|Rules, Regression and the ‘Background’:
Dreyfus, Heidegger and McDowell
The work of Hubert Dreyfus interweaves productively ideas from, among others, Heidegger and Wittgenstein. A central element in Dreyfus’ hugely influential interpretation of the former is the proposal that, if we are to – in some sense – ‘make sense’ of intentionality, then we must recognize what Dreyfus calls the ‘background’. Though Dreyfus has, over the years, put the notion of the ‘background’ to a variety of philosophical uses,1 considerations familiar from the literature inspired by Wittgenstein’s reflections on rule-following have played an important role in motivating the case for believing that we need to recognize the ‘background’ and thus also in identifying precisely what it is about the intentional that supposedly needs to be ‘made sense of’.
Dreyfus argues that what he calls ‘representationalism’ will land us with an unstoppable ‘regress of rules’. In this paper, I first argue that there are actually two different arguments that Dreyfus invokes; I then go on to evaluate quite how, in the light of the problems that those arguments reveal, our position might be thought to be improved by our recognizing the ‘background’. Given that various philosophical positions designed to deal with these problems have emerged within the Wittgensteinian literature, an obvious question to ask is whether the position that Dreyfus would have us adopt is essentially one of those positions. If it isn’t, then how does it differ? There is surely a variety of ways in which such a comparison might be carried out and what I offer is only one. I argue that if, through a recognition of the ‘background’, we are thought to have acquired solutions to those problems, then it’s not at all clear that the supposed solutions that emerge work. So I explore instead the possibility that that recognition forms part of an attempt to ‘dissolve’ those problems. In order to bring some clarity to that possibility I consider a number of different ways in which Dreyfus’ proposals might be interpreted by drawing on ideas set out by John McDowell (and I suggest that his view of one of the ‘regress’ arguments is anticipated by Heidegger himself). I then identify and assess some of the consequences of adopting such McDowellian readings.
My sense is that Dreyfus is on the side of the angels, so to speak. But if what is right in his proposals is to become clear, and if he is to be spared some obvious objections that those proposals may elicit, we need to be clear about just what kind of contribution those proposals are meant to make. In pursuing that clarity, I am attempting to follow through on the comparison of Wittgensteinian and Heideggerian ideas that Dreyfus and his supporters have initiated: what has yet to be clarified is how and why recognizing the ‘background’ will allow us to ‘cope better’ with the puzzles in the rule-following literature that they have cited in making a case for the need to recognize the ‘background’. Ultimately, I will argue that assessing this matter may require a yet broader comparison of Wittgensteinian and Heideggerian themes, one which raises questions about what we take ‘doing ontology’ and ‘doing phenomenology’ to be.
1. What is the Dreyfusian ‘Background’ meant to do for us?
1.1 An Initial Sketch
In Heidegger’s discussion of ‘being-in-the-world’, Dreyfus sees described 'a mode of awareness' (BITW 68, HC 34) that is intentional in that it 'reveals entities under aspects' (BITW 68) but which does not consist of having ‘representations’ (HH 9). This ‘background’ is not a matter of 'mindless’, ‘mechanical’ or 'zombie-like behaviour' (BITW 68, HC 34); but it is, nonetheless, ‘nonmental’ (BITW 76), ‘non-cognitive’ (HH 9), ‘unthinking activity’ (HC 35). Dreyfus can propose that the ‘background’ is neither ‘mental’ nor ‘mindless’ because he also maintains that the ‘traditional’ vocabulary of philosophy is inadequate for describing this ‘background' (BITW 7):
The Cartesian/Husserlian ontology of the brute physical world and the intrinsic intentionality of individual minds just is not rich enough to explain how we are able to act. We may just have to grit our teeth and countenance body-intentionality and being-in-the-world as a third way of being. (R 336)2
What is it then that a recognition of this ‘third’ ‘background’ way of being is meant to do for us?
According to what Dreyfus calls ‘representationalism’, ‘the mind is defined by its capacity to form representations’:
On this view all that we know – even our general know-how for getting around in the world and coping with things and people – must be mirrored in the mind in propositional form. … Representationalism assumes that underlying everyday understanding is a system of implicit beliefs. (WCSCD xvii)3
In opposition to this ‘view of the mind and its relation to the world’, Dreyfus proposes that:
[R]elating to objects by way of intentional states such as desires, beliefs, perceptions, intentions, etc. [is] a derivative and intermittent condition that presupposes a more fundamental way of being-in-the-world that cannot be understood in subject/object terms. (BITW 5)4
As Wrathall puts it, Dreyfus argues ‘that intentional states can only have a content against a non-intentional or pre-intentional background’ (2000 p. 93). These negative characterisations of this ‘more fundamental way of being-in-the-world’ are filled out using a variety of different - though clearly related – notions that include ‘skill’, ‘coping’, ‘know-how’, ‘body-intentionality’, ‘custom’ and ‘practice’: 'For Dreyfus … the background is a set of practices, skills, and activities' (Wrathall 2000 p. 93). Recognition of this ‘background’ is meant to have wide import, ‘help[ing] us to see how phenomena as diverse as consciousness, intentionality, rule-following, knowledge, and representation presuppose skills, habits, and customs’ (Stern 2000 p. 53). What emerges is typically thought of as a novel ‘account of intentionality’ (Rouse 2000 p. 7).
1.2 Some specific arguments
But exactly what is it about intentionality that needs accounting for, that needs to be 'made intelligible' or shown to 'make sense'? A natural way to approach this question is by working our way backwards from the arguments that have been presented as demonstrating the need for the Dreyfusian ‘background’; in this way we come to see needs that it is meant to satisfy and thus what it is that is supposedly in need of explanation.
Although Dreyfus states that Heidegger does not try to prove his ‘theses’ (BITW 60, 120), in commentaries on Dreyfus’ discussion of the ‘background’, three particular arguments are attributed to him again and again. One is phenomenological and is illustrated by the following passage from BITW 86:
Heidegger can and does claim to have given a concrete demonstration of his position, by showing that when we carefully describe everyday ongoing coping activity we do not find any mental states.5
A second consideration which is frequently cited is the claim that ‘knowing-how is not reducible to knowing-that’.6 Though there has been a recent revival of interest in this claim,7 Stern has argued – plausibly, I think – that it is less than clear how this claim contributes to a case for the existence of a Dreyfusian ‘background’; in particular, Stern argues that what support it provides seems best characterised in other terms, for example, as embodying a version of the phenomenological argument above or the observation that we have failed thus far to produce analyses of forms of knowing-how in terms of knowing-that, an observation which is, of course, compatible with the possibility that we might produce such analyses in the future.
