Kant on Apperception and the Unity of Judgment∗

Download Kant on Apperception and the Unity of Judgment∗

Post on 09-Dec-2016




0 download


This article was downloaded by: [University of Stellenbosch]On: 08 May 2013, At: 12:31Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954Registered office: Mortimer House, 37-41 Mortimer Street, London W1T3JH, UKInquiry: An InterdisciplinaryJournal of PhilosophyPublication details, including instructions forauthors and subscription information:http://www.tandfonline.com/loi/sinq20Kant on Apperception and theUnity of JudgmentTimothy Rosenkoetter aa New York University, USAPublished online: 25 Oct 2006.To cite this article: Timothy Rosenkoetter (2006): Kant on Apperception and theUnity of Judgment , Inquiry: An Interdisciplinary Journal of Philosophy, 49:5, 469-489To link to this article: http://dx.doi.org/10.1080/00201740600937989PLEASE SCROLL DOWN FOR ARTICLEFull terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditionsThis article may be used for research, teaching, and private studypurposes. Any substantial or systematic reproduction, redistribution,reselling, loan, sub-licensing, systematic supply, or distribution in any formto anyone is expressly forbidden.The publisher does not give any warranty express or implied or make anyrepresentation that the contents will be complete or accurate or up todate. The accuracy of any instructions, formulae, and drug doses shouldbe independently verified with primary sources. The publisher shall notbe liable for any loss, actions, claims, proceedings, demand, or costs ordamages whatsoever or howsoever caused arising directly or indirectly inconnection with or arising out of the use of this material.Kant on Apperception and the Unityof Judgment*TIMOTHY ROSENKOETTERNew York University, USA(Received 3 August 2005)It is safe to say that more has been written in an attempt to come to termswith Kants second-edition Transcendental Deduction (TD) than has beenwritten on any other similarly sized section of his theoretical philosophy. Asmight be predicted from this fact, new contributions to the secondaryliterature are often characterized by the slight revisions that they make toearlier forays. Yet the array of available interpretations is sufficiently diversethat a survey can easily lead one to despair of progress. A monograph whichsystematically applies bold and in many cases unusual claims regardingrepresentation, objectivity, and apperception to the interpretation of the TDis a particularly welcome addition in this context. Whats more, A.B.Dickersons Kant on Representation and Objectivity (KR&O) paints anattractive picture. Kant is reconstructed as having a sensible, even inviting,position that lacks any obvious confusions or hopeless arguments.Though KR&O is advertised as a study of the B-deduction, Dickersonarranges pursuit of that goal so that he first lays out his positions on a numberof the core issues that confront any interpretation of Kants theoreticalproject (chapters 1 and 2). This general interpretation is then put to the test ofmaking sense of the B-deduction (chapters 3 and 4). It is an ambitious work.Indeed, one of its strengths is that its form and coverage make it appropriatefor use as one of the main secondary sources in a course on Kants theoreticalphilosophy, something that cannot be said of most studies of the TD.Furthermore, though it can and will be read with profit by specialists, this*A.B. Dickerson. Kant on Representation and Objectivity (Cambridge: Cambridge UniversityPress, 2004).Correspondence Address: Timothy Rosenkoetter, Department of Philosophy, New YorkUniversity, 1000 Washington Square East, Silver Center 503, New York, NY, 10003, USA.Email: tr37@nyu.eduInquiry,Vol. 49, No. 5, 469489, October 20060020-174X Print/1502-3923 Online/06/05046921 # 2006 Taylor & FrancisDOI: 10.1080/00201740600937989Downloaded by [University of Stellenbosch] at 12:31 08 May 2013book is sufficiently readable, and its author sufficiently patient in hisexplanations, that it should not overwhelm talented undergraduates. Thisability to speak to both audiences is as much a virtue as it is a rarity.Dickerson tells us that he intends KR&O to be a contribution to whatRobert Sleigh calls exegetical history, rather than just a variation on aKantian theme. As such, his ambition is to produce sentences such thatwe know what propositions those sentences express and those propositionsare the very ones our author accepted.1 With few exceptions Dickersonquite successfully avoids producing terminology-laden sentences to whichwe can attach no determinate proposition. However, I have morereservations about whether they are in fact the propositions that Kantaccepted. In some cases my disagreements are relatively minor. There areplaces where Dickerson has simplified to the point of distortion. (Theauthors quite successful efforts at producing a very readable work haveperhaps exacted a toll here and there.) But even in cases where I think thatDickersons interpretation is on the wrong track, it often manages to beprovocative in a way that can be quite useful in spurring those who disagreeto explain why it is wrong.Dickerson belongs to the camp of interpreters that understands the TDnot as a response to skepticism but rather as primarily an analysis of theconcept of human cognition (p. 206). This should not, Dickerson hastensto add, lead us to assume that its goals are modest. It aims to prove that ifour minds are receptive in cognition or constrained by an independentreality, then our cognition must be governed by the categories (p. 208).Dickersons most general take on the goal and method of the TD is notparticularly new or startling (though not for that reason any less welcome).For this reason, I have chosen to focus the following discussion on thebackground theory which Dickerson employs in his detailed interpretationof the argument of the TD, since it contains several original theses that formthe basis for his detailed claims regarding the TD. Now, though I will bepassing over much of the latter, I want to stress that there is much that isinteresting and useful there. (Those who desire an informative overviewmight begin with the master presentation of the argument in twenty-six stepsat p. 201.) Further, as Dickerson stresses, the success of the backgroundtheory in accounting for details of Kants text has the potential to provideevidence of its correctness. However, I believe that there are some questionsthat are usefully posed in relative abstraction from Dickersons reconstruc-tion of the argument.I. Representationalism, pictures, and apperceptionThe starting point for Dickersons interpretation is his claim that Kant is arepresentationalist. Kant clearly believes that objects produce representa-tions in us, which when considered merely as modifications of our sensibility470 T. RosenkoetterDownloaded by [University of Stellenbosch] at 12:31 08 May 2013are termed sensations.2 Kant is a representationalist, according toDickerson, because these internal modifications or determinations arethen the immediate objects of awareness (p. 6). Dickerson takes himself tobe opposing direct realist readings of Kant and he believes that this moveshould not be controversial. The exegetical historian simply has no room tomaneuver on this point, for the language of Kants texts plainly indicatesthat representations are the objects of our mental actsas objects ofconsciousness or awareness (pp. 910). In the first of several straightfor-ward analogies that Dickerson employs throughout the book, he comparesthe Kantian subject to a viewer who is within a hollow globe of opaqueplastic. As the globe is depressed by external forces the subject can seenothing but the internal surface of the globe.It should be noted up front that Dickerson takes Kants representation-alism to explain a commitment that he believes is Kants: that sensations assuch are intrinsically unavailable to the subjects awareness because they arenot self-revealing (p. 85) and the subject, in relating to them, is aware ofthem as objects (e.g., buckets and chairs) rather than as sensations. Thus, itis more than a little misleading for Dickerson to refer (as he routinely does)to sensations as objects of consciousness or awareness (pp. 910). Thinkof them, instead, as