Practice poster for INPP 2011

I have never made a poster (here) before but have to for the INPP 2011 meeting. This has already taken all afternoon. Hmph.

Meaning, understanding and explanation

This short pre-conference course before the 2011 INPP conference in Sweden examines the relation of understanding and explanation starting with Jaspers views, then more modern accounts of the distinction, Bolton and Hill’s attempt to draw them together and finally John Campbell’s very recent criticism of the assumption that causation in psychiatry is rationally structured.

Adrian Moore on ineffable knowledge

[Another work in progress. Very rough.]

A different connection between Wittgenstein’s regress argument and tacit knowledge can be drawn from Adrian Moore’s book Points of View by charting his arguments for ineffable knowledge. (Moore himself retains the phrase ‘tacit knowledge’ for something else, something for whose existence he offers no explicit argument.) Whilst in one paper, Moore argues directly that conceptual mastery is a form of ineffable knowledge, in his book, this is placed in a broader context of responding to Wittgenstein. We will first follow that latter route.

Moore suggests that, in his discussion of understanding a rule or grasping a concept, one of Wittgenstein’s targets is the idea that our concepts answer to a ‘super-physical landscape’. Discussing the idea that there is a necessary connection between the concept of aunt and being female, for example, he suggests that Wittgenstein rejects the idea that such concepts ‘were things we just stumbled across, the one an inseparable part of the other’. But instead of charting such an independent super-physical or platonic realm, Wittgenstein’s discussion makes it clear instead that: ‘It is on our own contingent practices that we are focusing’ [Moore 1997: 128].

The connection between meaning or rules and contingency is this. Recall our summary of Wittgenstein’s regress argument and also Kripke’s reconstruction of it. It seems that nothing that can come before the mind’s eye, nor anything that can be put into words, nor any finite examples of past practice can determine a rule or a concept. What then explains our ability to go on in the same way?

Moore quotes, approvingly, a famous passage from Stanley Cavell who says:

That on the whole we... [make, and understand, the same projections of words into further contexts] is a matter of our sharing routes of interest and feeling, modes of response, sense of humour and of significance and of fulfilment, of what is outrageous, of what is similar to what else, what a rebuke, what forgiveness, of when an utterance is an assertion, when an appeal, when an explanation – all the whirl of organism Wittgenstein calls ‘forms of life’ [Cavell 1976: 52 quoted in Moore 1997: 128-9]

It is because of this shared background that we react in similar ways to explanations of rules and concepts and make the same projections of word use into the future. Lacking something like a platonic landscape to chart or a signpost that needs no further interpretation, it is a shared whirl of organism that underpins the conceptual order.

But if so, then this suggests that what seem to be necessary features of our concepts themselves depend on a background of contingencies. The contingencies do not merely concern the fact that, for linguistic historical reasons, the word for aunt is ‘aunt’ and female is ‘female’. Rather, the very idea that aunts are female seems to depend on the whirl of organism. Similarly the truths of mathematic and logic.

On a Wittgensteinian view, not only does 2 + 2 equal 4, but 2 + 2 must equal 4. ‘2 + 2 = 4’ is a rule. And yet – it is a rule only because of our contingent linguistic practices (and not just in the sense that we might have used different sounds or inscriptions to express it). [Moore 1997: 132 italics added]

But that idea seems simply false. As Bernard Williams puts it:

[I]f our talk of numbers has been determined by our decisions, then one result of our decisions is that it must be nonsense to say that anything about a number has been determined by our decisions. [Williams 1982: 95]

One response to this, which Moore considers but rejects, is to attempt distinguish between an empirical and transcendental interpretation of the role of contingency. Whilst within the empirical realm it seems simple false to say that the truths of mathematics or the greenness of grass, for example, depend on our whirl of organism, perhaps there is a way to advance such a claim at a transcendental level, off stage. On this approach, conditionals such as, had our language been different then grass would not have been green, do not express empirical possibilities or point to alternatives which are alternatives for us. But this leaves the thoughts apparently expressed as incoherent, as pure and utter nonsense.

Moore’s own response to the tension is nuanced and lies mainly outside the scope of this book. But one element connects to his claim that conceptual understanding is ineffable. In the face of the tension outlined, we are inclined to ask:

‘But what, ultimately, does somebody’s being an aunt consist in? What does something’s being green consist in?’ We cannot help asking these questions because we cannot help wondering about the basic form of that to which our representations answer. [Moore 1997: 134]

Such questions presuppose that our concepts answer to something: the underlying form of the world, its necessary background logical structure. Given the apparent insight from Wittgenstein’s regress argument that necessary features of our concepts themselves depend on a background of contingencies, answering these questions in their own terms leads inevitably, Moore says, to transcendental idealism but that is mere nonsense. Moore suggests that, instead, the questions should be rejected. But this is not just in order to try to escape the tension. Rather, it is because grasp of concepts or rules does not answer to anything.

Focusing self-consciously on our understanding, we recognize the deep contingencies that sustain it… [But] Our understanding has nothing to answer to. It is part of how we receive the world… If we do achieve such clarity, then what we actually get into focus is an arrangement of interlocking, mutually supporting practices that are grounded in one another’s contingency, a complex knotted structure that might easily have been different. [Moore 1997: 162]

So part of Moore’s response to the regress argument and the tension it seems to set up between necessity and contingency is to deny that conceptual mastery answers to anything. It is not representational knowledge. For that reason it is ineffable.

My understanding of English is a prime example. I would certainly count that as ineffable, even though it includes large tracts of effable knowledge such as... that the word ‘green’ denotes green things.
Understanding, of the sort that I have in mind, has nothing to answer to. Of course, I may think that I know what a particular word in English means and be wrong: I may think that the word ‘rabbit’ denotes hares as well as rabbits. If that is the case, then what I understand is strictly speaking an idiolect distinct from English. But I do still have my understanding... a mode of reception. It is not itself a reception. It includes my knowing how to exercise the concept green, for instance, which in turn includes my knowing what it is for something to be green. But this is not the same as my having an answer to any question. (Still less is it the same as my having an answer to the pseudo-question, ‘what is it for something to be green’.) [Moore 1997: 184]

Elsewhere Moore advances a similar line of argument more directly but which also more clearly connects back to the regress argument and the role of Platonism. (This is the direct argument mentioned above.) As reported in chapter 2, Moore criticises the view that there is any neat semantic marker for a distinction between knowledge-how and knowledge-that and argues that in many cases knowledge-how can be put into words, in accord with Stanley and Williamson’s views. Nevertheless, he also argues that there is a form of knowledge or understanding which is ineffable and which cannot be put into words. Conceptual mastery is such a case.

