1

The Unity of Kant’s Critical Theory of Real Possibility1

(to appear in The Sensible and Intelligible Worlds: New Essays on Kant’s Metaphysics and Epistemology. Ed. Karl Schafer and Nicholas Stang (Oxford University Press, forthcoming))

Nicholas Stang University of Miami

nick.stang@gmail.com

The Pölitz transcripts of Kant’s lectures from the 1780s 2 on rational theology contain a

fascinating discussion of modality:

On this point rests the only possible ground of proof for my demonstration of God's existence, which was discussed in detail in an essay I published some years ago. Here it was shown that of all possible proofs, the one that affords us the most satisfaction is the argument that if we remove an original being, we at the same time remove the substratum of the possibility of all things. - But even this proof is not apodictically certain; for it cannot establish the objective necessity of an original being, but establishes only the subjective necessity of assuming [annehmen] such a being. But this proof can in no way be refuted, because it has its ground in the nature of human reason. For my reason makes it absolutely necessary for me to assume a being which is the ground of everything possible, because otherwise I would be unable to know what in general the possibility of something consists in [worin etwas möglich sey]. (Pölitz RT, Ak. 28:1034)3 The ‘essay published some years ago’ to which Kant here refers is his 1763 work The only

possible ground of proof in support of a demonstration of the existence of God (henceforth, OPG). While

1 This paper is a condensed treatment of the material in chapters five and eight of my book, Kant’s Modal Metaphysics (KMM). It was presented, under a different title, at the Kant and Modality conference at Humboldt Universität zu Berlin in July 2012 and at the Eastern APA in December 2012. I want to thank Tobias Rosefeldt, Toni Kannisto, Jessica Leech, Andrew Chignell, Sam Newlands, and Colin McLear for helpful comments. I also want to thank Robert Hanna, my respondent at the APA, whose remarks were quite enlightening. 2 Kant lectured on rational theology during the winter semesters 1783-4 and 1785-6, and it is unclear which set of lectures is the basis of the Pölitz transcripts. See Kant (1996), 337-338 for more. 3 In the rational theology lectures, see also Volckmann RT, Ak. 28:1176; and Danziger RT, Ak. 28:1259. The translation of the Pölitz text is from Kant (1996), with minor modifications.

2

the exact interpretation of that work is controversial4, it is clear that he there defends an

inter-connected set of claims about possibilities and their grounds:

(1) It is logically possible that p iff p does not entail a contradiction. This is the definition of logical possibility. (Ak. 2:77)

(2) Not all logical possibilities are ‘real’ possibilities (Ak. 2:77, 85). Kant does not

explicitly offer a developed theory of what ‘real’ possibility is, but it is clear that ‘real’ possibility is more metaphysically ‘robust’ than mere logical possibility. He also lays down certain conditions on it.5 For instance,

(3) If it is really possible that p then some existing being grounds the fact that it is

really possible that p (Ak. 2:78, 79). As with (2), Kant does not explicitly offer a developed theory of the relation between real possibilities and their grounds.

(4) All real possibilities must be grounded in a ‘first’ or ‘highest’ ground of real

possibility (Ak. 2:85-87). This first ground of real possibility is said to exist with ‘absolute’ real necessity, which Kant defines to mean: if this being did not exist, nothing would be really possible. (Ak. 2:83).

(5) Since there are real possibilities, there is a first ground of real possibility (Ak.

2:157). Kant argues that this being has the traditional divine attributes (simplicity, omniscience, omnibenevolence, etc.) and identifies it with God.

Of course, Kant’s actual theory is much more complex than that, but for now this brief

outline will suffice.6

In this passage from the 1780s Kant appears to be saying that the 1763 argument does

not constitute a proof (or even a ‘ground of proof’)7 of the existence of God or of a first

ground of real possibility. I take this to mean that by following the steps of the 1763

4 Cf. Adams (1994), Watkins and Fisher (1998), Chignell (2009) and (2012), Stang (2010), Yong (forthcoming), and Abaci (forthcoming). 5 One of the most important elements of Kant’s modal theory in OPG, the principle that real possibility requires a “material element,” is not as important for understanding the Pölitz passage, so I omit it here. Cf. Ak. 2:77, 79-81. 6 I examine all of these claims in more detail in Stang (2010) and in KMM, ch. 4. 7 See Chignell (2009), 160-162 for a discussion of the distinction between ‘Beweis’ and ‘Beweisgrund.’

3

argument and concluding on that basis that there is a first ground of real possibility we do

not thereby come to know that there is such a being. Nonetheless, it is rationally necessary

for us to hypothesize the existence of such a being that grounds all real possibility and to

hypothesize it for the very reasons given in the 1763 argument. Whereas in 1763 Kant

argued that the fact that something is really possible entails that there exists a ground of all

real possibility, he now appears to think that this fact rationally requires us to seek an

explanation of it, and the only rationally satisfying explanation is that there is a ground of all

real possibilities, so we are rationally required to assume – I want to remain neutral on

exactly what epistemic attitude we are supposed to adopt so I am going to use the non-

technical term ‘assume’8 – that there is such a being.9

This raises a number of questions, including, what is the nature of this rational necessity,

the necessity that we assume an original being? And what epistemic status does this resulting

assumption have, if it is not known to be true? However, these questions will not be my

focus in this paper; I address them elsewhere. 10 I want to focus on the necessity of the

being that is assumed, which Kant goes on to characterize:

For in addition to the logical concept of the necessity of a thing (where something is said to be absolutely necessary if its nonexistence would be a contradiction, and consequently impossible), we have yet another rational concept of real necessity. This is where a thing is eo ipso necessary if its nonexistence would cancel all possibility. Of course in the logical sense possibility always precedes actuality, and here I can think the possibility of a thing without actuality. Yet we have no concept of real possibility except through existence [Existenz], and 8 I will also sometimes speak of ‘hypothesizing’ this being, or the hypothesis that there is such a being. I mean this to be synonymous with assuming that there is such a being. 9 In Kant’s technical terminology, I am remaning neutral on which mode of Fürwahrhalten we have towards this proposition: opinion (Meinung), belief (Glaube), or knowledge (Wissen). I am in broad agreement with the interpretation of Chignell (2007a) and (2007b), and Hanna (2006), 258-361. In particular, I agree that this is a case of what Chignell calls ‘theoretical’ belief (2007b, 347-354); I argue for something very similar in KMM, ch. 8. 10 Chapter eight, sections three and four, of KMM.

4

in the case of every possibility that we think realiter we always presuppose [setzen . . . voraus] some existence [Daseyn]; if not the actuality of the thing itself, then at least an actuality in general which contains the data for everything possible. Hence all possibilities presuppose [voraussetzen] something actually given, since if everything were merely possible, then the possible itself would have no ground; so this ground of possibility must itself be given not merely as possible but also as actual. But it must be noted that only the subjective necessity of such a being is thereby established, i.e. that our speculative reason sees itself necessitated to presuppose [vorauszusetzen] this being if it wants to have insight into why something is possible, but the objective necessity of such a thing can by no means be demonstrated in this matter. (Pölitz RT, Ak. 28:1036)11

Kant attributes necessity not just to the hypothesis of an ‘original being,’ but to the content

of that hypothesis: it is rationally necessary that we hypothesize an absolutely necessary

being. He says we have a logical concept of necessary existence: a being exists necessarily if

and only if the proposition that it does not exist entails a contradiction. This conception of

necessary existence was the basis of the ontological argument in Leibniz, Wolff, and

Baumgarten. In OPG and, more famously, in the Critique of Pure Reason (henceforth, CPR)

Kant rejects this conception of necessary existence because it presupposes (incorrectly,

according to him) that existence is a determination or ‘real predicate.’12

I want to focus on what Kant calls the ‘other’ concept of absolutely necessary existence:

a being exists absolutely necessary just in case that being exists, and were it not to exist,

nothing would be really possible. This is the very same definition of absolutely necessary

existence Kant gave in 1763 in OPG (see (4) above).13 He also introduces a concept of ‘real’

possibility and a principle governing real possibility — if it is really possible that p then the

fact that it is really possible that p is grounded in something that actually exists — that are 11 This crucial text is also cited and discussed in Adams (1994), 182, note 9 on that page; and Chignell (2007b), 349. 12 A599/B627. This objection is already made in OPG, Ak. 2:72-3. I examine this famous Kantian doctrine in detail in Stang (forthcoming-a) and in KMM, ch. 2. 13 Ak. 2:81-3. See Stang (2010) for a discussion of this interpretation of absolutely necessary existence.

5

familiar from that earlier work (see (3) above).14

Readers of CPR know that Kant frequently distinguishes between logical and real

possibility and claims that we cannot always infer from logical possibility to real possibility.

Recently, some scholars have argued that this distinction is crucial to understanding Kant’s

Critical epistemology and his restriction of knowledge to the empirical world.15 However, it

is often unclear what the Critical Kant means by ‘real’ possibility and whether, in fact, he

means one thing, or whether he has a number of distinct notions of possibility which he

simply groups under the heading ‘real’ to distinguish them from logical possibility.

In this paper I am going to argue that in the Critical period Kant does have a unified

theory of real possibility and that this theory has the same general structure as his pre-Critical

theory of real possibility. My stalking-horse will be the Pölitz passages quoted above, and, in

particular, a set of questions about the content of the hypothesis of an absolutely necessary

being:

(Q1) What is the necessity in the content of the assumption, the necessity of the being that is thereby assumed? What is this concept of necessity that Kant claims we have?

(Q2) Similarly, what is this concept of real possibility of which he says, “in the case of

every possibility which we think realiter we always presuppose some existence”? In the first section I consider what it could mean, within Kant’s Critical system, to apply

modal concepts like <necessary>16 to a hypothesized being that cannot be an object of

experience; in section two I begin answering the second question by arguing that the concept 14 Ak. 2:77-78. See the secondary literature cited above for discussions of these ideas in the context of OPG. 15 E.g. Chignell (2010) and Stang (2011). I discuss the role that modality plays in Kant’s epistemology, and criticize Chignell’s interpretation, in detail in KMM, ch. 5. 16 I follow Anderson (2004) and use expressions flanked by angle brackets to denote the corresponding concept. E.g. ‘<necessity>’ denotes the concept corresponding to the English (German) expression ‘necessity’ (Notwendigkeit).