the forever-hidden raw materials of consciousness.If we are trapped within a globe with modifications of our sensibility (i.e.,sensations) as our sole epistemological aid, how shall we account for ourknowledgeknowledge which is, at least in part, of independent objects?Two historically influential strategies suggest themselves. First, one can inferfrom ones awareness of representations to their external causes (e.g.,Descartes). Second, one can insist that objects are in fact identical to someset of representations, so that an immediate awareness of objects is in factretained (Berkeley). At least part of what leads interpreters to deny ordownplay Kants representationalism, on Dickersons diagnosis, is theirunderstandable wish to avoid reading Kant as employing either one of thesemodels. Dickerson thinks that there is an unconsidered alternative open tothe representationalist. The key to seeing the alternative is an analogy withpictures.Suppose that you are looking at a pencil drawing of Smokey the Bear.What is the relation of Smokey to the various lines and shadings that makeup the drawing? It would be wrong to say that Smokey is identical to theselines and shadings. After all, Smokey has a thick coat of fur and is wearing ahat, whereas none of that is (or could be) true of the pencil markings.Dickerson also insists that you do not infer from the pencil markings toSmokey, for that assumes that Smokey is a separate object that liesbehind or outside of the configuration of lines (p. 14). Dickersoninstead wants to sayin a phrase that recurs again and again at key pointsin the bookthat Smokey is in the picture.Kant on Apperception and the Unity of Judgment 471Downloaded by [University of Stellenbosch] at 12:31 08 May 2013Now if we substitute a Kantian subjects sensations for the pencil-markedpaper in this example, we get Dickersons recipe for admitting Kantsrepresentationalism while avoiding the pitfalls of Cartesian and Berkeleyanstrategies. We also get the key to the meaning of Kants terminology ofsynthesis and apperception. The act of synthesis is the act of seeing anintentional object in ones sensations. This is importantly different from theatomistic model of synthesis that is, Dickerson would admit, most readilysuggested by many of Kants texts, as in his definition of synthesis in themost general sense as the action of putting different representationstogether with each other [zu einander hinzutun] and comprehending theirmanifoldness in one cognition.3 Synthesis is not accomplished through thecombination of independently intelligible pieces into a larger whole. This isnot to deny that the representational medium in which you see Smokeycontains a manifold composed of many different lines, variations in shading,etc., and that it may be impossible to see Smokey unless your visualapparatus receives information from all of these lines. Dickersons claim isonly that your seeing Smokey is not properly understood as the result ofadding together various smaller seeings (e.g., seeing his color, claws,snout and ranger hat) (p. 125).Now we must immediately note an obvious fact, which Dickerson doesnot discuss. It may be the case that Dickerson is quite right to make holisticclaims about our seeing of certain objects in pictures, but it would seem thatthere are seeings which are much more like combinations of smallerseeings. It may, for instance, be the case that my cognitive relation to adepicted bear must be understood by starting with my seeing the whole bear(color, texture, snout and all the rest). But perhaps if I next look at anotherwise similar bear who is wearing a beret, my seeing of that bear will bethe result of something more like combining my seeing of a bear with myseeing of a beret. Now, I have no idea whether this hypothesis accuratelydescribes our visual cognition, and my use of something more like as afudge should certainly give pause. What is important in the present context,however, is that Kant might recognize something like this distinction (i.e.,between seeing a bear and seeing a bear-with-beret) in his account ofobjective cognition. We will return to this point below. What is important inthe meantime is that on Dickersons reading, it is precisely a grasping of thewhole, prior to the grasp of any of the parts (such as we find, I amallowing, in the case of seeing a plain old bear) which Kant is trying toexpress with his use of the term synthesis.How, then, does Dickerson propose to interpret apperception? In terms ofthe picture analogy, Dickerson would say that you perceive Smokey in thepicture by apperceiving the representational medium (the pencil marks, etc.)(p. 87). Applying this to the cognizing subject we derive the claim thatapperception is that relation which a subject bears to her sensations byvirtue of which she cognizes objects in them. These sensations are of course472 T. RosenkoetterDownloaded by [University of Stellenbosch] at 12:31 08 May 2013her own representations, but she is not relating to them or aware of them asher own. Rather, she is relating to them as, e.g., a shovel or a bucket ofwater (their intentional object). In short, accompanying a representationwith the I think means thinking about the object of that representation(p. 93). Readers even minimally familiar with the secondary literature onKant will recognize that this is a highly unusual take on apperception. Hereare two reasons.First, Dickersons account straightforwardly entails that it is impossibleto apperceive a representation without thinking of its object. This will bethought by some to conflict with Kants account of analytic judgments, butthis depends on assumptions that are at least controversial.4 What will bemore widely regarded as problematic is the fact that Kant obviously allowsthat subjects can think concepts without thereby involving any intuitions.While it is open to Dickerson to deny that subjects are thereby apperceiving,this would be difficult to square with textual evidence.5 So I assume thatDickerson would instead concede that all thinking of concepts isapperceptive (and therefore refers to an object). But it is entirely unclearwhat this could mean on Dickersons account in these limit cases. In the casein which no intuition is involved there is no representation that is beingobserved on the internal surface of my globesurely we cannot conceive ofconcepts in that wayand so there is nothing to apperceive as an object.6Second, many interpreters will feel that Dickersons model does notadequately reflect Kants understanding of pure apperception as (somevariety of) self-consciousness. Dickersons core position is that apperceptionqualifies as self-consciousness only in the sense that it is a relation to onesown representationsthough not, it is essential to repeat, as ones ownrepresentations. Apperception is self-awareness only in this attenuatedsense: All perception involves apperception, or all awareness a certain sortof self-awareness, just as seeing the depicted object [in a picture] involvesseeing the representational medium.7 Accordingly, Dickerson argues againsttwo of the most natural (and well-represented) rival interpretations,according to which apperception is Kants term for my ability either:(A) to be aware that a representation is mine (or to ascribe it tomyself); or(B) to make a first-person propositional-attitude judgment (I judgethat q).One of Dickersons worries about these more standard interpretations isthat they will make the TD (116, in particular) depend on the controversialassumption that all finite cognizers have the capacity to do things that arevery sophisticated things (p. 94). Dickerson cites with approval Guyersobjection that there very well could be subjects who judge about objects andyet lack the abilities described in (A) and (B).8 Dickersons interpretation, asKant on Apperception and the Unity of Judgment 473Downloaded by [University of Stellenbosch] at 12:31 08 May 2013we have it here, can be couched in a simple motto: apperception is cognition,not self-consciousness.Now if Dickersons interpretation of apperception were as simple as itfirst seems, then it would qualify as at mosta variation on a Kantiantheme (p. 10) rather than as exegetical history, for the weight of textualevidence against such a reading is simply overwhelming. It