Consider my knowledge of what it is for an object to be green. On [Stanley and Williamson’s] view this is knowledge, concerning something, that that thing is what it is for an object to be green. But concerning what? A simple reply would be: ‘What it is for an object to be green.’ But what kind of thing is that? If I try to express my knowledge by indicating a green object and saying, ‘This is what it is for an object to be green,’ what can I be referring to by ‘this’? There does not seem to be any good answer. Nothing short of an unacceptable Platonism, it seems to me, can subserve the extension of their account to this case. I do not think that my knowledge of what it is for an object to be green is knowledge that anything is the case. Nor, crucially, do I think that it is effable. [Moore 2003: 177]

In the case of grasp of a concept such as green, Moore argues that the demonstrative approach fails to work. It could only work if something like the whole use of the word were available for demonstrative singling out. But failing platonism of that form, that cannot be the case. He goes on to suggest that where knowledge is ineffable, the attempt to put it into words can only result in nonsense. Further such nonsense really is nonsense. There can be no ‘suggestion that there is nonsense that captures in some more or less obscure way how things are’ [Moore 1997: 201]. Nevertheless, that attempt can produce something that has a role in showing even whilst it cannot say anything. Showing, however, is not merely a shorthand for saying in some more or less obscure way how things are. Rather: ‘To say of some piece of nonsense that it is the result  of attempting to express the inexpressible is something like making an aesthetic evaluation’ [ibid: 202].

We will not follow Moore’s thought further along this trajectory of using nonsense to show something about our grasp of rules but rather examine the stage-setting already in play. Two elements inter-mingle in the claim that grasp of concepts is ineffable. One is the failure of a demonstration to express what one grasps when one grasps the concept of greenness. The other is the diagnosis of this that it is because that concept does not answer to or represent anything independent of it. Understanding the meaning of a word is not an instance of representing something as the case but rather a general precondition of any such representation.

My understanding is knowledge of how to acquire knowledge., then. But it is not itself true representation of how things are. It is not a representation at all...

My understanding is not true, nor true of anything, nor yet true to anything. But the fact that other people communicate with me as they do is reason for my having an understanding that will enable me to make good sense of them (as mine does). More generally, the fact that the world is the way it is is a reason for my having an understanding that will enable me to make good sense of it. And as for what ‘good’ means here: it means, not ‘right’, but... something more like useful. This is not to say that, granted the concepts I have, there is no right or wrong in how I use them to arrive at my interpretations. The point is rather that there is no right or wrong in the concepts I have. [ibid: 185-6]

The claim that grasping a rule does not answer to anything suggests a worry that such understanding cannot be a form of knowledge. Moore considers what he calls ‘the effability argument’ to the effect that answering to something independent is an essential feature of knowledge. Thus for example, strength is a capacity that enables one to do particular things in particular circumstances. But its success conditions are ‘simply the conditions in which the subject is in that state’. It is more or less useful but does not get anything right. Strength is thus not knowledge. By contrast, practical knowledge of how to make an omelette answers to facts about eggs and temperature. Had those been different, a given state of practical knowledge would fail. Thus, the latter is a form of representational knowledge. It is thus effable, according to Moore, because it can be articulated through suitable demonstratives.

How then can something which does not answer to anything independent of it count as knowledge? Moore aims to earn the right to call understanding a concept a form of knowledge by identifying three general marks of or indicators of knowledge: versatility (there is no relevant foreclosing of the possibilities it affords a subject), performance transcendence (evidence for its possession must be more than someone simply ‘bringing something off’) and rationality (it stands in logical relations to other cognitive states). Now whilst states which answer to something independent of them can meet these three conditions so also can ineffable knowledge because, roughly, by being the right sort of precondition of representational knowledge, it can inherit these three marks.

Moore is free to define ‘ineffable’ knowledge the way he wishes: as practical knowledge which is non-representational because it does not answer to anything independent of it. But it is not meant to be a purely stipulative definition. It is tied to a pre-philosophical sense of ‘ineffable’ because the meaning of ‘green’ cannot, he argues, be expressed in words. This in turn is reinforced by something like Wittgenstein’s regress argument. Only if platonism were true could one use a platonic conception of the real, underlying extension of our concepts both to explain to what they answer but also to be the object of a demonstrative to express conceptual mastery in words (as ‘green is that!’). Lacking platonism, the regress argument shows that no other attempt to capture one’s understanding in words will succeed. Any utterance will stand in need of an appropriate interpretation.

Even though this is not Moore’s own account of tacit knowledge it serves to locate a possible response to the regress argument which could be used to support a role for tacit knowledge. Concentrating on the negative moral of the regress argument, it seems that grasp of a rule or the meaning cannot be made explicit because any utterance stands in need of interpretation and that initiates a regress. Equally it cannot consist in any mental talisman akin to a signpost because that will also stand in need of interpretation. Kripke’s response to this accepts that, properly speaking, nothing is grasped in the way originally assumed. Understanding meaning is indeed tacit, though not in any clear sense knowledge, because it is a matter of projection based on not being out of step with a community. (Blackburn suggests that a similar account could be given for an individual.)

Moore’s response also accepts the negative thrust about what can be put into words (although in other cases of practical knowledge, such as omelette making, he happily endorses demonstrative expression). It is ineffable. We might say: tacit. But he nevertheless wishes to preserve the idea that it is knowledge even though this puts under strain the idea that it has content because it does not answer to anything and is thus (unlike omelette making) not representational knowledge.

In the next section we will attempt to outline more directly how an account of tacit knowledge can be drawn from Wittgenstein which shares the general pattern of Moore’s account as summarised here. (

Two claims will be key. One, is that Wittgenstein’s discussion allows more to be expressed than either Kripke or Moore accepts and that helps undercut Moore’s claim that conceptual understanding cannot be expressed. Second, part of the attraction of a substantive ineffable account of tacit knowledge stems from an only partial rejection of platonism. A more thorough going rejection of platonism removes this spurious support. But it will be helpful here to mention a further point of disagreement specifically with Moore’s account.

We said above that Moore mingles the claim that conceptual grasp cannot be expressed with the idea that it is non-representational. He says: ‘This is not to say that, granted the concepts I have, there is no right or wrong in how I use them to arrive at my interpretations. The point is rather that there is no right or wrong in the concepts I have.’ [ibid: 186]. This latter claim reflects a central theme in Wittgenstein’s later work referred to by commentators as the ‘autonomy of grammar’ [eg Hacker **]. It expresses the view that an explanation of conceptual connections in independent terms is impossible. They do not, for example, track independent platonic extensions. Following a rule is not a matter of going over in bolder pencil moves already somehow made.

One can, however, concede that claim whilst insisting that understanding a concept does answer to something: a normative pattern of use which prescribes correct instances which is reflected in the first part of the quotation above: This is not to say that, granted the concepts I have, there is no right or wrong in how I use them. That is the content of the substantive knowledge one has when one knows the meaning of a word or a rule. There is no simple link from the autonomy of grammar as a whole – the fact that it does not represent an underlying platonic structure – to the inexpressibility of what one understands when one understands a concept. One would need a further argument for this and in the next section we will attempt to undermine just such an argument.

