I largely agree with Dennis Schulting’s book, Kant’s Radical Subjectivism, so I will not be challenging any of his main claims. However, a number of things that he does not discuss or discusses only very briefly either provide additional support for his views or challenge them. It may be a bit begrudging to ask for even more in a book that as it stands makes a rich and detailed contribution to the study of Kant’s ‘theoretical’ philosophy so perhaps the points that I will raise are best thought of as areas of possible future work.
1. Agreements, and Discussions Related to them that I would have Welcomed
In my view, it should not be controversial that Kant was a radical subjectivist in Schulting’s sense, that is, ‘that Kant’s Critical project is concerned … to formulate a manner in which a rational structure can be discerned … within experience, hence in judgement – which … determines that relationship as … objectively valid’ (p. 11). However, this view or an even more radical subjectivism seems to flow immediately from Kant’s view that space and time are ‘forms of intuition’, that is to say, forms of organization that the mind imposes on sensory input, and are not properties of mind-independent reality. Yet these claims about space and time do not make an appearance until late in the book. Likewise, it seems to flow from Kant’s view of matter in at least the A-edition: ‘Matter … is only a species of representation’ (A370).Footnote
1
(And the B-edition? Because of the Refutation of Idealism new to that edition, the issue is more complicated. See below.)
Schulting does a fine job of reconciling Kant’s claim that beliefs can be objectively valid with the ‘Copernican turn’ that Kant introduces in the B-edition Preface, with, that is to say, his claims that our beliefs are largely the product of our judgemental, concept-applying activities and that we are aware of nothing more than our own representations. (The latter is the essence of Kant’s subjectivism for Schulting, as we just saw.) However, this has to result in a concept of objective validity that is, to an analytic philosopher anyway, strange. For Kant, objective validity cannot be a matter of getting the mind-independent world right and must become something to do with something being out of the mind’s control. Schulting talks about objective validity in the early going but I would have welcomed a fuller account.
In my view, Schulting is right that Kant held that there is a non-conceptual component to beliefs and the other results of judgement. Indeed, a non-conceptual component enters into belief-formation and plays a crucial role in adjudicating which beliefs are warranted, even have content, and which beliefs are not and maybe do not, or so the famous saying ‘thoughts without concept are empty’ (A51/B75) would suggest. Content is clearly provided by intuitions. Two things about this.
-
1. It is not clear that Schulting wants to go this far – he limits non-conceptualized content to ‘content that is not conceptualized or even subject to conceptualization’. This would restrict non-conceptualized content to what Kant called dark representations (dunkelen Vorstellungen), that is to say, representations of which we are not aware. In CPR, Kant mentions dark representations only twice, both times in obscure footnotes and very briefly (A117n., end; B415n.), though he discussed them many times in his lectures and other more popular works (Brook Reference Falkensteinforthcoming). He seems to have thought that they play no important role in unified conscious knowledge. However, a non-conceptual component of some kind does, or certainly seems to do so, as we just saw.
-
2. Kant may have changed his mind about the role of non-conceptual knowledge from the A- to the B-edition. In particular, §26 of the B-edition Transcendental Deduction (TD) contains an argument that even space and time have to be synthesized, presumably using concepts.
Schulting treats Kant’s views of interest to him as changing little from the A-edition TD to the B-edition version and cites the two versions interchangeably. Yet Kant completely rewrote the chapter for the B-edition and many things changed. The threefold doctrine of synthesis almost disappeared – though arguably the new figurative or productive imagination is part of the old synthesis of recognition in a concept under new names. As Schulting clearly recognizes, the starting point of the deduction changes, at least if one takes §16 to be the starting point. (Like many commentators, Schulting gives short shrift to §15.) And most importantly for Schulting’s interpretation, objectivity, i.e. objective validity, is much more prominent. We need an argument that the A- and B-edition versions of TD take the same line on this topic. On the subject of Kant changing his mind, however, the new Refutation of Idealism is even more important. It is hard to see how some of the things that Kant says there are consistent with either version of TD. The relationship of this passage, which Schulting does not discuss, to TD is one of the issues on which I will focus in the remainder of these comments.
