Fichte and the Contemporary Transcendental Arguments Debate

I. Thomas-Fogiel : The North American Fichte Society, may 2012

Abstract. The major debate in late twentieth-century analytical philosophy turns on the questions of the purpose, uniqueness and results of transcendental arguments. I pursue this debate in reference to Fichte the first thinker after Kant (and before Cohen, Cassirer, Apel, etc.), to claim the title of “transcendental philosopher”. In this respect, it is to be noted that while only two instances of transcendental argument may be taken seriously in the works of Kant, Fichte's Science of Knowing is in all significant respects the most accomplished and systematic development of what Strawson has called transcendental argumentation. More than that, Fichte's mode of reasoning is what unifies and gives continuity to the successive versions of his science of knowing, thus suggesting a form of transcendental argument that potentially has great significance for the contemporary debate. In this perspective, it is useful to reassess in the light of Fichte's philosophy  a series of major questions raised by analytical philosophy: How is transcendental argument related to skepticism? How original and specific is it? And, last but not least, what does it accomplish?

From Strawson to Hintikka, from D. Henrich [1] to Rorty [2], from Stroud to Apel, and more recently from Bitbol [3] to Zahar, a great many authors have taken part in what can now conveniently be called the “transcendental arguments debate.” Rather than providing a long series of references, let me offer a brief history of this recent debate. My focus will be on the three problems that are still being discussed today.

In The Bounds of Sense (1966) [4], Strawson showed that the Critique of Pure Reason employs two different strategic approaches. On the one hand, Kant frequently favors a description of how our faculties work together to produce experience. On the other, he occasionally relies on a type of argument that purports to show that if we do not accept this or that concept, we would be able to think or to act in the manner that our thinking and acting actually occur. Thus, in the “Analytic of Principles”, he justifies recourse to the concepts of cause and of substance by showing they are necessary conditions for our experience of succession. This is what Strawson regards as the core, or the molecular structure, of “transcendental argument”, namely, a demonstration that certain concepts or series of concepts are necessarily implied by the cognitive operations that we actually carry out. Strawson employs this same pattern of argument, which he detects in Kant’s “Analytic of Principles” and in the “refutation of idealism”, in his own philosophical investigations. He removes the schema from its original context and takes it to be a valid form of reasoning, which thus can be used outside the context of transcendental psychology. For instance, in Persons, Strawson showed that we are unable to differentiate the objects that we perceive if we are not capable of thinking of ourselves and of objects as entities coexisting in space. In The Varieties of Reference [5], G. Evans, who follows Strawson, still has in view a general type of constantly active cognitive operations (such as perceiving objects, referring to something, even predicating), and he tries to establish the presuppositions that are necessarily implied by them. In the same vein, K.O. Apel undertakes to reveal what sort of rationality must necessarily underlie our ethical evaluations. Transcendental argument always involves teasing out presuppositions that are necessarily implied by actual thinking or speaking. In this methodological approach, even skeptical adversaries are forced to acknowledge the legitimacy and validity of concepts that they otherwise explicitly deny. Starting from the definition of transcendental argument derived from its original context (that is, Strawson’s reading of the Critique of Pure Reason), we may lay out the forty years’ history of the transcendental arguments debate by means of three questions: first, the question of their purpose; second, that of their uniqueness; finally, the question of what they achieve or accomplish.

I. Three questions about transcendental arguments

The question about the purpose of transcendental argument is the first question raised against Strawson. In an influential  article (1968) Stroud sets out to contest the legitimacy of transcendental argument. With Kant’s text as his point of departure, Stroud shows that transcendental argument us meaningfull only in opposition to skepticism [6]. In typical analytic manner Stroud reduces the skeptical challenge to nothing more than doubting the existence of things outside  us. This is how he explains that for Kant, transcendental argument amounts to “responding definitively to those who are skeptical of the existence of things outside us.” [7] Having reduced transcendental argument to this single theme, Stroud then shows that, so defined, it misses its mark because, in the course of its development, it must have recourse to a verificationist premise. It is to this interpretation of the purpose of transcendental arguments that D. Heinrich objects  in noting : “the problem of Kantian transcendental deduction has nothing to do with the problem of metaphysical realism.” [8] The first question for debate, accordingly, is as follows: is the purpose of transcendental argument limited to the rejection of skepticism where the latter is reduced simply to the position of doubting the existence of things outside ourselves? And must we conclude, following Stroud and then Rorty, that such reasoning loses all meaning and value because it cannot be maintained without resorting to a petitio principii?

The second question, concerning uniqueness, stems from a position taken by Hintikka. In a short 1972 article [9]., he contests the specific Kantian character of this type of argument.. The fact that there are few texts by Kant on this explicit topic strongly suggests that the very conception of transcendental argument is a Strawson’s invention, lacking any relationship to Kant’s definition of the transcendental. Hintikka thinks that the only way in which Kant establishes that an a priori concept must apply necessarily to phenomena is by means of his constructivist theory of knowledge. In brief, Kant’s pattern of argument cannot be dissociated from his theory of the faculties. To deny the Kantian character of the argument becomes the same as denying its uniqueness. This is how it came about that, following Hintikka’s intervention in the debate, many thinkers have attempted to reduce the supposed “transcendental argument” to a simple deductive sequence of the type P implies Q; P; therefore Q. (To put the point in Kantian terms: since the experience of succession implies causality, and we have the experience of succession, we must posit causality.

The third question, formulated earlier by Apel and very recently by Bitbol in Les Arguments transcendantaux en physique moderne (“Transcendental Arguments in Modern Physics”), takes up their accomplishments, that is, their capacity to produce new statements. Indeed, , because transcendental arguments are often based on exposing a self-contradiction in the opposed thinker’s claims (in the form of “You cannot avoid conceding X without contradicting yourself”), they seem to be unable to produce positive statements. Can transcendental arguments be more than simple devices for refuting past systems or propositions? And if yes, then what exactly do they accomplish?

Now, we can ask to Fichte these three questions, which give the “transcendental arguments debate” its structure and tenor. In fact, it is interesting to retrace the debate starting with Fichte for the reason that, while only two instances of the transcendental argument can really be taken seriously in Kant’s writings, Fichte’s entire science of knowing (Wissenschaftslehre) [10] is arguably the fullest, most nuanced rendering of what Strawson called transcendental argument. Again, this structure of argument – which gives unity and continuity to different versions of Fichte’s science of knowing and therefore defines and qualifies the essence of his philosophy – allows us to thematize a sort of transcendental argument that clarifies the contemporary debate. This debate focused on a question of realism (the question of the relation between my representation and the facts of world). But there is another question more “metaphilosophical” : question of the relation between the content of my philosophical propositions and  status of philosophical discourse [11].

We shall show that this specifical question interests Fichte and highlights another aspect of the contemporary debate about transcendental arguments. To show it, I suggest that we raise for Fichte, three basic questions which dominate the debate in contemporary analytical philosophy, namely: what is the link that connects transcendental argument with skepticism? What is its purpose? What is unique about it? And finally, what does it accomplish?

