LUCY ALLAIS | Manifest Reality: Kant’s Idealism and his Realism | Oxford University Press 2015


By Colin McLear 

In her book Manifest Reality, Lucy Allais articulates a series of compelling criticisms of both epistemological and phenomenalist readings of Kant, while advancing a new ‘moderate’ metaphysical interpretation of Kant’s transcendental idealism. Along the way, Allais provides a substantive elaboration of Kant’s theory of cognition, and in particular, on the nature of, and relation between, the cognitive faculties of sensibility and understanding, and their characteristic representational output of intuitions and concepts.

One of the central contributions of Allais’ discussion of intuition is her emphasis on the importance of taking seriously Kant’s conception of intuition as a singular and immediate representation that gives us objects. According to Allais, intuitions “give us acquaintance with objects” (p. 147), by which she means that spatio-temporal particulars are perceptually presented in intuition, and it is in virtue of our being acquainted with such particulars that we are in a position to form thoughts concerning them. I think that there is much that is correct in Allais’s interpretation of Kant’s notion of intuition, and much to admire in her expression of it. The view she articulates is a clear and forceful expression of central elements of Kant’s mature thought on our cognition of reality.

I focus here, somewhat narrowly, on Allais’s conception of intuition as acquaintance. I discuss three points. First, I review Allais’s interpretation of intuition as providing acquaintance with a perceptual particular. I raise some questions for this view including how the epistemological and metaphysical aspects of the view are supposed to fit together. Second, I examine Allais’s discussion of the cognitive role of intuition. I raise two worries. First, that her exposition of intuition’s cognitive role is undercut by her view that referential thought requires merely possible acquaintance. Second, that the acquaintance condition on referential thought is implausibly strong, and is better considered as part of a condition on cognition. Finally, in section three, I raise some questions concerning the role of the categories in the having of intuition. I object that the semantic role which intuition is supposed to play stands in tension with Allais’s conception of the role of the categories.

1. The Nature of Intuition

Allais argues that interpretations of Kant’s mature view often fail to take seriously his account of intuition (p. 146). As is well known, Kant distinguishes intuitions from concepts by virtue of the former’s ‘immediacy’ and ‘singularity’ in contrast with the latter’s ‘mediacy’ and ‘generality’ (A320/B377; see also A19 and A50/B74). Intuitions are what give us objects, while concepts allow us to think about them (A19/B33, A239/B298, A719/B747). Allais provides a very specific gloss on what we should understand Kant to mean by the characterisation of intuition as an immediate, singular representation which gives an object. Allais writes:

I argue that the singularity of intuitions should be understood as their presenting perceptual particulars, that the immediacy of intuitions should be understood as their giving us acquaintance with or presenting the particulars they represent, and that this is what it means to say that intuitions give us objects. Intuitions give us acquaintance with objects. (p. 147)

So, on Allais’s view of intuition, its singularity consists in its presenting a ‘particular’, while its immediacy consists in the acquaintance relation that the subject stands in to the intuited particular. Let’s unpack a bit further the notions of a ‘particular’ and of ‘acquaintance’ as Allais understands them. According to Allais, a “perceptual particular” is

a thing which a subject singles out as a perceptual unit—a distinct, bounded thing to which the subject can pay perceptual attention. This could be a causally unitary object, but could also be less than, or more than, an object. A subject may intuit, for example, a desk and the lamp attached to it, or may attend just to the light bulb in the lamp. A spot of light moving on a wall could be a perceptual particular—it is something outside of and other than the subject, that the subject can pick out as a unit. (p. 147n.2)

Allais, following Golob (2011), construes a perceptual unit in terms of Kant’s notion of a “basic measure” (Grundmaß) (KU, AA 5:251ff.; see Allais’s discussion at p. 171n.46). The basic measure “refers to the explanatorily primitive capacity for the representation of a determinate spatial extent, or a unit of spatial representation” (Golob 2011:509). Depending on context, the basic measure can vary.

