfitting Casati and Varzi’s first example. Consider now their second example. Suppose that an invisible man is walking on the sand and leaves footprints behind. According to Casati and Varzi we see the invisible man indirectly by seeing (not in an indirect manner) his footprints in the sand. However, this second example doesn’t work either: the footprints that the invisible man leaves behind are holes in the sand. Thus, if the example tells us anything, it tells that we see holes directly (i.e. not in an indirect manner), after all. Let me then modify the example slightly, by supposing that the invisible man starts walking after dipping his feet in red paint. We might claim that we see directly the red footprints and, indirectly, we see him. Likewise, as Casati and Varzi would put it, we might claim that when we see holes, there is something we see directly. A good candidate for this ‘something’ is the surroundings of the hole. Does the second example tell us anything significantly different from the first one? I am not sure that it does. It does not straightforwardly show in which way we can combine the idea that seeing holes is a visual achievement with the idea that we see them indirectly – as Casati and Varzi hold. In fact, its most natural interpretation involves belief: seeing the invisible man is seeing the footprints and believing that there is a walking invisible man. Likewise, seeing a hole would be seeing the hole’s surroundings and believing that there is a hole there. Once again, this runs counter to the claim that seeing a hole is visual all the way through, which Casati and Varzi endorse. At this point we ponder whether the claim that we see holes indirectly (i.e. we see them in a different manner than we see bona fide objects around us) and the claim that seeing a hole is a visual achievement are even compatible. But Casati and Varzi have an argument for indirect seeing. Before presenting their argument, let me say that Casati and Varzi agree with the common psychological claim that the visual field has a figure-background structure. Figures are what the observer focuses the attention on, and have boundaries. Boundary detection depends on light information and it goes hand in hand with shape attribution. Thus, given that figures have boundaries, they also have shape. Instead, the background has no boundaries; thus, it does not have a shape. Rubin’s vase is a case in point. As we know, we can shift from seeing the vase into seeing the two faces and vice versa. Although the figure-ground reversal does not change the boundaries per se, it changes the information encoded: curvature polarity is reversed and the inside becomes the outside.1 When we switch from seeing the vase into seeing the two faces, the faces only

96

Natural and Artifactual Objects in Contemporary Metaphysics

have boundaries. To see the faces is to detect their boundaries, and to see the vase is to detect the vase’s boundaries. To what boundaries belong depends on what is the figure. With respect to holes, Casati and Varzi contend that we attribute shape to them and hence our visual system detects their boundaries, yet boundary detection does not bring about direct seeing. Questions loom. If the boundaries belong to holes, why detection of their boundaries does not bring about direct seeing? Interestingly enough, Casati and Varzi predict that the shape of a visual hole is immediately available to the observer. If their prediction is correct, the visual system considers holes like ordinary objects by assigning boundaries to them and ‘there is no qualitative difference between the perceived shape of a hole and the perceived shape of an object when such shapes are congruent’.2 If visual differentiation is brought about by boundary detection and we detect the boundaries of holes, we differentiate them visually and, hence, we see them directly, after all. But this is not what Casati and Varzi contend: while claiming that there is a relation between boundary detection and visual differentiation and hence a relation between boundary detection and direct seeing for bona fide objects, they claim that, in the case of holes, although we detect their boundaries, boundary detection does not bring about direct seeing. Why not? Given the above characterization of direct seeing, the reason must be that in the case of holes detection of their boundaries does not bring about their visual differentiation. Again, why not? Casati and Varzi’s answer: because of the metaphysics of holes. Recall that for Casati and Varzi holes are immaterial particulars. Thus they appear to accept the following argument: 1. For any observer S and any object x, if S detects x’s boundary then S sees x directly only if x is a material object (by assumption). 2. Holes are not material objects. 3. Thus, we do not see them directly. We want a principled argument to dispel our intuition that we see holes directly. Casati and Varzi give us such an argument, a worry of question begging lingers.

Seeing holes We should start anew on firm ground. Given that we haven’t said much about seeing, our starting point will be the semantics of the verb to see. To see is a verb of perception and, like other verbs of perception, it occurs in sentences

The Eye of the Needle

97

that can have either a direct object construction or a propositional construction. Moreover, verbs of perception are factive verbs. This claim amounts to the following theses: 1) Any sentence of the form ‘A sees F’, where ‘F’ is a proper name, a common noun, or a definite description, entails ‘There is F’. 2) Any sentence of the form ‘A sees that p’ (where ‘p’ is a sentence) entails ‘p’. For example, the sentence ‘Maria sees a cat’ entails ‘There is a cat,’ and the sentence ‘Maria sees that the cat on the couch is black’ entails ‘The cat on the couch is black’. If someone says ‘I see a pink elephant, but there is no pink elephant’ or ‘I see that the cat on the couch is black, but it is not black’, this is an improper use of the verb to see, and he or she should use instead the locution ‘it seems to me that I see’ or ‘it is as if I were seeing’. Compare the verb to see with the verb to imagine, which can equally well occur in sentences with a direct object construction and in sentences with a propositional construction. ‘To imagine’ is not a factive verb. This means that ‘Maria imagines a unicorn’ does not entail ‘There is a unicorn’, and ‘Maria imagines that the cat on the couch is black’ does not entail ‘The cat on the couch is black’. Finally, the above theses on the meaning of the verb to see find their reflection in two theses on the nature of seeing, namely: 3) For any x and for any y, if x sees y, then y exists. 4) For any x and for any p, if x sees that p, then p. The distinction between the two constructions corresponds to two different types of seeing, namely direct seeing and indirect seeing. Directly seeing an object is a belief-independent visual achievement. In other words, if I see directly F, I may have no beliefs to the effect that there is F or that there is something that looks like F. In contrast to direct seeing, indirect seeing is a cognitive state, involving beliefs of the above kind. For example, if I see that a cat is on the mat, I have some kind of visual experience on which ground I believe that there is a cat and that cat is on the mat. My focus here is on direct seeing only. Both the layman and the theorist agree that the sense of sight allows us to single out objects and localize them. By seeing them, we know by and large where they are by the way they look. Of course, in order to see an object, we do not have to see it entirely. For example, while facing an opaque square box on the floor, we see its front side only. We do not see its backside, the inner sides and the side on the floor. Nevertheless we see the box. We may also see only one part of its front side, if that box is in part behind another box, and, again, we see the box. Notice further that given the observer’s viewpoint and particular

98

Natural and Artifactual Objects in Contemporary Metaphysics

light conditions, any object that we see must have a particular visual aspect, different from that of the background. For example, I can see a sheet of paper hanging on the wall insofar as I visually differentiate that sheet of paper from the wall. Of course, if both the paper and the wall are white, there is no visual contrast available and I do not see the paper. Finally, I may see an object even without differentiating it entirely from its background. Suppose that I am facing a big wall. I do see it, even if I do not single it out entirely against a background, if I detect, say, the surroundings of a crack on it. Those surroundings are part of the wall and by seeing them I see the wall. Fred Dretske captures the gist of the above account of seeing by saying that there is one type of seeing, which is belief-independent visual differentiation.3 Thus he makes a distinction between this type of seeing, which he calls ‘simple seeing’ and another type, namely ‘epistemic seeing’. Whereas epistemic seeing depends on beliefs, simple seeing does not. In terms of the terminology we have previously introduced, simple seeing is a visual achievement. We called it ‘direct seeing’. Here is then Dretske’s definition of direct seeing.4 Direct Seeing: Directly seeing an object consists of the visual differentiation of the object or some of its parts from its immediate background. I said that in order to see an object we do not have to see it entirely. But is seeing any un-detached part of an object sufficient for seeing the whole? Interestingly enough, Dretske’s answer is ‘No’: How much of a thing S must visually differentiate in order to simply see it is a question that, in the abstract, divorced from the sort of thing he is seeing and the circumstances, under which he is seeing it, cannot be answered. We can say, as a necessary condition, that some part of D … must look some way to S, but this will hardly do as a sufficient condition. … Trying to answer such questions is a bit like trying to decide when a boy, any boy, becomes a man. (Dretske 1969, p. 27; emphasis mine)

Dretske’s idea is that conversational context as well as the observer’s interests and skills matter in establishing whether S sees enough of D for simply seeing D. He makes a distinction between visually differentiating D by visually differentiating one part of it and visually differentiating one part of D only. In certain circumstances and from a particular point of view, it may be that visual differentiation of one part suffices to differentiate visually D itself; in other circumstances, it may not be enough. S’s interests, focus of attention and the conversational context are part of the background conditions enabling S to identify an object. If seeing the whole requires that the parts one sees be significant

The Eye of the Needle

99

for identifying the whole, then the conditions for identification must be part of the conditions for simply seeing a partially obscured D. For S the viewing conditions must be such that they allow identification of D, that is, S visually differentiates D if and only if, given certain background conditions, S visually differentiates a non-detached part of D, and that part characterizes D enough for identifying it. However, identification is conceptual. If direct seeing is but a visual achievement, it does not involve any capacity for conceptual identification. To see an object is to track it by the way it looks, that is, to know by and large where it is by the way it looks. This takes us to the following definition: Direct Seeing Revised: For any x and any y, x directly sees y if and only if x would know by the way y or any non-detached part of y now looks to x, where y is. Thousands of difficult cases lurk just round the corner. Here are two of them. You get lost in the fog. Later on, someone asks you whether you saw some dark patches. You recall that you saw them. They were the windows of a house. Did you see the house? Given that directly seeing an x is visually differentiating it from its immediate background, does the man lost in the fog visually differentiate the house by visually differentiating its windows? The question is whether the man lost in the fog would know where the house is on the basis of how it looks to him. I think he would. The house occupies an area around those dark shadows. Consider this other case. You are in front of a white wall and there is a white sheet of paper on it with a tiny red spot. You see some spotted area of the paper. Are you seeing the paper? Let me put it this way. If the red spot is an un-detached part of the sheet of paper, the paper is by and large where the spot is: it starts from there. If the windows are an un-detached part of the house, the house is by and large where the windows are: it starts from there. In this sense, we just need to see some parts to see the whole, and no special constraint is due on which parts these should be – any part would do. One may not know that this is a house or that this is an A4 sheet of paper, but this is not an issue. Epistemic seeing does require that the parts sufficiently characterize the whole, but direct seeing does not.5 I said that by sight we localize and single out objects from the background – we know by and large where they are by the way they look. Suppose now that you are about to enter a tunnel. Do you see it? Of course you do. You detect its boundaries, since you distinguish a back area in front of you from its surroundings. Once you are in the tunnel, do you see it? Are there boundaries that you detect? Suppose that all you see are two walls of concrete that appear to

100

Natural and Artifactual Objects in Contemporary Metaphysics

extend to the horizon, and meet at a vanishing point, and a dark asphalt strip. The concrete walls and the asphalt strip are not part of the tunnel. Thus, you do not see the tunnel. Suppose now that you see luminous patch at the end of what looks like a dark tunnel. Do you see the tunnel? You see some of its boundaries and hence one part of it. Hence, you see the tunnel. Casati and Varzi contend that holes have boundaries, but boundary detection does not bring about visual differentiation. Along this line Bertamini and Croucher argue that in the presence of a hole, an observer judges many aspects of the hole indirectly, by looking at the object enclosing it.6 There is no reason to say that we do not see holes directly, since the visual constraints on direct seeing that apply to bona fide objects apply to holes also. Note that I am not committed to saying either that the boundaries belong to the hole or that they belong to the enclosing object. Nor am I committed to endorsing immaterialism. My only assumptions are that holes exist and that we see them. Susanna Siegel (2009, p. 539) makes the following point: It is clearly possible to see holes, and to represent them experientially as holes. Here we seem to have a case in which there is a property – the property of being a hole – that is represented in experience, even though we do not represent in experience any more specific feature that distinguishes holes from non-holes. Suppose you see some cheese with a hole in it, and your experience is veridical. … When the cheese appears to have a hole in it, does it appear to host an immaterial particular, or does it appear to have a material part – or neither, or both? If the cheese appeared to have both of these more specific properties, then the experience could not be veridical after all, since the hole cannot be both material and immaterial. The best answer seems to be ‘neither’: visual experience just seems neutral on whether the hole in the cheese is a material but negative part or an immaterial part hosted by a material particular. Visual phenomenology does not seem to count for or against either of these theories, even if there are faults to be found with the theories on other grounds. If so, it seems likely that visual experiences do not take a stand on the metaphysics of holes. Suppose some metaphysical theory of holes turns out to be correct […]. If experience is neutral in the way I have described, then we have a case in which an object has a distinctive feature H if it has a hole, and experience represents the property of having a hole, but does not represent distinctive feature H.

In this chapter, I have defended a similar claim. If the visual system treats holes as bona fide objects by assigning boundaries to them, we see them directly after all. The immaterialist may be right, that is, holes qua immaterial entities may not be bona fide objects. If she is right, holes have the distinctive feature of

The Eye of the Needle

101

being immaterial, but the experience does not represent that feature. It is not even obvious that it represents them as absences. Thus, I do not see any reason for making ontology wear the trousers and impose the claim that we see holes indirectly.

Notes 1 Bertamini and Croucher (2003, p. 35). 2 Bertamini and Croucher (2003, p. 37). 3 Dretske identifies it with visual differentiation, which is ‘a pre-intellectual, pre-discursive sort of capacity which a wide variety of beings possess. It is an endowment which is largely immune to the caprices of our intellectual life. Whatever judgements, interpretations, beliefs, inferences, regrets, memories or thoughts may be aroused by the visual differentiation of D, the visual differentiation of D is, itself, quite independent of these accompaniments. It can take place with or without them (or, better, with or without any particular one of them), although with human beings it would be extraordinary indeed if it took place without any such concomitant’ (Dretske 1969, p. 29). 4 See also Mulligan (2003). 5 I discussed these cases and provided a detailed argument in favour of Direct Seeing Revised in Calabi (2012). 6 ‘Our conclusion is that the shape of a hole is known indirectly via the shape of the object enclosing it. Therefore, our representation of a hole is qualitatively different to our representation of a figure possessing the same contours, and this is due to the drastic differences in part structure between the two. Let us take the example of a fish-shaped hole (discussed in Casati and Varzi, 1994). It is true that people can recognize this as a fish, but our claim is that to do that people must be doing one of two things; either they enforce a figure/ground reversal so that they do not see this region as a hole, or, alternatively, they rely on a slower process that compensates by means of our vast cognitive resources for the fact that perceptually that figure is not the figure of a fish (albeit it is a figure containing enough information to infer the presence of a fish)’, p. 34.

References Bertamini, M., and Croucher, C. (2003), ‘The Shape of Holes’, Cognition, 87: 33–54. Bertamini, M., and Casati R. (2015), ‘Figures and Holes’, in J. Wagemans (ed.), Oxford Handbook of Perceptual Organization, Oxford University Press, Oxford, 281–93.

102

Natural and Artifactual Objects in Contemporary Metaphysics

Calabi, C. (2012), ‘Cats! Michotte’s New Enigma of Perception’, in Calabi (ed.), Perceptual Illusions. Philosophical and Psychological Essays’, Palgrave McMillan, London, 124–41. Casati, R., and Varzi, A. (1994), Holes and Other Superficialities, MIT Press, Cambridge (MA). Dretske, F. (1969), Seeing and Knowing, University of Chicago Press, Chicago. Lewis, D., and Lewis, S. (1970), ‘Holes’, Australasian Journal of Philosophy, 48: 206–12. Mulligan, K. (2003), ‘Seeing, Certainty and Apprehension’, in H. Fossheim, T. M. Larsen and J. Rickard Sagend (eds), Non-Conceptual Aspects of Experience, Oslo, Onipub vorlag, 27–44. Siegel, S. (2008), ‘The Visual Experience of Causation’, The Philosophical Quarterly, 59: 1–22.

6

Bona Fideness of Material Entities and Their Boundaries Lars Vogt

Introduction Natural science investigates the material realm, its underlying laws and repeatable processes and the properties and dispositions of various types of material entities. Investigating the material realm usually requires partitioning matter into smaller parts, that is, distinct material entities that can be observed, studied and interacted with. Because partitioning matter into manageable parts is crucial to most practices of empirical research, as it underlies almost every experiment and every scientific observation, natural scientists are interested in whether it is possible to partition the material realm into natural units, that is, units that exist independent of any human partitioning activities. Such natural units are considered to be material entities that exist and that possess specific properties independent of the existence of other material entities. The existence of such natural units would justify a reductionist approach to empirical research, if and only if we have operational/empirical criteria for their identification. Boundaries have been suggested to be essential for identifying natural units and demarcating them from their environments (Smith and Varzi 1997a, 2000; Smith 1994, 1995, 2001). Material entities, such as the moon, an apple, you and I, are not only extended in space, but are clearly demarcated from their respective environments. They possess a continuous outer bona fide boundary, which usually can be recognized as their outer surface. This boundary symmetrically demarcates the object from its complement, and vice versa (Smith and Varzi 1997a). Bona fide boundaries are natural or mind-independent boundaries (Smith 1995, 2001; see also ‘Natural Boundaries’, Varzi 2015). They are physical boundaries in the things themselves

104

Natural and Artifactual Objects in Contemporary Metaphysics

that exist independently from human perception (Smith 1994, 1995, 2001; Smith and Varzi 1997a, 2000). Bona fide boundaries can be demarcated on grounds of ‘some interior physical discontinuity or some qualitative heterogeneity among the parts of the object (some sharp gradient of material constitution, color, texture, electric charge, etc.)’ (Smith 2001, p. 132). The surface of an apple, the surface of our skin and the surface of the moon are examples of bona fide boundaries of material objects. Because the moon, an apple, you and I are material entities, they consist of divisible matter that can be partitioned along inner boundaries into their spatial parts, either in reality or virtually. Some of these inner boundaries are bona fide boundaries, but most of them are fiat boundaries. Fiat boundaries are artificial or mind-dependent boundaries and thus non-physical boundaries that depend on human decision – they are the products of mental and linguistic activities (Smith 1994, 1995, 2001; Smith and Varzi 1997a; see also ‘Artificial Boundaries’, Varzi 2015). Fiat boundaries cannot be directly seen (Bittner and Smith 2001). They are arbitrarily imposed by human partitioning activities and thus independent of an autonomous mind-independent world (Smith 1995, 2001; Smith and Brogaard 2000; Varzi 2015). Fiat boundaries owe their existence merely to some conventional laws, agreements, political decrees and habits, but do not separate anything in reality (Smith 1995, 2001; Smith and Varzi 1997a, 2000). The inner boundary between head and trunk, the inner boundary between heart and aorta and the equator are examples of fiat boundaries. Whereas a bona fide boundary is considered to belong to the physical object and to symmetrically demarcate it from its complement (Smith and Varzi 1997a), thereby considering the object to be closed and its complement to be open, a fiat boundary is considered to be shared by the two fiat parts it demarcates (see ‘Coinciding Boundaries’, Smith and Varzi 1997a, 2000). The distinction between bona fide and fiat boundaries is considered to be categorical and absolute: no instance of the type fiat boundary instantiates the type bona fide boundary, and vice versa (Smith and Varzi 1997a). The underlying account of boundary sticks to a mathematical understanding of boundaries, in which a three-dimensional physical object possesses a single continuous outer boundary, that is, its two-dimensional surface – the boundary is not a threedimensional body itself. According to this mathematical notion, boundaries are necessarily crisp in the sense that they are always of a dimensionality one less than the object they bound. When confronted with concrete examples, like, for instance, deserts, dunes or the Caribbean Sea, which seem to have no crisp

Bona Fideness of Material Entities and Their Boundaries

105

surfaces as their boundaries, but rather indeterminate (i.e. fuzzy, vague) border zones or boundary-like regions that share the dimensionality with their host, the interpretation favoured by various authors is that the region only specifies where the crisp boundary must be located, but we simply cannot narrow down to its actual location (e.g. Sorensen 1988; Williamson 1994; Smith 1995, 2001; Smith and Varzi 1997b, 2000). This argumentation traces the indeterminacy back to conceptual issues owed to linguistic and epistemological problems rather than to an ontological origin (cf. Varzi 2015). The distinction between bona fide and fiat boundaries is important, because it allows for distinguishing bona fide objects (i.e. natural units) from fiat material entities (i.e. artifactual entities). Bona fide objects possess a single continuous bona fide outer boundary, whereas fiat material entities possess some fiat outer boundary (Smith and Varzi 1997a, 2000; Smith 1994, 1995, 2001). The distinction between bona fide objects and fiat material entities in turn is important not only for the identification of natural units but also for organizing, structuring and managing domain knowledge. Top-level categories of bona fide objects have been used for organizing the material realm according to compositional hierarchies (e.g. Woodger 1929; Novikoff 1945; Oppenheim and Putnam 1958; Simon 1962; Wimsatt 1976, 1994; Mayr 1982; Eldredge 1985; Riedl 1985; Levinton 1988; Craver and Bechtel 2007; Winther 2011; Vogt 2017). The distinction between bona fide and fiat entities is also important for the development of formal top-level ontologies,1 such as the Basic Formal Ontology (Arp et al. 2015; Smith et al. 2015), and domain reference ontologies, because it provides basic criteria for distinguishing top-level categories of material entity. In the following, I discuss the role that the distinction between bona fide and fiat boundaries has played in characterizing bona fide objects and thus natural units and the specific problems connected to this direction of characterization. I discuss a new approach of characterizing the bona fideness of natural units independent of their boundaries, but instead based on different types of causal unity. I relate this approach to the role of different frames of reference in the life sciences and distinguish a spatio-structural from a functional and an historical/ evolutionary frame of reference. I propose two additional types of causal unity that are associated with the latter two frames of reference. I conclude by proposing to reverse the characterization of different boundary types by characterizing bona fide boundaries as the three-dimensional boundary regions that bound natural units and fiat boundaries as mathematical boundaries with a dimensionality one less than the entity they bound.

106

Natural and Artifactual Objects in Contemporary Metaphysics

Problems with bona fideness Problem I: Bona Fideness is granularity dependent According to the above mentioned role of boundaries we would have to say that whether a given boundary is bona fide or fiat, and thus whether a given material entity is a bona fide object or a fiat material entity respectively, only depends on where to draw the line between physical continuity and discontinuity and between qualitative homogeneity and heterogeneity. Unfortunately, physical discontinuity is a matter of scale and therefore of granularity, because with increasing resolution spatial boundaries of material entities become increasingly fuzzy (Vogt et al. 2012; see Unger 1980; Lewis 1983). This fuzziness cannot be completely traced back to a purely linguistic or epistemological problem. At the subatomic level, for instance, any material entity resolves into a swarm of subatomic particles with no clearly determinable boundaries. Therefore, whereas at very fine levels of granularity the notion of abrupt physical discontinuities is questionable and thus the distinction between bona fide and fiat boundaries not decidable anymore, at coarser levels this dichotomy can be maintained (Smith 2001; Smith and Varzi 1997a,b; Vogt et al. 2012). Physical discontinuity also depends on physical connectedness, which in turn comes in degrees due to an underlying continuity problem (Schulz and Johansson 2007; Smith and Varzi 1997b). Part of the problem with connectedness concerns the fact that what can be considered to be maximally self-connected is granularity dependent: except for maybe whole organisms, every anatomical entity of the cellular and supra-cellular level of granularity is connected to neighbouring entities via vessels, conduits, tunnels, ducts, nerve cords, channels, intercellular spaces, junctions or pores (Schulz and Johansson 2007; Smith et al. 2005; Vogt et al. 2012). If bona fide objects are characterized as being maximally self-connected (e.g. Arp et al. 2015), delineating bona fide objects above the cellular level based on physical discontinuity is problematic, to say the least, because at supra-cellular levels physical continuity is always present, but not always relevant, turning bona fideness based on physical discontinuity into a granularity-dependent property (for a detailed discussion see Vogt et al. 2012). Like physical discontinuity, qualitative homogeneity also comes in degrees. Biological entities are the product of evolution and therefore exhibit a high degree of variability and diversity. Therefore, biologists are dealing with a

Bona Fideness of Material Entities and Their Boundaries

107

continuum of forms and functions, in which usually no two individual objects are completely identical regarding their morphological properties (Schulz and Johansson 2007). If no two cells are the same, then any two neighbouring cells within a given aggregate of cells exhibits qualitative heterogeneity, with the consequence that delineating objects above the cellular level based on qualitative heterogeneity is as problematic as with physical discontinuity, because at supracellular levels qualitative heterogeneity is always present but not always relevant, turning bona fideness based on qualitative heterogeneity into a granularitydependent property (for a detailed discussion see Vogt et al. 2012).

Problem II: Bona Fideness is perspective dependent and needs the consideration of frames of reference In its strictest meaning, fiatness implies mind-dependence and bona fideness mind independence. Regarding boundaries, a fiat boundary is thus a boundary that is determined by human fiat, lacking any natural indication, whereas a bona fide boundary bounds an entity that exists independent of human partitioning activities (Vogt et al. 2012). Based on the strict meaning of bona fideness one must conclude that the criteria of physical discontinuity and qualitative heterogeneity are not only problematic because they are granularity dependent but they also fail to sufficiently and unambiguously characterize the boundaries of those entities that are delimited on grounds of their causal properties and dispositions or their historical relations to other biological entities. Biological material entities actively participate in many different ways as causal agents in various biological processes, from all kinds of physiological processes within an organism to all kinds of ecological and social interactions. Moreover, each biological material entity has its own particular history. Evolution, individual development, embryogenesis, death, birth, growth, reproduction, locomotion and so forth – in all these processes specific biological material entities are involved and act as causal units, independent of any human partitioning activities. Although respective functional, developmental and evolutionary units often do possess particular characteristic spatio-structural properties, they do not possess them necessarily so, because these spatiostructural properties do not represent their defining properties. Functional, developmental and evolutionary units thus cannot be characterized exclusively on grounds of their spatio-structural properties, but they nevertheless are

108

Natural and Artifactual Objects in Contemporary Metaphysics

natural units that can be characterized by real properties of a causally dynamic reality – by their dispositions to actively participate in some causal processes or their historical relations to other entities (Vogt et al. 2012). Often, spatio-structurally separable entities coincide with causal units, as, for instance, an apple, which is a unit of reproduction, but at the same time also a spatio-structurally separated entity after it has fallen from the tree. Spatio-structurally separated entities can also coincide with evolutionary and developmental units (see, for example, ‘Character Identity Networks’, Wagner 2014). However, more often they do not coincide: for instance, your left thumb, which is a unit of locomotion that is physically continuous with your left hand; the active centre of an enzyme, which is a physiological–biochemical unit that is physically continuous with the remainder of the enzyme or a bicellular eye, which is a sensory unit that is physically an aggregate of two cells that are embedded in an aggregate of other cells. All these examples demonstrate that in the life sciences we are dealing with entities that can be spatio-structurally fiat but that are nevertheless natural units and thus bona fide entities in the strict sense of bona fideness mentioned above. Whether we understand a given material entity as a natural unit or a fiat entity thus directly depends on the entity’s properties we are interested in and therefore on our frame of reference. As a consequence, the bona fideness of material entities is not only granularity dependent but also dependent on the given frame of reference. A frame of reference is used for modelling a given domain and specifies a way to partition the space–time continuum into spatial and temporal regions based on a specific set of criteria. Moreover, a frame of reference can be applied to function like a filter that defines a focus by specifying a set of properties that are relevant for the frame of reference while filtering out all other properties. Because a granular perspective (see Keet 2008, 2006a,b) can be understood as a specific frame of reference, the bona fideness of material entities is perspective dependent. A direct consequence of the perspective dependence of the bona fideness of material entities is that a given material entity can have more than one bona fide mereological partition, because it may represent a natural unit according to more than one perspective. In other words, how a given material entity is partitioned directly depends on the perspective one is interested in.2 Different perspectives yield different partitions (Vogt 2010), with each perspective potentially having its own associated bona fide mereological partition and thus its own bona fide granularity tree (for a detailed discussion of basic perspectives in biology see Vogt et al. 2012; Vogt 2017).

Bona Fideness of Material Entities and Their Boundaries

109

A new approach to bona fideness of material entities Causal unity as a criterion for bona fideness of material entities The problem of the granularity-dependence of physical discontinuity that affects the practical applicability of the concept of bona fide boundaries can be circumvented elegantly by characterizing bona fide objects in reference to causal unity. According to Smith et al. (2015; see also Arp et al. 2015), bona fide objects can be characterized as entities that are causally relatively isolated (Ingarden 1983; Smith and Brogaard 2003), that is, they are both (1) structured through and (2) maximal relative to a certain type of causal unity (Smith et al. 2015). Smith et al. (2015) characterize a causally unified entity b as a material entity that has material parts that are tied together in such a way that if some sufficiently large part c of b is in the interior of b at time t, and then moved to some location outside of b, then either b with all its other parts will be moved in a coordinated fashion along with c or b will be damaged during c’s movement. Smith et al. (2015) characterize three different types of causal unity: 1. Causal unity via internal physical forces: The parts of the unified entity are causally unified through physical forces (e.g. fundamental forces of strong and weak interaction, covalent bonds, ionic bonds, metallic bonding, etc.) that are strong enough to maintain the structural integrity of the entity against the strength of attractive or destructive forces from its ordinary neighbourhood. Examples are atoms, molecules, portions of solid matter, as, for instance, grains of sand, lumps of iron. 2. Causal unity via physical covering: The parts of the unified entity are causally unified through a common physical covering, as, for instance, a membrane. This covering may have holes, but must be completely connected3 and must still serve as a barrier for entities from inside and entities from outside that are above a certain size threshold. Examples are organelles, cells, tissues, organs. 3. Causal unity via engineered assembly of components: The parts of the unified entity are causally unified through screws, glues and other fasteners. Often, the parts are reciprocally engineered to fit together (e.g. dovetail joints, nuts and bolts). Examples are cars, ballpoint pens, houses, shoes, power grids. Unfortunately, this characterization of bona fide objects does not cover all the objects that can be recognized in reference to qualitative heterogeneity. In fact, many natural units identified in biology are not covered by the three types of

110

Natural and Artifactual Objects in Contemporary Metaphysics

causal unity characterized in Smith et al. (2015).4 This applies to functional units in particular. For instance, eyes, many larval eyes and some adult eyes are composed of only two cells, a photoreceptor cell and a supportive cell with shading pigment (Eakin and Hermans 1988; Purschke 2005; Jékely et al. 2008; see also Richter et al. 2010). These bicellular eyes (ocelli) are located adjacent to each other within an aggregate of other cells. Such a bicellular eye forms a functional sensory unit and exists as such independent of any human partitioning activity. It therefore represents a natural unit, but it is maximal neither with respect to some physical covering5 nor with respect to internal physical forces. It is thus not covered by any of the cases of causal unity listed above. However, a bicellular eye can be demarcated as an eye on grounds of qualitative heterogeneity, because of the specific structure of the photosensitive microvilli of the photoreceptor cell and the characteristic opaqueness of the shading pigment granules of the supportive cell (see Richter et al. 2010). Regarding boundaries, with the focus on causal unity Smith et al. (2015) retain the mathematical notion of boundaries, but drop the distinction between bona fide and fiat boundaries, because if boundaries of bona fide objects are conceived as external surfaces in this mathematical sense, then they are something fiat (Smith et al. 2015; Arp et al. 2015). As a consequence, all boundaries are considered to be fiat, assuming no coincidence with physical discontinuities and associating the ontology of boundaries more closely to the ontology of regions (Smith et al. 2015). However, if the boundary of every material entity is a fiat boundary, then any material entity, whether it is a bona fide object or a fiat material entity, is demarcated from its complement by a two-dimensional fiat boundary. This is insofar problematic, as Smith et al. (2015) define fiat object part as being demarcated by a two-dimensional fiat boundary, but if all material entities are demarcated by two-dimensional fiat boundaries this necessarily implies that bona fide objects are fiat material entities, just like fiat object parts. In this context it is interesting to see that, although the distinction between fiat and bona fide boundary plays no direct role anymore regarding the characterization of bona fide objects, Smith et al. still contrast fiat object parts with bona fide object parts in reference to bona fide boundaries: [Bona fide object parts] are themselves objects (for example a cell is a bona fide object part of a multi-cellular organism), and are marked by bona fide boundaries, or in other words by physical discontinuities. (Smith et al. 2015, p. 39)

Therefore, whereas Smith et al.’s (2015) new focus on causal unity represents a promising framework for ontologically characterizing natural units in biology,

Bona Fideness of Material Entities and Their Boundaries

111

the distinction between bona fide and fiat boundary still seems to be important. And it probably will remain to be important, because boundaries are particularly relevant in the epistemological situation of practical research: in order to be able to actually partition an entity into its parts, a researcher needs a diagnostic framework that provides practically applicable (i.e. operational) criteria for identifying instances of bona fide objects. How do we identify that a specific type of causal unity applies? How do we identify internal physical forces taking effect? How do we identify some physical covering? No matter how you look at it, in the end it often depends on identifying ‘some interior physical discontinuity or some qualitative heterogeneity among the parts of the object (some sharp gradient of material constitution, color, texture, electric charge, etc.)’ (Smith 2001, p. 132), because in many cases researchers cannot conduct the experiments required for adequately testing the relevant causal properties. However, due to their granularity-dependence, physical discontinuity and qualitative heterogeneity are not strong but weak criteria that have to be applied and decided on a caseby-case basis.

The need for additional causal unity types and a consideration of frames of reference When discussing their characterization of causal unity via internal physical forces Smith et al. (2015) provide examples of only physical forces that apply to the atomic and molecular scale. They do not explicitly include also physical connections from coarser scales, as, for instance, cell–cell connections, like, for instance, tight junctions or septate junctions. I suggest extending Smith et al.’s characterization of causal unity via internal physical forces to include all kinds of physical connections between material component parts, including screws, glues, bolts and cell-cell connections, because ultimately they all go back to the physical forces discussed by Smith et al. (2015). Moreover, the three different types of causal unity that Smith et al. (2015) discuss with respect to the role of causal unity in the characterization of bona fide objects are ontologically not independent from one another. Both causal unity via physical covering and causal unity via engineered assembly of components existentially depend and thus supervene on causal unity via internal physical forces. The physical integrity and stability of a physical covering depends on the physical forces that hold the component parts of this covering together. Also, screws, glues and other fasteners used to hold together the components of engineered assemblies depend on physical forces.

112

Natural and Artifactual Objects in Contemporary Metaphysics

It is also important to take into account that the relevance of a specific type of causal unity is context dependent: it depends on the frame of reference. Each type of causal unity is connected to a specific general frame of reference. For instance, both causal unity via internal physical forces and causal unity via physical covering, at least as conceived by Smith et al. (2015), are associated with the spatio-structural frame of reference. They share the same general frame of reference because the latter supervenes on the former. One of the reasons for applying a spatio-structural frame of reference lies in inventorying what is given in a particular point in time by focusing on the spatio-structural properties of a given entity (see ‘Spatio-Structural Perspective’ in Vogt et al. 2012). The spatio-structural frame of reference considers reality at a particular point in time, filtering out its dynamical aspects. An apple that has fallen from the tree, for example, is a bona fide object in reference to causal unity via internal physical forces, and thus according to the spatio-structural frame of reference. It can be mereologically partitioned into two types of bona fide entity parts: each seed and the remainder of the apple. But since the spatio-structural frame of reference is not the only frame of reference relevant in the life sciences, this is not the only way an apple can be partitioned into its natural unit parts. Obviously, the three types of causal unity listed by Smith et al. (2015) do not cover all cases of causal unity relevant in the life sciences, nor does the spatiostructural frame of reference cover all relevant frames of reference. Functional units and historical/evolutionary units are neither structured through nor maximal relative to causal unity via physical covering, via physical forces or via engineered assembly of components, nor do they belong to the spatio-structural frame of reference. Regarding functional units, causal unity via engineered assembly of components comes closest, because it involves reference to functions, but it is nevertheless not suited to cover biological functional units. Therefore, I suggest two additional types of causal unity along with their associated frames of reference that are suited to cover the missing cases.