II) transcendental argument and the challenge of skepticism in the “wissenschatslehre”

If we consider Fichte’s Wissenschaftslehre, we must acknowledge immediately the close connection between recourse to the transcendental argument and the fight against skepticism. Indeed, Fichte says repeatedly, his Wissenschafstlehre is born from a confrontation with the skepticism of Aenesidemus. [12] Let us recall in this connection the well-known letter of December 1793 in which Fichte writes to Stephani: “Have you read Aenesidemus? He has troubled me for a while, made me reject Reinhold, made me suspicious of Kant (…) Rebuilding is what’s needed, I’ve thrown myself into it for the past six weeks.” He comes back to this theme in a letter of January 1794:  “Aenesidemus brought me to the firm conviction that in its present state, philosophy is not a science.” [13] It follows that the challenge of skepticism forces Fichte to modify his relationship with Kantianism. This is the case for his relation to Maimon, another skeptic, as well [14]. How are we to understand this impact of skepticism on Fichte ? Some observes, such as Cassirer or in France  J. Rivelaygue believe that Aenesidemus’skepticism is not particularly threatening. According to these authors Aenesidemus’s arguments fall short of their critical objective and betray a complete misunderstanding of the Kantian position. Yet how are we to understand this gulf between current commentators who note the weakness of Aenesidemus’ arguments, and the explosive effect that this skeptical position provoked on Fichte? [15] Why does Fichte hold this skeptical reaction to the the Critique in such high esteem, to the point of making it the starting-point of his own philosophical enterprise?

To respond to this question, we have to understand what Fichte took from Aenesidemus. This text can be taken in two ways. The first is to isolate the theses opposed to Kant and to expose their often dogmatic character on the one hand and their blatant incoherence on the other, ultimately demonstrating how they misconstrue the transcendental. [16] By contrast, the second approach might show how the skepticism of Aenesidemus emerges through his pattern of reasoning, his conceptual approach, and not from the particular theses that the author develops, theses that are, by design and in virtue of their tone, totally contradictory. The central chapter of Schulze’s Aenesidemus, “Was Hume’s skepticism refuted by the Critique?”, clearly illustrates the type of argument he employs against Kant. According to Aenesidemus, Kant’s theses can only be proven by presupposing propositions that Hume dismissed [17]  and can  only be affirmed by accepting an internal contradiction. It follows Kant can only prove his theses by invoking premises he rejects.

For Fichte, Aenesidemus shows the fundamentally self-contradictory character of Kantian discourse, which in the final analysis cannot elaborate its most important theses without transgressing against principles on which they should be based. [18] Exposing the self-refuting character of the critical stance is the over-arching aim that gives the Aenesidemus’ text its structure and its ultimate inner consistency. Kant’s philosophy is self-refuting.. In other words, if the Critique is true, then it is false. [19] This contradiction, stated in many ways in the Aenesidemus, is summarized masterfully by the author in another of his works, the Critique of Theoretical Philosophy: “Thus if what the Critique claims to know about the sources of experience is itself an instance of real knowledge, the claim in this same Critique that all true knowledge available to us is limited exclusively to objects of experience, is completely false. If on the other hand, this affirmation were true, then all knowledge of the sources of experience must be  as an empty illusion.” [20]

With respect to this position, it would be utterly illusory to object that Kant proceeds through regression towards conditions of possibility and that, thanks to this regression, manages to uncover the very essence of representation. There are actually two reasons to abandon this artificial claim:

1) On the one hand, similar to Reinhold, Aenesidemus notes that a critique that requires reasoning which is entirely dependent on something that is itself highly open to question, namely the truth of mathematical and physical judgments, assumes the truth of that which must be proven.

2) On the other hand, a similar argument concerning the way that the Critique proceeds simply reinstates the dilemma. Indeed, in his employment of regression towards conditions of possibility, Kant would be admitting implicitly that there is a mode of validity other than the connection of an intuition and a concept. Therefore, his limiting of truth exclusively to experience would be utterly improper. If following Kant’s line of philosophical reasoning requires us to accept a mode of truth apart from the exclusive application of experience, then the principle that underpins the entire Kantian opposition to metaphysics simply crumbles. So it is truly the entire Kantian edifice that is implicated here, since in the final analysis what is contested is the very possibility of stating the theses of the Critique. Since it cannot sustain its claims to validity, Kantian discourse is forced back to its own question: Quid juris?

What should we take from this analysis of skepticism in relation to the beginnings of the transformation of transcendental philosophy into the Science of Knowing (Wissenschafstlehre)?

In the first place, there is this singularly important account of the profound transformation of skepticism after Kant. Skepticism before the Critique deals with the link between knowing and being, or subject and object. What Hume objects to is going from a causal relationship in our minds to its objective reality. [21] Accordingly, the skepticism to which the Critique responds is about the possibility of the subject-object relationship. Yet, after Kant, this question is put into perspective in the refutations or skeptical readings of the Critique. Post-Kantian skepticism proves to be as novel as it is destructive, precisely in that it considers the “metaphilosophical” question to be more important. To answer the question concerning realism, the philosopher must, first, ask about the relation between the content of philosophical propositions (for example the proposition about the definition of truth) and the possibility of stating this definition. The question “what becomes of the objectivity of our representations?” is no longer the immediate issue because the guiding question of Aenesidemus concerns the relation of Kant to his own claims. What is the status of philosophical discourse? The philosopher’s aim is to draw out how consciousness functions in its capacity with respect to knowledge of objects, but he tells us nothing about the processes by which he arrived at knowledge of this functioning. Kant invalidates his own discourse by the very fact of his discursive activity. Thus, beyond the minimizing efforts of later Kantian commentaries, we are at the point of understanding the shock-wave that the skeptical attack provoked in the early post-Kantians. In effect, the key contribution of Aenesidemus is to have shown that the Kantian discourse that proposes to define science, which is to say, to determine what is valid and what is not, cannot establish its own claim to validity. Unable to achieve a “self-reflection” that would consist in enabling its definition of validity to apply to itself, this discourse appears condemned to drowning in self-contradiction and self-refutation.

In the second place, in light of this reconstruction, we can refine our understanding of the meaning of the task that Fichte assumes when he undertakes in 1793 to “rebuild everything.” It will mean his moving beyond the self-refutation of Kantian criticism in order to establish philosophy as a science, that is, a discipline capable of justifying its claim to validity: “Aenesidemus brought me to the firm conviction that in its present state, philosophy was not a science. But he reinforced my conviction that it will be able to become one within a short period of time.” [22] It is this term : “science”, that gives coherence to the philosophy of Fichte from 1790 to 1814. [23] The question of scientific character is always highlighted by Fichte as the line demarcating the difference between his philosophy and that of Kant. However, most importantly, because of the nature of the skepticism of Aenesidemus, this question of scientific character acquires a very particular aspect, namely the question of philosophical reasoning that is not self-refuting. To attain the status of a science means to develop a particular type of reasoning.

This is why this analysis allows us once again to shed light on the dispute about transcendental arguments and to renew the terms of the discussion. Having dwelt on the origins of the science of knowing, we see that the link between the struggle against skepticism and the development of a transcendental form of reasoning is historically proven. In this sense, Stroud is on perfectly safe grounds in suggesting that the uppermost target of the transcendental argument is refutation of skeptical positions. This is moreover what Fichte himself notes in saying that “the critical skeptical such as that of Hume, Aenesidemus and Maimon (…) points out the inadequacy of the grounds so far accepted and shows in doing so where better are to be found (…) And if knowledge gains nothing as to content from this, it certainly does as to form and the interest of knowledge are but poorly recognized in denying to the sharp-sighted skeptic which respect which in his due.” [24] Nevertheless it would not do to misunderstand the nature of this skepticism. Since Kant, being skeptical does not consist at all in doubting the equivalence of our representations or mental schemas to things. So to challenge skepticism here in no way implies the wish to establish some sort of realism. Contrary to what Rorty advocates following Stroud, the purpose of transcendental arguments could not be to prove only the legitimacy of realism against skepticism. [25] Rather, the question that is posed, with Fichte, is about the status of the discourse of the philosopher whose avowed quest is for a type of reasoning that is capable of avoiding the shoals of Kantian self-refutation. To investigate the norms governing a type of discourse that hopes to be valid is not at all reducible to inquiry into the links between discourse and things, to which Stroud, like Rorty, would reduce the purpose of transcendental arguments in order to debunk them with greater ease. Thus, studying the philosophy of Fichte allows us on the one hand to understand that skepticism is not simply doubts concerning things outside ourselves – and were it so, then the definition of skepticism in Stroud or in Rorty would be actually a narrower definition– and on the other, allows us to circumscribe precisely what is so at stake in transcendental argument: the consistency and coherence of propositions that proclaim the validity of what they state.

