Mind: Ontology and Explanation: Collected Papers 1981-2005 9783110327151, 9783110326727

Citation preview

Laird Addis Mind: Ontology and Explanation

Philosophische Analyse Philosophical Analysis Herausgegeben von / Edited by Herbert Hochberg • Rafael Hüntelmann • Christian Kanzian Richard Schantz • Erwin Tegtmeier Band 25 / Volume 25

Laird Addis

Mind: Ontology and Explanation Collected Papers 1981-2005

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

Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.d-nb.de.

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

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

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

2008 ontos verlag P.O. Box 15 41, D-63133 Heusenstamm www.ontosverlag.com ISBN 978-3-938793-86-2

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

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

CONTENTS INTRODUCTION

1

ONTOLOGY Mind, Structure, and Time

7

Natural Signs

23

Pains and Other Secondary Mental Entities

49

Intrinsic Reference and the New Theory

67

The Ontology of Emotion

87

The Simplicity of Content

107

The Necessity and Nature of Mental Content

127

EXPLANATION Dispositions, Explanation, and Behavior

147

Behaviorism and the Philosophy of the Act

173

Parallelism, Interactionism, and Causation

199

Dispositional Mental States: Chomsky and Freud

221

Review of Laurence D. Smith’s Behaviorism and Logical Positivism: A Reassessment of the Alliance

243

Human Action and the Humean Universe

255

REFERENCES

273

AUTHOR’S NOTE

285

INDEX OF NAMES

287

INTRODUCTION

T

he papers of this book were originally published over a period of about a quarter century. For better or worse, my views throughout these years as reflected in these papers have remained virtually unchanged. Besides saving me the trouble of having to explain why I changed my mind about this or that, it also makes it easier to summarize what I do believe about the nature of mind. I have divided the papers into two groups, the one focusing primarily on the ontology of mental phenomena, the other on the role of mind in the explanation of behavior, especially but not only in explanations by way of dispositional mental states. Within each group, the papers appear in their order of original publication. The theses that dominate my views on the ontology of mind are these: (1) There is a distinct realm of mental properties that are not, in any relevant ontological sense, reducible to physical properties of the brain or body, to behavioral properties, to functional properties or causal connections, or to anything else; thus a dualism of properties is true. (2) Mental phenomena are usefully understood as falling under the three categories of (a) awarenesses (also called occurrent mental states or mental acts) which, for reasons argued at length, I characterize as the primary mental entities; (b) sensations such as pains and itches, emotions, moods, images and afterimages, which I characterize as the secondary mental entities; and (c) dispositional mental states, which I characterize as the tertiary mental entities. (3) Every awareness has a property that by its very nature grounds what the awareness is an awareness of, and so is what I call a natural sign; thus, there is (internal) mental content. (4) Consciousness just is intentionality; to be in a conscious state is to be in an intentional state. This thesis, important though it be, is never argued for directly. But it follows, more or less, from my argued-for theses that all of the primary mental entities are

intentional states, that the secondary mental entities occur only as the objects of intentional states, and that dispositional mental states are not, ontologically speaking, intentional states. The main theses regarding explanations of behavior are these:

dispositions

and

dispositional

(1) Dispositions necessarily involve grounds and laws; it is ontologically impossible for something to have a disposition without there being a ground of that disposition that enters into some relevant law or laws with respect to the behavior that is the, or a, manifestation of the disposition. (2) Dispositional mental states are not, ontologically speaking, intentional states although, just as dissolving is involved in the definition of solubility without being a constituent of it, intentional states are involved in the definition of dispositional mental states without being constituents of them. (3) Explanations of behavior by dispositional mental states (including explanations by reasons) are legitimate even though they are not causal explanations. (4) But even though they are not causal, every dispositional explanation of behavior presupposes the possibility (by way of knowing the ground of the disposition and the relevant law) of a causal explanation of that behavior. Underlying my development of these theses is the belief, argued for sporadically, that they are, without exception, consistent with the scientific worldview. By the scientific worldview I mean, roughly, the theses that everything that happens, including all mental phenomena, has a purely physical explanation (that is, that what I call scientific materialism is true), that methodological behaviorism (as contrasted with logical behaviorism) is the proper approach to the scientific study of behavior, and that only evolutionary biology can give us whatever understanding is possible of how mental phenomena came to exist My commitment to the scientific worldview is deeper that any particular view I have of the nature of the ontology of mind and its role in explaining behavior; and should any of the

2

latter views be shown to be inconsistent with the scientific worldview, I would take that as a conclusive reason for rejecting that particular view. But, putting my view negatively and in very broad strokes, I believe that the widely-held beliefs among both philosophers and scientists that everything is physical (that is, that what I call absolute materialism is true), and that only absolute materialism in consistent with the scientific worldview, are both false. Indeed, I suggest that an uninhibited empiricism (also part of the scientific worldview) leads one to that ontologically distinct realm of mental properties. The necessary permissions by editors and publishers have been granted and are gratefully acknowledged. A table indicating where each piece was originally published will be found near the end of the volume. L.A. Iowa City, Iowa, USA January 2008

3

ONTOLOGY

MIND, STRUCTURE, AND TIME

I

n his The Idea of History, R. G. Collingwood writes the following intriguing words: In the case of a machine, we distinguish structure from function, and think of the latter as depending on the former. But we can do this only because the machine is equally perceptible to us in motion or at rest, and we can therefore study it in either state indifferently. But any study of mind is a study of its activities; if we try to think of a mind absolutely at rest, we are compelled to admit that if it existed at all (which is more than doubtful) at least we should be quite unable to study it. . . . Hume was therefore right to maintain that there is no such thing as ‘spiritual substance’, nothing that a mind is, distinct from and underlying what it does. (1966, 221-222)

In the sentences I have omitted from this passage, Collingwood appears to suggest that the distinction between structure and function is merely contextual. In the final sentence of it, however, he affirms the manifestly ontological proposition that the mind has no structure. The meaning and the implications of this crucial distinction and crucial proposition constitute the main theme of this essay around which many subthemes will also be developed. I shall proceed as follows: first, I shall sketch a certain view of the ontology of the mind; second, I shall develop an ontological basis for the crucial distinction and make a first application of it to conscious mental states; third, I shall investigate further the meaning and plausibility of the crucial proposition with special attention to the allegation that the crucial proposition implies the impossibility of the scientific explanation of mental life; fourth, I shall set out an account of dispositional mental states to which the crucial distinction will also be applied; and fifth, I shall engage in some closing reflections concerning the connection of mind and time. I What is a mind? There are many different contexts in which this question may be raised. For the present purpose, however, an appropriate kind of

answer and also, I believe, the correct answer is that a mind consists of (1) conscious mental states, (2) dispositional mental states, and (3) those intentional objects. if there are any, that exist only when they are intended in a certain way. A sudden remembering, a feeling of pain, the hearing of a sound are all examples of conscious mental states. Conscious mental states constitute the “stream of consciousness,” bringing something at any moment at which one is conscious “before” the mind. Conscious mental states, however analyzed philosophically, constitute mind in the primary sense: a being who did not have any conscious mental states is a being with no mind at all. Examples of dispositional mental states are: holding that kicking dogs is wrong, desiring world peace, and knowing that light travels at 300,000 km/sec. A dispositional mental state is a mental state only because of its being a disposition to have certain conscious mental states as well as to engage in certain kinds of behavior. Probable examples of intentional objects that exist only when they are intended in a certain way are: a pain in one’s toe (as distinguished from the feeling of pain in one’s toe), a tickle in the middle of one’s back, images, and afterimages. Even more than those of the other two categories, the idea of this one presupposes certain philosophical judgments. But what justification I shall give for it in this essay will come only at the end of this section.1 Consider the conscious mental state of imagining that Winston Churchill is president of the United States. (As you read these lines you yourself are in that very state.) Suppose that it is my own imagining. Then the essentials of an adequate analysis of this imagining must derive from the following two considerations: (1) 1 know, directly and with certainty as long as my imagining lasts, first that I am imagining and not, say, doubting or perceiving or remembering; and second, that it is Churchill’s being president of the United States I am imagining and not, say, that e=mc2 or that the moon is made of green cheese. (2) We know, not from any single conscious mental state but from a number of them, that a person can imagine (or doubt or remember or, as in dreams and hallucinations, 1

One should not suppose that I am here speaking of what some call “intentional existence” or “inexistence” or the like: unicorns are not mental because they only have inexistence, that is, can be thought about. Unicorns do not exist at all and so are not dependent for their existence on conscious mental states.

8

perceive) what is not the case. The first consideration requires that the mode of my awareness as well as the differentia of its content be grounded in monadic, descriptive properties. (Purely scientific considerations lead to the same conclusions.2) The second consideration requires that the link between me (or my state or some feature of it) and that which I imagine be a polyadic, logical entity of a sort that can hold between something that does exist and, as in the example at hand, something that does not exist. The following analysis, in sketch only, satisfies these requirements: Some particular exemplifies the mode property of being-an-imagining and also the intentional property that-Winston-Churchill-is-president-of-theUnited-States, and that intentional property intentionally means that Winston Churchill is president of the United States. Intentional meaning is a primitive logical connection, the mode property and the intentional property monadic and descriptive. The notion of intentional meaning need concern us no further in this essay.3 As for the properties, we may note that neither is of a kind that characterizes any physical object and neither therefore is of the sort that is recognized in traditional empiricist philosophy of mind.4 Yet they are, on the view here advocated, the essential properties of this conscious mental state. This analysis I now extend, mutatis mutandis, to all conscious mental states. To give this extension some plausibility against a tradition that insists on (1) the essential role of images in imagining (and the same or similar entities in other kinds of mental states) and (2) the nonintentionality, even in the pre-analytic sense, of pain and other sensations, I shall indicate briefly how I would deal with each. As for images: they may or may not accompany an imagining. When they do, then there is, in addition to the imagining as already analyzed, another act of being directly aware of an image which image may or may not actually resemble that 2

I have in mind that a person’s behavior varies as to both the mode and the content of a person’s conscious mental state. The frame of reference of modern science requires that the properties by which that behavior varies be descriptive rather than logical (logical properties have no efficacy) and monadic rather than polyadic. The conclusion holds whether one is considering the properties of conscious mental states themselves or only their parallelistically-connected properties of the brain or central nervous system. For some more details, see Addis, 1982. 3 For a detailed defense of this notion, see Bergmann, 1960. 4 Brentano, Moore, and Broad are each, in his own way, a mild exception to this claim.

9

which is imagined. Take the example used above. In this case the mode property is being-a-direct-awareness and the intentional property is thatthis-particular-is-(whatever properties the image has). Then there may also be a third act whose mode property we might call being-a-judgment and whose intentional property is that-this-is-an-image-of-WinstonChurchill. The image then is something that is intended and not something that intends. So images are mental only in the sense of being causally dependent for their existence on conscious mental states (if they are) and belong therefore to the third, attenuated category of the mental. As for sensations: dialectical economy as well as careful phenomenological scrutiny permits one even in them to distinguish act and object. So, on the analysis here advanced, if one has a pain in one’s toe (whether one has a toe or not), there is some particular that has the mode property of being-a-feeling and the intentional property that-my-toe-hurts. The pain is something that is intended; it does not in itself, any more than an image, intend anything. If pains can, causally, exist only when they are felt, then, like images, they are mental only in the attenuated sense. And if they can exist without being felt, then they are not mental at all; and so for all sensations proper. This concludes my sketch of the analysis of mind. It remains only to emphasize that when henceforth I speak of the essential properties of a conscious mental state, I mean its mode property and its intentional property; and to repeat that conscious mental states are mind fundamentally: a being without them has no mind at all. II A watch has a structure and a function. Its structure is, or is expressed in, its shape, composition, and the arrangements of its parts. lts function is, or is expressed in, the motions of certain of its parts and therefore of the changes it undergoes. Furthermore, the functioning of a watch or, for that matter, its malfunctioning has its proximate cause in properties of its structure. Is the distinction between the structural properties and the functional properties of a thing absolute? Is it ontological? The answers to these questions can only be open on the basis of what has been said so far but are, as many ordinary examples would show, probably negative. So I

10

shall draw a somewhat different, clearly ontological distinction between two kinds of properties, a distinction that more or less corresponds, as will be apparent without argument, to the commonsense one between structures and functions. Among the properties that are exemplified by particulars the following distinction may be made: Some of those properties could be exemplified in a universe without time; the others could not. (‘could’ of course refers not to what is lawfully but only what is ontologically possible.) For convenience, let us label a property that requires for its exemplification a temporal universe a T-property and a property that could be exemplified in a non-temporal universe an N-property. Since it is not a distinction between properties that require time and properties that require space that is being made here, it follows that among properties that are exemplified by particulars (in contrast to those exemplified by other properties), the distinction between T-properties and N-properties is exclusive and exhaustive. How do we know for any given property whether it is a T-property or an N-property, whether it is a property that characterizes function or rather one that characterizes structure? The answer can only be that we know it directly by considering the nature5 of the property itself. This conclusion is forced by the facts that (1) we do know, at least in most cases, which kind of property a given one is, and (2) there is no other way we could know it except directly. As to the latter, we obviously could not know of a given property whether it is the one or the other by considering some other property; nor by considering or knowing some complex of which the given property is in fact a constituent since in the actual universe all such complexes are in time; nor, finally, by considering or knowing any other complex of which the given property is not a constituent. It follows that it is only by considering the property itself that we know, when we do know, that it is a T-property or an N-property. In our universe every particular is in time; some particulars are not, or at least are not given as being, in space. The reverse of this situation is ontologically impossible; it is altogether inconceivable that there should be 5

This use of ‘nature’ is both helpful (I hope) and logically redundant (I believe) since on it only a simple entity “has” a nature which is however nothing other than the entity itself.

11

a world in which some but not all particulars are in time. This impossibility shows a deep asymmetry in the connection of particularity to each of space and time. It may be argued by an objector that this asymmetry is simply one reflection of the fact that a particular without a duration is inconceivable. If so, then a timeless world itself is impossible and the distinction between T-properties and N-properties is empty in the sense that there are not and cannot be any of the latter. To this challenge I reply as follows: The possibility of a universe without change is uncontroversial. My critic, therefore, in insisting on the impossibility of a timeless world believes that there can be duration without change. Rather than argue this exceedingly difficult matter, I suggest, for the purposes of this paper, the following accommodation: let my distinction be conceived rather as one between properties that can and properties that cannot be exemplified in a world without change; for the mental properties in which I am here interested, the difference makes no difference. But having accommodated my objector to this degree, I shall continue to speak with my convictions of temporal and non-temporal universes. 1 now submit that we do know, by considering their natures, that the essential properties of mental states are T-properties. Mode properties and intentional properties can be exemplified only in a temporal universe.6 Moreover, a conscious mental state has no N-properties.7 6

If a conscious mental state can occur only in a temporal universe, then a being that is not in time cannot have a mind. (And a being, if it is to have any descriptive properties at all, that is in neither space nor time is impossible.) Some philosophers and even theologians, perhaps being more or less aware of this, took the step of asserting that to God belongs not any particular thought but only the generic property of being-(a)-thought. One thinks of Spinoza, possibly of Aristotle, and even of some who, incomprehensibly, claim to believe in a personal god. I am prepared to grant with Spinoza that, loosely speaking, (1) the universe as a whole is not in space and time and (2) that thought and extension are the two most general characteristics of the universe under which all particular properties “cluster.” But, strictly speaking, thought as a generic property can only be either a property of those particulars that also exemplify specific mode properties or else a property of the mode properties themselves. In either case, the generic property can only be exemplified by some constituent of a particular conscious mental state which itself can only be in time. So even in the sense of merely generic thought, a thinking atemporal being is impossible. 7 It may usefully be recalled that images, pains, and so on are never constituents of conscious mental states but rather the intentional objects of some of them and are

12

But what now can be said to another critic who, even while granting the existence of mode properties and intentional properties, simply denies that they are T-properties? Strictly speaking, I can do no more for this critic than for someone who denies, for instance, that blue cannot exemplify green. But I would stress to him my conviction that the essential properties of conscious mental states not only require duration for their exemplification but also, unlike some other T-properties, are capable only of a momentary, though perhaps immediately repeated, exemplification. It is because of this latter feature, or my belief in it, that I was willing so easily to accommodate my other critic who recognizes the possibility of a changeless but not a timeless universe. To my present critic, if he is not yet persuaded, 1 offer a different kind of truce: that he read what follows in the conditional (“If mode properties and intentional properties are T-properties and if conscious mental states have no N-properties, then . . .”). But he may yet be persuaded by some other considerations to be advanced in the next section concerning space and time. I further submit that the two propositions that the essential properties of conscious mental states are T-properties and that a conscious mental state has no N-properties are what lie at the heart of the crucial proposition that the mind has no structure. But do the two propositions entail the crucial proposition? And if they do not, is the crucial proposition nonetheless true? And if it is true, what is further implied about the nature of the mind and the lawful explanation of mental phenomena? Let us explore the issues these questions raise. III Abstractly speaking, there are four possibilities: (1) the mind as mind does have a structure after all only it is forever hidden from our inspection; (2) the body is the structure of the mind in the same sense in which it is the structure of its physical T-properties; (3) the body is the structure of the mind in a looser, possibly unique, and yet-to-be-specified sense; and (4) the mind has no structure in any sense. mental, if they are, only in the attenuated sense. Thus their N-properties are irrelevant to the present argument. Of course the intentional properties that intentionally mean pains, images, and so on are themselves T-properties.

13

The first of these possibilities, which is essentially that of a mental substance, requires there to be N-properties with which we have no acquaintance but which nevertheless are exemplified for as long as a person has a mind. Of such a possibility anyone in the empiricist tradition should be extremely dubious. And insofar as we are acquainted with no mental N-properties, one should also be dubious of the claim, which is made by some in the empiricist tradition, that each of us is acquainted with a continuant self. But let us look into this matter more closely. Consider a red, round spot moving across one’s visual field. How does one know, as surely one does know, that it is the same object that is red and round as the one that moves? How for that matter does one know that it is the same object that is red that is round? The answer can only be, in both cases, that it is so given to one: we see directly and immediately that the object that is red is the same one that is round and the same one that moves. The example at hand is intended to be a clear and simple one for being able to tell immediately that an object that has certain N-properties (redness and roundness) also has a certain T-property (motion) although in this case the possession of those particular N-properties is not likely to be relevant to the explanation of its possession of this T-property. In other cases we establish that different properties belong to the same object not by a single, direct perception of their coincidence but rather by a series of perceptions of property coincidence. The properties of being-theMorning-Star and being-the-Evening-Star may, despite certain complications, serve as an example. In general we may say that in the case of material objects there is no problem in principle in establishing whether or not it is the same object that exemplifies a certain property as exemplifies a certain other property. And so therefore there is no problem in establishing that an object of a certain function is the same as one of a certain structure and vice-versa. Consider again the mind in its primary sense of conscious mental states. In this case something that is given to us as having certain T-properties is not ever given to us as having also a structure. Nor are the properties of conscious mental states by any series of directly perceived coincidences linked to structural properties of the body or of anything else. Phenomenologically, the fundamental aspect of this situation is that conscious mental states are not given to us as being in space. This is not

14

the same as saying that they are given to us as being not in space, but it is sufficient to ground the impossibility of our ever, either by a single, direct perception or even by a series of perceptions of coincidences, being able to establish in this way that the mind in any sense whatsoever has a structure. And insofar as we are allowed to assume that (the particulars in) conscious mental states are not only not given as being, but in fact are not, in space, we may wonder at the very intelligibility of the notion of a mental substance insofar as it would be that of a persistent structure that is not spatial. Having rejected the possibility of a forever-hidden mental structure we are left with the body if we are to avoid the conclusion that the mind has no structure at all and thereby virtually grant the essential lawful unexplainability of mental life. We must now investigate the possibility that the body is, after all, either in the literal or in some looser sense, the structure of the mind. These possibilities may now be stated somewhat more precisely as follows: (1) the body, or some part of it, is literally the structure of the mind, only, as we have just noted, we are forever barred from perceiving directly that the object that is “grey matter” also has the property of, say, imagining that Winston Churchill is president of the United States. The brain has a structure and some of its T-properties are perceivable as being of the same object as its N-properties while other of its T-properties—those that characterize conscious mental states—are not perceivable as being of the same object as its structure; or (2) the structure of the body, or some part of it, while not literally the structure of the mind, is nonetheless (in each particular case) so intimately related to the conscious mental states of the person whose body it is that one may, loosely but reasonably, call it the structure of the mind. In this case the mind in its primary sense may be thought of as either a noncontinuous series of “objects” or as a noncontinuous “object” which exemplify(ies) only those properties that conscious mental states are given to us as exemplifying. If we now ask what reason we could have for supposing that the body is the structure of the mind in either of these two senses, the answer can only be that we have established or have reason to believe in a systematic correlation (not coincidence!) between and among certain properties of the body and those of conscious mental states. That, in a sense, is only to say

15

that we have reason to suppose that something is caused if we can find the causes.8 Still, there is ample empirical evidence—some of it being, as it were, casual and practical, learned by each of us from the experience we have of our own minds and bodies, some of it from science proper—for believing that every conscious mental state is lawfully correlated to a “state” of the body. For the purpose of rejecting the alleged lawful unexplainability of mental life due to there being no structure of the mind itself, it is not even required that the correlation be one/one but only, as it almost certainly in truth is, many/one from body to mind. For by knowing all the relevant laws one could from the information that the body is in a certain state deduce that conscious mental state if any at all that person was in. Of the two possibilities themselves—the body is literally the structure of the mind and the body is only in a weaker sense the structure of the mind—which is more plausible? How shall we choose? It is important to see that aside from one relatively minor consideration, we should expect no help from empirical science in deciding the matter. The question essentially involves the ontology of particularity and thus is philosophical. The empirical consideration derives from the fact that all the properties of a conscious mental state are T-properties. Suppose that we have identified the relevant properties of the brain or central nervous system to which those T-properties are lawfully correlated. Will those relevant properties themselves be T-properties or N-properties? If they were all Nproperties then for obvious reasons it would seem possible to maintain that the body is literally the structure of the mind. But if, as seems more likely, the relevant properties of the body include some T-properties, then one may have some ground for believing that the mind is a temporally discontinuous “object” or series of “objects” distinct from the body but whose structure is not in itself but rather in “its” body. __________ The dialectic of particularity is rich and complex. I cannot hope even to summarize it here much less reproduce or add to it. At best I can only 8

It is not necessary for this point that causation consist only of lawful connection plus various contextual factors—the correct view, in my opinion—but only that lawful connection be essential to causal connection.

16

recall some fragments, suggest a point or two with respect to the question before us. A human being, like a star or a dog or a coconut, is a commonsense continuant. For nearly all everyday purposes a person is regarded as and is described as being one “thing” persisting through time. John who runs today is the same person as John who was sick yesterday and he who tomorrow will remember his first visit to the dentist. Thus a human’s being a commonsense continuant is essentially a cultural or linguistic or perhaps legal fact, though not one altogether without natural foundation. At this level of discourse, mistaken by some philosophers as being itself ontological, the question of whether a human being with all its various properties is one “thing” or rather more than one contains its own trivial answer. If we choose rather to examine the matter ontologically there are four basic answers in the tradition to the question of what, in its particularity, a human being is; namely, (1) the one-substance view, (2) the two-substance view, (3) the momentary-particulars view, and (4) the bundles-ofproperties view. On the two-substance view, which I have already rejected, the mind is said to have its own, intrinsically mental structure. There are, I believe, due to certain features of substances themselves and most importantly the independence feature, other difficulties if not a certain repugnance in the two-substance view of human beings. Hence, it one were to be a substantialist at all, since the one-substance view is much to be preferred, one would hold that the body is literally the structure of the mind. But is that not just the one side of the coin which on its other reads that the mind is the form of the body? And with the distinction between structure and function—N-properties and T-properties—we grasp clearly why even in the human case the one substance should be thought of as something essentially physical even if it is also the subject of some irreducibly mental properties. Momentary-particulars views and bundles-of-properties views share a feature which permits their being treated in common with respect to the question whether the body is literally or not (if at all) the structure of the mind. A set of properties that are exemplified by the same particular or that occur in the same bundle must be capable at least of being given together to a perceiver. It follows that, with the possible exception of the so-called commonsensibles, no properties given to different senses (including

17

introspection) can be exemplified by the same particular or be in the same bundle. It further follows that those particulars or bundles that exemplify or contain the essential properties of conscious mental states are never the same particulars or bundles that exemplify or contain the properties of bodies. Thus it would appear that on these views the body can be the structure of the mind only in the weaker, non-literal sense. This conclusion however is somewhat complicated by the circumstance that on the views under consideration the body itself is constituted of a series of sets of momentary particulars or bundles. Except for the crucial spatial relations, which hold only among the bodily particulars or bundles, the same relations—the temporal and the lawful--that relate these bodily particulars or bundles to each other will also relate the mental ones to them. It is not important to choose here from among these various ontologies. For the record, I believe that only an ontology of momentary particulars and properties as non-Platonic universals is ultimately defensible. And on that ontology, as we just saw, the question of whether the body is literally or less than literally the structure of the mind gets somewhat blurred. It is not that the ontology itself is somewhere vague but rather that the basis of all unity over space and time and across the senses is that of certain relations holding and never that of literal sameness.9 IV Since dispositional mental states are mental only by virtue of their definitional connections with conscious mental states, it may he thought that they raise no special problem with respect to their relation to time. Consider belief. Since, as we speak, a person can believe something while asleep or even unconscious, belief is, on the face of it, an N-property; and so for all dispositional properties. This is, in my judgment, essentially correct, but it does not imply what to some it may seem to imply—that even in a non-temporal universe there could be at least rudimentary minds. That judgment, naturally, rests on a certain view of the nature of dispositions. Let us look more closely. 9

From this we must except the properties themselves which, as non-Platonic universals, may be literally the same in states of affairs as separated in space and time as can be.

18

According to a view I have developed elsewhere (Addis, 1981a),10 for something to have a disposition is for it to have a non-dispositional property that enters into a law of a certain, only partly specified content. For example, to be soluble is to have some non-dispositional property unspecified except that it is a property that satisfies the law-statement that whatever has it and is placed in water dissolves. (There could be several such properties and an object could have more than one of them.) The crucial features of this account then are (1) a disposition essentially (analytically) involves and therefore cannot exist without what is commonly called its “ground” and (2) a disposition essentially (analytically) involves and therefore cannot exist unless certain laws of nature obtain. Dissolving, let us suppose, is a T-property. Does it follow that being soluble into (the name of) whose definition ‘dissolving’ enters is also a Tproperty? The answer would appear to be that it does not follow since just as something can exemplify the property being red or falling without falling (and hence exist in a non-temporal universe despite the fact that falling is a T-property) so something can, despite the definitional connection, exemplify the property being soluble without exemplifying the property dissolving. This much, since it rests only on the commonsense notion that something can have a disposition without its ever being exercised, everyone can grant. But if a dispositional property is not made into a T-property by virtue of its possible realizations11 involving T-properties, may not its other features force that conclusion? The answer is negative provided (1) that the relevant non-dispositional property that is its ground is not a T-property, (2) that it makes sense to ascribe laws of succession to a non-temporal universe, and (3) that the mode of acquisition of a disposition and especially of a dispositional mental state is not considered essential to its being the disposition it is. As to (1): in general we may assume that the ground of a disposition will be an N-property, but if there are cases in which this is not so, then for this reason alone, such dispositions could not exist in a nontemporal universe. As to (2): if such laws are said not to hold, then there 10

The view is also stated and defended briefly in Addis, 1976. This may include the “condition” property as well as the “consequence” property— being placed in water as well as dissolving. 11

19

are no dispositions; but it would still be possible for the non-dispositional properties (that are themselves N-properties) which in a temporal universe would be the grounds of dispositions themselves to be exemplified As to (3): return to the case of belief. Since belief is a dispositional mental state, the form of its (perhaps only partial) analysis may be expressed as: to believe so-and-so is to have some non-dispositional property such that (it is a law that) anyone with that property will behave in certain ways under certain conditions and will have certain conscious mental states under certain conditions. We know that the conscious mental states that are, or would be, its partial realization can exist only in a temporal universe and are what make a dispositional mental state mental at all. We know also that the relevant non-dispositional property may itself be an N-property and so may be exemplified in a non-temporal universe even if it can never be realized and even if the relevant laws cannot really sensibly be said to obtain. Without these laws, as we just saw, it is not a disposition at all and so not a belief. But suppose the laws are said to obtain. Some who might accept the essentials of this analysis of dispositions generally and of dispositional mental states in particular as so far developed may insist that in the case of the latter, the occurrence of certain conscious mental states in their acquisition is essential to their characterization as dispositional mental states. I reply only that it is a matter of choice and not of discovery or “conceptual” analysis whether or not to include such factors. But we are now in a position to appreciate that even though dispositional mental states were said to be mind only in a secondary sense, even that may have a stronger implication than is desired when it is realized that in a changeless universe such dispositions, if they are said to exist at all, must necessarily be shorn of both their characteristic mode of acquisition and their most important kind of realization. For both of these involve the existence of conscious mental states. But what sort of property is the relevant non-dispositional property in a dispositional mental state? Of what, exactly, is it a property? For it is the property that is really there, the disposition itself—even those that are realized—being, in an ontological sense nothing in addition to the nondispositional property that is the ground of the disposition and the fact that that property is mentioned in a law of a certain sort. Since this nondispositional property is, in the case of belief at least, exemplified in a

20

person even when that person is asleep or unconscious, there are limits to the possibilities. There are, of course, only the obvious two: either a mental structure or the body. Since I have rejected the first. that leaves only the second. The properties that are literally exemplified in the mere having of a disposition are properties of the body. Thus should change cease, all that would remain of the human world is the body and its N-properties including the nondispositional grounds of dispositional mental states. When we recall that what makes a dispositional mental state mental at all is only that it may be realized in conscious mental states, even the exemplification of those Nproperties that ground those dispositional mental states in a timeless world should not prevent us from saying that it is also a mindless world. V A universe without time is one without conscious mental states. That that is so we know by considering the natures of the properties involved. Can we also ask why it is so? Or is it, given that the natures of simple entities are involved, simply an ultimate fact of reality? The answer is that while it is an ultimate fact of reality that conscious mental states require time, it may be possible to cast further illumination on the character of these and similar facts that are, ontologically speaking, not further explainable. While indubitably given to us as particularized, conscious mental states are not given to us as being in space. We saw earlier that the idea of a particular is inextricably bound up with space and time. Since nothing occurs that is literally instantaneous, the having of a duration is essential to the particularity of a conscious mental state. Yet a given act of awareness even while having a duration does not, unlike a sound, admit a division such that its temporal parts are separately acts of awareness or conscious mental states. There is a minimum duration for there to be, and not merely for us to be aware of, a conscious mental state. One may speculate that among simple properties only those of conscious mental states require a certain minimum duration and not just some duration or other for their exemplification. If that is so, the dependency of mind on time is even more intimate than has hitherto been conceived.

21

NATURAL SIGNS

A

n entity is a natural sign if by its very nature it represents some other entity or would-be entity.1 Many different kinds of things are said to represent other things, and in many cases it is recognized that the connection is purely conventional, in others that it is partly conventional being based in some sense on natural relations, and perhaps in yet others purely natural. My thesis is that a thought (as a generic feature of every state of consciousness) and a thought alone is, or contains as a constituent, an entity that is a natural sign. This could be denied by rejecting either the existential claim or the uniqueness claim: most who deny it, which appears to be the overwhelming majority of contemporary philosophers who worry about such matters, say that there are no natural signs. As for the uniqueness claim, 1 shall for the moment merely register my conviction that an entity that is or contains a natural sign is necessarily a state of consciousness, and indeed the common feature of states of consciousness. I While I have no intention of stating, and even less one of defending, a general “theory” or catalogue of the apparently manifold forms of representation, a few general comments will be useful in setting the frame of reference for discussing the specific, and for the philosophy of mind fundamental, notion of a natural sign. It is clear that the simple distinction between conventional signs and natural signs is far too simple. This may be seen by considering the following examples. The word ‘red’ considered as either sound or inscription (or any physical encoding) is said to be a purely conventional sign for the property of which it is the sign. By this is meant at least that the one entity—the word—has no inherent relation to the other entity—the

1

We can, apparently, represent non-existent things and facts to ourselves. This is the only time that I shall take note of that fact. Hereafter I shall write simply of the represented entity and the representing entity. But there is a place in section III where

property—that it does not also have to a number of other entities and which is the relation that is or grounds its representative function. It is a matter of human choice that the one entity is a sign for the other in the sense that we could simply decide that in the same sense in which ‘red’ does represent a certain color, it will hereafter be otherwise; and by so doing it would be otherwise. Other entities represent, manifestly, in part because of the inherent resemblance in some respect to the entity that is represented. What emphasis is to be put in such situations on the inherent similarities of the representing entity to the represented entity and what on the intentions, purposes, and other properties of human beings will depend on the context. To consider an example that will be used later: if I haven’t seen a friend for twenty years, I may discover that my image of him more greatly resembles his son than him when I meet them both once again. Was my image then of the son rather than the father? Was it of either or anything at all inherently? By assumption, my imagining was of the father, my friend. And if my child’s drawing more closely resembles a horse than the cow she insists she has pictured, do the crayon marks represent a cow or a horse? Or has the question been answered just by drawing a distinction between (1) the physical object (or kind of physical object) to which the representing entity has the greatest similarity, and (2) the physical object (or kind of physical object) that the creator of the representing entity intends or claims to be representing? And if the first, who shall say, independent of context and tradition, which similarities are relevant and which are to be ignored? And if the second, does that mean that what is represented by a picture or an image or a photograph is anything whatsoever its creator intends? Nor, of course, must we forget that pictures can also represent emotions, beliefs, ideas, and so on. About such matters there are long and interesting stories to be told, both historically and analytically. The only moral I wish to draw here—or rather to remind us of since it is by now a truism—is this: that resemblance is never a sufficient condition for one entity to represent another, and that the role resemblance plays in many different modes of representation itself varies along many dimensions Another kind of representation that is relevant to fixing the frame of the fact that we can think of non-existing things enters my argument.

24

reference is of a sort that one might reasonably call “natural” representation and which therefore I want to be especially careful to distinguish from what I mean by natural representation. It may be introduced by recalling Schopenhauer’s claim that while all other forms of art represent certain particularized goals, music represents the striving of the will itself (1966, 255-267). Let us focus on the thesis about music. 1 am not interested in the truthvalue of this theory nor in the idea of will that this is involved, but rather in the notion of representation that is assumed. Does Schopenhauer mean that music, by its very nature and therefore independently of human or any other minds, represents the striving of the will and so is a natural sign as I originally defined the notion? Perhaps he does, but more likely he means that music represents the striving of the will to the human mind, music itself being a product of the human mind. But I wish now to modify the theory in a way that makes it one that I am myself strongly disposed to believe but, more important, relieves us of Schopenhauer’s suspicious notion of a universal will and makes it easier to clarify the notion of representation that may here be involved. It may well be that, given the nature of the human mind (and possibly of all minds or even all lawfully possible minds or all actual or possible minds to which music has any meaning at all), a particular piece of music or part of it represents to the mind a possible state of mind. I do not say that the mind that is so represented to must be in that state of mind itself in order to have represented to it a certain state of mind. Nor do I say that every state of mind that the music causes in him or of which he is reminded is represented to him, any more than when one reads that “Upon learning of his rejection by the university, John grabbed a glass from the table and flung it to the floor” the state of mind that it causes in the reader, which might be amusement, or causes him to think of, which might be the reader’s own quite different reaction in a similar situation in the past, detracts in the least from the fact that it is anger and frustration that is represented (although in this case conventionally) by the sentence. 1 am suggesting that, whether we know it or not, certain possible states of mind are represented to us by music which may or may not cause us to think of or to be in that state of mind. But whether this suggestion has any merit or not—and it has many difficulties despite its virtuous attempt to sketch an account of the

25

fundamental appeal that music has for human beings, something to which little serious attention has been given—it makes use of a notion of representation that may now usefully be both generalized and weakened. Generalized, the idea is that, in the strictest sense and in the particular case, one (kind of) entity may for all humans represent just one other (kind of) entity because of our nature and in a way therefore that cannot easily be changed by choice. This is therefore a kind of “natural” representation, but because it requires the human mind as intermediary, it is not the purest and most fundamental form of representation that is conceivable. We may recall too that in the purely stipulatory sense any entity can be made to represent another entity. But the fact that Wagner’s Götterdämmerung can be stipulated to represent the chocolate bar in my refrigerator in no way diminishes the possibility that that wonderful music “naturally” represents certain possible states of mind. Weakened, the idea is that, except in the purely stipulatory sense, there may be, because of the nature of the human mind, certain ranges of possibilities that make for limits on what, to the human mind, can represent what. This may even be true of language itself. To the extent to which it may be true, the distinction between conventional representation and “natural” representation as presently being discussed may be more of the character of a continuum than is commonly supposed. There is, of course, an almost but not quite utterly trivial sense in which all representation (known to us) depends on the nature of the human mind, but that truism does not obliterate the distinction itself between conventional and “natural” representation. But how shall we understand “natural” representation even in this sense in which it is really a three-term relation among the entity that represents, the entity that is represented, and the human mind? I am not now asking what it is to be represented to, although this is a mysterious enough notion insofar as it is apparently being distinguished from being caused to think of or be in the state of mind that is represented. For although music may have some kind of resemblance to or even isomorphism with states of consciousness, we already know that such relations in themselves can never make one entity represent another. Indeed, when we are considering any non-mental thing—whether it be music or painting or language (considered as sounds and inscriptions)—as representing, there exists no

26

inherent characteristic of it that makes it represent something else. It requires in some sense the intervention of the mind whether through choice or by its nature to complete the conditions for one non-mental entity to represent any other entity. Thus we have not yet come upon any purely natural signs. But we may now strongly suspect, as Brentano said many decades ago (1973, 88-89) and as Chisholm (1958) argued more recently in an extended exchange with Sellars, that intentionality pertains nonderivatively only to the mental. I now take this to mean, minimally, that if there are (pure) natural signs, they exist only in the realm of the mental. II There are natural signs then if a state of consciousness has some constituent that intrinsically represents something else. Abstractly, that appears to leave three possibilities among simple entities as to what kind of entity could be a natural sign: particulars, properties, and relations.2 Furthermore, it may safely be said that if some complex entity such as an ordinary object or a fact or a complex property is intrinsically representative, it can only be in virtue of the nature of some one or more of its simple constituents; for there is not and cannot be any form of a fact or other complex that could make it intrinsically representative of anything else. (Wittgenstein’s notion that the form of the sentence in an ideal language must “correspond to” or “picture” the form of the fact—a crucial foundation of twentieth-century analytic ontology—is not a case of intrinsic representation. For, once again, a sentence considered as sounds or marks has no intrinsic representative nature at all; only as something by which humans mean do sentences mean, systematically speaking.) For the form of any mental thing or fact or other complex cannot be different from the form of innumerable non-mental complexes (with a possible, presently irrelevant exception) none of which is intrinsically representative of anything. Thus we must examine the simple kinds of entities in order to ask, first, 2

I shall not argue here but merely take for granted that the ontologically simple constituents of facts and things are particulars, properties, and relations. In this essay ‘property’ always means ‘monadic property’ and that whatever the linguistic form of the corresponding predicate may be.

27

which of them if any could conceivably be a natural sign and, second, whether or not any of them is a natural sign. A particular considered in itself differs from another particular only numerically. Since that is so, no particular is recognizable as just the particular it is apart from whatever properties and relations it in fact has.3 It follows that no particular could by itself mean or be about or intend or represent another entity. That leaves properties and relations. A property or relation is, by its nature as just the property or relation that it is, recognizable and therefore re-recognizable as such.4 That means that it does not differ only numerically from other properties and relations and that therefore each property and each relation has its own “individual” character. That is why various truths hold of some properties but not of other properties and of some relations but not of other relations in a way that no truths hold of some particulars but not of other particulars. For example, it is never meaningless to say or impossible to conceive of some particular considered apart from the properties and relations it in fact exemplifies that it has any (first-level) property or relation whatsoever. But the simple property of being-scarlet cannot be conceived as having the property of being-a-shape or being-a-pitch: its intrinsic character as being just the first-level property it is prevents it from exemplifying these particular second-level properties. The same goes for relations. Thus it would appear that some properties or relations, by being just the properties or relations that they are and not some other properties or relations, might be natural signs, and indeed that only a property or a relation could, at ontological base, be a natural sign. We shall see, as my argument progresses, that it is reasonable to suppose that only some properties and 3

This, of course, is the idea of a bare particular, that is, of a particular that, unlike the substances of another tradition, possess no characteristic so intimately that it cannot be analytically separated from each and every property and relation that it exemplifies. 4 In that sense only properties and relations have or are essences. No particular as such has an essence although, speaking derivatively, one could say that it is essential to a certain particular that it possess such-and-such property in virtue of its possession of some other property or properties. This ontological fact—that particulars have no essences—has profound implications for the theory of reference and, more important, the theory of intentionality and in particular for the issue of whether or not it is possible to refer to or, more important, to intend particulars directly or rather only by way of (individuating) descriptions.

28

not any relations are in fact natural signs. Now what reasons can be advanced for believing that there are in fact natural signs, that there are (exemplified) some properties (or relations) that by their very natures represent other entities? There are, as I see it, essentially three arguments—not necessarily of equal value or persuasiveness--that can be made in favor of the existence of natural signs. They are: (1) the dialectical argument that unless there are natural signs, there cannot be any representation at all; (2) the phenomenological argument that unless we are given entities that are, and can be immediately grasped as being, natural signs, there is no accounting for our certainty about certain features of our conscious states; and (3) the scientific argument that unless there are natural signs, we cannot explain the fact that the behavior of human beings varies systematically according to what they are aware of. Let us consider these arguments in reverse order for reasons that will later become apparent. III Consider two persons who are physically qualitatively identical in every lawfully possible respect except for whatever is lawfully implied by the fact that one of them is thinking about the Eiffel Tower and the other about the Great Wall of China. Both the materialist and the parallelist must insist that qualitatively different mental states always lawfully imply qualitatively different physical states, while the interactionist may deny it. But are thinking of the Eiffel Tower and thinking of the Great Wall of China really qualitatively different states of consciousness in the sense of involving different properties of the persons involved? It seems utterly obvious that they are, and so I think it is; but we must be careful especially in light of the views of certain philosophers who deny, or appear to deny, that there is anything at all, on the side of the mind, that makes a state of consciousness about some particular entity (or kind of entity) rather than another. Russell, for example, is explicit in denying that there are “states of mind” (1956. 169-173).5 For him, the only descriptive features (as con5

As Russell puts it: “At first sight, it seems obvious that my mind is in different

29

trasted with the merely logical features) of an intentional situation are in what he takes to be the relation—believing, imagining, doubting, or whatever—that one has to the object of one’s awareness; the only ones, that is, except for the descriptive features of the object itself. Sartre goes even further in claiming that all the descriptive features of an intentional situation are in the object of awareness itself (1957).6 This is, of course, part of what he means in saying that the mind is a “nothingness”: not only is the “content” of what one is aware of really just the object itself, as Russell would agree, but what we would ordinarily regard as our mode of awareness—imagining, desiring, wondering and so on—is also really just a feature of the object. The common feature of these two views is that there are no descriptive properties on the side of the mind that are essential to the full and proper characterization of any state of consciousness, and it is with this aspect in particular that I wish to argue. My present argument put succinctly is that the commonsense fact that behavior varies systematically according to what a person is thinking about requires that there be descriptive properties of occurrent mental states some of which properties satisfy the conditions of being natural signs. It is important to understand first that the connection between what a person is thinking of and that person’s thinking of it—the intentional connection-cannot possibly be a causally efficacious one even though, of course, one often also stands in a causal connection with one’s intentional objects even to the degree that being in that causal connection is the explanation at one level of being also in the intentional connection.7 Perception is the most obvious example. Indeed the intentional connection cannot really be a ‘states’ when I am thinking of one thing and when I am thinking of another. But in fact the difference of objects supplies all the difference required” (1956, 171-172). 6 I quite agree with Sartre that, in his language, there is no “ego” in “unreflected” consciousness, which is the main point of the book. But of course I disagree with his broader thesis that there are no contents to consciousness whatsoever. 7 So the view of Russell and many others that what distinguishes the modes of awareness—believing, imagining, perceiving, and so on—is a relation of a self to the intentional object is surely mistaken. The alternatives are to make such features properties of the state of awareness (the correct view, in my opinion) or, like Sartre, features of the intended objects. Had the fact that the intentional connection cannot possibly be a descriptive relation been more widely grasped, the ideas of intentionality and of (methodological) behaviorism would not so often have been perceived as incompatible. See Addis, 1982, for detailed argument on this point.

30

relation at all in the sense of an external connection that is causally efficacious. Unlike temporal and spatial relations, the intentional connection is not, as such, a relevant variable in any (true) law of nature. Less technically, a person does not have any effect on an (external) object simply by being aware of it whether through perception or imagination or any other way. But, even more than that, since we can think about objects, or at least kinds of objects, that do not exist, one’s connection to such objects, or kinds of objects cannot possibly be the immediate causal explanation of the variations in behavior that correlate systematically (though not always very precisely) with what a person is thinking of.8 One may reasonably conclude from these considerations that variation in behavior due to what a person is thinking of must derive most immediately from some property of the person. It is now reasonable further to conclude that, given that there is some relevant difference in property on the side of the person thinking according to what he is thinking of, this is also a difference on the side of the mind; that is, some property of his conscious state that uniquely “corresponds” to what he is thinking of. If interactionism is true, this must be the case since there is no lawful correlation of the coexistence type between a person’s conscious state and his bodily state and hence no possibility of this property’s being merely a property of the body. For the parallelist, it is not formally impossible, as far as the present argument goes, that the property be just a property of the body, that is, one that is not paralleled by any property of the state of consciousness itself. But if one wishes to preserve also the commonsense fact that what one is thinking of, as a state of consciousness, affects one’s behavior, then there must be some property of one’s state of consciousness that correlates uniquely with that of which one is thinking, the intentional object. Since the correlation, for reasons we have already covered, cannot possibly be a lawful one (or therefore a causal one), one is on the very verge of being forced to conclude that such properties of conscious states satisfy the conditions of being natural signs. I am largely persuaded by the force of this argument that I have labeled the scientific one. But I do not claim that it is, inherently, utterly 8

One might conceivably wish to maintain that non-existent objects can, in themselves, be causally efficacious. For the defense of a metaphysics that might underlie such a position, see Butchvarov, 1979.

31

conclusive. Let us turn then to the phenomenological argument for the conclusion that there are natural signs. IV When I am imagining (as I am as I write and you are as you read) that the star Sirius has ten planets, I do not know whether or not what I am imagining is the case. (Nor, of course, do I know when I am not imagining it.) But although I do not know whether or not what I am imagining is the case, I do know with utter clarity and certainty what it is that I am imagining. With the same degree of incorrigibility at the moment that some believe attends to knowing that one has a headache, anyone can have incorrigible knowledge at the moment that one is thinking of such-andsuch and not so-and-so or nothing at all. No evidence of any sort could override at the moment my conviction that I am imagining that Sirius has ten planets and not, say, that Winston Churchill is president of the United States or that the moon is made of green cheese or nothing at all.9 (The same utter conviction attends to my knowledge that 1 am imagining that Sirius has ten planets and not, say, perceiving or remembering it; but that is not my present interest.) How shall this fact be accounted for in a philosophically interesting way? My answer, as you will undoubtedly have guessed, is that I am, if I choose to introspect my conscious state of imagining that Sirius has ten planets, given a property that by its very nature is of or about the fact that Sirius has ten planets. I trust it may be agreed that if there is such a property of my conscious state and if I am given that property in introspection (and if, I suppose it should be added, one accepts the whole philosophical framework that lies behind such ways of thinking), then I would have accounted for my certainty in what it is that 1 am imagining even if one differently describes the phenomenon that I describe in that way. What are the possible alternatives to this account? As far as I can see, there are three possible ones if, in charity, we stretch the notion of possibility for the sake of the present argument. One of them holds that in 9

This has been denied, of course. See Rorty, 1965. Rorty has since adopted a different view that is even further from the truth. See the next note.

32

all or some cases in which a person claims such absolute certainty at the moment (so I don’t include dream telling, for example), one is not reporting a discovery about some non-linguistic fact about oneself so much as simply making it so by saying, so to speak. However this view might be qualified and refined, I reject it out of hand: even if, as some philosophers believe, my “inner” states can somehow be understood as “inner speech” or like “inner speech” there is still presupposed a difference between my “talking” and what I am “talking” about. It is unintelligible to maintain that to imagine something is simply to say or to be disposed to say that one imagines something.10 A second alternative holds that while I am directly acquainted with a property of my conscious state or otherwise of myself in introspecting my imagining that Sirius has ten planets, this property is not by its own intrinsic nature of or about the fact that Sirius has ten planets but rather only the property in me (it might be quite different in you) that I have discovered always to be there when I imagine that Sirius has ten planets and which, therefore, I learn to associate with that fact. The essence of this view is that I must learn by induction that this property is always present in certain kinds of situations and so is the property that, if one so insists, is about that fact; but it is no intrinsic feature of the property. This view, like the first, I would not even mention if it had not been held by someone, because it is so obviously flawed.11 The property in question cannot possibly be the property that makes my conscious state of just the particular fact that it is of, for if I am to discover a correlation between its presence and my imagining the fact that I am imagining, I must already know by other means what it is that I am imagining in order to discover the correlation. This is not to say that whenever I imagine something I do in fact know that I am imagining that thing or fact, for I 10

It is also, at the least, a paradoxical view that it is simply a matter of “social practice” that no one chooses to challenge our utterances about our own simultaneously occurring mental states. We would still want, in any case, an explanation of this highly invariant “cultural” fact. That the “incorrigibility” of such utterances is only a matter of “social practice” with no further philosophically interesting ground or explanation is the view to be found in Rorty, 1979, 174, and elsewhere. 11 Who holds this view? If he is not interpreted as an out-and-out materialist, Ludwig Wittgenstein may reasonably be said to subscribe to this position in his 1953.

33

may not actually introspect. But this view would have me learn to correlate the members of two sets when by initial assumption I can identify the members of one set only by the correlation. The true absurdity of this view may be brought out by observing that if it were true, any property of me whatsoever might be the property that is so correlated. For example, a pain in my elbow might always accompany my imagining that Sirius has ten planets. Would that property of me then be the property by which I have just that particular (kind of) conscious state? To the extent that this view has any merit at all, it is because it must be amended somewhat to embrace the third alternative. The third alternative, easily the strongest competitor to the theory of natural signs, holds that it is simply the fact that one is in an intentional connection to some object that enables a person to tell, with utter certainty at the moment at which one is in that connection, what it is one is thinking of. According to this view then, as I shall construe it, there is a particular that has the property of being-an-imagining and that also stands in some connection to the thing or fact that one is imagining, and that is all. Negatively, the essential claim of the third alternative is that there is nothing on the side of the mind, or at least need not be, in order to account fully for the certainty in question: it can be accounted for adequately by introspection of the mere fact that one is in the intentional connection to a given thing or fact. I find it difficult to deal with this view without some feeling of vertigo; but I believe nevertheless that it is utterly mistaken and derives from the misleading assumption, to put it rather roughly and somewhat paradoxically, that the “I” who knows what is being imagined is the same “I” who is imagining. I shall try to explain myself by beginning with what might seem to be an obvious objection to this alternative to the theory of natural signs. The “obvious” objection is that, in general, one cannot know that two terms stand in some relation or any other kind of connection unless either (1) one is independently aware of both terms of the relation and also of their being so related (as when I perceive that this table is about one meter from that chair) or (2) one or both of the terms contains some nonrelational mark or ground or sign of the other term of the relation and the fact that they are in that relation or connection. Hence, for example,

34

although I do in fact stand in some particular spatial relation to the Eiffel Tower at this moment, I cannot discover that fact by any examination of myself as I sit here in Iowa City even if, by the law of universal gravitation, the existence of the Eiffel Tower in just the place it is has some specific effect on my body. (Its effect is not unique and I would also need to know the laws.) Analogous reasoning would seem to reveal that if, according to a certain analysis of conscious states, to be aware of something is merely to be in some connection to it (and have the property of being-an-imagining or whatever), one could not possibly tell by introspection what it is to which one stands in that connection. But of course we can tell. I believe that this reasoning is entirely correct and demonstrates its desired conclusion. Since, by assumption, there is absolutely nothing “in” either term of the connection (the terms being a particular on the side of the mind and, let us say, a fact on the side of the intentional object)12 that shows or reflects or even gives the slightest evidence of the fact that those terms stand in that connection (presumably either can exist without standing in that connection whereas a natural sign merely by being exemplified does stand in the requisite connection), then being directly aware of only one term of the connection—the particular—could not possibly ground the certainty one may have of what it is one is aware of. “But wait!” I imagine a critic to say. “Have you not overlooked the crucial fact that is peculiar to the intentional situation? You are not (my critic continues) being asked to examine some other term that may stand in some relation or other connection in order to discover whether or not it does in fact do so; that would be impossible to be sure. Even your body is not, in the relevant sense, you yourself. But in the case at hand, surely you who in fact stands in that connection at the moment can tell immediately by inspecting that fact alone that you stand in that connection. The ‘I’ who knows that he stands in that connection is none other than the ‘I’ who stands in that connection, and that is quite different from, and crucially different from, all other situations in which some connection and 12

Nothing I say in this essay depends on my occasional assumption that only facts, as opposed to things, are the objects of awareness. Nor is any distinction between actual facts and only potential or otherwise non-existent or non-obtaining or false facts relevant to the present discussion.

35

knowledge of that connection are involved.” The commonsense fact from which we started is the apparent utter certainty that attaches at the moment that one is aware of something that it is just that and not something else or nothing at all (though “it” may be non-existent) of which one is aware. Dismissing the view that this is not really knowledge at all (and therefore not certainty either) and the view that such certainty derives from one’s acquaintance with a property whose correlation with the intended object is one that is learned by induction, we came to the third alternative. I now restate my argument to take into account this “sameness-of-self” argument against the theory of natural signs, and I do so essentially by presenting in finer detail the argument I made in the paragraph before last. If my certainty is to derive from what I am given in introspection, it is not correct to say that I am, or even possibly could be, given the fact that a certain particular—the self, for present purposes—stands in some connection to something else. The reason is very simple: there is at least one constituent of that fact that I most certainly am not given; namely, the fact that I am thinking about. What is as certain as the fact that I am imagining that Sirius has ten planets is that what I am imagining—that Sirius has ten planets—is not given to me, else I could answer the astronomer’s question about Sirius without difficulty. Since I am not given that term of the connection, I cannot be given the fact of which it is one term. This way of putting the matter may succeed in sidestepping the “sameness-of-self” argument. I permit myself two final comments or rather speculations about the view according to which there is nothing on the side of the mind that is intrinsically about what is intended. The first comment has to do with what may be seen as a defect of the theory of natural signs. I put it in the form of an accusation: that the theory of natural signs is a form of representationalism with all the epistemological and other difficulties that attend that discredited position; saying that there are natural signs “in” the mind whenever one is aware is to duplicate unnecessarily and even disastrously the objects of awareness. So the commonsense fact to be accounted for must not be taken care of by claiming that there are natural

36

signs of which we are (sometimes) directly aware.13 The proper reply to this accusation is to point out that it is based on a misunderstanding. Representationalism is the view that we are, in the first instance, aware of something “in” the mind which is then said to stand in some relation or other connection to something outside the mind. A crude view, sometimes ascribed to Descartes, has it that to see a chair, for example, is simply to be directly aware of an “idea” that is “in” the mind, which idea in fact resembles the chair. There is on this view no simple intentional connection between me and the chair. (This, in fact, is the fundamental objection to any such view and not the usual epistemological one about “knowing” that there are any physical objects: coherence of my perceptions and coherence alone is the only possible criterion of such knowledge and that does not depend on whether or not what I am directly aware of is mental.) The view I have been advocating holds that one may, if one chooses to introspect, become aware of a property that is exemplified by one in imagining that Sirius has ten planets and that intrinsically represents that fact; the imagining itself does not involve the awareness of that property but only its exemplification. So there are not, on this view and contrary to representationalism, two different connections involved—one of the self to the “idea” and another of the “idea” to the original intentional object—but only, if one thinks of the self as the particular (which is part of the source of the confusion) that of the exemplification of the property by the self which exemplification just is the awareness of the intentional object, the chair or Sirius’s having ten planets or whatever. So there is no cause here for suspicion of the theory of natural signs. In my second comment, I return an accusation: that the persuasiveness, such as it may be, of the third alternative derives in part from a lingering remnant in the minds of some philosophers of the traditional substantialist account of the mind. That account has two features that are of present relevance: (1) The self is essentially a particular and indeed an intrinsically mental particular. I say ‘essentially’ because the traditional notion of substance is of an informed or natured particular which cannot be analyzed, at least not without distortion, as a bare particular (or particulars) 13

Russell makes this complaint in his 1981, 160-161.

37

exemplifying certain properties and relations. In any case, this view seems to make any temporary properties (“accidents”) that a person might exemplify irrelevant to the being and the mental character of the self. (As you will also remember, on the Thomistic account of awareness, this self— not yet officially accredited as a substance in the 13th century—takes on the form of what one is aware of—an earlier version, one might say, of the theory of natural signs as developed in something like the form I am defending in the 14th century.)14 (2) But more directly to the point, a substance is active in a metaphysical sense, at least in some forms of awareness. This notion, not fully intelligible though it may be, has been a very powerful one in our tradition and even now continues to exert its sway, most explicitly in some contemporary philosophers’ notion of agency. The mind conceived of as active in the metaphysical sense leads, I suggest, to the belief that it needs no “help” from natural signs or similar entities in order to grasp its objects; it just reaches out and grabs them of its own power and nature. Be all this as it may, I conclude that the only satisfactory explanation of the commonsense fact with which we started is the phenomenological one—that we are given natural signs. It is time to turn finally to the dialectical argument which I have left to last not because I find it the most compelling but because it leads us naturally to another related matter. V The dialectical argument for natural signs also depends for one premise on a commonsense fact; namely, that there is representation at least some of which is conventional. But, the argument continues, there cannot be conventional representation unless there is also, at some level, nonconventional or natural representation; for deciding that or coming to agree 14

My direct inspiration, terminologically at least, is William of Ockham who writes of the “signum naturale” or, more often, the “terminus conceptus” and says that “a conceptual term is a mental content or impression which naturally possesses signification” (1957, 47). But my notion of a natural sign may also be recognized by many as very similar to Gustav Bergmann’s notion of a proposition (1960), And of course Frege’s notion of Sinn essentially fulfills my characterization of a natural sign even though it is an entity (if there were any) that is not a constituent of a state of consciousness and that would exist even if there were no minds.