Wittgenstein's regress argument and tacit knowledge

Wittgenstein’s regress argument appears on a first reading to support a role for tacit knowledge runs as follows. The examples of correctly determining the direction that a sign-post points, or of the possibility of deviant reactions to explanations of how to continue, correctly, a mathematical series, suggests that everything that can be said still allows for misunderstandings. Since everything that can be made explicit apparently underdetermines the correct understanding, such understanding must instead be based on something unsaid and implicit. It must depend on tacit knowledge of the rule. Hence, on this account, the regress argument is stopped by an appeal to tacit knowledge.

This connection seems also to fit Wittgenstein’s own conclusion:

What this shews is that there is a way of grasping a rule which is not an interpretation, but which is exhibited in what we call “obeying the rule” and “going against it” in actual cases. [Wittgenstein 1953: §202]

Nevertheless, whilst there is something right about that line of thought, there is something misleading about it. That there is something misleading can be by three problems it faces.

The first problem with it as an interpretation of Wittgenstein is that it accepts part of what he criticises: a platonic picture of rules as rails ‘invisibly laid to infinity’ fundamentally distinct from our capacity to articulate them. That picture is easily prompted by the case of the deviant pupil. What that case, and others like it, seems to show is both that any finite set of examples underdetermines a correct understanding of the rule but also that such correct understanding must involve grasp of an extra-human or supernatural pattern. Since no actual human enumeration of the pattern seems enough to determine it, it must be extra-human. Hence the metaphor of rails laid to infinity.

With that picture of the way rules determine correct moves in place, there is a substantial role for tacit knowledge to bridge the gap between what can be made explicit in the sublunary realm and the ideal platonic rule. But if so, it seems that Wittgenstein offers support for a platonic picture he also seems to criticise. (To put this point in the terms used by McDowell: such a picture of tacit knowledge presupposes a rampantly platonic picture of rules [McDowell 1994].)

A second problem concerns the communication of knowledge of rules. Since explanations are insufficient explicitly to fix a unique rule, the tacit grasp of a particular rule cannot be a matter of knowledge even if it were, as a matter of fact, of the rule intended. Nothing could justify the selection from the infinite range of alternative options. But even this is problematic because of a third problem.

The final problem is accounting for the idea that tacit knowledge of a rule or the meaning of a word has some content to be known. As far as what can be made explicit, this approach is in the same position as Kripke’s sceptical account. But it differs from that in attempting to invoke something tacit. The problem, though, is that this means that nothing can be said by way of positive account of what the tacit knowledge amounts to since any attempt will fall prey to the Kripkean objections to explicitness. But if that is the case, what reason is there to think that what remains tacit is a ‘something’ at all? It may justify the label ‘tacit’ but only at the cost of undermining the idea of knowledge.

These three problems all stem from the idea that tacit knowledge is needed to plug a gap between what can be explained, or otherwise made explicit, and the full grasp of a rule which can be understood as a result. Wittgenstein undermines that gap, however, and thus that account of the support his argument gives for tacit knowledge. He suggests, instead, that there is a close connection between what a teacher can express and what a student can grasp in the examples which manifest the teacher’s meaning.

“But do you really explain to the other person what you yourself understand? Don’t you get him to guess the essential thing? You give him examples, – but he has to guess their drift, to guess your intention.” – Every explanation which I can give myself I give to him too. – “He guesses what I intend” would mean: various interpretations of my explanation come to his mind, and he lights on one of them. So in this case he could ask; and I could and should answer him. [ibid: §210]

“But this initial segment of a series obviously admitted of various interpretations (e.g. by means of algebraic expressions) and so you must first have chosen one such interpretation.”–Not at all. A doubt was possible in certain circumstances. But that is not to say that I did doubt, or even could doubt…[ibid: §213]

In §210. the interlocutor expresses the worry that since an explanation fails to determine the rule to be explained, a listener has to guess – from an infinite range of options – what rule was intended. The guess is needed to bridge the gap between what is actually expressed and what was really intended. But Wittgenstein’s response is to equate the what can be explained to another person and what might have been assumed to be epistemically optimal: what a speaker can explain to him or herself.

This equation might be thought – optimistically – to offer in the third person case the happy circumstances of the first person case: what one knows one intends in one’s explanation. But it might also be thought – pessimistically in the context of an inquiry which undermines the efficacy of mental templates to underpin one’s own grasp of a rule – to limit what is available to others to what is available to oneself. Either way, the connection undermines the idea that a guess is necessary to bridge a gap between first and third person cases.

§213 applies the moral of §210 to the explanation of a rule. Whilst some explanations can fail that is not the general case. (Just as in general sign-posts succeed in pointing.) Although Wittgenstein rejects substantive explanations of our grasp of rules, via mental mechanisms, he does not claim that there is a gap between what can be manifested and what must be understood, a gap that has thus to be filled by a tacit element.

Recognising that our understanding can be expressed in examples undermines the gap between the sublunary and the platonic and thus that potential role for tacit knowledge. It also blocks the worry raised above that such a model of the tacit understanding of rules or meanings would put under pressure the idea that there is something to be known, a content grasped. There is a content which can be expressed in examples or ongoing practice.

That might suggest that, on a proper understanding, Wittgenstein’s regress argument offers no support for tacit knowledge. And indeed, an interpretation, perhaps inspired by Kripke, which concentrates on the potential failures of explanations or the lack of efficacy of signposts seems to have things almost exactly the wrong way round. A pointing sign can be a paradigm case of what it is to make a direction explicit not implicit. Some ostensive examples can explicate the meaning of a word.

But there is, nevertheless, a tacit dimension for two reasons.

First, such explanations work for those with eyes to see or ears to hear. It is because we are the kind of subjects we are, with our shared routes of interest, perceptions of salience, feelings of naturalness etc., that we are able to use finite explanations, as we do, to communicate unending rules [Lear 1982: 386]. Because we share what Stanley Cavell calls our ‘whirl of organism’ we can respond to explanations in a way which does not threaten a regress of interpretations.

Wittgenstein’s deviant pupil illustrates what it would be to lack the right background and thus fail to understand or react to explanations by example as we do. Under such hypothetical circumstances, examples would need to be bound together under an interpretation but – again without the right background – there would be no way to encode that interpretation without a vicious regress.

Now it might be tempting to obviate the need for such a background by trying to articulate or encode how understanding depends on one’s interests, saliences and perceptions of naturalness. But any attempt to articulate that would also have to presuppose the background that underpins understanding of any such explanation. If we take full articulation to be an articulation which does not require possession of the shared background, then no full articulation is possible (because of the regress of interpretations). But for those who do share that background, it can be put into words piecemeal.

That is one reason to think of the background of shared routes of interest, perceptions of salience, feelings of naturalness as tacit. Making elements of it explicit is only possible against a background which cannot be simultaneously articulated to someone outside it. An explanatory project which attempted to take nothing for granted would be bound to fail. But on the other hand, articulating the bounds of sense from within is possible. Thus it is a nuanced view of the tacit.