Having made this first point, it is now appropriate to see how Fichte responds to our second question, namely the specificity of transcendental argument. What does “transcendental” mean in the science of knowing?

III) The Uniqueness of Transcendental argument

In order to shed light on this second problem, while continuing with Fichte’s orientation, we need to answer three questions. First of all, there is the question of the “laicization of transcendental arguments”: [26] may we, as Strawson advocates, detach this type of argument from Kantianism, abstracting it from his doctrine of faculties and from his constructivist theory of knowledge (intuition + concept)? Or in the manner of Hintikka, must we rather treat Kant’s critical philosophy as an organic whole, a system from which one may not detach particular components in order to use them in another context? To put it in starker terms, can we separate the description of the work of the faculties as a function that synthesizes diversity (Kant’s constructivist psychology) from the transcendental philosophy? Further, we will need to understand what Fichte means by “transcendental” in order to finally elucidate the specific character and profound originality of this Fichtean reasoning, neither dependent strictu sensu on the Kantian position nor yet reducible to a type of deductive reasoning, the modus ponens, that has already been discredited.

 (1) To our first question then: Even though the very term “transcendental philosophy” occurs in Fichte with surprising frequency, he does not fundamentally adopt any of its divisions, namely transcendental aesthetics, analysis and dialectics. Indeed, he discounts the very idea of a transcendental aesthetics in that, in the science of knowing, space and time are produced (like the categories) by activity and are later thought as arising from acts of mind. Furthermore, he denies the transcendental dialectic and the notion that grounds it, that of the illusion of reason – in fact, he declares in the Second Introduction to the Science of Knowledge that there is no “illusion of reason”, which is why “Kant could have been convinced when writing the Critique of pure reason”. [27] Finally, he rejects transcendental analysis with exceptional vehemence. In light of such a position, it would not be exaggerated to say that Fichte appears to be the least Kantian of all the philosophers who laid claim to transcendental philosophy. Indeed, if we consider this Fichtean position in which the notion of transcendental aesthetics is rejected, in which the very content of analysis and dialectic is attacked, in which ultimately the division of transcendental logic into analytical and dialectical is considered to be purely and simply useless – then we feel compelled to recognize that readings such as that of Heidegger who highlights the transcendental aesthetics to the detriment of analysis or that of H. Cohen who, to the contrary, devalues aesthetics in favor of analysis, are closer to the Kantian text because both of them reconstruct that text on the basis of a division that actually exists in the Critique; but this is not at all the case with Fichte. Accordingly, we would be justified to ask: if we remove from pure reason the transcendental aesthetics, transcendental analysis and transcendental dialectics, what remains? Nothing, apparently, replies Fichte, unless it is the transcendental philosophy itself. Furthermore, Fichte demonstrates here the absence of a necessary connection between the transcendental argument and the Kantian theory of faculties. In the argument between Strawson and Hintikka he would be for Strawson on the possibility of examining transcendental argument independently of Kantian psychology.

(2) This being the case, what does Fichte mean by “transcendental”? If we enumerate and synthesize all occurrences of the different versions of the science of knowing, it appears that the transcendental in that context is defined on the basis of two characteristics: first of all, it is always opposed to “ontological”; further, it defines a certain manner of demonstration, of “making operational”, to employ an expression of Fichte’s and not a doctrinal point.

Regarding the first characteristic, one could easily cite numerous texts. I will take just one of these, from his latest writings. In his Science of Knowing of 1813, Fichte writes: “The science of knowing (Wissenschaftslehre, henceforth Wl) is not a doctrine of being. Being of any sort, to the extent that it is subject to observation, can only be apprehended by us as known; therefore being is always situated in knowing … This is why it is only through error that we can speak about a doctrine of being, when we are inattentive to our knowing and our thinking” [28] . Thus, to the extent that the Wl comprehends that its exclusive object can only be knowing, in other words that it is a doctrine of knowing and it abandons being; and to the extent that it clearly recognizes that there cannot be a doctrine of being; to that extent, it is also a transcendental idealism, which is to say, the absolute exclusion of being through the attention directed at it. Whoever affirms that there is no doctrine of being and that the only possible absolute doctrine and science is the science of knowing; such a person is a transcendental idealist because he recognizes that knowing is the very highest object of knowing.” [29]

This “apposition” between the two expressions “transcendental philosophy” and “exclusion of being” shows how much the Wl aspires to avoid ontology but rather is an epistemology in the literal sense of the term (that is, a science of science or a science of knowing)[30]. Still more precisely, the doctrine is not a doctrine of the finite subject or of freedom, even less a doctrine of the infinite or of the absolute – in a word, not a doctrine about something or other but a doctrine about itself; it concerns, in Fichte’s words in a text of 1813,  “attention directed at itself”, which means it directs attention to knowing. In a certain sense, this definition recalls literally the Kantian text in which transcendental philosophy “is never a relationship of our knowledge to things but rather to the faculty of knowledge alone”;[31] but it connects equally with the definition of transcendental argument in use since Strawson. In effect, unlike ontological arguments that aim at establishing claims about objects themselves and their properties (thus arguments of the form S is P), the transcendental argument must establish propositions concerning concepts implied necessarily in our thinking. As Bitbol, whose interest is the use of transcendental arguments in quantum physics, emphasizes: “strictly speaking, transcendental arguments tells us nothing on the ontological level.”

The second characteristic in Fichte’s definition of the transcendental distances him even more decisively from the Kantian approach. In fact, comparing his transcendental idealism with that of Kant on numerous occasions, Fichte anchors the difference in their different manners of reasoning. In this vein he writes in the First Introduction to the Science of Knowledge (§7): “transcendental idealism can be deployed in two ways.” The first “deployment”, that of Kant, leads to ““unproved und unprovable” propositions because it shows itself to be unable to answer the crucial philosophical question: “from where and how do we know it?” Evading this question concerning the status of its discourse, Kant’s “imperfect” transcendental idealism is exposed as unable to justify its claim to validity. So Fichte exclaims that “if he knows it, the mystery is how does he know it”. In this passage and many others that echo it, it is evident that the term “transcendental” in Fichte denotes a type of proof intended to answer the question that leaps out so often from this text: “but from where do we know it?” Epitomizing the shift that has entered the notion of the transcendental, the question “from where do we know it?” puts all philosophers on notice that they must account for the conditions of their discourse. The philosopher, in this case Kant, must show “from where he speaks” and “from where he knows.” An anachronistic reading could interpret this question as a demand that the surrounding conditions of the work be investigated. For instance psychoanalysts, hearing some statement, tend to cast it in terms other than its own (the unconscious, personal traumatic experiences, an unusual life and other impedimenta); again, a sociologist could say that Kant is speaking in the context of the triumph of the bourgeoisie at the time of the French Revolution; and finally, an analytic philosopher such as Austin could understand the “speech context” as the account of the contingent, particular, exterior circumstances of the act. [32] However, opposed to all these cases, the Fichtean question “from where” is not conceived as an appeal to a context external to the statement; it seeks to achieve a return upon the statement itself. [33] Thus, what is required here is a sign not of an ad extra reference, but of a reference by the proposition to itself. Kant gave a definition of validity that, unable to apply to itself, was rendered invalid the moment it was issued. The question “from where?” leads accordingly to a return to the statement itself and not to an inquiry about the external circumstances where it arises. And so the definition of the “transcendental” will call for the possibility of a self-inclusion or of a self-application of the proposition to itself. If the philosopher cannot achieve this return to his statement in order to establish its conditions of validity, then his propositions, like those of Kant, will remain “unproved und unprovable” [34].”

