¹ I have identified Kant's individual works using the following abbreviations:

CPrR: Critique of Practical Reason

CPR: Critique of Pure Reason

MFNS: Metaphysical Foundations of Natural Science

OP: Opus Postumum

All Kant citations refer to the Akademie edition of Kants Gesammelte Schriften (Berlin: G. Reimer [now de Gruyter], 1902), 29 vols. Abbreviated Ak. The only exception is CPR, for which I use the A/B edition notation. All quotes from CPR are taken from Immanuel Kant, The Critique of Pure Reason, trans. P. Guyer and A. Wood (Cambridge, Cambridge University Press, 1997).

² Vaihinger, Hans, Commentar zu Kants Kritik der reinen Vernunft Vol. II (Stuttgart: W. Spemann, 1892), p. 53.Google Scholar

³ Most commentators make the same assumption and their interpretations break down into two main camps. The first defends a ‘one-world’ interpretation of Kant's distinction between appearances and things-in-themselves. For example, Henry Allison argues that one can consider the same object from two different standpoints or aspects, both as appearance as well as thing-in-itself. See Allison, Henry, Kant's Transcendental Idealism: An Interpretation and Defense, second edition, (New Haven: Yale University Press, 2004), pp. 52f.CrossRefGoogle Scholar The second camp defends a ‘two-world’ interpretation of Kant where appearances and things-in-themselves are ontologically distinct from one another. For example, James van Cleve argues for a virtual object interpretation of Kantian appearances whereby they are constructed from noumenal subjects and their cognitive acts. See van Cleve, James, Problems from Kant (Oxford: Oxford University Press, 1999)Google Scholar, especially pp. 5f and 150f. Although Langton adopts a one-world view, she understands appearances not as particulars but rather as relations.

⁴ It is important to distinguish between the different senses of ‘property’ and ‘relation’ that I will be using. Monadic or non-relational properties are one place properties (e.g. X is simple). When intrinsic, these properties are essential to the objects that have them. Relational properties are multi-place properties (e.g. X is to the left of Y) which may or may not be essential to the objects so related. Extrinsic relations are inessential to the objects so related. Intrinsic relations are, however, essential to the objects so related. What this means, for my purposes, is that not all relations are extrinsic, nor are all intrinsic properties non-relational. As I will argue below, appearances are both relational (multi-place properties relating phenomenal objects and subjects) and intrinsic (essential to both phenomenal objects and subjects).

⁵ Langton, Rae, Kantian Humility: Our Ignorance of Things in Themselves (Oxford: Oxford University Press, 1998), p. 20. Although Langton also talks about phenomena as things-in-themselves in relations to other things, for the sake of simplicity I will use the locution of ‘relational property’ or ‘relation’ when describing phenomena (or appearances as a subset of phenomena). It is important to note, however, that phenomena are always relational properties of things-in-themselves for Langton.Google Scholar

⁶ See CPR A274/B330.

⁷ Langton, Kantian Humility, p. 12.

⁸ Langton spends some time discussing two arguments against the reducibility of relations. Although she admits that there are problems with both arguments, she adopts the version that rejects the bilateral reducibility of relations where the latter are understood as causal powers (i.e. causal powers that are irreducible to the intrinsic properties of the relata of causal relations), given the fact that God could have chosen natural laws that would not allow for causal powers even if the intrinsic properties of the relata were held fixed. See ibid., chapter 5.

⁹ Ibid., p. 19.

¹⁰ Ibid., p. 23.

¹¹ Ibid., p. 13.

¹² See ibid., p. 48f.

¹³ CPR A146–7/B186.

¹⁴ See Allison, Kant's Transcendental Idealism, pp. 17f. Critics have posed a number of objections to Langton's humility thesis. For example, and in contrast to Allison, Van Cleve believes that Langton's thesis is far too humbling arguing that it precludes knowledge even of our own mental content. See van Cleve, James, ‘Receptivity and Our Knowledge of Intrinsic Properties’, Philosophy and Phenomenological Research, 65 (1) (2002), 218–37, p. 232.CrossRefGoogle Scholar

¹⁵ CPR A51/B75.

¹⁶ Langton, Kantian Humility, p. 13.

¹⁷ See ibid., chapter 1.

¹⁸ Ibid., p. 13.

¹⁹ For the logical table of judgements, see CPR A70/B95.

²⁰ Ibid. A147/B186–7. Kant gives an example using the unschematized category of substance which he characterizes as signifying nothing more than the relation between subject and predicate, i.e. the first relation of thinking in a judgement from the logical table of judgements.

