Sameness beyond numerical identity. A defence of the One Object View of Kant´s transcendental idealism

Some Kant scholars argue that appearances and things in themselves are distinct things (Two Objects View). Others argue that they are the same things (One Object View). This last view is often understood as the claim that appearances and things in themselves are numerically identical (Numerical Identity). However, Walker (2010) and Stang (2014) show that Numerical Identity clashes against Kant’s claim that we lack knowledge of things in themselves (Noumenal Ignorance). I propose a weaker version of the One Object View that is not couched in terms of Numerical Identity and, consequently, avoids the problem raised by Walker and Stang. My case is based on a sustained analogy with perceptual experience that aims at showing that appearances and things in themselves are the same things in the following sense: the very same things can be presented under the mode of sensory intuition or (possibly) under the mode of intellectual intuition. Those things presented under the mode of sensory intuition are appearances; presented under the (possible) mode of intellectual intuition are things in themselves. This way of construing appearances and things in themselves preserves the core insight of the One Object View. At the same time, as it does not entail any isomorphism between appearances and things in themselves, it does not clash against Noumenal Ignorance.


Introduction
The debate about how to make sense of Kant's distinction between appearances and things in themselves started as soon as his first Critique appeared and still goes on today. Usually, interpreters are divided in two camps. The first camp is that of the B Mattia Riccardi riccardi@letras.up.pt 1 Faculdade de Letras da Universidade do Porto, Instituto de Filosofia, Porto, Portugal Two Objects View, which says that appearances and things in themselves are distinct things. The second camp is that of the One Object View, which says that appearances and things in themselves are the same things. There is widespread agreement among Kant scholars that the textual basis offers only limited guidance in adjudicating the dispute. On the one hand, some passages seem to straightforwardly support the Two Objects View, as when Kant describes appearances as "mere representations, which … have outside our thoughts no existence grounded in itself" (CPR A490-91/ B518-19). For how can "mere representations" with "no existence grounded in itself" be the "same" as mind-independent things in themselves? On the other hand, however, Kant sometimes seems to say that appearances and things in themselves are the same things, as when he describes their distinction as that "between things as objects of experience and the very same things as things in themselves" (CPR BXXVII; see also B306). Consequently, the issue is often treated as a purely theoretical problem generated by ambiguities and, arguably, by a certain tension in Kant's thought. That will also be my approach in this paper.
But what is the dialectic of the dispute (in its current shape)? The formulation of the One Object View provided above is obviously vague, for it is not clear what it means to say that appearances and things in themselves are the "same" things. The most straightforward way of fleshing out this claim is in terms of numerical identity. Taken in this way, the One Object View reads: Numerical Identity: appearances and things in themselves are numerically identical. 1 Another famous claim held by Kant is the following: Noumenal Ignorance: we lack cognitive access to and, consequently, knowledge of things in themselves.
Again, it is not clear how we should interpret this claim. Noumenal Ignorance can be read as ruling out any kind of knowledge about things in themselves. A weaker reading would allow that we can acquire some purely conceptual knowledge about things in themselves generaliter, for instance that they are not spatio-temporal (see Willaschek & Watkins, 2020 for a recent discussion of these issues). However, what Noumenal 1 I am not sure whether One Object theorists really mean to endorse Numerical Identity. Both Marshall (2013: pp. 521-522) and Stratmann (2018, fn. 19) distinguish Numerical Identity from the weaker claim according to which appearances and things in themselves are "two aspects" of the same thing, and thus not numerically identical. While Marshall maintains that "most interpreters seem to take the 'identity' formulation to be more metaphysically perspicuous" (2013: pp. 521-522; see also the illustrative sample of quotes in fn. 5), Stratmann assumes that the contrary is the case. In this paper I am interested in articulating and defending a version of the One Object View which is weaker than Numerical Identity, not in defending Numerical Identity. Marshall (2013) is the only other scholar I am aware of who explicitly pursues the same strategy. (I will come back later to his proposal.) Ignorance clearly rules out is that we can enjoy cognitive access to and, consequently, knowledge of particular things in themselves, including that we can knowledgeably construe "ways of mapping from the one set [of appearances] to the other set [of things in themselves]", as Walker (2010: pp. 839) puts it (see also Ameriks 2003: p. 83 for a very similar formulation of the problem). Thus, we can formulate: Particular Noumenal Ignorance: we lack cognitive access to and, consequently, knowledge of particular things in themselves.
Only Particular Noumenal Ignorance will be relevant to our present discussion.
But why is it at all relevant? Walker (2010) and Stang (2014) argue that Numerical Identity is inconsistent with Particular Noumenal Ignorance. (See also Marshall 2013: p.539, who calls it the "noumenal counting problem"). Given that Kant's endorsement of the latter claim seems undisputable, rejection of the former is thus mandatory. Their argument to this effect is what I turn to now.

The argument
Why should we think that Numerical Identity is inconsistent with Particular Noumenal Ignorance? In this section I examine the simplified version of the argument that Stang presents as the most damaging one for the One Object View and which he calls the "synchronic individuation argument". 2 The argument is a reductio and works as follows.
