Published in Commonsense Realism: Critical Essays on the Philosophy of Everett W. Hall, edited by E. M. Adams, The Southern Journal of Philosophy 4 (1966): 103-15. Reprinted in Philosophical Perspectives (Springfield, Illinois: Charles C. Thomas, Publisher, 1967;


Wilfrid Sellars


Everett Hall's intentional realism is an example of systematic philosophy at its best. It is no myopic sequence of small scale analyses strung together like beads on a string. Yet its foundation was laid over the years by painstaking and scrupulous probings into the many problems and puzzles with which a systematic philosophy must deal. Again, though it is rooted in a sympathetic and perceptive interpretation of the philosophical classics, it is as contemporary as the latest issues of Mind. Few philosophers have taken as seriously the obligation to keep in touch with the best work of their contemporaries. He recognized that it is only by submitting ideas to the constant challenge of other lines of thought that philosophers can gain assurance that their speculations are not sheltered idiosyncracies. Everett Hall's philosophy is thoroughly empiricist in temper, but completely lacking in the procrustean urge which has marred so many recent ernpiricists. Above all, it is in a most important sense self-conscious or self-referential in that it includes as an essential component a theory of the philosophical enterprise, a theory which faces up to the ultimate challenge which any systematic philosophy must face: What is the status of your philosophical claims, and what are the criteria by which you distinguish them as true from the false and unacceptable claims made by rival philosophies? Thus, by no means the least important of his achievements is the way he found between the horns of Hume's dilemma, which in modern dress, reads as follows:

Philosophical statements are either analytic (in which case they tell us nothing about the world) or synthetic (in which case they fall within the scope of empirical science).

Inspired by this dilemma, Hume was willing to throw all distinctively philosophical statements into the flames. (One wonders whether he would have done so with his own.) The Wittgenstein of the Tractalus denied that there are any distinctively philosophical statements. What purport to be such he found to be therapeutic devices which can be cast aside, perhaps into the flames, once they have served their purpose. Hall offers instead a conception of philosophy as "neither a priori nor empirical."1 By "empirical" he has in mind, I take it, the inductive methods of the empirical and theoretical sciences. He argues for "a third kind of knowledge" which he calls "categorial" (Hall, p. 6). The test of claims failing within this "third enterprise" is to be found "in the forms of everyday thought about everyday matters in so far as these reveal commitment in some tacit way to a view or perhaps several views about how the world is made up, about its basic 'dimensions'" (Hall, p. 6). "We find," he continues, "these forms of everyday thought chiefly in the grammatical structures (in a broad sense) of daily speech, in what may be called the resources of ordinary language, although they are also present in the ways in which we personally experience things. . . . The latter," he adds, "reflect, to a great extent, the formative influence of our mother tongue," (Hall, p. 6) a theme to which I will return.

This characterization of the philosophical enterprise illustrates once again the catholicity (i.e., the universal sweep) of Everett Hall's philosophy, for, in my opinion, this conception of philosophy is the truth to which both the descriptive phenomenology of Husserl and the conceptual analysis of the developing phase of Oxford philosophy are halting approximations.

But my concern in this essay is not the Hallian metaphilosophy, but with certain specific theses in his philosophy of mind -- theses pertaining to his realistic analysis of perception. The central theme of Hall's intentional realism is his interpretation of perceptual acts as sentence-events in a natural language (i.e., as items which are analogous to sentence-events in conventional languages). This move enables him to put such concepts as meaning, reference, predication, ascription, truth to effective use in clarifying the relation between the act of perceiving and the object perceived. It must be emphasized in this connection that although sentence-events in natural language are conceived by analogy with sentence-events in conventional language, this does not mean that the meaning, reference, and truth which pertain to natural language are derivative from the meaning, reference, and truth which pertain to conventional language. We must distinguish here, following Aristotle, between priority in the order of knowledge and priority in the order of being. We come to understand the nature and function of natural language through conceiving of it by analogy with conventional. But in the order of being, natural language is prior. As Hall sees it, if there were no such thing as natural language with its modes of meaning, refrence, predication, and truth, there could be no such thing as conventional language.

Hall's conception of natural language is in the tradition of William of Ockham and revitalizes insights which were lost in the heyday of British Empiricism. A glance at Hume will be instructive, for although, like all empiricists, Hall has learned much from Hume, the guiding thread of the intentionality of the mental (i.e., of those features of mental acts by virtue of which they are analogous to conventional speech) enables him to avoid a key confusion which occurs at the very beginning of the Treatise and infects the argument on almost every page.