Since the phenomenological argument above also invites an obvious response - that we manage to deal skilfully and intelligently with the world around us without entertaining explicit beliefs about the objects with which we are dealing does not rule out the possibility that this feat rests on certain implicit or unconscious beliefs8 - commentators seeking a compelling argument for the existence of the Dreyfusian ‘background’ have been forced to look elsewhere. An argument to which commentators on Dreyfus repeatedly turn is an argument which turns on what might be called a ‘regress of rules’; but as I will show, there are, in fact, at least two different arguments there to be unearthed.9
2. A first ‘regress of rules’
2.1 Rules for applying rules
Wrathall gives a brief but clear statement of the first argument:
[T]he application of rules itself depends on skills for applying rules. If we try to capture those rules in terms of the application of further rules, then … (Wrathall 2000 pp. 96-97, ellipsis in the original)
The philosopher who believes that one can explain a person’s behaviour on the basis of, so to speak, ‘rules alone’ faces an explanatory gap that cannot be crossed; the system of rules envisaged would need to be ‘capable of expressing how to apply its own concepts and rules’; however, as Rouse puts it,
That is not possible, because no rule or concept can determine its own correct application. (2000 p. 17)10
We do indeed find something like this argument in Dreyfus. The philosopher of ‘rules alone’
will either have to admit a skill for applying these rules or face an infinite regress. Or, if he says that one doesn’t need a rule or skill for applying a rule, one simply does what the rule requires, … [then] why not just accept that one simply does what the situation requires, without recourse to rules at all? (HH 8-9)11
The argument is perhaps best known today through its occurrence in Wittgenstein’s rule-following considerations;12 but the fact that the crucial notion upon which, as Dreyfus makes clear, the argument turns - the notion of one’s needing a rule for applying a rule - will lead to trouble has been recognized at least since Kant, who, in a well-known passage (Kant 1961 [1781/1787] A133/B172), cites the need to halt the regress described as revealing the need for the ‘peculiar talent’ he calls ‘judgment’, ‘the power of rightly employing’ rules, ‘distinguishing whether something does or does not stand under a given rule’.
As Sec. 1.1 made clear, Dreyfus’ ‘background’ provides a ‘background’ not only to rules but also to all intentional states; so how might the above argument about rules be generalised? Robert Brandom, another important reader of Heidegger, sets out what might be a suitable generalisation in the following reflection, which turns on an analogue of the crucial notion about rules set out above; the analogous notion here is that one’s grasp of the ‘representational purport’ of a belief might itself come in the form of a belief:
Taking something as a representation must not be parsed in terms of the adoption of explicitly contentful attitudes or intentional states such as belief. If being a consumer of representational purport, taking something as a representation of something, is understood as believing of it that it correctly represents (or equally if the purport is understood as intending that it do so), then an infinite explanatory regress is generated by the possibility of querying the nature of the representationalist purport (‘that …’) and success (‘of …’) such a belief exhibits. There must be some way of understanding something as a representation that consists not in interpreting it (in terms of something else understood as a representation) but in taking, treating, or using it as a representation. (1994 p. 74)
Without the further, ‘supplementary’ kind of understanding to which the argument seems to point, a representation seems to stare back at us as ‘just another bit of worldly furniture, like what it represents’:
Why is not confronting a map as well as terrain just adding one more thing to be baffled about? (1994 p. 74)
But before considering whether Dreyfus’ ‘background’ ought to be understood as such a ‘supplement’ to ‘representations’, let us examine whether the need for such a ‘supplement’ is quite what the above argument demonstrates.
2.2. A Problem to be Solved or Dissolved?
Consider a second formulation of the generalised form of the argument, one offered by Stroud in an attack on the notion that 'thinking, or intentional phenomena generally' might be 'accounted for or even fully described by speaking exclusively of "representations" or "intentional contents" … present to the mind':
It seems obvious that ‘representations’ alone are not enough because, at the very least, the person also has to … ‘grasp’ the ‘content’ that is represented there. That suggests that thinking involves both ‘representations’ or ‘contents’ and a grasp or an understanding of them. … To say that [those graspings or understandings] too consist of nothing more than the presence before the mind of … ‘representations’ would lead eventually to a regress. (Stroud 1991 p. 245)
What Stroud’s presentation begins to make apparent is that this argument turns on a very particular understanding of ‘representations’.
Crucial questions, to phrase them in terms of rules once again, are: What understanding of a rule entails that one ‘needs a rule or skill for applying a rule’? Why not say instead that what it is to know a rule simply is to be able to apply it? In saying that only someone who has ‘grasped’ the rule can apply it, what is it that the addition of a ‘supplementary’ ‘grasp’ is meant to bring? And to what exactly? In our first ‘regress’ argument, we are being asked to imagine someone who in some sense ‘knows the rule’ but at the same time does not ‘grasp’ it: as Stroud puts it, it is ‘present before the mind’. But what kind of thing is this ‘rule’ that it can be ‘known’ in this way? To reject the notion that what it is to know a rule just is to be able to apply it seems to involve embracing a notion of ‘knowing a rule’ as ‘having present before one’s mind’ something like a string of signs whose meaning remains to be established, a version of what McDowell has called the ‘master thesis’:
[T]he thesis that whatever a person has in her mind, it is only by virtue of being interpreted in one of various possible ways that it can impose a sorting of extra-mental items into those that accord with it and those that do not. (McDowell 1993 p. 270)
Just as one might feel that one must not only have a rule ‘before the mind’ but also ‘grasp’ it, or that real mastery of a rule requires that one also master a rule for applying that rule to the world, Brandom’s argument turns on ‘the possibility of querying the nature of the representationalist purport (“that …”) and success (“of …”) [that] a belief exhibits’. The McDowellian would suspect that this ‘possibility’ rests on the ‘master’ notion - the ‘extraordinary idea’, as McDowell puts it – that one can ‘have’ a thought ’before the mind’ without also ‘having before the mind’ the kind of worldly situation that that thought represents. In opposition to this ‘possibility’, McDowell proposes that
[A] thought, just as such, is something with which only certain states of affairs would accord. (McDowell 1993 p. 270)
2.3 Heidegger on the First ‘Regress’ Argument and the ‘Master Thesis’
It is easy to suspect – because it may be true - that our first ‘regress’ argument is often read into Heidegger’s work by readers familiar with it in Wittgensteinian or Kantian guise. Nevertheless, we do actually find a version of it in the 1925 lecture course, History of the Concept of Time; but what is most striking about its presentation there is the attitude that Heidegger adopts towards it: what we find there is an anticipation of McDowell’s diagnosis.
In the passage in question (HCT 42-43), Heidegger attacks ‘efforts … to take the apprehension of a picture as the paradigm by means of which … any perception of any object can be illuminated’. Heidegger insists that the ‘real reason’ for rejecting this proposal is that ‘it does not correspond to the simple phenomenological findings’; but he does throw in a further ‘difficulty’ which he will ‘only mention without exploring’. 'If knowledge in general is an apprehension of an object-picture as an imminent picture of a transcendent thing outside', then we face not only the familiar, proto-sceptical question, 'how then is the transcendent object itself to be apprehended?', but also a puzzle about why confronting the object-picture as well as the transcendent object isn't, to echo Brandom, 'just adding one more thing to be baffled about':
If every apprehension of an object is a consciousness of a picture, then for the immanent picture I once again need a picture-thing which depicts the imminent picture for me etc. etc.
But Heidegger continues to insist that this is a ‘secondary factor’ and not ‘the main thing’:
It is not because we fall into an infinite regress, and so explain nothing, that the infrastructure of the consciousness of a picture for the apprehension of an object is to be rejected. It is not because we arrive at no genuine and tenable theory with this infrastructure. It is rather because this is already contrary to every phenomenological finding. It is a theory without phenomenology.
Exactly what we are to make of the charge with which this paragraph ends is a question to which Sec. 4.2 will return. But a natural construal of his point here is that the ‘infinite regress’ alluded to only arises if we first endorse a theory which, as Heidegger puts it, ‘goes against all the plain and simple findings’; those in the grip of such an ‘extraordinary’ theory fail to recognize that
[c]onsciousness of a picture is possible at all first only as perceiving, but only in such a way that the picture-thing is actually apprehended beginning with what is pictured on it.
