Evidently, there is no seeing without having some sense impressions; in the standard case, we see that an object is colored by having an impression of its color. But how, in general, do perceptions depend upon sense impressions? How, in general, do sense impressions relate to, or figure in, the sensuous judgments we make? From critics of an abstractionist account of concept formation there have come a range of answers to these questions. The answers largely have been unchallenged. Surprisingly, they all assert that sense impressions play no conceptual role in perception at all. Apart from necessary fussing with the difficult concept of 'conceptual role', the point has been put variously in theses of varying strength. Professor Aune, for example, has suggested that occurrences of sense impressions are adventitious. Aune wrote:

. . . there is nothing essentially conceptual about sensory experiences, and . . . such experiences are no more essential to thought than sensuous imagery. ([3], 235). . . Also, it is essential to note that while "red" and similar primitive predicates of the common-sense scheme are observation terms, this is strictly an accidental fact about them. ([3], 248n.)

Aune's remarks are surprising. Surely, it is a necessary truth that red is a color. But to be a color is to be a quality which appears in the visible spectrum; color is that "quality in virtue of which objects present different appearances to the eye". [Entry from The Shorter Oxford English Dictionary.] But if so, then it is not after all an accidental fact that 'red' is an observation term, if that means that it is a term for an observable quality. It is true of course that facts might have been different. It might have been that we were so constituted as to be unable visually to discriminate red things by their color. But it is important to recognize that this possibility shows something about us, and not about the concept red. That 'red' is an observation term for most of us is strictly an accidental fact about us. It is not an accidental fact about the term.

For another thing, it is not true, as the remarks quoted from Aune may suggest, that sense experiences are related to perceptions and sensuous judgments as images are related to non-sensuous thought. We may think or believe, truly or not, that something is the case, and we may think this quite independently of the occurrence of any imagery. But we cannot perceive or sensuously believe that something is the case quite independently of the occurrence of any sense experiences. Judging that my book is red in the absence of visual experiences is not seeing what color it is. So sense impressions bear a closer relation to perceptions than do images to non-sensuous judgments. Judgments without the occurrences of related sense experiences are not perceptions.

Occurrences of sense experiences are necessary conditions for occurrences of perceptual judgments. This is, no doubt, a truism. Granting that this dependence exists, the interesting question remains: In what does the dependence consist? How are sense experiences related to perceptual judgments? Professor Sellars wrote:

. . . the direct perception of physical objects is mediated by the occurrence of sense impressions which latter are, in themselves, thoroughly noncognitive. Furthermore, this mediation is causal rather than epistemic. Sense impressions do not mediate by virtue of being known. ([8], 90-91.)

Sellars is right of course in saying that sense impressions do not mediate the occurrences of sense perceptions by virtue of being known. It is (in standard, veridical occurrences) an object which is perceived and concerning which we know something of its visible character, not a sense impression. Sense impressions are had, not known. They are neither cognitions, nor, in their standard occurrences, are they objects of a judgment. Moreover, it is also true that occurrences of sense impressions are themselves, in the standard case, the product of causal interactions with our environment. These facts, together with our truism (that occurrences of sense impressions are necessary conditions for occurrences of perceptions), combine to suggest that occurrences of sense impressions causally mediate the occurrences of perceptions. These facts and truism may suggest that it is upon just this that the dependence of perception upon experience rests.

This putative consequence is, I think, plausible. After all, much of what is special about perceptual judgments consists in the way in which they are tied to the contexts of their occurrences in a fashion that non-sensuous thoughts are not. But plausible or not, the conclusion, even if true, is not a consequence of the cited facts and truism. And if it is not, then our acceptance of the putative conclusion requires further grounds extending beyond our acceptance of the cited facts and truism. Thinking about it, it is by no means obvious that, since sense impressions are not themselves cognitions, they play no conceptual role in perception. And, thinking about it, the fact that sense impressions are themselves products of causal interactions with our surroundings does not at all indicate that their relation to perception is itself that of causally mediating, i.e., bringing about occurrences of, separate perceptual acts as logically distinct psychical events.

It certainly does not follow from the fact that sense impressions are not themselves objects of epistemic acts in standard cases of perception, nor cognitions, that they play no conceptual role in sensuous judgments. For it is equally true that concepts are themselves neither the objects of epistemic acts in standard cases of non-sensuous judgment, nor cognitions, yet none of us would want to claim that concepts play no conceptual role in such judgments. Why then should we think this of sense impressions? Some further justification is needed.