is just not tenableto claim that the only sense in which apperception constitutes or involves aform of self-consciousness is that the apperceiving subject is entering into areflexive relation with its own representations (though not as its ownrepresentations). Fortunately, though it provides the core of Dickersonsaccount, the foregoing has abstracted from one important claim. In order toavoid confusion, lets call the foregoing rationale for why Kant would be ledto associate apperception with self-consciousness the ownership rationale.9Of course, it competes directly with (A) and (B), which providestraightforward rationales of their own. Understanding Dickersons moreplausible interpretation of apperception will require isolating a furtherrationale for that association. We will see, however, that shining light on thisrationale will have the effect of making it less clear whether Dickersonsinterpretation of apperception really is fundamentally different from thepropositional-attitude interpretation (i.e., (B)) that he criticizes. Explainingthis further rationale will require that we return to the analogy with pictures.Early on Dickerson draws attention to the familiar fact that a subjectsaccurate apprehension of all of the lines (etc.) in a representational mediumdoes not suffice to guarantee that she will see what is depicted in the picture.This conceptual gap is especially clear in the case of trick pictures orpictures in which a face is hidden in the midst of many distracting lines, butit holds in principle for all pictures (pp. 2021). Dickersons preferred wayof expressing this point is to say that a representational medium, in order todepict, must be a picture for me. Correlatively, he points to certain errorsthat are apt to arise if we consider pictures from (what he refers to as) athird-person perspective. If I consider your relation to a representationalmedium without adopting your standpoint I may err in the intentionalobject that I ascribe to you. More seriously, I may even fail to grasp that themedium is a pictureunless, that is, I adopt a first-person standpointtoward the medium and it also becomes a picture for me. Dickerson is,however, most interested in the sort of general error that a theorist ofpictorial representation might make by adhering rigidly to a third-personstandpoint. Two types of error are possible, analogues of which weencountered above in the guises of Descartes and Berkeley. It is natural, onthe one hand, for the third-person observer to understand the observingsubject as relating to two separate things (in a suitably broad sense ofthings): representational medium and depicted object (p. 18). This missestheir intrinsic connection. More deleterious, however, is the error of simply474 T. RosenkoetterDownloaded by [University of Stellenbosch] at 12:31 08 May 2013identifying representational medium and depicted object. This is the resultof failing to allow the medium to become a picture for oneself.Now, turning back to our main topic, we find Dickerson using thesepoints about pictures to argue that a representation is not merely amodification of the subject but something more like the subjectsperspective or point of view upon the object of the representation. To beconscious of my representation as representing is thus for it to come tofunction as a point of view for me, or for it to become my view ontosomething. In this way a representation as it were intimates the point fromwhich that point of view is had (pp. 8283). This would seem to giveDickerson a new rationale for why Kant would have been led to associateapperception and self-consciousness. I will call it the point of view rationale.It is not a dubious piece of introspective reportage but rather theconceptual point that it is impossible for me to cognize my sensations asan object without having some awareness that this cognition is from mypoint of view (p. 82). At the same time, this is not to make the subject intoa further object within the field of the point of view (p. 83). Nor hasDickerson given back what he denied Kant in rejecting (A).10While I think that Dickersons remarks on pictures can be made fruitful inunderstanding Kant; and while I think that there is much to be said for thepoint of view rationale, it seems to me that Dickersons attempt to supportthat rationale by recourse to the picture analogy is confused. The basic pointis the following. It is consistent with Dickersons observations concerningseeing in a representational medium that I could see a particular set ofpencil markings as Smokey the Bear without having any clue that this is myparticular view on these markings. It might help to vivify this possibility ifwe imagine a subject for whom seeing the markings as Smokey isautomatica subject for whom, in other words, there could be noquestion of whether the markings are Smokey or Bambi, or whether they arenothing but meaningless scribbling (i.e., no intentional object is cognized). Isuspect that Dickerson has been misled because it is easy for human subjectsto vary their relation to (at least some) pictures so that at one moment theysee only some pencil markings and at the next moment they see an object inthose markings. His discussion also evokes the possibility that multipleobservers could become aware that they have differing cognitive relations tothe same representational medium (e.g., one sees nothing in it and the othersees Smokey). Now, perhaps it would be impossible for these subjects not torealize that their seeing an object in the medium is in each case my view onsomething which admits of alternative interpretations. After all, they areaware that other subjects (or themselves at different times) are interpretingthe medium differently. Perhaps awareness of the possibility (or actuality) ofconflicting interpretationsincluding the limit case of not seeing anintentional object in the mediumbrings with it awareness that ones owninterpretation is a point of view. Or perhaps not (in which case, so much theKant on Apperception and the Unity of Judgment 475Downloaded by [University of Stellenbosch] at 12:31 08 May 2013worse for Dickersons case). In either case, however, there is no reason tosuppose that the former sort of awareness (lets call it awareness of divergentinterpretations) belongs essentially to the experience of pictures. Thus, thereseems to be no reason to suppose that a subject lacking self-consciousnesswould not be able to see Smokey in a picture. This is, in effect, to turn avariation of Guyers objection against Dickerson: the ability to have mypoint of view-thoughts is a relatively sophisticated ability, and nothing inDickersons thesis that cognizing objects is analogous to seeing picturesshows that the former requires a capacity for self-consciousness.Looking closer, we see that the problem is even more serious. It is not justthat the picture analogy fails to show that cognizers must have someintimation that their cognitions are, in each case, from my point of view. Ifanything, the analogy points in the opposite direction. This is because it is apillar of Dickersons reading of Kant that our spontaneous use of concepts,which Dickerson understands as the imposition of content that is not given,can qualify as objective only if all finite judgers would judge identicallygiven identical inputs (sensations) (cf. p. 73). Dickerson suggests that wethink of Kants cognizers as applying a function (made up of the sum totalof the categories) to given arguments (sensations). This function cannotderive its objective validity from the fact that it mirrors reality, for it doesnot. Yet the resulting cognitions will nonetheless be objective so long as allpossible finite cognizers are applying the same function.11 Hence, Kantiansubjects are on Dickersons reading analogous to a group of subjects who,no matter what pencil markings (etc.) are laid before them, always agree asto which objects they see in the representational medium. This is to say thatfar from providing a reason to think that cognizing subjects will possess anawareness of divergent interpretations, the picture analogy, when combinedwith Dickersons understanding of Kantian objectivity, gives us reason toexpect that cognizers will precisely not be provided with the conditionswhich might be expected to render