Nous sommes tous néokantiens.Michel Foucault1In the 1980s and early 1990s, when poststructuralism set the agenda in cultural theory and shaped the way in which theorists from other traditions were received, the work of the Bakhtin Circle was often seen as anticipating contemporary concerns to a quite uncanny extent. While some adopted Bakhtin as a poststructuralist avant la lettre, others seized on Bakhtinian ideas as an alternative way of dealing with the very issues poststructuralism had raised without disappearing into the poststructuralist void of Derridaʼs ʻoutside textʼ or partaking of Foucaultʼs metaphysics of power. Gradually, it became apparent that despite a reiterated adherence to the ʻconcrete eventʼ and ʻsocial contextʼ, Bakhtinian theory was itself as thoroughly anti-realist as the poststructuralists themselves.
Few were prepared to search for the grounds of perceived correspondences in intellectual history, partly because, in a common effort to maintain a politically radical public proﬁle, all three theorists kept their own philosophical sources well out of sight. Indeed, despite Foucaultʼs above-mentioned invocation of neo-Kantianism, the extent to which he or Derrida realized the traditions behind their own ideas is unclear, since they only considered their immediate theoretical ancestors. Bakhtin was rather more aware of his place in intellectual history and it is recent research into the sources of his ideas that has made the current investigation possible.  I shall argue here that these three ﬁgures share roots in a common philosophical tradition: neo-Kantianism, speciﬁcally that of the Marburg School.
Few in the Anglophone world are primed to recognize neo-Kantian traits in cultural theory these days. Sustained works on neo-Kantianism in English are Neo-Kantianism in cultural theory Bakhtin, Derrida and Foucault
rather scarce, and those that trace its inﬂuence on social and cultural theory have focused on the inﬂuence of Wilhelm Windelband and Heinrich Rickert on Max Weber. The Marburg Schoolʼs inﬂuence on Durkheim via Charles Renouvier (1815–1903), Léon Brunschvicg (1869–1944), Octave Hamelin (1856–1907) and Emile Boutroux (1845–1921) has been given much less attention than the inﬂuence of Comte and Spencer.  Whilst a recent upsurge of interest in the work of Georg Simmel has given rise to work on his sociological writings, his distinctive and no less inﬂuential forays into neo-Kantianism and Lebensphilosophie have been rather less well served. The work of the late Gillian Rose is a notable exception, providing learned assessments of the traces of both Baden and Marburg School neo-Kantianism in some areas of classical sociology, Western Marxism and poststructuralism.  Many of Roseʼs critiques are, however, little more than sketches, and the unfamiliarity of neo-Kantian ideas – coupled with her own very dense, Hegelian proclivities and juridical focus – mean that her work has not been widely received. 
Bakhtin studies have, however, bought neo-Kantianism and the philosophies it spawned back into focus, allowing us to reassess recent intellectual history and better diagnose the malaise afﬂicting much contemporary theory. While the Baden School has been relatively well served by translations in English, with the exception of the work of Ernst Cassirer and the late Jewish writings of Hermann Cohen, that of the Marburg School remains largely untranslated. The earliest translated work by Cassirer, Substance and Function, marked the beginning of his divergence from the School and his convergence with a speciﬁc variety of neo-Hegelianism.  Bakhtin began reading the central Marburg School texts quite early, encouraged by his friend Matvei Kagan, who had studied in Marburg under Cohen and Paul Natorp and with Cassirer.  In the 1930s, Bakhtin followed Cassirer towards Hegel, and also maintained a deﬁnite connection with key elements of the Marburg method. The Marburg inﬂuence on Derrida and Foucault come via different routes – including Husserl, Heidegger, Brunschvicg and classical sociology – so materializes in different ways. There are deﬁnite points, however, where the shared heritage is apparent.
When Roy Bhaskar notes that Derrida has an ʻunfortunate tendency to elide the referent in the semiotic triangle … which deconstructs his own practice which cannot thereby be theoretically sustainedʼ, it is the neo-Kantian basis of the method that has been identiﬁed.  Even the corporeally insistent historical works of Foucault and Bakhtin are rather different to the materialist works they once appeared. The subtitle of Foucaultʼs 1963 The Birth of the Clinic illustrates this well: An Archaeology of Medical Perception. The book deals with the production of an object of knowledge according to ʻnew forms of the mathematical a prioriʼ, a consideration of which also brings his methodological opus, The Order of Things, to a close. In the case of both Bakhtin and Foucault, the body is a historically changing object of knowledge. In the Renaissance the individual body is a microcosm of the universe, with life and death having a cosmic signiﬁcance. Over the next two centuries the individual life is separated and the cosmic co-ordinates lost.  This is the way the transformation of perception – medical and otherwise – occurred, and all other considerations are beyond knowledge. This characteristic ʻbracketingʼ is clearly stated in the following comment by Bakhtin:
Three centuries ago the ʻwhole worldʼ was a unique symbol that could not be adequately represented by any model, by any map or globe. In this symbol the ʻwhole worldʼ, visible and cognised, embodied-real, was a small and detached patch of earthly space and an equally small and detached chunk of real time.
Everything else unsteadily disappeared into the fog, became mixed-up and interlaced with other worlds, estranged-ideal, fantastic, utopian. The point is not that the other-worldly and fantastic ﬁlled-in this impoverished reality, combined and rounded reality out into a mythological whole. The otherworldly disorganized and bled this present reality. 
The world literally was that symbol, bled and disorganized by mythical thinking. The symbol did not represent the world badly, but it was the world itself. That symbol could not be represented, it could not appear for itself but existed only in itself. Culture had no self-consciousness because the power of mythical thought was sufﬁcient to prevent any such objectiﬁcation. To understand the nature of these intellectual enterprises we must turn to their philosophical roots.