To think about our knowledge of knowing immediately requires that we address this question head-on: “How do we know that we know?”; or putting it a bit differently: “How can a philosopher say what he says?” The latter formulation is favored by the Wl of 1804: “We must ascend from what [the philosopher] says to what he does not say but which he had to presuppose in order to be able to say what he says.”  The process of reasoning through which “we must rise from what must be presupposed in order to say what is said” consists of restoring the conditions on the basis of which a proposition – or a series of propositions – acquires coherent and consistent meaning. In fact, every proposition refers back to a set of implicit statements that are the conditions of its meaning. What Fichte intends to reveal in a methodical manner in every philosophical system, his own as well as those of his opponents, are the tacit propositions presupposed “in order to say what is said.” Reasoning therefore comes down to bringing out the elements that are intrinsically and implicitly connected to a proposition that claims to be valid. Ultimately this approach to reasoning is defined as uncovering ( de-implying) what is contained within “what is said.”

(3) This shows how far Fichtean signification is from the term “transcendental” in relation to Kant’s philosophy and its proximity to what today we call the structure of transcendental argument. All that remains for us to do, to respond to the second question concerning the uniqueness of this type of reasoning, is to outline how the argument now in play, while not dependent on the Kantian theory of faculties, is also different from classical reasoning in the modus ponens manner.

The many who contest the uniqueness of transcendental argument all share a common viewpoint or a single position: “A valid argument can only be deductive, that is, its conclusion must contain a logical content that is just as elevated as that of its premises.” [35] In this vein E. Zahar writes that “Logicians of genius such as Frege, Russell, Hilbert, Gödel and Tarski have taught us that any inference that adds something (ampliative, amplifying) is necessarily invalid” [36] – echoing A. Boyer who says on the same theme of reasoning that “reasoning is either deductive or inductive, tertium non datur.” [37] For Boyer, since the transcendental argument  can’t boil down  to classical and formal deduction or induction, it  isn’t valid.

This resounding verdict, rendered with the backing of such illustrious names as Frege, Russell, Tarski, Gödel and Hilbert, may seem difficult to counter, yet this is what Fichte does. In effect, the structure of his reasoning does not depend at all on the relationship of premises to conclusions (in respect to the formal and classical definitions) nor is it the result of induction or generalization. The heart of the Fichtean argument is self-reference[38]. Fichte spoke of this self-reference in many ways throughout his oeuvre. Without any doubt, it is this definition that accounts for unity and coherence among the different versions of the Wl that Fichte first introduced and then constantly reviewed, revised and elaborated over a twenty-year period. He expresses it as “the accord between what one says and what one does” in the Wl of 1804 as well as in the Second Introduction to the Science of Knowledge. Further, he approaches it in the WL of 1794 as an accord between “what is done” and “the doing” [39]. In this context, what is to be expressed is the identity produced between the relationship of an x and a y (or as Fichte says, “we reflect only upon the possibility of members involved in the relation”) and “production of the relationship itself.” This “production of the relationship” is “the reflection of the philosopher on the modality of observation”, that is, the taking into consideration of the philosopher’s “doing” when relating two terms to each other. Yet again, he portrays the principle of reflexivity by this syntagma, found at the end of the theoretical part of the WL 1794: “identity between what is explained and the ground of explanation.” [40] He also expresses it very often as the identity between the form and the matter of a proposition, for example in the Wl of 1813. Finally he formulates it in less technical texts such as the Bestimmung des Gelehrten as ‘non-contradiction with the self’. [41] Briefly, if the formulations are many, the idea remains the same: self-reference is defined as congruence between statement and stating, the content of a proposition and the fact of expressing it, the “saying” and the “doing” (the “Sagen” and the “Tun”). The expression “what is said” points to the content of a philosophical system, namely the totality of propositions with which a philosopher attempts to state what knowledge is, or the human, God, nature, existence, and so on. The expression “what is done” covers the totality of the procedures that philosophers use as philosophers in order to be able to express that content. [42]

This line of reasoning does not consist of drawing the consequences of a given premise (deductive reasoning), but aims to ensure that there is congruence between the content of what is said and the actual activity of saying it. What is involved is two things at the same time: becoming aware of the actual content of a proposition; and inquiring into the discourse procedures underlying the statement – or again, as the beginning of the 1804 Wl said, it involves rising towards “what is presupposed in order to be able to say what one says.”

This identity (as congruence between the doing and the saying) is an identity that Fichte discovers; if he is not the first to touch on this theme [43], at least he is the first to position it as the foundational principle of an entire system. This sort of non-contradiction or identity is novel – it has nothing to do with contradiction in the area of formal logic; nor with physical contradiction between two antagonistic forces (which Kant, following Newton, called opposition); nor even less, with contradictions between a proposition and the fact that it ought to be conveying (these being the contradictions that for Kant arose from dialectical propositions). The contradiction of which Fichte speaks is a contradiction between the act of saying X and what is said about X, a contradiction that nowadays we should call a performative or practical contradiction in the sense that, for example, regarding the proposition “I am not talking,” the fact of the stating (“Tun” or “effectuation” or “status of speech”) contradicts the content of the statement (the “Sagen” or “what is effected” or the “matter of that which is spoken”).

At the end of this analysis, it is evident that the structure of Fichtean reasoning is truly specific and original. In no way can it be reduced to the tropes of logical or scholastic reasoning such as modus ponens. But for all that, it also cannot be relegated (as some adherents of a strict logicism would prefer) to the status of invalid propositions because they are neither deductive nor inductive. Tertium datur, a third way is given by the self-reference that Fichte placed at the heart of reasoning, the same self-reference that Russell spurned “to the point of phobia” [44]. Indeed, Fichte employs a type of reasoning that was never truly thematized as such before him. So in effect, were we required to sketch the history of the various ways of proving a philosophical statement, nothing else would resemble what Fichte calls the “deployment” of the transcendental. Proof does not consist here in deducing the consequences that flow from a starting-point that is certain (this would be the Cartesian schema of the “Regulae” or the Discourse on Method). Nor is proof a regression from fact to the anhypothetical (Plato) or to conditions of possibility (Kant). It is not a matter, as Aristotle occasionally advocates, of a method of reasoning that proves a position by demonstrating that its opposite is impossible. And proof is also not demonstration of the logical non-contradiction of any statement (as Leibniz would have it). The challenge for proof is now to harmonize what is said with the fact of saying it; knowing with the known; reflection with reflecting; positing with what is posited. Attaining what Fichte calls “understanding of understanding” or “knowing of knowing” is thus to sketch a particular style of reasoning, a certain structure of reasoning that may be considered to be the highest realization of transcendental argument. In order to complete the demonstration of this thesis, we must now turn to the third question: what does transcendental argument accomplish.