unavoidable an awareness that theircognitions reflect a point of view.12What is emerging, I would suggest, is that Dickerson incorrectly locatesthe point in Kants project to which the putative first-person nature ofseeing in a picture is relevant. There is a sense in which all acts of finitecognition arise from a point of view, but that point of view is not my pointof view as opposed to yours. Instead, all acts of finite cognition togetherreflect a single point of view. To the extent that finite cognition is analogousto seeing in a picture, thinking about this analogy may help elucidate thenature of Kants transcendental idealism.13 As Dickerson briefly suggests,reading Kant as a phenomenalist may be tantamount to holding,incorrectly, that facts about Smokey the Bear are reducible to statementsabout the spatial arrangements of pencil markings (p. 74). What couldbenefit from more investigationthis is not a topic that Dickerson tries topursue hereis how this defense depends on conceiving of transcendental476 T. RosenkoetterDownloaded by [University of Stellenbosch] at 12:31 08 May 2013philosophy itself as adopting a first-person perspective as opposed to theperspective of ontology.14 Dickerson quotes the following passage from theA-edition Paralogisms: it is obvious that if one wants to represent athinking being, one must put oneself in its place, and thus substitute onesown subject for the object one wants to consider.15 He then comments:Since the Critique is itself a representation of a thinking being, Kantsremark is an instruction for reading the book (p. 18). Well put. Theimplications of this thought await development.Lets take a step back and review our progress. One of the primarysources of the originality of Dickersons approach is that while mostcommentators take apperception to be one aspect of a whole objectivecognition, Dickerson insists that apperception be identified with the wholeof cognition itself.16 Any comprehensive account of Kantian cognition mustpresent apperception as some sort of self-consciousness; otherwise it will notbe textually supportable. Dickerson attempts to show that the model ofobjective cognition that he has brought over from the picture analogyprovides a direct explanation for the self-conscious quality of apperception.If this were successfulI have argued that it is notthen the interpreterwould not need to introduce a notionally separate source for self-consciousness. That is, on Dickersons model, any explanation of the self-consciousness of objective cognition will simply amount to an account ofobjective cognition itself. Beneath the surface of this accountandsupporting this approach to apperceptionI detect what I will callrepresentation-positivism, i.e., the position that anything which makes acontribution to an objective cognition must itself be a representation,understood as an internal object along the lines that we see in Dickersonstreatment of sensations. On this model, anything of cognitive significance iseither a representation (internal object) or an entire cognition. WhileDickerson is able to show that some interpretative advantages accrue tothose willing to take Kants representationalism seriously, I believe that hisexample demonstrates the perils of going beyond this and applyingrepresentation-positivism to Kant. This will become clearer in the nextsection.I believe that we ought to admit that apperception is for Kant notionallyseparate from objective cognition. Consider an interpretation that connectspure apperception closely to a subjects consciousness of its own action.17A controlling idea for this account would be that merely being in a state isnot an action. As such it is not something for which one could takeresponsibility. However, subjects are able to relate to an identical contenteither as a state (in inner sense) or as something that they are thinkingunderstood as an action that they are performing and therefore as an actionthat they are responsible for performing correctly. What kind of action isthinking? It is the action that aims ultimately at truth. So, once we observethat for Kant the truth of a cognition is its correspondence to thatKant on Apperception and the Unity of Judgment 477Downloaded by [University of Stellenbosch] at 12:31 08 May 2013cognitions object, we can explain why the grain size of judgment isprivileged and plays a special role in Kants account.18 Namely, thefundamental building block of all judgmentI think of object x as F (thisclaim will be explained in the next section)is the smallest unit which caneither correspond or fail to correspond to an object.This account can also motivate the association of apperception with anawareness that the resulting cognition is a point of view on the objects.Namely, it is plausible that in order for an act to count as an assertion thesubject must be aware that its negation might instead be the case.19 This isnot to say that the subject must hold that this assertion represents nothingmore than her own idiosyncratic point of view about which judging actionis correct. Instead, the subject is taking a stand that this is the correct actionwith the awareness that it is in principle possible that x is not F.This model is neutral with respect to Dickersons claim that seeing inpictures is a good model for objective cognition. However, it parts companywith him by admitting that there can be acts of apperception in which thereis no cognition of an object. As I said above, I think that this view is forcedupon us by Kants texts. Yet while intuitions are not involved in allapperception, I will explain in the following section why I join Dickerson inmaintaining, not uncontroversially, that intuitions are an ineliminableconstituent in what is properly considered the central act of cognizingsubjects, judgment. Furthermore, Dickerson is correct to see that intuitionsare necessary to the central act of cognition if Kant is to have a satisfactoryaccount of the unity of judgment. This will be a result of the next section.II. Apperception and the unity of objective cognitionsWe have seen how Dickerson combines the claim that Kant espousesrepresentationalism with a set of observations concerning pictures to arriveat a novel interpretation of Kants key notion of apperception. Lets nowshift our attention to the other main piece of background theory thatDickerson deploys in his interpretation of the TD. It is the hypothesis thatKants conception of apperception allows him to solve a problem, which isvariously termed the representationalist equivalent of the semantic problemof the unity of the proposition and the representationalist parallel of thesemantic question of what it is to understand a complex sign.20 Thisproblemwhich I will henceforth simply call The ProblemstructuresDickersons reading of 111619 of the TD. In simple terms, what Kant issupposed to be doing in those sections is contrasting a picture of objectivecognition, as holistic apperception allows us to conceive of it, with a pictureof cognition as mere building up through the addition of simple,independently intelligible representations, i.e., the picture that is typical ofempiricist models. Kants insight, we are told, is that the philosopher musttake the holistic route of explaining the properties of the parts by appeal to478 T. RosenkoetterDownloaded by [University of Stellenbosch] at 12:31 08 May 2013the properties of the whole (p. 119). Contrary to empiricist atomism, thestarting point for my conscious experience is my consciousnessof thecomplex unity as a whole (p. 126).There is, indeed, a striking parallel to be found between the difficultiesfaced by representationalists when trying to account for the cognition of anobject and the difficulties faced by Bertrand Russell in The Principles ofMathematics (1903), whose case Dickerson invokes. The problem thatRussell faced was to account for the difference between a proposition suchas A is different from B and a mere list of its constituents, {A, difference,B}. The reason the challenge took just this formwhy Russell should haveexpected that the proposition would be reducible to a listis that herecognized only one ontological category (terms), whose independenceand lack of intrinsic relatedness make them object-like.21 Given thispreference for ontological parsimony, Russell