Beatrice Han-Pile on Early Foucault
On Thursday the 27th of October the Transcendental Philosophy and Naturalism project hosted its second talk. Béatrice Han-Pile (Essex) gave a fascinating and thought-provoking paper entitled ‘Early Foucault and Transcendental History’. This paper was written as part of an ongoing debate between her and Gary Gutting over whether the early Foucault’s work should be seen as primarily historical or philosophical. In the paper, Béatrice cast the debate in terms of whether we should view the early Foucault as a transcendental philosopher interested in delineating what she called a “…‘transcendental history’ of the conditions of possibility of knowledge in the West,” or whether we should simply view him as an innovative historian of ideas. She argued that we must see Foucault as the former rather than the latter, and thus as primarily a philosopher. Béatrice made clear that the paper was a work in progress, and that she was merely presenting a hypothesis about how we might fruitfully understand Foucault’s early work.
The strategy of the paper was to make a case for seeing Foucault as holding (some kind of) transcendental understanding of history by examining what he himself says about the modern ‘historical a priori’ of ‘History’. The ‘historical a priori’ (or ‘épistémè’) is Foucault’s term for the underlying non-empirical framework of intelligibility that makes possible our knowledge of ourselves and objects within a given epoch. Presumably, Béatrice’s point was that Foucault’s interest in ‘History’ is motivated by a desire to uncover the transcendental conditions of possibility of the modern era in the West, rather than by a desire simply to describe the empirical formation of various types of knowledge (which seems to be Gutting’s view of what the early Foucault is doing).
Béatrice pointed out that there is a potentially fatal confusion in Foucault’s understanding of the role of the historical a priori of History. For he states both that History is an epistemic framework through which we understand entities in the world, and an ontological property of such entities. It seems prima facie a non sequitur to infer from the fact that we conceive entities according to certain non-empirical constraints, that entities themselves are ontologically constituted according to such constraints. Béatrice suggested that we avoid attributing such a non sequitur to Foucault by articulating his understanding of the historical a priori of History in terms analogous to Kant’s. According to Béatrice:
“…what Foucault means by saying that History is both the condition under which entities are known and their mode of being, then, is that History has become the very framework (in his terms, the historical a priori) under which they are constituted as entities. It does not mean that their historicity is an arbitrary construction of the mind, as Foucault makes it clear that we do not choose our épistémè; on the contrary, we are bound to construe beings as historical by the fact that we are governed by History, and this until a new épistémè arises (such as the “Return of language”).”
However, Béatrice qualified Foucault’s understanding of transcendental history vis-à-vis Kant’s transcendental turn with the following three caveats: (1) transcendental history is not premised on an ontological claim that there is a way the world is ‘in itself’, (2) neither is it premised on anything like Kant’s transcendental psychology, nor (3) does the historical a priori refer to a universal set of non-empirical constraints on experience and knowledge.
According to Béatrice, the problem over the conflation of the epistemological and ontological orders in Foucault’s early work can be solved by adverting to the understanding of the historical a priori of History that he explicates in the Order of Things. History, as a historical a priori, is an ‘origin’ which precedes the chronological ordering of events it makes possible. As Béatrice put it, History “…logically pre-exists the time of succession and does not belong to it.” The idea seems to be that the possible range of understandings of empirical objects available to us in the modern era is constituted by the historical a priori of History. So in uncritical mode, we think of ourselves as embedded in economic, biological and socio-cultural processes over which we have no control. But according to Béatrice’s reading of Foucault, at the critical level we come to see that the very idea of our being so embedded, is not simply the result of brute contingencies, but rather a transcendental understanding of the possible limits of our world. To put the point another way: at the uncritical level, that empirical objects are formed by ineluctable historical processes is taken to be an empirically discoverable fact. But at the critical level, this ‘fact’ is exposed as constituted by an épistémè ― the historical a priori of History.
So the épistémès that Foucault’s early work describes are to be considered, Béatrice suggested, as non-empirical constraints on what forms practices can take in a given epoch. The épistémès that transcendental history describes would not then be reducible to contingent historical practices. If this idea of épistémès as non-empirical constraints is coherent, then Béatrice is certainly right (contra Gutting) to view the early Foucault’s work as transcendental history.
However, Béatrice pointed out that there are three problems with this understanding of épistémès as non-empirical constraints on empirical practices.