The second main reason to think that understanding a rule is a species of tacit knowledge is as follows. Consider a rule which can be partly codified in an informal statement such as that the digits always follow the pattern: ‘0, 2, 4, 6, 8, 0 etc’ or more fully codified in an explicit mathematical formula or principle. Someone who understands such a rule may understand such a general principle or perhaps a set of related principles using some of them to explain others. They may thus be able to articulate what they understand the rule to be in general and context-independent terms. Nevertheless, even with such a codifiable rule, understanding it cannot be independent of understanding its instances. One needs to know, in Wittgenstein’s phrase, how to go on. It is a form of practical knowledge. Wittgenstein gives an example of someone who grasps a series either with, or without, having a formula in mind:

It is clear that we should not say B had the right to say the words “Now I know how to go on”, just because he thought of the formula – unless experience shewed that there was a connexion between thinking of the formula – saying it, writing it down – and actually continuing the series...
We can also imagine the case where nothing at all occurred in B’s mind except that he suddenly said “Now I know how to go on” – perhaps with a feeling of relief; and that he did in fact go on working out the series without using the formula. And in this case too we should say – in certain circumstances – that he did know how to go on. [ibid: §179]

In addition to this connection to practice, however, there is something implicit in this example. It is connected to particular cases. To understand the rule requires a grasp that this particular number is the next number in the sequence, for example. So grasp of a general concept, even one which can be explained in general terms, implicates a situation-specific sensitivity. But as the regress argument establishes, situation-specific judgements cannot be reduced to or captured in merely general terms.

Because grasp of a rule involves knowing how to go on, it is a form of practical knowledge. But it is, additionally, situation specific. This is what we take to be the best understanding of tacit knowledge.

Travis on determination / rule following

A couple of years ago, I took some time off work to read Charles Travis’ Thought’s Footing and found it a difficult but rewarding read [Travis 2006]. I didn’t take any notes about how it works so I’ve had to go back to recall Travis’ version of the rule following considerations which has a number of interesting features. The main chapter is chapter 4 on determination although a later section (pp189-93) helps to shed light on the approach here.

In chapter 4, Travis takes as a central clue a passage later than the conventional rule following passages (I would say [Wittgenstein 1953 §§139-239]):

In §459 Wittgenstein notes,
We say ‘The order orders this – ‘ and do it; but also ‘The order orders this: I am to...’. We translate it at one time into a proposition, another into a demonstration, and at another into action.
It is the transition from order to action that matters here. For a proposition, the transition is from it, or understanding it bears, to particular facts as to things being as they are (for the various ways that might be) being, or not being, things being as the proposition says. What might effect such a transition? [Travis 2006: 122-3]

I think that, by taking this as his clue, Travis stresses two things. First, there is a transition or translation from the order to what we do with it (to what I would have said ‘satisfies’ it). That suggests a gap to be crossed in some way. It is more obvious when what has to be connected is an order and an action as they seem to be of different natures or ontological orders. (This reminds me of the other crucial passages nearby in the 400s which present a solution to the connection of thought and world.) But it is true of statements as well.

On occasion we see given such translations as right. For us to mean or understand words in some given way (as we do on an occasion) is for us to be prepared to see some such translations as right, others as wrong. [ibid: 129]

The second stressed aspect is that the translation is to something particular or individual. ‘In the imagery of §459, we make translations from proposition to action (more generally, to treatments of particular cases)’ [ibid: 129].

Travis then presents two rival accounts of this:

We...have two contrasting models of meaning and understanding. They differ importantly as to what it is to be prepared to make given translations of the sort of which §459 speaks. ...[ibid: 129]

The first account assumes that to mean words in some particular way is to be related to an explicit abstract or disembodied representation.

On that expansion [of what are in themselves, pre-philosophically, innocent banalities about meaning], an explicit specification of an understanding someone means words to bear is always possible in principle. To give it would be, inter alia, to produce a representation of things being as meant... That representation would be explicit, which is to say that it would have a very special property: unlike our ordinary representations, it would not admit of understandings (of things being as thus represented). [ibid: 121]

Travis puts initial pressure on this account by stressing that, in general, our words ‘admit of understandings’. So giving the example of Sid saying that the shoes are under the bed, Travis argues that this might mean any of many options. They might have to be completely under the bed. Or having their ends protruding might be allowed. Or, rather less likely for us or for Sid, ‘in a plumb line with the bed, but three floors down’ [ibid: 120]. So a condition of adequacy of the first account is that it singles out just one such interpretation.

One the first model... for Pia to understand the sign as she does is for her to relate in a particular way to a very special sort of (disembodied) representation. That representation requires the particular transitions it does independent of anything Pia is prepared to recognise, and independent of any understanding we may happen to share as to what it requires...[ibid: 129]
Such a representation would thus be a synopsis of the solutions to indefinitely many problems; a synopsis from which all those solutions are recoverable. One would not need Pia’s form of worldliness, or any other particular form, to manage the recovery. It might be a task for an idiot-savant. [ibid: 124]

Although initial pressure is put on this model, Travis goes on to introduce a rival, second account before assembing his key Wittgensteinian objections to the first. Those objections are that nothing could do the work that the disembodied representation is supposed to do. I’m less interested in those (because they are familiarly Wittgensteinian) than with the novel set-up. So to the second model:

On the second model... for Pia to understand the sign as she does is for her to be positioned to see the solutions to an indefinite range of novel problems as to which translations it requires – problems as to whether it requires doing this to the door now. For her to be so positioned is for her to share a competence we have to see how to take the sign, except where, for one or another special reason, she deviates in her understanding from that. [ibid: 129]

On this second account, to mean something by one’s words depends on the contingencies of the situation, or the ‘occasions’ in Travis’ key word, on the practical competences of speaker and hearer (‘Pia’s worldliness places her to deal with an indefinite range of potentially relevant, and sometimes unexpected, considerations’ [ibid: 124]) and the relation between the speaker and facts about what it would be reasonable to be taken to mean in the situation. That third point is one of the ways in which Travis is a kind of communitarian Wittgensteinian.

Travis suggests one is pushed towards the first, Fregean and disembodied representation by stressing the question: What is it that Pia thus meant? Or which way is the way she meant it? But the second account develops from ‘rotating “the axis of reference of our examination... about the fixed point of our real needs” [§108]’ [ibid: 128].

This ‘rotation of our axis of reference’ brings with it two crucial new features. One is that there is no longer any supposition of a unique right answer to the question in which way Pia meant her words. If a given candidate for a way she meant them merits, by the above sorts of considerations, counting as a way she did in fact mean them, that does not exclude any other substantively different candidates from meriting the same answer once they do come into consideration. The other is that, given the role of what is most reasonable is to play in the truth of an answer for a given candidate, there is room, or more, for personal meaning to be an occasion-sensitive affair. Whether Sid is most reasonably classified as one who meant ‘gold watch’ in such-and-such given way is entirely likely  to depend on the circumstances in which, or purposes for which, that classifying is to be done. [ibid: 128]

So by p129 Travis has outlined two rival accounts of the way in which words are meant and understood, stressed the occasion sensitivity of meaning, strongly hinted at where the first will have difficulty and suggested something of the way the second will cope. All this comes in part by noticing that there is a translation between an order and an action and between a statement and a state of affairs. (The connection between these was already there in §459.)

