1. Introduction

In “Study of Concepts” (1992a) and subsequent works (1992b, 1998, 2001), Christopher Peacocke advances a non-conceptualist view of perception. One of the main arguments Peacocke puts forward to make his case derives from the possession conditions of observational concepts. According to Peacocke, the possession of observational concepts—i.e., concepts the application of which is based on observations—does necessarily rest on the exercise of perceptual discriminatory capacities. However, the exercise of perceptual discriminatory capacities cannot rest, in turn, on the possession of conceptual abilities, on pain of circularity. For, if perceptual capacities were conceptual, then exercising these capacities would depend on possessing observational concepts, which, in turn, would depend on exercising perceptual discriminatory capacities, in a vicious circle. Therefore—Peacocke concludes—perceptual discriminations must be non-conceptual.

Much has been written about the nature of observational concepts and their relationship to perception. Empiricists were the first philosophers who explicitly took the acquisition and possession of observational concepts (“simple ideas” in Locke’s vocabulary) to depend on perceptual capacities. After the linguistic turn, logical empiricists tried to explain the semantic content of observational predicates (such as “red”) through semantic rules that tie the application of these predicates to non-conceptual states of perceptual awareness (Ayer 1954). Currently, no philosopher denies that there must be a close relationship between observational concepts and perceptual discriminatory capacities. Recent discussions, however, have mainly focused on what kind of relationship this is. Some philosophers have denied that the relationship that ties observational concepts to perceptual capacities is semantic and epistemic, leaving the role perception plays in the possession condition of observational concepts to be merely causal (Brandom 2002, 2010; Davidson 1982, 1983). Others philosophers, including McDowell (1994, 2009), Peacocke (1992a, 1998), and Bermúdez (1997, 2003, 2007), have argued, in contrast, that the relationship between observational concepts and perception is both semantic and epistemic.[1] Thus, assuming that this relationship is so, McDowell, on the one hand, and Peacocke and Bermúdez, on the other, have disputed about the very nature of perceptual capacities. According to McDowell, perceptual capacities must be fully conceptual, for otherwise, an episode of the Myth of the Given would obtain (McDowell 1994,1998, 2009). According to Peacocke and Bermúdez, in contrast, perceptual capacities must be non-conceptual, for otherwise, a circular explanation would obtain (Bermúdez 2007; Peacocke 1992a, 1998).

In this paper, I will not get into the debate between Peacocke, Bermúdez, and M­cDowell on the nature of perceptual capacities. I am not going to argue in favor of either McDowell’s conceptualism or Peacocke’s and Bermúdez’s non-conceptualism. What I want to focus on here is the problem that follows from holding Peacocke’s explanatory model of the possession conditions of observational concepts.[2] My strategy will be Sellarsian. In effect, like McDowell (1998), I will argue that Peacocke’s explanation of the possession condition of observational concepts falls into the Myth of the Given. However, unlike McDowell, I will offer a much more powerful interpretation of the Myth of the Given.[3] In my view, Sellars’s Myth of the Given is not just the thesis according to which non-conceptual mental states cannot be reasons insofar as they lack—as McDowell holds (1994, 2009)—the structure required to participate in reasoning. The Myth of the Given, I argue, is an illusory or mythical explanation where, in order to explain the possession conditions —and, thus, the semantic and epistemic properties—of observational concepts, another phenomenon is brought into play—namely, perceptual discriminations—which, in turn, is taken to have the same properties to be explained (the semantic and epistemic properties of observational concepts); and where, in order to avoid circularity (i.e., that the semantic and epistemic properties of perceptual discriminations derive from observational concepts), these properties are taken to be given in perception (they are self-explanatory).[4]

The paper breaks into two sections. In the first section, I briefly present Peacocke’s two-level account of non-conceptual perceptual content: scenario contents and protopropositions. Then, I reconstruct Peacocke’s main argument for non-conceptualism derived from the possession conditions of observational concepts. In the second section, I raise an objection to Peacocke’s account of the possession conditions of observational concepts. To do so, I follow a two-step strategy. Firstly, I argue that if non-conceptual perceptual contents are understood in terms of scenario contents, then perceptual experiences cannot present perceivers with the circumstances specified by the application conditions of observational concepts—i.e., the circumstances which must obtain for a concept P to be applied correctly—and,[5] consequently, they cannot play the semantic and epistemic roles Peacocke wants them to play in the possession conditions of these concepts. Secondly, I argue that if non-conceptual perceptual contents are understood in terms of protopropositions, then Peacocke’s account of the possession conditions of observational concepts falls into the Myth of the Given. In the end, we will see that Peacocke non-conceptualism is a dead-end.