(1) As épistémès follow one another successively, then surely Foucault must be committed to some kind of universal transcendental order of which the épistémès count as modifications. Béatrice suggested that Foucault had a conception of a ‘naked experience of order’ which could play the substratum role his transcendental history seems to necessitate. However, this conception remains somewhat obscure (which is something Béatrice herself pointed out).
(2) Even if one accepts that there is a mysterious ‘experience of order’ of which épistémès count as modifications, there is a question over how épistémès change. If such change is driven by socio-political conditions (which it would seem to have to be, given Foucault’s general anti-metaphysical stance), then épistémès are subject to empirical causality. If the latter is the case, then it is difficult to view Foucault as describing non-empirical constraints on the general form empirical practices can take within an epoch.
Despite these difficulties, Béatrice suggested we see Foucault’s early work as implicitly committed to there being universal transcendental constraints that act as a ‘meta-historical a a priori’ (that which I referred to earlier as playing some kind of non-metaphysical substratum role). The latter makes possible the succession of épistémès. But she also suggested that the early Foucault rejects the idea that universal transcendental constraints could ever be adequately described, due to his insistence that such constraints would always only be describable from within the purview of a particular épistémè.
According to Béatrice then, Foucualt’s conception of transcendental history has two critical upshots. First, it reveals a purely empirical approach to the analysis of the historical to be naive, because the role of the historical a priori (that is, of épistémès) in determining practices cannot be accounted for by such an approach. Second, although transcendental history presupposes a non-metaphysical ‘substratum’ of order (a meta-historical a priori), Foucault would deny that the latter could ever be adequately described, due to the fact that any description would unfold from within the purview of a specific épistémè.
So what we seem to get on Béatrice’s reading of the early Foucault is;
(1) An insistence that épistémès are historical a prioris that non-empirically constrain practices.
(2) An implicit commitment to a substratum of order that is modified by the various épistémès, yet which it is impossible to describe in a perspicuous manner.
Béatrice let it be known that the idea of épistémès presupposing some kind of substratum of order that remains the same throughout ‘epistemic shifts’ in the historical a priori, was a hypothesis aimed at saving Foucault’s understanding of transcendental history from incoherence. She emphasised that it is not entirely clear what Foucault himself thought on these matters.
All this raises several questions, some of which came up at the talk:
(1) Just because it is a conceptual truth that alteration requires something alterable, why not simply see the ‘substratum’ which transcendental history seems to logically presuppose, as constituted by brutely contingent physical processes which get interpreted by us in differing ways? If we do see the ‘base-superstructure’ relationship in the latter manner, Foucault seems to be as deflationary about transcendental philosophy as Gutting says he is.
(2) Why should we even think there is a ‘substratum’ of order that the various historical épistémès modify, when we can view that which is altered by new épistémès as nothing more than the web of practices that we inherit from the past?
(3) On Béatrice’s account, transcendental history shows that a purely empirical understanding of history will not do, because such an understanding cannot account for the existence of either the historical a priori, or the ‘meta-historical a priori’ of order the latter presupposes. But to anyone not already convinced of the need for transcendental history this will seem circular: i.e. a purely empirical approach to understanding history cannot be ruled out on the grounds that it cannot account for presuppositions necessary for a transcendental understanding of history, because the former understanding denies the intelligibility of the latter. What is needed here is an independent argument for the necessity of historical a prioris and the concomitant commitment to some kind of substratum of order.
(4) The problem over épistémès being both non-empirical yet changes in them being driven by empirical causality seems serious. Without addressing this problem it is hard to consider Foucault’s understanding of transcendental history as coherent.
The strategy of the paper was to make a case for seeing Foucault as holding (some kind of) transcendental understanding of history by examining what he himself says about the modern ‘historical a priori’ of ‘History’. The ‘historical a priori’ (or ‘épistémè’) is Foucault’s term for the underlying non-empirical framework of intelligibility that makes possible our knowledge of ourselves and objects within a given epoch. Presumably, Béatrice’s point was that Foucault’s interest in ‘History’ is motivated by a desire to uncover the transcendental conditions of possibility of the modern era in the West, rather than by a desire simply to describe the empirical formation of various types of knowledge (which seems to be Gutting’s view of what the early Foucault is doing).