What a creature tends to attend to as a basic perceptual unit can vary, depending on context (I could see the hill and the trees on it as a unity, from a distance, and a leaf as a singular thing up close) and can be determined by biologically driven principles of association (what size, for example, its prey typically is). (p. 171n.46)

So the perceptual particulars given to a subject via intuition can vary, depending on a variety of factors including the subject’s relation to the object, the subject’s biological needs, and (presumably) the nature and sensitivity of the subject’s sensory organs. Perceptual particulars may in fact be comprised of multiple distinct objects, or they may not be ‘objects’ in any robust sense at all, as with shadows and patches of light (see pp. 147n.2, 154n.18, 156). Hence, the singularity of intuition is based on the fact that what is presented in a single intuition is one perceptual unit, not in the fact that this perceptual unit itself consists of one object or thing. One initial worry about this construal of the singularity of intuition is that it focuses on the singularity of presentation rather than the singularity of what is so presented. We’ll return to this point in the final section below.

For Allais, the immediacy of intuition consists in the fact that via intuition a subject is “acquainted with an object” (or more accurately, with what is presented as a single perceptual unit). It isn’t clear whether Allias considers acquaintance to be an epistemic relation to the object (as would be indicated by using it to gloss Kant’s notion of immediacy) or as a metaphysical feature of the psychological state one is in when intuiting an object, or as both. Certainly, Allais repeatedly speaks of intuition’s putting a subject “directly in touch” with a perceptual particular (pp. 13, 113, 147, 269). This suggests the epistemic gloss, whereby intuition provides a subject with non-inferential awareness of a perceptual particular. But Allais also construes the immediacy of intuition in terms of its being “object-dependent”, in that “we have an intuition of an object only when that object is in fact present to us: a dream or a hallucination of an object does not count as an outer intuition” (p. 156). It is not entirely obvious how these two positions—viz. the epistemic and the metaphysical—relate to one another.

Since Allais ties the notions of immediacy and acquaintance together, there is a similar ambiguity in her discussion of the immediacy of presentation as to whether it is an epistemic or metaphysical feature. The gloss of immediacy in terms of the object-dependence of intuition seems metaphysical, in that it is making a claim in terms of a necessary condition of a mental state’s being an intuition. This is further supported by the explanation of object-dependence in terms of constitution. Intuitions depend on the particulars they present because those particulars are part of the intuition (pp. 12, 105–7, 117, 197). In this way intuitions are relations, which have conscious subjects and environmental objects (and perhaps more besides) as relata. So just as Sue cannot be taller than Tom without Tom’s existing, so too one cannot have an intuition of a red ball without that ball’s existing, being red, and being a part of the subject’s current environment (perhaps glossed in terms of some suitable causal relation between subject and object). The problem is that this is all so much metaphysics. It tells us nothing as to why something that is a part of a mental state should be something that is also epistemically (i.e. non-inferentially) immediate to the subject of that state.[1]

For example, the limbic system is a part of me (a part of my brain), but this fact gives me no privileged epistemic access either to it or to its functions. This issue of the relation between constitution and privileged epistemic access is not a problem peculiar to Allais’s reading of Kant, to which I am in fact very sympathetic, and has been raised with respect to contemporary forms of perceptual relationalism.[2] But it would be helpful to hear more about how constitution is supposed to do its epistemic work in the Kantian context she discusses.

Further, it isn’t clear what object-dependence has to do with singularity as Allais conceives of it. Allais takes the two notions as going together (pp. 153, 158). But recall that singularity for Allais is the presentation of a single perceptual unit, and that the unit can itself be constituted by any number of objects (e.g. a desk and a table), depending on the animal and its sensory and environmental relations to the object. But then it is not clear how the perceptual unit that is a constituent of a perceiving subject’s consciousness is singular in the sense that contrasts with the generality of concepts, viz. with the fact that a concept can apply to a potential multiplicity of objects (p. 157). This worry will also be relevant to assessing Allais’s claims concerning the cognitive role of intuition, to which we’ll now turn.

2. The Cognitive Role of Intuition

As we’ve seen, Allais construes the givenness of an object, or perceptual particular, in intuition in terms of a subject’s being acquainted with that object. I suggested that Allais’s notion of acquaintance had both an epistemic and a metaphysical aspect, though it is unclear exactly how the two are related. But why construe giveness in these terms at all? What cognitive requirement does acquaintance satisfy?