Causal unity via bearing a specific function and the functional frame of reference According to Smith et al. (2015), a function is a disposition that exists in virtue of the bearer’s physical make-up and this physical make-up is something the bearer possesses because it came into being either through evolution (in case of natural biological entities) or through intentional design (in case of artifacts), in order to realize processes of a

Bona Fideness of Material Entities and Their Boundaries

113

certain sort. ... Functions are realized in processes called functionings. (Smith et al. 2015, p. 59)

A disposition, in turn, is an internally grounded realizable entity, which has a material entity as its bearer. If a disposition ceases to exist, then its bearer is physically changed. Moreover, a disposition is realized ‘when and because the bearer is in some special physical circumstances, [and] this realization occurs in virtue of the bearer’s physical make-up’ (Smith et al. 2015, p.58). In other words, a function is integral to its bearer and represents a property of the bearing material entity. The function can thus be used for demarcating its bearer (Vogt et al. 2012; Vogt 2017). Moreover, functions and thus their bearers exist independent of any human partitioning activity and therefore represent bona fide entities. In analogy to Smith et al.’s (2015) general characterization of bona fide objects, a functional unit is both structured through its function and maximal relative to its function. In other words, it is causally unified via bearing a specific function. If the corresponding type of functioning b1 of a functional unit c1 can be partitioned into sufficiently autonomous spatio-temporal parts b2−n, the material entities that bear the disposition to participate in the corresponding types of subfunctionings b2−n are functional component parts c2−n of the functional unit c1 and are thus functional units of a finer level of functional granularity. The component parts c2−n, however, must be spatio-structurally proper parts of the functional unit c1. This notion of functional parthood relates to the mechanismbased approach to levels, in which Craver (2007) referred to the relata of a mechanistic level as acting entities that are neither entities nor activities, but active component parts that contribute to the mechanism so that it properly functions. A functional unit with all its functional component parts thus can be demarcated on grounds of Craver’s mutual manipulability criterion: A part is a component in a mechanism if one can change the behavior of the mechanism as a whole by intervening to change the component and one can change the behavior of the component by intervening to change the behavior of the mechanism as a whole. (Craver 2007, p. 141)

As a consequence, a functional unit can be a material entity that is a physically connected spatio-structural unit, like, for instance, the bicellular eye discussed earlier, but it could also be a physically disconnected spatio-structurally discontinuous spatio-structurally fiat material entity, as, for instance, a metanephridial system that is composed of a spatially scattered group of material entities.6 Functional units thus may lack physical connectedness, but they do exhibit functional connectedness (Vogt et al. 2012).

114

Natural and Artifactual Objects in Contemporary Metaphysics

The here proposed causal unity via bearing a specific function applies to biological as well as artifactual material entities, because both types of entities can be bearers of functions. Causal unity via bearing a specific function is thus more general than Smith et al.’s (2015) type of causal unity via engineered assembly of components. The latter represents a sub-type of the former by restricting it to material entities that are (1) made by humans (artifacts) and that are (2) physically self-connected, which excludes spatially scattered entities. Only the former treats the spatially separated combination of a radio transmitter and a radio receiver as a functional unit of communication. Causal unity via bearing a specific function is associated with the functional frame of reference. The functional frame of reference may be applied for making reliable predictions of what can happen in the future by focusing on dispositional/functional aspects of reality (see ‘Predictive Perspective’ in Vogt et al. 2012). It considers reality as a dynamic system and models an entity’s causal space of possible interactions. Considering the example of an apple discussed earlier; by applying a functional frame of reference that focuses on reproduction one can also identify an apple as a functionally bona fide unit of reproduction that can be further partitioned mereologically into functionally bona fide parts.

Causal unity via common historical/evolutionary origin and the historical/evolutionary frame of reference The example of metanephridial systems discussed earlier demonstrates that more principles of connectedness exist than only physical connectedness. Common historical/evolutionary origin can serve as another principle of connectedness that allows the delineation of historical/evolutionary units. Developmental lineages (e.g. cell lines), genealogical lineages (e.g. families) and evolutionary lineages (e.g. homologs, synapomorphies, species, monophyletic taxa) represent different kinds of historical/evolutionary units in the life sciences (Vogt et al. 2012; Vogt 2016). An historical/evolutionary unit is both structured through and maximal relative to the historical/evolutionary origin of its parts. In other words, it is causally unified via common historical/evolutionary origin. An historical/ evolutionary unit is demarcated so that all of its component parts share the same historical/evolutionary origin, with no material entity not belonging to it sharing the same origin. Most historical/evolutionary units are spatiostructurally scattered material entities, but they nevertheless, just like functional

Bona Fideness of Material Entities and Their Boundaries

115

units, represent units that exist independent of any human partitioning activities and therefore represent bona fide entities (for a detailed discussion see Vogt et al. 2012; Vogt 2016). Causal unity via common historical/evolutionary origin is associated with the historical/evolutionary frame of reference. This frame of reference may be applied for making reliable retrodictions of what has happened in the past by focusing on using a set of known types of repeatable processes to reconstruct the sequence of events that may have lead to the currently observable situation (Vogt 2016; see also ‘Retrodictive (Diachronic) Perspective’ in Vogt et al. 2012). It also considers reality as a dynamic system, but models an entity’s causal space of possible past histories. Again, considering the example of an apple discussed earlier, by applying an historical frame of reference that focuses on the particular developmental origin one can also identify an apple as an historically bona fide unit of development that can be further partitioned mereologically into historically bona fide parts, depending on its developmental sub-units: stalk (developed from receptacle), fruit flesh (developed from hypanthium), blossom end (remains of petals, calyx and stamen), core (developed from ovary) that in turn contains the seeds.

Conclusion Many of the problems regarding the conceptual role and the distinction between fiat and bona fide boundaries discussed above result from a strictly mathematical conception of boundaries. I believe that whereas fiat boundaries can be considered to be mathematical boundaries in the sense that they have a dimensionality that is always one less than the entities they bound, bona fide boundaries cannot resemble such mathematical boundaries. A bona fide boundary of a threedimensional material entity must necessarily be a three-dimensional material entity itself, thus being a boundary region rather than a crisp surface. After all, it only makes sense to consider bona fide boundaries to be physical boundaries, if they are three-dimensional material entities themselves. Whereas a boundary region bounds a bona fide object from its environment, one can still point to the object’s surface as its mathematical fiat border (to use a different term than boundary) that demarcates the object from its environment. If the surrounding environment is itself a material entity, it also possesses a three-dimensional boundary region that demarcates it from the object. In any case, the fiat border

116

Natural and Artifactual Objects in Contemporary Metaphysics

itself is shared by both the bona fide object and its environment, and it is defined by the qualitative difference of its adjacent boundary regions. Considering bona fide boundaries as three-dimensional boundary regions makes sense insofar, as bona fideness of an entity always has causal implications, not only for the entity it bounds but also for us to be able to identify and demarcate the entity as a bona fide unit. Bona fideness of a three-dimensional entity is only detectable for us if its boundary can be physically experienced (directly or mediated through instruments, techniques and experiments), which depends on processes that causally involve a three-dimensional boundary region, and not a two-dimensional mathematical surface. In morphology, we can only demarcate a cell from its neighbouring cells by identifying the cell membrane, which itself is a three-dimensional material entity. Whenever we detect physical discontinuity or qualitative heterogeneity within an object, it always involves one or two inner boundary regions, not only epistemologically but also ontologically – a surface cannot contain physical discontinuities or qualitative heterogeneities. This is an important aspect, as it provides bona fide objects with what Wimsatt called robustness7 (Wimsatt 1994; see also Levins 1966). By characterizing bona fide objects based on strong criteria of causal unity together with weak criteria of interior physical discontinuity and qualitative heterogeneity, instead of the possession of bona fide boundaries, we can now characterize bona fide boundary regions as the boundaries of bona fide objects. The characterization of bona fide boundaries would thus depend on the characterization of the bona fide object that it bounds, thereby reversing the initial direction of characterization. As a consequence, one must first characterize a specific kind of natural unit based on criteria that are independent of the characterization of bona fide boundaries, before its bona fide boundaries can be characterized. This, however, would allow utilizing the concept of bona fide boundaries as operational criteria for the identification of instances of a specific type of natural unit, like, for instance, cells that can be identified by their cell membrane.

Acknowledgements I thank Thomas Bartolomaeus, Peter Grobe and Björn Quast for commenting on an earlier draft of this chapter. It goes without saying, however, that I am solely responsible for all the arguments and statements in this chapter. This study was supported by the German Research Foundation DFG (VO 1244/7-1; VO 1244/8-1). I am also grateful to the taxpayers of Germany.

Bona Fideness of Material Entities and Their Boundaries

117

Notes 1 Formal top-level ontologies are supposed to provide domain- and purposeindependent theories within a formal framework of axioms and definitions for most general terms and concepts, which can be used as a top-level template and formal framework for developing domain reference ontologies and terminology-based application ontologies (Smith, Munn and Papakin 2004; Rosse et al. 2005). 2 Craver 2015, p. 7: ‘The same thing can be carved into parts in many ways depending on what one is interested in describing or explaining’ (Kauffman 1971; Wimsatt 1972). 3 Connected in the sense that a continuous path can be traced between any two points on the surface and that path has no gaps and does not leave the surface. 4 Smith et al. (2015) do not claim completeness regarding their list of cases of causal unity. 5 Contrary to human anatomy, in which organs are usually organized within epithelial compartments, many functional units in invertebrates are not physically separated from neighbouring entities through some membrane or epithelium. But these functional units nevertheless exist independent of human partitioning activities and thus represent natural units. 6 The filtration site is formed by podocytes situated at the blood vessels, whereas the nephridial duct drains the coelomic space that stores the filtrate. 7 ‘Things are robust if they are accessible (detectable, measurable, derivable, definable, producible, or the like) in a variety of independent ways’ (Wimsatt 1994, p. 210f).

References Arp, R., Smith, B., and Spear, A. D. (2015), Building Ontologies with Basic Formal Ontology, The MIT Press, Cambridge, MA, 248. Bittner, T. and Smith, B. (2001), Granular partitions and vagueness. Proc Int Conf Form Ontol Inf Syst – FOIS ’01 2001: 309–320. Available: http://portal.acm.org/citation. cfm?doid=505168.505197. Craver, C. F. (2007), Explaining the Brain, Oxford University Press, Oxford. Craver, C. F. (2015),. ‘Levels’,. in T. Metzinger, Group, vol. 8, p. 26. Craver, C. F. and Bechtel, W. (2007), ‘Top-Down Causation Without Top-down Causes’, Biology and Philosophy, 22: 547–63. doi:10.1007/s10539-006-9028-8. Eakin, R. M. and Hermans, C. O. (1988), ‘Eyes’, in W. Westheide and C. O. Hermans (eds), The Ultrastructure of Polychaeta. Microfauna Mar, vol. 4, 135–56. Eldredge, N. (1985), Unfinished Synthesis: Biological Hierarchies and Modern Evolutionary Thought, Oxford University Press, New York. Ingarden, R. (1983), Man and Value, Philosophia Verlag GmbH, München, 184.

118

Natural and Artifactual Objects in Contemporary Metaphysics

Jékely, G., Colombelli, J., Hausen, H., Guy, K., Stelzer, E., Nédélec, F. and Arendt, D. (2008), ‘Mechanism of Phototaxis in Marine Zooplankton’, Nature, 456: 395–99 Kauffman, S. A. (1971), ‘Articulation of Parts Explanation in Biology and the Rational Search for Them’, Boston Studies in the Philosophy of Science 8: 257–72. Keet, C. M. (2006a), Enhancing Biological Information Systems with Granularity. Knowledge Web PhD Symposium (KWEPSY06). Budva, Montenegro. Available: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.100.2613&rep=rep1& type=pdf. Keet, C. M. (2006b), A Formal Approach for Using Granularity in the Subject Domain of Infectious Diseases. Bozen: Faculty of Computer Science, Free University of Bozen-Bolzano. Available: https://www.inf.unibz.it/krdb/pub/TR/KRDB06-2.pdf. Keet, C. M. (2008), A Formal Theory of Granularity – Toward enhancing biological and applied life sciences information system with granularity Bozen: Free University of Bozen – Bolzano. Available: http://www.meteck.org/files/AFormalTheory OfGranularity_CMK08.pdf. Levins, R. (1966), The Strategy of Model Building in Population Biology. American Scientist, 1966, 421–31, reprinted in E. Sober, ed., Conceptual Issues in Evolutionary Biology (Cambridge, MA: MIT Press 1984). Levinton, J. (1988), Genetics, Paleontology and Macroevolution. Cambridge University Press, Cambridge. Lewis, D. (1983), ‘New Work for a Theory of Universals’, Australasian Journal of Philosophy, 61 (4): 343–77. Mayr, E. (1982). The Growth of Biological Thought: Diversity, Evolution, and Inheritance, Harvard University Press, Cambridge, MA. Novikoff, A. B. (1945), ‘The Concept of Integrative Levels and Biology’, Science (80) 101: 209–15. Oppenheim, P. and Putnam, H. (1958). ‘Unity of Science as a Working Hypothesis’, in H. Feigl, M. Scriven and G. Maxwell (eds), Concepts, Theories, and the Mind-Body Problem, Minnesota Studies in the Philosophy of Science II, Minneapolis, MN: University of Minnesota Press, 3–36. Purschke, G. (2005). ‘Sense Organs in Polychaetes (Annelida.)’, Hydrobiology, 535/536: 53–78. Richter, S., Loesel, R., Purschke, G., Schmidt-Rhaesa, A., Scholtz, G., Stach, T., Vogt, L., Wanninger, A., Brenneis, G., Döring, C., Faller, S., Fritsch, M., Grobe, P., Heuer, C. M., Kaul, S., Møller, O. S., Müller, C. H., Rieger, V., Rothe, B. H., Stegner, M. E. And Harzsch, S. (2010), ‘Invertebrate Neurophylogeny: Suggested Terms and Definitions for a Neuroanatomical Glossary’, Frontiers in Zoology 7: 29. Riedl, R. (1985), Die Spaltung des Weltbilds. Biologische Grundlagen des Erklärens und Verstehens, Parey Verlag, Hamburg, Berlin. Rosse, C., Kumar, A., Mejino, L. V., Cook, D. L., Detwiler, L. T. and Smith, B. (2005), A Strategy for Improving and Integrating Biomedical Ontologies. AMIA 2005 Symposium Proceedings. pp. 639–43.

Bona Fideness of Material Entities and Their Boundaries

119

Schulz, S. and Johansson, I. (2007), ‘Continua in Biological Systems’, Monist 90: 499–522. Simon, H. A. (1962), ‘The Architecture of Complexity’, Proceedings of the American Philosophical Society, 106: 467–82. Smith, B. (1994), ‘Fiat Objects’, in L. V. Guarino and S. Pribbenow, (eds.) Parts and Wholes: Conceptual Part-Whole Relations and Formal Mereology, 11th European Conference on Artificial Intelligence. European Coordinating Committee for Artificial Intelligence, Amsterdam: 15–23. Available: http://ontology.buffalo.edu/smith/ articles/fiat1994.pdf. Smith, B. (1995), ‘On Drawing Lines on a Map’, in A. U. Frank, W. Kuhn, D. M. Mark, (eds.) Spatial Information Theory: Proceedings in COSIT ’95, Springer, Berlin/ Heidelberg/Vienna/New York/London/Tokyo, 475–484. Smith, B. (2001). ‘Fiat Objects’, Topoi 20: 131–48. Available: http://ontology.buffalo.edu/ smith/articles/fiatobjects.pdf. Smith, B. and Brogaard, B. (2000), ‘A Unified Theory of Truth and Reference’, Logique Et Analyse, 169–170: 49–93. Smith, B. and Brogaard, B. (2003), ‘Sixteen Days,’ Journal of Medicine and Philosophy 28: 45–78. Available: http://jmp.oxfordjournals.org/content/28/1/45. Smith, B. and Varzi, A. C. (1997), ‘Fiat and Bona Fide Boundaries: Towards an Ontology of Spatially Extended Objects’, in Spatial Information Theory: A Theoretical Basis for GIS, Springer, Berlin, Heidelberg, 103–19. Smith, B. and Varzi, A. C. (1997b), ‘The Formal Ontology of Boundaries’ Electronic Journal of Analytic Philosophy 5: 1–23. Available: http://ejap.louisiana.edu/ EJAP/1997.spring/smithvarzi976.html. Smith, B. and Varzi, A. C. (2000). ‘Fiat and Bona Fide Boundaries’, Philosophy and Phenomenological Research, 60: 401–20. Available: http://www.jstor.org/stable/26534 92?origin=crossref. Smith, B., Munn, K. and Papakin, I. (2004), ‘Bodily Systems and the Spatial-functional Structure of the Human Body’, in D. M. Pisanelli, (ed.), Medical Ontologies, IOS Press, Amsterdam, Vol. 102, 39–63. Smith, B., Mejino, Jr. J. L. V., Schulz, S., Kumar, A. and Rosse, C. (2005), ‘Anatomical Information Science’, in AC Cohn and D Mark, (eds.) Conference on Spatial Information Theory (COSIT 2005). Lecture Notes in Computer Science, Springer, Berlin, 149–64. Smith, B., Almeida, M., Bona, J., Brochhausen, M., Ceusters, W., Courtot, M., Dipert, R., Goldfain, A., Grenon, P., Hastings, J., Hogan, W., Jacuzzo, L., Johansson, I., Mungall, C., Natale, D., Neuhaus, F., Petosa, A., Rovetto, R., Ruttenberg, A., Ressler, M. and Schulz, S. (2015), Basic Formal Ontology 2.0. Available: https://github.com/BFOontology/BFO/raw/master/docs/bfo2-reference/BFO2-Reference.pdf. Sorensen, R. A. (1988), Blindspots, Clarendon Press, Oxford:. Unger, P. (1980), ‘ The Problem of the Many’, Midwest Studies in Philosophy, 5: 411–67.

120

Natural and Artifactual Objects in Contemporary Metaphysics

Varzi, A. C. (2015), Boundary. Stanford Encycl Philos (Winter 2015 Ed). Available: http://plato.stanford.edu/archives/win2015/entries/boundary/. Vogt, L. (2010), ‘Spatio-Structural Granularity of Biological Material Entities’, BMC Bioinformatics 11. Doi.org/10.1186/1471-2105-11-289. Vogt, L. (2016), ‘Assessing Similarity: On Homology, Characters and the Need for a Semantic Approach to Non-evolutionary Comparative Homology, Cladistics EarlyView : 1–27. DOI:10.1111/cla.12179 Vogt, L. (2017), ‘Levels and Building Blocks – Towards a Domain Granularity Framework for the Life Sciences’, PeerJ Preprints 5: e2429v2 Vogt, L., Grobe, P., Quast, B. and Bartolomaeus, T. (2012), ‘Fiat or Bona Fide Boundary – A Matter of Granular Perspective’, PLoS One 7 (e48603). DOI:10.1371/ journal. pone.0048603. Wagner, G. P. (2014), Homology, Genes, and Evolutionary Innovation, University Press Group Ltd, Princeton, New Jersey. Williamson, T. (1994), Vagueness, Routledge, London. Wimsatt, W. C. (1972), ‘Complexity and Organization’, PSA: Proceedings of the Biennial Meeting of the Philosophy of Science Association, Springer. pp. 67–86. Wimsatt, W. (1976), ‘Reductionism, Levels of Organization, and the Mind-Body Problem’, in G. Globus, I. Savodnik and G. Maxwelll, (eds.), Consciousness and the Brain, Plenum Press, New York, NY, 199–267. Wimsatt, W. C. (1994). ‘The Ontology of Complex Systems: Levels of Organization, Perspectives, and Causal Thickets’, Canadian Journal of Philosophy, 20 (Suppl.), 207–74. Winther, R. G. (2011), ‘Part-Whole Science’, Synthese, 178, 397–427. Woodger, J. H. (1929), ‘Biological Principles: A Criticial Study’, K. Paul, Trench, Trubner and Co., London.

7

A Conceptualist View in the Metaphysics of Species Ciro De Florio and Aldo Frigerio

The species concept is one of the central concepts in biological science. Although modern systematics speculates about the existence of a complex hierarchy of nested taxa, biological species are considered particularly important for the active role they play in evolution. However, neither theoretical biologists nor philosophers of biology have come to an agreement about what a species is. In this chapter, we address two questions pertaining to biological species: (1) are they individuals or universals? and (2) are they bona fide or fiat entities? In section The Species-asIndividuals View, we illustrate the reasons that have led many scholars to support the view that species are individuals. In the next two sections, we show that the relational concepts of species – on which the species-as-individuals view is based – provide neither necessary nor sufficient conditions for species membership. This seriously undermines the species-as-individuals view. In the section A Conceptualist Model for the Metaphysics of Species, we advance the proposal that species are fiat concepts (and thus, universal entities partially dependent on the human mind) carved in a multi-dimensional space representing the properties that the biological organisms possess. The final section concludes.

The species-as-individuals view In the Aristotelian tradition, species are regarded as kinds, that is, as universal entities, predicated of individual substances that characterize their essence (for instance, the essence of being a dog). An individual substance (consider our cat Kurt as an example) instantiates a natural kind. The existence of the natural kind ‘cat’ depends on the existence of particular cats; however, the fact that Kurt

122

Natural and Artifactual Objects in Contemporary Metaphysics

lacks gills, for example, depends on his essence − his being an instance of the kind ‘cat’. In this view, the offspring of a couple of organisms belonging to a species can diverge from their ancestors only by contingent characters. Essential and defining characters must remain the same through generations. Therefore, a species is defined as the set of traits that are regarded as sufficient and necessary for being a member of that species. Darwin’s theory of evolution has provided a framework in which the Aristotelian concept of biological species seems odd for many reasons (Hull 1965a, b). In Darwin’s evolutionary view, in principle, every character of an organism can disappear in its offspring. Moreover, the distinction between defining and contingent characters disappears; every character can change, and none is necessarily constant through generations. When a character changes, it is difficult to decide whether it was a defining character of the species (so that the species evolved into another one) or whether it was contingent (so that the species remained the same). Every character is not stable in principle; therefore, the distinction between defining and contingent characters is not easy to trace. In the evolutionary framework, it is difficult to define a species as a set of necessary and sufficient traits because every trait can vary in an almost limitless way, and none seems to be essential. Therefore, some alternative definitions of biological species have been formulated, and these are not grounded in the characters possessed by singular organisms but rather in the relations among members of the same species. In particular, two kinds of relations have received attention: interbreeding relations and genealogical relations among organisms. Interbreeding relations are at the basis of the biological concept of species.1 According to this concept, a species is a group of individuals that can breed together and are reproductively isolated from other organisms.2 The gene flow among members of the same species and the lack of gene flow with members of other species contribute to preserving the genetic pool of the species as relatively constant by recombining the genes of deviant individuals with those of conspecific individuals and protecting the genetic pool from the introgression of genes belonging to individuals of other species. In this way, the most favourable combinations of genes are preserved, and the production of too many disharmonious, incompatible gene combinations is prevented. Another concept based not on the intrinsic properties of the members but on their relationships is the phylogenic concept. The basic idea is that an individual is a member of the species S only if it descends from another individual of S. However, we cannot speak of a unique phylogenetic concept of species but rather a family of concepts related to each other but having remarkable differences. In particular, a phylogenetic concept is incomplete unless it is not specified

A Conceptualist View in the Metaphysics of Species

123

when a speciation episode occurs, that is, when an individual descending from a member of S is no longer a member of S. If this is not specified, the definition has the consequence that every descendant of an individual belonging to S is a member of S, and thus, that speciation is impossible. In this framework, the arising of a new species can be indicated either by the reproductive isolation of a community of individuals from the other descending members3 or by the appearance of a new particular trait that is not present in the ancestors of an individual.4 These relational concepts of species favoured a new ontological view of species that are no longer regarded as abstract kinds instantiated by singular organisms but as complex individuals composed of singular organisms (Ghiselin 1966, Hull 1976, Sober 1980). As Ghiselin (1974) argues, species are similar to firms. To decide whether a person is employed at a certain firm, we do not have to observe his or her features but rather the relations between him or her and the other employees of that firm. Consequently, a firm is not a species instantiated by its employees – it is a complex individual whose employees are parts linked by some relation. Indeed, the unity of the firm is due to the relations among employees. The same can be said about species. In order to evaluate whether an individual belongs to a biological species, its features are merely an indication. What is decisive is the relations between the individual and the other members of the species. It is, therefore, natural to consider a species as a complex individual whose cohesion is determined by the relations among its members. Besides firms, biological species can be compared to other complex individuals whose unity is due to the relations among their members, such as musical bands, associations or families.5 To conceive of species as individuals has further advantages. According to the theory of evolution, biological species have a temporal beginning and a temporal end, evolve and change. Abstract entities (such as universals) are usually conceived as timeless and, thus, as entities that have no beginning and no end and cannot undergo changes and transformations. Many scholars have concluded that biological species cannot be kinds and that every attempt to deny this was anti-evolutionist and fixist because it would have implied the negation of the thesis that species evolve. A natural conclusion is that species are individuals because, unlike universals, individuals can have a temporal beginning, can have a temporal end and can change. Peter Simons sums up the question as follows: Modern evolutionary biology is based on populations of organisms. In this respect it is bottom–up or nominalistic, since populations are particular, not universal. It constitutes a definitive rejection of the platonistic conception

124

Natural and Artifactual Objects in Contemporary Metaphysics

of biological kinds as ideal archetypes initiated by Plato, retained in modified form by Aristotle, and perpetuated by scholasticism, Linnæus, Goethe, Cuvier, and Owen. The gradual acceptance of evolution following Lamarck and the subsequent explanation of its mechanism by Darwin and Wallace mean that biologists can no longer sustain a top–down or typological conception of what constitutes a species. As to what species are, even under a single conception, theorists differ. The common-sense idea is that they are certain pluralities or collections of organisms, delimited by their interrelations. (Simons 2013, p. 278)

It is interesting to note that even a friend of natural kinds, Jonathan Lowe, distinguishes the notion of biological species – which he believes to be a complex and collective individual – from that of natural kind, which is a metaphysical category having to do with the structure of reality: But perhaps we need to make a distinction, which can best be brought out by analogy with a related case, that of biological species. These too are said to come into existence and undergo change – indeed, that they do so is crucial to the theory of evolution. How then can species names denote universals, which are abstract entities and so on the present proposal timeless? The solution is to distinguish between biological species, which are concrete particulars consisting at any time of the mereological sum of their currently existing members (individual tigers or individual oaks), and biological sorts or kinds, which are universals instantiated by the members of those species. (Lowe 1998, p. 75)

To summarize, the naive interpretation of the concept of biological species as natural kind has been substantially revised in light of the theory of evolution. However, we will see in the next two sections that the relational concepts of species fail in many respects, and this seriously undermines the species-asindividuals view.

The biological concept provides neither necessary nor sufficient conditions In this and the next sections, we show that relational criteria provide neither necessary nor sufficient conditions for being a member of a species. Consequently, any biological taxonomy must, at least in part, turn to the similarity among organisms in their classification. Thus, one of the main arguments for regarding species as complex individuals fails. If relations are neither necessary nor sufficient to ascribe an individual to a species and we need to rely on similarity among organisms, the analogy between biological species and complex

A Conceptualist View in the Metaphysics of Species

125

individuals such as firms or musical bands must be dismissed. Let us consider the biological concept first. We will show that it provides neither necessary nor sufficient conditions for being a member of a species.

Not necessary Critics of the biological concept often highlight that it can be applied only to sexually reproducing organisms (Vrba 1985, Templeton 1989). This is a severe restriction, because most organisms on Earth are asexual but seem to display the same patterns of phenotypic cohesion within and discontinuity between species as do sexual species. Taxonomic practice has always subdivided the asexual world into easily defined biological taxa, as it has always done with the sexual world. Advocates of the biological concept can react in three ways to this restriction: they can claim that asexual organisms are members of no species (Mayr 1987); they can affirm that every asexual organism is the unique member of its own species since it is reproductively isolated from every other organism (Ghiselin 1974) or they can hold that a concept of species different from the biological concept should be applied to asexual organisms (Ereshefsky 2001). The first two solutions are hopeless in light of the fact that asexual organisms can be classified in species as easily as sexual organisms. Thus, the only way out seems to be maintaining that interbreeding relations are a necessary condition for being a member of a species for just a portion of the biological world: the sexual one. However, this solution does not solve all the problems of the biological concept. Ehrlich and Raven (1969) argue that many conspecific sexual populations have little or no gene flow between them, particularly in plants, due to the physical distance between them. Moreover, ‘selection can override the effects of gene flow’ (p. 64) so that the cohesive force binding species together is for the most part ‘similar selective regimes’ (p. 63). The biological concept would overrate the importance of the interbreeding relations as a mechanism for preserving the unity of the species, which, on the contrary, would be largely caused by environmental factors. Ehrlich and Raven’s thesis is questioned by Ridley (1993), who states that it is not clear if selection is more important than gene flow; the two are not necessarily mutually exclusive, and, depending on the situation, one might be more important than the other. He concludes that ‘the evidence … hardly warrants any firm conclusion’ (p. 399) but that ‘selection and genetic flow are probably not usually opposed forces in nature’ (p. 397). Be that as it may, genetic flow is not a unique cohesive mechanism in sexual species, and perhaps it is not even the most important.

126

Natural and Artifactual Objects in Contemporary Metaphysics

There is another, more serious, problem with the biological concept – the members of a species S that cannot mate with the other members of S because of malformation or malfunctioning of the reproductive system, or sterility. If potential interbreeding were a necessary condition for being a member of a species, these organisms should be judged as non-members of the species S. This is an unacceptable conclusion. Therefore, potential interbreeding among conspecific members is not a necessary condition for being a member of a species, even if the species is a sexual one. Could advocates of the biological concept appeal, as they often do,6 to particular isolating mechanisms to solve this problem? These mechanisms are of several kinds and are usually divided in two classes, ‘namely mechanisms preventing the appearance of hybrid offspring (incompatibility of the parental forms), and mechanisms making the hybrids sterile and, consequently, incapable of propagating further (hybrid sterility)’ (Dobzhansky 1935, p. 350). One could maintain that the members of S that are reproductively isolated from the other members of S are isolated for mechanisms that are not typical of S. In other words, the isolated mechanisms that characterize S are not those mechanisms that isolate the members of S that cannot mate and reproduce with other members of S. However, since the malformed and sterile members of S share the particular isolating mechanisms that reproductively isolate the species S from other species, they can be considered members of S. Nevertheless, this answer is very risky for advocates of the biological concept. It seems to imply that being a member of S means to possess a particular isolating mechanism (e.g. particular forms of genitalia that prevent mating with organisms of other species, a particular structure of chromosomes that prevent meiosis with chromosomes of different species). These isolating mechanisms are structural properties of organisms, not relational properties. Then, species membership would be determined by the possession of certain structural properties. The biological concept would not be a relational concept but would be grounded in the possession of some intrinsic properties of the members of a species. This would be against the intents of the advocates of the biological concept, who have always claimed that relationships – and not intrinsic properties – determine the membership to a species.

Not sufficient Many species, especially among plants,7 are not reproductively isolated but can cross-breed with other species, giving rise to fertile and viable hybrids. An oftenmentioned case is that of oaks, a genus with more than 400 species that hybridize

A Conceptualist View in the Metaphysics of Species

127

each other, creating new, fertile species. The new species can then hybridize with the parent species but remain distinct from them for long periods of time because of geographic barriers and a low rate of hybridization among the species (Burger 1975, Van Valen 1976). The frequency of the hybridization phenomena among different species shows that potential interbreeding is not a sufficient condition for being of the same species. Another phenomenon that leads to the same conclusion is that interbreeding relationships are not transitive; population A can easily cross-breed with population B, and population B can interbreed with population C, but this does not mean that A can cross-breed with C. A particular case of this phenomenon is that of the ring species. Ring species are continuously connected chains of races of a species that surround a geographic region than cannot be occupied because it is a mountain massif, ocean or polar region. Therefore, they show a ring-shaped distribution around the inhospitable region. At the endpoints of this geographic ring, the most distant populations encounter each other secondarily. Here, they are no longer genetically compatible and are unable to cross-breed. Although they still belong to the same species, they have become genetically different because their common ancestor dates back far into the past and the mutual exchange of alleles has become scarce or has ceased entirely. A good example of a ring species is the greenish warbler (Phylloscopus trochiloides), a passerine that spread from the south of the Tibetan Plateau to both the western and eastern sides of the plateau. Today, the Tibetan Plateau is surrounded by several races of greenish warblers that are all reproductively connected to each other. On the northern side of the Tibetan Plateau, however, the races from the west and the east encounter each other; here, they are mutually genetically incompatible. Between the two races in the contact zone, gene exchanges no longer occur (Irwin, Bensch and Price 2001). In many cases of ring species, the different races or sub-species end up being considered different species that can cross-breed with the species that are geographically close but not with those that are geographically distant. Again, potential cross-breeding does not imply being of the same species.

The phylogenic concepts provide neither necessary nor sufficient conditions In this section, we will argue that phylogenic concepts provide neither necessary nor sufficient conditions for being a member of a species.

128

Natural and Artifactual Objects in Contemporary Metaphysics

Not sufficient The proponents of phylogenetic concepts admit that these concepts do not provide sufficient conditions for species membership. The reason for this can be easily grasped − all living organisms likely descend from a unique organism or from a few organisms that lived some billions of years ago, but they obviously do not belong to the same species. Thus, in some cases, some organisms that descend from a member of the species S are not members of S. This happens when a speciation event occurs. Therefore, a phylogenetic concept is incomplete unless it is not specified when a speciation episode occurs, that is, when an individual descending from a member of S is no longer a member of S. If this is not specified, the definition has the consequence that every descendant of an individual belonging to S is a member of S, and thus, that speciation is impossible. Phylogenetic lineage is not a sufficient criterion for species membership. The issue is that the conditions that the advocates of the phylogenetic concepts add require that one either belongs to the same interbreeding community (which, as we have seen, is not a sufficient criterion) or possesses some phenotypic or genotypic characteristics. If this second route is followed, the criterion for species membership is not purely relational anymore; rather, it is based also on some intrinsic properties, that is, on the resemblance among members of the species.