Nor does it follow from the dependence of perception upon experience, or from the causal conditions of the occurrences of impressions, that impressions are themselves mere causal mechanisms triggering the further occurrences of perceptions. For not every dependency is a causal dependency. The causal conditions for the occurrences of sense impressions might, for all we so far know, thereby be conditions for the causal occurrences of perceptions. This would be so, for instance, if sense impressions were, say, literal constituents of sensuous judgments, rather than being external and separate, although related, occurrences. In fact, having reached this point, it is now puzzling to see how it is that perceptions are after all thought to be causally mediated by sense impressions. How exactly do sense impressions literally cause occurrences of perceptions? Why are these causal occurrences also necessary in determining what the judgment is about? Why is it that in knowing what we perceive we thereby know how we are sensuously impressed? Why is it that none of these things is true of other causes of other non-sensuous judgments?

So far we have collected a few background conditions for the characterization of the relation of sense impressions to perceptions. We have, common to the views at hand, the assertion that sense impressions have no conceptual status. We have that sense impressions or their occurrences, are, however, necessary conditions for the occurrences of perceptions. Presumably, we also have that perceptions, by contrast to sense impressions, are a species of judgment. (We see what is the case; we see, say, that the book is red. The expression of perceptions is propositional and assertive in a manner in which the expression of sense impressions is not.) Given these background conditions, Sellars has argued that the relation of impression to perception is a causal relation. We can see now that the relation is not only this, if it is at least this, from some remarks by Professor Geach. We need after all to explain not merely the conditions of the occurrence of the perception. We need to explain as well how its particular reference is secured. Geach wrote:

The content of the judgment is always intelligible and conceptual -- acquaintance with a particular sensible thing is no part of the judgment itself -- but an act of judgment performed in a particular sensory context may thereby be referred to particular sensible things. It is clear, indeed, that the act of judgment must bear a closer relation than mere simultaneity to the context of sense-perception that gives it its special reference to these particular sensible things; I am not prepared to characterize this special relation it must bear to its context. . . . the problem is not how we advance from judgments like this is before that to more general judgments, but contrariwise how a judgment inherently general can be tied down to referring to particular things. ([5], 64-65.)

What constitutes this reference to definite particulars? Or again, what is the difference between "there are white cats", "some cats are white", on the one hand, and on the other hand "these cats are white"? How do the judgments they express differ? What constitutes the reference to a particular set of cats?

In all such cases, I should maintain, there is no difference to be found on the side of the judgment itself. What we may call the intelligible content of the judgment is the same in all judgments expressible as "that flash was before this bang", regardless of which flash and bang are in question. ([5], 63.)

. . . the judgment expressed in "The cat is eating the liver" did not differ as regards its own content from one expressible as "a cat eats liver"; the references to a particular cat and a particular bit of liver and a particular time come in because the judgment is made in a particular sensory context, of seeing a cat eat liver. ([5], 122.)

Perceptions, then, seem to be judgments plus sensory contexts. That is, they are judgments occurring in, and if Sellars is right, they are judgments causally provoked by, (elements of) sensory contexts. Given the gap between perceptions as conceptual and sensory stimulations as not, we have now Geach's Puzzle: how is a sensuous judgment referred to a particular thing on a particular occasion? There are two aspects to Geach's Puzzle. One aspect is quite general and runs beyond, although it includes, sensuous judgments. It involves the view that what is conceptual and intelligible is always really general. Paradoxically, there really are no singular judgments at all. Singular references which cannot be explained away in terms of the conceptual mechanisms adequate for general knowledge and belief, e.g., the quantifiers, are really not conceptual items at all. This general view applies of course to perceptions as a special case, for perceptions are paradigms of apparent singular judgments. They are psychical occurrences which are immediately, often demonstratively, tied to the sensory contexts of their occurrences. (I see for example that something, that thing, has moved.)

The other, central, aspect of Geach's Puzzle resides in this, the fact that two sensuous judgments (whether really singular or not) may have the same conceptual content and yet be referred to different things. Presumably they may on occasion even have opposite truth-values. This is not just, although it includes, the question of characterizing indexical and demonstrative references. (On the surface at least, the intelligible content of the two assertions, 'I can beat you up', made by two youngsters, fists clenched., is the same although the references and truth-values of each are different. But equally, on the surface at least, the intelligible content of our judgments, 'something moved', when, walking together you and I catch a flicker of change at the periphery of our visual fields, is also the same. It need not however be true that it was the same movement we each detected.)