According to Allais’s Kant, we need acquaintance with objects in order to have referential thoughts about them (pp. 154–5, 168, 269–70, 289). Since we cannot have knowledge of objects without being able to refer to them, acquaintance is a necessary condition of all knowledge. I am very sympathetic to Allais’s strategy of articulating the nature of intuition by appealing to its cognitive role in Kant’s Critical philosophy. However, I think there are some important questions surrounding some of the positions she ultimately ascribes to Kant.

In the next two subsections I discuss two issues. First, I try to get a bit clearer as to whether it is actual or merely possible acquaintance that is necessary for referential thought, and raise problems for Allais’s view with respect to the two possible positions she might take. Second, I look at the acquaintance condition itself. I argue that Allais’s interpretation of the acquaintance condition is too strong, and suggest that it is better construed as concerning cognition rather than thought.

2.1 Acquaintance: Actual or possible?

Allais’s account of the importance of acquaintance appears at times to vacillate between claiming that actual acquaintance is necessary for referential thought and claiming that merely the possibility of acquaintance is necessary. Russell, with whom she compares Kant’s view, required actual acquaintance for reference, famously saying that “[e]very proposition which we can understand must be composed wholly of constituents with which we are acquainted” (Russell 1910:117). Allais suggests that she endorses a similarly strong view at times, saying, for example that “[t]o have successful referential thought it is necessary that I am presented with perceptual particulars” and that “concepts will lack relation to objects unless they apply to objects which are given to us in empirical intuition” (p. 270). She also says that intuition “is a relation to an object that […] is necessary for us to be in a position to have thought about the object” (p. 163).

These remarks suggest that Allais interprets Kant as endorsing a version of what Gareth Evans has called “Russell’s Principle”, according to which (singular) reference requires possession of identifying knowledge of that to which one refers (Evans 1982: 44). In the above passages, Allais seems to attribute to Kant a similar principle, viz. that a necessary condition of referential thought of a particular requires that one be in a position to distinguish that particular from all others. Acquaintance with the relevant particular in intuition is then, on her view, understood to fulfil that role.

However, Allais also clearly says that it is only the possibility of acquaintance that is required for cognition (pp. 14, 163, 303). A footnote makes this explicit:

Kant does not require that we have actual acquaintance with all the things we can successfully think about; rather, it must be possible for us to have acquaintance with an object, and the object must be causally connected to one with which we have direct acquaintance. (p. 270n.11)

Allais identifies here two conditions on referential thought of particulars: viz. (i) possible acquaintance; (ii) causal connection with something with which one is “directly” acquainted. I’m inclined to think that this is her considered view, though it would be helpful to see this point clarified. As I understand Allais, then, referential thought requires either acquaintance with the subject matter of the thought or acquaintance with something that is appropriately causally connected to the thought’s subject matter. However, if my interpretation of her is correct a problem arises.

Recall that the motivation Allais presents for the acquaintance view turns on the ‘direct’ or ‘immediate’ way in which it puts a subject ‘in touch’ with an object. This is part of what is supposed to separate her interpretation of perceptual experience (intuition) as acquaintance from an interpretation of perceptual experience as always being of some causal intermediary (p. 159). Acquaintance is supposed to be both epistemically secure, by guaranteeing the existence of its object (p. 157) and the best explanation of the singularity of intuition, since according to Allais mental intermediaries, such as images, cannot be singular (p. 158).

