Kant's a priori history of metaphysics: Systematicity, progress, and the ends of reason

This paper explores Kant's conception of the relation between philosophy and its history. The idea that philosophy must account for its historical development is often associated with German Idealism. On the traditional view, the German Idealists departed from the ahistorical Kantian framework by conceiving of reason in a developmental manner, thus initiating a “ historical turn ” within philosophy. How-ever, I argue that Kant's sketch of the history of metaphysics in the final chapter of the Critique of Pure Reason and in the so-called Progress Essay anticipates this turn. Specifically, I argue that in this essay Kant articulates a developmental conception of reason as a basis for understanding the history of metaphysics. My aim is to change the prevailing view of Kant as an ahistorical thinker and to challenge the canonical account of the genesis of the historical turn in philosophy. an a priori plan for organizing the body of knowledge of past philosophies. Second, this schema not be static, as on the Fülleborn model, but should be able to account for the succession of distinct historical

Instead, what is required for a genuinely historical conception is that reason itself rather than its manifestations in the empirical world be conceived as subject to development throughout time. According to Breazeale, this is precisely what Reinhold's "astonishingly original conception" of the "historicity of reason" consists in (Breazeale, 2010, p. 90). Namely, Reinhold "defends a thoroughly historical conception of the 'interests" or 'needs" of reason as changing over time and in different circumstances" (2010, p. 90). To the extent that such needs and interests are constitutive of reason, and that these needs and interests develop (qualitatively) over time, Reinhold (like Fichte, Hegel, and Schelling after him) conceives reason itself in developmental terms.
For Kant, on the other hand, the rational is the realm of the a priori, necessary, universal, and atemporal. This same a priori status attaches also to the needs and ends of reason, which thus transcend any given historical epochas Kant states, these know no temporal or national differences (P 20:321). If Reinhold's historical turn consisted in temporalizing the needs of reason so as to make them responsive to given national-historical circumstances, then Kant's a priori framework seems particularly unsuited for this task. Consequently, most commentators find a developmental conception of reason to be incompatible with the tenets of Kant's critical philosophy. For example, Yovel (1980, p. 21) concludes that "no mediation between reason and empirical history is conceivable" within the Kantian framework. Though he can accommodate empirical developments in morality or religion, Kant is thus taken to defend an ahistorical conception of reason (see, among others, Breazeale, 2010, pp. 91-97, Ameriks, 2006, p. 5, or Beiser, 2011. Consequently, Ameriks concludes that a conventionally rationalist style, together with the ahistorical view of reason that underpins it, mean that Kant is "not committed to the historical turn" (Ameriks, 2012, pp. 144, 147).
The third and final characteristic of the historical turn is a closer attention to theoretical issues related to the history of philosophy and, ultimately, the articulation of a historiographical model able to capture the development of philosophy throughout history. Ameriks argues that also in this respect Reinhold was the pioneer because he was "the first to devote an essay to a genuinely modern and philosophical treatment of the nature of the history of philosophy itself" (Ameriks, 2006, p. 9). The work in question is not the Letters but the 1791 essay "On the Concept of the History of Philosophy", in which Reinhold attempts to define "the history of philosophy" and to articulate a theoretical model for capturing this history. Though Ameriks admits the essay is sketchy in some of the crucial details, and that at times it relies too heavily on atemporal elements, he nonetheless considers Reinhold's essay to be the "best example from this period of a philosophical approach that makes history central" (Ameriks, 2006, p. 25). Consequently, the essay holds an "epochal status as a first attempt at self-definition by the first era of serious history of philosophy" (2006, p. 204). Thus, the final mark of Reinhold's historicisation of philosophy is his willingness to seriously engage with past thinkers and to reflect theoretically on the nature of this engagement.
Kant, on the other hand, is said to be "condescending in an obvious way" to his historical predecessors, believing that "the grand validity of [his] own secure and eternal results would be ill served by stooping to distract readers by offering a detailed account of the specific arguments of earlier thinkers" (Ameriks, 2006, p. 10). As Pippin puts it, because he believed himself to be uncovering the universal and atemporal nature of human reason, "the history of philosophy is not important to Kant"-indeed, from the secure perspective of pure reason, a Kantian history could be nothing more than "a history of uncritical dogmatism and errors" (Pippin, 2015, p. 4). However, as suggested above, it is not merely a question of a willful decision on Kant's part to take up an attitude of condescension toward his predecessors. Rather, Kant's allegedly ahistorical conception of reason is seen as not only rendering a historical approach to philosophy unnecessary, but also as representing an obstacle that made it in principle impossible to adapt the critical framework to a fruitful exploration of philosophy's past. Thus, Grandjean argues that because pure reason is atemporal it is inconceivable that reason have a history in any genuine sense; to the extent, furthermore, that philosophy is the expression of reason par excellence (as Kant believes that it is), a Kantian history of philosophy would be "a history, that is not a history at all" (Grandjean, 2017, p. 218). Brandt and Yovel both highlight the apparent incoherence of attempting to compose such a history, arguing that any attempt to shoehorn pure reason into an empirical history of philosophy will necessarily end in failure due to reason's atemporal nature (Yovel, 1978, p. 116;Brandt, 2017, p. 198).
Given the assumption that Kant's atemporal conception of reason does not leave space for a philosophical approach to the history of philosophy, it should be somewhat surprising that in jottings pertaining to drafts of the so-called Progress essay, composed around 1793, Kant asserts that "a philosophical history of philosophy is itself possible, not historically or empirically, but rationally, i.e., a priori" (P 20:341). 2 This assertion cuts right to the heart of the issue: the possibility of an a priori, rational treatment of the history of philosophy seems to presuppose a developmental conception of reason itself. The idea of a Kantian a priori history should not be completely unexpected, however. First, Kant's historical writings from the critical period betray not only an interest in the philosophy of history, but also an attempt to come to terms with the possibility of grasping historical developments from an a priori, rational perspective. Second, in the first Critique Kant had already claimed that a history of pure reason represents "a place that is left open in the system and must be filled in the future" (A852/B880). Though Kant did not wish in this brief sketch of the history of pure reason to "distinguish the times in which this or that alteration of metaphysics occurred" (A853/B881), and so presented philosophical positions in abstraction from their historical development, the above assertion nonetheless provides a promissory note for the future execution of such a history.