IV) What Transcendental argument Accomplishes

Though simple, this third question is still challenging. Most often, in effect, transcendental argument consists of showing that a thesis X cannot be put forward without its author immediately falling prey to contradiction of the pragmatic sort. [45] Accordingly, as D. Heinrich remarks, this argument always has an ad hominem character because making it presupposes an actual adversary. Thought of in this way, the argument’s force would only be refutational. It would not be able to produce new truths. In fact, this thesis can draw some initial credibility from an examination of Fichte’s own writings. A large part of the Science of Knowing is actually devoted to demonstrating the self-refuting character of opposed theses. In this connection we may refer to the Wl of 1804 which carries on at great length in this purely refutational mode. There Fichte explains, in effect, that we only arrive at a “definit result” [46], after having dismantled the errors of earlier systems. All of these philosophical systems share a common characteristic of being “derived knowledge and not science of knowledge.” [47] Fichte states: “I say : in all derivative knowing , or in appearence, a pure and absolute contradiction exists between enactment (Tun) and saying (Sagen): propositio facto contraria. Let me add here by the way as I thought previously on an appropriate occasion, a thoroughgoing skepticism must base itself on just this and give voice  to this ineradicable contradiction in mere consciousness. The very simple refutation of all systems that do not elevate themselves to pure reason, I.e their dismissal and the presentation of their insufficiency even though their originator is not thereby improved, is based on just the fact that on points out the contradiction between what they assert in their principles and what they actually do, in asserting them : as has been done with every system that we have tested so far, and yesterday with Jacobi’s as well.” [48] This passage contains all the traits of the transcendental that our analysis has highlighted, namely the link with skepticism, the definition of self-reference as “agreement between the saying (Sagen) and the doing (Tun)”, and finally the refutational power of employing the transcendental because it always is a matter of saying to an adversary “You cannot say X without contracting yourself.” Must we conclude from this statement that the argument’s entire force is expended in this purely destructive outlook? Doubtlessly not, since immediately after this passage Fichte say that if a contradiction between saying and doing inhabits all of derived knowledge”, by contrast absolute reason will distinguish itself from this relative knowledge by the fact that what it does is expressed in it and it does what is expressed. With this proposition Fichte would move from the proposition “you cannot say X” to this other one which, in spite of the double negative, is pregnant with positive propositions, namely “you cannot NOT say that Y.”

To understand this movement from negative to positive, we must first explore more deeply the special style of Fichtean reasoning. Let us first recall that the point of departure, this matching or identity as congruence between what is said and what is done, is not a fact (Tatsache) but an action. This proposition, always rendered in a reflexive form (Tathandlung in 1794, knowing of knowing in 1804, understanding of understanding in 1813), is a task to be carried out, not a psychological or factual given. The most obvious trait that characterizes the science of knowing is that it is not defined in terms of a description of some X; rather, it constructs itself. It is a construction because to reflect is not a matter of reflecting or giving an image of some prior thing but of creating concepts or constructs (Bildung) based on a task that one has assigned to oneself. On this point we may cite the most important formulations of the Wl of 1813: “Our first task is to construct the unity mentioned earlier;” [49] “this unity that we must produce (Erzeugen)”, which Fichte associates with “reflexivity as understanding of understanding,” is “an achievement that is required;” [50] and again, “the unity of understanding” is “a constructed unity.” [51]  Reasoning consists really in constructing or producing concepts with regard to a goal or a task or an aim. In a word, reasoning is a process that is structured according to a model to be realized: the identity between the saying and the doing. Another way to put it is this: a goal is targeted, namely identity, and contradiction is given as the factual point of departure – the contradiction of the not-I in the Wl of 1794, contradiction of the three Kantian absolutes in the Wl of 1804, etc. The path or process will consist of overcoming contradiction by means of intermediate concepts or schemas until we arrive at a proposition X that contains no contradiction whatsoever. Therefore, at each step of the process, it is a matter of producing a proposition that is untainted by any contradiction. This is the way that a bridge occurs from a negative statement of the sort “you cannot say X” to a positive statement of the sort “you cannot NOT say that Y.”

Two examples illustrate this Fichtean style of reasoning, namely “check” (Anstoss) [52], and the relation of finite to infinite.

Let us recall that the basis for responding to the strange thesis of the check – an abrupt sense of being blocked in one path and propelled into another – in the WL of 1794 is the need to make “what is explained” coincide with “the ground of explanation.” It is clear that the principle of congruence between saying and doing influences the analyses that guide the author in working out what we need to understand by this thesis of a “check.” This principle allows us to determine both what the check cannot be and what it is – or more exactly, what it cannot not be. For instance, the requirement for performative non-contradiction dictates that the check could not be considered as an external impulse or stimulus, an accidental occasion of activating the faculties or “the task of determining oneself.” By a series of negative demonstrations, Fichte thus shows that the check cannot arise from classical realism, nor from the theory of conditional causality (cf. Malebranche), nor from that of things in themselves. The reason for these rejections is simple: in such a doctrine, the explanation is not contained in the principle of explanation; what this type of realism explains enters into contradiction with the fact of explaining it. The error therefore of realists is always the same, consisting in the inability to report on what they say without contradicting themselves in the act of speaking. At the end of this negative proof, Fichte shows how this check can be conceived in one way alone, as a feeling. The feeling is what affects the subject. It is not something external, a thing in itself causing feelings but simultaneously “subject-object”. This is the only coherent manner, in the sense of avoiding self-contradiction in the act of saying it, to give substance to the notion of feeling or of “check.” Through this example we see that the proof contains not only a negative or refutational side but also a positive thesis that later develops as a doctrine of sensation or feeling. Accordingly, we have truly passed here from a formulation of the type “you cannot say that X” to another of this sort: “if you wish to respect the congruence between saying and doing, then you cannot not say that X”.