38

that one entity—say ‘red’—shall represent another—being-red-presupposes that one is already able to represent each, not of course by words (which form of representation would be circular) but, ultimately, in some non-linguistic way that is non-conventional. The idea of the argument is that if every representation required a “third”—something that makes the one entity represent the other—then we have a vicious infinite regress. This regress will appear even if we countenance the kind of representation that may occur in music and elsewhere in which, because of the nature of the mind, one entity represents another to it even though there is nothing inherent in the representing entity that makes it represent uniquely the entity it does represent. For there is still a third involved that must somehow represent each entity to itself independently in order for the one entity to represent the other to it. So, concludes this argument, in order to avoid this regress, we must suppose that there is a kind of representation in which one entity, by its very nature, just does represent another without the presupposition or intervention of a “third” which must in turn represent both. How good is this argument? Formally, it is very persuasive and, I think, also establishes its conclusion. But of course it rests on a certain picture (as one says) of the human activity of language use and representation, a picture that many contemporary philosophers regard as outdated and wrongheaded. This picture, these philosophers say, treats language as a kind of “external” activity that is unintelligible without the presupposition of an “inner” realm that it reflects (but does not, in the sense we have been meaning, represent!) while that “inner” realm is supposed itself to be intelligible without the notion of language. It is the picture rather that is unintelligible, these philosophers continue, and so the dialectical argument for natural signs, since it rests on this picture for its own intelligibility, is to be discounted. Furthermore, some of these philosophers may say, once it is clearly understood that the realm of thought can be made intelligible only as something that occurs in a language-using being and that, in an important sense, is a consequence of, rather than a prior condition for, the use of language, it will be seen that appeals to the given in introspection and the like are fundamentally illusory; and while in what is called “introspection” something is genuinely occurring, it is a wholesale error to maintain that what is occurring is a series of non-physical, inspectable,

39

private events with properties that can, as it were, be “read off” in their intrinsic character. So say the critics. Let us see. It is said that schizophrenics “hear” voices that they may rather abstractly recognize as the products of their own disordered minds but which present themselves as the words of others. Jaynes speculates that our not-so-distant ancestors were in a similar but not exactly similar situation in that they too did not recognize their thoughts are their own but “heard” them as the voices of the gods (1976, book III, chapter 5). But since, by Jayne’s hypothesis, these ancestors did not have the sense of self that the schizophrenic retains even in his worst moments, they did not suffer: while the schizophrenic may describe his own suffering as being or as resulting from a loss of the sense of self (and who is entitled to quarrel fundamentally with this description?), it must be said that the very description presupposes at least a residual sense of self that is aware of the loss and that suffers therefrom. Be that last matter as it may, that a person’s own thoughts could present themselves as the words of others in a systematic fashion (or, less controversially put, that a person could systematically hallucinate his own thoughts as the voices of others) presents powerful evidence for the role of language in the formation and the very nature of thought in some sense or other. It is also said—I first read it in Ryle’s The Concept of Mind in a characteristically unfootnoted aside (1949, 27)--that it was only in the Middle Ages that people learned to read to themselves, earlier having to say or at least to mouth what they were reading as if they were unable to think what was being conveyed in writing unless, in speech, someone (in this case, themselves) were actually saying the words at the time. This fact also provides a significant evidence for the importance of language, and especially of spoken language, in the formation and nature of thought in some sense or other. On the other hand, people have now learned to read without mouthing the words (or even moving the larynx or being significantly disposed to move the larynx), and people who have never heard anything in their lives learn to speak and to write and presumably also to think, although the intrinsic character of their mental lives may be quite different from that of most other humans. One might, however, reasonably conclude that some

40

kind of language, whether written, spoken, sign, kinesthetic, or some other that first presents itself to one of the outer senses must be learned in order (1) to come to recognize thinking as something in the universe and especially one’s own thinking as just that, that is, to gain a sense of one’s self and one’s private mental life; and (2) to be capable of having certain kinds of thoughts that we generally regard as essential to a certain kind of intelligence—the ability to think of something in its absence, the ability to think of negative facts, the ability to think counterfactually, and so on.15 I mention these facts in order to acknowledge and even to insist on the profound role that language learning and language use, especially in all its social dimensions, plays in determining the character of our thinking in some very important senses. No doubt much more evidence already exists and yet other evidence will come to exist to support the same general conclusion. And when this evidence is combined with the facts about the way in which a child first learns the language of mental life and the way in which, very probably, our ancestors first introduced “mentalistic” terms into their discourse, it is not difficult to come to the conclusion that thought and even mind generally can be understood only by way of language and that, therefore (to return to the official theme of this essay), it is entirely a mistake to suppose that the representation that occurs in and through the use of language is dependent on the allegedly intrinsically representative capacity of this mind whereas the power of the mind to represent is not in turn dependent on language and its use. Not difficult indeed, but not correct either. That, of course, is my thesis. Provisionally, I venture to assert that at one level the view that language is primary in some philosophically interesting sense rests on a confusion between causal dependency and what we may call logical dependency. Or, equivalently, one could say that different senses of ‘understand’ or ‘explain’ are being confused, or one sense being pressed on us as the “real” meanings of these terms or as being of the greater philosophical 15

One may wish to exclude written language from this list for both causal and analytic reasons: the written language is not a language at all in the sense of consisting of momentary occurrences that are easily and naturally perceived as expressing what at the moment of their occurrence a particular person wishes to express; and it may well be that a person must learn a genuine language before he can either think in the sense I have been meaning or learn a written “language.”

41

significance, all to the conclusion, on the large scale, that language is primary and, on a smaller scale, that there are no natural signs. I shall try to establish these points and others in the following series of comments which may also be regarded as the steps of an argument. First, consider the following passage from a recent defender of the view that language is primary, both for its brief but useful piece of structural history and for its summary of the view under consideration: “For if thought is a representational system analogous to public language, then it cannot be appealed to explain how representational systems succeed in representing a world. Whatever the merits of agent-semantics as a component in an account of public linguistic performances the analysis of representation must be conducted at a level undercutting the distinction between the overt and the covert, between public language and thought. This, as I read it, is one of the main lessons of the Philosophica1 Investigations, and it is paradoxical on that account that the Investigations has also provided the main incentive for intentionalistically grounded theories of agent-semantics” (Rosenberg, 1974, 28-29).16 The argument of this passage, a summary of what has been discussed over several preceding pages, is formally invalid. In fact, Rosenberg’s main thrust against the view I am defending seems to be that it is only by reference to language that we can understand how it is possible to think of what does not exist. And since this essential feature of thinking, or at least of a certain kind of thinking, is thus otherwise unintelligible, we must take the representative function of language as necessary to “explain” the power of the mind to think of what isn’t there. Rather than respond directly to this argument, I say here only that this argument does seem to confuse the relevant senses of ‘explain’ to the issues at hand. For the sense in which it may be true that the ability to think of the non-existent can be explained only by appeal to language appears to be merely causal. Second, in this argument, at least at the level I am treating of it, it is or ought to be agreed on all hands that language use conceived as the overt production and perception of sounds, marks, and possibly other kinds of physical things is one set of things in the universe and that thought, even of 16

Rosenberg also mentions Bruce Aune, Peter Geach, and Wilfrid Sellars in addition to Wittgenstein as defenders of the view that linguistic representation and not mental representation is primary.

42

the highly conceptual sort upon which the opposing theorists tend to focus, and conceived as something “inner” and “private,” is another set of things in the universe. Understanding mind through language or language through mind then is not, in either case, assimilating the one fully to the other, maintaining that thinking just is the production and perception of marks and sounds, for example; there really are two distinct sets of phenomena in the universe which are not probably best described as mind and language or as thinking and using language respectively, but (if you like) “inner” or covert mental episodes and “outer” or overt production of sounds and marks respectively. Third, the excessive and self-conscious focus of the opposition on conceptual thought betrays a certain weakness in their account or, at least, a presupposition of a certain view of the nature of non-conceptual thought or mental life.17 It is not necessary for me (or them) to attempt any precise account of the distinction between conceptual and non-conceptual thought in order to grant the distinction and also to realize that the opposition is apparently committed, both historically and logically, to the thesis that non-conceptual mental life is also non-representational, at least in the sense that is relevant to the present dispute. There are easily crossed verbal bridges as well as certain causal and phenomenological facts that may seem to support this thesis while the thesis itself may seem to give strong support to the main contention of the opposition. But I called it a weakness. So I must take it apart. Conceptual thought is often described also as “symbolic” or “abstract” thought, and while the latter term may also imply a kind of direct grasp of something non-particular, both notions carry with them the idea that one is using symbols, that is, doing what one is doing representationally. Furthermore, whether we are thinking to ourselves or with our hands, so to speak, many of us are acutely aware that we either, in the latter case, write down, arrange, and rearrange the “symbols” until we arrive at what we meant to think or, in the former case, think “in” or picture to ourselves images of these written symbols. But the apparent strength of these observations to the cause of the view that language is primary can be shown actually to be a weakness by the following two additional 17

See especially Sellars, 1956; Aune 1967, chapter VIII; and List, 1981.

43

observations—one mainly on conceptual thought and the other mainly on non-conceptual mental life. (1) Although symbols may be necessary in some sense to what we call conceptual thought, their occurrence is not, except by trivial stipulation. analytically either necessary or sufficient for such thought to occur. For while it may be causally impossible for a person to think about, say, certain mathematical relations unless he also has certain images in mind, whether they be images of written symbols or of objects standing in certain spatial relations or of anything else whatsoever, the thinking about those mathematical relations cannot possibly consist in the having of those images. They are, apart from a mind that treats them as images of something—that is, a mind that independently intends the objects of which the images are images—merely objects among objects that may more or less resemble or have some isomorphism with or a shared form with the objects of which they are made to be the images as well as any number of other objects. As we saw earlier in the example of the friend who hadn’t been seen for twenty years, the thought is indubitably of the father even if the image (however having an image is analyzed philosophically) more nearly resembles the son. In short, what makes my conceptual thoughts what they are of has absolutely nothing to do, analytically speaking, with whatever images, symbols, words, or other conventional and semiconventional signs may occur concomitantly with or in relation to my thinking. I have already argued independently that the only thing that can account for those thoughts being what they are of, and be knowable as being of what they are of, is that each contain a property that is a natural sign. Thus it would appear once again that the occurrence of the conventional representation that is causally necessary to our ability to have certain kinds of thoughts presupposes natural representation. (2) But the defenders of the view that language is primary curiously seem to ignore non-conceptual mental life. I believe there is a certain basic confusion involved in how these philosophers treat matters at the outset that infects their entire argument. First there is a kind of identification of the conceptual with the symbolic and the representational. Then there is a view of their opponents’ position according to which, if conceptual thought is to be inherently representational, we have a form of representationalism. This view is then interpreted to mean that we are always in conceptual

44

thought looking at or attending to our own inner life. It is, in fact, a persistent mistake—blunder, I should say—to suppose that those who maintain that there is an inner life whose character can be known by introspection do or must in consistency also maintain that a person constantly is introspecting or that any form of awareness consists in a primary awareness of or attention to something “in” the mind. Ryle, for one example, makes this blunder in his general attack on dualism (1949, 14, and elsewhere). But the view defended in this essay is that in any form of awareness of something, whether it be conceptual thought or the lowliest form of perception, there is exemplified a property that by its intrinsic character is of or about that of which the awareness is said, preanalytically, to be of. I have also argued, if it needs any argument, that some beings who are aware also sometimes become aware of their awarenesses and may thereby also come to know that they are, say, imagining that Sirius has ten planets or that Winston Churchill is president of the United States. But in principle I may have any of those awarenesses without also introspecting and without therefore knowing that I have them. The dog goes through its life as a being that is aware of various things in perception (including dreaming), in sensation proper such as feeling pains and itches, and possibly even in memory and imagination. In doing so, according to the view I am defending, the dog exemplifies certain properties that are natural signs. But never once during its whole existence does the dog become aware of those properties it exemplifies that are natural signs; he is not aware of his awarenesses. The view I am opposing would apparently restrict the notion of representation to conceptual thought, then shift its origin to language by pointing to what is in fact a mere causal dependency of such thought on having learned a language. But representation in the sense that is relevant here occurs in all awareness: intentionality is a pervasive feature of occurrent mental life. We saw that what is peculiar to conceptual thought—the occurrence of symbols and images—does not and cannot possibly account for its intentionality nor therefore for its representative character in any interesting sense. Fourth, I have argued in effect that “language explains mind” only in an essentially causal manner and then only in certain respects. This idea may now usefully he summed up in the following empirical propositions. (1)

45

Our ancestors probably first used “mentalistic” language in essentially explanatory contexts. (2) Our children learn “mentalistic” language in essentially explanatory contexts. (3) Learning a language is a causally necessary precondition of being capable of conceptual thought. (4) Learning a language is a causally necessary precondition of being able to introspect one’s conscious mental states and hence also of knowing that one is a being that is conscious. It is therefore true that much of what is characteristically human about our mental life would not and lawfully could not exist unless we were also beings with language. But the following propositions are also true, and constitute the basis for saying that “mind explains language.” (1) Language when considered as sounds and marks has no intrinsically representative or intentional nature. (2) The representative functions of language, or language-like items that sometimes accompany certain conscious states, are unintelligible apart from minds that “already” intend, that are capable in some degree of thinking about that which pieces of language or language-like items are made to represent. (3) Conscious states can and do occur in beings with no language, and also in us with no apparent connection to the fact that we are beings with language. Thus we may say that “mind explains language” in a logical or philosophical sense: that while it is perfectly intelligible to imagine beings who have no language also having much the same kinds of conscious states that we have including introspections of other conscious states, it is unintelligible to imagine beings who are using language in all of its representative functions and who are also lacking in conscious states. The very notion of language as a representational system presupposes the notion of mind but not vice-versa. Thus I quite agree with Chisholm who, in his dispute with Sellars on this matter, maintained that “Thoughts would be intentional even if there were no linguistic entities. . . . But if there were no thoughts, linguistic entities would not be intentional. . . . Hence thoughts are a ‘source of intentionality’—i.e., nothing would be intentional were if not for the fact that thoughts are intentional” (1958, 533).18 18

The fundamental question here is, I believe, whether or not intentionality can, so to speak, reside in an individual occurrence such as a particular occasion of imagining oneself as President; or rather, as the functionalists and some other latter-day materialists claim, is really at root a property of a system and applies only derivatively to individual occurrences. For those of us who believe the former, it is incumbent to

46

It may he useful, finally, to draw an imperfect analogy in which, oddly, mind is compared to matter and language to mind. If what 1 call philosophical materialism19 is true, then matter can exist without mind but mind cannot exist without matter even though (assuming that parallelism is true) it is lawfully impossible for certain states of matter to exist without concomitant states of mind. Similarly, mind can exist without language but language cannot exist without mind even though it is lawfully impossible for certain states of mind to exist in beings without language. (The analogy is imperfect since the ‘cannot’ in the second sentence is logical while in the first sentence it is only causal.) And while there is a proper place even in philosophy for stressing that certain states of matter cannot exist without minds, one would naturally describe the whole relation as one of the primacy of matter to mind. Similarly, given the relations that hold between mind and language, it is natural to characterize the whole as one of the primacy of mind to language. I conclude that the dialectical argument for natural signs is, or at the least may be, a sound one: given that anything represents at all, there must be something that represents intrinsically. That something cannot be language. That leaves, in all plausibility, only mind. VI Many problems remain. Two of the most important are: (1) what is the nature of the connection between a natural sign and that of which it is the sign? (2) What kinds of facts—singular, general, negative, and so on—can one intend in the various modes of awareness? But even in the absence of answers to these questions, the arguments I have presented make a strong case for the existence of natural signs and for the proposition that every conscious state has at least one property that is a natural sign.

try to characterize exactly in what the intentionality of an individual act of awareness consists. 19 For further definition and discussion of the various kinds of materialisms, see Addis, 1975, 198-199.

47

PAINS AND OTHER SECONDARY MENTAL ENTITIES

T

here are, or very much appear to be, a number of things in the universe with the following characteristics: they are “private” to any person who is aware of them in a certain way; they seem very possibly to exist only if someone is aware of them in a certain way; they are not literally constituents of states of consciousness; they are not composed of atoms and molecules but may have some “physical” characteristics such as shape, spatial location, and color; and they are commonly referred to as something mental. Anything that more or less satisfies these conditions I call a secondary mental entity. Secondary mental entities fall into three moderately distinct subcategories. They are (1) sensations proper such as pains, tickles, and itches; (2) emotions proper such as anger, fear, and exhilaration; and (3) perceptionrelated entities such as sense data, images, and afterimages. Those of the third category are obviously more theory-dependent notions than those of the other two; and my calling the members of all three categories entities should be taken initially as only a convenient way of speaking and not as an endorsement or denial of any particular theory as to their nature or existence. In the empiricist tradition in the philosophy of mind, secondary mental entities have usually been taken to be the very essence, or at least paradigmatic, of the realm of the mental while states of consciousness have been largely ignored. Or, perhaps it would be more nearly accurate to say that many have supposed that states of consciousness can be exhaustively analyzed by secondary mental entities. For this lamentable state of affairs there are very good historical and very poor philosophical reasons; or so at least I have argued in detail elsewhere (Addis, 1982, 410-414). But for good or ill, it is easy to understand how taking mind to consist only, or paradigmatically, of secondary mental entities could facilitate the slide into materialism that has largely befallen our tradition. For secondary mental entities are much more plausibly, if yet wrongly, analyzed as “identical to”

physical things than are states of consciousness. States of consciousness are the primary mental entities. By this I mean, in part, that a being of any kind that has no states of consciousness is a being with no mind at all, however subtle or sophisticated (by external description) its behavior may be. According to the analysis of states of consciousness that I and others have defended elsewhere,1 every state of consciousness consists minimally of a particular (whether momentary or continuant does not matter here) exemplifying instances of two kinds of properties: an intentional property which indicates what the awareness is of or about, and a mode property which is the sort of awareness it is—an imagining, a perceiving, a sudden remembering, or whatever. The intentionality of the situation is captured primarily in the notion of the intentional property which, at least on my view, intrinsically represents what, preanalytically, the thought is said to be about.2 While I shall present my account of secondary mental entities through this analysis of states of consciousness, most of what I shall say does not depend on its details even as just briefly sketched. Dispositional mental states such as belief and hope are the tertiary mental entities. They will be entirely ignored in this essay. So much then for largely classificatory and terminological matters. Now to theory. It is widely believed that the existence of secondary mental entities shows that Brentano was mistaken in maintaining that intentionality is the “mark” of the mental because, it is said, such entities are mental but they are not intentional; that is, they are not about anything.3 The truth of the matter, I shall argue, lies somewhere between, but more nearly with Brentano than his critics. Indeed, the account that follows may be taken as a partial defense of the act/object analysis of states of consciousness in particular and of mental life in general. Ignoring numerous important differences among the various kinds of secondary mental entities (including the fundamental question of their very 1

See especially Addis, 1982; Bergmann, 1960; and Grossmann, 1965. See Addis, 1983, for an extended defense of the thesis that there are intrinsically representing entities. 3 For just one example of this argument, see Rorty, 1979, 22. Brentano’s own treatment of pain in the context of his view that the mental is the intentional is not, to be sure, transparently clear. Some crucial passages can be found in his 1973, 89-90, 2

50

existence), I shall proceed by taking (bodily) pain as representative of all sensations proper and more broadly of all secondary mental entities, Some of what I say would have to be modified mutatis mutandis to be true, even in my opinion, of the members of the second and third subcategories of secondary mental entities, but such requirements will ordinarily be obvious. On the theory I shall defend, there is a special dialectical advantage to taking pain as representative, for pain has a nearly unique feature that makes it more resistant on the surface to that theory. Thus, if I can make the account plausible in the case of pain, I shall have made it even more so for the other secondary mental entities. I propose simply to present a succinct summary of my account at the outset, then elaborate it and defend it by considering and replying to a series of likely objections. A (felt) pain (or a fear, or an image) is an intentional object. To feel a pain is to be in a certain conscious state that intends a pain. Thus, if I feel a pain in my right toe, there is on the side of the act the conscious state of feeling the pain, and on the side of the object the pain in my right toe. Like all act/object situations, this one involves on the side of the act the occurrence of a particular exemplifying an intentional property and a mode property. The intentional property may reasonably be referred to as thethought-that-my-right-toe-hurts and the mode property as simply being-afeeling. (It does not matter particularly that I find exactly the right word or even a less ambiguous word than ‘feeling’ for this mode property; it I am right, we all can discover by introspection what the property is.) Thus the pain as such is never something that intends but only something that is intended.4 Generalizing from the case of pain, I maintain that every (occurrent) state of mind involves a mental act, even though some mental things—secondary mental entities—are not themselves states of consciousness or even constituents of states of consciousness. Thus, Brentano was essentially though not literally correct in maintaining that and 1981, 16 and 19. 4 Even though (some) emotions are sometimes said to be “object-directed” or even to “intend” something, it is not correct to say that they, in the primary sense, are or contain intentional entities. When I fear something, there is, of course, in me a mental state that intends what I fear, but it is not the fear itself that does the intending. For details, see Addis, 1995.

51

intentionality is the “mark” of the mental. Apart from a general suspicion of mental acts, which are not here at issue, why should anyone resist this account of pain and other sensations as well as the other secondary mental entities? First objection: “If you are correct, it should be possible in principle for a pain to exist unfelt. For if a pain is merely an object of some mental act, then like all intentional objects it should be able to be, if it can be at all, without being intended. But it is absurd to suppose that there are or could be pains that are not felt; our very concept of pain precludes the possibility of unfelt pains by identifying a pain with the feeling of pain.” I reply: Claims about “our very concept” of anything whatsoever are very difficult to evaluate both for their accuracy and for their relevance since they are, I take it, essentially anthropological theses. I cannot speak, perhaps, for anyone else’s “concept” of pain or anything else, very strictly speaking; but I, for my part, do not have the conceptual apparatus my objector would ascribe to all of us. That however goes only to the lesser part of the objection. For the rest, I have much more to say. If a pain, considered apart from someone’s feeling it or otherwise intending it, just is a particular that exemplifies certain properties, then there can be no denying the ontological possibility of the existence of a pain that no one is aware of.5 For, when considered not simply as a particular but as one exemplifying certain properties, a pain is an ontological fact (where a fact, at least in the simplest cases, is a complex consisting of the exemplification of a property by a particular or of a relation by two or more particulars); and facts, if not facts alone, are ontologically capable of

5

So I am one of those people of whom Rorty complains in his 1979, 30, who make pains into particulars. But then, through a line of reasoning whose cogency escapes me, Rorty says that these particulars are “really” universals after all because, presumably according to these mistaken philosophers, these particulars are ones whose esse is percipi. That Descartes, Berkeley, Hume, maybe Russell, and certainly many others were confused in thinking of “ideas” or “impressions” or “sense data” sometimes as particulars and sometimes as properties and only rarely as what they must be, if there are any such things, namely, particulars exemplifying certain properties, does not justify forcing such a theorist into holding such things to be (Platonic) universals and then—even more dubious—offering this as an explanation why some philosophers think of such things as “immaterial.”

52

independent existence.6 But is it causally possible for a pain to exist unfelt? And is this really all the issue of the possibility of unfelt pains comes down to? Or is there not some stronger sense in which it is impossible for there to be unfelt pains? I shall try to do justice to the phenomenology of the situation (again, the anthropology of it I find of little interest) while nevertheless preserving a view that is both true in my judgment and also of some substantial, controversial content. I contend that not only is it ontologically possible for pains to exist unfelt, not only is it causally possible for pains to exist unfelt, but even more: there are unfelt pains! This theory, I believe, is the best explanation of certain phenomenological and other observational data, but has the additional virtue of making intelligible on the act/object analysis of pain here defended certain things that on most alternative accounts are not intelligible. The relevant phenomenological and other observational data may be represented by the following kind of situation: a friend has a moderately severe headache of which he complains when you meet him on the street. As you talk, the conversation turns to philosophy and soon becomes quite animated. Throughout this part of the conversation your friend, uncharacteristically, rubs his head and furrows his brow as if in pain. As the philosophical talk abates a few minutes later, your friend again complains of his headache but adds that he completely “forgot” it during the philosophical exchange. Thus the phenomenological data as reported by your friend are that he first felt his pain, for a while did not feel any pain, then again felt an exactly similar pain. The observational data available to you are (1) the verbal reports of the phenomenological data which, let us assume, you have no reason to doubt as to their accuracy, and (2) the fact that your friend exhibited headache-pain behavior throughout the meeting. I submit that the best way to understand this situation is to suppose that the friend continued to “have” his headache throughout the

6

By way of contrast, neither particulars as such nor properties as such are ontologically capable of independent existence. A particular must have properties and properties must be exemplified in order to be. For a thorough and incisive treatment of the issues of facts, independence, and ontological possibilities, see Grossmann, 1983, chapter 7.

53

meeting even though it was not felt by him part of the time.7 But what then is it to “have” a headache that is not felt? Is it some kind of intentional state? Perhaps some further distinctions need to be drawn. In order to do so, however, something must be said first, in a longish digression, about the very important notion of attending-to, a notion almost entirely ignored by philosophers of mind.8 Let us consider for a moment the case of visual perception or, more properly, of what does and may go on with respect to vision. In a typical visual situation one may distinguish (1) that which is in one’s visual field, (2) that upon which one is focusing, and (3) that to which one is paying attention. Sometimes one is not attending particularly to anything in the visual field whether or not one is focusing on anything; and sometimes one is attending to something in the visual field other than what one is focusing on. (Furthermore, with respect to each of these three aspects one may distinguish that which is given from that which is perceived, properly speaking; but here we may ignore this complication.) Is that which is in the visual field but upon which one is neither focusing nor attending nevertheless intended by some mental act? This is one question one may ask with respect to these distinctions. Another is: what is this matter of attending to something? For we all know that under exactly the same external conditions—that is, what light rays strike my eyes — I may attend to this and then to that and again to this. Before I attempt some answer to these questions, we may take notice of the fact that the phenomenon of attending-to is peculiar neither to visual perception nor to human beings. Among a cacophony of voices we are able to attend to only one, not necessarily the loudest; among several distinguishable features in the taste of a delicate wine, we can attend first to its bodiness, then to its dryness, and so on. Obviously, animals can and do make similar distinctions; the very survival of any animal depends on its being able to attend selectively to some aspects of what it perceives and 7

David Armstrong expresses a similar view in his 1968, 114. One can find some oblique comments about attending-to in two of the classics in philosophy of mind: Broad, 1925; and Price, 1953. And David Palmer, 1975, draws some distinctions that involve something like this notion. But there is little else, at least in the analytic tradition, known to this writer. Alan White has a book with the promising title of Attention (1964), but it turns out to be about how people speak rather 8

54

simultaneously to ignore others. One may reasonably suspect that the phenomenon of attending-to is a feature of all states of consciousness in all beings. But what is attending-to, ontologically speaking? Is it some additional mental act, or some additional mental property of some mental acts, or just a heightened “intensity” of certain properties that mental acts already have, or something else altogether? What, to concentrate on a particular case, happens when as I sit here focusing on the paper I allow my attention momentarily to fall on the cupboard on which I am not focusing and which is at the edge of my visual field? My “sensory input” remains unchanging by assumption, and I surely continue to focus on and to perceive the paper to which I was a moment earlier also attending. The answer, I believe, is either (1) that to every or nearly every mode of awareness there corresponds a “heightened” form of it—call it the supermode—which, when it occurs in a given awareness, makes impossible the simultaneous occurrence of the supermode of any other awareness, or (2) that there is a separate mode of attending-to, which act of awareness can intend only what is simultaneously intended by some other mode of awareness and which can at a moment accompany only one other mode to the same intentional object. Assuming that the supermode has a very close similarity to its corresponding mode, I find no way to choose between these possibilities on phenomenological grounds; and no other is relevant in this case. (I reject as being too “external” the idea of simply a “third” property—in addition to the intentional property and the mode property— of some mental act, just as I reject all other possibilities I can think of.) But, finally, to return to the case of pain: we are now in a position to see the possibility of maintaining that an unfelt pain is not necessarily one of which a person is totally unaware, although that remains a possibility too. In fact we now have three possibilities to consider in accounting for the data of the situation I imagined earlier: (1) the pain itself has simply ceased to exist during the time that the conversation was animated, possibly just because the person was not attending to it but instead to the subject of the conversation, but the essential physiological ground of the pain continued to exist and continued to produce the pain behavior; or (2) the pain than what they speak about.

55

continued to exist throughout the conversation although the person was in no sense aware of it; or (3) the person continued to “have” the pain, that is, in some sense continued to be aware of it much like something’s merely being in one’s visual field, but was not attending to it. This last possibility suggests that to feel a pain may best be characterized as both “having” a pain and attending to it, the latter of which, depending on the circumstances, may or may not be “forced” on one. (Having a pain and attending to it are, both separately and jointly, to be distinguished from disliking the pain. Some people enjoy certain even rather severe pains, and all of us seem to welcome certain slight ones especially in overcoming severe itches, for example; some people find that even rather intense pains do not bother them, and other bodily sensations such as itches, for example, are more discomforting than many pains. These facts provide additional grounds for distinguishing pain from the various modes of awareness of it insofar as one may have or feel or attend to or dislike one’s pains in various degrees and combinations. And of course these same facts make ridiculous those crude pleasure/pain calculi so beloved of some earlier ethicists.) Having brought matters to this point, and understanding clearly that there is no ontological or other purely philosophical reason why a pain should not exist without any awareness of it whatsoever,9 it may not be very important to choose from among these possibilities insofar as very little else seems to depend on which one of them, if any, is true. I am extremely dubious about the first possibility according to which there may be pain behavior without pain (except of course when it is faked or otherwise simulated). As between the second and the third, I find no ground for choosing, the only relevant kind being phenomenological, because by assumption, one could not possibly introspect as to the nature of an hypothesized awareness that is at the outer edge of consciousness, just as one cannot introspect directly as to the nature of merely having something in one’s visual field. And with this I take myself to have answered the first objection. Second objection: “If pains are to be distinguished from the awareness 9

Could a pain exist if there were no consciousness at all? Despite the probably universal inclination to believe not, I can think of no good reason in support of that belief when the modal term is ontological rather than causal.

56

of them, then either pains are where they are experienced as being or, like some other forms of awareness on the act/object analysis of mental life, being aware of a pain does not ensure that a given pain actually exists. The first possibility entails that pains may exist where no part of a body is as in the familiar case of the amputee who experiences pain “in” his no longer existing limb, an implication which is absurd; and the other possibility is absurd on its face: existing but unfelt pains are already bad enough for your theory; now you must admit also the possibility of nonexisting felt pains.” I reply: Before I tackle this objection directly, it is worth absorbing more fully the crucial fact that pains, like some other sensations and secondary mental entities, are usually experienced as being of a definite spatial location, as being in a certain place. This fact seems to be entirely unaccountable on theories that deny any distinction between pain and the awareness of it. Is my awareness of a pain in my right toe itself supposed to be in my right toe? Or is the pain itself, like awareness generally, not in space at all but merely caused by a physical disorder in my right toe or somewhere else? And if the physical cause is elsewhere, then just what, on these theories, does my right toe have to do with what is preanalytically described as a pain in my right toe? For I experience the pain as being in my right toe and do not (which is all that appears to be left for these theories) merely believe that the pain has its cause there. With this said, let me turn directly to the objection. It is true that on the general analysis of intentionality here defended there is no requirement that the intentional object exist or that, if it does exist, it be in every respect just as it is intended. (This distinction, raising as it does the non-issue of when an intended object is the “same” object as some real one, is not absolute but only the two ends of a continuum.) With respect to pains and other sensations, there are commonsense presumptions that they are not always where they are experienced as being as in the case of the amputee, that if someone experiences a pain then there is a pain that he experiences, and that nevertheless some pains are not “real.” Accommodating all of these presumptions simultaneously as it were is, I believe, neither possible nor necessary insofar as ordinary discourse sometimes distinguishes, sometimes denies any distinction between, and sometimes simply confuses, pain as something experienced on the one side

57

with the experiencing of it on the other. So I can proceed only by setting out the matter as I see it and then determining, if there remains any impulse to do so, how it fits these commonsense presumptions. Once more, however, I set out only alternatives that are consistent with the act/object analysis of pain here advocated. With respect to the case of the amputee, I at any rate have no difficulty in believing that the pain exists just where it is experienced as being even though no part of anyone’s body is in that place, and this whatever the physical cause of the pain is (which seems to me here, as in all other cases, to be entirely irrelevant to the philosophical analysis of the situation insofar as the physical cause of a given experience of pain might be anything whatsoever). There can be no basic ontological reason for denying this possibility, which in fact is the one I prefer on the ground that it is always preferable to assume and to believe that the world is as it is experienced as being unless there is some compelling reason for believing otherwise. I, for one, simply do not grasp the widely-assumed absurdity in this possibility; and in my answer to the next objection I may make it sound somewhat less absurd to those who think they do so grasp. It is also consistent with my general analysis to allow that pains are not always just where they are experienced as being. My specific reason (as contrasted with the general principle just adumbrated) for finding this alternative possibility somewhat less appealing than the theory that pains are just where they are experienced as being is that there then seems to be no criterion at all for deciding where a pain really is. A pain surely is not always where its physical cause is, if only because some pains have no direct physical causes. Of course, the awareness of a pain always has, in humans, a lawfully corresponding and temporally simultaneous brain state which, if there is no direct physical cause, may be regarded as the indirect cause of the pain itself. But the hypochondriac’s recurrent stomach aches are surely not in his brain. As for the matter of how, on the act/object analysis of pain, one can guarantee that a pain genuinely exists when a person is aware of it, the answer can only be that there is no such guarantee if that means that the analysis is supposed to show that it is ontologically impossible for someone to be aware of a pain and the pain not exist. But this lack of a guarantee applies, I believe, to awareness of any kind and thus to all forms of

58

direct awareness including the introspection of one’s own states of consciousness. It is the case, I emphatically believe, contrary to Rorty and others, that there are certain kinds of awarenesses that are such that nothing else one could discover at the time could or would override one’s belief in the existence of their objects. Thus the existence of a pain in my right toe, to which I stand in the form of awareness we call feeling, is as “certain” at the time as is that of a sudden remembering which I am aware of by introspection. More precisely, my “knowledge” that a pain exists in my right toe derives from the facts that I feel a pain in my right toe and that I introspect that feeling. The dog who is in pain never “knows” that it is feeling a pain since it never introspects its feeling of pain, and so is neither certain nor uncertain that the pain it feels really exists. Thus in the sense in which the objector raised this complaint, no guarantee exists and none is called for. Or, to turn the question around, how on any alternative account is it guaranteed that one actually is in a state of pain, however analyzed and whether taken to be physical or mental or some combination? Whatever measure of certainty comes from whatever form of awareness on any alternative account can just as well come from the awareness of pain and the awareness of that awareness, on my analysis. With this I take myself to have answered the second objection. Third objection: “But what then, on the act/object analysis of pain, and especially on your version of it that allows for the possibility that pains may exist where no body is, makes pains mental as they are commonly supposed to be? And if they are not mental after all, why are they both by commonsense and by most of the relevant philosophical literature taken to be, and often paradigmatically to be, mental? Are they not just objects in space and time on your view, and is that not a sufficient condition for being something physical?” I reply: It is important to see first that it is not, at least not directly, the dualism/materialism dispute that is at issue here. For, except for the eliminative materialists whose views may safely be ignored, everyone makes a distinction among existents between the physical and the mental even if some regard the latter as a subset or a “part” or in some sense an aspect of the former. The objection before us then is, I take it, that pains ought to be mental in the minimal sense in which both the dualist and the (non-eliminative) materialist countenance mental things, but that my act/object

59

analysis appears to make pains physical in a sense in which they are not also mental. My answer to this objection is largely contained in what I said earlier in the characterization of secondary mental entities. First, even on the act/object analysis of feeling pain that I have given, pains are “private.” That is, for me to feel a pain is for it to be “mine”, and no one else can (as a matter of empirical fact) be aware of that particular pain in that way. The “privacy” thus preserved is, in a sense, only causal rather than ontological (or “conceptual” or “logical” as some would say); but that, I believe, is as it should be insofar as, to speak very roughly, it is a broadly causal fact that states of consciousness tend to cohere in a way that allows for their groupings into “selves.” These groupings occur partly through the exclusive introspections of certain conscious mental states, partly through the exclusive objects of certain conscious mental states, and partly through other processes. But to the degree that pains are “private” and to the degree that what is “private” is also mental, one has here one ground for saying that even on the act/object analysis pains are mental. The reverse side of the “privacy” of pains is the fact that they cannot be perceived through the outer senses: they can be neither seen nor heard nor smelled nor tasted.10 Nor can they be felt in the same sense in which one feels a bug in one’s bed or a chair that one sits on. Furthermore, pains cannot be analyzed physically into atoms and molecules. These facts too provide reasons for saying that pains are mental if only because, while they are in space and time, they are not physical. In saying this however (and now the dualism/materialism issue does obtrude), I am assuming that the (simple) property of being-a-pain-of-such-and-such-a-character is not the same property as any physical property. In that sense, because that property is one that is known neither by introspection directly nor by the outer senses, it is not strictly speaking either physical or mental. It is known through being felt, a form of direct awareness; and while it is entirely proper and natural to characterize the mental formally so as to include secondary mental entities, it is important also to stress that these entities 10

Can any secondary mental entities be perceived through the outer senses? I think not, in the strict sense, even ignoring the fact that I more or less defined them so they cannot. But it is by no means utterly obvious that neither emotions nor sense data can never be perceived.

60

fall on the far side, so to speak, of the ontological divide that is the deepest among things at this level: those that are intentional and those that are not. Intentionality is the mark of the mental in the primary and preanalytic notion of the mental which is that of states of consciousness. But pains and other secondary mental entities are not constituents of states of consciousness. It will be seen too that this account apparently assumes that the particular that exemplifies the property that makes it a pain, while located in space and time and sometimes even being of a certain shape, is not physical. For if it were a physical particular, that is, if it also exemplified some properties known by the outer senses, then there is no reason why it should not admit of physical analysis. For some the idea of a non-physical particular is alone grounds for deep suspicion; and others may join them when these particulars are claimed also not to be minds or literal constituents of minds. For others the apparent necessity of this act/object analysis of pain will only reinforce their willingness to believe in non-physical particulars and therefore to imagine that those particulars that exemplify intentional properties and mode properties may also be non-physical. But that is another matter beyond the scope of this paper. Fourth objection: “When I look inside myself I cannot distinguish phenomenologically, as you say, between my pain and my feeling of pain. The act/object analysis locates my pains at too great a “distance” from me. In very severe pain when all else is blotted out, I am my pain; and your analysis cannot account for what is thereby expressed.” I reply: In many respects this objection (which goes to the special feature of pain I referred to near the beginning) has already been answered. I have pointed to a number of facts that lend support to a distinction between what is experienced and the experiencing of it; for example, that the pain is experienced as spatially located. But I do not wish to deny the initial force of other phenomenological data which seem to tend to the opposite conclusion. For when I “am” my pain, all distinction between me as one who experiences and that which I experience may seem to have disappeared. And no doubt many other kinds of experiences occur—in mystics, in schizophrenics, in users of certain drugs, and in others—that can be and are described as ones in which the subject/object distinction simply evaporates.

61

While not the subject of any such experiences myself, I nevertheless do not believe that in fact anyone has ever had a state of consciousness in which what is imagined to be the complete disappearance of these distinctions occurs; and I believe further that careful phenomenological scrutiny, when it is possible at all, would show that to be so. Certainly, when I “am” my pain, I still have no trouble distinguishing my pain from my awareness of it. But even discounting religious and other ideological motives for wanting to believe or being disposed to say that the self has disappeared into its object in some states of mind, there remains, I also believe, a distinctly philosophical motive, especially in the case of pain and other sensations, for thinking and being disposed to say that the act/object analysis is wrong. This motive has to do with a remnant of the old active/passive distinction according to which the mind is passive in sensation but active in cognition. In general, and especially in the case of severe pain, we cannot help but feel it; it forces itself upon us and we are powerless to resist. Like the culturally-determined distinction between essential and accidental properties, the active/passive distinction has no fundamental ontological basis; but that does not mean that it has no natural foundation. In particular, if the act is somehow identified with or simply associated with the active, one has an apparent reason, even with phenomenological support, for denying an act/object distinction in the case of pain. But it is merely a causal fact that we cannot begin and end our pains at will; and “I am my pain” is a poetic way of describing that unfortunate circumstance in which one seems unable to attend to anything other than one’s pain, in which nothing seems to matter except one’s pain, and in which one feels helpless before the forces of nature. Such situations are not, I submit, ones in which one cannot still distinguish what is experienced from the experiencing of it even if it does in some sense overwhelm one and, as it were, try to force a union with one. _________ I do not claim to have answered all possible objections to the act/object analysis of sensations and other secondary mental entities. Nor do I claim that no loose ends remain or that it accords in every detail with all the ways of thinking and speaking about pain and other sensations that occur in

62

multifarious ordinary contexts. But as a unitary account it does a better of job of explaining the facts than any alternative I know of. Behaviorist and functionalist accounts, as philosophical theories of the nature of pain and other sensations, are patently inadequate by simply ignoring what everybody knows to be their essential characteristics and the ones by which, eventually, we identify such sensations to ourselves.11 Other materialist accounts—central state or eliminative—in addition to their general defects seem unable to accommodate the fact of the experienced spatiality of some secondary mental entities. And on adverbial accounts according to which a pain, for example, is just a property of the mind, there seems to be no distinguishing what is experienced from the experiencing of it. But these are very general and roughly stated complaints, and I do not pretend to have refuted any alternative account to the one defended here. Three general comments will serve as a conclusion First, there is nothing in my view, or so I submit, that is inconsistent with either the presuppositions or the findings of empirical science, as might initially be supposed. As to presuppositions: the most important and relevant one is that the physical world is causally closed.12 Provided that, as I do hold, there is no occurrence or entity in the realm of secondary mental entities (including the experiencing of them) that is not “parallelled” by some occurrence in the physical world in a way that makes it lawfully impossible for just that physical state to recur without that mental one’s recurring, the presupposition is preserved. To put it the other way around: there is nothing in my view that requires that secondary mental entities or the experiencing of them be interacting things or events with the physical realm in a way that would make some occurrences in the physical realm not fully explainable except by appeal to those mental things or events. As to findings: I cannot think of any finding that might 11

On these matters it is useful to compare the views of a leading functionalist, Daniel Dennett, who, despairing of finding any coherent account of what even he seems to recognize as something that is there, tell us in his 1978, 228: “I recommend giving up incorrigibility with regard to pain altogether, in fact giving up all ‘essential’ features of pain, and letting pain states be whatever ‘natural kind’ states the brain scientists find (if they ever do find any) that normally produce all the normal effects.” Could there be a better example of the fundamental nihilism at the heart of functionalist theories of mental life? 12 For a detailed discussion of this notion, see Addis, 1982, 407-410.

63

plausibly be thought to pose a difficulty for my view, and I have already suggested that some of those findings reinforce what anyone can discover for oneself anyway—that a pain (or any other sensation or secondary mental entity) and the experiencing of it are always two and not one (or none, as the eliminative materialists and some others would appear to hold). Second, I have alluded in my reference to functionalism and elsewhere and otherwise assumed that we can and do know the nature of at least some secondary entities by experiencing them in a direct way that is private to each of us. This does not prevent me from holding with Wittgenstein and so many others that, in some good sense, “the inner stands in need of outer criteria.”13 Because I have stated my views on this matter in some detail elsewhere (Addis, 1983, 560-569), I shall here only repeat the point that even if, as is probably true, we come first to identify (in the sense of being able to recognize and possibly learn the words for) some mental properties by their outer criteria, it does not follow, nor is it true, that we are not directly aware of some of these properties or that we are unable eventually to identify them by such awareness.14 This is, of course, the simple and direct way of saying what more fashionable philosophers these days express by saying that pain, for example, is the “same” in all possible worlds or that ‘pain’ is a so-called “rigid designator”. Third and finally, I have made use in my account of at least one and probably more ontological presuppositions, and it may be desirable to make them explicit. A pain or other secondary mental entity, if it is indeed an entity that satisfies the theory advocated in this paper, is something that exists in time and in some cases also in space. Thus it is particular thing or event. Speaking more ontologically, it is a particular that exemplifies certain properties; and while the ontology of neither particularity nor universality is at issue here, it is easy to see that this theory fits especially comfortably with an ontology of momentary particulars and non-Platonic universals, that is, particulars that cannot survive change of quality and properties that are literally shared by, but incapable of existing independent of, those 13

For a “Wittgensteinian” account of some of what I call secondary mental entities, see Kenny, 1963. 14 For a detailed and very useful discussion of the epistemology of pain, see Baier, 1962.

64

particulars. Rejecting particulars as substances, that is, as natured continuants, makes it easier to accept the notion that a pain might be where no part of a body is and that some other secondary mental entities are not in space at all. But the defense of such an ontology, while perhaps systematically required, cannot transpire here.15

15

There is, as everyone knows, a vast literature on this topic. My own contributions to the dialectic can be found mainly in Addis, 1967, 1974, and 1981b.

65

INTRINSIC REFERENCE AND THE NEW THEORY

T

he new theory of reference, as Howard Wettstein has called it, while bidding well to become the dominant mode of thought among philosophers of language, contains at the same time, as the work of Wettstein himself reveals, the seeds of its own dissolution, at least within philosophy. This is one point I hope to demonstrate in the pages that follow. And while I refrained from putting ‘theory’ in double quotes in my title in the context of ‘new theory’, part of my argument will be that the new “theory” is not really a theory of reference at all; first, because it is not about reference to begin with (to put my thesis rather more strongly than I shall defend it, for the sake of emphasis) and, second, because it is not a theory so much as a description, better left to empirical anthropologists, of what external conditions must obtain in order that, as we speak, a person may correctly be said to have referred to something, or else a prescription as to how to use the word ‘refer’. These points, too, I hope to make plausible. But I shall continue to refrain from putting ‘theory’ in double quotes, and treat ‘the new theory of reference’ as a name that is, however, a disguised description for something like ‘the account that Howard Wettstein calls “the new theory of reference”.’ In a forthcoming book and an earlier article (Addis, 1989, 1983), I have defended the existence of entities of a sort that, following William of Ockham, I call natural signs. The theory of natural signs contains or suggests a theory of reference because, first, it is about reference and not the conditions external to it and, second, because it is a theory insofar as it defends the existence of entities of a certain kind that other philosophers would say cannot, or at least do not, exist. It would probably be agreed all around that the key feature of the new theory is that what determines reference cannot (always? ever? usually?) be found in a person’s state of consciousness, at least if such states are conceived as “private” to the person who has them in the sense of not being available to public observation. This will be said to be so especially

in the case of so-called ‘‘singular’’ reference, and so the philosopher must look ‘‘outside’’ the mind—whether to causal conditions or to context or to something else—in order to specify the, or the remaining, ‘‘conditions” of reference. The theory of natural signs, on the other hand, entails or strongly suggests that what, in the analytic sense, determines reference is entirely contained in a person’s state of consciousness. Thus the question of whether or not it is possible, and what it means, to refer to a particular person or thing that is not present to one seems to be the locus of the dispute about whether or not “meanings are in the head.” I shall argue eventually that in a certain sense there is no real issue here, once one makes a distinction between referring in thought and referring in language. I shall proceed as follows: first, I shall set out, in sketch only, the theory of natural signs and the theory of reference it contains; second, I shall consider an objection, as formulated by Michael Devitt, to any such theory— an objection that I believe underlies much recent philosophizing about reference; third, I shall develop some general, almost metaphilosophical, ideas about the issue of reference; fourth, I shall examine the writings of one of the new theorists, Howard Wettstein, as an instructive and almost too explicit case of what I believe to be the inevitable direction of the new theorists; and, finally, I shall have some concluding remarks. I William of Ockham says that a natural sign (signum naturale) or what he more often calls a conceptual term (terminus conceptus) is “a mental content or impression which naturally possesses signification” (1957, 47). I prefer to express the idea by saying that a natural sign is an entity that is intrinsically intentional, that is, by its very nature as the entity it is, points to or is about or of or intends something else. Fregean senses, whatever else they are (or would be, if there were any), would seem to be natural signs; and while the many philosophers who talk about “propositions,” especially in the context of what it is that people believe, might seem to be committing themselves to natural signs—that is, propositions as natural signs of the states of affairs that make them true of false—few of them are explicit enough, ontologically, for one to be confident of any such judgment. In Frege’s case, however, we may say that he does countenance

68

natural signs, but as entities in a “third” realm, neither mental nor physical, and that exist eternally and independent of particular states of consciousness. Frege appeared to believe (falsely, I think) that if senses were constituents of individual states of consciousness, then truth and falsity would depend in a relativistic way on individual minds in a way that would entail an unacceptable psychologism. I shall not attempt to argue here, at least not directly, that Frege is wrong. Rather, I shall sketch some of the arguments I made in detail in the aforementioned book and article why it is reasonable to believe that individual states of consciousness contain natural signs. After looking briefly at the ontology of natural signs, we will be in a position to connect the theory of natural signs with matters of reference and meaning. Apart from the general consideration that the theory of natural signs is better than its alternatives by allowing one to avoid commitment to such “things” as third-realm propositions or false facts or non-existing particulars or possible states of affairs, there are three specific and somewhat overlapping arguments in its favor. One of them, what I call the scientific argument, starts from the fact that differences in behavior sometimes stem from (or at least correlate with) differences in what people are thinking about. For example, when I am thinking of the Eiffel Tower and you are thinking of the Great Wall of China, there must be some relevant difference between you and me, and in you and me, that would causally explain the different answers we would give to the question of what we are thinking about. Any theory that holds that to think about something is merely to be in some relation to it, whether causal relation or uniquely intentional relation, makes such differences in behavior impossible to explain. The argument from the idea that there must be some property of a person that correlates uniquely with what that person is thinking of is not quite, by itself, an argument for natural signs, but it does establish the notion of there being some kind of property of me that varies according to what I am thinking about. A second consideration in favor of natural signs derives from the felt certainty a person may have at the moment of thinking about something as to what it is that he or she is thinking about. If I now imagine that Sirius has ten planets, I have utter certainty what it is that I am imagining as well as, not so incidentally, that I am imagining and not, say, remembering. And

69

so, again, it would seem that there must be some property of me the awareness of whose exemplification, if I care to introspect, allows me to have such certainty. This I call the phenomenological argument, but it may be noted that it is not a simple argument as to what is given to us, which “arguments” are rarely convincing, but rather one as to what must be given to us when we consider the nature of a certain phenomenon—the felt certainty as to the object of one’s awareness. Further considerations show, I believe, that this property of me must be a natural sign of what I am thinking about and not a conventional or learned sign. Those further considerations blend into the third argument for natural signs, a partly a priori argument to the effect that if there were no natural signs, there could not be, as there uncontroversially are, conventional signs. The idea of this dialectical argument is that if A is a conventional sign of B, there must be a “third” that makes or takes or regards A as a sign of B, ultimately by thinking of A as a sign of B. But that presupposes already being able to think of A and B. Thus starts a regress that can be stopped only by supposing that one can just think of something without having a conventional (or learned) sign of it, and that means being able to “represent” it to oneself naturally. This argument, especially, takes for granted a certain frame of reference concerning the relation of language and mind that presupposes the primacy of mind to language in all philosophically interesting respects. But I cannot stop to discuss this further here. These arguments for natural signs, combined with the hint above about one’s knowing that one is imagining rather than remembering, suggest that an individual state of consciousness consists minimally of the exemplification of two monadic properties—one a natural sign (or intentional property) that specifies what the state of consciousness is a consciousness of, and the other a mode property that specifies the kind of awareness it is. The link or relation between a state of consciousness and what it is of, then, is precisely the connection between a natural sign and what it is a sign of; and that connection, unique to the intentional situation but common to all of them, I call simply the intentional connection.1 It is 1

In Addis, 1989, I argue in the fifth chapter that the intentional connection actually holds only when its object actually exists. But this is a complication that can be ignored here.

70

this connection alone that can reasonably be called the relation of reference. It is common in the literature to make a distinction between descriptive properties and relations on the one hand and logical properties and relations on the other or, to speak somewhat more ontologically, between the world’s content and the world’s form. It is also commonly supposed, and I agree, that whereas descriptive properties and relations are or may be causally efficacious in the sense of appearing as relevant variables in true laws of nature, logical properties and relations are not causally efficacious. So if I now assert that the intentional connection is, or is like, a logical relation while intentional and mode properties are descriptive properties, it will be evident that on this analysis all descriptive features of the intentional situation are properties and none relations. It is this feature of the analysis that preserves the scientific respectability of the theory of natural signs as will become clearer in the next section. It remains here, however, to say something more about reference. As a “neo-Fregean” I hold (1) that reference can be understood only by understanding the nature of thought and its intentionality and that, therefore, anything reasonably called “semantics” whether of thought or language necessarily involves intentionality; (2) that, in the primary sense, reference is “achieved” by a person only when that person has a thought of what is referred to—whether in thought or language—which thought uniquely picks out the referent; (3) that thoughts are, or contain, natural signs, that is, entities of a sort that by their nature point to, or are about, something else; (4) that natural signs are properties exemplified by people and thus (contra Frege) are not “eternal” except in the sense in which, strictly speaking, no property is in time; and (5) that meaning, when it cannot be simply identified with what is in the present consciousness of the person who speaks, is nevertheless a function of actual or possible states of mind. Analytically speaking, it is clearest to characterize the situation with regard to reference and meaning as follows: In the richest sense, to think of a particular person or thing that is not present2 is to exemplify a natural 2

I restrict my discussion to these cases. I believe that Russell’s acquaintance/description distinction, which is smudged or denied by the new theorists and others in their accounts of singular reference, well accounts for the obvious phenomenological difference between perceiving someone or something and all other

71

sign of it alone, that is (to speak more commonly), to have in mind a definite description of that person or thing; to the extent that one lacks such a sign or description, one is not thinking of just that person or thing to the exclusion of all others.3 So too, to speak of a particular person or thing that is not present is, in the richest sense, also to exemplify a natural sign of it alone while speaking, that is, to have in mind a definite description of that person or thing. But now comes the asymmetry between thought and language, for, as we use the language, we are indeed prepared to say that a person may speak of (that is, refer to) a person or thing that is not present even if the speaker does not have in mind a definite description of the person or thing and even if, perhaps, the speaker has little or no thought of the person or thing in mind. In fact, little more seems to be required than that the person be barely conscious and be someone who has some knowledge of the language that he or she is using. But whatever the anthropological details may be, it is obvious that in such cases, we are concerned with the public meaning of the words, with what is “asserted” or “communicated” whatever the speaker may have had in mind. It is this kind of situation on which the new theorists focus and which, they believe, somehow presents a challenge to the neo-Fregean. Nothing seems clearer to me that the fact that, whatever the details, public meaning is a function of the thoughts that people do have, would have, and might have. If we choose to say that a person who is drunk or almost asleep or otherwise with little awareness of what he or she is saying nevertheless referred to Socrates when mumbling the words ‘Socrates was happy’, then granted that this “reference” is not determined by the speaker’s possession of a successful definite description of Socrates or even of the speaker’s possession of such a description under certain modes of awareness of persons and things. This is another reason for believing that there is no such thing as “singular” reference to particular persons and things that are not present. 3 Another account is that of Searle, who would locate the “missing” intentional content elsewhere in the mind—in what he calls the Background and the Network. Part of the reason I cannot follow him in this is that I do not share his view that beliefs and other dispositional mental states are literally intentional states. For more detailed discussion of Searle’s views on this matter, see the fifth chapter of Addis, 1989. For his account of the Background and the Network, see Searle, 1984, 65-71, 141-159.

72

unrealized conditions (although there is, necessarily, some condition under which he or she would possess it, for what it matters), it is nevertheless determined by the possibility of such a thought and implicit reference to that possibility. By that I mean that if anyone—speaker, listener, or anyone else—is pressed as to what or whom we the linguistic community mean by, say, ‘Socrates’, it will not suffice, as some seem to imagine, to keep saying ‘Socrates’ in an increasingly louder voice. We can fully satisfy ourselves only if we come to believe that we can produce a definite description that picks out Socrates alone (assuming that in historical fact there was exactly one person who did the things customarily ascribed to Socrates). Definite descriptions, as linguistic items, are usually characterized syntactically, but if they are to serve their semantic function(s), they must necessarily be conceived as the expression of actual or possible thoughts. And it is not possible to think of Socrates to the exclusion of everyone and everything else except by way of what in thought corresponds to the linguistic notion of a definite description.4 II Opponents of the philosophical theory (and commonsense belief) that what a person is thinking about is determined, to the extent that it is determined, by the contents of that person’s state of consciousness sometimes profess to be mystified by the very idea that, as Devitt puts it, “an intrinsic property of an object can determine its relation to a particular object external to it” (1984, 83). Devitt goes on to say that “This is no more possible for the relation of reference than it is for the relations of kicking, teaching, being taller than, or being the father of” (1984, 83). He links this objection to the theory of intrinsic reference with Putnam’s ‘‘Twin-Earth’’ argument which is based on the supposition that under certain conditions, two persons with identical thought contents would be referring to different things and thus follows Putnam in adopting the largely rhetorical device of 4

In the fifth chapter of Addis, 1989, I make a detailed argument why it is ontologically impossible to think of a particular except by way of description. The basic idea is that a particular, as contrasted with some properties and conceived as independent of the properties it happens to exemplify, is only numerically different from some or all other particulars, and so is not recognizable or thinkable as such.