Given its focus on the history of metaphysics, which represents the highest expression of pure reason, the Progress essay seems to be the closest Kant comes to cashing in this promissory note. A serious obstacle to reconstructing Kant's account, however, arises due to the composition and publication history of the essay. Kant drafted the essay as a response to a prize question set by the Royal Academy of Science and Letters in Berlin on the topic of the progress that metaphysics has made since the age of Wolff and Leibniz. The essay contest was first announced in 1790 with a deadline in 1792, which was later extended to 1795. For unknown reasons, Kant never submitted the essay and left the text unfinished. He passed the various drafts and jottings pertaining to the essay to his friend Rink around 1800, who edited and published them several months after Kant's death in 1804.
The incompleteness of the text, together with Rink's questionable skills as an editor, poses significant difficulties in reconstructing Kant's a priori history of metaphysics. Consequently, the few commentators that have addressed the topic view Kant's reflections on history as excessively sketchy as to yield a coherent account or as inconsistent with the critical framework. The difficulty relates primarily to the question of how precisely to understand the developmental stages of metaphysics that Kant identifies. Brandt, for example, writes that problems in the "conception and execution" of this aspect of the work were ultimately the cause of Kant's failure to submit the essay (Brandt, 2017, p. 198). 3 The alleged incoherence of an essay that represents Kant's most detailed treatment of history thus serves to reinforce the prevailing view of Kant as an ahistorical thinker and to sharpen the divide that is seen to separate Kant from his German Idealist successors on the topic of philosophy's relation to its history. 4 In my view, however, closer attention to Kant's drafts and jottings pertaining to the Progress essay is reveals an outline for a coherent model of an a priori history of philosophy underpinned by a developmental conception of reason. In what follows, I reconstruct Kant's model based on the jottings and examine its application to the history of metaphysics in the drafts of the essay. By doing so, I aim to bring out the continuity between Kant and his successors on the question of the relation between philosophy and its history. My plan is as follows: In the first part I discuss the relevant architectonic considerations from the Doctrine of Method of the first Critique to contextualize Kant's undertaking in the essay. In the second I examine Kant's model for an a priori history of philosophy from the jottings in light of these considerations. In the third part I reconstruct Kant's history of metaphysics from the drafts of the Progress Essay and assess the extent to which this history instantiates Kant's proposed model for an a priori history. I end with a discussion of the epistemic status of the ends of reason in Kant's account in relation to Reinhold's, with the aim of establishing Kant's relation to the "historical turn" based on its three defining characteristics I listed above.

| ARCHITECTONIC CONSIDERATIONS
There are several forms that a philosophical history of philosophy could take within the strictures of Kant's critical system. Kant often defines historical cognition as cognitio ex datis, or cognition from that which is empirically given, and contrasts this with rational cognition, which is ex principiis, or from principles (A836/B864). On this definition of historical cognition, an empirical history of philosophy could consist of a chronology of the philosophical doctrines that were put forth at a particular time along with a description of their influence-in short, such a history would focus on the context of discovery and transmission of philosophical ideas.
However, Kant sharply rejects such an endeavor as unphilosophical. A philosophical history is instead of "a special kind" in that it deals with "reason developing itself from concepts" (P 20:343). Because this development is rational, according to Kant, a philosophical history must strive to do significantly more than merely give an account "of the opinions which have chanced to arise here or there" (P 20:343). Kant's claim is thus that the philosophical historian cannot merely recount the succession of philosophical ideas chronologically as they arose. Instead, the historian must impose an order onto the history of philosophy that could in some sense track the development of reason within the successive presentation of philosophical doctrines. In other words, the collection of historical data must be considered to constitute a body of knowledge organized around principles that are provided by the historian.
Kant's name for such a body of knowledge is "system" (see A833/B861 and MFNS 4:467). Consequently, a philosophical history of philosophy must be systematic. Furthermore, Kant holds that systematizing a body of cognitions requires a schema, that is, a sketch of the form that such a body of cognitions should take. Kant defines a schema in this sense as "an essential manifoldness and order of the parts [of a system] determined a priori from the principle of the end" (A833/B861). The "end," presented schematically, sketches in advance the "form of the whole," allowing, for example, "the absence of any part" to be noticed and ruling out any contingent additions (A832/B860). 5 Kant's metaphysics of corporeal nature as presented in the Metaphysical Foundations of Science offers an example of such a system. The particular idea underpinning this system, Kant writes, is that of "all determinations of the general concept of matter" . This concept is an idea because it contains a representation of a totality, in this case of the possible a priori determinations of matter. This representation also serves as the principle of the end, because it tells us what is to be achieved. The schema, on the other hand, tells us how the end is to be achieved.
In this case, Kant tells us, the schema is the table of categories, which sketches in advance the plan for the construction of the system and so can be used as a guiding thread. Specifically, each set of categories allows the philosopher to approach the general concept of matter from a particular angle and, in this way, to elaborate its necessary determinations in a systematic manner. This schema not only determines the structure of the system, but also guarantees its completeness-as Kant tells us, "there is no more to be done, or to be discovered, or to be added here, except, if need be, to improve it where it may lack in clarity or exactitude" (MFNS 4:476).
Because a philosophical history of philosophy needs to be systematic, and because a system requires a schema in the above sense, in the jottings related to the Progress Essay Kant raises the question of whether a schema can be provided for the history of philosophy with which "the epochs and opinions of the philosophers so coincide, that it is as if they had had this very schema themselves before their eyes, and had progressed by way of it in knowledge of the subject" (P 20:342). Based on the principle of the end, such a schema would impose a unified order onto the totality of philosophical doctrines developed throughout the empirical history of philosophy.