The marburg school project
Neo-Kantianism was a rather misleading term since the revisions of Kant were fundamental. Where Kant argued that concepts are validated in their a priori application to the empirical world, the neo-Kantians, following the work of R.H. Lotze, argued that the validity (Geltung) of propositions is established independently, by logic. The realm of values and validity is now akin to Platoʼs Ideas so that what has validity (was gilt) is quite different from what is (was ist) and there is no point of transition between them. The ramiﬁcations of this move were brought out in the work of Cohen, the leader of the Marburg School, who argued that it no longer made sense to speak of a ʻthing in itselfʼ if validity was autonomous. ʻObjectiﬁcationʼ meant not the relationship between the mind and an empirical world, but the production of the object by and in thought. Cohen was uneasy about the term ʻproduceʼ since this implied production of and from ʻsomething placed outsideʼ; instead he adopted the term ʻoriginʼ (Ursprung). Cohenʼs intention was later clariﬁed by his student Cassirer:
The primacy of activity over possibility, of the independent-spiritual over the sensible-thinglike, should be carried through purely and completely.
Any appeal to a merely given should fall aside; in place of every supposed foundation in things there should enter the pure foundations of thinking, of willing, of artistic and religious consciousness. In this way, Cohenʼs logic became the logic of the origin. 
Marburg neo-Kantianism was militantly antipsychologistic, and this made it a powerful inﬂuence on the development of ʻtranscendental phenomenologyʼ. Paul Natorp, the second major Marburg neoKantian, argued – in an article that Husserl cited as the catalyst of his own anti-psychologistic turn – that the ʻthing in itselfʼ was at best a limiting concept which organized thought, ʻan unknown X that we endlessly deﬁneʼ. However, he also argued that the ʻobjective validityʼ of that deﬁnition is ʻindependent of the subjectivity of knowledge … what is to be objectively valid, is to be valid apart from the givenness of its representation in this or that consciousness.ʼ  Validity is established in accordance with the ʻfactual validityʼ (faktische Geltung) of the mathematics that underlies each science, with the consequence that, as Cohen put it, genuine actuality consists of science in ʻpublished booksʼ. 
Where Kant argued that dogmatism begins with an analysis of being, and criticism with an analysis of knowledge, the Marburg School argued that the difference lay in the perceived task of cognition. The dogmatic method regards the object of knowledge as given (gegeben) and takes the task of cognition to consist in the cognition of that object, drawing closer to it, revealing what is given in experience from the given object. The critical method, on the other hand, regards experience only as the occasion for the production of the object; the object is not given but set as a task (aufgegeben). This is a never-ending task (unendliche Aufgabe) because all thought is in becoming (das Werden). Despite the Marburg Schoolʼs humanism, it is possible here to see the seeds of Heideggerʼs critique of ʻpresenceʼ, which plays such a central role in Derridaʼs work.  In the Marburg formulations Kantʼs transcendental logic now becomes a pure logic (Cohen) or a general logic (Natorp), methods in which the object can and should be ʻproducedʼ from and in thought.
This is an uncompromising anti-realism incompatible with any form of reference to the empirical world, no matter how self-critical or falliblistic. Being is now equated with being known; ʻmetaphysicsʼ, as the French neo-Kantian Brunschvicg put it, ʻmay be reduced to the theory of knowledgeʼ.  The divergence between, on the one hand, the neo-Kantian approach to language and experience and, on the other, analytical philosophy with its concentration on reference, can be traced back to the initial divergence between Lotze and Frege.  But it is in the work of the Marburg School that the implications of the divergence become apparent: all knowledge of the empirical world is excluded in principle. As we shall see, the Bakhtinian and poststructuralist philosophies of language are typical developments of neo-Kantianism, which in their own ways transform the object of cognition – the signiﬁed – into a never-ending task validated by the object domains of individual sciences.
Derrida: the origin of différance
With his famous – or notorious – declaration that ʻthere is no outside-textʼ, Derrida undertakes a neo-Kantian reworking of structuralist linguistics startlingly reminiscent of Marburgian Geltungsphilosophie. Where Brunschvicg had developed a Marburgian ʻmodality of judgementʼ in which knowledge constitutes a world for us, Derridaʼs ʻIl nʼy a pas de hors-texteʼ does not posit an ontological ʻnothingʼ outside the text17 but claims that language constitutes the world for us beyond which nothing can be justiﬁably posited. Both Foucault and Bakhtin agree with this banishment of ontology and attack on representation as mimesis in their own respective neo-Kantian moves; all three effectively present their work as ʻCopernican revolutionsʼ in philosophy, deconstructing the ʻcopy theoryʼ of relations between language and a ʻgivenʼ object in favour of an investigation into the discursive constitution of the object of knowledge.  For all three, the object can only be known as signiﬁed, but never ﬁnally and entirely known since it is never actually present. In the case of Derrida, presence is eternally disrupted by the ʻplay of absence and presenceʼ within a constantly shifting signifying system.
Rose has shown that Saussureʼs Course in General Linguistics proceeds along Kantian lines, with the langue/langage distinction corresponding to that between the a priori (precondition) and the empirical (conditioned). Saussureʼs distinction between ʻconceptʼ and ʻsound-imageʼ further corresponds to Kantʼs distinction between concept and (given) intuition/representation; Saussureʼs account of linguistic value to Kantʼs move from ʻdescriptionʼ to transcendental reconstruction where concept (signiﬁed) and sound image (signiﬁer) are shown to be a construct of signiﬁcation.  The Derridean critique of Saussure actually follows many of the moves of the neo-Kantians in their critique of Kant. Langue (a priori) now ʻproducesʼ language (empirical), but the subject is no longer able to halt the play of difference to achieve a stable meaning (object of cognition). That meaning is produced by the play of difference (the process of becoming). Where the neo-Kantian object of cognition is never ﬁnally deﬁned (present) but is constantly being produced, Derridean presence is endlessly deferred. Difference-ruled deferral famously becomes différance, a Derridean version of Cohenʼs ʻoriginʼ, which also ʻdifferentiatesʼ.  This is the ʻoriginary differenceʼ between Kantʼs quid facti and quid juris recast in Lotzeʼs contrast of what is and what has validity to become an opposition between being and meaning, the elision of which constitutes metaphysics. As Vincent Descombes – one of the only writers to recognize the inﬂuence of neo-Kantianism on Derrida – argues, the identity of being and meaning is, for Derrida, only ʻat inﬁnityʼ, inﬁnitely deferred like neo-Kantian ʻinﬁnite tasksʼ and ʻinterminable “teleologies”ʼ. Derridaʼs notion of ʻtraceʼ, which highlights ʻthe sign left by the absent thing after it has passedʼ, is a mark of this originary delay. As Descombes puts it, without reference to Cohen: ʻif every present bears the trace of an absent which circumscribes it (and by which, in this sense, it is constituted, produced and given to be what it is), then paradoxically an “originary trace” must be conceived of; that is, a present trace of a past which never took place – an “absolute past”.ʼ  Furthermore, Habermas, another neo-Kantian thinker, refers to Derridaʼs temporalized Ursprungsphilosophie (philosophy of origin) that is rooted in Jewish mysticism.  This is again reminiscent of Cohen, whose philosophy was a particular convergence of neo-Kantianism and Jewish mysticism, the messianic aspects of which one also ﬁnds in Bakhtinʼs notion of carnival.