To actually think of something as a picture – to ‘apprehend’ a ‘picture-thing’ – is to ‘begin with what is pictured on it’. To ‘have a picture before the mind’ is to have before the mind something which - as McDowell puts it, ‘just as such’ - is something which ‘shows something, [namely,] what is pictured itself’ (HCT 42). Heidegger’s explanation of how we come to our ‘infinite regress’ is that it emerges only if we begin with the ‘extraordinary’ notion that a picture is 'a thing like a natural thing or another environmental thing' (HCT 42); the argument does not lead us to, but instead begins with, that ‘master’ notion that a picture is just another piece of 'terrain … to be baffled about’.
2..4 The Implications of a Critical McDowellian Reading of Dreyfus
The McDowellian and, I have argued, Heideggerian diagnosis offered here suggests that it is only if one first embraces something like the ‘master thesis’ that one comes to believe in a ‘gap’ between rules and representations, on the one hand, and the world, on the other, a gap which a ‘supplementary’, non-representational ‘grasp’ or ‘judgment’ would bridge. But if the ‘master thesis’ is a confusion, the puzzle to which it seems to lead is an illusion, and so is the ‘grasp’ which seems necessary if that puzzle is to be solved: that ‘grasp’ would meet an illusory need.
So how, then, does Dreyfus see things? Is his invocation of the ‘background’ part of a confused attempt to solve this first ‘regress’ problem? His remarks about the need to ‘grit our teeth and countenance … a third way of being’ might well suggest that the ‘background’ plays something like the ‘supplemental’ role sketched. By embodying 'a mode of awareness' which is itself non-representational but which 'reveals entities under aspects', the Dreyfusian ‘background’ would seem to fit the job description of something that would bridge the ‘gap’ which, according to our first ‘regress’ argument, ‘representations’ leave us facing. In this way, the postulation of the 'background' could be presented as 'accounting for intentionality', the need for such an ‘account’ emerging from the ‘fact’ that - as Dreyfus does at one point state - ‘rules are, by themselves, meaningless’ (BITW 118): but, as I have made clear, the McDowellian/Heideggerian response would be that only an ‘extraordinary’ idea of rules would lead one to think that.
Another reason for adopting this critical McDowellian reading of Dreyfus is that it would provide an explanation of how he comes to feel driven to claim that ‘everyday skillful engagement with familiar things’ does not involve ‘representations’, a claim which, Christiansen has argued, we can in no way take literally:
Surely, when I am routinely hammering away, I do see that or how the nail is going as it should, namely, straight, as I intend. Surely, I quite literally perceive, come the appropriate moment, that the nail has been hammered in as required, so that it is time to stop hammering. (Christiansen 1998 p. 66)13
If Dreyfus were, on some level, under the influence of the ‘master thesis’ and saw his ‘background’ as playing the ‘supplemental’ role that our first ‘regress’ argument ‘reveals’, that would explain why he might feel the need to deny Christiansen’s seemingly truistic claims. The first ‘regress’ argument may seem to show that our most immediate dealings with the world must be non-representational: the final step whereby one recognizes, for example, that a spade before one is indeed a spade would seem to be something that a further ‘supplemental’ ‘skill’ has to make possible, despite the fact that we would ordinarily talk quite happily of ‘seeing that the object was a spade’. But, from the McDowellian perspective, this inclination to withhold a perfectly normal use of the propositional attitude idiom arises out of a residual commitment to a ‘master’ understanding of propositional attitudes such as perceiving that …, recognizing that …, etc. It is only on that understanding that those attitudes leave us confronting the ‘gap’ described. If one rejects that understanding, then the gap evaporates along with the need for non-representational ‘skills’ to step in and assume roles that we would ascribe – in ‘ordinary language’ - to seeings that …, perceptions that ..., etc.
Now ordinary language could, of course, be confused and we might find that we need a new descriptive vocabulary. But the McDowellian diagnosis we have examined offers an alternative, and I think more appealing, explanation of how and why that ‘need’ might seem to arise, of how and why we may find ourselves driven to view with suspicion utterly natural ways of describing utterly familiar activities of ours. According to that diagnosis, in ascribing a hitherto unappreciated and wider role to ‘the non-representational’, an earlier ‘move in the conjuring trick’ (Wittgenstein 1967 sec. 308) - an earlier confusion about ‘the representational’ - has influenced us: our descriptions of familiar phenomena have become forced and peculiar, as we wedge ‘the non-representational’ into areas which we - in our ‘master’ confusion - see ‘the representational’ as having forfeited but which we - with what remains of our good sense - will still naturally want to describe in the representational terms that ‘ordinary language’ presents to us.14
I have parallel Wittgensteinian worries about Dreyfus’ phenomenological case for thinking that ‘expertise’ is ‘a way of coping in which reasons play no role’; Dreyfus bases this claim on observations such as that the expert ‘depend[s] entirely on perception and not at all on analysis and comparison of alternatives’ (OMM 53, cf. also MOM). But as Wittgenstein argued, the relationship between the presence of introspectively available ‘mental events’ and our sense of an agent’s acts being rule-guided is really quite loose: for instance, we have little reason to think that for an agent to be following a rule some particular formula must be ‘present to the mind’.15 From my point of view, what drives Dreyfus to deny - incorrectly - that rules ‘play a role’ in such situations is his recognizing - correctly - that certain simplistic images of rules and our life with rules really tell us very little about those rules or that life.
A fair question for Dreyfus to ask now is: So just why do I want to say that rules are involved in these cases? One – no doubt disappointing - Wittgensteinian reason would be that there are enough similarities between – and continuities with – straight-forward cases of rule-following for it to make sense: for instance, although chess masters ‘may be at a loss to reconstruct a reasoned account of [their] actions’ (OMM 54, my italics), they also often can talk illuminatingly – and in general terms - about just what led them to make the moves that they made, and our sense of the brilliance of chess masters and the fact of their success do not typically - as Horgan and Timmons (unpublished) have recently put it - ‘jar’ with subsequent reason-based reconstructions of what these players did. The absence of enough such similarities and continuities would be my answer to the other, equally fair, Dreyfusian question, ‘So just what would it take for you to say that rules aren’t involved?’ Adopting this policy spares us - as above – the need to attack a lot of intuitively natural ways of talking, such as that chess masters are engaged in a ‘mental’, ‘cognitive’, ‘reason-based’, ‘rule-governed’ activity.
2.5 A More Sympathetic McDowellian Reading of Dreyfus, and its Remaining Critical Implications
But might one yet argue that Dreyfus’ remarks on the ‘background’ are precisely meant to point out how ‘extraordinary’ some of our thinking about rules and representation is? He observes that:
The essential characteristic of representations according to the tradition is that they are purely mental, i.e. that they can be analysed without reference to the world. … Heidegger rejects this traditional interpretation of the mental. Even deliberation is not the pure detached theoretical reflection described by the tradition. (BITW 74)
In remarks such as these, Dreyfus seems to be distinguishing representations as they really are from what those in the grip of a certain philosophical myth think of representations as being, from ‘the sort of self-sufficient entities philosophers since Descartes have supposed’, 'self-contained representations’ (BITW 74, 75). So might we instead read Dreyfus as making precisely McDowell’s point? I think we can, but at a price.
Firstly, the first ‘regress’ argument would no longer give us any reason to propose that ‘everyday skillful engagement with familiar things’ does not involve ‘representations’; the ‘gap’ that that argument exposed is between the world and ‘master’ ‘representations’, not representations as they actually are.