In the first Critique, Kant states that a schema, furthermore, can be outlined either a priori in accordance with necessary principles, as in the case of the metaphysics of nature, or empirically in accordance with "aims occurring contingently" (A833/B861). In the latter case, the given body of cognitions acquires a merely "technical unity" rather than a properly "architectonic unity" (A833/B861). The upshot of this distinction is that science in the proper sense cannot arise technically, but must have an architectonic unity (A834/B862). A body of cognitions organized around empirical principles, Kant tells us in the Metaphysical Foundations, does not constitute a "proper science," or is only knowledge (Wissen) "improperly so-called" (MFNS 4:469). An example of an improper science, for Kant, is chemistry, because its "principles are merely empirical" (MFNS 4:468).

| KANT'S MODEL IN THE JOTTINGS
As regards the history of philosophy, the question is thus whether this body of knowledge allows for an architectonic unity based on a priori elements or for merely a technical unity. In other words, the question is whether Kant considers it possible to develop a history of philosophy that is properly scientific. The answer in the jottings seems to be resoundingly in the affirmative. Regarding his question of whether "a philosophical history of philosophy is itself possible, not historically or empirically, but rationally, i.e., a priori" (P 20:341), Kant responds with an uncharacteristically exclamatory "Yes!": he believes that the science of metaphysics "lies wholly prefigured in the soul, albeit only in embryo" (P 20:342). In other words, at least in the case of metaphysics, a philosophical history establishes "facts of reason," which it does not "borrow" from a historical narrative but "draws from the nature of human reason" (P 20:341). We can see that the term "history" here can no longer refer to that which is given ex datis, and perhaps for this reason Kant instead christens such an endeavor as "philosophical archaeology" (P 20:341). REICHL Though Kant affirms the possibility of an a priori and thus scientific history of philosophy, the question of the specific nature of its principles, and thus of its ultimate form, remains. Kant provides a somewhat cryptic clue when he writes that such a history occupies itself not with "what has been reasoned out" but instead with "what reasoning has achieved (ausgerichtet) through mere concepts" (P 20:343). In the case of metaphysics, Kant writes in the drafts that what is important is what can be achieved (ausgerichten) by reason when it is directed ( gerichtet) toward its final end (P 20:296). Presumably, this means that an a priori history should not focus solely on the explicit content of philosophical concepts, but also on the functional contribution of the concepts toward the final end of metaphysics. This seems to be confirmed by Kant's further claim that the history of philosophy, to the extent that it is rational, cannot recount what has happened "without knowing beforehand what should have happened, and also what can happen" (P 20:343). The claim is thus that the history of philosophy must be constructed with an eye to what paths reason was able to take and what paths it should have taken toward its goal. In other words, such a history would not only be about what a given philosopher said, but also about what he or she should have and could have said in light of the final aim of metaphysics.
Because, according to Kant, we can know a priori what is possible for reason in light of its end, we have available an a priori standard against which we can classify the development of reason through the successive presentation of philosophical doctrines. In the context of this discussion in the jottings, Kant mentions the name "Fülleborn," without further explanation (P 20:343). Given the lack of context it is difficult to ascertain what precisely Kant had in mind with this reference, but it is nevertheless instructive to briefly examine Fülleborn's model as a potential historiographical option on the post-Kantian philosophical landscape. As a historian of philosophy confronted with an ostensibly chaotic succession of past doctrines, Fülleborn recounts that what was needed was "a sure principle according to which we could evaluate and arrange" these doctrines. Inspired by Reinhold's application of Kantian principles to history in the Letters, 6 Fülleborn developed a model for an a priori history of philosophy that employs Kant's analysis of reason in the first Critique as the a priori standard for constructing the history of philosophy. Namely, past doctrines could be classified under the various wrong turns that reason takes in attempting to achieve cognition of unconditioned objects, as diagnosed by Kant in the Dialectic. In this manner, when put into a systematic form Kant's analysis of reason could provide "a priori a complete vision of the whole of all possible philosophies, hence of all the history of philosophy" (Fülleborn, 1791, p. 57).
Such a model would fulfill Kant's joint criteria that the philosophy of history be a system and that such a system be organized by a priori principles. 7 Nonetheless, it would be somewhat of a disappointment if this were Kant's intended model. Fülleborn's model, namely, retains an ahistorical conception of reason, in which both the ends of reason, as well as the errors into which reason falls into in its striving after these ends, "have in all times been the same" (Fülleborn, 1791, p. 55). The negative consequences that result from this conception of reason for Fülleborn's historiographical model are twofold. First, to the extent that reason is non-developmental, and to the extent that the development of the history of philosophy is merely empirical, it is not possible to understand the nature and sequence of historical development philosophically. The result is a static snapshot of past doctrines that are understood in terms of atemporal options of reason. Second, all past doctrines become failed attempts to satisfy the eternal needs of reason, and so Fülleborn's model becomes a sort of error-theory in which all historical doctrines are strictly speaking false. The resulting picture would lend credence to the criticisms of Kant's own historiographical model that, because there was no genuine philosophy before the Critique, a history of philosophy based on Kant's static conception of reason can be nothing more than catalog of errors that displays no genuine historical development.