The Bakhtin Circle: neo-Kantian philosophy of language
The Bakhtinian linguistic development of neo-Kantianism is very different, not least because it is not based on Saussurean linguistics. The Bakhtin Circle subscribed to a philosophy of language that stressed subjective spontaneity in communication while opposing the psychologism of Romantic linguistics. In a recently published plan for Valentin Voloshinovʼs Marxism and the Philosophy of Language, Cassirerʼs 1923 work The Philosophy of Symbolic Forms vol. 1: Language is hailed as initiating a new phase of neo-Kantianism:ʻthe wordʼ becomes a partition [sredostenie] between transcendental validity and concrete actuality, a ʻthird realmʼ, as it were, lying between the cognising psycho-physical subject and the empirical actuality surrounding him on the one hand, and the world of a priori, formal being on the other. At the same time, the form of the sign and signiﬁcance (symbolic form) is common to all regions of cultural creativity, uniting them.
Such is the systematic place of the word in the teachings of the neo-Kantians.… It is precisely on the ground of the philosophy of language that the Marburg Schoolʼs abstract scientism and the Freiburg Schoolʼs abstract ethicism is now being overcome. By means of the inner form of language (as a semi-transcendental form) movement and historical becoming is being introduced into the petriﬁed realm of transcendental-logical forms. It is also on this basis that an attempt to reestablish the idealist dialectic is being made. 
Voloshinov here signals the neo-Kantian appropriation and adaptation of Wilhelm von Humboldtʼs organicist philosophy of language in its evolution towards a type of neo-Hegelianism. In this new adaptation, language becomes a ʻpartitionʼ between ʻlifeʼ as understood by Lebensphilosophie and objective culture (validity). Where in Bakhtinʼs early work the ʻact of our activity, of our experience, like the two-faced Janus, looks to both sides: at the objective unity of cultural realms and at the unrepeatable singularity of experienced lifeʼ, so now language embodies the ʻpartitionʼ between these ʻmutually imperviousʼ worlds.  This development is key to overcoming what Simmel had called the ʻtragedy of cultureʼ resulting from the confrontation of life and objective culture. The circularity of precondition and conditioned was now embraced; as Cassirer put it, the ʻliving process of speechʼ is one in which ʻindividuality and universality are contained as equally justiﬁed and equally necessary aspects of the process. The universal is that through which the individual constructs the world, and the universal is what constructs the world of the individual.ʼ  The linguistic structure now shifts, ʻimperceptiblyʼ, with every utterance and this shifting structure ʻconstructsʼ the individual utterance. As Bakhtin put it in 1961, the utterance ʻcreates something that before it never was, something absolutely new and unrepeatable.… But something created is always created from something given.… Everything given is transformed into what is created.ʼ While ʻpositivistʼ scholarship studies only the ʻgiven in the createdʼ, the object is nevertheless ʻcreated in the process of creation, the poet himself is created, and his world view, and the means of expressionʼ.  Here we have the Bakhtinian recasting of the Marburg never-ending task: thought as thought of origin.
Where Cassirer recast Humboldtʼs ʻinner-formʼ as the ʻlaw of signiﬁcationʼ, Bakhtin concentrated on the relationality of discourse and called this ʻdialogismʼ, the ongoing and never-ending process of intersubjective thought-exchange, the units of which are utterances. Dialogism is the logic of thought-exchange on the basis of language and, like all the ʻhuman sciencesʼ, has an ethical signiﬁcance. For Cohen, the ʻmathematicsʼ of ethics is jurisprudence, and it is for this reason that dialogism is both juridical and normative. In the novel, as the most self-consciously dialogic form of complex (meta-)utterance, sociospeciﬁc discourses are brought to an inquest, with the novelist himor herself not ʻspeakingʼ directly.  The utterances of the hero are deﬁned by their relations towards (a) the object; (b) the given linguistic system of possibilities for expression; and (c) other utterances within the sphere of intercourse. In this way the discourse can be judged ʻcorrect (or incorrect), true, right (or false/wrong), beautiful, just etc.ʼ. The author thus plays the role of the assumed but unheard ʻsuperaddresseeʼ who, like Kantʼs theoretical reason, sits in judgement. Outside the literary work, however, the presence of such a judge is simply assumed as a structural requisite of dialogue, the ʻthirdʼ the ʻabsolutely just responsive understanding of whom is presupposed either at a metaphysical distance or in distant historical timeʼ. 
Linguistics: the regional sciences of discourse?
The Marburg School sought to justify the existence of distinct object domains by establishing the ʻfactual validityʼ of mathematical thinking underlying each individual science. Sciences, they thought, were essentially regional, with a priori boundaries and principles, the erasure of which is fatal to science itself. Neo-Kantianism was thus the ultimate self-justifying academic philosophy. According to the Marburgers, jurisprudence, the science of legal concepts, is the ʻmathematicsʼ of the cultural sciences. Bakhtin and Voloshinov sought to establish a science of discourse as a cultural science, which is validated according to the ʻmathematicsʼ of jurisprudence, producing its objects but not implying that the existing normative linguistic structure (like the existing law) is the embodiment of rationality. Linguistics is a speciﬁc object domain, but one whose method is that of the natural sciences; this lies behind Bakhtinʼs hostility to Saussurean linguistics. La langue is a system of norms that cannot be questioned: a method appropriate for the natural sciences is imposed on the cultural sciences so that linguistics claims to explain cultural phenomena. The existing standard of value must be able to be questioned if the production of the object is a never-ending task, and this is especially the case in the science of culture, the object of which is a ʻteleological formationʼ. 