73

labeling such theories as “magical’’ ones. Putnam’s arguments have been effectively attacked elsewhere; here I wish to focus on Devitt’s rather general complaint, for, if I am not mistaken, it underlies much of the resistance to the theory of natural signs. That complaint, once again, is that the very idea of a property that intrinsically determines a relation to something else is one of something impossible. It is, to use some ever receding metaphysical language, the idea of an internal relation, one that is in some sense determined by the natures of one or more of its terms, as opposed to a purely external relation which is one that things just happen to have to each other in the sense that while their being in that relation is, to be sure, causally determined by something or other including, probably, some properties of the things in the relation, in no case can one deduce or otherwise figure out that those things are in that relation just from knowing what properties they possess. More precisely, it is, in the case at hand, the idea of an internal relation to some other object; and yet more precisely, the idea of an internal relation not only to some other object but, to use Devitt’s language again, “to a particular object external to it.” And just why is it impossible that there be natural signs, that is, properties of a sort that do determine a relation to particular objects external to them? Devitt gives no reason at all except to list a number of other relations of which it is uncontroversially true that no property of any object so determines their particular exemplifications. But surely the fact that some relations are external relations is no reason at all for supposing that there cannot be a relation that is not an external relation. Nor, to my knowledge, does Devitt or any of the advocates of a causal “theory” of reference give any argument against the possibility of such a property; its impossibility is merely asserted (as in Devitt’s case) or otherwise presupposed. So we must look behind the scenes. I strongly suspect that what does lie behind the unargued-for assumption is the belief that for us in the empiricist, analytic tradition from Hume through Mill and Russell and the logical positivists, it is a settled matter that all relations are external. That is something that is imagined to follow from the “atomistic” way of thinking that characterizes our analytic modes of philosophizing and thinking about the world. To believe otherwise, some might say were they to give full voice to their assumption, is to be

74

“unscientific,” to revert to ‘‘holistic’’ or ‘‘organic,’’ not to say ‘‘idealistic’’ or even ‘‘religious’’ ways of thinking. Now I have no wish, for it would be contrary to my own deeply rooted impulses, to be ‘‘religious’’ rather than ‘‘scientific’’ or ‘‘holistic’’ rather than ‘‘analytic.” But I do wish to suggest that the idea of a natural sign and thus of a property that, in the disputed way, determines a relation of an object to another particular object is not only coherent in itself but accords fully with a scientific, analytic worldview and philosophical method. Part of Putnam’s reason for calling the theory of natural signs a “magical” theory is his notion that any such theory is linked somehow with those superstitions according to which by intoning certain words or thinking certain thoughts one can manipulate or otherwise affect, either directly or through the gods, those objects and events that those words and thoughts are about (1981, 3). Putting aside all suggestions of any commission of the genetic fallacy, it is nonetheless important to be clear that act theory, as embodied in a theory of natural signs, does not in any way presuppose or entail or lend any credence to the proposition that our thoughts have any effect on their objects or, even more important, that anything occurs in this universe that does not admit of a purely physicalistic explanation.5 Had this requirement, the satisfaction of which goes a long way to demonstrating the compatibility of natural sign theory with the scientific worldview and in particular with scientific psychology, been clearly kept in mind by Devitt, he would have seen at the outset that the relation of reference, the intentional connection, is radically different from the others he lists, all of which are such that being in such a relation to someone or something else does or may involve having a causal effect on it. The fact that the relation of reference is of a sort that, as such, no causal effects are involved suggests that this relation, whatever it is, is not a descriptive relation but instead is, or is very much like, a logical relation. This suggests further that the worry about internal relations may be somewhat misplaced insofar as the arguments from the analytic tradition directed against such relations have usually been against internal,

5

For details, see Addis, 1982.

75

descriptive relations 6 If the relation of reference is more like, say, beinglarger-than-by-five (between numbers) or being-identical-to than, say, being-taller-than or being-the-father-of, then it is much more plausible to say that it, like the first two, is determined in a particular case by the nature of one or more of its terms. Questions about the ontological status of logical relations and about whether or not, in the case of reference, the relation actually holds when the object of thought does not exist are, to be sure, very important ones; but the first task is correctly to have located the ontological category to which that relation belongs. Is there then any remaining argument against the possibility of natural signs, of properties that by their nature represent something else? I think not. Arguments against such properties based on their evolutionary improbability apply to all properties and those that are based on their causal idleness (which impotence I am prepared to grant, in a sense) only reflect their proponents’ commitment to an ontology of the basic—the belief that only the properties of basic science exist. But the notion of causal idleness and its connection with the question of whether or not natural signs fit into a scientific worldview is worth further remark. By causal idleness, as applied to a category of properties, I mean the characteristic of being such that whatever can be explained by the invocation of any of those properties (or, more precisely, their exemplifications) can just as well or better be explained otherwise. Explanation, in this context, means lawful explanation (of which so-called ‘‘causal’’ explanations are a subset). If both a dualism of properties and psychophysiological parallelism are true, as I believe them to be, then whatever can be explained—behavior, physiological states, social phenomena, other mental states—by anyone’s mental state can he explained just as well or better by the brain or other physiological states that ‘‘parallel’’ through laws of coexistence those mental states. (By ‘can’ I mean, of course, that the laws are such that those explanations are ‘‘there’’ 6

Even some descriptive relations are at least initially recalcitrant to treatment as purely external relations. Some examples are: being-darker-than as in the fact that red is darker than pink and being-higher-in-pitch-by-a-major-third as in e is a major third higher than c. For what is probably the best treatment of these kinds of cases and thus for the argument that all descriptive relations are external, see his paper “Synthetic A Priori” in Bergmann, 1967.

76

to be discovered.) Natural signs are mental properties, and while their invocation is reasonable in common explanations of behavior (as is ‘She reversed her steps in suddenly remembering that she needed to stop at the grocery store’), one may reasonably also believe that all such behaviors admit as well of full, purely physiological explanations. Thus, if it is part of the scientific worldview to believe that everything that occurs admits of purely physical explanation—and I believe that it is—then the theory of natural signs, provided that it is coupled with a parallelistic view of the connection of mental with physical properties, is entirely consistent with that critical component. I submit then that it is not part of the scientific worldview to believe that only physical properties exist, that is, to be an absolute materialist. Indeed, insofar as that worldview requires us to be, broadly speaking, empiricist in deciding what exists, we should be prepared to accept as existing whatever we do or seem to come across in our experience provided that there are no compelling reasons to believe otherwise (as with ghosts and gods). And because mental properties that are unknown to physics are among the things that we come across in our experience and because there are no compelling reasons to deny their existence, we should accept them as existing and be uninhibited in recognizing among them those that are natural signs. III What are the new theorists really up to? Are their “theories” really about the same phenomenon as those of the Fregeans and the neo-Fregeans? When the new theorists argue, both abstractly and by example, that it is often the case that when a person refers to someone or something, that person has no description or other content in mind that uniquely picks out the person or thing referred to, they take entirely for granted that, in such a situation, one can still truly say that the person “succeeded” in referring. Indeed, in general, it is customary to assume (and not only among the new theorists) that with respect to reference, a person either ‘‘succeeds” in referring or “fails” to refer in any particular case of what I suppose would have to be ‘‘trying’’ to refer, and that is all there is to it. But is the assumption correct? Or rather, what is the nature of the

77

assumption and how does it shape the new theorists’ accounts of reference? Suppose one starts instead with the idea suggested earlier that while granting that a person often does not have in mind a uniquely determining description of anything, to the extent that that is so one is simply not referring to any particular person or thing alone. If all that one has in mind is something that might be put into words as ‘that old Greek philosopher’ when asked whom he or she means by ‘Socrates’ (and ignoring the fact that with these two notions together one might get a successful definite description), why shouldn’t we say that that person has only to a certain extent referred to Socrates in that while that person has, as it were, conceived a relatively small class of which Socrates is a member, he or she has not really picked him out uniquely? The answer that many philosophers, including the new theorists, would give to that question is, I believe, that we know whom the speaker is talking about and, in any case, we would say that the person has referred to Socrates unless some other very special circumstances are known also to obtain (such as that the speaker believes that ancient Greece was a myth or that a philosopher is a kind of plant). So, assuming the conditions are “normal,” has the speaker really referred to Socrates, after all? What I want to suggest is that the answer to this question is not very interesting. Or, more precisely, it is of no philosophical importance once one sees clearly that to answer it one way or the other is to do no more than either to report on what one takes to be the linguistic habits of certain communities or else to stipulate how the word ‘refer’ is to be used. The important philosophical question is, rather, what is the nature of aboutness in its primary sense—the sense in which if there were no aboutness of that kind, there would be no reference and no “semantics” at all? The rest really is just linguistic anthropology or stipulation. Let us explore this idea in more detail. If one were to attempt to set out in a relatively uncontroversial way what facts obtain in a situation in which we say that a person has, or might have, referred to someone or something, we might come up with something like this: that person is conscious and probably is in some state of mind that is somehow relevant to his or her having referred to one person or thing instead of at least some other persons or things or nothing

78

at all; that person and more particularly the speech act by which he or she was said to have referred stand in certain causal connections to the person or thing to which he or she was said to have referred; and this person’s speech act took place in a certain context which may include the rules of the language, the particular histories of speaker and listeners, and anything else. States of mind, causal connections, contexts: these seem to be the “materials” out of which contemporary philosophers of language would forge their “theories” of reference. But many of these “theories” are not theories at all. When Putnam tells us, at great length (1981, 1-21), that a brain in a vat cannot really refer at all to most of the things that it (the brain) is “thinking about” because the usual causal connections do not obtain and no matter what the nature of the brain’s conscious states, what really are we being told except how the word ‘refer’ does or ought to lay on the world? It is certainly not a theory about how the extralinguistic reality is. Of course, the ruminations of the new theorists start from a certain account of how the world is in an interesting respect; namely, in insisting that many if not all states of consciousness fail to contain contents adequate to pick out a particular person or thing. That is a genuine theory about a philosophically interesting part of the world. The theory may be false, or at least vastly overdrawn, but it is not a piece of ‘‘concept analysis” so much as a piece of phenomenological analysis. But having convinced themselves of this theory, the new theorists feel called upon to deliver an account of what reference in language “really” is, given that it does not consist, presumably, in the having of certain states of consciousness while uttering certain words. And so we are told about causal connections and contexts. One can labor as one will with trying to specify accurately the external conditions that must obtain in order for it to be the case that, as we speak, someone has “succeeded” in referring to someone or something. This is essentially a matter of behaviorist anthropological linguistics, for once one clearly separates the two questions of (1) What are the external conditions that must obtain for it to be the case that, as we speak, someone has “succeeded” in referring? from (2) What is the nature of that kind of state of consciousness that is the “presupposition” for anyone to have referred, in thought or language?, one comes to see, first, that the new theorists have been largely confused in trying to have a “theory” of reference that would

79

somehow at the same “level” combine external features such as causal conditions and context with internal features such as contents of consciousness; and, second, that whatever the new theorists say about the alleged paucity of states of consciousness with respect to traditional theories of reference, this is largely irrelevant to what they should really be about. One new theorist, and perhaps so far the only one, who seems to have arrived at much the same conclusions is Howard Wettstein, and a closer examination of his journey will further advance my argument. IV In a series of papers over the last few years, Howard Wettstein has increasingly separated some kind of intellectual enterprise said to involve the “semantics” of language and deriving from the new theory of reference from what he takes to be the Fregean project of understanding the “structure” of thought (albeit, at least in Frege’s own case, through the study of language). And although, in at least one crucial respect, Wettstein is not in accord with most of the new theorists, a brief survey of his papers will be instructive with respect both to my quasi-historical prediction about the future of the new theories and my systematic claims about reference. One theme that has not changed in Wettstein’s thinking is his emphasis on what is asserted or communicated in contrast to what is thought when a person is said to have referred in speech to someone or something. His first published paper, ‘‘Can What Is Asserted Be A Sentence?” (1976) while more or less assuming the existence of propositions, something he comes later to question, argues that sentences are not what we assert while implicitly assuming that the “content” of what is asserted depends on the public meaning of one’s words and not on one’s individual thoughts. In “Proper Names and Referential Opacity” (1977), the emphasis on communication in thinking about reference remains strong while at the same time Wettstein begins his defense of what is one hallmark of the new theory—the account of singular reference. This defense continues in his third paper “Indexical Reference and Propositional Content” in which he undertakes to “defend an account of indexical expressions which sides with Mill against Frege with regard to its model of singular reference” (1979, 91). And in rejecting any Fregean-type account of reference, Wettstein

80

again sounds the theme that while a person may make a “complete and determinate” reference to something or a “successful” reference to something in uttering certain words, what goes on in the speaker’s mind may be quite incomplete and indeterminate. So reference cannot be wholly determined by the speaker’s conscious state. Yet again, in “Demonstrative Reference and Definite Descriptions,” we are told that “Russell’s theory [of reference] fails as an account of what is communicated” (1981, 247). And in this paper, following an attack on Russell’s theory when descriptions have either a “referential” or an “attributive” use, Wettstein begins to separate himself from most of the other new theorists by maintaining that it is context and not causal connection that supplies the additional elements for “determinate” reference, given that thought contents are not sufficient. In his next paper, “The Semantic Significance of the Referential-Attributive Distinction,” Wettstein again invokes context as the crucial additional element while still retaining the assumption that, in many cases in which no uniquely specifying description is operating and thus no uniquely specifying thought involved, nevertheless “determinate references are made and determinate propositions [are] asserted” (1983, 189). In a short paper, “Did the Greeks Really Worship Zeus?” (1984b), Wettstein addresses the ancient problem of reference to non-existents, while in “How to Bridge the Gap Between Meaning and Reference” he distinguishes three distinct accounts of reference—the causal theory, the intentional theory, and the context theory (1984a, 64). Advocates of causal and context theories, by locating some or all of what I just called the “elements” of reference outside states of consciousness, tend to agree with Wittgenstein that “an individualistic or agent-centered picture of language and thought is inadequate and needs to be replaced or at least supplemented by a picture that sees language as a social institution” (1984a, 65). This emphasis on the “social” aspect of reference becomes more and more pronounced as Wettstein begins to doubt that the new theorists and the Fregeans are even trying to answer the same questions. This, indeed, is the main theme of “Has Semantics Rested on a Mistake?” in which Wettstein proposes, after arguing that neither “linguistic meanings” nor anything else will account adequately for the ‘‘cognitive significance’’ of certain utterances within the resources of the

81

new theorists, ‘‘that we make a more radical break with Frege’s outlook. The new theorist should reject Frege’s adequacy condition outright” 1986, 200). (Frege’s adequacy condition is that the correct account of reference must explain how identity sentences with proper names can be both true and informative.) And in language that I also believe gets at the heart of the matter, Wettstein tells us that ‘‘Frege, unlike the new theorist, was not concerned—at least not primarily concerned—with what we might call the anthropology of those institutional arrangements which constitute natural language, the uncovering of the semantic rules that govern our linguistic practices” (1986, 200). Even if the new theorists cannot answer the questions that Frege raised about reference, Wettstein wants to maintain that “there is no reason to suppose, in general, if we successfully uncover the institutionalized conventions governing the references of our terms, we will have captured the ways in which speakers think about their referents” (1986, 201). Thus, too, ‘‘An account of linguistic meaning is no longer to be seen as an account of anything like what the competent speaker understands by his terms, but rather as an account of the practices he has mastered” (1986, 204). In his most recently published paper (as of this writing), “Cognitive Significance Without Cognitive Content,” Wettstein further develops this idea of the new theorists as engaged in an almost altogether different enterprise from that of the Fregeans while arguing that the latter are not, however, merely doing something different but are wrong or confused in many particulars, whereas the former can, after all, give an adequate account of the cognitive significance of some utterances without appeal to anything mental. As for the purposes of the new theorists, now conceived as anthropologists, “the aim of the anthropological semanticist is not, after all, to solve Frege’s problems” for “Reference . . . has little to do with the head of the speaker” (1988, 5). And it is appropriate to this clarified conception of his enterprise that Wettstein will no longer have any use for propositions: “a truly social and naturalistic conception will want to do without propositions” (1988, 14) for it is no longer a question of the thoughts of the person but of “the practices he has mastered.” This brief sketch hardly begins to indicate the richness and subtlety of much of Wettstein’s writing, nor does it make clear that most of his detailed arguments are directed not at Fregeans but rather at other new

82

theorists. Its primary value in the larger scheme of things, as my review has tried to show, is the emergence of the conception of the enterprise of the new theorists as an anthropological one that has everything to do with externally observable behavior of human beings and little or nothing to do with what goes on in their minds. This, in my judgment, is exactly as it should be. But, at the same time, I believe that Wettstein’s admirable journey has not reached its end, and that certain confusions hinder its completion; and so I shall now argue. In one sense what I want to argue is that as the role of consciousness is an ever-retreating one in the anthropological enterprise that the new theorists are engaged in, so will the notion of reference, with which their challenge to the Fregeans began, itself begin to recede. This is a partly predictive, partly analytic chain. As to the latter, the idea is that if Wettstein and the other new theorists were to carry through fully on their enterprise as anthropological and as treating language (I now pointedly do not say ‘‘reference’’) as ‘‘social” and ‘‘naturalistic’’ (where ‘naturalistic’ apparently means something like: ignoring consciousness, something profoundly “unnaturalistic’’ in another sense), it would come to be seen that while the word or, better, the sound (or mark) ‘reference’ is part of the anthropologist’s data because it is a sound that is produced by certain humans, reference itself, at least as preanalytically conceived as some kind of connection between a person and something else, will have disappeared from the scene. To put the notion a bit more precisely: the student of observable human behavior will not find anything reasonably called reference; there is no relation in ‘‘nature” (in their sense) that connects words with things. The reason is that, strictly speaking, there are no words in “nature”! There are human organisms behaving in certain ways including making certain sounds. To be sure, the anthropologist of such behavior may also discover that those sounds arc produced under certain causal conditions and in certain contexts, and one could say that when one of these organisms does produce a certain sound or sequence of sounds under a certain causal condition or context that that organism has “referred” to something. Perhaps the reason for saying that would be that some other organisms of that species sometimes make certain sounds that include the sound ‘refer’ in the “same” situation. (We cannot, of course, if we are strict behavioral

83

anthropologists, say that those other organisms were talking about the first one.) Now, despite what may sound almost like a parody of what a strict behaviorist anthropologist (where ‘behaviorist’ means here: really ignoring the realm of the mental and relying only on what is strictly observed) of human soundmaking should say about the subject matter, I have no objection in principle to proceeding in that way. If one is going to approach the subject of human behavior, including “linguistic” behavior, scientifically—that is, by allowing only the observable and what can in appropriate ways be understood through the observable but excluding the mental—this is the approach one should take. That is what it would really be to look at “language” from a “social” or “naturalistic” viewpoint. And I, for one, do not doubt the possibility or the value of such inquiry. But Wettstein and the new theorists want to have it both ways: on the one hand, they reject, or tend to reject, the role of consciousness and the intentional connection in understanding philosophically what, at root, it is for a person to refer to something and instead look increasingly to discovering the external conditions under which we say that a person has referred to something, Brushing aside the important fact that such an enterprise should be a purely empirical and inductive one, I observe that, on the other hand, the new theorists also still wish to preserve the notion of reference in their cogitations in something like its original, relevant sense of aboutness. Otherwise, there is little point, if any at all, in continuing to say that they are giving accounts of reference. Let Wettstein and the other new theorists say very clearly (1) whether or not, if a certain population of humans were suddenly to become devoid of consciousness altogether but continued to behave as before, they would be sometimes referring to people and things; and (2) if not, why not. In short, is there really any such thing as semantics for the “anthropological semanticist’’? V Wettstein prefaces his “Cognitive Significance Without Cognitive Content” with a passage from Walker Percy’s “The Delta Factor” about how an intelligent Martian without language would regard human speech; and while Wettstein a couple of pages later says that “Making the

84

anthropological semanticist a Martian—that is, one who is not a participant in our, or perhaps any, linguistic practices—raises its own problems that, for the present, I would rather avoid” (1988, 3), he clearly regards the thrust of the passage approvingly. Here is most of it: Imagine how it must appear to the Martian making his first visit to earth. Let us suppose that he too is an intelligent being whose intelligence has, however, evolved without the mediation of language, but rather, say, through the development of ESP. . What is the first thing he notices about earthlings? That they are forever making mouthy little sounds—clicks, hisses, howls, hoots, explosions, squeaks—some of which sounds name things in the world and are uttered in short sequences which say something about these things and events in the world.

Instead of starting out with such large, vexing subjects as soul, mind, ideas, consciousness, why not set forth with language, which no one denies, and see how far it takes us. (Percy, 1975, 11-12)

Is the first thing the Martian would notice about humans that they make sounds “some of which sounds name things in the world and are uttered in short sequences which say something about these things and events in the world”? Is it even the second or the third thing the Martian would notice? Indeed, would the Martian ever notice this thing? I have already suggested that, in the strict sense, there is nothing in what is observed that is the naming of things by sounds or the saying of things by sequences of sounds; there are just the sounds and their sequences. But would the Martian eventually deduce or otherwise reasonably conclude that, as we might now put it, these sounds have a semantics; and if so, on what basis would the conclusion be reached? Why would the Martian ever suppose that there is anything “semantical” going on at all? The Martian has no language; the Martian observes only the outward behavior of soundmaking by the humans. Surely the only way the Martian could come to conclude that there is something “semantical” going on would require it first to have come to conclude that the soundmaking of these humans is part of a language, that is, an instrument for purposeful communication among themselves. Although it is not completely clear how the Martian would arrive at that conclusion, it seems clear that only if that happens will the Martian ever arrive at the further or more specific notion that there is anything ‘‘semantical’’ involved, that is, at some notion of aboutness. But where will the Martian’s notion of aboutness have ever

85

come from in the first place? Not from its own language, for it has none. The obvious answer is that it gets it, if it has it at all, from the intentionality of its own thoughts, and nowhere else. And if our intelligent alien does conclude that we humans are speaking a language and not just making sounds, it will naturally understand this phenomenon as one in which the meanings of the sounds are derived somehow from the intentionality of the thoughts of which those sounds are somehow the expression. One “hears” someone refer to something not by hearing sounds or knowing the causal conditions that led to producing the sounds or taking notice of the context in which the sounds are produced but by taking the sounds as expressive of the speaker’s actual or possible intentional state. In everyday life we all know that very well and could not possibly believe otherwise.

86

THE ONTOLOGY OF EMOTION

I

t seems that some emotions are about something as when, for example, a child is afraid of the dark, while other emotions are not about anything as when, for example, someone is just happy. Philosophers who think about the emotions have generally responded to the fact of this seeming in one of two ways: either they have denied what seems to be the case by affirming that every emotion is about something, or they have accepted what seems to be the case as so and gone on to assert that this fact shows the act/object analysis of at least some states of consciousness to be false. I shall argue that, while it may seem that one or the other of these responses must be correct, both are wrong. I shall not argue for the proposition that some emotions are and other emotions are not about something or for the proposition that consciousness is always consciousness of something. The thesis I shall defend is the compatibility of these two propositions. I do say, however, that both of them are ones that a philosopher of mind should want to hold if at all possible—the first because it is the way things seem, and the second, if not also for that reason, because it seems impossible to understand what a state of consciousness that is not a consciousness of something could be. What my compatibility thesis requires is a certain account of the relation of emotion to intentionality. The general idea of this account is that the intentionality of emotions, when it exists, is quite unlike the intentionality of conscious states proper. This clearly suggests that an emotion as such is not a conscious state; and indeed the most important feature of the theory of the nature of emotion I shall defend is that an emotion is not a way of being aware of the world or even, in a strict sense, a way of feeling, but instead only an object of a certain kind of awareness. This further suggests that the intentionality of an emotion is not, as it were, inherent in the emotion itself but is somehow derivative from the intentionality of an attendant conscious state. But it will be well to canvass what have gone before as philosophical theories of emotion in order to see more clearly what the meaning and significance of this new theory is.

I Philosophical theories of emotion I shall classify (somewhat following Lyons1) into the following four groups: (1) feeling theories, according to which an emotion is a feeling; (2) behaviorist/functionalist theories, according to which an emotion is a disposition to, or a cause of, a certain kind of behavior; (3) cognitive/evaluative theories, according to which an emotion essentially involves a belief or an evaluation or a judgment; and (4) significance theories, according to which an emotion is a way of perceiving things, a way that encodes deep facts about the person with the emotion. The theory I shall champion is a kind of feeling theory, but unlike any other that I am aware of. It is not my purpose here to attempt any general refutation of any competing theory of emotion. But it will be useful to indicate briefly why I reject any of the other approaches as likely to result in an adequate philosophical account of emotion. Behaviorist/functionalist theories of emotion are surely plausible only in the context of behaviorist/functionalist theories of mind generally. What all such theories have in common is the denial that mentality consists to any degree of an “inner life,” that is, of things and events that are “private” to the person who has them in the sense that only that person is able to be aware directly of their existence or occurrence. (Or, if any such theory does admit of “private” things in this sense, then it denies that their intrinsic features are essential to their being the kinds of mental things they are or even to their being mental.) Denying that there is an “inner life” is thought by some to be required either by adherence to the principle of ontological economy or by acceptance of the scientific worldview. Because the principle of ontological economy becomes relevant only when all other rational means for deciding between competing theories have been exhausted without resolving the matter, examination of those other means is a prior requirement. Needless to say, I do not believe that their invocation does leave the matter unresolved. As for the requirements of the 1

Lyons, 1980, especially the first two chapters. His actual labels for the four types are: feeling theories, behaviorist theories, cognitive theories, and psychoanalytic theories.

88

scientific woridview: while I do accept the necessity of such adherence, I reject the widespread belief that it compels materialism in the sense of denying the existence of “private,” non-physical phenomena, that is, absolute materialism.2 Because behaviorist/functionalist theories deny what seem to be the everyday facts about the nature of mentality, they should be rejected if nothing else recommends them except the principle of ontological economy and compatibility with the scientific worldview. And I know of nothing else that does recommend them. Cognitive/evaluative theories of emotion take many forms, and there are specific arguments that can be made against this or that version, some of which are made by defenders of other versions. I have an overriding reason for rejecting all such theories which is a principle that guides my theorizing about emotions generally. That principle is the assumption that animals have emotions, which are in some cases very much like emotions in humans. One thinks first, probably, of fear; but there seems to be good reason to suppose that animals also experience sadness, euphoria, desire, satisfaction, and other emotions. An adequate theory of emotions should, therefore, be applicable to animals as well as humans. But there is another fact about animals that is relevant, a fact to which there will be fewer subscribers than the fact that animals do have emotions. It is the fact that animals do not have beliefs or make evaluations or render judgments. If this is a fact, then cognitive/evaluative theories of emotion are, if not just wrong, at least not general theories of emotion. I want a general theory of emotion. The trouble with significance theories of emotion is that they are not theories about what emotions are, but theories about their causes and consequences. Freud’s theory is one such theory, and Sartre’s theory of the emotions is also essentially of this character. While one might possibly derive a theory of the intrinsic nature of emotion from Sartre’s general theory of consciousness, his monograph on the subject of emotion really tells us little about what emotions are, philosophically speaking, and a lot about what we might achieve by having a certain emotion in a certain kind 2

It may, however, compel acceptance of scientific materialism, the thesis that everything that occurs has a physical explanation. How often one finds absolute materialism and scientific materialism confused in the literature!

89

of situation.3 Sartre himself, by the way, links his theory to Freud’s in many important respects including that of what kind of theory it is. So I dismiss significance theories of emotion as not answering, or even trying to answer, the philosophical question of what an emotion is. Feeling theories are difficult to characterize generically, partly because, for example, the behaviorist might agree that a feeling is involved in an emotion, but go on to analyze having feeling behavioristically. That is a natural result of the fact that it is a piece of common sense that an emotion is a feeling. But in the sense in which it is controversial that an emotion is a feeling and in which many, possibly most, theorists have denied that an emotion is a feeling, we must think of a feeling as something “private” to the person with the emotion, something probably non-physical in nature or at least not given as the sort of thing that might consist of atoms and molecules, something without apparent spatial location, and something that seems probably to exist only as long as it is “had.” Perhaps the general idea can be made clearer by summarizing briefly some of the standard objections to feeling theories of emotion4 with the idea of returning to them later in order to evaluate the version I shall put forward. One objection derives from the philosophy of the later Wittgenstein concerning the impossibility of private languages and the necessity for “outer” criteria for talk about “inner” phenomena. It is simply that if an emotion were a feeling and therefore “private” to the person who has it, then the names of emotions could not acquire their meanings. This objection, obviously, involves difficult and complicated issues of meaning and language acquisition that are well beyond the scope of this paper, but I shall eventually indicate at least the sketch of a reply to it. A second objection is that if emotions are feelings, then they are “passive” and like sensations, which is counter to the datum that a person can be in an emotional state and not be aware of it; while a third and related objection holds that if the causes and consequences of some state are irrelevant to its proper characterization as a certain kind of emotional state, then there is no way to distinguish emotions from sensations, and this, too, is counter to common sense and the obvious facts of the matter. 3

Sartre, 1948. Sartre’s theory might also be said to fall under the cognitive accounts to the extent that he maintains that beliefs are essentially involved in emotions. 4 These are drawn, with modifications, from Lyons, 1980, 5-16.

90

A fourth and a fifth objection also derive from that feature of feeling theories according to which there is no “conceptual” connection between an emotion and behavior of any particular kind: that the feeling theory makes it impossible to explain how, as common sense holds, emotions can give rise to behavior (the etymology of “emotion” is shared with that of “motion” itself); and also that it makes it impossible to explain why, as common sense holds, emotions can sometimes be characterized as reasonable or unreasonable. II A final preliminary step is worthwhile before I present the new theory of the ontology of emotion. This step has to do with philosophical method. I do not take my task, as an ontologist of emotion, to be the elucidation of what some philosophers would call the “concept” of emotion. It is emotions, and not any “concepts” thereof, that I wish to understand. The “concept” of emotion, if that means how people actually think and talk about emotions, is certainly of intellectual interest, but that interest is primarily psychological and anthropological and not philosophical. The “concept” comes into play in the philosophical treatment of emotion to the extent that we must identify with less or greater precision just what phenomena in the world we are subjecting to philosophical scrutiny, and I shall do that a few paragraphs hence. But that is using the concept of emotion and not “analyzing” it. My method is “analytic” insofar as I shall attempt to say accurately what the simple constituents of emotions and the having of emotions are, but not in the usual sense of our philosophical culture according to which one proposes “definitions” that give “necessary and sufficient conditions” for something to be what it is, that is, a “conceptual” analysis that is supposed to yield analytically equivalent sentences (or sentence forms) to some original sentence (or sentence form) and then invites so-called “counterexamples.” It is ontological analysis in which I am engaged, and the defense of a proposed ontological analysis rarely has to do with analytical equivalence or its absence. It is true, to be sure, that an adequate philosophical theory, whether of the emotions or anything else, must be in significant accord with our common ways of thinking and talking about what the theory is about, else

91

there is the danger that theory is not about what it purports to be about. But, on the other hand, it is wrong to expect that any adequate and reasonably comprehensive theory will be congruent in every respect with common ways of thinking and speaking about its subject, just because those ways are not themselves usually consistent or even entirely clear. That is another reason why “conceptual” analysis is not to the point, at least as anything more than a preliminary to understanding what the “concept” is a concept of. Or, better put, there is no such thing as the concept of emotion (or of knowledge or of causation or of anything else) so incompletely understood, both scientifically and philosophically, as the emotions. Nor, finally, is there any single way to judge the adequacy of a philosophical theory and certainly not by “conceptual” means in the sense just indicated. Often the best argument for a theory is that, despite its rough edges and lack of demonstrable sufficiency, it is better than any of its competitors in accounting for certain data in the sense in which the objections I listed earlier charge any version of the feeling theory of emotion with not being able to account for certain uncontroversial facts. Let us see. III There is a use of “emotion” in which love is said to be an emotion, but also a narrower use in which no durable state like love is an emotion, even though love would be said to have certain characteristic emotions such as desire and euphoria and nostalgia associated with it. In this narrower use, in addition to those just mentioned, fear and anger and disgust and joy and sadness are emotions. It is emotion in the narrower sense that I am interested in. But there is another distinction, not quite coextensive with the first, the invocation of which will further specify the subject matter of my theory. Like the terms for many mental phenomena, at least some of those for emotions have both dispositional and occurrent uses. When a person is said to be sad even when unconscious, the use of ‘sad’ is clearly dispositional. But the disappointment one feels on getting a rejection slip from a potential publisher is just as obviously not something merely dispositional, although dispositions are no doubt involved, at least causally. It is emotion in the

92

occurrent and not the dispositional sense that I am interested in. This distinction is not, however, to be confused with that of emotions one is aware of and emotions one is not aware of: it is a datum, I believe, that a person who is in the throes of (occurrent) rage may not be aware of, and may even sincerely deny, the existence of that rage. So the aware/not aware distinction quite cuts across the occurrent/dispositional distinction (as it does also the narrow/broad distinction) in the use of ‘emotion.’ It is strongly tempting to say, therefore, that it is emotions in the sense of feelings that I am interested in, but that would be somewhat unfairly loading my case despite the fact that the dictionary lists ‘feeling’ as a synonym for ‘emotion’ and that, or at least so I believe, it is the commonsense view that emotions just are feelings of a certain kind. It is important to understand clearly that the different theories of emotion are not about different phenomena, as if (so one might suppose) cognitive/eva1uative theories are really about emotions in the broad sense of ‘emotion’ and behaviorist/functionalist theories are really about emotions in the dispositional sense. In fact, those theories too are about occurrent emotions in the narrow sense of ‘emotion’; and one should not suppose that behaviorist/functionalist theorists, for example, cannot and do not make the occurrent/dispositional distinction even if, on their view, occurrent mental states are just dispositions to behave in certain ways or states apt to produce certain behaviors. I state the new theory initially as consisting of five major propositions. (1) An emotion is an intentional object, that is, an object that is something of which one is aware and that probably exists only insofar as there is an awareness of a certain kind of it. (2) Although an emotion is an intentional object and probably exists only insofar as there is an awareness of a certain kind of it, one may not be attending to one’s emotion; and so, in one common use of the word, one may be unaware of one’s own emotional state. (3) An emotion in itself is not about anything. (4) What is called the “aboutness” of an emotion derives from the literal aboutness of an accompanying mental state. (5) The connection between the “having” of an emotion and its accompanying mental state (when there is one) is only that that obtains between any two states of consciousness occurring simultaneously in the same mind with the addition that in this case the accompanying mental state is the cause of the emotional state.

93

Ad 1: It is necessary first to say a few words about the general theory of consciousness that I am assuming here and have defended in other publications.5 A state of consciousness involves a particular (because it is a particular state or occurrence) which may be thought of, in this context, as the self. This particular exemplifies instances of two kinds of properties: a mode property, which determines what kind of awareness it is, an imagining or a perceiving or a remembering or whatever; and an intentional property, which determines what the consciousness is a consciousness of. It is intentional properties that I have taken to calling natural signs in order to emphasize that each is an entity that intrinsically represents what, preanalytically, the state of consciousness is said to be a consciousness of. Finally, when the object or state of affairs of which one is aware actually exists, there is a connection, the intentional connection, between the natural sign and that object or state of affairs. When that of which one is aware does not exist, there is no such connection, merely the exemplification of the two properties by the self. States of consciousness I call the primary mental entities for the reason that a being without any would naturally be said to be a being without a mind. Those entities that are objects of states of consciousness but that seem to depend for their existence on awareness of them, I call secondary mental entities. On my view, of which this paper is a partial defense, the realm of the secondary mental entities includes sensations, emotions, moods, and such perception-related entities as images, afterimages, and (if there are any) sense data. Dispositional mental entities, such as hope and belief, constitute the tertiary mental entities. An emotion, then, is an intentional object.6 To feel an emotion is to be in a certain conscious state that intends that emotion. (Eventually, I shall cut things finer, distinguishing between feeling an emotion and having an emotion; between, for example, feeling angry and merely being angry.) Thus, if a person feels angry, there is on the side of the act the conscious 5

See especially Addis, 1983 and 1989. Anything whatsoever can be an intentional object. Some things, such as chairs, exist whether or not they are also intentional objects. Other “things,” such as unicorns, do not exist but are sometimes intentional objects. My thesis is that emotions are things that exist in exactly the same sense as chairs exist but only when they are also intentional objects. 6

94

state of feeling the anger, and on the side of the object the anger itself. Like all act/object situations, this one involves on the side of the act a particular exemplifying a mode property and an intentional property. The mode property may reasonably be referred to as being-a-feeling, and the intentional property (natural sign) as being-of-anger. Thus the anger as such is never something that intends or is about anything; it is only something intended, something of which a person can be aware but never something by which a person is aware, at least in the strict sense. Ad 2: A person can be angry and not know it, and even sincerely deny it. This I take to be a datum, and so not something I need to argue for. Another way to put this idea is the way I did in the last paragraph: a person can be angry without feeling angry. But what is it to be angry or, as I shall also put it, to have anger without feeling angry? Is that also an act/object situation in which the anger is the object of some kind of awareness? The answer to the latter question is, I believe, in the affirmative, and I shall now try to explain what is involved by turning first to the important but neglected notion of attending-to. Because I have already said something about this elsewhere,7 I shall here limit myself to a very few comments. Perhaps the easiest way to grasp the idea is to consider one’s own visual field. In it there are many things of which, by assumption, one is aware in a minimal sense but to which one is not attending. If one is attending to anything in one’s visual field it may be, but it also may not be, that upon which one is (visually) focusing. In short, much of which we are aware of we are not paying attention to, and it would seem to be necessary to our survival and to that of any sentient being to be able to ignore at the level of attending-to most of what, all the same, one is aware of. If we ask what attending-to, which we now see to be a familiar and pervasive phenomenon, amounts to ontologically, we seem to be limited to two possibilities: (1) that every form of awareness has, corresponding to it, a “heightened” form, which we might call its “supermode,” being such that when it occurs in a given awareness, it makes difficult or impossible the simultaneous occurrence of the supermode of any other awareness; or (2) 7

See Addis, 1986, especially 63-65. More recently John Searle has given very similar treatment to what he calls “attention” in his 1992, 137-139.

95

that there is a distinct mode of awareness which just is that of being-anattending-to and is such that it can intend only what is simultaneously intended through some other mode of awareness and that makes difficult or impossible the simultaneous occurrence of this or another act of attendingto accompanying any other awareness. (A lot of this verbiage is just a way of saying that it is difficult or impossible, depending on the particular circumstances, to attend to more than one or two things at a time.) As to choosing between these two possibilities, I can find no basis, phenomenological or otherwise, for preferring the one to the other. But because the first is, by not requiring any more conscious states, a simpler view, let us assume that it is correct in what follows. It is a curious fact, remarked on in different language and to different ends by other commentators, that when a person has one of the desired emotions such as joy or pleasure, that person is likely to be attending to or absorbed in the activity or object that produces that joy or pleasure, whereas when a person has one of the undesired emotions such as sadness, that person is likely to be attending to or absorbed in the emotion itself. Naturally, there are numerous exceptions to this, and I refer only to probabilities. Furthermore, even when a person does have one of the undesired emotions, that person may, at least temporarily, “forget” the emotion by being absorbed in some unrelated matter. In any case, to return to the main point: a person may have an emotion in the narrow and occurrent sense and not be aware of that fact. This we may now express by saying that while in such a case one is, at some level, aware of one’s emotion (for to have an emotion is just to be in some intentional connection to it), one may not be attending to one’s emotion and so in that sense not be aware of it. This is no more paradoxical than saying that one is not aware of, in the sense of attending-to, much of what is in one’s visual field, even though to be in one’s visual field is also to be something to which one is in some intentional connection. With this, then, I take myself to have explained the sense in which a person may be having an emotion, even be expressing the emotion quite visibly in language and behavior and facial expression, and yet be unaware of being in that emotional state. Ad 3: The mere fact that an emotion is an intentional object is not, by itself, ground for saying that emotions are not, in themselves, about

96

anything: states of consciousness can themselves become the objects of other states of consciousness and as such are intentional objects that are, in themselves, about something. But in fact, so I assert, an emotion no more than a sensation such as a pain is a state of consciousness; and so emotions no more than sensations are about anything. It is commonly thought that emotions must be radically unlike sensations just because while no sensation is, in any sense, about something, at least some emotions are about something. But this is a mistake. My theory of emotions, to be sure, is intended quite deliberately to treat emotions as very similar to sensations; but this I regard as an advantage of the theory, especially insofar as the distinction between emotions and sensations seems to be in fact a rather nebulous one. Another advantage is just that it makes no mystery at all of those situations in which we do very much seem to have an emotion that is not, in any sense, about something. But it remains to explain what goes on in those cases in which our emotions are said to be, in what for me is a derivative sense, about something. Ad 4: If a person or a gazelle is afraid of something, then there is the fear that is had, whether attended to or not. But of course there is also something else, one’s intentional connection to the object of which one is afraid. We must be careful, at least insofar as we wish to say that the gazelle is afraid of something, not to suppose that this intentional state is a belief that the object might harm one (although there may be such a belief) or any similar cognitive state of consciousness, for animals have no beliefs or at least no beliefs of the requisite kind. In a simple case, it will be only the perception through sight or other external sense of the object that is the relevant intentional connection. Thus, to be afraid of a lion that a person or a gazelle sees charging oneself is to perceive the lion and to feel afraid. There are two distinct states of consciousness, the one a feeling with fear as its object, the other a perceiving with the lion as its object, which together are what it is to be afraid of the lion.8 Either state of 8

I am by no means the first to put forward a theory according to which the aboutness of an emotion derives from an accompanying mental state instead of the emotion itself. One example of such an account is to be found in Aquila, 1975. What, to my knowledge, makes my account different from Aquila’s and all other previous accounts lies in also regarding the emotional state as an intentional state, with the emotion as the intentional object.

97

consciousness can exist without the other, for one can be afraid without being afraid of anything, and one can perceive a lion without fear or other emotion. But to be afraid of a lion is, I repeat, simply to have these two, analytically distinct states of consciousness, albeit as connected in a certain way. Ad 5: What then is the connection between these two states of consciousness that together constitute being afraid of a lion? Although it is tempting to think that this connection must be of some very intimate kind, it is in fact, so I believe, only that of simultaneously occurring mental states in the same mind in which the one state is the cause of the other— the perceiving of the lion the cause of the feeling of the fear. But as for the “togetherness,” it may be granted that in the typical human case there will be other states of consciousness besides the two already mentioned in the kind of situation we are imagining: there will very probably be the thought that the lion may do one harm, the desire that one not be harmed, the thought that perceiving the lion is the cause of one’s fear, and so on. And if someone wants to insist that the fear is a fear of the lion only in case some or all of these or even some other conscious states obtain, I will not resist very strenuously. I do think it is preferable on analytic grounds to characterize the situation as I did, at least to the extent that we choose to say that the emotions of animals can be about something; for example, that a gazelle can be afraid of a lion. But I am not interested in how, precisely, the words do or might lay on the world, that is, in the “conceptual” analysis of the situation. The crucial point is that the intentionality of those emotional situations that have intentionality does not reside in the emotion itself or in the conscious state of which it is the object but in some other state or states of consciousness, whatever it or they may be. And, again, if one then wants to say that in that case, there is no intentionality in the emotions of animals (beyond the having of the emotion itself, on my theory), that is only a choice on how to use words and not a substantive theory. ________ If an emotion is an object, what kind of object is it? Because, on the theory here proposed, an emotion, like a chair, is not literally part of any act of awareness itself while nevertheless being something, unlike a chair, that

98

exists only as long as there is an act of awareness of a certain mode of it, the particularity of an individual instance of emotion—this person’s anger—must lie in the emotion itself. And what makes it a case of anger must derive as well from some intrinsic feature of it. We are thus led to the conclusion that an individual instance of an emotion is a momentary particular exemplifying a certain property. Thus the ontology of the entire situation of a person’s being angry is as follows: on the side of the act there is a particular (which, I said earlier, could be regarded in this context as the self) exemplifying the mode property of being-a-feeling and the intentional property of being-of-anger which intentional property stands in the intentional connection to the anger itself. On the side of the object, the anger itself, there is a momentary particular exemplifying the property of being-anger. If the person is angry with someone or something, there will also be another mental act consisting of a particular exemplifying the mode property and the intentional property appropriate to the nature of the awareness of that someone or something. Some will object that if one wants a theory of emotion in the spirit of the one here proposed, it would be more reasonable to hold that an emotion is just a property of the self. Indeed, most feeling theories of the emotions, while rarely being ontologically explicit, seem probably best understood in just that way. And one may well suspect that many philosophers have not conceived any other ontological possibility. If anyone did conceive the theory I am defending here, that philosopher probably rejected it on either the phenomenological ground that it makes the emotion too “distant” from the person who has it or the ontological ground that the idea of an emotion as an “independent” entity is absurd. As for the phenomenological objection: it is true that when a person is angry, that person has the property of being angry. That, however, is not an ontological account of the situation but only another way of saying the same thing. But let us consider the genuinely ontological theory that to be angry is merely for some particular to exemplify the property of beingangry whether that be a momentary particular or a mental substance or the brain or the entire body. Agreeing, as I do, with Husserl and Sartre that consciousness is always consciousness of something (but not, of course, always of something that exists), my objection to any such theory is that it fails to account for the fact that to be angry is to be in a kind of conscious

99

state. Where is the act, and what is its object? And if someone suggests that each different emotion is a distinct mode of awareness, I still ask what those conscious states that are emotions and that seem to have no objects are conscious states of. Let us consider further, however, a theory that would make the emotions modes of awareness and that would ascribe “the human condition” or “the world as a whole” or “the null class” as the object of any emotion for which there is no ordinary object. This theory would have the virtue of being formally satisfactory in that it makes having an emotion a conscious state, that is, a state consisting of a particular exemplifying both a mode property and a natural sign and that is able, in every case, to specify an object of that natural sign. But unless one is prepared further to maintain that having a sensation, which is also just as certainly a conscious state, is about “the human condition” or whatever, one has both an incomplete account of mental life and one that fails to explain the similarity and even overlap between the emotions and sensations. Emotions and sensations alike are, to make a direct phenomenological appeal, things that are felt, things of which one is aware. And treating emotions and sensations as objects of awareness instead of “forms” or “kinds” or “modes” of awareness is, as we now understand, entirely consistent with the commonsense observation, sometimes mistaken for a philosophical theory, that to have a certain emotion or sensation is to have a certain property or to be in a certain state. As for the ontological objection: It is true that my theory makes it ontologically possible for an emotion to exist without any awareness of it. At least this is so on the plausible ontological principle that while neither a particular alone (that is, without any properties) nor a property alone (that is, without being exemplified) is ontologically capable of existence, any complex consisting of a particular exemplifying a property is ontologically so capable. But we have every reason to believe that what is ontologically possible is, for emotions, causally impossible. My emotion is still my emotion just because, first, its existence is contingent of my being aware of it in a certain way and, second, only I can be aware of that (instance of) that emotion in that way. If I have established nothing else in this essay, perhaps I may claim to have shown that a theory of the emotions can be defended adequately only

100

in the context of a comprehensive theory of the mind and of consciousness in particular. Most extant theories of the emotions, I submit, have not been so defended. The task is to secure the best theory, everything considered. I further submit that this new theory of the emotions is that theory. IV What are the advantages of this new theory of emotion? I shall summarize and elaborate on those that have already been mentioned. (1) It accords well with the commonsense view that emotions are feelings. Common sense, like most philosophers, makes no clear distinction between a feeling in the sense of that which is felt and a feeling in the sense of the primary kind of awareness of it; we may say that an emotion is a feeling, but not the only kind of feeling, in the sense that it is an object of that mode of awareness we call feeling (although there are other modes of consciousness that are also called feeling). Even here, though, there is ambiguity. For we may say that a person feels an emotion (or a pain) without being aware of it, that is, without attending to what is felt; but we may also say that genuinely to feel something is just to have a characteristic awareness of it and to be attending to it. Again, the words matter little once we grant that there is a distinct mode of awareness that we have to our emotions and that grounds their being called feelings. (2) Perhaps the biggest advantage of the theory is that it accounts clearly, by showing just how it is possible, for the fact that some emotions seem to have no objects while others do seem to have objects. And it accounts for the fact by maintaining that what seems to be the case is the case. So we are no longer stuck with those artificial views that invent objects for those emotions that seem not to have any or that attach them to “the world as a whole” or “the human condition” or whatever. Some emotions have no objects in any sense. At the same time, although some may doubt that this is genuinely an advantage of the theory, it retains, or at least does not contradict, the thesis that an act of consciousness is always a consciousness of something. Emotions are not acts of consciousness but objects of some acts of consciousness. (3) A third advantage of the theory is that it recognizes the significant similarities between emotions and sensations. Some would regard this as a

101

disadvantage of the theory, believing that because many emotions are in some way object-directed while no sensations are, they must be quite different sorts of things. I have given some reasons for thinking otherwise, but more detail here on the similarity of emotions and sensations will be valuable to my case. Some people have thought that some emotions, especially strongly-felt ones (in the narrow, occurrent sense) essentially involve, or even just are, certain bodily sensations such as feelings of being flushed, or of labored breathing, or of increased pulse rate, and so on. This, I am convinced, is a mistake even though certain interesting experiments seem to suggest that people sometimes identify the emotions they are having more by way of the sensations they have than by attending to what, on my account, is the emotion itself. But even if emotions do not essentially involve sensations and even if there is sometimes difficulty in categorizing a given feeling as an emotion or a sensation, there are certain important similarities between the two sorts of things even to the extent that some sensations are more like emotions than they are like most other sensations. Perhaps most important is just the fact that both kinds are thought of as something “mental.” This is, to be sure, a somewhat vague fact, at least as to what it is that makes the thought true, if anything. I am inclined to think that the essential ingredient is the fact that neither an emotion nor a sensation can exist without a mind that is, at some level, aware of it. Whether this “can” is ontological, as some believe, or instead only causal, as I believe, does not matter for the point at hand. Sensations and emotions do not exist “in” things without minds. At the same time, common sense vaguely realizes that emotions and sensations are not like perceivings and rememberings and imaginings and the like: very roughly, the former are things that are, or can be, felt within oneself while the latter are ways of getting outside oneself, as it were. Emotions and sensations are occurrences “in” one that one likes or dislikes, again quite unlike acts of awareness, at least typically. Emotions and sensations are very much of a kind. How, then, are emotions and sensations different? In addition to the semi-intentionality of most emotions, there is the more definitely bodily and spatial aspect of most sensations. But—and this only strengthens my account—some sensations, such as feelings of embarrassment, have little discernible bodily location or identification as felt. In my replies to the ob-

102

jections I listed earlier, I shall discuss another point in connection with the similarities and differences of emotions and sensations. V Let us turn finally to consider the standard objections to any feeling theory of the emotions. The first objection, you will recall, is the Wittgensteinian one about the possibility of naming “private” entities: if an emotion just is something to which only the person who has it can have direct access, then there can be no possibility of being right or wrong in the label one gives to any particular instance of such “private” entities, for there is no “check” on one’s use of such labels, either by oneself or by others. That, roughly, is the drift of the argument, and many who object to any version of the “feeling” theory of emotion have made some version or other of it.9 I think the best point to make first about this argument is that in fact, as we all know, there are things to which each of us has one’s own privileged access, and about which one does manage to talk intelligibly and easily. So if some theory, Wittgenstein’s or anyone else’s, entails that that is impossible, so much the worse for the theory. The convoluted theories of Ryle and other metaphysical behaviorists and functionalists testify to the absurdity of attempting to construct an adequate philosophy of mind that does not recognize the most obvious facts about mental life. But a reply to the argument itself would begin by acknowledging and even insisting that in learning the language of mental life, children must have “outer” criteria for their “inner” experience. If our private, mental life were not connected in a systematic way with some aspects of what is public and external (roughly in the way demanded by psychophysiological parallelism), there would be no possibility of mastering a public language about mental life.10 But it does not follow from that, nor is it true, that the referents themselves of “mental” language are not “private,” nor that we are unable, eventually, to identify the occurrences of such things and events simply by experiencing them and being aware of such experiences. There always 9

This line of argument, especially with respect to the emotions, is developed, for example, in Kenny, 1963. 13-14. 10 This point is argued in detail in Addis, 1984, 341-343.

103

remain those same “outer” facts that provide all that is required by way of a “check” on one’s continuing use of words that refer to “private” things and events. It may be agreed also, in the spirit of Sellars’s myth of our Rylean ancestors, that mentalistic language may have arisen originally in the context of the explanation of outer behavior, and thereby has an essential connection to it. And we might suppose further that without this connection and the language to express it, we would never have been able, through introspection, to be aware directly of our “private” mental states. But we are able to be aware directly of some of our mental states and we are able to talk easily and intelligibly about those mental states, however that situation has come about. Stories about how ancestors and children learn mentalistic language are interesting and not completely irrelevant to the issue at hand, but they do not entail or even strongly suggest how the modern, adult speaker of a language uses it to refer to features of one’s own “inner” life. The second objection holds that if emotions are feelings, then they are “passive” like sensations, and one cannot account for the fact that a person can be in an emotional state and not be aware of it. I have already largely had my say on this matter, but it may be worth adding that the objection seems to presuppose that it is impossible for a person to have a sensation without being aware of it, and that, as I have argued elsewhere, is probably false.11 Just as a person can have an emotion and not be attending to it, so a person can have a pain or an itch without attending to it. As for “passivity”: whatever sense can be given to this notion and the objection that goes with it will be examined in consideration of the fifth objection. The third objection, too, decries the similarity of emotions to sensations on the theory here defended, and expresses doubt about their actually being distinguishable on that theory. I remind the reader of what was said on this point earlier, and emphasize again that it is a virtue of the theory that it promotes a close similarity and even overlapping, as it were, between emotions and sensations. And what remains on this point also may be treated in connection with the fifth objection. According to the fourth objection, there is a difficulty in accounting for 11

See Addis, 1986, 62-66.

104

the fact that emotions are causally related to behavior if the feeling theory is correct. The idea seems to be that feelings alone cannot cause behavior, that there must be desires involved, and that insofar therefore as we do sometimes explain a behavior by appeal to an emotion, the emotion must somehow include the desire. Thus the “concept” of emotion involves the causes and consequences of whatever feelings may also be involved in emotional states. But this, I submit, is just playing with words. Nothing one cites in any particular case of explaining behavior is going to be a sufficient condition, not because there are no prior sufficient conditions for behavior but because no one knows specifically what they are in a particular case. So the fact that a feeling alone is not a sufficient condition is no evidence against the feeling theory of emotion even granted our common talk that sometimes would “explain” a piece of behavior by that person’s emotion. Especially insofar as the occurrence of the feeling and its “internal” consequents (including, in many cases, the relevant desire) are the only relevant changes in the macro-state of a person, it is natural and correct, given our ordinary causal talk, to attempt to explain someone’s behavior by the occurrence in that person of some emotion, even when emotions are thought to be just feelings. The fifth objection—that if emotions are feelings, then no sense can be given to the commonsense datum that at least some emotions are reasonable or unreasonable—is trivial on its face, but does involve some deeper points. The rather trivial point is that in fact it is not the emotions themselves that we ordinarily regard as reasonable or unreasonable but only the behavior that results from them. I do not wish to deny that we use some normative language in discourse about emotions, but I do suggest that this is not a very interesting fact to the extent that it seems to tell not at all for or against any particular extant theory of emotion. Further, it would seem that such judgments as we do make about emotions are usually either judgments about the relation of an emotion to something else—behavior, the situation, the person’s character—or else judgments about the person’s allowing or preventing some emotion to be or to exist in a certain degree and the like. There is nothing in my theory of emotion or any feeling theory that wrongly delimits the extent to which human beings may control or indulge their emotions, or that entails any irrationality in the practice, to whatever extent it is useful, in making normative utterances or otherwise

105

indicating one’s attitude toward someone’s emotion in any of the many dimensions in which we do such things. And it is an obvious fact that we do have some effect on the capacity of others (and ourselves) to have and to feel emotion by the language we use in talking about emotions as well as, if not especially, the behavior that is associated with the emotions. True, there is a sense in which most of us regard it as either irrational (to the extent to which a person lacks control over the possession of an emotion at a time) or unwise (because it may stunt the ability to express or even to feel emotion and so produce useless guilt and mental suffering) to make direct normative judgments about a person’s feelings. But this fact, to the extent that it is a fact, would seem to accord well with the feeling theory of emotion just to the extent that we refrain from making direct normative judgments about a person’s emotions. There is obviously a strong cultural factor in this matter, and in our culture we have Freud to thank more than any other one person for having released most of us from being overly judgmental about the having of certain feelings including emotions and desires. And, finally, the very considerable extent to which we do refrain from making direct normative judgments about emotions tends, for obvious reasons, once more to support the idea of significant similarity between emotions and sensations. _________ I conclude that the best theory of the emotions is the one defended here, a form of feeling theory that takes emotions as objects of awareness and locates the intentionality of an emotion, when it exists, in an accompanying mental state.

106

THE SIMPLICITY OF CONTENT

S

ome philosophers hold that minds do not exist. With these philosophers one cannot, or in any case I shall not, argue. A larger group—the largest, almost certainly, among contemporary philosophers of mind, as I categorize--maintain that the mind exists but is, in one way or another, the same thing as something physical. It may be brain states, or behaviors, or patterns of or dispositions to behaviors; but, like the first group, these philosophers also hold that a human being is no more and no less than a physical organism. I shall argue only indirectly that these philosophers are mistaken in their materialism. My argument in this paper is, instead, with some of a smaller group of philosophers of mind who allow that there is, or may be, mental content. I put it this way because I want to distinguish within this group those philosophers who, while dualistic in spirit, in fact doubt or deny the existence of mental content from those who affirm the existence of mental content. Against the skeptic of content within the context of dualism, I shall make some arguments for the existence of mental content. The main argument of this essay, however, will be yet more narrowly focused. It will be against those philosophers who agree that there is mental content but who maintain that there is in all such contents a kind of complexity that in some way corresponds to the complexity of the objects of awareness. My positive thesis is that every mental content, in an important ontological sense, lacks complexity; I will be arguing for the simplicity of content.1 I What is a mind? What is mental content? What is simplicity (complexity)? Before we turn to some history and then to the arguments, we need at least some measure of answers to these questions. 1

This essay is, in part, an elaboration and extension of ideas and arguments made in the second and third chapters of Addis, 1989. The reader is invited to look there for discussion of other aspects of the issues discussed in this essay.