It will be remembered that the human mind, according to Hume, is a system of impressions and ideas. If we leave aside the impressions and ideas which constitute the life of feeling and emotion and concentrate on those which pertain to our experience of the physical world, we can take as our paradigm of a Humean impression a visual impression of a red triangle. Now a careful study of the Treatise makes it clear that there are two incompatible themes in Hume's account of the sort of item referred to by the italicized phrase:

  1. An impression of a red triangle is a red and triangular item which is immediately and non-inferentially known to exist and to be red and triangular.
  2. An impression of a red triangule is a knowing that a red and triangular item exists.
According to (1), the impression is an object of knowledge. It is literally red and triangular, and is directly nown to be so. It is an object of knowledge, but scarcely, it would seem, a knowing. For, even if a knowing could be literally red and triangular, its character as a knowing could not consist in its being red and triangular. Furthermore, if there were such a thing as a red and triangular knowing, it could scarcely be its own object. The object known must surely be other than the act by which it is known.

According to (2), the impression is a knowing rather than a known, and it is the known rather than the knowing which is literally red and triangular.

Alternative (1), by using the term "impression" for objects of immediate knowledge, has no category for the acts by which they are known, and hence no place for the intentionality of perception. Alternative (2), by using the term "impression" for acts of knowing, has no category for the objects of immediate knowledge. The conflation of these two alternatives provides both knowings and objects to be known, but at the expense of a radical ambiguity in the use of the term "impression" and a confusion, which Hall points out, of the relation between a perceiving and a certain attribute -- e.g., triangularity --by virtue of which the perception signifies or intends the attribute with the relation which holds between something which is literally triangular and the triangularity it exemplifies.

Against the background of this radical confusion, the strategy of Hall's intentional realism stands out in bold relief. What Hall calls perceptions correspond by and large to Hume's impressions in those respects which take intentionality into account. With this in mind, we can draw up the following points of similarity and difference:

  1. Hall's pcrceptions -- like Hume's impressions -- are states of the percipient which are causally evoked by the action of external objects on his sense organs.
  2. Hall's perceptions -- unlike Hume's impressions -- are not the objects of perceptual knowledge. They can, indeed, be known in what is misleadingly called introspection. But they are not perceived. They are perceivings rather than perceiveds.
  3. Hall's perceptions -- unlike Hume's impressions -- are not knowings. The term "know" in its primary use implies
    1. truth (i.e., to say that a sentence (conventional or natural) expresses knowledge is to endorse the sentence as true) and
    2. conclusiveness or certainty. But on Hall's analysis perceptions can be false, and even true perceptions are usually only highly probable, supported by their consilience with other perceptions.2 Hall's perceptions might well be called perceptual convictions about physical objects and processes. They are perceptual convictions that something is the case, e.g., that a certain object is red and triangular.

Before continuing with our enumeration, it might be noted that there is a strand in Hume's treatment of impressions according to which they would be knowings only in a weakened sense of this term in which it amounts to "convictions." According to this strand, "impressions" are intrinsically indistinguishable from very vivid "ideas." Hume's account of "ideas" is at least as confused as his account of "impressions;" thus sometimes the "idea" of a red triangle is a red and triangular item deserving the name "impression." But, as in the case of impressions, the theme of intentionality is not lacking. Here "ideas" are thoughts which, according to their degree of forcefulness, range from "merely thinking of," e.g., the existence of a red triangle, through various degrees of belief in the existence of a red triangle, to the conviction, indistinguishable from an impression, that a red triangle exists. The fact that impressions are located in a continuum with ideas suggests that when impressions are described as knowings, the latter term is to be construed in a way which does not imply endorsement or certainty in any other than the subjective sense of conviction. This, of course, would be the point of view of Hume within his study; outside, the normative overtones of the credibility and authority of perception would reappear.