2. Non-conceptualism and observational concepts

In “Study of Concepts” (1992a), Peacocke defends a non-conceptualist view of perception. According to Peacocke, perceptual non-conceptual contents are of two types: scenarios and protopropositions. A scenario content is a non-propositional, non-conceptual content that presents the subject with her environment in a spatial way. For example, if I open my eyes right now, my visual experience informs me that there is a laptop in front of me, that the laptop is sitting on a library table, that there is a cup of coffee right by the laptop, that the cup of coffee is white, etc. A scenario content is a type of egocentric spatial representation where surfaces and their properties (features) are located around the perceiving subject. In other words, a scenario content is a way of filling out the space around the perceiver (1992a, 61). To specify a scenario content, we have to fix an origin—normally in the perceiver’s chest—and axes—normally drawn from body directions such as right/left, up/down, front/back of the perceiver—and to specify a way of filling out the space around this origin—normally by determining whether there is a surface with some properties (e.g., solidity, texture, saturation, brightness, hue) for each point identified along the axes. Of course, perceptual experiences with a scenario content have correctness conditions. A perceptual experience is correct if and only if the scenario content matches the perceiver’s immediate environment, that is, if real objects (with their properties) are located in the space around the perceiver in the way that is specified by the scenario content. Peacocke calls the way objects and their properties are actually located in the perceiver immediate environment a scene. Thus, according to Peacocke, a scenario content is correct if and only if the scene falls under the spatial type specified by the scenario content (1992a, 64).

Now, it might be possible that two perceivers, or even the same perceiver on different occasions, perceive the same spatial type but a different content. Suppose, for example, that when I perceive the floor tiles of my bathroom, I perceive them as diamonds. Suppose, however, that when my wife perceives the same floor tiles, from the very same angle, she perceives them as squares. Clearly, scenario contents cannot help us account for the difference in content between my wife’s visual experience and mine. After all, specifying an origin, axes, and way of filling out the space around this origin fails to capture the fineness of grain (aspectuality) involved in my wife’s and my visual experiences. Peacocke introduces protopropositions to account for such cases. When I perceive the floor tiles as diamonds—Peacocke argues—my visual experience conveys the protoproposition that the figures are symmetrical about the bisectors of their angles. When my wife perceives the floor tiles as squares, her visual experience conveys the protoproposition that the figures are symmetrical about the bisector of their sides. Protopropositions contain individuals (particular objects), spatial properties, and relations, rather than concepts thereof (1992a, 72). As a result, protopropositional contents represent spatial properties or relations themselves as holding of the individuals they also contain (1992a, 77). As such, they have a subject-predicate structure. However, they are—according to Peacocke—non-conceptual. When a subject perceives a surface as a square instead of as a diamond, what she perceives is that certain spatial property, say the symmetry about the bisectors of its sides, holds of the surface. Accordingly, the difference between perceiving a surface as a square and perceiving it as a diamond is a matter of which symmetries of the surface are perceived. The perceiver, of course, “does not need to know that this is the nature of the difference” (Peacocke 1998, 381). Moreover, she need not possess the concepts of “symmetry”, “surface”, and “side”, in order to enjoy a visual experience conveying the protoproposition that the surface is symmetrical about the bisector of its sides. All she needs is just to perceive such a symmetrical relation itself. Thus, although protopropositions have the subject-predicate structure, which is characteristic of conceptual contents, they are essentially non-conceptual.