Béatrice pointed out that there is a potentially fatal confusion in Foucault’s understanding of the role of the historical a priori of History. For he states both that History is an epistemic framework through which we understand entities in the world, and an ontological property of such entities. It seems prima facie a non sequitur to infer from the fact that we conceive entities according to certain non-empirical constraints, that entities themselves are ontologically constituted according to such constraints. Béatrice suggested that we avoid attributing such a non sequitur to Foucault by articulating his understanding of the historical a priori of History in terms analogous to Kant’s. According to Béatrice:
“…what Foucault means by saying that History is both the condition under which entities are known and their mode of being, then, is that History has become the very framework (in his terms, the historical a priori) under which they are constituted as entities. It does not mean that their historicity is an arbitrary construction of the mind, as Foucault makes it clear that we do not choose our épistémè; on the contrary, we are bound to construe beings as historical by the fact that we are governed by History, and this until a new épistémè arises (such as the “Return of language”).”
However, Béatrice qualified Foucault’s understanding of transcendental history vis-à-vis Kant’s transcendental turn with the following three caveats: (1) transcendental history is not premised on an ontological claim that there is a way the world is ‘in itself’, (2) neither is it premised on anything like Kant’s transcendental psychology, nor (3) does the historical a priori refer to a universal set of non-empirical constraints on experience and knowledge.
According to Béatrice, the problem over the conflation of the epistemological and ontological orders in Foucault’s early work can be solved by adverting to the understanding of the historical a priori of History that he explicates in the Order of Things. History, as a historical a priori, is an ‘origin’ which precedes the chronological ordering of events it makes possible. As Béatrice put it, History “…logically pre-exists the time of succession and does not belong to it.” The idea seems to be that the possible range of understandings of empirical objects available to us in the modern era is constituted by the historical a priori of History. So in uncritical mode, we think of ourselves as embedded in economic, biological and socio-cultural processes over which we have no control. But according to Béatrice’s reading of Foucault, at the critical level we come to see that the very idea of our being so embedded, is not simply the result of brute contingencies, but rather a transcendental understanding of the possible limits of our world. To put the point another way: at the uncritical level, that empirical objects are formed by ineluctable historical processes is taken to be an empirically discoverable fact. But at the critical level, this ‘fact’ is exposed as constituted by an épistémè ― the historical a priori of History.
So the épistémès that Foucault’s early work describes are to be considered, Béatrice suggested, as non-empirical constraints on what forms practices can take in a given epoch. The épistémès that transcendental history describes would not then be reducible to contingent historical practices. If this idea of épistémès as non-empirical constraints is coherent, then Béatrice is certainly right (contra Gutting) to view the early Foucault’s work as transcendental history.
However, Béatrice pointed out that there are three problems with this understanding of épistémès as non-empirical constraints on empirical practices.
(1) As épistémès follow one another successively, then surely Foucault must be committed to some kind of universal transcendental order of which the épistémès count as modifications. Béatrice suggested that Foucault had a conception of a ‘naked experience of order’ which could play the substratum role his transcendental history seems to necessitate. However, this conception remains somewhat obscure (which is something Béatrice herself pointed out).
(2) Even if one accepts that there is a mysterious ‘experience of order’ of which épistémès count as modifications, there is a question over how épistémès change. If such change is driven by socio-political conditions (which it would seem to have to be, given Foucault’s general anti-metaphysical stance), then épistémès are subject to empirical causality. If the latter is the case, then it is difficult to view Foucault as describing non-empirical constraints on the general form empirical practices can take within an epoch.
Despite these difficulties, Béatrice suggested we see Foucault’s early work as implicitly committed to there being universal transcendental constraints that act as a ‘meta-historical a a priori’ (that which I referred to earlier as playing some kind of non-metaphysical substratum role). The latter makes possible the succession of épistémès. But she also suggested that the early Foucault rejects the idea that universal transcendental constraints could ever be adequately described, due to his insistence that such constraints would always only be describable from within the purview of a particular épistémè.
According to Béatrice then, Foucualt’s conception of transcendental history has two critical upshots. First, it reveals a purely empirical approach to the analysis of the historical to be naive, because the role of the historical a priori (that is, of épistémès) in determining practices cannot be accounted for by such an approach. Second, although transcendental history presupposes a non-metaphysical ‘substratum’ of order (a meta-historical a priori), Foucault would deny that the latter could ever be adequately described, due to the fact that any description would unfold from within the purview of a specific épistémè.
So what we seem to get on Béatrice’s reading of the early Foucault is;
(1) An insistence that épistémès are historical a prioris that non-empirically constrain practices.
(2) An implicit commitment to a substratum of order that is modified by the various épistémès, yet which it is impossible to describe in a perspicuous manner.