²¹ In CPrR, Kant suggests that practical reason can use the unschematized category of causation to generate knowledge claims which cannot be made from the perspective of theoretical reason alone, e.g. concerning the practical efficacy of the will. See CPrR, Ak. 5: 50–7. This addendum cannot help Langton, however, in so far as she is operating strictly within the realm of theoretical reason.

²² CPR A20/B34.

²³ Ibid., B66–7. See also A251–2, A265/B321, B306–8, and A284–5/B340–1.

²⁴ See Langton, Kantian Humility, chapter 9. In the third section of this paper, I will argue that the thing-in-itself can indeed be viewed as the ground of appearance though only as the logical ground.

¹⁵ Ibid., A264/B320. See also A44–5/B61–2.

²⁶ Ibid., A264/B320.

²⁷ Ibid., A255/B310.

²⁸ Allison poses this objection to double affection in Allison, Kant's Transcendental Idealism, p. 66.

²⁹ It is important to note that outer relations need not be extrinsic relations. The concept of an extrinsic relation has nothing to d o with spatial character. As I will argue below, these outer relations are, in fact, intrinsic relations.

³⁰ CPR A265/B321. For Kant's explicit discussion of ’substantiae noumena’ see A276/B332.

³¹ Ibid., A277/B333.

³² Ibid., B306.

³³ Ibid., A265/B321.

³⁴ See also ibid., A166. In the MFNS, Kant talks about phenomenal objects affecting us through their motion rather than their moving forces. See MFNS, Ak. 4: 476. Even so, in the OP, Kant returns to the idea that phenomenal substances affect the subject through moving forces. For example, see OP, Ak. 22: 300, 22: 318, and 22: 458.

³⁵ CPR A20/B34.

³⁶ Ibid., A276/B332. See also A267/B323.

³⁷ Appearances can be viewed as relations in a two-fold sense: (1) Appearances are ordered sets of causal-structural relations that obtain, in the phenomenal world, between affecting objects and receptive cognitive subjects. Affection is then one relational element within this ordered set of causal-structural relations, whose specific function it is to reveal structural features of the affecting object to the receptive subject. (2) Appearances are spatiotemporal relations that obtain between phenomenal objects (including embodied cognitive subjects). Spatiotemporal relations in this sense are a necessary condition for the set of causal-structural relations, in so far as receptive cognitive subjects can only be affected by objects in space and time.

³⁸ Robert Hanna describes this position as ‘weak transcendental idealism’. This is the position that space and time can exist in a possible world (including the actual world) even if no cognitive subjects like us existed in that world provided that if cognitive subjects like us existed in this world they could correctly represent space and time. See Hanna, Robert, Kant, Science, and Human Nature (Oxford: Oxford University Press, 2006)CrossRefGoogle Scholar, chapter 6. One might still push the modal issue, however, claiming that, under my view, it makes little sense to talk about actual objects without reference to actual subjects (e.g. the actual world before cognitive subjects) since actual objects must bear intrinsic relations to actual subjects and vice-versa. At this point, I think it is important to note Kant's rather unique way of understanding modality. For Kant, the domains of the necessary, actual, and possible are coextensive and the difference between these modalities is simply intensional. The modal categories have different meanings in so far as they describe different cognitive relations or ways of considering objects. Consequently, regardless of whether you are talking about the necessity, actuality, or possibility of a given object, you are always presupposing some relation to the subject and, for Kant, it makes little sense to consider the modal status of some object without reference to some subject doing the considering. In the above example, Kant would say that one can consider actual objects in the distant past prior to the existence of cognitive subjects via the connection that the former have to perception in accordance with natural law. This is the upshot of Kant's Second Postulate of Empirical Thought describing the category of actuality. For Kant's discussion of the coextensive nature of the modal categories and their intensional differences see CPR A230–5/ B282–7. See also Sebastian Gardner's discussion of the modal categories in Gardner, Sebastian, Kant and the Critique of Pure Reason (New York: Routledge, 1999), pp. 197Google Scholarf.