Let us assume Numerical Identity and consider two spatially distinct appearances, a1 and a2-imagine, for instance, a book and a pen on your desk, occupying two distinct regions of space at a certain time t. Furthermore, let t1 be the thing in itself supposed to be numerically identical to a1 and t2 the thing in itself supposed to be numerically identical to a2. The following (valid) argument can be construed.
(3) a2 is not in location L.
The problem here is that Kant's Particular Noumenal Ignorance excludes that we can know propositions such as (6). This means that Kant cannot allow a (valid) argument the conclusion of which is (6). Of course, what makes the argument go through is Numerical Identity, which works as its underlying assumption. Thus, Numerical Identity cannot be a view Kant endorses. I think Stang is right that Numerical Identity leads to the kind of unacceptable conclusion he draws from it. 3 Anyway, I will assume that Numerical Identity must go. This means that my strategy for rescuing the One Object View will be oblique. Instead of confronting Stang's reductio head-on, I shall propose an alternative, weaker version of the One Object View that proves immune to it. To work it out, I shall proceed by analogy with the case of perceptual experience.

The perceptual analogy
A key insight of the One Object View is that Kant's distinction between appearances and things in themselves is (at least in part) 4 epistemic: appearances are things considered in their relation to our cognition, while things in themselves are the same things considered independently from their relation to our cognition (see CPR B307). That Kant's distinction has an epistemic dimension becomes even more evident when we consider how it maps onto the distinction between "phenomena" and "noumena". Being a phenomenon means being the object of sensory intuition. Consequently, the identification of appearances with phenomena is straightforward. Things are more complicated when we turn to the relation between things in themselves and noumena. Kant distinguishes between noumena in the negative and in the positive sense. Noumena in the negative sense are just things in themselves. Noumena in the positive sense are objects "of a non-sensory [intellectual] intuition" (CPR B307). Given that, Kant argues, we cannot know whether there are creatures equipped with a capacity for non-sensory (intellectual) intuition, we have no reason to assume that things in themselves are noumena also in the positive sense. However, Kant also stresses that the notion of intellectual intuition and, consequently, that of noumenon in the positive sense do not involve any contradiction. Thus, we can at least say that it is logically possible 5 that things in themselves are the objects of intellectual intuition (in fact, Kant seems to think that God is capable of intellectual intuition). 6 3 Stang considers possible maneuvers by the One Object theorist, starting with rejection of (5). However, he argues that these moves are ultimately untenable, because they lead to a position that is only nominally different form the Two Objects View. Later in my paper I come back to this issue. 4 I add this qualification because some defenders of the One Object View claim that the distinction between appearances and things in themselves is not only epistemic, but also metaphysical. That is the main difference between so-called epistemic versions (originally proposed by Prauss and Allison) and so-called metaphysical versions (defended for instance by Langton, Allais and Rosefeldt) of the view. As I defend that the difference between appearances and things in themselves also concerns distinct sets of properties, my position counts as a metaphysical one. 5 The "possible/possibly" qualifier in relation to intellectual intuition should always be read in terms of logical possibility throughout the paper. 6 This is a relatively standard way of understanding how the relation appearances-things in themselves maps onto the relation phenomena-noumena. See, for instance, Stang 2016: "things in themselves, the objects whose existence is "ground in itself", and which appear to us in space and time, cannot be objects of any sensible intuition, so they are negative noumena. Whether, additionally, they are also objects of an intuitive intellect, is a separate matter." For a more detailed discussion along these lines, see Onof 2019. This reading has been recently challenged by Jauernig (2021: pp. 338-339), who defends a "fictionalist" account of noumena. Her argument, however, seems unconvincing to me. For instance, she fails to address What this shows is that Kant construes appearances and things in themselves (at least in part) as objects of distinct modes of epistemic access. More precisely, Kant conceives of appearances as objects of sensory awareness (intuition) and of things in themselves as (possible) objects of intellectual awareness (intuition).
What does distinguish distinct modes of epistemic access? The difference is in the way things are cognitively presented to epistemic subjects. Thus, that the same thing can be presented under different cognitive modes of presentation means that there are distinct ways to access it epistemically. Of course, the notion of "mode of presentation" is of Fregean provenance. Frege famously noted that terms with different senses (Hesperus and Phosphorus) can refer to the same thing (Venus). What distinguishes the two senses is that they are different "modes of presentation" of the same thing. More recently, Frege's notion of mode of presentation has been applied to discussions concerning the content of conscious experience (see, for instance, Chalmers 2010, especially Chaps. 11 and 12). This implies a substantial departure from Frege's original conception, for it allows that modes of presentation are not only ways in which things are presented in thought, but also ways in which things are presented in perceptual experience. The notion of "cognitive mode of presentation" I employ here is equally liberal, for it allows for awareness-like modes of presentation (ways things may be presented in episodes of awareness, both sensory and non-sensory) as well as for thought-like modes of presentations (ways things may be presented in episodes of thought, discursive or non-discursive).
