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


By Lucy Allais 

I would like to thank Critique for hosting this discussion, and especially to thank Colin Marshall and Colin McLear for their really fantastically interesting, penetrating, thoughtful and detailed engagement with my book. They have given me lots to think about.

Précis of Manifest Reality

Manifest Reality presents an interpretation of a central part of Kant’s Critical philosophy, particularly as this is presented in the Critique of Pure Reason: Kant’s transcendental idealism. As I see it, making sense of transcendental idealism centrally involves giving an account of Kant’s distinction between things as they are in themselves and things as they appear to us, and then making sense of the claims he makes about each side of this distinction. With respect to things as they appear to us, Kant claims that they are mere appearances, that exist in us, and that do not exist apart from a connection to possible perception. With respect to things as they are in themselves, he claims that they ground appearances and that they cause appearances, but that cognition of things as they are in themselves is impossible for us.

A starting point for my account is that I aim to steer a middle course between two possible extremes in interpreting Kant’s position. Kant sometimes says things that sound realist, and that sound like his concern is with conditions of knowledge rather than presenting a metaphysical system, and he denies that he is asserting the existence of a supersensible reality. On the other hand, he sometimes says things that sound strongly idealist, and also sometimes expresses a commitment to the existence of an aspect of reality we cannot cognise. Many interpretations of transcendental idealism emphasise one of these sides of Kant’s position at the expense of the other, leading to what I see as a polarisation between extreme idealist interpretations, on the one hand, and extreme deflationary, anti-metaphysical interpretations, on the other. I aim to present an interpretation that does justice to both sides of Kant’s position while avoiding the problems of both interpretative extremes. In the first part (of the three parts) of the book, I present and examine the textual evidence that has been taken to lead to the extreme views and also present objections to what I take to be representative accounts of the extreme views in the literature. I argue that an interpretation which accommodates the maximum textual evidence will be one which sees Kant as holding some kind of idealism or mind-dependence with respect to the spatio-temporal world of appearances, but which does not see this as extreme phenomenalistic idealism which understands the world of appearance as a construction out of what exists (actually or potentially) merely in minds. With respect to things in themselves, I argue that doing justice to the textual evidence requires an interpretation which sees Kant as committed to the existence of something independent of us that grounds the reality we experience, as well as to the claim that we cannot cognise this something, while not holding that we should understand this as a commitment to the existence of supersensible, non-spatio-temporal objects, distinct from the things of which we have experience. The rest of the book presents my positive account of the position.

In part two of the book I present two central components of my positive approach. I argue that to understand both Kant’s argument for transcendental idealism and transcendental idealism as a position, it is important to note his concern with cognition (as opposed to knowledge), as well as the significance of his account of the necessary ingredients of cognition: intuitions and concepts. I argue that clarifying Kant’s technical notion of intuition is crucial for understanding both the nature of Kant’s idealism as well as his argument for it. Kant says that intuitions are singular and immediate representations that give us objects, and he contrasts them with concepts, which he says are general, mediate representations which can never give us objects. As I read Kant, intuitions are representations which give us acquaintance with perceptual particulars; this relation is crucial for cognition to connect with a world (to be about objects in the right way). On this account, when one has an intuition of an object that object is actually present to one’s consciousness, so a hallucination would not count as outer intuition. I argue that understanding intuition in this way enables us to make sense of Kant’s central argument for his idealism: it is because he holds that our representations of space and time are both intuitions and a priori that he concludes that they cannot present us with features of things as they are in themselves. In addition to my analysis of the role and nature of intuition, the other central aspect of my presentation of my positive account of Kant’s idealism involves sketching an account of perceptual experience that allows that we can be directly presented with objects outside us in space while not necessarily perceiving these objects as they are independent of the possibility of their being perceived by us. I first present such an account of perception, independent of Kant, and use this to give an account of a kind of mind-dependence that does not involve existence merely in the mind. I argue that we can make sense of qualities of things which they have only in relation to the possibility of their being present in conscious experiences, and I call such qualities essentially manifest qualities. I argue that Kant’s idealism consists in limiting the spatio-temporal world to what is essentially manifest.

In the final part of the book I put my understanding of the mind-dependence of appearances together with my account of Kant’s commitment to things as they are in themselves and give an account of the relation between my interpretation and the role of the categories in the Transcendental Deduction. I close with a discussion of the relation between Kant’s transcendental idealism and his account of the possibility of metaphysics, as I read it.