³⁹ See Langton, Kantian Humility, chapter 2. One of the passages that Langton uses to support the idea that substances must have intrinsic non-relational properties is CPR A274/B330 from the Amphiboly section. An interpretative difficulty with this section, however, is that it is often hard to disentangle Kant's view from the Leibnizian view he is criticizing. When it comes to the passage that Langton quotes, Graham Bird argues that Langton mistakes the Leibnizian position that Kant rejects for Kant's own position. See Bird, Graham, ‘Review of Kantian Humility’, The Philosophical Quarterly, 50 (198) (2000), 105–8, p. 106.Google Scholar

⁴⁰ For different contexts in which Kant makes use of empirical affection, see CPR B124–5, A213/B260, and B520. One of the most thorough examinations of Kant's views on empirical affection can be found in Adickes, Erich, Kants Lehre von der doppelten Affection unseres Ich: als Schlüssel zu seiner Erkenntnistheorie (Tuebingen: Mohr, 1929), chapter 1.Google Scholar

⁴¹ CPR A144/B183.

⁴² Ibid., A182.

⁴³ Ibid., A210/B257–8.

⁴⁴ This real ground should be distinguished from the logical ground of appearance. I will return to the latter in the third section of this paper.

⁴⁵ One might even want to go further and claim that appearances for Kant are super-intrinsic relations, since there are some cases where a relation can be essential to the relata in so far as each relatum would lack some essential property without the other but the relatum would still exist though without this property. A good example of this might be Robert Boyle's discussion of the first lock and key. Put simply, he claims that without the existence of the key, the lock is not really a lock, but rather just a hunk of metal organized in a certain way. The same thing goes for the key. See Boyle, Robert, ‘The Origin of Forms and Qualities According to the Corpuscular Philosophy’, in Stewart, M. A. (ed.), The Selected Philosophical Papers of Robert Boyle (Indianapolis: Hackett, 1991), pp. 1–96Google Scholar, esp. p. 23. An example of a non-philosophical super-intrinsic relation might be the relation that obtains between points and line segments in the first postulate of Euclid's Elements. Whereas a straight line segment is uniquely determined by its two endpoints, these points are uniquely determined by the straight line that is drawn between them. Even so, a point is not reducible to a line or vice-versa. Although Euclid has a primitive definition of ‘point’ as ‘that which has no parts’ and ‘line’ as ‘length without breadth’, the geometrical objects that these terms denote are completely indeterminate without reference to one another. See Euclid, , The Elements of Euclid, trans. Simson, R. (Philidelphia: Desilver, 1821), Book I, pp. 7 and 11.Google Scholar

⁴⁶ Hume, David, A Treatise of Human Nature, Nidditch, P. H. (ed.) (Oxford: Oxford University Press, 1978), p. 636.Google Scholar

⁴⁷ CPR A226–7/B279–80.

⁴⁸ Berkeley's view on existence can be interpreted in two different ways: (1) to be is to be perceived or to be a perceiver, or (2) to be is to be capable of being perceived or to be capable of perceiving. The former interpretation is suggested by Berkeley's discussion in the Dialogues where he claims that all sensible things must be perceived by God in order to exist. See Berkeley, George, Three Dialogues between Hylas and Philonous, in Luce, A. A. and Jessop, T. E. (eds), The Works of George Berkeley Bishop of Cloyne (London: Thomas Nelson and Sons, 1949), vol. II, pp. 212–13 (11.18). The latter interpretation is suggested by Berkeley's discussion in the Principles where he claims that sensible things must only be capable of being perceived in order to exist (e.g. that the table in his study exists since it would be perceived if he were in his study to perceive it). See George Berkeley, A Treatise Concerning the Principles of Human Knowledge from The Works of George Berkeley Bishop of Cloyne, vol. II, Part I, §3. If nothing else, the second interpretation would pose problems for Berkeley's argument for God via the first interpretation. In any case, for the purposes of this paper, I assume the first interpretation.Google Scholar

⁴⁹ See Berkeley, Principles, Part I, §25.

⁵⁰ See ibid., §27–32. According to Berkeley although we can produce ideas of imagination through our own will, the ideas of sense are produced by the will of God. In either case, however, the ideas themselves are causally inert.

⁵¹ Van Cleve, Problems from Kant, p. 170. It is important to note that although there is ’stuffing’ according to Langton (intrinsic non-relational properties), this stuffing is irrelevant to the ‘structure’ (relational properties) in so far as the relational properties are irreducible to the intrinsic non-relational properties of the objects. Consequently, there is a certain way in which phenomenal objects (relational properties) are all structure without stuffing under Langton's view as well.

⁵² CPR A248–9.

⁵³ Even if one admits that appearances are intrinsic relations between phenomenal objects and subjects, are there relations that phenomenal objects bear to one another independently of the relations they bear to phenomenal subjects? Although phenomenal objects are related spatially and causally to one another, phenomenal objects underdetermine the relationships they bear to one another. In order for these relationships to be determined, the subject must make judgements concerning these objects (e.g. causal). Consequently, determining the relations between phenomenal objects requires the subject's synthetic activity, much as determining the phenomenal objects themselves requires this synthetic activity.