Certainly, Allais’s emphasis on the importance of acquaintance for referential thought, as opposed to a view based on the representation of causal intermediaries, comports well with Kant’s emphasis on the inherent uncertainty of causal inferences and of the immediacy of that which we cognise (A368, B276). But it is not clear how, on Allais’s view, the mere possibility of acquaintance, plus causal connection, could do the cognitive work Allais wants it to make referential thought possible. For example, while it is true that Kant does allow for the legitimacy of inferences from our immediate experience to the existence of objects that are only mediately available to us—this is the point of Kant’s claim concerning the possibility of cognising the existence of “magnetic matter” (A226/B273)—it isn’t at all clear that such inferences are certain (Allais seems to agree on this point; see pp. 157–8). And here a difficulty awaits. On the one hand, if it is the cognitive intimacy of acquaintance, which is certain, that is supposed to secure the possibility of reference, then it is difficult to see how mere causal connection with an object of acquaintance could provide the requisite security for singular referential thought of something other than that object of acquaintance. On the other hand, if causal connection rather than occurrent acquaintance were sufficient for referential thought of an object, then it wouldn’t be clear why we would need acquaintance with external objects in order to have referential thoughts about them. Acquaintance with sense-data that stand in the proper causal connection to external objects should be enough, but this is a position that Allais strongly rejects as descriptive of Kant.

Allais might respond here that I am ignoring the fact that causal connection isn’t sufficient, but rather necessary, along with the possibility of acquaintance itself. In reply I would argue that there are various aspects of empirical reality that are impossible for one to be acquainted with (e.g. Kant’s “magnetic matter”, reality at the quantum level, etc.). Causal connection would thus seem to be one’s only cognitive link in such cases. Would that be sufficient for cognition? Again, as above, I think that neither a negative nor an affirmative answer here is particularly satisfactory.

Thus, though I believe that according to Allais’s interpretation only the possibility of acquaintance is necessary for referential thought, I think that her adoption of this weaker condition potentially undercuts the motivation for endorsing the acquaintance view in the first place. Alternatively, if she chooses to endorse only the stronger requirement of actual acquaintance, then she is going to have trouble with texts in which Kant seems to allow referential thought, and perhaps even cognition, without acquaintance (e.g. magnetic matter, fundamental repulsive forces, etc.).

2.2 Reference and aboutness

Apart from this worry about possible vs. actual acquaintance, I also have doubts that acquaintance (or even its possibility), as conceived by Kant, is a necessary condition of referential thought. According to Allais’s Kant, referential thought is a condition of cognition, and ultimately of knowledge, of empirical objects.

I will argue that Kant’s concern in the Deduction is with something specific: what it takes for thought (concepts) to refer to an object […]. For Kant, referential thought is not the same as acquaintance with objects, and it is not required to have acquaintance with objects, but rather, is dependent on our having acquaintance, but is also dependent on the application of the categories. (p. 168)

Since acquaintance is a necessary condition of referential thought, it is therefore a condition on cognition and knowledge. What does Allais mean by the phrase “referential thought”? Consider her discussion of the following example.

Suppose I dream about a man living in Australia who wears red socks every day and teaches Anthropology. Suppose further that there is exactly one man living in Australia who wears red socks every day and teaches Anthropology. This coincidence, one might think, is not enough to make my dream about that particular man. My dream lacks the right kind of connection to the man to be about him. Kant’s concern is with how we establish the kind of connection needed for thoughts to refer to objects. He thinks that conceptual thought on its own cannot connect to objects, which is why it cannot constitute cognition proper: it lacks relation to an object. (p. 269)

From this example involving a dream, Allais clearly indicates that it is not enough to have a referential thought that one have a thought which stakes a claim about some subject matter.[3] She instead pushes the alternative and stronger claim that referential thought is thought of things with which one is (possibly) acquainted (p. 270).

In what follows I make two points. First, I argue that Allais conflates (or does not justify the connection between) two arguably distinct but still broadly referential notions, viz. satisfaction and constitution. There are good philosophical reasons for keeping these notions distinct. Second, I argue that Kant plausibly recognises this distinction, and thus allows that there is a form of referential thought—or ‘aboutness’ more generally—without acquaintance. In the end, I think Allais’s point about the cognitive role of intuition, and acquaintance in particular, is best construed in terms of cognition rather than (referential) thought. I take these points in turn.