  1. According to Hall's account, the perception of a red triangle is neither red nor triangular. That is to say, it does not literally exemplify either of these attributes. On the other hand, the attribute is present in the perception in another manner (i.e., a manner other than that of being exemplified). As Hall sees it, one of the primary sources of skepticism and phenomenalism is the idea that "the only way properties are present in the world" (Hall, p. 37) is by being exemplified.
  2. Since perceivings, on Hall's account, are perceptual convictions about physical objects and processes, the problem of skepticism is not that of what authorizes an inference from impressions as directly known objects to their putative causes -- a source of skepticism which is rooted in the first strand of Hume's account of impressions. The problem is rather: Are impressions as perceptual convictions (i.e., the impressions of the second strand in Hume) capable of being true? And, if so, can we have reason to suppose that any of them are? The Hume we were sketching a moment ago would say, on Berkeleyan grounds, that they are not capable of being true and would add that, even if these Berkeleyan grounds were mistaken and they could be true, we could have no reason to suppose that any of them are. Hall's answer to both these questions is categorically affirmative.


Let me now develop a second background against which it will be helpful to view Hall's intentional realism. I have in mind the Cartesian distinction, derived from Scholasticism but given a distinctive twist, between formal and objective reality. This distinction concerns two modes of being or reality which entities can have:

  1. Things and their attributes can have "formal" being; things, including minds, by existing in the world and attributes by being exemplified by things.
  2. Things and their attributes can have "objective" being in acts of thought. Thus Socrates has objective being in any mind that thinks of him, and wisdom in any mind that thinks of something as being wise.
Notice that an object, e.g., Socrates, which has objective being in a certain mind is not in that mind in the sense of being a part or aspect of the mind -- let alone being spatially within it. The object is not, as the metaphor has it, "bodily present" in the mind. Succinctly put, objective presence is not bodily presence. Again, an attribute, e.g., triangularity, which has objective being in a certain mind which is thinking of something as triangular, is not in that mind in the sense of being exemplified by the mind. Obviously the thought of something as triangular is not a triangular thought.

This distinction between the "objective" presence of things and attributes in thought and their formal presence in the world permits the formulation of a simple theory of the truth of at least certain kinds of thought. Roughly, a thought to the effect that a certain object O has a certain attribute A is true if and only if the object O, which is objectively present in the thought, is formally present in the world, and the attribute A, which is obiectively present in the thought, is exemplified by O.

Now, in the case of the attributes involved in perceptual convictions, this distinction between "objective" and "formal" presence corresponds to Hall's distinction between "present as experienced" and "present as exemplified" (Hall, P. 35). Thus, if I have a perception of a red object and the perception is veridical (i.e., there really is a red object before my eyes), then red is present as experienced (i.e., objectively present) in the perceiving and this same attribute, red, is also present as exemplified in the object perceived. Again, Hall emphasizes that the objects of perception are not "bodily present" in the act of perception, but only present as experienced (Hall, pp. 72, ff.). In Cartesian terms, the objects of perception as well as the attributes they are perceived to have are only objectively present in the act of perception. All this is compatible, as Hall emphasizes (Hall, pp. 35-36, 74 ff.), with the idea that perceptual and other mental acts exemplify attributes, e.g., temporal and spatial relations, and even neurophysiological properties, in their own right.


The Cartesian also distinguishes between the act of thought qua having things and attributes objectively present in it and the act of thought qua existent exemplifying attributes in its own right. His account of acts of thought in the latter respect, however, is remarkably thin. Acts of thought exemplify temporal relations, and they exemplify the relation of belonging to or occurring in a or occurring in a mind as well as such attributes as tending to be followed by certain other thoughts. They exemplify the derivative attribute of having certain items objectively present in them. But, one wonders, what is their qualitative character in their own right? Here one runs up against the puzzling view that qua acts of thought, though they may differ in their relations and their derivative attributes, they are intrinsically indistinguishable. One is put in mind of the dialphanou character which G. E. Moore ascribes to awareness -- though Moore's general approach is quite different.3

Hall avoids this puzzling conception by rejecting the classical interpretation of self-awareness on a perceptual model (Hall, p. 16). To construe direct self-knowledge as akin to sense perception is to look for qualitative features in mental acts as the objects of introspection which would correspond to colors as qualitative features of things seen and sounds as qualitative features of things heard. Since no such qualities are to be found, the fictitious quality of "diaphanousness" is invented to fill the gap. This generates the classical problem of how mental acts with their diaphanousness are related to non-diaphanous states of the body. By refusing to construe self-knowledge on a perceptual model, Hall is enabled to suggest that in their intrinsic being perceptual acts are neurophysiological events. For since self-knowledge does not assign them any qualitative character, let alone a diaphanous quality inconsistent with the qualitative character of neural events, it is open to us to hold that their qualitative character of a neural events, it is open to us to hold that their qualitative character is that of a neural event.