It is important to clarify why I appeal to such a non-standard version of the Fregean notion. This notion will help me to show how we can construe Kantian appearances and things in themselves as (at least in part) epistemic objects, i.e. as objects tied to specific modes of cognitive presentation. 7 More precisely, given this terminology we can say that appearances are objects under the mode of presentation of sensory intuition, while things in themselves are objects (possibly) under the mode of presentation of intellectual intuition. Hence: Footnote 6 continued the distinction between noumenon in negative and in positive sense. As I cannot offer any detailed discussion of Jauernig's reading here, I leave it for another occasion. 7 This also explains what may seem a puzzling feature of my strategy. Its goal is to offer a version of the One Object View that avoids construing the difference between appearances and things in themselves in terms of numerical identity. What may seem puzzling is that I pursue this strategy by using a notion-Frege's notion of mode of presentation-that was introduced to make sense of numerical identity. The puzzle disappears by appreciating that I appeal to (a liberal version of) the Fregean notion to construe appearances and things in themselves as (at least in part) epistemic objects, i.e. as objects under a specific mode of presentation. Frege never intended to use it in this way. In fact, there is simply no place in his ontology for an epistemic notion of object. It is helpful to compare the use I make of (a liberal version of) the Fregean notion in order to introduce an epistemic notion of object with the use Chalmers makes of it in order to introduce an epistemic notion of state of affairs involving subjective data. As Chalmers argues (2010: p. 37, fn. 1), states of affairs of this kind "should be individuated in a fine-grained way that is sensitive to the mode of presentation, as befits the epistemological role of data". Accordingly, ""The glass contains water" (observed by tasting) and "The glass contains H2O" (observed by chemical test)" can be seen as different states of affairs for they involve different subjective data. Given that appearances and things in themselves are (at least in part) epistemic objects, they should also be "individuated" in a way "sensitive to the mode of presentation". The remainder of this paper aims at explicating what this exactly means by analogy with the case of perceptual experience. (Many thanks to an anonymous referee for raising this puzzle and for urging me to clarify my usage of the Fregean notion.) possessing the capacity for sensory intuition means having epistemic access to appearances, for appearances are objects under the mode of presentation of sensory intuition, on the one hand; possessing the (possible) capacity for intellectual intuition means having epistemic access to things in themselves, for things in themselves are objects (possibly) under the mode of presentation of intellectual intuition, on the other hand.
With these conceptual distinctions in place, we are finally in a position to appreciate how reflection on perceptual experience can help us to see the sense in which Kantian appearances and things in themselves are the "same" things. Such a perceptual analogy is what the remainder of this section spells out.
Imagine holding a tennis ball in your hand. You see it with your eyes and feel it with your hand. As an item of visual awareness the ball is round and yellow (among other things). As an item of tactile awareness it is round and hard (among other things). Each sense modality provides awareness of some of the ball's properties, some of which can be apprehended by both. So you see the ball's yellowness and feel its hardness. But its roundness is a property you see and feel. Thus, we can say that vision and touch afford awareness of distinct, though overlapping sets of a thing's sensible properties. Nonetheless, what you see and what you feel is the same thing: the tennis ball.
Should we conclude that the object of touch-let's call it the tactile object-and that of vision-let's call it the visual object-are numerically identical? The answer to this question depends on what we mean with terms such as "tactile object" and "visual object". On the one hand, we can mean them non-epistemically. In this sense, we are entitled to say that the tactile object I feel and the visual object I see are numerically identical: they are both just the same thing-the same tennis ball. On the other hand, however, we can also mean "tactile object" and "visual object" epistemically. Understanding them in this way means construing them as (at least in part) conceptually tied to a specific cognitive mode of presentation. Accordingly, a visual object is necessarily an object under the specific visual mode of presentation, while a tactile object is necessarily an object under the specific haptic mode of presentation. This has two consequences. On the one hand, we can no longer say that a certain visual object is numerically identical with a certain tactile object, for whereas the former is necessarily an object under the visual mode of presentation, the latter is not (and vice-versa). On the other hand, it is still possible that the same thing be experienced under the visual as well as under the tactile mode of presentation. In fact, this is just what we have been imagining all the time: that I am seeing and feeling the same tennis ball. What might be harder to grasp is that the visually experienced ball is not numerically identical with the haptically experienced ball. But this is just another way to say that vision affords awareness of some of the ball's properties, for instance its yellowness, that cannot be apprehended by touch (and vice-versa). Thus, although the seen-ball-the ball under visual mode of presentation-is not numerically identical with the touched-ball-the ball under tactile mode of presentation-, what I am perceiving in each case is the same tennis ball. (For the sake of clarity, I will always use "object" as an epistemic term and "thing" as a non-epistemic term in the sense explained here.) 8 A point already touched upon is the fact that visual experience and tactile experience provide awareness of some of the ball's sensible properties. (There is overlap, we saw, between the properties I am visually aware of and those I am haptically aware of, such as roundness.) Accordingly, we can construe the visual (tactile) object as the subset of the ball's properties awareness of which is afforded by visual (tactile) experience. Thus, visual ball and tactile ball are just distinct (albeit overlapping) subsets of the properties constituting the tennis ball. The latter is a thing that can be presented as an object of visual awareness or as an object of tactile awareness.