A person’s mind is, first and foremost, that person’s “stream of consciousness,” composed of his or her awarenesses—perceivings, rememberings, imaginings, feelings, doubtings, desirings, and so on— those entities that define us as conscious beings. These entities I call the primary mental entities. Comprising the secondary mental entities are those mental phenomena that seem not to be literal constituents of awarenesses but to depend on awarenesses for their existence—emotions, moods, bodily sensations and feelings, images and afterimages. And, further, as the tertiary mental entities the mind is also those states that, unlike awarenesses, a person may have even while asleep or unconscious— dispositional mental states such as beliefs, hopes, doubts, desires, and so on. Some philosophers believe that “behind” all these particular mental entities there is also, in each of us, a mental substance that is the self, while others believe that the self just is the “bundle” of some or all of the more particular kinds of mental entities already specified. For our purposes, we needn’t resolve that issue or even those of my threefold categorization of mental entities; for the questions I intend to deal with have to do exclusively with the first group of entities; namely, awarenesses. While many philosophers also speak of content with respect to dispositional mental states, I assume and have argued that, given the nature of dispositions, this cannot be literally true; and that the only kinds of entities that do, or could, in that literal ontological sense possess mental content are awarenesses. This rather cavalier insistence on the uniqueness of content to awareness will become both clearer and more plausible if we now ask just what mental content is. There is mental content if when a person is aware of something, that awareness is or contains an entity that correlates uniquely with the thing, or kind of thing, of which the person is aware. Suppose you are thinking about the horse that you rode yesterday. If there is “in” your awareness any entity that correlates uniquely with either the particular horse that you were riding or with the kind horse (that is, the property of being-a-horse), then there is, “in” your mind, mental content. But what is it to be “in” the mind or, more narrowly, “in” an awareness or “contained” by an awareness? I have used the double quotes, which I henceforth drop, because in its literal sense, ‘in’ is a spatial notion that is, for the most part anyway, inapplicable to mind. But we may begin

108

clarification—call it stipulation, if you prefer — by saying that to be in the mind is just to be a property of the mind, assuming that the mind itself has been antecedently identified. And, in this context, we may say that to be a mind is just to be whatever particular or particulars exemplify the properties that characterize mental phenomena. More precisely, and very importantly, we shall say that to be in the mind is to be a monadic property of the mind, that is, a property that is not, in whole or in part, a relation. This assumes, as is proper, that there is an ontological distinction to be made between monadic and polyadic properties, a distinction that is reflected in, but not exactly paralleled by, the distinction between monadic and relational predicates. (The predicate ‘tall’ is grammatically monadic but refers to an inherently polyadic property.) But what counts as a genuine property of the mind? Here again one enters a realm of confusion and complication in which one would want, in a full treatment of the matter, to distinguish descriptive from logical properties, substantive from formal properties, general from specific properties, and so on. For our purposes we may ignore most of these distinctions and say, first, that to be a property of the mind is to be a property of an awareness; and, second, that to be a property of such an awareness is to be a constituent of that awareness that distinguishes it, qualitatively, from some other actual or possible awarenesses. Thus, if your thinking about the horse you rode yesterday has a (monadic) property that correlates uniquely to that horse or to the property of being-a-horse, then it has a property that qualitatively distinguishes that awareness from that of, say, imagining that the moon is made of green cheese. And, if any awareness has any such property (I believe that all do), then there is mental content. Before we turn to the notion of simplicity, there remains to be said something about the matter alluded to in distinguishing, as I did above, between a mental entity that correlates uniquely with a particular horse from one that correlates only with the kind horse. The tradition seemingly has found it easier to understand how a person can be aware of the kind of thing one is aware of than of the particular thing itself. This is most evident, perhaps, in medieval abstraction theory according to which to be aware of a horse is for the mind to take on what was called the form horse (without, however, thereby becoming a horse). But taking on the form

109

horse would have something in the mind that correlates with (or just is?) the property of being-a-horse and not with the particular horse of which the person was aware. How the medievals accommodated the datum that the person was aware of a particular horse and not just horsehood (it has to do, in Aquinas for example, with “phantasms” and the distinction between the senses and the intellect) is not my concern here. It is, instead, to establish that the most useful notion of mental content will include either kind of property—one that would correlate uniquely to the particular horse as well as one that would so correlate only to horsehood. In most of what follows, however, I shall take for granted that if there are mental contents, some of them are of particular things as particular and not only of their “forms.” What is simplicity? More exactly, what is it for a property to be simple and not complex? Is this really an ontological distinction or only a misleadingly circumspect, if unintentional, way of reflecting the distinction between undefined and defined predicates, respectively—a distinction that is plausibly regarded as language-relative? I shall not argue that issue here, but simply assume that there is a difference in reality itself, independent of mind and language, between those properties that do, and those that do not, have other properties as constituents. Thus, as plausible examples, we may say that a specific shade of red is a simple property while the property of being-a-horse is a complex property. For a specific shade of red (and I’m referring, of course, not to light waves or dispositions to produce certain light waves or anything other than the property we see), while it may exemplify hue and brightness and other properties, has no other properties as literal constituents of itself. But the property of having-two-eyes is not exemplified by the property of being-a-horse; instead it partly constitutes that property without itself being that property. Thus, assuming what has been here at best well-illustrated—that there is an ontological distinction between simple and complex properties—we may say initially that if mental content is simple, then the property of the awareness that correlates uniquely to what that person is aware of is a simple property. That is exactly the position I shall defend in this paper. The alternative is that mental content is complex, but we must now understand that there are at least two ways in which mental content might be complex, both of which are in contradiction to the thesis of the

110

simplicity of content. One way for mental content to be complex, the way that would perhaps most naturally suggest itself, is for there to be a single particular (substance, momentary particular, bundle of properties) that exemplifies a complex property, which complex property is the mental content. But another possibility is that there are multiple particulars (momentary particulars or separate bundles but not, of course, substances) each of which has a property—simple or complex—which properties jointly are the mental content. Each of these possibilities has its own peculiar flaws, as I shall argue later, but they share a broader defect in being unable adequately to account for the unity of an act of awareness. Indeed, at one level, one may say that any view that denies the simplicity of mental content must fail to safeguard the unity of thought. Can it not be said that almost from the beginning of Western philosophy (there isn’t much philosophy of mind in the pre-Socratics, as we know them) until at least as late as the nineteenth century, perhaps even only the twentieth century, it has been assumed nearly without argument that when a person is aware of something, there is in that person’s mind something that correlates uniquely to that of which the person is thinking or at least to its kind? Some of the medievals, as we noted, spoke of the mind’s taking on the form of the object of awareness; some of the later medievals, most notably William of Occam, spoke of natural signs or conceptual terms (in contrast with conventional signs and linguistic terms, respectively). The early modern philosophers preferred to speak of “ideas” while the language of “content” became dominant in the nineteenth century. (More precisely, because most of the important thinkers on the topic in this century were German-speaking, it was the language of “Inhalt”, which is usually rendered as ‘content’ in English.) Early analytic philosophers, on the other hand, especially under the influence of phenomenalism, tended to speak of “sense data” (or “sensa”) or, more rarely, of “percepts”, taking perceptual awareness as the paradigm for all awareness. The twentieth century saw also the first fully explicit challenges to the general notion of mental content—most notably in the thought of two of the century’s most important but in most respects quite different philosophers, Bertrand Russell and Jean-Paul Sartre. I close this fragment of philosophical linguistic history by noting that the language of “content”

111

has made significant return in the later twentieth century along with that of “representation”.2 My use of ‘content’ then is by no means idiosyncratic, even if, in some eyes, the view in which it is embodied may seem so. The first philosophers who felt compelled to make, or who were able to conceive, arguments for the existence of mental content were Kazimierz Twardowski and Alexius Meinong.3 Even though I emphatically agree with their conclusion—that mental content exists—I have tried to show elsewhere that the arguments, as formulated by Meinong, are seriously flawed (Addis, 1989, 36-42). More recent philosophers who have explicitly defended the existence of mental content are, from quite different traditions, Edmund Husserl and Gustav Bergmann.4 But unlike the two earlier thinkers, they made no arguments for the existence of mental content, being apparently satisfied to rest their case on the claim or mere assumption that we are, or can be, directly acquainted with mental content. For my part, I have made three detailed arguments for mental content (Addis, 1989, chapter 3), or what I call natural signs, following William of Occam. Here I will summarize those arguments very briefly. One of them, the scientific argument, holds that differences in behavior that are due to a difference in what people are aware of can be explained only if there is a relevant difference in some monadic properties of the people involved. If you are thinking of mermaids and I am thinking of unicorns, then any difference in our behaviors that are due to the difference in what we are thinking about (for example, our different responses to the question, “What are you thinking about?”), obviously cannot—contra Russell and Sartre— be explained by the different objects of our awarenesses; for, in the examples at hand, they do not exist. A second argument, the phenomenological argument, does not, as might be initially supposed, simply appeal to what is given to us in experience; such arguments, as we well know, are rarely convincing to anyone not already convinced. Instead, this argument maintains that the best and possibly the only plausible explanation of the psychological and, perhaps, also rational certainty one ordinarily has about what it is one is aware of is 2

Other expressions that have been used in the twentieth century to capture the idea of mental content include ‘intentional property’ and ‘proposition’. 3 See especially Twardowski, 1977, and Meinong, 1899. 4 See especially Husserl, 1900, and Bergmann, 1967.

112

that one is, or can become, directly aware of a monadic property of one’s mind, the having of which property just is what it is to be aware of whatever one is aware of. I know with utter certainty that I am now imagining a unicorn; and the fact that it is a unicorn I am imagining and not a loaf of bread or a paramecium or anything else is surely due to some property of me, and one of which I can become directly aware. Finally, the dialectical argument urges that the existence of any kind of representation other than that of mind itself—be it by language or maps or artworks or anything else, any situation in which a first thing represents a second thing for a third thing (a person)—presupposes the existence of a kind of entity that, by its inherent nature, represents what it does represent. And that is, and can only be, mental content. More briefly put: any kind of conventional representation presupposes natural representation, and natural representation presupposes mental content. And because there is conventional representation, one can be confident that there is mental content. In what follows, then, the existence of mental content understood as monadic properties of awarenesses that correlate uniquely with the objects, or kind of objects, of awareness—will be regarded as established. II There are really three relevant possibilities with respect to the property or properties that constitute the mental content of a given act of awareness. Either (1) the mental content is a simple property (and therefore, necessarily, of a single particular), or (2) the mental content is a complex property of a single particular, or (3) the mental content is made up of different properties of multiple particulars. Because there would be no point to it, we may ignore the formal possibility of multiple particulars all with the same property constituting a single act of awareness. Furthermore, because the emphasis of the third possibility is on the multiplicity of particulars, we may also ignore the question of the simplicity or complexity (or mixture of the two) of the multiple properties involved. To establish the simplicity of content will be, first, to show that the second and third possibilities both have insuperable difficulties not shared with the first possibility; and, second, to demonstrate that certain apparent

113

difficulties in the first possibility are not insurmountable. Let us begin with the second possibility. Suppose then that a mental content is a complex property of a single particular. If the act of awareness is the thought that Kristin is to the left of Karin, it would be natural to suppose that the complex property—call it the property of being-the-thought-that Kristin-is-to-the-left-of-Karin—consists of the simple or simpler properties of being-the-thought-of-Kristin and being-the-thought-of-Karin and being-the-thought-of-to-the-left-of. For the argument that follows, it doesn’t matter, as will become evident, if these properties themselves could or would be given further analysis or even if somewhat different properties are said to make up the complex property. In order to see the fatal flaw in the thesis that a mental content is a complex property, we need to examine a little more closely the very idea of a complex property. Let us take a clear and unproblematic example—the complex property of being-red-and-round. For something to have this complex property is, as I assume is evident, just for it to have the property of being-red and also the property of being-round. Some ontologists might wish to assign a separate and possibly different ontological status to the complex property in addition to that of each of its simpler, constituting properties; but the joint exemplification of red and round will still be the ontologically necessary and sufficient condition for the exemplification of the complex property, whatever its own ontological status. Returning to the case of a putatively complex property that is a mental content, we understand that to exemplify the complex property of beingthe-thought-that-Kristin-is-to-the-left-of-Karin would be for the particular jointly to exemplify the properties of being-the-thought-of-Kristin and being-the-thought-of-Karin and being-the-thought-of-to-the-left-of. I have spelled out these matters in pedantic detail so that we may grasp very clearly why the view that a mental content is a complex property of a single particular cannot possibly be correct. But in order to do that, we must introduce a second situation; namely, that in which someone is thinking, not that Kristin is to the left of Karin, but that Karin is to the left of Kristin. If we now ask what it is, on analysis, for the putatively complex property of being-the-thought-that-Karin-is-tothe-left-of-Kristin to be exemplified, the only answer can be for the particular jointly to exemplify exactly the same three simpler properties as

114

in the case of the thought that Kristin is to the left of Karin. The difficulty is obvious and insuperable. If a proposed ontological analysis yields the same result for both of two qualitatively different situations, then it cannot possibly be correct. The thought that Kristin is to the left of Karin and the thought that Karin is to the left of Kristin are obviously qualitatively different mental contents. Yet, on the theory that mental contents are complex properties, they yield to identical analyses. The conclusion can only be that the theory is false. Underlying my argument is the assumption of another basic ontological principle which, I should think, admits of no plausible challenge: that joint exemplification of properties by a particular admits of no order. Two spots each of which is at the same time red and round cannot be such that, by some ordering principle, the one is “first” red and “then” round while the other is “first” round and “then” red. There is no such ordering relation and no sense can be made, I submit, of the very idea of any such relation. That means that if there is to be found a difference in analysis in the two putatively complex properties that are mental contents, it could come only by some difference in the simpler properties—some such property that is exemplified in the one case but not in the other. But what, in our examples, could this property possibly be? There is no answer to this question, and so the conclusion stands that a mental content is not a complex property of a single particular. Let us turn to the other form of the theory that mental contents are not simple properties—that a mental content consists in properties exemplified by multiple particulars. In this case, we will be more importantly faced with the idea of an ordering relation, as is perhaps already evident. Remaining with our examples, we may suppose that there is, in both situations, one particular that exemplifies the property of being-thethought-of-Kristin and a second particular that exemplifies the property of being-the-thought-of-Karin. Whether or not there is a third particular that exemplifies the property of being-the-thought-of-to-the-left-of will depend on which variation of this theory is before us. One of them would hold that the thoughts of each of Kristin and Karin are really images of the two women, and that what distinguishes the two situations is that in the one case, the image of Kristin is to the left of the image of Karin in the person’s phenomenal space while in the other case, the image of Karin is to

115

the left of the image of Kristin. The other variation, as I have indicated, would have a third particular that exemplifies the property of being-to-theleft-of and maintain that just as there is in the state of affairs of, say, Kristin’s being to the left of Karin some ordering relation that makes it that state of affairs and not its “opposite” so there is among the three thoughts that are the components of the thought that Kristin is to the left of Karin some ordering relation that makes it the thought that Kristin is to the left of Karin and not its “opposite.” Let us consider each variation in turn. With the first variation there are two fundamental difficulties, the second of which points to the other variation. The first difficulty is just that of the basic idea of the theory itself—that awareness consists in an essential way in the having of images. I cannot here make any sustained argument against the theory of awareness as images. Instead, I will urge that however causally important image-having may be to our species in our being able to have certain kinds of awarenesses, the having of images is, in general, ontologically neither necessary nor sufficient to having any particular awareness. A person who is conceiving that e=mc2 may have any from what are probably as many different images as there are people who conceive that fact. But in what images—visual, auditory, olfactory, kinesthetic—could the thought that e=mc2 possibly consist? We must insist that conceiving that e=mc2, whatever its proper ontological analysis may be, is wholly distinct from any images that may accompany its occurrence, however helpful the having of any such images may be to being able to do the conceiving. But let us accede temporarily to the theory of awarenesses as images in order to see that, even it if were true, it could not account for the idea of mental content. For while it might be just barely plausible in the case of the thought of Kristin’s being to the left of Karin to suppose that, in the person’s phenomenal space, the image of Kristin is to the left of the image of Karin, thus grounding its qualitative difference from the thought of Karin’s being to the left of Kristin, this theory cannot serve as a general account of what it is to think of something as being in some relation to something else. Keeping in mind that the theory before us is that there is, in the person’ mind, not the image or thought of the relation in the state of affairs of which the person is aware but the relation itself, we need only consider the thought of the arithmetical fact that ten is larger than five to

116

grasp that, whatever images may be involved for this or that person, those images cannot possibly stand in the arithmetical relation of being-largerthan. For images, however broadly conceived, are not numbers and cannot exemplify relations that only numbers can exemplify. At this stage of the argument, the defender of images as essential to the nature of awareness might suggest that the ordering relation of the images need not be the same relation as that in the state of affairs of which one is aware, that while the images involved in thinking that ten is larger than five admittedly do not and cannot stand in the arithmetical relation, they do stand in some other relation that secures the distinct qualitative nature of the thought—perhaps, in this case, the geometrical relation of “beinglarger-than.” But this suggestion raises a possibility, or seeming possibility, that is essential to the second variation of the theory that mental content is complex by virtue of a multiplicity of particulars, to which variation I now turn. The form of the theory now before us is at once the formally most sophisticated and therefore abstractly most plausible form of the theory that mental content is complex, but also, in the end, the most obscure and therefore concretely the least plausible form of the theory. It holds that a mental content consists of multiple particulars each with its own appropriate property standing not in the relation that the constituents of the object of awareness stand in but in some other relation. This relation is to serve to secure the difference in analysis between two thoughts of states of affairs that differ only in the order of their constituents as in the case of our examples but also, in general, to bind simple or simpler contents into a complex content. There may be many such relations, for there are thoughts of many different kinds of states of affairs with many different kinds of structures. These relations may be called “structural relations” (Meinong speaks of “similarity of structure” between thought and object), or as “certain relations” (as in Herbert Hochberg’s argument (1981, 157) against the simplicity of content), or perhaps only as the relations, whatever they are in their inherent nature, that do the ontological job that is required to be done on the theory. In any case, the relation in a particular case is, on this theory, never the same relation as that in the object of awareness. No one—neither Twardowski nor Meinong nor Hochberg nor any other defender of this theory—has ever given an example of any such structural

117

relation. The reason, I suggest, is that there are no such relations. There are, to be sure, relations between and among particular awarenesses that make them into a unity of a sort that might be described as a “line of thought” or a “sustained” or “concentrated” thought, but these relations are only temporal and causal and, looked at another way, logical. Thus there just aren’t any identifiable relations of the sort the theory requires, and the theory is therefore to be rejected.5 But this is obviously all too facile as a refutation of the theory, and it may be objected in particular that just because we have no labels for any such structural relations and cannot, therefore, identify any one of them specifically, it doesn’t follow that there are no such relations. Why can’t we just say, as was suggested above, that these structural relations are just the relations, whatever they are inherently, that serve to unify simpler contents into more complex contents in the way that is required in order to solve the ontological issues at hand? Why not indeed? There should be no objection, in the abstract, to the possibility of there being entities for which we have no names, at least in any of our natural languages, but for the existence of which there may be compelling arguments and reasons. Nor, in the end, is my decisive objection to this theory based on the considerations I have already mentioned, although they should give one pause. No, there are other, what we might call “dialectical”, reasons for maintaining that no relations of the requisite sort exist, as we shall now see. We may approach the problem by noting that Twardowski, an explicit defender of the complexity of (some) mental contents, was insistent that the relation in the awareness that ties together the simpler mental contents into a complex content was not itself part of the content. Restricting the expression ‘structural relations’ here to just those putative relations required by the theory, let us proceed by asking whether or not structural relations are themselves part of mental content. My argument will be that neither answer is defensible, from which it follows that there are no structural relations. In order to get fully into the argument, we must remind ourselves of the very idea of mental content with an emphasis that has not hitherto been 5

John Findlay makes a very similar point in his 1963, 16-17.

118

present in this essay. We said that a mental content is just that constituent of an awareness that uniquely correlates with the object of awareness or at least with its kind. But what is this “correlation” and how does a content manage to be, or get itself, so “correlated?” Although I cannot argue the matter here, I am going to assume that the answer to this question is that it is due to the inherent nature of a mental content that it does uniquely correlate with—or let us now say, with the tradition, intend—its object; in other words, that a mental content is a natural sign of its object. Although there are philosophers who might reasonably be said to be defenders of mental content who dispute or deny this assumption, I think it can truly be said that all of those who enter explicitly into the simplicity/complexity argument about mental content share the assumption. But the argument that follows, if I am not mistaken, partly supports as well as assumes the theory of natural signs. Returning to our prime example, we may say, then, that the property of being-the-thought-of-Kristin is a natural sign of Kristin, that of being-thethought-of-Karin a natural sign of Karin, and that of being-the-thought-ofto-the-left-of a natural sign of being-to-the-left-of. On the theory, these simpler thoughts are combined into the complex mental content expressed by being-the-thought-that-Kristin-is-to-the-left-of-Karin (and not its “opposite” or any other thought) by virtue of some structural relation that parallels the relations in the state of affairs itself that make it just the state of affairs it is (and not its “opposite” or any other state of affairs). Let us simplify by supposing that there is only one such relation in the awareness and one in the state of affairs that is the object of awareness so that we may ask: is or is not the structural relation a natural sign of the relation in the state of affairs? Suppose, first, that the structural relation is a natural sign. The supposition probably will seem absurd in itself—that one relation could be a natural sign of another relation—but leaving that aside we may proceed by asking if this relation that is itself a mental content is a simple content or a complex content. If it is a complex content, then on the theory before us it should yield to an analysis into simpler components that will also require a structural relation. This, obviously, leads immediately to a vicious infinite regress, and so cannot be correct. Perhaps, then, a structural relation is a simple natural sign of the relation in the state of affairs that is

119

the object of awareness. But in this case (and probably also in the case of its being complex), there will need to be, on the theory’s own principles, a combining relation that ties this natural sign to the other given natural signs in the thought. But that line of thought, just as obviously, also leads to a vicious infinite regress, and so must be rejected. The conclusion, the same as Twardowski’s so far, is that a structural relation cannot be a natural sign of the relation in the object of awareness. Suppose then that the structural relation is not a natural sign of the relation in the object of awareness. The question then becomes what makes a certain set of natural signs bound by that structural relation into the complex mental content in the thought that Kristin is to the left of Karin instead of its “opposite” or, for that matter, any other mental content. The answer can only be that while the structural relation is not a natural sign of its parallel relation, it nevertheless is correlated to that relation, at least in the particular case, and not to any other relation, and in the appropriate order. But how could it correlate without being a natural sign? With this question we come, finally, to the very heart of the issue—both analytically and historically, the latter insofar as this is the form of the theory of complex mental content that its main defenders all hold. Considering the only possible answer will reveal the falsehood of the theory. That answer is that we do the correlating of the structural relation in the awareness to the parallel relation in the state of affairs; or, to put the emphasis in a usefully different way, that we decide that the structural relation shall stand for the parallel relation; or, in yet other words, that the correlation is one of convention, somewhat like the way in which a word is only conventionally related to its referent. But in this case it is radical conventionality because it can only be individually conventional; unlike the case of language, there is no possibility of learning from others what structural relation correlates with what relation in the objects of awareness. One cannot learn; one must decide for oneself. Putting aside the phenomenological absurdity of this theory, we may note that it requires that a person, in making the correlation, already be able to think separately of the entities to be correlated. But this means in turn that there must be awarenesses not only of other awarenesses but also of mind-independent states of affairs prior to and independent of the making of the correlation. Yet the whole point of the theory was to explain how we

120

can have any such awarenesses only by virtue of the existence of such correlations. It is precisely at this point that the theory falls into incoherence beyond all possibility of resurrection. Or is there, possibly, one last hope for the theory if it is formulated somewhat differently: that while the structural relation is neither a natural sign of, nor something that a person correlates to, the relation in the object of awareness, the “whole” that is the mental content just does, by its inherent nature, uniquely intend its object. Thus, in our example, the “whole” that is the property of being-the-thought-that-Kristin-is-to-the-leftof-Karin is just different, but not in a way that can be shown by analysis, from the “whole” that is the property of being-the-thought-that-Karin-isto-the-left-of-Kristin. But this is either insistently to refuse to face the issue of the analysis of what is claimed to be a complex of a certain kind or else to agree, after all, that mental contents are simple. III The positive thesis that all mental contents are simple does not face the difficulties that I have exposed in the various forms of the contrary thesis just because all of those difficulties derive from the idea of complexity. But there may also be some problems with the idea of simplicity, and we must now consider what objections can be and have been raised against it. Some philosophers have had a problem with the very notion of simplicity, not only (or even, necessarily) in the context of mental content, that can be expressed by asking: if two putatively simple properties—red and green, for example—do not admit of analysis, then what is it that accounts for their difference? The question will seem especially forceful in those cases in which the properties (or relations) have all second-order properties in common. The answer to the question, in all such cases, is: they are just different. This does indeed amount to a denial of the principle of the identity (sameness) of indiscernibles (having all properties in common) at a certain level—here, at the level of properties and not, as it is usually considered, at the level of individuals. Some things are, or well may be, just different from each other without there being some constituent or property of the entities that makes them different. Or, if you prefer, it is the inherent nature of each (simple) property itself that makes any two of

121

them different from each other. So we should not be bothered by the very notion of a simple property in a way that would, for that reason, cause us to doubt that mental properties are simple properties. Another question that may be asked, in light of what has just been said, is this: if mental contents are simple properties that, as such, have nothing literally in common with each other (although they may exemplify, as contrasted with being constituted of, some common properties), what accounts for the obvious, pre-analytic “similarity” of the two properties of being-the-thought-that-Kristin-is-to-the-left-of-Karin and being-thethought-that-Karin-is-to-the-left-of-Kristin? On the theory that these are simple properties, they have no constituents in common at all; they are just two, different properties. And that seems to deny them their obvious “similarity”. The answer to this query is perhaps obvious: the pre-analytic “similarity” of the two properties is fully accounted for by the fact that the states of affairs intended by them have many, indeed most or even all, of their constituents in common. Kristin, Karin, and being-to-the-left-of are common constituents of the different facts intended by the two properties even though the two properties themselves do not have any constituents in common. Some philosophers will find this puzzling, but there is really nothing mysterious or paradoxical about it once one has accepted the very idea of a simple property. Another possible objection to the theory that mental contents are simple is that if it is correct, no sense can be made of the making of inferences, whether they be practical inferences leading to behavior or theoretical inferences leading to conclusions. Let us consider only the latter, for the point to be made is the same in both. If a person reasons from the thoughts that all humans are mortal and that Socrates is human to the thought that Socrates is mortal, we say, loosely speaking, that the fact that the first two thoughts entail the third both explains and justifies the person in that reasoning. But if we go on to explain how the entailment itself is justified, we are likely to appeal, in part, to common elements (as in a logic class, with the appropriate sentences on the blackboard) in the “propositions” involved. We would point out that there are three “terms” each of which occurs twice, and so on. But if each of the thoughts is a simple entity— merely different from each other with no elements in common—then it is

122

unintelligible, it may be said, how they could stand in the requisite or any logical relations to each other. The reader will probably, from my reply to the last objection, anticipate at least part of my response to this objection. It is that insofar as these three thoughts intend states of affairs that have, or would have, common or overlapping constituents, we can satisfactorily explain what is to be explained. For logical impossibility (to cast the point in that way) is first and foremost a relation among states of affairs themselves: it is logically impossible that the states of affairs of all humans’ being mortal and Socrates’ being human obtain and the state of affairs of Socrates’ being mortal not obtain. But I say further that I see no problem in maintaining that simple mental contents, just because they are natural signs, do by their intrinsic natures themselves stand in logical relations to each other insofar as they are entities with truth values. Indeed, it is natural signs and only natural signs that possibly could be the ultimate bearers of truth values. A final pair of objections have to do with the question of whether or not adherence to the simplicity of mental content commits one to Platonism, the theory that unexemplified properties exist. If so, and if, further, it is assumed that Platonism is false, one seems to have an argument against the simplicity of content. These issues are raised by Hochberg in his criticisms of Gustav Bergmann’s claim that mental contents, or what Bergmann calls “propositions,” are simple. The context is a very narrow one involving other claims of Bergmann’s, a particular form of the “ideal language” he used, and the transcription of sentences about awarenesses into that language. Thus, I am not concerned with all aspects of Hochberg’s criticisms, but we can usefully abstract from them some possible problems for the theory as I have formulated and defended it. Hochberg’s arguments can be regarded as making these two essential and related points. First, if mental contents are simple, then unless they exist they cannot be the subjects of necessary truths. But not all possible mental contents are exemplified in actual awarenesses; yet every mental content—exemplified or not—is the subject of some necessary truth. So some such contents must exist without being exemplified, and that is Platonism (Hochberg, 1978, 353). Second, an unexemplified mental content is understood only by description, as “the property that intends soand-so,” and never by acquaintance. This is intelligible only if either the

123

property is complex and can be defined through properties all of which are exemplified or the property exists without being exemplified, which is Platonism. So, Hochberg concludes, if mental contents are simple, then Platonism is true (1981, 156-157). Neither point is convincing. If the theory of simple mental contents commits one to Platonism because of there being necessary truths about them, exemplified or not, then the truth of ‘All golden mountains are golden’ does likewise. For this sentence is a truth, even a necessary truth (if that matters) that is about things that don’t exist. But then we would have a completely general argument, if an argument at all, for Platonism, one that is independent of the context of mental contents. Does the truth of ‘All golden mountains are golden’ entail Platonism? I think not (and I’m confident that Hochberg would agree), but that question takes us to issues outside the scope of this essay. My point here is that there is nothing in asserting the simplicity of mental content that is peculiar to the context that should lead one to Platonism. The other point, if I have understood it, tells us that for a predicate or referring expression such as a definite description to be intelligible, it must pick out either a property that does exist or one that can be defined through properties that exist. But if the property is never exemplified—as in the case of what would be the thought of the precise mass of the universe in grams—then either it is simple and so exists unexemplified, which is Platonism, or it can be defined, which is to deny its simplicity. So, the argument concludes, if Platonism is false, the property must be complex. Let it be granted that mental contents—all mental contents, if you wish --can be understood, in the sense of being referred to in language, only through definite descriptions. Can a definite description pick out its intended referent except by mentioning the constituent elements of the referent (all of which constituent elements presumably exist)? The answer is that there are other ways: the description can instead pick out its referent either by mentioning some unique property or relation that the referent exemplifies or by mentioning some entity to which the referent is uniquely related. For an example of the latter, for it is the relevant sort here: a particular shade of red, which presumably is a simple property, can be identified as the property that is lighter than such-and-such shade and darker than such-and-such shade. The point stands even if the simplicity of

124

shades of color is disputed, for no appeal or assumption as to its status as simple or complex has been made in order to pick it out in the way indicated. In the case of mental contents, the relevant description is that of (the form of) being the property that intends a certain object. And, contrary to Hochberg, the complexity of the object gives us no reason whatsoever to conclude to the complexity of the thought (or any other entity) that is identified by way of being the only entity that stands in this or that connection to the object. A particular shade of red might be uniquely identified as the color that is the favorite of some particular person, but the complexity of the person tells not at all as to the simplicity or complexity of the shade of color. To this reply it might he objected that an unexemplified property does not actually stand in the relation or connection that allows its unique identification. One should say instead that it is what would stand in that connection if anything did. And this, it might be suggested, is not adequate. My reply is that, adequate or not—and the issues of what makes counterfactuals true are difficult and important—the same point applies to identification by constituents: the property of being-a-present-king-ofFrance should, by the same reasoning, be specified as the property that would consist of the properties of being-a-king, being-France, being-inthe-present, and so on, if that property were exemplified. So there is no problem here, either, with the thesis that unexemplified mental contents can be understood by way of the connections that they would have if they were exemplified, and so may well be simple. I conclude that all mental contents are simple. IV The tradition of ontology as analysis, despite the goal inherent in its name, has been curiously reluctant with respect to certain particular ontological issues to allow that one actually has arrived at the simplest level, the level at which that is just the way things are. When the issue is that of thought, many philosophers have been overly impressed, to say the least, by the fact that the objects of awareness are, at least sometimes and maybe always, complex. Still not having broken entirely free of the old belief that there

125

must be some kind of important similarity of thought and object, these philosophers have remained indentured to the thesis that the thought also must be complex. If, instead, one does liberate oneself entirely from that thesis and focuses on the nature of awareness in itself, attending especially to the phenomenon and the datum of the unity of thought, one will be more receptive to the thesis of the simplicity of content.

126

THE NECESSITY AND NATURE OF MENTAL CONTENT1

T

he philosopher of our time who would attempt to understand the nature of mind is, if he or she is a thinker of reasonable sensitivity, inevitably moved by two overarching and, so some may think, contrary constraints, one very old and the other more recent, at least in details. The older one is that consciousness is something unique, something quite different from everything else in the known universe, that a universe with consciousness and one without it are radically different kinds of worlds. Even if the only consciousness in existence were the perceptions of simple insects, there would still be a phenomenon in the world that seems to be quite unlike any in a world altogether without consciousness. The apparent implication of this constraint or, seen positively, this imperative is that any adequate philosophy of mind must reflect, at the basic ontological analysis of consciousness, the uniqueness of mind. The other constraint, or imperative, is that consciousness is of a piece with nature generally and seems to exist only in intimate conjunction with or, some would say, in identity with highly organized pieces of matter. More generally, any adequate philosophy of mind must be fully consistent with the presuppositions, methods, and findings of empirical science and especially with those of evolutionary biology and the processes of natural selection. I would add, more contentiously, especially in these times of socalled intelligent-design theory, that the best theory of mind must be consistent with a conception of the mind as a product of purely natural processes, that is, of evolution as an unintended, lawfully explainable sequence of events that has taken, and is taking, place on our planet and, possibly, elsewhere. The apparent implication of this constraint is to deny or to disregard the uniqueness of mind and to seek an ontology of mind whose basic categories are no more than those of a world without mind. Although it is not my main topic here, I state without hesitation or 1

An earlier version of this paper was presented at a conference on intentionality held in Riskolc, Hungary in June 2002, organized by Gabor Forrai and George Kampis.

doubt my belief that the best philosophy of mind must reflect the uniqueness of consciousness at the basic ontological level but in a way that is fully consistent with a purely naturalistic worldview backed by the theory and fact of evolution. This sort of strategy guides the remarks that follow. If the phenomenon of consciousness is essential to what it is to have a mind, then the essence of consciousness is, in turn, characterized by its intentionality. Let us see. I Let us all imagine that the planet Mars has just exploded. If you have heard and understood what I just said, an event took place “in” you in some sense of “in” that is the event of your imagining that Mars has just exploded. You knew, and still know, that this event took place, and you know, too, that it was only an imagining and not, say, a perceiving of Mars exploding. You knew, and still know, that it was Mars and not, say, Jupiter you were imagining to have exploded. What is the nature of this event that took place in each of us of imagining that Mars has just exploded, and what significance should we give to the immediate knowledge or, if you prefer, the semblance— genuine or not—of the immediate knowledge that each of us had of the occurrence of that event “in” us? My focus in this paper is on that aspect of the kind of event that took place in each of us that, still pre-analytically speaking, is its content, that part somehow alluded to by the words “that the planet Mars has Just exploded” in the slightly longer expression “imagining that the planet Mars has just exploded.” The distinction, in a mental state of this kind, between its content on the one hand and its mode—in this case its being an imagining—on the other is a familiar one and indeed one of common sense. Yet only a few philosophers, especially among those of the analytic tradition, have attended explicitly to the ontological analysis of both of these features of events like imagining that Mars has just exploded, that is, of intentional states and events. Let us ignore for the time being those accounts of mentality that treat intentional states as mere states of the brain either directly as in simple identity theory, or more circumspectly as in functionalism or any other forms of causal theories of the mind, or merely as ascribed states to others

128

in order to explain their behavior. I will proceed by stating and defending an ontological account of intentional states and, more narrowly, of mental content, and then later examine some direct competitors to, and some direct criticisms of, theories of mental content like mine. I begin with a specification of just what mental states and events are intentional states and events. Because when I do philosophy of mind, I speak ontologically and not linguistically. I exclude dispositional mental states as being genuinely intentional states. Persons who are asleep or even unconscious have dispositional mental states, such as the belief that whales are mammals. I treat such mental states as dispositions both to some kinds of behaviors and to some kinds of occurrent mental states. Occurrent mental states figure in the characterization of dispositional mental states in much the same way that the property of dissolving-in-water figures in the property of being soluble: not as constituent (for something can be soluble without dissolving), but as definitional realization. In its inner nature, a dispositional mental state is, in us anyway, just a state of the brain as subject to laws, and as such, is lacking in literal intentionality. I have argued elsewhere that other occurrent mental states including the having of sensations, moods, and emotions are the only mental entities that have genuine, or literal, or what some call intrinsic intentionality (Addis, 1986). Indeed, a major thesis of this paper is that literal and fundamental intentionality pertains only to those occurrent mental states in us and other conscious beings that constitute what we call “the stream of consciousness.” So I affirm the unique intrinsic intentionality of occurrent mental states, not only in contrast with dispositional mental states, but also with the many other kinds of things and events to which we ascribe aboutness—language above all, but also pictures, maps, blueprints, artworks, religious and national symbols, gestures, computer programs, and much more. My main purpose of the moment is to delimit the class of mental things that are literally intentional; and along these lines, I want to note that we sometimes refer to occurrent mental states as conscious mental states in order to contrast them with dispositional mental states, but that ‘conscious’ is contrasted not only with ‘dispositional’ but also with ‘unconscious.’ The notion of unconscious mental states is, in turn, sometimes taken to mean a

129

subset of dispositional mental states, those that are realized in some kinds of abnormal behaviors and occurrent mental states, but sometimes also as a subset of occurrent mental states themselves, those of which the person is not, or perhaps cannot become, consciously aware. We thereby have the awkwardness of unconscious mental states being among the set of conscious mental states when we take the latter as the set of all those events that constitute the “stream of consciousness.” Once we are fully aware of this terminological oddity, living with it is easier than attempting to remove it. Returning to the notion of mental content itself—that feature of occurrent mental states that is or grounds their intentionality—we all know that it has, after a long period of dormancy in the analytic side of the tradition, again become popular in philosophy even if it sometimes takes quite bizarre forms. When we look at the subject historically, until quite recently the notion of mental content has always been that of something in the mind of the person who is aware that correlates uniquely with the object of awareness. So characterized, content may be what is called form by Aristotle and Thomas Aquinas (although the form gets us only to the kind of thing of which we are aware), ideas (idées) by René Descartes and many others of the early modern period, Inhalt by Alexius Meinong (1899) and others of the Austrian school of the act, matter (Materie) by Edmund Husserl (1971), propositions by Gustav Bergmann, (1960), or representative content as in John Searle (1983), and so on. Still speaking only historically, until the twentieth century no philosopher explicitly and probably none even implicitly denied that there is mental content as something in the mind of the person who is aware. Therefore, nearly none of them felt called upon to argue for the existence of mental content so conceived. Meinong is sometimes credited, as I myself have done, with having made the first explicit arguments for mental content in his famous paper of 1899, but probably the honor for this effort should go to Kasimir Twardowski in his rich and unjustly ignored book of 1894 (1977). Meinong’s and Twardowski’s arguments for mental content are not good arguments, and I will not rehearse them. Instead, I will present what I think are good arguments for the existence and specific nature of mental content. But first I want to remark on the fact that three of the most

130

important philosophers--Husserl, Bergmann, and Searle--who have explicitly affirmed an analysis of intentional states as consisting of, or involving the exemplification of, two kinds of properties—a mode property (Husserl’s Qualität, Bergmann’s species, Searle’s psychological mode) and an intentional property—never made explicit arguments for this analysis. In Husserl’s and Bergmann’s cases, they appear to have believed that we are, or can be, directly acquainted with instances of these kinds of properties, and so no argument would be necessary. In Searle’s case, whether he is speaking ontologically or only grammatically is unclear. He may be doing no more than merely calling attention to, and reaffirming, the commonsense way of speaking about intentional states, in which case again no argument would he necessary. This stance is puzzling because there is a widespread and quite natural tendency among philosophers to treat the mode feature of intentional states as a relation between the person and the object of awareness, as indeed grammar strongly and unfortunately suggests, instead of, as all three philosophers maintain, a monadic property of awarenesses. More important for our purposes, this stance is puzzling because two of the most important philosophers of the twentieth century—Bertrand Russell (1956) and JeanPaul Sartre (1957)—each in his distinctive way denied the existence of mental content. It becomes imperative, even if none of the three did it, to provide arguments for the existence of mental content while at the same time providing a precise and defensible ontological analysis of what they have argued exists. Here I can only sketch three arguments that I have made in detail elsewhere (Addis, 1989), wanting to have time also to criticize the Russell/Sartre theory and any theory of purely “external” content and to reply to some criticisms of any theory like mine, especially those of Hilary Putnam. I begin by noting a crucial feature of my conception of mental content and introduce an expression to capture that feature. The feature is that of being such that the content by its intrinsic nature and not by habit, convention, causal or other relation to something else, represents whatever it does represent. The expression, taken from William of Occam’s signum naturale (1957), is that of natural sign. So I will speak of the three arguments for the existence of natural signs. Let us keep in mind that this

131

is not the much more common use of the expression in which, for example. smoke is a natural sign of fire, and of which Ruth Garrett Millikan (1993) makes significant use in her theory of mind. I have labeled my three arguments as the scientific, the phenomenological, and the dialectical; and I will take them up in that order. The scientific argument for natural signs rests on two fundamental assumptions, widely but not universally held, about the causal explanation of human behavior or, indeed, of the behavior of any macro-objects whatsoever. Restricting ourselves to the human case and putting aside any notions of indeterminism at this level, the assumptions are: (1) If two persons behave differently in exactly the same external circumstances, there must be some prior difference in them that explains the difference in behavior (the principle of different effects, different causes). (2) The nature of those differences in the persons, in their role as immediate causes, must be that of being monadic properties of the persons and not of any relations they have to anything else. Such relations may well be mediate causes, but they can affect subsequent behavior only insofar as they result in non-relational properties of their relata. The point I am about to make stands whether or not the objects of awareness are existents, but an example in which those objects are nonexistents will be easier to grasp. So let us consider two persons who are alike in all respects except that one is thinking of mermaids and the other is thinking of unicorns. Speaking commonsensically, we would attribute their different responses to the question “What are you thinking about?” to the differences in their thoughts which, again speaking commonsensically, we would take to consist in part of particular properties of them. The different prior states that seem to explain the differences in behavior—in this case, linguistic behavior—just are the states of mind of the two persons, the one thinking of mermaids, the other of unicorns. In any case— and that is what this example shows most convincingly even though the argument holds also when the objects of awareness are existent and even

132

when they are also the causes of the awarenesses themselves—it cannot be the objects of awareness per se that are the causes of the differences in behavior, for they do not exist. Even if the objects of awareness are existent, they cannot he the immediate causes of differences in behavior if only, in some cases, because they are too distant in space or time. The conclusion to be drawn is that if, commonsensically speaking, two persons have qualitatively different thoughts, the thoughts exemplify some different monadic properties. These properties satisfy one of the conditions of natural signs in that they correlate uniquely with the objects of the thoughts. More generally, for every qualitative difference in what two people might he thinking, there will need to be a difference in their monadic properties in order to account, causally, for subsequent differences in actual or dispositional behavior. This argument does not show that these monadic properties are, in a strong sense, signs—natural or otherwise--of the objects of awareness. It does establish the minimum condition of so-called content internalism-that there must be monadic properties of the person that are unique correlates of the objects of thought. In this connection--and as a preparation for the phenomenological argument for natural signs—we may consider briefly the views of G. E. Moore, one of the great defenders of the act in the analytic tradition, in his ultimately unsuccessful attempts to understand the nature of consciousness. Moore was convinced, in light of what he called the “diaphonousness” of consciousness, that there are no mental contents; a state of consciousness has no intrinsic feature that indicates its object. At the same time, considerations of the sort I have just advanced as the scientific argument for natural signs occurred to Moore. Both of these impulses are evident in this brief quote from his still-important paper, “The Subject Matter of Psychology,” where he says: it cannot be the different objects which produce the different effects; and therefore there seems to me to be some force in the argument that there must be some internal difference in my consciousness of the one and of the other, although I can discover none (1910, 56).

133

This tension led the usually sober Moore to suggest a view in which the existent and the non-existent combine to explain differences in behavior, a theory he happily never mentioned again. But what should we make of Moore’s inability to find contents in his own states of consciousness? This sort of consideration leads me to the phenomenological argument for natural signs. It does not suffice to declare that Moore and others have not looked carefully enough, that the requisite entities—natural signs—are there for anyone to find in his or her stream of consciousness. That may be true, but to say so is not convincing, nor should it be, to anyone who doubts the existence of natural signs. The phenomenological argument is instead the assertion that immediate acquaintance with natural signs is the best explanation of the feeling of certainty that each of us has right now that we are imagining Mars having exploded and not some other state of affairs or nothing at all, Yet, some philosophers, in the grip of causal theories of reference, have argued— Putnam, for example, in his brain-in-the-vat fantasy—that a person might be radically mistaken about the objects of his or her thoughts. But skepticism about the objects of your own thoughts is one of those theories that no one believes or could believe. All of us, like it or not, are Cartesians; all philosophizing, indeed all reflective conscious life, begins with the Cartesian assumption of certainty about the objects of thought— which is an excellent reason for rejecting at the outset any causal theory of reference or any other theory that affirms or implies the contrary So we all know that Mars exploding is what we are now imagining and not something else. Our certainty attaches only to the fact that we are imagining that Mars has exploded, and not to Mars actually having exploded. The best explanation of that feeling of certainty is, I submit, that each of us is directly acquainted with a monadic property of ourselves that, by its intrinsic nature, “points to” or, as we philosophers like to say, intends, the state of affairs imagined; this property is a natural sign of that state of affairs. To be aware of that state of affairs--Mars having just exploded--is just to exemplify such a property; to be sure that one is aware of that state of affairs is to be aware of exemplifying such a property. My thesis does not require that you know that you are aware of that property; for here, as in many other kinds of situations, the best explanation of some behavioral or mental phenomenon is that the person

134

had an awareness of a certain character even if the person was not aware of having that awareness, does not remember having had the awareness, and, in extreme cases, denies having such awareness. People with blindsight can see, for that is unquestionably the best explanation of some of their behavior and other mental states, even though some such people vehemently insist that they are blind. This third argument for natural signs—and this one is most clearly directed to the idea of mental content as consisting of intrinsically intentional entities—rests on the claim that if there is to be any kind of representation in addition to natural representation, that is, if there be either purely conventional representation or what I call quasi-natural representation, then there must be natural representation, a kind of representation that involves entities that by their intrinsic nature represent whatever they do represent. By quasi-natural representation I mean a kind of representation that occurs, or so I have argued (Addis, 1999), in music and possibly also in dreams and religion, due to human nature and the kinds of representing entities involved, that is, lawfully necessary representation. But we may restrict ourselves to the contrast between conventional representation as instanced by language, for example, and natural representation as I specify my argument in more detail. The word ‘red,’ we say, only conventionally represents the color red. There is no resemblance or any other two-term relation that connects the word to the property that the word does not also have to many if not all other properties. How then does the word ‘red,’ for one subset of human beings, manage to represent a specified color? The answer at one level is that those human beings just make it so represent, by choice, habit, or otherwise, but by choice in the sense that English speakers could, in principle, simply decide otherwise I want to express this situation by saying that conventional representation requires a third term, that is, conventional representation is a minimally three-term relation; a first thing (a word) represents a second thing (a color) because of a third thing (a human being). The third thing cannot be just any kind of thing; it must he a conscious being who is able to be aware of both the first thing and the second thing in order for the first to represent the second. More generally, and as most philosophers would probably agree, without conscious beings, there would be no such

135

phenomenon as representation or intentionality whatsoever; a universe without minds is a universe altogether lacking in that asymmetrical connection of aboutness, whatever resemblances or any other relations its constituents may have to each other. But then, my argument continues, if a person’s awareness of, in our example, the word ‘red’ and the color red required in each, or either, case a third, we would be off on an infinite regress of the vicious variety. For then we would need yet a further explanation of how that third was aware of both the thought and what it represents, and so on. The only way to avoid this regress is to suppose that thoughts. or some constituents of them, are such that they represent by their inherent nature, without the intervention of a third. They are natural signs. Dale Jacquette, in his book Philosophy of Mind, links me with Noam Chomsky and Jerry A. Fodor as a defender of a “language of thought.” On the contrary, I must stress that natural signs, as I conceive them, are utterly non-linguistic items. To say of my view, as Jacquette does, that I “explain thought in terms of language, rather than the other way around,” is a mistake (1994, 105). Indeed, I am one who insists on the primacy of mind to language. At one level, the main opposition to the argument comes from those who, like Wilfrid Sellars and his students, following Ludwig Wittgenstein, hold that thought in its intentionality can he understood only through the “intentionality” of language. These philosophers based their line of thought, I surmise, on the false belief that language is a wholly public phenomenon as contrasted with thought, and the further thesis that the “private” becomes philosophically respectable only by way of the public. This belief is false because the essence of language—its aboutness, its semantics—is not a publicly observable feature of it and is only the intentionality of’ thought itself. What would it he like to observe as a public phenomenon the representation of the color red by the word ‘red’? II So there you have a theory of mental content, with some brief arguments in its favor. In recent language, some writers would refer to it as a theory of internal content. Alternatively, if there is any philosophical significance in

136

the notion of “broad” content, this theory might he called a theory of “narrow” or psychological” content. For the moment, I do not want to consider those many theories that would add something to psychological content in order to arrive at what their proponents would consider more nearly adequate accounts of intentionality. Instead, I will address those more radical theories that we might call either no-content theories or theories of wholly external content. These theories deny that there is anything at all, on the side of the mind, that correlates to, or otherwise indicates, what, or any part of what, a person’s object of awareness is. We already saw this view hesitantly expressed by Moore, but we also know that Russell and Sartre defended the theory with more confidence, Russell expresses the theory by saying, “At first sight, it seems obvious that my mind is in different ‘states’ when I am thinking of one thing and when I am thinking of another. But the difference of object supplies all the difference required” (1956, 171-72). For Sartre, ascribing any (descriptive, monadic) properties to a state of awareness (he does not express it that way) would make the mind into a “thing” and so make freedom impossible. More recent theories of wholly external content have other motivations, but all such theories, I suggest, are incapable of answering the following two criticisms. The first objection—what we might call the empirical objection— reiterates what was said in the scientific argument for (internal) mental content. That is, if two persons in the same situation behave differently because, speaking commonsensically, they have different objects of awareness, there must be a difference in the persons themselves to explain the differences of behavior and not merely a difference in the objects of their awarenesses. Unlike Moore, Russell and Sartre never faced up to this obvious difficulty, to my knowledge. Sartre, perhaps, with his radical libertarianism might have welcomed the consequence, but neither Russell nor recent defenders of wholly external content could allow that there just is no any scientific explanation of the differences in behavior. Let them then tell us just what the mechanism is, or could be, for the explanation of the differences. I believe that no coherent answer is possible. The same is true, I believe, of what we may call the ontological objection. Unlike the empirical objection, it requires us to consider

137

awarenesses whose objects do not exist, cases in which what is, or would be, the external content is non-existent. So let us now all think of mermaids. Now let us all think of unicorns. On the theory of’ wholly external content, there is nothing in us, in our minds that distinguishes the one thought from the other. But equally, and this is the point, there is nothing outside our minds either such that, in each case, we might be said to stand in some relation to the one but not to the other, nothing in the actual world that distinguishes the one case from the other. On the theory of wholly external content there is nothing that distinguishes one situation from the other. Or to put the point in the arcane language of “possible worlds”: on the theory of wholly external content, a possible world in which a person is thinking of a mermaid is, all other things being the same, qualitatively identical to a possible world in which that person’s counterpart is thinking of a unicorn. That, surely, is an absurd consequence, which, I would imagine, no one would deny. Therefore, it remains for the defender of wholly external content to try to show that this is not a consequence of the theory. I believe that any such attempt, as, for example, by invoking Russell’s so-called multiple-relation theory according to which to think of a unicorn is to think separately of the exemplified properties that constitute the unexemplified property of being a unicorn, will fail. Although these arguments are all too brief here, I conclude that the theory of wholly external content is helpless in the face of these objections and, therefore, false. I turn now to a different kind of objection to natural sign theory, one that directly challenges the idea of a natural sign, especially as expressed in the writings of Hilary Putnam. In a section entitled “The Reasons for Denying Necessary Connections between Representations and Their Referents” from the first chapter of his book, Reason, Truth and History, Putnam maintains that the theory of intrinsically intentional entities is, “a survival of magical thinking” (1981, 3) and the idea that things have “true” names, knowledge of which gives a person power over those things. I suspect that this piece of historiography, if that is what Putnam intended his claim to be, is false. Regardless, this is completely irrelevant to the plausibility of the theory. Putnam does make arguments even though the topic of the chapter section shifts mysteriously and apparently unconsciously on Putnam’s part

138

from what it is to have an occurrent awareness of something to what it is to understand a sentence. He directs part of his argument against the thesis that images or sense data are intrinsically intentional. I have no disagreement with that. From there he moves, all too hastily, to the conclusion that concepts are the only remaining candidate for intrinsically intentional entities and then argues that concepts are not mental items at all, for to have a concept is only to have a behavioral disposition or capacity that is caused in a particular way. In this context, Putnam presents his now well-known Twin-Earth example, which is also part of his argument against necessary connections between thought and object. He presents his argument in terms of concepts which he has already denied are mental items; but I assume, with others and probably Putnam himself, that we can take it to apply to any kind of entity or putative entity whatsoever that is claimed to be intrinsically intentional. In my criticism of that argument, I too will use the language of concepts. Let us orient ourselves a little more precisely. We must surely agree-for this is obvious from considering a simple disjunctive awareness—that the actual truth conditions and, in that sense, the actual referent for a given thought may be different in different possible worlds. I take Tyler Burge’s argument for social content (1979) to be but an elaborate variation on this idea along with his implicit assumption of the falsity of a descriptionist theory of reference. If all Burge, Putnam, or anyone else means, in saying that what is in the mind does not wholly determine its referent, is that the referents of qualitatively identical thoughts will vary in different possible worlds, we should reply, yes, that is true and we already knew it. But Putnam purports to describe a situation in which, in the same possible world, exactly similar thoughts have different referents, This, if true, would be quite interesting indeed. Whether Putnam has succeeded is highly questionable, as I will now try to show. As we all know, Putnam asks us to imagine that each of us has a Doppelgänger, or exact twin, on a distant planet whose thoughts are qualitatively identical to our own. But, according to Putnam, if it happened that on that planet what is called water is composed of elements XYZ and not of H2O as water is here, then the referents of absolutely qualitatively identical thoughts of those who are thinking of what each calls water here

139

and on Twin Earth are different because their extensions are different. Putnam expresses this conclusion by saying that “contrary to a doctrine that has been with us since the seventeenth century, meanings just aren’t in the head” (1981, 18-19, his emphasis). But, on the assumption that would make the conclusion quite interesting— that Earth and Twin Earth are in the same possible world— the conclusion just does not follow. For, either your and your twin’s concepts of water do include its chemical composition or they do not. If they do—yours being that of water as H2O and your twin’s as being XYZ—then we obviously do not have exact similarity of concept (content) to begin with. Suppose instead that the concepts do not include the chemical composition, in which case we might say that water is the concept of “the clear, odorless liquid that quenches my thirst.” Ignoring the possible different effects of H2O and XYZ, and, for the moment, any other indexical aspects of the situation, in that case our concepts do have exactly the same extension; namely, all of what is composed of H2O, and all of what is composed of XYZ, and all of whatever else there is, if anything, that is a clear, odorless liquid that would quench my thirst. In neither case, therefore, do we have sameness of concepts with difference of extension. That can happen, indexicals aside, only across, but never within, possible worlds. Putnam’s implicit reply to this argument appears in one of his much more extensive discussions of meaning, “The Meaning of ‘Meaning’” (1975), in which he insists that so-called natural-kind terms like ‘water’ are implicit indexicals. The thesis is that even if a person’s concept of water does not include its being composed of H2O molecules, it does include the notion that all water has the same (or perhaps highly similar) composition. If this liquid in the glass before me is water, and is composed of H2O, then all water is composed of H2O. In addition, insofar as I take this liquid to be water, the extension of my concept of water is all and only what is composed of H2O. Finally, according to this line of thought, even if for me water is only ‘‘this clear, odorless liquid that quenches my thirst,” any liquid with those properties that is not composed of H2O is not in the extension of my concept of water. I can find no reason to agree that my or anyone’s concept of water,

140

insofar as it does not include its chemical composition, includes the notion that all water is of the same (or highly similar) chemical composition. Nor can I find any direct argument for the thesis in Putnam’s writings. The closest thing I find to an argument are two assumptions that are, ironically, contrary to the spirit, if not the main theses, of the essay and to Putnam’s thought as a whole. Those two assumptions are: (1) that there is a neat category of concepts, of which the concept of water is a member, to be called natural-kind concepts; and (2) that the concept of a natural kind is such that the members of a natural kind must all be of the same or highly similar chemical (or other) composition. If I were to submit that my concept of water does not require that all water be of the same chemical composition, I imagine it would be objected that this is to put dubious phenomenology ahead of more reliable . . . what?--neat categorization and conceptual analysis? But can we discover through conceptual analysis, as Putnam appears to claim to have done, that our concepts of water and relevantly similar notions must, as so-called natural kinds, include the mono-composition feature? I doubt it in the extreme, but must here leave the argument and conclude that we have no reason from Putnam to deny the possibility of natural signs conceived as entities that specify only their possible truth conditions but which, if they are multiply-exemplified, will have the same actual truth conditions within the same possible world, I find it curious that analytic philosophy is, in its name, committed to the idea of a mode of understanding the world by way of the simples that, in some sense or senses of ‘simple,’ make it up while so many of its practitioners resist the suggestion that, in any particular case, one might actually have arrived at that level. Therefore, in many quarters some will continue to consider it impossible that one might have, that one could have, arrived at entities that just, by their nature, represent and are, therefore, the ultimate terms of analysis in the theory of intentionality. It just has to be more complicated than that, these critics will say. I will not try to argue my case any further in this regard but, in keeping with my title, close by outlining an argument that natural signs are simple not only in being the final terms in the analysis of intentionality but also in the sense of not being complex properties. Even though they are monadic properties that intend complex states of affairs, they are themselves such

141

that they have no properties as simpler constituents. My thesis is, in short, a rejection of the widely-held dogma that the complexity of the object of awareness must be reflected in an ontological complexity of the awareness itself. The basic idea of my argument, made in greater detail elsewhere (Addis, 2000), is that as soon as one allows a complexity of property in any property that intends, there is no way to account for what is traditionally known as the unity of thought and no way to distinguish qualitatively different thoughts. Let us see. There are various ways in which a thought might be, or be conceived as being, complex. But all of them, including Russell’s “multiple-relation” theory, are unequipped, so I claim, to explain what must be explained in the following example. There is the thought that A is to the left of B, and there is the thought that B is to the left of A, both of which thoughts we all now have. These are clearly qualitatively distinct thoughts, as I will assume, if A and B are, and are thought of, as qualitatively distinct. Also obvious is that not both, and therefore not either, can he analyzed as a complex consisting only of the thought of A, the thought of B, and the thought of to-the-left-of. So the putatively complex intentional property of being-the-thought-that-A-is-to-the-left-of-B, for example, cannot consist only of the putatively simpler intentional properties of each item. But it will be and has been said that something has been left out—a relation that orders the three, putatively component thoughts one way in the thought of A’s being to the left of B and the other way in the case of B’s being to the left of A. Following Meinong, who speaks of a “similarity of structure” between thought and object, we may call such a relation a structural relation. I can now reformulate my thesis as the claim that there cannot be any such thing as a structural relation, so conceived. The reason is that both the assumption that the relation is a constituent of the intentional property, and the assumption that the relation is not a constituent of the intentional property, leads, in different ways, to vicious infinite regress, as might perhaps be half-suspected even without thinking it through. Twardowski made an argument for the impossibility of a structural relation’s being part of the content itself but failed to see, additionally, that it cannot be something that is there and not part of the content,

142

So, all too quickly here, I conclude that the natural sign that is the intentional property that-A-is-to-the-left-of-B is, like all natural signs, a simple property that is just different from the natural sign of B’s being to the left of A, much as the simple property of phenomenal red is just different from that of phenomenal green. As intentional properties, and unlike red and green, natural signs point to objects that, in my example, clearly do have constituents in common. But the natural signs, as simple monadic properties, themselves have no shared constituents just because they have no constituents and yet are numerically distinct. III I have argued that natural signs, conceived as entities that intend by their inherent nature, exist; that they belong to the ontological subcategory of simple, monadic properties; and that every state of awareness has such an entity as a constituent. What I have not done is to discuss the nature of the connection between a natural sign and that of which it is a sign, as a full discussion of the topic would require. In concentrating on mental content as natural signs, I hope, nevertheless, to have contributed to our understanding of that curious and endlessly fascinating phenomenon that is the subject of our conference on intentionality.