Béatrice let it be known that the idea of épistémès presupposing some kind of substratum of order that remains the same throughout ‘epistemic shifts’ in the historical a priori, was a hypothesis aimed at saving Foucault’s understanding of transcendental history from incoherence. She emphasised that it is not entirely clear what Foucault himself thought on these matters.
All this raises several questions, some of which came up at the talk:
(1) Just because it is a conceptual truth that alteration requires something alterable, why not simply see the ‘substratum’ which transcendental history seems to logically presuppose, as constituted by brutely contingent physical processes which get interpreted by us in differing ways? If we do see the ‘base-superstructure’ relationship in the latter manner, Foucault seems to be as deflationary about transcendental philosophy as Gutting says he is.
(2) Why should we even think there is a ‘substratum’ of order that the various historical épistémès modify, when we can view that which is altered by new épistémès as nothing more than the web of practices that we inherit from the past?
(3) On Béatrice’s account, transcendental history shows that a purely empirical understanding of history will not do, because such an understanding cannot account for the existence of either the historical a priori, or the ‘meta-historical a priori’ of order the latter presupposes. But to anyone not already convinced of the need for transcendental history this will seem circular: i.e. a purely empirical approach to understanding history cannot be ruled out on the grounds that it cannot account for presuppositions necessary for a transcendental understanding of history, because the former understanding denies the intelligibility of the latter. What is needed here is an independent argument for the necessity of historical a prioris and the concomitant commitment to some kind of substratum of order.
(4) The problem over épistémès being both non-empirical yet changes in them being driven by empirical causality seems serious. Without addressing this problem it is hard to consider Foucault’s understanding of transcendental history as coherent.
posted by matt at 9:27 am
12 Comments:
Hi Matthew,
Many thanks for your report and for your questions. (1) and (2) are variations of the same problem, so I’ll answer them together. Re: (2), I don’t think that from an archaeological perspective it is right to frame the question as you have done, and to suggest that épistémès “modify” or “alter” each other (or the meta historical a priori): when Foucault speaks of modifications of the épistémès, the genitive is objective: he does not attribute to them any form of causal efficacy (this would be a gross mistake from his part). Still, I see what you mean but even so I think that Foucault would reject this idea because it would be tantamount to denying that there is a specific level at which épistémès perform their determination and must be analysed. Your notion of a “web of practices” would need clarification, but it does suggest that in the end épistémès would be homogeneous to practices. If this was the case, then it would be difficult to make sense of Foucault’s resistance to the Marxist idea of ultimate determination by the infra-structure, and of all his efforts in OT and AK to identify the specific level at which the historical a priori should be studied (I’m using “historical a priori” and “épistémè” as if they were synonymous, which is not completely the case, but here the difference is irrelevant to the question you ask). The most one could say is that while changes in practices may cause epistemic change, it still remains that the latter is not reducible to changes in practices. In short, your point could only be conceded if one stepped outside of the context of Foucault’s thought, and conceding it would be tantamount to giving up from the start on the strong notion of an historical a priori which he is attached to in the early work. It would also mean giving up on the idea of transcendental determination (as all that would remain would be causal determination, practices acting on practices). Whether there would be good reasons to do this refers to your question (3), but certainly from the exegetic point of view which was mine when I introduced the hypothesis, your suggestion would have to be rejected.
Re: (1), one preliminary remark: on my reading Foucault IS deflationary about transcendental philosophy, although not as much as Gutting would have it. Gutting thinks that he rejects ALL forms of transcendentalism; in my view, he wants to play the notion of the historical a priori against the possibility of identifying transcendental constraints, although I have tried to show that this very notion still commits him to acknowledge the existence of the latter, something which he may or may not have liked to hear. Now, re: the substance of your question, I think that the answer must be “no” because for Foucault there is no “brute contingent physical processes” which would exist first and then be interpreted in different ways, or which could be thought of independently of the way in which they appear to us. This is the point of the Allisonian position which I have suggested makes the best sense of his claims about the historical a priori being both a space for knowledge and a mode of being of entities. Foucault would not deny that there are physical processes, but he would see them as an empirical realist, and thus would reject the idea that the “base” you mention is detachable from epistemic frameworks, so the possibility you suggest could not arise.