However, Kant's jottings indicate he had a more ambitious conception of history in mind. Kant's philosophical history of philosophy was apparently to be based on a schema that maps not only the totality of a priori possible philosophical theories, but also their historical sequence. A common strategy in the philosophy of world history during Kant's time was to employ a stage-based model in order to provide a developmental account of the progress from one historical period to the next (see Mah, 2002, pp. 147ff.). An example of this is Fichte's epochs of world history, each of which is derived from a developmental stage of reason. History is teleologicaly structured based on a final end, which for Fichte is the "dominion of reason through freedom" (see Reichl (2020b). However, there are several "intermediary conditions" on reason's attainment of this end, and each of these are indexed to a particular historical epoch. Thus, for example, a condition on reason achieving dominion through freedom is the liberation from the dominion of reason through instinct, which corresponds to what Fichte calls the epoch of liberation in human history. The point to take away from this kind of model is that reason itself is conceived as developmental in the sense that it must necessarily pass through several temporally ordered stages, each of which directly results from the preceding one and prepares the ground for the succeeding one. 8 Kant's division of philosophy into stages seems to follow suit. In the Progress Essay, Kant asserts that the "temporal sequence" (Zeitordnung) of the stages of the development of philosophy is "founded in the nature of man's cognitive capacity" (P 20:264). In the jottings Kant asks accordingly "whether a history of philosophy might be written mathematically; how dogmatism must have arisen, and from it skepticism, and from both together criticism" (P 20:342). Such a history would not merely classify a given philosophical doctrine under the general typology of dogmatism and skepticism (or criticism) in the manner of Fülleborn, but also show that (and perhaps why) the one "must have" arisen from or succeeded the other. On this picture, the history of philosophy would account not only for the form of the complete system of all philosophies but also for the necessity of their temporal succession. This conception of the philosophy of history would thus depart from Fülleborn's static model and anticipate, to some extent, the Hegelian model, which carries over the basic form of the Fichtean developmental concept of reason from world history to the history of philosophy.
This Hegelian model can be characterized in terms of a twofold criterion. First, the temporal order of philosophical systems must exhibit some form of a priori necessity, and second, each such system must be a part of an overarching meta-system (see De Boer, 2017, p. 630). The first criterion points to the afore-mentioned dynamical principle that would account for the temporal order of historical doctrines. The second would solve the problems associated with the conception of past doctrines in terms of an error-theory, because these could now be conceived as (necessary) steps leading up to the establishment of the Critical philosophy rather than as blind alleys. While an interpretation that aligns Kant with Hegel on this point may seem fanciful, some commentators understand Kant's intentions in this manner. For example, Yovel (1978, p. 120) argues that for Kant "all the important doctrines in the history of philosophy are… tacit members of one systematic whole, each stressing… special aspects of the final pure system". Indeed, Kant himself gives this impression when he writes in the first Critique that past philosophical systems are to be "purposively united with each other as members of a whole in a system of human cognition" (A835/B863).
The danger with this model, however, is that it threatens to prove too much. Kant worries about the question of how it is possible to "bring a history into a system of reason, which requires the contingent to be derived, and partitioned, from a principle" (P 20:342). This problem is analogous to the problem of accounting for the lawfulness of the contingent with respect to organized matter in the third Critique. As is clear from Kant's treatment there, subsuming the contingent under necessary laws would amount to the abolition of contingency, and, more importantly, a constitutive principle which would allow such a move is beyond our cognitive ken. Kant would thus say that Hegel's model is only possible on the condition of making a teleological principle constitutive, which Kant rules out. The question is thus whether Kant himself is able to provide a philosophical model for the history of philosophy that is both based on an a priori schema and consistent with the critical framework in general.

| THE STAGES OF METAPHYSICS
The jottings thus present several conditions that a history of philosophy should comply with. First, as was discussed above, Kant holds that such a history must be systematic and scientific, which means that there must be a schema capable of providing an a priori plan for organizing the body of knowledge of past philosophies. Second, this schema should not be static, as on the Fülleborn model, but should be able to account for the succession of distinct historical doctrines or stages. Third, the underlying schema needs to remain within the bounds of the critical philosophy. However, it is unclear how this succession of doctrines can be comprehended rationally, and hence a priori, without lapsing into a necessitarian model of the Hegelian stripe.
To answer these questions I now turn to Kant's sketch of the history of metaphysics on the pages of the drafts for the Progress Essay. Because Kant's schema here indeed relies on elements from the first Critique, another way to put the question is whether the framework of the first Critique can be put into a dynamic form for historicalphilosophical purposes. Though, as mentioned, the history of pure reason presented at the end of the Critique is sketchy and abstracted from empirical history, Kant earlier in the work hints at a stage-based, developmental history of reason by equating Dogmatism with childhood and Criticism with adulthood (A761/B789). As I shall show, Kant's Progress Essay fleshes out this schema by focusing on a more detailed presentation of these stages of development.
The question is thus whether the history of metaphysics Kant there presents is sufficiently filled out so as to yield a tenable schema for a philosophical history of philosophy capable of fulfilling the desiderata of the jottings.
Kant's first step in adapting the conceptual framework of the first Critique for the purposes of a philosophical history of philosophy is to make a nuanced but important modification to the definition of metaphysics. Whereas in previous works metaphysics was referred to in more static terms as a "science" or a "system" of pure reason (Prol 4:371; CPR AXXI), Kant now characterizes it as a teleologically structured activity directed at a final end (Endzweck).
More specifically, metaphysics is defined as "the science of progressing (fortschreiten) by reason from cognition of the sensible to that of the supersensible" (P 20:260). Sensible here refers to "that which can be an object of experience" (P 20:316). Transcending the sensible in order to reach the supersensible is called the final end "to which the whole of metaphysics is directed" (P 20:260). However, due to the "immeasurable gulf" (P 20:319) between the sensible and the supersensible, metaphysics can reach its end neither by a "constant and unending" progress like the empirical sciences (P 20:259) nor by a gradual approximation (20:299) as in the case of the historical development of morality toward the highest good (RHI 8:65). Metaphysics does not progress toward its goal in steps (fortschreiten) but rather through a sudden leap that must transcend (überschreiten) the whole of the sensible (P 20:317).