143

EXPLANATION

DISPOSITIONS, EXPLANATION, AND BEHAVIOR1

T

he theory of dispositions defended in this paper was first advanced, to my knowledge, by Kaila (1945). Pap (1958) urged a closely similar view and the one espoused by Hochberg (1967) retains some of its essential features. In my judgment, while this theory is the most plausible account of the nature of dispositions, it has not received the kind of defense it, or any theory of dispositions, ought to have. Furthermore, the earlier proponents of this theory, in common with nearly all other advocates of this or that theory of dispositions, failed to relate it to an account of dispositional explanation. Finally, if I am correct, this theory is applicable in a philosophically illuminating way to the issue of reasons and causes and to a related controversy concerning so-called “rational” explanations of human behavior. I The philosophical problem of dispositions arises from the circumstance that dispositional properties are not observable ones. Yet many of them such as being magnetic and being acid are of systematic importance in science. The dispositions of things are often the very terms of scientific classification. Observable properties of substances such as their colors, textures, odors and so on play but a secondary and derivative role in chemistry, for example; the real measure of what a substance is comes from discovering its dispositions. This fact, which on some theories of dispositions should be entirely mysterious, is on the one here defended just as it should be. For, on it, to have a disposition is, roughly speaking, to be

1

Earlier versions of this paper were read at University of Minnesota-Minneapolis, Mankato State University, State University of New York at Binghamton, Drake University, and the University of Iowa. I received constructive criticism from people at all these places. I would mention especially the written comments of Peter Arneson (1978) which engendered significant revision of some of the arguments on my part.

subject to a certain law (or number of laws); a little less roughly, to have some property by which the object said to have the disposition is subject to a certain law. At the same time, we may be confident that dispositional notions have been part of human thought as long as there have been language users. Persons have surely been thought of and described as sleepy, for example, as long as humans have had even the rudiments of language; and being sleepy is a dispositional property. The same can be said for being edible, being dangerous and many others. This fact—that dispositional concepts are by no means peculiar to science or the context of scientific explanation—reminds us that any adequate theory of dispositions must accommodate not only the dispositional concepts of science but also those of everyday life and language use. But what should a theory of dispositions attempt to do? The problem of dispositions, I noted, arises from the fact that dispositional properties are not observable properties. One cannot look at an object and see its acidity as one sees its color. Thus any adequate theory of dispositions should provide a schema for the treatment of individual cases2 of ascriptions of dispositions such that all of the property terms that occur within a specific analysis are themselves—at least within the context—the names of observable properties. Little would have been achieved otherwise. Eliminability in principle of dispositional terms is a feature of the singular material implication analysis, for example, according to which: Dx “means” O1x ⊃ O2x that is, some object x’s having a certain disposition D is analyzed as: if x has some observable property O1, then it has the observable property O2. It has sometimes been assumed that if one doesn’t accept the singular material implication analysis of dispositional terms, but is also committed to the kind of theory that would permit their systematic elimination from ordered discourse, one is necessarily accepting a non-Humean view of 2

I shall deal throughout only with the analysis of singular disposition statements. That way the analysis can be applied directly to the treatment of dispositional explanations of the behavior of individual things and persons. But as will be obvious, the analysis of dispositions proposed here is easily applied to general disposition statements.

148

causation. The idea is that on any alternative to the singular material implication analysis the connection will be “stronger” and so no longer be the constant conjunction of the Humean. This belief is not quite correct for, as we shall see, on the theory here advocated the counterfactuals problem with respect to dispositions, while not entirely eliminable, can be shifted entirely to laws themselves, with the result that the problem of the object which lacks a certain disposition but is never put to the test becomes only the problem of distinguishing genuine laws of nature from accidental generalities. This I take to be a strength of the theory. Here is the theory. Letting ‘D’ be the dispositional predicate, ‘O1’ and ‘O2’ be observational predicates, and ‘f’ be a predicate variable, we have: Dx “means’” (∃f)(fx . (y)[(fy . O1x) ⊃ O2x]) that is, some object x’s having a certain disposition D is analyzed as: there is some property that x has and (it is a law that) anything that has that property and has O1 also has O2. Illustrating this in a particular case will make the idea clearer. Taking the sentence form ‘x is brittle’ and letting ‘BR’ stand for being brittle, ‘S’ for being struck with force, and ‘B’ for breaks, we have: BRx “means” (∃f)(fx . (y)[(fy . Sy) ⊃ By]) that is, some object x’s being brittle is analyzed as: there is some property that x has and (it is a law that) anything that has that property and is struck with force breaks. Before I consider what I take to be the strongest objections to this proposal, a series of comments will bring out the most important features of it. First, both the strength of the theory, I believe, and its peculiarity, all may agree, lie in the use of the predicate variable. This use implies that, in ascribing a disposition to a particular thing, reference is being made to a property which is unspecified except to say of it that it stands in a certain lawful connection with some other properties that are specified. It is obvious that what one hopes to get at by this notion is that of the (nondispositional) ground in the thing or person said to have the disposition.

149

The occurrence of a variable in the analysis as well as this reference to the ground of the disposition pose certain problems which I shall consider later in the form of replies to objections. Second, another peculiarity of the analysis resides in the expression in the English paraphrase of the formal analysis: ‘it is a law that’. Bergmann (1951b) has argued that the notion of a disposition cannot be captured faithfully in a systematic reconstruction at all, since such an attempt would involve an illegitimate mixture of levels of discourse. I agree. A more faithful idiomatic rendering of the view I am espousing would look something like this: Dx “means” (∃f)(fx . ‘(y) (fy . O1x) ⊃ O2y’is a true law)] This is indeed an illegitimate conjunction of an object-language statement and a meta-language statement. So I would say, granting the inaccuracy, that my view, as originally stated, comes as close as one can, in a formal reconstruction, to capturing the notion of a disposition. More precisely, as long as one insists that a lawful generality (as opposed to a mere accidental generality) cannot as such be represented faithfully in a formal reconstruction, one cannot capture with absolute fidelity the notion of a disposition in a formal reconstruction. But having once taken note of this untidiness, I propose to ignore it henceforth and proceed as if it were not there. It is simply one of the limits of formal analysis that one must be prepared to live with.3 Third, although in this analysis the use of the variable is essential and irreducible, one will naturally want to know what would count as a legitimate substitution for it or, more precisely, what is to be excluded as a legitimate substitution,4 and why. There are properties5 of three classes that, 3

This apparently cavalier dismissal of an apparent flaw in my analysis is only apparent. If someone wishes to insist that a lawful generality can be represented as such in a formal reconstruction through some connective other than material implication, he or she is invited to make that substitution. In short, I am not in this paper concerned with the correct analysis of lawfulness itself. 4 For more discussion of this problem, see Pap, 1958. 5 I shall write indifferently of substituting (or excluding) either predicates or properties for the variable, just as I write indifferently of analyzing either dispositions or dispositional terms.

150

for different reasons, must be excluded: (1) any dispositional property, (2) any property that would make the generalization logically true, and (3) any property that would, in conjunction with O1, produce O2 but would not, as we use the language, lead us to affirm that an object with it has the disposition being considered. As to (1): the reason for excluding dispositional properties is obvious. The most important one to be excluded, if such a matter admits of degrees, is, of course, the property that, in a particular case, is itself being analyzed. As to (2): since the resulting generalization from the substitution must be an empirical and not a logical one, it is necessary to stipulate that no property may be substituted for the variable which would make it a logically true sentence. O2 is one such property; O1 ⊃ O2 is another; and of course D itself is yet another. Of others, there are an infinite number. As to (3): consider a hard rubber ball with an explosive inside that causes the ball to break when struck with force. Is the ball brittle? If one decides to answer no, then one must exclude the property of containing an explosive as a legitimate substitution for the variable. The same principle applies in all logically similar cases. A few paragraphs hence I shall consider an objection based on the necessity of these exclusions. Fourth, there is nothing in my account that requires that the property that grounds a given disposition of something be a property of its finer structure. Hence, lacking a finer structure is no bar to an entity’s having dispositions. Assume that electrons are fundamental entities in the sense of having no finer structure and consider what is expressed by ‘This electron is posistructive’ (where being posistructive is some dispositional property concerning the electron’s behavior under impact with some other atomic particle). If the statement is not taken to be definitionally true (even ignoring its existential component), then the property of being an electron is or is not itself the ground of the property of being posistructive. If it is not the ground, then some other property of the electron is the ground and the analysis proceeds as hitherto described. If, on the other hand, the ground of the disposition is the property of being an electron itself, we have a statement that is simply an instantiation of the law that every electron is posistructive. This last-mentioned possibility gives rise to a problem that I shall consider in my replies to objections. Fifth, one may ask precisely what sort of statement the lawlike formula

151

‘(∃f)(y)[(fy . O1y) ⊃ O2]’ is. As the logicians speak, it is an existential generalization. So one may wonder whether it should even be called ‘lawlike’. For myself, I cannot see that it makes any difference what one calls it. Some philosophers would consider it to be “ontologically stronger” than a corresponding proposition in which the predicate variable is replaced by a constant, on the ground that it quantifies over predicates whereas genuine laws do not. Again, for myself, I do not accept this criterion of ontological significance: if anything ontological—in the sense of whether or not there are properties—is implied by the truth of sentences of this kind, then the occurrence of the names of properties (or, more neutrally, of predicate constants) as much implies their existence as quantifying over their corresponding variables. Sixth, a distinguishing feature of this theory is that it appears to eliminate entirely the problem of the object which lacks a certain disposition but has never been put to the test. This problem has been historically the basis of the crucial objection to the singular material implication analysis, and the one that moved Carnap, Hempel, and others to adopt the notion of “reduction sentences.” On the theory here espoused, the truth or falsity of a dispositional statement (about a particular object) depends not at all on whether or not the object is put to the test. Rather, all that is required is that it possess the ground of the disposition and that a certain lawlike statement be true. The significance of this point will become clearer in my reply to the first objection. The four objections I shall consider are, in summary, as follows: first, that, contrary to my claim and intention, the problem of the untested object that lacks the disposition remains; second, that the theory is inadequate on purely formal grounds, because it fails to provide a schema for the statement of sufficient conditions for the application of a disposition term to an object; third, that the relation of a ground to a disposition is purely empirical and not, as the theory maintains, “definitional”; and fourth, that the theory entails a vicious infinite regress of properties and laws. Consider first the argument of someone who objects as follows: “But what of O1 ⊃ O2? Since, on your theory, from the fact that something has D it follows by simple logic that it also has O1 ⊃ O2, we may conclude that the latter too is a genuine property of the object. And whether we call it a disposition or not, doesn’t there remain the problem of the untested object?

152

If, for example, D is the property of being soluble, then whatever has it also has the corresponding O1 ⊃ O2 property of dissolving if put into water. But doesn’t the non-soluble, untested object also have this latter property? So what have you gained?” I reply as follows: It is not my intention to deny that what is expressed by the singular material implication is a property. But it is not a dispositional property. In the example just mentioned, what is expressed by the singular material implication is not the property of being soluble but rather, more perspicuously expressed, simply the property of not being put in water or dissolving. Consider my desk which, by assumption, never has been, is not, and never will be in water. I see no objection to saying, once one sees through it all, that it is true of my desk that if it is put in water it will dissolve, but false of it that if it were put in water it would dissolve. Only the latter expresses the property of being soluble and is, once again, letting ‘a’ stand for the desk, ‘W’ for being put in water and ‘DI’ for dissolving, analyzed as follows: (∃f)(fa . (y)[(fy . Wy) ⊃ DIy]) Since, by assumption, my desk lacks any property that is a legitimate substitution for the variable, this sentence is, as it should be, false. In short, the way to take the ‘were-would’ out of a singular, unanalyzed ‘if-then’ is to “postulate” another property actually possessed by the object and to point to a certain lawful connection between it and the properties explicitly mentioned in the ‘if-then’ sentence. A second objector reasons as follows: “Any adequate analysis must provide both necessary and sufficient conditions for the truth of the sentence being analyzed. Formally, the theory provides a (schema for stating a) necessary condition, but insofar as it does not and cannot allow for unrestricted substitution for the variable, it fails to provide a (schema for stating a) sufficient condition for the truth of any given dispositional statement. Furthermore, one cannot, even in principle, list all the exclusions—all the predicates which, without defeating the whole idea of the analysis, cannot be permitted as substitutions for the variable. So, on formal grounds alone, the theory is inadequate.” I reply as follows: I do

153

take analysis of the present kind6 to be the (providing of a schema for the) specification of truth-conditions, that is, an indication of what the world must be like in terms which are, at least in the context, antecedently unproblematic in order for the sentence being analyzed to be true. Such a specification of truth-conditions must be consonant with the meaning ordinarily attached to the problematic notion (or to the sentences in which they occur), but need not be implied by it (or them). (So I am not analyzing what people necessarily do have, or even on reflection would have, in mind when they use disposition sentences.) It is true that, on the theory, the schema for the analysis of any particular disposition sentence would provide a (schema for specifying a) sufficient condition only if unrestricted substitution for the variable were permitted. Since it is not, one would have a statement of a sufficient condition in the particular case (as contrasted with the schema) only if the particular exclusions for every dispositional statement were actually listed or if a way could be found in the schema itself to represent a mechanism for generating the necessary exclusions for each particular case. Neither of these is, even in principle, possible. But there is a simple way to state the exclusion of any given property, from whichever of the three classes, one may wish to exclude: For any such property G one simply conjoins to the analysis of any particular disposition statement the formula ‘f ≠ G’. It is not up to the philosophical analyst to decide for any particular disposition exactly which properties in particular are to be excluded, but only at most to indicate, as I already have, the kinds of properties to be excluded and the reasons why. In a broader sense of analysis then, this theory of dispositions, at least as, put forth by me, includes the characterization of the excluded classes. In that sense it does, after all, provide a sufficient condition. But in the sense in which it does not, that is in the very nature of any schema which includes a variable with restricted substitution. So one should not worry. The third objection, while less fundamental in so far as it does not concern the nature of analysis itself, is to my mind more interesting. This time my objector argues as follows: “Your theory is simply wrong in that it 6

By ‘of the present kind’ I mean what might be called “same-level” analysis as opposed to “depth” or ontological analysis. Russell’s theory of descriptions is an example of same-level analysis. A sense-data analysis of ordinary objects is an example of a depth analysis.

154

assumes that something cannot merely possess a disposition without there being any lawful connection of the sort indicated or indeed of any other sort. Surely it may be true of some object that it would dissolve if it were put into water without assuming that all objects like it would do so. Or, to put the point in another way, although we may all believe and have good reason to believe that every disposition must have a ground in some nondispositional property or properties, this is not a logical or definitional necessity but only, if it is true at all, itself a general empirical fact which cannot be simply taken for granted in any analysis.” I reply as follows: This objection truly does go to the heart of what a disposition really is. Others have insisted, correctly, that the notion of a disposition does have connotations of causal or lawful connections. But I do not believe that that observation, true and even universally agreed upon though it may be, is a satisfactory answer to this objection. For the question is whether the correct analysis of dispositions involves their grounds in particular, and not just whether there is some lawfulness connotatively involved. In order to pursue this matter, let us make a rigid distinction between the non-dispositional properties of an object and its dispositional properties. Let us assume that among the former are all properties of the sort having to do with chemical composition, molecular structure, and the like, as well as the observable macro-properties of the object such as its color, shape, odor, and so on. We might then imagine two objects—two cubes of sugar, for example—that are alike in all their nondispositional properties. Is it possible that they differ in their dispositional ones? The answer to that question may, obviously, depend on the sense of ‘possible’ being employed. No doubt it would be difficult to show that it is formally contradictory to assert of two objects which are exactly similar in their non-dispositional properties that they nevertheless have some different dispositions, unless one has already assumed something like my account of dispositions. On the other hand, it is difficult to conceive of what it would be for it to be true of something that it would dissolve if it were placed in water when, as we might ordinarily speak, there is nothing about it that makes it true. Indeed, according to the objection at hand, something might be soluble even though in principle there is absolutely nothing one could discover about the object at the time that would, no matter what else one knew,

155

show it to be so! This, I submit, borders on the incoherent. It is more than simply the frame of reference of modern science, but a matter of intelligibility itself, that a disposition of something be connected with some non-dispositional property of the object. This idea is captured in such linguistic expressions as ‘There is something about this object that would make it dissolve if it were put in water’, or ‘This object is such that it would dissolve in water’, and so on. The idea of determinism which requires that a state of a system be specifiable and discoverable would fall into incoherence as well. Or, more cautiously, one might say that a system in which it is true of two objects, exactly similar in their non-dispositional properties, that they may behave differently under exactly similar conditions is not a deterministic one to begin with. Finally, and most important, accepting the force of this objection would once again require us to collapse the distinction between a ‘were-would’ property and a ‘willif’ property; and one would, once again, be faced with the problem of the untested object that lacks the disposition. Consider, finally, the objection of one who argues as follows: “According to you, every singular disposition statement is a subjunctive conditional and involves in its analysis an unmentioned property and a law. But surely this law, like every law, itself entails singular subjunctive conditionals; and these, on your account, must involve another property and another law, which law in turn entails another set of subjunctive conditionals, and so on—an infinite regress of properties and laws. For example, if we arbitrarily call F the property that in some object x grounds its solubility, we get, according to your analysis that Sx “means” Fx . [(y)(Fy . Wy) ⊃ DIy] But from that it follows that if your desk were F and were put into water, it would dissolve. And that disposition—call it F-solubility—needs a ground and another law, and so on. Since this infinite regress of properties and laws is vicious, your account must be rejected.” I reply as follows: Let us say that solubility corresponds to the generalization that (x)(Wx ⊃ DIx)

156

It is because this generalization is not a true law that there arises the problem of the untested, non-soluble object. F-solubility, then, corresponds to the generalization that (x)[(Fx . Wx) ⊃ DIx] which is a true law. Let us call a disposition that corresponds to a true law a universal disposition, since everything has such a disposition, and a disposition that corresponds to a non-law a specific disposition. I now concede that my analysis of dispositions, as so far characterized, strictly speaking applies only to specific dispositions: whenever something has a disposition that some other things do not have, then it has that disposition by virtue of some other property that enters into a law of a certain form. But what about the universal dispositions? What makes my desk Fsoluble? The answer is, and can only be: the fact that it exists. (Nonexistent “things” have no properties at all, not even dispositional properties, for there are no non-existent particulars to exemplify any properties.) Being universal, the disposition requires no differentiating ground. Or, equivalently, one could say that the ground of a universal disposition in a particular thing is any property of it whatsoever, especially since to exist and to exemplify some property or other is, for a particular, one and the same. Hence, by this way of speaking, my desk is F-soluble by being brown or by being one meter tall or. . . and soon. But this does not entail that my desk has an infinite number of properties, for in asking what the ground of each newly generated universal disposition is, one may (as only one formal possibility) simply alternate between any two properties of the thing. For example, one may say that the ground of being F-soluble in my desk is its brownness. Hence it will be a true law that (x)[(Bx . Fx . Wx) ⊃ DIx] while the ground of the new universal disposition (being B/F-soluble) is its being one meter tall, from which it follows that the generalization (x)[(Mx . Bx . Wx) ⊃ DIx]

157

is a true law, while the ground of the new disposition determined by this law is again brownness; and so on. This regress of laws, infinite though it be, is obviously trivial and non-vicious. With that I take myself to have answered the objection.7 Let us now consider some other views on dispositions in comparison with the one here championed. Carnap (1936-7) and following him Hempel (1965) treat dispositions by means of so-called reduction sentences. Both of them admit that they have given up the attempt to analyze dispositions in a way that would permit of their systematic elimination from ordered discourse. Yet their motive, if I am not mistaken, is misplaced. Consider Hempel in particular. The only analysis he even considers that would seem to permit the eliminability of dispositional terms is the singular material implication analysis. But Hempel is immediately overcome by the problem of the object that lacks the disposition but is never put to the test. So, in an effort to avoid the difficulties dependent on the meaning of material implication, he turns to reduction sentences. The theory here advocated avoids the problem of counterfactuals as applied to dispositions in particular, and shifts it to that of laws in general. This, some will say, is no real advance at all, whatever other merits the proposal may have. Be that as it may, what must be remarked is that those who have turned to reduction sentences because of the counterfactuals problem with respect to dispositions have not done something analogous with laws, whatever that would be. This would seem to indicate a certain inconsistency in the practice of those who would analyze dispositional statements by reduction sentences. In point of fact, however, as Wilson (1969) has shown, reduction sentences do not allow one to avoid the counterfactuals problem anyway. So the sole motive for turning to them is misconceived. Wilson’s own analysis is of the singular material implication type. His extended defense of this analysis need not concern us here. But the following comments may be added to what has already been said about this type of analysis. Even if, as Wilson attempts to do, the problem of the 7

If this reply is adequate, it also obviates the ontological need—at least in the analysis of dispositions—some have seen for a connective stronger than material implication. See Lewis, 1973, and Stalnaker, 1968.

158

untested object which lacks the disposition could be solved, there remains the fact that the singular material implication analysis does not make any explicit reference to lawful connections. While it is widely agreed that there is some causal “implication” in the notion of a disposition, the ‘ifthen’, whether of material implication or even something stronger, does not capture this causal feature as long as generality is not brought into the analysis. Causality is captured only in the notion of lawfulness, and whether the ‘if-then’ of material implication suffices in the formal representation of lawfulness or not, generality is in any case a necessary part of that representation. Thus only if there is a generalization or, as on the theory here embraced, the closed schema of a generalization, have we arrived at an adequate analysis of dispositions. Brief mention may be made also of Popper’s “theory” of dispositions. It, too, I believe to be misconceived, even though it may be based on a true premise. Popper, like Hempel and Carnap, seems first to “wish” the singular material implication analysis were satisfactory. But being persuaded that it is not, he, unlike them in their “theory” of reduction sentences, claims that it doesn’t matter. For this he appears to have two reasons. First, the concern with definition and a distinction between observation terms and theoretical, including dispositional, terms is, he believes, misguided to begin with. Or rather, it is not the distinction that is at fault; there simply aren’t any members of the first class: “all terms are theoretical to some degree, though some are more theoretical than others” (Popper, 1965, 119). His second, more concrete reason is that every term is a disposition term and that nothing is achieved by attempting to define one disposition term by other terms that are themselves necessarily also dispositional. The truth in this is, I believe, as follows. While it is true that every property may be treated in a purely formal sense as if it were a dispositional property, and while an absolutely rigid distinction between dispositional properties and observational properties may be one that cannot be sustained, nevertheless in a given context the distinction can be justified. Popper himself grants an important difference between the notion of being soluble on the one hand and the notions of being put into water and dissolving on the other. Relative to any dispositional term, certain others are non-dispositional or observational. I have just indicated

159

such a context. That dissolving in some other context might itself be treated as a disposition term is of no particular significance. Popper’s view derives in part from his contention that there are no primitive or simple properties given in experience, since, he argues, all observation depends in an important sense on one’s beliefs. Obviously, to explore these issues here would take us far afield. I can only say that I believe both of these claims to be false. But even if either or both were true, the point stands that dispositional terms may admit of analysis within a certain context and that, therefore, the theory here adopted is both to the point and possibly correct. A theory which is close in certain respects to the one here defended is that advanced by Hochberg (1967). He brings lawfulness into his theory, but in the following way. First, he tells us that a disposition can properly be ascribed only to something of a certain kind, and that in saying of some particular thing that it is soluble, for example, one is really saying that things of its kind are soluble. Thus ‘x is soluble’ to be analyzed must first be written as ‘x which is sugar is soluble’ or ‘x which is salt is soluble’ or whatever. Then the analysis proceeds as follows: Taking ‘x which is sugar is soluble’ we get: SUx . Sx “means” SUx . (y)[(SUy . Wy) ⊃ DIy] that is, ‘x is sugar and soluble’ is analyzed as ‘x is sugar and anything which is sugar and is put into water dissolves.’ (This is not strictly the way Hochberg formulates the matter, but it gets at the essential features of it without distorting it unduly.) Now the problem with this analysis, as Hochberg himself recognizes, is that of knowing and being able to specify the relevant kind in every case. The virtue of the analysis, in my opinion, is its explicit mention of generality and its consequent shifting of the problem of the singular counterfactual to that of lawfulness itself. Yet clearly one might know that a certain object has a certain disposition without knowing what kind-property of it is relevant, or even, if some properties are not kind-properties, whether it has a lawfully relevant kind-property. Especially in the case of human beings in whom many dispositions are by no means universal in the species, there arises a genuine problem of specifying the relevant kind without mentioning the dispositional property

160

itself. To what relevant kind does Jones who is sleepy belong, for example? Or Smith who is depressed? Or Doe who is patriotic? The theory here proposed avoids this problem altogether while retaining the virtues of Hochberg’s. There is no way to prove, once and for all, that a particular analysis of dispositions is the correct one. Perhaps ‘correct’ isn’t even the right word: ‘adequate’ might be more to the point. I have attempted to show the defects of certain other major accounts of dispositions while at the same time arguing that my own theory does not share those defects. Furthermore, I have attempted to answer what I take to be the most serious objections that might be raised to this account. Beyond this, one can only show that the application of the theory to other, related philosophical issues sheds some illumination on them. In the remainder of this essay, I intend to show that this theory of dispositions does that when applied to some problems concerning the nature of dispositional explanation in general, and to the related issues of rational explanation and reasons and causes. If I am successful, I believe it may be said that strong, additional grounds for this theory of dispositions have been adduced. II One may agree with Hempel (1965) that the following is a perspicuous way to represent the real logical form of a so-called rational explanation: A was in a situation of type C. A was a rational agent. In a situation of type C, any rational agent will do X. Therefore, A did X. and that, accordingly, Dray (1957) is wrong in his view of the matter. I disagree with Hempel that such an explanation or any dispositional explanation is of the covering-law type. But I do believe that to every dispositional explanation there corresponds what may be called a potential explanation of the covering-law type. I shall proceed in this section first by giving my reason for denying that any dispositional explanation is of the covering-law type, then by discussing Hempel’s defense of the view just

161

disputed, and finally by explaining the notion of a potential explanation of the covering-law type. In order to pursue the question whether or not a rational explanation is of the covering-law type, let us forget human beings for the moment and consider instead a brittle window that has just been shattered and the following dispositional explanation of that occurrence: This window was brittle. This window was struck with force. Brittle objects when struck with force break. This window broke. If this is to be an explanation of the covering-law type, then it is, patently, the third premise of it that is the required empirical generalization. The obvious objection to this is that the premise, while having the same grammatical form as an empirical generalization, is not one at all but rather a definitional truth: brittle objects are ones which, by definition, break when struck with force and anyone who knows the language knows this to be true. If the statement were an empirical one, then it is one which, one may presume, experience has so far tended to confirm. And if that is so, then, by the usual canons of empirical meaningfulness, the proposition must also be falsifiable by experience. Yet no experience could falsify this proposition. Even more, if it were an empirical proposition that brittle objects break when struck with force, one would have to suppose that the world might have been such that it would be lawfully impossible for brittle objects to break. (I take for granted that if two properties are only lawfully related to each other, they might have had any such relationship whatsoever.) I conclude that the proposition in question and all others like it in the relevant respect are not empirical at all. And if that is so, then any explanation in which such a proposition occurs essentially is not an explanation of the covering-law type. Hempel, wanting to demonstrate that any kind of explanation which might occur in science really is of the covering-law type,8 replies to this 8

On this point and the proper kind of defense of the covering-law “model,” see the exchange between Hempel, 1976, and Addis, 1976.

162

objection seemingly by attempting to prove directly that, despite appearances to the contrary, propositions of the sort in question are empirical generalizations. I say ‘seemingly’ because Hempel grants that “this objection carries some weight when a dispositional characteristic represents just one kind of law-like behavior, such perhaps as breaking under specified impact” (1965, 462). And while it would seem that, given his general mode of defense of the covering-law “model,” he should have to show also that explanations by “single-track” dispositions are of the covering-law type, in fact he proceeds to argue solely on the basis of “multi-track” dispositions. Thus it would appear that even if this argument were satisfactory, he has not established what he needs to establish and that there still exists good reason to believe that at least some dispositional explanations are not of the covering-law type. But let us look at his argument about “multi-track” dispositions. His own example is that of being magnetic. To this notion correspond a number of “symptom sentences” which express the “necessary and sufficient conditions” for something’s being magnetic. (There being such a number is what makes it a “multi-track” disposition.) For example, ‘If x is magnetic then if iron filings are placed close to it, they will cling to its ends’ expresses a necessary condition while ‘If x is in the area of a compass needle and if one of x’s ends attracts the north pole of the needle and repels the south and the other end shows the opposite behavior, then x is magnetic’ expresses a sufficient condition. But these jointly imply that if x satisfies the compass needle condition it also satisfies the iron filings condition. And this proposition is surely empirical and not definitional. Hence, Hempel concludes, “the total set of symptom statements is more appropriately regarded as part of the system of general laws governing the concept in question” (1965, 462). To this argument it may be immediately replied that, systematically speaking, there are really several concepts of being magnetic to each one of which only one “symptom” corresponds and that all that is empirical is the relations among them. That is to say that the general notion of being magnetic (and so for all “multi-track” dispositions) can be analyzed into a set of lawfully related “single-track” dispositions. Hempel anticipates this objection to his account and attempts to answer it as follows:

163

[T]he general physical statements pertaining to the property of being magnetic include, besides such symptom statements, also certain general laws which represent no dispositional tendencies, and which are no less characteristic of the concept of being magnetic than are the pertinent symptom statements. Among them is the law that a moving magnetic field will produce an electric field, which implies that in a closed wire loop near a moving magnet an electric current will be induced, which in turn implies a general statement concerning the response made by an ammeter which is put into a closed wire loop near a moving magnet. This last statement may be regarded as a further symptom statement for the property of being magnetic, but it should be noted that the symptom here specified is associated with the property of being magnetic by virtue of theoretical principles connecting the given characteristic with other theoretical concepts, such as that of electric and magnetic fields and their interrelations. Thus, when a concept like that of a magnet functions in a theory, then, in applying it to some particular object, we are not simply attributing to this object a set, however extensive, of dispositions to display certain kinds of observable response under given, observable stimulus conditions: the assignment also has various theoretical implications, including the attribution of other “broadly dispositional” characteristics. (1965, 462-463)

It is not easy to see what all of this amounts to with respect to Hempel’s primary aim, which is to show that dispositional explanations “really” are of the covering-law type. For, in the first place, he appears to be telling us that being magnetic, presumably like all other “multi-track” disposition terms, is not a purely dispositional term after all. But if this is so, then the relevance of the argument is in question. It was already so, since Hempel seemed to grant at the outset that explanations that make use of “singletrack” dispositions may plausibly be claimed not to be of the covering-law type. So it would appear that in any case this argument cannot prove, or even make plausible, the claim that dispositional explanations in general are of the covering-law type. But even narrowly construed, the argument does not show what it is intended to show, namely that “multi-track” dispositions cannot be reconstructed systematically as a set of “single-track” dispositions. The argument begins by in effect begging the whole question, that is, by taking the notion of being magnetic as unanalyzed and as a “whole”. To be sure, the scientist may also do so, having established certain empirical connections. Furthermore, we may grant that the non-dispositional

164

property that, as I have argued, must “underlie” every dispositional property may well be the same for each of whatever is expressed by the various “symptoms”. But it doesn’t matter for the question whether or not being magnetic can be analyzed into a set of lawfully related “single-track” dispositions. Would an object that attracts iron filings to its ends but has no effect on a compass be magnetic or not? There is no point in agonizing over the answer, but the fact that the question can be asked shows that magnetism “really” is only a set of lawfully related “single-track” dispositions which in practical scientific discourse are put under the heading of being magnetic. At its crucial point, Hempel’s argument appears to rely on the fact that the ascription of being magnetic to some particular object has “various theoretical implications” and so is not to be construed as “a set, however extensive, of dispositions to display certain kinds of observable response under given, observable stimulus conditions. . .” But is this any more than to say that when a scientist uses the term being magnetic, it connotes to him various empirical and theoretical principles and implications due to his knowledge of various lawful connections? I cannot see that it does say any more than that. If it does not, then the argument is once more beside the point, for the question is what the truth-conditions are for something’s being magnetic and not what goes on in a scientist’s head when he uses the term ‘magnetic’. To be sure, in analyzing any notion we must begin with what people mean in a general sort of way, but not with all the many connotations of a term that occur to this or that person. If we found that the lawful connections between, say, attracting iron filings and affecting a compass in a certain way hold only under special conditions and that under other conditions an object may exhibit one but not the other behavior, we would have to decide how to use the term ‘magnetic’. There is no genuine issue there, much as some, Hempel not included, appear to believe that there are real meanings of words. But it shows once again that however words may be used, whatever “holistic” connotations they may have given the knowledge of certain lawful connections among certain properties, all there is in the world, for the present purpose, is a set of externally connected, “single-track” dispositions. Words as meaningful “objects” may be open-ended, “holistic”, and the like, but the world itself is not. Hempel then has failed to show that dispositional explanations in

165

general are of the covering-law type. Yet, if I am not mistaken, although dispositional explanations are not of the covering-law type, there does “correspond” to every such explanation what I shall call a potential explanation of the covering-law type. It is time now to explain this idea. Consider once again the dispositional explanation of the window’s breaking: This window was brittle. This window was struck with force. Brittle windows when struck with force break. This window broke. This explanation, I have just argued, is not one of the covering-law type, for it does not contain an empirical generalization despite grammatical appearances to the contrary in the third premise. But, given the theory of dispositions here advocated, the first premise, although not itself an empirical generalization, contains “reference” to a law of nature. For ‘this window is brittle’ really “says” that this window has some (nondispositional) property and (it is a law that) anything with that property that is struck by force breaks. Taking out the (existential) generalization from this analysis of the first premise of the original explanation, we can construct the following potential explanation of the covering-law type: There is some (non-dispositional) property such that anything with that property that is struck with force breaks. This window has that property. This window was struck with force. This window broke. Since this explanation is, by the principles I have indicated, mechanically constructible from its corresponding dispositional explanation, it is apparent that such a corresponding explanation of the covering-law type is constructible for every dispositional explanation. It should be apparent, therefore, that an explanation of the covering-law type, and therefore a causal explanation, is always compatible with a dispositional explanation of the same phenomenon. So let us now, finally, apply these considerations

166

to the issues of rational explanation and of reasons and causes for, if I am not mistaken, we have the materials to say something useful about them. III Can reasons be causes? I begin by making two fairly obvious distinctions. First, it is necessary to distinguish a reason considered as something abstract and not connected with anyone in particular from a reason or, better, the having of a reason, that is connected with some particular person. In the former sense a reason may be said logically to imply another reason or to contradict it and so on. Reasons as propositions can stand in purely logical relations to one another. But the having of a reason, as an occurrence or disposition in the life of a person, never logically implies the having of another reason, either in the same person or in any other. It is the failure to make this distinction or at least to maintain it systematically that gives much of the undeserved plausibility to Winch’s thesis, in his The Idea of a Social Science (1958), that social inquiry is essentially an a priori enterprise that blends into philosophy itself since, on his account, its subject-matter is (ultimately) ideas and ideas stand only in logical and never causal relations to each other. Be that as it may, let this stand as our first distinction. But what is it to have a reason? This question introduces the second distinction. To have a reason is either (1) to have a certain kind of disposition or (2) to have a certain kind of conscious mental state. In the situations represented by ‘Jones waved the flag as the parade went by because he is patriotic’ and by ‘The reason the general withdrew his troops under the circumstances is that he is rational’, the reasons being cited— being patriotic and being rational--are dispositions. But in ‘Jones waved the flag as the parade went by in order to signal a friend’ and ‘The reason the general withdrew his troops was that he suddenly realized the gravity of his situation’, the reasons referred to—an intention to signal and a sudden realization—are occurrent or conscious mental states. (Since certain dispositions are sometimes called dispositional mental states, I shall, by way of contrast, write of conscious mental states.) Consider first, briefly, conscious mental states. They are, on my view, occurrences among occurrences for most purposes. While they are not

167

physical occurrences, each of them is connected parallelistically with a lawfully relevant physical occurrence--some state of the brain or central nervous system. So conceived, there is no special reason why one should not say that reasons as conscious mental states are among the causes of behavior. (To be sure, if parallelism is true, as I believe it to be, anything that can be explained with conscious mental states can also be explained without them.9 Yet, given the parallelistic laws, if certain conscious mental states did not occur, then certain brain states would not occur, and so certain behaviors would not occur.) Some find it odd to say that, for example, Jones’s intention to signal a friend was the cause or a cause of his signaling his friend since the intention seems to be described in terms of the action (or vice versa). It would seem that if the behavior is described instead as waving a flag, then that problem disappears. If, however, one wants to say that actions as contrasted with mere behaviors cannot be the effects of their intentions, at least insofar as they are described in the same terms as their intentions, I have no objection. But I find this trivial and, contrary to Winch (1958), Maclntyre (1967), and many others, no evidence whatsoever against the idea of the Humean universe.10 The more interesting question is whether reasons as dispositions can be among the causes of behavior. Let us construct the potential explanation of the covering-law type that corresponds to Hempel’s schema of a rational explanation in order to help ourselves see what is implied by it and the theory of dispositions underlying it. It will be so: There is some (non-dispositional) property such that everyone with that property who is in a situation of type C will do X. A had that property. A was in a situation of type C. A did X. What the idea of this schema implies about human beings is as follows: First, if two persons who were qualitatively identical, and therefore (on this account) also dispositionally identical, were in exactly the same situation, 9

In Addis, 1965, this claim is clarified and defended in detail. Further discussion of this point can be found in the third chapter of Addis, 1975, and especially in Addis, 1998. 10

168

then the resulting behaviors would be the same.11 The converse is, of course, not true—that two people whose behaviors are the same in some situation are necessarily qualitatively the same. In fact, that is certainly never the case. Secondly, a person no more than an object can have a disposition to do something unless there is some non-dispositional ground for it. However, since the “same” disposition may have more than one ground, two persons who were dispositionally identical would not necessarily be, even in the relevant respects, qualitatively identical. Just as solubility may be grounded in a number of different molecular structures, so rationality or patriotism may be grounded in a number of different states of the brain and central nervous system. Is solubility one of the causes of dissolving? It is not, because the proposition that whatever is soluble dissolves when put in water is not a law but a definitional truth. This I argued earlier. Let us generalize. No disposition as a disposition is among the causes of that to which it is a disposition, but ascribing one to an object or person implies that that to which it is a disposition has a cause, some molecular structure or state of the central nervous system or whatever, in the object or person in question. Having generalized, we may now instantiate. So far as (the having of) a reason is the having of a certain disposition to behave in certain ways, (the having of) that reason is not a cause of any of those behaviors. But ascribing the disposition to a person implies that a given behavior has among its causes some non-dispositional property or state of that person. A reason as a disposition is not among the causes of the behavior for which it is a reason. In this sense and, I believe, in this sense only is it true that reasons, as actually had by someone, are not causes. But the fact that dispositions and therefore reasons are not causes of that to which they are dispositions and for which they are reasons does not imply that dispositions and therefore also reasons cannot be causes at all. For while the name of a disposition cannot significantly appear in the antecedent of those laws that are relevant to the explanation of that or those occurrence(s) to which it is a disposition, it can significantly appear in the antecedent of some law which is relevant to the explanation of some kind of behavior or property to (the name of) which it is definitionally 11

For further discussion of this kind of possibility, see Addis, 1982.

169

unrelated. Solubility is not a cause of dissolving since ‘Whatever is soluble and placed in water dissolves’ is a definitional and not an empirical truth. But ‘Whatever is soluble and rubbed with wax explodes’ is, though false, a non-trivial, empirical proposition and would if true mean that solubility is among the lawfully possible causes of explosions. Similarly, in the case of human beings, a person possessed of a disposition of a certain sort will have other properties to which the (name of the) disposition is definitionally unrelated. Dispositional explanations then are not causal explanations since they are not of the covering-law type. And in physical science we usually regard them as somehow unsatisfactory. ‘This dissolved because it is soluble’ is both true and at least in some measure an explanation. Yet we want to know more. But in the explanation of human behavior we don’t always seem to be faced with the same obvious dissatisfaction. ‘Jones waved the flag because he is patriotic’ seems to be satisfactory in a way in which the other is not. Why is that? There are, I believe, three reasons. First, all dispositions relevant to the explanation of voluntary human behavior are “multi-track”. Thus in explaining Jones’s flag-waving by ascribing to him the disposition of being patriotic, one is referring to something that is much more than the inclination to act in this particular way in this particular kind of situation. We are locating the behavior in a multifaceted pattern of actual and potential behaviors, and not simply, as in the case of the sugar that dissolves in water, identifying it as one of a number of identical behaviors. Second, qualitatively identical human behaviors may issue from altogether quite different dispositions. Jones may wave the flag because he is patriotic, but he may also wave it for exercise. Third, there is also the fact, of which in some views of the human sciences altogether too much is made, that the explainers of human behavior are themselves humans. But it does account for the circumstance that we are simply more interested ordinarily in knowing from what disposition a person acted (or as a consequence of what conscious mental state) than in what state of that person’s brain or other physiological state brought on the behavior to be explained. This is further strengthened by the fact that we can, I believe, in some measure introspect our own dispositions.

170

There is no genuine philosophical question concerning what the “real” kind of explanation of human behavior is—dispositional or physiological or even environmental. There is a sense, as I take myself already in effect to have shown, in which physiological explanations are more deeply rooted ontologically than dispositional ones. Yet explanations in history and the human sciences will more often be of the latter than of the former kind. And we have just seen why such explanations, unsatisfying though they often be in physical science, have a central role in understanding human behavior.

171

BEHAVIORISM AND THE PHILOSOPHY OF THE ACT1

A

n adequate philosophy of mind must satisfy a number of criteria. The two most important among them are, first, that it take account of and be entirely consistent with the phenomenological data; and, second, that it accord with both the presuppositions and the findings of scientific psychology. Many philosophers as well as many philosophically-minded psychologists have believed that it is impossible at once to satisfy both of these criteria—at least as their requirements are widely interpreted. Consider two notable examples. The psychologist John Watson believed that the requirements of a scientific psychology imply the denial of the efficacy and possibly the very existence of private mental states. Denying the existence or the efficacy of private mental states is inconsistent with the phenomenological data. The philosopher Jean-Paul Sartre believed that an adequate phenomenology of the mind implies the impossibility of the causal explanation of human behavior. Asserting the impossibility of the causal explanation of human behavior is not in accord with the presuppositions of scientific psychology. Among the relatively few philosophers and psychologists who were moved by the impulse to account fully for the phenomenological data were a yet smaller number who defended the act. The most elaborate analysis and defense of the act among twentieth-century analytic philosophers is to be found in the writings of Gustav Bergmann.2 Yet Bergmann is also known as one who, as a logical positivist, helped to develop and defend the philosophy of psychology known as behaviorism.3 Behaviorism and the philosophy of the act: the dearth of those who would defend both is almost intimidating to one who, like myself, agrees with Bergmann that both must 1

An earlier version of this paper was presented at one of two special symposia on the contributions of Gustav Bergmann to the philosophy of psychology at the meetings of the American Psychological Association in Washington, D. C. in September 1976. 2 See especially Bergmann, 1954a, 1955, and 1960. 3 See especially Bergmann, 1940a, 1940b, 1940c, 1941, 1942, 1943, 1945, 1951a, 1952, 1953, 1956, and 1957.

be defended. Indeed, I believe that behaviorism and the philosophy of the act alone jointly satisfy the requirements of a scientific psychology and of respect for the phenomenological data, respectively. The historical tension that has existed between scientific psychology in general and behaviorism in particular, on the one hand, and dualistic philosophy of mind in general and the philosophy of the act in particular, on the other, is without doubt based on the belief that there is a logical tension between the two. And as the ideas of each are often presented, it is not surprising that there should appear to be such a tension. Yet more careful analysis shows, I believe, that there is no incompatibility whatsoever between behaviorism and the philosophy of the act. In the remarks that follow it will be my primary intention to show that that is so. I shall proceed as follows: First, I shall discuss behaviorism, distinguishing what is from what is not important in it from a philosophical point of view. Second, I shall show what behaviorism presupposes about the relation of mind to body. Third, I shall do some very selective history of the philosophy of the act, trying to show how it came to be seen as incompatible with the notion of a scientific psychology. Forth, I shall sketch and defend the philosophy of the act, with heavy emphasis on Bergmann’s particular version of it. And fifth and finally, I shall show precisely how the philosophy of the act complements behaviorism. I Behaviorism is a methodological approach to the study of human behavior that assumes that, in principle, every behavior has a full, lawful explanation that does not require mention of the realm of the mental. Or, to put it somewhat more technically: a complete set of relevant, independent variables for the lawful explanation of behavior exists among the physical variables alone.4 Like any other occurrence, a behavior is given a full, lawful explanation if and only if initial conditions and laws are cited such 4

For the main ideas of this essay, the crucial idea in the fundamental assumption is the superfluity of the mental in the explanation of behavior and not the logically separable presupposition that every behavior has a full, lawful explanation. Yet for reasons both of history and of clarity of explanation, it is desirable to proceed under the deterministic presupposition.

174

that whenever and wherever precisely those conditions obtain, by those laws that precise behavior does or would occur. It is important to understand for a point that will be stressed in the next section that, like any other occurrence, one and the same piece of behavior may have more than one full, lawful explanation. That will be so in one sense if certain different initial conditions from those that actually did obtain would also have brought about the behavior to be explained. But the sense I wish to stress here is somewhat different: that from among the conditions that actually did obtain, more than one subset is such that by the laws of nature whenever and wherever it occurs, the behavior of the kind to be explained occurs. If, for example, a, b, c and d are four conditions such that condition c occurs if and only if condition d occurs, and if a, b, and c constitute a set of conditions that fully explain the occurrence of some given behavior, then a, b, and d also constitute such a set. Either of these senses may be expressed by saying that there may be more than one (minimum) complete set of relevant variables. It is evident then that what I shall call the fundamental assumption of behaviorism—that the mental variables need not be taken into account in the full, lawful explanation of behavior—does not imply that the mental variables cannot be invoked in the explanation of behavior. It would also appear to be evident—the reason for my caution will become clear momentarily--that this fundamental assumption is a very general, empirical hypothesis; for what is lawfully connected with what, and in what manner, is always a matter of fact and never one of meaning or of logic or of values. That it is an empirical matter is reinforced by the observation that, in the opinion of many, it is a hypothesis that is patently at odds with our everyday experience of the way in which our minds and bodies affect each other. But it will be granted, even by the most hardline of positivists, that even if it is an empirical hypothesis ultimately, it is one that, since it does involve presuppositions about the nature of mind and its relation to body, naturally recommends itself to philosophical reflection and clarification. But is the fundamental assumption really an empirical hypothesis after all? One might be led to believe from the writings of the logical positivists including Bergmann that behaviorism can be defended philosophically not merely as a consistent and possibly true hypothesis about how certain variables do and do not lawfully connect with one another, but, much more

175

radically, as the only “hypothesis” that is even meaningful, all others being ruled out as being non-empirical. This is what, very roughly, was meant by the idea of logical behaviorism. Still roughly, the theory, as expounded in Ayer’s Language, Truth, and Logic and elsewhere, is that the very notion of an unobservable variable that interacts with any of the set of observable variables is unintelligible, whether that unobservable be (the property of) an atom, or a god or a mind. Obviously, this very crude account disregards a world of differences. Put a little less crudely and with the essentials of the argument, the reasoning was as follows: In order for any descriptive predicate notion to have empirical content (and thus to be eligible for appearance in scientific contexts), it must either refer directly to a property that is observable or else be wholly (or in some versions only partially) definable by predicates for such properties. Since the mental variables, considered as properties of private mental states, are not (publicly) observable, their corresponding predicates either are not to be admitted to systematic scientific discourse at all or else must be “defined” by predicates that refer only to publicly observable—that is, physical and especially behavioral—properties. So it was made to appear that the fundamental assumption of behaviorism is guaranteed by philosophy by arguing that any alternative to it is meaningless. Logical behaviorism, is, I believe, if not simply false, at the least misleading. Furthermore, I hesitate to ascribe logical behaviorism without qualification to all the logical positivists and especially to Bergmann (who in any case was inclined to apostasy among the logical positivists very early). The main reason for my hesitation is as follows: unlike many of the more stringent positivists (such as Feigl and Carnap), Bergmann recognized, correctly, that the sense in which the (names of the) mental variables can, and even possibly must for the purposes of systematic scientific discourse, be “defined” by the (names of the) behavioral or other physical variables is not definition in the literal sense. (That is why here and in the next to final sentence of the last paragraph I put ‘defined’ in double quotes.) In the literal sense, a definition shows that the notion to be defined is an abbreviation--an expression that stands for or refers to strictly the same property or state of affairs as the expression for which it is the abbreviation. Thus, because the expression ‘being an unmarried human male’ defines the expression ‘being a bachelor’ in the literal sense, the

176

latter therefore being an abbreviation of the former, the property of being an unmarried human male and the property of being a bachelor are strictly one and the same property. If the same conception of definition is applied to the case of the mental and the physical, mental predicates being therefore conceived as abbreviations for physical predicates, one gets the result that the mental variables are strictly the same as some physical variables. That—the thesis that the physical variables alone exhaust the universe--is materialism. Although, to judge by his earlier writings alone, the evidence is not clear, I think Bergmann always believed, as he came later more emphatically to assert, that materialism is false. Thus my reservation. In any event, materialism is false. It is false on phenomenological grounds (which, by the way, don’t apply to atoms and gods), and the mental variables are a set distinct from, and not merely a subset of, the physical variables. Since this is so, and since each of us is acquainted with some members of the set of mental variables, logical behaviorism must be rejected as false or misleading. I now proceed to refine further my characterization of behaviorism by discussing four ideas of what behaviorism is not or, better, is not, for analytical purposes, usefully thought of as being. First, behaviorism is not the thesis, nor does it presuppose the thesis, that lawful explanation is the only possible or even, apart from any context, the only “appropriate” kind of explanation of behavior. I have never understood the impulse to argue, from whatever quarter, that there is but one “real” kind of explanation to be given or understanding to be had of human behavior. That lawful explanation has a certain kind of superiority to all other kinds I would insist upon, if it is in need of any defense, in that it is that kind of understanding of the phenomena that enables prediction and control (about which more momentarily). The essential logical feature of lawful explanation that gives it this characteristic is, of course, that it is, in principle, deductive in nature. Since that is so, a particular explanation of the lawful type tells us, as we expect, why what it explains occurred rather than did not occur. All the defender of behaviorism need argue is that the lawful understanding of behavior is possible and that nothing in philosophy or in our experience of ourselves shows that such understanding cannot be had through the physical variables alone.

177

Second, behaviorism is not the thesis that the prediction and control of behavior are the goal of scientific psychology. That proposition, which one associates primarily with B. F. Skinner, might be taken in any of three senses: (1) as a statement about the nature of science itself, or more precisely, one that follows from such a statement; or (2) as a statement about the possibilities of a science of human behavior; or (3) as a statement about our responsibilities deriving from our present or future knowledge of the causes of human behavior. Nature, possibilities, responsibilities: the words suggest that a normative or modal statement is disguised as a categorical, factual one. Yet I can think of no other reasonable, precise meaning to attach to the statement. Be that as it may, and whatever the truth value or reasonableness of each of these three statements, I wish to dissociate each of them from what I mean by behaviorism in the following way. I characterized behaviorism as the thesis that only physical variables are needed in the lawful explanation of human behavior. Scientific prediction and therefore, where possible, effective control of any kind of phenomena, are logically more superficial than, because they are logically dependent upon, scientific explanation, at least when one is relying upon knowledge of the causes of the phenomenon to be predicted or controlled. Within certain limits, as we know, reliable prediction is possible without knowledge of causes just because the phenomenon is regular in its occurrences. The succession of the seasons is an obvious example. Control, however, does ordinarily require knowledge, however imperfect, of causes. That I take to be obvious. The essential point, however, is that, historically and logically, the “goal” of science has been, and remains, lawful explanation. More accurately, it is the search for laws and theories that permit the explanation and sometimes the prediction of particular occurrences and contain the explanation(s) of kinds of phenomena. The asymmetry between what is “permitted” and what is “contained” shows the logically more fundamental nature of explanation as compared to prediction. (We don’t predict kinds of phenomena.) Prediction as such is not the “goal” of any science and not therefore of scientific psychology. Third, behaviorism is not the thesis that behavior is the proper subject matter of psychology. I do not wish to disagree with this essentially normative proposition and for anyone who has a role in decisions concerning the use of resources in academic and professional psychology,

178

its importance is obvious. Obvious, too, is the motive force in treating this essentially normative proposition as one of fact, or of meaning or of “philosophy.” Insofar as it derives from or even embodies the nonnormative proposition that mental phenomena cannot be scientifically studied in just the way that pre-behavioristic psychologists supposed, it is not merely a value judgment. But whatever its precise status is or is even decided to be within reasonable limits, it is usefully separated from propositions about what is possible in the lawful explanation of behavior. Fourth, behaviorism is not the thesis that a full set of relevant, independent variables in the lawful explanation of behavior exists among the so-called environmental variables alone. Probably no advocate of behaviorism ever held this thesis in such a stark form whatever its critics may think and, alas, whatever reason some of its advocates may have given them for so thinking. Which among the physical variables are essential in the full, lawful explanation of behavior is just science in a way in which the exclusion of the mental variables is not just science but necessarily involves certain philosophical considerations. Whether, for example, certain near-universal characteristics of the human brain (or the dispositions or capabilities of which those characteristics are the “ground”) are essential, as they almost certainly are, in the full, lawful explanation of language acquisition and employment is of no obvious philosophical significance, Chomsky and some of his critics to the contrary notwithstanding. Nor does the extent to which the full, lawful explanation of differences in human capacities and behavior requires mention of genetically-determined physiological differences among us in itself raise any interesting philosophical questions. Once again, that is or ought to be science and only science. But the fact, and indubitable fact I presume everyone would take it to be, that no full, lawful explanation of behavior is to be found among the environmental variables alone gave rise to an issue among philosophers of which I must take some extended note before I turn directly to the mind/body issue. Given that there are bound to be “gaps” in explanations of behavior that countenance only environmental variables, the question arises what should be assumed to fill those “gaps”. Ignoring entirely the possibility that nothing fills them, that is, the possibility that determinism is false, there are, within the context of our discussion, three possibilities: (1) mental

179

states, (2) physiological states, and (3) behavioral dispositions. Behaviorism rules out the first or, more precisely, implies that the second or the third will suffice. Whether it should be the second or the third will be recognized as the by now somewhat hoary issue of hypothetical constructs (the physiological) vs. intervening variables (behavioral dispositions). As to this matter I have three comments. First, when the classical positivists spoke of “defining” the mental variables through certain physical ones, what they should have meant is not, as we have already seen, that a mental predicate, given its already existing sense, can be regarded as literally an abbreviation for some physical predicate, but rather that one can, for the purposes of science, substitute for any mental notion a behavioral-dispositional one. To feel thirsty “is”, by such a substitution, to be such that, under certain conditions, one drinks. But should the behavioral-dispositional notion in turn be replaced by a physiological one even while always keeping firmly in mind that a mental state, a behavioral-dispositional state, and a physiological state are three, and not two or one? That brings me to my next comment. The answer to the question I just asked may depend on the answer to the philosophical question as to the nature of dispositions in general. I believe that it does, especially since, in my view, dispositions are, in a special but important sense, nothing.5 Indeed they are, to speak most succinctly, materially though not formally, identified with their “grounds.” Be that as it may, it is enough to remark here that the discussion of the old issue I mentioned was seriously impaired by a failure to bring to it an adequate theory of dispositions. Another confusion is the point of my final and longest comment. The hesitancy many philosophers and psychologists had about filling the “gap” with the physiological was due to the “unobservability” of the physiological. We cannot, ordinarily, look inside living human bodies, and it is the behavior of conscious, alert, intelligent human beings that, one assumes, the psychologist is most interested in understanding. Yet, research strategies are one thing, matters of principle and especially of philosophical principle quite another, even if the latter do, as they ought, to 5

See Addis, 1981a, for details.

