Kantian appearances and intentional objects

While disputes about the interpretation of Kant's transcendental idealism continue, there are perhaps currently some tendencies toward convergence in the literature. A number of recent interpretations agree on the need to make sense of Kant's transcendental idealism as a metaphysical position which holds that appearances are grounded in really existing things in themselves and are genuinely mind-dependent.1 Dispute is focused on how exactly to understand the mind-dependence of appearances and their relation to things in themselves. In her dense, rich, bold, combative book, The World According to Kant, Anja Jauernig presents an intentional object interpretation of Kantian appearances. While she is not the first to understand Kantian appearances in this way,2 the detail with which she develops the account makes a welcome and important contribution, which enlarges our interpretative options, including because this detail helps bring out differences between intentional object and phenomenalist readings. Jauernig sometimes explains her view through formal statements of a mind-dependent relation, such as that ‘[t]he being of an entity E is mind-dependent if, and only if, necessarily, if E has being, there is a finite mind, M, and E has being, at least partly, in virtue of the fact that M has, or could have, a possible complex representation whose presentational content includes E’ (Jauernig 2021: 44). This is compatible with a number of different views (for example, both hers and mine), so there is much more to be said in philosophically explaining the mind-dependence at issue, which is where she appeals to intentional objects. Jauernig aims (as do I) to explain a kind of mind-dependence which does justice to the empirical reality of appearances and to their being grounded in things in themselves.3 While there is a wealth of detail in her account, I will focus here on the basic idea of appealing to intentional objects as an explanation of the mind-dependence of appearances. Neither ‘intentional object’ nor ‘in-existence’ are terms Kant uses, and ‘intentional in-existence’ is not obviously less in need of explanation than ‘appearing’. I don't take this to be an objection to Jauernig's approach, since I take it that the aim is to explain the mind-dependence of Kantian appearances in terms of an independently philosophically understood notion.4 But of course this means we need an account of intentional objects to work with. Intentional objects are what it is that mental states are about or represent. They feature in discussions about both thoughts and perception, though not always in the same ways, potentially raising different issues in the context of explaining Kant's account (as I note below). Both intentional theories of perception and accounts of the intentional objects of thought take it that in the case of veridical perception or thoughts about existent objects the intentional objects of perception and thought are the ordinary (mind-independent) objects in the world that we perceive and about which we think. This is a straightforward sense of ‘intentional objects’, but comparing Kantian appearances with intentional objects understood in this sense would result in a purely realist account of transcendental idealism.5 This is not what Jauernig has in mind of course, but it is worth noting that it is a central meaning of intentional object. In order for the notion of intentional objects to explain a kind of ideality, what is needed, as Jauernig says, are merely intentional objects: objects which exist only insofar as they are the contents of mental states, or, as she puts it ‘in-exist’ in thoughts. A central way merely intentional objects are understood philosophically is when they are appealed to in accounts of hallucination and in explaining thoughts about non-existent objects, though they are controversial in both contexts. The most obvious thing to note about merely intentional objects as they typically feature in such accounts is that it is usually taken that they do not exist.6 Pegasus, Vulcan, Santa Claus and phlogiston can be regarded as the merely intentional objects of thoughts – as what thoughts are about – which do not exist. This might seem to make merely intentional objects not promising for an account of Kant's idealism that includes his empirical realism, since it would seem to put appearances on an ontological footing with hallucinations, fictions and objects mistakenly postulated by incorrect scientific theories. Despite this, merely intentional objects have been argued to be helpful for explaining Kant's account because they supposedly enable us to give an account of what it is for the existence of appearances not to be something over and above the existence of mental states with certain representational contents, but without identifying appearances with mental stuff.7 This might be thought to give intentional object accounts an advantage over phenomenalist ones, because while phenomenalist accounts which identify appearances with mental states might seem to make appearances private, Kant takes appearances to be public. Intentional objects accounts can deal with this by noting that it is no part of the content of our representations of objects in space that they are private, and therefore what is represented is not something private. As Jauernig says, on this account appearances exist in virtue of the contents of thoughts (or representational mental states), and she says that in talking about what is represented in our mental states Kant means to be referring to the intentional objects that are immanent to our mental states and conceptions … not to these mental states or conceptions themselves (Jauernig 2021: 33–4). Kant expresses the mind-dependence of appearances when he says that appearances are representations that exist in us. Interestingly, while Jauernig takes ‘in us’ to be literal, and obviously so, she does not take literally Kant's repeated claim that appearances are representations, since she takes them to be representeds (Jauernig 2021: 139). The distinction between representations and representeds is not directly presented in Kant's text, and if there is a literal reading of representations in us, it suggests mental states themselves. I don't take this to be an objection to her account, since I don't think ‘in us’ should be read literally either – or, more accurately, I don't find it obvious that there is a straightforward literal sense to appeal to, since the various ways in which things are ‘in’ mental states is the subject of much philosophical controversy – including in Kant's time.8 For example, directs realist about perception hold that perceptual states present us with