Bakhtinʼs most systematic comments on this matter are in his late notes on establishing a ʻmethodology for the human sciencesʼ and his major 1953 essay on ʻdiscursive genresʼ, the latter of which takes advantage of Stalinʼs recent consignment of linguistics to the natural sciences.  The neo-Kantian underpinnings of Bakhtinʼs argument are especially clear in recently published archival notes for the discursive genres essay, where Bakhtin tries to justify this new regional science of discourse by arguing that only this can analyse the ʻrelatednessʼ of linguistic meaning (znachenie) to ʻobjective validityʼ – that is, the ʻtruthfulness, beauty, veracity, necessity, expressiveness, sincerityʼ of the utterance.  These categories are what Kagan, following Cohen, called ʻundoubted facts of actualityʼ, the ʻspheres of objects and problems of separate scientiﬁc and cultural disciplinesʼ into which ʻactuality is broken upʼ.  The same feature found its way into the neo-Kantianism that shaped French philosophy. As Durkheimʼs teacher, Hamelin, put it, quoting Renouvier, such ʻcategories are necessary phenomena in relation to our minds … they are necessary truths … which are the condition of [the understanding] being used.ʼ  Like Bakhtin, Derrida seeks to found a new regional science of language in use, though on rather different premisses. His characterization of science and of writing, the science of which is grammatology, is typically neo-Kantian: science is a ʻtaskʼ in which ideal objects are produced, while writing is ʻthe condition of the possibility of ideal objects … the condition of the epistémè.ʼ  Derrida draws heavily on Husserlʼs essay ʻthe Origin of Geometryʼ, which, as Rose shows, is itself a ʻclassic piece of neo-Kantianism in the Marburg styleʼ:
It seeks to justify an exemplary or regional science not knowledge as such; it drops the distinction between appearances and things in themselves; it turns the question of transcendental possibility into the delineation of a productive origin; and it deﬁnes the a priori metacritically as ʻcultureʼ or ʻhistory.ʼ ʻCultureʼ or ʻhistoryʼ becomes the name for the source of signiﬁcation which repeatedly creates or posits its idealities or validities: this historical beginning is deﬁned as ʻorigin in an accomplishment, ﬁrst as a project and then as a successful executionʼ. 
Derrida sets writing as a metacritical a priori one stage back from that of Husserl: ʻbefore being the object of a history – the object of an historical science – writing opens the ﬁeld of history – of historical becoming. And the former (Historie in German) presupposes the latter (Geschichte).ʼ  Writing, the world of and in signs, now becomes the mathematics of history as science (method) and the object of history is ʻproducedʼ by différance. Derrida criticizes the notion that experience ʻalways corresponds to a certain type of factual or regional experience (historical, psychological, physiological, sociological, etc.), giving rise to a science that is itself regional and, as such, rigorously outside linguisticsʼ  for its phonologism. Reversing the hierarchy between speech and writing means that regional sciences themselves become conditions of ʻexperienceʼ: regional sciences give rise to (objectify) regional experience.
Bakhtin makes a similar point when he argues that ʻcommunication requires objective validity (in all its various forms depending on the sphere of intercourse); without it communication would degenerate and decay. All utterances have dealings with objective actuality regardless of the consciousness or will of people (speakers, those engaged in communication), and regardless of communication itself.ʼ  Derridaʼs notion of ʻarche-writingʼ as the ʻpure movement which produces differenceʼ  is thus akin to Cassirerʼs ʻlaw of symbolizationʼ but without the organicism which the latter inherited from von Humboldt and bequeathed to Bakhtin. History is transferred wholesale to the realm of culture and made a productive origin. As Habermas recognizes with reference to Derrida, the printed form of language severs the text from the context in which it arose, from concrete connections with individual subjects and gives it an autonomy in relation to all living contexts, so that the text becomes readable in all changing contexts. The semantic content (Natorpʼs ʻobjective validity) is thereby saved from psychologism.  Derridaʼs embrace of writing is thus similar to Cohenʼs notion of genuine actuality being science in published books.
Life, culture and the inbetween
Foucault presents the most systematic poststructuralist attempt to develop a neo-Kantian scheme of regional validities and to relate this to discourse in life. The following passage from The Archaeology of Knowledge shows the connection with the principles of Cohen and Derrida but it also provides a useful bridge to Bakhtinʼs work:
Different oeuvres, dispersed books, that whole mass of texts that belong[s] to a single discursive formation – and so many authors who know or do not know one another, criticize one another, … pillage one another, meet without knowing it and obstinately intersect their unique discourses in a web of which they are not the masters, of which they cannot see the whole, and of whose breadth they have a very inadequate idea … they communicate by the form of positivity of their discourse.… Thus positivity plays the role of what might be called a historical a priori. 
Objective validity is here an agitated ﬁeld of discursive interaction strikingly reminiscent of Bakhtinʼs mature work. In a well-known passage, Bakhtin describes how every utterance is a dependent but active participant in a constantly (re)forming discursive mesh of intersecting utterances of which individuals have but partial knowledge:
Every … prose word – everyday, rhetorical, scientiﬁc – cannot but be oriented on the ʻalready saidʼ, the knownʼ, on ʻcommon opinionʼ etc. The dialogic orientation of the word is of course a phenomenon common to every word. It is the natural condition of every living word. On each of its routes toward the object, in all its directions, the word meets the alien word and cannot but enter into a living tension-ﬁlled interaction with it. 
Though Bakhtinʼs passage displays a much stronger phenomenological coloration than that of Foucault, there is a close relationship between the notions of ʻdirectednessʼ and ʻpositivityʼ, since in each case the object of discourse is not given but, in classic neo-Kantian style, posited in the discursive act. In each case the object is posited not by a subject in complete control of his or her positing discourse, but according to a discursive a priori with historical being.
In the preface to The Order of Things Foucault distinguishes three levels at which his analysis will work. The ﬁrst level is constituted by the primary codes of culture:
The fundamental codes of culture – those governing its language, its schemas of perception, its exchanges, its techniques, its values, the hierarchy of its practices – establish for every man, from the very ﬁrst, the empirical orders with which he will be dealing and within which he will be at home.