Notice that although perceivings would be neural episodes, which requires that they have properties which we do not ordinary thunk of them as having, this identification also requires that certain neural episodes have properties which we do not ordinarily think of them as having. For the neural episodes in question would have the property of having objects and attributes objectively present in them in the mode of being experienced as contrasted with that of being exemplified. Hall, as far as I know, explicitly applies this analysis only to perceivings and bodily feelings (i.e., to mental acts which belong to the natural language of perception and feeling). It is, however, readily extensible to all mental acts without exception, and so extended would construe them as neural events which "have" the derivative property of having objects and attributes of various kinds and types objectively present in them. This would raise the interesting question of what correlation might exist between these derivative properties of neural events and what might be called the natural characteristics of neural events (i.e., their character as objects of scientific description and explanation).


We are now in a position to examine more closely Hall's conception of a perceptual act, or perceiving. We have seen that, although perceivings are neural events, they are perceivings by virtue of the objective presence in them of objects and attributes (i.e., the presence of the latter as experienced). Let us relate this to our earlier account of them as sentence-events in the natural language of perception. The reference to sentence-events reminds us of Peirce's distinction between linguistic types and their tokens. (The word "the" as type is a repeatable; there are many tokens of "the" on this page.) Hall refers to a sentence-type in natural language as an intention (Hall, p. 74). Thus a sentence-event in perceptual language is, so to speak a token of an intention. Intentions, he tells us, "are like properties in being incomplete, by nature dependent on something else, and universal in the sense that the same one may belong to several events" (Hall, p. 74). He distinguishes the sense of "intention" he has in mind from the sense which pertains to action by calling them "references" (Hall, p. 74). This, however, is misleading in that it suggests that intentions are the natural language counterpart of referring expressions rather than sentences. Not only is this not the case, but one of our problems will be to see in what sense there are any referring expressions in natural language as construed by Hall.

Intentions, he tells us, have a "certain inherent complexity. They may be said to include ordinary properties" (Hall, p. 75). Thus the intention tokened by the perception that an object is triangular includes the attribute triangularity.

Now it is clear that if the intention were simply a complex property, as being-red-and-triangular is a complex property, its relation to the mental act could not be that of exemplification. For if the intention were a complex property of which triangularity were a part, and if the perceiving exemplified the intention, then the perceiving would a fortiori be in part literally triangular -- which is exactly the thesis he is most concerned to reject. At this point, Hall could make either of two moves:

  1. He could hold that an intention is a complex property but deny that the relation between the perceiving and the property is that of exemplification;
  2. He could hold that the intention is a unique kind of complex which, although it includes properties, is not a complex property.
Actually he makes both moves. He characterizes the relation between an act and the intention which it "has" or which "belongs" to it as analogous to but not the same as exemplification. And he emphasizes that intentions are analogous to properties but unlike them in key respects. Finally, he stresses that although "they may be said to include ordinary properties," they do so "neither as exemplified by themselves (i.e., the intentions) nor by what has them (i.e., the neural events), but, instead, by their objects." He concludes that "intentions can themselves be called natural or radical signs, identical with their objects (in veridical perception) in the quality or character of those objects' properties, but non-identical in the factor of exemplification (in place of exemplification intentions have ascriptions)" (Hall, p. 75).

This important passage is rather cryptic, but I take it that the term "ascription" characterizes the inner structure of the intention, and is intended to differentiate the place of triangularity in an intention from its place in a complex property.

Perceptions, then, ascribe properties to objects. How is this to be understood? We might begin by construing a perceptual intention as a perceptual sentence of the form .(object) exemplifies (property). where the dot-quotes indicate that the intention is analogous to a sentence in conventional language, while the fact that they are special quotes indicates that after all it is not a sentence in conventional language. If we assume provisionally that the intention contains an expression which refers to an object and an expression which refers to a property, then for the intenfion to ascribe the property to the object is for the intention to consist of three items uniquely related: a natural word for the property, a natural word for the object, and a natural word which means exemplified.