Response to McLear

Colin McLear’s detailed, interesting and challenging comments focus on my account of intuition as giving us acquaintance with objects. As McLear notes, I read the singularity of intuition in terms of the idea that intuition presents us with perceptual particulars. I allow that what a subject is presented with as one perceptual particular might not in fact be one object in the full-blown Kantian sense in which an object is a relatively persisting causally unified subject of properties which is made up of stuff that persists through its coming into and going out of existence, and is in causal interaction with all other objects. For example, a perceptual particular could be made up of two objects (in the full-blown Kantian sense of object), or could be something which is not an object, such as a patch of light. McLear thinks this creates a worry for the ‘fit’ between what is given in intuition and what is cognised as a result of cooperation between intuitive and conceptual representation. He points out that, on the one hand, I allow as singular perceptual particulars things that don’t correspond to singular objects; on the other, I argue that a central role of the categories is to exclude ‘gerrymandered’ concepts and properties as candidates for cognition, because such concepts are not excluded by what is given in intuition. McLear argues that it is not clear how intuition can make a determinate semantic contribution to thought while being unable to determine how it is that one should think of the object so presented.

On my view, the determinate contribution intuition makes is putting us in direct cognitive contact with the things about which we are able to think, the things to which we apply concepts. What is presented in intuition has determinacy in terms of spatio-temporal location and boundaries. Further, it is the actual things (empirically real objects) that determine both what is presented to us in intuition and what concepts apply to these things, but the fact that things are presented to us in intuition is not enough to explain what puts us in a position to think about objects, just as the existence of distinct spatially bounded and located objects does not solve the binding problem—the problem of how we succeed in perceptually representing distinct objects on the basis of a mass of sensory input. The question with which I am concerned is not what determines how one should think of the object present to one, but what puts one in a position to think about it at all. As I read him, Kant’s concern in the Deduction is with what it takes for us to be in a position to think about objects (to apply empirical concepts to them); he argues that this requires the application to objects of the a priori concepts he calls the categories; and he argues for this on the basis that without concepts representing necessity and universality (and thereby, as I read it, ruling out gerrymandered objects and gerrymandered properties) we would not have a determinate object of thought and would not be in a position to apply empirical concepts to anything. The categories enable us to think about the objects given in intuition and this may involve discovering that what we initially took to be one object is in fact two. We couldn’t discover this without the categories because their role of limiting what counts as a property and what counts as a subject of properties is crucial in enabling us to learn things’ properties, including learning their principles of causal unity, which is part of what is involved in learning that what we at first took to be one thing is in fact two separate causal unities.

As McLear notes, I say that for Kant, when we have an intuition of an object, that object is present to consciousness, and I explain this in terms of the notion of acquaintance and the notion of direct presentation. McLear worries that I am not clear as to whether this is an epistemic relation to an object or a metaphysical feature of the state one is in when intuiting an object. Kant does not explain his notion using these terms but I think that it is fair to say that his notion has both epistemic and metaphysical aspects. It is not epistemic in any sense directly involving knowledge, since, as already noted, cognition, for Kant, is not the same as knowledge, and intuition is not cognition, but merely one ingredient in cognition. McLear suggests that the epistemic component in my reading is in my saying that intuition puts a subject ‘directly in touch’ with, and so provides non-inferential awareness of, a perceptual particular. However, he worries that I also construe the relation as metaphysical by seeing it as being ‘object-dependent’, in that we have an intuition of an object only when that object is in fact present to us. He sees similar ambiguity in my talk of ‘presence’ since a mere requirement on the existence of the object for some relation does not tell us anything about the epistemic relation.

In response to the last point, my account is not simply that the object of intuition is present to the subject in the sense of existing in the same space as the subject, but that it is present to the subject’s consciousness. This, in my view, brings out how the epistemic role and the metaphysical requirements are connected, in the same way as in the relational accounts of perception I draw on. According to such accounts, a conscious state of perceiving an object intrinsically involves the presence to consciousness of the object perceived; on this analysis, a state of merely seeming to perceive an object is metaphysically different from an actual perceptual state and this difference is key to the role perception plays in cognition and knowledge: that of putting us directly in touch with a world of objects.