Not necessary The proponents of the phylogenetic concepts insist on the necessity of belonging to the same lineage for being a member of a species. Hence, Hull (1978) offered a thesis in which an extinct species cannot reappear again. When an individual has ceased to exist, it cannot begin to exist again: ‘If a species evolved which was identical to a species of pterodactyl save origin, it would still be a new, distinct species’ (p. 349). Since the two lineages have two different origins, they must be considered two different species. Modern taxonomic practices tend to follow the origin criterion for organizing the taxa to which biological organisms belong. It often happens that a certain taxon is moved from a superordinate taxon to another one in order to obtain only monophyletic taxa.8 However, the thesis that phylogenetic concepts provide necessary conditions can be put into question. Consider the following thought experiment. Suppose that on a distant planet, environmental conditions very similar to Earth’s allow for the evolution of organisms that are genetically and phenotypically

A Conceptualist View in the Metaphysics of Species

129

indistinguishable from terrestrial lions. Moreover, suppose that some of these animals are brought to Earth and begin to cross-breed with terrestrial lions. At a certain point, one cannot know whether an organism comes from the Twin Earth planet, from Earth or from a hybridization of the two races.9 According to the phylogenic concepts, terrestrial lions and twin lions are two different species, and the organisms descending from cross-breeds are hybrids of the two species. Moreover, the hybrids that cross-breed with the parent species give rise to new species. We would have several species without any possibility (genetic tests, morphological measurements, anatomical and physiological studies) to assign any individual to the species that are supposed to exist. This is an odd result. Let us consider another thought experiment. Suppose that progresses in genetic engineering allow for the production of artificial life. Moreover, suppose that laboratory A produces a new life form starting from a certain genetic structure. Laboratory B buys the patent from laboratory A and produces organisms starting from the same genetic structure. Again, these two kinds of organisms are not distinguishable, both from genotypic and phenotypic points of view. Suppose that several laboratories produce this kind of organism and that they are scattered over environments. If sexual, they can freely interbreed. Again, according to the philological concepts, the organisms built in different laboratories are different species because they have different origins and belong to different lineages. When they cross-breed, they give rise to hybrids. But again, there is no possibility to distinguish the individuals when they are scattered in the environment. There is no test for establishing whether an organism originates from laboratory A rather than laboratory B, whether it is a hybrid between the lineages of laboratory A and laboratory B, or whether it is a hybrid between an organism of this kind and an organism coming from laboratory A and so forth. Thought experiments have never been very popular among philosophers of biology. Perhaps the reason for this is that biology is a science more empirical than physics, where thought experiments are very common.10 However, it is not possible to dismiss arguments of this kind simply on the ground that thought experiments are irrelevant in biology because there are actually very similar cases. Furthermore, these cases are very numerous. Many species hybridize and give rise to other species. There could be many hybridization events between two species A and B that are separate in time and space. They can give rise to several lineages of a species C that can also be separate in time and space. Some of these lineages can go extinct. Others can evolve in a parallel and independent way if they are subject to similar selective

130

Natural and Artifactual Objects in Contemporary Metaphysics

regimes. However, it is difficult to deny that these different lineages belong to the same species, although they have different origins. Furthermore, a hybridization episode between A and B can give rise to a lineage of the species C, which can go extinct. Afterward, another hybridization episode can give rise to another lineage of C. Thus, C is a species that went extinct and then began to exist again. Similar situations are probably very common among allopolyploid plants11 (i.e. plants in which homologous chromosomes are more than two and descend from different species), given the high number of allopolyploid species among angiosperms. More rarely, two or more different and isolated populations of a species A, if subject to very similar selection regimes and to very similar environmental conditions, can separately evolve into a species B in parallel ways. In this case, different lineages of B have different origins. An actual example is the threespine stickleback. These fish live in different lakes in British Columbia and have followed very similar but independent evolutionary paths, which have led to three parallel speciation events in three different lakes of the area (Rundle 2000).12 In addition to these biological reasons, there is another flaw in the proposed criteria. If species membership can be defined exclusively by the relations it bears with other individuals, it is conceptually impossible to determine the species of a hypothetical organism that is the only one of its species. Let us imagine that Sgrunt is the unique specimen of a kind of living being, a sort of monstrum. Now, Sgrunt does not exist, and very probably, mono-species organisms have never lived. Moreover, according to working biologists, it is simply meaningless to affirm that there exists a species with just one example. However, from the metaphysical and conceptual point of view, it would seem possible to ascribe a species to Sgrunt. In the end, it has a series of features that characterize it. But the definitional criteria of the species-as-individuals approach block a priori the possibility to collocate Sgrunt into a species, and this is suspicious. If an organism cannot belong to a species solely by virtue of its relations with other individuals, it does belong to a species because of certain intrinsic features. But this is precisely what the species-as-kinds advocates state – it is because an organism has certain properties (i.e. it is an instance of a certain kind) that it is a member of a species. The insufficiency of relational criteria undermines the basis of the conception of species as complex individuals, since it pushes towards the view that living beings are collected in species not because of their relationships but by virtue of having certain features.

A Conceptualist View in the Metaphysics of Species

131

A conceptualist model for the metaphysics of species In a series of articles (2000, 2001), Barry Smith introduced an important distinction between, roughly, fiat and bona fide entities. In particular, he distinguishes between fiat boundaries, which are the outcome of an individual or collective intentional act, and bona fide boundaries, which are (in a sense) grounded in the reality of things. There is no doubt that political boundaries (for instance, those of some states of the United States) are fiat, while the boundaries of an object like a table seem to be bona fide. Behind this distinction, there is the old-as-philosophy topic of carving the nature at joints, that is, the question of realism. Achille Varzi (2001), for instance, maintained a strong conventionalist position: there are no bona fide boundaries, and thus, there are no bona fide objects. This debate is also relevant for our analysis. Barry Smith (2001, p. 34) states that biological species are a typical example of fiat objects. We believe that two different philosophical theses are condensed in this view: the first has a metaphysical character and concerns the nature of the entities under scrutiny; the second has an ontological character and concerns the existence and modality of existence of the entities in question. Therefore, according to the first subclaim, biological species are individuals, that is, spatiotemporally collocated items. Moreover, the species are scattered, not connected, individuals. The second subclaim – the proper ‘fiat’ part – has to do with the ontological dependence of species. If they are fiat entities, they strongly depend on cognitive human acts. In a nutshell, a world with no intentionality is a world with no biological species. Now, if the arguments we advanced above are correct, Smith’s thesis in which biological species are fiat objects cannot be entirely endorsed. We argued that there are good reasons to believe that species are not individual entities.13 However, the problems that stem from evolutionary theory and cast doubt on the alleged identification of biological species with natural kinds still persist. We arrive at a sort of stalemate. On one hand, there are good reasons not to consider species as individuals. On the other, even if universals, species seem to show no necessary and fixed features as the natural kinds. In the following, we will sketch a solution that tries to unify both a realist intuition (departing from Varzi–Smith’s conventionalism) and the data from biology. We will assume that in the world there exist at least some biological properties that are bona fide entities; a nominalist would disagree, of course. But it is meaningful, from a philosophical point of view, that the very biologists advancing the species-as-individuals view admit the existence of biological

132

Natural and Artifactual Objects in Contemporary Metaphysics

properties and relations.14 Any living organism exemplifies certain clusters of properties, intended as bona fide entities. Now, an advocate of the speciesas-kind view could identify some privileged clusters with natural kinds. This manoeuver bumps into the well-known problems we previously discussed. Our idea is to exploit the distinction between fiat and bona fide by applying it not to individuals or properties but to concepts. Smith (2001, pp. 141–2) discusses fiat concepts. We wish to extend his considerations to the metaphysics of species. We assume that concepts are epistemic entities (they are what we understand), and in this sense, they are dependent on the human mind. Classically, concepts can be individual (for instance, the concept associated with the proper name Kurt), universal (for instance, concepts expressed by predicates such as ‘is white’ or ‘is a cat’) and, finally, logical (as those expressed by particles such as ‘or’, ‘and’ or ‘not’). The ontological tie between universal concepts and individuals is not instantiation but falling under. One could state that the reason Socrates falls under the concept of being wise is that Socrates instantiates the property (if any) of wisdom. Universal concepts are the clippings by which the world can be categorized. In Smith’s words, Suppose that each concept is associated with some extended region in which its actual and possible instances are contained, and suppose further that this is done in such a fashion that the prototypes, the most typical instances, are located at distances in the center in proportion to their degree of non-typicality. (Smith 2001, p. 141)

Now, Smith’s view is that concepts are typically fiat entities, since they depend on the intentional acts of the subjects. This is true, but it grasps just one side of our proposal. What we want to maintain is that biological species are universal concepts and that they are grounded in the biological nature of the organisms. A very simplified example follows: let us imagine that there are just three biological properties (P, Q, R) that come up as graduated (they are determinable properties). For convenience’s sake, we will indicate the measure of the property under consideration with a number in the interval 1–1,000. Moreover, let us imagine that in our domain, we have just three individuals (a, b, c) that exemplify these properties and we have to classify them in species. The first scenario we consider is the following: a b c

: : :

P235 P233 P236

Q28 Q28 Q27

R177 R176 R177

A Conceptualist View in the Metaphysics of Species

133

Given this scenario, it is wholly plausible to consider the three individuals as belonging to the same species.15 Thus, we will say that a, b and c fall under species S because it is a fact of the matter that they instantiate a significant cluster of biological properties. Unfortunately for working biologists, things are very rarely so easy. Let us consider another scenario: a b c

: : :

P670 P668 P31

Q750 Q751 Q750

R800 R20 R802

What happens in this case? Clearly, based on these data, a, b and c cannot be subsumed under the same concept of species. Indeed, we have that a and c are different concerning property P. a and b are different concerning property R. b and c are different concerning properties P and R. Now, we can notice the typical fiat feature of the concept of species – it is a biologist’s choice to decide what the most relevant properties are. They could, for instance, maintain that property P is, above all, less important and then consider a and c as belonging to the same species. Alternatively, they can reason the same way about R. Another possibility is to focus on property Q and then declare that a and b and c belong to the same species, their differences notwithstanding. There are so many available alternatives open to biologists to categorize this scenario. However – and this is the trait that differentiates the world of living beings from, say, the world of geography – not all the divisions are possible. In this case, it is quite unreasonable to consider b and c to belong to a species (say, S) and a not to belong to S. And this is because it is a fact of the matter that some clusters of properties are instantiated and others are not. It is plausible to think that in cases such as the political boundaries between Wyoming and Colorado, there are no choices that must be disregarded in principle. The toy model we have provided can easily be extended to a much higher number of properties. In particular, we can represent any determinable property as a dimension. So we get an n-dimensional16 space, where n indicates the number of properties taken into account. In our framework, if we model P, Q and R as dimensions, we obtain a tridimensional space in which we can place the individuals. Each living being instantiates the properties P, Q and R to a certain degree; this means that in

134

Natural and Artifactual Objects in Contemporary Metaphysics

the model, any organism occupies a certain position in the space. Very close organisms constitute clouds in the space. It is now clear why the concepts are seen as clippings – it is natural to consider close individuals as belonging to the same species, while it is odd to make partitions of the space that includes nonhomogeneous regions, empty spaces or even disconnected areas. Nevertheless, there are doubtful cases. For example, some individuals can be far from the centre of a cloud; they are borderline cases. In this scenario, biologists have to decide whether to cut the space with a concept of species that includes the case under scrutiny or consider it as belonging to another species. Then, although they are grounded in re, the concepts of species are vague. Let us see how our framework accounts for the evolutionary phenomena. From a merely ontological point of view, what mutates from generation to generation is the exemplification of some biological properties (think about determinable properties as, for example, ‘having an olfactory system that is sensitive to a certain kind of pollen’). With the passage of time, the conceptual clipboard, that is, the species Felis Catus, no longer encompasses the majority of the descendants that now instantiate different properties. Therefore, it is natural to select the cluster of individuals in a different way. The origin of a new species is precisely the acknowledgement, on a conceptual level, of a new clipping in the space of biological properties and individuals. Analogously, extinct species, such as Tyrannosaurus Rex, are concepts under which no living organisms now fall. They are classifications of properties with no instances. But, if in a Jurassic Park laboratory, a scientist created animals ‘in vitro’ that exemplify the cluster of properties subsumed under the species Tyrannosaurus Rex, these animals are really members of the species in question. A similar consideration could be made for the hypothetical case of alien species.

Conclusion In this chapter, we discussed the metaphysical interpretation of the concept of biological species. One of the reasons to adopt the species-as-individuals view is the preference for relational species criteria over structural ones. However, we have shown that relational criteria provide neither necessary nor sufficient conditions for species membership. In order to group biological individuals in taxa, we need to use structural properties and the similarity among individuals.

A Conceptualist View in the Metaphysics of Species

135

Thus, one of the main arguments in favour of the species-as-individuals view fails. Then, we provided a framework to consider biological species as concepts instead of individuals or natural kinds. These concepts show both fiat and bona fide features. On one hand, they depend on the real instantiation of biological properties by living beings. However, on the other hand, they reflect the decisions made by biologists when evaluating the importance of properties or considering borderline cases. We argued that this model can overcome some difficulties that affect the previous views.

Notes 1 In Chapter 6 of this volume, Vogt considers biological entities as ‘causal agents in various biological processes’. Among them, reproduction and evolution are important in defining their identity. 2 For the modern biological concept of species, see Dobzhansky (1935), Dobzhansky (1937), Mayr (1942), Mayr (1949) and Mayr (1970). 3 Therefore, this view mixes elements of the phylogenetic concept of species with elements of the biological concept. Compare Henning (1966). 4 These phylogenetic concepts, which are grounded on traits that are not present in the ancestors of a population, mix elements of the phylogenetic concept of species with elements of the classical morphological concept as a set of individuals that instantiate some properties. Compare Rosen (1978), Eldredge and Cracraft (1980) and Nelson and Plamick (1981). 5 It is important to note that the biological and phylogenetic concepts of species do not have the species-as-individuals view as a necessary consequence. In fact, the species-as-kinds views that hold relations among members as a criterion of membership are also possible (cf. see Okasha 2002 and LaPorte 2004). However, it should be noted that, if the membership criterion is based on relations among members, it is much more natural to consider species as complex individuals whose cohesion is a result of the relations among the organisms that are parts of the species (for this thesis, cf. Crane 2004). 6 Mayr (1991) defines a species as ‘a reproductively isolated aggregate of populations which can interbreed with one another because they share the same isolating mechanisms’ (p. 186). 7 Templeton (1989) affirms that hybridization is not rare even among animals, as his study on mammals would show. 8 Some proposals aim even at questioning Linneaus’s hierarchy and to substitute it with a classification in which every taxon is monophyletic (cf. de Queiroz and Gauthier 1992, Ereshefsky 2001).

136

Natural and Artifactual Objects in Contemporary Metaphysics

9 As it often happens with biological species, one can draw an analogy between this example and an example concerning languages. Suppose that on a very faraway planet a population speaks a language indistinguishable from English (same phonology, morphology, lexicon, grammar, etc.). Suppose that terrestrial English and the English of this population are not genetically linked, but evolved independently. Are terrestrial English and the twin English the same language or two different languages? Suppose, for example, that this population moves on Earth and mixes with English speakers. Is it possible to hold that the two populations speak two different languages similar in every aspect? 10 For a criticism of the notion of thought in biology and philosophy of biology, compare Hull (1989). 11 For an argument along these lines, which uses allopolyploid species, compare Stamos (2003, pp. 237–9). 12 Vogt (see Chapter 6 of this volume) maintains that common historical origin can serve as a principle of connectedness that allows the delineation of bona fide objects. However, if the criticisms advanced here are corrected, species cannot be considered bona fide individuals on the basis of a common historical origin. Another reason why species might be seen as bona fide individuals according to Vogt’s criteria is that they are often considered as the units of evolution, what might be described as ‘the engines of evolution’ (Eldredge and Cracraft 1980; Wiley 1981; Mayr 1982; Ghiselin 1987). Thus, their unity might be causal or functional. However, that species are evolutionary units is a controversial claim. Molecules, genes, cells, organisms, populations have been proposed as possible units of evolutions (cf. Brandon and Burian 1986 for discussions on this topic) and it is not clear which are the levels at which selection acts. For a criticism of the idea that species are evolutionary units, compare Ereshefsky (1991). 13 Obviously, there is a sense in which we can non-ambiguously refer to a biological species by meaning the individuals of that species collectively considered. But the problem is metaphysical − does that species coincide with the conglomeration of individuals? 14 Tuomas Tahko (2012) argues contra Varzi and Smith that fundamental natural laws suggest the existence of bona fide entities. Tahko’s examples have to do with quantum physics, and in particular, with the properties of fundamental particles. It is an open question if the same considerations can also hold for levels of reality as complex as those of living beings. 15 In our example, the differences among individuals are graduated using an interval 1–1,000. There is no reason to prefer this level of distinction to one higher (or lower). In practice, this choice also crucially depends on the fine-tuning of the instruments. 16 Compare Gardenfors (2000).

A Conceptualist View in the Metaphysics of Species

137

References Brandon, R. N. and Burian, R. (eds) (1986), Genes, Organisms, Populations. Controversies over the Units of Selection, MIT Press, Cambridge, MA. Burger, W. (1975). ‘The Species Concept in Quercus’, Taxon, 24: 45–50. Crane, J. (2004), ‘On the Metaphysics of Species’, Philosophy of Science, 71: 156–73. Dobzhansky, T. (1935), ‘A Critique of the Species Concept in Biology’, Philosophy of Science, 2 (3): 344–55. Dobzhansky, T. (1937), Genetics and the Origin of Species, Columbia University Press, New York. Ehrlich, P. and Raven, P. (1969), ‘Differentiation of Populations’, Science, 165: 1228–32. Eldredge, N. and Cracraft, J. (1980), Phylogenetic Patterns and the Evolutionary Process, Columbia University Press, New York. Ereshefsky, M. (1991), ‘Species, Higher Taxa, and the Units of Evolution’, Philosophy of Science, 58 (1): 84–101. Ereshefsky, M. (2001), The Poverty of the Linnaean Hierarchy: A Philosophical Study of Biological Taxonomy, Cambridge University Press, Cambridge. De Queiroz, K. and Gauthier, J. (1992), ‘Phylogenetic Taxonomy’, Annual Review of Ecology and Systematics (23): 449–80. Gärdenfors, P. (2000), Conceptual Spaces: The Geometry of Thought, MIT Press, Cambridge, MA. Ghiselin, M. (1966), ‘On Psychologism in the Logic of Taxonomic Controversies’, Systematic Zoology, 15: 207–15. Ghiselin, M. (1974), ‘A Radical Solution to the Species Problem’, Systematic Zoology, 23 (4): 536–44. Ghiselin, M. (1987), ‘Species Concepts, Individuality, and Objectivity’, Biology & Philosophy, 2: 127–43. Henning, W. (1966), Phylogenetic Systematics, University of Chicago Press, Chicago. Hull, D. (1965a), ‘The Effect of Essentialism on Taxonomy-Two Thousand Years of Stasis (I)’, The British Journal for the Philosophy of Science, 15 (60): 314–26. Hull, D. (1965b), ‘The Effect of Essentialism on Taxonomy-Two Thousand Years of Stasis (II)’, The British Journal for the Philosophy of Science, 16 (61): 1–18. Hull, D. (1976), ‘Are Species Really Individuals?’, Systematic Zoology, 25 (2): 174–91. Hull, D. (1978), ‘A Matter of Individuality’, Philosophy of Science, 45 (3): 335–60. Hull, D. (1989), ‘A Function for Actual Examples in Philosophy of Science’, in M. Ruse (ed.) What the Philosophy of Biology Is, Kluwer, Dordrecht, 309–21. Irwin, D., Bensch, S. and Price, T. (2001), ‘Speciation in a Ring’, Nature, 409: 333–7. LaPorte, J. (2004), Natural Kinds and Conceptual Change, Cambridge University Press, Cambridge. Lowe, E. J. (1998), The Possibility of Metaphysics: Substance, Identity, and Time, Oxford University Press, Oxford.

138

Natural and Artifactual Objects in Contemporary Metaphysics

Mayr, E. (1942), Systematics and the Origin of Species, Columbia University Press, New York. Mayr, E. (1949), ‘The Species Concept: Semantics Versus Semantics’, Evolution, 3: 371–2. Mayr, E. (1970), Populations, Species, and Evolution, Harvard University Press, Cambridge. Mayr, E. (1982), The Growth of Biological Thought, Harvard University Press, Cambridge. Mayr, E. (1987), ‘On the Ontological Status of Species: Scientific Progress and Philosophical Terminology’, Biology and Philosophy, 2: 145–66. Mayr, E. (1991), One Long Argument, Harvard University Press, Cambridge. Nelson, G. and Plamick, N. (1981), Systematics and Biogeography: Cladistics and Vicariance, Columbia University Press, New York. Okasha, S. (2002), ‘Darwinian Metaphysics: Species and the Question of Essentialism’, Synthese, 131: 191–213. Ridley, M. (1993), Evolution, Blackwell Scientific Publications, Boston. Rosen, D. (1978), ‘Vicariant Patterns and Historical Explanation in Biogeography’, Systematic Zoology, 27: 159–88. Rundle, H. et al. (2000), ‘Natural Selection and Parallel Speciation in Sympatric Sticklebacks’, Science, 287: 306–8. Simons, P. (2013), ‘Vague Kinds and Biological Nominalism’, Metaphysica, 14: 275–82. Smith, B. (2001), ‘Fiat Objects’, Topoi, 20 (2): 131–48. Smith, B. and Varzi, A. C. (2000), ‘Fiat and Bona Fide Boundaries’, Philosophy and Phenomenological Research, 60: 401–20. Sober, E. (1980), ‘Evolution, Population Thinking, and Essentialism’, Philosophy of Science, 47 (3): 350–83. Stamos, D. (2003), The Species Problem, Lexington Books, Oxford. Tahko, T. E. (2012), ‘Boundaries in Reality’, Ratio, 25 (4): 405–24. Templeton, A. (1989), ‘The Meaning of Species and Speciation. A Genetic Perspective’, in D. Otte and J. Endler (eds), Speciation and its Consequences, Sinauer associates, Sunderland. Vrba, E. (1985), ‘Introductory Comments on Species and Speciation’, in E. Vrba, (ed.), Species and Speciation, Transvaal Museum, Pretoria, ix–xviii Van Valen, L. (1976), ‘Ecological Species, Multispecies, and Oaks’, Taxon, 25: 233–9. Varzi, A. (2001), ‘Vagueness in Geography’, Philosophy and Geography, 4: 49–65. Wiley, E. (1981), Phylogenetics: The Theory and Practice of Phylogenetic Systematics, Wiley and Sons, New York.

Part Three

Where Do Tools Come From?

8

Artifacts and Fiat Objects: Two Families Apart? Massimiliano Carrara and Marzia Soavi

Introduction Fiat objects may come into existence by intentional explicit definition and convention or they can be the result of some spontaneous and unintentional activity resulting in tracing fiat spatial boundaries. Artifacts and fiat objects seem intuitively to be correlated: both artifacts and fiat objects depend for their existence on agents and their intentions. Is it possible to consider fiat objects as artifacts and to what extent? Or else can we conceive at least some artifacts as fiat objects? In order to draw a map of the possible answers to these two questions we will take into account various definitions of artifacts stemming from the two classical approaches: the intentional and the functional one. We start our analysis by considering the following two questions: Question (1): Is there a relation between being an artifact and being a fiat object or are they two completely independent notions? In this first section of the chapter we adopt a sort of top–down approach; that is, we confront the two definitions associated with the general expressions ‘artifact’ and ‘fiat object’. We may cast the second question in the following form: Question (2): Does being an artifact imply being a fiat object? And, conversely, does being a fiat object imply being an artifact? Clearly, the bi-conditional expressed in question (2) is false in one direction: being an artifact does not imply being a fiat object. The other direction is instead much more interesting: in fact, it is debatable if being a fiat object implies being an artifact. We argue that the notion of ‘intentional production’ plays a crucial role in answering the question.

142

Natural and Artifactual Objects in Contemporary Metaphysics

In the second part of the chapter we consider a bottom–up approach: instead of considering the general definition of artifacts and fiat objects we will examine the possibility of a functional characterization of kinds of fiat objects in line with that of artifact kinds, hence the individuation of fiat objects through functional criteria. By way of conclusion one could argue that even if artifacts and fiat objects are not strongly related as ontological categories, they are not two families apart: both kinds for fiat objects and artifact kinds can belong to the family of functional kinds.

Artifacts and fiat objects: A preliminary characterization In this first section we start from a top–down approach, that is, we start by comparing the definitions of ‘artifact’ and ‘fiat object’. Where artifacts are concerned, we can isolate two key elements in the most discussed definitions in the philosophical literature: intentions and purpose or function. For example, Baker writes, ‘Artifacts are objects intentionally made to serve a given purpose’ (Baker 2007, p. 49; emphasis in the original).

Hilpinen writes, An artifact may be defined as an object that has been intentionally made for some purpose. (Hilpinen 2008, p. 1)

But what do we mean exactly by stating that artifacts are intentional products? On the one hand intentionality marks the intuitive − but as we will see misleading − distinction between artifacts and natural objects; on the other hand being an intentional product marks the difference between objects that are merely results of some intentional human action and objects that are the intended product of an action. To exemplify, pollution, crumbs, garbage and the greenhouse effect are all products of human intentional actions but for sure they are not the intended products of our actions. This is a crucial distinction in the literature concerning human action. Following a suggestion by Dipert (1993) we can label objects that are merely the result of human action: ‘artificial objects’. Thomasson (2003, p. 600) gives particular importance to this intentional character of the artifacts as products. In her perspective conditions for belonging to artifact kinds are the following:

Artifacts and Fiat Objects

143

Necessarily, for all x and all artifactual kinds K, x is a K only if x is the product of a largely successful intention that (Kx), where one intends (Kx) only if one has a substantive concept of the nature of Ks that largely matches that of some group of prior makers of Ks (if there are any) and intends to realize that concept by imposing K-relevant features on the object. (Thomasson 2003, p. 600)

Casting aside the complexity of this definition, we can briefly say that a spoon is a spoon only if its author intended to produce a spoon, and was largely successful in producing it, and this is, in our view, exactly the condition for an object to be considered an intended product. If someone carves a spoon out of a piece of wood he will for sure leave on the working table a lot of scrap material, but this material is not to be considered an artifact but is something artificial, something that exists in consequence of human action. Indeed, it seems plausible to say that our intuition regarding ‘artificial’ includes under this label not only objects that are the products of human action but products that from a certain perspective we can consider as objects that would not exist without the intervention of human action, like scrap, the ozone layer hole, rise in sea levels and so forth. It is plausible to think that there is no clear-cut distinction between artifacts and artificial objects, nor between artificial and natural objects. This type of remark concerning the requirement of intentional production for artifacts will be central for evaluating fiat objects that don’t seem to be produced by the fiat of intentional decisions. The second component of the definition, that of purpose or function, doesn’t play a fundamental role at the level of the general definition of the category of artifacts but, as we will see, it is central to the formulation of the conditions for being an artifact of a certain kind. So we will examine it in detail in the second part of the chapter, where we will adopt the bottom–up approach. Consider now the term fiat object. In the following passage Smith and Varzi give a short and precise definition for fiat objects: Once fiat boundaries have been recognized, it becomes clear that the bona fide–fiat opposition can be drawn not merely in relation to boundaries but in relation to objects also. Examples of bona fide objects are: John, the moon, a lump of cheese. Examples of fiat objects are: Dade County, the State of Wyoming, the North Sea. ... Broadly, it is the drawing of fiat outer boundaries in the spatial realm which yields fiat objects. (Varzi, Smith 2000, p. 403; emphasis mine)

Roughly, the idea is that if there are discontinuities in the material world corresponding to the external boundaries of an object, then the object is

144

Natural and Artifactual Objects in Contemporary Metaphysics

bona fide, while if there are no such material discontinuities the boundaries of the object are fiat boundaries and the object is a fiat one. The boundaries of bona fide objects depend crucially on topological features of the material world (and if they are artifacts also on the intention of some agents), while the boundaries of fiat objects depend exclusively on the intentions of the agents but they are not arbitrarily drawn. In the most prototypical cases, fiat boundaries and hence fiat objects are created for very good reasons that can be of economic, political, social and cultural nature. In the words of Varzi and Smith, Moreover, there are normally perfectly good reasons – reasons of topography, economy, or military strategy – why these and those fiat objects are created rather than others. Fiat objects thus owe their existence not exclusively to human fiat: real properties of the underlying factual material are involved also. (Varzi and Smith 2000, p. 403; emphasis mine)

Indeed, we are perfectly aware that the burden of the distinction is completely charged on the notion of physical discontinuity. There are many problems about strictly defining what a physical discontinuity is supposed to be, but in this work we take for granted that such a distinction can be traced in a relevant way. On this argument many interesting observations are made by Vogt.1 Casting aside the difficulties on clearly distinguishing fiat and bona fide objects we can say that in this top–down inquiry the answer to the first question, whether being an artifact implies being a fiat object, is obviously negative: there are a lot of artifacts that do not have fiat boundaries and so a lot of artifacts that are not fiat objects. The answer to the second one, whether being a fiat object implies being an artifact, is less straightforward and involves crucially the requirement of intentional production.

Comparing artifacts and fiat objects We can be driven by our intuitions to believe that fiat objects are indeed artifacts basically because human intentionality is essentially involved in their coming into existence. Despite this belief the way we understand ‘production’ in the definition of artifacts plays a central role in two different respects: 1. To the extent that we have reasons to believe that the concept of ‘production’ involves the idea of some kind of physical or material modification, we have

Artifacts and Fiat Objects

145

reasons to exclude fiat objects from the category of artifacts: fiat objects are just produced by fiat. We can illustrate different reasons found in the literature to be more or less restrictive on the notion of production. 2. If we accept the distinction between artifacts and artificial objects described above and we see an artifact as the intended product of an intentional action, we might find that there are fiat objects that are not artifacts. With regard to point 1, a quite intuitive approach would take production as an action involving some kind of material modification, but such a stance on production does not fit well with some intuitions we may have concerning artifacts as well. There are strong reasons for admitting abstract objects among artifacts; for example, if we choose certain metaphysical options, we may want to enumerate not only mathematical and logical proofs among artifacts but also words, pieces of music, scientific theories and even such social objects as institutions.2 Whether to include or exclude a requirement of material modification for intentional production depends on our preferences concerning the extent of the concept of artifact. Given this double possibility concerning the notion of production, it would be perfectly reasonable either to adopt a less restrictive notion of artifact including fiat objects or to opt for a more restrictive one, excluding fiat objects. The conclusion that the relation between fiat objects and artifacts crucially depends on the more or less restrictive interpretation of the notion of ‘production’ is focused just on one of the components of the concept of artifact. One further consideration on the general notion of artifacts may in fact lead us to reconsider question (2) from a quite different perspective (Does being an artifact imply being a fiat object? And, conversely, does being a fiat object imply being an artifact?) We can place in doubt the relevance of a category of artifact as classically defined by Baker and Hilpinen (see the above definitions). As a matter of fact such definitions impose very few restrictions on the type of objects that can figure as artifacts: in the end, the only restriction in place depends on the limits of our ability to produce something. This results in a hodgepodge class of objects with no relevant similarities that could help us in finding some common nature. At this point we might be tempted to take into account as an alternative approach a naive functional characterization of artifacts, but what we obtain are some odd results, as Wiggins (2001, p. 87) observed. Take, for example, a functional characterization of a pen in terms of ‘any rigid ink-applying writing implement’, a clock in terms of ‘any time-keeping device’ and so forth.

146

Natural and Artifactual Objects in Contemporary Metaphysics

Now, observe that it is part of our way of individuating artifacts to accept a vast range of possible changes in them – for example, parts replacement, dismantling, interruption of functioning and so forth – while leaving the identity of the artifact untouched. In the case of a damaged clock, in order to fix it we may send it to a watchmaker, who may stop it, open it and replace its damaged parts. The very same clock starts to function again. Hence one could argue that the principle of functioning is not related to the conditions of persistence for a clock. Moreover, the functional nature of an artifact is insufficient to specify any common properties in relation to objects belonging to the same artifact kind: Clocks, for instance, may be made of a variety of different kinds of material and may function by radically different kinds of mechanisms and are collected up not by reference to a theoretically hypothesized inner constitution but under functional descriptions that have to be indifferent to specific constitution and particular mode of interaction with environment. (Wiggins 2001, p. 87)

Hence, the category so defined seems to be completely irrelevant from a metaphysical point of view or for any other theoretical interest one may have. Before deriving a conclusion concerning the possibility for fiat objects to be considered a kind of artifacts let us examine point 2, concerning production. If artifacts are intended products of intentional action is it possible that there are fiat objects that are not the intended product of intentional action or even not the product of intentional action at all, that is, fiat objects that are not artifacts and not even artificial objects? So it might be that the question of whether fiat objects are or are not a subclass of the class of artifacts so defined is not only ontologically uninteresting but also questionable. Can we conclude that they are two unrelated notions despite the fact that they have some point of similarity? Is there any further consideration we can do to tackle the problem of the relation between artifacts and fiat objects?

Artifacts and fiat objects: A bottom–up strategy Our proposal is to adopt a different strategy for characterizing the relation between artifacts and fiat objects, a sort of bottom–up one. We start from some considerations concerning artifact kinds leaving in the background the general category of artifacts. We take into account different notions of artifact kinds, even if all of them developed around the notion of function. Broadly speaking

Artifacts and Fiat Objects

147

artifact kinds constitute a particular type of functional kinds, and we propose to treat kinds of fiat objects as functional kinds on par with artifact kinds. Functions are commonly recognized as the most adequate properties to be used for classifying and individuating artifacts. For example, we can say that given an object a at time t and an object b at time tʹ, Identity criterion (IC): If a and b belong to a functional kind K, then a is the same as b iff there is a continuous material path between a and b and a and b are able to perform the same function. The intuition behind IC is that the identity of an artifact is strongly connected with its ability to perform a certain function and the function gives the spatiotemporal identity conditions of an artifact. We propose to consider a similar IC for fiat objects also. There are, as a matter of fact, a lot of fiat objects undoubtedly endowed with functional unity: chairs made of one piece of wood have fiat parts as legs, backs and seats that are functionally individuated. Other interesting examples can be taken from the biological domain: roots, trunk and leafs of a tree are all physically continuous to one another and so they are fiat objects, but are clearly functionally distinguished parts of the entire organism. The same can be concluded for many functional parts of other living organisms. It is interesting to note that Vogt (see Chapter 6 of this volume, esp. pp. 109–13) – as we have done – concludes that there are many fiat objects having functional unity, but, unlike our point of view, by exploiting an innovative characterization of bona fide objects he takes these to be not real fiat objects. As regards the role of function as an essential feature of artifacts, it is important to notice that there are different conceptions of functions, and that they are differently applied to artifacts. Following Carrara and Vermaas (2009), it is possible to distinguish four conceptions of artifact function. 1. The designer/creator intentions account, in which the technical functions of an artifact are the capacities or goals for which agents designed the artifact. This is a position Philip Kitcher (1993) proposed. In the design view of function one holds that what defines an artifact a as an F is that a is a token of the designed artifact kind F. 2. The user intentions account, in which the technical functions of an artifact are the capacities or goals for which agents use the artifact. This so-called use view holds that what defines an artifact a as an F is that a is being, or could be, used as an F, irrespective of whether a was designed as an F in the first place. This conception of function has been proposed, for example, by Neander (1991) and McLaughlin (2001). Note that if one takes

148

Natural and Artifactual Objects in Contemporary Metaphysics

the designer/creator of an artifact as one of its users, the use view may well subsume the designer/creator intentions account. It is a position Dennett, for example, considers when he says that ‘the inventor is just another user, only circumstantially and defeasibly privileged in his knowledge of the functions and uses of his device. If others can find better uses for it, his intentions, clear-headed or muddled, are of mere historical interest’ (Dennett 1990, p. 186). 3. The causal-role account, in which the functions of an artifact are the capacities by which it causally contributes to the capacities of larger and more complex systems (Cummins 1975). For example, the function of a carburettor is the capacity by which it causally contributes to the overall capacities of a car it is a part of. According to the original definition provided by Cummins, a function of a component is a certain kind of disposition to interact with the other components of a containing system so that it produces certain effects. The conception of dispositional property we intend to adopt then becomes crucial, because the category of fiat objects surely includes abstract entities. Cummins adopted a plain causal notion of disposition, and its functional analysis is meant to be a type of explanation that can be applied to a wide range of scientific disciplines: from physical to psychological and even social systems. The notion of disposition presupposes the notions of capacity and causal power and this might raise some problems to the extent that we admit abstract fiat objects; the same problem arises for artifacts as well. 4. The etiological account, in which the functions of an artifact are the capacities for which the artifact is reproduced in a long-term sense (Millikan 1984, 1993 and Preston 1998). For example, the function of aspirin in the twentieth century was painkilling, because painkilling was the reason why aspirin was reproduced. Some supporters of the etiological account of functions, in particular Neander (1991), proposed inclusion of in the etiological characterization of function also the capacities for which artifacts are reproduced in a short-term sense. The stability and the historical persistence of fiat objects clearly suggests that they are actively maintained or reproduced by human community. If this analysis ends positively, we can defend the thesis that, after all, there is some relevant similarity between artifacts belonging to functional kinds (of a certain type, the type depending on the notion of function adopted) and fiat

Artifacts and Fiat Objects

149

objects. This doesn’t allow us to give a straightforward positive answer to the second question of question (2) (Does being a fiat object imply being an artifact?) but we can say that both questions of question (2) (Does being an artifact imply being a fiat object? Does being a fiat object imply being an artifact?), even if natural and intuitive, are not useful in revealing important similarities between artifacts and fiat objects.