Secondly, this McDowellian reading affects how we ought to hear the claims that Dreyfus makes, and are made on his behalf, about having revealed ‘conditions of’ – having ‘accounted for’ – the ‘possibility’ of intentionality. On this reading, the only sense in which we have established ‘how intentionality is possible’ is by recognizing how a confused, ’extraordinary’ idea made us think that it was impossible. From the McDowellian point of view, when, in rejecting the ‘master thesis’, one comes to state that ‘a thought, just as such, is something with which only certain states of affairs would accord’, what one is stating is a truism. One is returning to a piece of lost common-sense rather than completing a piece of ‘constructive philosophy’ (1994 p. xxiii). Though what one ought to make of that claim is a topic to which Sec. 4.1 will return, it does seem that, according to the present, more sympathetic reading, Dreyfus does not explain how real, non-‘master’ ‘representations’ - ‘representations as they actually are’ – ‘are possible’: no question of ‘possibility’ was ever successfully raised about them. Rather he shows how a fantasy of representation seems to uncover the need for a ‘supplementary’ ‘grasp’ which would then make those (fantastical) ‘representations’ ‘possible’ once again.
The idiom of ‘grounding’ and ‘presupposition’ that is so characteristic of Dreyfus (and Heidegger) can still play a part in reflection that is meant precisely not to provide ‘constructive’ explanations but to show that our demand for such explanations is confused.16 So, for example, one might still articulate the point of the preceding paragraph by saying that a ‘background’ is ‘presupposed’ by ‘the representational’ (BITW 5). If we took ‘the representational’ to refer to the ‘master’ myth, we cannot take the ‘presupposition’ in question to reveal an explanation of how that is ‘possible’ or is 'founded', because, as a myth, it is neither. But one might instead say that that myth presupposes a ‘background’ in that those who are in its grip, if they are to believe that they can make sense of intentionality, must take for granted further forms of understanding which will ‘reach outside’ the mythical ‘closed’ ‘inner space’ that is the home of the ‘representations’ that the ‘master thesis’ conjures up. Such a sense of having discovered further non-representational forms of understanding would be the natural way for one still in the grip of that thesis to react; but ultimately, they ought to recognize and reject that thesis.
If this is how we interpret the ‘presupposition’ in question, what emerges with the ‘background’ is not a novel and coherent ‘account of intentionality’. If the ‘inside’ by reference to which such a ‘reaching outside’ is defined is a confusion, so too is the imagined need to ‘reach outside’ and the ‘capacities’ or ‘skills’ that we imagine meeting that need. The point of showing that this myth needs to be supplemented in the way described would then be to show that it is indeed a myth and the puzzle to which it seems to lead an illusion. In the light of that puzzle, intentionality had seemed ‘unaccountable for’ and the ‘job description’ emerged which a hitherto unacknowledged ‘background’ was believed to fit; but if the puzzle is an illusion, then that undermines at least one reason for seeing intentionality as needing ‘accounting for’ and for believing that there is a real task for the envisaged ‘background’ to perform. So a third critical consequence of this more sympathetic, McDowellian reading is that our first ‘regress’ argument gives us no reason to think that representations as they actually are need to be supplemented by a non-representational ‘background’ that illustrates a third way of being; indeed it is in our thoughts about the ‘second’, about ‘the intrinsic intentionality of individual minds’, that our confusion lies and what we need to do is identify and eliminate that confusion – the ‘master’ myth - not add a ‘third way of being’ to our ontology.
I will end this section with the suggestion that, on this reading, a recognition of what is right in Dreyfus’ remarks might require a recognition of what could be called their ‘reactive’ or ‘dialectical’ character. For example, in saying that ‘deliberation is not the pure detached theoretical reflection described by the tradition’, we need to recognize that actually nothing is ‘pure’ or ‘detached’ in the way that ‘the tradition’ imagines; when we think with ‘the tradition’, we have no determinate notion of ‘purity’ or ‘detachment’ in mind. But if this is so, saying that deliberation is not ‘pure’ or ‘detached’ ought not to be our final conclusion: instead we need to recognize that we are operating with confused notions of ‘detachment’ and ‘involvement’, ‘purity’ and ‘impurity’. Heidegger recognizes this difficulty in the following remarks:
Inasmuch as Dasein exists qua being-in-the-world, it is already out there with beings; and even this manner of speaking is still distorted [schief]17 since 'already out there' presupposes Dasein is at some point on the inside. Even if I say, Dasein's intentional activity is always already open towards beings and for beings, there is still at bottom the supposition that it was once closed. (MFL 167)
Drawing on an interpretation of Heidegger’s proposal that all philosophical concepts are ‘formally indicative’ (FCM 293),18 Dahlstrom (1994 p. 787-88) has suggested that Heidegger’s ontological proposals might need to be ‘climbed and thrown away’, as Wittgenstein declares those who understand him ought to treat the propositions of his Tractatus (1922 sec. 6.54). Though Dreyfus interprets the notion of ‘formal indication’ radically differently (in his CTT), he does seem to see Heidegger’s supposed ‘prioritising of doing over knowing’, his ‘reversal’ of the tradition’s ‘usual priority’, as possessing something like the ‘peculiar “transitory” character’ that Heidegger ascribes to ‘formal indications’ (PIA 87).19 Dreyfus does not ascribe that ‘character’ to the core proposals he himself wants to make about the ‘background’; but our more sympathetic McDowellian reading suggests that perhaps he should.
3. A Second ‘Regress of Rules’
3.1 The Impossibility of ‘Spelling Out’ an Open-ended Web of Ceteris Paribus Clauses
Despite its heritage, its familiarity, and its invocation by Dreyfusian commentators and – on occasion - by Dreyfus himself, our first argument is not the ‘regress of rules’ argument that Dreyfus typically offers. Consider the following passage:
When I am acting transparently – for example, making a promise – I do not need any rules at all. I have learnt from imitation how to promise, and I am a master promiser. But if something goes wrong, I may have to invoke a rule – for example, the rule that one must keep one’s promise. But the important thing to notice is that this is not a strict rule whose conditions of application are stated in the rule itself. It is a ceteris paribus rule. In the case of an unfulfilled promise there are allowable excuses, such as I was sick, or I saw that what I promised would hurt you. The rule ‘always keep your promises’ applies ‘everything else being equal’, and we do not, and could not, spell out what everything else is nor what counts as equal. Moreover, if we tried to define each exception, such as being sick, we would again have to bring in further ceteris paribus conditions.20
These ceteris paribus conditions never capture, but rather presuppose, our shared background practices. These practices are an aspect of our everyday transparent ways of coping. Thus, understanding is not in our minds, but in Dasein – in the skilful ways we are accustomed to comport ourselves. Thus even when mental content such as rules, beliefs, and desires, arise … they cannot be analyzed as self-contained representations as the tradition supposed. Deliberative activity remains dependent upon Dasein’s involvement in a transparent background of coping skills. (BITW 75)
This passage has a number of interesting features. First of all, the crucial issue does not seem to be one with rules as such. Dreyfus distinguishes ‘a strict rule whose conditions of application are stated in the rule itself’ from ‘ceteris paribus rules’ and his point would seem to be that the majority of the familiar rules we follow in life cannot be translated into the former. In principle, the passage above seems to allow that some (real enough) rules could be ‘strict’ in the above sense and could be ‘spelt out’; if so, it also gives no reason to think, as Rouse puts it above, that ‘no rule or concept can determine its own correct application’. Similarly, when Wrathall proposes that if ‘the application of rules itself depends on skills for applying rules’, then ‘[i]f we try to capture those rules in terms of the application of further rules, then …’, pace Wrathall and his ominous ellipsis, no endless regress will result – at least according to this second argument – as long as we come eventually to ‘strict’ rules. Dreyfus’ point seems to be that he doesn’t think that that will happen when we consider ‘rules for applying’ familiar rules such as ‘one must keep one’s promises’.21
To begin the task of evaluating this argument, a natural worry that might be expressed at the close of the first quoted excerpt is: ‘So how on earth do we manage?’ What kind of response to this worry does the second excerpt embody? As it stands, one can imagine it eliciting just more worry: ‘Is Dreyfus just saying that we just do manage? If he is, then this misses the real worry, which is: How do we manage?’ When Dreyfus proposes that ‘our commonsense understanding … is not a propositional knowing-that … but rather consists of dispositions to respond to situations in appropriate ways’ (BITW 117) and invokes ‘a direct way of responding appropriately to the solicitations of the environment in which the agent is inextricably embedded’ (MPC 302), it is easy to hear some sense in the response, ‘But how can there be such things?’ To cite ‘know-how’, ‘sensitivities’, forms of ‘awareness’, ‘skill’ and ‘familiarity’ thanks to which we do supposedly manage to ‘manage’ or ‘cope’ will strike some as merely giving a (rather misleading) statement of what the problem is. So in what way does the citing of the ‘background’ provide us with more than a ‘We just can!’ response?22
We face a similar worry if we understand the argument as a reductio ad absurdum of a certain ontological austerity, the project of making sense of intentionality by reference to ‘representations alone’. The implications of the fact that such a project cannot work are unclear: it might be seen as showing that we must have certain further ‘capacities’ or ‘skills’, but to be told in this way that ‘there must be more’ will strike many as – again - simply identifying our difficulty not eliminating it.