I do not mean to suggest that what I have said so far constitutes a plausible analysis of perceptual experience and its object. Rather, I want to suggest that Kant's relation between appearances and things in themselves should be framed in a way analogous to the way in which I have framed the distinction between visual and tactile objects. (There are also some important asymmetries. Nonetheless, as I argue in due course, they do not undermine the main point I aim at drawing here.) As we saw, Kant understands appearances and things in themselves (at least in part) in epistemic terms: appearances are objects under the mode of presentation of sensory intuition, while things in themselves are objects (possibly) under the mode of presentation of intellectual intuition. Therefore, although we cannot say that a certain appearance is numerically identical to a thing in itself, it is possible for the same thing to be intuited both ways, just as it is possible for a tennis ball to be both seen and felt. In this sense, an appearance I am aware of via sensory intuition can be the same thing (possible) creatures equipped with the capacity for intellectual intuition would become aware of in itself.
As anticipated, there are important disanalogies between the perceptual case and how things are with Kantian appearances and things in themselves. To appreciate a first disanalogy, consider again the tennis ball in your hand. Touching it and seeing it afford awareness of some of its properties. But as we noticed above, there is overlap between the properties I can see and those I can touch, such as its roundness. When we turn to the Kantian distinction, however, there is no such overlap. For according to Kant all properties sensory intuition affords awareness of are mind-dependent and spatiotemporal, while all the properties intellectual intuition (possibly) affords awareness of are mind-independent and non-spatio-temporal. 9 Thus, while the visual ball and the tactile ball share some properties, such as the ball's roundness, there are no properties shared by appearances and things in themselves.
Does this undermine the entire strategy by analogy I have been pursuing so far? I do not think so. Consider a piece of cheese and the way we see and smell it. Arguably, there is no overlap between the properties I see and those I smell. But this does not mean that in the two cases I am not perceiving the same piece of cheese (under distinct sensory modes of presentation, of course).
But the disanalogy is deeper, one might argue, for two interrelated reasons. First, all perceptual examples I have given involve sensory capacities that make us aware of Footnote 8 continued Jauernig 2021: pp. 10-11). I discuss how my own account would be classified according to Jauernig's taxonomy in fn. 14. sensible properties, while Kant contrasts the capacity for sensory intuition überhaupt with the essentially different (and merely possible) capacity for intellectual intuition. Second, whereas the properties all sensory capacities make us aware of are properties of the same kind-they are all sensible properties-, the Kantian capacities make us aware of essentially different kinds of properties (mind-dependent and spatio-temporal vs. mind-independent and non-spatio-temporal properties). Note, however, that my previous discussion remained silent about the metaphysical nature of the sensible properties we become aware of in perceptual experience. In particular, I have nowhere assumed that they need to be properties of the same kind. The point I wish to draw here does not require any such assumption.
But in case you remain unconvinced, let us consider Sellars' famous pink ice cube example-an example that also involves a non-sensory mode of cognitive presentation (Sellars, 1991). The pink ice cube is an object of ordinary perceptual experience. Can we say it is identical to a system of physical particles? In general, can we say that a manifest object and a scientific object are the same thing? Again, numerical identity does not seem to work-or so argues Sellars-, for while the pink ice cube "presents itself to us as ultimately homogeneous" (Sellars, 1991: p. 30), the system of physical particles with which it is supposed to be numerically identical does not. Thus, if there is at all sameness, it must be couched in different ways. Extending my originally purely perceptual analogy to this example and adopting Sellars' terminology, we can say that the pink ice cube is a certain thing as it figures within the manifest framework, while the relevant system of particles is the same thing as it figures within the scientific framework. Thus, while a manifest object is not numerically identical to a scientific object, we can say that the same thing can figure in both the manifest and the scientific framework.
But what is "figure" supposed to mean here? To start with, it means that the same thing is presented under different cognitive modes of presentation. While the mode of presentation associated to the manifest framework is that of first-person sensory experience, the mode of presentation associated to the scientific framework is that of third-person physical description. Chalmers draws a similar distinction when he says that "color experiences attribute colors under a distinctive mode of presentation, one quite distinct from a physical mode of presentation of a reflectance property" (Chalmers, 2010: p. 391). (Again, I am not suggesting this way of solving the puzzle posed by Sellars is the right one; as before, I am just using this further example to push the analogy which-in my view-can help us to solve the Kantian puzzle.) Another disanalogy between the perceptual case and the Kantian distinction is as follows. All forms of epistemic access to things I have been considering so far (different sensory modalities as well as, in the last example, Sellars' manifest and scientific frameworks) are available to human beings. However, according to Kant's story we are only capable of sensory, and not of intellectual intuition. For as we saw, Kant claims that an appearance-an object we are aware of sensorily-also possesses a "constitution in itself" (CPR B306) independent of the fact that we are sensorily aware of it; and that we can non-contradictorily conceive a capacity for intellectual intuition that would afford awareness of such a "constitution in itself", even if we lack such a capacity entirely, nor have any reason to assume that there are creatures actually possessing it. 10 But crucial as these claims are to Kant's transcendental idealism, we do not need to bother about them here, for the point I am trying to make does not turn on them. It suffices to see that Kant construes appearances and things in themselves as (at least in part) epistemic-as objects presented under proprietary modes of intuition (even if one such mode remains for us a mere logical possibility). Realizing this, and pursuing the perceptual analogy worked out in this section, is enough to see that although an appearance cannot be numerically identical to a thing in itself, they can be the same thing presented under two distinct modes of intuition.