Benefits and costs of using functions for artifact and fiat object classification One could argue that the above-described four options work differently as criteria for artifact classifications. In the following, we briefly examine the benefits and the costs of adopting each of them in turn as the basis for artifact classifications and then see if it is plausible to adopt each of them also for the classification of fiat objects. The first two options deal with intentional notions of function and will be discussed here together. The attribution of function according to the designer or user intention can capture quite well many of our intuitions concerning artifact classifications. Some artifact kinds seem to bear in their names their functional nature: can opener, corkscrew, pencil sharpener and screwdriver are a few prominent examples. But, in fact, even for artifacts such as chairs, cars, phones and so forth, it seems intuitive to regard their nature as that of objects designed to perform a certain function. Nonetheless, there are two main problems regarding this option: first, functions commonly associated with kinds, such as chairs or cars, seem to be too loosely described for the individuation of artifacts of the same kind; secondly, we may have objects produced for a certain function that are – as a matter of fact – systematically used to perform a different function and in many circumstances we seem to favour a function attribution related to their use and not to their design. The result is that an object can be classified in different ways according to its different uses. Furthermore, if we try to define functional identity conditions associated with artifact kinds we obtain an ontology of very fragile artifacts that can jump in and out of existence whenever a new use is established for them. Let us examine the first problem. Consider the first phone designed by Alexander Graham Bell and the later phones that succeeded it. On the designer intentions account of artifact functions the first telephone had the function of

150

Natural and Artifactual Objects in Contemporary Metaphysics

aiding the hard-of-hearing, since history has it that Bell designed his original phone for that capacity. Later phones designed by Bell or by others were however designed for long-distance communication and thus have this type of communication as their function on the designer intentions account. The first phone is thus of a different functional kind from the successor phones. Given the later developments in telephone technology, the consequence is that Bell’s original phone and a modern twenty-first-century cellular phone are not of the same functional kind. Is this really a problem? Couldn’t it be acceptable to consider artifacts that we commonly classify as objects of the same functional kind as such even when in fact they are not? Might it not simply be that for our everyday communication we don’t need a fine-grained classification of artifacts? As soon as we move to a technical level we need a some more refined way of describing functions and it is at this level that we can find the tools for such a description. In some sense, we commonly misclassify artifacts, in the same way as we misclassify fruits and vegetables or stars and planets. The way an agent thinks of the artifact he is designing is for sure much richer than a simple description such as ‘keeping time’ for a watch or ‘sitting upon’ for a chair. As is well known, what is characteristic of functions is that they can be described at different levels of abstraction, from the more abstract with an input–output formulation to more detailed descriptions including elements of material constitution and functioning: the more abstract characterization of functions seems to be inadequate to describe the function intended by the designer. The price to be paid is that of regarding our common classifications of artifacts as in need of revision in favour of a more fine-grained and sophisticated classification. It seems to be a small price to pay, which simply places common functional kinds of artifacts on par with common kinds of natural objects: trees, stars, fruits and so forth are all commonly associated with criteria that we have no problem in discarding in favour of scientific ones. The second problem, connected with the proposal of classifying artifacts according to their use, concerns essentially ontological benefits and costs. This problem leads us to connect directly with the discussion of the second criterion of function attribution mentioned. As a matter of fact no one has defended the general thesis that functions are attributed in relation to users’ intentions excluding the possibility of considering designers as users: all the more so if our objective is not that of resolving the general problem of function attribution to artifacts but that of merging it with the individuation of artifact kinds − objects would too easily change their kind. A chair would become all of a sudden a shelf

Artifacts and Fiat Objects

151

and then go back to being a chair even tens of times a day and this would sound as a weird classification already from a commonsensical point of view, but if we add the metaphysical requirement that criteria of classification are or give rise to criteria of individuation, then we would obtain a plethora of objects going in and out of existence too easily. There are no reasons in principle to think that such an option is impossible to defend; maybe someone could take such an inflationary ontology seriously. Of course, one may wonder what is exactly the notion of use at work here; probably a more manageable way to consider it is to understand ‘use’ as ‘standard use’ or ‘common use’. However, we are handling some very pre-theoretical and vague notions3 and their indeterminate nature does not allow us to arrive at a good standard criterion of classification for artifacts. The user intention account is maybe the one that more easily can be applied to fiat objects as well: they are put to uses and their existence does not always start with any easily identifiable event that could demarcate their production. In this regard it seems obvious that a relevant distinction to make is that between institutional fiat objects and non-institutional fiat objects. While Italy may be regarded as a fiat object that originated on 17 March 1861, the northern hemisphere of Earth can be individuated as something with a geographical use but for sure it does not have any clear event that signals its origin. We already remarked on the centrality of the notion of production for understanding the relation between artifacts and fiat objects but we placed in question only one aspect of the notion of production involved in the definition of artifacts, that of material modification; the other important aspect is that of intentionality. Production of an artifact might be a long-lasting process involving the action and thought of a lot of people, but it seems natural to attribute a crucial role to some final decision: something is intentionally produced if at a certain point the processes of production are deemed complete, and this can be the mark of intentional directedness of the process of production. To the contrary we think it is possible to leave open the possibility that fiat objects are created by habits and traditions and, even though these can be regarded as kinds of production processes, they are not intentionally directed processes. To accept a more liberal notion of production including collective and unintentional production is a strategy for treating both artifacts and fiat objects as objects produced by human agents and regarding their functions as those assigned by their users/makers. Nonetheless the attribution

152

Natural and Artifactual Objects in Contemporary Metaphysics

of function in relation to this more liberal notion of production gives rise to an extravagant and inflated ontology, as we already mentioned at the end of the last paragraph. As regards the third approach, the causal-role account, its adoption for artifacts has one main consequence: in order to work out the criteria of classification we need to make reference to a containing system and a functional analysis. In the classification of parts of artifacts we need to appeal to a certain functional analysis of the object and for the classification of whole artifacts we need to consider some external system, typically the system of use, in order to evaluate the function and classify the object used. In such a perspective, the resulting classifications would be contextually dependent on the overall system referred to, and relative to the functional analysis adopted. Hence, the causal-role function doesn’t seem to be the best choice for a functional classification that could work as a taxonomy; nonetheless there seem to be no specific problems for fiat objects in this case. In the fourth account – the etiological one – in order to obtain a functional criterion we need to supply an historical reconstruction of maintenance or (re-)production in which some particular properties of a given artifact play a distinctive causal role: such properties playing that role favour the re-production of objects of a certain type. Quite often an etiological account of function will involve a notion of copy; that is, the items with a certain function are reproduced in the sense that they are copies of other items of the same type. For artifacts it gives rise to some common problems concerning either the attribution of function to the first instances of a certain type, that is, instances that cannot be considered as the result of some history of maintenance, or re-production or else the possible shift in causal role. The first problem can be overcome by accepting that the first exemplars produced are part of a sort of mental history of maintenance4 or accepting a mixed criterion of function attribution. As regards the second one, consider again the case of aspirin (the example is taken from Houkes and Vermaas 2004). The aspirin produced during all of the first half of the twentieth century was intended to be a painkiller; only later, in the second half of the century, was its property as a blood-clot preventer discovered. In the etiological account, in which artifact functions are the capacities for which artifacts are re-produced, the first tablets had painkilling as their function. Indeed, at that time, painkilling was the known capacity of aspirin. A tablet of aspirin produced nowadays has both the function of painkilling and that of blood-clot prevention: this has generated a shift in how to classify aspirin. Shifts in classification are not

Artifacts and Fiat Objects

153

per se problematic but they have the undesirable consequence of allowing different kinds with the same name and with instances with the same physical structure. Let us consider now fiat objects. Their existence depends on cognitive acts of agents intending to refer to portions of physically continuous objects. This intention is socially recognized and adopted and it gives rise to different types of convention. These are established for their practical utility but can persist even despite a complete lack of it. They persist as long as the conventions are active. Hence the most natural way of extending the etiological account to fiat objects is in relation to the production version of it: they come into existence in consequence of certain conventions and these fix their function. As a matter of fact the notion of re-production is sensitive to the same problems of the notion of production and also to further problems concerning the notion of a copy.

Conclusions Our choice to compare artifacts and fiat objects was mainly motivated by the fact that both ontological categories involve a certain dependence of the existence and nature of the objects on human intentions. They seem to be two kindred ontological categories. We started from a classical top–down perspective analysing the main implications of their general definitions, but this survey didn’t give us any conclusive result: we simply noted that a certain problematic aspect in the general definition of artifacts – that is, if production involves material modification – is also crucial for the inclusion of fiat objects among artifacts. Some further considerations led us to question the ontological benefit of such an approach; it is evident, in fact, that a general definition of artifacts doesn’t lead to the individuation of a category of objects significantly similar to one another. We proposed then to shift to a different approach trying to consider fiat objects as well as artifacts as objects that are best classified according to their function, that is, objects that belong to functional kinds. Functional classifications depend on the criteria for function attribution we decide to adopt and so we briefly examined four different classical options. The aim of our analysis is not to give reasons to choose one among these options, but simply to give a critical assessment of each of them in order to give credibility to the idea that fiat objects might be considered objects belonging to functional kinds. Regarding artifacts we have few reasons to doubt that they are best viewed as items of functional

154

Natural and Artifactual Objects in Contemporary Metaphysics

kinds and so a strict comparison between function attribution and classification of artifacts and function attribution and classification of fiat objects may lead to the adoption of a functional classification also for the latter.

Notes 1 2 3 4

On this see Chapter 6 of this volume, on granularity and frame of reference. On words as artifacts, see Chapter 9 of this volume. For a more detailed analysis of the notion of use, see Houkes and Vermaas (2004). See Millikan (1984), (1993).

References Baker, L. R. (2007), The Metaphysics of Everyday Life: An Essay in Practical Realism, Cambridge University Press, Cambridge. Carrara, M. and Vermaas, P. E. (2009), ‘The Fine-Grained Metaphysics of Artifactual and Biological Functional Kinds’, Synthese, 169: 125–43. Cummins, R. (1975), ‘Functional Analysis’, Journal of Philosophy, 72: 741–65 Dennett, D. C. (1990), ‘The Interpretation of Texts, People and Other Artifacts’, Philosophical and Phenomenological Research, 50: 177–94. Dipert, R. (1993), Artifacts, Artworks and Agency, Temple University Press, Philadelphia. Hilpinen, R. (2008), Artifact, The Stanford Encyclopedia of Philosophy, E. N. Zalta (ed.), http://plato.stanford.edu/archives/fall2008/entries/artifact/ Houkes, W. and P. E. Vermaas (2004), ‘Actions Versus Functions: A Plea for an Alternative Metaphysics of Artifacts’, Monist 87: 52–71. Kitcher, P. S. (1993), ‘Function and Design’, in French, P. A., Uehling, T. E. and Wettstein, H. K., (eds), Midwest Studies in Philosophy, Volume XVIII, University of Minnesota Press, Minneapolis, 379–97. McLaughlin, P. (2001), What Functions Explain, Cambridge University Press, Cambridge. Millikan, R. G. (1984), Language, Thought, and Other Biological Categories: New Foundations for Realism, MIT Press, Cambridge, MA. Millikan, R. G. (1993), White Queen Psychology and Other Essays for Alice, MIT Press, Cambridge, MA. Neander, K. (1991), ‘The Teleological Notion of “Function”’, Australasian Journal of Philosophy, 69: 454–68. Preston, B. (1998), ‘Why is a Wing like a Spoon? A Pluralist Theory of Functions’, Journal of Philosophy, 95: 215–54

Artifacts and Fiat Objects

155

Santambrogio, M. (forthcoming), The semantics of artifactual words. This volume. Smith, B. and Varzi, A. C. (2000), ‘Fiat and bona fide Boundaries’, Philosophical and Phenomenological Research, 60 (2): 401–20. Thomasson, A. (2003), ‘Realism and Human Kinds’, Philosophical and Phenomenological Research, 67 (3): 580–699. Vogt, L. (forthcoming). Bona fideness of Material Entities and Their Boundaries. This volume, 103–20. Wiggins, D. (2001), Sameness and Substance Renewed, Cambridge University Press, Cambridge.

9

The Semantics of Artifactual Words Marco Santambrogio

Is the semantics of artifactual words, like table and pencil, externalist? The semantics of a referential expression is externalist if it is not the case that its reference is not determined by some uniquely identifying marks, some unique properties satisfied by the referent and known or believed to be true of that referent by the speaker (Kripke 1980, p. 106). Saul Kripke, Hilary Putnam, Tyler Burge and others have given arguments, which most philosophers have found convincing, to show that the semantics of proper names and of the names of natural kinds is externalist. What about artifactual words? Kripke never paid much attention to them.1 Putnam claimed, at some point, that they too have an externalist semantics (Putnam 1975, pp. 242–4). More than a decade later, David Kaplan suggested that all terms obeying the same syntactic principles ought to share the same semantics and ‘is a bachelor’, for example, ought to be treated in the same way as ‘is a horse’.2 If this is so, then – bachelor being the paragon of terms for which a property uniquely identifying the reference seems to be readily available – bachelor, table and horse are on par and externalism holds true across the board. More recently, semantic uniformity has been forcefully challenged. As far as I can tell, only a minority of philosophers would now be prepared to grant that the semantics of artifactual words is externalist. Some hold a hybrid view, to the effect that ‘against generalized externalism, … artifactual words come in (at least) three different semantic varieties: a few have an externalist semantics, others have an internalist semantics, still others have neither but rather behave as “family names” in Wittgenstein’s sense’ (Marconi 2013, p. 1). I am convinced, to the contrary, that generalized externalism holds true, if it holds true for proper names and names of natural kinds. Arguments in favour of it, like those put forward by Kornblith (1980, 2007), have been underestimated,

The Semantics of Artifactual Words

157

in my opinion. However, I do not attempt a reappraisal here. Instead, I try an utterly different strategy. By radical externalism concerning artifactual words, I mean the following two theses: (a) Members of an artifactual word’s extension share a common nature, that is, a set of necessary features. (b) Possession of such features determines the word’s extension independently of whether the linguistic community is aware of them (ignorance) or can accurately describe them (error). There is some vagueness in these theses, due to the term nature. If it is taken to mean something like essence, then (a) is unnecessarily strong. Externalism is not committed to that. However, under another reading to be explained shortly, I will defend both theses as phrased above. Here is my strategy. First, I argue that the words of all languages, even though they are unlikely to occur to anyone as primary instances of artifacts, are in fact prototypical artifacts. Secondly, since, in most cases, it is a determinate matter which uttered sounds and inscriptions count as words (as opposed to, say, mere noises, such as grunts and hiccups), it can hardly be supposed that some count as words for one reason, some for another, words are likely to have a common nature. However, even though they are most familiar to us, and very easily produced by us, their metaphysics is still very much disputed among philosophers. Since philosophers are as competent speakers as any, it is uncontentious that the nature of words is unknown to some – in fact, to most, if not all – competent speakers. Conditions (a) and (b) above are therefore satisfied for some artifactual words, namely those referring to the words themselves. I also tend to endorse David Kaplan’s common currency conception of words, without claiming to know that it is correct.3 The gist of Kaplan’s view, which stands opposed to a Platonic conception of words as abstract forms, is that a necessary condition for two inscriptions or utterances to be of the same particular word – any word, be it a proper name such as David, or a noun, such as colour – is for them to have a common ancestry, in a suitably defined sense. This can be put in the form of a principle – the Principle of Common Ancestry: Two utterances or inscriptions belong to the same word only if they have some common ancestor. Without amounting to anything like an essence of words in general, the fact that each one of them obeys this principle is a significant trait of their common nature.

158

Natural and Artifactual Objects in Contemporary Metaphysics

What is most important is that a corresponding condition holds for any two specimens to belong to the same biological species, according to most biologists. I will show that such a structural condition is precisely what is needed for the main arguments for externalism to carry through in the case of species words. It is debatable that species essentialism makes much sense. But Twin Earth thought experiments do not depend on essences. In order to prove that what is known to competent speakers does not determine membership of a natural species, all they have to show is that at least one case is imaginable where two organisms have exactly the same traits, which are known to all competent speakers, and yet do not belong to the same species. This is the case with, for example, tigers and Twin Earth tigers, precisely because, ex hypothesi, they have no common ancestry. Nor do Kripke’s arguments for externalism depend on essences. (I am assuming here that his arguments, as well as Putnam’s and Burge’s, are conclusive.) In order to complete my defence of externalism concerning artifactual words in general, I only have to show that the metaphysics of all artifacts obeys something like the Principle of Common Ancestry. In order to show this, I draw on the works by Paul Bloom and Jerrold Levinson, who have defended some version of the Principle, albeit using different terminology. In the following sections, I set out to present my argument.

Words as prototypical artifacts By any reasonable standard, the words of all languages, both natural and formal, belong to the category of prototypical artifacts. They are produced by humans and, unlike, for example, homes and shelters, which many kinds of animals are also able to build, only by humans, either directly or indirectly (e.g. by means of speech synthesizers).4 Unlike a large number of artifacts, such as paths, villages and cities, which are mostly produced by humans not intentionally, words are almost invariably produced intentionally.5 Unlike some artifacts that are intentionally produced but not intended, such as pollution, scraps and footprints,6 words are almost always intended.7 Unlike some borderline artifacts such as objets trouvés, and stones used as doorstops, which are not produced by humans, words are almost always the product of conscious and purposeful activity. Dan Sperber characterizes prototypical artifacts as follows: ‘Prototypical artifacts are middle-sized, spatially and temporally continuous material objects’

The Semantics of Artifactual Words

159

(Sperber 2007, p. 125). Whisky, drugs, plastic and so forth are not spatially and temporally continuous. Cities and the DNA of GMOs are not middle-sized. Multiplication tables, institutions of all kinds and possibly money itself are abstract or at least not invariably material objects. On the other hand, written words – inscriptions – are middle-sized, spatially continuous material objects. If they are not continuous, for example, when some fonts are used, they can easily be made continuous, with no kind of loss. Utterances are temporally continuous acoustic events. Words in a sign language are corporeal events. Now, of course an event is not, properly speaking, an object. But I wonder if Sperber really means to exclude all kinds of music, cinema, performances, not to mention poetry, audio books and the rest, from typical artifacts. Note that we directly perceive all words by means of our senses – we see, hear, even touch them (e.g. if they are in Braille). Not all artifacts have proper functions to discharge. There are breeds of dogs that owe their existence to human intervention, and are in this sense artificial, even though they have neither a function nor a proper function.8 Dishes of all kinds have no specific function, except that they are all for eating. We would be hard put to find the function of artworks and I am not even sure one could use them.9 Insofar as they have no specific function or use, one could take them to be atypical artifacts (even though I wonder if one could think of anything more clearly artifactual than a work of art). But, in any case, words are not like that: they have a very clear function, insofar as they are meant to be used as a means of communication and are mostly irreplaceable in this function. Needless to say, words are by far the most widespread artifacts on Earth.10 A basic distinction has sometimes been made concerning the kind of dependence artifacts have on our thoughts and intentions. Amy Thomasson distinguishes dependence of origin from dependence of continuing existence and properties (1999, 2007). A knife, for example, owes its origin to a designer and a factory but the rest of its existence is independent of them. Money, on the other hand, depends for its very existence on the continuing attitude of an indefinitely large number of people. To which side of the divide do words belong? They belong to both, as they depend on the thoughts and intentions of the speakers of their respective languages both for their origin and for their continuing existence. I shall only mention in passing that the whole point of distinguishing natural entities from artifacts metaphysically must have something to do with the contrast between nature and culture – that is, between what could exist even if mankind had never come into existence and what owes its existence to human

160

Natural and Artifactual Objects in Contemporary Metaphysics

minds, intentions and activities. If the contrast makes any sense at all, there is little doubt that language is the paragon of all that falls on the side of culture.

The nature of words In the debate on artifacts and externalism, the metaphysical notion of nature looms large. What is generally meant is something like Locke’s real essence. As far as natural substances and species are concerned, it is clear that their natures can be known to none, or almost none, of the speakers in a linguistic community. In fact, most natures are normally unknown until the community reaches a fairly advanced stage in scientific development. It then follows that an individual speaker, or even the totality of speakers as a whole, can be competent in the use of a natural-kind word even if nothing much is known concerning the conditions for belonging to its extension. This is enough to show that requirement (b), in the definition of externalism in general, is satisfied for natural-kind words, if requirement (a) is. It is generally agreed that natural kinds have a common nature, in the sense of a set of necessary features for belonging to them. As for artifactual words, some philosophers claim that the reasons for externalism for natural substances and species do not apply to artefacts. First, artifacts have no nature that can be hidden.11 Since artifacts are made by us, we – perhaps not individually, but rather as the linguistic community taken as a whole, in which experts are bound to exist for each kind of artifact – must know what they are, that is, their function, their use, how they are manufactured and so forth. Moreover if, in the case of artifacts, nature means more or less what it does in the case of natural kinds, then how can artifacts be possessed of a nature if the specimens of most artifacts are made of different materials (think of, for example, spoons), have different shapes and work according to different physical principles (e.g. telephones) and the same artifact can even have different proper functions at different times (e.g. aspirin before and after 1950, when people began using it to prevent blood clotting)?12 Since artifacts have no hidden nature; speakers competent in the use of an artifactual word must grasp what it takes to fall within its extension – such conditions being contained in the meaning of the word which they know ex hypothesi. Words like spoon, telephone and aspirin are like bachelor on this score. Much as a competent speaker of English knows that a bachelor is an unmarried man, so must she know that a spoon is an instrument

The Semantics of Artifactual Words

161

with a certain shape and function and so forth. Thus, neither condition (a) nor condition (b) in the definition of externalism is satisfied in the case of artifactual words. That was the main argument against generalized externalism. I shall now show that the argument fails. Note, first of all, that the notion of nature, as used here, is far from clear. If it is assumed that most natural kinds have a nature, much as chemical elements have atomic numbers and a natural phenomenon like heat is the motion of molecules – that is, if it is assumed that their nature is, or is related to, their essence – this is likely to be false. Species, which loom large in Kripke’s and Putnam’s arguments for externalism, don’t have essences according to present-day biologists. In a pre Darwinian age, species essentialism made sense. Such essentialism, however, is out of step with contemporary evolutionary theory. Evolutionary theory … tells us that a number of forces conspire against the existence of a trait in all and only the members of a species. (Ereshefsky 2001, section 2.1)13

Because it varies considerably even within the same species, not even DNA is of much use in this respect. Kripke’s and Putnam’s arguments, which are convincing, as I have assumed, cannot therefore rely on such a simple-minded notion of nature. The notion they use must be distinct from that of essence. Now, whatever it is, it has not been shown that artifacts lack it. In fact, they do have their own nature, as I now proceed to show. I begin with words. Whether or not words have essences, they certainly have nothing comparable to either DNA or atomic numbers. But it is quite likely that they do have a nature, in the sense that necessary features exist for being a word (as opposed, for example, to mere noise). A few distinguished philosophers are busy discussing what it is and of course their discussion makes good sense.14 The very fact that the discussion is still going on is proof enough that the nature of words in general is not yet known, or not generally known. In his seminal work ‘Words’ (1990), David Kaplan contrasts two main conceptions of the metaphysics of words, the orthographic conception and the stage-continuant model, which is the one he favours. I shall present the outline of Kaplan’s views here, since it will prove very useful when I consider artifacts other than words. The orthographic conception is the conventional token/type theory, according to which words are abstract objects, types, exemplified by utterances and inscriptions, which are their tokens. Types are identified as strings of letters corresponding to sounds. ‘Expressions of the language consist of strings of

162

Natural and Artifactual Objects in Contemporary Metaphysics

atoms called “letters”, certain strings form words’ (98).15 Meanings, and semantic properties in general, are not appealed to in order to identify words. The view that Kaplan himself upholds is the stage-continuant model or the common currency conception.16 Let us begin with how Kaplan envisages the creation and the transmission of a word like Hesperus. I have a story about how the word ‘Hesperus’ came to us. The story is surely incorrect … but it gives the flavor of my views. I imagine that at some point some Babylonian looked up in the sky one evening and said (in Babylonian) ‘Oh, there’s a beauty. Let’s call it …’, and then he introduced the name. What he did was to create a word. He created a word as a name, a tag, for this beautiful heavenly body. He then passed that word on to other people through inscriptions and utterances. Those people passed that word on to others and so on. … As it went through different communities, the way this word was pronounced and written changed in very dramatic ways, through whatever processes account for dialectal variation. The presupposition of these processes of change are the principles of continuity in accordance with which a changing word retains its identity. … Changes in pronunciation and spelling need not suggest the notion of replacement of one word by another, which then takes up the task in the manner of a relay race. Rather, we can use the notion of a single entity undergoing change. (100–01)

The ‘principles of continuity’ that Kaplan is appealing to here can be stated concisely as a single principle: Principle of Common Ancestry: Two utterances or inscriptions belong to the same word only if they have some common ancestor – where an utterance or inscription is an ancestor of another if whoever uttered or wrote the latter intended to repeat the former or some intermediate utterance or inscription that was similarly intended as a repetition of it.17

Even though Kaplan does not place much emphasis on the artifactuality of words, it is only to be expected that the intentions of the maker have to be taken into account in order to determine which artifact she produces.18 The difference in sound or shape or spelling of any two utterances or inscriptions can be just about as great as you would like it to be and yet they can be of the same word (e.g. color/ colour, centre/center), provided the intentions of their users are appropriately related. The contrast with the orthographic conception is stark. Note that, normally, that is, if no misunderstanding occurs, the histories of two words that originated in distinct events stay distinct, in the sense that they have no utterance or inscription in common. In particular, names that originated

The Semantics of Artifactual Words

163

in distinct baptism ceremonies are distinct. Suppose that two distinct persons were named David. Clearly, their baptisms are distinct events, in which their names originate. Since the origins are distinct, the names are distinct. Thus we have two names David. It is incorrect to say that David Kaplan and David Israel, for example, share the same name. They have distinct names.19 The model emerging from all this is the stage-continuant model. A word is a four-dimensional continuant with various utterances or inscriptions as its stages: ‘utterances and inscriptions [i.e. the actual concrete sounds and figures] are stages of words, which are continuants made up of these interpersonal stages along with … intrapersonal stages’ (Kaplan 1990, 98).20 A particular utterance or inscription of a word w thus turns out to be a spatial-temporal part of w: w is a fusion of its spatial-temporal parts or ‘stages’. Kaplan’s conception is subtle and rich with interesting consequences but further details are not needed for our purposes. He is clearly aiming at establishing the metaphysics, that is, the nature, of words. The main tenet of his conception is the Principle of Common Ancestry. It is to be noted that this Principle gives us nothing like a notion of essence for words, the reason being that an essence for any kind of entity consists in some property or other that is common to all and only the individual members of that kind, whereas the Principle of Common Ancestry only sets a necessary condition for pairs of utterances and inscriptions to belong to the same word. (Some relations produce interesting properties. For example, from an equivalence relation, such as being parallel for lines, one can define, by abstraction, the notion of direction, and then the property of having a definite direction. However, the relation of having a common ancestor is definitely not an equivalence relation.)21 Even so, the Principle amounts to a substantial requirement concerning the nature of words in general. It is important to realize that being intentionally related in the manner specified above is a necessary, not a sufficient, condition for two utterances or inscriptions to belong to the same word. All sorts of changes, in meaning as well as in reference, might intervene in between. For instance, incense, in present-day English, is certainly related to the Latin word incensum, insofar as an uninterrupted chain exists from users of the latter to present-day users of the former, all the proper intentions being in place. In this sense, therefore, they are the same word. However, as they have different meanings, for whatever reason, there also is a sense in which they are not the same word. As another example, it is debatable whether the name Madagascar, as used by some of Marco Polo’s informants to refer to a region on the African mainland, is the same word as our Madagascar referring to the great African island.

164

Natural and Artifactual Objects in Contemporary Metaphysics

I stressed above that words are artifacts, and prototypical ones at that. However, there is also a sense in which they are entirely natural objects.22 There is no contradiction here, as there is no reason to suppose that natural entities and artifacts belong to disjoint realms.23 There is a substantial reason why words are best thought of as natural entities. We saw above that the Principle of Common Ancestry holds for words in general – if the stage-continuant conception is on the right track. (Note that nothing in the argument I am presenting depends on this conception being exactly right. For the argument to go through, the mere possibility that it is approximately correct concerning the nature of words is quite sufficient.) The same structural principle holds – according to most biologists – for all biological species, which can, but need not, be taken as the fusion of their spatial-temporal parts; that is, their specimens, much as words, are the fusion of all their utterances and inscriptions and all of their intra-personal stages. Principle of Common Ancestry for Species: For two individual organisms to belong to the same species it is a necessary condition that they share some ancestors.

Here, ancestor is to be understood in its ordinary, biological, sense. Of course, having ancestors in common is a necessary, not a sufficient, condition for belonging to the same species, as shown by the fact that we belong to a different species than monkeys, even though some of our ancestors are also theirs. This falls far short of giving anything like the essence of biological species in general, for the same reasons specified above – namely, it gives no property that is necessarily common to all and only the specimens of a species and only sets a necessary condition for pairs of specimens to belong to the same species.24 It is also perfectly compatible with all sorts of variations in size, shape, behaviour and even DNA, occurring within the same species. Still, it is a condition that tells us something quite important concerning the nature of species.25 It is important to realize that the Principle of Common Ancestry both for words and for species is independent of the stage-continuant conception. Perhaps, even the type/token conception can accommodate the intuition that historical-relational features are crucial for an individual to belong to a particular word or species. The only thing that the Principle of Common Ancestry stands opposed to is the idea that what determines membership in a word or species are overt properties like form, sound and physical appearance – that is, that an animal looking just like a tiger cannot fail to be a tiger. Note that the Principle of Common Ancestry for species is sufficient to validate the arguments given by Kripke and Putnam for externalism in the case of natural-kind words. I consider here only the Twin Earth thought experiment.

The Semantics of Artifactual Words

165

Suppose that, on Twin Earth, a species of animals exists, each specimen of which is molecule for molecule indistinguishable from some of our tigers. Therefore, even their DNA is the same. Would they be tigers? Of course not, since, as can be assumed, they have no common ancestor. The corresponding argument carries through in the case of nouns referring to any one of those particular artifacts that are the words of any language, natural or artificial. Suppose that on Twin Earth a word exists that is spelled exactly as our word tiger and is pronounced exactly the same way. Is it the same word as our tiger? If the stage-continuant conception is on the right track and the Principle of Common Ancestry holds for words, it is not – unless some linguistic exchange took place between Earth and Twin Earth, and those words have a common origin. But even if the stage-continuant conception is not on the right track, the two words must differ as to their referents, because there are no tigers on Twin Earth, for the reason given above. If sameness of reference is a necessary condition for being the same word, this shows that our tiger and their tiger are not the same word. What holds for tiger holds for all words referring to biological species and, of course, for proper names referring either to Earthlings or to Twin Earthlings. What about those Twin Earth words that coincide with some English words in spelling, pronunciation and also reference? Suppose that, by some fantastic coincidence, a remote Indonesian tribe has a word that is pronounced like Sun and also refers to the Sun. Would we say that it is the same word? I think we would say that it is just a curious coincidence, unless it can be proved that the tribe had some contact in the past with English speakers. This, of course, confirms that the Principle of Common Ancestry is on the right track. Let us take stock. I claimed that words are prototypical artifacts. It is also apparent that, whether or not they have an essence, words have an objective nature. Since philosophers are generally competent speakers and are still debating what their nature is, it cannot be claimed that every competent speaker as such knows what it takes for an inscription or an utterance to belong to the extension of a word. Thus the main argument against externalism fails, at least as far as the names of some artifactual kinds – namely, words in quotes – are concerned.

Other artifacts and their nature Words are artifacts. Each one of them is a species of sorts. What about other artifacts? Before considering whether any general principle can be given,

166

Natural and Artifactual Objects in Contemporary Metaphysics

comparable to the Principle of Common Ancestry and encompassing all artifacts, let us see if the story told by Kaplan, concerning the creation and transmission of a word like Hesperus, can be repeated, mutatis mutandis, for other artifacts. Here is a story concerning that ancient implement for writing, the stylus. The story is certainly incorrect but it gives the flavour of my views. I imagine that at some point some Egyptian scribe thought, ‘Hieroglyphs could be drawn on wax-covered tablets more easily by using something like this …, rather that the twigs we have been using so far.’ He managed to have the instrument he had in mind made for him by some craftsman, after explaining what exactly he intended as to the shape, the material to be used and so forth. What he did, with the help of the craftsman, was to create an artifact. The craftsman, even though he was illiterate and did not understand much about the function of the artifact and how to use it, realized that he could sell copies of it to other scribes. So he made copies which went around and were sold to other scribes, who passed them on to other craftsmen to copy and so on and so forth. As it went through different communities – I am echoing here Kaplan’s words, quoted above – the ways this instrument was made and perhaps also used changed in very dramatic ways, through whatever processes that account for this sort of variation. The presupposition for these processes of change is the principles of continuity in accordance with which a changing instrument retains its identity. Changes in shape and use need not suggest the notion of replacement of one instrument by another, which then takes up the task in the manner of a relay race. Rather, we can use the notion of a single entity undergoing change. The principles of continuity involved here amount to this: if a craftsman intends to make a specimen of the same artifact as the one he takes as a model, then, if all goes well, he makes a specimen of the same artifact, whether or not he knows anything about its function, its use, the name given to it by other speakers, and even if he makes some significant changes, either unwittingly or on purpose, for example, in order to make it more efficient or cheaper. It must be borne in mind that repetition, which is as important for artifacts as it is for words, requires very little knowledge and understanding on the part of those doing it, as shown by children, who can repeat any word at a very early age, long before they show any tendency to philosophize and to pore over the nature of words. The same holds for other artifacts. Note that a craftsman making an artifact of any given kind – for example, a stylus – need not even be aware that a kind already exists, or will exist, over and above the specimen he is working on and the one he takes as his model.26

The Semantics of Artifactual Words

167

The conception of artifacts in general that is closest in spirit to the one I defend here is Paul Bloom’s. Bloom generalizes to all artifacts an idea put forward by Jerry Levinson concerning artworks: something is a member of the kind artwork if it was successfully created with the intention that it should belong to that kind.27 In general, ‘We construe the extension of artifact kind X to be those entities that have been successfully created with the intention that they belong to the same kind as current and previous Xs’ (Bloom 1996, p. 10). As Bloom points out, this is not meant as a definition of the kind word X. Take for instance the kind chairs. It would not do to define chairs as being those objects that are successfully created with the intention that they be chairs, if only because ‘chair’ occurs on both sides of the ‘definition’. By the same token, the characterization above does not give the essence of chairs. If the intention to make something like the artifact one takes as a model is a necessary condition for the two products to belong to the same artifactual species – I use the term species here on purpose, of course – then a Principle of Common Ancestry for Artifacts has a prospect of holding true in general. Once again, it does not amount to a sufficient condition. A craftsman might intend to make a stylus only a bit different from its model, which can leave marks not only on wax-covered tablets but also on other surfaces, and the product of his work might end up being classified by the linguistic community under a different name – for example, ballpoint pen. Another craftsman might intend to make a caravel and, for lack of wood, end up being the inventor of the sloop and so on.28 How much change in function, shape, material, working principles and so forth is allowed (or required) for the name of an artifact to be kept (or changed) is completely unpredictable. Different linguistic communities have different policies. For instance, the difference between a blender and a hand processor has been taken by the English-speaking community to be conspicuous enough to require a change of name, whereas the Italian-speaking community has retained the former name frullatore (only adding the qualifier phrase a immersione).29 It is nonsense, of course, to suppose that the English approach is better than the Italian one, or vice versa. There is no fact of the matter concerning how an artifact ought to be named. Sometimes, a change in name matches no change whatsoever in the thing: it is conceivable that the same firm producing aspirin might change its name into, for example, Cardiolin, in order to better advertise the beneficial effects it has on the heart, without changing the chemical composition. In this case, it would be true to state that aspirin is Cardiolin. A hand processor, on the other hand, is definitely not a blender.