was unable to find anythingto account for the unity of a proposition. As Peter Hylton helpfullydescribes Russells bind, [his] attitudewas that any component of theproposition would bewell, just one more component with the same statusas the others.22 Now, the problem that Frege faced in accounting for theunity of the proposition is importantly different, for the simple reason thathe was willing to analyze propositions into two irreducibly differentconstituents (concept and object), one of which is unsaturated oruncompleted and as such intrinsically requires an object, which is saturated,for its completion. Unlike Russell circa 1903, who could offer nothing evenapproaching a solution to the problem of the unity of the proposition,Freges analysis of propositions into concept and object at once makesobvious progress.23 While that move does not eliminate all problems inaccounting for unity, the important points for our purposes are that theproblems Frege encounters are of a different sort, and they are of a sort thatKant did not take himself to face.24The difficulty with reading the TD as containing Kants solution toRussells problem is that Kant, unlike Russell circa 1903, recognizes twoirreducibly different kinds of objective representations: intuitions andconcepts. To the degree that intuitions and concepts are such as to form aunity when combined, Kants approach to The Problem will be more likeFreges than it is like Russells. Accordingly, though Locke and Hume facedthe representationalist parallel of Russells problem, Kant may not have. Itall depends on whether concepts, which are classed along with sensationsand intuitions as representations, are the same sort of things as thesesensible representations.25 If they are, then it is a mistake to read Kant asanticipating, in however rudimentary a fashion, Freges approach. While itwould be sanguine to deny that Kant sometimes treats concepts as if theywere just another representation upon which the subject can gaze, there arealso discussions in which Kant attempts to break free from the empiricistsKant on Apperception and the Unity of Judgment 479Downloaded by [University of Stellenbosch] at 12:31 08 May 2013approach to concepts. I will be working with material from this more radicalstrain in what follows.How, then, might concepts and intuitions be made for each otherandKant, to that degree at least, be anticipating Frege? Lets suppose that ajudgment is unified by an act that acts upon intuitions and further thatconcepts in some sense are such acts. Hylton made a similar suggestion in anearly, groundbreaking article, to which we read this reply:this does not solve the unity problem, for what representationalcontent does this special elementthis actcontribute to the unifiedwhole? If that content is specifiable independently of the wholeas thepresupposition of representational atomism demandsthen thisspecial element is simply a further representation (call it what youwill).26Though I think that there is at least one other commitment that ismotivating Dickersons reaction here,27 it is clear that he is also treatingKant as representation-positivist, for whom both intuitions and concepts,whatever their other differences, must as representations both be immediateobjects of awareness with a certain representational content (p. 96).28 It isdifficult to make sense of how concepts function along these lines, and soconcepts more or less drop out of his account of cognition. It is a symptomof this absence that Dickerson provides no account of how Kantunderstands analytic judgments or the mere thought of a conceptunconnected to intuition.None of this shows, of course, that Kant has an entirely unproblematicaccount of conceptual content! However, worries about its shortfalls shouldnot lead us to attribute to Kant a blank holism consisting in the claim thatthe grasp of the whole (whatever it is) is prior to the grasp of the parts(whatever they are). Kants attempt to account for judgment by analyzing itinto concept and intuition does indeed bear some resemblance to Fregesanalysis employing concept and object; and in both cases the specificstructure provided by these twin notions iswellimportant. Thecognition-swallowing conception of apperception that Dickerson attributesto Kant leaves him unable to do justice to this structure.This is reflected in the striking fact that nothing in Dickersons argumentsprovides a genuine explanation for why the holistic demand for the priorityof our grasp of the whole should be satisfied with a whole that is the size ofa judgmentas opposed to alternative wholes such as that of a syllogism, atheory, or even an entire lifes experience. It might at first seem that the factthat judgment is the proper grain size is explained by the fact that (i) it isapperception that solves The Problem, together with the fact that (ii) theapperception of a sensation just is the cognition of an object, and that (iii)any such cognition is a judgment. The problem with this response (which I480 T. RosenkoetterDownloaded by [University of Stellenbosch] at 12:31 08 May 2013take to be Dickersons) is that apperception in its role in (i) has come apartfrom apperception in its role in (ii)(iii). As a consequence, Dickersonsde facto treatment of them as one and the same apperception is purelystipulational. This will be clearer if we consider one of the arguments withwhich Dickerson motivates his holistic approach.The guiding question in this particular discussion is how we canunderstand a complex representation whose content is vblue and redw.To demonstrate that an atomistic approach will not suffice, Dickerson asksus to imagine a subject who has a representation of blue as well as arepresentation of red.There would be a point of view on blue (where this exhausts the fieldof awareness) and a point of view on red (where this exhausts the fieldof awareness), but there would be no unified point of view on, orawareness of, blue and red together (p. 118).The demonstration is completed with the claim that whatever else I mighttry to add in order to unify these separate representations, it will be simplyyet another separate point of view needing to be unified with the others(p. 118). Clearly, this argument can be run identically if we replace blue andred with two judgments or even with ones experiences before and after theage of 40much as Dickersons points about seeing in a picture were madeso that they applied identically to seeing a bear and seeing a bear-with-beret.Yet since the TD most definitely does not demonstrate that ones grasp ofones whole life is prior to ones grasp of its pieces, something is amiss. Iwould suggest that Dickerson is here running together the question of whatmakes for a unified awareness (or point of view) in general (answered by (i))with the question of what makes for a unified cognition of an object(answered by (ii)(iii)). I doubt that Kant attempts to solve the formerproblem in any general way. The latter problem he attempts to solve, Iwould argue, specifically at the level of the judgment.This brings me to a general claim regarding The Problem: Kantsapproach to the problem of the unity of representations cannot be fullyunderstood without an understanding of his approach to the problem of theunity of a judgment. We must be able to understand Kants grounds forholding (iii) and should not accept otherwise unsupported assurances thatan objective cognition is equivalent to an act of judging. This is becausejudging is not an empty term, to which Kant may attach whatever sense hepleases. Rather, the Metaphysical Deduction has provided an a prioriaccount of what judgment is and the forms that it can take. That accountmust fit with the account that he has given of the unity of representations inthe cognition of an object. At first glance the prospects look decidedlyunpromising. After all, Kant is working with a conception of judgmentthat allows it to fit into traditional syllogistic logic. Yet it is well known thatKant on Apperception and the Unity of Judgment 481Downloaded by [University of Stellenbosch] at 12:31 08 May 2013traditional logic is based upon a symmetric understanding of the role thatsubject- and predicate-terms play in judgment. In addition, judgment firstbecomes an issue for traditional logic when there is a