The interpretation I am proposing shares important features with two readings put forward in recent years. On the one hand, it shares with Marshall (2013) the explicit goal of articulating a weaker (but still metaphysical) version of the One Object View that does not take appearances and things in themselves to be numerically identical. Marshall does that by exploiting the notion of "qua-object". Accordingly, he argues that "Kant's distinction between appearances and things in themselves is a distinction between things-qua-appearances and things-qua-in-themselves, for the same things" (Marshall, 2013: p. 534). One advantage of my reading over Marshall's one is that it dispenses with such a "semi-technical" notion (526), as he himself characterizes it. Of course, I also make use of the liberally Fregean and, therefore, non-Kantian notion of "(cognitive) mode of presentation". But it seems to me that this notion imports less foreign theoretical weight into Kant's thought, for it is simply introduced as the determinable concept of which Kant's own notions of sensory and intellectual intuition are determinates. Therefore, interpreters who are skeptical about the notion of "quaobject" employed by Marshall may be more inclined to accept the weaker version of the One Object View presented here. 11 What about the position I am defending? Where does the metaphysical difference between appearances and things in themselves lie according to it? As I suggested above, the difference is between two different sets of properties: the mind-dependent and spatio-temporal properties, on the one hand, and the mind-independent and non-spatiotemporal properties, on the other hand. One Object theorists have proposed different ways to understand these two different sets of properties. Langton (1998) understands 10 Considering a thing in its "constitution in itself" (independently of its relation to our cognition, i.e. as not being the object of sensory intuition) means considering it as a noumenon in the negative sense. Considering it as being the object of a non-sensory intuition means considering it as a noumenon in the positive sense. See CPR B306-307 and the discussion above at the beginning of Sect. 3. 11 Another potential problem for Marshall's reading comes from his claim that "appearance qua-objects are the.same things as things in themselves" in virtue of being "substance-sharing qua-objects" (Marshall, 2013: p. 542). Even admitting that there is a sense in which the notion of substance-which Kant takes to be a category and, therefore, only applicable to appearances-can be meaningfully applied to things in themselves, it is surely not the same sense in which the notion is applied to appearances. Marshall distinguishes two possible ways in which substance can be understood here: either as the bare bearer of properties or as already possessing a set of "primary attributes" (see section III.2 of his paper). The second option does not seem suitable to me, for there is no property shared by appearances and things in themselves. Thus, if they are "substance-sharing qua-objects", the shared substance cannot be the bearer of any property. This leaves us with the alternative option according to which the shared substance is just a bare bearer of properties. As Marshall recognizes, it is "less clear that Kant even has a term for a featureless particularityconferring ingredient of substances" (536). It is even less clear how this notion would apply both and, so to speak, simultaneously to mind-dependent and spatio-temporal appearances, on the one hand, and to mind-independent and non-spatio-temporal things in themselves, on the other hand. What would such a bare substance crossing the phenomenal and noumenal realms be like? the distinction in terms of extrinsic vs. intrinsic properties; Allais (2004) in terms of relational vs. non-relational properties; Rosefeldt (2007) in terms of dispositional vs. categorical properties. Here, I wish to remain neutral on how the two different sets of properties should be characterized, for the general strategy I propose to save the One Object View from the problem of Numerical Identity is compatible with alternative views of the matter. Importantly, all the alternatives I just mentioned agree that the mind-independent and non-spatio-temporal properties we lack cognitive access to ground the mind-dependent and spatio-temporal properties we have cognitive access to. In other words, the mind-dependent properties awareness of which is afforded by sensory intuition are grounded in properties the (possible) capacity for intellectual intuition would afford awareness of. As I identify appearances and things in themselves as epistemic objects constituted by these distinct, non-overlapping sets of properties, it is clear that they are not numerically identical. However, as the perceptual analogy suggests, what I become aware of under the mode of sensory intuition can still be the same thing a (possible) creature equipped with the capacity for intellectual intuition would become aware of. For that creature would become intellectually aware of the set of mind-independent properties grounding the set of mind-dependent properties I am sensory aware of. 12 On the other hand, I agree with Onof (2019) in suggesting that the distinction between appearances and things in themselves should be understood (at least in part) by appeal to two distinctive cognitive "aspects" or "perspectives", the "perspective of an intuitive intellect that would grasp them as they are in themselves … and that of a discursive intellect that thinks and cognizes their relation to a discursive cognition" (211). However, instead of explaining how a view of this sort may help to solve the worries raised by