168

Natural and Artifactual Objects in Contemporary Metaphysics

The unpredictability of these variations and their effects on how artifacts are classified, for example, by what names, parallels that concerning organisms. How conspicuous can variations in DNA, shape, colour, behaviour and so forth be without affecting membership of the same species as one’s ancestors? Not being a biologist myself, I have no firm opinion, but I suspect that in drawing the boundaries between species many factors have to be taken into account and weighed against one another. Ample room is left, therefore, for the biologists’ educated but not completely constrained judgement.30 How are we to think of the kinds of artifacts – chairs, pencils, blenders, aspirin and so forth? Once again, I suggest that biological species offer us a useful model: thus, the kind of chairs, for example, might be conceived of as a four-dimensional continuant with the individual chairs as its stages, that is, as the fusion of the archipelago of its specimens. As to the semantics of their names, are there thought experiments of the Putnam kind available to show that it is externalist in kind? Let us see. I pointed out above that even without resorting to the notion of essence for tigers, Putnam’s main argument for the externalist semantics of the word tiger holds since, there being no ancestors common to our tigers and to Twin Earth tigers, Twin Earth tigers are not tigers, even though they are physically indistinguishable from them. The Principle of Common Ancestry is the only assumption needed. Now, let us consider any artifact we like. The Pearsall Mousetrap will do. (The Pearsall Mousetrap is an artifact, the function of which can easily be guessed, invented by one Ralph E. Pearsall. It is carefully described in a 1968 US patent.) Suppose that Twin Earth is even more similar to Earth than Putnam makes it out to be, insofar as, not only is twin water H2O but also for every Pearsall Mousetrap existing on Earth an object exists on Twin Earth that is physically indistinguishable from it. Would the latter be a Pearsall Mousetrap? We are entitled to assume that Mr Pearsall and his invention are completely unknown on Twin Earth. Two cases are to be distinguished: the Pearsall Mousetrap could either be a definite description or be a name. It might be thought that, in both cases, the obvious answer is that the Twin Earth artifact is a Pearsall Mousetrap. I shall dispute that. Suppose that the Pearsall Mousetrap is a definite description. Then it cannot denote the Twin Earth artifact which, for one thing, was not invented by Ralph E. Pearsall – or, in any case, not by our Ralph E. Pearsall – and, for another, is not a mousetrap and must have a different proper function than

The Semantics of Artifactual Words

169

our Pearsall Mousetrap, since there are no mice on Twin Earth, the Twin Earth rodents indistinguishable from our mice not being mice. Now suppose that the Pearsall Mousetrap is the name of a kind of artifact. Do the Twin Earth artifacts belong to the same kind as our Pearsall Mousetraps? They are physically indistinguishable, to be sure. But this is not sufficient.31 Even forgoing the difference in their proper functions, a further difference is that, while our Pearsall Mousetraps are all related insofar as they are ‘descendants’ of the first mousetrap produced by Ralph E. Pearsall, the Twin Earth artifacts are not so related. If the Principle of Common Ancestry holds for artifacts, then this fact clearly militates against conflating the Twin Earth artifactual kind with ours. But does that Principle hold? The obvious objection is that on many occasions the same artifact was independently created in different places by different makers. So, why should it not be possible for the Pearsall Mousetrap to have been independently created both on Earth and on Twin Earth? First of all, I would dispute the supposition that independent inventions of the same artifacts are not rare events. However, this is not the place to discuss historical matters.32 In a more philosophical vein, let us consider a supposed case of independent inventions of the same artifact, the telephone, credit for which is frequently disputed. Alexander Graham Bell was the first to be awarded a patent, but Charles Bourseul, Antonio Meucci, Elisha Gray and Johann Reis, amongst others, have also been credited with the invention of it. We are not interested here in the priority issue. Let us pretend that they all worked independently of one another – which is what really matters to us.33 Can we say that they invented the same artifact? The likelihood that two poets write two poems word for word identical is fantastically low.34 Still, I would say it is definitely possible. But it is not possible for two persons independently to produce the same material object. As a matter of fact, the differences between the artifacts produced by Bell, Bourseul, Meucci, Gray and Reis were quite substantial. They had different shapes, they were made of different materials and even the principles of physics applied were different. Let us call their respective inventions, Bell’s telephone, Bourseul’s telephone and so forth. (The inventors themselves used different names, such as speaking telegraph, sound telegraph and so forth, which would eventually be replaced by the newer, distinct name, telephone.) Now, what about the telephone, with no qualification? Is it not obvious that since Bell’s telephone, Bourseul’s telephone and so forth were all telephones, the Principle of Common Ancestry stands falsified? I do not think it does.

170

Natural and Artifactual Objects in Contemporary Metaphysics

Tables, chairs, cupboards and so forth are all pieces of furniture. Is furniture an artifactual kind or species? As tables, chairs and all pieces of furniture are so different from every point of view, except for the fact that we group them together and use them to furnish houses, I would say that furniture is an artifactual family or cluster, but neither a kind nor a species. Tools, vehicles and so forth are clusters, too. Clearly, no such principle as the Principle of Common Ancestry holds for furniture, tools and vehicles. The Principle holds true only for species. Similarly, I claim that Bell’s telephone, Bourseul’s telephone, Meucci’s telephone and so forth were distinct artifactual species, all belonging to the cluster of telephones. (Modern telephones, on the other hand, do belong to the same kind or species, insofar as they come from Bell’s master patent.) As another example, it is possible that noodles were independently invented in China and in Europe. But they were not quite the same artifact, if only because Chinese noodles are made of rice, whereas European noodles are made of wheat. Thus, noodles are a cluster of artifacts. The analogy with words is useful at this point, too. I pointed out above that, according to the stage-continuant conception, David Kaplan’s given name and David Israel’s given name are not the same common currency name, insofar as they originated in different baptisms, even though they belong to the same generic name. Generic names might be taken to be clusters of common currency names. Note, incidentally, that Kripke’s and Putnam’s arguments were only meant to show that externalism holds true for such names as tiger, which names a species, not for bird, tree and quadruped – the latter being, possibly, the name of a cluster. Now, even though clusters, unlike species, fail to satisfy the Principle of Common Ancestry, this is not to say that the semantics of cluster names is not externalist. Consider, for example, the term furniture. The entry for it in a standard dictionary is, ‘The movable articles that are used to make a room or building suitable for living or working in, such as tables, chairs, or desks’. Clearly, reference to tables, chairs and other prototypical instances of furniture is not dispensable in giving the meaning of the term, because, due to the variety of the artifacts that can serve as furniture today, not to mention those that may so serve in the future, no attempt to give necessary and sufficient conditions in completely general terms, without bringing in particular instances, is likely to succeed. If semantics as a whole is compositional then, the semantics for table and chair being externalist, the semantics for furniture is bound to be externalist as well. Moreover, a simple argument shows that the same holds for the word

The Semantics of Artifactual Words

171

artifact itself. The nature of artifacts is as unknown to most, or all, competent speakers as the nature of words is. Since, as we have seen above, widespread ignorance of the nature of something is a sufficient reason for the semantics of its name to be externalist, the conclusion follows. Parallel arguments can be used to show that the semantics for bird, tree, quadruped, animal and so forth must be externalist, too – contrary to what some of Kripke’s and Putnam’s commentators have claimed. The notion of cluster, as introduced here, might turn out to be useful in biology as well. Suppose that, by some fantastic coincidence, a species of animals exists on Twin Earth undistinguishable from our tigers. Or suppose – what is more likely – that it becomes feasible to synthesize DNA and then to give birth to completely artificial tigers and other animals and plants. Would they belong to the natural species? If they did, the Principle of Common Ancestry for species would stand falsified. But I think we would prefer to say that they belong to the same cluster, not to the same species. This is the place to comment on some differences between words and artifacts in general. Kaplan has a principle for words which might be called the Principle of the Authority of the Speaker: in order to establish what words exactly were uttered or written, ‘ask what the speaker would say of his or her own performance.’ In other words, as far as the identity of words is concerned, it is only the speaker’s intentions that matter. The linguistic community has no say in these matters and no authority to force the speaker to abide by prevalent standards. This is not to say that things always go smoothly. The speaker herself might harbour more than one intention in uttering a word. As Kripke convincingly showed, cases exist in which a speaker uses an extant name, but neither intends to give it a new referent nor merely repeats it with whatever referent it may already have. In these cases, the speaker intends to use the name to refer to a particular referent that she somehow has in mind. If, as may happen, that particular referent is not the same as the name’s original one, it is not quite determined what exactly happens. The name, particularly if the speaker sets a precedent for others, might change its referent in due course. Or – this is another possibility – it might be substituted by a new name with a new referent, retaining the appearance of the former.35 In the case of artifacts, something similar is not unlikely to happen. The intentions of craftsmen and engineers always count in establishing what they make or design. However, as a matter of course they harbour both the intention of making a new specimen of an extant artifact and that of improving on the prevalent design. It is up to others – the sales department, the users, the linguistic

172

Natural and Artifactual Objects in Contemporary Metaphysics

community as a whole – to decide whether the innovations are sufficiently significant to conclude that a new kind of artifact, deserving a new name, has come into existence. We can imagine that this is what happened with the hand processor. Possibly, its inventor’s intention was only to produce yet another blender with minor changes. However, the linguistic community as a whole, possibly influenced by the manufacturer’s sales department, decided otherwise: in the end, that artifact comes to be known as the hand processor, which is something other than the blender. In this case, both the maker’s intentions and those of the linguistic community matter equally, even though the latter prevail in the long run. Are there cases in which the maker’s intentions count for nothing from the start and only those of the community matter? Artworks, it seems to me, are a case in point. Children and amateurs often intend to make artworks and fail. Why they fail is not quite clear, but it is possible that our current notion of the artistic has a lot in common with the notion of what is fashionable.36 Let us briefly go back to the hand processor, the pita, the armchair and the different policies of different linguistic communities. Let us consider, for example, a particular armchair and ask, Is it a chair? It can be supposed that most English speakers would answer that it is – an armchair being a chair – whereas no French competent speaker is likely to assent, as there is a sharp divide, in French, between fauteuils and chaises. This is only natural if the speakers’ intentions and intuitions are to count in identifying artifactual kinds, and it is conceivable that intuitions are partly moulded by linguistic usage. We have here a kind of linguistic relativity, albeit of a benign kind. In itself, this is no more surprising than the fact that the question whether a particular action by a particular person is a legal crime might receive different answers in different countries.37 Of course, this also raises the problem of the metaphysical reality of artifactual kinds, but we are interested here only in the semantics of artifactual words and we can place it aside. To conclude, since competent speakers usually (but not invariably) have some idea concerning the function and possibly even the shape of an artifact, but are normally unaware of the intentions of its maker (not to mention those of the linguistic community as a whole), they are in the same position with respect to artifacts as they are concerning natural kinds. Thought experiments of the Kripke and Putnam kind can always be run concerning artifactual names. I conclude that externalism holds true for the latter as much as for the names of natural kinds.

The Semantics of Artifactual Words

173

Acknowledgements I wish to thank for their insightful comments Massimiliano Carrara, Giulia Felappi, Andrea Iacona, and all the participants to a seminar in the Philosophy Department, University of Bergamo, where I was given the opportunity to discuss an earlier version of the present paper.

Notes 1 At least in his published works. 2 ‘I would treat “is a bachelor” in the same way as “is a horse”. While acknowledging the metaphysical differences between a species and bachelorhood, the syntactical unity of “horse” and “bachelor” suggests an analogous semantical treatment’ (Kaplan 1989: 581 n. 30). 3 Kaplan’s conception is set forth in his 1990 paper. Hawthorne and Lepore criticize it in their (2011) work. Kaplan (2012) is a response to that work. 4 See, for example, Gould (2007) for a vast array of animal artifacts. Most of what is discussed in this section applies to both word types and word tokens. The type/token distinction itself, in the case of words, has been questioned, as I explain shortly. For the time being, however, this does not matter much. 5 See Thomasson (2007). 6 See Thomasson (2007). 7 Instances of unintended words are rare. ‘It seems to me that cases in which one utters a word other than the one intended must be exotic indeed’ (Kaplan 1990, p. 105). 8 I use proper function in the sense of Ruth Millikan’s ‘direct proper function’, as some function that past members of a kind successfully performed, and such that because of that functioning other members of the kind were produced. See Millikan (1993, pp. 13–14). 9 Kant famously held that what is proper to artworks is that they lack both function and purpose. 10 It must be mentioned that Sperber hesitates to include words among typical artifacts: ‘Is a multiplication table an artifact, in spite of being an abstract object? Is a word?’ (2007, p. 125). But he seems to take it for granted that words are abstract. I shall come back shortly to this particular conception of words as abstract forms. 11 ‘I believe … that there is no such underlying nature of pencils, nor is there a presumption of such a nature. What makes something a pencil are superficial characteristics such as a certain form and function. There is nothing underlying

174

12 13

14 15

16

17

Natural and Artifactual Objects in Contemporary Metaphysics about these features’ (Schwartz 1978, p. 571). See also Schwartz (1977, Introduction), Schwartz (1980), Schwartz (1983). For this kind of argument, which is fairly frequent in the literature, see, for example, Marconi (2013). That species do not have essences is cited by Richard Grandy as ‘a complication of the standard story’, that is, ‘the explanation of how our biological terms refer (independently of the descriptions we may associate with the term)’ (Grandy 2007, p. 23). I will show, to the contrary, that it is no complication at all. For scholarly references to the sparse literature on the topic, see Hawthorne and Lepore (2011). See also Sainsbury (2015). Among those who uphold this view, Hawthorne and Lepore quote Davidson (1979, p. 90); ‘[Expressions are] either a pattern which similar tokens exemplify, or a class of similar tokens’ (Haack 1978, p. 75); ‘Expressions … can be construed as classes of perceptible particulars similar in some physical respect to given perceptible particulars’ (Hugly and Sayward 1981, p. 184); ‘An expression is a phonic or graphic form: the former has acoustic properties and the latter spatial ones’ (Gillon 2004, p. 162) and (Lyons 1977, p. 18). This conception owes much to the picture presented by Kripke in Naming and Necessity of how a name originates and is circulated within the linguistic community. A few years before Kaplan, Ruth Millikan had given a unified theory about language and biological categories that is reminiscent of that conception: ‘What makes tokens of a word … tokens of the same word … is, in the first instance, history, not form or function’ (Millikan 1984, pp. 72–3). This principle is not to be found in Kaplan (1990) as such, but I take it to be implicit, for example, in the following passage: The identification of a word uttered or inscribed with one heard or read is not a matter of resemblance between the two physical embodiments (the two utterances, the two inscriptions, or the one utterance and one inscription). Rather it is a matter of intrapersonal continuity, a matter of intention: Was it repetition? We depend heavily on resemblance between utterances and inscriptions … in order to divine these critical intentions. If it sounds like ‘duck’, it probably is ‘duck’. But we also take account of accent and idiolect and all the usual clues to intention. It is the latter that decides the matter. (Kaplan 1990, p. 104) Also ‘this notion of repetition is central to my conception’ (104). Kaplan (2011) states the principle thus: ‘What makes two utterances utterances of the same word is that they descend from a common ancestor’ (509). In this work, Kaplan also advances a closely related principle, which might be called the Principle of the Authority of the Speaker: in order to establish exactly what words were uttered or written, ‘Ask what the speaker would say of his or her own performance.’ Kaplan contrasts it with another principle that was put forward by Hawthorne and

The Semantics of Artifactual Words

18

19 20

21

22 23

24

25

175

Lepore (2011, p. 463) – ‘Tolerance*: Performance p is of a word w only if p meets relevant local performance standards on w’. I shall come back to the relation between the Principle of Common Ancestry and the Principle of the Authority of the Speaker. There is nothing in the orthographic conception that depends on words being artifacts. Strings of sounds and marks as such could be produced by other animals besides humans, or even by chance, as far as the orthographic conception is concerned. On the other hand, once intentions are taken to be crucial in identifying words, as on Kaplan’s conception, it becomes clear that words must be human artifacts. They have the same generic name, though. I shall not consider here the notion of generic name. As a matter of fact, even though his terminology suggests the adoption of a four-dimensionalist metaphysics, which takes things to be space–time worms or, at any rate, things that do not change but only have temporal parts, Kaplan disavows any such adoption. ‘This was certainly not what I had in mind. In fact, my paper emphasises the notion of change for words, especially changes in spelling and pronunciation’ (Kaplan 2012, p. 508). It is crucial to the stage-continuant conception that some of the stages of a word are mental stages, when the word is stored in memory. It is easy enough to give examples, involving misunderstandings of various kinds, of three utterances of proper names a, b and c, such that a and b have a common ancestor as well as b and c, even though a and c have none. This can happen, for example, when two common currency names are taken to be one by some speaker and then linked and stored together in a single location. ‘Common currency names (and other common currency words) are not abstract constructions, they are natural objects’ (Kaplan 1999, p. 116; my emphasis). It might be thought that whether artifacts and natural kinds are disjoint depends entirely on how one defines the former. For instance, if we follow Risto Hilpinen (2011) in characterizing an artifact as ‘an object that has been intentionally made or produced for a certain purpose’, then of course artifacts and natural kinds are disjoint, by definition. However, we would be hard put to deny that canine breeds are entirely natural, even though they were intentionally originated by breeders. In view of the fact that all living organisms on Earth are likely to come from the same primeval cells, it might look as though the relation of having a common ancestry is transitive in general. However, instances of non-transitivity can be concocted resorting to science fiction. Nowhere does Kaplan explicitly say that common currency names and words in general are species. However, I believe this comes pretty close to what he has in mind. For one thing, he points out that the Platonic, pre-Darwinian conception of species stands opposed to the naturalistic, Darwinian conception much as the

176

26

27 28

29

30

31

32

Natural and Artifactual Objects in Contemporary Metaphysics orthographic conception of words stands opposed to his own, common currency conception. Also, he explicitly says that generic names are species. To repeat, for the overall argument in favour of externalism to carry through, no robust notion of nature, or essence, of a kind, either natural or artifactual, is needed, as distinct from the set of necessary conditions for membership in it. The robust notion of nature is what Kit Fine is after in his 1994 work. Fine points out that, for example, everyone necessarily belongs to the singleton of which one is the only member, but this is no part of one’s own nature. On the other hand, any necessary condition for membership of an individual in a kind, that competent speakers might be unaware of, is quite sufficient for externalism. Even less, necessary conditions for two individuals to belong to the same kind are sufficient. Thomasson (2007) maintains that a substantive conception of an artifact kind K must be involved in the intentional production of an artifact of that kind. My little story concerning the origin of the stylus seems to go against her claim. See Levinson (1979), (1993). To see that the relation of having a common ancestry for artifacts is not transitive, consider an old fashioned telephone, a smartphone and a computer. The first and the second have a common ancestor, as well as the second and the third. But the first and the third do not. The example is taken from Marconi (2013). I simply assume that Marconi is right in holding that the intersection between the kind of hand processors and that of blenders is empty – that is, that no hand processor is also a blender. As another example, in some Mediterranean cuisines pita bread can be either stuffed or topped with all sorts of vegetables and meat without the name of the resulting dish changing in the least: ‘In Palestinian, Lebanese, Israeli, Egyptian and Syrian cuisine, almost every savoury dish can be eaten in or on a pita, from falafel, lamb or chicken shawarma, kebab, omelettes such as shakshoula (eggs and tomatoes), hummus and other mezes’ (Wikipedia, ‘Pita’). In Italian cuisine, on the contrary, pizza (the word is clearly related to ‘pita’ – it might even be the same word) is clearly to be distinguished from calzone, which is made of exactly the same ingredients stuffed inside the bread, instead of topping it. As a further example, consider the armchair. The English classify it as a chair – if the composition of the word is to be taken seriously. The French do not, as fauteuils are definitely not chairs. I understand that the contemporary notion of biological species differs from the traditional one, which was defined by interbreeding. Even the notion of a planet admits of no simple definition. I understand that the rights legally granted to the author of a patented artifact do not cover artifacts even physically indistinguishable but independently originated. This seems to confirm the point made in the main text. As a matter of fact, independent inventions of the same artifact are much less common than one might suppose. Apparently, the alphabet was invented only once

The Semantics of Artifactual Words

33

34

35

36

177

in the history of mankind. See Diamond (1997). The same might hold true even for noodles. At this point, it is appropriate to point out that the general case of artifacts resembles that of words more closely than is generally realized. At first sight, artifacts and words look very different. For one thing, one might suppose that most artifacts are due to inventors who can be identified relatively easily. It is indeed true that we actually know the names of those who invented the ballpoint pen, the zip, the safety razor, aspirin, the blender – not to mention the computer and every other patented invention. On the other hand, we would be hard put to find almost any inventor of a word – proper names and scientific names being an exception. The vast majority of words come from other words by slight and hardly perceptible variations. Is this not a significant difference between words and other artifacts? Not really. We are overly impressed, I think, by the huge number of new, patented inventions that have come into being in the last century or so. The vast majority of artifacts come from other artifacts by slight and hardly perceptible variations. Almost everything that pertains to food, clothing, agriculture, and traditional ways of life has a long history, the origin of which is lost in a distant past. This is only pretence. Bell’s US Patent no. 161,739, ‘Transmitters and Receivers for Electric Telegraphs’, was granted in 1875. As early as 1844, Innocenzo Manzetti had mooted the idea of a ‘speaking telegraph’. Ten years later, Bourseul published an article, in French, ‘Transmission électrique de la parole’ (Electric transmission of speech), describing a ‘make-and-break’ type telephone transmitter later created by Johann Reis (Wikipedia, ‘Telephone’). In 1861, Reis publicly demonstrated the Reis telephone before the Physical Society of Frankfurt. The early history of the telephone is intricate, but it is doubtful that the inventors were unaware of one another. Thus, it might not be inaccurate to say that the telephone was collectively invented, as a development of the telegraph. Would they be the same poem? In ‘Pierre Menard, Author of the Quixote’, Jorge Luis Borges says they would not be. Jerrold Levinson (1979, 1980, 1993) gives convincing reasons in support of this intuition. ‘I find it plausible that a diachronic account of the evolution of language is likely to suggest that what was originally a mere speaker’s reference may, if it becomes habitual in a community, evolve into a semantic reference. And this consideration may be one of the factors needed to clear up some puzzles in the theory of reference’ (Kripke 1977, p. 271). Levinson (2007) takes a nearly opposite view: As regards artworks, ... it is far from clear that any such non-purely-historical conception of arthood is in play, or that there is any minimal success condition of a substantive sort on the making involved. Why is that so? It helps to recall what, in the post-Duchampian era, are two salient features of art-making in contrast to standard artifact-making. First, one can more or less simply declare

178

Natural and Artifactual Objects in Contemporary Metaphysics something a work of art, and it becomes such. Or at least one can, in certain contexts, or with certain standing, do so. Second, anything, whatever its material constitution, cultural, category, or ontological status, can become, or can be incorporated into, a work of art. (79)

37 ‘According to the ordinary use of language, an action a of a certain person is called legal crime if the penal law of the country in which that person lives places the description of a kind of action to which a belongs in the list of crimes’ (Carnap 2000, p. 308). See also the half-serious remark attributed to Arnold Schoenberg who, in his Hollywood exile, was once asked by a journalist if it was true that he played ping-pong. ‘No!’, replied the composer, ‘When I play ping-pong, it is table tennis. But when my wife plays table tennis, then it is ping-pong.’

References Bloom, P. (1996), ‘Intentions, History, and Artifact Concepts’, Cognition, 60: 1–29. Burge, T. (1979), ‘Individualism and the Mental’, Midwest Studies in Philosophy 4: 73–121. Reprinted in T. Burge, ed., Foundations of Mind. Oxford: Clarendon Press, 2007. Carnap, R. (2000), Logical Syntax of Language, Routledge, London. Davidson, D. 1979, ‘Quotation’, in Inquiries Into Truth and Interpretation. Oxford: Oxford University Press, 2009: 79–93. Diamond, J. (1997), Guns, Germs, and Steel, W. W. Norton, New York. Ereshefsky, M. (2001), The Poverty of the Linnean Hierarchy: A Philosophical Study of Biological Taxonomy, Cambridge University Press, Cambridge. Fine, K. (1994), ‘Essence and Modality’, Philosophical Perspectives, 8: 1–16. Gillon, B. (2004), ‘Ambiguity, Indeterminacy, Deixis and Vagueness’, in B. Gillon and S. Davis, (eds), Semantics, Oxford University Press, Oxford, 157–87. Gould, J. (2007), ‘Animal Artifacts’, in E. Margolis and S. Laurence, (eds), Creations of the Mind: Theories of Artifacts and Their Representation, Oxford University Press, Oxford, 249–66 Grandy, R. (2007), ‘Artifacts: Parts and Principles’, in E. Margolis and S. Laurence, (eds), Creations of the Mind: Theories of Artifacts and Their Representation, Oxford University Press, Oxford, 18–32. Haack, S. (1978), Philosophy of Logic, Cambridge University Press, Cambridge. Hawthorne, J., and Lepore E. (2011), ‘On Words’, Journal of Philosophy, CVIII, 9: 447–85. Hilpinen R. (2011), ‘Artifact’, in E. Zalta, (ed.), Stanford Encyclopedia of Philosophy, http://plato.stanford.edu/entries/artifact/ Hugly, P. and Sayward, C. (1981), ‘Expressions and Tokens’, Analysis, 41: 4.

The Semantics of Artifactual Words

179

Kaplan, D. (1989), ‘Afterthoughts’, in J. Almog, J. Perry, and H. Wettstein, (eds), Themes from Kaplan, Oxford University Press, Oxford, 565–614. Kaplan, D. (1990), ‘Words’, The Aristotelian Society, Supplementary Volume LXIV, 93–119: 117–8. Kaplan, D. (2012), ‘Words on Words’, Journal of Philosophy, CVIII, 9: 504–29. Kornblith, H. (1980), ‘Referring to Artifacts’, Philosophical Review, 89: 109–14. Kornblith, H. (2007), ‘How to Refer to Artifacts’, in E. Margolis and S. Laurence, (eds), Creations of the Mind: Theories of Artifacts and Their Representation, Oxford University Press, Oxford, 138–49. Kripke, S. A. (1972), ‘Naming and Necessity’, in D. Davidson and G. Harman (eds), Semantics of Natural Language, Reidel, Dordrecht. Reprinted with a new introduction as Naming and Necessity. Oxford: Blackwell Publishers, 1980 (page numbers given relate to this volume). Kripke, S. A. (1977). ‘Speaker’s Referencer and Semantic Reference’, Midwest Studies in Philosophy 2: 255–76. Reprinted in S. Kripke, 2011, Philosophical Troubles, Oxford and New York: Oxford University Press. Kripke, S. A. (1979), ‘A Puzzle about Belief ’, in A. Margalit, (ed.), Meaning and Use, Reidel, Dordrecht. Kripke, S. A. (1982), Wittgenstein on Rules and Private Language, Harvard University Press, Cambridge. Levinson, J. (1979), ‘Defining Art Historically’, British Journal of Aesthetics, 19: 232–50. Levinson, J. (1980), ‘Autographic and Allographic Art Revisited’, Philosophical Studies, 38, 4: 367–83. Levinson, J. (1993), ‘Defining Art Historically’, Journal of Aesthetics and Art Criticism, 47: 21–33. Levinson, J. (2007), ‘Artworks as Artifacts’, in E. Margolis and S. Laurence, (eds), Creations of the Mind: Theories of Artifacts and Their Representation, Oxford University Press, Oxford. Lyons, J. (1977), Semantics, Cambridge University Press, Cambridge. Marconi, D. (2013), ‘Pencils Have a Point: Against General Externalism About Artifactual Words’, Review of Philosophy and Psychology 4 (1): 497–513. Millikan, R. (1984), Language, Thought, and Other Biological Categories: New Foundations for Realism, MIT Press, Cambridge, MA. Millikan, R. (1993), White Queen Psychology and Other Essays for Alice, MIT Press, Cambridge, MA. Putnam, H. (1975), ‘The Meaning of “Meaning”’, in K. Gunderson, (ed.) Language, Mind and Knowledge, Minnesota Studies in the Philosophy of Science, 7. University of Minnesota Press, Minneapolis. Reprinted in H. Putnam, Philosophical Papers, vol.2: Mind, Language and Reality. Cambridge, U.K.: Cambridge University Press. Sainsbury, M. (2015). ‘The Same Name’, Erkenntnis 80 (S2): 195–214. Schwartz, S. P. (1977). ‘Introduction’, in S. P., Schwartz, (ed.), Naming, Necessity, and Natural Kinds, Cornell University Press, Ithaca and London, 274–87.

180

Natural and Artifactual Objects in Contemporary Metaphysics

Schwartz, S. P. (1978), ‘Putnam on Artifacts’, Philosophical Review, 87: 566–74. Schwartz, S. P. (1980), ‘Natural Kinds and Nominal Kinds’, Mind, 89: 182–95. Schwartz, S. P. (1983), ‘Reply to Kornblith and Nelson’, Southern Journal of Philosophy, 21: 475–9. Schwartz, S. P. (2006), ‘General Terms and Mass Terms’, in M. Devitt and R. Hanley, (eds), The Blackwell Guide to the Philosophy of Language, Blackwell Publishers, Oxford. Sperber, D. (2007), ‘Seedless Grapes: Nature and Culture’, in E. Margolis and S. Laurence, (eds), Creations of the Mind: Theories of Artifacts and Their Representation, Oxford University Press, Oxford, 124–37. Thomasson, A. (1999), Fiction and Metaphysics, Cambridge University Press, and Thomasson (2007). Thomasson, A. (2007), ‘Artifacts and Human Concepts’, in E. Margolis and S. Laurence, (eds), Creations of the Mind: Theories of Artifacts and Their Representation, Oxford University Press, Oxford.

10

Are Linguistic Objects Fiat or Bona Fide? An Ancient Proposal Maddalena Bonelli

Preliminaries The primary aim of this chapter is to reconstruct some considerations attributable to Alexander of Aphrodisias regarding the problematic relation between two levels of linguistic objects and the things they stand for in light of the distinction that Barry Smith has been elaborating since 1994 between bona fide and fiat objects.1 Though the text of Alexander’s own commentary on Aristotle’s On Interpretation has been lost, it is worth trying to understand what he must have been saying on the basis of what we find in Boethius and Ammonius not only because he was the most sensitive and faithful reader of Aristotle in late Antiquity (he held the chair of Peripatetic Philosophy at Athens at the cusp of the second and third centuries CE2) but also because he not infrequently anticipates modern concerns. Before turning to the ancient theories we may look at how the conception of bona fide and fiat objects handles linguistic facts. This is a rather intricate business, pursuing which would take us too far from the level of detail required for our dicussion here, but we can hardly do without an outline. On Smith’s account, there are two basic reasons why ordinary language has a great deal to do with fiat objects. The first is that, when they are uttered, words and sentences are processes that are demarcated by fiat on the basis of a concrete material sound that is not itself articulated into bona fide linguistic units. In this respect, language itself is a bona fide process, with a beginning and an end; but its internal articulation requires that there be, as Smith puts it, internal ‘delineations’, which are fiat.3 The second reason is that, even if

182

Natural and Artifactual Objects in Contemporary Metaphysics

language adapts itself to the external world and is the guarantor of the truth or falsity of sentences, it is also the case that the external world is, in a certain sense, tailored to fit our language and our linguistically generated expectations. In short, natural language contributes to the making of boundaries and, so, of fiat objects. And, in the long run, this is why we can express a ‘matter of fact’ in more than one way.4 It is important to stress that the existence of fiat linguistic boundaries depends on a bona fide reality that is made of objects among which we can trace fiat boundaries. The sentences we pronounce are responsible to truth and falsehood. Language, after all, is a cognitive and not a creative matter, and the things that we grasp by means of language are indeed grasped by way of our linguistically expressed concepts, but this does not mean that these things exist merely in virtue of those concepts.5

Expressed words: The Peripatetic theory In a well-known passage very early in the De Interpretatione, Aristotle asserts: Sounds are symbols of affections of the soul; written words [are symbols] of sounds. As writing is not the same for all, nor are the sounds. But those [affections] of which these [sounds] are primarily signs are the same for all, and those things of which these [affections] are likenesses are also the same [for all].6

Ἔστι μὲν οὖν τὰ ἐν τῇ φωνῇ τῶν ἐν τῇ ψυχῇ παθημάτων σύμβολα, καὶ τὰ γραφόμενα τῶν ἐν τῇ φωνῇ. καὶ ὥσπερ οὐδὲ γράμματα πᾶσι τὰ αὐτά, οὐδὲ φωναὶ αἱ αὐταί· ὧν μέντοι ταῦτα σημεῖα πρώτων, ταὐτὰ πᾶσι παθήματα τῆς ψυχῆς, καὶ ὧν ταῦτα ὁμοιώματα πράγματα ἤδη ταὐτά

This passage has been commented on not only by classicists but also by philosophers of language as it is regarded, along with Plato’s Cratylus, as the starting point of a reflection on the relations between language and reality.7 In the absence of a follow-up by Aristotle himself, I would like to pursue some thoughts on the cited passage that we find in Alexander of Aphrodisias. In the first place, I aim to clarify his considerations about the status of the articulated vocal sounds (ta hen tè phoné), which is to say, words; I leave to a second moment Alexander’s reflections on their relations to the affections of the soul (pathemata tes psyches), which in this context are identifiable as thoughts,8 and to things (pragmata).

Are Linguistic Objects Fiat or Bona Fide?

183

As will have been noticed, there are many puzzles about Aristotle’s position that are not adequately explained (nor indeed clarified) by the author himself. Setting to one side written signs, which further complicate matters, we may raise various questions about these few lines from Aristotle: (1) In what respect (or respects) are sounds symbols and signs of thoughts? What differences might there be between a symbol and a sign? (2) In what respect(s) are thoughts the same for all? (3) In what respect(s) are thoughts likenesses of things? From what has come down to us of Alexander of Aphrodisias,9 we may say somewhat generically that he shares Aristotle’s realism, according to which things ‘produce’ the corresponding thoughts in the soul (presumably thanks to experience and its conceptualization, but also thanks to language10), and these thoughts are in turn ‘signified’ by sounds, which are symbols and/or signs of thoughts. Over and above this much, Alexander presents some extremely interesting reflections on the conventionality of language and on the question of understanding why Aristotle should say that articulated sounds are primarily symbols (or signs) of thoughts and not directly of things. These are matters to which we shall return, but we can already see how they fit into what we have already said about the fiat–bona fide distinction as regards linguistic facts. On the one hand, the question is raised of the less-than-total naturalness or physicality of language, which, as we shall see, is a claim that is not shared by all ancient thinkers. On the other hand, there is the interesting question of the role of thought in the constitution of language and, to some extent, of reality. Indeed, in the question raised by Alexander, we hear an echo of two views expressed in Antiquity, one according to which there is a direct correspondence between words and things in the world and the other that saw only an indirect link between words and things mediated by thoughts.11 It is widely agreed that in the De Interpretatione Aristotle treats articulated sounds (phonai) as signs or symbols insofar as they are conventional,12 while thoughts and things are by nature.13 The standing of phoné is thus exactly that of a relation between a bona fide outer boundary, made up of a material process with a beginning and an end, and fiat inner delimitations, insofar as the sequence is conventional because established by human fiat in the act, as Aristotle would say, of meaning and thus of fixing the relation between thought14 and reality. In his commentary on this passage, Ammonius carries a syllogism that perhaps appears in Alexander’s lost commentary on De Interpretatione but that

184

Natural and Artifactual Objects in Contemporary Metaphysics

appears to point in the opposite direction, namely towards the total naturalness of names and verbs.15 The syllogism runs as follows: (1) Names and verbs are sounds. (2) Sounds exist by nature. (3) Therefore names and verbs exist by nature.16

ἑξῆς δὲ τούτοις ἐπισκεψώμεθα τὸν συλλογισμόν, ὃν ὁ Ἀφροδισιεὺς ἐξηγητὴς ἐκτίθεται, κατασκευάζειν δοκοῦντα μόνως εἶναι φύσει τὰ ὀνόματα καὶ τὰ ῥήματα· τὰ γὰρ ὀνόματα, φησί, καὶ τὰ ῥήματα φωναί, αἱ δὲ φωναὶ φύσει, τὰ ἄρα ὀνόματα καὶ τὰ ῥήματα φύσει.