One of the main reasons Peacocke puts forward a non-conceptualist view of perceptions is to explain the individuation, acquisition, and possession of observational concepts (Peacocke 1992a, 108). According to Peacocke, it is essential for concepts to play a role in thoughts or propositional attitudes. Concept possession, then, is defined as the capacity of thinkers to form propositional thoughts by applying those concepts (1992a, 5). Now, observational concepts are concepts the application of which is based on observations. Color and shape concepts are typical examples of observational concepts. Thus, the possession of observational concepts may well be defined as the capacity of thinkers to form propositional thoughts containing concepts the application of which is based on perceptual discriminations. According to Peacocke, an accurate explanation of concept possession must avoid circularity, that is, the explanation must not presuppose that the thinker already possesses the concepts the possession of which is sought to be explained (1992a, 116). Since the possession conditions of observational concepts involve perceptual discriminatory capacities, perceptual discriminatory capacities—Peacocke argues—must be non-conceptual. Otherwise, exercising perceptual discriminatory capacities would depend on possessing observational concepts, which, in turn, would depend on exercising perceptual discriminatory capacities, in a vicious circle. For example, an account of the possession conditions of the concept “red” must not presuppose that, in order to perceptually discriminate red stuff, the subject already possess the concept “red”. After all, for a subject S to possess the concept “red” is—according to Peacocke—for S to be (feel) primitively compelled to apply the concept “red” in present-tense demonstrative thoughts such as “this is red” whenever S perceives the circumstances specified by the application conditions of this concept—i.e., whenever she perceptually discriminates red objects (1992a, 117).

It is not surprising that Peacocke explicitly commits himself to perceptual reasons. Insofar as perceptual experiences are contentful mental states on which the possession of observational concepts rests, perception provides compelling reasons for forming beliefs involving those concepts. Quoting Peacocke, “experiences give a thinker who possesses the relatively observational concept square not merely reasons but good reasons for forming the belief that the demonstratively presented object is square” (1992a, 80). What is behind Peacocke’s claim is that, since perceptual beliefs and judgments are made of observational concepts, and observational concepts are, in turn, concepts the application of which is based on perceptual discriminations, when perceptual discriminations—which are non-conceptual representational mental states—are correct, they provide thinkers with good reasons for applying the correct observational concepts and, consequently, for forming true beliefs involving those concepts. In Peacocke’s terms, “that they are good reasons is intimately related to the condition required for the belief ‘That’s square’ to be true. If the thinker’s perceptual systems are functioning properly, so that the non-conceptual representational content of his experience is correct, then when such experiences occur, the object thought about will really be square” (Peacocke 1992a, 80). This is an empiricist thesis with which many defenders of non-conceptualist agree (Heck 2000; Peacocke 1992b; Bermúdez 2007; Hanna 2011). Indeed, Empiricism is the view that empirical knowledge derives from sense experience. Peacocke’s non-conceptualism is a sophisticated form of Empiricism in that, when a non-conceptual perceptual experience, say of a square surface, is correct, it presents perceivers with the circumstances specified by the application conditions of the concept “square” (i.e., with square surfaces) and, consequently, it provides them with good reasons for believing that the surface is square.[6] I will return to this point soon.

3. Concept possession and the myth of the given

Sellars’s Myth of the Given (1956) is often construed as holding that no mental state can be both non-conceptual and the sort of cognitive episode that might serve as a reason for holding beliefs. Hence, it has been concluded—somewhat hastily, I believe—that non-conceptual approaches to perceptual experience fall into the Myth of the Given (McDowell 1994, 1998, 2009). However, Sellars’s critique of the Given is originally concerned with the sense data theory and the foundation of empirical knowledge. As Sellars himself admitted, the Given “is a piece of epistemological shoptalk” (Sellars 1956, 13). Non-conceptualism, on the other hand, is the view that the representational content of perceptual experiences is, unlike those of beliefs and judgments, partially or entirely non-conceptual. Non-conceptualism, broadly speaking, is a content rather than an epistemic view. Thus, it is far from clear to what extent non-conceptualism falls into the Myth of the Given.

Now, McDowell is famous for his strong rejection of non-conceptualism. According to McDowell, non-conceptualist views of perception cannot help but fall into the Myth of the Given (McDowell 1994, 1998, 2009). I believe McDowell is right to a certain extent. Nevertheless, his conclusion derives from an interpretation of the Myth of the Given which, to my mind, is weak. In effect, McDowell usually refers to “the Given” as episodes of awareness that do not involve conceptual capacities. However, when he is asked to elucidate the reasons why such episodes might pose a problem to non-conceptualists, he is not so clear. More often than not, McDowell argues that episodes of perceptual non-conceptual awareness cannot be reasons for beliefs and judgments insofar as (i) they lack conceptual structure, and (ii) only conceptually structured episodes of awareness have the logical or inferential properties required to play a role as premises in reasoning (McDowell 1994, 1998, 2009). Many non-conceptualists, however, have rejected (ii) by arguing that reasons are not to be confined to conceptually structured mental states (Hanna 2011; Heck 2000; Peacocke 1998). In what follows, I will argue—like McDowell—that Peacocke’s account of the possession conditions of observational concepts falls into the Myth of the Given. But, unlike McDowell, I will do so by bringing into play a different interpretation of Sellars’s Myth of the Given. In my view, the Myth of the Given is an illusory explanation where, on the one hand, perceptual discriminations are introduced to explain the semantic and epistemic properties of observational concepts, but on the other hand, perceptual discriminations end up having the very same semantic and epistemic properties that observational concepts are thought to possess; and where in order not to fall into a vicious circle (i.e., the semantic and epistemic properties of perceptual discriminations ultimately derive from conceptual moves), these properties are taken to be given in the perception. But let us not waste more time and get straight to the heart of the matter.