6

of real possibility of which Kant speaks in the Pölitz lectures is the modal category of

possibility and that this is a generalized version of the pre-Critical concept of real possibility.

It is this concept that unifies the different kinds of real possibility Kant accepts in the

Critical system. In the third section I answer the first question by arguing that for each kind

of Critical real possibility there is a corresponding kind of absolute necessity and identifying

which kind of absolute necessity is being spoken of in the Pölitz lectures.

1. Schematized and unschematized categories

One important feature of the Pölitz passages quoted earlier is that these modal concepts —

absolutely necessary existence and real possibility — are being applied beyond the range of

experience. We are rationally necessitated to hypothesize an absolutely necessary being, but

no object of experience is absolutely necessary.17 Without opening the Pandora’s box of

competing interpretations of Kant’s transcendental idealism, I would like to say a few things

about what kind of object we are hypothesizing.18 Kant’s term for the objects of experience

is ‘phenomena’ (in some contexts, appearance [Erscheinung])19, so the hypothesized absolutely

17 A559/B587, A561/B589 18 I explore various such interpretations and the textual and philosophical bases for them in depth in Stang (forthcoming-c). My own interpretation is a variant of the traditional metaphysical “two objects” or “phenomenalist” reading. (For the now standard distinction between ‘one object’ and ‘two object’ readings see Ameriks (1982)). However, I am not presuppoing that reading here. While my formulations are most naturally read on ‘two object’ lights, they can also be recast in ‘one object’ friendly ways (e.g. Allais (2004)). Where I talk about phenomena and noumena as distinct kinds of objects, the ‘one object reader can interpret this to mean: objects considered as they appear to us (phenoemna) and those very objects considered as they are in themselves (noumena). At various points in the paper, I will indicate in footnotes how to interpret my formulations in ‘one object’ friendly ways. 19 In the Transcendental Aesthetic, Kant defines appearance as “the undetermined object of an empirical intuition” (A34/B20) and later he defines phenomena as “appearances to the

7

necessary being is not a phenomenon. Our form of intuition is sensible: we can only intuit

existing objects in virtue of the causal influence of objects on our sense organs. Any object

of our sensory intuition is a phenomenon (or appearance)20, so the hypothesized necessary

being cannot be an object of sensory intuition. Kant’s calls such objects ‘noumena in the

negative sense,’ which means “a thing insofar as it is not an object of our sensible intuition”

(B307).21 I will simply call them ‘noumena’ for short. Noumena are ‘supersensible’

[übersinnliche] (a term Kant sometimes uses) because they are not the right kind of thing for us

to sense or experience; because they are not in space and time, we could not sense them,

even if our sense organs were vastly finer, and we cannot infer their existence from any

empirical causal laws.22 In hypothesizing the existence of an absolutely necessary being that

extent that as objects they are thought in accordance with the unity of the categories are called phenomena” (A249). I take this to mean that appearance is the genus of which phenomenon is the species. All objects of empirical intuition are appearances, but only those that are “thought in accordance with the unity of the categories” are phenomena. For instance, if I have a visual after-image or highly disunified visual hallucination, that perception may not represent its object as standing in cause-effect relations, or being an alteration in an absolutely permanent substance. These would be appearances but not phenomena. 20 See previous note. 21 Some readers will wonder whether any supersensible being could be an object for us. By ‘object’ here I mean only ‘object of thought.’ I am not claiming that such beings could be objects in a more substantive sense, e.g. objects of cognition or of experience. See A105, A108, B137. 22 See A226/B273, and Ak. 8:205. In this paper, I remain neutral on the relation between ‘noumena in the negative sense’ and what Kant calls ‘things in themselves’ [Dinge an sich (selbst)] and, indeed, on whether ‘things in themselves’ functions referentially as a name for a kind of object, or adverbially for a way of considering objects (as they are in themselves), as, Prauss (1974) famously argued. For relevant texts see A254/B310, A256/B312, A259/B315; Prolegomena §30 (Ak. 4:392), §32 (314), §33 ( 315); §59 (360); On a Discovery (Ak. 8:208); and the Progress essay (Ak. 20:292, 308). In Stang (forthcoming-c) I argue that ‘things in themselves’ and ‘noumena in the negative sense’ are co-extensive terms with different meanings.

8

grounds all real possibility, we are using modal concepts to think about a noumenon.23

Although Kant denies that we cognize noumena through the categories, he maintains

that we not only may think of them using the categories but that we must do so. In fact, we

cannot think about any objects without thinking of them under the categories.24 So it should

not be in itself surprising that Kant asks us to think of a supersensible object under the

modal categories of <necessity>, <possibility>, and <existence>.

My question in this section is: what are we doing when we apply a modal category to a

noumenon? This is an instance of a larger question in Kant’s philosophy: what are we doing

when we apply any category whatsoever to noumena? In other words: what is the content of

a category when it is applied to noumena? 25

23 Some readers might object that the necessary being hypothesized at Ak. 28:1034-6 is not supersensible; this being is merely the ‘totality of experience’ that grounds all possibility according to Kant at A581-2/B609-10. However, that claim occurs in the context of Kant’s discussion of the idea that God is necessary as the ground of the complete determination of all objects. The OPG argument does not concern complete determination, so Kant’s discussion in the Pölitz lectures, which refers back to that text, is not about the idea of complete determination. See Stang (2012) for a discussion of the complete determination principle in CPR. 24 A253, Ak. 5:54, 103. When Kant writes at B148 that, when applied beyond possible experience, categories are “empty concepts of objects, through which we cannot even judge whether the latter are possible or not — mere forms of thought without objective reality” I take him to mean that categories, as applied to super-sensible objects, do not allow us to cognize anything; I do not take him to mean that they are literally senseless or meaningless. It is implausible to read him as claiming that they are literally senseless or meaningless because Kant himself frequently subsumes non-empirical objects under the categories. For instance, Kant repeatedly claims that noumena are among the causes of the sensory matter of experience. See especially A190/B235, A387, A494/B522, Ak. 4:289, 4:314, 4:318, 4:451 and 8:215. 25 This question does not presuppose that such a non-sensible object exists, for it only appears within an intensional context: my thought of such an object using a modal concept. Thus, I am not falling afoul of Kant’s dictum that “to this extent the categories extend further than sensible intuition, since they think objects in general without seeing to the particular manner (of sensibility) in which they might be given. But they do not thereby determine a greater sphere of objects, since one cannot assume that such objects can be given without presupposing that another kind of intuition than the sensible kind is possible,

9

Although we can (and must) think of all objects under the categories, there is an

important difference in the epistemic status of these thoughts for objects we can experience

(phenomena) and those we cannot experience (noumena).26 Kant claims that the categories

have ‘objective validity’ when applied to phenomena, but lack it when applied to noumena.27

We cannot cognize a domain of objects – we cannot know much of anything positive about

them28 – using a set of concepts unless those concepts have objective validity for those

objects.29 So, since we cannot think about any objects without using the categories30, if the

categories lack objective validity for some domain of objects, then we cannot cognize those

which, however, we are by no means justified in doing.” (A254 – my underlining) I am not assuming that there are supersensible objects, but only that we can (and must!) have thoughts whose contents represent such objects. 26 On the ‘one object’ interpretation, empirical and non-empirical conceptioins of objects. 27 A93/B126, A111, A128, A239/B298. 28 Notoriously, Kant claims that things in themselves (which are noumena) are non-spatiotemporal and that they appear to us by causally affecting us. Hence the qualificaiton ‘much of anything positive’ in the main text. How these two doctrines -- non-spatiotemporality, and noumenal affection -- are compatible with Kant’s denial that we can know noumena through theoretical means lies far outside the scope of this paper. 29 See Bxxvi, A90/B122, A239/B298. 30 At B110 Kant distinguishes between the mathematical categories (of quantity and quality) and the dynamical categories (of relation and modality), a distinction to which he returns at A160/B199. The distinction seems to be that in thinking about objects of mere intuition (mere appearances in the terminology developed above, e.g. objects constructed in pure intuition) we must use the mathematical categories but not the dynamical categories, but we must represent phenomena (objects of experience) using both sets of categories. So it is not the case that we must use all moments of the table of categories in thinking about every object. However, Kant’s reasons for thinking that we must use dynamical categories in thinking about phenomena is those categories “concern the existence of the objects of a possible empirical intuition” (A160/B199). I take this to mean that we must use dynamical categories in thinking of phenomena because phenomena exist indepndently of any particular intuition of them (unlike objects constructed in pure intuition). However, in thinking about noumena we are also thinking about objects that exist independently of our thinking about them, so we must also use dynamical categories. So the point in the main text holds.

10

objects. 31 We cannot know anything about them, aside from logical truths and perhaps

some negative facts about them (e.g. that they are not in space and time). Categories as

applied to empirical objects have objective validity in part because they can be schematized, i.e.

provided with a priori rules for application to objects given in experience. I will henceforth

refer to them as ‘schematized categories.’

Categories applied to empirical objects can be schematized, the categories in general

cannot, and because of this, categories have objective validity for empirical objects, while

categories lack objective validity for objects überhaupt; consequently, in applying categories to

empirical objects we cognize those objects, but in applying categories to supersensible

objects (noumena) we do not cognize anything about those objects. Sometimes, rather than

categories as concepts of objects in general, I will just talk about ‘unschematized’ categories.

The question from earlier can be rephrased: what are we doing when we think of an object

under the unschematized categories <necessary> or <possible>? What is the content of

such thoughts?

Before going any further in trying to answer that question, I want to investigate further

what is lacking or missing or defective about unschematized categories. The section of the

CPR in which Kant goes into the most detail about the distinction between categories as

concepts of empirical objects (phenomena) and as concepts of objects in general is ‘On the

ground of the distinction of all objects in general into phenomena and noumena.’ It

contains the following interesting discussion of the difference between schematized and

unschematized categories:

31 ‘Cognition’ translates Erkenntnis, a technical term in Kant’s epistemology. Whether cognition/Erkenntnis is equivalent to knowledge/Wissen is a matter of scholarly debate. I am convinced by the arguments of Schafer (forthcoming) that Kantian cognition is not knowledge, but in this paper I suppress that complication. Accordingly, I will treat cognition and knowledge interchangeably here.