Ammonius proceeds then to refute the syllogism and, though he does not attribute the refutation to Alexander, there are good reasons to think that it is indeed his.17 The argument according to which the naturalness of sounds implies the naturalness of names and verbs is based on a misundertanding of the first premise, for names and verbs are not sounds simpliciter, but only ‘in relation to their matter’ (kata ten hylen).18 In effect, Alexander equates the naturalness of the phoné with the naturalness of any artifact. As we cannot say that a wooden door is natural merely because wood is natural,19 so we cannot say that the phoné exists by nature merely because the matter of which it is made is natural. Here, Ammonius is picking up on a polemic in Aristotle’s Physics, probably against Antiphon and other materialists, who recognize only the material cause in the generation of things.20 Another argument, this time securely attributable to Alexander, upholds the conventionality of names on grounds that we have by nature the capacity to give names to things, whereas, if names existed by nature, this capacity would have no role to play.21 This is to say that we make use of matter to forge names, under the necessity of expressing thoughts and reality or, rather, as we shall see, of expressing thoughts about reality.

A second Peripatetic view: Sounds are symbols of thoughts and not of things Also referring to De Interpretatione (16a3–8) we find, this time in Boethius, some comments of Alexander regarding the relation, which Aristotle does not find problematic, among words, thoughts and things. Let us first see what Boethius attributes to Alexander22: But, since Aristotle says this: ‘but those of which these are in the first place signs (quorum … primorum notae), namely the affections of the soul, are the same for all’: if names are names for things, why does Aristotle claim that sounds23 are in

Are Linguistic Objects Fiat or Bona Fide?

185

the first place signs for thoughts (ut primorum intellectuum notas esse voces)? In fact, the name is given to the thing, for example when we say ‘man’ we signify (significamus) a thought, but the name is the name of the thing, i.e. the rational mortal animal. So why are sounds not in the first place signs for the things (non primarum magis rerum notae sint voces) to which they have been given, rather than for thoughts? But Alexander says that this has been said [by Aristotle] perhaps for the following reason: though the sounds are names of things (rerum nomina sint), nevertheless, we do not use the sounds to signify things (ut res significemus) but [to signify] those affections of the soul that are produced in us by the things.24

As already noted, Boethius is commenting on 16a7, of which he gives a verbatim translation: quorum autem haec primorum notae, eadem omnibus passiones animae sunt (‘of which these are primarily signs are the same affections of the soul for all’).

Here then is the question that Alexander poses, which can be subdivided into three phases of which we shall look first at the first two: (1) Names are names of things (and, as we have seen, humans are able to give them to things, which is what makes them conventional). (2) Nevertheless, in the De Interpretatione Aristotle says that the voces (which are expressed or pronounced terms) are in the first place notæ of the affections of the soul. Let us look at the first phase (1): to clarify what it means, Alexander (or Boethius) gives the following example: when we say ‘man’ we signify a thought, yet the name is the name of the thing, that is, the rational mortal animal. To be a bit clearer, let us take the example of a proper name: ‘Socrates’ is the name of the thing Socrates, and names in general are names of objects. Now, in the case of a common noun (as in the example of ‘man’), the problem arises of which is the corresponding object. One reply might be ‘the set of men’, so that, in the example, ‘man’ applies to ‘rational mortal animal’.25 Regarding the second phase (2), in the De Interpretatione Aristotle asserts that articulated vocal sounds (voces) are in the first place signs (quorum … primorum notae) of the affections of the soul. We should bear in mind that Boethius (or Alexander) says twice that the voces signify thoughts; in return, he never says that a name signifies a thing, going no further than to say that a name is of the thing or is given to the thing and adding that we do not use a vox to signify a thing. This means that, of itself, a name does not signify anything and that its sense or significance is given by the thought produced in us by the thing and not

186

Natural and Artifactual Objects in Contemporary Metaphysics

directly by the name. This is to say, we create names on the basis of our thoughts and, hence, on the basis of our cognitive activities.26 The fact that the voces are in the first place notæ of the affections of the soul would seem to imply that they are in the second place notæ of things. From this follows Alexander’s question in the third phase: (3) If names are the names of things, why did Aristotle not say that voces are in the first place notæ of things? In other words Alexander seems to have spotted a contradiction between (1) (names are names of things) and (2) (names are signs of thoughts); Aristotle ought to have said that names are at least in the first instance signs of objects. Alexander’s solution lies in saying that, while it is true that the voces (considered as names) are names of things, it is nevertheless true that we use them not to signify objects, but to signify our thoughts about objects. Thus, Alexander seems to read (2) as meaning in the first instance: when I emit a phoné, I signify one of my thoughts. For instance, when I proffer the sound ‘Socrates’, I think of Socrates and signify my thought of Socrates. And this in turn can be read as (2b): if I produce a phoné, I give a public sign of the fact that I am thinking of something of which the phoné is a name. For instance, if I produce the sound ‘Socrates’, I give a sign of the fact that I am thinking of an object of which the sound ‘Socrates’ is a name. The same will hold for a proposition, which in its minimum form is made up of at least two voces: a name and a verb. At least in most cases, Alexander’s theory works, because it seems reasonable to think that if I pronounce the name ‘Mummy Mummy’, this vox is taken as a sign that I am thinking of my mother. But what are we to say of common nouns? Much the same: if I assert ‘I want oysters this evening’, I am giving a public sign that I am thinking of oysters. Nevertheless, Alexander’s account has been subject to criticism.27 In the first place, objections have been raised against the idea that when I speak, my primary aim is to indicate what I am thinking. If, for instance, I say ‘crabs scuttle backwards’, my aim is probably not to manifest a thought of mine, but rather to give information about crabs. Furthermore, even if the primary aim were to communicate thoughts, this does not imply that words signify thoughts. If, for instance, I say ‘I am thinking of Claudia’, the word ‘Claudia’ does not refer to one of my thoughts, but rather to Claudia. When, on the other hand, I say what I am thinking, I should refer to the objects of my thought and not to the thought about those objects.

Are Linguistic Objects Fiat or Bona Fide?

187

Though Alexander’s argument is very cramped, these objections are not entirely well founded. As to the first, neither Alexander nor Aristotle speaks of a primary intention or aim. All we are told is that when voces are used, a thought is signified, such as the thought that crabs scuttle backwards. (It makes no odds whether the thought is autobiographical or not.) This is a theory that is all the more interesting from the point of view of the fiat–bona fide distinction, whose predecessors we are seeking to elucidate. For, here it seems that Alexander’s grasp of the complexity of linguistic facts comes out clearly: he was aware that we have to take into account both a material fact and its cognitive ‘elaboration’. As to the second objection, from what Boethius tells us, we may certainly assert that Alexander would not deny that in the sentence ‘I am thinking of Claudia’ the word ‘Claudia’ should refer to Claudia. Indeed, this is precisely what he asserts: the name refers to the object, but it signifies a thought.28

The proposition: Problems for the Aristotelian conception As has been justly noted, in the De Interpretatione, Aristotle is above all concerned to establish a parallel between the realm of language and the realm of thought,29 which is what we have seen in the initial remarks about the relations of words to thoughts and reality. If we are to take seriously what Aristotle and – even more so – Alexander say about the separate phonai (name and verb), we have to accept as a consequence that likewise propositions, from the simplest (as we know, made up of a name and a verb) to the most complex, do not directly signify things, but rather our thoughts or judgements about them.30 This could be explained in various ways,31 but Alexander appears to opt for the idea that meaning depends on something that is peculiar to the thinker and that is transitory. In this sense, it concerns a particular thought that crops up in the thinker in the given circumstances.32 The parallel between the realm of language and the realm of the mental then continues in the passages in which Aristotle considers the kind of proposition that is the main concern of the De Interpretatione, namely the assertoric statement (logos apophantikos). In this connection, he says that he is interested only in speech that has truth or falsity in it (17a2–3),33 specifying that the true and the false have to do with the combination or division of names and verbs, which, taken on their own are neither true nor false (16a9–16).

188

Natural and Artifactual Objects in Contemporary Metaphysics

On these grounds, Aristotle establishes that assertoric statements can be divided into affirmations, which attribute something to something, and negations, which subtract something from something (17a25–6). In turn, this distinction is connected to truth and falsity understood precisely in terms of combination and division, which as we shall see is a source of difficulties for the theory. But, in this sense, to attribute something to something (or to subtract something from something) means to combine (or divide) names and verbs (which function logically as subjects and predicates) with truth or falsity. The interest of Aristotle’s position lies in the fact that truth is not located in reality,34 but in our uniting or separating terms that refer to things that are really united or separated (in the case of the true) or uniting or separating terms that refer to things that are not really united or separated (in the case of the false).35 Once more, a direct link between language and thoughts is established in which reality’s role is (or should be) as a guarantor of our linguistico– mental operations. A further point that seems to me important concerns the presuppostion that underlies the theory of the relation between language and reality, which is the well-known theory of truth as adequatio intellectus et rei. The point is that there must be an isomorphism between the structure of the simple proposition, made up of a subject and a predicate bound by the copula, and that of reality, made up, in Aristotle’s view, of substances and properties. This isomorphism (as well as the structure attributed to propositions) has come under attack from various angles,36 and has stimulated discussion of the constitutive power of language with respect to reality. We shall now see how, in this light, we can locate a problem raised by Alexander of Aphrodisias. As is well known, at the outset of the Prior Analytics, Aristotle gives the following definition of the premise (protasis) of a syllogism: A premise is a proposition that affirms or denies something of something.37

Πρότασις μὲν οὖν ἐστὶ λόγος καταφατικὸς ἢ ἀποφατικός τινος κατά τινος.

In his comment on this, which has come down to us, Alexander seeks to show the coherence of the definition of premise in the Prior Analytics with the one that we have already seen of an assertoric statement at De Interpretatione 17a2–3: Not every proposition is assertoric, but only that in which the true and the false are to be found.38

Are Linguistic Objects Fiat or Bona Fide?

189

Commenting on the Prior Analytics passage, Alexander asserts, A definition of a premise (protasis) would be an assertion (apophansis) that Aristotle has given in the De Interpretatione: a proposition (logos) ‘in which the true and the false are to be found’. But in fact, [here] he defines more precisely; indeed, even if the premise (protasis) and the assertion (apophansis) are the same in relation to the substrate (kata to hypokeimenon) they differ in their definition: to the extent that they are true or false, they are assertions (apophanseis); to the extent that they are expressed affirmatively or negatively, they are premises (protaseis). In other words: to be an assertoric statement (apophantikos logos) for something is simply [a matter of its being] true or false, whereas to be a premise (protasis) for something is [a matter of] how it contains the true or the false. For this reason, premises (protaseis) that do not contain the true or the false in the same way are the same proposition (logos), but not the same premise (protasis). Indeed, the premise (protasis) that justice is good is like the one that affirms that injustice is bad, for they are both true and are both affirmations (kataphaseis), but they are not the same premise (protasis) because their substrates and their predicates are different; further, the affirmation [justice is good] and the negation [justice is not bad], which are true, are alike in this respect and are the same statement (logos) but are not the same premise (protasis) because the quality of the assertion (apophansis) is different39: they are not the same protasis, but they are the same apophansis.40

Εἴη μὲν ἂν λόγος τῆς προτάσεως κἀκεῖνος, ὃν τῆς ἀποφάνσεως ἀπέδωκεν ἐν τῷ Περὶ ἑρμηνείας· λόγος γάρ, “ἐν ᾧ τὸ ἀληθεύειν ἢ ψεύδεσθαι ὑπάρχει“. οὐ μὴν ἀλλὰ καὶ ἰδίως τὴν πρότασιν ὁρίζεται· καὶ γὰρ εἰ ταὐτὸν κατὰ τὸ ὑποκείμεν ον πρότασίς τε καὶ ἀπόφανσις, ἀλλὰ τῷ λόγῳ γε διαφέρει. καθόσον μὲν γὰρ ἢ ἀληθής ἐστιν ἢ ψευδής, ἀπόφανσίς ἐστι, καθόσον δὲ καταφατικῶς ἢ ἀποφατικῶς λέγεται, πρότασις. ἢ ὁ μὲν ἀποφαντικὸς λόγος ἐν τῷ ἀληθὴς ἢ ψευδὴς εἶναι ἁπλῶς τὸ εἶναι ἔχει, ἡ δὲ πρότασις ἤδη ἐν τῷ πως ἔχειν ταῦτα. διὸ αἱ μὴ ὁμοίως ἔχουσαι αὐτὰ λόγοι μὲν οἱ αὐτοί, προτάσεις δὲ οὐχ αἱ αὐταί. ἡ μὲν γὰρ λέγουσα πρότασις, ὅτι ἡ δικαιοσύνη ἀγαθόν, ὁμοίως ἔχει τῇ λεγούσῃ, ὅτι ἡ ἀδικία κακόν, ἀληθεῖς δὲ ἄμφω καὶ καταφάσεις, προτάσεις δὲ οὐχ αἱ αὐταί, εἴ γε διαφέρουσιν ἐν αὐταῖς οἵ τε ὑποκείμενοι καὶ οἱ κατηγορούμενοι. ἀλλὰ καὶ ἡ κατάφασις καὶ ἡ ἀπόφασις αἱ ἀληθεῖς ὁμοίως ἔχουσι κατὰ τοῦτο καὶ εἰσὶ λόγοι οἱ αὐτοί, προτάσεις δὲ οὐχ αἱ αὐταί, εἴ γε διάφορον τὸ ποιὸν τῆς ἀποφ άνσεως ἐν αὐταῖς, καὶ εἰσὶ προτάσεις μὲν οὐχ αἱ αὐταί, ἀποφάνσεις δὲ αἱ αὐταί.

Alexander was the first commentator to spot trouble here, as, in his wake, many would bring out the problematic relation between the syllogistic premise (which is in any case a proposition) that we find in the Prior Analytics and the definition of an assertion that we find in the De Interpretatione. Alexander begins from the definition of ‘premise’ as affirmative or negative (which we find in the Prior Analytics, and also, as we have already noted, in De Interpretatione (17a25–6))

190

Natural and Artifactual Objects in Contemporary Metaphysics

and of ‘assertion’ as true or false (once more at 17a2–3) to show that they stand in a complex relation of equivalence. In the early part of the passage just cited, Alexander explains that a premise and an assertion can be identified ‘in relation to the substrate’ (kata to hypokeimenon), while their definitions differ. What Alexander is pointing to is that premise and assertion are equivalent in that where there is the one there is also the other, that they predicate of the same subject and that perhaps they are the same proposition. But the former is defined in terms of the true and the false, while the other in terms of affirmation and negation. We might go as far as to say that Alexander is prefiguring the distinction made famous by Frege between sense and reference: premise and assertion have the same reference, but not the same sense. By way of example, we might say that the proposition ‘justice is good’ is a premise if we consider its quality (in this case, affirmative), but it is an assertion if we consider its truth-value. This is not the only time that, perhaps in light of Stoic logic, Alexander presents a distinction of this sort: to explain the equivalence of one-being that we find in Metaphysics, IV, ii (1003b22–6), he appeals to the identity of subject, but difference of meaning.41 More interesting still is the continuation of the text, in which Alexander explains the difference between premise and assertion calling on the true and the false and applying it to pairs of propositions. Considering the axis true–false, an assertion (apophansis, which Alexander treats as equivalent to apophantikos logos) has only need of the true or the false to be defined, while a premise (protasis) is defined by how it contains them (affirmatively or negatively). Where we have a pair of premises (protaseis) that do not contain the true and the false in the same way, we may find ourselves faced with the same utterance (logos42), but these will not be the same premise (protasis). To clarify the point, Alexander offers the following cases: (1) ‘Justice is good’, ‘justice is not bad’ (2) ‘Justice is good’, ‘injustice is bad’ In case (1), the propositions are the same apophansis: both are true; they are the same logos (they say the same thing); they are the same ‘in relation to the substrate’ because they have the same subject (namely, justice); but they do not amount to the same premise because the one is affirmative and the other negative. In case (2), Alexander seems to want to say (though I am unsure that I want to follow him) that the propositions are the same apophansis because they

Are Linguistic Objects Fiat or Bona Fide?

191

are both true; they are also both affirmative; but here the accent is on their not being the same protasis because their substrates (which is to say their subjects) and their predicates are different; and we might add, that they say different things (each has a logos of its own). There are two reasons why Alexander’s comment on the passage from the Prior Analytics seems to me to have a certain weight. One is that he employs a distinction that would be resonant in modern logic, namely that between the meaning or sense of a proposition and its reference. The other is the canniness with which Alexander handles the difficulties built into a language that seeks to be scientific and responsive to the formalization of knowledge, but that is stuck in a model of the inner structure of propositions that raises more problems than it solves. We may again recall Frege’s criticisms of the subject–copula–predicate form, which leads to logically different treatments of propositions that have the same meaning, as in the pair presented in case (1).

Conclusions The aim of this chapter has been to pinpoint in Aristotle and in Alexander of Aphrodisias the origin, the development and the problems related to the distinction between bona fide and fiat objects as regards language. An initial question that we raised is that of the double nature of natural language, considered as phonic material that is nevertheless articulated in conventional sequences that are determined by human beings. A second question considered regards the crucial role of thought in the constitution of language and reality. These questions are fundamental to Aristotle’s theory of language and underlie the treatment of linguistic facts in terms of bona fide and fiat objects, and they led Alexander to develop arguments in favour of Aristotle’s own theory, which was presented in so succinct a way as to be rather obscure. Alexander’s first argument makes it clear once and for all that human language is not natural, even though it it is made up of sounds that exist by nature. The second argument brings out the sense of Aristotle’s debatable claim that words are not signs of reality but of thoughts about it. The treatment of the isomorphisms among language, thought and reality in relation to the assertoric propositions turns out to be more problematic. Despite the clearly realist starting point of the Peripatetic view, according to which reality is a guarantor of the

192

Natural and Artifactual Objects in Contemporary Metaphysics

truth (or falsity) of assertoric propositions, the creative power of language, both over language and, to some extent, over reality, emerges clearly both from the observations of Aristotle and from its developments in Alexander. We may add to this that the treatment of propositions is fixed by fiat in different, and perhaps even irreconcilable, ways according to the logico–inferential roles they play or how they ‘reflect reality’. It is worth stressing the importance that, in their very foundations, these theories concede to cognitive activity. Indeed, language is regarded as the medium through which thoughts are fixed and communicated and, despite the efforts of the Peripatetics to demonstrate the opposite, by which reality is moulded.

Notes 1 Smith (1994, pp. 15–23); we refer to the pagination of the expanded online version available at ontology.buffalo.edu (pp. 1–31). 2 See Chaniotis (2004a, pp. 388–9); (2004b, pp. 79–81). 3 Here Smith (1994, p. 5) refers, among other sources, to Kenny 1975. 4 Smith 1994, pp. 11–12. Language ‘articulates reality’ not only because it groups things together but also because it brings to salience and discriminates reality in the terms set by our concepts. 5 Smith 1994, pp. 13–14. 6 Aristotele, De interpretatione 16a3–8. 7 The great scholar of ancient philosophy Pierre Aubenque has recently returned to analyse this passage and the De Interpretatione as a whole: (2009, pp. 37–50). See also J. Barnes’ important article (2012, pp. 582–601), especially pp. 584–90. 8 See Aristotle, De Interpretatione, 16a10 (henceforth de int). 9 Indeed, his commentary on the De Interpretatione has been lost, but we have fragments of it in the commentaries by Ammonius and Boethius on the Aristotelian text. See A. Busse (ed.), Ammonius in Aristotelis De interpretatione Commentarius (CAG IV.5), Berlin, Reimer, 1896 (henceforth in de int); for Boethius, see C. Meiser (ed), Anicii Manlii Severini Boeti Commentarii il librum Aristotelis PERI ERMENEIAS, editio secunda, Leipzig, Teubner, 1880 (henceforth in de int, ed. sec.). 10 For language ‘fixes thoughts’, see Aristotle, de int, 16b20. 11 On these matters and on Boethius’s position, which is our main source for Alexander’s views, see Marenbon (2003, pp. 34–5). 12 See de int, 16a26–29, where it is said that names exist by convention: ‘For sure, inarticulate sounds, such as those of the beasts, mean something, but none of them

Are Linguistic Objects Fiat or Bona Fide?

13 14 15

16 17 18 19 20

21 22

23 24

25 26

193

is a name,’ likewise for verbs (16b6–7) and, consequently, also for propositions (16b33–17a2). See, for instance, Ammonius, in de int, 19, 3–4. See, again, Aristotle, de int., 16b20. As Todd observes, the argument is not a mere aporia, but perhaps has Stoic roots; see R. B. Todd, ‘Alexander of Aphrodisias on De interpretatione 16a26–29’, Hermes, 104, 1976, pp. 140–6, n. 2. For a theory of language as natural, see also the first part of Plato’s Cratylus. Ammonius, in de int, 39, 13–17. On this, see Todd, ibidem, pp. 142–5. Ammonius, in de int, 39, 19: πρὸς ταῦτα τοίνυν ῥητέον ὅτι τὰ ὀνόματα καὶ τὰ ῥήματα οὐχ ἁπλῶς ἂν λέγοιντο εἶναι φωναί, ἀλλὰ κατὰ τὴν ὕλην. Ammonius, in de int, 39, 20–22. Aristotle, Physics, II, i, 193a8–15. See also Physics, II, ii, 194a16–19 (= Metaphysics Delta, 1013b17–21), where it is said that letters are the material cause of syllables. On this view and on Alexander’s comment on it, see Bonelli (2009, pp. 1–17), especially pp. 4–7. Alexander, Quaestio 3. 11, in I. Bruns (ed.), Alexander Aphrodisiensis. Praeter commentaria scripta minora, (CAG suppl. II 2), Reimer, Berlin 1892. Boethius very frequently cites Alexander’s lost commentary, often to criticize it in favour of Porphyry, who was himself the author of a lost commentary on De Interpretatione, of which Boethius makes abundant and explicit use. The fact that Boethius generally cites Alexander’s commentary along with Porphyry’s criticisms of it has led some scholars to suggest that Boethius was not quoting directly from Alexander but only by way of Porphyry’s commentary. In this direction, see Shiel (1990, p. 358). Meaning obviously articulated vocal sounds, which is to say the words expressed. Boethius, in de int, ed. sec., p. 40, 28–41, 11: sed quoniam ita dixit Aristoteles: quorum autem haec primorum notae, eadem omnibus passiones animae sunt, quaerit Alexander: si rerum nomina sunt, quid causae est ut primorum intellectuum notas esse voces diceret Aristoteles? rei enim ponitur nomen, ut cum dicimus homo significamus quidem intellectum, rei tamen nomen est id est animalis rationalis mortalis. cur ergo non primarum magis rerum notae sint voces quibus ponuntur potius quam intellectuum? sed fortasse quidem ob hoc dictum est, inquit, quod licet voces rerum nomina sint, tamen non idcirco utimur vocibus, ut res significemus, sed ut eas quae ex rebus nobis innatae sunt animae passiones. On the commentary to this passage, see also Bonelli (pp. 64–7). The employment of these two examples is justified by the fact that, for Aristotle, things (πράγματα) can be either individual or universal; see de int, 17a38–b1. On the importance that Aristotle attributes to language in the construction of our scientific (and hence cognitive) understanding, see Mario Vegetti’s chapter on

194

27 28

29 30

31 32 33 34

35

36

Natural and Artifactual Objects in Contemporary Metaphysics seeing and hearing the world in Vegetti and Ademollo (2016, pp. 47–57), as well as Francesco Ademollo’s on logic and epistemology in the same volume, pp. 59–83. Barnes (2012, p. 590). Not even lies are counter-instances to Alexander’s theory, insofar as when I pronounce a sentence, I signify a thought and not a belief, in the sense of a proposition in which I believe. Alexander’s account does not cover all cases, however, for it can happen that I produce the sound of something without thinking about it, as when I use the name of one thing while thinking about another: I say ‘Socrates’ but was thinking of Plato, so I add ‘I meant to say “Plato”’. See Ademollo, op. cit., p. 61. At de int, 24b1–2 Aristotle is explicit: ‘If this is the case with judgement, and the voiced (en te phonè) affirmations and negation are symbols of those mental judgements, it is clear…’. Barnes (2012, pp. 587–90). Barnes (2012, p. 589). We have already seen the criticisms of this theory in the preceding section. Aristotle, de int., 17a2–3: ἀποφαντικὸς δὲ οὐ πᾶς, ἀλλ᾽ ἐν ᾧ τὸ ἀληθεύειν ἢ ψεύδεσθαι ὑπάρχει. See for instance Aristotle, Metaphysics IV, vi, 1027b17–1028a6. Alexander accepts the view without reservations; see M. Hayduck (ed), Alexandri Aphrodisiensis in Aristotelem Metaphysica Commentaria (CAG I), Berlin, Reimer 1881, 148, 15–16, where it is clearly stated that truth does not reside in things but in thought. In point of fact, the matter is more complicated, as Paolo Crivelli shows well in Aristotle on Truth, Cambridge, Cambridge University Press 2004. See also Cavini (2011, pp. 371–402). To be a little more precise, we should say that we unite or separate terms that signify our thoughts, which would therefore present themselves first in the form of separate concepts that are not themselves subject to truth and falsehood, and only then in a propositional form (as a judgement) that is subject to truth and falsehood. These thoughts are in turn images of reality. Aristotle shows that he has a theory of this sort when he observes that ‘verbs by themselves are nouns and they signify something, indeed, a speaker fixes thought’ (de int, 16b20), or again when he claims that ‘as sometimes there are thoughts in our souls without being true or false, while there are others that are necessarily one or the other, so it is also with speech, for the true and the false concern combination and division. Thus names and verbs resemble (?) thoughts without combination or division’ (de int, 16a9–14). See again de int 24b1–2, quoted above (n. 30). We may think of Frege’s famous critique of Aristotle’s account of the proposition in terms of subject, predicate and copula, which is of no logical use

Are Linguistic Objects Fiat or Bona Fide?

37 38 39 40 41 42

195

(G. Frege, Begriffsschrift, Halle 1879: English translation as Conceptual Notation, by T.W. Bynum, Oxford 1972, pp. 107, 112, 113). See also Barnes (pp. 147–71), which offers a point-by-point comparison between Aristotle’s and Frege’s grammars. Aristotle, Prior Analytics, 24a16–17. See note 33. The one is affirmative and other negative. Wallies (ed) (1883, 10, pp. 13–28). See Hayduck (ed) (1881, pp. 247, 8–32). The choice of ‘utterance’, to stress that we have to do with something that has been actually expressed, is due to Barnes et alii, (tr.) (1991, p. 55).

References Aubenque, P. (2009), ‘Sens et unité du traité aristotélicien De l’interprétation’, in S. Husson (ed.), Interpréter le De interpretatione, Vrin, Paris, 37–50. Barnes, J. et alii, (tr.), (1991), Alexander of Aphrodisias on Prior Analytics 1.1-7, Duckworth, London, 55. Barnes, J. (2012), ‘Meaning, Saying, and Thinking’, in M. Bonelli (ed.), Jonathan Barnes: Logical Matters, Essays in Ancient Philosophy II, Oxford University Press, Oxford, 582–601. Barnes, J. (2012), ‘Grammar on Aristotle’s Terms’, in M. Bonelli (ed.), Jonathan Barnes: Logical Matters, Essays in Ancient Philosophy II, Oxford University Press, Oxford, 147–71. Bonelli, M. (2009), ‘Alexandre d’Aphrodise et la cause matérielle’, Journal of Ancient Philosophy, III: 1–17. Bonelli, M. (2009), ‘Alexandre d’Aphrodise et le De interpretatione’, in S. Husson (ed.), 64–7. Cavini, W. (2011), ‘Vero e falso nelle Categorie’, in M. Bonelli and F. Masi (eds), Studi sulle Categorie di Aristotele, A.M. Hakkert, Amsterdam, 371–402. Chaniotis, A. (2004a), ‘New Inscriptions from Aphrodisias’, American Journal of Archaeology, 108: 388–9. Chaniotis, A. (2004b), ‘Epigraphic Evidence for the Philosopher Alexander of Aphrodisias,’ Bulletin of the Institute of Classical Studies, 47: 79–81. Hayduck, M. (ed.), (1881), Alexandri Aphrodisiensis in Aristotelem Metaphysica Commentaria (CAG I), Reimer, Berlin, 247, 8–32. Kenny, A. (1975), Will, Freedom and Power, Oxford. Marenbon, J. (2003), Boethius, Oxford University Press, Oxford, 34–5. Shiel, J. (1990), ‘Boethius’ Commentaries on Aristotle’, in R. Sorabji (ed.), Aristotle Transformed, Cornell University Press, Ithaca-New York, 358.

196

Natural and Artifactual Objects in Contemporary Metaphysics

Smith, B. (1994), ‘Fiat Objects’, in N. Guarino, S. Pribbenow and L. Vieu (eds), Parts and Wholes: Conceptual Part-Whole Relations and Formal Mereology, Proceedings of the ECAI94 Workshop, Amsterdam: ECCAI, 15–23. Vegetti, M. and Ademollo, F. (2016), Incontro con Aristotele. Quindici lezioni, Einaudi, Turin, 47–57. Wallies, M. (ed.), (1883), Alexandri in Aristotelis Analyticorum priorum librum I commentarium (CAG II.1), Reimer, Berlin, 10, 13–28.

Part Four

What Does Mind-Dependency Depend On?

11

Leibniz’s Principle and Psycho-Neural Identity Andrea Bottani and Alfredo Paternoster

Leibniz’s principle as an epistemic criterion of numerical difference Things are said to be qualitatively identical, or indiscernible, just in case they have exactly the same properties, while they are said to be numerically identical just in case they are no more than one. Leibniz’s law states that qualitative identity and numerical identity are equivalent: a and b can neither be numerically identical while being qualitatively different nor be qualitatively identical while being numerically distinct. Accordingly, Leibniz’s law can be analysed as the conjunction of two distinct principles, currently referred to as the ‘identity of indiscernibles’ (indiscernibility entails numerical identity) and the ‘indiscernibility of identicals’ (numerical identity entails indiscernibility). There are prima facie good reasons to treat the latter principle (henceforth IND/ID) as holding in general, with no restriction to categories, kinds or sorts of entities. The main reason is the intuition that, if an a and a b whatever are numerically identical (i.e. if they are just one entity), a necessary condition for a having properties that b lacks is that some entity c (the particular entity that both a and b are) has properties that it itself lacks. Thus, if IND/ID were not valid for all entities in general – ordinary objects, events, types, universals, tropes, real entities and possibilia, present, past and future things and so forth – some contradictions would then be true. There are ‘non-Leibnizian’ approaches to identity that do not take seriously the above intuition and treat IND/ID as invalid. Examples include dialetheism (Priest 1987/2006; see also Priest and Berto 2013), the theory of occasional identity (Gallois 1990) and the theory of relative identity (Geach 1967). We shall not discuss these approaches here. On the contrary, we shall take seriously the intuition and assume that IND/ID is unrestrictedly valid.

200

Natural and Artifactual Objects in Contemporary Metaphysics

If IND/ID is unrestrictedly valid, its contra-positive is unrestrictedly valid too: if numerical identity entails qualitative identity, qualitative difference entails numerical distinction. Whenever a and b are demonstrably different in qualities, then, they are also demonstrably different in number. Since qualitative difference entails numerical difference, qualitative difference can in principle be used as a reliable criterion of numerical difference. This is why qualitative difference has often played an important philosophical role as an antidote against reductionism about many kinds of entities. Whenever some Fs are tentatively reduced to some Gs, but there is some property that demonstrably the Fs have but the Gs lack (or vice versa), the tentative reduction is demonstrably flawed. It has been claimed, for example, that a tree is not (identical with) a collection of molecules because it can survive a loss and gain of molecules, while the collection cannot (Wiggins 2001); that pain is not (identical with) a stimulation of C-fibres because the stimulation has a frequency of firing while the pain has none (at best, the pain has an intensity that correlates with such a frequency); that persons are not human animals because animals have parts while persons have none (Lowe 2000) and so forth. Unfortunately, there are several problems with using qualitative difference as an epistemic criterion of numerical difference. 1) Sometimes, as Dummett has shown, the order of epistemic priority is indeed reversed, for in some cases the only way of discovering that an a has a property P is to discover that a is identical with a particular b which is already known to have P. (For example, the only way of coming to know that in the evening the morning star is located in a particular region of the sky consists in discovering that the morning star is identical with the evening star.) Similarly, in some cases, one can discover that a lacks a property P only by discovering that a is different from a particular b which is already known to have P. For instance, the only way we can discover that the adult Alfredo lacks the property of being less than one meter tall in 1962 is by discovering that he is not identical to the child he was, but only a different temporal slice of the same four-dimensional worm; the only way we can discover that the morning star is not in that position in the evening consists in discovering that the morning star is different from the evening star, and perhaps the only way of discovering that pains have no frequency of firing consists in discovering that pains are not stimulations of C-fibres.1 2) Sometimes, differently than in the previous cases, we may have definite evidence that both ‘… Pa …’ and ‘… Pb …’ are true, but still doubt whether

Leibniz’s Principle and Psycho-Neural Identity

201

there are two distinct things called a and b, one having and the other lacking a property expressed by the predicate P. This is the case exemplified by sentences such as ‘9 is necessarily greater than 7 but the number of planets isn’t’ in opaque reading. This sentence is manifestly true, but one cannot infer from it that there are two distinct things: the number 9 on the one hand, and the number of planets on the other, even though the property ‘being necessarily greater than 7’ is predicated of the former but not of the latter. (Note that in transparent, or de re, interpretation of the sentence is false – as Kripke contends – or perhaps meaningless, as Quine is inclined to say.) A similar example is, ‘The statue has always been of that shape but the clay that constitutes it has not’, which does not allow one to conclude that the statue is not the clay. The same point can be stated from a slightly different point of view: ‘being necessarily greater than 7’ is an intensional predicate; hence, it applies or does not apply to an object depending on how the object is described. Therefore, one must be careful, in applying Leibniz’s law, not to take for granted that there are two distinct objects because of the occurrence of an intensional predicate. The relevant predicates occurring in Leibniz’s law must be only the extensional ones. Note that this can be a prima facie reason for being sceptical as regards the application of Leibniz’s principle to psycho-neural identities: several authors follow Davidson in taking mental predicates as intensional predicates. We will come back to this point later. 3) In other cases, it is not clear what properties are expressed by a predicate in one or more of its occurrences. Consider ‘The old Beethoven was deaf but the young Beethoven wasn’t’, which does not entail that the old Beethoven and the young Beethoven are two different entities unless it is assumed that the predicate expresses the same property in both conjuncts (say, the property of being deaf simpliciter in both conjuncts, rather than the property of being deaf in one’s old age in the affirmative conjunct and the property of being deaf in one’s youth in the negative conjunct). Problems of this kind prevent one to use the diversity of discernibles as an easy key to solve well-known problems of identity across time, or across possible worlds. In sum, discernibility can be used as an ‘indicator’ of numerical identity just in case it is both semantically transparent and epistemically decidable, or at least more decidable than identity.