Re: (3), I did wonder about this myself, although I didn't incorporate it in the paper because it was more than long enough already (!). I have given an argument as to why we need the meta- historical a priori, but it is premised on the existence of the épistémès. F's starting point is indeed that there are epistemic constraints. However in his view it is not an arbitrary starting point. In fact, I think that the implicit structure of his thoughts on this is that of a typical transcendental argument: he starts from a study of empirical discourses and goes back to what they require so as to have been possible at all. His answer is that it is the historical a priori they belong to, which the archaeology tries to describe. Perhaps a more specific argument about this can be constructed from his remarks about order at the beginning of the Preface, I'll think about it and will get back to you if I have one. However your question opens up to the more general problem of why any form of transcendental account is needed at all, and of whether naturalism or naïve realism could be satisfactory, which is rather central to the TP&N project, so if anybody wants to join in, please feel welcome!
Re: (4), this is the real problem (in my view). It’s the main reason why in my book I had to conclude that Foucault’s transcendental history fails, which is rather ironical given that it is precisely the criticism he addresses to the analytic of finitude. I didn’t have the time to go into this in the paper, but in OT Foucault argues that Man as an empirico-transcendental double is an impossible structure because the distinction between the empirical and the transcendental cannot be maintained, and the transcendental is ultimately submitted to empirical determinations (this is largely grounded in his criticism of Kant’s transcendental psychology). I have suggested in the paper that it does not follow from this rejection that he wants to reject any form of transcendentalism, and that in fact he does need the idea of transcendental constraints. But while this may solve the problem of the possibility of epistemic change, it does not address the question of its modality, and of a possible conflation between transcendental and causal determination which would ruin the possiblity of the former. I have argued that the genealogical work was designed to work on this issue (hence the well-known shift to the articulation between discursive practices and non discursive practices) but I’m not sure it solves it. The idea of a power/knowledge nexus may merely make the problem more complicated and more interesting, but nothing more. I’m still working on this and trying to prove my earlier interpretation wrong.
I hope that this helps.
Beatrice
Forgive some speculations and questions from one whose knowledge of Foucault is less than slender...
I'm having some difficulty is seeing exactly what the problem is re: 4. It seems that on Foucault's view, whatever épistémès are it is certainly not a necessary truth that we have one rather than another. For we have different épistémès at different times (and in different places? I notice the use of "...in the West" cropping up). I can see that thinking of épistémès as contingent pushes us towards asking for an explanation of why we have one rather than another, since (arguably) contingent facts require explanations. But cannot Foucault maintain that the objection (4) is mistaken in the following way: It's not as if an empirical thing causes an épistémè-thing to come about, for épistémès are not things. What is meant by calling épistémès non-empirical? Does it mean, “not known by empirical means”, or “not an entity in the empirical world”? Take the former: I see no contradiction between saying that something is contingent and subject to (empirical) causality, yet our knowledge of it is non-empirical (a priori). Take the latter: we must distinguish between the (non-empirical) épistémè itself and the (entirely empirical) fact that some particular individuals operate according to that épistémè. This empirical fact can be subject to empirical causality, but that does not involve some weird causal relation between empirical states of affairs and non-empirical épistémès.
I guess another way of putting my confusion is this: Beatrice says that Foucault, "starts from a study of empirical discourses and goes back to what they require so as to have been possible at all". But why is it surprising or problematic that different empirical discourses, in different epochs, will require different things (épistémès) to account for their possibility?
Hi Joel,
Thank you very much for your comment. It would be very nice if you were right. Let’s see: to call épistémès “non empirical” does mean “not an entity in the empirical world” indeed. For Foucault they are conditions of possibility of entities, not entities. It is also true that one must distinguish between épistémès and the empirical fact that individuals operate according to them. And, finally, you’re right that there is nothing surprising in the fact that different practices in different epochs (and places: in OT Foucault “only” makes claims about the West indeed) will point to different épistémès. But the problem is not about the difference in épistémès. As long as one considers each épistémè in isolation, everything you say works. The trouble occurs when one tries to think the change from, say, épistémè A to B. A is necessary to practices A’ (in the sense that without A, the practices would not be possible), and B to practices B’. Practices change (empirically). The épistémès change accordingly. The question is whether one can maintain the two levels, empirical and transcendental, separate in thinking this change.
I can see two possibilities:
(a) One is to say (which is the most intuitive view) that the change in practices BROUGHT ABOUT the epistemic change, in the causal sense that without the former the latter would not have happened. If this is the case, then the problem referred to in (4) does obtain, for it would mean that the change in epistemic determination is an effect of empirical changes (albeit an indirect effect, since no one can set about to change their own épistémè; they can’t even know it). If this was the case, then it would mean that the two levels cannot be thought of as separate.