11

For every concept there is requisite, first, the logical form of a concept (of thinking) in general, and then, second, the possibility of giving it an object to which it is to be related. Without this latter it has no sense, and is entirely empty of content. (A239/B298)

This is the familiar Kantian point that in order to cognize objects using some concept we

must first be able to prove the real possibility that there are objects falling under that

concept32. The logical possibility of a concept being instantiated consists in that concept

being logically consistent (not containing any marks that are mutually self-contradictory), but

logical possibility does not suffice for real possibility (whatever that is, exactly). Kant here is

claiming that the reason unschematized categories do not have objective validity for

noumena is that we cannot prove the real possibility of noumena falling under them; we can

only prove the real possibility of phenomena falling under them, so only empirical

(schematized) categories have objective validity.

However, Kant goes on in the next paragraph to suggest that unschematized

categories — the categories as applied to all objects in general, empirical as well as non-

empirical objects — are defective for cognitive use by us in a more radical sense:

That this is also the case with all categories [that we cannot use them to cognize anything about objects in general] however, and the principles spun out from them, is also obvious from this: That we cannot even give a real definition of a single one of them, i.e. make intelligible the concept of their object, without immediately descending to conditions of sensibility, thus to the form of the appearances, to which, as their sole objects, they must consequently be limited, since, if one removes this condition, all significance, i.e., relation to the object, disappears, and one cannot grasp through an example what sort of thing was really intended by concepts of that sort. (A241/B300 – underlined material added in B edition)

Kant here identifies the lack of objective validity of the categories in general with our

inability to give real definitions of them for all objects in general (phenomena and noumena).

He identifies a real definition of a category as making ‘intelligible’ [verständlich] to ourselves

how it is really possible for objects to fall under them; this is most naturally read as meaning 32 Cf. the footnote to Bxxvi.

12

that a real definition of a concept reveals not merely that it is really possible for objects to fall

under it, but reveals or makes comprehensible to us why it is really possible. In his own copy

of the A edition, Kant wrote in the margin of this passage: “we cannot explain their [the

categories’] possibility.” Giving a real definition of a concept is proving that an object of that

concept is really possible by eliciting the grounds of that real possibility and thereby explaining

that real possibility. When we really define a concept we come to know what it is in virtue of

which objects of that concept are really possible.

This connection between real definitions and knowledge of the grounds of possibility is

further substantiated in Kant’s lectures on logic. Following standard logical practice of the

day, he distinguishes between nominal and real definitions. 33 However, Kant’s distinction

between nominal and real definitions is slightly different than Leibniz’s or Wolff’s. A

nominal definition of a concept reveals what we ‘think’ in the concept, the marks we have

assembled together to form that concept, what Kant sometimes calls the logical essence of

the concept. Kant thinks that nominal definitions are relatively trivial for empirical scientific

purposes. It may be important to make sure we all mean the same thing by our words – that

we are all using ‘water’ to denote the clear, drinkable liquid in rivers and lakes – but this does

not by itself tell us anything substantive about water, the stuff our concept refers to. 34 Water

itself, rather than the concept <water>, is the object of a real definition. This means that

when Kant says that real definitions are definitions of things [Sacherklärungen] 35 he does not

33 The principal discussions of real and nominal definition, and the associated notions of real and logical essence, are in the Jäsche Logic (Ak. 9:61, 140-145) and the transcripts of Kant’s logic lectures (Ak. 24): Blomberg Logic (113-118, 268-273), Vienna Logic (838-840, 913-925), Dohna-Wundlacken Logic (727-729,756-760), Philippi Logic (408-9, 456-9), Busolt Logic (634-5, 656-60), and Pölitz Logic (535, 573-5). 34 Ak. 24:116, 271, 757, 839, 918. Cf. A728/B756. 35 Jäsche Logic §106 (Ak. 9:143); cf. Ak. 24:839, 757.

13

mean to restrict real definitions to individual things, which Kant thinks are never susceptible to

real definition36; real definitions, if they are possible at all, would be of the generic kinds of

individual objects that fall under our concepts. A real definition reveals the essential inner

marks of the object that make it possible and that explain its other properties. For instance,

a real definition of water (if we could give one, which Kant denies) would reveal the inner

properties of water that are essential to it and which explain its manifest properties (clarity,

drinkability, etc.). In doing so, a real definition would reveal not only that water is really

possible, but why. It would reveal the grounds of the real possibility of water. Kant thinks our

ability to give real definitions is quite limited. We can really define concepts we arbitrarily

make of empirical objects, because in the case of such artificial concepts, to be an instance of

the concept just is to have the marks that constitute our concept. Likewise, we can give real

definitions of most mathematical concepts, because mathematical concepts like <triangle>

or <polynomial> are arbitrarily generated by us: the nominal definition that combines a set

of marks together (e.g. <three-sided> and <plane figure>) to form the concept also tells us

what it is to be an instance of that concept and, provided we follow correct mathematical practice

and construct an instance of the concept in pure intuition prior to using the concept in our

demonstrations, then we know a priori both that the concept is possibly instantiated and why.

However, he is quite skeptical about our ability to give real definitions of ‘given’

concepts, either a posteriori given concepts of natural kinds like water, or a priori concepts

given by the nature of our understanding itself, the categories. Kant allows that for ‘given’

concepts, both empirical and a priori, we may be able to analyze the concept to determine a

few marks contained in it, although in general we will not be able to fully analyze the concept

36 Ak. 24:118, 268.

14

and thus give a complete nominal definition.

This is evidence that, in CPR, by denying that we can give real definitions of the

categories for all objects whatsoever, Kant is denying that we can know the grounds of the

real possibility of objects in general falling under the categories. In the A edition, Kant

expands upon the lack of a real definition of the categories as concepts of objects in general:

Above, in the presentation of the table of the categories, we spared ourselves the definitions of each of them, on the ground that our aim, which pertains solely to their synthetic use, does not make that necessary, and one must not make oneself responsible for unnecessary undertakings that one can spare oneself. [. . .] But now it turns out that the ground of this precaution lies even deeper, namely, that we could not define them even if we wanted to [. . .]. (A242)

The reason Kant did not define the categories earlier in CPR is not, as he earlier claimed,

that putative definitions “would distract us from the chief point of the investigation by

arousing doubts and objections that can well be referred to another occasion” (A83/B109),

but because real definitions of the categories in general are impossible for us; we can only give

real definitions of categories restricted to empirical objects, if at all.

Kant admits that we can give definitions of a sort of the categories for all objects überhaupt,

both phenomena and noumena, and gives a series of examples of such pseudo-definitions:

If I leave out persistence (which is existence at all times), then nothing is left in my concept of substance except the logical representation of the subject, which I try to realize by representing to myself something that can occur solely as subject (without being a predicate of anything). (A242/B300) From the concept of a cause as a pure category (if I leave out the time in which something follows something else in accordance with a rule), I will not find out anything more than that it is something that allows an inference to the existence of something else. (A243/B301) No one has ever been able to define possibility, existence, and necessity except through obvious tautologies if he wanted to draw their definition solely from the pure understanding. For the deception of substituting the logical possibility of a concept (since it does not contradict itself) for the transcendental possibility of things (where an object corresponds to the concept) can deceive and satisfy only the

15

inexperienced. (A244/B302)37

The ‘obvious tautologies’ to which Kant alludes in the third passage are principles like: the

necessary is that whose non-being is not possible.38 The third passage also tells us what is

incomplete in these putative definitions—it is not that they are false, but they do not show

that it is really possible for an object to fall under the concept, much less why. At most they

show that the concept is logically consistent. Kant does not deny, for instance, that a cause

in general is something from which the existence of something else follows, but this principle

does not show us what makes it really possible for there to be a cause. Similarly, it is surely

the case that a necessary being is one whose non-existence is not possible. But this does not

tell us what would make such a being necessary. Given that Kant has just invoked the

notion of a real definition—and thus implicitly contrasted the real definition of the

categories with mere nominal definitions—it is natural to think of these as partial nominal

definitions. They unpack conceptual connections between categories (e.g. the inter-

definability of the necessary, the possible, and the contingent) but they make no progress in

making intelligible to us the real possibility that objects fall under these concepts.

2. Real Possibility and the Categories

I have argued that by not being able to give real definitions of the categories as concepts of

objects in general we are not able to identify what in general grounds the possibility that

37 Cf. A459/B487 and, especially, A593/B621, where Kant explicitly describes as a nominal definition the principle that a necessary being is one whose non-existence is not possible but criticizes this for not making ‘verständlich’ to me “whether through the concept of an unconditionally necessary being I am still thinking something or perhaps nothing at all.” 38 This standard inter-definition of possibility and necessity is repeatedly asserted by Kant in his metaphysics lectures. See Ak. 28: 418, 498, 556, 557, and 633.

16

objects fall under these categories, e.g. we are not able to answer perfectly generally what

makes it possible that something is a cause, or a substance. Now, this raises an important

question: what is the relevant notion of possibility here? In the passage quoted above

(A244/B302) Kant makes it clear that knowing the logical possibility of a concept’s being

instantiated—knowing that no contradiction follows from the hypothesis that the concept

has an instance—does not suffice to give a real definition of the concept, so clearly logical

possibility is not the answer here. Later in the same paragraph Kant reiterates the point that

showing that a concept is logically possibly instantiated is not sufficient to show that it is

really possibly instantiated. This strongly suggests that the relevant notion of possibility is

real possibility. But what does ‘real possibility’ mean here?