Now Hall emphasizes that in the natural language of perception, "red as experienced names, designates, means, immediately the red as exemplified; the red of the perception refers to the red of the sunset, the red in one capacity or status means the red in the other" (Hall, p. 35). Here, he tells us, "we, have the basis of designation, designation as it occurs originally, prior to conventional rules" (Hall, p. 35). Thus the intention includes in a unique way the property itself and the nexus of exemplificcation. If we assume that the intention includes in the same unique way the object itself, then we could think of natural words as the objects and attributes themselves as occurring in intentions. Intentions, then, would be unique wholes of which objects and attributes would, in a unique way, be parts. I suppose that this is true enough, for the term "unique" is elastic enough to cover a great deal of ground. Yet, at least as far as objects are concerned, we run into the difficulty that an intention may contain an object, even though -- as in the case of hallucination -- there is no such object. But I shall not press these points, since they lead to some of the most intricate issues in ontology. Suffice it to say that Hall is well aware of this difficulty (Hall, p. 75) and that the above remarks are not intended as a criticism.

But there are other aspects of perceptual intentions as characterized by Hall which I find quite puzzling and which lead me, not to doubt the truth of his intentionalistic analysis of perception, but to wonder if his natural language is rich enough to do the job.

To bring this out, it will be helpful to explore the structure of perceptual intentions by scrutinizing the conventional sentences which give them expression. In this connection, Hall draws a most important distinction between "perception proxies" and "perception depicters" (Hall, pp. 38 ff.). Failure to observe this distinction he finds to be another source of subjectivism and skepticism. Perception depicters are biographical statements which assert the occurrence of a perceiving -- e.g., "First X saw a bright red flash, then X heard a rumbling noise." Notice that, although perception depicters can be about anybody at any time or place, the most important subset for our purposes consists of those which concern myself here and now -- for it is these which can be confused with perception proxies.

Perception proxies are, in the first instance, translations (Hall, p. 38) of selected features of the total perceptual experience into conventional language. They are, indeed, the closest conventional counterpart of sentences in the natural language of perception. An example might be "This is red and triangular."

Now perception proxies have several features which are not present in the perceptions they translate. Thus, to begin with, perceptions, unlike perception proxies, "contain neither demonstratives nor proper names" (Hall, p. 42). This raises the question: How is the reference of a perception to its object to be understood? Surely, we are tempted to expostulate, there must be expressions in natural language which are the counterparts of referring expressions in conventional language, as properties-as-experienced are the counterpart of conventional characterizing expressions.

One might be tempted to suppose that, lacking demonstratives and proper names, natural language contains the counterpart of definite descriptions. Perhaps the perception which translates into "The book over there is red" has a natural language definite description as its subject. But Hall rejects this move on the ground that Russell's analysis (and, for that matter, Strawson's) shows definite descriptions to involve such notions as "all," "identical with," "if. . ., then. . .," in short, all the conceptual apparatus of propositional and functional logic, and these, as we shall see in a moment, are also not present at the perceptual level proper, but belong exclusively, as Hall sees it to conventional langyage.

Hall emphasizes that conventional language, once we have it, blends with natural language, so that perceptual experience proper is enriched with "in the head" use of conventional expressions. But, he writes:

[Pure perception] contains neither demonstratives nor proper names. This is perhaps partly just a verbal specification. When I meet with perceptions containing demonstratives or proper names, I refuse to call them "pure." But it is in part observational also. I find that I have perceptions which are descriptive or predicative or predicative throughout and are in fact quite devoid of conventional symbols. (Hall, p. 42, italics mine.)

What puzzles me about this is how a pure perception can be a sentence, and yet be "predicative throughout." Are not predicative expressions the correlatives of referring expressions? Must not pure perceptions contain expressions referring to an object in order to be able to characterize an object? Hall writes:

In pure perceptual experience, we have batteries of predications, so to speak. These are "of" individuals; and one individual is selected in perception for predication simply by the whole battery of predications made of it, the "it" being that which is asserted by the perception to exemplify them all together with an indefinite number of unspecified other properties from which the perception has selected just the ones whose assertion constitutes that perception (Hall, pp. 42-43).

This suggests that while perceptions contain neither demonstratives nor proper names, they do contain expressions which are the natural language counterparts of an indefinite "it" and have the form

·It exemplifies /properties/·

Yet it must be granted that this interpretation is not easily reconciled with the above claim that pure perceptions are predicative throughout. On the other hand, it is confirmed by such statements as the following:

Although I have no way of getting at [the chair] perceptually save through properties I experience it as having, still it is not these properties. My experience is of the properties as exemplified. . . . (Hall, p. 44, my italics.)