Finally for these introductory comments, if the mind contributes as much to its experiences, beliefs and the like as Kant suggests, including even their spatial and temporal structure, it is hard to see how the ‘manifold of intuition’ could play any role in controlling the truth-status of experiences, as Kant insists that it must. If on arrival the manifold of intuition has no temporal or spatial structure and no properties that the categorial concepts name, what structure could it have? Yet without such structure, how could it guide us in determining which propositions and other experiences to accept and which to reject as expressing ‘capricious and incongruous fictions’ (A96, A376)? When this concern focuses on spatial and temporal structure, it is often called the Localization Problem. If the Refutation of Idealism is primarily a problem for Schulting’s assumption that Kant’s views were consistent across the A- and B-editions of CPR, the localization problem is primarily a problem for Kant, though a complete account of his picture of knowledge acquisition must address it. Since not only Schulting but most commentators on Kant on knowledge do not address it (Falkenstein Reference Falkenstein1995 is a partial exception), we will take it up first.
2. Localization
The localization problem in its full generality goes like this. If we do not distribute the manifold of intuitions in space and time and apply concepts to them by virtue of spatial, temporal and property structures which they have on arrival, how do we do so? The slightly less general version of the problem was just discussed: in virtue of what can we decide that some orderings and applications of concepts are warranted, others not? We cannot localize and conceptualize elements in the manifold of intuition as we please – this is part of what Kant had in mind when he said that we are passive or merely receptive to intuitions. So what is it that controls, what could possibly control, how we do so?
For Kant, localization is a severe problem. Once space, time and concepts are separated from intuitions, it seems that neither perceptual pathways or filters nor acts of synthesis nor anything else would have the information needed to distribution elements of intuitions across space, time and concepts. (Kant once recognized the need for ‘reliable information’ (A60/B85) but did nothing with the idea.) There would have to be something else in the incoming intuitions that our perceptual system uses to distinguish them. (Falkenstein Reference Falkenstein1995: 85 makes a related point.) Otherwise, to insist, as Kant does, that intuitions have nothing spatiotemporal or conceptual about them on arrival would be an a priori proof of the impossibility of spatial, temporal and conceptual localization. There must be something else in intuitions as they arrive that resists some forms of organization and facilitates others. This is an especially acute problem for time and concepts. We might be able to use, say, shades and saturations to organize particulars in space but it is hard to think of any non-temporal property from which we could infer temporal location or anything both non-temporal and non-conceptual which could guide us in the application of concepts.
In fact, the localization problem is even harder for Kant than this, harder indeed than it would be, or need be, for other representational idealists. Not only are temporal and spatial properties conferred by the mind, space and time themselves are very different from anything that a particular contingent representation could have. Time and space are ‘pure forms of intuition’, which means that they are a priori in their origins, are known a priori (prior to, or at least independently of, experience) and are a priori in their modal status (they are features of how experience must be and confer necessity on some bodies of spatial and/or temporal knowledge); for these distinctions, see Brook Reference Brook1992. Being a priori in these ways makes time and space very different from particular, contingent properties, even temporal or spatial ones. It is not just that the forms exist prior to experience whereas particular intuitions are received by the senses. As the modal status of space and time make clear, the difference is much bigger than that. The forms of intuition allow us to discern necessity about the spatial (geometry: B16–17) and the temporal (in mechanics: B17–18; arithmetic: A103/B15–16, A103; and algebra: A717/B745, A734/B702). For Kant, ‘experience tells us … what is, but not that it necessarily must be so’ (A1). For that we need something a priori. All this makes time and space so different from particular intuitions that it is not easy to see how the latter could come to be located in the former at all. A cognitive system whose form is enough by itself to establish necessary truths will have a very special relationship to a posteriori empirical sensory inputs.
The issue of localization has received little attention in the literature on Kant and in this Schulting not an exception. As we saw, Falkenstein takes it up (briefly): he urges that Kant had to have been a realist about the time-order of representations because otherwise he would have had ‘no way to explain why we place matters in one location rather than another’ (Falkenstein Reference Falkenstein1995: 85; see also p. 250). Falkenstein takes the matter no further. Schulting and other commentators do not even get that far.