It follows from Peacocke’s assumptions that representing facts of the form “s is red” in conceptual thinking—where “red” is an observational concept—depends on non-conceptually representing red stuff in perception. After all, Peacocke is for the view that the possession of observational concepts such as “red” depends on being able to discriminate red stuff in perception, and to perceptually discriminate red stuff is, in turn, to non-conceptually represent red stuff in perception. It is worth noting that, insofar representing facts of the form “s is red” in conceptual thinking involves non-conceptually representing red stuff in perception, circularity disappears. For, on this view, the ability to correctly apply the observational concept “red” in order to form empirical thoughts depends on an ability which, in turn, does not depend, to be exercised, on the application of the concept “red”. However, another problem arises. For, if perceiving is a matter of representing the perceiver’s immediate environments non-conceptually, then perceiving red stuff either does not involve an awareness of objects as being red and, consequently, does not present perceivers with the circumstances specified by the application conditions of the concepts “red”, or does involve an awareness of objects as being red, rendering the explanation a myth. Let me expand this point.

Peacocke argues that the possession of observational concepts such as “red”—and consequently, the ability to think (believe, judge) facts such as “s is red”—depends on non-conceptually representing red stuff in perception. However, non-conceptually representing red stuff in perception may be interpreted, following Peacocke, in two different ways:

1. 1. Non-conceptually representing red stuff in perception is a matter of representing surfaces and instances of the property RED as occupying the same region (point) in an egocentric spatial representation.

   2. Non-conceptually representing red stuff in perception is representing the property RED itself as holding of places, lines, or regions in an egocentric spatial representation, or of objects perceived as located in such places (Peacocke 1992a, 77).

Thus, based on the distinction drawn above, we may legitimately ask now: can non-conceptual perceptual states in (1) and (2) play the semantic and epistemic roles Peacocke wants them to play in the possession conditions of the concept “red”? In what follows, I will argue that neither (1) nor (2) can do so. To make my case, I will proceed by analyzing (1) and (2) respectively.

3.1. Scenario contents

If perceiving red stuff is a matter of representing surfaces and instances of redness as occupying the same regions in an egocentric space, then—as Sellars has taught us— perceiving red stuff is not a form of perceptual knowledge (Sellars 1956, 16). For, being perceptually aware of particulars, say of objects and instances of properties, does not amount to perceptually representing those objects as having the properties the instantiation of which one is perceptually aware of. In other words, perceptually representing a surface and a particular instantiation of redness as occupying the same region in an egocentric space does not amount to representing that particular surface as having the general property of being red. Co-instantiation—we may say—is not predication. But if perceiving is not a form of knowledge in the sense described above—that is, if perceiving does not involve any kind of predication—then perception cannot be the kind of mental state upon which the application of observational concepts may be based. Let me show why.

We know that observational concepts are those the content of which expresses, when correctly applied, perceptual knowledge of the form “s is red”. Thus, in order to conceptually represent facts such as “s is red”—that is, that “red” is the appropriate concept to apply to s on a particular occasion—one must perceptually discriminate the circumstances specified by the application conditions of the concept “red”. However, the circumstances specified by the application conditions of the concept “red” are precisely those where objects are red (i.e., objects as having the property RED). Hence, in order to represent facts such as “s is red” in conceptual thinking, one must perceptually discriminate s as being red—in other words, one must perceptually represent s as being red. But representing red scenario contents does not amount to perceptually representing objects as being red. Co-instantiation, we have seen, is not predication. Therefore, perception cannot present us with the circumstances specified by the application conditions of the concept red obtain. Since perception cannot present us with the circumstances specified by the application conditions of the concept red, perception cannot be the basis on which the possession of observational concepts rests.