Perception is the predication of these properties, the ascription of them to an obiect (Hall, p. 73).

. . . Perceptual experience . . . always affirms that there is something that exemplifies the multiplicity of its groups of predicated properties. . . (Hall, p, 47).

The last of these would have to be construed as a misleading formulation of the above analysis, for taken as it stands it suggests that perceptions have the form

·(Ex) x exemplifies /properties/·
which is ruled out by the absence, on which Hall insists, of logical connectives and quantifiers from natural language.

To sum up this first comparison of perceptions with perception proxies, Hall thinks of pure perceptions as containing a primitive or rudimentary mode of reference, though the account is not unambiguous. On the other hand, the sophisticated devices of demonstratives, proper names, and definite descriptions are lacking at the level of pure perception, and make their way into perceptual experience only through the blending of pure perceptions with conventional language.


What, then, of logical conneclives and operators (i.e., such expressions as "not," "if . . . then . . .," "all," etc.)? Hall insists, as has already been pointed out, that these are not to be found at the level of pure perception. He claims, however, that their use in conventional language does have perceptual underpinning. Let us begin our exploration of this contention by considering the case of "not." Here the thesis is by no means lacking in plausibility, for surely, as Hall puts it, "to all appearances our perceptual experience is positive throughout" (Hall, p. 49). Whether or not this is actually the case, and if so in what sense the phrase "perceptual experience" is being used, are questions for subsequent exploration. But certainly there are many who would agree with Hall on this point. Let us grant, then, at least for the moment, that we see that a certain object is red, but do not see, at the level of pure perception, that it is not green.

Yet, as Hall points out, singular negative perception proxies do occur. Consider the perception proxy, "This is not green." Now that something is not green can be inferred from the fact that it is red, for these attributes are incompatible. But how do we know this? As a staunch empiricist, Hall replies -- by generaliztion. Of course, the generalizations -- roughly, "(x) x --> ~green x" and "(x) green x --> ~red x" -- cannot be from pure perceptions by tke route

O1 is red, round, not green . . .
O2 is red, square, not green . . .
O3 is green, oval, not red . . .
for there is no pure perception of objects as not green or not red. Rather, as far as I can make out, the generalizations take their point of departure from positive perception depicters, and instead of being instantial inductions are of the nature of hypotheses to explain the fact that we never perceive green objects which are red or red objects which are green. I suspect that lurking in this ingenious solution is a corresponding problem about negative self-knowledge, thus "I am not perceiving a red object which is green;" but it is certainly an interesting move and far more penetrating than most discussions of the subject.

But even granting the success of Hall's move with respect to "not," there are other problems to be met. Thus while philosophers since the time of Parmenides have been bothered by the little word "not," they have paid less attention to "or" and far less to "and." Yet it is a familiar feature of modern logical theory that "and," "or" and "not" are the same sort of word. Thus one would expect that if there are there are no negative pure perceptions, there are also no conjunctive pure perceptions. What would this involve? To come immediately to the point, it would seem to conflict with the account of perceptual "sentences" at which we arrived in the preceding section. It will be remembered that according to this account, these sentences are of the form

O exemplifies /properties/
But surely this must be interpreted as
O exemplifies P1 and P2 . . . and Pn
which is equivalent to
O exemplifies P1 and O exemplifies P2. . . and O exemplifies Pn.
Hall speaks on p. 44 of a "group of predications about a single individual." On p. 42 he equates a "battery of predications" being made of an individual with the object's being "asserted to exemplify them all together." But this identification is debatable, to say the least, for it amounts to identifying a set of sentences with the single sentence which is their conjunction. The point at which I am driving is, in Kantian terms, that the perception of a manifold must not be confused with a manifold of perceptions. Thus, if we leave "and" out of the language of pure perceptions, one could not pure-perceive an object to exemplify a number of properties all together. This conjunctive togetherness would, on Hall's account, have to be grasped through conventional language. The perception proxy
(O is P1 and P2 . . . and Pn)
would have to be interpreted as involving a move from the more basic perception proxies, "O is P1," "O is P2," etc., which do translate pure perceptions, via what logicians call "Conjunction Introduction;" schematically,
p and q

Another apparent consequence of the denial that the logical connectives are present in pure perception would be the restriction of the attributes which could be ascribed to objects in pure perception to what Aristotelian's call the "proper sensibles." It would exclude not only negative and conjunctive attributes, but also disjunctive attributes. Since I think that determinable or generic attributes are a special case of disjunctive attributes, this would mean that no pure perception could be generic. I believe that this consequence would disturb Hall, since I know from personal conversation that at one time (1955) he thought they could.