In the case of mathematical concepts and schematized categories the relevant notions of

real possibility are constructability in pure intuition and compatibility with the forms of

experience (the former may be an instance of the latter), respectively.39 But those notions of

real possibility clearly do not apply to all objects in general. When Kant claims that we

cannot give a real definition of a single one of the categories as concepts of objects in

39 A239-240/B298-9. At A228/B265 Kant defines the possible as “whatever agrees with the formal conditions of experience (in accordance with intuition and concepts).” If this were the only notion of real possibility in Kant’s Critical philosophy, then we could prove that it is really impossible for noumena to fall under the categories; because they are not in space and time, they are not compatible with the forms of experience. This in turn entails that we are not ignorant about the real possibility of objects in general falling under the categories: we know that phenomena can fall under the categories, and we know that noumena cannot. But, as we saw earlier, throughout the ‘noumena and phenomena’ section Kant claims that we cannot know whether it is really possible for objects in general (and in particular, noumena) to fall under the categories. Immediately after the definition of possibility quoted above, Kant points out that “the principles of modality are nothing further than the definitions of the concepts of possibility , actuality, and necessity in their empirical use” (A219/B266 – my underlining). Kant goes on to deny that modal categories, or any categories for that matter, have a “transcendental use” (for noumena), but later, in ‘Phenonmena and noumena,’ he makes clear that by denying a transcendental “use” of the categories he means merely to deny that we can use them in synthetic a priori cognition of noumena (A247/B304).

17

general, which I have interpreted to mean that we cannot identify the grounds of the real

possibility of an object in general falling under the categories, he, presumably, does not have

these notions of real possibility in mind. So what kind of real possibility is relevant to the

question of the real definition of the categories as concepts of objects in general? In what

sense of real possibility do we lack knowledge of the grounds of real possibility of objects in

general falling under the categories?

Before I address that question directly, I want to point out that the modal categories

should be treated slightly differently than the other categories. The problem of the real

definition of <cause-effect>, for instance, is to determine the ground of the real possibility

of objects in general standing in cause-effect relations, i.e. one object being such as to posit

the existence of an object distinct from itself. But applying the same scheme to the problem

of the real definition of <possibility> we get: the problem is to determine the ground of the

real possibility of objects in general being possible. In other words, we get an iterated

modality: the ground of the real possibility of possibility. This is very indirect and artificial.

It is much more natural to think that what we want from a real definition of the category

<possibility> is to determine the grounds of real possibility of objects in general; no modal

iteration is necessary. After all, possibility is already included in the idea of a real definition

of a concept, so it makes sense that when we come to the question of the real definition of

<possibility> itself we do not ask about the grounds of the real possibility of possibility, we

ask about the grounds of real possibility period. To take a related example, Kant’s complaint

about what he calls the ‘nominal definition’ of necessary existence (A593/B621) 40—a being

exists necessarily just in case its non-existence is impossible—is not that this fails to tell us

what makes a necessary being possible (iterated modality) but that it fails to tell us, if there is

40 Kant makes the same point in OPG at Ak. 2:81.

18

a necessary being, what grounds the necessary existence of that being, i.e. why it is not

possible for that being not to exist. Likewise, what is missing in our grasp of the category

<possible> is an account of what grounds the possibility of all objects in general – not an

account of what grounds the real possibility of the possibility of all objects in general.

But this means that the question being raised about the non-modal categories—e.g. what

is the ground of the real possibility of objects in general being causes? —is the same kind of

question being raised about the modal category—what is the ground of the possibility of

objects in general?—only if the notion of ‘real possibility’ involved in the first question is the

same as the notion of possibility involved in the second question, which is itself the modal

category <possibility>. This means that the sense of ‘real possibility’ involved in a demand

for a real definition of the non-modal categories is the same notion of possibility involved in

a demand for a real definition of the modal category <possibility>. In other words, the

unschematized category <possibility>, applied to all objects in general, is the sense of real

possibility at issue when Kant denies we have insight into the grounds of real possibility of

objects in general falling under the categories.

The categories, including the modal categories, are concepts we apply when we think

about objects in general. So if the unschematized category <possibility> (henceforth,

<possibility>UC) is a concept of real possibility, then it must be the highest or most general

concept of real possibility. If there were a more general concept of real possibility than

<possibility>UC then in thinking about that kind of real possibility we would not be using the

unschematized category, and this is impossible; any thought about objects requires thinking

about them under the categories.

Let me offer another argument for the same conclusion. If there were a concept of real

possibility more general than <possibility>UC then that concept would be the concept of real

19

possibility relevant to the question of the real definition of the non-modal categories. The

question would be, what grounds the real possibility, in this broader sense, that objects fall

under the non-modal categories? But then either that would not be the concept of real

possibility involved in the question of the real definition of <possibility>UC—in which case

the question of the real definition of the non-modal categories would be substantially

different than the question of the real definition of the modal categories, rather than parallel

questions, as Kant presents them—or the question of the real definition of <possibility>UC

would be a question of iterated modality: what grounds the real possibility, in the broader

sense, that something fall under the more restricted unschematized category of (real)

<possibility>UC?

Conversely, if the most general concept of real possibility were less general than

<possibility>UC—if logical possibility were a species of the unschematized category—then

Kant’s claim that we cannot give a real definition of a single one of the categories as

concepts of objects in general would be misplaced with respect to <possibility>UC. For in

this case, we could give a real definition of a species of <possibility>UC for all objects in

general, logical possibility, for we know quite generally what logical possibility is (not

containing mutually contradictory marks).41 The categories are concepts at exactly the level

where Kant thinks we cannot give real definitions for all objects in general. If there is a

more general genus of the concept of which we cannot give a real definition for all objects in

general, then we have too narrowly specified the category. And if there is a species of the

concept for which we can give a real definition for all objects in general, then we have

identified the category too generically.

What does it mean that the unschematized category is the most general concept of real

41 Ak. 8:195.

20

possibility? First of all, it means that <possibility>UC is a concept of real possibility, rather

than of logical possibility. This means that <possibility>UC is a different concept than

<logical possibility>. Consequently, it is logically possible that they pick out distinct

extensions42 in any given domain; it cannot be determined through conceptual analysis alone

whether, for instance, all and only the propositions43 about non-empirical objects that are

logically possible are also really possible in the most general sense.44 For instance, they may

(for all we know through conceptual analysis) pick out the same extensions for non-

empirical objects; it may be that every logically possible proposition about non-empirical

objects is really possible in the most general sense (i.e. falls under <possible>UC). There are

synthetic, hence logically contingent, really necessary truths about empirical objects,45 so

42 Some readers might balk at the idea of modal concepts having ‘extensions,’ especially given Kant’s claim that “the categories of modality . . . do not augment the concept to which they are ascribed in the least, but rather express only the relation to the faculty of cognition” (A219/B266). I take this to mean that the extensions of modal concepts (the ‘things’ to which they can be correctly applied) are not objects (<possibility> does not pick out a class of objects, the possible objects, as opposed to the impossible objects). We can talk about the ‘extension’ of a modal concept by talking about the judgments that this modal concept can be correctly applied to; the extension of <possible> is the collection of all judgments which are possibly true. Since ‘judgment’ carries with it the connotation of a particular mental act or event, while Kantian judgments are the inter-subjectively shareable contents of such acts, I will talk about propositions instead (without assuming the contemporary orthodoxy that propositions are Platonic entities). So to say that p is in the extension of, e.g. the modal concept <possibility>, will mean that the judgment <possibly p> is true. For more on Kant’s technical notion of ‘determination’ and why modal concepts are not determinations, see Stang (forthcoming-a) and KMM, ch. 2 43 By ‘proposition’ I mean merely the content of a judgment rather than the act of judging itself. See previous note. 44 It is, I take it, a conceptual truth that everything that is really possible is logically possible. 45 This is a consequence of Kant’s most famous doctrine: the existence of synthetic a priori judgments. If they are a priori, they are necessarily true (B3-4). Since they are synthetic, they are not logically necessary; no contradiction follows from the negation of such a judgment (A150/B189-A153/B192). For more on the relation between the a priori and the necessary in Kant, see Stang (2011).

21

<possible>UC and <logically possible> have different extensions in the empirical domain.

Consequently, they have distinct extensions simpliciter. Whether they have distinct extensions

in the non-empirical domain cannot be determined through conceptual analysis. In other

words, we cannot assume at the outset that everything that is logically possible is really

possible for non-empirical objects; because the category <possible>UC is not a mark of the

concept <logically possible> (the category is a concept of real possibility) we need some way

to synthesize a connection between them, some ‘X’ to connect them, in order to judge that

everything that is logically possible is really possible for non-empirical objects.46 This, I take

it, is Kant’s point when he repeatedly distinguishes real possibility from logical possibility,

and denies that we can always infer from logical possibility to real possibility,47 even in the

case of non-empirical objects.48

The other thing I mean by claiming that the unschematized category of possibility is a

concept of real possibility is that it is part of the unschematized category of (real) possibility

that a (real) possibility is grounded in some aspect of actuality. This is what Kant means in

the Pölitz lectures when he writes: “yet we have no concept of real possibility except through

existence, and in the case of every possibility which we think realiter we always presuppose

some existence” (Ak. 28: 1036).49 I interpret this as a claim about the unschematized

category of (real) possibility in general, <possible>UC. I take this to be a partial nominal

46 A9/B13. 47 E.g. Bxxvi, A244/B302 (and the footnote added to that page in the B edition), and Ak. 20:325. 48 E.g. A596/B624. In context, Kant’s point is that we cannot infer that God is really possible just because the concept <God> is logically consistent. This only makes sense if real possibility also applies to non-empirical objects, for if it did, we could conclude that God is not really possible because he is not the object of any experience. 49 See also Ak. 28:310, as well as Refl. 3931 and 5034.

22

definition of the category similar to the principle: the necessary is that whose non-being is

not possible (A244/B302). This is why our concept of ‘real possibility’ carries with it the

rational necessity to inquire into the ground(s) of real possibility: in thinking about noumena

we have to think about them using the unschematized category of (real) possibility and this

carries with it the thought that (real) possibilities for noumena—as for all objects in

general—have grounds in actuality.50 To anticipate slightly, I will argue that Kant implicitly

distinguishes several different kinds of real possibility. What makes them all kinds of real

possibility is that they all instantiate the nominal definition of the highest concept of real

possibility, <possibility>UC; they are all conceptually distinct from logical possibility and they

are grounded in actuality. More precisely,

(Real Possibility) For any kind of possibility ◇xp and its associated kind of necessity □xp (where ◇xp iff ~□x~p), ◇x is a kind of real possibility (and □xp is a kind of real necessity) only if (i) Non-logicality. It is not a conceptual truth that ◇Lp ⊃

◇xp (and it is not a conceptual truth that □xp ⊃ □Lp) and,

(ii) Groundedness. If ◇xp then the fact that ◇xp is grounded in some feature of actuality.