As McLear notes, I understand Kant’s talk of our being ‘given’ objects in terms of our being acquainted with them or directly presented with them; McLear asks why one should construe givenness in these terms. This excellent question focuses on the crucial question of how to understand the cognitive requirement intuition satisfies: what is the unique contribution to cognition made by intuition—a contribution that could not be made by concepts? As McLear says, I read Kant as holding that acquaintance is required for referential thoughts (on a certain reading of ‘referential’). However, while I do hold this, I hold that the contribution intuition makes to cognition has broader significance than this. The point is not just that we need to be able to have successful referential thoughts—that we need to be able to refer to objects if we are to have knowledge about them—but that we will not have what Kant calls cognition without the possibility of being in direct conscious contact with the objects of cognition. I argue that Kant’s specific concern in the Deduction (his concern with what it takes for the categories to have ‘relation to an object’) is a concern with the conditions of a certain kind of referential thought or referential concept application, and that he argues that this requires that the categories be brought into relation to what is given in empirical intuition. Without the possibility of bringing concepts in relation to intuitions, a certain kind of thought about objects is not possible. As I read him, Kant thinks that concepts never uniquely individuate or pick out an object, that having a concept is always compatible with the possibility that there exists no object corresponding to the concept, and that merely having a concept cannot, on its own, give you the kind of direct connection to an object that guarantees that there is an object. This is why conceptual thought, on its own, lacks what Kant calls objective validity. The objective validity of cognition requires concepts and intuitions together, with intuitions putting us directly in touch with the objects about which concepts enable us to think. On my account, the way intuitions guarantee that there are objects corresponding to our concepts is by presenting us with them, and this is also what enables intuitions to individuate or pick out particulars.

McLear’s interesting and challenging discussion pushes me on the point as to whether it is actual or merely possible acquaintance that is necessary for referential thought and cognition on my account. It may be that I am not always sufficiently clear about this, as I introduce and explain my account with respect to the basic case of referential thought of something with which the subject is actually acquainted. My view is that actual intuition is required for (theoretical) cognition, but that actual intuition of the specific object that is being cognised may not be needed, so long as it is both the case that this object is causally connected to something which is actually presented in intuition and also the case that the object of cognition could be presented in an intuition. Both possible and actual intuition feature here: it must be possible for the object being cognised to be presented to consciousness in an intuition and this object must be causally connected to something that is actually presented in an intuition. As I read Kant, he holds that concepts lack relation to an object unless they apply to objects which are given to us in empirical intuition, or to objects which are causally connected to something given to us in empirical intuition and also possibly presentable in intuition. Whenever we have (theoretical) cognition and whenever we have referential thought, there is something of which we have an actual intuition (an object actually present to consciousness), though this may not be the object referred to in the thought. Thus, acquaintance immediately puts a subject in contact with its object, and thereby guarantees the existence of its object, but this does not mean that we have referential thoughts only of the particulars we are acquainted with; this is also possible with respect to particulars with which it is possible for us to be acquainted, where these are causally connected to something with which we are actually acquainted.

McLear worries that the mere possibility of acquaintance, plus causal connection, would not be enough to do the cognitive work I need it to do in my account, to make referential thought possible. Once I understand this cognitive work as providing certainty, McLear argues, I cannot understand it in terms of the kind of inferential connections causal connections would provide. As I see it, the crucial point is not that acquaintance provides certainty but that it provides presence—an actual connection to objects that concepts cannot provide. McLear wonders why, once I have allowed causal connections to be enough, I am in a position to disallow acquaintance with sense-data that stand in the right causal connection to external objects. The problem with this is that (at least on standard understandings of sense-data) there would be no objects with which we are actually immediately presented. But as I read Kant, he holds that cognition requires that at least some of the objects of (theoretical) cognition are actually presented in intuition, and that other objects of (theoretical) cognition are the kinds of things that could be presented in intuition and also are causally connected to things which are actually presented in intuition. This account allows for uncertainty in a number of ways. We can be wrong about the kind of thing we take ourselves to be presented with in intuition, we can be wrong about whether we are actually presented with something in intuition, and we can be wrong about what is causally connected to the things directly presented to us in intuition.