What about Kant himself? Usually the issue is deep in the background in his writings and unacknowledged. Sometimes it creeps close to the surface, as in: ‘the manifold to be intuited must be given prior the synthesis of the understanding and independently of it’; unfortunately, he then merely says, ‘How this takes place remains here undetermined’ (B145). Interestingly, he did walk right up to it at least once. (Whether he saw that he was doing so is another question.) In subsection II of §8 of the Aesthetic, the first sub-section of new material that he added near the end of the chapter for the B-edition, in the course of discussing what is presented by intuitions, he says this:
intuition … contains nothing but relations, namely, of locations in an intuition (extension), of change of location (motion), and of laws according to which this change is determined (moving forces). (B66–7)
How he could not then ask, ‘And how do items end up in the locations which we represent them as having? What controls where we represent an item as located?’, is difficult to fathom. But he did not. This passage comes the closest of anything that I know of in his writings to identifying and addressing the localization problem.
One response to Kant’s silence on localization would be to say that he was an idealist who held that one cannot be aware of anything outside the content of one’s own mind, so it is not surprising that he has nothing to say about what controls the element in our mental repertoire to which we are passive, what controls location in time, space and concepts. This response would be problematic – Kant was also an empirical realist and believed that we have a number of bodies not just of knowledge but of universal and necessary knowledge (B4). Physics and mathematics are the two main examples. So he cannot avoid the question, ‘In virtue of what that is non-temporal, non-spatial and non-conceptual would the way in which a physical or mathematical judgement temporalizes, spatializes, and conceptualizes make it an item of knowledge or a false belief?’ That said, if he has an answer better than the inadequate coherentism of A376, I do not know what it was.
As I said, a complete account of Kant’s view of knowledge acquisition needs to address the localization problem. Given the remarkably interesting claims that Schulting makes about other aspects of Kant on knowledge, I hope that at some point he turns to it.
3. Refutation of Idealism
Most commentators believe that Kant added the Refutation of Idealism to the B-edition to refute the charge that he was a Berkeleian idealist, a charge that his attack on the fourth Paralogism in the A-edition particularly encouraged. It may refute too much. The claims that Kant makes in the passage may undermine his own brand of idealism – and Copernican constructivism – too. This would make it a problem for Schulting. At minimum, his story would not be the whole story about Kant on knowledge. What is going on in the Refutation overall and how the passage relates to the rest of CPR is complicated and I have explored these issues elsewhere (Brook Reference Brook2011). Here I will restrict myself to the implications of the Refutation for the ideas that temporal, spatial and conceptual order are given to experience by the mind and that all we are aware of are experiences so structured.
The key element is a new doctrine of matter. In the first edition, Kant distinguished between matter as a feature of appearances – a feature that consists of the objects of these appearances having extension, impenetrability, cohesion and motion (A358) – and things as they actually are (A268/B324):
Matter is with [the transcendental idealist], therefore, only a species of representations (intuition), which are called external, not as standing in relation to objects in themselves external, but because they relate perceptions to the space in which all things are external to one another, while yet the space itself is in us. (A370)
What matter really is, what ‘inwardly belongs to it’ (A277/B333; I will not go into what Kant meant by this Leibnizian term ‘inward’), is hidden. All I can be aware of are its outer relations, its effects on my representations. This view is also expressed in the Appendix on the Amphiboly of Pure Reason and is a strong form of subjectivism and of the doctrine of the unknowability of the noumenal.
In the Refutation, something very interesting happens. Having argued that we must have immediate awareness of something permanent that is not just outside us but other than ourselves (‘an immediate consciousness of other things outside me’: B276, my emphasis) – something denied by implication in the passage quoted just above – Kant says in Note 2 that ‘we have nothing permanent … save only matter’ (B278, his emphasis). He gives objects on the earth as examples of the permanent, saying that we can see the sun move, for example, by comparing it to their permanence. If the permanent things other than oneself are matter, the sun and the earth, for example, and one is aware of the sun and the earth, then one is ‘immediately aware’ of things other than oneself. To our great frustration, that is all that Kant says but it is enough to indicate that he at least floated the idea here that matter is both not a feature of oneself and knowable.Footnote
2
On this new view of matter, Kant could not continue to hold that time, space and any order that can be captured in concepts have no extra-mental existence, though he may well not have seen this. He could retain the idea that we impose a temporal, spatial and conceptual matrix on our intuitions. However, now matter is not part of the mind and actually has temporal, spatial and conceptual order. It is not just represented as having these properties. If that is now Kant’s view and we put it together with the new realism about awareness of permanent objects other than oneself that we saw Kant advancing, he is now espousing not idealism but realism about temporal, spatial and conceptual order in at least some things other than oneself.