The second example that can be selected here is that of the relationship of the finite and the infinite. As we know, Fichte rejected the thesis of radical finitude as deployed by Kant or Maimon because of a single consideration: radical finitude cannot be uttered from the standpoint of its stating. In a long section of the WL 1794 (Gamma 1 and 2, devoted to the error of Kant and Maimon), Fichte undertakes the analysis of what he himself terms the systems of the finite I. He writes  that “the  basic principle of such a system would run : the self is finite simply because it is finite” [53]. Regarding this system, he notes: “ But now whether or not such idealism makes a higher ascent, it does not ascent so high as one should ; to the absolutely posited and  unconditioned. To be sure, a finitude is to be absolutely posited; but everything finite is limited, in virtue of its concept, by what is counterposited thereto : ans absolute finitude is a self contradiction” [54] All finitude can only be stated on the basis of the hidden consideration of what it is not, on the basis therefore of its contrary, the infinite. We cannot posit the finite except in relation to the infinite, the latter being conceived (in the manner of Maimon) as a necessary fiction. It follows that radical finitude is an empty concept, its utterance a performative contradiction. Because quantitative idealism cannot reconcile what it says with what it does, it locks itself within what Fichte calls here, very significantly, a “manifest contradiction.” Neither Maimon nor Kant succeeds in reflecting on the conditions of stating the finite. The “manifest contradiction” that Fichte condemns is a contradiction the essence of which consists in not being able to account for the status of what it says when finitude is affirmed as radical. The finite can only be stated on condition of supposing the non-finite: this is the condition for a meaningful statement of finitude. To repeat the formula of 1804, this is what is presupposed by being able to say what one says. We do well to be attentive to the nature of Fichte’s reasoning on this point. It is not a matter of emulating the path of the third Cartesian Meditation, which concludes from an experienced finitude that an actual infinite exists external to oneself. It is not a question of knowing if the infinite or even the finite exists. Rather, the question is this: what conditions allow us, in a non-contradictory statement, to determine what finitude is. The only way to do this is via the subversive positing of its contrary. This is the point that Fichte seeks to highlight in rejecting the finitude of Kant or Maimon. Kant defines finitude as radical without ever being able to justify the status of this proposition, which thereby becomes a matter of unchallenged dogma, an embargo against thinking that could unsettle our most settled certainties. For his part, Maimon, taking lessons from the Critique of Judgment, admits the necessity of positing an infinite in order to allow humans to reflect on themselves as finite; but he defines this infinite as “fiction.” Now, what necessary line of reasoning requires that the infinite be thus defined as “fiction”? This characterization in terms of “fiction” is just as arbitrary as that which consists in claiming that the infinite is an actual reality. Even more, if the infinite is posited as “fiction”, why would the finite not be so defined as well, unless some undisclosed evidence stops us from asking? Based on this refutation, Fichte posits the necessity of thinking the relation between the finite and the infinite in order to escape from the manifest contradiction.  He will therefore show that establishing the congruence between the saying and the doing supposes what one could call an ideation of the infinite. Obviously this ideation would not consist of demonstrating the infinite in concreto nor positing it as a subsisting reality external to ourselves. This ideation will consist in casting our statements in infinite mode, or un-restricting the limit[55]. This ideation of the infinite is a requirement that produces concrete and positive statements. For example, this strategy explains the moral judgment “be magnanimous” which is so important in Fichtean ethics. This magnanimity arises from what Descartes called the generosity through which the infinite reveals itself within the finite. Magnanimity or generosity is a moral norm resulting from the work of reason understood as a need to make infinite. It corresponds on the ethical side to the process of ideation of the “infinitely great” aesthetic of the Critique of Judgment. Magnanimity exceeds all “fixation” and thus fulfils the need for infinity.

From all this it is clear that Fichte’s reasoning is far from simply being a machine for refuting arguments; it produces positive statements that are founded in reason.

Accordingly, we have replied to the three questions that arise in the transcendental arguments debate. Contrary to the dominant position, the purpose of transcendental argument cannot be reduced to a rejection of doubts concerning the existence of things outside of us. According to Fichte, the question of the relation between subject and object (propositions and the world) isn’t the sole or even the principal question ; or more precisely, there is a another question, a more question “metaphilosophical”, which the philosopher must study and which is very important for the philosophical system. This question bears on the relation between the philosophical discourse and the act of enunciation (the relation of language to oneself and not the relation with respect to any fact of world). While transcendental argument may well be deployed in connection with skepticism, this is because skepticism challenges the claim to validity found in our philosophical propositions. Rorty too is obliged to address this claim when he explicitly denies all truth value to our philosophical statements. A philosopher, realist or not realist, must be able to say “How he can say what he says”. Subsequently we were able to arrive at a better understanding of transcendental argument by showing that it cannot be reduced either to the Kantian doctrine of faculties nor to classical deductive or inductive reasoning, as A. Boyer, in Russell’s wake, defines it. With transcendental argument, the core of which is self-reference, Fichte deploys what could be called a grammar of effectuation, which is no less rigorous than other logical programs. Thus, just as the Analytics of Aristotle rests finally on the principle of non-contradiction, from which the different laws of reasoning can then develop; and just as the Wissenschaftslehre of Bolzano lays out the rules and operations of logic based on the method of variation, which latter supports definition of analytical judgment on other grounds than the Kantian inclusion of the predicate in the subject; so too the Wissenschaftslehre of Fichte builds a “logic” or method of reasoning with foundations that will later allow the network of isolated truths to be deployed. These truths, arising from the employment of reasoning, show the degree to which this reasoning is capable of producing concrete and positive statements.

What is more, however, is that this analysis of transcendental argument in Fichte also allows us to respond to two recurring objections against transcendental argument. [56] The first consists of challenging the validity of the “premise” in a transcendental argument. For what Fichte teaches us is that the “premise” or the “starting point” is not a material or mental fact (the validity of which would effectively require justification) but an act. And what’s done is done; the actuality of an act that has occurred cannot be denied.  By positing this act as act, Fichte anticipates and responds to the question of the acceptance of premises.

This analysis also allows us to reply to the naturalist objection against any recourse to transcendental argument. This objection, found in particular in Quine, consists of saying that if we cannot NOT posit X or Y, this is due to our nature. The necessity of the argument would not arise from anything other than our natural makeup. As Quine says, it is enough “to accommodate the transcendental concepts of a naturalist viewpoint.” [57]  Now with Fichte, the identity which constitutes the starting-point is an ideal, an obligation that we freely adopt, a goal at which we aim and not a natural thing that we experience. Identity as the congruence of saying and doing, a pillar of the Fichtean position, is in no way the expression of a nature that is given; nothing obliges us to conform to it, and many types of discourse run against it. In effect, this identity is the very act of detaching ourselves from the natural by prescribing a non-natural purpose for ourselves. As an act, this identity is the beginning of rationality, the initiation of a purpose that one adopts freely for oneself. The exhortation “be rational” is not nature; it is rather a requirement. In the act of knowing, there is no question of the natural; on the contrary, it is a question of detaching oneself from nature so that freedom may appear. Knowing is not an occurrence of nature; rather, as Fink put it, what “is involved in that primordial act of knowing is this – the freeing of freedom” [58]

=======

[1] Henrich D. (1979) “Challenger or Competitor? On Rorty’s Account of Transcendental Strategies”,  Transcendental arguments and science (Bieri, Hortsmann, Krüger, editors; Boston)

[2] Rorty,  R (1984) « Arguments transcendantaux, autoréférence et pragmatisme » Cahiers sciences techniques et sociétés, 04,  p.61-87.Translation of the article in the collection Transcendental Arguments and Science, Boston 1979.

[3] Bitbol, M (2000) In his  article « Arguments transcendantaux et physique moderne » Bitbol undertakes to test their bearing on quantum physics. La querelle des arguments transcendantaux (2000), Cahiers Philosophiques de l’université de Caen, N° 35 (henceforth CPUC, 35) (Caen, France, 2000), p. 81 à 101

[4] Strawson, R (1966), (London, Methuen and Co), especially in pages 72 to 89.

[5] Evans, G (1982) (Oxford, Clarendon press).

[6] Stroud, B (1968) « Transcendantal arguments », Journal of Philosophy,  45, 9, (Columbia), p. 241- 256  (Translated in French in Kant analysé, Cahiers de philosophie de l’Université de Caen, 1999, N° 33).  Stroud argues that transcendental arguments are thus committed to demonstrating the impossibility or illegitimacy of the skeptical challenge by showing that certain concepts are necessary for thought and experience.

[7] Op. Cit. p. 28.

[8] D. Heinrich (1979) points out that even though Kant may be a metaphysical realist and although the paradigmatic case of a so-called  transcendental argument (no doubt the only instance in Kant) may occur in the refutation of idealism, his epistemological program is quite independent of these traits, in Challenger or Competitor ?, Op. cit p.114.