That the exclusion of logical connectives from pure perception would also exclude the pure perception of causal attributes, dispositions, and propensities is obvious, but would not disturb Hall, since he would regard all concepts of this kind as involving inductive generalization, and hence as derivative from basic perceptual experience. Thus the fact that we perceive objects as stones, pieces of paper -- classifications which clearly involve reference to causal properties -- would simply highlight the pervasive role played by conventional languag in sophisticated perceptual experience. We have already noted that he emphasizes on many occasions that, although the core of perceptual experience is the innate natural language of pure perception, as soon as the child learns a conventional language, the latter becomes almost seamlessly intertwined with the former. It becomes a second-nature language of perception.

. . . Now that we have this sharp distinction [between natural and conventional language] it is necessary to admit that in everyday situations we frequently have mixtures and cases that are difficult to classify. I remember witnessing a floating pile-driver capsize in the harbor of our yacht club and, while seeing a man scrambling down the scaffolding, suddenly hearing myself exclaim, "Why, it's Ivan Thorp and he can't swim!" Now it would be easy enough to separate this into two parts, one in the natural language of perception, the other in conventional English, and this would be justified. But such a procedure is not wholly fair to the situation. My perceptual identification of the man was in part effected by my uttering of his name; the latter was not just a translation of a recognition of him that had already occurred. Indeed, most of us talk to ourselves most of the time, so that pure perceptions, perceptions devoid of all conventional expressions, are very rare (Hall, p. 33).


Let me pull some of these considerations together. As I see it, they add up to something like the following. While I wholeheartedly agree with the fundamental thesis of Hall's intentional realism, the thesis that perceivings are to be regarded as tokens of inner or mental sentences, I doubt that the language to which these sentences belong can do the job required of it unless it contains the fundamental logical apparatus found in conventional language. I think that this point would stand out even more clearly if HalI had developed an account of self-awareness in terms of an inner language of what, for lack of a better word, I will call introspection. Kant's decisive point about the unity of apperception or self-consciousness, namely that the consciousness that

I who have experience A am identical with I who have experience B
requires not just the occurrence of the two self-awarenesses
I have experience A
I have experience B
but the single "combinative" awareness
I have experience A and (I have) experience B.
It is around this (and related) points that Kant built his transcendental deduction of the logical forms of judgment. In this general respect my position vis-a-vis Hall is analogous to that of Kant vis-a-vis Hume.

Thus it is my conviction that even the most basic "inner" or "mental" language must contain the logical operation of coniunction, and, once this is granted, there is no place for boggling at the remaining logical forms.

It is important to note, however, that to allow that the inner language of perception contains logical expressions does not require one to say that negative, conjunctive, disjunctive, or quantified percepfions are completely on a footing with singular affirmative perceptions. It is compatible with the claim that there is a sense in which rock-bottom perception is a matter of singular, affirmative sentences which attribute perceptible attributes in a strict sense to objects. Thus it could be argued that conjunctive perception involves a "conjunction introduction move" from the component perceptions. The important thing would be that the inner language of perception proper contained the logical apparatus for making this move, so that the move could occur in it rather than after the translation of the conjuncts into another language ("conventional perception proxies").

Again it might be true that, although the inner language of perception contained quantification, no "rock-bottom" perception is quantified, quantification coming in as a move from the perception ·O is red· to ·(Ex) x is red·; but, as I see it, these "rock-bottom" perceptions would simply be the ones we have when our mental set is cautious and skeptical. When less cautious, we simply perceive in richer terms without "making moves from rock-bottom." (We see that the there is no cheese in the refrigerator, to use a well-worn example.)

Thus it is my conviction that the inner language of perception and, I would add, self-awareness is a language in the full-blooded sense; it requires the presence of logical expressions and principles of inference, and I must confess that I do not understand why Hall committed himself too thoroughly to the opposite position. I have two suggestions to make.