Start with the notion of ‘aboutness’. There are many cases in which a subject might essay a thought about something in the world even if she is not in a position to independently indicate that which her thought is about. Nevertheless, the thought has a subject matter and either correctly or incorrectly stakes a claim about that subject matter. For example, one might think, upon entering St. Mark’s Basilica, that the tallest person in Italy would seem small in such a place. Assuming that there is just one such person, one’s thought would make a claim, correctly or incorrectly, concerning how tall that person would seem when standing in St. Mark’s. However, grasping the thought—understanding the claim to reality that the thought makes—does not require any antecedent grasp of the particular Italian person (assuming there is one) who satisfies that thought. In this sense a thought is about an object—one might even say it ‘refers’, at least in a weak sense—just in case the object satisfies the condition set out in the thought. These conditions of satisfaction, along with the world, determine whether such thoughts are true or false. In the case where there is no unique object for the thought to be about, either because of a plurality of satisfiers or because of a total lack of satisfiers, the thought would be graspable but simply false.

Going back to Allais’s dream example, suppose the dream was not of a man who wears red socks, is an anthropologist, etc. but rather was of the man who wears red socks, etc. In the case of there being just one such man, then there is a sense in which the dream was about that man, and specifically because that particular Australian man satisfies the conditions set out by the content of the dream. In the case there was no such man, then the dream would simply be making false claims.

Contrast this ‘satisfactional’ way in which a thought might be about the world with a much more stringent notion, according to which a thought is about something only if that thing constitutively enters into the content of the thought, such that one could not grasp the thought without thereby knowing which thing in the world the thought is about.[4] In the case where there is no such thing, there is no corresponding thought to think. This is in contrast to the notion of satisfactional thought outlined just above. Satisfactional thought, lacking a unique thing to be about, ends up merely false. In the case of more stringent constitutive notion of thought, when there is no relevant object there is no thought at all. Call this the ‘constitutive’ view of thought.

Now, it seems to me to be a largely terminological issue as to which notion—viz. satisfaction or the constitutive view—should properly receive the moniker ‘referential thought’. Perhaps they both should. Allais seems to want to deny that satisfactional thought deserves the honorific ‘referential’. Perhaps that is right. But be that as it may, I am not clear that any argument is provided by her for the position that satisfactional thought is wholly lacking in aboutness. Nor is there a clear argument from the position that thought is satisfactional in nature to the conclusion that satisfactional thought depends on referential thought, and perhaps ultimately on acquaintance. Certainly, the popularity of such a view amongst contemporary philosophers, from Strawson to McDowell, attests to its attractiveness,[5] but it would be helpful to see the view set out in more detail.

So far, I have discussed a way in which I think Allais’s use of ‘referential thought’ is somewhat ambiguous, and presented some argument for why one might think that a thought can be about something whether or not it is ‘referential’ in Allais’ stringent constitutive sense. But there is also the question of whether this distinction is plausibly attributible to Kant. I believe that it is.

Allais is, I think, correct to claim that Kant construes the capacity for referential thought (in the stringent ‘constitutive’ sense outlined above) of objects as, in some way, dependent on a more fundamental capacity to intuit objects in sensory experience. This is at least in part due to the fact that Kant denies that there are any genuinely singular concepts (B683–4; Log, AA 9:91, 97), so in singular judgments (e.g. that a is F) the subject concept a is not itself a singular concept, but only one that is playing the subject role in that particular judgment—it is “singular in use” in Kant’s phrase (Log, AA 9:91).[6] But since Kant never clarifies what he means by making a concept singular in use, it isn’t clear how descriptive thought about the world—thought that latches on to its subject matter via satisfaction conditions—must depend on singular thought—thought that latches on to the world in a constitutive manner.[7]

We moderns now have a relatively clear explanation of the difference between satisfactional and constitutive thought based on the manner in which an object figures in the proposition constituting the content of the thought.[8] But this distinction is not obviously one that we can attribute to Kant. Since, for Kant, the content of thought (or, in his terminology, of judgement) is always conceptual (KU §35, AA 5:287; see also B146, B283; Log, AA 9:101; V-Lo/Wiener, AA 24:928), and concepts never have objects as constituents,[9] there is no way in which Kant could allow that there is singular thought in our contemporary sense. The closest that Kant might get to our contemporary conception is his notion of cognition as a complex form of representation requiring a combination of intuition and concept. Intuition might then be the psychological analogue to what we now call ‘devices of direct reference’ (e.g. demonstratives and other indexicals, names, etc.), which would contribute some specific individual to cognition, and which satisfies the conditions (in successful cognition) set by the concept or conceptual judgement. So, according to this model, unlike the conceptual content of thought, which determines its object by means of setting conditions which the object may or may not satisfy, in cognition intuition contributes its referent directly to the propositional content of the cognition, which is a combination (in some manner) of both concepts (judgement) and intuition.[10]