From the view that metaphysics is not "a continuous progression in the same order of principles" follows, Kant tells us, its division into stages (P 20:273). Kant thus takes the above definition to entail that metaphysics is a process of stepping over toward the supersensible and to justify its division into stages. Presumably, the stages of metaphysics represent distinct ways of moving from the sensible to the supersensible, where the orders of principles represent distinct methodological contexts for the execution of this move. 9 A closer look at Kant's account of these stages of metaphysics should thus tell us which order of principles each stage employs and provide an understanding of both the reasons for the failure of a given stage and of the logic of transition to a subsequent one.
Providing a description of the external form of the historical stages of metaphysics is a relatively straightforward task. Not only does Kant employ the well-known schema Dogmatism-Skepticism-Criticism, but he indexes each stage to a well-known section of the first Critique. Kant's description of the first stage, Dogmatism, deals with several aspects of Leibniz's philosophy, namely the principle of identity of indiscernibles, the principle of sufficient reason, the system of pre-established harmony, and the monadology. Together, these are said to constitute "the new element" that Leibniz "attempted to introduce into the metaphysics of theoretical philosophy" (P 20:285). Kant sharply rejects these principles, claiming that they are among the "strangest figments ever to be excogitated by philosophy" (P 20:284) and set up a "sort of enchanted world" (P 20:285). The reason for such a judgment is familiar: Leibniz lacked the "principle of intuition a priori" (P 20:282), which means that everything was to be "explained and made intelligible by concepts" alone (P 20:284). Consequently, Leibniz is said to have "intellectualized" intuition (P 20:282), a criticism that echoes Kant's language in the appendix to the Transcendental Analytic on the Amphiboly of Concepts (A271/B327).
Kant associates the second stage of metaphysics with skepticism. During this stage, metaphysics attempts to extend its knowledge to the complete series of conditions given in experience by progressing to the unconditioned (P 20:287). Kant's presentation of these attempts and their failure tracks the four antinomies of the Transcendental Dialectic of the first Critique. As in the Dialectic, Kant concludes that the mathematical antinomies are both false, and that the dynamical ones can both be true. The dynamical antinomies open up the possibility of a practical cognition of freedom, and Kant dedicates some space to a discussion of the way in which freedom can provide practical access to the supersensible. His point is, however, that metaphysics at this stage realizes that it cannot achieve theoretical cognition of the supersensible, which is what metaphysics at this stage demands.
However, Kant warns that the transition to the supersensible may not proceed morally practically by way of a metaphysics of morals, for this would mean "to stray into a wholly different field" (P 20:293). Instead, the third stage of metaphysics, which Kant associates with Criticism, discusses the purposiveness of nature and so takes its bearings not from the second but from the third Critique. In his discussion of this stage, Kant alludes to the argument from the Thus concludes Kant's philosophical history of metaphysics since the age of Leibniz and Wolff. In a link that is not immediately clear, Kant further associates each stage with ontology, cosmology, and theology, respectively.
While there is a clear connection between, for example, the antinomies and cosmology in the second stage, it is initially unclear why each historical stage should be indexed to a division within metaphysics. A second, more pressing puzzle arises with respect to the second stage, that is, skepticism. It is unclear that this label can be employed within a history of metaphysics, especially given Kant's definition of the latter. The problem is that skepticism does not seem to attempt to progress to the supersensible; on the contrary, it rules out such a progression as impossible. 11 How, then, can it represent a particular stage within metaphysics?
This puzzle can be resolved if we focus on Kant's presentation of the antinomies, both in the Critique and in the Progress essay. It is clear that both parties to the conflict present their arguments dogmatically, that is, by developing them from ostensibly certain principles. When it emerges that neither side is able to get the upper hand (because both are able to construct valid arguments) the antinomy is represented from the perspective of skepticism, which attempts to resolve the standoff by "withdrawing from the quarrel" completely (A757/B785). Thus, the label of skepticism for this second stage of metaphysics is slightly misleading, as in this stage too metaphysics proceeds dogmatically. Kant affirms this point when he writes in one of the drafts to the Progress essay that the "skeptical standstill" of the second stage of metaphysics actually "contains no skepticism, i.e., no renunciation of certainty in the extension of our rational knowledge beyond the limits of possible experience" (P 20:329). It is thus not the case that metaphysics itself becomes skeptical about the possibility of access to the supersensible, but rather that metaphysics is in this stage "skeptically assailed" (P 20:294).
Stage two thus consists of a dogmatic metaphysics carried out under the assault of skepticism. This, however, raises a new puzzle. Namely, if stages one and two are both cases of dogmatic metaphysics, what is the inherent difference between them, apart from the external circumstance that one is threatened by skepticism? Here Kant's association of the first two stages with ontology and cosmology provides the clue to answering this question. Namely, as Kant writes, the first stage "proceeds solely within the bounds of ontology, the second within those of transcendental or pure cosmology" (P 20:281). Kant, furthermore, tells us why reason at this stage fails: what transpires in the first stage takes place "through understanding and judgement", whereas in the second stage metaphysics attempts to extend its knowledge "by means of reason" (P 20:286). This differentiation harks back to Kant's claim, mentioned above, that reason at each stage attempts to reach the supersensible based on a distinct order of principles. Reason's attempts would thus proceed based on principles of the understanding in the first stage, and on principles of reason in the second. Each stage would thus represent a distinct manner of attempting, and of failing, to reach the supersensible.
Again, the cause of reason's miscarriage in the second stage is straightforward, namely, the cosmological idea of the world as a whole cannot find a corresponding intuition, leaving the concept empty and unusable for purposes of theoretical cognition. Reason is thus unable to progress to the supersensible by means of cosmological concepts, and so its failure is in this stage indexed to this particular division of special metaphysics. The case of the failure of the first stage, ontology, is more complicated. Kant seems to hold that cosmology fails because its concepts are empty in the sense that they do not relate to intuition, that is, its concepts are mere ideas of reason. However, Kant showed in the transcendental deduction that at least some a priori concepts of the understanding do have a necessary relation to intuition. The concept of cause, for example, is of non-empirical provenance and yet has objective validity-is this concept thus not able to provide us with knowledge of the supersensible? If so, then the failure of (Leibnizian-Wolffian) ontology would amount the failure to provide a deduction for such concepts establishing their dependence on intuition.