The subscript ‘x’ indicates the different kinds of possibility, e.g. ‘◇L’ refers to logical

possibility. This nominal definition presupposes an account of what it is for something to

be a kind of possibility (or what it is for a concept to be a modal concept); I do not have

space here to give an analysis of these notions (e.g. of what makes logical possibility and real

possibility both kinds of possibility).51

Some readers might note that in the passage quoted from the Pölitz lectures, Kant says

50 This is only the beginning of such an explanation. My complete story of this ‘ratioanl necessity’ is told in KMM, ch. 8. 51 I attempt to answer this more general question in KMM, ch. 6.

23

that a real possibility must be grounded in existence [Dasein, Existenz] while I claimed it was

grounded in actuality [Wirklichkeit]. Kant sometimes uses these terms interchangeably52, and

sometimes he distinguishes them. For instance, in Refl. 6324 (Ak. 18:647) Kant claims that

space and time are ‘etwas Wirkliches’ but ‘nichts Existierendes.’ In that context, I take the

distinction to be between something that actually is the case (e.g. we actually have space and

time as our forms of intuition) and something that, in addition, exists as a substance with

causal powers.53 Formulated in terms of this distinction, I think Kant’s view in the Pölitz

lectures is “in the case of every possibility which we think realiter we always presuppose some

actuality,” for clearly there are kinds of real possibility that are not grounded in an existing

object. At A218/B265 Kant defines possibility as “whatever agrees with the formal

conditions of experience (in accordance with intuition and concepts).” I will argue below

that this is a notion of real possibility (I call it ‘formal real possibility’). Formal real

possibility is grounded in the forms of experience, including space and time, which,

according to Refl. 6324, are actual but do not exist. Not all kinds of real possibility are

grounded in existence.54

52 Arguably, the passage from Pölitz RT quoted at the beginning is such a passage; Kant seems to use ‘Wirklichkeit’, ‘Existenz,’ and ‘Daseyn’ interchangeably there. In the initial table of the categories in the first Critique he gives ‘Dasein-Nichtsein’ as the second category of modality (A80/B106), but in the Postulates of Empirical Thought, the later section devoted to the modal categories, the second modal category is now the category of ‘Wirklichkeit’ (A218/B266). See also the von Schön metaphysics lectures, where Kant says: “real actuality is here the category of existence [Existenz], in contrast to the possibility of a thing” (28:493). 53 Readers who know German will note that this is somewhat ironic, because ‘Wirklichkeit’ (much like the English ‘actuality’) has ‘wirken,’ to act or to affect something, as its root; it would be much more natural for ‘Wirklichkeit’ to refer to objects that are causally efficacious. 54 Some readers will note that Dasein is a category (and Wirklichkeit sometimes seems to be treated as a category as well--see note 47 above) and ask: in the main text, am I talking about the unschematized category of Dasein (or Wirklichkeit) or the schematized category? My short answer (defended more fully in KMM) is that the unschematized category of possibilty is grounded in the unschematized category of existence, which is the concept of ‘absolute

24

This also shows us how to relate Kant’s Critical theory of real possibility in general (his

theory of <possibility>UC) to his pre-Critical theory of real possibility. As I explained in the

Introduction, in OPG Kant argues for two principles about real possibility: (i) not all logical

possibilities are real possibilities, and (ii) all real possibilities must be grounded in an actually

existing being. In particular, Kant argues that all real possibility must be grounded in a

substance, God. In terms of the distinction between ‘existence’ and ‘actuality’ drawn in the

previous paragraph (cf. Refl. 6324), the Critical Kant drops the requirement that the grounds

of real possibility must be existent (they need not be causally efficacious substances) but

maintains the condition that they must be actual. Kant’s Critical theory of real possibility in

general (his theory of <possibility>UC) is a generalization of the pre-Critical theory that

maintains the same overall structure, a point to which I shall return in the next section.

We can now return to Kant’s claim that we can only give real definitions of the

categories for empirical objects. I argued earlier that the problem of the real definition of

the modal categories is the problem of knowing the grounds of the real possibility of objects

überhaupt. We can now see that Kant’s claim that we can give real definitions of the

categories only for empirical objects means, in the case of the modal categories, that we only

know the grounds of real possibility for empirical objects. If p is some proposition, and p is

really possible, then we can know the grounds of the real possibility that p only if p is a

proposition concerning empirical objects only. But empirical objects have two kinds of

grounds: ‘immanent ones,’ i.e. the forms of experience of cognizing subjects and causal

grounds in other empirical objects, and ‘transcendent’ ones, i.e. noumenal objects that positing’ (A596/B624); it is this concept of existence that Kant has in mind when he denies that existence is a ‘real predicate.’ The schematized concept of existence is the concept of a causally efficacious item (a substance or a mode of a substance); this concept is a real predicate. So the distinction I am drawing in the main text between actuality and existence is really the distinction between the unschematized category of existence (actuality, in the main text) and the schematized concept of existence (existence, in the main text).

25

ground empirical objects by causally affecting subjects and producing the sensory matter of

experience.55 We only know the ‘immanent’ grounds of real possibility for empirical objects,

so the only kinds of real possibility for which we can give real definitions are those with

immanent grounds.

To illustrate this point, consider two kinds of real (non-logical) possibility Kant invokes

in CPR without explicitly distinguishing them:56

(Formal) It is formally really possible that p iff p is compatible with the forms of experience.57

(Empirical) It is empirically really possible that p iff p is compatible with the

natural laws and the past up to time t, where p is the proposition that alteration E occurs at t.58, 59

For instance, it is formally possible both that I go to the movies tonight, and that I do not,

55 Kant talks of phenomena having noumenal grounds at A380, Ak. 4:314-5, 453, 8:203, 205. Elsewhere, he claims that noumena causally affect us, thereby producing the matter of experience: A190/B235, A387, A494/B522, Ak. 4:289, 4:314, 4:318, 4:451 and 8:215. On my view, noumena are grounds of phenomena in that they causally affect our minds, producing sensations, which are synthesized into experience of phenomena. For a defense of the claim that noumena causally affect us, see the classic Adickes (1924), as well as the Introduction to Ameriks (2003). 56 These are discussed in much greater detail in Stang (2011). 57 This is the kind of real possibility Kant defines at A218/B266. This is the kind of real possibility Kant typically has in mind in CPR when he talks about ‘possibility’ in general or ‘real possibility’ in particular. E.g. A111 and A127. 58 This is the kind of possibility that corresponds to the kind of necessity Kant defines at A218/B266 and which he discusses further at A226-7/B279-80. Those are discussion of empirical necessity, which is related to what I am calling ‘empirical possibility’ in the standard way: it is empirically possible that p iff it is not empirically necessary that ~p. See also Kant’s lectures on metaphysics (Ak. 28:417, 29:814) and Refl. 4298 and 5177. 59 I am departing from Kant’s terminology here. Kant typically defines possibility in terms of the possibility of an object – see his metaphysics lectures (28:410, 543 and 29:811, 960). For ease of exposition, I have defined possibility in terms of the possibility of a proposition. These are inter-convertible definitions: an object a is possible just in case the proposition that a exists is possible.

26

since either state of affairs is compatible with the a priori forms of experience. However, if

the natural laws are deterministic, as Kant thinks they are60, they determine a unique

empirically possible future; so the proposition that I go to the movies tonight is either

empirically necessary or it is empirically impossible. These are both kinds of real possibility

because both of them are distinct from logical possibility; not all logical possibilities are

compatible with the forms of experience (e.g. the propositions of non-Euclidean geometry)

and many logically possible propositions about the future are not compatible with the laws

and the past up to this moment (e.g. any logically consistent but false proposition about the

future). Both kinds of real possibility are grounded in actuality: our actual forms of

experience (space, time, and the schematized categories) and the actual natural laws and past

events.

In both cases, we know what grounds real possibility. With respect to formal real

possibility and empirical real possibility, we do not know merely that certain things are

possible; we also know why. For instance, we can construct a triangle in pure intuition and

thereby know why it is formally possible: it is compatible with our spatial form of intuition.

Likewise, by coming to know the natural laws we come to know not merely that various

actual events are empirically possible but why; we come to know why they are continuations

of the past that are compatible with those laws. But we do not know the noumenal (non-

immanent) grounds of the real possibility of empirical objects, because we in general do not

know the noumenal (non-immanent) grounds of the real possibility of objects. So we only

know one ‘half’ or ‘side’ of the real possibility of empirical objects. And we cannot give a

real definitions of <possibility>UC for all objects überhaupt because we do not know in

general the grounds of the real possibility of objects. We only know the immanent grounds 60 A108, A113, A114, A127-8, B165, A159/B198, A216/B263, and Prolegomena § 36; for a different interpretation, see Allison (1983), 228-34; and (1994), 298.

27

of real possibility for empirical objects.

We have the unschematized category of real possibility in general, <possibility>UC. We

can give real definitions of at least two61 distinct kinds of real possibility, but these kinds of

real possibility only apply to empirical objects and, furthermore, only apply to kinds of real

possibility that have immanent grounds (empirical objects, or subjective conditions thereof).