Even if something like this were the correct thing to say about Kant’s conception of cognition, it would not show that satisfactional thought depends on constitutive thought (or cognition, as I have now suggested). It would, rather, only show that cognition depends on the semantic contributions made by intuition, which is important, but plausibly less demanding than Allais’s original interpretive claim.

I think that similar issues beset Allais’s contention that, for Kant, satisfactional thought depends ultimately not just on constitutive thought, but also on acquaintance. Again, this seems to be a condition that sits better with Kant’s conception of cognition rather than thought, and it need not deny what Allais seems set on denying, that a thought could be about its subject matter even if one fails to be acquainted with that subject matter.

Concerning these points, consider Kant’s stance concerning the theoretically unverifiable but still rationally legitimate status of beliefs concerning God and the non-empirical self. The fact that we cannot have any sensory consciousness of such entities does not entail, for Kant, that we cannot have thoughts about them. If it did then how would it be possible to hope (i.e. bear a distinctive kind of doxastic attitude towards a content), for example, that one’s soul survives the death of one’s body, or that God apportions happiness to virtue? That Kant did find such thoughts coherent, and as delimiting a specific subject matter, suggests that while we may not have any acquaintance with our own (or others’) soul(s) or God’s existence, we may nevertheless be able to think of those things by staking a claim about the relevant subject matter, that is, by specifying conditions under which those objects, if they exist, would fall.

For all these reasons I believe that Kant (i) primarily conceives of thought satisfactionally; (ii) allows that we might have satisfactional thought of a subject matter even under conditions in which no intuition—and so no acquaintance—with the relevant subject matter is possible; (iii) cannot plausibly construe thought as dependent on acquaintance. However, I think much of Allais’s view could be retained if we read it as setting conditions on cognition and its relation to knowledge, rather than on thought more generally.

3. Idealism and Individuation

A final worry concerns what we might call the ‘fit’ between what is given in intuition and what is cognised as the result of cooperation between intuitive and conceptual representation. Allais says that

[a]cquaintance (unlike merely having a concept) is a relation to an object that guarantees the existence of the object and which individuates a specific particular. (p. 14)

Recall, however, that the particular so presented may in fact consist of multiple objects. It might, for example, consist of a desk and a table—a ‘desk-table’—as far as the intuition is concerned. This is compatible with Allais’s claim that

subjects have ways of discriminating individual things from other things, of perceiving their boundaries, and of representing them as unified, discrete, and distinct from each other and from the subject that are independent of conceptualising. (p. 167)

The “unified, discrete, and distinct” objects may be things like desk-tables and rock-trees. Objects need not obviously be presented in intuition as the discrete particulars that we think about in everyday life.

This point might seem to lend support to Allais’s interpretation of the role of concepts, specifically the categories, in our thought of objects. According to that view, the categories exclude “gerrymandered” properties and objects as candidates for cognition, including “mereological sums which do not correspond to what we ordinarily think of as objects” (pp. 277–81). The categories are supposed to set limits concerning “what counts as a subject of properties” as well as “what counts as more than one property belonging to the same object” (p. 282).

However, Allais also wants to deny that the categories’ playing this role entails any further form of idealism. Kant’s idealism, on her view, is set at the level of intuition—all givenness to cognition is conditioned by our forms of intuition, so all objects of cognition must ipso facto be ideal. Intuition thus plays the metaphysically crucial role in determining the idealist status of the phenomenal world. The role of the categories in providing “relation to an object” is thereby solely epistemological and not metaphysical (p. 285).