Stang and Walker concerning Numerical Identity, Onof simply dismisses them by stating, first, that "the very characterization of them [Dual Aspects versions of the One Object View] as 'identity' theories already involves an appeal to a view from nowhere that does not exist"; second, that according to his own reading 12 This position comes close to what Jauernig describes as a reading of the distinction between appearances and things in themselves in terms of "genetic sameness" (Jauernig does not endorse it herself, for she defends instead a Two Objects View on reasons different from those I am concerned with in this paper and which I therefore leave aside): "While the appearance and the thing in itself so conceived are numerically distinct, they are genetically the same in that they are derived from the same thing by means of an "as it appears" operation and an "as it is in itself" operation, respectively, and in that they share an ontological ingredient, namely, the bare particular. The "as it appears" operation in this case consists in deleting all mind-independent properties of the original thing to arrive at the appearance; the "as it is in itself" operation consists in deleting all mind-dependent properties of the original thing to arrive at the thing in itself. On views on which the relation between appearances and things in themselves is conceived as genetic sameness, appearances and things in themselves are the same things in the sense that every appearance is genetically the same as a thing in itself or several things in themselves, but there may be no one-to-one mapping between appearances and the things in themselves that ground them" (Jauernig, 2021: pp. 9-10). Of course, in my reading the "as it appears" and "as it is in itself" are epistemic operators tied to distinct modes of intuition affording awareness of two distinct sets of properties. The only thing in which my view differs from what Jauernig describes here is that I do not appeal to the notion of a bare particular as common ontological ingredient between appearances and things in themselves. In my reading, what glues together appearances and things in themselves is simply the fact that the set of properties individuating the latter ground the set of properties individuating the former. (As noted while discussing Marshall's reading in the previous footnote, the idea of a bare particular conceived of as a substance somewhat crossing the phenomenal and the noumenal realm seems hard to make sense of.) "there is an asymmetry between the intelligible and the empirical aspects of things that would not arise were identity at stake" (215). I am not sure I understand these two points. Concerning the first one, Onof seems to be saying that couching the One Object View in terms of Numerical Identity simply means to renounce the very fundamental insight of that view, namely that the distinction between appearances and things in themselves should be understood (at least in part) as an epistemic one (in the terminology he favors, as the distinction between two different "aspects" or "perspectives" of the same things). Concerning the second point, he seems to be saying that if appearances and things in themselves were numerically identical they could not possibly display the distinct sets of properties (sensory and intelligible) Kant ascribes to them. Taken together, these two points seem to suggest that Onof not only rejects as misguided any version of the One Object View couched in terms of Numerical Identity, but also that he takes his own version of that view to be immune from worries concerning Numerical Identity. But even if Onof is right about his reading's immunity, he still owes us an explanation as to how we are to conceive of appearances and things in themselves as being the "same" (in a non-numerically-identical sense) things. So even if I have also appealed, as Onof did, to the distinct epistemic perspectives-in the terminology I have been using, (cognitive) modes of presentation built into Kant's distinction-, my aim in so doing was to answer a question that he left unanswered.
But are we sure that my answer succeeds in laying to rest the worries raised by Stang and Walker? That is the issue I turn to now.

Back to the argument
Let us consider the following argument, which is a tactile pendant to Stang's argument. Imagine a subject who is simultaneously experiencing two distinct tactile objects (as characterized above), to1 and to2. (In Kantian terms, we could say: a subject who is being aware of two spatially distinct tactile appearances). Let us further assume an adapted version of Numerical Identity according to which tactile objects and (physical) things are numerically identical. Accordingly, assume that to1 and to2 are numerically identical to (physical) things-call them thing1 and thing2, respectively. The "tactile" argument would run as follows: (1) to 1 is cold.
(3) to 2 is not cold. (4) thing 2 feels non-cold. (5) It is not the case that thing 2 feels cold. (6) Therefore, thing 1 = thing 2 . This argument does not work. Suppose that, at t, I am feeling a tactile object to 1 with my right hand and that to 1 feels cold. Suppose further that, at t, I am also feeling a tactile object to 2 with my left hand and that to 2 does not feel cold. That does not entail that the (physical) thing I am feeling with my right hand is not the same (physical) thing I am feeling with my left hand. Even if I am presented with two distinct tactile objects-two distinct objects of tactile awareness-, for all I know I may be touching the same (physical) thing. Substitute "hot" for "cold" and you're describing an experience I often make in cold winter days by touching my house's inefficient radiators with both hands.