There are, it seems to me, at least three different directions which one might take now, turning on just what kind of response one takes the citing of the Dreyfusian ‘background’ to provide. The first, which I will mention only briefly, is to say that Dreyfus is simply shifting the burden of proof on to representionalists, making clear that they have an awful lot of explanatory work to do and that the prospects don’t look good. Though there may well be something to be said for this proposal and there are points at which Dreyfus seems drawn to it,23 it, by itself, entitles us to no stronger conclusion than that both ‘sides’, as it were, in this debate might be equally stumped. But I believe that Dreyfus’ overall view is meant to be more positive than that, his invocation of the ‘background’ being meant to help us resolve – in some way - the problems that our second ‘regress’ argument raises
3.2 Problems with Providing a Solution by Filling Out the ‘Background’
A second option is that one might hope to be able to fill out one’s story of ‘skills’ and ‘familiarities’ in such a way as to provide a substantial solution to the problem identified above. So what precisely is the problem that the second ‘regress’ reveals? The representationalist seems to be committed to a ‘spelling out’ of the demands embodied in the rules that we follow that requires a kind of ‘closure’ that cannot be secured. This problem, which is, once again, one familiar from the rule-following literature,24 is also sometimes articulated using the concept of ‘relevance’: if the conditions that bear on whether a particular concept or rule should be applied are open-ended in the way described then there would appear to be no end to the task of determining whether ‘all relevant considerations’ have been taken into account because that set of ‘considerations’ is open-ended.25 The present section examines how Dreyfus’ work might be seen as providing a ‘filled-out’ solution to this problem; the picture that emerges is a largely critical one and the section that follows will explore a more sympathetic, dissolving and - once again - McDowellian reading.
Any attempt to fill out the story of the ‘background’ must address a version of the ‘We just can!’ complaint: there seems very little reason to think that ‘skills’ or ‘familiarities’ have any prospect of achieving a kind of ‘closure’ that cannot be achieved with rules. For example, one way in which the first excerpt might make one worry about our capacity to ‘manage’ is by raising the question of why I ought to believe that ‘I have learnt from imitation how to promise’ rather than qu-omise, a practice quite like promising but for failure to keep a qu-omise being excusable if it is made on May 5th 1973. Since this might turn out to be a clause buried in the imagined web of yet-to-be-articulated certeris paribus clauses, why ought I to think that I am indeed familiar with - skilled in - promising rather than qu-omising? Reference to clear instrumental skills, such as riding a bike, may provide us with a comfortable sense that, with skills, what constitutes mastery is not a difficult issue. But it is actually unclear why skills are any more able to deal with such ‘gerrymandering’ worries (to use Brandom’s term (1994 p. 28)) than dispositions are typically thought to be.26 (Sec. 3.3 will consider the possibility that the invocation of ‘skills’ etc. is precisely not meant to bring ‘closure’; the question then will be: how can we drop this demand?)
In considering these matters, an issue that arises is quite what kind of ‘filling out’ the ‘background’ story needs and what kind of thinker – which university department, as it were - is to provide it. Dreyfus has indicated that he saw the emergence of neural networks as a promising development for those interested in his issues but also that that optimism has waned; his reasons for both reactions are interesting.
His initial enthusiasm seems to have been prompted by the notion that neural networks might in some way provide an explanation of our senses of sameness without an (apparently doomed) recourse to rules. (Neural nets represent a possible answer to the question, ‘If not symbols and rules, what else?’ (WCSCD xiv-xv).)27 Suitably trained networks will respond appropriately to examples of a category with which they have never been presented and this may seem to suggest that such networks have mastered the similarity in question without their ‘containing’ anything that might be thought of as a representation of a rule that articulates what that similarity is.28
One way of articulating Dreyfus’ later pessimism is as arising from the realisation that such nets can be expected to produce the appropriate responses only as long as one restricts the inputs with which it is provided to ones sufficiently similar to those using which it was trained; however, if one steps outside of that domain – as we humans who are ‘able to project our understanding into new situations’ can (WCSCD p. xxiv) - one has no grounds for expecting the net to generalise in the way in which we intuitively would. This problem, I suggest, is the familiar gerrymandering problem: any sequence of examples that one might provide – any ‘history of training input-output pairs’ (WCSCD p. xv) – is consistent with any number of different generalisations. (‘Everything’, Dreyfus notes, ‘is similar to everything else in an indefinitely large number of ways’ (WCSCD p. xxvi).) Thus, similarities between the ways in which two nets are disposed to respond to a particular range of inputs do not give us reason to expect that they are disposed to respond in the same way to all inputs.29
What is interesting about Dreyfus’ disappointment here is that it seems to show that what he thinks we need is a theory that will solve problems such as the gerrymandering problem. But, given that we don’t have such a theory, it is unclear now why he still feels happy to say, for example, that ‘background sensitivities … determine what counts as similar to what’ (WCSCD p. xxix). In BITW, we read:
Referring is a shared social skill, and Dasein is socialized into this practice of pointing. We can either stop with this claim and do empirical work in developmental psychology to find out how people get socialized into a practice, or we can attempt to lay out the general structure of the phenomenon. Heidegger, of course, chooses the latter approach. (BITW 270)
One might propose that Dreyfus’ talk of ‘background sensitivities’ is a contribution to the latter project too and his interest in neural nets essentially an interest in the former.30 But if problems such as the gerrymandering problem are laid at the door of that ‘empirical’ project, then it’s unclear what philosophical work the ‘laying out’ project is to be expected to do and why the most just assessment of what it is that ‘determines what counts as similar to what’ is not that ‘background sensitivities’ do but that neither representationalists nor their Dreyfusian opponents can say: both camps would then, again, be revealed as equally stumped.31
3.3 The Possibility of ‘Dissolving’ our Second ‘Regress’ Problem and of a Second, More Sympathetic, McDowellian Reading of Dreyfus
Another way of interpreting how a citing of the ‘background’ might bear on the ‘spelling out’ problem would be to see this citing as part of an attempt to dissolve that problem. One ‘dissolving’ approach I will consider – and there certainly could be others - would be to argue that the sense that we face a puzzle in seeing how we ‘manage’ with capacities like promising turns on a very particular and questionable – one might say ‘extraordinary’ - understanding of what ‘seeing how’ must involve; one might argue, for example, that that sense of puzzlement precisely presupposes that ‘seeing how’ will come in the form of an account using which someone who has no mastery of the concept of ‘promising’ could still identify what counts as ‘a promise that must be kept’. McDowell’s plea for ‘modest’ theories of meaning and his rejection of the need for ‘a view from sideways-on’ is a rejection of such a demand.32 Might then we understand Dreyfus’ views in a similar way?