[9] Hintikka, J (1972)  See  “Transcendental Arguments: Genuine and Spurious”, Noûs, Vol 6, 3,  p. 27-54.

[10]Several translations of Wissenschaftslehre are possible ( the science of knowledge or the science of knowing).I accept the reasoning of Walter E. Wright in the introduction to his translation The Science of Knowing: J.G. Fichte’s 1804 Lectures on the Wissenschaftslehre (State University of New York Press: Albany, 2005), p. 10-12;  but I will write more often the German expression Wissenschaftlehre, like D. Breaezeale (see for example his introduction  to the translation of The system of ethics (Cambrige University press, 2005).

[11] In recent years, a number of historians of philosophy have studied the roots of transcendental arguments in German Idealism, see : Franks, P (2005) All or Nothing: Systematicism, Skepticism and Trasncendental Arguments in German idealism (Harvard),  Forster, M. (2008) : Kant and Skepticism (Princeton 2008) ; Sacks, M (2005) : “Kant’s First Analogy and the Refutation of Idealism”, Proceedings of the Aristotelian Society, and  “The Nature of Transcendental Arguments”, International Journal of Philosophical Studies. See, for the French -language discussion, Thomas-Fogiel, I (2004) : Fichte, reflexion and argumentation, (Paris, Vrin) and  (2006) Référence et autoréférence, (Paris, Vrin) (American translation by Lynch, R.A, Reference and Self-Reference: On the "Death of Philosophy" in Contemporary Thought, Columbia University Press, 2010). For the Anglo-Saxon approach of the roots of transcendental arguments, it should be noted that the debate mainly concerns Kant or Hegel, but less often Fichte. This debate turns on the questions of realism, namely the question about the relation of my propositions and the world. Others concerned with this theme including Stroud, Rorty, and Mc Dowell, J (1994), Mind and World, (Harvard, p. 42) focus their analysis on the relation of propositions and the facts of world (problem of realism). The economist Lawson, T (2002) raises  a similar question in relation to economics in Economics and reality, (London and New-York, Routledge, Taylor and Francis). We will show, in this article, that, for Fichte, isn’t the question more important. His philosophy makes possible a new metaphilosophical approach to transcendental arguments. For a application of this point of view in economics, see, in France, Parthenay, C (2008) Vers une refondation de la science économique, économie et argument transcendantal (Paris, Cerf).

[12] Schulze G.E [1792] (1996) , Aenesidemus, (when it appeared without the author’s name), republished by Meiner, Hamburg, 1996. It presents a dialogue between Hermias and Aenesidemus; the former defends Kantianism while the latter, speaking for the author, defends skepticism against Kant. When I refer to Aenesidemus in this essay, the reference is to this work by Schulze rather than to the historical model for Schulze’s construct, the first-century founder of Pyrrhonian skepticism.

[13] [1794] (1964) Letter to Reinhardt, in Gesamtausgabe der Bayerischen Akademie der Wissenschaften, directed  by R. Lauth and H Gliwitzky, ed. F. Frommann, Stuggart, (henceforth G.A)  III, 2, 39. He adds: “But he reinforced my conviction that it (philosophy) will be able to become one (a science) within a short period of time.”

[14] On the subject of Maimon another skeptic, Fichte writes: “I am in total admiration of Maimon. I believe, I am ready to prove, that he completely shatters, top to bottom, the whole of Kantian philosophy.” G.A. III, 2, 275. In WL  (1794), Fichte will declare Maimon to be the greatest thinker of his time. For this importance of Maïmon in German Idealism, see Breaezale, D (2003) "Reinhold über Maimon über den Gebrauch der Fikzionen in der Philosophie" in Die Philosophie Karl Leonhard Reinholds (Rodopi). And so Thomas-Fogiel, I. (2000) Critique de la représentation, (Paris, Vrin), in particularly chapitre II, and the articles  (2009) “coalition des systèmes et topologie des contradictions. La pratique herméneutique de S. Maïmon »,  Revue germanique internationale, Haskala et Aufklärung, (CNRS éditions), 9, p. 35-52 and 2010 « réversibilité du scepticisme ou la raison illimitée, le scepticisme de Maimon face à celui des nouveaux humiens, Revue de métaphysique et de morale, N°1, 2010.

[15] For the influence of Aenesidemus in our day, read Manfred Frank’s introduction to the re-publication of Aenésidemus, Hamburg, Meiner, 1996. On the nature of skepticism before Kant, see Thomas-Fogiel, I (2007) “Leibniz or Thomasius ? An Enquiry Concerning German Modern Philosophy, in Idealistic studies, 3.

[16] Traditionally, these are the recurring arguments of Kantian commentators against Aenesidemus. On this topic see the chapter devoted to Aenesidemus in E. Cassirer’s Systèmes postkantiens and in Leçons de métaphysique allemande by Rivelaygue, J (1990) Volume I, (Paris, Grasset), p. 153 ff. The French book : Fichte, Réflexion et argumentation, ( Thomas-Fogiel, I, 2004) defends Aenesidemus in more extensive and precise terms.

[17] For instance: “The Critique of Pure Reason attempts to refute Hume’s skepticism only by presupposing, as already certain and affirmed, the  propositions that were the exact target of Hume’s skeptical doubts.” Op.cit. p.100

[18] “The Critique of Pure Reason provides final proof of this proposition: necessary synthetic judgments issue forth from the spirit and its specific mode of operation; or again, from a use of the causal principle that contradicts its own principles regarding the applicability of categories; or again, they are completely lacking in proof according to their own principles.” Op.cit p.99.

[19] The alternative is as follows: 1) Either what the Critique says is true, then for an instance of knowledge to be true, an intuition and a concept are required. Yet the account of the conditions of the possibility of knowledge does not satisfy this criterion of truth since these conditions cannot be represented by intuition and concepts. It follows that if what the Critique of Pure Reason says is true, then the Critique of Pure Reason is false. 2) Alternatively, there exists a type of truth other than the connection between a concept and an intuition. In this case, Kant is in error when he says that the only truth is that of judgments in mathematics and the physical sciences. Here too it follows that the Critique of Pure Reason is false.

[20]Aetas kantiana (1968), Tome 244, in two volumes, II, p. 578, Brussels

[21] This can be restated in the terms of the letter of February 21 1772 to Marcus Herz: if the problem of philosophy can be explained as the problem of the equivalence between our representations and things, various solutions are then possible: either we deduce the subject from the object (classical realism or realism in the manner of Spinoza); or we deduce the object from the subject (the idealism of Leibniz); or again, we deny that there can be any equivalence, that is, we dismiss all connection between subject and object.

[22] Op.cit, II, 275.

[23] In effect, it is this “need for science” that spurs his 1790 conversion to criticism. What drives his urge in 1793 to “rebuild everything” is the question of the unscientific character of Kantian philosophy. Finally, all his later works will set out his relationship to Kant in terms of additional scientific character. He writes in this connection in 1798: “With Kant, the laws of human thinking are not deduced in a rigorously scientific manner; this however is what the science of knowing is intended to achieve.” In 1805, in The Way Towards the Blessed Life, the Kantian point of view is conceived of as a common-sense perspective that does not meet the requirements of “grounded knowledge.”

[24]  The science of Knowledge [1794] (2003), edited and translated by P.Heath and J. Lachs, (Texts in German Philosophy) (Cambridge university press, New-york), p. 118.