However, Kant's approach to this stage follows a different path, and here a caveat with respect to the definition of supersensible becomes relevant. In line with the Transcendental Deduction, Kant notes that the sphere of the sensible includes "that whose representation is considered in relation, not merely to the senses, but also to the understanding, so long as the pure concepts of the latter are thought in their application to objects of the senses, and thus for purposes of a possible experience" (P 20:260). In this case, the a priori concepts of the understanding (i.e., the categories) remain within the bounds of the sensible because they have a necessary reference to intuition, as was demonstrated in the transcendental deduction: "thus the nonsensible, such as the concept of cause, which has its seat and origin in the understanding, can still, as regards cognition of an object by means of it, be said to belong to the field of the sensory, that is, to objects of the senses" (P 20:260). Kant thus distinguishes between the nonsensible, which includes the pure concepts of the understanding, and the supersensible; it is only cognition of the latter and not of the former constitutes the proper end of metaphysics. To the extent that ontology concerns the former, "it makes no allusion to the supersensible, which is nevertheless the final aim of metaphysics" (P 20:260). For this reason ontology belongs to metaphysics "only as a propaedeutic, as the hallway or vestibule of metaphysics proper" (P 20:260).
In the first stage ontology thus seeks to progress beyond the sensible to the nonsensible by means of concepts of the understanding, and not yet by concepts of reason, as in the second stage. Though the concepts of the understanding employed in ontology are "admittedly a priori in origin", they are nonetheless "valid only for objects of experience" (P 20:262). For this reason, the first stage is ultimately "unsuited to the final goal of metaphysics, of attempting to pass over from the sensible to the supersensible" (P 20:286). Instead, the supersensible is only "to be met with in the second stage of pure reason's attempts", which proceed by means of concepts of reason (P 20:286).
While metaphysics in this second stage reaches out toward the super-sensible for the first time it does not attain it, for the reasons given above. Kant's argument is thus that the first stage miscarries not because its concept of the supersensible is empty, but because it cannot yet disentangle itself from the sensible so as to form a genuine concept of the supersensible.

| CONCLUSION
Kant's a priori history of metaphysics, as shown above, aims to determine what reason does, what it can do, and what it ought to do at each stage of its development in light of its final end. In the first stage, reason attempts to forge a direct path to the supersensible by means of ontological principles. However, given that these principles need to be related to sensibility, reason can at this stage only produce an unsatisfactory concept of the supersensible.
There can thus be no shortcut to the supersensible by means of the understanding, and a leap based on ontological principles of the understanding through which objects of the senses are constituted is destined to miscarry; in Kant's pithy formulation, "one cannot build on the far shore with any materials of sensory representation" (D 8:213). What reason ought to do at this stage is realize that ontology is a mere vestibule of metaphysics, and transition to the second stage. In this stage, reason attempts to grasp the supersensible by means of concepts of reason, and, in the case of metaphysics of nature specifically, by means of cosmological concepts. The advantage over the previous stage is that its concepts are now purified of any sensible residue, and thus appropriate for achieving knowledge of the supersensible. However, in this case reason can only form empty concepts of the supersensible that provide no cognition of objects. Consequently, reason ought to take its concepts as merely regulative rather than constitutive, and move to the final stage. In the final stage, that of Criticism, reason achieves symbolic cognition of the supersensible by means of the teleological conception of nature as a whole. Because it contains "a totality of cognition of the super-sensible", this stage, finally, "comprises everything that can suffice for the needs of [pure] reason" (P 20:300).
As on Fülleborn's model, reason retains its eternal and necessary need of achieving cognition of the supersensible. However, unlike on Fülleborn's model, reason is now conceived as possessing distinct subordinate needs on the path to its final end, which are equally rational and necessary but not eternal. These subordinate needs are indexed to distinct historical periods: what reason needs to do in the first stage, that of Leibnizian ontology, is distinct to what it needs to do in the third stage, that of Kantian Criticism. These stages now account for the "temporal sequence" (Zeitordnung) of the stages of reason as expressed in the history of metaphysics, and are founded a priori "in the nature of man's cognitive capacity" (P 20:264). For this reason, Kant now defines metaphysics in developmental terms as "the science of progressing by reason from knowledge of the sensible to that of the super-sensible" (P 20:260; italics added).
The above model, furthermore, should not be viewed as a hastily abandoned departure from Kant's familiar framework. Rather, a comparison with Kant's more condensed discussion in the first Critique over a decade earlier reveals remarkable similarities. Past systems, Kant tells us, "seem to have been formed, like maggots… garbled at first but complete in time, although they all had their schema, as the original seed, in the mere self-development of reason" (A835/B863). Philosophy, furthermore, is said to be an idea that "one seeks to approach in various ways until the only footpath, much overgrown by sensibility, is discovered" (A838/B866). 12 Kant's Progress Essay develops these thoughts, by articulating the nature of the schema that corresponds to the original seed of metaphysics in the self-development of reason, and by illustrating the "various ways," in the form of stages of development, reason takes on the path to the satisfaction of its needs.
Finally, the Progress essay demonstrates that Kant's philosophy is not as foreign to the subsequent historical turn as previously thought. The precise sense in which this is so can be shown by determining Kant's relation to the historical element in Reinhold's philosophy, and thus to the historical turn in post-Kantian philosophy as it has recently been understood, in light of the three criteria discussed at the beginning of the article. These included (a) the employment of a historically oriented style of expression, (b) the advancement of a developmental conception of reason, and (c) the articulation of a sophisticated historiographical model for understanding philosophy's past.
In terms of (a), Reinhold indeed surpasses Kant in the quantity of writings that employ a style of expression focused on the historical (especially the social and cultural) contexts in which philosophical arguments are presented.