At the second level, and ʻother extremity of thoughtʼ are ʻthe scientiﬁc theories or the philosophical interpretations which explain why order exists in general, what universal law it obeys, what principle can account for it, and why this particular order and not some otherʼ. Between these two extremes, however, is a third ʻmore obscureʼ level that he calls the ʻpure experience of order and its modes of beingʼ, or the ʻepisteme.ʼ This is regarded as the most difﬁcult sphere to analyse and also the most fundamental, since it occupies the space between the ﬁrst and second levels and provides the basis for the construction of the second level. This realm is described as a sort of ineffable protoreﬂexivity ﬁrmly linked to time and space, shifting and anarchic but also more ʻtrueʼ than the second level. 
Foucaultʼs fascinating exposition is remarkably similar to an analysis developed by Voloshinov in his books on Freud and on the philosophy of language, where he makes a distinction between ʻlifeideologyʼ (zhiznennaia or zhiteiskaia ideologiia, misleadingly rendered as ʻbehavioral ideologyʼ in the English translation) and ideology proper as the realm of objective validity. As Galin Tihanov has shown, this derives from an inventive development of the work of Simmel and Nikolai Bukharin so that ʻlife ideologyʼ becomes a ʻsupply chamberʼ for ideological systems (ʻethics, science, art and religionʼ), displaying a greater ﬂuidity and sensitivity due to a closer contact with social situation; inﬁltrating, inﬂuencing and transforming ideological systems. Voloshinov also distinguishes between different ʻstrataʼ of life ideology, with the lower regions having a greater proximity to ʻlifeʼ than the upper levels, which are more systematized.  The category of life and its relations with the realms of ʻobjective validityʼ was similarly the central question throughout Foucaultʼs career, though his approach to this question changed considerably. Foucaultʼs ʻthird realmʼ roughly corresponds to Voloshinovʼs ʻupper levelʼ of life-ideology, while the ﬁrst level corresponds to the latterʼs lower level. In Bakhtinʼs mature work the relatively stable forms that the different levels of ʻlife-ideologyʼ take are deﬁned in terms of discursive genres, a category which Foucault rejects, with ʻprimaryʼ genres being embedded in direct social interaction:
Secondary (complex) discursive genres – novels, dramas, all types of scientiﬁc research, large publicistic genres etc. – arise in more complex conditions and relatively well-developed and organized cultural communication (primarily written): artistic, scientiﬁc, socio-political etc. In the process of their formation they absorb and rework various primary (simple) genres, which have been composed in direct discursive intercourse. On becoming ingredients of the complex ones, the primary genres are transformed and acquire a speciﬁc character: losing their direct relation to real actuality and towards real alien utterances. 
Here we have a simple bifurcation into what we might call ʻlife-genresʼ (Foucaultʼs ﬁrst level) and the genres of ʻobjective cultureʼ (second level), but in Bakhtinʼs essays on the novel he introduces a third realm consisting of the small parodic genres which are themselves incorporated into the larger, secondary, meta-genre of the novel; this is further expanded in his discussion of the Menippean satire in the 1963 Dostoevsky book. Indeed, Bakhtin develops a scheme in which the genres associated with popular culture (ʻlowʼ humorous genres, and especially carnival) are preconditions of science in the same way as Foucaultʼs third level of ʻpure experience of orderʼ.  These genres stand midway between life and the ʻofﬁcialʼ culture that, like Foucaultʼs second level, is an object of suspicion for its association with authority. In each case the ʻthirdʼ realm represents an ʻorder which divides systems of positivities “before presenting them to the understanding”ʼ, delineating ʻa classic Geltungslogikʼ. 
Foucaultʼs ʻarchaeologicalʼ project is to analyse the relationship between third and second levels:
to rediscover on what basis knowledge and theory became possible; within what space of order knowledge was constituted; on the basis of what historical a priori, and in the element of what positivity, ideas could appear, sciences be established, experience be reﬂected in philosophies, rationalities be formed, only, perhaps, to dissolve and vanish soon afterwards. 
Though very different from those of Foucault,
Bakhtinʼs historical works follow precisely this pattern, focusing entirely on the forms of pre-literary culture which made the novels of, say, Rabelais and Dostoevsky possible. Where Foucault attempts to outline ʻrules of formation which were never formulated in their own rightʼ but were nevertheless intuitable, Bakhtin develops the notion of the ʻchronotopeʼ, or time–space intuitions characteristic of a particular life-form which make certain types of literature possible, and which correspond to speciﬁc discursive formations.  The age of the epic corresponded with Athenian monoglossia, and it was only the collapse of this that made the development of science and ʻrealisticʼ literature possible. The ﬁrst pre-novelistic works were made possible by the ﬁrst inter-animation of cultures and languages – polyglossia – which resulted from the collapse of Athenian democracy and Hellenic expansion. The modern novel was born with the breakdown of the feudal order, which allowed the inter-animation of different social groups and their discourses (heteroglossia).
Rabelais and Dostoevsky were both writing at ʻthresholdʼ times when the old order was in decline and the new order had yet to form, times when the inter-animation of social discourses was particularly acute, leading to a ﬂood of ʻunofﬁcialʼ culture into their novels. The carnivalesque plays a particularly important role here, constituting a sort of genre of the third realm, a transition point between cultureʼs ʻprimary codesʼ and reﬂexive forms. In each case there is no attempt to assess the adequacy of any historically generated form of social consciousness against a world existing independent of knowledge, but only an attempt to uncover the preconditions for historically speciﬁc oeuvres. This is a neo-Kantian project which both Bakhtin and Foucault share.
Shifting positions on a neo-Kantian base
With these methodological parallels it is no surprise that there are so many areas in which Foucaultʼs and Bakhtinʼs historical works can be compared, but it is almost solely on this basis that comparative work has so far progressed. Foucaultʼs pervasive anti-humanism and subsequent turn from archaeology to genealogy has overshadowed attempts to discern philosophical common ground at a more fundamental level. As Rose shows, Foucaultʼs reversion to genealogy was a defection from one school of neo-Kantian thinking to another, with the Marburgian primacy of validity being replaced by a Nietzschean version of Baden School neo-Kantianism in which the primacy of value in the production of knowledge takes the form of power.  It is signiﬁcant that it is this move that draws the ﬁre of the more Marburgian Habermas, who argues that ʻthe internal aspects of meaning, of truth-validity, and of evaluating do not go without remainder into the externally grasped aspects of practices of powerʼ. 