202

Natural and Artifactual Objects in Contemporary Metaphysics

Can Leibniz’s criterion be used as an argument against the identity theory? In the philosophy of the mind, the diversity of discernibles has been extensively used as a tool for a purported reductio ad absurdum of the psycho-neural identity theory, the idea that mental states or events are nothing else than neurophysiological states or events. To what extent can the principle be successfully employed to argue against reductionism about the mind? Note, first of all, that we are not in a position to apply the principle in real cases, since no one has yet discovered a brain process (or event) that could be regarded as a plausible candidate to be identified with a mental process (or event). Indeed psycho-neural identity is (at least until today) a largely programmatic thesis. Nevertheless, it is worth discussing in principle whether the application of Leibniz’s law to a pair constituted by a typical mental event and a hypothetical neural event (taken as a good exemplar of the class of neural events candidate to the identification) implies the failure of psycho-neural identities, as a consequence of the apparent discernibility of these two events. If this were the case, we would have an a priori reason to reject identities of this kind. Prima facie, the application of the principle seems actually to undermine the identity thesis: there seem to be some extensional properties that can be predicated of a neural event but can hardly be predicated of a mental event and vice versa. For instance, a given pain has the property of being sharp, but a stimulation of C-fibres cannot have the property of being sharp, or so it seems at any rate. Or, the other way around, the activation of a neural pattern has the property of being spatially located whereas a mental state whatsoever can hardly have such a property. In general, it seems plausible to say that phenomenal properties cannot be predicated of neural processes. However, as Place (2004) has persuasively argued, this proves nothing, since the fact that something is not known to have a given property is no evidence that it does not have that property. According to Place, discarding the possibility that an experience has a non-experiential property is a ‘phenomenological fallacy’, that is, involves the mistaken assumption that one’s introspective observations report ‘the actual state of affairs in some mysterious internal environment’. Thus, for instance, pain has actually the property of being caused by some internal lesion whose nature and existence is yet entirely unknown to the sufferer. To this, one could reply that, while there is (perhaps) room to claim that phenomenal states can have ‘hidden’ or surprising properties, it is far less

Leibniz’s Principle and Psycho-Neural Identity

203

plausible to say that neural states can have phenomenal properties. Even if far less plausible does not mean that we are in a position to discard definitely this possibility, the burden of proof seems to be on identity theorists. However, there is another main problem faced by those who want to exploit Leibniz’s principle against the identity theory, namely, the individuation of the relevant properties. We argue that individuating the relevant properties is not less difficult than individuating the relevant events. As we saw earlier (see point 3), we can be in doubt as to what property is exactly specified by a given predicate. In particular, even when we share intuitions on what properties are expressed by two predicates, we can still have doubts as to whether the property expressed by one predicate is identical to or different from that expressed by another. Take again the case of pains, C-fibre stimulations, the property of intensity that pains have and the property of firing at a certain frequency that C-fibre stimulations have. On the anti-reductionist’s stance, intensity is a determinable property of pains, but the frequency of firing is not, so pains are not C-fibre stimulations. But this can work only on the assumption that intensity and frequency of firing are different determinable properties, which seems to require an independent argument. This casts light on a very important point. Using discernibility (i.e. difference of properties) as a criterion of numerical difference for particulars requires a criterion of numerical difference for their properties. (More generally, using discernibility as a criterion of numerical difference for entities of type n requires a criterion of numerical difference for entities of type n+1.) But no criterion of numerical difference for properties can be neutral with regard to what sort of entity a property is assumed to be. If, for example, properties are assumed to be resemblance sets, lack of co-extensionality is a necessary condition of difference, whereas it is not so if they are assumed to be universals, for two different universals can accidentally be instantiated by the same objects. In both cases, the very idea of discernibility as a general criterion of numerical difference yields problems. For, if properties are resemblance sets, properties are different just in case they are possessed by different individuals, which makes it circular to assume that individuals are numerically different just in case they have different properties. By contrast, if properties are universals or the like, they may be said to be different just in case they have some different higher-level properties, for example, causal efficacy or categorical nature, which potentially yields an infinite regress. This is not to say that discernibility can never be used as a test of numerical difference of course. But certainly it can be used this way only provided there is some evidence about the identity of the involved properties, which is far from being granted.

204

Natural and Artifactual Objects in Contemporary Metaphysics

Therefore, our first conclusion is that Leibniz’s law cannot by itself provide a definite reason to reject the identity theory. At most, one could argue that it is the reductionist who should provide some reasons to justify that the relevant mental property is one and the same with the relevant neural property.2 Let us stress now that there is another extremely interesting consequence of Leibniz’s law for the identity theory.

Why token identity and type identity are in the same boat It is commonly believed that the psycho-neural identity theory can take two forms, depending on whether psycho-neural identity is interpreted as an identity of tokens or instead of types. It is also commonly believed that the latter version of the theory is stronger than the former, since the identity of types entails the identity of tokens, but not vice versa. Yet IND/ID invites scepticism about that. Suppose that a neural token n is numerically identical with a mental token m and suppose that n belongs to a type N. Then, by IND/ID, m belongs to N too. (After all, n and m are just one entity!) This seems to suggest that the identity of tokens entails the identity of types just as the reverse, which means that there is no distinction between the token identity version and the type identity version of the psycho-neural identity theory. Indeed, suppose that the token identity theory holds, and the set E(N) contains all and only the neural events of type N (so that E(N) is the extension of N). Suppose that n is an element of E(N), and n is identical with a mental event m (say, a certain toothache). Then, m must be an element of E(N), for m and n are no more than one thing, and n belongs to E(N). Therefore, n and m are one and the same element of a certain neural type: if n is an element of the neural type N, then m is of that type too. The same must be said, mutatis mutandis, of the mental types to which m belongs; that is, if m = n, then n is an element of the set E(M), where M is a mental type to which m belongs.3 So, if there are mental as well as neural types and the neural–mental token identity holds, mental types and neural types are not distinguished according to whether their tokens are mental or neural, and one can wonder what makes a type neural instead of mental (or vice versa) in such a case. The natural, intuitive answer is that a type T is mental just in case it corresponds to the extension of some fundamental predicate of naive psychology (i.e. there is some predicate P of naive psychology such that something is a token of T just in case it belongs to the

Leibniz’s Principle and Psycho-Neural Identity

205

extension of P), while it is neural just in case it corresponds to the extension of some fundamental predicate of neuroscience. Along these lines, the dichotomy of mental and neural types can be established on a semantic rather than an ontological basis (while it cannot fail to be also ontological if mental and neural tokens are distinct). Establishing the dichotomy on a semantic basis, however, does not necessarily deprive it of any ontological value. Take the distinction between sortal and non-sortal properties. It seems to be established on a semantic basis, grounded as it is on the different semantic behaviour of sortal and non-sortal predicates. But this does not prevent a number of philosophers from treating it as ontologically fundamental. It is not uncommon, after all, to assume that semantic differences reflect ontological differences. Usually, the distinction between type identity and token identity, together with an endorsement only for the latter, is motivated by the argument from multiple realization. The intuition underlying multiple realization is that two mental events m1 and m2, both of type M, need not be respectively identical to two neural events n1 and n2 belonging to one and the same type N. In fact, it could be the case that, while m1 is identical to n1 (therefore m1 belongs to N), m2 is identical to a neural event n* belonging to the type N* (different from N). Thus m1 and m2 turn out to be respectively of two different neural types N and N*. For instance, let m1 be X’s pain at t1 and m2 X’s hamster’s pain at t2. m1 and m2 are both instances of one and the same type, the kind ‘pain’. Now, ex hypothesi, m1 is identical to a neural event n1 and m2 is identical to a neural event n2. But why should we be forced to claim that n1 and n2 belong to one and the same kind N? It could be the case that n1 is X’s C-fibre stimulation at t1 whereas n2 is the hamster’s CH-fibre (different from C-fibre) stimulation at t2. If this were the case, while n1 is an instance of a kind N, n2 is an instance of a kind N*, so m1 and m2 are also members of two distinct classes of neural events. Let us make the same point in a slightly different way, in order to be further more explicit. Suppose, as we did above, that m1 (= n1) is an event belonging to the class N, while m2 (=  n2) belongs to N*. What prevents us from saying that there are two mental types M and M* such that M  =  N and M*  =  N*, hence concluding that type identity is true after all? The answer is that m1 and m2 are both instances of pain, and there is not a unique neural kind candidate to be identified with pain. The kind ‘pain’ is orthogonal to N and N*: it has only a subset in common with each of them. A familiar way to avoid the argument from multiple realization (hence, to assimilate token identity to type identity) is the following: from a reductionist

206

Natural and Artifactual Objects in Contemporary Metaphysics

point of view, it makes little sense putting together two mental events (such as two pain events) in a same kind if there were not a (at least) similar neural basis. As the above example shows, clear cases of token identity without type identity usually take place when we individuate mental events in a coarse way. However, if our aim is finding psycho-neural correlations, we should rather cut the mental much more finely, for instance, by relativizing mental kinds to biological species: pain in human beings, pain in cats, pain in hamsters and so forth. (It is just an example, and there are several other possibilities.) If mental types are properly treated as fine-grained entities, mental tokens are far more easily identified with neural tokens of a same type, and we will likely find that, as a consequence of Leibniz’s law, token identity entails type identity. Hence (given that type identity also entails token identity), there is no difference between token identity and type identity. A slightly different way to make the same point hinges on the relation between determinable and determinate properties.4 An ideally adequate neurophysiological theory should be in a position to recognize that Nʹ and N* are different determinates of one and the same determinable N  (=  M), and conversely an ideally adequate psychological theory should be in a position to recognize that M is a determinable, of which Mʹ  (=  N) and M*  (=  N*) are two different determinates (just as carmine and scarlet are two different determinates of red). Multiple realization might simply be multiple determination.5 And the apparent non-coincidence between mental and neural types might be simply produced by the mistake of comparing neural determinates with mental determinables. The distinction between determinables and determinates also explains why in many contexts (e.g. when some functional law is at stake) it may turn out useful to consider mental and neural kinds in a coarse rather than fine-grained way. (What typically figures in functional laws are indeed maximally general determinables like mass, speed, momentum, etc.) However, there may be reasons other than multiple realization to embrace a token identity cum type diversity theory of the mental–neural relations. Two languages may indeed be such that their quantifiers range on one and the same domain, yet no extension of some predicate in one language is a (proper or improper) subset of the extension of some predicate in the other language. Suppose that we classify humans according to (1) their income class, (2) their height, (3) their race, (4) their disposition to develop cancer, (5) their age and (6) their personality. If each of these ways to develop a taxonomy of human

Leibniz’s Principle and Psycho-Neural Identity

207

beings can ground general laws concerning humans (e.g. functional laws that connect the variation of certain determinable properties with the variation of other determinable properties), none of them can vindicate the ownership of the nature of the classified entities. If, on the contrary, some of the above kinds of purported theories can ground no nomologically interesting generalization, its kinds should perhaps be dismissed, that is, completely eliminated. In any case, what seems to be clear is that no finer-grained way to articulate the different taxonomies, or some of them, could make one coincide with another. For no category in one taxonomy contains either the same entities as some category in another taxonomy or a proper subset of them. So no item of one taxonomy can be treated either as a determinate or a determinable of some item in another taxonomy, for the extension of a determinate is necessarily a subset of the extensions of all its determinables. For instance, the set of people nineteen years old is necessarily a subset of all people. Is not this a case of token identity cum type diversity? It depends. First, the various theories do not quantify over the same domains, but (partially) quantify over different tokens. (Some quantify over institutions and currencies, others over cells and viruses, still others over temporal intervals, over genetic groups, over human behaviours, etc.) Perhaps, the various different domains partly overlap, but there is much in each of them that there is not in the others. Relatedly, there is no chance to reduce one taxonomy to some of the others – no chance to say what is said about human incomes in terms of human heights, or human personalities, for example – nor the reverse. This is not surprising, because they quantify over entities of different kinds, and classify what they call ‘human beings’ according to their relations with entities of different kinds. One can wonder whether the different domains of quantification all include at least human beings or not. If they do not, there is no token identity (no overlap between the items classified by one taxonomy and those classified by each of the others). If they do, either what one theory calls ‘human beings’ has the same properties of what the other theories call the same way or Leibniz’s law is false. Therefore, in this case, provided that Leibniz’s law is true, token identity entails type identity. In sum, as far as we can see, there is after all a distinction between type and token identity, but this distinction has no ontological import. Note that nothing in our argument for metaphysical equivalence between token and type identity proves or suggests that the psycho-neural (type) identity is true. What we have argued for is just that Leibniz’s principle entails the metaphysical equivalence

208

Natural and Artifactual Objects in Contemporary Metaphysics

of token identity and type identity. This, in turn, shows that the token identity theory (as it is usually assessed) is not an available option for the supporter of the so-called non-reductive physicalism, that is, for the view according to which all particular things are physical, but there are non-physical properties (or, at least, there are properties that cannot be reduced to physical properties). Antireductionists should rather ground their position on weaker ontological views, such as supervenience.6 To sum up, it is not clear that IND/ID can say a definite word on the tenability of the psycho-neural identity theory. But it seems likely that IND/ID makes very hard to distinguish, in point of metaphysics, between two different versions of the psycho-neural identity theory, the weaker ‘token identity’ theory and the stronger ‘type identity’ theory.

Arguments for identity If Leibniz’s principle can be used neither in favour of nor against type identity, is there another compelling argument? Who would bear the burden of proof? Let us start by assuming, arguably with the majority of philosophers, that the burden of proof is upon identity theorists. The idea is that one has to provide reasons for identity rather than merely coping with objections. There are three kinds of arguments (cf., for example, Kim 2005), each corresponding to a slightly different way to conceive of psycho-neural reduction. Each of them assumes that there are some empirically verified psycho-neural correlations (though in the current state of the research it is not so trivial to find them), since, of course, absent these correlations, we would not have any reason to believe in the psycho-neural identity. Under this assumption, identity is a way of explaining why there are these correlations, for instance, why C-fibre stimulation is felt as pain rather than as a tickle. As Kim points out, giving an answer to this why question is what allows us to move from mere inductive predictions (only based on repeated observations) to theoretical predictions. Let us shortly survey the general structure of the arguments for identity.

(i) Argument from simplicity (Smart 1959) Identities must be true because they are the easiest way to account for psychoneural correlations. Indeed alternatives to physicalism, such as emergentism or

Leibniz’s Principle and Psycho-Neural Identity

209

epiphenomenalism, are obscure. It is worth noting that, in contrast to what is usually found today, identities were regarded by Smart as being contingent and a posteriori. However, this argument seems to assume what its opponents (explicitly) deny. In fact the argument assumes that the physicalist ontology is the most elegant and parsimonious one, but a dualist could complain that such ontology does not save the relevant phenomena. The question is just what phenomena must be saved in an overall ontology of the world.

(ii) Argument from (inference to best) explanation (Block and Stalnaker 1999)7 Here identity, which is regarded as a necessary relation, is considered as being the best explanation of psycho-neural correlations. Since the relevant identities are metaphysically necessary, they can account for the nomological character of correlations (insofar as metaphysical necessity entails nomological necessity). The model of identity that Block and Stalnaker have in their minds is provided by the theoretical identities established between macrophysics (thermodynamics) and microphysics (statistical mechanics). The point is that, taken together, these identities allow for explaining a collection of observational facts by establishing a reduction between theories; that is, certain causal relationships between macro properties hold because a corresponding set of relations between micro properties holds. This view is effectively presented in this piece from Block (2010, pp. 121–2): If we want to know why water = H2O, freezing = molecular lattice formation, heat = molecular kinetic energy, temperature = mean molecular kinetic energy, etc., we have to start with the fact that water, temperature, heat, freezing and other magnitudes form a family of causally inter-related ‘macro’ properties. This macro family is mirrored by a family of ‘micro’ properties: H2O, mean molecular kinetic energy, molecular kinetic energy and formation of a lattice of H2O molecules. … The key fact is that the causal and explanatory relations among the macro properties can be explained if we suppose that the following relations hold between the families: that water = H2O, temperature = mean molecular kinetic energy, heat = molecular kinetic energy and freezing = lattice formation. For example, why does decreasing the temperature of water cause it to freeze? … Here is a sketch of the explanation: The oxygen atom in the H2O molecule has two pairs of unmated electrons, which attract the hydrogen atoms on other H2O molecules. When the kinetic energy of the molecules decreases, (i.e. the temperature decreases) each oxygen atom tends to attract two hydrogen atoms

210

Natural and Artifactual Objects in Contemporary Metaphysics

on the ends of two other H2O molecules. When this process is complete, the result is a lattice in which each oxygen atom is attached to four hydrogen atoms.

Therefore the idea is that assuming that identities are true yields overall the most explanatory picture. In other words, the epistemology of theoretical identity is just a special case of inference to the best explanation. Without identities, explanation of macro facts would be unsatisfactory or much more complicated. According to Kim (2005), however, in cases like this the burden of the explanation is at the microphysical level; the identity merely allows for rewriting in ordinary language the microphysical statement. Even assuming, for the sake of the argument, that the analogy with the psycho-neural case is correct (although this cannot straightforwardly be taken for granted), rewrite rules are not capable of generating explanations of their own. To put it shortly, it is not the set of identities that plays the explanatory role; there is reduction without reductive explanation, so to speak. Indeed, according to Block and Stalnaker, identities are not explainable (Block and Stalnaker 1999, p. 24).

(iii) Argument from causal role (Kim 1998; 2005) Here the idea is that we are justified in claiming, for example, that water is H2O for the causal roles of water are realized by its molecular constitution. In the psycho-neural case identity is entailed by the conjunction of the causal closure of the physical world and the causal efficacy of mental states. Identities are justified by functional analysis. For instance, M1 (say, pain) causes M2 (distress) because P1: N1 causes N2 P2: M1 = N1; M2 = N2 M1 = N1 is justified by the fact that M1 has the causal role F1, and N1 realizes the role F1. The same goes for M2 = N2. The crucial point of this strategy is functionalization: in order to establish an identity, one must in advance provide a causal–functional role for a given mental property – if something is a mental state then it plays a certain causal role. The next step is finding the physical property that realizes that role. Therefore, if one is unable to provide a functional role for a mental state, there is no chance for reducing that mental state to a cerebral state. (Incidentally, according to Kim this is the case of qualitative properties, such as pain, which have a non-functional aspect.) It is worth pointing out that Kim’s proposal works only provided it

Leibniz’s Principle and Psycho-Neural Identity

211

is a given fine-grained individuation of mental properties, as we saw earlier discussing multiple realization. Indeed, if one cuts coarsely the mental, there will be a collection of causal roles, not only one. The force of this argument is that the link between the mental property and the physical property is conceptually warranted: it is an a priori truth that a mental property is defined by a certain causal role, while it is an empirical matter what physical property realizes that role. Only functionalization allows assimilation of the psycho-neural case to cases such as water = H2O. Yet, according to Block (2010), functional reduction of this sort is metaphysically incoherent since, from a metaphysical point of view, a functionalist account cannot be a physicalist account. Consider, for instance, two realizations of a phenomenal or qualitative property (i.e. take two phenomenal events that are very similar): Qme and Qyou. Even if both realizations are themselves physical, if the similarity between Qme and Qyou can be accounted for only on the basis of the functional similarity (albeit physically realized), then the real metaphysical account is functionalist, not physicalist. In other words, Block contends that Kim cannot have it both ways: (reductive) physicalism and functionalism about mental properties. A metaphysical physicalist must have a physical account of the shared phenomenal property. Let us try to take stock. All the proposed justifications of identity face some difficulties. To put it briefly, identity seems not adequately justified in the argument from explanation, while the account based on the causal role, even discounting Block’s charge of metaphysical incoherence, requires revision of mental predicates and works only for states that can be characterized as causal roles. Since the nature of many (perhaps most) mental states seems not to be exhausted by functional role, Kim’s proposal falls short. It seems to us that the extensional psycho-neural identity can be properly justified only by an inter-theoretical reduction, as showed in Block’s example of thermodynamics. This could seem to be circular: identity requires reduction and reduction requires identity. However, we could reply invoking a sort of reflective equilibrium: if we have a psychological theory (or a piece of a theory) that, taking a few identities as hypotheses, can be reduced to a neurological (piece of) theory, then we can accept the hypothesized identities and hopefully try to draw further ones. Yet, if this picture sounds plausible, chances for identity and reduction turn out to be meagre, since folk psychology is hardly reducible to any neurological theory, for familiar reasons such as externalism, holism, normativity. As to qualitative properties, both Block and Kim do not believe they are reducible,

212

Natural and Artifactual Objects in Contemporary Metaphysics

for different reasons. According to Kim qualia are not susceptible to functional analysis; according to Block reduction is in principle possible but we lack evidence for correlations. Perhaps, ‘if superscientists of the future were to tell us what consciousness is, we probably would not have the conceptual machinery to understand, just as the pre-Socratic would not have the conceptual machinery to understand that heat is a kind of motion or that light is a kind of vibration’ (Block 2012, p. 1114). This claim is of course compatible with (and perhaps to a certain extent suggests) psycho-neural identity, but, at the same time, it shows that, as far as we can currently tell, we have no reason to believe in psycho-neural identity. Perhaps physicalism does not require identity: it could be enough that physical states necessitate mental states; that is, the physical necessarily entails (a posteriori) the mental, but not the other way round. (This amounts to saying that the mental supervenes in a strong sense to the physical.) Yet, this is another issue. In this chapter we are interested in identity, not in what is required for physicalism.

Conclusions Leibniz’s law cannot be used as an argument against the identity theory, even if it suggests that there is no metaphysical difference between token identity and type identity. On the other hand, the ontological equivalence of token identity and type identity does not make better the perspectives for psycho-neural identity. We need an argument for identity, but none of the arguments explored here appears to be conclusive. At most we can say that psycho-neural identity is a possibility that, in the current state of the research, we are not able to grasp properly. The equivalence between token identity and type identity implies the failure of non-reductive physicalism, but it has yet to be proven that (forcefully reductive) physicalism is required for the scientific image of the world.8

Notes 1 Though one could arguably say that the thesis that pains have not a firing frequency can more easily be inferred from their experiential (phenomenal) nature rather than from the assumption that pains are not C-fibre stimulations. More on this point later.

Leibniz’s Principle and Psycho-Neural Identity

213

2 One reason could be that identities allow one to provide better explanations. We will come back to this point in the last section. 3 Provided that there are mental types at all, that is, something more than gerrymandered sets of incongruous elements, having no sparse properties in common (cf. infra, p. xx). 4 Johnson (1921, vol. 1, ch. XI). 5 We assume this without arguments, even though it is not self-evident. For an argument to the effect that determination is not realization, see Funkhouser (2014). 6 Interestingly, Susan Schneider (2012) has proposed a different argument for the same conclusion: she has argued, on purely metaphysical grounds, that the token identity theory can be true just in case type identity is also true. 7 Arguments based on the explanatory role of identities have been also proposed by Hill (1991) and McLaughlin (2001); see also Hill and McLaughlin (1999). Here we take Block and Stalnaker’s argument as exemplar case. 8 For instance, there is nothing non-scientific in an emergentist view, over and above its vagueness. After all, the most recent developments in philosophy of science go clearly in a non-reductionist direction (cf., for example, Horst 2007): explanation is not generally derivation, and indeed the majority of scientific explanations cannot even be successfully reconstructed as reductions.

References Block, N. (2010), The Canberra Plan Neglects Ground, in T. Horgan, M. Sabates, E. Sosa (eds), Supervenience in Mind, MIT Press, Cambridge, MA, 105–33. Block, N. (2012), Comparing the Major Theories of Consciousness, in M. Gazzaniga (ed.), The Cognitive Neurosciences, IV, MIT Press, Cambridge, MA, 1111–22. Block, N. and Stalnaker, R. (1999), ‘Conceptual Analysis, Dualism and the Explanatory Gap’, The Philosophical Review, 108: 1–46. Funkhouser, E. (2014), The Logical Structure of Kinds, Oxford University Press, Oxford. Gallois, A. (1990), ‘Occasional Identity’, Philosophical Studies, 58 (3): 203–24. Geach, P. (1967), ‘Identity’, Review of Metaphysics, 21 (1): 3–12. Hill, C. (1991), Sensations, Cambridge University Press, Cambridge, MA. Hill, C. and McLaughlin, J. (1999), ‘There are Fewer Things in Philosophy than Are Dreamt in Chalmers’ Philosophy’, Philosophy and Phenomenological Research, 59 (2): 445–54. Horst, G. (2007), Beyond Reduction. Philosophy of Mind and Post-Reductionist Philosophy of Science. Oxford University Press, Oxford. Johnson, W. E. (1921), Logic, 3 vol., Cambridge University Press, Cambridge. Kim, J. (1998), Mind in a Physical World, MIT Press, Cambridge, MA.

214

Natural and Artifactual Objects in Contemporary Metaphysics

Kim, J. (2005), Physicalism, or Something near Enough, Princeton University Press, Princeton. Lowe, E. J. (2000), Introduction to the Philosophy of Mind, Cambridge University Press, Cambridge. McLaughlin, J. (2001), ‘In Defense of New Wave Materialism. A Response to Horgan & Tienson’, in C. Gillett, B. Loewer (eds), Physicalism and Its Discontents, Cambridge University Press, Cambridge, 319–30. Place, U. T. (2004), Identifying the Mind. Selected Papers, Oxford University Press, Oxford. Priest, G. (1987/20062), In Contradiction, Oxford University Press, Oxford. Priest, G. and Berto, F. (2013), ‘Dialetheism’, in E. Zalta (ed.), The Stanford Encylopedia of Philosophy, https://plato.stanford.edu/entries/dialetheism/. Schneider, S. (2012), ‘Nonreductive Physicalism Cannot Accept Token Identity’, Philosophy and Phenomenological Research, 85 (3): 719–28. Smart, J. J. (1959), ‘Sensations and Brain Processes’, The Philosophical Review, 68: 141–56. Wiggins, D. (2001), Sameness and Substance Renewed, Cambridge University Press, Cambridge.

12

Do We Exist? Mereological Nihilism, Collective Thinking and Dualism Alfredo Tomasetta

Are human persons bona fide, objective entities, or are we, in some sense, fiat entities, which result from useful conventional constructions? This question may strike some as preposterous – of course we are not established by conventions! – and yet, there are serious and very able philosophers who would indeed say we are not bona fide objects, the idea being that properly speaking (that is, putting aside useful conventions) we do not really exist. This chapter focuses on a way to defend this strongly revisionary thesis. In particular, the first section deals with the general metaphysical standpoint – mereological nihilism – which is presupposed by the argument against our own bona fide existence. In the second section the argument is presented along with a Cartesian objection to it. Then I consider a prominent reply to this objection, offered by Dorr and Rosen (2002), and show how this reply is in fact wanting. I conclude, in the third section, with some remarks on a more promising way to argue for our non-existence as bona fide objects.

Mereological nihilism As I have said, the chapter deals with an argument against our own existence (as bona fide objects). Given that this argument is crucially based on the metaphysical doctrine known as ‘mereological nihilism’, I am going to start by briefly clarifying this essential assumption of the argument I am concerned with.1 According to mereological nihilism composition never takes place (Cameron 2010; Dorr 2002; Dorr and Rosen, 2002; Sider 2013).

216

Natural and Artifactual Objects in Contemporary Metaphysics

That does not mean that nothing exists, just that the only things that exist are not composed of anything, that is, they have no proper parts – proper parts of x being, of course, parts of x other than x itself. So the only things that exist – according to mereological nihilism – are not composed of anything (they are things with no proper parts whatsoever), and these things are called ‘mereological simples’, ‘simples’ or ‘mereological atoms’ (which are not the ‘atoms’ of physics, which do have parts). Usually contemporary nihilists take those things – mereological simples, that is – to be the smallest subatomic particles there are, and I will stick to this commonly held idea given that it is also shared by the nihilists I am concerned with here. While science has not understood for certain what the smallest particles are, nihilists say that whatever those things turn out to be, they are all that exists – they are the only things, the only objects, that exist. So, the nihilist denies that there are composite objects: there are no tables, no chairs, no trees, no mountains, no dogs, no cars and so on and so forth. This is a startling thesis indeed, and, one may be tempted to say, it is surely and obviously a false one. In fact (1) it cannot be doubted that tables, chairs, trees and dogs exist: they do exist, and they are ultimately composed by subatomic particles arranged in certain ways; and of course (2) we see every day composite objects of these kinds; moreover, (3) wouldn’t it be absurd to say that ‘there are tables’ is false just like ‘there are dragons’? But the nihilist has clever responses to these qualms.2 (1ʹ) We commonly say that there is a table composed by many particles arranged in a certain way. But this is not exactly true. The sober truth is that there are just a lot of simples arranged table-wise: these simples however do not compose a single thing, the table. And the same goes for every alleged composite object: strictly speaking, there are no chairs, but simples – partless subatomic particles – arranged chair-wise, no trees but simples arranged tree-wise and so forth. (2ʹ) But why talk of simples arranged chair-wise and deny the existence of chairs while everyone is able to just see chairs? But, the nihilist says, this objection is not a good one. It won’t do to defend the claim that there are composite objects with nothing more than ‘I can simply see the object composed of the atoms arranged chair-wise’ (or tree-wise, or dog-wise). One’s visual experience would be exactly the same whether or not the simples composed something. The debate over whether chairs, tables and composite objects generally exist is not an empirical one.

Do We Exist?

217

(3ʹ) Nihilists do say that strictly speaking the statement ‘There are tables’ is false. But according to the nihilists it is not the case that all falsehoods are on the same footing. ‘There are tables’ is a false sentence indeed, but it is an appropriate and correct one: it is appropriate and correct to say that there are tables if, as in fact it is the case, there are simples arranged table-wise. And making statements like these is not only appropriate and correct but also useful and even indispensable in our everyday life. Saying ‘There are dragons’, instead, is not at all appropriate and correct, and this is because there are no simples arranged dragon-wise. Yet, the usefulness of ‘There are tables’ notwithstanding, there really are no tables, and our common talk of composite object is just a useful convention with simple (i.e. non-composite) fundamenta in re. That said, mereological nihilism may still look like mad metaphysics and one in need of strong arguments to be upheld. While some interesting arguments have indeed been put forward, this is not the place to discuss them, so I just hint to one, the causal redundancy argument, to which I’ll make reference to in the following section.3 There may be different versions of the causal redundancy argument, but the central, and quite simple, idea goes like this. Let us consider baseballs as an example of composite objects. Now, every alleged causal action of a baseball (breaking a window, say) is completely accounted for by its constituents (simple) parts acting in concert. So the baseball is a causally redundant entity and we have no reason to admit its existence. More generally, we have no reason to posit composite objects as entities existing over and above their alleged parts. Having sketched the (quite desert) metaphysical landscape of mereological nihilism, let us now turn to the argument for our non-existence.

Do I exist? The collective thinking question In order to properly state the argument for my (and your) non-existence, consider the following two points: First point – Nihilists are ordinarily concerned with denying composite objects and this, I take, is to be intended as a denial of composite individuals or substances. Yet, if nihilists don’t want to be completely arbitrary mereologists, they are bound to deny composite properties, events, actions and, more generally,

218

Natural and Artifactual Objects in Contemporary Metaphysics

the existence of composite entities tout court. How could one justify the idea that there are composite events, say, but not composite objects? So sensible nihilists will say that there are no composite entities. Second point – The vast majority of contemporary philosophers subscribe to the thesis that if persons exist, then they are composite entities. If this conditional were false, then persons would have to be simple entities such as immaterial souls or elementary particles, and few philosophers are happy to endorse these ideas. (Consider, for example: how could a person be an elementary particle?) The two previous points immediately give us the two premises of the argument against our own existence – and here is the argument (see Dorr and Rosen 2002; Sider 2013; van Inwagen 1990): N (1) If I exist, I have proper parts – and so there is at least one composite entity. (2) There are no composite entities. Conclusion: I do not exist. N is a valid argument: if premise (1) and premise (2) are true, then the conclusion must be true as well. And N’s conclusion claims that I – and we – do not exist. This does not mean that sentences such as ‘I exist’, or sentences presupposing the existence of various persons, are to be abandoned. Just like ‘there are tables’, many personinvolving statements are appropriate, correct and useful, and indeed they are even indispensable in our everyday life. But, strictly speaking, human people do not exist: we are, in a sense, useful artificial constructions. But how could this idea be true? Certainly my thoughts exist, and if they do, so do I: cogito ergo sum. And if so, either premise (1) or premise (2) – that is mereological nihilism – must be false. Against this simple and straightforward Cartesian critique to N, Dorr and Rosen (2002, p. 159) have an answer: You do not exist but the things you used to think of yourself as doing get done all the same. Certain [mereological] atoms jointly think those thoughts, dream those dreams, and so forth.

In other words, ‘your’ thoughts do exist, but you do not. There is no single entity, you, that is the bearer of the thoughts, but simple objects collectively perform the thoughts themselves, or, in other words, the activities of thought are carried out jointly by partless things.

Do We Exist?

219

The activities I am used to ascribing to myself – thinking these thoughts, feeling these feelings and so forth – are not performed by any single thing: rather they are jointly performed by those things that I used to call ‘the particles that compose me’ (Dorr 2002, p. 74). Unless this idea can be showed to be incoherent, Dorr and Rosen say, N may well be a valid argument with true premises. Now, van Inwagen (1990, p. 118) has said that collective thinking is an impossibility and an incoherent idea, and this is because he thinks that thinking, desiring or feeling is not the sort of thing that several things could do collectively. But van Inwagen has not said why thinking could not be a cooperative activity. He simply finds it obvious. But, confronted with this alleged obviousness, a defender of N may well remain sceptic. One could insist that saying that ‘many simples think’ is just ungrammatical, or that it conflicts with the meaning of verbs like ‘think’ or ‘believe’. But, it seems, a nihilist could easily answer: ‘How could this be a truth of deep metaphysical or epistemological significance, as opposed to a shallow fact about the way our language works, to which one might respond by advocating a minor linguistic reform?’ (Dorr 2002, p. 75). A third objection to collective thinking could be the following: if A, B and C think, then A thinks, B thinks and C thinks. Therefore, if many simples think, then every single simple thinks, and this is absurd. But against this point, the nihilist may simply repeat the previous answer: ‘You said, “if A, B and C think, then A thinks, B thinks and C thinks”. But this is simply a shallow fact about the way our language works, to which one might respond by advocating a minor linguistic reform.’ Now, let us grant that from the somewhat vague idea that thoughts are collectively performed by a plurality of simples one could not draw any contradiction – at least not in an obvious way. And let us grant that the ‘grammatical’ objections are not sound. But conceding the prima facie consistency and tenability of Dorr and Rosen’s proposal is not to say that, as it stands, it is a convincing one: lacking some more details on the metaphysics of collective thinking, Rosen and Dorr’s suggestion is little more than a gesture towards a possible reply to the Cartesian objection to N, and the Cartesian objector is entitled to be unmoved until some more details are given. So how exactly, metaphysically speaking, do the simples perform collectively the acts of thought? What exactly are, in this perspective, the acts of thought

220

Natural and Artifactual Objects in Contemporary Metaphysics

themselves? Rosen and Dorr do not offer clear answers to questions like these and, indeed, they do not even attempt to give one. But I think they are obliged to give some more details; otherwise, as I have said, a Cartesian could simply ignore their far-fetched metaphysics of thinking. In order to advance the debate, I consider here three different ways to understand the nature of collective acts of thought – and show them to be wanting.

Thinking is like putting on a play with many actors. Consider a play, Hamlet, for example, and putting aside nihilist qualms, let us focus on some ordinary things one could say about it. Each actor, of course, does something less than performing the play, and these lesser activities together make it the case that the play is performed: one may plausibly say that the joint actions of the actors add up to, or compose, one ‘big action’ – the play, that is (Olson 2007, p. 188). In the same way, one may suggest, the joint actions of some simples could compose one big action, a single act of thought which would be collectively performed: a thought without a single thinker. This could indeed be an elegant account, if it were not the case that it is obviously not available to the mereological nihilist – and, a fortiori, to the supporter of N. As I have said, while nihilists are ordinarily concerned with denying composite objects, if they don’t want to be completely arbitrary mereologists they are bound to deny composite actions as well (and, more generally, the existence of composite entities). So, by nihilist lights, collective acts of thought cannot be collective actions composed by actions performed by simple objects.