What confuses me about the resulting view is how intuition is both supposed to provide a determinate semantic contribution to thought, in virtue of a particular’s being presented in intuition, while simultaneously being unable to determine how it is that one should think of the object so presented. Allais might be thought to anticipate this objection when she says that

Kant’s claim that combination is not given can be understood as saying that how we group objects in regarding them as falling under a common property is not presented in perception; there are indefinitely many predicates corresponding to possible ways we could group or classify the objects that are given in intuition […]. Crucially, this point is fully consistent with thinking that intuition presents us with perceptual particulars: being presented with a red apple in intuition does not determine whether or not we will group it together (unify it) with other things by bringing it under the concept ‘red’, the concept ‘is 50 miles from a burning barn’, or the concept ‘round if red and square if green’. (p. 278)

This way of putting her position seems to try and have things both ways. On the one hand, Allais needs it to be the case that the mind is presented in intuition with determinate perceptual particulars (e.g. the red apple), for it is the fact of our acquaintance with such particulars that makes it the case that we can have referential thoughts of them. On the other hand, the particulars so presented do not themselves determine which of a class of gerrymandered vs. non-gerrymandered concepts is correct. For that we need the categories. While I don’t think there is anything logically inconsistent about adopting these two positions—perhaps reference is one thing, and predication another—they do strike me as being in tension. The issue is that the semantic function that Allais seems to see acquaintance playing requires that a determinate object be contributed to the content of cognition. This is part of what makes intuition an immediate, direct apprehension of a thing rather than a merely specificational representation of some (putatively) possible object. What the categories do, on Allais’ view, is then provide the boundaries of what could be an acceptable conceptualization of what is already determinately given in intuition.

The problem is that, as Allais admits, intuition doesn’t give a perfectly determinate object or ‘particular’. Instead, the particular given in intuition is determined by the subject’s basic measure, which can vary with context and might be quite coarse-grained in extent. As discussed in Section 1 above, the particular might perhaps consist of multiple distinct objects, so that a rock and the tree growing from it are presented in intuition as one particular—a ‘rock-tree’ fusion, so to speak. Now it is no longer clear how intuition performs its required semantic function. What is presented—i.e. what is contributed to the semantic content of cognition—is a rock-tree, not a rock and a tree (as distinct objects). But the presentation of a rock-tree presumably licenses different referential thoughts than presentation of a rock and a tree. In order for intuition to present two distinct and determinate objects—the rock and the tree—we would need to appeal to the kinds of individuating conditions offered by the categories. But that means that the categories don’t just (epistemically) constrain what kinds an object may fall under, but that they actually determine what kinds of semantic contribution an intuition might make to cognition. And this seems to push in the direction of the sort of “sensory mush” view (pp. 148–9, 153) that Allais wants to deny.

Invited: 8 July 2015; received: 23 November 2015.


[1] I assume here that Allais indicates that, at least, the token mental state one is in depends on the existence and presence of the relevant object, not merely that the representational content of the state depends on the object. Call the former ‘object-involvingness’ and the latter ‘object-dependence’ (see Martin 2002). Since Allais construes Kant as a kind of ‘relationalist’ about perception (p. 12), she denies that representational content plays any significant explanatory role in accounting for perceptual experience. So she cannot explain the dependence of intuition on its object in terms of a representational content, but rather must do so in terms of the direct connection between the object and the experience itself. This is also one of the ways in which Allais’s view contrasts with Robert Hanna’s. According to Hanna, an intuition is object-involving in virtue of the object-dependent singular representational content it possesses.

[2] See, for example, Cassam’s objections to Campbell in Cassam (2011).

[3] Compare David Kaplan’s discussion of “Newman 1”, the name of the first child born in the twenty-second century, in Kaplan (1968:201) and his subsequent denial in Kaplan (1979:397) that one need be “en rapport” with an object to assert something of it. For recent treatment of these issues see the essays collected in Jeshion (2010). Concerning Allais’s use of the dream example in particular, one might object that it is not obvious that one can assert something in a dream. Hence, it isn’t clear that what drives our intuitions concerning whether such a case is or isn’t illustrative of referential thought is the fact that the content of one’s thought in such a case isn’t referential, or that one is not (and perhaps never could be) in a position to assert the content. Hence there is an ambiguity—perhaps difficult to resolve—as to whether it is the content or the attitude that is doing the work in the proposed case.