McLear understands Kant’s account to allow that there are aspects of empirical reality with which it is impossible for us to be acquainted; my reading of transcendental idealism denies this. In particular, with respect to magnetic matter (the example McLear cites with respect to this possibility), Kant says that “in accordance with the laws of sensibility and the context of our perceptions we could also happen upon the immediate empirical intuition of it in an experience if our senses, the crudeness of which does not affect the form of possible experience in general, were finer’ (A226/B273; emphasis added). Notice that here Kant acknowledges that our senses are not able to give us an empirical intuition of magnetic matter but also says that it is something which could in principle be given in an empirical intuition: something that could be present to a conscious experience. Crucially, he says of magnetic material that “we could have an immediate empirical intuition” of it. I take this to mean that for something to be empirically real, for Kant, it must be the case that it could be presented in an empirical intuition to a possible sense and that it is causally related to things which are actually present to our consciousness. Thus, I do not take the discussion of magnetic matter to undermine my reading of the centrality, in Kant’s account, of presence to consciousness in intuition.

McLear also worries about the claim that acquaintance is necessary for referential thought. He argues that I fail to distinguish between two distinct but both broadly referential notions—satisfaction and constitution. I do not think that it is easy to map onto Kant all the distinctions that might be drawn with respect to reference in contemporary philosophy, and of course the term can be used in a number of different ways. I therefore explain that,  as I am using the term, it is not enough for a referential thought that there is in fact an object that falls under the concept in the thought (satisfaction), but rather that the subject must be in a position to connect the concept with the object. The most straightforward case of this would be applying it to something immediately given in intuition. This is why, on my use of the term ‘reference’, Kant would deny that we have referential thoughts about monads even if there are in fact (unknown to us) monads and our concepts of monads are coherent—we think about them using coherent concepts. We are not able to have acquaintance with them and therefore are unable to connect with them—to be in a relation to them in which we apply concepts to them through picking them out. While I do not argue that this is the only way in which contemporary notions of reference feature in Kant, or that this is the only thing philosophers might mean by the term ‘reference’, this is a possible sense of referential thought, and, in my view, it is what Kant is concerned with in the Deduction, when he talks about what it takes for concepts to have ‘relation to an object’.

A problem for this account is that Kant seems to allow that we can have some kind of cognition, and therefore, one might think, referential thought (in my sense) of objects like God. Kant’s account of practical cognition is a complicated and, arguably, under-explored notion. In order to assess this worry we would need a detailed account of what it is in practical cognition that makes good the lack of intuition, and how this is done. This is a fascinating issue and not one on which I have anything to say in the book. All I will say about it for now is that, as I see it, the access to God, the soul, and freedom that practical cognition gives us is special, and requires some special account of how this fills the role played by intuition; further, I think that such an account is not, on Kant’s view, available with respect to notions such as that of monads. Thus I think the fact that Kant holds practical cognition to give us some cognitive access to the supersensible is compatible with denying that Kant allows us to have referential (in the specific sense of reference introduced) thoughts about monads. Further, whatever account we have of the way practical cognition fills the gap with respect to objects that cannot be given in empirical intuition, it is still significant that filling this gap does not provide us with theoretical cognition. A developed account of practical cognition should explain both how the gap filling works and the differences in the resultant cognition. This is a fascinating project, and one on which I wish I had more to say.

Response to Marshall

While Colin McLear pushes at the details of one central part of my account (my reading of intuition), Colin Marshall raises a wide variety of worries about a number of different aspects of the position. Marshall pushes more on the philosophical coherence of the picture I present than on its accuracy as an interpretation of Kant and in doing so raises what seem to me to be very interesting questions for the position.

Marshall notes that my reading falls within the tradition of ‘one-world/two-aspect’ interpretations of Kant’s idealism, and that a central claim in this tradition is to say that each appearing object of experience is, in some sense, the same thing as some thing in itself, sometimes construing this as a claim about numerical identity. I do not make this numerical identity claim: I think it is a misleading way of expressing the point. One problem I raise with making the numerical identity claim is that, on Kant’s account, we are not in a position to pose the question as to whether some thing in itself is an identical thing to some thing as it appears, because we are never in a position to individuate things in themselves. Further, I think saying that every appearance is a thing that is numerically identical with a thing that exists in itself problematically sees both appearances and things as they are in themselves as things, whereas they are, rather, aspects of things. It would be like saying that the round shape of an object and its red colour are numerically identical things. Rather, what we should say is that there is a thing which has a round shape and a red colour; further, we could say (for example on a Lockean view) that the thing is red only in our experience of it, and not as it is in itself, without asserting that redness and roundness are numerically identical things.