(b) The other, less intuitive possibility would be to claim that changes in practices and epistemic change happen IN PARALLEL, or in other words that epistemic changes are simply the mirroring, at another level, of the changes in practices (as different practices require different conditions of possibility). All that is causally brought about is a change in practices from A’ to B’, and there would be no causal link between the latter and the change from A to B. This view would be in keeping with your suggestion, I think.
Now I am becoming rather excited as it looks like (b) could indeed fit very well Foucault's position during the archaeological period! The reason for this is that during that time, Foucault is very keen to insist on the possibility of isolating discursive practices from non discursive practices. He holds that one can bracket (the parallel with phenomenology is intentional) the question of the empirical genesis of (and changes in) discourses and study their conditions of possibility in isolation. So I think that he would be quite happy with the idea that epistemic change is just the parallel reflection of empirical change, and this would indeed show that the problem identified in (4) needs not arise. As it was the most serious one, this would be good news!
I need to think more about this, I wanted to write a quick reply to keep in with the spirit of the blog, which (I assume) is that of a fairly spontaneous, almost conversational exchange. The obvious questions are whether (b) is a plausible view (beyond its formal coherence and suitability to Foucault’s position) and also whether it would still hold for later Foucault, after the turn to genealogy.
It seems to me that question 4) is best understood as a specific way of understanding question 3). This question is one that all transcendental arguments face: What reason do we have for accepting the idea that A necessarily presupposes B? Setting aside the general problem of how we can know modal truths without presuming that conceivability exhausts possibility, the question ask us to justify why we suppose that A is not possible without B. So an anti-transcendentalist like Gutting will want to know why empirical practices aren’t possible without the prior existence of a ‘historical a priori’. For surely we can see the ground of empirical practices as nothing more than the practices themselves. Anyway, there is certainly no inconsistency in taking this line, so why, in the spirit of Ockham’s razor, need we multiply theoretical posits?
It is to the issue of why we should even countenance the idea of ‘historical a prioris’ as transcendental conditions that question 4) is intimately connected: if such ‘epistemic frameworks’ are simply caused by the woof and wharf of empirical practices then it is hard to see them as conditions of the possibility of the latter. It would seem more natural to view them as constellations of concepts and principles that reflect the ‘family resemblances’ shared by the salient features of the clusters of empirical practices that are dominant at a given time. Under the rubric of such a view, ‘epistemic frameworks’ would be non-empirical abstract objects that have empirical causes. But if we do view epistemic frameworks in such a way, it becomes incoherent to think of them as somehow preceding empirical practices in the sense of making them possible. This incoherence, I take it, is what attaches to Foucault’s holding that epistemic frameworks are caused by empirical practices, whilst at the same time insisting that they act as the conditions of possibility of such practices.
Perhaps after certain clusters of practices are maintained for a certain amount of time we might want to say that the ‘epistemic framework’ that emerges from them, makes possible new practices, and so acts as their ‘condition of possibility’. For example, coming to understand how certain general features of language-use are governed by syntactic rules has made possible the practice of studying grammar (which is the study of a non-empirical constellation of concepts and principles, which arguably, is caused empirically). But it would be putting the cart before the horse to say that the study of grammar made possible language use. Of course, grammar can be studied in isolation from its empirical origin, just as Foucault suggests that epistemic frameworks can be studied in isolation from their empirical origins. But I am confused about why the fact that we can isolate a phenomenon in order to study it bears on the issue of that phenomenon being a transcendental condition of some other phenomenon.
I should like to take up Béatrice’s invitation in the third paragraph of her first comment/post. So what I have to say bears on the need for any form of transcendental philosophy. Yet my point of departure is the idea labelled ‘b’ near the end of Béatrice’s last (second) comment.
I should add that I wrote what follows before seeing Matt's latest post. Having seen that post now, it seems to me that my comment, while somewhat different from his, is in a comparable spirit.
Here's the comment.