The first account will not be able to bridge the gap that the need for translation reveals. That sounds familiar from any account of §§139-239. But Travis brings it out in the context of a neo-Fregean / neo-Russellian distinction about thought.

Sid tells Pia that her shoes are under the bed. Pia understands Sid’s words in a certain way. In particular, she takes him to speak on a certain understanding of shoes being under a bed. Now she enters the room. She encounters things being relevantly as they are (supposing how they are three floors down not to matter). She learns something, perhaps enough, of the conditions (circumstances) which then obtain. Three understandings of Sid’s words now become available. There is an understanding of them on which things being that way just is their being as Sid said. There is one on which it just is not. And there may be a third on which that much leaves the issue undecided. Other than the mentioned differences, these understandings may be very much alike. It may be that just one of them is the one that Pia’s understanding of Sid’s words requires. In that sense, just one of them is part of that understanding...

There is now a point about the availability of these three understandings. I start with another instance of just the same point. But for a certain successful enterprise on the part of Frege’s parents, there would have been no singular thoughts about him. Before 1848, there were none that anyone could think...

Some thoughts are only available to us given suitable acquaintance with our environment. And so it is with those understandings of Sid’s words I just mentioned. Pia’s shoes are positioned as they are with respect to the bed. There is then this understanding of Sid’s words: what they say is such that things being that way is things being as they said. Someone may thus understand them. One may only so understand them if one is suitably acquainted with things being as they then were. It is to things so being that one must be responding in having that understanding. An understanding thus unavailable to someone before a given time I will call novel (for that person at that time), and an understanding available anyway, even when that other one was not (or if it were not) prior relative to that novel one. [ibid: 129-30]

With this latter distinction in play, the gap covered by translation (of, for example, an order into an action) is between prior understanding and novel understanding. As the comparison with singular thoughts about Frege (and his account of such thoughts earlier in the book is excellent, by the way) makes plain, one is simply in no position to have the novel or situation-dependent thought in advance. That thought simply cannot come to mind, or be entertained, when one is framing a prior thought or understanding. 

Elsewhere he gives this example:
A man in Ulan Bator is now standing before his yurt, sipping tea. (Make it ten o’clock his time.) I cannot think a thought, of him, that he is doing that—a thought which presents him as the one who must be some way for the thought to be true, and sipping tea before his yurt at ten as what he must be doing. I cannot do this, since I neither know, nor know of, anyone in Ulan Bator (though I am sure some people live there). I can, to be sure, think that everyone in Ulan Bator stands before his yurt at 10 and sips tea. What I thus think will be false if this man does not do that. The thought I thus think has a certain kind of generality which allows it to be true, or false, in this way. But as Frege points out (1914: 108-109, different example), that man falsifes my statement only given that he is in Ulan Bator—in present terms, only given that his being as he is participates in the instancing relation with that way for a thing to be. And it is just this last that I am not in a position to think—a corollary of not being able to think of him at all. Thinking a thought which is false, given his being as he is, is not the same thing as thinking a thought of him. [Travis 2011: 310-11]

In the context of Wittgenstein’s discussion of the expectation of an explosion and the explosion which satisfies, one cannot have had a situation-dependent thought about that explosion (of the form that! explosion) in advance. Still, there has to be some connection between prior and novel understandings.

[I]f an understanding of Sid’s words is one on which things being as they are... is things being... as Sid says, then that understanding should contain some understanding which requires the corresponding novel understanding(s) relative to it. It should be an understanding relative to which understanding Sid’s words as speaking of what is instanced by this... is a novel understanding, but, moreover, one which that prior understanding mandates... The issue it raises is how prior understanding can require novel ones. [Travis 2006: 130-1]
[W]ords can be true (or false) only where they bear a prior understanding that requires the right novel one. If they are made true by things being as they are, that is by virtue of an understanding available anyway... [131]

So this way of setting up the problem stresses in terms one doesn’t find in the Investigations just what is the gap between order or intention or understanding of a rule and action. There are two different kinds of thought in play (prior and novel). So how can the one determine the other? But unless they do, there can be no such thing as truth and falsity.

There follows a discussion of why a disembodied representation could not do the work. This culminates in this passage:
Where Sid meant that things were a certain way, acquaintance with conditions in the bedroom makes three thoughts (so understandings) available which otherwise are not: that things being thus is their being that way; that it is their not being that way; and that it, so far, leaves the matter undecided. Each corresponds to a different way Sid might have meant his words. Suppose these understandings to be novel for Sid at the time he spoke. On the first model, for Sid to have meant his words as he did is for him to relate to a certain disembodied representation. Suppose that things being as they are in the bedroom is their being as Sid meant.  Then, on the model, for him to have meant his words as he did is for him then to have related to a representation of things as F, where but one of these three novel understandings is an admissable understanding of things being F (namely, the first one). What being F is rules the others out tout court. Suppose there is such a thing as being F. Then, the present point is, there is another – call it F* - which agrees with being F in point of all the understandings available to Sid prior to these novel ones, particularly, all those which are, as the above, translations from ‘order to action’ (from things being such-and-such way to things being thus), but which disagrees with F in re the novel understandings (the ones Sid could not have had, or entertained, at time of speaking). Things being as they are would be their being F, where that is a thought then available just w[h?]ere it would be their being F*; a then available understanding would be an admissable understanding of being F just where it was also an admissable understanding of being F*. What, then, could there be in Sid’s being as he then was which makes it a (disembodied) representation of things as F, and not one of things as F* that he related to in meaning what he said as he did? [ibid: 137]

I must say I find myself reading back from how I already think of the regress of interpretations argument to this dense passage. Thus, I take it that the gap between prior and novel thoughts permits a variety of bent interpretations which agree up to just this! novel, situation-dependent, thought but differ at just this point. The disembodied representation is something that can be a prior understanding and yet, somehow, fixes, by itself, just this! novel thought. But how could that be? Any concrete representation will stand there like a signpost and admit of a variety of interpretations. (I am not sure that there is textual evidence to suggest that Travis thinks this but there is a dilemma here. The embodied representation must either fix the novel understanding as it were mechanically or just contain the novel understanding. Neither works. The former for familiar Wittgensteinian reasons. The latter trivially once one buys the neo-Fregean distinction.)

(There is another element of the criticism to which I will return. On the criticised approach, the prior understanding somehow names the circumstances such that the later encounter - in the novel understanding - is not really novel but a kind of re-cognition of those circumstances. This is connected to the idea that properties are ingredients in situations that can be re-encountered as Frege the man was and involves a blurring of naming and predication. It is connected to the way Travis unpacks §429: If I say falsely that something is red, then nonetheless, for all that, it is not red. But I will try to summarise this thought some other day.)

One other difference from a conventional set up. All the talk of shoes under the bed is akin to all the examples of deviant pupils and does similar work. If those are to be ruled out by a proper account of understanding, that presents a challenge for either Fregeanism (Travis) or platonism (more familiarly). A deviant pupil does appear in discussion of §185 on p133 and Travis points out the luck that we can find a pattern in the way that he mistakenly hears the instruction: we might have made nothing of it. But in the main occasion-sensitivity stands to Wittgenstein’s pupil as the empirical does to the transcendental. Travis thinks that occasion-sensitivity is real and rampant and thus plays a more than merely formal role in shaping his account of meaning.