180

some extent shape the former. Let me explain. The notion of unobservability is really at least four notions. Or, perhaps better said, the various kinds of unobservability derive from such fundamentally different circumstances that its core meaning is insufficient to its useful application. When confused, the four notions lead to bad arguments. The four “meanings” are: (1) something’s being unobservable because it is covered up, too far away, too small. Examples are: my pineal gland, the most distant star in the Andromeda galaxy, a virus; (2) something’s being unobservable because of its nature as a disposition or other kind of defined property. Examples are: being soluble, being a trade union; (3) something’s being unobservable because of its privacy as a mental phenomenon. Examples are: being a sudden remembering, being a sharp pain; and (4) something’s being unobservable because of “its” being utterly non-existent. Examples are: unicorns, mermaids, and infallible philosophers. Although behavioral dispositions are unobservable in the second sense, this was felt by many philosophers not to be a difficulty of principle because such notions are, one assumes, definable by way of observable properties: solubili1ty, for example, by the observable properties of being put into water and dissolving. Being unobservable in the sense in which the mental is unobservable is really, it is worth stressing, only unobservabilityto-others: not being publicly observable and not being observable at all are two quite different matters. This point is worthy of emphasis because for an empiricist all property notions that are to be intelligible and therefore eligible for scientific discourse must be either observable or (literally) definable by way of the observable. The mental is not publicly observable. Yet each of us has observed a remembering. How we know we are all observing (instances of) the same property doesn’t concern me for the present. Somehow we do know it. But it does make it at least intelligible (1) to insist that an empiricist may hold that mental properties are a distinct set of properties from the physical, and (2) to imagine that those properties could be interacting ones with the physical, that is, that the fundamental assumption of behaviorism could be false. Once again, it is, therefore, an empirical hypothesis, and logical behaviorism is false or misleading. The unobservability of the non-existent needs no further comment here, even though, in an obvious sense, it is of the most radical sort. But the

181

unobservability of the physiological is, philosophically speaking, the most superficial of all. Whatever unobservability it has is, in part, a function of the lack of an adequate technology for its observation combined with, of course, moral and legal prohibitions on the unrestrained observation of living human beings in any sense. Furthermore, and partly for the same reasons, especially keeping in mind the “technology of behavior” some psychologists wish to develop and apply as the result of psychological research, the environmental variables are often easier to modify in order to shape or modify behavior. Finally there is a tendency to polarization— either the environmental or the physiological. The result was an entirely misplaced distrust of the physiological from Watson on among the behavioristic psychologists with a left-handed boost from the logical behaviorists. Yet the shallowness of filling the “gaps” by behavioral dispositions was evident to many despite the heroic efforts of some to show (what is true, in some cases) that the behavioral disposition could be defined and even identified logically independently of and, in particular instances, temporally prior to the behavior for which it was to be the partial explanation. And in some cases, as Skinner has argued, the role could just as well be taken by the earlier environment of the person whose behavior was to be explained. So, for example, rather than invoke the behavioral disposition being thirsty, one could point instead to the fact of water deprivation for several hours. Neither, however, as we already know from common-sense and science, will serve systematically. As a research strategy, for the reasons mentioned above, one may reasonably sometimes ignore the physiological.6 But as a matter of principle based on what we already know of the empirical facts and the logic of dispositions, it is reasonable to conclude that only the physiological will suffice systematically to fill the “gaps” in the full, lawful explanation of human behavior. I shall assume so in what follows. II It will be useful in the sequel if I generalize the fundamental assumption of 6

The laws that actually are discovered may be only what I call “polychronal” laws. A polychronal law is one in which (kinds of) events from two or more different times are mentioned in the antecedent. For an analysis of such laws, see Addis, 1975, 90-92.

182

behaviorism as follows, although, of course, it is not strictly implied by it: that every occurrence in the physical world, including therefore all human behavior, has a full, lawful explanation solely by other occurrences in the physical world. That means that for every event in the physical world there is what some call a “deterministic” explanation: laws and a set of purely physical prior initial conditions from the statement of which laws and initial conditions a statement of the event to be explained can be deduced and which initial conditions are such that were they to occur again, the result would be precisely the same. Let us, following Bergmann, express with succinctness this idea of the non-necessity of invoking any nonphysical occurrences in the explanation of physical phenomena by saying that the physical world is “causally closed”.7 Commonsense holds, and phenomenology confirms, that (1) minds exist in the sense that there is a set of non-physical variables (that materialism is false); and (2) minds are efficacious in the sense that what we desire, believe, remember, imagine and so on make a difference to our behavior (that fatalism is false). Both minds and bodies are there; neither is, in the ontological sense, reducible to the other. And they do affect each other: if I didn’t feel thirsty, I wouldn’t drink water now; if I hadn’t just been pricked by a pin, I wouldn’t feel pain now. Some behaviorists, we know, believed that, in order to be consistent with the fundamental assumption, they had to deny the efficacy of mind. The most radical way to do that is to deny its existence. And some philosophers continue to support this barely plausible but fundamentally mistaken line of reasoning by adopting as the criterion of existence whether or not the candidate is “needed” for scientific explanation. Are the phenomenological data (and therefore commonsense) consistent with the idea that the physical world is causally closed? Does holding the fundamental assumption commit one in logic to materialism or fatalism? Let us see. First, as to the mere existence of a distinct set of mental variables, there is, fairly obviously, no problem unless we adopt, as we should not and shall not, the scientistic criterion of existence. For the fact that a given set 7

Obviously a world can be causally closed without being deterministic. But once again it is easier to introduce the idea under the deterministic presupposition.

183

of variables constitutes a causally closed set implies nothing whatsoever about the existence or non-existence of any other variables. If we assume then, as is assuredly the case, the there is a distinct set of mental variables, there are three abstract possibilities as to their lawful connection with the physical variables. Those possibilities are captured roughly in the notions of (1) no lawful connection at all, (2) connection by laws of succession, and (3) connection by laws of coexistence. Let us consider these possibilities in appropriate detail by imagining the somewhat repugnant situation of two human beings who are physiologically identical, two persons whose brain states, neurological states, and all other bodily states are qualitatively indistinguishable. Assume that they are conscious. Are their mental states of lawful necessity also qualitatively indistinguishable? If, for example, the one is thinking that 2 + 2 = 4, is the other also of lawful necessity thinking that 2 + 2 = 4? The answer is either yes or no. If the answer is no, two of our three abstract possibilities remain. One is that the mental variables have no lawful connection with the physical ones at all. That is fatalism and fatalism is inconsistent with the phenomenological data. The other is that by certain laws of succession the mental variables interact with the physiological ones much as, in mechanics, mass, position, and velocity interact with each other. But then, just as no two of the latter constitute a causally closed set, so too, if this possibility is correct, the physical world is not causally closed. Interactionism is precisely this thesis—that the values of the mental variables make a difference to those of the physical variables in a way that makes the physical world not causally closed. If, however, the answer to our question is yes, then for every mental state, there is a simultaneous physiological state to which it is related by a law of coexistence such that whenever that particular physiological state occurs, that specific mental state occurs. That, precisely, is the thesis of psychophysiological parellelism. It does not imply, of course, that for every physiological or more radically, as Spinoza believed, for every physical state, there is a lawfully corresponding and simultaneous mental state. But that accords with commonsense and science too, insofar as they hold that bodies can exist without minds but minds cannot exist without bodies. This kind of “materialism”—the primacy of matter—one denies only at the risk of absurdity.

184

Thus, of the four “isms”--materialism, fatalism, interactionism, and parallelism—only interactionism implies that the physical world is not causally closed. (Some may wonder whether my categories are truly exclusive. Does not the materialist in denying the existence of mind also deny its efficacy thus making him a fatalist as well? The answer, of course, is negative insofar as the materialist regards the mind not as non-existing but as a subset of the physical variables, whether as, in the case of Ryle, behavioral and behavioral-dispositional ones or as, in the case of the socalled “identity” theorists, certain states of the brain or central nervous system. Thus mind, while non-existing in a certain sense in nonetheless efficacious. So there is no inconsistency.) For the behaviorist that leaves materialism, fatalism, and parallelism as consistent with the fundamental assumption. Materialism and fatalism are, on phenomenological grounds, absurd, the first in denying the existence of the mental variables, the other in denying their efficacy. Parellelism is thus the only reasonable position for the behaviorist to hold with respect to the connection of mind and body. Parallelism implies the fundamental assumption of behaviorism. For whenever one might wish to explain the occurrence of some behavior by the prior or simultaneous occurrence of a certain mental state (and, contrary to Skinner, such explanations are sometimes accurate as well as appropriate), one can, as it were, “substitute” the lawfully corresponding physiological state (or behaviorial disposition or earlier environment) in the explanation. But—and this is the crucial point with respect to the phenomenological data—it does not follow from parallelism that the values of the mental variables make no difference to the values of the physiological and therefore of the behavioral variables. Quite the contrary, parallelism entails that some differences in the values of the mental variables will, of lawful necessity, be paralleled by differences in the values of the physiological variables. If I did not have the desire I do now have, my brain state would be different from what it is, and so my behavior different from what it is or will be, and so on. So it is not paradoxical as it first appears that parallelism is consistent with the propositions both that the physical variables constitute a causally closed set and that the values of the mental variables make a difference to those of the physical variables. Yet some may wonder whether or not this is not just a modern version of epiphenomenalism according to which the mental variables simply ride

185

“piggy-back” on certain physical ones while themselves having no real causal efficacy? And if they have no real causal efficacy, is this not really fatalism after all? My answer to the first of these questions is yes and no, and to the second, therefore, no, while observing at the outset that the labels don’t really matter but only accurate description. There are several reasons why some may wish to say that on parallelism, at any rate, brain states are the cause of mental states and not the other way around. I shall mention only the two most important. The first is that brains and some of their states can exist with no lawfully corresponding mental states while no mental state can exist without a lawfully corresponding brain (or some physical) state. This is again but the primacy of matter and embodies, in an obvious sense, a contextual notion of what makes a certain lawful connection a causal one. This, in my opinion, is as it should be. Yet, our universe being what it is, if parallelism is true, it is lawfully impossible for certain brain states to occur unless certain mental states occur. Could the universe have been such that those brain states could occur without those or any mental states at all? Yes, but it isn’t that way. The second reason lies in the fact that the parallelistic connection is almost certainly many/one, not one/one, from body to mind. If that is so, then, in an obvious sense, mind is “deducible” from (not: reducible to!) body, but not the other way around. More precisely, knowing all the laws permits the calculation of specific mental states from specific bodily states, but from specific mental states only to ranges of bodily states. Parallelism is true. Indeed, properly clarified, it is commonsensical. By that I mean that it is the only hypothesis that is consistent with our experiences of ourselves and our presumed knowledge of others. In its dualism, it recognizes the existence of a distinct set of mental variables. But unlike the other dualisms, interactionism and fatalism, it is what is presupposed by the practice of medicine in general, and anesthesiology and attempts to relieve pain in particular, by the practice of lie detection, by all attempts to locate precise neural substrata for each of the various mental activities, by our belief that we often know what another person is thinking, and, fundamentally, by the constant experience each of us has of the relation of his or her own body and mind.8 And since parallelism is 8

See Addis, 1965, 44-56, for a more detailed defense of parallelism.

186

true, behaviorism is justified in its fundamental assumption. I turn to the history of the act. III Speaking commonsensically, there are mental acts of remembering, imagining, doubting, perceiving, and so on. Yet many philosophers and psychologists have denied that there are acts. Obviously, they were not denying that people sometimes remember, imagine, doubt, perceive, and so on. What, then were they denying? As is usually the case in philosophy, some understanding of the history of the issue is essential to grasping the issue itself. For the issue at hand there is, however, an unusual complication. On the one hand, the notion of the mind as “active” in cognition is, unquestionably, the historical source of the notion of the act itself. Denying that there are acts in this sense is denying that the mind is “active” in cognition (or in any other of its states). This notion of “activity” is, of course, a metaphysical one and is tied historically and structurally to a substance ontology. On the other hand, the question of the existence of acts has been raised, particularly in the last century and a half, entirely outside a substance ontology. In this case, I submit, the fundamental question was: Are there any simple entities that are intrinsically intentional, that is, that by their very “nature” are about something else? If so, there are acts; if not, there are none.9 Historically, the two notions were not always distinguished nor, what amounts to almost the same, conceived and characterized clearly. So there were many bad arguments on both sides of the issue. Eventually I shall, as it were, “side” with the notion of an intrinsically intentional entity as being both more defensible than that of “activity” and structurally the more important notion in the analysis of mind. For the moment I observe only that whether we rely on the one notion or the other—“activity” or intrinsically intentional entity—the issue concerns, as it should, not the existence or non-existence of some ordinary object or state of affairs such as an act of perceiving a tree, but rather the existence or non-existence at the basic ontological level of a category or 9

The notion of an intrinsically intentional entity is to be found among the medievals too in their notion of a natural sign.

187

subcategory of entities of certain characteristics. It is not surprising then that the historical review of the issue that follows should be built around the notions of relations, substance, and the theory of ideas. As early as his Theatetus Plato raised one of the fundamental questions of intentionality, a question that has plagued philosophers to the present day: how can we think about what is not the case or what does not exist? It is essential to my task that we understand both why that raises a problem and why a certain answer to it is not acceptable. My desk is in the relation being-two-meters-from to my bookcase. In order that it stand in that relation to the bookcase it is necessary that the bookcase, like the desk itself, exist. In general, if any object A stands in any relation whatsoever to another object B, it is necessary that both A and B exist. Now it is natural to suppose that to think of something or to perceive it is to be in some relation to that something. But how can someone stand in any relation to something that does not exist as when one imagines a centaur? Putting aside materialist “solutions” to this problem which are vacuous and irrelevant, there would seem to be three abstractly possible kinds of answers for it: (1) one may insist that there is, after all, always something there to be related to—a proposition, a false fact, a sentence, perhaps a Cartesian idea; or (2) one may claim that, on analysis, the appearance of being in a relation to something disappears altogether, or (3) one may hold that the “relation” is not really a relation but a link of another sort such that it can, after all, hold between an existent and a non-existent. If there are intrinsically intentional entities, the first or the third of these solutions would appear to be the right one. Yet the second probably has the most initial attraction. How tempting then to take it and thereby assert that there are no acts. Aristotle introduced into Western philosophy the ontological category of substance. Among the several features ascribed to a substance by philosophers is its capacity for spontaneous “activity,” although always in accord with its nature. For the medieval philosophers, even though the mind is still officially only the “form” of an essentially physical substance, it is in itself inherently “active”—this being the beginning of a line of thought that culminated in Berkeley and the idealists according to whom only the mind is “active”. For the medievals this “activity” was expressed primarily in knowing, something in which, according to their theory of

188

abstraction, the mind takes the form of something else and makes it temporarily its own. Officially, the mind is a form. In practice, as it were, the mind is “active” and takes on a form. The language was already that of a mind as a substance itself and not merely the “form” of another. Descartes made it explicit. The mind is itself a substance, and mental acts are the “activities” of these mental substances. If there are no such “activities” or if they are of no efficacy, then the existence of acts is, or was eventually taken to be, in doubt. Two developments, one from science and one from philosophy, tended to that very conclusion. In the first instance, the scientific revolution spread the idea of “mechanical” explanation, that is, of the physical world’s being causally closed. When mental acts are thought of as spontaneous “activities” impinging in a causal manner on the physical world, the grounds for their efficacy and hence their existence (since their very conception as “activities” has to do as much with their efficacy as with their intentionality) is undermined. But in the second instance, Descartes also taught that to be aware of something is for the mind to “have” an idea of that something where an idea is conceived as something like an image or picture. Yet we know, on phenomenological grounds, that we can think of things with very imperfect images of them or none at all. My thought yesterday of my friend whom I had not seen for several years was no less of him and of no one else even though my image of him in fact resembled his son more than him. In short, so far as ideas are something like images, we make them stand for what we are thinking of. Hence the intentionality of the thought cannot lie in the relation of resemblance that the idea or image has to anything. (Nor, of course, and for a similar reason, can it lie in any causal relation which the image or anything else in the mind has to the object thought about or perceived. If that is what the so-called causal theory of perception asserts, it is absurd. If it does not, then why should the causes of our mental states, whether of perception or imagination or remembering, be mentioned systematically at all in a philosophical account of their intentionality?) Images or ideas are not (always) object-specific. Yet my thought is undoubtedly specific: it is directed to one definite person and to no other. With something of a historical paradox, Descartes who was to affirm most explicitly the two-substance view of human beings (which is not, Ryle and many others to the contrary notwithstanding, the same as dualism

189

as such) by his theory of ideas initiated a development that led, in one major school of thought, to the demise of substance altogether. Locke was uncertain whether or not there are mental substances and his physical substances are but shadows of Aristotelian-Thomistic ones. Berkeley was convinced that there are no physical substances, only mental ones. Yet each, while subscribing to the theory of ideas in his own way, retained greater or fewer remnants of the act, Locke in talking about the operations of the mind, and Berkeley in his assertion that only minds are “active.” It remained for Hume, who recognized no substances, to declare that the mind is just ideas. There are no acts then because there are no “activities” and there are no “activities” because there are no substances. In such a manner, if my history is reasonably accurate, were the philosophically disastrous associations established: acts with a substance ontology and a notion of freedom as spontaneous “activity” incompatible with determinism and mechanism, all with an overtone of distaste for empirical science characteristic of post-l7th-century rationalism; empiricism with determinism, the theory of ideas, and an actless philosophy of mind, all deriving in part from the critique of substance. The associations were disastrous among other reasons because the critique of substance and hence of “activity” was mistakenly taken to cast doubt on the existence of intrinsically intentional entities. Modern psychology began a little over a century ago with a conception of its task that derived essentially from philosophy: to discover the nature of and connections among the “elements” of the mind. An element is, of course, a “simple” entity. A simple entity is one with no constituents. So the argument persisted: Are there among those elements any that are intrinsically intentional? Are there acts? Given the background I have just sketched, it is not surprising that those who defended the act as being among those elements—the school of intentionality—should also typically have been substance philosophers, although in the case of Brentano, only minds and never bodies are genuine substances. Brentano insisted that the act is the distinguishing mark of the mental, and Meinong and his student Ehrenfels explicitly (but deplorably) gave the act its old “active” role. The unfortunate associations also persisted. Meanwhile, the founders of the first laboratories in psychology, thinking of themselves as good empiricists, maintained that a mental state

190

consists entirely of “ideas” or rather, as they had now come to be called, “contents.” Contents, hovering uncertainly between minds and physical objects, at least have, or were described as having, the same properties that we ascribe to physical objects—colors, shapes, pitches, and the like. This, it was felt, is at most a very mild dualism, whereas to make the act the distinguishing mark of the mental is to maintain a radical dualism of mind and body that is not in accord with the modern outlook. Indeed, for many of the members of the experimentally-oriented tradition, the insistence on the act came to be seen as downright pandering to religion and superstition. As Boring stresses: “Act fits the Church better than content, for it tends less toward a mechanistic and therefore a deterministic psychology” (Boring, 1957, 453). Thus, the suspicion of and even hostility to the act has several sources. First, the act is conceived at the outset as being or involving a relation between a knower and that which is known. But then, according to its critics, there is no way to understand how there could be such a relation in the situation in which, as we ordinarily speak, we think of what does not exist or obtain. Second, the act is thought of as being or depending upon an inherent “activity” of the mind, a notion that, it will be (rightly) said, derives from an anthropomorphic and outmoded substance ontology. Third, and closely related to the preceding point, the act is often advanced as the basis of a presumed “freedom” of a sort that would make determinism false and psychology “non-mechanistic”. And fourth, the act is described, unlike mere contents, as being radically unlike physical objects or their constituents. This, it is objected, introduces an extreme and unacceptable dualism into the world. In short, it may be said by way of summary of this line of thinking, the act is “unscientific” especially for anyone committed to behaviorism. I turn to the philosophy of the act. IV As I walk down the street I suddenly remember that my grandmother wore glasses. What in this situation are the phenomenological data? First, it is given to me that something has happened in the universe and something of which, despite the several people around me, only I am or could be directly

191

cognizant. No matter what changes in my brain or central nervous system accompany this occurrence, it is given to me as something in addition to those or any such changes, that is, as a non-physical occurrence. But the falsity of materialism is not my present point and has been labored enough already. Second, it is given to me that I am indeed remembering (whether accurately or not) and not, say, doubting or imagining or perceiving. This feature is what Bergmann calls the species of the mental state. And third, it is given to me that it is that my grandmother wore glasses I am remembering and not, say, that I had strawberries for breakfast or that Hubert Humphrey became vice-president of the United States in 1965. It is not of course the intention itself of my remembering that is directly given—I don’t have my grandmother herself in my mind—but rather that the intention is my grandmother’s having worn glasses. Let us call these latter two features which are directly given to one in any typical mental state simply the species feature and the intention feature. How shall these two features be grounded in the ontological analysis of mind? To answer that, we should first have at least the sketch of an ontology before us. In Bergmann’s ontology as in many others, there are the two fundamental categories of particulars and universals. The universals are of two major subcategories--what I shall call simply properties and relations. That properties and relations are universals rather than so-called “perfect particulars” is of no present, direct relevance. As for the particular, they are, unlike the substances of the traditional act philosophers, momentary and not inherently “active”. Their being momentary, while obviously of importance in any treatment of the problem of the self, is of no relevance in the context of this discussion. Their not being “active”, however, is crucial. Bergmann properly rejects the notion of “activity” (reflected in some contemporary notions of “agency”) as unduly anthropomorphic and “irremediably confused”. The world, including mind, is in this respect Newtonian and not Aristotelian. Consider again my sudden remembering and its two crucial features. Like every particular occurrence in the history of the universe, it will consist at least in the being of one particular. The two features, being, broadly speaking, descriptive aspects of the situation, must be grounded in either properties or relations. Which shall it be? If we recall that the features are directly given ones, the question almost answers itself. A

192

relation cannot be given to me unless both (or all) of its relata exist and are also given to me. Yet when I think of centaurs and mermaids, the intention does not exist, and when I think of my grandmother, she is not directly given to me. Nor is there any other entity that could plausibly serve, in addition to the particular that “is” my mind, as the other term of a descriptive relation that would ground either or both of the two features. So both the species feature and the intention feature must be grounded in properties of the particular and not in relations that particular has to the intention or to anything else. What are those properties? The most natural way to refer to the ones of our example are being-a-remembering and being-the-thought-that-mygrandmother-wore-glasses. My inscribing their names with hyphens is designed to indicate that each is a simply property, that is, has no other properties, of the same or of a different kind, as constituents.10 Being simple, they both have the first qualification for being an intrinsically intentional entity of a sort that would entail that there are acts. Neither, of course, is of the character of an “idea” or content or sense datum. Neither therefore is either an image of or a constituent of the intention. It is obviously the second of these properties, the one that grounds the intention feature, that is an intrinsically intentional entity. It is such that by an “inspection” of it, one can tell what specific state of affairs it points to. It is therefore what the medievals would call a natural sign unlike the purely conventional ones of language and the partly voluntary ones of both mental and physical images. Nor can the intentional property (as I now call it) be decomposed into simpler properties none of which is a natural sign. Thus if intentional properties exist (are exemplified), then there are acts. I have already given the sole, fundamental “argument” for maintaining that there are such properties—the phenomenological fact of my knowing immediately what the intention of my remembering is. So, on this analysis, which is essentially one first proposed by Bergmann and defended in greater detail by Grossmann (1965) and this author (1965) elsewhere, my remembering consists of a particular’s exemplifying the species property being-a-remembering and the inten10

On other notions of analysis, remembering might be said to be a complex insofar as it will be maintained that, no doubt, ‘S remembers P’ is to be analyzed as ‘. . .’ and ‘. . .’ and, etc.

193

tional property being-the-thought-that-my-grandmother-wore-glasses. And it is clear that this analysis can be extended to all mental acts, commonsensically speaking; even, in my opinion, to the experience of pain and similar sensations. No relation is involved in the analysis. To remember something is not, ontologically speaking, to stand in the relation of remembering to that something. Yet it remains to be answered how, after all, a mind is connected with what it thinks about. If I am in no relation to my grandmother when I remember that she wore glasses, surely there is yet some kind of link between her and me that does not, at that moment, hold between me and anything else. It will be apparent that for the analysis here outlined and using the example I have chosen, the problem concerns precisely the connection between the intentional property being-the-thought-that-my-grandmother-wore-glasses and my grandmother’s wearing glasses, whether or not that state of affairs that is the intention of my remembering actually obtains (or did obtain or will obtain). What is the connection between a natural sign and that of which it is the sign if it is not a descriptive relation? In Bergmann’s language, the answer to the question is that the intentional property means the state of affairs intended. For our present purposes the word itself is not particularly important although it is by Bergmann naturally very carefully chosen. What is important for us is its general ontological character. Let us consider. Or, as in the fact stated by ‘It’s raining or it’s not raining’ is a logical connection. Two-meters-from as in the fact stated by ‘My desk is two meters from my bookcase’ is a descriptive relation. A descriptive relation, in order to be exemplified, requires the existence of its relata. In the fact stated by ‘It’s raining or it’s not raining’ the logical connection or holds between an existing state of affairs and non-existing state of affairs. This possibility is one feature of a logical connection. Two-meters-from is causally efficacious in that by the law of universal gravitation distance is a relevant variable in explaining certain phenomena. All relations are of lawful significance. Or, like all indubitably logical connections, is of no lawful significance, is not causally efficacious. That is another feature of a logical connection. The link between an intentional property and what is intended is a logical connection. By assimilating this link to the general category of logical connection, two palpable advantages are obtained: (1) it

194

is at least somewhat easier to understand in the manner appropriate to philosophy how we can think about what does not exist, and (2) it accounts for the commonsense fact that to think about something even when it does exist is not to affect it in any way. There may, of course, also be a descriptive relation between the knower and the known which may indeed be the cause of the knowing. But the intentional link itself is logical. Bergmann is, I believe, the first to see clearly that the link between the mind and what it thinks about must be a logical connection, cannot be a descriptive relation.11 I also believe that this insight and its articulation in his writings constitute his most important contribution to philosophy. Be that as it may, it is well to remind ourselves that the analysis carried to this point leaves entirely open the answer to, but invites attention upon, the difficult questions of the ontological status of logical connections and of non-existent states of affairs. Bergmann himself has devoted considerable energy to those tasks.12 For what I am about however we need not pursue them. If that is not already evident, it will become so in the final section of this paper. V Before I state positively the fundamental point of this paper, permit me to rehearse the essential features of behaviorism as I described it and of the philosophy of the act also as I described it and defended it but in its basics as developed by Bergmann. As to behaviorism: its fundamental assumption is that for the purposes of scientific explanation of human behavior, a full set of relevant, independent variables exists among the physical ones alone, and for the same purposes the non-physical ones may therefore be ignored. Thus, the defender of behaviorism must deny that there are non-physical variables that interact with the physical ones. If that defender also wishes to be consistent with commonsense dualism and the everyday facts about and scientific knowledge of the effects of minds and bodies on each other, he will also subscribe to parallelism rather than, as would also be consistent with the fundamental assumption, materialism or fatalism. As to 11 12

Brentano may also have glimpsed this point. In this connection see Arnauld, 1975. See especially Bergmann, 1967.

195

the philosophy of the act: despite its historical antecedents in a substance philosophy and the corresponding notion of spontaneous “activity,” the essential characteristic of the act, that of an intrinsically intentional, simple entity, can be consistently maintained without being committed to those historically antecedent notions. Furthermore, all of the descriptive features of the act are properties of the mind and none relations between it and something else, the intentional link itself being a logical connection. One more brief preparatory step is necessary. Temporal relations are descriptive relations. Few would dispute that claim and I shall not argue for it here. They are therefore of lawful significance. Yet their significance is unlike that of all other properties and relations in that their holding is the very precondition for any sort of lawful connection whatsoever between and among the phenomena. They are pervasive; indeed, in Bergmann’s words, time is the “great binder” of the realms of the physical and the mental. So one (of simultaneity) holds between any given mental state and its lawfully corresponding bodily state. Having taken note of the special status of the temporal relations (which, by the way, does not apply to the spatial relations), I may now ignore them in the following decisive paragraph of this paper. If the act did involve a descriptive relation of lawful significance between the mind and what it intends (or anything else non-mental), then the physical world would not be causally closed and the fundamental assumption of behaviorism would be false. But since, on the philosophy of the act as described and defended, all the descriptive features of the act are properties and none relations, all such features may be held to be connected by laws of coexistence to features of the physical world. Hence, the philosophy of the act is consistent with the denial of interactionism. Thus is the fundamental assumption of behaviorism safeguarded within the philosophy of the act. One may fully recognize the act but deny that there are “activities”. To do so is to separate the philosophy of the act from its historical origins in substance philosophy. And as we have seen, the recognition of the existence of intrinsically intentional, introspectively given entities need in no way diminish one’s commitment to the possibility of a deterministic and “mechanistic” psychology, that is, to behaviorism. Only intellectual fashion, not logic or phenomenology, says otherwise. So it is not surprising

196

after all that Gustav Bergmann, who began his career as a member of the Vienna Circle and as such was a proponent of logical positivism and behaviorism, should emerge eventually as the most distinguished contemporary representative of the Austrian school of the act.

197

PARALLELISM, INTERACTIONISM, AND CAUSATION

O

ne may gather from the arguments of two of the last papers (Mackie, 1979, 1981) published before his death that J. L. Mackie held the following three theses concerning the mind/body problem: (1) There is a distinct realm of mental properties, so a dualism of properties at least is true and materialism false. (2 All bodily movements probably have sufficient causes in physical facts and properties, but mental facts and properties are not causally irrelevant to human action. (3) At the same time, the view that there are not sufficient causes in the physical realm alone for all bodily movements has no good and adequate empirical or philosophical reasons against it. In this paper I wish (1) to register my strong agreement with the first thesis by way of simply taking it for granted, (2) to defend the second thesis in greater detail and in a manner somewhat different from Mackie’s, and (3) to show the third thesis to be false. I If a dualism of properties is true, there are fundamentally three abstract possibilities: the mental properties are related to the crucial physical properties by (1) laws of coexistence, (2) laws of succession, or (3) no laws at all. These views may reasonably be labeled as (1) parallelism, (2) interactionism, and (3) fatalism, respectively. Although there have been people who, crippled by their theological commitments, have thought they believed in fatalism, it is phenomenological absurdity to maintain that what one desires or chooses or values never makes a difference to one’s behavior; and no one ever acts that way either (whatever it could possibly be to act as if one’s mental life made no difference to one’s behavior). That

leaves only parallelism and interactionism as realistic contenders for the dualist’s allegiance. But are parallelism and interactionism, as characterized above, really exhaustive of the remaining possibilities? What about a theory according to which some mental properties are tied by laws of coexistence while others are tied by laws of succession to the crucial physical properties? My reply is that if any mental properties are tied to the physical world essentially by laws of succession, then the view of the interactionist is correct. The reason is that in such a case the physical world is not causally closed, and the issue between the parallelist and the interactionist is really just the question of whether or not that is so. So it is time to explain clearly what is meant by saying of a system that it is causally closed.1 Assume that determinism is true. Eventually I shall remove the assumption, but it makes it easier to state the basic ideas initially. Determinism is the thesis that for every occurrence or state of affairs in the past and present and future of the universe, there exists some earlier occurrence(s) or state(s) of affairs that is lawfully sufficient for its occurrence at the time at which it in fact occurs. Otherwise put, determinism is the thesis that identical conditions not acted on from without produce of lawful necessity identical consequences. Citing a lawfully sufficient condition for a given occurrence as well as the relevant law(s) is to provide a full explanation of that occurrence. (I say this stipulatively and do not thereby rule out other kinds of explanations of occurrences—by reasons or dispositions or constituents or whatever.) It is evident then that a given occurrence will have more than one full explanation if, for example, the occurrence of either the set of properties a, b, and c or the set a, b. and d is lawfully sufficient for that occurrence. This will be so for one reason if c and d are themselves bound by a law of coexistence such that the occurrence of either lawfully implies the simultaneous occurrence of the other. To put the idea somewhat more informally: the fact that one has found a full explanation for some occurrence in the sense stipulated does not imply that the occurrence or nonoccurrence of anything else is lawfully irrelevant, especially insofar as that something else is itself lawfully related to that which is offered 1

Some of what follows can be found also, but in more detail, in Addis, 1982.

200

originally in explanation. A kind of system is specified by listing the properties (or variables, as some say) that are taken to characterize the kind of things in the system. A particular system is specified by additionally listing the names of or otherwise indicating all the particulars in it. A state of the system is specified by describing for an instant the values of each of the variables for each of the particulars of the system. A deterministic system is causally closed if for every state of the system there exists some earlier state of the system that is lawfully sufficient for its occurrence, that is, that given the laws of the system constitutes its full explanation. If one must go “outside” the system to properties2 that are not part of it (or of its kind) to obtain a full lawful explanation of any occurrence in the system, then the system is not causally closed. Notice here for later reference the use of ‘must’ rather than ‘may’ in the last sentence: a system’s being causally closed does not preclude the possibility of legitimately citing some occurrence “outside” the system as at least part of the full explanation of some occurrence within the system. If a, b and c are within a system and (some values of them) are advanced correctly as a full explanation of (some value of) e, which is also within the system; and if further (the values of) c is (are) tied by a strict law of coexistence to (the values of) d, which is “outside” the system, then a, b, and d also constitute a full explanation of e. Thus my use of ‘must’ rather than ‘may’. It will now easily be seen that the general idea of a causally closed system can be captured even when the system is nondeterministic provided that we allow some nonformalized notion of “degree of explanation” into our thinking. If for some occurrences in the system there do not exist any full explanations (that is, these occurrences have no lawfully sufficient antecedent conditions no matter what is taken into account either within or without the system), then the system is causally closed but not deterministic if whatever degree of explanation those occurrences do admit of can be found within the system. This circumstance too does not preclude the possibility of occurrences or properties outside the system being legitimately cited in an explanation of some occurrences within the system. 2

One might go “outside” the system also in the sense of having to mention other particulars; but this is of no theoretical interest, in science or philosophy.

201

Before I use these ideas to attempt a precise characterization of parallelism and interactionism, one preliminary remains. Even the dualist, such as myself, agrees that some of the properties that are commonly called mental are (also) physical properties. But since the theses of the parallelist and interactionist apply crucially to (“occurrent”) conscious mental states, it is desirable for analytic purposes to treat as mental properties only those that are exemplified by such states and to consider all other properties as physical. So doing, we may now say that contemporary parallelism and interactionism consist of the following propositions, respectively: Parallelism (1) Mental properties and physical properties constitute exclusive and exhaustive sets of the properties of the universe. (2) The physical world is causally closed. (3) The mental world is not causally closed. (4) Every mental property is tied to some physical property (or disjunction of physical properties) by a law of coexistence. Interactionism (1’) Mental properties and physical properties constitute exclusive and exhaustive sets of the properties of the universe. (2’) The physical world is not causally closed. (3’) The mental world is not causally closed. (4’) At least some mental properties are tied to physical properties only by laws of succession. Three comments on these characterizations are necessary before I turn to evaluation of the views. First, probably the most famous parallelist of all, at least on one reasonable interpretation of his words, denied (3) explicitly and possibly (4) as well. In short, this philosopher held that the realm of the mental is also of such a scope that everything that happens in it has a full explanation by some other occurrence(s) in it. Whether Spinoza should really be treated as holding (4), when he explicitly insists that the connection is one of perception, we need not bother with. In any case, the contemporary parallelist need not, nor to my knowledge does any, hold that the mental is of such a scope. But (3) of course implies (given also (l), it must be added)

202

the necessity of the physical in the explanation of the mental. Since therefore this “parallelism” is by no means a simple, one/one connection (and I shall weaken it further a few sentences hence), one may wish to question the propriety of the label. And further, since this implication of the asymmetry of the relation of the mental to the physical in general seems to suggest that the physical causes the mental but never the other way around, some may ask why we don’t forthrightly admit that parallelism socalled is really epiphenomenalism and be done with it? For the moment, however, I wish to characterize the alternatives of what I shall continue to call parallelism and interactionism only in terms of the kinds of lawful connection that each involves and, despite my use of ‘causally closed’, reserve all questions of causation in the relation of the mental to the physical until later. This way of proceeding, despite its rarity in the literature, is, I am convinced, the more fruitful in grasping clearly what is involved. Second, it is important to see that my use of ‘only’ in (4’) above is not superfluous: if every mental property is related by a law of coexistence to some physical property, then whatever laws of succession may also apply to the system, the physical world is causally closed and parallelism is true. For, at any point at which one might cite some mental occurrence in the explanation of some later physical occurrence, there will always be a lawfully simultaneous physical occurrence that can serve instead. But the interactionist insists that some physical occurrences require mention of the mental in their full (or maximally possible) explanations, that there are “gaps” in the physical realm as far as the explanation of some occurrences in it are concerned. Hence the use of ‘only’. Third, I have so far given the impression that the parallelist would hold or even must hold that the lawful correspondence of any given mental property to some physical property is one/one. It is now time to loosen this assumption and to understand clearly that the parallelist is not so bound, consistent with the four propositions that define the position; and that the parallelist can therefore allow for the possibility of different physical grounds for qualitatively identical states of consciousness. Given the asymmetry noted earlier—the physical but not the mental realm is causally closed—all that is required for parallelism is that the laws permit the “deduction” of the mental from the physical but not the other way around.

203

In short, the parallelist may allow a many/one connection from body to mind, but not a one/many from body to mind nor, obviously, a many/many connection. This will always imply that two persons or any things or beings whatsoever that are in qualitatively identical physical states will have qualitatively identical mental states. Indeed, with this idea, the parallelist’s position may usefully be contrasted with the interactionist’s as follows: the parallelist affirms and the interactionist denies the lawful impossibility of two persons or other beings or things being in the same physical state but having different mental states. Putting the matter this way will permit me eventually to formulate a very serious objection to interactionism. II “Parallelism implies fatalism, and fatalism is absurd.” That is, both historically and analytically, the most serious objection to parallelism. The objection goes that if the physical world is causally closed, as the parallelist claims, then it really makes no difference what goes on in a person’s mind, what states of consciousness a person has, including the conscious states of desiring, willing, and so on; and that is fatalism. Surely, it may be said, if we are going to be dualists at all, we must also be interactionists in order, like the central-state materialists, to give the mind its proper explanatory role in human behavior. The materialist, to be sure, holds with the parallelist that the physical world is causally closed; but then, according to the materialist, that world is all the world there is and already includes whatever one may wish to call “mental.” What should the parallelist say in response to the charge of holding a position that implies fatalism while surely agreeing that any view that implies fatalism is itself absurd? Not only does parallelism not imply that it makes no difference what goes on in one’s mind to one’s behavior, but in fact implies just the opposite. If I did not have the mental state I now have, then by the law of coexistence that ties that mental state to some state of my brain, that brain state would not be occurring and so my behavior would be different, and so on. It is lawfully impossible, by the parallelist’s very position, for that brain state to occur without that state of consciousness also occurring; and

204

if someone wonders why, on the parallelist’s view, the universe couldn’t just as well have been exactly as it is physically but without the occurrence of any mental states at all, the answer is that the universe well could have been that way in the sense that its laws could have been different from what they are without contradiction, but that is not the way it in fact is. The way it is makes it lawfully impossible for me to write this essay without thinking about what I am doing, desiring to write it down, and so on, because unless those states of consciousness occur in me, the relevant brain states won’t occur either. None of this, however, contradicts the original assumption that the physical world is causally closed and that those brain states also have a full lawful explanation (or maximum degree of explanation) in the physical world alone. To make use once more of the abstract symbols for the sake of clarity: let a, b, and c be physical properties with c being the relevant brain state: let d be a mental property; and let e also be a physical property, the person’s behavior. Then while, by assumption, a, b, and c jointly explain e (a and b perhaps being the state of the rest of the physical universe at the moment), so do a, b. and d. But c lawfully cannot occur unless d occurs since, also by assumption, they are tied together by a law of coexistence. Hence e will not occur (in the particular case: I have not assumed that a, b, and c are lawfully necessary for the occurrence of e) unless d occurs. This also illustrates how the parallelist may say, at least consistently with what has been said so far, that the explanation of a person’s behavior lies in his mental state insofar as everyone allows that a particular context permits one to cite but one of the factors of a set that only jointly are sufficient for any given occurrence. Many other arguments have been made for and against parallelism and interactionism. Many of them, I believe, carry very little weight, such as direct appeals to common sense, calling attention to various facts about evolution, arguments that involve the principle of the conservation of energy, and a priori arguments about the ontological possibility or impossibility of either view. 1 shall comment briefy on the first and the last of these matters, taking the latter first, however. Without going into much detail, let it be said here that none of the positions mentioned so far—parallelism. interactionism, fatalism, and even materialism—is ontologically impossible. By this I mean that, consistent

205

with the basic principles and findings of general ontology, the universe could have been as each view says it in fact is. The only amendment I would make to that stark claim concerns materialism: insofar as the materialist says simply that everything is physical, what he says involves no obvious ontological impossibility, Berkeley and some other idealists to the contrary notwithstanding. But insofar as the materialist maintains that mental properties are literally also physical properties, or that no properties exist (and, paradoxically, therefore no mental properties), or that only the basic properties of physics exist in a world that is given to sentient beings in that world as having many other properties as well, he does speak ontological nonsense. Fatalism’s absurdity, as I said earlier, is phenomenological and not ontological: there could be a world with two lawfully unrelated sets of variables—the mental and the physical—with the noncausal ties of intentionality and time being the only “links” between members of the two sets. But experience shows conclusively that this is not our world. Finally, since I have defined ‘parallelism’ and ‘interactionism’ by way of the kinds of lawful connections that may hold between mental and physical properties, there should not be and is not any ontological difficulty with either view. That judgment does indeed presuppose something like a Humean ontological principle that between and among simple properties, any lawful connection or lack thereof whatsoever is possible, a principle which, although I shall not argue for it (I really wouldn’t know how to argue for such a fundamental principle), I firmly believe is true.3 So no basis for choosing between parallelism and interactionisin is to be found among the principles of ontology themselves. It is a piece of true common sense, firmly grounded in the phenomenology of the relevant situations, that what happens to and in our bodies affects what goes in our minds and also that what goes on in our minds affects what happens in and to our bodies. Call this, if you will. “commonsense interactionism.” But we have already seen that parallelism can account for, and even requires, the fact that some of what happens in each realm makes a difference to what happens in the other. Hence, commonsense interactionism is fully consistent with parallelism. So a 3

For some discussion of the principle in the context of the philosophy of the human sciences, see the third chapter of Addis, 1975.

206

direct appeal to this kind of common sense as an argument for interactionism is simply a mistake. But how then shall we proceed in an evaluation of the relative merits of these two positions, one of which is almost certainly the truth about the relation of mind to body? When one considers again a certain feature of the interactionist’s view and investigates its implications, we find, I believe, good and adequate reasons for rejecting interactionism. Since there are no similarly good reasons for rejecting parallelism (although I still have the crucial matter of causation to deal with), we must conclude that it is rational to believe that parallelism is the truth. That feature of interactionism that leads to its difficulties is the fact that on it, any one of a range, perhaps an unlimited range, of mental states lawfully may accompany a given physical state whether the latter be a brain state or a behavioral state. Without this feature, which is shared with fatalism, the view is not interactionism but parallelism, that is, if the “range” is one. It is again the question of whether or not two physically identical persons could, lawfully, have different conscious states. What this feature of interactionism entails is that it is not calculable by the kinds of laws that the interactionist claims to hold between mental and physical variables what the state of a person’s mind is from the states of his body. To see this idea more clearly, consider the analogy of mechanics in which an interacting set of variables of mass, position, and velocity is such that given (the values of) any one or two of those variables for some object or objects, (the values of) the third remain unknown, and this independent of temporal relations. To apply the laws of succession that are the laws of mechanics, one must independently ascertain the values of all three variables at some time in order to calculate the present, past, or future values of any one variable (except, trivially, the value of the same variable in the same object at the same time). When we apply this consequence to the mind/body problem, it has grave consequences for interactionism, as we now shall see. When anesthesiologists do their job correctly, they and everyone else assume that by putting the brain of the patient in a certain state, that patient’s consciousness ceases temporarily. When a suspected criminal is given a lie detector test and it shows a certain pronounced pattern when certain questions are asked, both those involved and most others assume

207

that when the suspect’s body is in a certain state as shown by the machine, the suspect has the mental state of intending to deceive. When a parent gives a child a spanking, that parent and everyone else assume that the child has the mental state of feeling pain in its posterior. These facts, and innumerable others like them, are, I submit, incompatible with interactionism and intelligible only on parallelism if the assumptions involved are true. All of them presuppose that when the body is in a certain definite state, the mind is in a certain definite state. This general fact, which we may label as “commonsense parallelism,” does support its philosophical namesake, parallelism, for, unlike the relation of commonsense interactionism to parallelism, commonsense parallelism entails the falsity of interactionism. Commonsense parallelism is in fact just the vaguely perceived fundamental claim of the parallelist: that it is lawfully impossible for two persons who are in every respect physically alike to have qualitatively different mental states at that same moment. Same body, same mind. The interactionist may try to avoid these damaging implications by limiting to a very narrow range the lawfully possible mental states that can accompany a given brain or other physical state; by stipulating that the members of the set of possible mental states that accompany a given brain state must be very similar to each other; by allowing that some but not all mental states or features of them are, after all, tied by strict laws of coexistence (which, again, may be many/one from body to mind) to properties of the physical world; or by some combination of the foregoing. Any such move would obviously be a significant step in the direction of parallelism proper. Can the interactionist leave things there and be left with a defensible position? The answer, I suppose, is that it depends on how closely the interactionist comes to resemble the parallelist. But for an interactionism that remains strong enough to be of any interest and that has any sense of internal coherence (that is, is not modified just to meet every example on an ad hoc basis), there remains a very serious objection which is my fundamental argument against interactionism. Common sense holds that each of us often knows not only that another person is conscious and awake but what in particular, at least in part, is, as we say, “going on” in that other person’s mind. Common sense also

208

realizes that, at nearly any moment, there may be more, even much more, going on in a person’s mind than anyone else knows and that sometimes we have no good idea at all what the conscious state of another person at a certain time is. Finally, common sense also acknowledges what science likewise takes for granted, that the only access we have to the mind of another person is through the observation of that other person’s body and the physical objects that he or she produces, such as books, works of art, conversations, and so on. Precise formulation of this idea is as difficult as it is unnecessary; however, the description of what we observe may be as sophisticated as and of whatever scope anyone wants, provided that it does not include the properties of conscious mental states as exemplified by anyone other than oneself. Now ignoring the matter of the very existence of other minds, how on the interactionist’s view could anyone ever know or even make a reasonable guess what another person is thinking? For, on this view, the fact that that person is in some particular bodily or behavioral state does not lawfully entail (or in any other way entail) that that person is in any specific mental state. Certainly we cannot ask the person what he or she is thinking, for taking the answer seriously would presuppose a connection of the coexistence kind between linguistic behavior and states of mind. (Thus the fact of communication and its presupposition that there is a systematic but not unbendable correlation between what a person says—in the sense of what sounds or marks are produced—and what a person thinks is another important aspect of commonsense parallelism, although this connection is, of course, not “natural” in the sense that one learns a language with all of its conventional aspects.) In short, nothing I can observe about a person at a moment can give me even a clue as to what that person is thinking, if interactionism is true. But, it may be said in reply, we don’t ordinarily rely only on what we observe at the moment anyway in order to calculate or come to believe what is going on in another person’s mind. We rely, varying widely as to the person and the sort of situation we are in and other factors, on shared and unshared cultural traditions, on what we know about human beings in general and what we know of the person’s past in particular, and probably much more besides. So if the interactionist is forced to deny that we can simply “read off” what another person is thinking from any observation of

209

only that person’s roughly simultaneous physical properties in addition to knowledge of laws or lawlike generalizations, he too only affirms common-sense truth. The parallelist, in retort, may immediately grant that while, according to his view, there exist laws that would in principle allow a person to calculate anyone’s mental state from a full description of his bodily and especially his brain state, in fact we often come to know what other persons are thinking in the manner just described. Not only do we not know all (or perhaps any, strictly speaking) of the relevant laws, but actually determining the precise state of anyone’s brain or body is factually impossible. But, the parallelist must insist, at certain crucial points and in certain crucial respects, we do know what is going on in another person’s mind on the basis alone of simultaneous circumstances and knowledge of or belief in certain generalizations. As to what causes us to have such knowledge or beliefs, we may speculate that much of it is part of our genetic endowment; for example, taking certain gestures (bodily motions) as indicative of friendly intentions or of fear or of submission. As to what justifies or could justify such knowledge and belief, something like the so-called argument from analogy must suffice. (But I do not argue here that, by some rigorous philosophical standard, we ever do know what is going on in another person’s mind. I take for granted that we do, in any ordinary sense of ‘know’, sometimes know what another person is thinking. My point is that interactionism entails the impossibility of any such knowledge, or even plausible guess.) I can know that another person is in pain on the basis of his behavior only if (1) I know from my own experience what pain is, and (2) I know from my own experience what behavior typically accompanies pain (which, of course, I may learn to inhibit; but that presupposes a natural accompaniment). Again, it is probable that we are “preprogrammed” to react in certain definite ways to “pain behavior” in others—to regard it with alarm, sympathy, and so on. It is difficult to understand how the species could survive when its newborns need such extended attention if it were otherwise, that is, all “learned.” But this is irrelevant to the question of what, if anything, could justify the rationality of such reactions even if they are causally impossible not to have. But then we cannot help but believe that others have minds, are sometimes in pain, feel desire, and so on.

210

The reason that the parallelist must insist on the existence of certain crucial situations of the kind in which one does virtually “read off” the other’s mental state from his behavior of the moment is in order to be able to avoid the same scepticism that the interactionist is necessarily faced with. Consider again the case of mechanics., Mass, position, and velocity constitute a set of interacting variables (properties). By assumption and in fact, the values of no one of those variables is, by any true law or generalization, calculable even within a certain range from the values of the other two either at the same time or any different time. In short, if I wish to make any reliable predictions about any future states of a system with respect to these properties, I must independently ascertain, that is, by separate observation of each, the values of all three variables at a time. The interactionist says, in effect, that a person “is” or “has” a set of interacting properties, some mental and some physical. But if another person can neither observe, and thereby independently ascertain, a state of another person’s mind at a time nor by any law of coexistence calculate that state from observation of the other person’s body (or anything else whatsoever), he could not possibly know or have the slightest good reason to believe that any other person is in any particular state of mind rather than any other, either at this moment or in the future. Nor, by the way, could he ever make any rational prediction about the future behavior of any other person. It would be comparable to trying to predict any of the future positions, velocities, or masses of some bodies on the basis of knowing only the present positions and velocities and the laws of mechanics—a sheer impossibility. Now it is true that we sometimes, in fields of science that do not involve the mind, come to believe that there are unobservable properties that apparently interact with the properties we do observe and even that we can sometimes calculate the values of those variables. The most obvious example is atomic and subatomic physics. Would my argument, therefore, not be subject to the reductio ad absurdum that if it were sound, it would also prove the impossibility of atomic and subatomic physics? The correct answer to this objection, probably already apparent to many readers, in fact strengthens my point and my argument. For any time a scientist wishes to test an hypothesis about the existence of an entity of a certain kind (that is, as having a certain property or set of properties), he

211

must assume a connection of the coexistence kind between that entity as exemplifying certain properties and some feature of the world of everyday experience even if it be only a reading of a dial (with many intermediate steps, of course, each of which assumes a connection of the coexistence kind). Every difference that is to be knowable and therefore testable for in the realm of the nonobservable must be paralleled by some difference under some condition in the realm of the observable. We need not make this a condition of meaningfulness in order to insist that an hypothesis about causes (or apparently so, such as “God wills all”) that yields no concrete predictions about what. under any specifiable condition whatsoever, will occur in the realm of the observable is, at least as an explanatory account, idle and empty. But this, it now seems clear, is exactly the position each of us would be in, with respect to the contents of everyone else’s mind, if interactionism were true. Only parallelism even makes possible the knowledge each of us supposes ourself to have about what is going on in the minds of others. Parallelism therefore appears to be the only position that is in accord with common sense correctly understood, with the experience each of us has of himself or herself, and with any attempt of science to discover as precisely as it can those features of the brain and central nervous system that are correlated with, say, dreaming or doing mental mathematics or willing to wiggle one’s ears. In its dualism, parallelism fully accepts the commonsense view confirmed by introspection that there are certain properties of the universe known to us only by introspection, that is, that there is a distinct set of mental properties, while in its parallelism proper, it and only it satisfies the requirements of common sense, inner experience, and science. III There remains the matter of causation. I call the reader’s attention more fully now to my conscious design to leave this matter until last in a discussion of the present issue in the belief that much contemporary discussion of the matter is flawed by a too hasty and unexamined introduction of causal (as opposed to merely lawful) notions into the argument. How often does one read that parallelism is patently absurd in

212

denying any causal role to the mind in behavior and so may be safely and quickly dismissed from further consideration? Is it not, after all, the contention of the parallelist that every physical occurrence has a purely physical cause? Now we have already taken notice of the fact that while on the one hand parallelism may imply to some (as in fact Spinoza explicitly held) that each realm has only its own causes, the modern parallelist is certain to deny that the realm of the mental is in fact causally closed and therefore to insist that only the realm of the physical can complete the full explanation of occurrences in the realm of the mental. But since the parallelist also holds that the physical realm is causally closed, the ground is laid for the charge that the modern parallelist is really better called an epiphenomenalist and a believer in one-way causation from the physical to the mental. This is not simply to repeat the earlier objection based only on the lawful connections involved, but rather to insist that even granted that it is lawfully impossible for certain brain states to occur without certain mental states also occurring on parallelism, that is not to say that mental states ever cause any physical and especially any behavioral states. And what common sense and possibly morality require is just that: that by having certain mental states such as desiring or willing or choosing, we cause certain effects in the realm of the physical. Since the parallelist must deny that anything in the realm of the mental ever really causes anything in the realm of the physical, parallelism must be false. So charges the critic. Causation is lawfulness plus context. That is my formula and my thesis. Clarifying it and defending it to some degree will at the same time allow me to defend parallelism against the charge that it denies some patent fact about the connection between mind and body. (So I do not take the line that some defenders of parallelism have—that it is merely an illusion that our mental states sometimes cause our behavioral states.4) At the same time I do not propose to make an elaborate defense here of any particular account of causation but only to state briefly what the theory of causation involved is and how it can rescue parallelism from the charge. If we begin by asking what lawfulness itself is, I have no new answer to that question. My view is that it is mere regular connection, understood to mean that a true law of nature is a generalization that does or would 4

See, for example, Honderich, 1981.

213

survive the tests of Mill’s methods. That of course does not imply, contrary to the beliefs of some, that for two properties A and B to be lawfully connected, there must be a true law that connects them alone in some regular fashion, that is, as either “If A, then B” or “If B, then A” or “A if and only if B.” A and B are lawfully connected in a law of the form “If A and C, then B,” for example, which implies none of the three previous sentences. That is one qualification to the notion of “regular connection.” Another is that since laws are best understood as not having existential import, a law may hold and even be importantly true of our universe even though it has no instantiations. Newton’s first law, the law of inertial motion, may be an example. Then the “regular connection” is or may be hypothetical—what would occur if certain conditions were to obtain rather than what does occur when certain conditions do obtain. This distinction between the subjunctive and the indicative is the source of many familiar difficulties and disputes which, however, I shall not discuss here. I merely register my agreement with what I take to be the essential insight of Hume—that there is not, in anyone’s experience, any additional entity such as a “necessary connection” to be included in the proper understanding of lawfulness or, now to move on, causation. If, as the examples of many philosophers might show, we can in a sense “observe” a causal connection in the particular instance, such as when a rolling boulder smashes a hut (to repeat an example from the literature), nothing very important philosophically follows. We may be simply “preprogrammed” to take certain kinds of occurrences as instances of causal connections immediately, so to speak, rather than, as the regularity view might seem to suggest, only by realizing those occurrences to be instances of lawful connections. Indeed, the survival of the species would seem to require such “preprogramming” at least in the behavioral responses to such situations, if not also in their cognitive evaluation. The work of Chomsky, of the so-called sociobiologists, and of others such as the traditional ethologists has made it increasingly plausible to believe that certain beliefs that we have are partially “preprogrammed” in us in the sense that these beliefs, while possibly true and even rationally justifiable, come to be held on the basis of extremely slight evidence which, by the usual canons of induction, and at the time and by the means that they typically come to be held by a particular person, have an entirely

214

inadequate foundation in the justificatory sense. It is easy to see, in the case of beliefs about causal connections, how, provided that it is biologically possible for it to be so, there could be considerable survival value in having the genetically-based predisposition to have such beliefs under relatively “weak” circumstances. Observations such as these are in any case, I believe, the proper way to discount the otherwise somewhat perplexing examples of the sort I mentioned. Negatively, these philosophers have as yet failed to make clear what additional entity of an ontologically interesting kind is supposed to be present in such experiences. It is this failure and not any utter lack of difficulty in the regularity view that makes the regularity theory the most plausible account of lawfulness and the ontology of causation. But causation, I said, is lawfulness plus context. For, at least as we speak, not all lawful connections are causal connections. Which ones are, which not, and why? I believe that, despite the attempts of some philosophers to have a “theory” of what additional features must or must not be present in order that a certain lawful connection be a causal one, it is altogether a mistake for the philosopher to attempt any formal or systematic account of these features. Even the most commonly mentioned one of the temporal relation (as a necessary, not a sufficient, condition) of priority of cause to effect is reasonably disputable insofar as specification of contexts is largely a description of linguistic behavior (or, as some philosophers like to call it, “intuitions”). What remains therefore is to specify certain contexts that are relevant to talk about, or allusion to, the relation of mind to body. Specifying such contexts up to a certain point will, if I am successful, achieve the goal of reducing the charge against parallelism to a harmless one that can be lodged, once one sees through it all, against all of its alternatives as well—that the sense of the view seems incongruous with certain ways of speaking and thinking when, in a certain context, some features of the lawful situation are being treated as causal and others not while in a different context, the opposite or something like it might be true. I begin with the easier task, mainly in order to illustrate the idea of specifying contexts, of showing circumstances in which it would be natural to say that the physical is the cause of the mental but not the other way around. Almost anyone can be caught in a frame of mind in which he or she will

215

readily agree that while there can be lifeless and therefore “mindless” bodies, there cannot be disembodied minds. This view, which simply asserts the primacy of matter, is sometimes called philosophical materialism; but it might reasonably also be called “commonsense materialism,” provided that it is clearly seen that commonsense materialism does not entail absolute materialism, the theory that there is nothing non-physical. More specifically, we do believe (apart from religious and other even more dubious motives) that while a brain can exist without a mind, a mind cannot exist without a brain or other, somehow comparably complex physical ground. This fact, or one’s awareness of it, provides a context for saying that while, on the parallelist’s view, a certain brain state lawfully cannot occur unless a certain mental state occurs simultaneously, and even while, if the connection is many/one from body to mind, that mental state might occur under different physical conditions, the brain state is the cause of the mental state and not the other way around. Another, related context is provided by the fact just mentioned— the probable many/one relation of body to mind—which, given also the relevant laws, makes the mental calculable from the physical but not the other way around. And of course a third context has already been mentioned in formulating the objection—the claim of the modern parallelist that the physical realm but not the mental is causally closed with its implication that the full (or maximally possible) lawful explanation of mental phenomena requires mention of the physical but not the other way around. But all of this, while it identifies contexts for saying that on the parallelist’s view the physical is sometimes the cause of the mental, is only the less important half of the story as far as most critics of parallelism are concerned. For the great fear as we know, is that if parallelism is true, it is never the case that anything mental causes anything physical. How, finally and precisely, should this charge be dealt with? One may begin by pointing out that it is usually subsequent behavior and not simultaneous brain states that anyone is anxious to be able to regard correctly as being sometimes the causal consequences of having certain mental states. Of course, insofar as there is a chain of intermediate physical occurrences between one’s choosing to do something and doing it, one may wish to say that the mental state of choosing must also be the

216

cause of those intermediate physical occurrences. But that is, philosophically speaking, irrelevant detail. It may also be noted that any talk about the cause of an occurrence, including an instance of human behavior, apart from a specific context sounds a little odd, although perhaps not outright misleading or false. There is, of course, in a different philosophical context, a long story to be told about reasons, motives, intentions, and so on in which more subtle distinctions and finely grained contexts would be drawn upon than we need to invoke here.5 In a sense, therefore, after all this preparation it remains to point out only the same fact that was earlier stressed when it was alleged that our mental states make no difference to anything physical if parallelism is true, a charge that I was careful to distinguish initially from the one now under examination. The fact is, to remind ourselves, that on parallelism, it is lawfully impossible for certain behaviors to occur at least as constituents of certain patterns of behavior unless certain mental states also occur and occur temporally prior to and “in” the same person as the behavior in question. This fact, when combined with the general conception of causation that is the only ultimately intelligible one, gives a sufficient context and a sufficient reason for being entitled to characterize such situations, even on the parallelist’s view of the lawful connections involved, as ones in which a mental state causes a physical state. The “particularists” about causation often point to our experience of willing and doing as another and, for some of them, the most important kind of case in which causation is immediately experienced. They may be right. As one strongly committed to the importance of the phenomenology of experience in philosophical musings. I find it difficult to deny the strength of this claim. But if I am also right, this fact is entirely consistent with the regularity account of causation and with parallelism. And if am right about that, then, I submit, we have removed the last and most difficult barrier many have encountered in believing what is surely so: that parallelism is true.