Unlike the poststructuralists, Bakhtin remained a humanist throughout his career, his philosophy of language being based on an organicist model rather than the intellectualism of Saussure. His idea of dialogue was a linguistic rendering of an ethic of intersubjective relations developed from the Marburg School and the phenomenology of Max Scheler, whose own work was to some extent an attempt to overcome the fragmentation of subjectivity described by Nietzsche. In the Bakhtinian mix there are the developments of neo-Kantianism in the work of Simmel and Cassirer, with the attempt to overcome the dichotomy of life and objective culture through the notion of unfolding symbolic forms. Derridaʼs and Foucaultʼs attachment to Nietzscheʼs fragmentation of the subject and a neoSaussurean philosophy of language is in sharp contrast to Bakhtin, while the French theorists both effectively took the opposite side to Bakhtin in the famous 1929 Davos disputation between Heidegger and Cassirer. Foucault and Derrida diverge on their attitude to phe-nomenology and to Heideggerʼs Ursprungsphilosophie, but their points of agreement are more pervasive. There are nevertheless distinct points of convergence between the French and Russian theorists. Derridaʼs encounter with Levinas on the basis of intersubjective ethics, pioneered by Scheler, led Derrida to move toward a delineation of an ethics of deconstruction which in some respects echoes the ethical charge of Bakhtinʼs dialogism. Similarly the later Foucaultʼs gravitation toward an ethics which reactivates the sollen (ought) at the basis of Baden neo-Kantianism brings him back into a common problematic. Common to all three theorists is a tendency to efface politics and ultimately to replace it with ethics, itself a classic neo-Kantian move. Foucault ultimately transforms the will to power into a vitalistic will to life, revealing his own connection to Lebensphilosophie, taking us back to his formulations in The Order of Things and to the parallels with Bakhtinʼs attempts to mediate the ʻworldsʼ of life and culture through a ʻthird realmʼ. These developments are all explicable with reference to roots in a shared philosophical tradition which in the case of Bakhtin was direct but in that of the French theorists was gained through a tradition of academic philosophy and sociology founded on German neoKantianism.
The neo-Kantian logic of presenting the world as not ʻgivenʼ but ʻset as a taskʼ asserted itself in all three cases with the wholesale replacement of politics by ethics. While Marburg neo-Kantian philosophers were predominantly reformist socialists, their prime aim was to establish a philosophy of regional sciences based on the ʻproductionʼ of objects of knowledge according to mathematical logic. In the realm of the human sciences this meant treating society exclusively as a ʻmoral realityʼ and human beings as ʻjuridical personsʼ, while consigning consideration of the physical structures and biological requirements that impose given parameters on, and act as forces within, social life to the natural sciences. Thus, the socio-economic structures existentially presupposed by a speciﬁc and emergent mode of political rule or cultural formation were neglected in favour of a sui generis logic of culture. Bakhtin, Derrida and Foucault were all, in their own way, compelled by the logic of their own premisses to misrepresent one element of a relationship that is integrated at a ʻmolecularʼ level by hypostasizing that element. The consequences for analysis are seriously debilitating and the political consequences are nothing short of disastrous.
Not only does this return us to a fundamentally idealist philosophy; in the case of poststructuralism it reverts back to the Marburg-type metaphysics of the production of the object through mathematical categories that the thinkers which mediated the Marburg inﬂuence had tried to undermine. Where the intervening generation of sociologists, existentialist philosophers and Western Marxists had sought to outline the relationship between the history of science, philosophy and productive relations, the poststructuralists re-establish originary productive categories. After all, if validity is established independently of an unknowable existence, then a disembodied ethics of object-constitution can leave the world unknown and unchanged. It is easy to see how this ﬁts in with the perceived need of disillusioned post-1968 intellectuals to justify their withdrawal from collective politics. The resulting ʻnew ethicsʼ that now dominates social theory is, predictably, reminiscent of that with which incipient Western Marxism struggled. It also reminds us of the roots of cultural studies in Kulturkritik.  This ʻnew ethicsʼ and Kulturkritik was Bakhtinʼs point of departure, with the presence of Simmel, Cohen and Scheler dominating his early ethical work. However, in Stalinʼs Russia Bakhtin was not able to settle into what Rose calls the ʻslumber of the mathesisʼ,  even through a movement from philosophy to literary studies. The apparent separation of economic life from politics and citizenship that characterizes bourgeois democracy and on which neo-Kantian ethicism rests, could not be maintained in Stalinʼs Russia, even within the realm of literary criticism. Bakhtin and the poststructuralists thus meet on the road going in opposite directions: the former struggles to accommodate the increasingly apparent effects of political economy on culture within a compromised neo-Kantian framework, while the latter respond to the dissipation of socio-economic struggle and consolidation of bourgeois democracy by absorbing culture into a neo-Kantianized realm of validity.
Whatever the virtues of their respective works, which are often considerable, Bakhtin, Derrida and Foucault are seriously compromised by the neo-Kantian philosophy that underlies them. Just as economic life and politics are intimately intertwined and cannot be adequately explained and criticized in isolation, so the ʻrealmsʼ of existence and validity, fact and value, must be related in any adequate social and cultural theory. This requires a theory of reference principally excluded from neo-Kantianism and a politics that strives to overcome the division of social life in practice. Theories based on neo-Kantianism can reﬂect such a division but they can neither explain nor be a guide to changing it.
This article is based on research carried out under the project ʻThe Russian and European Contexts of the works of Mikhail Bakhtin and the Bakhtin Circleʼ, funded by the Arts and Humanities Research Board of the British Academy.
1. ^ Quoted in Frédéric Vandenberghe, Comparing Neo-Kantians: Ernst Cassirer and Georg Simmel, University of Manchester Department of Sociology, Occasional Paper No. 49, 1996, p. 2.
2. ^ I have in mind the work of, among others, Natalia Bonetskaia, Askol´d Muratov, Brian Poole and Galin Tihanov, I am indebted to Poole and Tihanov for comments on an earlier version of this article.
3. ^ Notable exceptions include Steven Collins, ʻCategories,
Concepts or Predicaments? Remarks on Maussʼs Use of Philosophical Terminologyʼ, in Michael Currithers et al., The Category of the Person: Anthropology, Philosophy, History, Cambridge University Press, New York, 1985; and Terry F. Godlove, ʻIs Space a Concept? Kant,
Durkheim and French Neo-Kantianismʼ, Journal of the History of the Behavioral Sciences, vol. 32, no. 4, 1996, pp. 441–55.