The general aspect of Geach's Puzzle raises issues larger than the dimensions of this paper. But it seems to me quite evident at minimum that the intelligible content of "some cats are white" and "these cats are white" is not after all the same; and it seems to me quite evident that "the cat is eating the liver" and "a cat eats liver" do differ in content, on any reasonable characterization of the content of a judgment. Surely, the negations of judgments with the same intelligible contents must themselves agree in their intelligible content. But that is not true in these instances. And surely, judgments which imply things which other judgments do not, must differ in their contents from those other judgments. But a judgment the expression of which is that this cat is white, incorporating the demonstrative 'this', implies, as the judgment that some cats are white does not, that there is a cat in the surrounding environment. (The truth-conditions for, and the semantic characterization of, statements incorporating demonstrative terms are different from those of quantified statements. Cf. [7], esp. pp. 112-113.)

However this larger issue might go, there remains in any case the central, and more circumscribed, aspect of Geach's Puzzle. We agree, that is, at least for the sake of polemic, that two judgments may indeed have the same intelligible content but, in a given context, have various references. How is this possible? Borrowing from Geach, we think, perhaps, of perceptions as pairs, the members of which are a judgment and a sensory context. Perceptions are thus complex entities with a conceptual, cognitive component and with a separate, non-conceptual, sensory component. Sensory contexts in turn must at least be certain triples, occurrence points the elements of which are psychical agents, objects of perception, and occasions. My Tuesday perception, then, that the cat is eating liver is different from my Thursday perception of the cat's doing so precisely in the difference in the occasion of the occurrence, the other factors remaining constant: me, cat, and judgment. The rub is that every judgment occurs in some sensory context, even non-sensuous ones. But not every judgment is a perception. We require, evidently, some further conditions serving to link the subject of a sensuous judgment with the second element of the sensory context of the judgment. A causal link does not ensure that what is thought of is what is seen. And even the identity of the object of reference with an object of perception in a sensory context is neither sufficient nor necessary to ensure that a judgment is a perception. It is not sufficient; I might, for example, judge that the cat is eating liver (knowing perhaps her regular habits and that it is 4 o'clock), see the cat, but not perceive that she is at the liver. I judge, then, albeit non-sensuously, what I might have perceived in the context in which I might have perceived it. Neither is the identity of the object of reference with an object of the sensory context a necessary condition for the occurrence of a sensuous belief. For I may look, judge, but be mistaken. You may know that what I think I saw is not, in fact, what was before me. The referent of my judgment and the object of my perception fail to coincide. Yet I did sensuously judge. I merely (mis)perceived what was the case.

So Geach's Puzzle remains. The way in which perceptual judgments are tied to their objects of reference in the sensory context remains a mystery. There is something right about the appeal to the sensory context of the judgment in securing the link of judgment to surroundings, but what exactly? Help, I think, is at hand. All of the philosophers whose remarks have been twisted to our polemical purposes share two important views. The first, the view that sense impressions are non-conceptual, is already before us. The second can now be introduced and, I think, exploited to deal with Geach's Puzzle. The second view is the view that thought is to be understood by analogy to speech. We shall attempt to exploit this, but at a price. The first view, the view that sense impressions are, after all, non-conceptual, must be given up.

The analogy between thought and speech is based upon some very basic features common to each. Judgments and assertions are each intentional, each has truth-values. The logic of the ascription of a judgment to an agent exhibits formal features analogous to those which characterize the ascription of assertions to him. There exist common to each failures of the substitution of identities and of the interchange of equivalents to preserve truth. There is the mutual failure of Existential Generalization. And acts of judgment, like acts of assertion, can themselves be the subjects of immediate, first-person, non-inferential awareness. (I need not listen to myself to know what I am saying.) There is, thus, some basis for the belief that attention to the overt, communal acts of assertion may give some insight into the nature of the covert, personal acts of thought. Let us suppose, for now at least, that this is true. Judgments, we suppose, are to be understood by analogy with acts of overt speech. But perception is, or incorporates, a kind of judgment. Presumably, then, we can bring understanding to the concept of perception by analogy with the concept of speech. And this, too, may seem plausible. Perception, like non-sensuous thought, issues in knowledge or belief. Perceptual judgments, like non-sensuous judgments, have truth-values, are intentional, and are the subjects of reflexive, non-inferential knowledge. The ascription of perceptual acts to agents exhibits formal characteristics reminiscent of those which attach to the ascription of thoughts or assertions to him. So the question arises: Can we exploit the thought-speech analogy in a way that gives insight into Geach's Puzzle and its possible resolution?