5

I have told some of this story in Addis, 1981a.

217

`

APPENDIX: A SPECULATION

Mental properties are not publicly observable. Yet each of us has observed (by introspection) a sudden remembering, for example. How does any of us know that in observing a sudden remembering he is observing the same property as others do? How, for that matter, does any of us know that he is observing the same property on successive occasions of his own sudden rememberings? I am willing to allow a certain force to these Wittgensteinian questions with their implicit answer that “the inner stands in need of an outer criterion” provided that it is also realized that, since at some point a person simply does take some property or properties as being the same without a “criterion,” there is no reason in principle, that is, by the intrinsic nature of the mental as mental, why that person cannot so take mental properties. Certainly the “privacy” of their particular instantiations is no such reason, but only, if at all, the causal fact that one cannot learn to identify them as such without an “outer” criterion and without language. Furthermore, whether the property that is given to me when I am aware of my sudden remembering is the same property that is given to you when you are aware of your sudden remembering is a matter of no greater significance than whether the properties given to me (or otherwise observed by me) when I see a cow or hear someone speak are the same properties observed by you in similar situations. Every experience that involves a cognitive aspect, whether of something “inner” or “outer,” requires a judgment or a presupposition or (not to be too intellectualistic about it) a simple taking of some property or properties as the same properties one has experienced before. (And I am not talking about culturally determined “family resemblances” but literally the same simple property, such as a particular shade of red or a certain pitch.) I have already argued that if interactionism were true, we could never know what particular state of mind, if any, another person was in at a particular time. That, of course, is quite a different sort of doubt from the kind that Wittgenstein’s musings are supposed to engender. But his are, I believe, entirely misplaced except insofar as they may show how very deeply rooted in common sense parallelism really is. If there were not some reasonably systematic connection of the coexistence kind between our inner and our outer lives, we probably would never learn the language

218

of mental life. But that is “merely” a causal fact and goes little deeper than the fact that while most of us learn to identify colors as such, that is, without any “criterion,” few of us can do so with pitches. Mental properties, I suggest, are somewhat “between” colors and pitches in this dimension: we are able at first to identify mental properties, or some among them, only by outer “criteria,” but eventually we are able to identify them in themselves. Since we can, one way or another, come to identify at least some mental properties without “criteria,” it is intelligible for a philosopher, including the empiricist, to hold that mental properties are a distinct set from the physical and to imagine that the world could have been such that mental and physical properties interact, that is, that parallelism could have been false. My suggestion is that it is only because the physical world is causally closed that we can have, as we do, a conception of a distinct mental realm. For if the physical world were not causally closed, that is, if interactionism were true, then we would never be able to learn the language of mental life and would have no systematic idea of a distinct realm. There would be no outer criteria of inner episodes. So we may be assured that parallelism is true by the simple fact that we can imagine it.

219

DISPOSITIONAL MENTAL STATES: CHOMSKY AND FREUD

L

et it be granted that while some mentalistic notions refer to conscious mental states and others to (what I call) secondary mental entities,1 yet others refer to dispositional mental states. Even so, it is not my intention in this essay to argue that certain mentalistic notions, such as knowledge or belief, really are dispositional ones or that others, such as remembering and perceiving, are not. Detailed analyses of particular notions in which one tries to demonstrate, for example, that there is or that there is not such a thing as “occurrent” belief and to identify the kinds of behavior and conscious mental states that are the realizations of the dispositional mental state of believing something, are not the business of fundamental philosophy of mind. Its task is only to identify the basic categories and nature of mind and ordinarily does not, except for purposes of illustration, pursue the further detailed analyses that are appropriate in other, usually epistemological or metaethical contexts. Yet there are specific dispositional notions whose detailed consideration is of more than local interest. The theories of Chomsky and Freud provide rich examples of mentalistic notions whose exact status in an articulated metaphysics of the mind is worth pondering, especially insofar as they or their followers have sometimes drawn dubious philosophical conclusions from what were initially scientific theories. Treating of their views with this idea in mind thus provides an agreeable way of elaborating and defending a certain theory of the nature of dispositional mental states. I I propose to begin by summarizing in a few paragraphs a general theory of

1

The general notion of a secondary mental entity is that of one that is “private” and causally dependent on conscious mental states but is not a constituent of conscious mental states. Sensations, emotions, and images are plausible examples. For details, see Addis, 1986.

dispositions that I have defended in detail elsewhere (Addis, 1981a). Patently, an adequate theory of dispositions is essential to an adequate theory of dispositional mental states; and the failure to bring the former to the latter has led, if I am not mistaken, to some faulty thinking concerning dispositional mental states, especially in their role in the explanation of behavior. The general philosophical problem of dispositions arises from the circumstance that dispositional properties are not, as such, observable properties. One cannot look at an object and see its acidity or solubility as one sees its color or its shape. Thus, at least within the empiricist tradition of analysis, any adequate theory of dispositions should provide a schema for the treatment of individual cases of ascriptions of dispositions such that all of the property terms that occur within a specific analysis are themselves—at least within the context—the names of observable properties. The analysis that most naturally suggests itself is the singular material implication analysis, according to which, for example, ‘x is soluble’ “means” ‘If x is put into water, then x dissolves’. The standard objection to this analysis, as everyone knows, is that if the ‘if-then’ is indeed that of material implication, then any object that is not put into water is soluble by reason of a false antecedent, so to speak. Two major responses to this problem have been made within the tradition of empiricist philosophy of science. One is the way of “reduction sentences” adopted by Carnap and Hempel with its accompanying notion of “implicit definition”. This account simply gives up on the attempt to provide explicit analyses of the kind that would allow the elimination, in principle, of dispositional terms from ordered discourse. The other response, of which my view may be considered a variant, is to regard the connective as other than or “more” than material implication. This may be accomplished either by introducing some nontruthfunctional connective or, as I do, by introducing greater complexity into the analysis. In particular, one must explicitly introduce the idea of lawful generality into the analysis of dispositions. The idea of this analysis may now be expressed by considering again the example of ‘x is soluble’. According to the proposed analysis, this sentence(-form) “means” ‘There is a property that x exemplifies and (it is a law that) anything with that property dissolves when put into water’. Using the

222

familiar notation that is so often employed in such matters and letting ‘S’ stand for the property being-soluble, ‘W’ for the property being-put-intowater, ‘D’ for the property dissolves, and ‘f’ be a predicate variable, we get: Sx “means” (∃f) (fx . (y)[(fy . Wy) ⊃ Dy]) The distinguishing features of this analysis are: (1) that it has the peculiarity of using a (bound) predicate variable in the analysis; (2) that it solves the problem of the untested object without either introducing some nontruthfunctional connective or embracing the unfortunate notion of “reduction sentences” but at the price of bringing into the analysis the old problem of distinguishing so-called “accidental” generalities from genuine laws of nature. As to this problem, I have nothing to say here but only assume that the solution has something broadly to do with context, and that in the perspicuous representation of a generality, nothing can (or should) show whether it is accidental or lawful;2 (3) that it requires every disposition to have a ground without formally identifying a dispositional property with its ground but with the implication that any two objects or persons that are absolutely identical in all their non-dispositional properties will of lawful necessity be identical in all their dispositional ones; and (4) that, in effect, for something to have a disposition is for it to be such that it is subject to a certain law. The most important objections to this account of dispositions that must be answered are: (1) that it is beside the point insofar as it apparently says nothing about the property that is represented by the words ‘if placed in water, then dissolves’ (Wx ⊃ Dx); (2) that it fails to provide both necessary and sufficient conditions for the truth of the sentences to which the theory is supposed to apply; (3) that the preanalytic notion of something’s having a disposition does not analytically require that there be any ground of a disposition even if, in empirical fact, every disposition does have a ground; and (4) that this account leads to a vicious infinite regress of properties and laws. I have attempted to dispose of these objections elsewhere (Addis, 2

Thus, to put it in the linguistic mode, ‘is a law’ is on a par with ‘is a tautology’, ‘is true’, ‘is well-formed’, and so on, even if some of these properties can be defined by, or correlated with, syntactical properties of the written or spoken instances of the propositions of which they are properties

223

1981a, 210-214), and shall not repeat my replies here. One should remember, too, that some philosophers, such as Popper, object to the very idea of a significant distinction between dispositional properties and nondispositional properties. But I propose to continue my treatment of dispositional mental states on the basis of this, what I believe to be the only adequate, account of dispositions according to which, to apply the account to humans, to have any disposition, be it a dispositional mental state or only a behavioral or any other kind of disposition, is to have some (nondispositional) property that makes one subject to certain laws of nature. II What needs to be said here about the intrinsic nature of dispositional mental states in general can now be said rather briefly. To have a certain dispositional mental state, such as knowing that e=mc2 or desiring world peace or disapproving of kicking dogs, is to have some non-dispositional property such that one is subject to certain laws. In another paper I have argued that this non-dispositional property can be a property only of the body (Addis, 1981b, 50-51). (Thus, only a being with a body could possibly have dispositional mental states. And this ‘could’ is ontological, not merely causal.) What then makes a certain dispositional state of a person a dispositional mental state given that not all the dispositions we possess are “mental”, not even, perhaps, all of the ones that are realized in complex patterns of behavior (as contrasted with simple, “single-track” dispositions such as being-soluble-in-molten-iron)? The answer, I believe, (and I intend this as broadly descriptive of our linguistic habits rather than analytic in a narrow sense) is that a dispositional mental state is either (1) one that has its realization in certain conscious mental states as well as certain behaviors, or (2) one that has its realization in certain patterns of behavior that are as if they were “guided” by certain conscious mental states and occur in beings that in fact have conscious mental states. Perhaps these two ideas really come down to the same in the end, but my motive in so characterizing the matter is to exclude as having dispositional mental states those artifacts that do exhibit patterns of behavior that are as if they were “guided” by the conscious mental states of perceivings and desirings (in the “occurrent” sense) on the part of the artifact; only a being

224

that has conscious mental states can also have dispositional mental states. What is important is not to be able to specify the exact conditions under which one may correctly ascribe a dispositional mental state to something, but rather to appreciate that in causal fact to have a property that disposes one to behave in certain overt ways is often at the same time to have a property that disposes one to have certain conscious mental states; and that in anthropological (that is, linguistic) fact, Ryle and others to the contrary notwithstanding, we would not consider such dispositions to be mental without their having certain relations, in this way or another, to conscious mental states. It may now be grasped fully what might reasonably be meant by saying that a dispositional mental state is nothing additional to what has already been identified as existing in the physical and mental realms. The only “thing” in addition to the physical properties of the body and the particulars and properties of conscious mental states and the secondary mental entities are the laws themselves. As to the metaphysics of lawfulness, I have nothing to say here, but if it should involve some unique simple entity of a kind one might label as a “necessary connection” then I would have only slightly to amend my claim. In any case, dispositional mental states involve no additional, peculiar mental entities, either simple or complex. It should be interesting and worthwhile to see how this conception of dispositional mental states can be applied to some of the theories of Freud and Chomsky, perhaps most interestingly to Chomsky’s efforts to secure a fundamentally creative or “active” aspect to the human mind. III Freud and Chomsky both appear to maintain that there are mental things (or events or states or structures or processes) that are in no sense bodily but which, at the same time, are not either conscious mental states or secondary mental entities or constituents of either kind. Both thinkers acknowledge or appear to acknowledge the possibility of physiological explanations of these mental phenomena, although Chomsky is somewhat ambiguous on this point. Be that as it may, I have argued that among that which exists and may reasonably be called mental are to be found only (1)

225

the primary mental entities; namely, conscious mental states, (2) the secondary mental entities; namely, sensations, emotions, and perceptionrelated entities such as sense data (if there are any) and images, and (3) the tertiary mental entities; namely, dispositional mental states. My major argument over the next several pages—with several asides—will be that the phenomena that the theories of these two men are designed to account for as well as, and far more important, the explanations of those phenomena embodied in those accounts can be accommodated fully within the framework of these categories even though the ontological sense of that accommodation will be quite different from what either of them had in mind. The theories of both Freud and Chomsky are designed to account, in the first place, for certain aspects of behavior and especially of linguistic behavior; and, in the second place, for certain aspects of conscious mental states. Calling it behavior that is to be accounted for may give quite the wrong impression, for it is not the fact that human organisms move their limbs or even their larynxes and facial muscles in certain ways and emit sounds of certain acoustical properties that is to be explained as such; it is rather, to use some contemporary jargon, those behaviors as actions, that is, as behaviors with meaning that are to be accounted for. This is not the place to develop or to defend any particular theory of what an action is, so conceived. But I would say here, most programmatically, that in the final analysis an action is a behavior that is being considered in its relation to certain actual or possible conscious mental states that, even if they are actual, are not necessarily occurring at the same time as the behavior and are not even necessarily, in total, one’s own.3 So analyzed, there cannot be any a priori or other philosophical reasons why actions should not have full, lawful explanations. At the level of discourse that is appropriate to the theories of Freud and Chomsky, the notion of understanding a language, which is central to their accounts, is entirely taken for granted, and rightly so. In the present context, we may say that to understand a piece of behavior is to know what its purpose is. (In other contexts, understanding behavior need involve no reference to anything mental. One should never argue what it is “really” to 3

For a discussion of this slightly paradoxical idea, see Addis, 1975, 51-53.

226

understand an action or a behavior, but only that certain kinds of understanding, including causal understanding, are always possible—or impossible—in principle.) Hence, to understand a piece of linguistic behavior is to know what its purpose is. But that usually presupposes that one also understands it in another sense; namely, that kind of understanding that is peculiar to understanding a language. In one important kind of case—possibly the most important kind—to understand an utterance is to know what fact would have to obtain in order for the utterance to be true, whether it be true or not. But beyond these few observations we need not go in characterizing the general nature of the phenomena to be accounted for. As to the specific nature of the phenomena to be accounted for we need mention only the most important aspects of human action for each of the two thinkers. In Chomsky’s case, the crucial fact to be explained is that a human being can both create and understand utterances that he has never heard before; a person knows, at least sometimes, whether or not a new string of words is a grammatically correct sentence and, at least sometimes, what fact obtains if the utterance is true. In Freud’s case (and I am concerned here only with his theory of neurosis, not his more general theories of human psychosexual development), the most important facts to be explained have to do with behavior that is often at odds with the purposes of the person whose behavior it is as well as with what much of society regards as normal, and that is generally distressing to the person. (Here I stretch the notion of behavior to cover such things—at one extreme—as Freud’s own neurotic symptom of fainting when he approached his father’s gravesite and—at another extreme--excessive daydreaming, unusual forgetting, fixation, and so on, which are really conscious mental states proper. Here the difference makes no difference.) Beyond these few observations we need not go in characterizing the specific nature of the phenomena to accounted for. It will be helpful for the moment to adopt the terminology of the computer in characterizing the problems. Given that the human “machine” has a certain (kind of) “input” and a certain (kind of) “output”, what is the “machine” itself like on the “inside”? In Freud’s case it would be more nearly accurate to say that all that was given was the “output”—the neurotic symptoms—and that the task he set himself was to discover both

227

the (relevant) “input” and the nature of the “machine”. His theories are in fact about both as everyone knows; and as he changed his mind about the one, he naturally also revised his views about the other. Remember the dramatic effect of his view of the nature of the “machine” from his discovery that the memories of female patients having been violated by their fathers were false memories. In other words, the “input” was not what he had previously supposed, and thus did he arrive at his theory of infantile sexuality. In Chomsky’s case, the nature of the relevant “input” seems to be (and I have no reason to challenge its legitimacy) entirely taken for granted; namely, hearing sentences of the language of the “output” and, of course, being aware of the other immediate circumstances in which the sentences of the “input” are uttered. For the present purpose, the relevant part of their answers to the question of the nature of the “machine” is that its mind has unconscious mental states (for Freud) and that its mind is of the structure of a transformational grammar device (for Chomsky).4 Chomsky also refers to his theory as “mentalism” (and in its epistemological aspect as “rationalism”) and rails against behaviorism by which he means a view according to which, to put it extremely for clear contrast, all is to be explained by the “input” and none by the nature of the “machine”. Freud also resisted behavioristic interpretations of his theories by which he meant nothing special about the relation or ratio of “input” to inner nature but rather what it is to have an unconscious mental state. In short, Chomsky’s anti-behaviorism is really anti-environmentalism while Freud’s is antimaterialism.5 4

I have not thought it necessary to provide references in the writings of Chomsky and Freud to the aspects of their theories discussed here, partly because those theories are so well-known and partly because they pervade their writings. I don’t know that Chomsky ever uses the expression ‘transformational grammar device’, but it seems succinctly to capture the key idea. One could, following Leiber who expresses Chomsky’s view by saying that English is not a “finite-state grammar”, say that the theory is that the human mind has the structure of an “infinite-state machine”, but this expression is less precise. See his very clear and useful discussion of these ideas in Leiber, 1975, especially 72-94. 5 See especially in this connection his chapter called “Psychical Qualities” in Freud, 1970. None of this is to deny, of course, that Freud was a faithful adherent of scientific materialism, believing that everything that happens, including all mental phenomena, has a purely physical explanation. But of course scientific materialism

228

IV It is not my purpose in what follows to cast any doubt on the existence of unconscious mental states in a preanalytic sense or on the idea that a human mind is of a structure that makes it a transformational grammar device. My interest is ontological and therefore only on what may, and not on what in fact does, explain what. Speaking only as an outsider, I find the scientific arguments of both Freud and Chomsky largely convincing.6 But the ontological form in which they have cast their theories and the philosophical consequences they draw from them I find often to be confused and unacceptable. Having now posed my interest in their views with what I hope is both decent clarity and respect, I may now have my philosophical say. Once again, all there is to a human being in this context is a body that is subject to certain laws, conscious mental states, and secondary mental entities. How then do we fit Freud’s unconscious mental states and Chomsky’s transformational grammar devices into such a view? My answer is, in a sense, at once simple and complete: unconscious mental states and transformational grammar devices are dispositional mental states and structures of the body respectively. And because a dispositional mental state is a structure of the body plus a law, we may say, with only slight inaccuracy, that the inner states that Freud and Chomsky invoke in their explanatory accounts are simply states of the body considered as subject to does not entail absolute materialism, the theory that everything that exists, including all mental phenomena, is physical; and it is clear that Freud, at least in his mature thought, was not an absolute materialist. In “Psychical Qualities” a crucial sentence reads: “Nevertheless, if anyone speaks of consciousness we know immediately and from our most personal experience what is meant by it” and he footnotes this sentence with: “One extreme line of thought, exemplified in the American doctrine of behaviorism, thinks it possible to construct a psychology which disregards this fundamental fact!” (Freud, 1970, 14). 6 That is, their arguments that there is something reasonably called the unconscious and that the mind is a transformational grammar device are plausible. I offer not even my layman’s opinion as to the role of either in larger explanatory contexts although both were postulated for the purposes of explanation. In particular, I offer no opinion concerning the role of the unconscious in the etiology of the neuroses, especially since the appearance of Grünbaum, 1984.

229

certain laws. Thus the real content of their theories lies not in the intrinsic nature of these properties (which, because they are bodily in fact, are of interest only to the physiologist) but only in the kinds of laws and “principles” into which these properties enter. It is altogether a mistake to believe that a more careful examination of the body or the mind would reveal the sorts of things that these thinkers are talking about. Let us examine this idea more closely after first considering an objection. Someone may point out, in a somewhat formalistic manner, that to be subject to a certain law is to have the property of being subject to a certain law. What then, it may be asked, is this alleged distinction between having a property of a certain intrinsic nature and being subject to a law? My reply, of course, is that this is precisely the distinction between a nondispositional property and a dispositional property. To know that something is soluble is to know that it is subject to a certain, perhaps only partially specified, law. To know why something is soluble is to know the “ground” of the disposition, probably some molecular structure of the object, as well as the law.7 I do not propose to argue here for a real distinction between non-dispositional and purely dispositional properties of a thing or person, or for the claim that the “real” properties of a thing or person are its non-dispositional properties which properties fully determine all of its dispositional properties. But the reader may, if he chooses, substitute for ‘having a property’ and ‘being subject to a certain law’ the expressions respectively of ‘having a non-dispositional property’ and ‘having a dispositional property’ in what follows. And I too shall sometimes continue to speak of dispositional properties, as in the next paragraph. It is said that some modern computers are so complex that no one person fully understands their workings. More to the point and more certain is the fact that some computers “do” things appropriate to computers for which they were not either wired or programmed to “do” by intention and which, therefore, no one anticipated the possibility of their so “doing”. In short, some modern computers have capacities as computers 7

In his attack on my theory of dispositions in Wilson, 1985, the author seems to confuse the “that” with the “why” to the somewhat startling conclusion that one must know both the ground and the law before one is justified in ascribing a disposition to someone or something.

230

that were not previously known. Having discovered such a case, we might say that the computer is operating by “principles” different from and possibly more subtle than those by which and in mind of which the computer was built and programmed. A computer has no conscious states.8 It is a purely material object. Furthermore, all of its relevant properties as a material object are known except, as in cases of the sort just described, some of its dispositional properties. No one doubts, I trust, that the machine is or, more precisely, is part of a deterministic system. Each state of the computer is, ordinarily, fully determined by its immediately preceding state and its input—electricity and, if any, new “instructions”. In the case of the computer then, it is clear that to understand how the computer “does” things for which it was not built or programmed is to know which of its structural or other non-dispositional properties are the “ground” of the particular capacity and that these properties enter into certain laws in certain ways. The crucial point is that we do not need to postulate any additional properties to those that the computer is already known or assumed to possess. What is “new”, so to speak, are not the properties and the physical structure they determine but only that these hitherto known properties, perhaps in certain previously uncontemplated combinations, enter into perhaps hitherto unknown laws. Freud and Chomsky discovered that human beings have certain dispositions and nothing more.9 To discover that a human being has certain dispositions is to discover that it is such that under certain, perhaps extremely wide-ranging and variable conditions, it acts in certain ways 8

As a general metaphysics of the mind, functionalism can be dismissed immediately. But granted that no conscious state is merely a so-called “functional” state, I do not mean to say that it is lawfully impossible that an artifact could have conscious states even if none, at least around this part of the universe, does. We simply don’t know. 9 As Ryle, 1949, has pointed out, the notion of a disposition actually covers a number of different kinds of characteristics variously called inclinations, traits, capacities, potentialities, strengths and weaknesses, powers, and so on according to various contextual factors having to do with, among other features, their strength, their duration, their scope, their distribution in the species, their importance, and the manner of their acquisition. All of this we may ignore except to note that most of the dispositions that Freud discovered might most naturally be called inclinations whether of the sort common to the species or of the sort acquired by some people only, whereas those Chomsky discovered might most naturally be called capacities. The specific reasons for this need not detain us here.

231

and, sometimes, also that it has certain conscious mental states. Some of these dispositions may be assumed to be universal or nearly so in the species such as the capacity to acquire a language while others, such as the unconscious desire to kill one’s sibling, are not universal (although the second-order capacity to acquire that desire may be). The conditions, at least in their most interesting aspects, may have to do more with the earlier “input” (and its “preservation” in the structure of the body) than with present circumstances, as in the Freudian account of neurosis. But it is of no philosophical significance whether “input” and “output” (which names I always put in double quotes when applied to humans because we are not, or not only, computers) are separated in time or roughly simultaneous or, as is more commonly the case depending on what one takes to be the “original” state of the organism, that the “input” is spread out over time in a way that makes some of it much earlier, some of it only a bit earlier, and some of it nearly simultaneous to the “output” to be explained. So we may now express the main idea by saying that Freud and Chomsky each has discovered that the human body is such that under certain conditions (of “input”), it behaves in certain ways or ranges of ways and/or that it has certain conscious mental states. V Strictly speaking, I have now done my job: shown how, at least verbally as one says, the theories of Freud and Chomsky can be accommodated within a framework in which the only mental things are conscious mental states and secondary mental entities. Yet certain important details of the matter remain due to features that are peculiar to each theory, and the two italicized expressions in the last sentence of the last paragraph call attention to them in Chomsky and Freud respectively. Let us consider Chomsky first. Chomsky has, in his theorizing, emphasized the creative aspects of language use, especially the fact that nearly any person can, at quite a young age and in his native language, produce sentences that have never been uttered before but whose meaning and grammatical correctness (or incorrectness) are recognized instantly by both speaker and listeners. To this claim and this emphasis no reasonable person should have any

232

philosophical or scientific objection. Yet the capacity for creativity and novelty which, as we now see, are fundamental properties of human beings in their linguistic capacities, or a disproportionate stress on that capacity, is, in the minds of many, linked with a non-deterministic view of the world and especially of human choice and behavior. But philosophical libertarian or not, one may resist the notion, implicit in my reconstruction, that what Chomsky really discovered is that human beings are subject to certain laws; for if it doesn’t strongly imply a deterministic view of human thought and behavior, it at least seems to stress being “bound” or “restricted” or “subject” rather than “free” and “creative” and “innovative”. And hasn’t Chomsky shown that human beings are naturally more creative than has previously believed? I have already tried to weaken the force of this seeming implication (or rather connotation) of my reconstruction by my use of ‘range of ways’, But it will be valuable to look at this complaint a little more closely all the same. The rather natural assumption that a person is shown to be only further “bound” rather than, say, having a greater interesting or important capacity by the discovery that certain laws apply to that person is, I submit, nevertheless a mistaken assumption. I submit further that this mistake derives in part from the oft-noted confusion between the logically quite different notions of a law of nature on the one hand and a law of the State on the other.10 But whether this diagnosis is correct or not, it is an error or at least highly misleading to assert that the discovery of more and more laws that are applicable to human beings or to which we are subject, as we revealingly say, further and further diminishes our otherwise presumed creativity or freedom or capacity for innovation.11 For, (1) whether or not some human product—a poem, a musical composition, a scientific theory, 10

Of course, even the latter often increases our political and social freedom when enforced or abided by; and the modern conservative’s complaint that more laws and regulations necessarily further restrict one’s freedom is also a mistake. Even when such laws and regulations do restrict someone’s freedom, they may increase someone else’s. And, of course, when they do restrict everyone’s freedom, it may be desirable to do so. 11 For someone who thinks otherwise, and in connection with Chomsky’s views, see Sampson, 1979, especially the second and third chapters. Sampson connects Chomsky’s views about language acquisition and use with quite specific political and social values as, to be sure, does Chomsky himself.

233

or just a previously unspoken sentence—is regarded as novel or original will ordinarily depend on criteria of description and not of explanation. By that I mean that the characterization of something as novel or as not novel will result from its comparison with previous human products and other naturally occurring items and perhaps the creator’s knowledge of those earlier products and items rather than from its explanation or lack thereof at whatever level or of whatever kind the explanation may be. But, (2) it should now be apparent that in any case the occurrence of a novel human production may be fully explainable. Let us look at this idea more closely. It is customary in the literature to distinguish between being explainable in practice on the one hand and being so only in principle on the other. Yet, for the analysis of the matter at hand, this distinction does not cut quite finely enough. We say that something can be explained or predicted only in principle if, while there are laws and initial conditions that, if they were known, would permit the explanation or prediction of the event in question, in fact are, either or both, unknown. But what about a situation in which, while there are laws and initial conditions that are relevant to the explanation or prediction of some occurrence, the laws that are relevant not only are not but cannot be known prior to the occurrence of the event to be explained? The nature and importance of this possibility will, of course, depend on the meaning of, and the reasons for using, the modal term. The situation I have in mind produces a certain asymmetry between explanation and prediction, but in doing so accommodates the idea of a novel and “unpredictable” occurrence within a deterministic framework. It is very simply that an event that involves significant novelty and creativity—for example, Stravinsky’s composition of Le Sacre du Printemps—is not predictable because, prior to its occurrence, we cannot know the law or laws of which it is almost certainly the single instance. More than that: it would be altogether paradoxical, as has long been recognized, to predict such an occurrence in detail insofar as it would entail doing just that whose first occurrence is being predicted. It no longer matters whether we call this kind of unpredictability that of principle or of practice as long as we grasp completely that this kind of unpredictability does not in the least exclude the possibility of full lawful explanation even in practice and therefore does not exclude the possibility of a deterministic universe. Because what we cannot know in advance are not the relevant

234

conditions but only the relevant laws, and because we now know that under certain, in principle specifiable conditions Stravinsky did compose Le Sacre, we may reasonably believe that were precisely the same (qualitatively) conditions to arise again somewhere in the universe the outcome would be the same. And determinism is the thesis that under precisely the same conditions, whenever and wherever they occur, the result will be the same. All of that emphasizes what cannot be known in advance, at least in the particular case. What can be known is that only certain kinds of beings (ourselves, around this part of the universe) are able to write poems or compose music at all. Furthermore, we know some of the conditions under which this is more likely to happen than not and certain conditions under which it won’t happen at all. If we specify only some of the conditions under which people utter new sentences, to return specifically to Chomsky’s problem, then even if we did know all the laws, we could only reliably predict a certain kind of result (just as we can rather reliably guess of what general character a new composition of a certain composer will be or a newly-found composition of Mozart). Moreover, the actual result may come from any of a number of different prior conditions. (That is, future/past determinism is almost certainly false while past/future determinism is, or at least may well be, true. In that sense, it would be much more nearly accurate to say that relative to the present, it is the past that is “open” and the future that is “closed” rather than, as many would have it, the other way around.) None of this entails, or even makes more likely, that (past/future) determinism is not true; but it should relieve us of any fear that a day may come in which human behavior would be entirely predictable in practice. We don’t even have to worry that the beginnings of speech in a three-year old will be predictable in practice, at least in their detail. It is thus possible reasonably to maintain, consistent with all that “needs” to be preserved about human creativity and capacity for novelty, that to have such powers and capacities is to be such that one is subject to certain laws. We are the beings in whom there can and do arise just those conditions that make for interesting new consequences. I conclude therefore that Chomsky’s discovery is that human beings are subject to certain laws of nature and that their behavior and their conscious mental

235

states are the realizations or consequents of those laws whose antecedents are that human bodies that are in certain physiological states are also in certain external conditions. VI What I call certain properties or structures of the body and the facts that they enter into certain laws whose consequents are about behavior and conscious mental states, Freud calls unconscious mental states. Why? For the points at hand we may consider only the notions of unconscious desire and unconscious memory, and we may say that probably Freud had two basic motives for insisting that these things are irreducibly mental in themselves, so to speak, and not to be analyzed behavioristically or even, as I am doing, as dispositions to act in certain ways and to have certain conscious mental states. Those motives are (1) the somewhat vague idea that the unconscious, while not nearly so subject to change and variation as one’s everyday consciousness, is nevertheless something “dynamic” in which there are constant “goings-on” combined with the assumption that the body is relatively passive and unchanging in the necessary and relevant aspects; and (2) the feeling that terms such as ‘desire’ and ‘memory’ are simply forced on us in characterizing adequately the situations in question, and that their sense cannot be captured fully by treating unconscious mental states as dispositions grounded in the body to behave in certain ways and/or to have certain conscious mental states. The first of these reasons we may dismiss immediately. For, as C. D. Broad pointed out (Broad, 1925, 475) several decades ago in his too rarely read classic, whatever processes are assumed to take place in the unconscious considered as something irreducibly mental can just as well be “carried” by the body, especially if one also assumes, as Freud himself apparently believed, that for every different mental state there is a different physiological state, that is, if one assumes psychophysical parallelism. There are, therefore, lawfully sufficient “goings-on” in the body to account for, and therefore in a sense to constitute, apparent changes in the unconscious over time. As for the second of these reasons, I hope my more detailed comments on unconscious desire and unconscious memory will show that Freud need not have worried on this score either. But I may

236

register my strong agreement with his belief that a purely behavioristic reconstruction of the notion of an unconscious mental state is philosophically inadequate insofar as it ignores the fact that, if they are dispositions, they are dispositions to have certain conscious mental states as well as to behave in certain ways, that the disposition to have certain conscious mental states is what makes them mental in any significant sense, and that conscious mental states do not themselves admit of any philosophically significant behavioristic analysis. None of this of course denies the reasonableness of methodological behaviorism and the necessity, for scientific purposes, of treating unconscious mental states, conscious mental states, and anything else mental as behavior, patterns of behavior, and dispositions to behavior.12 What Freud really discovered is not that human beings have certain previously unknown properties or mental states but rather that certain laws are applicable to them. Both unconscious desire and unconscious memory are dispositional mental states.13 Both therefore are dispositions to exhibit certain behaviors and to have certain conscious mental states. In that respect they do not differ from ordinary memories and desires. For an ordinary memory or desire can be either conscious (or “occurrent” as some say) or else dispositional. As Freud usually spoke, even ordinary dispositional desires or memories that a person may be said to have when asleep but that can easily be brought to mind when the person is awake are said to be “conscious”. So, to avoid a purely terminological squeeze, let us speak of ordinary memories and desires which can be either conscious or dispositional, and unconscious memories and desires. This, or course, already suggests a certain direction to the analysis, but that cannot be helped. Consider desire. An ordinary desire in the dispositional sense is a desire because (1) the behavior that is its realization is “goal-directed” and (2) among the conscious mental states that are also its realization is that of intending to bring about the goal or at least of desiring its existence. Naturally the actual situation is much more complex than this, but as a 12

For detailed discussion, see Addis, 1982, 400-407. Pap, 1959, undertakes a somewhat similar treatment of these and other psychoanalytic notions. Furthermore, and independent of the Freudian context, Pap’s account of dispositions is similar to mine, as I acknowledge in Addis, 1981a.

13

237

schema it will do. The notion of a pattern of behavior or even a single behavior’s being goal-directed can be characterized purely externally, that is, without reference to conscious mental states that must actually occur. This is our cue. An unconscious desire is the disposition to act in certain goal-directed ways with neither the conscious mental states of intending to bring about that goal or the desire that it come about nor the ordinary disposition to have either or both of those conscious mental states. One may say then that the person’s behavior is as if he intended to bring about the goal; thus reference is made to a kind of conscious mental state but not one that actually occurs. Of course, among the realizations of an unconscious desire may be quite powerful conscious desires, even of the “craving” sort. In general, an unconscious mental state is the disposition to have all sorts of conscious mental states, but the mode property and the intentional property of no one of them can both be the same as those ascribed, analogically, to the unconscious mental state.14 The crucial point here is that what makes an unconscious mental state a desire is not at all the intrinsic nature of the property or state that, as it were, persists in the person who has it and which is something purely physiological, but rather the kinds of behavior and conscious mental states that are its realization. Hence, to affirm that humans are the sort of beings who have unconscious desires is to say that their bodies are structured in a way whatever it is that under certain conditions, they act in certain ways and have certain conscious mental states. Now consider memory. An ordinary memory, in contrast to a remembering, is a disposition. It is the disposition to have the conscious mental state of remembering, under certain conditions. That is, a person with an ordinary memory of so-and-so is one whose body is structured in a way such that under certain conditions that person has a conscious mental state of remembering so-and-so. What then is an unconscious memory? At the level of description I just used in characterizing ordinary memory, there 14

The mode property is that which makes it the kind of mental state it is: a perceiving or an imagining or whatever. The intentional property is that which specifies what it is of: that my grandmother wore glasses or that 2+2=4 or whatever. Such properties, and therefore intentionality itself, pertain only to conscious mental states. Dispositional mental states are not, therefore, literally intentional states. See the last section of this paper for further discussion of this idea.

238

is no difference between ordinary memory and unconscious memory. The difference then can lie only in the nature of the conditions that must obtain in order that the relevant conscious mental states occur. This, in my judgment, is exactly as it should be, for it suggests what I believe phenomenology confirms: that there is no sharp break but only a continuum between ordinary and unconscious memories while being entirely consistent too with the fact that the unconscious may have memories of occurrences that ordinary memory not only doesn’t have but contradicts. (I always use ‘memory’ and ‘remembering’ as Freud did, so that there may be either true or false memories and rememberings.) Thus, a certain conscious mental state of remembering may occur only under the condition of hypnosis or extended free association; and having once occurred may, variously, puzzle the person who has it by being in contradiction to other conscious states of remembering that the person has had and possibly continues to have, or be of such force that the person now easily sees that it but not the other rememberings that contradict it is probably veridical (although “force” is only one evidence of veridicality and frequently overridden by other evidence), or continue to reject this remembering as veridical (and possibly be correct in doing so, because an unconscious memory need not be veridical either). Of course, the analyst, if there is one, may have reason to believe that the person has an unconscious memory of some event even if hypnosis and/or extended free association or any other technique used for some time fail to evoke it. In this case, we may interpret the analyst’s claim as either (1) that there is in fact some condition under which this person would have the conscious mental state of remembering that event even though he, the analyst, has not succeeded in bringing it about, perhaps by choice; or (2) that this person in historical fact witnessed that event (in which case the analyst is presupposing that the “memory” is veridical) and, possibly, that the person’s witnessing the event left some permanent effect on his behavior and everyday mental life. At all events, we can now see how, consistent with all the data and in sufficient harmony with the ways the words are actually used, we may say that an unconscious memory, like all unconscious mental states, has to do not with the intrinsic nature of some property or state of a person—for the only relevant ones, and that only in a causal sense, are purely

239

physiological—but rather with the kinds of conscious mental states and, to a lesser degree, behavior that would occur under certain conditions. Hence, to affirm that human beings are the sort of beings who have unconscious memories is to say that their bodies are structured in a way whatever it is that under certain conditions, they have certain conscious mental states. It may reasonably be asked why I do not give serious consideration to the view that what we call an unconscious mental state is simply what I call a conscious mental state of which one not only is not but cannot, causally, be aware. On this account, all the mental states of dogs, for example, would be unconscious mental states because dogs are incapable of being aware of their own awarenesses. This is the view of Sartre, I believe; but then is it not really, after all, also the view of Freud himself? (I have never been able to discern any more than a terminological difference in Freud’s distinction between the conscious and the unconscious on the one hand and Sartre’s corresponding distinction between knowledge and consciousness on the other despite Sartre’s vigorous opposition to Freud’s notion of the unconscious.) It is also a view that Broad once considered at some length (Broad, 1925, Section C) but eventually rejected as making a superfluous assumption given the existence of the relevant physiological processes in any case. That is one reason, and perhaps a sufficient reason, for rejecting the position. But it is, as it were, merely one of the Razor. Another reason goes deeper, and it has to do with the connection of the mental in the primary sense with time. But this matter I have addressed elsewhere.15 VII A dispositional mental state, be it a belief or a desire or a capacity of a being who is, in part, a transformational grammar device, is a disposition to behave in certain ways and to have certain conscious mental states. According to the theory of dispositions held here, such a human disposition is a property of a person that enters into certain laws. Because such a property is necessarily a physiological one, intentionality does not and 15

See Addis, 1981b, in which it is argued that the properties that characterize conscious mental states are ones that can be exemplified only in a temporal universe.

240

cannot pertain literally to dispositional mental states even though the notion of such a state analytically involves that of intentionality. Let me explain briefly. On my view, intentionality literally obtains only when there is an actual awareness, when it is correct to say that a person is conscious of something. Thus a person who is unconscious or in dreamless sleep and thus conscious of nothing at all does not literally exemplify intentionality, whatever dispositional mental states may truly be ascribed to that person. On a view I have defended in another paper (Addis, 1983), to be in possession of intentionality in the strict sense is to exemplify a property of a certain special kind—what I, following William of Ockham, call a natural sign, a natural sign being any entity that intrinsically means or intends or represents that of which one is aware. Beliefs and desires and other dispositional mental states are naturally said also to have their objects, to be about something, to mean or intend or represent, and so on. But these ways of speaking, it would now appear, do not express literal truths; rather, the ascription is analogical. One could say that intentionality pertains to dispositional mental states only derivatively and to conscious mental states intrinsically, but this way of speaking invites confusion with the radically different way in which intentionality pertains to mind intrinsically and to language only derivatively. This will be evident if we say, despite the difference in levels of generality the example imposes, that intentionality is related to a given dispositional mental state in the way in which dissolving is related to the disposition of being-soluble. Just as something can be soluble without dissolving, so someone can possess a certain dispositional mental state without having the particular conscious mental states that are its partial realization and without, therefore, ever exemplifying the forms of intentionality that are particular to, and even enter into the notion of, that particular dispositional mental state. The objects of one’s dispositional mental states are, therefore, only the objects of which one would be aware in certain conscious mental states under certain conditions. That, precisely is the sense in which intentionality pertains to dispositional mental states only derivatively or analogically and

241

not intrinsically or literally.16

16

In Searle, 1983, the author, like many others, treats beliefs as paradigmatic of intentional states. I have just explained why I think that is, at the least, misleading. But then Searle does say that he is dealing with the “logical” aspects of intentionality as contrasted with its metaphysical ones. See especially the first chapter of the book.

242

REVIEW OF LAURENCE D. SMITH’S BEHAVIORISM AND LOGICAL POSITIVISM: A REASSESSMENT OF THE ALLIANCE

T

he connections that existed between two of the century’s more important intellectual movements, behaviorism and logical positivism, are on the received view relatively unproblematic: it was a partnership in which psychologists received advice or even instruction from philosophers on how to go about their business—in theory construction and methodology no less than in specific research strategies. Even though behaviorism existed in name and deed many years before the official founding of logical positivism, it was to receive its justification only through the philosophical developments emanating from the Vienna Circle—so much so that behaviorism and logical positivism stand or, rather, have fallen together. According to some, the failure of behaviorism in psychology has demonstrated the inadequacies of logical positivism; according to others, the demise of logical positivism shows the absurdity of behaviorism. Apart from the logical connections of ideas (although these are, in the end, the heart of the matter), there were close personal, academic, and other social ties that constituted the historical association between the representatives of these two movements that for a time so dominated their respective disciplines, at least in the United States, that they seemed to many to be psychology itself and philosophy itself—now and forevermore. In his superb new book, Laurence D. Smith, a psychologist at the University of Maine, shows through a wealth of historical detail and sophisticated philosophical considerations that the received view of the alliance, as just described, is, while not entirely mistaken, seriously defective in various ways. Succinctly put, while there was of course some kind of alliance, neither the historical influences nor the logical connections of ideas were, or are, nearly as intimate as the received view would have it. Smith’s book is structured as follows: In an introductory chapter in

which he uses the work of Sigmund Koch as his major foil,1 Smith provides accounts of both the received view and his revised view as well as indicating the nature of his method and evidence. The second chapter is a useful and generally accurate summary of the logical positivist view of science. The essential argument of the book is developed in the third through the ninth chapters in which the work and thought of each of the three most important behaviorists who had anything at all to do with logical positivism—Edward Tolman, Clark Hull, and B. F. Skinner—are separately treated. These treatments are undertaken through examination and analysis not only of each psychologist’s published works but also— and it is in these regards, Smith says, that his study is importantly different from, and more reliable than, previous ones—through his unpublished works, letters and other archival materials, and correspondence and interviews with members and students of members of both movements. A final chapter serves as summary and conclusion while also indicating in very sketchy form the author’s sympathy with the “new” philosophy of science and what he calls the “reemergence of psychologism” (Smith, 1986, 324). I shall proceed by summarizing the historical argument first and then turn to the more strictly philosophical arena where the few criticisms I have of Smith’s important achievement are to be made. It should be said at the outset, however, that it is the historical side that, properly, dominates Smith’s narrative; but, as Smith’s own account well illustrates, the detailed exposition of an important thinker’s ideas can never be merely historical I Smith’s treatment of Tolman, Hull and Skinner appears to show, and to this reviewer’s satisfaction does show, in each case, first, that the psychologist under discussion came to his behaviorism independently of, and prior to, his association with logical positivism; second, that this association was neither deep nor lengthy; and third, that in any case his conception of behaviorism, especially as it might be applied to theories of 1

The main work of Koch that he cites is Koch, 1964. A second foil derives from MacKenzie, 1977.

244

knowledge, was in crucial respects antithetical to basic tenets of logical positivism. It appears that the only personal association between major representatives of the two movements that was of significant duration and intensity, and that involved genuine mutual understanding and fertility took place at the University of Iowa in the late 1930s and early 1940s between Gustav Bergmann, a member of the Vienna Circle, and Kenneth Spence, a student of Hull’s. Otherwise, Smith says, the alliance “was generally much more limited in scope than is commonly supposed” (1986, 18), although the articles co-authored by Bergmann and Spence are cited by Smith as one of three major sources of the belief, more widespread among psychologists than philosophers, in a closer association of behaviorism and logical positivism than actually ever existed. The other two sources, according to Smith, are (1) the fact that the eminent historian of psychology E. G. Boring and his student S. S. Stevens “conspicuously identified behaviorism with logical positivism” (1986, 317), and (2) published statements by the philosophers H. Feigl and C. G. Hempel that inadvertently and mistakenly gave an impression of significant influence of the ideas of Carnap on Tolman and other psychologists. But in fact neither Tolman nor Hull nor Skinner was much influenced by Carnap or, for that matter, by any other member of the logical positivist movement. Tolman had early rejected the “reflex” or “molecular” behaviorism of Pavlov, Thorndike, and Watson, but arrived at his “molar” behaviorism (an expression he took from the philosopher D. C. Williams) long before he had any association with logical positivism. Part of this idea of taking the “molar” rather than the “molecular” as the proper unit of analysis was that behavior inherently involves purpose and cognition; and under the influence of the neorealists, especially Holt and Perry, Tolman at first believed that at least these crucial aspects of mind were directly observable. His studies of rats in mazes, which were to become in the minds of many the hallmark of behaviorism, were originally intended in part to demonstrate this belief. But under the influence of the Chinese behaviorist Z. Y. Kuo, Tolman came to reject this view and instead to adopt the notion of cognitive concepts as “intervening variables,” that is, as unobservable factors of a certain kind standing “between” the observable independent variables and the dependent variable of observable behavior. (Of this fundamental matter of the unobservable I shall say more later.)

245

The problem of how to understand unobservable variables was to become the source of Tolman’s main link with logical positivism, but he had in fact worked out his own version of operationism some years prior to his association with that movement. Having met Schlick at Berkeley in the early 1930s, he spent his sabbatical of 1933-34 in Vienna talking with members of the Circle and apparently attending some of their meetings. From these contacts and others, Tolman adopted the idea of the ineffability of immediate experience, with its correlative thesis, as developed mainly by Schlick, that psychology could therefore deal only with the structure and never the content of experience. Tolman was also for a while impressed by the confused notion of “probabilistic functions” that he got from Brunswick and Reichenbach; and all of these connections and interests led to his being invited to, and accepting, a role in the Unity of Science movement which was, following the dispersion of the Circle, the institutional heart of logical positivism. But despite all this, Smith tells us, Tolman’s “view of science as an uninhibited operation of exploration and discovery made little room for the careful logical distinctions that were characteristic of the logical positivist view” (1986, 143). His view of science, like those of Hull and Skinner, was, as Smith puts it, “psychological” rather than “logical” and he tended, also like the others, to treat even traditional philosophical issues concerning truth, theory, and knowledge as psychological ones in a way that both derived from certain pragmatist influences and anticipated present-day “causal” and “naturalized” theories of this or that. As for logic itself, once more like Hull and Skinner, Tolman tended to the radical empiricist views of Mill—a position directly at odds with a fundamental tenet of logical positivism. From these and many other considerations, Smith is able to conclude that “whatever influence logical positivism had on Tolman’s thought, it was more a corroborative than formative nature” (1986, 130) and that “Tolman did have sympathies with logical positivism but those sympathies were limited by the fact that he viewed science from the perspective of his own psychology” (1986, 131). Hull’s behaviorism and views of science were probably closer to logical positivism than Tolman’s or Skinner’s but even in his case, Smith says, that while he “adopted snippets of the jargon of logical positivism . . . he was in no way a serious student of its philosophical stances and appeared

246

not even to follow important developments in its basic doctrines” (1986, 195). All the same, Hull did nothing to discourage the appearance of a close association and, as has already been remarked, his student Spence did in fact have such an association, personally and intellectually, with Bergmann, the results of which association naturally often got back to Hull. Hull himself came to his behaviorism in the 1920s partly through the influence of Watson whose views, however, he considered seriously flawed. One such flaw was in Watson’s treatment of consciousness itself which Hull believed had to be taken seriously by the psychologist though never in an explanatory role. More metaphysically minded than the other behaviorists and indeed than the logical positivists at least took themselves to be, Hull considered himself a materialist, a mechanist, and a determinist. Of these, his mechanism—the human as machine—was to be the most important in shaping his research programs as well as his later views; and it was connected in his mind with that for which he was best known as a psychologist—his emphasis on theory and formal rigor; in short, his “deductivism”. His materialism supported a belief in the unity of method for all the sciences as well as a commitment to a kind of “objectivity” in psychology that presumably already existed in the physical sciences. All of this put him clearly on the side of the angels as far as the logical positivists were concerned even if, as Smith points out, Hull’s positivism was really much closer to that of Comte than that of the Circle (1986, 221). Furthermore, these views were largely developed prior to any contact Hull had with logical positivism as were his views regarding the symmetry of explanation and prediction and a distinction between the context of discovery and the context of justification—both to be cardinal features of the logical positivists’ view of science. Like Tolman, Hull was challenged by the issue of intervening variables and their definition, a challenge which led to his most sustained contact with any member of the logical positivist movement in the person of the British biologist and logician J. H. Woodger in early 1938. (Other members of the movement, including Neurath, Bergmann, and Arne Naess, visited Hull at his Institute of Human Relations at Yale.) But the relationship with Woodger, like Hull’s earlier association with the Unity of Science movement, yielded little in concrete results for either party. This interaction and its consequences Smith sees as “representative of the

247

relationship between behaviorism and logical positivism” (1986, 204). And again like Tolman, when it came to specific views regarding the nature of knowledge, truth, and especially logic (“Logical and mathematical theory is nothing but certain habits.... Thus computation is but a series of stimulus response combinations.”2) Hull’s views were in full opposition to those of the logical positivists. Smith concludes that “there was sympathy with the scientistic spirit of logical positivism and its rhetoric, as well as agreement on broad issues regarding the nature of science, but there was very little in the way of genuine intellectual exchange, mutual understanding, or common scholarly interests” (1986, 227). In some respects, Skinner’s differences with logical positivism were the greatest of the three, even if he himself sometimes ignored the differences or regarded them as only minor as he tended to do in the 1930s. An acquaintance of Carnap’s, friend and colleague of Feigl’s at Minnesota and later of Quine’s at Harvard, Skinner had come to his behaviorism originally through reading Bertrand Russell’s account of the ideas of Watson and Pavlov in his little-known book called simply Philosophy. But Skinner’s basic conception of science came from Mach and was, as Smith puts it, Baconian rather than Newtonian (1986, 259), in its generally antimethodological, anti-theoretical bent. From his anti-formalism (logic as empirical) to his notion of explanation as description (also from Mach); from his unorthodox not to say entirely original operationism to his curious view that laws of nature are really descriptions of the behavior of scientists (1986, 291), Skinner was in fact at severe odds with logical positivism, and in the 1940s and l950s these differences were to become clearer to all parties. A paper of his from 1945 was even characterized by one critic as “a forceful indictment of logical positivism” (quoted by Smith, 1986, 286), and the well-known (and widely-criticized) Verbal Behavior of 1957, by presenting a detailed account of his “behaviorist epistemology” as applied to scientists themselves, showed how far removed Skinner was from the logical positivist notions of a logic of justification and its distinction from the “logic” of discovery among other matters. As Smith puts it, “Skinner’s earlier limited sympathy with logical positivism was due to its limited 2

Hull as quoted by Smith, 1986, 245-246. Smith reports (1986, 254) that Bergmann learned “better than to discuss [such matters of logic] with Hull”.

248

overlap with his Machian positivism; but this same Machian positivism called for the development of a descriptive epistemology and thus also led to his antipathy to logical positivism” (1986, 287). Of course Skinner’s behaviorism was by now itself at considerable remove from the Watsonian reflexology with which he began; and his crucial concept of the operant embodied in what was necessarily a form of molar behaviorism, was bound to be viewed with suspicion by the logical positivists who were always much more sympathetic to, if not simply conceiving real behaviorism as, molecular behaviorism. In short, Skinner’s positivism was that of the psychologist and not the philosopher. “Far from being a logical positivism, Skinner’s positivism is grounded in biological expedience. It arises from a biological conception of organismic behavior, the same conception that has guided his empirical research on the process of adaptation by operant conditioning” (1986, 275). So it is a mistake to hold (1) that any of the major behaviorists in psychology came to his behaviorism as a result of his association with logical positivism (and even Spence got his behaviorism from Hull, although it was, to be sure, greatly reinforced and somewhat modified through the association with Bergmann); or (2) that there ever were, with the single exception just alluded to, sustained and fruitful associations between major representatives of the two movements; or (3) that the thought or the research policies of the behaviorists were significantly shaped by their associations with or other influence from the logical positivists. All of this would seem to make it unlikely (4) that behaviorism in psychology gets its justification only from the philosophy of logical positivism and that the two, consequently, stand or fall together. But this point, being logical and philosophical rather than historical, is not so readily established. Nor does Smith address this point as clearly as one might wish. His haziness on this point and the general confusion that persists on this matter can be seen, in one sense, as deriving from not getting sufficiently clear on the notion of the unobservable and especially in failing adequately to distinguish its several relevant senses for the issues at hand. For it is in this crucial notion, I believe, that the logical locus of the connection between behaviorism and logical positivism lies and not, as Smith and many of the subjects of his book—especially the behaviorists— seem to believe, in general conceptions of theory or the nature of science

249

or logic or even epistemology none of which, at least in most aspects, has much to do with psychology as distinguished from the other sciences. Let us see. II Behaviorism in psychology, that is, methodological behaviorism, began as a reaction to the almost complete failure of the introspectionist, associationist schools to come up with any consistent and secure results in what it took to be its tasks—to identify the kinds of elements of conscious states and the connections (“associations”) between and among those elements. The functionalists began the move away from these approaches, but to Watson goes the merit of having first given relatively clear formulation of the two core ideas of methodological behaviorism. The less important, semi-normative, and almost systematically dispensable core idea is that it is behavior and not consciousness that is the proper subject matter of scientific psychology. Given a few, relatively trivial premises this less important idea follows from the more important core idea that in the causal explanation of behavior, one may ignore the realm of the mental: a complete set of relevant, independent variables for the lawful explanation of behavior exists among the physical variables alone.3 This assumption, that the mental is not needed in the explanation of behavior (which is not the same as saying either that it is non-existent or that it cannot properly be invoked in the explanation of behavior), is itself a broadly empirical 3

Watson’s formulation of this crucial idea in his seminal paper, “Psychology as the Behaviorist Views It” goes as follows: “It is granted that the behavior of animals can be investigated without appeal to consciousness. Heretofore the viewpoint has been taken that such data have value only in so far as they can be interpreted by analogy in terms of consciousness. The position is taken here that the behavior of man and the behavior of animals must be considered on the same plane; as being equally essential to a general understanding of behavior. It can dispense with consciousness in a psychological sense” (1913, 176). There is also in Watson and later in Skinner an unfortunate emphasis on the prediction and control of behavior in the characterization of behaviorism. For reasons that are set out in Addis, 1982, 403, I believe that this emphasis is a mistake. (In that essay I also argue that it is better to leave the idea of the proper subject matter of psychology out of the formal characterization of behaviorism.) The reasons given there are of an analytic nature; but one may, as many have, also question the strategic and even the moral value of this emphasis.