We can then think of the noumenal grounds of the real possibility of objects, even for these

kinds of real possibility (formal, empirical) that apply only to empirical objects. For instance,

a series of events that is different than the series of actual events is compatible with our

forms of experience, if it is a series of law-governed alterations in permanently existing and

reciprocally interacting substances.62 Consequently, it is formally possible.63 Likewise, such a

series of non-actual events is empirically impossible because it is not a continuation of the

past that is compatible with actual laws (since these laws are deterministic). But there is

another kind of real possibility we might ask about: does this non-actual empirical series

have a noumenal ground? To make this more concrete, could we be noumenally affected so

that we would experience this non-actual empirical series? It is compatible with my forms of

experience that I receive a completely different set of sensory matter, and thus have an

experience with the same form as my actual experience but a different matter, but that is not

the question. What is at issue is a question about the noumenal ground of the real possibility

of my experiencing a different empirical series. I take it that is what Kant has in mind in the

61 In KMM, ch. 7 I argue that nomic necessity, the necessity that attaches to laws, is neither formal necessity nor empirical necessity. 62 I will leave unanswered here whether it is compatible with the forms of experience that these events be governed by natural laws different than the actual laws. The question of the modal status of natural laws for Kant is difficult; I address it in detail in KMM, ch. 7. 63 More precisely: the proposition that such a series occurs is formally possible.

28

‘Postulates of empirical thinking in general’ when he writes “whether other perceptions than

those which in general belong to our entire possible experience and therefore an entirely

different field of matter can obtain cannot be decided by the understanding, which has to do

only with the synthesis of that which is given” (A231/B283). It is not something we can

know; we do not know the noumenal grounds of real possibility, even of the real possibility

of propositions exclusively about empirical objects. This is the notion of real possibility that

I have called ‘noumenal1’ in figure one. I also think this is the notion of real possibility

involved in Kant’s theory of transcendental freedom and his reconciliation of our freedom in

action with the empirical necessity of all of our actions, but I do not have the space to

defend that claim here. 64

64 I argue for this in Stang (2011), 459-460 and in greater detail in KMM, ch. 6. Very roughly, the idea is that I am morally responsible for an action only if it is really possible for me to omit it, and the ground of that real possibility is my power as a noumenon to cause that action as a phenomenon to be omitted. See A537/B565, A541/B569, A536/B564, and Ak. 5:97-98.

29

Finally, for propositions involving noumena (as well as empirical objects) we might ask

whether they are really possible, using the unschematized category of (real) possibility,

<possibility>UC.65 In doing so, we would be thinking about a kind of real possibility whose

grounds must be noumenal, because noumena ground empirical objects, and not vice versa

(so empirical objects cannot be among the grounds of the real possibility of noumena or any

state-of-affairs involving noumena). We cannot in principle know what the grounds of this

kind of real possibility would be. This is the notion of real possibility that I have called

‘noumenal2’ in figure one. Some readers might be skeptical that Kant accepts any such

notion of real noumenal possibility. However, note the following claims Kant makes in the

65 Kant seems to be applying modal concepts directly to noumena—rather than merely considering noumena as putative grounds of the possibility of empirical objects—at Refl. 5184, 5723 and 5177. I also think that this kind of possibility is involved in Kant’s resolution of the Fourth Antinomy, but I do not have space to defend that claim here.

Concepts of possibility

Logical possibility Real Possibility = unschematized category of possibility for all

objects in general

•!Must have a ground in (unschematized) actuality

•!Interdefinable with real necessity (though not with absolutely

necessary existence)

Real Possibility for Empirical

Objects = unschematized category

of possibility applied to empirical

objects

Where the actual ground of real possibility is

‘immanent’:

-!Formal possibility. Grounded in actual

forms of experience.

-!Empirical possibility. Grounded in actual

laws and actual empirical events up to a time.

Real Possibility for Non-Empirical

Objects = unschematized category of

possibility applied to non-empirical objects

(Noumenal2) It is noumenally2 possible that

p iff p concerns noumena and there is a

noumenal ground of the possibility that p.

Where the actual ground of real

possibility is ‘transcendent’:

(Noumenal1) It is noumenally1

possible that p iff p concerns only

empirical objects and there is a

noumenal ground of the

possibility that p.

(Noumenal) It is

noumenally possible

that p iff there is a

noumenal ground of

the possibility that p. (Figure One)

30

Pölitz lectures:

[. . .] if we remove an original being, we at the same time remove the substratum of the possibility of all things [. . .] my reason makes it absolutely necessary for me to assume a being which is the ground of everything possible [. . .] otherwise I would be unable to know what in general the possibility of something consists in [. . .] a thing is eo ipso necessary if its nonexistence would cancel all possibility [. . . ] in the case of every possibility which we think realiter we always presuppose some existence [. . .] hence all possibilities presuppose something actually given [. . .] (Ak. 28:1034-6--underlining by NS). Kant’s repeated emphasis that the absolutely necessary being must hypothesized to ground

the real possibility of “all” things, or real possibility “in general” is naturally read as meaning:

the absolutely necessary being grounds the real possibility of both phenomena and noumena.

This only makes sense if there is at least one concept of real possibility that applies to

noumena, what I have called noumenal2 real possibility.

The Pölitz lectures are not the only places where Kant talks about noumenal2 real

possibility. In the section ‘On the ground of the distinction of all objects in general into

noumena and phenomena’ he writes:

we have no insight [Einsicht] into the possibility of such noumena, and the domain outside the sphere of appearances is empty (for us), i.e. we have an understanding that extends farther than sensibility problematically, but no intuition [. . .] (A255/B310)66 I take Kant’s point here to be that we do not know whether noumena are possible. This

agnosticism makes no sense if ‘possibility’ here means formal real possibility or empirical real

possibility, for, trivially, no noumenon is compatible with our forms of experience (formal

real possibility) or can exist in a continuation of the past given the laws (empirical real

possibility).67 Since the context of this passage is Kant’s explanation of why our knowledge

66 Cf. Kant’s discussion of “metaphysical possibility” at Ak. 28:182 and Refl. 5177 where he talks about “the metaphysical possibility of a thing in itself [Sache an sich], insofar as it cannot be an object of experience” (Ak. 18:109). 67 In Refl. 5184 Kant writes that “the synthetic conditions of the possibility of experience are at the same time condiitons of the possibility of the objects of experience. This is not the

31

of the real possibility of objects falling under the categories does not extend to non-empirical

objects, noumena, and in the next two paragraphs he will admit that thinking of noumena

under the categories is “free of contradiction” (A252, 253), it is also clear that “possibility”

here refers to a form of real possibility. Since Kant is claiming that, in the relevant sense of

possibility, we do not know whether noumena are really possible, this is clearly a kind of real

possibility that must be grounded in noumena, since noumena cannot be grounded in

phenomena. It follows that, whatever this notion of real possibility is, it is a case of what I

have called noumenal2 real possibility. Note, however, that we can assert this while

remaining completely agnostic about what noumenal2 real possibility is, which things are, and

which things are not, noumenally2 really possible, and how these possibilities are grounded in

noumena. My point is merely that it is something we can (without contradiction) think

about, and about which, according to Kant, in the Pölitz lectures, we are rationally required

to think.

Finally, noumenal1 and noumenal2 real possibility are species of a single conception of

real possibility, which I have relabeled noumenal real possibility simpliciter in figure one: the

concept of any kind of real possibility grounded in noumena. We cannot give a real

definition of this kind of possibility, for we do not know the noumenal grounds of real

possibility.

3. Absolutely Necessary Existence

To return to the original passage from the Pölitz theology lectures, I am now in a position to

offer my interpretation of what Kant means by ‘real possibility’ there: noumenal real possibility of things in themselves” (Ak. 18:111). I take this to mean: formal real possibility is not noumenal2 real possibility.

32

possibility, real possibilities for all things insofar as those real possibilities have noumenal

grounds, which we think through the unschematized category of (real) possibility,

<possibility>UC. The reason ‘real possibility’ there means ‘noumenal real possibility’ is that

Kant claims we must hypothesize an absolutely necessary being to ground all such real

possibilities and, as I pointed out at the beginning of section one, an absolutely necessary

being must be a noumenon (no phenomena are absolutely necessary). So the real possibility

at issue in the Pölitz text is grounded in noumena; it is noumenal real possibility, the genus

of which noumenal1 and noumenal2 real possibility are species. But Kant’s primary concern

in that passage, and the jumping-off point for our whole discussion, was his invocation of

the notion of ‘absolutely necessary existence’ and his Critical reinterpretation of the 1763

argument for the existence of an absolutely necessary ground of real possibility, or, as we can

now identify it, of noumenal real possibility.

The first thing we need to understand is what ‘absolute necessity’ means in the Pölitz

text. Fortunately, it means the same thing it does in OPG. Consider these two passages, the

first of which is from Pölitz, the second of which is from OPG:

In addition to the logical concept of the necessity of a thing (where something is said to be absolutely necessary if its nonexistence would be a contradiction, and consequently impossible), we have yet another rational concept of real necessity. This is where a thing is eo ipso necessary if its nonexistence would cancel all possibility. (Pölitz RT; Ak. 28:1036). Something may be absolutely necessary either when the formal element of all that can be thought is cancelled by means of its opposite [the ‘logical concept of the necessity of a thing’ – NS], that is to say, when it is self-contradictory; or alternatively, when its non-existence eliminates the material element and all that data of all that can be thought. (OPG; Ak. 2:82) The ‘data of all that can be thought,’ in the context of OPG, is a reference to real possibility.

Consequently, I interpret ‘absolute necessity’ in both texts as follows:

(1) It is absolute necessary that p =def were it not the case that p, nothing would be really possible.

The right-hand side of this definition is a counterpossible conditional: it says that per

33

impossibile if p were false, nothing would be really possible. This is how I interpret Kant’s talk

of the ‘cancellation’ of possibility in Pölitz (quoted above) and the ‘elimination’ of possibility

in OPG. Earlier in the same paragraph in OPG Kant writes “when I cancel all existence

whatever and the ultimate real ground of all that can be thought therewith disappears, all

possibility likewise vanishes” (Ak. 2:82). I take this to be clear support for the

counterpossible interpretation of absolute necessity in (1).68

One natural question is whether absolute necessity, so defined, can be simply identified

with real necessity, i.e.

(2) It is absolutely necessary that p =def it is really necessary that p.

It cannot; if it were, by the inter-definition of real necessity and real possibility, this would

entail that:

(3) It is absolutely necessary that p iff it is not really possible that ~p.

Given the definition of absolute necessity from above, (1), and the principle that if it is really

possible that ~p then there is a ground of the real possibility that ~p, this would entail that:

(4) If there is no ground of the real possibility that ~p then (if it were the case that ~p, nothing would be really possible).