mind-independent objects outside us; but perceptual states are mental states, so a direct realist can say that there is a straightforward sense in which the objects of perception are ‘in’, or constituents of, mental states. Interestingly, an initial difference between an intentional object account on the one hand, and phenomenalist and relational (essentially manifest) accounts on the other, is that the latter are able to take more literally Kant's claim that appearances are representations or presentations,9 while perhaps, at least for the relational account, less literally his claim that they are ‘in us’, though they can give an account of the latter.10 The idea of representational content is of course itself complex and disputed philosophical territory. As with ‘intentional object’ there is a straightforward realist reading of Jauernig's phrase ‘the content that representations present or disclose’ (Jauernig 2021: 33). Jauernig sometimes sounds like this is what she has in mind. She says, for example, that Vorstellung is ambiguous. It can refer to ‘what does the representing, that is, the content-bearing vehicle, for example, a perception that represents a tree or a dream that represents a unicorn. But the term can also be employed to refer to the content that is represented in a certain mental state or conception, for example, a tree or a unicorn. When Kant says that “appearances are representations” … he is using the term in the second sense’ (Jauernig 2021: 33–4). This might seem to take us back to, on the one hand, comparing appearances with content in the sense of things in the world like trees which are disclosed in perceptual experience (which doesn't give us idealism) or comparing appearances with things like unicorns (which doesn't give us empirical realism). The idea Jauernig wants to appeal to is the idea of contents which are represented in mental states, which have no reality apart from being so represented, but which are not fictions, but rather, constitute the empirically real world. One question that can be asked about this concerns the ontological status of merely intentional objects. Another question concerns how to distinguish the representational contents that constitute empirically real objects from those that are merely fictional or otherwise non-existent, as well as from merely individual representational contents. In terms of the first issue, Jauernig claims – and I agree – that, for Kant, from the standpoint of (what she calls) fundamental reality, empirically real objects really exist and have a different ontological status than non-existent intentional objects such as fictions. But, as Jankowiak argues, it is hard to see how this can be the case if appearances really are merely intentional objects.11 If empirically real objects are distinguished from fictions by the contents of certain representations, or by some kind of grounding relation to things in themselves, they still have the ontological status of ‘in-existing’ in representational mental states, which is the same ontological status as other fictional and other non-existent objects of thought.12 However, this is not an objection to the possibility of an intentional object account, just to one claim Jauernig makes about the ontological status of merely intentional objects. The second question is what determines which features of the representational contents of our myriad mental states constitute empirical reality. Of course Kant's a priori conditions of experience are part of what determines this, but equally obviously are not enough, since they are merely formal. Giving a detailed account of this is a major focus of Jauernig's book. My interest here is in a general feature of her account: the idea that what determines the content of the representations that constitute empirical reality depends on what is represented in (specified parts of) actual human representational states. Jauernig says that ‘an intentional object of a piece of empirical thinking ET actually exists at the empirical level of reality if, and only if, ET is a member of a collection of representations that amounts to experience’ (Jauernig 2021: 76). She sometimes describes experience as including things like a list of causal laws, a list of true inductive generalizations that have been made so far, and a record of all appearances that have been perceived to date, a collection of empirical judgments and judgments of experience made so far (Jauernig 2021: 94–5), and she uses the metaphorical expression of what is ‘on the collective books of human minds’ (Jauernig 2021: 362). This makes it sound like things written down, or more charitably, things at least consciously entertained or agreed on by some minds at some time (Jauernig 2021: 82). This brings out an interesting difference between her account, on the one hand, and phenomenalist and relational (essentially manifest) accounts, on the other, which is that the latter can characterize empirical reality in terms of what could be experienced or would be experienced, and not just what actually is experienced. I think Jauernig is correct to present her view in this way, as I don't see how merely intentional objects could have properties other than those they are in fact represented as having, or, more precisely, those that are grounded in some actually represented contents. A phenomenalist account and my relational account, in contrast, can take such properties to depend on the (partly mind-dependent) experiences that subjects like us could have and would have, in certain circumstances. I find this feature of her account very helpful in bringing out some of what is at stake in the differences between the various accounts of the ideality of appearances. As things the existence and nature of which depend on thoughts and other attitudes, merely intentional objects cannot have properties that go beyond what is represented in these thoughts or is somehow grounded in what is actually represented in thoughts. Truths which are fully ‘specified’ (in Jauernig's terms) by what happens in the actual course of the regress of experience cannot also be ‘specified’ by what would have occurred or could have occurred in the regress: what would have occurred or could have occurred in the regress is not made true by the contents of our representations. A phenomenalist or relational account can allow that Kant takes empirical reality to be thoroughly mind-dependent but also allows that there could be true claims about empirical reality that are never discovered but are there to be discovered because they could have been discovered. Saying that appearances are fully ‘specified’ by what is represented in the extent of human experience, as Jauernig does, is stronger than