250

hypothesis. For if interactionism (the mental is distinct from the physical to which it is connected systematically by laws of succession) were true, then the mental would be needed in the full, lawful explanation of some behaviors. The methodological behaviorist is, therefore, committed to either materialism (the mental is either nonexistent or merely an aspect of the physical) or parallelism (the mental is distinct from the physical to which it is connected systematically by laws of coexistence) or fatalism (the mental is distinct from the physical but lawfully unrelated to it) each of which possibilities allows that anything physical, including behavior, admits of purely physical explanation. The focus on behavior, and the broad empirical hypothesis which serves as a methodological principle about the explanation of behavior, derive both historically and systematically from a simple fact: the unobservability of the mental. And while one should agree with the behaviorists that, given this fact, the focus on behavior is entirely proper, the introduction of that fact into psychology had two unfortunate consequences, one in psychology and the other in philosophy. Within behaviorist psychology, the fear of dealing with the unobservable resulted in an entirely misplaced fear of invoking, at least in principle, the physiological in the explanation of behavior. Of course, this shrinking was partly a matter of research strategy and reasonable division of labor, and in any case was neither consistent nor universal; but the reluctance on the part of many of the behaviorists to say straight out that the “gaps” in purely “environmental” explanations of (differences in) behavior would have to be filled by physiological variables not only detracted unnecessarily from the very good sense in their research programs but gave rise to a number of confused discussions under the eventual broad heading of “intervening variables” versus “hypothetical constructs”4 about “defining” the “unobservable” variables. Meanwhile, in philosophy, it should have been realized that the unobservability of private mental states was only unobservability-byothers; that is, that while conscious states are not publicly observable, they are, in the relevant sense, sometimes privately observable; that the mental 4

As perhaps everyone knows by now and as noted by Smith, 1986, 118, this distinction was standardized only well into the discussion in an influential essay by MacCorquodale and Meehl , 1948.

251

variables constitute a realm of properties distinct from that of the physical; and that the mental variables might have been, even if in almost certain fact they are not, an interacting set of variables with the physical. Had all this been realized, there would have been little temptation to suppose that, in any philosophically interesting sense, the mental variables can all be literally defined through the physical (behavioral) variables; little temptation, that is, to believe that logical behaviorism is true. In fact, logical behaviorism is false as either an anthropological thesis à la Ryle (“This is what, know it or not, we really do mean by . . .) or as a philosophical thesis as advocated by the logical positivists (“This is what, given good empiricist principles, we only could mean by . . .). It was not the empiricist principles that were at fault but rather their misapplication in the failure to realize that introspection, however limited or even nonexistent its value is for scientific psychology and despite the buffeting it has taken from certain philosophers in the last few decades, is also an everyday mode of awareness alongside the outer senses by which we discover certain of the properties of nature. (The anthropological thesis, whether called logical behaviorism or functionalism as in its most recent incarnation, one may dismiss as absurd on its face—and also on further analysis.) We are now in a position to appreciate two crucial points that, had they been fully grasped by Smith, might possibly have changed the emphasis of some of his arguments and conclusions. First, despite the various associations, connections, and expressions of mutual admiration and support that did exist between representatives of behaviorism and logical positivism but in accord with Smith’s general thesis, methodological behaviorism and logical behaviorism are really quite logically distinct hypotheses. If some psychologists in fact embraced or thought they were embracing logical behaviorism, the fact remains that as behaviorist psychologists they were committed to nothing more than methodological behaviorism which came to be and still is the controlling methodological idea in scientific psychology. Thus at this crucial locus at which both the psychologist and the philosopher relied on the idea of the unobservable, the alliance was unstable even so, to the extent that the psychologist’s thesis was fundamentally one about what causes what while the philosopher’s was one about the actual or possible meanings of words.

252

But, second, and to develop the point just alluded to by italics: contrary to Smith and most others, the demise of logical positivism has not been the cause of, or the effect of, or even the accompaniment of the demise of behaviorism in psychology for the simple reason that methodological behaviorism has not only not found its demise; rather, its triumph in psychology has been so nearly total as to render itself almost trivial and imperceptible. To see this, consider the most important case. It is widely believed that cognitive psychology has today almost the status that behaviorism had in the past and, more important, that it has virtually supplanted, or greatly diminished the influence of, behaviorism. But this is a mistake.5 Far from supplanting behaviorism, cognitive psychology is itself done in the very spirit of methodological behaviorism as I have characterized it. What cognitive psychology is not, and has in significant measure supplanted, is conditioning theory: and while conditioning theory and cognitivism are certainly at odds in the relative importance each ascribes, broadly and very succinctly speaking, to the environmental and the biological variables, cognitivism no less than conditioning theory proceeds on the assumptions (1) that it is behavior that is to be explained, and (2) that the search for the relevant independent variables is undertaken on the basis of what can be observed. As Spiker puts it, the cognitivists “do indeed use patterns of stimulation and behavior in order to decide whether the subject has this cognitive state, is using that cognitive strategy, or is undergoing this cognitive process” (1977, 98). In any case there is no systematic appeal to the subjects’ descriptions of their inner states, and in that crucial respect one must agree with Spiker that it is a mistake to suppose that “these superficial similarities signal a return of psychology to the mentalistic cognitivism of classical structural psychology” (1977, 97). In short, the dispute now is only that of which among the physical variables are the relevant or the most important ones in the explanation of certain kinds of behavior. One might similarly wonder whether or not logical positivism, and more precisely logical behaviorism, is really dead. It is certainly true that while many particular philosophical analyses continue to be done in the 5

On this point, see the excellent essay by Spiker, 1977, from which some of the following material originates.

253

spirit of logical positivism, it would be difficult to find any philosopher today with a commitment to all of the original, major theses of logical positivism. But more to the point I think is the fact that logical behaviorism as such is dead. I put it this way because the matter is complicated by the fact that a very close cousin of logical behaviorism, functionalism, is, regrettably, the latest rage in contemporary philosophy of mind. Yet, by a curious and almost ironic twist, especially for those who falsely imagine that cognitive psychology is the logical enemy of methodological behaviorism, there exists today an alliance between cognitivism in psychology and functionalism in philosophy that is somewhat akin to that that existed, or was thought to exist, between behaviorism and logical positivism. But that is another story. If Smith has not always seen to the very logical heart of the issues of his story, if he has sometimes been too ready to take certain side issues or certain particular theses as essential to logical positivism,6 and if he has sometimes missed certain historical connections,7 still the achievement is impressive. Behaviorism and Logical Positivism will surely stand for the foreseeable future as the definitive account of its subject.

6

I think in particular of his taking Carnap’s and Hempel’s theory of so-called “reduction “ sentences as the logical positivist position on dispositions. 7 For example, his discussion of “objective reference” in Tolman seems to be noncognizant of the connection of the subject and its name with Brentano.

254

HUMAN ACTION AND THE HUMEAN UNIVERSE1

A

Humean universe is one in which all properties are only externally connected. Neither the full explication of this idea nor the defense of the thesis that the actual universe is a Humean one is my present task. Instead, I want to examine and refute one argument, or cluster of arguments, that the actual universe is not a Humean one based on the existence in it of human action. For it is the contention of several philosophers, mostly in the tradition of the later Wittgenstein, that because the connection between beliefs and the actions that are to be explained by them is not external but instead internal or logical or conceptual, ours is not a Humean universe. And some of these philosophers have drawn the further natural conclusion that some or all good explanations in the human sciences are not causal or lawful but instead logical or a priori. Maclntyre and Winch are two examples of such philosophers. It will already be evident that there are a number of overlapping issues involved in this topic allowing for a large number of variations of view; but it will be useful, I judge, to simplify by characterizing very broadly two contrasting metaphysics of internality/externality matters and causation/lawfulness matters as follows: there is the thesis that at the basic ontological level all properties are only externally connected and that causation is just lawful connection between and among (exemplifications) of properties, lawful connection itself being a species of external connection; and there is the thesis that at the basic ontological level at least some properties are internally or logically connected and that causation is either some kind of necessary (not external) connection that can, however, be expressed in laws (that is, causation is lawful connection, but laws express necessary and not just “regular” connections) or some kind of connection that can hold in the particular case independent of lawful 1

A shorter version of this paper was presented at the 9th International Congress of Logic, Methodology and Philosophy of Science held in Uppsala, Sweden in August 1991.

connection. It would take at least a long book to explore these possibilities thoroughly and, as I said at the outset, my purpose here is much more greatly limited. But it may be useful to say that I do not hesitate to align myself with the former of these metaphysics. I shall proceed as follows: first, I shall summarize MacIntyre’s arguments by use of his own main example; second, I shall elaborate somewhat on the idea of the Humean universe with its crucial distinction between simple and complex properties; third, I shall sketch an account I have developed elsewhere in detail of dispositions, dispositional explanations, and the relation of the latter to lawful explanations; fourth, I shall argue that explanations of actions by beliefs are a species of dispositional explanation and that being such they can be analyzed in a way that enables us to see that human actions and their explanations are compatible with a Humean universe; fifth, I shall say a few words about the connection of such explanations to explanation in the human sciences; and sixth, I shall state my main conclusions. I In his article, “A Mistake About Causality in Social Science”, MacIntyre makes the correct observation that if beliefs and actions were only externally connected or, as he puts it, “stand in a causal relationship (as that is understood by Hume)” (1967, 50), then any beliefs might go, or might have gone, with any actions. MacIntyre also gives this issue the unfortunate formulation of being the question whether or not beliefs and actions are “separate phenomena” (1967, 51); and it may be well to be clear at the outset that the question before us cannot reasonably be that of whether or not beliefs and actions, even beliefs and actions so connected that the beliefs explain the actions, are such that the one can exist without the other. It will become clear a few paragraphs hence that beliefs can exist without “their” actions and later that actions can exist without any particular beliefs as well as the precise reason why this is so. In any case, it is false that just any beliefs might “go with” just any actions in any social world that would be intelligible to us, and MacIntyre’s example from Weber illustrates this point very well.

256

Weber, as we all know, in his classic The Protestant Ethic and the Spirit of Capitalism argued that something he called “the spirit of capitalism” is a natural consequence of Christian belief in its Protestant form. And while Weber is not perfectly clear on the matter, it appears that he regards this connection as a causal one which MacIntyre, in turn, takes to mean an external one. For MacIntyre’s point, it doesn’t really matter exactly how we understand Weber on these points. Furthermore, we may agree with MacIntyre when he claims that Weber’s distinction between the “spirit” of capitalism and capitalist activities is, if not just spurious, in any case irrelevant; what we are really interested in is Protestant beliefs and capitalist actions. Now MacIntyre’s thesis is that the connection Weber uncovered between Protestant beliefs and capitalist actions is not a causal or any kind of external one, but one of a sort he in one place calls “rational” (1967, 55). Thus we are invited to think of this connection as an instance of what is known as the practical syllogism in which beliefs function as the premises and actions as the conclusion. Somewhat more precisely, by Weber’s account it is the peculiarly Calvinist form of Protestantism combined with a certain desire (which desire, however, can also be expressed as a premise) that “entails” capitalist actions, something as follows (this is my own formulation): Beliefs: Whether I go to heaven or hell is predestined by God, and nothing I do in my earthly life can affect God’s decision. My performance of “good works,” as required by God, would be a sign that I am going to heaven. Desire: I want to know before I die where I’m going after I die. Therefore, Actions: “Good Works” such as amassing capital, building factories, hiring workers, and so on. MacIntyre notes that Weber points out that Calvin himself denied the second belief-premise and so the satisfiability of the desire (at least in that way), and that it was up to Beza to modify Calvinism by introducing that belief-premise as something to be accorded adherence by the faithful (1967, 55). Calvin apparently had taught that God did command people to

257

perform good works, that it is of the highest importance to do what God commands, but that doing so is entirely irrelevant as either cause or sign of what is most important to believers—salvation or damnation. MacIntyre says that Calvin’s doctrine was logically at fault; this is not evident to me but it is certainly psychologically unstable—or would be, if it were not for the context. For in the context of supernatural religion and its psychological sources, the usual demand of the ego for consistency is greatly diminished by the power of the wishes at work. But whether or not the context is that of religious belief, we must keep in mind the following: As I have just implied, in order for the “entailment” and therefore the corresponding explanation to be plausible, one must add at least both that the person in question is rational in the relevant sense and that that person happens to see the connection between premises and “conclusion.” For it does happen that a person has the relevant beliefs and desires, knows the connections between them and the actions they “entail,” but fails to act accordingly out of irrationality. It also happens that a person has the relevant beliefs and desires, is rational in the sense of being prepared to act in ways that those beliefs and desires “entail”, but fails to know what they “entail.” It further happens that a person has the relevant beliefs and desires, is rational, knows what those beliefs and desires “entail” but fails to act not because of irrationality but because of having overriding reasons. To take the example before us, our “rational” Calvinist might also have the belief that becoming a capitalist would threaten the welfare of his children and the desire that the welfare of his children not be threatened. Finally, as I shall show later, it happens that a person has the relevant beliefs and desires, is rational, acts in the ways “entailed” by those beliefs and desires, but the explanation of the actions lies elsewhere. But despite these many complications that MacIntyre (not to mention Weber) largely overlooks, I am prepared to agree with him that in the case at hand at least, “the relation of belief to action is not external and contingent, but internal and conceptual.” All the same, I do believe that these considerations cast some doubt on the ultimate legitimacy or value, at least in science, of explanations by reasons alone; for in science we generally want to know why something occurred instead of not occurring; and explanations by reasons alone, like statistical explanations of individual events, would seem to fail that condition. But I am going to

258

proceed on the assumption that belief-explanations of actions in a modified form are a legitimate kind of explanation not just in everyday life but in science as well. (I couldn’t grant even initial credence to the thesis of Winch in his Idea of a Social Science that belief- or reason-explanations, modified or not, are the only legitimate kind of explanations in the human sciences (1958, 108,115). And making this assumption, I shall move on to the next stage of my argument. II A Humean universe, I said, is one in which all properties are only externally connected. One feature of the ontological property of being externally connected is that for any two properties that have that feature, it is ontologically possible for either one to exist without the other and ontologically possible for them to coexist—that is, neither requires or precludes the other. It is obvious, I trust, that this is only a necessary condition, not a defining or sufficient condition. But what does it mean for a property to exist and hence for properties to be able to exist one without the other or to coexist or, for that matter, to be unable to exist one without the other or to coexist? First, we may say, but without here arguing the issue, that for a property to exist is for it to be exemplified. (“Individual things exist only as exemplifying properties; properties exist only as exemplified.”) Even if “there are” unexemplified properties, the only meaning that, say, property exclusion could have would be with their exemplifications; for if “there are” unexemplified properties, then all possible properties are actual. But is it always co-exemplification—its necessity, possibility, or impossibility—that is at issue, that is, exemplification in one object? Whatever the answer is to that question, for our purposes we may say that if two properties are only externally connected, then it is ontologically possible for an object to exemplify either without exemplifying the other and ontologically possible for the object to exemplify both (ignoring all considerations of type, that is, assuming that it is possible at all for the object to exemplify either property considered independently). So, for example, the property of being-green and the property of being-happy may be only externally connected because it is ontologically possible for a

259

person to exemplify either property without exemplifying the other and it is ontologically possible for a person to exemplify both. Does one prove that our universe is not a Humean one by citing the example of, say, the properties of being-a-bachelor and being-unmarried? Those are two properties such that it is ontologically impossible for anything to exemplify the first without exemplifying the second. So they are not only externally connected. Was Hume, or anyone who might defend the thesis that the actual universe is a Humean one, simply unmindful of this or any similar example or else hopeful that these properties could really be shown to be, despite appearances, only externally connected? The answer is obviously negative, and suggests that in the thesis itself the expression ‘all properties’ doesn’t really mean all properties or, perhaps better put, assumes that not all predicates denote properties. What then does it mean, and what is intended to be the extension of the expression ‘all properties’ in the thesis that all properties are only externally connected? The answer to this question, as everyone will have anticipated, is that in the context of the thesis, the expression ‘all properties’ means ‘all simple properties’ so that the thesis becomes ‘All simple properties are only externally connected’. This is, surely, the way Wittgenstein, for example, took it in his metaphysics of logical atomism, and we know of his struggles to preserve the thesis by attempting to show that apparent examples of simple properties that are not only externally connected are really cases of complex properties. Simple and complex properties: some philosophers have argued that this is not a distinction in nature itself, but only—to the extent that it has any sense at all—a distinction relative to a language with its distinction between undefined (simple) and defined (complex) predicates. By a curious turn in the history of ideas, some of those— Macintyre and Winch, among several others—who are most eager to affirm or establish that the actual universe is not a Humean one are also anti-ontological to an extent that precludes their recognizing or claiming even to understand a de re distinction of simple and complex properties. Yet it seems clear that only with that distinction granted does either the affirmation or the denial of the thesis make sense. But what is a simple property? I think it’s better not to try to define the idea by way of linguistic notions, that is, as a property whose name is

260

undefined or undefinable; for ‘being undefined’ in some language is just an anthropological fact and ‘being undefinable’ in some or any language just invites the question of the reason for the modality. Instead, one should say that a simple property is one with no constituents, that is, with no other properties as literally constituting the given property. Thus, while being-abachelor has another property, being-unmarried, as a constituent and so is a complex property, being-green (taken as an exact shade) has no other properties as literally constituting it (though it may exemplify other properties), and so is a simple property. (We may recall here that it was the truths of color exclusions that so perplexed Wittgenstein—apparent cases of simple properties that are not only externally connected (1929).) The basic reasons for holding that there is in nature itself and not just relative to a language a distinction between simple and complex properties are these: (1) That is the way we experience the world: some properties are just given to us as having no constituents while others are given as complex. (We should always take the world as we experience it unless there are very compelling reasons, as occasionally there are, to do otherwise.) (2) There is a commonsense distinction between properties we must experience, through sense perception or introspection, in order to know what they are—particular colors, odors, pains, for examples—and properties that we need not experience and in some cases have not experienced in order to know what they are—complex shapes, imaginary animals such as unicorns and mermaids, for examples. The best way to account for this epistemological distinction is by the ontological one of simple and complex properties. (3) There is a commonsense distinction between properties that can be “pictured” only by themselves or other properties of the same kind and properties than can, and in some cases must, be “pictured” by properties of a different kind. Thus a picture of a horse is not a horse or any other animal, but a picture of a red ball, while neither a ball nor any other similar object, is either red or some other color. The best way to explain these facts is to suppose that being-red (taken as a particular shade) is a simple property whereas being-a-horse and being-aball are complex properties. But with that I must conclude my explication and defense of the distinction between simple and complex properties, and move on to a sketch of a theory of the nature of dispositions and dispositional

261

explanations. III Consider the following explanation of a window’s breaking: This window was brittle. This window was struck with force. Brittle windows, when struck with force, break. This window broke. It is tempting initially to suppose that this is a so-called “covering-law” and therefore, possibly, a causal explanation of the window’s breaking in which the first two premises state the initial conditions and the third premise a relevant law. I have argued elsewhere in detail (1981a, 218-221) that this is not so because the third premise is not a law but a definitional truth: being-brittle is a dispositional property (the name of) whose definition includes (the name of) the property of breaking-when-struckwith-force. And according to the theory of dispositions defended in the same article, the schema for the definition of any dispositional property is so: Dx “means” (∃f)(fx . (y)[(fy . O1y) ⊃ O2y]) where ‘D’ names the dispositional property, ‘O1’ and ‘O2’ name observational properties, and ‘f’ is a predicate variable. If we take brittleness as our dispositional property, we get BRx “means” (∃f)(fx . (y)[(fy . Sy) ⊃ By]) which in plain English says that ‘x is brittle’ “means” ‘There is a property that x has and (it is a law that) if anything has that property and is struck with force, then it breaks’. I cannot here restate my arguments for this theory of the nature of dispositions but will remark that it is designed to avoid the problems of both the “singular material implication” account according to which any object not submitted to the relevant test has any

262

given disposition and the “reduction sentences” account according to which dispositional predicates are not fully eliminable from ordered discourse. But even granted that our explanation of the window’s breaking was only a dispositional explanation and not a lawful explanation, (and therefore not a causal explanation either), one can, given this analysis of dispositions, construct a corresponding potential explanation of the covering-law type. For, on that analysis, the first premise by stating that the window possesses a certain dispositional property contains “reference” to a law because that premise really “means” that ‘This window has some (non-dispositional) property and (it is a law that) if anything has that property and is struck with force, then it breaks’. We can thus map our original explanation onto: There is some (non-dispositional) property such that (it is a law that) if anything has that property and is struck with force, then it breaks. This window has that property. This window was struck with force. This window broke. This potential lawful explanation of the window’s breaking becomes an actual lawful explanation when one can identify the property referred to by the variable in the first premise and replace that premise by a law in which the name of the property occurs relevantly in the antecedent. Thus we see that any dispositional explanation, whether trivial or not itself, is always replaceable in principle by a non-trivial lawful explanation. This result is, of course, exactly the opposite of that intended and defended by some who see dispositional explanations (whether recognized as such or not) of human behavior as the only proper sort of explanations in the human case. In any case, I am ready now to apply these ideas to belief-explanations and other reason-explanations of human actions. IV If we drop the idea of belief-explanations as practical syllogisms and regard such explanations as consisting instead entirely of sentences, we

263

may represent the schema of a belief-explanation or reason-explanation initially as follows: Bx Ax where ‘B’ represents belief and ‘A’ action. MacIntyre’s claim, we recall, is that such an explanation is never causal—at least in the Humean sense— because if it really serves as an explanation, the connection between B and A is logical or internal. And that is supposed further to show that the universe is not a Humean one. Let us see. A belief is a disposition. On this point we may agree with Ryle (1949, 133-135). But contrary to Ryle, I would insist that to have a particular belief is not only to be disposed to act in certain ways but also, and more importantly, to have certain conscious mental states (which themselves are not to be analyzed dispositionally or behavioristically, ontologically speaking); indeed that is what makes them dispositional mental states as contrasted with mere behavioral dispositions (Addis, 1989, 6-7). This fact, fundamental though it be in the philosophy of mind, I shall here put aside and treat beliefs as only dispositions to act in certain ways under certain conditions. If beliefs are dispositions to act in certain ways under certain conditions, then it is not implausible to suppose that belief-explanations are a species of dispositional explanations, with all that that implies. Applying the theories of dispositions and dispositional explanations that I sketched in the last section, we get the following results: Letting ‘B’ stand for belief and ‘A’ for action as above, and letting ‘E’ stand for the relevant internal and external conditions treated as a property of the one who would act and ‘f’ once more being a predicate variable, we get as our schema of belief: Bx “means” (∃f)(fx . (y)[(fy . Ey) ⊃ Ay]) And if we now regard ‘B’, ‘E’, and ‘A’ as specified for some particular belief, set of conditions, and action, we may express the idea in semi-plain English by saying that ‘x has belief B’ “means” ‘There is a property that x has and (it is a law that) if anyone has that property and is in conditions E,

264

then that person will perform action A’. And if we return to our schema of a belief-explanation, we may now reformulate it as follows: (∃f)(fx . (y)[(fy . Ey) ⊃ Ay) Ax With this analysis we are in a position to understand and, I hope, accept the following several points. First. We can now see clearly the precise sense in which the connection between belief and action is internal or logical. It is so in the sense of being definitional insofar as ‘A’ enters into the definition of ‘B’. Moreover, this definitional connection is exactly like that between brittleness and breaking or solubility and dissolving in the sense that just as breaking is the definitional realization of brittleness and dissolving the definitional realization of solubility, so, in a particular case, are certain actions the definitional realization of certain beliefs. Understanding clearly that the “internal” reduces to the merely definitional goes a long way, I think, to deflating what otherwise may appear to be an important metaphysical thesis. Second. Not all brittle things break, not all soluble things dissolve, and not all Calvinists, even post-Beza, become or try to become capitalists. Why not? Because, respectively, they are not ever struck with force, they are not ever put in water, and not all of the internal and external conditions E obtain. This is just as it should be. But it helps us to see that the schema of belief-explanations should be further modified, this time by way of adding an additional premise in order to be fully satisfactory formally as a kind of explanation of actions. For just as in our explanation of the window’s breaking, we had a premise that the window was struck with force—that is, that the condition for the realization of the disposition had occurred—similarly we need to add to make our explanation of the action a good one that the conditions E actually obtained. Doing so, we get, I submit, the correct schema of a belief-explanation as follows: (∃f)(fx . (y)[(fy . Ey) ⊃ Ay]) Ex Ax

265

which is a formally valid argument. Third. But this argument, we see, is not actually of the form of what I originally called a dispositional explanation but instead that of what I called a potential explanation of the covering-law type in which one removes the definitional generalization but makes explicit the definition of the disposition by way of reference to a law. We might then say, to speak both somewhat metaphorically and at the same time more accurately, that belief-explanations, when properly understood and reconstructed and in order to have any force, fall somewhere between pure dispositional explanations and causal explanations though crucially more similar to the former than the latter. Fourth. I do not purport to say what people have in mind or really have in mind or even, under some different conditions, would have in mind when they say this or that. But it is an implication of my view that to indulge in a belief-explanation of someone’s behavior is to commit oneself, know it or not, to the proposition that that person is subject to a certain law and that the behavior can in principle be lawfully and therefore causally explained. The causal explanation would, it is obvious, require first identifying the property for which the variable now takes the place and replacing the first premise of an instantiation of the final schema of beliefexplanations by a law. This implication of a deterministic frame of reference in the explanation of human behavior is one that I do not resist. But it is this point I had in mind when I said earlier that some philosophers who emphasize the importance of belief- or reason-explanations, which, to be sure, are not themselves causal explanations, do so in the belief that they are somehow rejecting or at least avoiding the deterministic frame of reference. But this is a mistake. Fifth. Dispositional explanations of the behavior of inanimate or other non-human objects often strike us as trivial and ones we want immediately to go beyond somehow. That this cube of sugar dissolved because it is soluble and was put in water, while both correct and in some measure an explanation, invites the scientist at any rate to look for the ground of the disposition, that is, the property of the sugar that will yield the relevant law for a causal explanation. In the human case, we are much more easily satisfied by belief- and other dispositional explanations of behavior.

266

Almost certainly the most important reason for this is that any human action, as actions are typically characterized, can be the realization of any one of a number of different dispositions: Joan may wave the flag because she is patriotic or because she likes exercise or in order to give a signal. (The latter is more likely an explanation by a conscious mental state than by a disposition, but it doesn’t matter here.) So it is an empirical question which disposition was in fact being realized by a given action and therefore in no way trivial to affirm that it was this and not that disposition that was involved. And this point, which entirely escapes MacIntyre and Winch as far as I can tell, leads to a couple of other, related ones: (1) that even if beliefs or reasons do in some extended way “entail” certain actions, there is never any guarantee that a person with those beliefs will perform those actions and, more important, (2) that it is entirely an empirical question (a) whether the person who performs those actions had those beliefs (because those actions could have come from other beliefs or dispositions) and (b) whether even if the person had those beliefs or reasons for acting, it is those beliefs or reasons that are being realized in these actions that they nevertheless “entail.” For Joan may desire exercise and be waving the flag yet not be waving the flag in order to get exercise. That is why, while nothing can be dissolving without being soluble, beliefs and actions are much more loosely tied: the belief (like any disposition) may exist without being realized, but so too the actions that would be the realization of a given belief may occur without the person’s having that belief (not all capitalists are Calvinists) or not as a result of having that belief even if it is present. Thus beliefs and actions are “separate phenomena” in the most important sense of either being able to exist without the other. Sixth. We are now in a position to understand clearly what is officially the main point of this paper—that the existence of human actions and their explanations by beliefs is fully consistent with the Humean universe. For while I granted and even insisted—it is at the very heart of my analysis of beliefs as dispositions, after all—that the connection between a belief and the actions that are its partial realization is internal and logical, that is, definitional, the connections between and among the simple, or simpler, properties whose names are the defining properties of ‘belief’ are purely external. More precisely, reminding ourselves of the definition of belief as follows:

267

Bx ‘means” (∃f)(fx . (y)[(fy . Ey) ⊃ Ay)] we see that while the connection between what is named by ‘B’ and ‘A’ is obviously internal, the connections between and among what is named by ‘f’ and ‘E’ and ‘A’ are purely external. Thus the possibility, so far forth, that the actual universe is a Humean one is preserved. V I have never shared the impulse to insist that there is only one proper kind of explanation of anything and certainly not of human behavior, whether it be Winch’s thesis that the only proper explanations in social science are those by reasons that the participants themselves would recognize (1958, 108) or the dogma of some positivists that the only proper explanations in social science are those that are causal and do not mention (private) mental phenomena. With respect to causal explanations, especially of the covering-law type, I have been disposed to argue that there are no good philosophical or empirical reasons for denying that human actions admit of such explanations. Thus I see no reason to doubt that social science and psychology can proceed, fruitfully, with the same methodology as the physical and biological sciences in the explanation of human behavior. And I believe further that such explanations will increasingly be forthcoming somewhat in the spirit of Wilson’s sociobiological program (1978), making use of findings from biology, especially evolutionary biology and eventually molecular genetics. But unlike Wilson and many others in the biological and the human sciences, I do not doubt for a moment the legitimacy of explanations by reasons and beliefs even if such explanations rarely have much predictive power. Neither, therefore, do I believe that the one kind of explanation precludes the other, from either direction. Yet it is possible to cut things finer (and far more finely than I shall cut them here) in the possible kinds of explanations of a given human action, and it may be of some value to consider some of them. As we shall see, much depends on what it is about the action that wants explaining. Consider then the action of a man’s shaking hands with a friend who

268

has just arrived for dinner: (1) We might ask the person whose action it is why he did it and get an answer to the effect that he wanted to greet his friend. That was his reason, which might possibly also be expressed by saying the he had the desire to greet his friend and the belief that offering his hand would be taken by his friend as a greeting. This explanation, or another of its kind, would appear to be the only kind that Winch would allow in understanding the behavior, at least as a social action. (2) We might want to know, given that the person wanted to greet his friend, why he chose to offer his hand instead of using some other form of greeting, say, a hug. In this case we are still in the realm of reasons or at least possible reasons that the person has himself, and he might say either that he or his friend doesn’t like to be hugged and shaking hands seemed the most appropriate mode of greeting or that he really didn’t have any particular reason for shaking hands instead of hugging or otherwise greeting his friend. (3) But we might possibly want to know why shaking hands is a form of greeting at all and why human beings ever put their palms one inside the other, move the joined hands up and down briefly, and then release them— as a form of greeting or of congratulations or of any social action. And here we must leave the realm of reasons that the man himself had though possibly not the realm of reasons altogether, depending on the answer. For it might be that those humans who first shook hands had reasons for choosing just that way to greet each other, and theories have been advanced as to what those reasons might he. Or it might be that this form of human behavior arose without any reasons on anyone’s part; it might be a cultural accident, so to speak, and we might look then to evolutionary biology for what might dispose human beings to behave in that way. (4) We might, however, in asking why that man shook hands want to know his personal cultural history. Here we have entirely left the realm of reasons considered as such and enter the realm of lawful explanations. For the schema of relevant laws for explanations of this kind would be that if a human being with standard perceptual and intellectual equipment has existed in such and such environments and is in such and such circumstances, that person performs the action we call “shaking hands.” Laws of this kind will be of the sort I call polychronal (1975, 52) in that

269

their antecedents mention occurrences from a multiplicity of times. But the frame of reference of modern science demands the eliminability in principle of polychronal laws in favor of monochronal laws (that is, if there are lawfully sufficient conditions for occurrences of the kind in question at all, there will be such a set of conditions in the immediately preceding state of the relevant system in which the occurrence to be explained takes place). And this demand leads to our last possibility. (5) The fact that a person has been in certain environments in the past shapes the dispositions of that person to present and future behavior. This is just a piece of common sense which, while seemingly denied by some defenders of causeless free will, is in fact presupposed by all of us in our roles as teachers, parents, friends, clerics, advertisers: if the probabilities of certain behaviors’ occurring is not affected by education, setting examples, sincere advice, religious training, and promotion of products, then our social life is simply unintelligible if not totally irrational. But how can having been in a certain environment in the past affect a person’s present behavior? The answer, of course, is that having been in that environment left a certain “trace” (as the literature has it), a lasting effect in the brain or central nervous system. But this trace in the brain is none other than the ground of the disposition to behave in certain ways in certain circumstances and indeed the very property (whose name is) to be substituted for the predicate variable in an instantiation of the schema (y)[(fy . Ey) ⊃ Ay] such that one has a true law for the explanation of behavior. This amounts to replacing a potential explanation of the covering-law type that corresponds to a dispositional explanation by an actual explanation of the covering-law type. And the polychronal law has now been replaced by a monochronal law, for the latter says that if a person is in such and such brain state and at the same time is in condition E, then the person does A. Thus letting ‘F’ stand for a particular relevant property of the brain, ‘E’ for a particular kind of set of conditions, ‘A’ for a particular kind of action, and ‘a’ for a particular person, a lawful explanation of the covering-law type of a’s having done A would be

270

(x)[(Fx . Ex) ⊃ Ax] Fa . Ea Aa And there is no good reason to doubt that every human action admits in principle of an explanation of this type. VI Here are my main conclusions. (1) A belief- or reason-explanation, when it includes a premise of the other internal and external conditions that must obtain, is a kind of dispositional explanation or else a potential explanation of the coveringlaw type. (2) Dispositional explanations, while not causal or even lawful explanations, are nevertheless informative and hence legitimate explanations. (3) Like all dispositional explanations, belief- or reason-explanations involve properties that are internally or logically connected. (4) But the fact of such internal connections, which are really just definitional connections, is no evidence that the actual universe is not a Humean one: the Humean thesis is that all simple properties are only externally connected and dispositions are necessarily complex properties. (5) Belief- or reason-explanations, being dispositional explanations, presuppose that what they explain can also he explained by causal, lawful explanations in which the properties involved are only externally connected. (6) Hence the thesis that only belief- or reason-explanations are possible or appropriate for human actions is either only stipulatively and so trivially true or else false and self-contradictory. (7) The human sciences can legitimately and fruitfully pursue explanations of human behavior of various kinds and at various levels. In doing so, it is likely that the human sciences will increasingly look to biology for many of its explanations.

271

REFERENCES Addis, Laird. 1965. “Ryle’s Ontology of Mind,” Moore and Ryle: Two Ontologists, Martinus Nijhoff, 1-101. Addis, Laird. 1967. “Particulars and Acquaintance,” Philosophy of Science, 34, 251-259. Addis, Laird. 1974. “Time, Substance, and Analysis,” The Ontological Turn: Studies in the Philosophy of Gustav Bergmann, edited by M. Gram and E. Klemke, The University of Iowa Press, 148-167. Addis, Laird. 1975. The Logic of Society: A Philosophical Study, University of Minnesota Press. Addis, Laird. 1976. “On Defending the Covering-Law Model,” PSA 1974 (Proceedings of the 1974 Biennial Meeting of the Philosophy of Science Association), edited by R. Cohen, C. Hooker, A. Michalos, and J. Van Evra, D. Reidel Publishing Company, 361-368. Addis, Laird. 1981a. “Dispositions, Explanation, and Behavior,” Inquiry, 24, 205-227. Addis, Laird. 198lb. “Mind, Structure, and Time,” Philosophical Topics, 12, 39-52. Addis, Laird. 1982. “Behaviorism and the Philosophy of the Act,” Noûs, 16, 399-420. Addis, Laird. 1983. “Natural Signs,” The Review of Metaphysics, 36, 543568. Addis, Laird. 1984. “Parallelism, Interactionism, and Causation,” Midwest Studies in Philosophy IX: Causation and Causal Theories, edited by

P. French, T. Uehling, and H. Wettstein, University of Minnesota Press, 329-344. Addis, Laird. 1986. “Pains and Other Secondary Mental Entities,” Philosophy and Phenomenological Research, 47, 59-74. Addis, Laird. 1988. “Dispositional Mental States: Chomsky and Freud,” Zeitschrift für Allgemeine Wissenschaftstheorie, 19, 1-17. Addis, Laird. 1989a. “Critical Review of L. Smith’s Behaviorism and Logical Positivism: A Reassessment of the Alliance,” Synthese, 78, 345-356. Addis, Laird. 1989b. “Intrinsic Reference and the New Theory,” Midwest Studies in Philosophy XIV: Contemporary Perspectives in the Philosophy of Language, edited by P. French, T. Uehling, and H. Wettstein, University of Notre Dame Press, 241-257. Addis, Laird. 1989c. Natural Signs: A Theory of Intentionality, Temple University Press. Addis, Laird. 1995. “The Ontology of Emotion,” The Southern Journal of Philosophy, 33, 261-278. Addis, Laird. 1998. “Human Action and the Humean Universe,” Philosophia, 26, 23-40. Addis, Laird. 1999. Of Mind and Music, Cornell University Press. Addis, Laird. 2000. “The Simplicity of Content,” Metaphysica: Zeitschrift für Ontologie & Metaphysik, 2, 23-43. Addis, Laird. 2005. “The Necessity and Nature of Mental Content,” Intentionality: Past and Future, edited by G. Forrai and G. Kampis, Rodopi Publications, 1-14.

274

Aquila, Richard. 1975. “Causes and Constituents of Occurrent Emotion,” The Philosophical Quarterly, 25, 346-349. Armstrong, David. 1968. A Materialist Theory of Mind, Routledge and Kegan Paul. Arnaud, Richard. 1975. “Brentanist Relations,” Analysis and Metaphysics: Essays in Honor of R. M Chisholm, edited by K. Lehrer, D. Reidel Publishing Company, 189-208. Arneson, Peter. 1978. “A Note on Addis’ Analysis of Dispositions,” unpublished. Aune, Bruce. 1967. Knowledge, Mind, and Nature, Random House. Baier, Kurt. 1962. “Pains,” The Australasian Journal of Philosophy, 40, 123. Bergmann, Gustav. 1940a. “On Physicalistic Models of Non-Physical Terms,” Philosophy of Science, 7, 151-158. Bergmann, Gustav. 1940b. “On Some Methodological Problems of Psychology,” Philosophy of Science, 7, 205-219. Bergmann, Gustav. 1940c. “The Subject Matter of Psychology,” Philosophy of Science, 7, 415-433. Bergmann, Gustav. 1941. “Operationism and Theory in Psychology,” Psychological Review, 48, 1-14. (with K. Spence.) Bergmann, Gustav. 1942. “An Empiricist Schema of the Psychophysical Problem, Philosophy of Science, 9, 72-91. Bergmann, Gustav. 1943. “Psychoanalysis and Experimental Psychology: A Review from the Standpoint of Scientific Empiricism,” Mind, 52, 122-140.

275

Bergmann, Gustav. 1945. “A Positivistic Metaphysics of Consciousness,” Mind, 54, 193-226. Bergmann, Gustav. 1951a. “The Logic of Psychological Concepts,” Philosophy of Science, 18, 92-110. Bergmann, Gustav. 1951b. “Comments on Professor Hempel’s ‘The Concept of Cognitive Significance’,” Daedalus, 80, 78-86. (also in his 1954b, 255-267) Bergmann, Gustav. 1952. “The Problem of Relations in Classical Psychology,” Philosophical Quarterly, 2, 140-152. (also in his 1954b, 277-299) Bergmann, Gustav. 1953. “Theoretical Psychology,” Annual Review of Psychology, 4, 435-458. Bergmann, Gustav. 1954a. “Bodies, Minds, and Acts,” in his 1954b, 132152. Bergmann, Gustav. 1954b. The Metaphysics of Logical Positivism, Longmans, Green, and Company. Bergmann, Gustav. 1955. “Intentionality,” Archivio di Filosofia: Semantica, 177-216. (also in his 1959, 3-38). Bergmann, Gustav. 1956. “The Contribution of John B. Watson,” Psychological Review, 63, 265-276. Bergmann, Gustav. 1957. Philosophy of Science, The University of Wisconsin Press. Bergmann, Gustav. 1959. Meaning and Existence, The University of Wisconsin Press.

276

Bergmann, Gustav. 1960. “Acts,” The Indian Journal of Philosophy, 2, 130 and 96-117. (also in his 1964, 3-44). Bergmann, Gustav. 1964. Logic and Reality, The University of Wisconsin Press. Bergmann, Gustav. 1967. Realism. A Critique of Brentano and Meinong, The University of Wisconsin Press. Boring, E. G. 1957. A History of Experimental Psychology (Second Edition), Appleton-Century-Crofts. Brentano, Franz. 1973. Psychology from an Empirical Standpoint, edited by O. Kraus, English edition by L McAlister, Humanities Press. (first published in 1874) Brentano, Franz. 1981. Sensory and Noetic Consciousness, edited by O. Kraus, English edition by L. McAlister, Routledge and Kegan Paul. (first published in 1929) Broad. C. D. 1925. Mind and Its Place in Nature, Routledge and Kegan Paul. Burge, Tyler. 1979. “Individualism and the Mental,” Midwest Studies in Philosophy IV: Studies in Metaphysics, edited by P. French, T. Uehling, and H. Wettstein, University of Minnesota Press, 73-121. Butchvarov, Panayot. 1979. Being Qua Being: A Theory of Identity, Existence, and Predication, Indiana University Press. Carnap, Rudolf. 1936-37. “Testability and Meaning,” Philosophy of Science, 3, 4 19-471 and 4, 1-40. Chisholm, Roderick. 1958. “Chisholm-Sellars Correspondence on Intentionality,” Minnesota Studies in the Philosophy of Science (Volume II), edited by H. Feigl, M. Scriven, and G. Maxwell,

277

University of Minnesota Press, 521-539. Collingwood, R. G. 1966. The Idea of History, Oxford University Press. (first published in 1946) Dennett, Daniel. 1978. Brainstorms, Bradford Books. Devitt, Michael. 1984. Realism and Truth, Oxford University Press. Dray, William. 1957. Laws and Explanation in History, Oxford University Press. Findlay, John. 1963. Meinong’s Theory of Objects and Values, Clarendon Press. Freud, Sigmund. 1970. An Outline of Psycho-Analysis, W. W. Norton and Company. (first published in 1940) Grossmann, Reinhardt. 1965. The Structure of Mind, The University of Wisconsin Press. Grossmann, Reinhardt. 1983. The Categorial Structure of the World, Indiana University Press. Grünbaum, Adolf. 1984. The Foundations of Psychoanalysis: A Philosophical Critique, University of California Press. Hempel, Carl. 1965. Aspects of Scientific Explanation and Other Essays, The Free Press. Hempel, Carl. 1976. “Dispositional Explanation and the Covering-Law Model: Response to Laird Addis,” PSA 1974 (Proceedings of the 1974 Biennial Meeting of the Philosophy of Science Association), edited by R. Cohen, C. Hooker, A. Michalos, and J. van Evra, D. Reidel Publishing Company, 369-376.

278

Hochberg, Herbert. 1967. “Dispositional Properties,” Philosophy of Science, 34, 10-17. Hochberg, Herbert. 1978. Thought, Fact, and Reference: The Origins and Ontology of Logical Atomisn, University of Minnesota Press. Hochberg, Herbert. 1981. “Intentionality, Logical Bergmann’s Ontology,” Noûs, 15, 155-164.

Structure,

and

Honderich, Ted. 1981. “Psychophysical Lawlike Connections and Their Problem,” Inquiry, 24, 277-303. Husserl, Edmund. 1971. Logical Investigations, Humanities Press. (Originally published in 1900) Jacquette, Dale. 1994. Philosophy of Mind, Prentice Hall. Jaynes, Julian. 1976. The Origin of Consciousness in the Breakdown of the Bicameral Mind, Houghton Mifflin Company. Kaila, Eino. 1945. “Wenn...So...,” Theoria, 11, 88-98. Kenny, Anthony. 1963. Action, Emotion, and Will, Routledge and Kegan Paul. Koch, Sigmund. 1964. “Psychology and Emerging Concepts of Knowledge as Unitary,” Behaviorism and Phenomenology: Contrasting Bases for Modern Psychology, edited by T. Wann, University of Chicago Press, 1-41. Leiber, Justin. 1975. Noam Chomsky: A Philosophic Overview, St. Martin’s Press. Lewis, David. 1973. Counterfactuals, Harvard University Press. List, Charles. 1981. “Images, Propositions, and Natural Signs,” Imagery,

279

Concepts, Results, and Applications, edited by E. Klinger, Plenum Press, 67-76. Lyons, William. 1980. Emotion, Cambridge University Press. MacCorquodale, Kenneth and Meehl, Paul. 1948. “On a Distinction Between Hypthetical Constructs and Intervening Variables,” Psychological Review, 55, 95-107. Maclntyre, Alasdair. 1967. “A Mistake About Causality in Social Science,” Philosophy, Politics, and Society (2nd Series), edited by P. Laslett and W. Runciman, Basil Blackwell, 48-70. Mackenzie, Brian. 1977. Behaviorism and the Limits of Scientific Method, Humanities Press. Mackie, J. L. 1979. “Mind, Brain, and Causation,” Midwest Studies in Philosophy IV: Studies in Metaphysics, edited by P. French, T. Uehling, and H. Wettstein, Minneapolis: University of Minnesota Press, 19-29. Mackie, J. L. 1981. “The Efficacy of Consciousness: Comments on Honderich’s Paper,” Inquiry, 24, 343-351. Meinong, Alexius. 1899. “Ǘber Gegenstände höherer Ordnung und deren Verhältnis zur inneren Wahrnemung,” Zeitschrift für Psychologie und Physiologie der Sinnesorgane, 31, 182-272. Millikan, Ruth. 1993. White Queen Psychology and Other Essays for Alice, MIT Press. Moore, G. E. 1910. “The Subject Matter of Psychology,” Proceedings of the Aristotelian Society, 10, 36-62. Ockham, William of. 1957. Philosophical Writings, edited by P. Boehner, Thomas Nelson and Sons Ltd.

280

Palmer, David. 1975. “Unfelt Pains,” American Philosophical Quarterly, 12, 289-298. Pap, Arthur. 1958. “Disposition Concepts and Extensional Logic,” Minnesota Studies in the Philosophy of Science (Volume II), edited by H. Feigl, M. Scriven, and G. Maxwell, University of Minnesota Press, 196-224. (also in Tuomela, 1978, 27-54) Pap, Arthur. 1959. “On the Empirical Interpretation of Psychoanalytic Concepts,” Psychoanalysis, Scientific Method and Philosophy, edited by S. Hook, New York University Press, 283-297. Percy, Walker. 1975. “The Delta Factor,” in his Message in the Bottle, Farrar, Straus, and Giroux, 3-45. Popper, Karl. 1965. Conjectures and Refutations, Routledge and Kegan Paul. Price, H. H. 1953. Thinking and Experience, Hutchinson House. Putnam, Hilary. 1975. “The Meaning of ‘Meaning’,” Mind, Language, and Reality, Cambridge University Press, 215-271. Putnam, Hilary. 1981. Reason, Truth and History, Cambridge University Press. Rorty, Richard. 1965. “Mind-Body Identity, Privacy, and Categories,” The Review of Metaphysics, 19, 24-54. Rorty, Richard. 1979. Philosophy and the Mirror of Nature, Princeton University Press. Rosenberg, Jay. 1974. Linguistic Representation, D. Reidel Publishing Company.

281

Russell, Bertrand. 1956. “On The Nature of Acquaintance,” Logic and Knowledge, edited by R. Marsh, The Macmillan Company, 129-174. (first published in 1914) Russell, Bertrand. 1981. “Knowledge by Acquaintance and Knowledge by Description,” Mysticism and Logic, Barnes and Noble Books, 152167. (first published in 1911) Ryle, Gilbert. 1949. The Concept of Mind, Barnes and Noble, Inc. Sampson, Geoffrey. 1979. Liberty and Language, Oxford University Press. Sartre, Jean-Paul. 1948. The Emotions: Outline of a Theory, Philosophical Library. (first published in1939) Sartre, Jean-Paul. 1957. The Transcendence of the Ego, The Noonday Press. (first published in 1937) Schopenhauer, Arthur. 1966. The World as Will and Representation, Dover Publications, Inc. (first published in 1819) Searle, John. 1983. Intentionality: An Essay in the Philosophy of Mind, Cambridge University Press. Searle, John. 1992. The Rediscovery of the Mind, MIT Press. Sellars, Wilfrid. 1956. “Empiricism and the Philosophy of Mind,” Minnesota Studies in the Philosophy of Science (Volume I), edited by H. Feigl and M. Scriven, University of Minnesota Press, 253-329. Smith, Laurence. 1986. Behaviorism and Logical Positivism: A Reassessment of the Alliance, Stanford University Press. Spiker, Charles. 1977. “Behaviorism, Cognitive Psychology, and the Active Organism,” LifèSpan Developmental Psychology, edited by N. Datan and H. Reese, Academic Press, 93-103.

282

Stalnaker, Robert. 1968. “A Theory of Conditionals,” Studies in Logical Theory, edited by N. Rescher, Basil Blackwell, 98-112. Tuomela, Raimo. 1978. Dispositions, D. Reidel Publishing Company. Twardowski, Kasimir. 1977. On the Content and Object of Presentations, Martinus Nijhoff. (first published in 1894). Watson, John. 1913. “Psychology as the Behaviorist Views It,” Psychological Review, 20, 158-177. Weber, Max. 1985. The Protestant Ethic and the Spirit of Capitalism, Unwin Paperbacks. (first published in 1904-05.) Wettstein, Howard. 1977. “Proper Names and Propositional Opacity,” Midwest Studies in Philosophy II: Studies in the Philosophy of Language, edited by P. French, T. Uehling, and H. Wettstein, The University of Minnesota, Morris, 187-190. Wettstein, Howard. 1979. “Indexical Reference and Propositional Content,” Philosophical Studies, 36, 91-100. Wettstein, Howard. 1981. “Demonstrative Reference and Definite Descriptions,” Philosophical Studies, 40, 241-257. Wettstein, Howard. 1983. “The Semantic Significance of the ReferentialAttributive Distinction,” Philosophical Studies, 44, 187-196. Wettstein, Howard. 1984a. “How to Bridge the Gap Between Meaning and Reference,” Synthese, 58, 63-84. Wettstein, Howard. 1984b. “Did the Greeks Really Worship Zeus?,” Synthese, 60, 439-449. Wettstein, Howard. 1986. “Has Semantics Rested on a Mistake?,” The

283

Journal of Philosophy, 83, 185-209. Wettstein, Howard. 1988. “Cognitive Significance without Cognitive Content,” Mind, 97, 1-28. White, Alan. 1964. Attention, Basil Blackwell. Wilson, Edward. 1978. On Human Nature, Harvard University Press. Wilson, Fred. 1969. “Dispositions: Defined or Reduced?,” Australasian Journal of Philosophy, 47, 184-204. Wilson, Fred. 1985. “Addis on Analyzing Disposition Concepts,” Inquiry, 28, 248-260. Winch, Peter. 1958. The Idea of a Social Science and Its Relation to Philosophy, Routledge and Kegan Paul. Wittgenstein, Ludwig. 1929. “Some Remarks on Logical Form,” Proceedings of the Aristotelian Society, Supp. 9, 162-171. Wittgenstein, Ludwig. Blackwell.

284

1953.

Philosophical

Investigations,

Basil

AUTHOR’S NOTE The papers in this volume appeared originally as follows: “Mind, Structure, and Time,” Philosophical Topics, 12/3/1981, 39-52. published by University of Arkansas Press. “Natural Signs,” The Review of Metaphysics, 36/3/1983, 543-568. “Pains and Other Secondary Mental Entities,” Philosophy and Phenomenological Research, 47/1/1986, 59-74. “Intrinsic Reference and the New Theory,” Contemporary Perspectives in the Philosophy of Language (Midwest Studies in Philosophy, XIV), eds. French, Uehling, and Wettstein, University of Notre Dame Press, 1989, 241-257. “The Ontology of Emotion,” The Southern Journal of Philosophy, 33/3/1995, 261-278. “The Simplicity of Content,” Metaphysica: Zeitschrift für Ontologie & Metaphysik, 2/1/2000, 23-43. “The Necessity and Nature of Mental Content,” Intentionality: Past and Future, eds. Forrai and Kampis, Rodopi Publications, 1-14. “Dispositions, Explanation, and Behavior,” Inquiry, 24/2/1981, 205227. “Behaviorism and the Philosophy of the Act,” Noûs, 16/3/1982, 399420.

“Parallelism, Interactionism, and Causation,” Causation and Causal Theories (Midwest Studies in Philosophy, XIV), eds. French, Uehling, and Wettstein, University of Minnesota Press, 1984, 329-344. “Dispositional Mental States: Chomsky and Freud,” Zeitschrift für Allgemeine Wissenschaftstheorie, 19/1/1988, 1-17. Review of Laurence D. Smith’s Behaviorism and Logical Positivism: A Reassessment of the Alliance, Synthese, 78/3/1989, 345-356. “Human Action and the Humean Universe,” Philosophia, 26/1-2/1998, 23-40.

286

INDEX OF NAMES Addis, Laird, 9, 19, 30, 47, 49, 50, 51, 63, 64, 65, 67, 69, 72, 73, 75, 94, 95, 103, 104, 107, 112, 131, 135, 142, 162, 168, 169, 180, 182, 186, 193, 200, 206, 217, 221, 222, 223, 224, 226, 237, 240, 241, 250, 262, 264, 269 Aquila, Richard, 97 Aquinas, Thomas, 38, 130, 190 Aristotle, 12, 130, 188, 190 Armstrong, David, 54 Arnaud, Richard, 195 Arneson, Peter, 147 Aune, Bruce, 42, 43 Ayer, A. J., 176 Baier, Kurt, 64 Bergmann, Gustav, 9, 38, 50, 76, 112, 123, 130, 131, 173, 174, 175, 176, 177, 183, 192, 193, 194, 195, 196, 197, 245, 247, 248, 249 Berkeley, George, 52, 188, 190, 206 Boring, E. G. , 191, 245 Brentano, Franz, 9, 27, 50, 190, 195, 254 Broad, C. D., 9, 54, 236, 240 Brunswick, Egon, 246 Burge, Tyler, 139 Butchvarov, Panayot, 31 Carnap, Rudolf, 152, 158, 159, 176, 222, 245, 249, 254 Chisholm, Roderick, 27, 46 Chomsky, Noam, 136, 179, 214, 221, 225-230, 231-233, 235-236 Collingwood, R. G., 7 Comte, Auguste, 247 Dennett, Daniel, 63 Descartes, René, 37, 52, 130, 189 Devitt, Michael, 68, 73-74, 75 Dray, William, 161 Ehrenfels, Christian, 190 Feigl, Herbert, 176, 245, 249 Findlay, John, 118

Fodor, Jerry, 136 Frege, Gottlob, 38, 68, 69, 80, 82 Freud, Sigmund, 89, 90, 106, 221, 225-230, 231-232, 236-240 Geach, Peter, 42 Grossmann, Reinhardt, 50, 53, 193 Grünbaum, Adolf, 229 Hempel, Carl, 152, 158, 159, 161-166, 222, 245, 254 Hochberg, Herbert, 117, 123-124, 147, 160-161 Holt, Edwin, 245 Honderich, Ted, 213 Hull, Clark, 244, 245, 246-248, 249 Hume, David, 52, 74, 190, 214, 256, 260 Husserl, Edmund, 99, 112, 130, 131 Jacquette, Dale, 136 Jaynes, Julian, 40 Kaila, Eino, 147 Kenny, Anthony, 64, 103 Koch, Sigmund, 244 Kuo, Z. Y., 245 Leiber, Justin, 228 Lewis, David, 158 List, Charles, 42 Locke, John, 190 Lyons, William, 88, 90 MacCorquodale, Kenneth, 251 MacIntyre, Alasdair, 168, 255, 256-259, 260, 264, 267 MacKenzie, Brian, 244 Mackie, J. L., 199 Meehl, Paul, 241 Meinong, Alexius, 112, 117, 130, 142, 190 Mill, John Stuart, 74, 80, 214, 246 Millikan, Ruth, 132 Moore, G. E., 9, 133-134, 137 Naess, Arne, 247 Neurath, Otto, 247 Palmer, David, 54

288

Pap, Arthur, 147, 150, 237 Pavlov, Ivan, 245 Percy, Walker, 84-85 Perry, Ralph, 245 Plato, 188 Popper, Karl, 159-160, 224 Price, H. H., 54 Putnam, Hilary, 73-75, 79, 131, 138-141 Quine, Willard, 249 Reichenbach, Hans, 246 Rorty, Richard, 32, 33, 50, 52 Rosenberg, Jay, 42 Russell, Bertrand, 29-30, 37, 52, 74, 81, 111, 112, 131, 137, 142, 154, 249 Ryle, Gilbert, 40, 45, 103, 185, 189, 225, 231, 252, 264 Samson, Geoffrey, 233 Sartre, Jean-Paul, 30, 89, 90, 99, 111, 112, 131, 137, 173, 246 Schlick, Moritz, 246 Schopenhauer, Arthur, 25 Searle, John, 72, 95, 130, 131, 242 Sellars, Wilfrid, 27, 42, 43, 46, 104, 136 Skinner, B. F., 178, 182, 185, 244, 245, 246, 248-249, 250 Smith, Laurence, 243-254 Spence, Kenneth, 245, 247, 249 Spiker, Charles, 253 Spinoza, Baruch, 12, 184, 202, 213 Stalnaker, Robert, 158 Stevens, S. S., 245 Thorndike, Edward, 245 Tolman, Edward, 244, 245-246, 247, 248, 254 Twardowski, Kasimir, 112, 117, 118, 130 Watson, John, 173, 182, 245, 247, 249, 250 Weber, Max, 257, 258 Wettstein, Howard, 67, 68, 80-85 White, Alan, 54 William of Ockham, 38, 68, 111, 112, 131, 241 Williams, D. C., 245

289

Wilson, Edward, 268 Wilson, Fred, 158, 230 Winch, Peter, 167, 168, 255, 259, 266, 267, 268 Wittgenstein, Ludwig, 33, 42, 90, 103, 136, 218, 255, 260, 261 Woodger, J. H., 247

290

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

8 Rafael Hüntelmann Existenz und Modalität Eine Studie zur Analytischen Modalontologie

2

9 Andreas Bächli / Klaus Petrus Monism

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

EditedBy • HerbertHochberg • RafaelHüntelmann ChristianKanzian • RichardSchantz • ErwinTegtmeier

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

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

19 Fred Wilson Acquaintance, Ontology, and Knowledge Collected Essays in Ontology ISBN 978-3-938793-58-9 XX, 726., Hardcover, EUR 159,00

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

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

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

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

24 Simone Gozzano, Francesco Orilia Tropes, Universals and the Philosophy of Mind Essays at the Boundary of Ontology and Philosophical Psychology ISBN 978-3-938793-83-1 ca. 180pp., Hardcover, EUR 69,00

EditedBy • HerbertHochberg • RafaelHüntelmann ChristianKanzian • RichardSchantz • ErwinTegtmeier