Frege, G. (1914) ‘Logik in der Mathematik’ in his Nachgelassene Schriffen: 219-270
Travis, C. (2006) Thought’s Footing, Oxford: Oxford University Press
Travis, C. (2011) Objectivity and the Parochial, Oxford: Oxford University Press
Wittgenstein, L. (1953) Philosophical Investigations, Oxford: Blackwell

See this entry on ‘A sense of occasion’, this on ‘Reason’s reach’, this on ‘The twilight of empiricism’, and this on the discussion of rule following in Thought’s Footing.

Housekeeping

I think that I am now on top of all writing except for finishing my book on tacit knowledge to which I will now devote myself exclusively.

Two draft chapters on explicit and tacit knowledge are now with an editor of a book on endovascular surgical techniques.

A review for Current Opinion in Psychiatry on ‘Recent developments for naturalising the mind’ is now in production.

A chapter on ‘Clinical judgement and tacit knowledge’ for Bill Fulford’s Oxford Handbook of Philosophy of Psychiatry is with him for the review process.

My paper from Paris last summer on ‘Radical liberal values based practice’ is in production with the Journal of Evaluation in Clinical Practice.

After some problems in the review process the paper ‘Delusional atmosphere, the everyday uncanny and the limits of secondary sense’ was accepted for Emotion Review although sadly Lois described it as a ‘pointless paper’.

After rather less difficulty ‘The recovery model, values and narrative understanding’ was accepted for the OUP collection The Recovery of People with Mental Illness. For some reason, however, whilst fiddling with it in the review process, I realised that it is overly burdened and hedged about by qualifications. I suspect I wrote it too slowly.

My contribution to the Association for the Advancement of the Philosophy of Psychiatry Bulletin, ‘Why taxonomise anti-psychiatry?’, rashly written on the day of the call for papers, is no doubt waiting for other contributions to come in with a little more thought.

There is no news at all about ‘Why teach the philosophy of mental health?’ submitted ages ago to the Journal of Mental Health Training Education and Practice.

Finally, ‘Capacity, mental mechanisms and unwise decisions’ is out in Philosophy Psychiatry and Psychology18: 127-32

Quirkier things. Paul Hoff told me at our UCLan INPP meeting at Manchester a year ago that a paper I'd written five years ago (‘Facts, values and meanings’ (invited speaker) Zürcher Symposium für Klinische Psychiatrie Zurich March 2006) will actually be published in a translated collection in Swiss.

I hear that my paper ‘Mind and World as transcendental anthropology?’ (invited speaker) McDowell, l’esprit et le monde, McDowell, Mind and World Amiens October 2007 will also come out in a collection with a reply from McDowell. Given that - as was pointed out on this blog by Paul Witcombe - it contained a glaring error that I never changed, I am a bit embarrassed by this.

Having cleared the decks, I have been invited by Yvonne Bonner to submit a paper the oldest Italian psychiatric review RSF (Rivista Sperimentale di Freniatria). She says with disarming candour: "I do hope you will accept as I do believe Italian readers would probably discover an interesting philosopher that they haven't come across yet." But would they even probably discover an interesting philosopher?

Explicit knowledge and its transmission

Another work in progress chapter.

Explicit knowledge and its transmission

The goal of knowledge, justification and the avoidance of luck

Medical science aims at truth. Why? Here is one reason. Medicine aims not just to understand the nature of the world but to intervene in it. For a given therapeutic aim and a belief – or set of beliefs – about how to bring it about, achieving that aim will in general depend on the truth of the belief (or set). Success given merely false beliefs would require the compensation of additional good luck. In general, truth unlike falsity is conducive to success. But if true beliefs explain therapeutic success, why aim not just for true beliefs but, more than that, knowledge? What is the value of knowledge?

In this chapter I outline some key issues in the analysis of knowledge which start from this question. One key problem is this. In the main knowledge is conveyed by testimony – through enculturation and upbringing, and explicit teaching and learning – and yet intuitive approaches to the analysis of knowledge which aim to capture personal responsibility for it would make this impossible.

One answer to the question raised comes from an intuition about the nature of knowledge. Knowledge can be undermined by luck. Suppose that a tennis fan forms an irrational conviction about who will be the winner of next year’s Wimbledon tournament which turns out, by chance, to be correct. His true belief, in such a case, is not knowledge. Whilst he may claim that he knew who would win, if the facts are as just described (that it was merely an irrational conviction), he did not know. Whatever has to be added to true belief to yield knowledge should address this widely shared intuition.

A longstanding traditional view holds that the addition is justification. Knowledge is justified true belief. The tennis fan lacks knowledge because he lacks a suitable justification for his belief, a justification that rules out the need for luck for the belief to turn out to be true.

Justification also plays a second role. It provides a means of aiming at true beliefs. It is one thing to worry that one’s beliefs about the efficacy of rival surgical techniques may not be correct, but quite another to work out how to avoid error. It would not be helpful to be told to replace any false beliefs with true beliefs. To hold a belief is to hold it to be true. (To hold that something is not the case is not to believe it.) Thus beliefs which are, in fact, false are not be transparently so to someone who holds them. But the advice that one should ensure that one holds only beliefs that are justified is helpful. And by aiming at justified beliefs one should in general succeed in reaching true beliefs since justification is, in general, conducive of truth.

This approach broad strategy is exemplified in the way Descartes approaches his epistemological project: the method of doubt.

Several years have now elapsed since I first became aware that I had accepted, even from my youth, many false opinions for true, and that consequently what I afterward based on such principles was highly doubtful; and from that time I was convinced of the necessity of undertaking once in my life to rid myself of all the opinions I had adopted, and of commencing anew the work of building from the foundation, if I desired to establish a firm and abiding superstructure in the sciences...

But, to this end, it will not be necessary for me to show that the whole of these are false--a point, perhaps, which I shall never reach; but as even now my reason convinces me that I ought not the less carefully to withhold belief from what is not entirely certain and indubitable, than from what is manifestly false, it will be sufficient to justify the rejection of the whole if I shall find in each some ground for doubt... [Descartes **: **]

Realising that some of his beliefs are probably false, but that he will not be able to determine which directly (for the reason above), Descartes decides to reject all those beliefs whose justifications are deficient. Since those cannot be knowledge this fits the role of justification outlined so far. But there are two further features of Descartes’ project which are attractive but misleading.

First, he takes his target to be anything that is not entirely certain and indubitable. Second, the justifications for beliefs he considers are those fully available to him (in his study) a view now called epistemic internalism. Taken together these drive him to conclude that the only thing that resists his method of doubt is expressed in the cogito: he thinks, therefore he must exist. Cogito ergo sum.