(Of course, since I also do that with open eyes, I do know that my hands are both touching the same radiator. Imagine, however, a being only equipped with the capacity for tactile experience and with two limbs (or two antennae) to touch around. Imagine that its first limb makes it haptically aware of something cold and that its second limb makes it haptically aware of something non-cold (simultaneously). For all that creature knows, it may be touching just one thing. Of course, Kant thinks our epistemic predicament concerning things in themselves is even more helpless than that of this imaginary creature regarding its imaginary environment. Not to mention the helplessness of my winter dealings with inefficient radiators!) The reason why the argument does not work is that-for the reasons spelled out in the previous section-the relation between tactile objects and (physical) things is not numerical identity. Of course-as we saw in Sect. 2-, Stang's original argument also goes through only if one assumes that the relation between appearances and things in themselves is numerical identity. Hence, Numerical Identity is the assumption that needs to be dropped in both cases. Still, there is a substantive sense in which we can say that-in the example-the tactile object is "the same as" the relevant (physical) thing. This is so because in becoming sensorily aware of the tactile object I am being presented with the relevant (physical) thing under the tactile mode of awareness. The tactile object is just one of the guises under which that thing can be epistemically accessed (by human beings). The same story applies to Kantian appearances and things in themselves: in becoming sensorily aware of an appearance, I am being presented with a certain thing under the mode of sensory intuition. A (possible) creature capable of intellectual intuition would become aware of that very same thing's "internal constitution"-it would become aware of that very same thing in itself . 13 13 One could insist that, in the tactile example, there really is not one (physical) thing, but two (physical) parts connected "radiatorwise"; and that, more generally, talk of "dry middle-sized goods", to use Austin's phrase, should be treated as a mere convention: what there really is at the physical level are (physical) parts, wholes being fictional or, at least, metaphysically parasitic on them. Consequently, so the objection would go, the "tactile" argument is sound: for it is the case that each tactile object is numerically identical to a (physical) thing, namely the relevant part of the radiator. I have two answers to this objection. First, if we assume that "dry middle-sized goods" do not really exist at the physical level, the same would also apply to "parts connected radiatorwise", for they would also not figure in any physical description. Nor, more generally, would we find in such a description any item that would be numerically identical to the tactile objects of the example (the relevant items would be, roughly, the two regions of the radiator surface touched by my left and right hand, respectively). Instead, we would find something like two sets of particles with different motion. This means that the radiator case would be similar to that of Sellars' pink ice cube: each tactile object with a certain manifest property (a certain felt temperature) corresponds to many particles with a certain physical property (motion). Thus, there would be no numerical identity between tactile objects and items identified at the physical level. Second, there is at least a cognitive mode of presentation familiar to us humans under which the environment is non-conventionally experienced as made up of "dry middlesized goods", namely visual experience. For the visual system is hard-wired to group visual properties together into "bounded, connected, cohesively moving three-dimensional constructions from surfaces" (see O'Callaghan 2008: p. 911; see also Scholl 2001 for a review about the cognitive processes underlying object perception in vision). Hence, if we draw the comparison between tactile objects and visual objects, many things we would experience haptically as distinct objects would be experienced visually as just one object. (I will come back to this kind of example later in the paper.) Again, numerical identity across distinct But are we sure that this picture is still a version of the One Object View? Stang argues that the maneuvers available to the One Object theorist once he abandons Numerical Identity are not satisfactory, because the resulting view ends up being only nominally different from the Two Objects View. Perhaps, the same happens also with my proposal: it offers only an apparent solution.
I do not think that is the case. To show this, let us resume the perceptual analogy. Consider again Sellars' pink ice cube. One could say that the manifest cube of visual experience is a purely mental object-a sense datum or a collection of sensations, saycompletely distinct from the scientific cube of microphysics. This would amount to a Two Objects View of perceptual experience, a typical version of which would be Indirect Realism. But nothing of what I have said so far about perception forces one to accept this picture. In fact, it is equally compatible with a One Object View of perceptual experience. Accordingly, seeing the manifest cube is being visually presented with the very same thing that physics describes as a system of particles. Similarly-and to repeat a point already made-, sensorily intuiting an appearance is being presented with the very same thing that certain (possible) creatures could intuit intellectually in itself. I think this is a robust enough notion of sameness to distinguish the account on offer here from the Two Objects View. If I am right, the One Object View can be saved from Stang's argument. 14 Let us consider another point. Walker affirms that "ascription of moral responsibility", although it provides "no reason to suppose an identity between a coffee-cup and an individual thing in itself", does demand "that there be an identity, or something very like an identity, between a human being and a single noumenal agent" (Walker, 2010: p. 836). 15 However, he goes on, reflection on the case of moral agency reveals that the relation obtaining between a noumenal agent and an (empirical) human being is "not an identity between an appearance and a thing in itself", but rather the "identity of a composite whole combining the intimately related pair" (837). I do not dispute, of course, Walker claims that Numerical Identity between noumenal agent Footnote 13 continued cognitive modes of presentations would not work. (Many thanks to two anonymous referees for raising this objection.) 14 According to Jauernig's taxonomy of readings of Kant's distinction, the one on offer here does not count as a One Object View in the standard sense of the label in virtue of rejecting Numerical Identity. Rather, it corresponds to what she calls a "same-things" view (see Jauernig 2021: pp. 10-11). This is, however, a merely terminological issue, for Jauernig agrees that "same-things" views qualify as "one world"/"two aspects" views for classifying "appearances and some things in themselves as the same things in some reasonable sense" (11; see also p. 4), which is precisely what my reading aims at. 15 Here Walker recognizes that the case of moral agency seems to require that we assume that the (empirical) human being and the noumenal agent are, in some sense, the same-i.e. it seems to entail some version of the One Object View in that specific case. What Walker points out, however, is that we have no reason to generalize from that specific case to all appearances (such as a coffee-cup). Thus, we have no reason to ascribe to Kant a generalized One Object View. Walker is right: but once we have a version of the One Object View that avoids Numerical Identity, why not take it to be Kant's general view of the distinction between appearances and things in themselves? Another issue the case of moral agency might raise concerns whether there really is no property shared by appearances and things in themselves. Cannot we say, for instance, that the (empirical) human being and the noumenal agent share the property of "being involved with free effects"? I would resist this conclusion: as empirical objects we are completely determined by causal laws and, therefore, do not contribute to the production of any free effect. Similarly, an empirical event is completely determined by causal laws and, therefore, does not count as a free action. (Many thanks to an anonymous referee for raising this issue.) and empirical human does not work. In fact, I am sympathetic with the mereological picture he puts forward. For my own proposal goes in the same direction. I suggest that, according to Kant, things instantiate two non-overlapping sets of properties, a set of mind-dependent and spatio-temporal properties and a set of mind-independent and non-spatio-temporal properties. Sensory intuition makes us aware of the former one, whereas intellectual intuition makes (possible) creatures possessing it aware of the latter one. But both forms of intuition afford awareness of the same things. This notion of sameness is all the One Object View requires as well as substantive enough to vindicate it.