Perhaps something like the classical Aristotelian dictum ("Nihil in intellectu. . . ") was operating in the form "Nothing in the basic inner language of perception that is not to be found in the world as perceptible by the senses." Thus Hall might argue that every feature of the basic language of perception must be something which is formaIly present in the world and derivatively capable of being objectively present in percipient organisms (i.e., present as experienced). Add to this the premise that logical words do not stand for anything in the world and the thesis would follow.

But logical words function so differently from referring and characterizing expressions that the principle

No feature can be objectively present in the basic language of perception which is not formally present in the world (i.e., "bodily" in the case of objects; "as exemplified" in the case of attributes)
might well be restricted to objects and attributes. In any case, to put logical words into the basic inner language of sense perception is not to commit oneself to the absurd idea that logical words can be defined in terms of words for perceptible qualities and relations.

Incidentally, Hall is, in my opinion, absolutely correct in attacking (Hall, p. 48) those who rugard logical words in conventional Ianguage as (in principle) dispensible. My contention is that they are equally indispensible features of even the most basic inner language of sense perception -- and, I would add, of introspection.

Is there any other reason why Hall would have felt uneasy about locating logical words in the inner language of perception? I think that there is, and I also think that it highlights one of the central problems in the philosophy of mind. It concerns the innateness of the basic inner language of perception. Hall emphasizes that this language is a natural language the use of which requires no learning, only maturation. The grasping of logical connections, on the other hand , is conceived by Hall as the sort of thing that does involve learning and, indeed, is acquired in the course of learning to use a conventional language.

We can sketch the situation in terms of an inconsistent triad:

  1. The understanding of logical words requires learning.
  2. The basic language of perception does not require learning.
  3. The basic language of perception involves logical words.
Hall accepts (1) and (2) and therefore rejects (3). I, on the other hand, accept (1) and (3) and, therefore, reject (2). Thus, whereas Hall sees the basic language of perception as an innate language which is the original on which is the original on which the richer conventional language of interpersonal communication is based, I see "inner language" in general, and perceptual inner language in particular, as nothing other than conventional language itself. It is conventional language as it exists in what Hall refers to as "talking to oneself," but which I prefer to think of as truncated thinking-out-loud. As we have seen, the notion of the occurrence of conventional language in "talking to oneself" is central to Hall's interpretation of perceptual experience as it exists in people who have learned to use a conventional language. But the very difficulty Hall finds (Hall, p. 33) in distinguishing between the natural language of pure perception and the conventional enrichment which is so seamlessly entwined with it suggests that there is something wrong with Hall's linguistic dualism.

On the other hand, it must be granted that perceptual experience does involve abilities which mature and are not learned. The abilities I have in mind enable the perceptible qualities and relations of physical things to be present in a unique way other than by exemplification -- in perceptual experience. But whereas Hall sees only two ways in which qualities and relations can be present in something,

present as exemplified
present as signified, meant, or intended
I think that there are three modes of presence,
representation by sense impressions,
and that the innate abilities involved in perception involve the third of these modes, a mode which is not to be construed by analogy with language and does not constitute an innate natural language of perception.4

But important though this difference between us may be, it is a minor difference compared to the major issue between realism on the one hand and skepticism, subjectivism, and phenomenalism on the other. For the essential thesis of Hall's intentional realism is that perceptions are to be construed as inner sentence events which refer directly to external objects and predicate perceptible qualities and relations of these objects themselves -- not copies or replicas of them. And this I think to be the very heart of a sound theory of perception.


1Everett Hall, Our Knowledge of Fact and Value. Chapel Hill, 1961, p. 5. References made solely by page number will be to this work.

2One of the most interesting moves Hall makes is to argue that each perception, merely by virtue of the fact that it occurs, has a certain intrinsic probability, to which additional credibility can be added by the mutual support of perceptions in accordance with principles of perceptual confirmation. These principles, as well as the principles of intrinsic probability, are given a metaphilosophical defense by pointing out that they are framework features of perceptual experience.

3For an elaboration of this point see the opening section of my "Being and Being Known," Proceedings of the American Catholic Philosophical Association, 1960, reprinted in my Science, Perception and Reality, London, 1963.

4For a constructive account of sense impressio along these lines, see the chapter on Phenomenalism in my Science, Perception and Reality, London, 1963; also "The Identity Approach to the Mind-Body Problem," Review of Metaphysics, 1964, and the Proceedings of the Boston Colloquium on the Philosophy of Science, Second Series, New York, 1964.