[25] On this point see his article “Transcendental Arguments, Autoreference and Pragmatism” in Transcendental Arguments and Science, Op. cit, p.78 ff.

[26]On this point see the introduction by Chauvier S (2000)  to “La querelle des arguments transcendantaux”, CPUC, 35, op.cit. p. 8..

[27] The Science of Knowledge, op.cit, p. 82

[28] Sämmtliche Werke ((henceforth SW), edition of Fichte I.H. (1845-46), reprinted in Fichtes Werke, (1971) W. de Gruyter and Co, Berlin,Volume X, p.3. Two sentences follow in which Fichte calls Spinoza the most coherent embodiment of ontology, immediately adding that he was not “attentive to the thinking about the formation of being”. The WL of 1813 is not translated in English.

[29] SW, (1971), op.cit, X, p. 3 and 4..

[30] On this importance of epistemological problematic, see for exemple Breazeale, D (2000)  :  "The Spirit of the Wissenschaftslehre," in The Reception of Kant's Critical Philosophy: Fichte, Schelling and Hegel  (Cambridge University Press).

[31] Prolegomena, paragraph 13, AK, IV, 293. The transcendental enquires into “all knowledge generally concerned not with objects but with our mode of knowing objects as they must exist a priori”. Critique of Pure Reason, Introduction, VII, AK III, 43.

[32] In the Austinian context, if a statement is made not as a description of a fact but as the actual doing of an act (a performative), such as “The meeting is open” or “I bless you,” then its pertinence (or success) supposes conditions provided by the contingent, exterior context in which it is made.

[33] Here we clearly and unambiguously see the self-reflection of the term “know” in the expression “if he knows it, it is a mystery to know from where he knows it.”

[34]  The Science of Knowledge, Op.cit, p 24.

[35] Zahar, E (2000) « Mathématiques applicables et arguments transcendantaux », CPUC 35, Op.cit. p. 104.

[36] ibid.

[37] Boyer, R, (2000) “Why are there transcendental arguments” (« Pourquoi des arguments transcendantaux »),  CPUC, 35, op.cit., p. 37.

[38] Russell, and after him, a certain part of the analytical tradition, rejects the notion of “self reference” . He doesn’t consider that the problem of self –reference is an important theme.. See on this topic,  Thomas-Fogiel. I, (2000) : “Autoréférence et autoréflexion, le problème de la relation d’un X à lui-même chez Russell et Fichte, in Revue de metaphysique et de morale, Paris, PUF, N°2. Even Austin refuses the concept of « self reference ». See on this point, for example Thomas-Fogiel, I (2004) : “Plaidoyer pour le langage philosophique. Du spéculatif à l’ordinaire et retour ? Austin, Searle, Cavell et la tradition spéculative”. Revue de metaphysique et de morale, N°2, in Usages d’Austin, dir. by S. Laugier and I. Thomas-Fogiel, and so, in English, a article which will publish in beginning 2010, in Fichte-Studien “Fichte and Austin”, presented at an international Fichte meeting, in Halle (Germany) in 2006.  In contrast to this tradition, an author like R. Brandom restarts  this problem in reference to Hegel’s problematic. On this point, see the presentation by I. Thomas-Fogiel of the French translation of Making it Explicit, Paris, Cerf , 2010. On the contrary , according to A. Boyer  : either an argument is an instance of classical modus ponens or an induction. But the performativ contradiction isn’t  a deduction in the classical sense. There isn’t a premise that differs from the conclusion, but rather an incompatibility between the content of a proposition (for exemple “I don’t speak) and the act of enunciation (I must speak so that I  can say : “I’don’t speak). In effect, we are faced in this instance with a pragmatic contradiction and not a classical deduction.

[39] See SW, paragraph 4, for example I, 211 et sq. Concerning this accord between “product” and “production”, and its great importance in the sequence of propositions in the WL1794, see Thomas-Fogiel, I (2000) Critique de la représentation, Etudes sur Fichte, (Paris, Vrin).

[40] SW, op.cit .I,  212  “Erkärende und Erklarungsgrunde”. ,English  translation p. 190 op.cit

[41] “Einige Vorlesungen über die Bestimmung des Gelehrten” (1794); "Some Lectures Concerning the Scholar's Vocation" in Fichte: Early Philosophical Writings,  Breazeale, D. trans., (Ithaca: Cornell U. P., 1988) pp. 144-84.

[42] For example, in the Critique of Pure Reason, the “saying” defines truth as the connection of a concept and an intuition; whereas the “doing” of Kant surreptitiously posits a different definition of truth (since his proposition is not the connection of a concept and an intuition), which invalidates the content of his proposition. Thus, what he says (Sagen) does not accord with what he does (Tun).

[43] A particular reading of Gamma 4 could show Aristotle making use of this type of reasoning in order to refute the sophist who denies the principle of non-contradiction (one of the principles of formal logic). In fact, to deny this principle, the sophist would be obliged to make use of it; his “doing” would not “accord with his saying” and, as Aristotle says, he would be unable to speak further. He concludes: “It is ridiculous to speak with someone who can say nothing further.” Nevertheless, even if this type of reasoning has been known for a very long time and is one of the tools for refuting skepticism, it remains in Aristotle an ontological argument – which it ceases to be as transcendental argument. Moreover, the uniqueness of Fichte consists in elevating it to the position of principle of all principles, the foundation of all philosophical discourse, which (as we shall see) allows us to issue new claims with respect to classical problems.

[44] This expression comes from Récanati, F (1979),  La transparence et l’énonciation (Paris, Seuil). On this point, see Thomas-Fogiel, I (2000) : “Autoréférence et autoréflexion, la relation d’un X à lui-même chez Russell et chez Fichte”, Revue de Métaphysique et de morale, 3, and the book  (2006) : Référence et autoréférence.

[45] We prefer this term to the term “performative  contradiction”, but whatever the formulation, this type of contradiction is defined as the cancellation of what is stated (“the effected”, says Fichte) by the fact or the act of stating (“the effectuation”).

[46] The science of Knowing 1804, op. cit p. 141

[47] Ibidem.

[48] Ibid. p. 141. We prefer to translate the German term “Tun” by the term “doing”.

[49] The Wl of 1813,  SW, X, p. 11. We translate. In this and succeeding citations, ‘unity’ could also be rendered as ‘oneness.’

[50] Ibid. p. 15.

[51] Ibid. p. 12.

[52] We follow the translation of this term by Breaezeale D.; see for example the translation of the System of Ethics, 2005, Cambridge University Press, Glossary p. 346.

[53] Op.cit p.169

[54] Ibid, p. 169.

[55] See The Science of Knowledge 1794, p.192 and 193, and the development “no infinity, no bounding (Linie), no bounding, no infinity…. But for all that it must also posit beyond this line…”

[56] About these objections, consult the presentation Stern, R (1999) Transcendental Arguments: Problems and Prospects (Oxford, Clarendon Press) ; as well as the article of  Virvidakis, S (2001)  “Les arguments transcendantaux et le problème de la justification de la normativité morale”, Philosophiques, 28/1, Printemps, p. 109-128.

[57] Quine (1995) “Naturalism; or, Living within One's Means.” Dialectica, 49, 264,  p.261.

[58] Fink, E  [1957] (1990) “ Réflexions phénoménologiques sur la théorie du sujet ”, 1990, Epokhe, 1 (Jérôme Millon, Grenoble),  p. 28.