Such realism would also have implications for localization. If things other than oneself have temporal, spatial and conceptual order, then those kinds of order are there to be ‘transduced’ into the experienced raw manifold of intuition. And if that happened, then those orders would be in the raw manifold ready to be read off it when we come to locate items of the manifold in time, space and concepts.
In short, the realism of the Refutation seems to be in serious tension with especially the A-edition of CPR. One commentator who takes the tension between idealism and realism in the B-edition seriously is Guyer (Reference Guyer1987: chapters 12 and 14, especially pp. 282 and 327–8; Guyer also documents a general tension between idealism and realism in CPR and later work in those chapters). Yet even Guyer does not see the Refutation as in tension with transcendental idealism. Instead, he suggests, Kant was claiming that we must presuppose, in our representational constructions, that there are ‘external objects’ (he must mean objects other than oneself), but that Kant continued to maintain that representations never present the objects being presupposed. Guyer claims that Kant’s new position is ontological realism with epistemological idealism. In my view, this response is too easy. Kant is clearly espousing epistemological realism of some kind as well – as the quotation above from B276 demonstrates. I hope that Schulting takes up the relationship of the Refutation to the rest of CPR in future work.
There is a great deal in Schulting’s rich and generative book that I have not discussed, including the applications of his general interpretative scheme to particular issues in TD, the subject matter of his later chapters. I hope that nonetheless these comments inspire some further thoughts – in him and in others – about the themes in Schulting’s book that I have discussed.
I largely agree with Dennis Schulting’s book, Kant’s Radical Subjectivism, so I will not be challenging any of his main claims. However, a number of things that he does not discuss or discusses only very briefly either provide additional support for his views or challenge them. It may be a bit begrudging to ask for even more in a book that as it stands makes a rich and detailed contribution to the study of Kant’s ‘theoretical’ philosophy so perhaps the points that I will raise are best thought of as areas of possible future work.
1. Agreements, and Discussions Related to them that I would have Welcomed
In my view, it should not be controversial that Kant was a radical subjectivist in Schulting’s sense, that is, ‘that Kant’s Critical project is concerned … to formulate a manner in which a rational structure can be discerned … within experience, hence in judgement – which … determines that relationship as … objectively valid’ (p. 11). However, this view or an even more radical subjectivism seems to flow immediately from Kant’s view that space and time are ‘forms of intuition’, that is to say, forms of organization that the mind imposes on sensory input, and are not properties of mind-independent reality. Yet these claims about space and time do not make an appearance until late in the book. Likewise, it seems to flow from Kant’s view of matter in at least the A-edition: ‘Matter … is only a species of representation’ (A370).Footnote 1 (And the B-edition? Because of the Refutation of Idealism new to that edition, the issue is more complicated. See below.)
Schulting does a fine job of reconciling Kant’s claim that beliefs can be objectively valid with the ‘Copernican turn’ that Kant introduces in the B-edition Preface, with, that is to say, his claims that our beliefs are largely the product of our judgemental, concept-applying activities and that we are aware of nothing more than our own representations. (The latter is the essence of Kant’s subjectivism for Schulting, as we just saw.) However, this has to result in a concept of objective validity that is, to an analytic philosopher anyway, strange. For Kant, objective validity cannot be a matter of getting the mind-independent world right and must become something to do with something being out of the mind’s control. Schulting talks about objective validity in the early going but I would have welcomed a fuller account.