4. ^ Hegel Contra Sociology, Athlone, London, 1981; Dialectic of Nihilism: Post-Structuralism and Law, Blackwell, Oxford, 1984.
5. ^ Christopher Norris anticipated ʻwidespread discussion among critics, philosophers and … theorists of a poststructuralist persuasionʼ arising from Roseʼs 1984 book on poststructuralism and law, but this does not appear to have materialized. Christopher Norris, Deconstruction and the Interests of Theory, Pinter Publishers, London, 1988, p. 245.
6. ^ The neo-Hegelian structure of Cassirerʼs central work was ﬁrst advanced by Donald Verene in ʻKant, Hegel and Symbol: the Origins of the Philosophy of Symbolic Formsʼ (Journal of the History of Ideas 30, 1969, pp. 33–46), and was elaborated in J.M. Krois, Cassirer: Symbolic Forms and History, Yale University Press,
New Haven and London, 1987.
7. ^ V.D. Duvakin, Besedy s M.M. Bakhtinym, Progress Publishers, Moscow, 1996, pp. 39–41.
8. ^ Roy Bhaskar, Plato etc., Verso, London, 1994, pp. 199–200.
9. ^ Michel Foucault, The Birth of the Clinic: An Archaeology of Medical Perception, trans. A.M. Sheridan Smith, Tavistock, London, 1973, pp. 170–73. Compare Bakhtinʼs account of medicine, largely borrowed from Cassirer without reference, in Rabelais and His World, trans. H. Iswolsky, Indiana University Press, Bloomington, 1984, pp. 359 ff. On Bakhtinʼs unattributed borrowing from Cassirer here, see Brian Poole, ʻBakhtin and Cassirer: The Philosophical Origins of Bakhtinʼs Carnival Messianismʼ, South Atlantic Quarterly, vol. 97, no. 3/4, 1998, pp. 537–78.
10. ^ ʻRoman vospitaniia i ego znachenie v istorii realizmaʼ, Estetika slovesnogo tvorchestva, Iskusstvo, Moscow, 1979, pp. 118–236, 224; ʻThe Bildungsroman and Its Signiﬁcance in the History of Realismʼ, Speech Genres and Other Late Essays, trans. Vern W. McGee, University of Texas Press, Austin, 1986, pp. 10–59, 43.
11. ^ Quoted in Jürgen Habermas, ʻThe German Idealism of the Jewish Philosophersʼ, Philosophical–Political Proﬁles, trans. F.G. Lawrence, Heinemann, London, 1983, pp. 21–43,
26. ^ See also Andrea Poma, The Critical Philosophy of Hermann Cohen, trans. J. Denton,
State University of New York Press, New York, 1997, p. 89.
12. ^ Paul Natorp, ʻOn the Objective and Subjective Grounding of Knowledgeʼ, trans. D. Kolb, Journal of the British Society for Phenomenology, vol. 12, no. 3, 1981, pp. 252–3. The most sustained examination of Husserlʼs debt to Natorp is Iso Kern, Husserl und Kant: Eine Untersuchung über Husserls Verhältnis zu Kant und zum Neukantianismus, Martinus Nijhoff: The Hague, 1964, pp. 321 ff. Derrida elides the continuities between Husserl, Heidegger and neo-Kantianism in his essay on Cohenʼs ʻDeutschtum und Judentumʼ, ʻKant, the Jew, the Germanʼ, New Literary History, vol. 22, no. 1, 1991, p. 41. Derrida here shows little familiarity with Cohenʼs earlier, more inﬂuential works.
13. ^ K.S. Bakradze, Ocherk po istorii noveishei sovremennoi burzhoiznoi ﬁlosoﬁi, Sabchota sakartvelo, Tblisi, 1960, p. 251.
14. ^ Heidegger succeeded Cohen in his chair in philosophy at Marburg. Commenting on the famous 1929 Davos disputation between Heidegger and Cassirer, Franz Rosenzweig argued that the former furthered the spirit of Cohenʼs philosophy more than did Cassirer. On this, see Gillian Rose, Judaism and Modernity, Blackwell, Oxford, 1993, p. 112; and Peter Eli Gordon, ʻRosenzweig and Heidegger: Translation, Ontology and the Anxiety of Afﬁliationʼ, New German Critique 77, 1999, pp. 114–16.
Rose also notes how, ʻlike Cohen, Heidegger begins by expounding time as the productive unity and difference internal to Kantʼs transcendental exposition of experienceʼ (Dialectic of Nihilism, p. 70).
15. ^ Vincent Descombes, Modern French Philosophy, trans.
L. Scott-Fox and J.M. Harding, Cambridge University Press, Cambridge, 1980, p. 19.
16. ^ On this, see Michael Dummett, ʻObjectivity and Reality: Lotze and Fregeʼ, in Frege and Other Philosophers, Clarendon Press, Oxford, 1991, pp. 97–125.
17. ^ Frank Lentricchia, After the New Criticism, Methuen,
London, 1983, p. 171. On this aspect of Brunschvicgʼs work see S.I.M. Du Plessis, The Compatibility of Science and Philosophy in France 1840–1940, Balkema, Cape Town, 1972, pp. 45–6 and 192–208.
18. ^ Compare Ernst Cassirer, The Philosophy of Symbolic Forms, vol. 1: Language, trans. R. Mannheim, Yale University Press, New Haven and London, 1955, pp. 105 ff.
19. ^ Rose, Dialectic of Nihilism, pp. 113–14.
20. ^ Ibid., 151.
21. ^ Descombes, Modern French Philosophy, pp. 144, 148.
22. ^ Jürgen Habermas, The Philosophical Discourse of Modernity, trans. F. Lawrence, Polity Press: Cambridge, 1987, pp. 167, 182.
23. ^ ʻLichnoe delo V.N. Voloshinovaʼ, Dialog Karnaval Khronotop 2, 1995, pp. 87–8.
24. ^ M.M. Bakhtin, Raboty 1920-kh godov, Next, Kiev, 1994, p. 12; Toward a Philosophy of the Act, trans. V. Liapunov, University of Texas Press, Austin, 1993, p. 2.