question of combiningconcepts, which are understood as independently intelligible atoms ofmeaning. (Similarly, the syllogism only becomes an issue when there is aquestion of combining independently intelligible judgments. Each levelpresupposes the lower level as in itself intelligible raw material for its higher-level operation.) In contrast, if Kant is to earn the right to a Frege-stylesolution, then concepts must be essentially unsaturated or uncompleted;concepts must be such as to make a unity only when completed by anintuition. Accordingly, judgment must be prior to both concept andsyllogism.On this nest of questions Dickerson makes a single significant move,contained in the suggestion that judgment for Kant is primarily anawareness of things as being thus and so, rather than an awareness thatthings are thus and so (p. 25). This suggestion preempts any uncomfor-table questions regarding the unity of judgments per se, since it simplyidentifies judgment with the representation of an object. I believe thatDickerson is on the right trackjudgments are for Kant essentiallycognitions of objectsbut that this solution does not come so easily.Dickerson connects the above suggestion with the supremely unconvincingclaim that Kant simply lacked any clear understanding of a judgment assomething that can be the content of a that clause. Yet rather thansimply dispose of the issue by denying that Kant had any comprehensionof the difference between Some F are G, on the one hand, and thecognition of an object as F&G, on the other hand, the challenge is toassess whether Kants undeniable characterizations of judgment as theformer really are as threatening as they at first seem to the project offinding unity in Kantian judgment. I will conclude with an all too briefsketch of this challenge.There is an obvious tension between the ways in which Kant conceives ofjudgment in discussions that fall within the purview of general logic, and theways in which he comes to describe judgment when he is trying to makesense of our cognition of objects. As a paradigm case of the former, considerthe definition of judgment in the Jasche Logic:(C) A judgment is the representation of the unity of the consciousnessof various representations, or the representation of their relationinsofar as they constitute a concept.29Because vtwo-sided polygonw is unimpeachable qua concept, it followsthat two-sided polygons are two-sided qualifies as a judgment. Nowconsider 119 of the TD, which opens with what I take to be a direct denialthat a definition such as we find in (C) is adequate:482 T. RosenkoetterDownloaded by [University of Stellenbosch] at 12:31 08 May 2013(D) I have never been able to satisfy myself with the explanation thatthe logicians give of a judgment as such: it is, they say, therepresentation of a relation between two concepts.30Kant then proceeds to explain that a judgment is nothing other than theway to bring given cognitions to the objective unity of apperception.31 Theensuing explication of this remark makes clear that judgment cannot beunderstood apart from objects. A judgment asserts that the representationsin the subject- and predicate-terms are combined in the object. Thisdiscussion recalls the definition of judgment that Kant had published theyear before as an action through which given representations first becomecognitions of an object.32Uncovering a Kantian solution to the problem of the unity of judgmentwould require making the case that this latter, object-centered conception ofjudgment is the fundamental, non-derivative sense of judgment for Kant.More specifically, it would require that the two following claims besubstantiated:(E) No more than one concept is required for judging; and(F) No genuine judgment is possible without a correspondingintuition.Each of these claims is contradicted by straightforward textual evidence, yeta case can also be made for each. Part of the appeal of (F) is that its denialseems to accord too trivial a role to intuition, as if intuitions were merelyadditional evidence for the truth of an independently intelligible relation ofconcepts. In embracing (F) one is instead according a semantic role tointuition. Of course, (F) is contradicted by any presentation of judgment asconsisting in the mere relation of concepts as in (C). Yet Kants owndiscussions in contexts such as 119 provide some grounds for discountingthese latter characterizations. A full defense of (F) would require workingout a theory of analytic judgments according to which they are somethingother than judgments in the strict sense.33(E) receives immediate support once we consider that Kant denies thatjudgments such as God is existent are properly understood as predicatinga concept of God. Kant analyzes the genuine logical form of suchjudgments as the absolute positing of a single concept (vGodw). So ifexistential judgments are genuine judgments, then there is somethingmisleading about Kants routine practice of speaking as if all judgmentsrequire at least two concepts. Indeed, when we look closely at theMetaphysical Deductions canonical definition of a judgment as themediate cognition of an object, hence the representation of a representa-tion of it, we see that it contains no justification for the necessity of twoconcepts in a judgment. This is because a concept that refers to an objectKant on Apperception and the Unity of Judgment 483Downloaded by [University of Stellenbosch] at 12:31 08 May 2013via an intuition is already the representation [5concept] of a representa-tion [5intuition] of [an object]. Of course, there still remains much to beexplained, starting with why Kant nonetheless presents the normal form ofa categorical judgment as containing both subject- and predicate-terms.Here a hint is given by Kants connecting of the categorical judgment-formto the category of substance, once the former is conceived as somethingmore than just a way to relate concepts. Once this change is madeonce Itake my judgment to be constrained by an objectit is no longer arbitrarywhether a concept belongs to the subject-term or the predicate-term.Nonetheless, no particular concept belongs essentially to the subject-termof such a categorical judgment. That is, any particular concept that is usedin the subject-term can be moved to the predicate-term (The stone is hardbecomes, e.g., The object is hard-stone-ish). Now if we repeat thisoperation, then in the limit case no intensional content will be expressedby the concept in the subject-term.34 All that will be left is the directreferential connection provided by intuition. Substance will always remainan I-know-not-what:I cannot cognize something of a thing other than through judgments,and predicates always underlie theseBut that we cannot comprehendthe substantial, but rather merely the accidents, comes from this:because we are much too short-sighted, and because the understandingcan think only through concepts, and concepts are nothing more thanpredicates.35The simplest building blocks of all objective cognition are judgments whichsubsume an intuition under a single concept:The categorical judgments are the basis of hypothetical and disjunctiveones. Since we cannot cognize anything without judgments, and eveneach concept is a judgment, the categorical judgments constitute anessential condition of experience.36It is clear, I hope, that I have been able to do no more than point in thedirection that a full Kantian account of the unity of judgment would have totravel. It would be no small undertaking. It should further be borne in mind,in all fairness to Dickerson, that it is probably not a project that can beundertaken wholly within the confines of exegetical history. Kant providessome surprising fertile materials to work with, but it is equally clear that hedid not fully appreciate the tensions inherent in his attempt both to retainprinciples of traditional logic and to justify his claim that we cantrace allactions of the understanding back to judgmentsand not originally toconcepts.37484 T. RosenkoetterDownloaded by [University of Stellenbosch] at 12:31 08 May 2013Notes1. A variation on a Kantian theme is Dickersons term (2). The rest of the quotes arefrom Sleigh (1990:4), where the typical sort of interpretation that the TD receives is citedas a paradigm case of unacknowledged