We need to know what features of the overt speech occurrences are to be carried back to the covert psychical acts. It is of course clear that all sorts of features of speech are quite irrelevant and adventitious so far as perception goes. Speaking, after all, consists in the linear production of conventional entities for communication. But thought, and particularly, perception, is not like this. Occurrences of perceptions are neither linear, produced by agents, conventional, nor communicatory.

Nonetheless, there remain important requirements necessary for the analogy between the overt, conventional productions of speech and the covert, sensuous judgments standardly produced by the stimulation of the senses. For in the simplest cases, in asserting what is so, we refer to something and ascribe something of it. It is in, and by, the acts of reference and ascription that we state what is the case. In doing so, we perform two, not three, conventional acts. It is in the occurrences of these constituent acts that the act of assertion consists. We distinguish of course acts of assertion from the assertions which result from them. Assertions in turn consist of referring and ascriptive expressions, those linguistic entities the production of the material tokens of which are the conventional occurrences required for stating what is the case.

If, then, thinking is significantly like speaking, and if thought is significantly like assertion, there must be acts on the covert, psychical side of the thought-speech analogy to match the manifest acts of referring and ascribing. There must be distinct constituents of thoughts to match the referring and ascriptive expressions of assertions.

In particular, if perceptions or sensuous judgments are a species of thought, and if thought bears, after all, an interesting analogy to overt verbal occurrences, then perceptions or sensuous judgments must also contain constituents which are the natural vehicles by which reference and ascription are carried in seeing what is the case. The act of seeing what is so, if it is like the act of saying what is so, consists in, and is constituted by, psychical acts the occurrences of which secure the object of reference of the sensuous judgment and ascribe to it what we manifestly see the object to be. Failing this, perceptions are not a species of judgment. Or, failing this, the thought-speech analogy is inadequate to characterize interestingly the features of thought which made it seem helpful to construe thinking as covert verbalization. But if the analogy is interesting, and perceptions are, or incorporate, judgments, where in a perception are its referring constituents, and where are its ascriptions? What constituent occurrent acts constitute the act of perception?

The role of sense impressions relative to perceptions has, I believe, been miscast. It is sense impressions which constitute the sensuous content of perceptions and whose occurrences determine their references. Perceptions are not, after all, complex occurrences the proper accounting for which requires double entries from the areas of conceptualization on the one side and from sensory contexts on the other. Sense impressions do tie our sensuous judgments to the contexts of their occurrence, but they do so not because they bear some special, but unintelligible, relation to what is essentially conceptual and intelligible but non-sensuous. They do so because they are, literally, constitutive of perceptions, at least basic ones, and because their occurrences have demonstrative force. Basic perceptions we understand to be those sensuous judgments which ascribe the proper and common sensibles to what is demonstratively before one. (We see that this, I know not what, has moved, or we hear that something rustled in the next room.) Most of what we perceive, and most of what is interesting in perception, outruns our basic sensuous judgments. (We hear, perhaps, that the carburetor needs adjusting.) The relation of sense impressions to these last, ordinary but complex perceptual judgments, remains undertermined here, though I believe that the notion of a basic perception is helpful in answering the question of how much we see, in an arbitrary act of perception, or, of how many perceptual acts have occurred on a given occasion. In any cas,e, the present point is this: sense impressions are fully constitutive of basic sensuous judgments. They are not the contexts in which such judgments occur. They make up such judgments. They are, then, after all, conceptual entities. The analogy is this: "as predicates are to declarative sentences, as ascriptions are to assertions, as concepts are to judgments, as ascribing is to stating, so, too, sense impressions are to basic perceptions" ([4], 12, and [6], Chs. 3 and 4). The occurrence, the having, of sense impressions in the natural order is the analogue of the production and use of material tokens of expressions in the conventional order. Seeing how something is, in a basic act of perception, is the ascription of the sensible content of a sense impression to what is before one. The occurrences of sense impressions in acts of perception are the vehicles for the psychical ascriptions of sense qualities to what is before one, just as the use of predicate expressions is the conventional means of characterizing an object of verbal reference. In seeing that an object before us is, say, red, we ascribe, in the standard case, the content of our sense impression to the object seen. The occurrence of an impression of a proper sensible, any at all, in any context, is a putative ascription. We come to learn to withold on occasions such ascriptions, and we come to learn to modify what it is that we ascribe. (Red things in ultra-violet light, we learn, do not stimulate sense impressions of red. Sense impressions in such occurrences are analogous to the presence of homonyms in conventional natural languages, and the recognition of homonymous occurrences is something we come to acquire. Cf. [4], 14-15.)