But this cannot be right, because if ~p simply lacks a ground of real possibility, it does not

follow that its non-obtaining would cancel all real possibility. For instance, if I do not have a

transcendentally free will then ,in the relevant sense of possibility, the possibility of omitting

an act I actually commit lacks a ground of real possibility; it does not follow that were I to

have done otherwise, nothing would be really possible!69 My actions do not have that

68 I further defend this interpretation in Stang (2010). 69 To flesh this out slightly, on my interpretation, it is a necessary condition on having a transcendentally free will that it is really possible that I omit actions I actually commit. The relevant notion of real possibility is real noumenal1 possibility: it is the possibility of a phenomenal event (the omission of my action) grounded in my nature as a noumenon.

34

exalted status; only God does.

This is a short argument to the effect that the concept of absolute necessity invoked in

the original Pölitz passage is not inter-definable with the concept of real possibility (which I

have identified with the unschematized category of possibility). But recall how Kant gets to

that concept of absolute necessity: it is part of the concept of real possibility that real

possibilities have grounds in actuality. But reason requires me to think of things as unified,

to impose in thought a unity that I do not know to obtain. This gives me reason to think

that there is a single unified ground of real possibility. While I am not attempting to

reconstruct Kant’s reasons in the 1780s for claiming that we are rationally necessitated to

hypothesize such a first ground of real possibility, this short sketch does suffice to make an

important point. Whereas Kant had argued in 1763 in OPG that real possibility could not be

collectively grounded by a plurality of distinct grounds70, he now claims less ambitiously but

more plausibly: it is rationally unsatisfactory if there are a plurality of distinct grounds of real

possibility, none of which has priority over any others. The most rationally satisfactory

conception of modal reality is one on which there is a unique ground of real possibility—or

at least a ‘first’ ground, one that grounds the grounding of real possibility in ‘lower level’

grounds, such that the non-existence of that first ground would cancel all real possibility.71

So rationally we have to hypothesize such a being, because it is part of the most rationally More specifically, the ground of the real possibility of this omission is the fact that, as a noumenon, I possess the causal power to have determined myself in such a way that this action would be omitted. Thus, the ultimate ground of the real noumenal1 possibility of the omission of my action is my causal power as a noumenon. See A532/B560 – A558/B586. I am in broad agreement with Wood (1984) and Pereboom (2007); my interpretation of Kant’s theory of free will is defended at greater length in KMM. 70 See Adams (2000), Stang (2010), and Yong (forthcoming) for different reconstructions of this argument. 71 I argue for this interpretation of the ‘first ground’ of real possibility in KMM, ch. 4.

35

satisfactory explanation of why anything is really possible.72

If my interpretation is correct up to this point, we should expect that—because quite

generally we are required to hypothesize an absolutely necessary first ground of real

possibility of all objects in general—for each lower level kind of real possibility, we should

hypothesize an absolutely necessary (in some sense) first ground of that kind of real

possibility for all objects of the relevant sort. In other words, for each of the lower-level

kinds of real possibility we should expect to find a conception of an absolutely necessary

first ground of real possibility of the relevant kind. Thus, for each lower-level kind of

possibility, we should expect to find some absolutely necessary principle p such that:

(1*) It is absolutely necessaryx (absolute-□x) that p =def were it not the case that p, nothing would be really possiblex (◇x).

73 The subscript x refers to the different kinds of possibility, from above. For instance, we

should expect that there is some p such that the right hand of this definition is true:

(1F) It is absolutely formally necessary (absolute-□F) that p =def were it not the case that p, nothing would be formally really possible (◇F).

We cannot say that the proposition p will be the proposition that the relevant first ground of

real possibility exists, for not all kinds of real possibility have existing grounds (e.g. formal

possibility—see above); but we can say that p will be the proposition that the relevant first

ground of real possibility (whatever it is) is actual or obtains (or something of that sort).

And in the case of ‘immanent’ real modalities—modalities with immanent grounds—we

should expect that we can know what this absolutely necessary first ground of real possibility

72 I reconstruct Kant’s argument for the rational necessity of positing a unique first ground of real possibility in detail in KMM, ch. 8, sections 3-5. 73 ‘Absolute-□x’ indicates that this is a definition of the absolute form of the relevant kind of necessity; the argument given above shows that this is not simply necessity of the relevant kind (□x).

36

is, for these are precisely the kinds of real possibility for which we can identify the grounds.

And this expectation is satisfied, at least in the case of formal possibility. Kant identifies

space as the absolutely necessary first ground of formal possibility for outer objects and

explicitly compares space in this respect to the rational idea of God as the absolutely

necessary first ground of real (noumenal) possibility:

All reality must be completely given, and so some actuality is prior to possibility, just as space is not merely something possible, but the ground of all possible figures. (Refl. 4119, Ak. 17:424)

From this it follows only that the ens realissimum must be given prior to real concepts of possibility, just as space cannot be thought antecedently as possible, but as given. But not as an actual [wirklich] object in itself, rather as a merely sensible form, in which alone all objects can be intuited. (Refl. 6290, Ak. 18:558)74 These texts suggest that some proposition about space will the right-hand side of (1F) true.

However, it is not clear which proposition about space this is. In OPG, the absolutely

necessary proposition is that God exists. This will not quite work for space, because space

does not exist in the same way that causally efficacious substances do; space is a form of

empirical objects, not a substance in its own right (see Refl. 6324). The absolute formal

necessity of space consists, then, not in the fact that space exists absolutely formally

necessarily, but in the fact that it is absolutely formally necessary that we experience outer

objects in space. If this is correct then, per impossible, were it the case that we did not

experience outer objects in space, then no outer object would be formally possible. It is

hard to know how to evaluate counter-possible conditionals like this. I have argued

elsewhere that we should evaluate claims like this by applying the following test: without

74 In the second passage, I take Kant’s point to be that space does not exist as a substance, but as a form in which objects can be intuited. So he here uses ‘wirklich’ in exactly the opposite sense of Refl. 6324, discussed above. See also Refl. 4515, 4570, 5723, and 6285, as well as Ak. 28:1259.

37

appeal to the relevant p can we explain facts about the given kind of real possibility? 75

Applying this to space and formal possibility, we should ask: without appeal to space, can we

explain facts about the formal possibility of outer empirical objects? It is a pillar of Kant’s

Critical theory of experience that we cannot. Therefore, for Kant, the representation of

outer objects in space is absolutely formally necessary. Space is the first ground of the

formal possibility of outer objects.76 For further confirmation of my view I would need to

search for evidence that something plays the role of an absolutely necessary first ground of

real possibility for the other kind of ‘immanent’ real possibility I distinguished earlier:

empirical possibility (likewise, for any other kinds of immanent real possibility). I argue

elsewhere that the absolutely necesssary first ground of empirical possibility is the actual laws

of nature and the past series of events, but I do not have the space here to go into that.77

I began by asking what we are doing when we use (unschematized) modal categories to

think about noumena. I think we can now answer that question, as follows: we are applying

the most general concept of real possibility, the unschematized category of possibility, of

which the following are conceptual truths (i) it is conceptually distinct from logical

possibility, (ii) real possibilities are grounded in actuality, and (iii) it is inter-definable with

real necessity. There are several different kinds of real possibility, lower-level species of the

genius, (real) <possibility>UC: formal real possibility, empirical real possibility, noumenal real

possibility, etc.. These conceptual truths (i)-(iii) also apply to each of these lower-level kinds

75 KMM, ch. 4 and 8. 76 One would expect Kant to say similar things about time. However, based on my overview of the relevant texts, I conclude that he more frequently draws the comparison between space and God as first grounds of (formal and noumenal, respectively) possibility. 77 See, however, KMM, ch. 8. I also distinguish empirical possibility from nomic possibility, and argue that the absolutely necessary first ground of nomic possibility is the real essences of empirical natural kinds.

38

of possibility. For each lower-level kind we also have the idea of a first ground of real

possibility of that kind that exists with absolute necessity with respect to that kind (as defined

above). With respect to some of these kinds of real possibility, we know what their grounds

are, and we can identify an absolutely necessary first real ground. For instance, we know that

space, time, and the categories are the grounds of formal real possibility, and that space is

absolutely formally necessary. But we do not know what grounds the real possibility of

objects in general; in particular, we do not know any of the noumenal grounds of real

possibility. For reasons I have not gone into in this paper, we are rationally required to

hypothesize that among the noumena there is such an absolutely necessary first ground of all

noumenal real possibility. But we do not know whether such a being exists, and we can only

have a quite indeterminate concept of it, and how it grounds real noumenal possibility.

This is not very satisfying. I have been arguing, though, that it is the most satisfying

explanation we can give, on Kant’s view. Kant’s view is that we cannot say more about real

possibility generally. We can say more about particular kinds of real possibility (immanent

real possibility) for empirical objects, but that is a specific application of the concept of real

possibility to a particular domain of objects.

But is contemporary modal metaphysics much more satisfying on this question?

Consider how modal metaphysics has developed since Naming and Necessity. Kripke

identified a class of examples of the necessary a posteriori (e.g. water=H2O) and convinced

philosophers that there was a kind of modality, metaphysical necessity, that comes apart

from a priori knowability and from analyticity. This introduced (or re-introduced) a

distinction in our concepts between <metaphysical necessity> and (what we might think of

39

as) <conceptual necessity>78 and showed that in one domain (e.g. propositions about the

identities and natures of natural kinds) they have a different extension; it did not, though,

show that in every domain they have a different extension. It may very well be that in

mathematics, the conceptually necessary and the metaphysically necessary coincide. Ever

since, the hunt has been on for an account of what metaphysical necessity is. More recently,

with the renewed interest in the metaphysical grounding relation, some philosophers have

focused on the question: what are the grounds of facts about metaphysical necessity? One

does not have to be very skeptically inclined to conclude that philosophers as yet have no

fully satisfactory answer to this question. What I take to be the best theory available—that

metaphysically necessary truths are grounded in facts about the essences of things79—is not

without its problems, because it presupposes the notion of essence, and, to be fully

satisfactory, would require some independent way of determining the essences of things

(rather than just packing into them everything you need to ground what you pre-theoretically

thought was necessary).