failure: lucid where Kant is lucid, degeneratingto mere paraphrase just where one most wants help.2. On this point see the so-called Stufenleiter passage (KrV A320/B376). I will use thetranslations of Kants works in the Cambridge Edition of the Works of Immanuel Kantand cite according to the following abbreviations: KrV 5 Critique of Pure Reason(according to the Guyer-Wood 1998 translation); MAN 5 Metaphysical Foundations ofNatural Science (Friedman 2002 translation, as contained in Theoretical Philosophy after1781); Jasche Logic (Young 1992 translation, as contained in Lectures on Logic); andindividual metaphysics by name (Ameriks-Naragon 1997 translation, as contained inLectures on Metaphysics). Reference will be made, where appropriate, to the standardDeutsche Akademie der Wissenschaften edition of Kants gesammelte Schriften (Berlin:Walter de Gruyter, 1902ff) according to the pattern Ak. 29: 276.3. KrV A77/B103.4. It conflicts with Kants account of analytic judgments if they do not (or, perhaps, neednot) refer to objects. I argue for the former, more stringent claim in Rosenkoetter(Forthcoming 1).5. Cf., e.g., KrV A123; B131 ft. in connection with 116; and Metaphysik Mrongovius 29:889.6. Surprisingly, Dickerson does not even come close to addressing this nest of issues, so thereader is left to guess how he might try to make sense of (e.g.) the thinking of a conceptfor which no intuition can be available (i.e., a real-impossibility). Perhaps, in line with apassage from Kants letter to Beck (July 3, 1792, Ak. 11: 347), the concept could beconsidered as subjectively given with respect to this particular act of thought (becauseits constituent concepts [Merkmale] have been made available by prior acts of thatsubjects objective cognition). Then, though the concept could not be considered amodification of sensibility (i.e., the result of direct causal impingement), it could betreated as a modification of our capacity for representation [Vorstellungsvermogen] andDickerson could hold that in apperceiving it we refer it to an object5x. This might be apromising line for Dickerson to take, though it should be noted that the seeing in-analogy plays no role in cases such as these in which there is no intuition to explain thereferring of the representation to an object.7. KR&O 89, italics added.8. Guyer (1987:141).9. The clearest instance of Dickerson supplying the ownership rationale for Kantsassociation of apperception and self-consciousness is to be found in the way he contrastshis position with Andrew Brooks, which he cites as the interpretation of apperception thatis closest to the one he is offering. Brook writes that apperception is the process offorming objects of awareness (1994:37). The obvious difference between their readings isthat Brook finds Kants association of apperception and self-consciousness to beunmotivated and misleading. Dickerson diagnoses the source of (what he regards as)Brooks error as follows: Brook fails to take Kants representationalism seriously, andtreats him as being some sort of direct realist. In other words, Brook thinks that for Kantrepresentations are states of awareness, rather than the immediate objects of awareness.Brook is thus not in a position to see that for Kant all cognition must involve the reflexivegrasp of our own representations, and thus self-awareness of a special sortnamely,apperception (KR&O 96, final italics added; cf. also 89, quoted above, and 97).10. While the ownership rationale cannot be supported by exclusive attention to picturessince the representational medium of a picture is in general something other thanKant on Apperception and the Unity of Judgment 485Downloaded by [University of Stellenbosch] at 12:31 08 May 2013oneselfDickerson tries to provide something like a free-standing argument for thepoint of view rationale in the case of pictures, only then exporting that argument to thecase of Kantian objective cognition. Unfortunately, Dickerson does not clearlydistinguish the point of view rationale from the ownership rationale. I make a pointof separating them not just because they are different, but also because, as I will nowargue, Dickersons case for ascribing an awareness of point of view to Kantian subjectsis flawed. This leaves him with only the ownership rationale.11. I lack the space to discuss this central pillar of Dickersons background theory. Notethat it is a reductive account of objectivity in the sense that agreement between allpossible finite subjects is by itself sufficient for objectivity. This raises all sorts ofinteresting questions, including why then the principle of non-contradiction andjudgments of taste are not objective in the same sense. I argue in passing against areductive account of objectivity in Rosenkoetter (Forthcoming 3), and also provide analternative to Dickersons associated interpretation of Kants objection to Berkeleyanidealism (cf. p. 59).12. It might be argued, alternatively, that apperceiving subjects will be aware that theircognitions are a point of view on objects because they can switch back and forth betweenawareness of their sensations as sensations and cognizing those same sensations asobjects. Whatever the merits of this suggestion, it is most definitely not available toDickerson, for a pervasive element of KR&O is its identification of consciousness (orawareness) with cognition (or awareness of an object). This is also, it must be said, one ofits more unsatisfactory features, for a clear argument for the equivalence is neverprovided. Indeed, the reader who is new to Kant could read the book attentively withouthaving any clue that there is a real question as to whether (objective) cognition is anaccurate gloss for conscious thought (KR&O 90). Inexplicably, Dickerson never evencites the locus classicus for the division of the genus Vorstellung, the Stufenleiter-passage(KrV A320/B376), in connection with this issue. This is all the more damaging becauseaccording to that text there are representations that are perceptions (representationswith consciousness) and yet are not objective perception[s] (cognitions), namely:sensations.13. It is a great virtue of Dickersons model that it can explain nicely Kants penchant forpresenting transcendental idealism as the doctrine that objects are mere representa-tions (e.g., A369). We can read Kant as really meaning that objects are what we cognizein our representations (in our sensations). To the extent that this strategy can succeedwith a number of such passages, this will be an interpretative coup, since they seem atfirst to point univocally toward phenomenalism.14. The distinction between first-person and third-person viewpoints shows up as thedistinction between: (i) the single point of view of all finite subjects, for whom aparticular sensation is object O; and (ii) a point of view from which that sensation ismerely a modification of the subject and the function which maps it to O is merely onefunction among many possible functions, none of which has intrinsic claim to objectivevalidity. The objects in view from (i) are phenomena. To attempt to cognize from (ii) isto judge about things [Dinge] in general. This is the project which bears the proudname of an ontology (KrV A247/B303).15. KrV A354.16. While this approach brings with, as Dickerson is able to show, some advantages in theinterpretation of 111620, it coheres far less well with 115. This is made all the moreimportant, I would argue, because in 115 Kant seems to be telling us why he isinvestigating apperception in those later sections. It frames the sections whichDickerson interprets. 