In my view, Schulting is right that Kant held that there is a non-conceptual component to beliefs and the other results of judgement. Indeed, a non-conceptual component enters into belief-formation and plays a crucial role in adjudicating which beliefs are warranted, even have content, and which beliefs are not and maybe do not, or so the famous saying ‘thoughts without concept are empty’ (A51/B75) would suggest. Content is clearly provided by intuitions. Two things about this.
1. It is not clear that Schulting wants to go this far – he limits non-conceptualized content to ‘content that is not conceptualized or even subject to conceptualization’. This would restrict non-conceptualized content to what Kant called dark representations (dunkelen Vorstellungen), that is to say, representations of which we are not aware. In CPR, Kant mentions dark representations only twice, both times in obscure footnotes and very briefly (A117n., end; B415n.), though he discussed them many times in his lectures and other more popular works (Brook Reference Falkensteinforthcoming). He seems to have thought that they play no important role in unified conscious knowledge. However, a non-conceptual component of some kind does, or certainly seems to do so, as we just saw.
2. Kant may have changed his mind about the role of non-conceptual knowledge from the A- to the B-edition. In particular, §26 of the B-edition Transcendental Deduction (TD) contains an argument that even space and time have to be synthesized, presumably using concepts.
Schulting treats Kant’s views of interest to him as changing little from the A-edition TD to the B-edition version and cites the two versions interchangeably. Yet Kant completely rewrote the chapter for the B-edition and many things changed. The threefold doctrine of synthesis almost disappeared – though arguably the new figurative or productive imagination is part of the old synthesis of recognition in a concept under new names. As Schulting clearly recognizes, the starting point of the deduction changes, at least if one takes §16 to be the starting point. (Like many commentators, Schulting gives short shrift to §15.) And most importantly for Schulting’s interpretation, objectivity, i.e. objective validity, is much more prominent. We need an argument that the A- and B-edition versions of TD take the same line on this topic. On the subject of Kant changing his mind, however, the new Refutation of Idealism is even more important. It is hard to see how some of the things that Kant says there are consistent with either version of TD. The relationship of this passage, which Schulting does not discuss, to TD is one of the issues on which I will focus in the remainder of these comments.
Finally for these introductory comments, if the mind contributes as much to its experiences, beliefs and the like as Kant suggests, including even their spatial and temporal structure, it is hard to see how the ‘manifold of intuition’ could play any role in controlling the truth-status of experiences, as Kant insists that it must. If on arrival the manifold of intuition has no temporal or spatial structure and no properties that the categorial concepts name, what structure could it have? Yet without such structure, how could it guide us in determining which propositions and other experiences to accept and which to reject as expressing ‘capricious and incongruous fictions’ (A96, A376)? When this concern focuses on spatial and temporal structure, it is often called the Localization Problem. If the Refutation of Idealism is primarily a problem for Schulting’s assumption that Kant’s views were consistent across the A- and B-editions of CPR, the localization problem is primarily a problem for Kant, though a complete account of his picture of knowledge acquisition must address it. Since not only Schulting but most commentators on Kant on knowledge do not address it (Falkenstein Reference Falkenstein1995 is a partial exception), we will take it up first.
2. Localization
The localization problem in its full generality goes like this. If we do not distribute the manifold of intuitions in space and time and apply concepts to them by virtue of spatial, temporal and property structures which they have on arrival, how do we do so? The slightly less general version of the problem was just discussed: in virtue of what can we decide that some orderings and applications of concepts are warranted, others not? We cannot localize and conceptualize elements in the manifold of intuition as we please – this is part of what Kant had in mind when he said that we are passive or merely receptive to intuitions. So what is it that controls, what could possibly control, how we do so?
For Kant, localization is a severe problem. Once space, time and concepts are separated from intuitions, it seems that neither perceptual pathways or filters nor acts of synthesis nor anything else would have the information needed to distribution elements of intuitions across space, time and concepts. (Kant once recognized the need for ‘reliable information’ (A60/B85) but did nothing with the idea.) There would have to be something else in the incoming intuitions that our perceptual system uses to distinguish them. (Falkenstein Reference Falkenstein1995: 85 makes a related point.) Otherwise, to insist, as Kant does, that intuitions have nothing spatiotemporal or conceptual about them on arrival would be an a priori proof of the impossibility of spatial, temporal and conceptual localization. There must be something else in intuitions as they arrive that resists some forms of organization and facilitates others. This is an especially acute problem for time and concepts. We might be able to use, say, shades and saturations to organize particulars in space but it is hard to think of any non-temporal property from which we could infer temporal location or anything both non-temporal and non-conceptual which could guide us in the application of concepts.