Such a reading would offer an interpretation of, for example, Dreyfus’ proposal that Heidegger ‘can only point out’ and ‘cannot spell out the background practices in so definite and context-free way that they could be communicated to any rational being or represented in a computer’ (BITW 4), Heidegger’s view being, according to Dreyfus, no worse off for that. Such a reading would also seem to chime with Dreyfus’ sense that the ‘background’ is ‘so pervasive in everything we think and do that we can never arrive at a clear presentation of it’ (BITW 32),33 a sense which commentators have struggled to articulate as a specific (and plausible) argument against representationalism.34 And the reading might also offer a way of understanding Dreyfus’ remarks about what one might call the ‘embedded’ character of the capacities that make up the ‘background’, remarks that would otherwise elicit particularly animated versions of the ‘We just can?!’ response.
Dreyfus suggests that our judgments of sameness are indeed a mystery if we imagine ourselves ‘stood outside the world’ (MOM 88).35 But rather than our being overwhelmed by the open-ended range of potentially relevant considerations that might influence whether one affirms or denies a claim such as ‘This is a promise that one ought to keep’, ‘[h]uman beings are somehow already situated in such a way that what they need in order to cope with things is distributed around them where they need it’ (WCSCD 260); our situation is ‘always already … permeated by relevance’ (OMM 49), ‘organized from the start in terms of human needs and propensities which give the facts meaning’ (WCSCD 262). In such a situation in which ‘man is at home’ (WCSCD 260):
[O]ur present concerns and past know-how always already determine what will be ignored, what will remain on the outer horizon of experience as possibly relevant, and what will be immediately taken into account as essential. (WCSCD 263)
Now it’s not hard to imagine a reader who would be exasperated by the above remarks. If the story concerning how ‘concerns’ and ‘know-how’ can do this ‘determinative’ work has to take for granted the fact that the world – ‘the human world’ - ‘is prestructured in terms of human purposes and concerns’ (WCSCD 261), then the pressing question simply now becomes ‘How does it come to be so prestructured?’
Clearly we do not want to ‘account for’ the ‘harmony’ between thought and world – and undermine our gerrymandering worries – through a hopelessly idealistic vision of objects somehow electing to present to us those aspects of themselves that correspond to the concepts we happen to use. I suggest that we might hear Dreyfus’ remarks on the ‘human world’ here as part of an attempt to reject any attempt to explain such a ‘harmony’. Dreyfus’ rejection of the attempt to see – from ‘sideways-on’ - what it is about an ‘inhuman’ world that calls forth our ‘human’ responses would then be a rejection of an ‘immodest’ attempt to understand the world in terms other than those that we use so as to explain why those terms are somehow suited to, or fit for, that world.36
3.4 Remaining Critical Implications of the Second, More Sympathetic, McDowellian Reading
But what price is there to be paid if we read Dreyfus’ proposals in this McDowellian spirit? For many philosophers of language, explanation just is immodest explanation; and McDowell agrees to the extent that he sees modest theories which, ‘by design, start in the midst of content’, as unable to ‘contribute to [the] task of representing content as an achievement’ (1987 p. 105). A McDowellian reading of Dreyfus here requires us once again, I suggest, to hear in a different way the claims that Dreyfus makes, and are made on his behalf, about having revealed ‘conditions of’ – having ‘accounted for’ – the ‘possibility’ of meaning and intentionality.
One might still articulate the McDowellian point by saying that the impossibility of the envisaged ‘spelling out’ reveals that a ‘background’ is ‘presupposed’ by ‘the representational’ (BITW 5), if one takes the latter to be something that can be understood ‘immodestly’ and the mode of ‘presupposition’ (as explained in Sec. 2.5) as once again one that makes clear the ‘architecture’ of what is, in fact, a confusion. The impossibility of such an immodest ‘spelling out’ could be said to reveal that we cannot stand outside of, and survey, our ways of making sense: at any stage in the project of ‘foregrounding’ - by attempting to ‘spell out’ - our understanding, there could then be said to remain a residual ‘background’ of understanding which we presuppose and have not ‘spelt out’. But what is crucial here is not a mistaken fixation on ‘representation’ at the expense of the ‘non-representational’, of a ‘third way of being’ (such talk might be best heard, echoing the discussion at the end of Sec. 2.5, as ‘reactive’ or ‘dialectical’); rather it is the confused notion of an immodestly ‘spelt out’ understanding of representation, of ‘the second way of being’.
It is by reference to that illusion of understanding that a challenge of explaining how representation ‘is possible’ seems to emerge. If one accepts that challenge, then, as Sec. 3.2 argued, it is unclear why the invocation of a background of ‘skills’, ‘familiarities’, etc. ought to help us meet it; if one rejects the challenge, then it seems misleading to describe what one has done as revealing the ‘conditions of’ – or having ‘accounted for’ – the ‘possibility’ of intentionality. In particular, one will not have provided a theory - based on ‘familiarities’ or ‘skills’, rather than, or in addition to, ‘representations’ – that will provide us with an explanation of how the world happens to succumb to human thought, or that is better placed to explain what immodest representationalist theories cannot. What has been revealed to us is not, as it were, the limitations of representation but the structure of a confusion about representation and of a wrong-headed explanatory project of ‘accounting for’ the ‘achievement’ of representation that that confusion inspires. On this reading and that presented in Sec. 2.2-2.5, what both ‘regress’ arguments fundamentally show is that - and how - our thoughts are informed by some crude notions about representation; when we attempt to think through these crude notions, we run into the regresses described, but that should lead us to recognize the crudity of those notions of the representational – of the ‘second way of being’ - and not to the need for a hitherto unrecognized, non-representational, third ‘way of being’.
We do indeed read in Dreyfus remarks such as that ‘the everyday world … cannot and need not be made intelligible in terms of anything else’ (BITW 122); 37 he declares that it makes no sense to imagine ‘background coping’ ‘failing’, that it is not even best thought of as a ‘relation’ (BITW 249, 347): as such, it makes no sense to imagine an account of how such ‘skills’ come to have traction on the world. In this sense, there is no deep ‘explanation of intentionality’ on offer that would ‘represent content as an achievement’.
But it is not clear to me that Dreyfus does really reject ‘immodest’ demands. As Sec. 3.2 argued, his disappointment with neural networks seems to be that they fail to explain a feat that ‘immodest’ theorists characteristically seek to explain. One might also think that Dreyfus’ thinking on another topic central to his work - but which we have yet to consider - commits him to ‘immodesty’. A major contribution of Dreyfus’ reflections on the prospects and appropriate methodology for the exploration of artificial intelligence is his revelation of unacknowledged and questionable a prioristic baggage carried by many scientists involved in that project. One reason why Dreyfus might present his own account as ‘laying out the general structure’ of an explanation of intentionality is surely because of an admirable concern to set AI research on the right track, combined with the belief that such research seeks to provide, in some sense, ‘an explanation of intentionality’.