To make my point clearer, I will put it in terms of a deductive argument:

1. (P1) Representing facts of the form “s is red” in conceptual thinking is an ability the exercising of which rests on applying the concept “red” correctly.

   (P2) Correctly applying the concept “red” depends, in turn, on perceptually discriminating the circumstances specified by the application conditions of the concept “red”.

   (P3) The circumstances specified by the application conditions of the concept “red” are those where objects are red.

   (P4) Perceiving red objects is a matter of representing red objects non-conceptually.

   (P5) Representing red objects non-conceptually is, in turn, a matter of being perceptually aware of (representing) surfaces and instances of RED.

   (P6) However, being perceptually aware of (representing) surfaces and instances of RED does not amount to perceiving those objects as having the property RED (co-instantiation is not predication).

   (P7) Therefore, non-conceptually perceiving red objects does not amount to perceiving the circumstances specified by the application conditions of the concept “red”.

   (C1) Non-conceptually perceiving red objects cannot be the basis on which representing facts of the form “s is red” in conceptual thinking rests.

(P1) is an intuitive thesis to which Peacocke is explicitly committed. (P2) is the well-known empiricist thesis according to which the application of observational concepts depends on perceptually discriminating the circumstances specified by the application conditions of such concepts, and to which Peacocke, as we have seen, is committed as well. (P3) is as intuitive as the first in that it just describes the circumstances which must obtain for the concept “red” to be applied correctly (i.e., objects as being red). (P4) and (P5) premises are just two Peacocke’s non-conceptualist assumptions. (P6) follows from Sellars’s famous distinction presented above between representing facts of the form “s is P” and being merely acquainted with particulars s and instances of properties P (Sellars 1956, 16). (P7) is just a consequence of assuming (P3) and (P6). Indeed, it states that being perceptually aware of objects and instances of redness does not amount to perceiving the circumstances specified by the application conditions of the concept “red”. For the circumstances specified by the application conditions of the concept “red” are precisely those where objects are red. Therefore, scenario contents cannot be the base on which the possession of observational concepts rests.

Peacocke might reply that the content of scenarios is not to be understood in terms of bare particulars (surfaces and property-instances) as occupying points in an egocentric space, but in terms of particulars as having general properties or relations in virtue of occupying certain points in this space. Thus, when I perceive two different surfaces, say a and b, as located one next to the other in an egocentric space—Peacoke might argue—I represent a as being on the left of b and b on the right of a. Similarly, when I perceive a surface and a color property-instance, say of RED, as occupying the same point in this space, I represent that surface as having the property RED. Of course—Peacocke might argue—I need not have the concept of either “being on the left of” or “being on the right of” to perceptually represent a and b as being related in such ways. Similarly, I need not have the concept “being red” in order to perceptually represent that surface as having the property RED. Placing such surfaces and properties in the right way is enough. Let us notice, however, that under this interpretation, scenarios are structured contents involving primitive forms of predication. After all, these contents not only include particulars (surfaces and property-instances) but also general properties as being related to these particulars predicatively—i.e., they take, so to speak, the predicative form “s is P”. Thus, under this interpretation, scenarios suffer the same fate as protopropositions.[7] In what follows, I will show why protopropositions are a dead-end.

3.2. Protopropositions

If perceiving red stuff is a matter of representing the property RED as holding of particular objects, perception seems to provide subjects with a perceptual awareness of the circumstances specified by the application conditions of the concept “red”. After all, perception seems to provide subjects now with representational contents of the form “s is red”, which mirror the circumstances specified by the application conditions of the concept “red” (i.e., objects as being red). But then, we are faced with the following problem:

1. (P1) Entertaining contents of the form “s is red” in conceptual thinking is an ability the exercising of which rests on correctly applying the observational concept “red”.

   (P2) Correctly applying the observational concept “red” depends, in turn, on perceiving the circumstances specified by the application conditions of the concept “red”.

   (P3) The circumstances specified by the application conditions of the concept “red” are those where objects are red.

   (P4) Perceiving red stuff is representing red stuff non-conceptually.

   (P5) Representing red stuff non-conceptually is, in turn, being perceptually aware of (representing) the property RED as holding of objects (surfaces).

   (P6) Being aware of (representing) the property RED as holding of objects (surfaces) amounts to entertaining contents of the form “s is red”.