115 is, as its title indicates, on the possibility of combination[Verbindung] in general. We learn that combination is the representation of thesynthetic unity of the manifold. As such, understanding the possibility of combination486 T. RosenkoetterDownloaded by [University of Stellenbosch] at 12:31 08 May 2013will require understanding: the manifold, its synthesis, and its unity. Kant focuses on thelatter: The representation of this unity cannotarise from the combination; rather, bybeing added to the representation of the manifold, it first makes the concept ofcombination possible. The unity in question is (we learn first) not the category of unitybut (as we learn in 116) original apperception. Now, the purpose of this recitation is topoint out that Kant seems to find it helpful (and quite possibly necessary) that the threeelements of combination be distinguished. On Dickersons reading, however, combina-tion, synthesis, and this original unity (apperception) are indistinguishable.17. Cf. esp. KrV A546/B574.18. Cf. KrV A68f/B93f.19. To relate this point to the terms of the above discussion: whereas there is no reason tothink that awareness of divergent interpretations belongs essentially to the ability to seesomething in a picture, it is plausible that a subject is not asserting if she does not realizethat what she is judging true might be false.20. KR&O, 107 and 1 (italics added in the former and elided in the latter).21. The following discussion draws on both Hylton (1984) and Linsky (1988).22. Hylton (1984:382).23. Dickerson quotes relevant texts from both Russell and Frege, but does not address thetwo versions of the problem separately.24. What is important if we are engaged in exegetical history is that there is no reason tothink that Kant was bothered by the representationalist parallel of the problems thatbedevil Freges concept-object approach to propositional unity. What are thoseproblems? The immediate cost of Freges progress over RussellRussell is unwillingto bear this cost and demursis that Frege must admit that we cannot judge aboutfunctions, for doing so would require that we be able to convert them to objects, and ifthat is possible then we find ourselves facing the question that Russell could not answer.Whats more, even if one is willing to accept this consequence, further problems arisewhen we try to use the metaphors of completeness and incompleteness to explain why afirst-level concept and a second-level function can together form a unified proposition.All that can be said is that the latter is being completed (or saturated) by an incomplete(unsaturated) thing. As Linsky remarks, It is as though putting one unsaturated spongewith another would produce a saturated pair of sponges. The metaphorscast no lightwhatever on this second-level case (1988: 265). Now, two counterexamples to my claimthat Kant did not confront these problems suggest themselves: existential judgments andanalytic judgments. It could be argued, namely, that Kant understood analyticjudgments as judgments about concepts. I argue against this reading in Rosenkoetter(Forthcoming 1). Second, existential judgments would indeed require a second-levelconcept if Kant had handled existential judgments in Freges manner as the predicationof the second-level property of a first-level concept. I argue that Kant has a differentanalysis of existential judgmentsan analysis which avoids second-level conceptsaltogetherin Rosenkoetter (Forthcoming 2).25. It would be superficial to pretend that the Stufenleiter classification of concepts asrepresentations answers this question. This is not primarily because Kant may be usingterminology loosely. The deeper issue is that Kant, in contrast to Russell and Frege,understands all of the entities at issue (concepts, intuitions, and that which results fromtheir unification) as belonging to the subject rather than to an independent ontologicalrealm. Consequently, they have at least that subjectivity in common, and the question oftheir difference is less stark than is the parallel issue between Frege and Russell, whichLinsky describes in terms that are felicitous for our discussion: For Frege, there is noconcept corresponding to Russells termsthere is no single widest word in thephilosophical vocabularyThere is nothing which both functions and objects are(1988: 247). Note that the very classification of concepts and intuitions as bothKant on Apperception and the Unity of Judgment 487Downloaded by [University of Stellenbosch] at 12:31 08 May 2013belonging to the genus of representation is plausibly a classification from (what wasabove termed) a third-person standpoint, the standpoint of ontology. Both aremodifications of the subject as opposed to something independent of the subject. It isworth considering whether from the standpoint of first-person theory (i.e., transcen-dental philosophy) there is anything which both concepts and intuitions are.26. KR&O 143, italics added. As far as I can see, Hylton is making no hiddenpresupposition of representational atomism, i.e., an assumption in his text that isnot captured in my gloss of that general position. Consequently, it is difficult to read thisas anything other than the stubborn demand a` la Russell that no more than a singlecategory be recognized.27. Namely, Dickerson is failing to distinguish two varieties of holism. Hyltons Kant cansubscribe to a weak variety by holding that a subject cannot understand G, and cannotuse G in analytic judgments, if it has not first cognized an object using G or itsconstituent concepts, an act which requires intuition. In this weak sense the contributionof a concept G to an objective cognition is indeed not specifiable independently of thewhole. Yet it seems clear that Kant took this bow towards holism to be entirelyconsistent with claiming that the contents of these acts (concepts) are in some sensespecifiable independently of particular objective cognitions. This raises interestingphilosophical issues, but I cannot assess here how problematic his account of conceptualcontent is.28. One is reminded that Russell circa 1903 thought of analysis as almost analogous tophysical decomposition (Hylton 1984: 376). Dickersons attribution of representation-alism to Kant has the effect of turning sensations into objects from the third-personperspective of ontology. Dickerson defends his Kant against the charge of reifyingrepresentations by arguing that since sensations are not entities that could existindependently of the mind they are not object[s] per se (KR&O 8). The category ofobject per se is wider than this.29. Jasche Logik 117, at p. 597 in the Young translation.30. KrV B140.31. KrV B141; following quote, B142.32. MAN Ak. 475f, at p. 190 in the Friedman translation.33. I attempt this in Rosenkoetter (Forthcoming 1). See also David Bells brief comments onso-called analytic judgments in (2001: 9).34. We could instead say that the remaining intensional content is nothing more than the apriori intensional content provided by the concept of substance. That just shows that thecategories are not like other, non-formal concepts. They cannot be understood bygrasping an intensional content that is separable from this formal feature of judgments.On the current discussion, see KrV A147/B186f and MAN Ak. 475, at p. 189 in theFriedman translation.35. Metaphysik L2 Ak. 28: 563, at p. 328 in the Ameriks-Naragon translation.36. Metaphysik Mrongovius, Ak. 29: 770, at p. 178 in the Ameriks-Naragon translation,italics added.37. KrV A69/B94.ReferencesBell, David. (2001) Some Kantian thoughts on propositional unity Aristotelian SocietySupplementary Volume LXXV: pp. 116.Brook, Andrew. (1994) Kant and the Mind (Cambridge: Cambridge University Press).Frege, Gottlob. (1969) Translations from the Philosophical Writings of Gottlob Frege P. Geach,et al., (Eds) (Oxford: Blackwell).488 T. RosenkoetterDownloaded by [University of Stellenbosch] at 12:31 08 May 2013Guyer, Paul. (1987) Kant and the Claims of Knowledge (Cambridge: Cambridge UniversityPress).Hylton, Peter. (1984) The Nature of the Proposition and the Revolt against Idealism inR. Rorty , et al., (Eds) Philosophy in History (Cambridge: Cambridge University Press).Linsky, Leonard. (1988) Terms and propositions in Russells Principles of MathematicsJournal of the History of Philosophy XXVI: pp. 621642.Rosenkoetter, Timothy. (Forthcoming 1) A Puzzle for Kants Theory of Analytic Judgement.Rosenkoetter, Timothy. (Forthcoming 2) On Kants alleged anticipation of Freges account ofexistence as a second-level concept.Rosenkoetter, Timothy. (Forthcoming 3) Truth criteria and the very project of atranscendental logic..Russell, Bertrand. (1903) The Principles of Mathematics (Cambridge: Cambridge UniversityPress).Sleigh, R. C., Jr. (1990) Leibniz and Arnauld: a Commentary on Their Correspondence (NewHaven: Yale University Press).Kant on Apperception and the Unity of Judgment 489Downloaded by [University of Stellenbosch] at 12:31 08 May 2013


View more >