The aforementioned idea ‘b’ is the following possibility for making Foucault’s account of epistemes plausible. We take Foucault to hold that changes in epistemes are, in Béatrice’s words, ‘simply the mirroring, at another level, of the changes in practices’. Call this ‘the Reflection Idea’ or ‘RI’ for short. Béatrice says RI is counter-intuitive (or anyway less intuitive than possibility ‘a’). It seems to me that RI may usefully be compared with another counter-intuitive doctrine: epiphenomenalism. Epiphenomenalism holds that the mind is caused by the body (brain) but is itself causally inert (like the smoke above a factory). Now while epiphenomenalism emerged as a modification to mind-body dualism, it remains dualistic. All that changes is that mind to body causation is dropped. The slightness of the change leaves epiphenomenalism facing some of the problems the more straightforward dualism faced. To wit: epiphenomenalism leaves mysterious both the nature of mind and the – albeit unidirectional – relation between the body and the mind. It seems to me that RI has the same problems. What kind of thing (in broadest sense of ‘thing’) is an episteme? And just how does it manage to mirror practices? A related challenge emerges from the limits of the RI- epiphenomenalism parallel. Epistemes are meant to have what we may call, at least loosely, ‘an effect’ on entities. For epistemes determine entities in some wise. On that point, RI resembles dualism rather than epiphenomenalism. Descartes was asked how the body could interact with the mind. At least somewhat similarly, we need to know just how – and perhaps how at all – epistemes affect (or condition or determine or make possible) entities.
In the context of the philosophy of mind, naturalism or materialism were the banners of doctrines meant to spare us the conundra just indicated. The promise of avoiding the parallel problems in transcendental philosophy (or at least Foucault’s philosophy) seems a strong motivation for adopting such a naturalism here too. On the other hand, naturalism may have its own equally serious problems . . . Moreover, evidently I have dealt hastily with difficult material. Further yet, I have done so with an inadequate knowledge of Foucault.
PS: The dating the system is putting on this comment is, in the preview mode, wrong. I write this on 1 November at 2:49 am.
Hi Matthew,
A preliminary remark: I think that there are two different issues implicitly intertwined in what you say. The first one is exegetic, and concerns the position that can be legitimately attributed to Foucault re: the transcendental (1). The second (as mentioned in my previous reply to you) is whether there is a justification for transcendental philosophy in general (2). I haven’t addressed (2) at all in the talk because Foucault himself doesn’t, he starts with the claim that it is possible and legitimate to go from empirical discourses to their conditions of possibility, and that these should be thought of as a historical a priori. As I said, there may be a more specific argument in the Preface but I haven’t had the time to look at it yet. My take on (1) is that seeing early Foucault as trying to historicise the Kantian project is the best way to make sense of his claims, and that such an endeavour faces deep but interesting problems (by contrast, Gutting’s claim is that Foucault is neither a transcendentalist nor a philosopher at all; but so far he hasn’t produced the sort of argument against transcendental philosophy you mention). Yet even if the idea of historical a priori (or of transcendental philosophy in general) turned out to be untenable, this would not affect the exegetic position I adopted re: (1). In other words, even if the anti-transcendentalist was right in rejecting transcendental claims, it wouldn’t mean that he is right about Foucault not having any such claims.
Re: the substance of (1), your post basically develops two of the objections I put to Foucault myself, namely the issue of empirical change/causality and that of the modality of his claims. Yet to be fair to him, to say that he “holds that epistemic frameworks are caused by empirical practices” is begging the question by attributing to him a claim he is careful not to make (in fact, he rejects it explicitly). Same thing with the idea that “the ground for empirical practices is nothing more than the practices themselves”, this is not an assertion that early Foucault would endorse (as shown by his analysis of History as underlying the practice of historiography). The question is whether he needs this sort of claim, hence the discussion with Joel and the suggestion (b) as a possible new way out. At any rate, his position is not obviously incoherent in the way you describe. But you’re right that (b) opens up to the more general problem identified above in (2), which does not pertain to Foucault only but to the grounding of any transcendental claims.
On this second issue, if I could give you a definitive answer right now we wouldn’t need the three year TP&N project! From a Foucaldian perspective, I’ll look at the Preface as soon as I get the time (apart from my other commitments, I’m discovering that answering posts is very time consuming…) and will get back to you. From a more general perspective, just a remark: as I understand it, the anti-transcendentalist argument you suggest (I don’t know whether it’s yours or whether you’re being the devil’s advocate here) is premised on the idea that epistemic frameworks emerge from practices and thus cannot pre-exist them. Although they can be studied in isolation from the practices they stem from, such frameworks can’t be said to perform the sort of determination that one would call transcendental. However the example given to illustrate this view does not work. You suggest that “coming to understand how certain general features of language use are governed by syntactic rules has made possible the practice of studying grammar (...), but that it would be putting the cart before the horse to say that the study of grammar made possible language use”. The transcendentalist’s rejoinder would be, presumably, that the STUDY of grammar certainly does not make language use possible, but that the existence of grammar does (insofar as without it there would be no language use at all, let alone any study of it). In the same way, the transcendentalist could claim that the genesis you described is fallacious because it is impossible to understand practices without epistemic frameworks in the first place (in the same way that Heidegger, for example, says that an understanding of being is involved in everything we do, and is a condition of possibility of our doing it).