But it is the occurrence of the sense impression in acts of basic perceptions, as distinct from their contents, which is finally crucial in determining the reference of those perceptions. Occurrences of sense impressions are fully demonstrative. A basic act of perception, consisting in the occurrence of a sense impression, is a putative ascription a schema for the verbalization of which might run like this: "This, sensuously before me, is qualitatively thus and so" ([4], 22).

We can further exploit the thought-speech analogy to characterize sense impressions semantically in the fashion in which verbal expressions are so characterized. Semantically, sense impressions are natural ascriptives which, like the predicates of conventional languages, can be construed as functions taking arguments into truth-values. Sense impressions are rather special functions for they are defined only over a special and limited domain. They are functions taking the demonstrative references embodied in their own occurrences as arguments. Jones' basic sensuous judgment that this, before him, is red is true just in case the object of the demonstrative reference carried by the occurrence of the sense impression has the sensible quality which the content of the sense impression ascribes on that occasion of its occurrence.

The semantical parallel is not particularly illuminating. It provides no special understanding of the relation of perceptions to what is perceived. Rather, it is invoked to reinforce two points. First, within the framework of the thought-speech analogy (shared by Aune, Sellars, and Geach), there is a natural answer to the puzzling question of what in thought and perception answers to the occurrence of referring and ascriptive expressions in assertive uses of sentences. Second, given that the analogy is extended in this natural way, the semantics of (the truth-conditions for sentences with) conventional demonstratives can quite literally be applied to perceptions. We have, thus, a resolution of Geach's Puzzle of how perceptions with the same intelligible content can get tied to different objects of reference. Since occurrences of sense impressions are, on the extended analogy, demonstratives in the natural, psychical order, two sensuous judgments with the same content will be referred to different objects precisely in the manner in which conventional demonstrative assertions may be. In the case of perception, this is, as Geach observed, a matter of the sensory context of the judgment. But if the sensory context consists in the occurrence of sense impressions, and if these have, in the natural order, demonstrative reference, then the manner in which the perception is referred to what is perceived is no more or less mysterious than is the manner in which demonstrative assertions refer to what is talked about.

The moral of the story is this: on the extended analogy, sense impressions are the psychical analogues of the quality predicates of conventional assertions; the occurrences of such sense impressions are the demonstrative acts of perceptual reference analogous to the use of demonstrative pronouns in overt acts of assertion. All this can be true only if sense impressions are conceptual constituents of perceptual acts of judgment.

[1] Aquinas, Saint Thomas, Basic Writings of Saint Thomas Aquinas, Volume One, edited and annotated with an Introduction by Anton C. Pegis (New York: Random House, 1945).

[2] Aristotle, Aristotle's De Anima in the Version of William of Moerbeke and the Commentary of St. Thomas Aquinas, translated by Kenelm Foster, O.P., M.A. and Silvester Humphries, O.P., M.A. with an Introduction by Ivo Thomas, O.P., M.A. (London: Routledge and Kegan Paul, 1959).

[3] Aune, Bruce, Knowledge, Mind, and Nature (New York: Random House, 1967).

[4] Clark, Romane, "The Sensuous Content of Perception," typescript, forthcoming in a Festschrift honoring Professor Wilfrid Sellars: 27 pp.

[5] Geach, Peter, Mental Acts (London: Routledge and Kegan Paul).

[6] Hall, Everett W., Our Knowledge of Fact and Value (Chapel Hill: The University of North Carolina, 1961).

[7] Montague, Richard, "Pragmatics", Contemporary Philosophy, edited by Raymond Klibansky (Firenze: La Nuova Italia Editrice, 1968): 102-122.

[8] Sellars, Wilfrid, Science, Perception, and Reality (London: Routledge and Kegan Paul, 1963).