Two worries that may be raised about this are, first, whether it problematically makes appearances into properties rather than things, and second, whether it problematically (in terms of Kant’s views about the limits of our cognition) allows us too much information about things in themselves.

In relation to the first worry, I don’t think that the fact that we can characterise things as they appear to us only in terms of mind-dependent properties makes these things properties. We characterise objects by characterising their properties; Locke, for example, characterises objects as having primary, secondary, and tertiary qualities (which have different metaphysical statuses, as secondary qualities are mind-dependent). This does not mean that Locke turns objects into properties, and we could express Lockean realism by describing the mind-independent nature of primary qualities, in his account. In giving an account of the metaphysical status of the properties of spatio-temporal objects, Kant gives an account of the metaphysical status of such objects. Kant thinks, as I read him, that the essentially manifest qualities of things we are able to cognise exhaust the characteristics of spatio-temporal empirically real objects. He thinks that empirical knowledge and science gives us knowledge only of essentially manifest objects and does not need anything more. This does not make spatio-temporal, empirically real objects into properties; rather, it gives us a description of the status of the properties of the objects we know, and thereby gives us an account of the metaphysical status of the objects, including the fact that, as I understand him, Kant thinks what is empirically real is in a crucial sense not complete, and not metaphysically fundamental.

In response to the second worry, I argue that to say that the things we experience have a way they are independently of their appearing to us is not to say that picking out objects at the level of experience tells us how they are individuated in themselves. By way of analogy, grant, for the sake of argument, that there is a distinction between the visual colour we experience and the properties picked out by science that ground visual colour, such as spectral reflectance properties. Empirical science tells us that the physical properties which ground red, for example, are enormously varied. This means that picking out redness, at the level of experience, does not individuate properties at the level of the non-experiential properties physics studies, even if the former are grounded in the latter.

Marshall has two challenges to my central notion of essentially manifest qualities. The first challenge is to explain how properties which are essentially manifest are qualities of mind-independent objects; the second is to explain how to make sense of qualities that are manifest to us in perception but also dependent on the possibility of our perceiving them. These two sides of essentially manifest qualities are central to my account: that they have these two sides is what enables my reading to do justice to both Kant’s empirical realism and his idealism. Essentially manifest qualities, as I introduce the notion, are genuinely qualities of objects (not of merely mental states), but are mind-dependent qualities of objects, and, in particular, depend on the possibility of their being perceived by creatures like us. Defending this account of essentially manifest qualities, and its role in explaining Kant’s idealism and his empirical realism, takes up two chapters of the book (Chapters 5 and 6); it will not be possible for me to give an adequate account of the central features Marshall challenges here. I shall simply say something brief about each of the two sides of his worry.

Drawing on a Lockean account of colour, Marshall allows that we can easily understand how a mind-independent object (a spruce) could have the mind-dependent property of being green. However, he thinks this kind of example is of no help with making sense of the case in which we have no independent grasp on the mind-independent thing. This seems to be because he takes it that spatial qualities are essential to our grasp of colour, and in the Lockean account, spatial qualities are mind-independent. The problem, Marshall says, is that without knowing what the metaphysically deep properties of mind-independent objects are, we don’t have grounds for saying that the essentially manifest qualities are qualities of those objects. I’m not sure I see the worry. For one thing, he has appealed to the claim that spatial qualities are essential to our grasp of colour, not to the claim that the mind-independence of spatial qualities is essential to this. So long as we can individuate objects within experience, we do not need to individuate mind-independent things which ground appearance properties, although we do, as I understand Kant’s view, need the idea that there is something which does this. I think the worry Marshall raises might have more force against an account of qualities like colour which understands them as existing literally in the mind; if these were all we had access to, this might make it hard to see how we could have grounds for thinking that they are properties of objects. But essentially manifest qualities, as I understand them, are qualities of objects, not of something that exists merely in the mind.

This leads to the other side of Marshall’s worry: how essentially manifest qualities can be both manifest in perceptual experience and mind-dependent. I hold that empirical reality, for Kant, is exhausted by what is essentially manifest, and that what is part of possible experience, for Kant, is either something directly present in experience or something which is both causally connected to something directly present in experience and also something which could be directly present in experience. In presenting my account of essentially manifest qualities, I start by presenting an account of colour understood as essentially manifest, which I argue is a possible account of colour. Marshall argues that the phenomenology of colour is in tension with colours having the metaphysical profile the essentially manifest account ascribes to them. He appeals to views which argue that the phenomenology of colour experience includes the fact that perceptual experience reveals to subjects what things in their environment are like independently of their experience of them. I have two responses to this.