But this, in many ways, is analogous to Kant’s view about real possibility. We know that

real possibility is conceptually distinct from logical possibility, and we know of some

examples where they come apart: mathematical truths are really necessary but not logically

necessary. But we don’t know on theoretical grounds if this difference in extension carries

beyond the empirical domain: for all we know through theoretical means, logical and real

78 I am not trying to commmit myself to any controversial views in post-Kripkean metaphysics here. By ‘conceptual necessity’ I mean, roughly, what is a priori knowable. I distinguish this from the analytic, because there may be a priori knowable truths that are not analytic (e.g. that nothing is red and green all over, or mathematical truths). For sophisticated recent discussions of conceptual possibility/necessity and its relation to metaphysical possibility/necessity, see the papers by Bealer and Chalmers in Gendler and Hawthorne (2002). 79 See Fine (1994) and (1995).

40

possibility might coincide for non-empirical objects. And most importantly we do not know

what in general grounds facts about real possibilities. We have what Kant claims is the most

rationally satisfactory explanation—real possibilities are grounded in God, who exists

absolutely necessarily—but this is not something we can know through theoretical means to

be true.

So while Kant’s metaphysics of real modality might ultimately be unsatisfying,

contemporary modal metaphysics is not much more satisfying on this question. In fact,

there’s a way of seeing the conclusion of Kant’s views about modality as even more similar

to contemporary metaphysics. Kant thinks there is no way to know what the grounds of real

possibility in general are; in other words, no way to know what the ultimately correct theory

of real possibility is. But he thinks we are rationally required to believe the most rationally

satisfactory explanation, which I sketched earlier. But post-Lewisian metaphysics has largely

abandoned the aim of decisively proving one theory to be correct and its rivals incorrect

(which many metaphysicians now take to be impossible in the case of the main competing

views). Instead, contemporary metaphysicians weigh theories against one another in terms

of large-scale theoretical virtues like agreement with pre-theoretic intuition, explanatory

power, elegance, simplicity, number of primitives, etc.80 It is somewhat implausible to believe

that any modal metaphysician knows her theory to be the correct one, even if her theory

overall is better than her contemporaries’ according to these measures. So we might think of

contemporary metaphysics as trying to determine the most rationally satisfactory theory of

real possibility and its grounds. Where contemporary philosophical practice and Kant differ

—and this is not a small difference—is in their conceptions of the univocity of reason. Kant

tends to view theoretical reason in its supersensible use as making a number of demands that

80 Cf. Lewis (1986), 3-5; and Sider (2001), xv.

41

are uniquely maximally satisfied by a set of hypotheses. If the methodology of

contemporary metaphysics shows one thing, it is that various desiderata on the best

theory—in Kantian terms, various demands of reason—stand in tension with one another,

and no theory can be said to maximally satisfy all of them. Instead, we must weigh various

demands of reason against one another, and which theory comes out the winner at the end

of the day may depend upon such non-Kantian considerations as what use to which we want

to put the theory, or individual preferences, for example, for simplicity over explanatory

power, etc. So really the difference comes down to a difference about reason: Kant thinks

that there is one hypothesis that uniquely satisfies the rational requirements on a modal

theory, while contemporary philosophers think we must balance different rational

requirements against one another.

42

Abbreviations for Works of Kant

A/B Kritik der reinen Verunft, Ak. 3 &4. Cited by page number in 1st edition of 1781 (A) and 2nd edition of 1787 (B). Translations from Kant (1997).

Ak. Kant (1900—). Cited by volume and page number (e.g. Ak. 29:1034). CPR Kritik der reinen Verunft, Ak. 3 &4. Cited by page number in 1st edition

of 1781 (A) and 2nd edition of 1787 (B). Translations from Kant (1997). OPG Der einzig mögliche Beweisgrund zu einer Demonstration des Dasein Gottes, Ak.

2:62-163. Translation: The only possible ground of proof in support of a demonstration of the existence of God in Kant (1992).

Pölitz RT Philosophische Religionslehre nach Pölitz, Ak. 28:989-1126. Translation:

Lectures on the Philosophical Doctrine of Religion in Kant (1996). Volckmann RT Natürliche Theologie Volckmann nach Bambach, Ak. 28: 1127-1226.

Danziger RT Danziger Rationaltheologie nach Bambach, Ak. 28:1227-1319.

Refl. Kants handschriftlicher Nachlass (‘Reflexionen’) in Ak. 14-18. Cited by

four digit number, and volume and page number in Ak. Where no translation is listed, translations are my own.

Works of Kant Cited Kant, I. (1900—). Kants gesammelte Schriften. Edited by the Berlin-Brandenburg (formerly Royal Prussian) Academy of Sciences. Berlin: Walter de Gruyter. -----. (1992). Theoretical Philosophy 1775-1770. Translated and Edited by D. Walford and R. Meerbote. Cambridge: Cambridge University Press. ----- . (1996). Religion and Rational Theology. Translated and Edited by A. Wood and G. di Giovanni. Cambridge: Cambridge University Press. -----. (1997). Critique of Pure Reason. Translated and Edited by P. Guyer and A. Wood. Cambridge: Cambridge Univesrity Press. Other Works Cited

Adams, R.M. (2000). “God, Possibility and Kant.” Faith and Philosophy 17, 425-440. Abaci, U. (forthcoming). “Kant’s Only Possible Argument and Chignell’s Real Harmony.” Kantian Review.

43

Adickes, E. (1924). Kant und das Ding an Sich. Berlin: Pan Verlag.

Allais, L. (2004). “Kant’s ‘One World’: Interpreting Transcendental Idealism.” British Journal for the History of Philosophy 12, 655-684. Allison, H. (1983). Kant’s Transcendental Idealism. New Haven: Yale University Press. (2004). Revised and Enlarged Edition. -----. (1994). “Causality and Causal Laws in Kant: A Critique of Michael Friedman” in Parrini, P. (1994), 291-307. Ameriks, K. (1982). “Recent Work on Kant’s Theoretical Philosophy.” The Philosophical Quarterly 19, 1-24.

-----. (2003). Interpreting Kant’s Three Critiques. Oxford: Oxford University Press. Anderson, R. L. (2004). “It Adds Up After All: Kant’s Philosophy of Arithmetic in Light of

the Traditional Logic.” Philosophy and Phenomenological Research 69, 501-540. Aquila, R. (1979). “"Things in Themselves: Intentionality and Reality in Kant.” Archiv Fur Geschichte der Philosophie 61, 293-307. Bird, G. (1962). Kant's Theory of Knowledge. London: Routledge & Kegan Paul. Beiser, F. (2002). German Idealism: The Struggle against Subjectivism 1781-1801. Cambridge, MA: Harvard University Press. Chignell, A. (2007a). “Belief in Kant.” The Philosophical Review 116(3), 323-360. -----. (2007b). “Kant’s Concepts of Justification.” Noûs 41(2), 33-63. -----. (2009). “Kant, Modality, and the Most Real Being.” Archiv für Geschichte der Philosophie 91 (2),157-192. ----- (2010). “Real Repugnance and Belief about Things in Themselves: A Problem and Kant’s Three Solutions.” In Krueger, J. and Lipscomb, B. (Eds). Kant's Moral Metaphysics. Berlin: Walter DeGruyter. ----- (2012). “Kant, Real Possibility, and the Threat of Spinoza.” The Monist 121 (483), 635-675. Erdmann, B. (1878). Kants Kriticismus in der ersten und in der zweiten Auflage der Kritik der reinen Vernunft. Leipzig: Voss. Fine, K. (1994). “Essence and Modality.” Philosophical Perspectives. Vol. 8: Logic and Language, 1-16.

44

----- (1995). “Senses of Essence.” in Sinnott-Armstrong, W., Raffman, D., and Asher, N. (Eds). Modality, Morality and Belief: Essays in Honor of Ruth Barcan Marcus. Cambridge: Cambridge University Press, 53-73.

Gendler, T., and Hawthorne, J. (eds). Conceivability and Possibility. Oxford: Clarendon Press. Hanna, R. (2006). Kant, Science, and Human Nature. Oxford University Press. Langton, R. (1998). Kantian Humility. Oxford: Oxford University Press. Lewis, D. (1986). On the Plurality of Worlds. Cambridge: Cambridge University Press. Parrini, P. (Ed). (1994). Kant and Contemporary Epistemology. Dordrecht: Kluwer.

Pereboom, D. (2007). “Kant on Transcendental Freedom.” Philosophy and Phenomenological Research 73, 537-567.

Prauss, G. (1974). Kant und das Problem der Dinge an Sich. Bonn: Grundman. Schafer, K. (forthcoming). “Kant’s Conception of Cognition.” In Schafer, K., and Stang, N. The Sensible and Intelligible Worlds: New Essays on Kant’s Metaphysics and Epistemology. Oxford: Oxford University Press.

Sider, T. (2001). Four Dimensionalism: A Study in Persistence and Ontology. Oxford: Oxford University Press.

Stang, N. (2010). “Kant’s Possibility Proof.” History of Philosophy Quarterly 27, 275-299. -----. (2011). “Did Kant Conflate the Necessary and the A Priori?” Noûs 45:3, 443–471. -----. (2012). “Kant on Complete Determination and Infinite Judgment” British Journal of the History of Philosophy 20:6, 1117-1139. -----. (forthcoming-a) “Kant’s Argument that Existence is not a Determination.” Philosophy and Phenomenological Research.

-----. (forthcoming-b) “The Non-Identity of Appearances and Things in Themselves.” Noûs. -----. (forthcoming-c). “Kant’s Transcendental Idealism.” Stanford Encyclopedia of Philosophy.

Watkins, E., and Fisher, M. (1998). “Kant on the Material Grounds of Possibility.” Review of Metaphysics 52, 369-395. Wood, A. (1984). “Kant’s Compatibilism.” In Wood, A. (Ed). Self and Nature in Kant’s Philosophy. Ithaca: Cornell University Press. Yong, P. (forthcoming). “God, Totality, and Possibility in Kant’s Only Possible Argument.” Kantian Review.

45