   (P7) Therefore, perceiving red stuff is a matter of entertaining contents of the form “s is red”.

   (C2) Entertaining contents of the form “s is red” in conceptual thinking depends on entertaining contents of the form “s is red” in perception.

But this poses a serious problem to Peacocke. For, when we push Peacocke’s explanation of the possession condition of observational concepts to the limit, we find that what is sought to be explained in the explanandum, say, the ability to entertain representational perceptual contents of the form “s is red”, is already presupposed in the explanans. In effect, (C2) shows that entertaining contents of the form “s is red” in conceptual thinking depends on entertaining contents of the form “s is red” in perception. But we can legitimately ask now: on which representational abilities does the ability to entertain contents of form “s is red” in perception rest? After all, we were told that entertaining contents of the form “s is red” in conceptual thinking relies on perceiving the circumstances specified by the application conditions of the concept “red”. And now, we are told that perceiving the circumstances specified by the application conditions of the concept “red” is a matter of entertaining those very same contents.[8] This, I argue, is the core of the Myth of the Given. For the Myth of the Given is just the assumption that the ability to entertain contents of the form “s is P” in perception does not depend, unlike in cognition, on the exercise of any other representational ability—those contents are, so to speak, merely given in perception—rendering the explanation trivial.[9] Paraphrasing Sellars’s famous inconsistent triad in Empiricism and the Philosophy of Mind (1956, 21), Peacocke is confronted with the following three inconsistent propositions:

1. A: Perceiving is entertaining contents of the form “s is P”.

   B: The ability to perceive is non-conceptual (it does not involve conceptual capacities).

   C: The ability to entertain contents of the form “s is P” is conceptual (it involves conceptual capacities).

A and B together entail not-C; B and C entail not-A; A and C entail not-B.

Peacocke might try to overcome this objection by rejecting C. Indeed, he might insist that non-conceptual perceptual contents involve protopropositions that take the predicative form “s is P” without bringing conceptual resources into play. Thus, the ability to entertain contents of the form “s is P” in perception is not constrained by conceptual moves. Let us remember that protopropositions consist of individuals (particular objects) and properties and relations rather than concepts thereof. So, although protopropositions share the same structure with conceptual contents, they do so without putting concepts to work. Let us notice, however, that this move goes against Peacocke’s initial assumption according to which entertaining contents of the form “s is P” in conceptual thinking is a function of correctly applying the observational concept “P”. For we were initially told that entertaining perceptual contents of the form “s is P” in cognition depends on perceiving the circumstances specified by “P”. And now we are told that perceiving the circumstances specified by “P” is a matter of entertaining in perception those very same contents where “P” was originally part of. Peacocke seeks to have his cake and eat it too. For, one the one hand, he wants protopropositions to be non-conceptual contents. But, on the other hand, he wants them to play the same role he had previously attributed to conceptually articulated contents. This tension, which is essentially the Myth of the Given, becomes more evident when we analyze what it is for someone to perceive protopropositional contents. Let me expand this point.

Protopropositions are non-conceptual perceptual contents that cannot be fully analyzed in terms of scenario contents. In truth, they are introduced to account for the possession conditions of observational concepts in cases where scenario contents fall short of accounting for them. As Peacocke argues:

if we are to have a noncircular and individuating account of mastery of the perceptual concept “straight”, that mastery must be related to some feature of experience that does not have to be explained in terms that presuppose possession of the concept. We cannot supply this by relating the mastery to experiences whose positioned scenarios require for their correctness that a certain line be straight, for that can be so without the straightness being perceived. Having the property STRAIGHT in the protopropositional content respects that point, without threatening a circular account of mastery (Peacocke 1992, 85).