In fact, the localization problem is even harder for Kant than this, harder indeed than it would be, or need be, for other representational idealists. Not only are temporal and spatial properties conferred by the mind, space and time themselves are very different from anything that a particular contingent representation could have. Time and space are ‘pure forms of intuition’, which means that they are a priori in their origins, are known a priori (prior to, or at least independently of, experience) and are a priori in their modal status (they are features of how experience must be and confer necessity on some bodies of spatial and/or temporal knowledge); for these distinctions, see Brook Reference Brook1992. Being a priori in these ways makes time and space very different from particular, contingent properties, even temporal or spatial ones. It is not just that the forms exist prior to experience whereas particular intuitions are received by the senses. As the modal status of space and time make clear, the difference is much bigger than that. The forms of intuition allow us to discern necessity about the spatial (geometry: B16–17) and the temporal (in mechanics: B17–18; arithmetic: A103/B15–16, A103; and algebra: A717/B745, A734/B702). For Kant, ‘experience tells us … what is, but not that it necessarily must be so’ (A1). For that we need something a priori. All this makes time and space so different from particular intuitions that it is not easy to see how the latter could come to be located in the former at all. A cognitive system whose form is enough by itself to establish necessary truths will have a very special relationship to a posteriori empirical sensory inputs.
The issue of localization has received little attention in the literature on Kant and in this Schulting not an exception. As we saw, Falkenstein takes it up (briefly): he urges that Kant had to have been a realist about the time-order of representations because otherwise he would have had ‘no way to explain why we place matters in one location rather than another’ (Falkenstein Reference Falkenstein1995: 85; see also p. 250). Falkenstein takes the matter no further. Schulting and other commentators do not even get that far.
What about Kant himself? Usually the issue is deep in the background in his writings and unacknowledged. Sometimes it creeps close to the surface, as in: ‘the manifold to be intuited must be given prior the synthesis of the understanding and independently of it’; unfortunately, he then merely says, ‘How this takes place remains here undetermined’ (B145). Interestingly, he did walk right up to it at least once. (Whether he saw that he was doing so is another question.) In subsection II of §8 of the Aesthetic, the first sub-section of new material that he added near the end of the chapter for the B-edition, in the course of discussing what is presented by intuitions, he says this:
intuition … contains nothing but relations, namely, of locations in an intuition (extension), of change of location (motion), and of laws according to which this change is determined (moving forces). (B66–7)
How he could not then ask, ‘And how do items end up in the locations which we represent them as having? What controls where we represent an item as located?’, is difficult to fathom. But he did not. This passage comes the closest of anything that I know of in his writings to identifying and addressing the localization problem.
One response to Kant’s silence on localization would be to say that he was an idealist who held that one cannot be aware of anything outside the content of one’s own mind, so it is not surprising that he has nothing to say about what controls the element in our mental repertoire to which we are passive, what controls location in time, space and concepts. This response would be problematic – Kant was also an empirical realist and believed that we have a number of bodies not just of knowledge but of universal and necessary knowledge (B4). Physics and mathematics are the two main examples. So he cannot avoid the question, ‘In virtue of what that is non-temporal, non-spatial and non-conceptual would the way in which a physical or mathematical judgement temporalizes, spatializes, and conceptualizes make it an item of knowledge or a false belief?’ That said, if he has an answer better than the inadequate coherentism of A376, I do not know what it was.
As I said, a complete account of Kant’s view of knowledge acquisition needs to address the localization problem. Given the remarkably interesting claims that Schulting makes about other aspects of Kant on knowledge, I hope that at some point he turns to it.