[4] See Evans (1982:44); cf. Hawthorne and Manley (2012:5–6).

[5] See, for example, Strawson (1959:19–23); McDowell (1998), Chap. 11; Brewer (1999:xiv–xv) and (1998), Chap. 2.

[6] See the classic discussion in Thompson (1972).

[7] Certainly, there is also a further constraint on our descriptive thought, at least as far as Kant is concerned, viz. the fact that it depends for its content on concepts that are abstracted from experience as “reflected representations” generated by acts of comparison, reflection, and abstraction (Log, §§ 8–9).

[8] See the overview of these issues by David Kaplan, who was perhaps one of the most influential contributors to this debate, in Kaplan (1989:568–9).

[9] There is some dispute as to whether and how a concept might have a non-logical extension (i.e. an extension which does not itself consist of concepts). For recent discussion see Anderson (2015:61–70).

[10] It is this model that drives Hanna’s conception of the semantic structure of intuitional content, as well as its contribution to cognition. See Hanna (1993), (2001:196–7) and (2005:258, 258n.29).


Anderson, L. (2015), The Poverty of Conceptual Truth: Kant’s Analytic/Synthetic Distinction and the Limits of Metaphysics (Oxford: Oxford University Press).

Brewer, B. (1999), Perception and Reason (New York: Oxford University Press).

Cassam, Q. (2011), ‘Tackling Berkeley’s Puzzle’, in J. Roessler et al. (eds) Perception, Causation, and Objectivity (Oxford: Oxford University Press), pp. 18–34.

Evans, G. (1982), The Varieties of Reference, ed. J. McDowell (Oxford: Clarendon Press).

Golob, S. (2011), ‘Kant on Intentionality, Magnitude, and the Unity of Perception’, European Journal of Philosophy 22 (4): 505–28.

Hanna, R. (1993), ‘Direct Reference, Direct Perception, and the Cognitive Theory of Demonstratives’, Pacific Philosophical Quarterly 74 (2): 96–117.

——— (2001), Kant and the Foundations of Analytic Philosophy (Oxford: Oxford University Press).

——— (2005), ‘Kant and Nonconceptual Content’,  European Journal of Philosophy 13 (2): 247–90.

Hawthorne, J. and D. Manley (2012), The Reference Book (Oxford: Oxford University Press).

Jeshion, R. (ed.) (2010), New Essays on Singular Thought (Oxford: Oxford University Press).

Kaplan, D. (1968), ‘Quantifying in’, Synthese 19 (1–2): 178–214.

——— (1979), ‘Dthat’, in P. French et al. (eds), Contemporary Perspectives in the Philosophy of Language (Minneapolis: University of Minnesota Press), pp. 383–400.

——— (1989), ‘Afterthoughts’, in J. Almog et al (eds) Themes from Kaplan (New York: Oxford University Press), pp. 565–614.

Martin, M. (2002), ‘Particular Thoughts & Singular Thought’, Royal Institute of Philosophy Supplement 51: 173–214.

McDowell, J. (1998), Meaning, Knowledge, and Reality (Cambridge, MA: Harvard University Press).

Russell, B. (1910), ‘Knowledge by Acquaintance and Knowledge by Description’, Proceedings of the Aristotelian Society 11: 108–28.

Strawson, P.F. (1959) Individuals: An Essay in Descriptive Metaphysics (London: Routledge).

Thompson, M. (1972) ‘Singular Terms and Intuitions in Kant’s Epistemology’, The Review of Metaphysics 26 (2): 314–43.

© Colin McLear, 2016.

Colin McLear is an Assistant Professor of Philosophy at the University of Nebraska-Lincoln, USA. He graduated with a Ph.D. in philosophy from Cornell University in 2013. His work focuses on the history of modern philosophy, especially Kant, and the philosophy of mind. He has published on these topics in a variety of places, including Journal of the History of PhilosophyMindPhilosophy Compass, and Philosophers’ Imprint.