25. ^ ʻ“Geist” and “Life”ʼ, in Ernst Cassirer, The Philosophy of Symbolic Forms, vol. 4: The Metaphysics of Symbolic Forms, trans. J.M. Krois, Yale University Press, New Haven and London, 1996, pp. 3–33, 16–17.
26. ^ M.M. Bakhtin, ʻ1961 god. zametkiʼ, Sobranie sochinenii V, Russkie slovari, Moscow, 1996, pp. 329–60, 330–31.
27. ^ See M.M. Bakhtin, ʻSlovo v romaneʼ, in Voprosy lit-eratury i estetiki, Khudozhestvennaia literatura, Moscow, 1975, pp. 72–233, 200; ʻDiscourse in the Novelʼ, The Dialogic Imagination, trans. C. Emerson and M.
Holquist, University of Texas Press, Austin, 1981, pp. 259–422, 388. The jurisprudential heritage is especially clear in the work of the Circleʼs other literary theorist Lev Pumpianskii: ʻIn the heroic novel there is an uninterrupted passing of judgement (sud) on the character. … The method [of the heroic novel] is the trial of the powers of contending sides in a well-considered social court through the construction of the walk of life, scenes of a life, accompanied by the uninterrupted interpretation of the authorʼ (L.V. Pumpianskii ʻRomany Turgeneva i roman “Nakanune”: istoriko-literaturnyi ocherkʼ, in I.S.
Tirgenev, Sochineniia VI, Moskva–Leningrad, 1929, pp. 9–26, 9, 11).
28. ^ Bakhtin, ʻ1961 god. zametkiʼ, pp. 333, 337. Note also Voloshinovʼs afﬁrmative comments on the ʻ“juridical” theory of tragedyʼ developed in Cohenʼs Ästhetik des Reinen Gefühls in ʻSlovo v zhizni i slovo v poeziiʼ, in V.N. Voloshinov, Filosoﬁia i sotsiologiia gumanitarnykh nauk, Asta, St. Petersburg, 1995, pp. 59–86, 82; English translation, ʻDiscourse in Life and Discourse in Poetryʼ, in Ann Shukman, ed., Bakhtin School Papers, Russian Poetics in Translation 10, 1983, pp. 5–29, 25.
29. ^ Hermann Cohen, Ethik des reinen Willens, Georg Olms Verlag, Hildesheim and New York, 1981, p. 309.
30. ^ On Stalinʼs Marxism and Questions of Linguistics, see V.M. Alpatov, ʻWhat is Marxism in Linguisticsʼ, in Craig Brandist and Galin Tihanov, eds, Materializing Bakhtin: The Bakhtin Circle and Social Theory, Macmillan: London, 1999, pp. 173–93.
31. ^ Sobranie, p. 251.
32. ^ Matvei Kagan, ʻGerman Kogen (4 iiulia 1842–4 aprelia 1918)ʼ, Nauchnie izvestiia sbornik vtoroi, R.S.F.S.R.
Akademicheskii tsentr narkomprosa, Moscow, 1922, pp. 110–124, 114.
33. ^ Quoted in Collins, ʻCategories, Concepts or Predicaments?ʼ, p. 61.
34. ^ Of Grammatology, trans. G. Spivak, Johns Hopkins University Press, Baltimore, 1976, p. 27.
35. ^ Rose, Dialectic of Nihilism, p. 150. Husserlʼs 1936 essay is reproduced in Thomas Luckmann, ed., Phenomenology and Sociology, Penguin, Harmondsworth, 1978, pp. 42–70. Derridaʼs ﬁrst major published work was his LʼOrigine de la géometrie de Husserl, Traduction et Introduction, Presses Universitaires de France, Paris, 1962; English translation Edmund Husserlʼs Origin of Geometry: An Introduction, trans. J.P. Leavey, Hays,
36. ^ Derrida, Of Grammatology, p. 27.
37. ^ Ibid., pp. 60–61.
38. ^ Sobranie, p. 251.
39. ^ Derrida, Of Grammatology, 62.
40. ^ Habermas, The Philosophical Discourse of Modernity, pp. 165–6.
41. ^ Quoted in Lentricchia, After the New Criticism, p. 188.
42. ^ ʻSlovo v romaneʼ, p. 92; ʻDiscourse in the Novelʼ, p. 278.
43. ^ Michel Foucault, The Order of Things: An Archaeology of the Human Sciences, Tavistock, London, 1970, pp. xx–xxi.
44. ^ Voloshinov, Filosoﬁia i sotsiologiia gumanitarnykh nauk, pp. 308–10; Marxism and the Philosophy of Language, trans. L. Matejka and I.R. Titunik, Harvard University Press, Cambridge MA, pp. 91–3. On this, see Galin Tihanov, ʻVoloshinov, Ideology and Language: The Birth of Marxist Sociology from the Spirit of Lebensphilosophieʼ, South Atlantic Quarterly, vol. 97, no. 3/4, 1998, pp. 599–621.
45. ^ M.M. Bakhtin, ʻProblema rechevykh zhanrovʼ, in Sobranie, pp. 159–206, 161; ʻThe Problem of Speech Genresʼ, in Speech Genres and Other Late Essays, pp. 60–102, 62.
46. ^ On this, see especially Bakhtinʼs ʻEpos i romanʼ, Voprosy, pp. 447–83, 466; ʻEpic and Novelʼ, The Dialogic Imagination, pp. 3–40,
47. ^ Rose, Dialectic of Nihilism, p. 185.
48. ^ Foucault, The Order of Things, pp. xxi–xxii.
49. ^ Though the term was allegedly taken from biology, one of the main sources of this idea was Cassirerʼs neoKantian writing on ʻintuitive expressionʼ in language in volume 1 of The Philosophy of Symbolic Forms, pp. 198 ff., and ʻmyth as a form of intuitionʼ and ʻlife formʼ in volume 2 on Mythical Thought (trans. R. Manheim, Yale University Press, New Haven and London, 1955).
50. ^ Rose, Dialectic of Nihilism, pp. 187 ff.
51. ^ Habermas, The Philosophical Discourse of Modernity, p. 276.
52. ^ See Francis Mulhern, ʻThe Politics of Cultural Studiesʼ, Monthly Review, vol. 47, no. 5, 1995, pp. 31–40.