PS: Nick, I'm not sure how far the analogy with epiphenomenalism (as "the idea that the mind is caused by the body (brain)") would go as Foucault rejects the idea that epistemic frameworks are caused by practices. Otherwise the general spirit of your comment is indeed similar to Matthew's.
Béatrice, thank you, the parallel with epiphenomenalism is indeed wrong and for the reason you give. Sorry. I was/should have been thinking of OCCASIONALISM, definable as follows (www.artsci.wustl.edu/~philos/MindDict/occasionalism.html). ‘A view popularized by Nicolas Malbranche whereby: (1) the mental and the material comprise two different kinds of substance; (2) neither has any direct causal effect on the other and; (3) all seeming interactions between the two are due to the continual intervention by God who brings about a change in one on the occasion of a change in the other.’ To connect with (your) Foucault we can the following substitutions. We replace 'the mental' with 'epistemes' and we replace 'the material' with 'practices'. (And we may need to strike the word 'direct' from 2.) Those substitutions prompt us to ask - and this is the upshot in the critical register - what we substitute for 'substances' and what we substitute for 'God'.
Hi again Nick,
Just a quick reply: I don't think that this second analogy will go very far either (sorry!) as occasionalism was introduced to explain (or rather reject, against Descartes, the possibility of) causal interaction between substances. But the épistémè is not a substance, and the relation suggested in (b) is not causal. As far as I can tell it's a different problem altogether, so I'm not sure that the analogy is enlightening. Malebranche is a metaphysician, Foucault isn't.
The Cartesian dualist faces the problem of how two very different kinds of thing, namely mental substance and material substance, can interrelate. Foucault does not deal in substances. Yet he does posit a relation between EMPIRICAL practices and NON-EMPIRICAL epistemes (whatever ‘empirical’ and ‘non-empirical’ mean here – and one wants to know). Further: like Malebranche Foucault appeals to a mirroring, as against causal, relation between his relata. If these points hold then the analogy has something in it. And I thought that the analogy helped frame some of the questions discussed on these pages.
Beatrice, you are right, the analogy with grammar doesn't work as I stated it. I realised that as soon as I posted. But there's no editing posts post-publishing!
You are right I am playing devil's advocate a little. But just to return to the analogy. Someone like Wittgenstein would say that grammar (now understood in the very broad sense of norms implicit in language-games) is indeed tacit knowledge that in some sense makes possible language use. But he would deny that grammar somehow pre-exists language use, as it only makes sense in terms of linguistic usage (hence the slogan 'meanings is use'). So here's a question from a Wittgensteinian perspective: would 'epistmic frameworks' make sense in isolation from the practices that instantiate such frameworks? If not, is there any reason to see such frameworks as anything more than explications of certain general features that practices share? This would seem to me the general worry that would have to be answered. You mention Heidegger's notion of an understanding being, and perhaps there is a case for saying there is some framework that is pre-understood by people living in a given epoch. But (1) I am wary of philosophical positions motivated solely in terms of a habitual philosophical way of thinking, such as 'there must be a more fundamental level that makes this level possible'; and (2) I still don't see why we can't see what is pre-understood by people living in a given epoch as simply the motley collection of the practices such individuals are initiated into.
Beatrice's paper is now available to download without an Essex password. See the link in Matt's original post.
I think this debate about the a priori will be aclared with a lecture of Historie de la Folie.
Han says that the first time of the aparition of this term is in Naissance de la Clinique, but the first ocorrence is in La Recherche Scientifique et la Psychologie (1957), and in HF.
In HF, the ocorrence of a "concrete a priori of all psychopatologies with scientific pretentions" can speak a lot about the diferences between a priori and episteme (in one side, and "episteme" is not tematized in HF), and about the debate of a confusion between the empirical and transcendental level.
Post a Comment
<< Home