3. Refutation of Idealism
Most commentators believe that Kant added the Refutation of Idealism to the B-edition to refute the charge that he was a Berkeleian idealist, a charge that his attack on the fourth Paralogism in the A-edition particularly encouraged. It may refute too much. The claims that Kant makes in the passage may undermine his own brand of idealism – and Copernican constructivism – too. This would make it a problem for Schulting. At minimum, his story would not be the whole story about Kant on knowledge. What is going on in the Refutation overall and how the passage relates to the rest of CPR is complicated and I have explored these issues elsewhere (Brook Reference Brook2011). Here I will restrict myself to the implications of the Refutation for the ideas that temporal, spatial and conceptual order are given to experience by the mind and that all we are aware of are experiences so structured.
The key element is a new doctrine of matter. In the first edition, Kant distinguished between matter as a feature of appearances – a feature that consists of the objects of these appearances having extension, impenetrability, cohesion and motion (A358) – and things as they actually are (A268/B324):
Matter is with [the transcendental idealist], therefore, only a species of representations (intuition), which are called external, not as standing in relation to objects in themselves external, but because they relate perceptions to the space in which all things are external to one another, while yet the space itself is in us. (A370)
What matter really is, what ‘inwardly belongs to it’ (A277/B333; I will not go into what Kant meant by this Leibnizian term ‘inward’), is hidden. All I can be aware of are its outer relations, its effects on my representations. This view is also expressed in the Appendix on the Amphiboly of Pure Reason and is a strong form of subjectivism and of the doctrine of the unknowability of the noumenal.
In the Refutation, something very interesting happens. Having argued that we must have immediate awareness of something permanent that is not just outside us but other than ourselves (‘an immediate consciousness of other things outside me’: B276, my emphasis) – something denied by implication in the passage quoted just above – Kant says in Note 2 that ‘we have nothing permanent … save only matter’ (B278, his emphasis). He gives objects on the earth as examples of the permanent, saying that we can see the sun move, for example, by comparing it to their permanence. If the permanent things other than oneself are matter, the sun and the earth, for example, and one is aware of the sun and the earth, then one is ‘immediately aware’ of things other than oneself. To our great frustration, that is all that Kant says but it is enough to indicate that he at least floated the idea here that matter is both not a feature of oneself and knowable.Footnote 2
On this new view of matter, Kant could not continue to hold that time, space and any order that can be captured in concepts have no extra-mental existence, though he may well not have seen this. He could retain the idea that we impose a temporal, spatial and conceptual matrix on our intuitions. However, now matter is not part of the mind and actually has temporal, spatial and conceptual order. It is not just represented as having these properties. If that is now Kant’s view and we put it together with the new realism about awareness of permanent objects other than oneself that we saw Kant advancing, he is now espousing not idealism but realism about temporal, spatial and conceptual order in at least some things other than oneself.
Such realism would also have implications for localization. If things other than oneself have temporal, spatial and conceptual order, then those kinds of order are there to be ‘transduced’ into the experienced raw manifold of intuition. And if that happened, then those orders would be in the raw manifold ready to be read off it when we come to locate items of the manifold in time, space and concepts.
In short, the realism of the Refutation seems to be in serious tension with especially the A-edition of CPR. One commentator who takes the tension between idealism and realism in the B-edition seriously is Guyer (Reference Guyer1987: chapters 12 and 14, especially pp. 282 and 327–8; Guyer also documents a general tension between idealism and realism in CPR and later work in those chapters). Yet even Guyer does not see the Refutation as in tension with transcendental idealism. Instead, he suggests, Kant was claiming that we must presuppose, in our representational constructions, that there are ‘external objects’ (he must mean objects other than oneself), but that Kant continued to maintain that representations never present the objects being presupposed. Guyer claims that Kant’s new position is ontological realism with epistemological idealism. In my view, this response is too easy. Kant is clearly espousing epistemological realism of some kind as well – as the quotation above from B276 demonstrates. I hope that Schulting takes up the relationship of the Refutation to the rest of CPR in future work.
There is a great deal in Schulting’s rich and generative book that I have not discussed, including the applications of his general interpretative scheme to particular issues in TD, the subject matter of his later chapters. I hope that nonetheless these comments inspire some further thoughts – in him and in others – about the themes in Schulting’s book that I have discussed.