On the one hand, it seems unlikely, on any interpretation of transcendental idealism, that the metaphysical status transcendentally ideal features of reality have is revealed in ordinary perceptual experience. If this were the case, transcendental idealism would be the default common-sense view—I assume that it is not. So I don’t take Marshall’s phenomenological point here to be enough to show that an analysis of qualities as essentially manifest could not capture Kant’s account of transcendental ideality.

On the other hand, part of my account of essentially manifest qualities is that these qualities are, as experience presents them as being, both qualities of objects (not of merely mental states or ideas) and presented or manifested in experience. The nature of redness is revealed in experience (whether or not its mind-independent/mind-dependent status is so revealed). Thus, this account does justice to a central part of the phenomenology of colours: their being manifest qualities. The further possibility that colours are not just manifest but essentially manifest is not, in my view, something we should expect to be part of the phenomenology. It is crucial to note, against Marshall, that essentially manifest qualities do not exist in mere possibility—bursting into action when perceived—rather, like manifest qualities they are qualities of objects presented in perceptual experience, but further, are qualities the existence of which is limited to the possibility of their being presented in perceptual experience. Marshall suggests that we need an account of Kant’s reasons for thinking that empirical reality has this status—my account of this is in Chapter 8, and it turns on both my account of intuition and Kant’s arguments for space and time being a priori intuitions.

Marshall has concerns about my account of Kant’s notion of sensation and its role in cognition and, in particular, its role in intuition. This is not something on which I have firmly settled views, and I find Marshall’s discussion here very interesting. I am not convinced that Kant’s account of sensation is either worked out enough or presented in enough detail to resolve all possible debates here. In my view, the way he talks about sensations is compatible with it being the case that, at least some of the time, they are theoretical postulates in a sub-personal processing system—ways in which objects causally affect us which play a role in our experience of objects without it necessarily always being the case that they are conscious representations. On my view, sensations are not intentional on their own (intuitions are intentional representations that present us with objects), but Marshall wonders how ordering something non-intentional could produce direct acquaintance with an object outside us; as I understand it, this is a problem for all direct realist or relational accounts of perception that (rightly) acknowledge that perceptual experience requires mental processing of a causally produced sensory input. The further complication for Kant’s account is that he holds that one part of what enables us to go from a causal input to experience of objects involves an ordering contribution from the mind which does not present the way reality is independently of the possibility of its being presentable to minds. This is, no doubt, a philosophically complicated position. It is not part of the view, as I read it, that the forms in which we organise the causal sensory input reflect something merely mental; they enable us to be presented with things outside us, but only with aspects of those things which are limited to the possibility of their being presentable to us. However, as I said, I do not have firm views on Kant’s account of sensation, I think that this is something that would benefit from more attention and I find Marshall’s suggestions here interesting.

Marshall argues that my account of Kant’s argument for his transcendental idealism (Chapter 8) neglects important alternatives to certain claims that are key to my account. I argue that, in the Transcendental Aesthetic, Kant does not establish transcendental idealism by postulating it as a theory that would close a justificatory gap, enabling us to reach conclusions about the way objects are in the spatio-temporal world on the basis of features of our minds. In particular, I hold that Kant’s transcendental arguments concerning the synthetic a priori conditions of cognition primarily enable us to reach conditional conclusions about the objects we can cognise, and that Kant does not argue for transcendental idealism on the basis of the claim that it would enable us to convert these conditional conclusions into conclusions about all spatio-temporal objects, but rather he argues for this latter move on the basis of already having established transcendental idealism. I don’t see that the possibility of transcendental arguments with other conclusions and premises affects this.

Marshall argues that my account of Kant’s argument for his idealism is open to a sceptical objection. The objection is that there could be states that are subjectively just like intuitions yet which do not in fact present any object; Marshall calls these quasi-intuitions. He asks what basis we have for thinking that we have intuitions rather than quasi-intuitions. It is not clear to me that Kant needs an answer to this question at this point: as I read him, his central concern is not in showing that we can rule out radical sceptical possibilities, but in giving an account of the nature of the world of which we have cognition, on the assumption that we do have empirical cognition. Further, I don’t find it obvious that the possibility Marshall is concerned with is really a possibility with respect to an a priori intuition. If, as I hold, the object an a priori intuition presents just is, thereby, present to consciousness, then there could not be a quasi-intuition, in Marshall’s sense, of space.