⁵⁴ Although the view may seem paradoxical, this has not stopped Langton's critics from entertaining it. For example, A. W. Moore considers the possibility that something could possess only relational properties. See Moore, A. W., ‘Review of Kantian Humility’, Philosophical Review, 110 (1) (2001), 117–20, p. 118.CrossRefGoogle Scholar

⁵⁵ CPR B146–7.

⁵⁶ Ibid., A105.

⁵⁷ Ibid., A109.

⁵⁸ Ibid., B133.

⁵⁹ See also Sebastian Gardner's discussion of how subject and object make one another possible in Gardner, Kant and the Critique, pp. 157f.

⁶⁰ A recurring criticism of Langton's position is how it at best ignores, or at worst is inconsistent with, Kant's transcendental idealism. For example, see Moore, ‘Review of Kantian Humility’, p. 118. Whereas Langton downplays the formal contribution of the subject to the objects of experience and commits Kant to substantive claims beyond the bounds of sense, my position emphasizes the formal contribution of the subject in determining the objects of experience while not committing Kant to any substantive claims beyond the bounds of sense. It is thoroughly consistent with Kant's transcendental idealism while also overcoming the trilemma that faces Kant's theory of affection.

⁶¹ See, for example, CPR Bxxvi, A249, A251–2.

⁶² Bird, ‘Review of Kantian Humility’, p. 105.

⁶³ CPR A372.

⁶⁴ Ibid., Bxxvi-ii. See also A190/B235, A280/B336, and A283/B339.

⁶⁵ Ibid., Bxxvi.

⁶⁶ Ibid., A227/B280.

⁶⁷ There are several places where Kant seems open to analytic judgments involving the concept of a thing-in-itself, e.g. ibid., B149, A433/B461, and A635/B664. Even so, as Kant makes clear in his criticisms of the ontological argument, such analytic judgments cannot have existential import. See A597/B625 as well as his comments at the end of the Phenomena/Noumena section at A258–9/B314–5.

⁶⁸ Ibid., A251. Gerd Buchdahl also recognizes the conceptual nature of the transcendental object. See Buchdahl, Gerd, Kant and the Dynamics of Reason (Oxford: Blackwell, 1992), p. 161.Google Scholar

⁶⁹ CPR A288/B344. See also A372.

⁷⁰ Ibid., A288/B345.

⁷¹ Allison, Kant's Transcendental Idealism, p. 63.

⁷² CPR A250–252. For other contexts in which Kant uses ‘transcendental object’ and ‘noumenon’ interchangeably, see CPR A358 and A288/B344–5. It is important to note that Kant claims in the paragraph immediately following the above quoted passage that the transcendental object cannot be called a noumenon. Buchdahl overcomes the seeming contradiction by noting that the distinction between negative and positive noumena is only implicit in this section of the A-edition. Once one has this distinction in mind, however, there is little problem seeing how the transcendental object could be equivalent to the negative noumenon while certainly different from a positive noumenon. See Buchdahl, Dynamics of Reason, 84f.

⁷³ In the B-edition, Kant characterizes the negative noumenon as ‘a thing in so far as it is not an object of our sensible intuition because we abstract from the manner of our intuition of it’. See CPR B307. This might suggest that the object has a way that it is or some transcendental reality independent of the way in which we intuit it. It is important to note that this is an act of abstraction, however, and so results only in the concept of an object in general considered independently of any particular features in sensibility. This concept need not have a referent, however, in order to function as a concept. In fact, Kant describes the negative noumenon in precisely this way before the passage quoted above. See ibid., B306.

⁷⁴ See Allison, Kant's Transcendental Idealism, pp. 64f.

⁷⁵ Allison, Kant's Transcendental Idealism, p. 55.

⁷⁶ Considering Allison's criticisms of Langton's argument, it is ironic that he would adopt the first horn while rejecting the second. By adopting the first horn, he opens himself up to the same objections that he levels against Langton's own position. For his explicit endorsement of the first horn and rejection of the second see Allison, Kant's Transcendental Idealism, pp. 68 and 73.

⁷⁷ I would like to thank Robert Hanna, Rae Langton and Walter Ott for their questions and comments on this paper. Thanks also to audiences in Milwaukee and Baltimore for their helpful suggestions on versions of this paper. Finally, thanks to the anonymous referees and editor of this journal whose comments were of great value in revising this paper for publication.