saying that they don't exist apart from the extent of (actual) possible experience (which is my understanding of Kant's view13), and it allows for a different kind of empirical realism.14 Consider the idea that there might, contingently, be determinate scientific truths that no humans will ever discover. A phenomenalist or relational account of Kant's empirical realism can allow that this is a possibility, but I think a merely intentional object account must deny this. It might be said that it is part of our collective representational contents that what is true of, for example, galaxies we could learn about if humanity survives another century, is independent of our survival. But while it could be part of our collective representations that there are such truths, if humanity is destroyed in the next thirty years, it will never be part of our collective representations what these truths are. If we allow that our ‘collective view’ (to use one of Jauernig's terms) is that objects in space and time are represented as having properties we may (contingently) never discover,15 an intentional object view can allow that empirical objects, as merely intentional objects have such properties. But it is unclear how it can account for there being any determinacy to what these properties are, since the objects of experience don't have properties other than those they are represented as having and those that are grounded in ones they are represented as having, and properties we never discover are not represented. This might be why an intentional object account appeals to those who want to emphasize the indeterminacy of experience in Kant's account.16 I agree that Kant takes empirical reality to be indeterminate in certain respects (as well as incomplete) but I take this indeterminacy to line up with what we couldn't ever or wouldn't ever experience (such as a smallest part, a complete division or a boundary to space), rather than simply what we happen never to experience. It seems to me that Jauernig's merely intentional object account makes empirical reality more indeterminate than this, which is helpful in showing that there are real differences at stake between merely intentional object and phenomenalist interpretations. I have suggested that Jauernig is right to treat an intentional object account as tied to actual experience. But it might be argued that a different version of an intentional object account could allow that there are truths out there to be discovered, even if we do not in fact discover them, because they are a function of the way things in themselves would ‘affect’ us. Truths about how things in themselves would affect us are not truths about empirically real objects (which, in the account, have no contents other than what is specified by the contents of experience), but perhaps it can be said that there are truths about the representational contents we could have given the way things in themselves would ‘affect’ us. This is a place where the difference between perception (which is more passive) and thought (which is more active) makes a difference. We can make sense of the idea that if humanity survives long enough to stand on some particular place on Mars there are things that are there to be seen, and therefore ways in which things would look. But it is less clear that, if we get to stand there, there are thoughts we would have – what thoughts we would have seems a lot more contingent, given the more active nature of thought. It seems to me that an intentional object view that wants to allow for contingently undiscovered truths needs to appeal to perceptual states we would have in other circumstances,17 but I find it hard to see how to make sense of perceptual representational contents (that are a function of the way things are in themselves would affect us) as merely intentional rather than relational (as in my account) or ‘caused’ by objects (as in a phenomenalist account).18 Such states seem to present things (whether or not they present them as they are in themselves) and to have content that is a function of the way things are in themselves, and therefore is not merely intentional. It may be that Jauernig can avoid this issue, through her actualist account. Jauernig sometimes makes strong claims which sound like none of the representational content of appearances depends on the way things are in themselves, for example, she says that ‘[a]ppearances exist and have all of their determinations and other ontological ingredients in virtue of being represented in … ‘human experience’ (Jauernig 2021: 64, 156). It seems to me she is right in presenting a pure merely intentional object view this way, but, while I lack space for discussion, this makes it difficult to accommodate Kant's view that appearances are appearances of things as they are in themselves on which they partly depend. Further, while intentionalist accounts of perception sometimes assume that there is a straightforward notion of intentional content shared between veridical perception and hallucinations, it is arguable that our understanding of how perceptual states have representational content depends on our understanding of veridical perception, and that we can understand merely intentional perceptual objects only derivatively. This is part of why noting mind-independent intentional objects, as I did at the start, seems to me important- we might illegitimately import into our understanding of merely intentional objects content we can understand them as having only by appeal to mind-independent intentional objects. Unless we want the appeal to merely intentional objects to be simply a notational variant of Kant's terms, rather than a philosophical explanation of them, it seems to me that philosophical work would be needed to defend the idea of perceptual states which are both merely intentional and which also are a function of the way things are in themselves. Some might take this to be an irrelevant appeal to contemporary philosophy of perception, but how to understand the very difficult notion of perceptual appearing was disputed in the early modern period, and is disputed today, and I see no reason why philosophical understandings of these disputes could not inform each other (it is a central hope of my research that they can). In my view, in the absence of a detailed philosophical defense of such merely intentional states, a merely intentional object account should, as Jauernig does, appeal only to actual representational contents, meaning that her view best represents the resources and trade-offs of a merely intentional object account. Bringing these out really advances the debate on interpreting transcendental idealism and is one of the many contributions of this dense book.