However, when we examine protopropositional contents in details, we end up facing the Myth of the Given. Let us take the concept “square” as an example. We were told that perceiving square surfaces is a matter of entertaining the protoproposition that the surface is symmetrical about the bisector of their sides. However, representing that a surface is symmetrical about the bisector of their sides involves entertaining a content of the form “s is P” which, to be sure, contains the concept of symmetry. Thus, it would seem that perceiving a surface as square involves the possession of the concept of symmetry. Peacocke, of course, does not want the capacity to perceive square surfaces to rely on any conceptual move (including that of symmetry) (see Peacocke 1992a, 76-77). So, in order to be consistent and avoid circularity, Peacocke makes the following move: to perceive the protoproposition that the surface is symmetrical about the bisector of their sides is not a matter of entertaining the propositional content “the surface is symmetrical about the bisector of their sides”, but just a matter of entertaining the protoporpositional content “SYMMETRY OF SIDES holds of the surface”. This move seems to guarantee that protopropositions be non-conceptual. For perceivers themselves need not have the concepts which figure in the description “the surface is symmetrical about the bisector of their sides” in order to perceive such a surface in such a way. Perceptually focusing on such a spatial relation is enough. At this point, however, what makes protopropositions non-conceptual contents is what makes them unable to play the role Peacocke wants them to do in the possession conditions of the observational concept “square”. For, either perceiving the property SYMMETRY OF SIDES as holding of a surface amounts to being aware of the fact that the surface is symmetrical about the bisector of its sides, rendering the explanation mythical, or perceiving the property SYMMETRY OF SIDES as holding of a surface does not amount to being aware of the fact that the surface is symmetrical about the bisector of its sides. But if perceiving the SYMMETRY OF SIDES as holding of a surface does not amount to being aware of the fact that the surface is symmetrical about the bisector of its sides, then perceiving the SYMMETRY OF SIDES as holding of a particular surface does not present perceivers with the circumstances specified by the application conditions of the concept “square”. For, the circumstances specified by the application conditions of the concept “square”—i.e., the circumstances that must obtain for the concept “square” to be applied correctly—are precisely those where surfaces are symmetrical about the bisector of its sides. To put it in a nutshell:

1. (P1) Entertaining contents of the form “that surface is square” in conceptual thinking is an ability the exercising of which rests on correctly applying the observational concept “square”.

   (P2) Correctly applying the observational concept “square” depends, in turn, on perceiving the circumstances specified by the application conditions of the concept “square”.

   (P3) The circumstances specified by the application conditions of the concept “square” are those where surfaces are symmetrical about the bisector of its sides.

   (P4) Perceiving square surfaces is a matter of entertaining the content “SYMMETRY OF SIDES as holding of particular surfaces”.

   (P5) However, entertaining the content “SYMMETRY OF SIDES as holding of particular surfaces” does not amount to being aware of the fact that the surface is symmetrical about the bisector of its sides—on pain of falling into the Myth of the Given.

   (P6) Therefore, perceiving the property SQUARE as holding of particular surfaces does not amount to perceiving the circumstances specified by the application conditions of the concept “square”.

   (C3) Protopropositions cannot be the basis on which entertaining perceptual contents of the form “s is P” in conceptual thinking rests.[10]

At this stage of the dialectic, the tension involved in Peacocke’s account of protopropositions should be clear. For, on the one hand, Peacocke wants protopropositions to play the same role as conceptual contents. But, on the other hand, he wants them not to be conceptually articulated mental states. In effect, Peacocke wants protopropositions to be an awareness of facts of the form “s is P” because only this kind of awareness can guarantee that perceptual experiences present perceivers with the circumstances specified by the application conditions of observational concepts. But at the same time, he does not want to pay the price of making such a move—namely, committing to conceptual capacities. This is especially clear when Peacocke tries to elucidate the non-conceptual nature of protopropositions, that is, when he is required to explain why, despite mirroring the conceptual structure, protopropositional contents are ultimately non-conceptual rather than conceptual. For, either he reduces protopropositions to episodes of perceptual awareness of fact of the form “s is P”, falling into the Myth of the Given, or he reduces protopropositions to episodes of perceptual awareness of particulars which lack the form “s is P”, rendering protopropositions incapable of presenting perceivers with the circumstances specified by the application conditions of observational concepts.

To sum up, Peacocke’s account of the possession conditions of observational concepts is a dead-end. Peacocke wants protopropositions to be non-conceptual contents of the form “s is P” (to hold A and B). But, at the same time, he wants the ability to entertain contents of the form “s is P” to be a function of conceptual moves (to hold C). So, when he is required to overcome this inconsistency, either he reduces protopropositional contents to episodes of the predicative form “s is P”, falling into the Myth of the Given, or he reduces protopropositions to episodes which lack the predicative form “s is P”—in other words, he rejects A—rendering protopropositions incapable of presenting perceivers with the circumstances specified by the application conditions of observational concepts and, consequently, incapable of being the rational base on which the possession of observational concepts may rest.