When reflecting on knowledge in the context of a philosophical discussion, even medical students tend to adopt both of these assumptions. They concede that they do not really know anything of which they are not certain and that certainty is a matter of, for what, they can personally vouch. This, however, rules out all but the most basic medical knowledge and threatens to undermine the point of much medical education such as lectures and textbooks. Whatever knowledge is, it can be shared and taught. In short, one can gain knowledge second hand through testimony: any form of transmission of knowledge through the reports of others contrasting with perception, reasoning, whether inductive or deductive, and memory. This obvious places constraints on how best to understand knowledge.

There is a second reason to be wary of the internalist version of the traditional picture. In the 1960s Edmund Gettier showed how to construct counter examples in which a subject had a true belief and a justification for it but the justification only worked through the intervention of luck. Smith and Jones have applied for a job [Gettier 1963]. Smith has good reason to hold that Jones will get the job and has ten coins in his pocket from which he concludes that the successful applicant has ten coins in his pocket. By chance, he himself gets the job and has ten coins in his pocket. Did he know that the successful applicant has ten coins in his pocket? No. A justification may fail to rule out the need for luck which undermines knowledge.

Diagnosis, testimony and internalism

Why does testimony not fit the internalist assumption made by Descartes (and the medical students) that justification depends only on factors which are reflectively accessible to the knower? The problem is that an individual cannot do enough to vouch for the status of knowledge transferred. For this to be possible, testimony would have to be justified in terms of, perhaps by being reduced to, processes which do fit an internalist analysis. But this is not possible.

Suppose, for example, that internalist accounts could be given of perception and induction (neither of which seems plausible). An internalist account of testimony would then be possible if testimony could be reduced to a combination of perception (of others, of their utterances, etc.) together with inductions from their previous reliability. David Hume attempted to outline just such a defence of testimony in his Enquiries Concerning Human Understanding [Hume 1975: 109-116]. But as the philosopher Tony Coady convincingly argues, no such attempt can work [Coady 1992: 79-100]. I will mention two criticisms of Hume which suggest the principled difficulty of any such attempt.

The first objection is that Hume’s defence depends on establishing inductive correlations between past instances of testimony and the truth of beliefs successfully communicated. But there is, in fact, much less evidence available to individuals than Hume supposes. Peter Lipton puts the point thus:

Hume’s discussion systematically hides the fact that our evidential base is far too slender to underwrite in this way even a small fraction of the testimony we rightly accept. Perhaps the main device Hume uses here is to appeal to the correlations we have observed to obtain between various types of testimony and the facts. This appeal to communal observation closes a vicious circle, since you can only in general know what others have observed on the basis of their testimony. The only evidence that you can legitimately appeal to consists of correlations between what you yourself have heard and what you yourself have seen, and this provides far less evidence than would be required to support inductively the wide range and variety of generalisations that would cover all the unchecked testimony you actually accept. [Lipton 1998: 15]

A second line of objection is that the observations that an individual might make are not themselves free from past testimony and thus cannot be used to justify it independently. The quickest argument for this is that observations are framed in language and language is taught through testimony.

Thus it seems that there is little hope of offering a non-question-begging justification of testimony, in internalist terms. Instead, whatever local checks might be carried out, we have to take the general reliability of testimony as a whole on trust. Successfully learning something by testimony is simply hearing in another’s utterance that something is the case. Whilst the ignorance or insincerity of a teacher or witness undermines such transmission of knowledge, one does not, and in general cannot, first ensure their knowledge and sincerity in non-question-begging, non-testimony-based terms.

The moral of this seems to be that one should reject internalism and argue that either justification, or something else playing that role, depends on factors that are external to a person.

Externalism

The most familiar version of epistemic externalism is reliabilism, according to which knowledge is true belief arrived at by a reliable method. Crucially, the knower need not know that the method by which she arrives at her beliefs is reliable (or anything about the method) as long as it actually is. And there is nothing akin to a justification for the knower (for which reason it is taken to be a rival to, not a variant of, the traditional justified true belief analysis). It goes some way to address the worry that individuals do not know enough of the pedigree of what they are usually taken to know. If a teacher is reliable, it is possible to gat knowledge from them second hand. But reliabilism faces the challenge of specifying just how reliable a method must be to deliver knowledge. Anything less than 100% reliable reintroduces the need for some luck for a resultant belief actually to be true. But restricting methods to 100% reliable threatens to make knowledge impossible.

A more promising recent version of externalism has recently been articulatedby John McDowell. Like traditional internalists he does think that reasons have an important role in knowledge. Internalists, however, construe justification as something under the complete control of a subject without need for luck, although Gettier cases suggest that luck is needed to promote a justified belief to a truth. McDowell rejects the view that:

reason must be credited with a province within which it has absolute control over the acceptability of positions achievable by its exercise, without laying itself open to risk from an unkind world. [McDowell 1998: 442]

On his account, even to have a justification - a ‘standing in the space of reasons’ - requires some luck. One is lucky not to be looking at an unrepresentative physiology or in distorting lighting conditions or not to be in the presence of a capricious lecturer. But given that initial luck, no further luck is required to transform one’s situation-based justification into true belief and hence knowledge. This idea is clarified by three further points:

1. A comparison with practical reason.
2. An anti-intellectual view of knowledge.
3. An anti-reductionist view of the kind of philosophical insight needed in epistemology.

Firstly, McDowell’s proposal about our authority over our justification can be compared with a view of practical reasoning which already seems natural:

The concept of what one does, understood as applying to one’s interventions in the objective world, cannot mark out a sphere within which one has total control, immune to luck. It is only if we recoil from this into a fantasy of a sphere within which one’s control is total that it can seem to follow that what one genuinely achieves is less than one’s interventions in the objective world. [ibid: 406 fn 16]

Although our actions are the result of an interplay between, on the one hand, our beliefs and desires and, on the other, contingent or lucky features of the world which shape the possibilities for action, this is not taken in general to undermine our responsibility for our actions. Likewise our epistemic status.

Secondly, McDowell combines his view that having an epistemological standing depends at least in part on a relation to the world with an anti-intellectualist view of knowledge. It can be brought out by considering his attitude to a contrast between what he terms ‘mediated’ and ‘unmediated epistemic standings’. If it existed, an unmediated standing would be one which was foundational, or an ‘absolute starting point’ of the sort hoped for by the Logical Positivists. That, sadly, is mere a myth and leaves only mediated standings, by contrast, which stand in rational relations to each other. Even perception is a ‘mediated standing’ because observation is theory dependent.

On one view of them, a mediated standing in the space of reasons is one for which an argument can be given, by the knower, from premises which do not beg any questions to a claim to knowledge. The argument might thus move from premises about how things look – their mere appearances about which which one supposedly cannot be mistake – to a conclusion to the effect that the subject can see that things really are thus and so. On this view, the space of reasons in general consists in the explicit arguments subjects can offer for their beliefs.

McDowell does not deny that there are some arguments relevant to one’s epistemological status. If a subject sees (or has seen; or hears; or has heard) that something is the case, then it must be the case. Furthermore, to be a subject capable of knowledge at all, he or she must be sensitive to the power of reasons. But he does reject the idea that the epistemic position of seeing or hearing that something is the case can be reduced to or constructed out of something more basic via an argument that the subject of the position could provide.