Thoughts are ‘n-ary actions’, collective actions involving n simple objects. Let us consider a different way to clarify the metaphysical nature of collective thoughts. When n horses surround a house, there exists an n-ary action of surrounding the house – or so one could say. Analogously, when n simples collectively think there exists a n-ary act of thought – this, at least, is the proposal: thoughts are ‘n-ary actions’, collective actions involving n simple objects. But, of course, each horse performs an action by itself when surrounding the house – and, by analogy, one should also say that each simple object involved in an act of thinking acts by itself.

Do We Exist?

221

What, then, is the relationship between these single actions – the actions, that is, performed by the simple objects involved in an act of thinking – and the n-ary action – the act of thinking itself? Not a relation of composition, of course: nihilists do not – or should not – allow composite actions. So the n-ary action, the thought, has to be considered a simple action which exists over and above the actions performed individually by n simple objects. But why postulate the existence of such a further action? In the case of the house surrounded by the horses, horses’ actions ‘do all the work’ and there is no need to admit a further n-ary action. And the same holds, it would seem, for the case of thinking. (This, notice, is a line of reasoning which is very similar to the causal redundancy arguments nihilists like to use.) But, given that the acts of thinking were identified with n-ary actions, then the acts of thinking themselves disappear from the inventory of the world, contrary to Rosen and Dorr’s assumption. Furthermore, let us grant, for the sake of argument, that n-ary actions are to be admitted in one’s ontology. If the single actions, a1 … an, and the n-ary action, A, were related by the composition relation, then arguably the existence of a1 … an would metaphysically necessitate the existence of A. But if, as we are supposing, A is not composed by a1 … an, it seems natural to say that A itself should be considered as contingently emerging, in our world, from the ‘basic’ actions of the simples: A is to be viewed as an emergent action, and as emerging as a matter of brute fact. But if one is willing to admit this, one will also be willing to admit the plausibility of the emergence of a substantial, immaterial, simple self, and if an emergent, simple self is a plausible metaphysical option, then premise (1) of N is no longer plausible.

Thoughts exist and they are identical to irreducibly plural collections of acts. Let us then consider a third way of making sense of collective thinking. Flocks exist, we ordinarily claim, and they are identical to plural collections of sheep. In the same way, one may say, thoughts exist, and they are identical to irreducibly plural collections of acts. But if so, then, arguably, things such as chairs, tables, human organisms and human brains also exist: they are identical to irreducibly plural collections of simples – plural collection of subatomic particles arranged chair-wise, table-wise, brain-wise and so forth.

222

Natural and Artifactual Objects in Contemporary Metaphysics

But this, of course, runs against what nihilists, and N’s supporters, want to say. Moreover, if human organisms and human brains exist, then, arguably, I exist as well – which, of course, contradicts N’s conclusion. And one may also add that claiming that thoughts are identical to plural collections of acts is at odds with what the phenomenology of thoughts themselves suggests about their very nature.

The challenge to be met – and a dualist-Buddhist suggestion Given the discussion in the previous section, the challenge for the supporter of N is as follows. Either one rejects Dorr and Rosen’s reply to the Cartesian objection (I) or one subscribes to it (II). (I) In the first case, the supporter of N has to come up with a plausible reply to the Cartesian objection not appealing (or not crucially appealing) to the idea of collective thinking. (II) In the second case, the supporter of N has to come up with a clarifying account of collective thinking. Otherwise the idea that thoughts are collectively done by simples would remain metaphysically obscure and could not provide a good answer to the Cartesian critique of N. Now, if I had to guess what is the more promising idea to meet the challenge just stated, I would say that non-physicalism concerning the mental is the right move to make. Yet not all kinds of non-physicalism would do. Let me explain. In Dorr and Rosen’s account, the simples, being elementary particles, are certainly physical, while their actions, which are responsible for the thinking, might be either physicalistically acceptable actions (as, I think, Dorr and Rosen would say) or, instead, non-supervenient and irreducible actions. If the latter were the case, one would have an explicitly non-physicalist version of Dorr and Rosen’s idea, and one may think that this could be relevant for meeting the second horn of the challenge. Yet embracing this kind of non-physicalism would not be a solution. Declaring the actions non-supervenient and irreducible does not by itself contribute much to clarify the metaphysical nature of collective thinking. And, moreover, the problems I emphasized for Dorr and Rosen’s original thesis do not depend on the physical or non-physical status of the actions involved in thinking, and so are also relevant to the non-physicalist proposal we are now considering.

Do We Exist?

223

Yet, I submit, a different form of non-physicalism could be a way to meet (the first horn of) the above challenge, providing a reply to the Cartesian objection which is different from the strategy suggested by Dorr and Rosen. Mentality, in this new perspective, should not be accounted for in terms of the collective activity of fundamental physical particles. The nihilist should instead admit in her ontology primitive and simple mental states, that is, primitive mental states with no proper parts. According to this dualist picture, there would be two quite different kinds of simples: the physical ones – including electrons, for example – and the mental ones – including, for example, pain states. It would be nonetheless true that you and I do not exist, but not because physical simples jointly think ‘our’ thoughts. There would be simple mental states without a bearer interacting with other mental and (non-mental) physical simples. And a proper interaction between certain mental and physical simples would constitute the appropriateness conditions for the use of words such as ‘person’, ‘I’ and ‘you’. Indeed what I am suggesting here is by no means a brand new position. The majority of early Buddhist philosophical schools (i.e. the majority of so-called Abhidharma schools) denied the existence of the ‘I’ embracing a version of mereological nihilism admitting physical and mental simples.4 And given the apparent difficulties the collective-thinking account – and the very similar account one finds in Sider 20135 – is confronted with, it seems that the ‘Abhidharma move’ may well be worth considering. Perhaps being a mereological nihilist, denying our existence and admitting only simple, non-mental, physical objects as relata or property bearers, is really too much. If you are a mereological nihilist willing to deny your own existence, then try Buddhist nihilistic dualism.

Notes 1 What follows is primarily intended for readers not familiar with mereological nihilism. Those already accustomed to this exotic idea will not discover anything new, but will have at least the occasion to know how I frame the issue. 2 Versions of the following responses can be found in Dorr 2002, Dorr and Rosen 2002, Merricks 2001, Sider 2013, van Inwagen 1990. 3 Causal redundancy considerations take central stage, for example, in Merricks’s (2001) quasi-nihilist metaphysics. 4 See, for example, Siderits (2015).

224

Natural and Artifactual Objects in Contemporary Metaphysics

5 Sider (2013, 268) sees no problem in accounting for mentality as a relation among many (simple) subatomic particles. While Sider’s very briefly stated proposal amounts to no more than what I have just said, it seems to me reasonable to view it as a variant of Dorr and Rosen’s idea – a variant subject to very similar difficulties.

References Cameron, R. (2010), ‘Quantification, Naturalness and Ontology’, in A. Hazlett (ed.), New Waves in Metaphysics, Palgrave Macmillan, Basingstoke, 8–26. Dorr, C. (2002), The Simplicity of Everything, Ph.D. thesis, Princeton University. Dorr, C., and G. Rosen (2002), ‘Composition as Fiction’, in R. Gale (ed.) The Blackwell Guide to Metaphysics, Blackwell, Oxford, 151–74. Merricks, T. (2001), Objects and Persons, Clarendon Press, Oxford. Olson, E. (2007), What Are We?, Oxford University Press, Oxford. Sider, T. (2013), ‘Against Parthood’, in K. Bennett and D. Zimmerman (eds), Oxford Studies in Metaphysics, volume 8, Oxford University Press, Oxford, 237–93. Siderits, M. (2015), Personal Identity and Buddhist Philosophy, Ashgate, Farnham. Van Inwagen, P. (1990), Material Beings, Cornell University Press, Ithaca.

Abstracts Carving Nature at the Joints Richard Davies, University of Bergamo Starting from a well-known passage in Plato’s Phaedrus, in which a contrast is set up between a skilled carver and a bungling butcher, the volume’s introduction seeks to set up in general terms the sorts of distinctions between the differentiations that are attributable to what is so in nature and those that depend in one way or another on human minds and determinations. In line with the thematic questions that provide titles for the volume’s parts, the first section seeks to bring out some of the interconnections among the chapters that follow. The second section is devoted to a brief excursus on social ontology, a topic not dealt with elsewhere in the volume. The suggestion is made that the leading schools of thought in this area may be unduly attached to the schemes for identifying social facts that are actually hegemonic in a given society and thus have underestimated the way that social objects can exist unrecognized at the times that they make their appearance.

Mental Acts, Externalism and Fiat Objects: An Ockhamist Solution Riccardo Fedriga, University of Bologna For a thinker like William of Ockham, a theory of human cognitive powers is subject to the following constraints: (1) the imperfection of our knowledgegathering apparatus; (2) the shifting borders between the knowing subject and the world as it is encountered and (3) the variety of objects that our conceptual tools allow us to grasp and categorize. Taking these constraints into account and acknowledging the historical perspective, we illustrate how Ockham’s causal – rather than representational – epistemological account of concept formation should be regarded as less decisive in determining the modal way we identify the objects of knowledge than the status and relevance of the objects themselves.

226

Abstracts

This seems to mean (1) that it is useful to mobilize Smith’s notion of a fiat object to capture Ockham’s position between internalism and externalism; (2) that the categorial distinction between substance and quality is ontologically foundational and, finally, (3) that there are close relationships between simplicity and parsimony in Ockham’s metaphysics.

Is the World Really a World of Objects? A Note on Quinean Ontology Antonio Rainone, “L’Orientale” University of Naples This chapter is aimed at highlighting a particular aspect of the Quinean views on ontology – the concept of a posit. According to Quine, all objects ‒ both the objects of common sense and those of the most sophisticated scientific theories ‒ are posits, that is, more or less deliberate results of our language and conceptual scheme. How is it possible, then, to reconcile this conception with the ‘robust realism’ that Quine defends? Should we perhaps say that the Quinean picture of the world involves a belief in fictitious entities ‒ a form of fictionalism or instrumentalism? Arguments from psychology of perception and evolutionism (in other words, naturalism) seem to rule out this interpretation.

Spatial Fictionalism. A Solution to the Grounding Problem Nicola Piras, University of Sassari The chapter considers a range of responses to the thesis that two distinct things cannot occupy the same region of space at the same time. This is often known as the No Co-occupation Law (NCL) and seems to be a common-sense view. Nevertheless, the NCL is hotly contested by many philosophers, who contend that there are counterexamples to it, for example, when we say that a statue and the clay of which it is made occupy the same region of space at the same time. If such counterexamples really do conflict with the NCL, then there is a problem about how to ground the differences between the two things in question. Even philosophers who wish to defend the NCL seem to have difficulties in responding to this grounding problem. The suggestion is then made that a solution can be found by formulating a theory that combines the bona fide standing of the stuff of which the world is made with the fiat nature of the boundaries that we

Abstracts

227

draw within that stuff. Insofar as things are such because of the boundaries that delineate them, all things are fiat. This is the doctrine of Spatial Fictionalism, among the merits of which is preservation of the intuitive principle of the NCL and not to give rise to any grounding problem.

Talking about Properties: A Couple of Doubts on Hofweber’s Internalist View Elisa Paganini, State University of Milan A couple of doubts are raised concerning Hofweber’s internalist view of the talk about properties. The first doubt relates to the argument used in support of the internalist view of the talk about properties: the suspicion is that one of the premises of the argument is not granted and therefore that the argument’s conclusion is undermined. My second doubt concerns a claimed consequence of Hofweber’s internalist view, that is, conceptual idealism. It seems to me that conceptual idealism is incompatible with the internalist view of talk about properties.

The Eye of the Needle: Seeing Holes Clotilde Calabi, State University of Milan I start with three nonnegotiable claims and one intuition. The claims are as follows: 1. Holes exist. 2. We see holes. 3. Seeing a hole is a visual experience of some kind. The intuition is that we see holes directly, as we see ordinary material objects, such as trees and chairs. Vision theory explains this intuition and supports it. Casati and Varzi argue instead that we see holes indirectly, because holes are not bona fide entities. This is interesting because it promises to show how ontology impinges on vision theory. If their argument is sound, then holes are unlike what the visual system takes them to be, and so we have a case in which ontology and vision theory come apart. In the chapter, I criticize Casati and Varzi’s argument and defend the contrary claim that, with respect to holes, ontology

228

Abstracts

and perception do not come apart. Then, I enquire whether this conclusion is neutral with respect to materialism and immaterialism about holes.

Bona Fideness of Material Entities and Their Boundaries Lars Vogt, University of Bonn After a brief introduction to the distinction between bona fide and fiat boundaries and their role in identifying natural units in the material realm, I discuss specific problems relating to their operational criteria that considerably limit their applicability in practical empirical research. I continue with discussing a new approach of characterizing bona fideness of natural units that is independent of boundaries, but instead focuses on different types of causal unity. I argue that the relevance of causal unity depends on the frame of reference. Each frame of reference has its own associated particular type of causal unity. Causal unity via internal physical forces is associated with the spatio-structural frame of reference. I introduce causal unity via bearing a specific function, which is associated with the functional frame of reference, and causal unity via common historical/ evolutionary origin, which is associated with the historical/evolutionary frame of reference. I conclude by discussing that boundaries are still important in this new approach, because they take in a central role in empirical research. I suggest that whereas fiat boundaries can be considered as mathematical boundaries, possessing a dimension one less than the entity they bound, bona fide boundaries of material entities must be physical boundaries and thus three-dimensional entities themselves – boundary regions instead of crisp surfaces. A bona fide boundary can be experienced in the physical realm only if it is a physical entity itself. Thus, only if we understand bona fide boundaries as physical boundaries, we will be able to identify natural units in a diagnostic empirical framework and only then bona fideness becomes an operational category.

A Conceptualist View in the Metaphysics of Species Ciro De Florio and Aldo Frigerio, Catholic University of Milan Barry Smith (Topoi, 20(2): 131–48, 2001) affirms that biological species, such as cats, are fiat objects created by moving from bona fide objects (the particular organisms) through a bottom–up process. This thesis is discussed, and it is split

Abstracts

229

into two different subclaims: (1) biological species are individual objects, not concepts and (2) biological species are fiat, not bona fide, entities. There are good reasons to reject the first thesis. Subclaim (1), species-as-individuals, is very popular among theoretical biologists and philosophers of biology because it seems to provide a concept of species compatible with evolution. However, this view is grounded in the relational concepts of species, which, as is shown, suffers from many problems. Subclaim (2) is correct, but requires some revisions. If it is true that species are sometimes cut for conventional reasons, it is also true that there are often objective reasons to isolate two different species (in contrast to what happens with political boundaries).

Artifacts and Fiat Objects: Two Families Apart? Massimiliano Carrara and Marzia Soavi, both from the University of Padua Fiat objects may come into existence by way of intentional explicit definition and convention or they can be the result of some spontaneous and automatic activity that results in tracing fiat spatial boundaries. Is it possible to consider artifacts as fiat objects and to what extent? Or else can we conceive at least some artifacts as fiat objects? In order to draw a map of the possible answers to these two questions we take into account different definitions for artifacts stemming from the two classical approaches: the intentional one and the functional one.

The Semantics of Artifactual Words Marco Santambrogio, University of Parma The semantics of a referential expression is externalist if its reference is not determined by some uniquely identifying marks satisfied by the referent and known or believed to be true of that referent by the speaker. In order to show that externalism holds true not only for proper names of bona fide objects, such as persons and names of natural kinds (as argued by Kripke, Putnam and others), but for artifactual words and names of fiat objects as well, I consider Kaplan’s conception of the metaphysics of words. I then discuss to what extent words can be considered as representative of all artefacts and all fiat objects so far as semantic theory is concerned.

230

Abstracts

Are Linguistic Objects Fiat or Bona Fide? An Ancient Proposal Maddalena Bonelli, University of Bergamo The chapter aims to pinpoint in Aristotle and in Alexander of Aphrodisias the origin, the development and the problems related to the distinction between bona fide and fiat objects as regards language. An initial question regards the double nature of natural language, considered as phonic material that is nevertheless articulated in conventional sequences. A second question considered regards the crucial role of thought in the constitution of language and reality. These questions are fundamental to Aristotle’s theory of language and underlie the treatment of linguistic facts in terms of bona fide and fiat objects, and they led Alexander to develop arguments in favour of Aristotle’s own theory, which was presented in so succinct a way as to be rather obscure. Despite the clearly realist starting point of the Peripatetic view, according to which reality is a guarantor of the truth (or falsity) of assertoric propositions, the creative power of language emerges clearly both from the observations of Aristotle and from their developments in Alexander. The treatment of propositions is fixed by fiat in different, and perhaps even irreconcilable, ways according to the logico-inferential roles they play or how they ‘reflect reality’.

Leibniz’s Principle and Psycho-Neural Identity Andrea Bottani and Alfredo Paternoster, both from the University of Bergamo Can the Leibnizian principle of the indiscernibility of identicals be used as an argument against the psycho-neural type identity, that is, against the theory according to which mental properties are nothing but neural properties? We argue that (1) Leibniz’s law cannot provide a reason to reject the identity theory and (2) it makes it very hard to distinguish between two different versions of the psycho-neural identity theory, the weaker ‘token identity’ theory and the stronger ‘type identity’ theory.

Abstracts

231

Do We exist? Mereological Nihilism, Collective Thinking and Dualism Alfredo Tomasetta, University School of Institute of Advanced Studies, Pavia Are we bona fide, objective entities? This chapter focuses on a way to defend the strongly revisionary thesis that we are not. In particular, it deals with an argument from mereological nihilism to our non-existence (as bona fide objects). After presenting the argument, and its presupposed metaphysical framework, a Cartesian objection to the argument itself and a nihilist reply by Dorr and Rosen are considered. I argue that, as it stands, the nihilist reply is unconvincing, and suggest that a better way to counter the Cartesian objection is offered by Buddhist philosophy.

Index of Names and Principal Subjects abstract entity (abstractum) 7, 51, 55, 62, 93–4, 98, 123–4, 145, 157–9 abstraction 6, 21, 27, 32, 48–9, 150, 163 abstractive apprehension (Ockham) 20–4, 27, 34–5 Adams, M. M. 43 n.35 Ademollo, F. 194 nn.26, 29 affections of the soul (pathemata) 9, 182, 184–6 affirmations and premises 188–91 Alexander of Aphrodisias 9, 181–96 Alston, W. P. 54 Ammonius Saccas 9, 181, 183–4 ancestor, biological 122–3, 127 ancestry, principle of common 157–8, 162–5, 168–71 angels (angelic cognition) 23, 30–1 animals (organisms) 2, 8, 50–4, 164–5, 171, 200 Antiphon 184 Apophansis, see affirmations apprehension 20–4, 26, 28, 35–6 Aquinas, St. T. 37 Aristotle of Stagira 9, 11, 14–16, 20, 33, 37–8, 71, 124, 181–92 Arp. R. 105–6, 109–10 artwork 70, 159, 167, 172, 173 n.9, 177 n.36 athens 11–12, 181 atomism 1, 2–3, 24, 48–9, 55, 72, 109, 111, 161–2, 209–10, 216, 218 Authority, Principle of Speaker’s 171, 174 Barnes, J. 192 n.7, 194 nn.27, 31, 32, 36 Basic Formal Ontology (BFO) 105 Bell, A. G. 149–50, 169–70, 177 n.33 Bennet, K. 58, 63, 64 Berkeley, G. 53 Bertamini, M. 93, 100 Berto, F. 199

bicellular eye 108, 110, 113 Bittner, T. 104 Black, M. 60 Block, N. 209–12 Bloom, P. 158, 167 Boehner, P. 34 Boethius, A. M. S. 9, 181, 184–5, 187, 192 n.9, 193 nn.22, 24 Boler, J. F. 27–8 bona fide boundary 72–5, 103–5, 107, 109–10, 115–16, 131 object 1, 8, 9, 10, 40, 44 n.47, 48–9, 59, 70, 72–3, 76, 93–6, 100–16, 131–2, 143–4, 147, 181–2, 191, 215 Bonelli, M. 181–96 Bottani, A. C. M. 199–214 boundaries bona fide 72–5, 103–5, 107, 109–10, 115–16, 131 fiat 72–4, 76, 104–6, 110–11, 131, 143–4, 182 Brandon, R. M. 136 n.12 Brogaard, B. 104, 109 Brower-Toland, S. 28–9, 43 n.35 buddhism 10, 222–3 Burge, T. 127, 156, 158 Burger, J. 127 Burke, E. 58, 70 Calabi, C. 93–102 Cameron, R. 215 Carnap, R. 81–2, 178 n.37 Carrara, M. 141–54, 173 carving nature at the joints 1–4, 16 n.1, 40, 44 n.47, 117 n.2, 131 Casati, R. 61, 93–6, 100, 101 n.6 category, categorisation 19–20, 25, 32, 42 n.13, 50, 105, 133, 145–6, 153, 199, 207

Index of Names and Principal Subjects causality 20, 22–5, 28–30, 32–4, 38–40, 43 n.32, 66–7, 116, 148, 152, 203, 209–11, 217 causal unity (biological) 105, 107–15, 136 n.12 Cavini, W. 194 n.34 cell 66, 106–8, 110–11, 113–14, 116 cellulose 62–5, 67–8, 136 n.12, 207 C-fibers 200, 202–3, 205, 208 charity (in interpretation) 52, 54 Chatton, W. 34, 43 n.35 Chisholm, R. M. 71 Chuang Tzu 16 n.1 cognition 21–9, 33–6, 41 n.12 coincidence, spatial 58, 65, 68–9, 75 community, linguistic 9, 52–3, 157, 160, 167, 171–2, 174 n.16 composition and identity 69, 105, 215, 221 conceptualism 40, 85–8, 121–35 conceptual scheme 25, 49, 51–5 contingency future contingents 33, 36–40 matters of fact 21–8, 34–5, 62 non-necessary characteristic 122, 209, 221 continuity and boundaries 103–4 and homogeneity 72, 104, 106, 127 in space and time 76, 108, 153, 158–9, 166 convention 1, 3, 6, 9–10, 30–3, 39–40, 48, 67–8, 72, 104, 131, 141, 153, 161, 183–5, 215, 217 counterfactual 34, 66 counterpart 64, 67–8 Craver, C. F. 105, 113, 117 n.2 Crivelli, P. 194 n.34 Croucher, C. 100 Cuvier, G. 124 Dalla, L. 14 Dalton, J. 4 Darwin, C. 8, 122, 124, 161, 175 Dasgupta, S. 74 Davidson, D. 52–3, 61, 201 Davies, R. W. 1–17 De Florio, C. 121–38 Democritus of Abdera 1–3

233

De Queiroz, K. 135 n.8 DeRosset, L. 63 Descartes, R. 215, 218–20, 220, 222–3 dimensionality of boundaries 7, 71, 74, 104–5, 110, 115–16 discontinuity, physical 72, 104, 106–7, 109–11, 113, 116, 143–4 disposition 34, 103, 107–8, 112–14, 148, 206 DNA 159, 161, 165, 168, 171 Dobzhansky, T. 126 Dorr, C. 215, 218–22 Dretske, F. 98, 101 n.3 Dummett, M. 200 Eakin, R. M. 110 earth, twin 128–9, 158–9, 164–5, 168–9, 171 Eco, U. 40 Ehrlich, P. 125 Einheuser, I. 63, 67 Eldredge, N. 105, 136 n.12 elementary particles 2, 218, 222 emergence/supervenience 59, 69, 208, 213 n.8, 221, 222 entity abstract 122–3, 148 entification in a theory 19, 21, 28–9, 49–50, 52, 55, 59, 69, 71–4, 76, 84–6, 94, 121, 131–3, 199–200, 207, 218 material/natural 103–17, 159, 164, 215 Ereshefsky, M. 125, 136 n.12, 161 essence, essentialism 27, 34, 54, 64, 67, 121–2, 147, 157–8, 160–1, 163–4, 168, 175–6 n.25 evolution 7–8, 49, 50, 52, 105–8, 112, 114–15, 121–4, 128, 131, 134, 136 n.12, 161 exist by nature 182–7, 191 existence 21–2, 24, 26–9, 33–4, 49–51, 55–6, 61, 65–6, 76, 81–8, 93–4, 97, 103–4, 111–14, 124, 128–30, 143–4, 151–3, 159–60, 166, 168, 215–20 externalism 19–35, 41 n.11, 156–8, 160–1, 164–5, 168, 170–2, 175–6 n.25, 211

234

Index of Names and Principal Subjects

facts – hard and soft 43 n.32 false judgement 34–6, 82–3, 182, 187–90, 194 n.35 Fedriga, R. 25–47 Ferraris, M. 12–15, 16 n.3 Fiat – objects 9–10, 19–20, 40, 44 n.47, 48, 59, 70, 75, 110, 131, 141–53, 181–2, 191 boundaries 70, 72–6, 104–7, 110–11, 115, 131, 143–4, 182 fictionalism, spatial 58–9, 70, 75–6 figure-background 95 Fine, K. 61, 63, 68–9, 175–6 n.25 four-dimensionalism 163, 168, 200 Frege, G. 190–1, 194 n.36 Frigerio, A. 121–38 function artifactual unit/kind 3, 7, 9, 33, 141–2, 145–54, 159, 166, 172 biological unit/kind 107–8, 110–14, 117 n.5, 211 law 206–7, 210 Funkhouser, E., 213 n.5 fusion, mereological 163–4, 168

Hume, D. 53 hylomorphism 68–9 hypothesis 39, 52

Gallois, A. 199 Gauthier, J. 135 n.8 Geach, P. T. 77 n.4, 199 gene 122, 125, 127 Gestalt 52–3 Ghiselin, M. 123, 125 Gilbert, M. 16 n.3 Gillon, B. 174 God 25–30, 34–5 Goethe, J. W. von 124 Goodman, N. 51 Grandy, R. 174 n.13 granularity 106–9, 111, 113, 154 grounding (problem (GP)) 38, 43 n.32, 58–9, 63, 65, 68, 76

Kant, I. 25, 55, 173 n.9 Kaplan, D. 156–7, 161–3, 166, 171, 173 nn.2, 3, 7, 174 nn.16, 17, 175 nn.18, 20, 22, 25 Karger, E. 26–7, 34–5 Keet, C. M. 108 Kemp. G. 56 n.1 Kenny, A. 192 n.3 Kilwardby, R. 38 Kim, J. 208, 210–12 Klima, G. 41 n.11 Korman, D. 16 n.2 Kornblith, H. 156 Koslicki, K. 68–9 Kripke, S. 156, 158, 161, 164, 170–2, 174 n.16, 177 n.35, 201

Haack, S. 174 n.15 Heller, M. 67, 74 hierarchy 105, 121, 135 n.8 Hilpinen, R. 142, 145, 175 n.23 hole 6–7, 93–102, 109, 143 Horgan, T. E. 66 Hugly, P. 174 n.15 Hull, D. 128, 136 n.10

identity token 204–8, 212 type 204–8, 212 identity conditions 64, 147, 149 identity of indiscernibles 60, 199 immaterial particular 96, 100 indiscernibility of identicals 60, 199 Ingarden, R. 109 instrumentalism 49, 55 intellect 20–1, 24, 26–7, 29, 31–2, 35–7, 41 n.2, 41–2 n.12, 101, 185, 188 internalism 19 intersubjective 50 intuition 22–30, 35, 58–9, 63, 75, 94–6, 131, 144–5, 147, 149, 164, 172, 199 intuitive knowledge (Ockham) 21–3, 26, 42 n.15 Irwin, D. 127 Johnson, W. E.

213 n.4

language 6, 9, 38, 49, 136 n.9, 161–2, 178 n.37, 181–2, 184, 187, 206 evolution of 177 n.35 mental 20, 30, 32–3 natural/ordinary 80, 85–6, 157–8, 165, 174, 181, 191–2, 210 and reification 53

Index of Names and Principal Subjects and scientific posits 49–50, 193 n.26 sign 159 spoken 31 and thought 187–91 Laporte, J. 135 n.5 Leibniz, G. W. von 199, 201–2, 204, 206–8, 212 Lepore, E. 173 n.3, 174 nn.14, 15 Levins, R. 116 Levinson, J. 158, 167, 177 n.36 Levinton, J. 105 Lévy-Bruhl, L. 54 Lewis, D. K. 67, 93, 106 Limonta, R. 44 nn.40, 44 Linnaeus, C. 124 Locke, J. 3, 16, 160 Lombard, see Peter Lombard Lowe, E. J. 124, 200 McLaughlin, J. 213 n.7 McLaughlin, P. 147 Marconi, D. 156, 174 n.12, 176 n.29 Marmo, C. 27, 42 nn.17, 23 marriage 11–14 Mayr, E. 105, 125, 135 nn. 2, 6, 136 n.12 mental act 19–20 mental state 33, 202, 210–12, 223 Menton 15 mereological nihilism 215–23 mereology 59–63, 69, 108, 112, 114–15, 124 Merricks, T. 66, 223 n.3 Meucci, A. 169–70 Millikan, R. 148, 154 n.4, 173 n.8, 174 n.16 mind-dependency 9–10, 104 modal properties 63–5, 67–8, 72, 75 money 13–14, 159 monism 1 Mulligan, K. 101 n.4 “Myth of the Given” (Sellars) 19, 24–5 names and verbs (onomata and rhemata) 184, 186–8, 194 n.35 nature 2–4, 16 n.1, 32, 34–6, 51–2, 80, 125, 131, 153, 157, 159–61, 164–5, 173 n.11, 175–6 n.25, 183–4, 191 Nelson, G. 135 n.4 neural state 203

No Co-occupation Law (NCL) 65–70, 75–6 Nomos 1–3, 10 Normore, C. 32 Notitia 20, 33, 41 n.12 Novikoff, A. B. 105

235 58–63,

object, see also Bona fide; Fiat natural 49, 80, 412–13, 150, 164, 175 n.22 social 10–14, 145 observation sentence (Quine) 50 Ockham, see William of Ockham Okasha, S. 135 n.5 Olson, T. E. 63, 69, 220 ontogenesis 50 ontological commitment 50, 58, 81 Oppenheim, P. 105 over-determination 66 Paganini. E. 80–90 Panaccio, C. 22–6, 28–31, 33, 41 n.11 Parmenides of Elea 1 parthood 59–60, 69, 113 Paternoster, A. 199–214 persistence conditions 6, 53, 63–5, 67–8, 72, 75–6, 146, 148 Peter Lombard 41 n.1 Phone (sound) 182–4, 186 Phone (= telephone) 60, 149–50, 160, 169–70, 176 n.28, 177 n.33 phylogenetic 50, 112, 127–9, 135 nn.3–5 Physis (nature) 1–2 Piras, N. 58–79 Place, U. T. 202 Plamick, N. 135 n.4 Plantinga, A. 43 n.32 Plato of Athens 1–2, 27, 44 n.47, 49, 123–4, 157, 175 n.25, 182, 193 n.15 Porphyry of Tyre 193 n.22 Posit 49, 52, 54, 217 Potrč, M. 66 Priest, G. 199 prophecy 36–8, 40 Protasis, see affirmations prototypical artifacts 132, 144, 157–8, 164–5, 170 Putnam, H. 55, 56 n.2, 105, 156, 158, 161, 164, 168, 170–2

236 Quine, W. V. O.

Index of Names and Principal Subjects 5–6, 48–56, 69, 81, 201

Rainone, A. 48–57 Raven, P. 125 Rea, M. C. 70 realization 94, 113 multiple 205–6, 211, 213 n.5 reduction 59, 200, 209–11 psycho-neural 208, 211 region, spatial 58–66, 68–72, 75–6, 94, 101 n.6, 105, 1156, 127, 132, 163, 200 reification 50, 52–5 reproduction, biological 107–8, 114, 135 n.1 Richter, S. 110 Ridley, M. 125 Rosen, D. 135 n.4 Rosen, G. 215, 218–23 Rosse, C. 117 n.1 Rubin, E. 95 Rudder Baker, L. 16 n.2 Rundle, H. 130 Russell, B. 77 n.12 Santambrogio, M. 156–80 Sayward, C. 174 n.15 Schaffer, J. 43 n.32 Schneider, S. 213 n.6 Schulz, S. 106–7 Scotus, J. D. 22, 27–8 Searle, J. R. 12–14, 16 n.3 seeing, direct and indirect 93–4, 96–100 selection, natural 54, 125, 130, 136 n.12 Sellars, W. 19, 24–6, 42 n.13 semantics 23, 30–2, 73, 81, 84, 96, 156–7, 159, 161–5, 168–70, 205 Shiel, J. 193 n.22 Sidelle, A. 63, 74 Sider, T. 62, 66, 68–9, 215, 218, 223, 224 n.5 Siderits, M. 223 n.4 Siegel, S. 100 Significatio 30–2 Simon, H. A. 105 Simons, P. 123–4 Smart, J. J. C. 208–9 Smith, B. 40, 48, 72, 76, 103–6, 109–14, 131–2, 143–4, 181

Soavi, M. 141–54 socrates (a philosophical example) 20–4, 132, 185–6, 194 n.28 Socrates of Alopece 1–2 sortal 43 n.13, 61–2, 64–5, 67, 70–2, 75, 205 sound (for communication) 9, 50, 157, 161–4, 169, 175 n.18, 181–6, 191, 192 n.12, 193 n.23, 194 n.28 species 51–2, 114, 121–35, 158, 160–1, 164–8, 170–1, 175 n.25, 206 Sperber, D. 158–9, 173 n.10 stage-continuant model (Kaplan) 161–5, 170 Stalnaker, R. 209–10 Stamos, D. 136 n.11 stimulus 50–5, 200, 202–3, 205, 208 stuff 59, 61–2, 65–6, 68–70, 72, 74–6 supervenience/emergence 59, 69, 208, 213 n.8, 221, 222 Suppositio 30, 37–9, 43 n.32 surface boundary 71, 73–4, 94, 103–5, 110, 115–16 inscription 167 sensory 35, 52–3 Sutton, C. S. 63, 67 symbol 13, 16 n.3, 182–4 syncategorematic 32–3, 39 Tachau, K. H. 26, 41 nn.1, 2, 12 Tahko, T. 136 n.14 Thomasson, A. 16 n.2, 142–3, 159, 176 n.26 thought experiment 23, 27, 128–9, 158, 164, 168, 172 Tomasetta, A. 215–24 truth-value 36–9, 190 Unger, P. 69, 77 n.13, 106 unit/unity causal/functional 7, 107, 109–16, 136 n.12, 147 evolutionary 107–8, 112, 123, 136 n.12 natural 12–13, 103, 105, 108, 181

Index of Names and Principal Subjects universal 21, 27, 41 n.5, 121, 123–4, 131–2, 193 n.25 universal quantification 50–1 vagueness 32, 73–4, 76, 105, 134, 157, 213 n.8 Van Inwagen, P. 69, 77 n.6, 218–19 Van Valen, L. 127 Varzi, A. 40, 44 n.47, 61, 72–3, 76, 77 n.8, 93–6, 100, 101 n.6, 103–6, 131, 136 n.14, 143–4 Vegetti, M. 193 n.26 Ventimiglia 15 verbs of perception 96–7 Vermaas, P. 147, 152

237

visual experience 93, 97, 100, 216 Vogt, L. 103–20, 135 n.1, 136 n.12, 144, 147 Vrba, E. 125 Wagner, G. P. 108 Wasserman, R. 58 Wiggins, D. 16 n.2, 61, 63, 71, 145–6, 200 Wiley, E. 136 n.12 William of Ockham 19–47 William of Sherwood 44 n.39 Williamson, T. 105 Wimsatt, W. C. 105, 116, 117 n.7 Wittgenstein, L. 156 Wodeham, A. 41–2 n.12