4. Conclusion

In this paper, I argued against Peacocke’s account of the possession conditions of observational concepts. Firstly, I showed that if perceiving is a matter of non-conceptually representing the perceiver’s immediate environments in a spatial way (scenario contents), then perception does not involve an awareness of fact of the form “s is P” and, consequently, cannot present perceivers with the circumstances specified by the application conditions of the observational concepts “P”. Secondly, I argued that if perceiving is a matter of entertaining protopropositions, then either protopropositions provide awareness of fact of the form “s is P”, rendering the explanation of the possession conditions of observational concepts a myth, or do not provide awareness of facts of the form “s is P”, rendering perception incapable of presenting perceivers with the circumstances specified by the application conditions of observational concepts.

References

Aguilera, Mariela. 2016. Cartographic systems and non-linguistic inference. Philosophical Psychology 29/3: 349-364.

Ayer, Alfred Julius. 1954. Basic propositions. In A. J. Ayer, ed., Philosophical essays, 105-129. London: M­acmillan Press.

Bermúdez, José Luis. 1995. Nonconceptual content: From perceptual experience to sub-personal computational states. Mind and Language 10: 333-369.

—. 2003. Thinking without words. Oxford: Oxford University Press.

—. 2007. What is at stake in the debate on non-conceptual content? Philosophical Perspectives 21: 55-72.

Brandom, Robert. 1994. Making it explicit: Reasoning, representing, and discursive commitment. Cambridge, MA: Harvard University Press.

—. 2002. The Centrality of Sellars’s two-ply account of observation to the arguments of “Empiricism and the philosophy of mind”. In Tales of mighty dead: Historical essays in the metaphysics of intentionality, 348-367. Cambridge, MA: Harvard University Press.

Brewer, Bill. 2005. Perceptual experience has conceptual content. In E. Sosa and M. Steup, eds., Contemporary debates in epistemology, 217-230. Oxford: Blackwell.

Camp, Elisabeth. 2007. Thinking with maps. Philosophical Perspectives 21/1: 145-182.

Castellano, Federico. 2014. Intellectualism against empiricism. Grazer Philosophische Studien, 90: 231-251.

Davidson, Donald. 1982. Empirical content. In Subjective, intersubjective, objective, 159-175. Oxford: Clarendon Press.

—. 1983. A coherence theory of truth and knowledge. In Subjective, intersubjective, objective, 137-153. Oxford: Clarendon Press.

Evans, Gareth. 1982. The varieties of reference. Oxford: Clarendon Press.

Fodor, Jerry. 2008. LOT 2: The language of thought revisited. Oxford: Oxford University Press.

Hanna, Robert. 2011. Beyond the myth of the myth: A Kantian theory of non-conceptual content. International Journal of Philosophical Studies 19/3: 323-398.

Heck, Richard. 2000. Nonconceptual content and the “Space of reasons”. The Philosophical Review 109/4: 483-523.

—. 2007. Are there different kinds of content? In J. Cohen & B. McLaughlin, eds., Contemporary debates in philosophy of mind, 117-138. Oxford: Blackwell.

McDowell, Jonh. 1994. Mind and world. Cambridge, MA: Harvard University Press.

—. 1998. Reply to commentators. Philosophy and Phenomenological Research 58/2, 403-431.

—. 2009. Conceptual capacities in perception. In Having the world in view: Essays on Kant, Hegel, and Sellars, 127-144. Cambridge, MA: Harvard University Press.

Peacocke, Christopher. 1992a. A study of concepts. Cambridge, MA: MIT Press

—. 1992b. Scenarios, concepts, and perception. In T. Crane, ed., The Contents of experience. Cambridge: Cambridge University Press.

—. 1998. Nonconceptual content defended. Philosophy and Phenomenological Research, 58/2: 381-388.

—. 1999. Précis of A study of concepts. In Eric Margolis & Stephen Laurence, eds., Concepts: Core readings, 335-338. Cambridge, MA: The MIT Press.

—. 2001. Does perception have a nonconceptual content? Journal of Philosophy, 98/5: 239-264.

Sellars, Wilfrid. 1954. Some reflections on language games. Philosophy of Science 21: 204-228.

—. 1956. Empiricism and the philosophy of mind. Robert Brandom, ed. Cambridge, MA: Harvard University Press, 1997.

Notes

Additional information

BIBLID: 0495-4548(2018)33:3p.-416