As Ameriks correctly points out, Kant's overriding stylistic tendency is to approach philosophical doctrines systematically (as he does, for example, in the Dialectic of the first Critique), while Reinhold tends more strongly to employ a more popular style that locates philosophical ideas in the changing circumstances of everyday life (as, for example, in the Letters). One possible reason for this is that Kant's primary aim was to present the Critical Philosophy while Reinhold's aim in the Letters was to popularize it. The latter aim inevitably lent itself to a more popular style that, in turn, was more amenable to the historical style of presentation (indeed, Reinhold's technical works are as ahistorical as Kant's). Furthermore, the one work that Kant did produce in the historical style-namely, the Progress essay-has often been overlooked, largely due to contingent factors relating to its complicated composition, editing, and publication history. Partly for these reasons, in the degree of influence on the popularization of a historical style of expression in philosophy, Reinhold's contribution outweighs Kant's.
As mentioned above, however, the historical turn cannot be reduced merely to its stylistic expression, and also covers the two remaining criteria. Here, a second obstacle for counting Kant's philosophy as historical resulted from the claim that Kant lacked (b) a developmental conception of reason-it is primarily in the sense that, as we have seen, Beiser and others accuse Kant of being an "ahistorical" thinker (Beiser, 2011, p. 147). As Ameriks puts this point, Kant's "imperialist" conception of philosophy is based on an atemporal "apodictic foundation" and, together with the resulting ahistorical style, constitutes the primary piece of evidence for the claim that Kant is "not committed to the historical turn" (Ameriks, 2017, pp. 144, 147). While Kant's writings at the time of the Critical turn may give this impression, the conception of reason presented in the Progress essay is remarkably similar to what Ameriks and Breazeale describe as Reinhold's "historicized" conception of reason. 13 Breazeale describes the Reinholdian conception in the following manner: "Reason itself develops and advances over time, and… it does so largely by means of conflict with itself and by coming to terms with and overcoming its own prior history" (Breazeale, 2010, p. 99). This dialectic of conflict and overcoming is the driver behind a stagebased understanding of reason, in which "at each stage of its development it faced specific needs that drove it to the next and higher stage of its development" (2010, p. 105). This sequence of stages, determined by the specific changing needs of reason in light of its eternal need, is not determined by contingent empirical factors but is itself rational and hence necessary: "it is not simply that the needs of reason actually have developed in the way just described, but rather that they necessarily had to develop in precisely this manner, in accordance with an internal logic of their own" (2010, p. 104). It is this teleological, stage-based framework that underpins Reinhold's developmental conception of reason that is, according to Breazeale, "a crucial feature of Reinhold's 'historical turn'" (2010, p. 104).
Recall that reason is conceived in a non-developmental manner if it is characterized solely in terms of general needs that are identical across all temporal moments-as we saw, Fülleborn conceived of reason precisely in this manner, with the unfortunate result that philosophy's history was reduced to nothing more than a catalog of errors.
Reason is conceived in a developmental manner, on the other hand, if its particular needs are taken to be distinct at distinct temporal moments, and if the temporal sequence of these needs is itself rational rather than contingent. This is how Ameriks and Breazeale conceive of Reinhold's developmental picture of reason: Reason still has essential and eternal needs but it is not characterized solely in terms of these. Rather, it is furthermore characterized in terms of particular, historically changing needs. For example, on Reinhold's account the most pressing need of reason of the pre-Christian age was to heal the rift between religion and morality and thus "to synthesize religion and pure morality in a manner accessible to all" (2010, p. 101). The most pressing need of reason in Reinhold's own age, on the other hand, was to heal the rift between faith and reason and thus to determine "the capacity of reason to provide any sure knowledge of God" (2010, p. 98). The sequence of stages established by the changing needs of reason is "necessary," according to Breazeale, for two reasons. First, this sequence is necessitated dialectically in the sense that "more recent developments absolutely presuppose and are thus made possible only by the earlier" ones (2010, p. 104). Thus, the need to adjudicate between faith and reason in Reinhold's age only arises once the earlier need to introduce morality into religion in the pre-Christian age has been satisfied and morality has developed to a satisfactory extent. 14 Second, the sequence is necessitated by the teleological nature of reason described above: in Ameriks' words, "reason's own 'expectations'"-which presumably come from reason's consciousness of its eternal needs-"disclose the shortcomings of past metaphysical attempts in a systematic way that indirectly points to" distinct needs and, hence, to new stages of reason's development (Ameriks, 2006, p. 177). Thus, while Reinhold argues Jesus' introduction of pure morality into religion satisfied the particular, historically indexed need of reason of the pre-Christian age, a gap nevertheless remained between this resolution and the perennial need of reason to "affirm the reality of God and a future life," as Breazeale (2010, p. 99) puts it. This gap led to a new particular need on the (long) path of reason's attempt to satisfy its perennial ones, a path that finally culminates with a (Kantian) doctrine of faith based on reason (rather than on feeling, as in the case of Jesus). In this manner, according to Breazeale, the "teleological necessity" of the sequence of reason's development is "rooted in the distinction between the unchanging or essential need of reason and its ever-changing needs along the path of its historical development," whereby each miss-step is equally an impulse for reason to re-orientate itself toward a new problem and move to the "next and higher stage of its development" (2010, p. 105).
In Kant's history of metaphysics, equally, reason's particular needs are distinct at distinct temporal moments and their sequence is rational rather than contingent. Thus, as mentioned above, the most pressing need of reason in the age of Leibnizian metaphysics is to purify purportedly supersensible concepts from sensual residue, while the most pressing need of the age of Humean skepticism is to show how such purified concepts can acquire objective reality.