Particular Noumenal Ignorance again
There is a last problem I would like to address. In his (2018) paper, Stratmann mounts a general argument targeting both One Object and Two Objects View. Moreover, the version of One Object View he considers is not Numerical Identity. Rather, it is the kind of weaker version I have been defending in the previous sections. So what is the trouble he thinks both camps are afflicted by?
The underlying issue is again that of Particular Noumenal Ignorance. As Stratmann (2018: p. 3) puts it, given this Kantian claim: we can know that there are things in themselves grounding appearances, but not which specific kind(s) of one-or two-object grounding relation(s) obtain(s) between them. Our ignorance of things in themselves therefore extends to their distinctness from appearances-pace both metaphysical one-object interpretations and metaphysical two-object interpretations. 16 If Stratmann is right, the entire debate between One Object and Two Object theorists is simply misguided, for there is no answer to the question they disagree about. What can we say here? I am not sure whether Stratmann's argument is sound. But even if it is, it does not affect the version of the One Object View I have been putting forward here. The perceptual analogy helps us again to see why.
Imagine again that you are having-both at t and with your eyes closed-a tactile experience with your right hand and another tactile experience with your left hand. This means you are aware of two tactile objects. Were you now to open your eyes, you would start to experience visually what you are already experiencing haptically. Does this mean that you would also experience two visual objects? Not necessarily. For how do you know that what is haptically presented to you as two objects would also be presented to you visually as two objects? Perhaps your right hand and your left hand are both placed on the same table (or on the same radiator, as in the previous version of this kind of example). In that case, what is now being presented to you haptically as two distinct tactile objects would be presented to you visually, once you open your eyes, as just one visual object.
My version of the One Object View works analogously. It simply says that being aware of an appearance means sensorily intuiting something that could also be intuited intellectually (not by me and you, of course, but by possible beings suitably equipped with the relevant cognitive capacity). This, however, allows that the noumenal ground of what we sensorily identify as one appearance be many things in themselves, just as the scientific ground of the one manifest pink ice cube may very well be many particles. As in the Sellarsian scenario we can say that the same thing is manifestly presented as one pink-ice cube and scientifically presented as many particles, in the Kantian scenario we can say that the same thing can be sensorily intuited as one appearance and (possibly) intellectually intuited as many things in themselves. 17,18 More generally, the weak version of the One Object View I have been defending allows that something that is sensorily intuited as n appearances be (possibly) intellectually intuited as being m things in themselves, where n and m can be either different or identical numbers. Therefore, it is immune to the kind of worry raised by Stratmann. 19

Conclusion
Numerical Identity clashes with Particular Noumenal Ignorance and should therefore be abandoned. This is no reason for Kant scholars sympathetic with the One Object View to despair, for they can happily embrace the weaker version of that view sketched above. According to Jauernig's taxonomy (see fn. 14 above), my weaker version of the One Object View would count as a "same things" view: there is only one set of things that can be considered either in relation to our cognition, i.e. as appearances (objects of sensory intuition); or independently from their relation to our cognition, i.e. as things in themselves ((possible) objects of intellectual intuition). This is the reading One Object theorists should embrace, for it not only survives the argument put forward by Stang and Walker against Numerical Identity without giving up the full spirit of the One Object View, but also avoids the more general worries raised by Stratmann about the very intelligibility of the dispute between One Object and Two Objects theorists. Consequently, the proposal defended here succeeds in fulfilling the desideratum-nicely described by Ameriks (2003: p. 83)-according to which a correct interpretation of Kant's transcendental idealism needs to involve "giving up