But I suspect – though cannot hope to prove here - that it is at least an open question whether ‘modesty’ clashes with Dreyfusian hopes for a suitably rethought (non-Good-Old-Fashioned) AI. I would suggest that ‘modesty’ ought not to be seen as leaving us with ‘something that we cannot do’ (Wittgenstein 1967 sec. 374). For instance, one ‘explanation’ which it ‘rules out’ is (very roughly speaking) an empirical explanation of why desires for apples are satisfied by apples and why expectations of explosions are fulfilled by the occurrence of explosions; but what it does not rule out is an empirical scientific exploration of the underlying mechanisms through which such desires and expectations are realised in particular individuals or species. ‘Modesty’ does, I suggest, cast doubt on the hope that we might produce an account of Intentionality-as-such in terms of a single mechanism that could be expected to be found in all creatures that happen to possess those desires and expectations; but there are philosophers who are obviously much more enthusiastic about the prospects of scientific investigation of intentionality than Heidegger or Wittgenstein who think that the pursuit of such an account is misguided.38 39
4. Phenomenology, Ontology and Therapy
As I hope has become apparent, Dreyfus’ discussion of the ‘background’ is rich. I have only considered it as a response to the two ‘regress’ arguments40 and each of the four readings - two critical and two more sympathetic – that I have considered has evidence upon which it might draw in its support, as well as evidence that seems to count against it. As a result, an over-arching critical suspicion of mine is that Dreyfus may be drawn towards all of the ideas presented - ‘solving’ and ‘dissolving’ - even though some are incompatible with others.
The final sections of this paper will consider a sketchy but important objection to the sympathetic, McDowellian readings of Dreyfus presented here: it is that the broadly ‘therapeutic’ spirit of those readings is at odds with Dreyfus’s overall sense of the kind of work he is doing, which is instead a matter of ‘doing phenomenology’ or ‘doing ontology’. Objections on this level of abstraction can’t be answered in a study of this length and all I will do here is make some very brief suggestions about how we might refine two preliminary questions that this objection raises.
4.1 Are the McDowellian thoughts presented here necessarily ‘anti-ontological’?
A pointed way of posing our first preliminary question is to ask why an ontologist – Dreyfus perhaps - couldn’t argue, not only that the McDowellian ‘dissolving’ thoughts could form part of an ontological project, but also that to reject the ‘master thesis’ and a ‘view from sideways-on’ is precisely to embrace the kind of ontology that Dreyfus seems to offer, the necessary alternative to a ‘master’ vision of ‘free-floating’ representations being one which presents us with representations that are, one might say, ‘contextually-embedded’ and ‘engaged’. Such an ontology could also be seen as ‘dissolving’ our ‘regress’ problems: in connection with the representations that such an ontology acknowledges, one can raise neither the first ‘regress’ problem (because of their ‘engagement’) nor the second (because of their ‘contextual-embeddedness’). As was noted above, from McDowell’s point of view, if, in rejecting the ‘master thesis’, one comes to state that ‘a thought, just as such, is something with which only certain states of affairs would accord’, what one is stating is a truism, not a new ontological insight; but is McDowell right? Can the crucial notions upon which our McDowellian readings turn only find a home within an outlook – like McDowell’s - that rejects ‘constructive philosophy’?
Certainly the belief that some questions are good and need answering while others are bad and don’t cannot be seen as the preserve of the ‘purely therapeutic’ philosopher. But our McDowellian readings still yield important questions for the kind of ‘partially-dissolving’ ontology envisaged. As they stand, the criticisms that those readings yield - despite their strong Wittgensteinian tone – do not amount to general criticisms of the attempt to meet the challenge of the two ‘regress’ arguments by refining our ontology; rather they are more specific, asking why those proposals have been characterised in certain, specific ways: most importantly, as explaining how intentionality ‘is possible’ and as revealing a third, non-representational and non-physical ‘way of being’ which is illustrated by capacities which ‘account for’ our most basic forms of intelligent interaction with the world. Our ontologist might concede that the citing of the ‘background’ ought not to be characterised in those ways but still propose that it ‘contributes to the ontology of intentionality’. One can, of course, say that; but the crucial issue now is ‘What is that project and what is it to contribute to it?’
4.2 How ‘anti-therapeutic’ is Heidegger’s ontology/phenomenology?
Our second preliminary question would take this issue further, requiring us to reflect on just what Heidegger takes ‘ontological problems’ to be and just what it is about ‘ontological matters’ that can be ‘problematic’. He declares that ‘ontological problems’ call for ‘phenomenal clarification’ (HCT 243): but what that and the corresponding ‘obscurity’ that it combats are is one of the deepest questions we can ask about his philosophy. My own suspicion is that Heidegger’s work has more in common with a Wittgensteinian ‘reminding’ of an understanding from which we are estranged than the kind of philosophizing we might be inclined to think of as ‘refining our ontological theories’.
In this piece, I have argued that, were one to abandon ‘immodesty’ and the ‘master thesis’, one would spare oneself the ‘regress’ problems described. But that, in itself, does not prove that those notions are actually confused. (And nothing has been said here that leads, I think, in any clear way to a dissolution of, for example, the qu-omising worry set out in Sec. 3.2.) But to indicate briefly the kind of confusion that I believe these notions involve (and to develop suggestions offered in Sec. 2.3 and 2.5), I suspect that one would not eliminate – but rather disguise by, as it were, burying - the influence of a confused vision of ‘disengagement’ on our thinking about representations by declaring that they are actually ‘engaged’; at best, such a declaration draws attention to confused notions of ‘engagement’ and ‘disengagement’ informing our thought. If so, the way one ought to challenge confusions such as the ‘master thesis’ is not by presenting alternative accounts of that which they mythologize, ascribing to it properties that those confusions lead us to deny it and vice versa; instead we need to show the particular forms of unclarity to which one must succumb if one is to ‘entertain’ such ‘theses’, elucidating the confusions that inform both the ascription of those ‘properties’ and their denial.
For example, I have argued elsewhere41 that the formulation of the first ‘regress’ argument rests ultimately on an equivocation over what we mean when we talk of ‘representations’ - in essence the equivocation that HCT 42-43 (discussed in Sec. 2.3) identifies - and our problem arises ultimately then, not out of an incorrect ontological theory of representation, but out of a lack of clarity about what we are thinking about, switching unwittingly back and forth between thoughts about two different things in the course of our reflections around the single word, ‘representation’.42 When we succumb to such illusions, our fundamental difficulty is that we don’t understand what we are saying; confusions have taken over the terms we use: our talk of ‘inside’ and ‘outside’, ‘transcendence’ and ‘embeddedness’, ‘engagement’ and ‘disengagement’ is informed by certain crude, confusing and inconsistent metaphors or models.
My suspicion is that, for Heidegger, our philosophical ‘forgetfulness’ – our ‘ontological indifference’ (HCT 216) - is a similar kind of inattention and his phenomenological call for a ‘return to the things themselves’ an attempt to make us aware of how unrecognized, crude metaphors and models – paradigmatically those embodied in the ‘metaphysics of presence’ – are embedded in the philosophical tradition in which we are ‘brought up’ and make problematic that which we already ‘pre-ontologically’ understand. The issues here are, of course, extremely complex, given important twists by Heidegger’s discussion of authenticity and BT’s ‘historical turn’; a further twist would be given by asking the (Platonic?) question of whether ‘the refining of our ontological theories’ might not itself be a form of ‘reminding’, whether that might not be the kind of knowledge that the ‘philosophically enlightened’ possess. That strikes me as a good question, and as another that Heidegger himself wants to raise. To hope to settle it, and the question of how ‘anti-therapeutic’ Heidegger’s - or indeed Dreyfus’ - ontology/phenomenology is, would require clarity about what we mean by ‘phenomenology’ and ‘ontology’, ‘therapy’ and ‘theory’, ‘reminding’ and ‘forgetting’.