Kant, like Reinhold, takes the sequence of these stages of reason to be necessary in sense that the satisfaction of more recent needs presupposes and is made possible by the satisfaction of prior ones: the need to provide objective reality to supersensible concepts in the Humean age only arises once these concepts have been satisfactorily purified of sensual residue in the previous age. Kant also highlights the significance of the teleological nature of reason for its historical development. As mentioned, this development is driven by the gap that opens up between the eternal needs of reason on the one hand and what it is possible for reason to achieve within the context of each stage of metaphysics on the other. Kant explains that his definition of metaphysics in terms of the eternal need of reason to achieve cognition of the supersensible "merely indicates what is wanted of metaphysics" but does not yet tells us "what there needs to be done in it" (P 20:261). What "needs to be done"-that is, the particular, changing need of reason-is determined by the historical situation based on the stage of development of metaphysics. Thus, while what needs to be done in the Leibnizian age is to purify the stock of metaphysical concepts, this does not yet satisfy reason's eternal need. This new gap between the satisfaction reason finds in achieving its particular aim of successfully purifying its metaphysical concepts and its lack of satisfaction with respect to its general need to attain cognition of the supersensible, in turn, drives reason the next stage of its development. The developmental conception of reason presented in Kant's Progress essay thus mirrors Reinhold's, not only in the sense that reason is viewed as responsive to the particular needs of the historical situation in light of the final ends of its necessary development, but that the sequence of the stages of this development is rooted in the teleological structure of reason itself.
In terms of (c), the articulation of a historiographical model based on the developmental nature of reason outlined above, the "epochal" text according to Ameriks is Reinhold's essay "On the Concept of the History of Philosophy" (Ameriks, 2006, p. 204). In this essay, Reinhold argues that a history of philosophy is only possible on the basis of the a priori "universal principles common to all human reason" (Reinhold, 1791, p. 34). With these principles in hand, the historian of philosopher is then able to separate the false from "the true" (die Wahre) in past doctrines and to uncover the "rational meaning" behind them (Reinhold, 1791, p. 32-4). Thus, the historian is able to "elevate himself above" the history of philosophy and reflect on the historical development of philosophical doctrines from the standpoint of the timeless laws of pure reason (Reinhold, 1791, p. 34).
The fact that Reinhold establishes his model for the history of philosophy on the basis of the timeless properties of reason against which past doctrines are to be measured makes him appear to take on board a position similar to Fülleborn's, together with its inherently ahistorical elements as identified above. Indeed, in this essay Reinhold defines philosophy explicitly in opposition to history-because that the latter is the sphere of "experience" and "sensory perception," Reinhold tells us that the former can thus be defined in terms of its "independence from anything empirical," as, namely, the sphere of "reason" or "pure thought" (Reinhold, 1791, p. 12-13).
Reflective of the more developmentally oriented story he told in the Letters, in "On the Concept of the History of Philosophy" Reinhold further draws a distinction between the history of philosophy and the history of the human spirit, writing that the latter is to provide an account of the "stages of the gradual development" of the cognitive faculties throughout the cultural, social, and political history of the human race (Reinhold, 1791, p. 22). However, though Reinhold's Letters paint a vivid picture of reason's changing needs throughout the history of spirit, focusing primarily on the moral and religious spheres of culture, Reinhold's historiographical essay does not manage to carry over this same developmental framework to the history of philosophy and instead seems to fall back on an ahistorical conception of reason.
In this spirit, Ameriks admits that Reinhold's essay on historiography of philosophy "can give itself the appearance of mechanically applying a pure ahistorical framework to a set of positions that might be best understood wholly apart from history" (Ameriks, 2006, p. 204). However, Ameriks points out that Reinhold's position is more nuanced in that he mentions that "philosophy's history needs to be divided in certain 'main epochs of change'" (2006, p. 204). Such a stage-based model would be commensurate to Reinhold's developmental conception of reason in its spiritual form as expressed in the Letters, and would furthermore mirror Kant's strategy for surpassing the Fülleborn model. Nevertheless, Ameriks concedes that Reinhold does not explain "why an epochal division of history is needed" nor how these epochs relate to general socio-historical conditions (2006, p. 204). In spite of this, Ameriks concludes that the essay outlines "very effectively" the manner in which the history of philosophy should be understood, and for this reason constitutes "an important, consistent moment in Reinhold's historical turn" (Ameriks, 2006, p. 204).
Without these further details, however, it is difficult to see how Reinhold would fill in the precise nature of his main stages of development within the history of philosophy or how he would determine the nature of philosophy's progression from one to the other, and thereby to move beyond Fülleborn's static model. It is no wonder that commentators view Reinhold and Fülleborn's respective models as essentially "the same", in that both understand the empirical historical development of philosophy as merely a succession of failed attempts to articulate the atemporal truths of the Critical philosophy (Micheli et al., 2015, p. 777).
The primary virtue of Kant's Progress essay, as I have argued above, is precisely that it is able to present a coherent and detailed picture of the development of reason within the history of philosophy. Overall, as I have shown above, Kant is able to translate the developmental conception of reason into a methodological reflection on the history of philosophy more skillfully than his younger successor, and the historiographical model he presents in the Progress essay exceeds Reinhold's in detail, sophistication, and historical acuity. The final marker of the historical turn (c), in Ameriks' words, was the ability to present a "genuinely modern and philosophical treatment of the nature of the history of philosophy itself." Pace Ameriks, however, the "epochal" text in this regard is Kant's rather than Reinhold's.
Due to the kinds of empirical contingencies Reinhold and Kant both tried to weed out from the history of philosophy, the Progress essay remained incomplete and unpublished during Kant's lifetime. Consequently, despite its philosophical superiority, the essay did not come to play the influential role Reinhold's works did in generating what has been recently called the historical turn. Rather than assessing the respective degrees of influence contributions of the thinkers at heart of this turn, however, my aim has been to bring attention to the particular form this turn took on the pages of the Progress essay, and to suggest that the continuity between conceptions of the relation between philosophy and its history found in Kant and in post-Kantian thinkers such as Reinhold is closer than has been generally accepted.