Marshall thinks I do not have a basis for ruling out a version of the so-called neglected alternative hypothesis which holds that the spatio-temporal structures present in our experience could also be present in reality as it is independent of us. I agree that in the Aesthetic Kant does not have an argument that rules out there being some structures in things in themselves that are somehow space and time-like. (I think he has further arguments for this in the Antinomies). However, I think that he does have an argument for the claim that whatever space-like or time-like structures there may be in things in themselves, these are not the structures that are present to us in our experience of space and time. In order to make my account of the argument in the Aesthetic work, I attribute to Kant an unstated premise that things that are independent of us get to be present to the mind only by affecting us. Marshall finds this premise implausible with respect to structures, such as the representations of space and time. My aim is not to defend the premise, but rather to defend the plausibility of Kant’s having held it, or something like it, as it makes his argument go through at exactly the place he presents it, in exactly the way he presents it. The plausibility of the premise is something on which I welcome further philosophical reflection, as it is needed to determine whether Kant’s argument is sound.

Like McLear, Marshall has some doubts about my account of the Deduction, and, in particular, my view that part of the argument of the Deduction is that the categories enable us to rule out gerrymandered objects and properties, and that this is a necessary condition of referential empirical thought. Marshall holds that someone could take herself to be in a very chaotic world, filled with gruesome properties and gerrymandered objects, and yet have no trouble in thinking about any possible combination of objects with properties. In my view, this claim is precisely what we need to deny to make sense of the part of Kant’s argument in the Deduction that I discuss. I argue that he holds that someone in this position would not actually be in a position to have a determinate object of thought, and to investigate the objects in the world around them to work out what properties they have. I agree with Marshall that generality itself does not suggest substantive restrictions on representation; I argue that what troubles Kant is the combination of generality and spontaneity, which threaten the possibility of determinate concept application. I don’t take myself to have sufficiently developed this account of the argument in the Deduction. In the book, I sketch it as a possibility; defending it as an interpretation of Kant and as an argument is an ongoing project.

Concluding Remarks

In closing, I want to respond to something Marshall says at the beginning of his piece. Marshall notes that I pay relatively little attention to Kant’s pre-Critical work and don’t discuss his contemporaries at all. I can only plead guilty to this: the book would have been a much longer book had I attempted to do this, and, in my view, this would have undermined from the close and continuous presentation of arguments based on philosophical concerns and primary texts, to present an interpretation of transcendental idealism. It would not, therefore, have been the book I wanted to write. This does not mean though that I disagree with Marshall’s pointing out the significance of this absence, or that I think investigating Kant’s contemporaries in relation to my reading would not be fruitful—I think it would be very fruitful indeed and I would welcome more work on it.

While I don’t think Marshall means this as large criticism, I will take responding to it as a brief opportunity to say something about my approach to the history of philosophy. My view is that a great variety of approaches, from the extreme of detailed contextual scholarship that pays attention to, for example, the typescript and size of original editions, to the other extreme of books like The Bounds of Sense, which are more like a riff on the primary texts, can all make useful and important philosophical contributions to the history of philosophy. Given the amount of time it takes me to bring philosophical arguments and Kant’s central Critical texts together (as well as my philosophical interests), I think my work will fall somewhat closer to The Bounds of Sense than the detailed contextual history extremes of the spectrum, but I have learned much from works with different approaches. Thus, I think that with respect to methodology in the history of philosophy, we should ‘let a thousand flowers bloom’; further, I think this is in fact happening, and that there is a wonderful amount of rapidly growing, vibrant and diverse new work on Kant (to which my two interlocutors are contributors), and that this is a fantastic time to be working on Kant.

Received: 14 August 2016.

© Lucy Allais, 2016.

Lucy Allais is Henry E. Allison Endowed Chair in the History of Philosophy at University of California, San Diego and Professor of Philosophy at Wits University, Johannesburg. Her work focuses on Kant’s philosophy and on topics related to wrongdoing and justice such as forgiveness and punishment. She has published widely in various journals such as Journal of the History of Philosophy, Philosophy and Phenomenological Research, Proceedings of the Aristotelian Society, International Journal of Philosophical Studies, Journal of Moral Philosophy and South African Journal of Philosophy