As I indicated at the start of the paper, my sense is that Dreyfus is trying to pull us in the right direction; but seeing how may require broader reflection on what we think philosophical confusion and philosophical illumination are; that might usefully involve the broader comparison of Heideggerian and Wittgensteinian themes that I just sketched, a comparison broader than a mapping of Dreyfus’ proposals on to proposals expressed in the rule-following literature might be expected to entail. Dreyfus is surely right when he modestly declares that ‘[t]he time is ripe to follow McDowell and others in putting aside the outmoded opposition between analytic and continental philosophy’ (OMM 61); I say ‘modestly’ because no one has a better claim than Dreyfus to have led philosophers to recognize how outmoded this opposition is. What I would add here is that this recognition may bring us to examine distinctions such as that between ‘doing therapy’ and ‘doing phenomenology’. Our sense that these activities are importantly different and that we understand what that difference is may be another legacy of our philosophical upbringing that we need to recognize and then assess.43
University of Southampton
BITW - H. L. Dreyfus, Being-in-the-World. Cambridge, Mass.: The MIT Press, 1991.
BT - M. Heidegger, Being and Time. Trans. J. Macquarrie and E. Robinson, Oxford: Blackwell, 1962. 
CAMIEI - H. L. Dreyfus, ‘Could Anything be More Intelligible than Everyday Intelligibility? Reinterpreting Division I of Being and Time in the light of Division II’’, in J.E.Faulconer and M.A.Wrathall (eds.) Appropriating Heidegger, Cambridge: Cambridge University Press, 2000.
CTT - H. L. Dreyfus and C. Spinosa, ‘Coping with Things-in-Themselves: A Practice-based Phenomenological Argument for Realism’, Inquiry 42 (1999), pp. 49-78.
FCM – The Fundamental Concepts of Metaphysics. Trans. W. McNeill and N. Walker, Indianapolis: Indiana University Press, 1995. [1929-30]
HC - H. L. Dreyfus, 'Heidegger’s Critique of the Husserl/Searle Account of Intentionality’, Social Research 60 (1993), pp. 17-38.
HCT - M. Heidegger, History of the Concept of Time. Trans. T. Kisiel, Indianapolis: Indiana University Press, 1985. 
HH - H. L. Dreyfus, 'Holism and Hermeneutics', Review of Metaphysics 34 (1980), pp. 3-23.
II – H. L. Dreyfus, ‘Introduction I: Todes’ Account of Nonconceptual Perceptual Knowledge and its Relation to Thought’, in S. Todes, Body and World, Cambridge, Mass.: The MIT Press, 2001.
MFL – M. Heidegger, The Metaphysical Foundations of Logic. Trans. M. Heim, Indianapolis: Indiana University Press, 1992. 
MOM – H. L. Dreyfus and S. E. Dreyfus, Mind over Machine. New York: Free Press, 1986.
MPC – H. L. Dreyfus, ‘A Merleau-Pontyian Critique of Husserl’s and Searle’s Representationalist Accounts of Action’, Proceedings of the Aristotelian Society 100 (2000), pp. 287-302.
OMM - H. L. Dreyfus, ‘Overcoming the Myth of the Mental: How Philosophers Can Profit from the Phenomenology of Everyday Expertise’, Proceedings and Addresses of the American Philosophical Association 79 (2005), pp 47-65.
PIA - M. Heidegger, Phenomenological Interpretations of Aristotle. Trans. R. Rojcewicz, Indianapolis: Indiana University Press, 2001. [1921-22]
PPLA – H. L. Dreyfus, ‘The Primacy of Phenomenology over Logical Analysis’, Philosophical Topics 27 (1999), pp. 3-24.
R - H. L. Dreyfus, 'Responses', in Wrathall and Malpas 2000b.
WCSCD – H. L. Dreyfus, What Computers Still Can’t Do. Cambridge, Mass.: The MIT Press, 1997.
References to these works are given by these abbreviations followed by page numbers.
Boghossian, P. (1990) ‘Naturalizing Content’, in G.Rey and B.Loewer (ed.) Meaning in Mind: Essays for Jerry Fodor, Oxford: Blackwell.
Brandom, R. (1994) Making It Explicit. Cambridge, Mass.: Harvard University Press.
Christiansen, C. B. (1997) ‘Heidegger’s Representationalism’, The Review of Metaphysics 51: 77-103.
(1998) 'Getting Heidegger off the West Coast', Inquiry 41: 65-87.
Crowell, S. (2001) ‘Subjectivity: Locating the First-Person in Being and Time’, Inquiry 44: 433-54.
(forthcoming) ‘Conscience and Reason: Heidegger and the Grounds of Intentionality’, in Crowell and Malpas (eds.) Transcendental Heidegger, Palo Alto: Stanford University.
Dahlstrom D. O. (1994) ‘Heidegger’s Method: Philosophical Concepts as Formal Indications’, Review of Metaphysics 47: 775-795.
Davidson, D. (1973-74) ‘On the Very Idea of a Conceptual Scheme’, Proceedings and Addresses of the American Philosophical Association 67: 5-20, reprinted in his Inquiries into Truth and Interpretation, Oxford: Oxford University Press, 1984.
Davies, P. S. (2000) ‘The Nature of Natural Norms: Why Selected Functions are Systematic Capacity Functions’, Nous 34: 85-107.
Horgan, T. and Timmons, M. (unpublished) ‘Making Room for Moral Principles: Reflections on the Phenomenology and Psychology of Moral Judgment’, British Society for Ethical Theory conference, Southampton University, July 2006.
Kant, I. (1961) [1781/1787] Critique of Pure Reason. Trans. N.Kemp Smith, London: Macmillan.
Kim, J. (1996) Philosophy of Mind. Boulder: Westview Press.
McDowell, J. (1987) ‘In Defence of Modesty’, reprinted in his (1998).
(1993) ‘Meaning and Intentionality in Wittgenstein’s Later Philosophy’, reprinted in his (1998).
(1994) Mind and World. Cambridge, Mass.: Harvard University Press.
(1997) ‘Another Plea for Modesty’, reprinted in his (1998).
(1998) Mind, Value and Reality. Cambridge, Mass.: Harvard University Press.
McManus, D. (2000a) ‘Boghossian, Miller and Lewis on Dispositional Theories of Meaning', Mind and Language 15: 393-99.
(2000b) 'Freedom, Grammar and the Given - Mind and World and Wittgenstein', Journal of the British Society for Phenomenology 31: 248-63.
(2005) 'Review of D. Charles and W. Child (eds.) Wittgensteinian Themes, A. Crary and R. Read (eds.) The New Wittgenstein and T. G. McCarthy and S. C. Stidd (eds.) Wittgenstein in America', Mind 114: 129-137.
(2006) The Enchantment of Words: Wittgenstein's Tractatus Logico-Philosophicus. Oxford: Oxford University Press.
(2007) ‘Heidegger, Measurement and the “Intelligibility” of Science’, European Journal of Philosophy 15: 82-105.
Miller, A. (1997) ‘Boghossian on Reductive Dispositionalism About Content: The Case Strengthened’, Mind and Language 12: 1-10.
(1998) Philosophy of Language. London: UCL Press.