In the case of testimony, it is particularly clear that a hearer is not in general in a position to rule out possible sources of error in what a speaker says or other factors that would imply that a speaker does not know what he or she affirms. Thus, in general, a hearer cannot provide an argument from what he or she hears said to its truth. Nevertheless when unbiased by mistaken assumptions about the nature of knowledge, it seems clear that testimony can indeed provide knowledge. (Likewise one may not be able to frame an argument from the nature of the lighting conditions to the reliability of appearances, but, in good lighting, seeing how things look can furnish one with knowledge.)

McDowell’s response to this tension is to suggest that the attempt to give a reductionist account of having justification is mistaken.

The idea is, then, that one’s epistemic standing with respect to what one comes to know by testimony consists in one’s, say, having heard from one’s informant that that is how things are; not in the compellingness of an argument to the conclusion that that is how things are from the content of a lesser informational state. [ibid: 436]

So – and this is the third point flagged above – the tenor of the analysis runs in the opposite direction to what is normal. Rather than attempting to build up an analysis of knowledge by asking what more primitive concepts need to be combined to yield it, he takes knowledge to be a basic concept and then explores its relation to other concepts such as reason, justification and truth.

Reasons and responsibility

Knowledge can be undermined by luck. Even true beliefs based on the wrong or no reasons do not amount to luck. And hence there has been a focus on the reliable methods of acquiring beliefs such as looking to justificatory reasons. But on an internalist view the way we acquire most of our knowledge of the world – second hand through explicit teaching and implicit enculturation – is impossible. A more nuanced account accepts that sensitivity to reasons plays a role in knowledge but that one can simply be in a position to see or hear what is the case.

This, however, does not mean that a knower can simply be passive in the face of experience. There may be reason to believe, for example, that lighting conditions are misleading or that a RCT has been discredited or is from a unreliable laboratory or that a once reputable textbook is now out of date. Even though one can acquire explicit knowledge second hand without being able to offer an argument for the reliability of the source, one still needs to exercise epistemic responsibility in keeping up to date with changing evidence.

Coady, A. (1992) Testimony: a philosophical study, Oxford: Clarendon Press.
Descartes, R. (1986) Meditations on first philosophy, Cambridge:Cambridge University Press.
Gettier, E.L. (1963) ‘Is justified true belief knowledge?’ Analysis 23: 121-123
Hume, D. (1975) Enquiries concerning the human understanding and concerning the principles of morals, Oxford: Clarendon Press.
Lipton, P. (1998) ‘The epistemology of testimony’ Studies in History and Philosophy of Science 29: 1–31.
McDowell, J. (1998) Meaning knowledge and reality, Cambridge, MA: Harvard University Press.

Tacit knowledge and practical and recognitional clinical skills

A work in progress chapter for a clinical book.

Tacit knowledge and practical and recognitional clinical skills

Whilst the emphasis in thinking of the content of natural sciences is generally on impersonal explicit knowledge, there is also an important ‘tacit’ element. This is especially true of sciences, or science-based disciplines, which contain a practical aspect. Surgery is a clear example of this. Nevertheless, exactly what tacit knowledge comprises is still a matter of debate. How can something count as knowledge but remain tacit? Clues include its connection to know-how or ability, the fact in some sense it cannot be put into words but that it depends on practical contexts.

This chapter will start by looking to a pair of arguments for tacit knowledge by the chemist turned philosopher of science Michael Polanyi who first drew attention to in the middle of the C20 in two books: Personal Knowledge and The Tacit Dimension.

But first, what does Polanyi mean by ‘tacit’ knowledge? A clear statement runs thus:

I shall reconsider human knowledge by starting from the fact that we can know more than we can tell. This fact seems obvious enough; but it is not easy to say exactly what it means. Take an example. We know a person’s face, and can recognize it among a thousand, indeed among a million. Yet we usually cannot tell how we recognize a face we know. So most of this knowledge cannot be put into words. [Polanyi 1967: 4]

The suggestion is that tacit knowledge is tacit because it is ‘more than we can tell’. We cannot tell how we know things that we know tacitly. But what argument does he give for this? What are the limits on what can be said still leaving something that can be known?

In Personal Knowledge, Polanyi’s strategy is to examine how what can be said or, more broadly, articulated both leaves room for and depends on something outside what can be articulated. There are two key arguments of relevance to this chapter. One depends on limits on the kind of representation available to summarise explicit knowledge in science, thus indicating a need for tacit knowledge. The other depends on an analysis of what is involved in recognition (an argument which promises to impact on diagnostic judgement), which also connects to Polanyi’s views of how linguistic representation in general is possible. I will suggest that this latter argument is the fundamental argument but start with the former.

To examine the limits of scientific representation, Polanyi considers the understanding that a skilled surgeon has of the spatial configuration and orientation of organs in the body. He argues that this cannot be captured in a representation.

The major difficulty in the understanding, and hence in the teaching of anatomy, arises in respect to the intricate three-dimensional network of organs closely packed inside the body, of which no diagram can give an adequate representation. Even dissection, which lays bare a region and its organs by removing the parts overlaying it, does not demonstrate more than one aspect of that region. It is left to the imagination to reconstruct from such experience the three-dimensional picture of the exposed area as it existed in the unopened body, and to explore mentally its connections with adjoining unexposed areas around it and below it.
The kind of topographic knowledge which an experienced surgeon possess of the regions on which he operates is therefore ineffable knowledge. [Polanyi 1962: 89]

The claim here is that three-dimensional spatial knowledge is ineffable, or tacit, because it cannot be captured in a representation. Polanyi goes on to argue that even if all human bodies were identical and even if there were a map comprising cross sections based on ‘a thousand thin slices’ of the body, that in itself would not articulate the knowledge of a trained surgeon. Someone knowing merely the former ‘would know a set of data which fully determine the spatial arrangement of the organs in the body; yet he would not know that spatial arrangement itself’ [89]. An additional act of interpretation or imagination is needed. But because that act cannot itself be encoded in a representation, according to Polanyi, it remains tacit.

This argument is a little surprising. Polanyi concedes that the set of cross sectional representations, presumably alongside some further information about their inter-relations such as their order and distance apart, ‘fully determine[s] the spatial arrangement of the organs’ and yet denies that this amounts to an articulation of the three-dimensional understanding.

Without the further information about the relations between the set of maps, the maps alone would not be an articulation of the skilled surgeon’s knowledge. But then neither would they fully determine the arrangement of bodily organs. With that addition, however, why would this not count as an articulation of their knowledge and thus imply that it could be explicit knowledge?

A further possible clue to Polanyi’s thinking runs thus:

The difficulty lies here entirely in the subsequent integration of the particulars and the inadequacy of articulation consists altogether in the fact that the latter process is left without formal guidance. The degree of intelligence required from the student to perform the act of insight which ultimately conveys to him the knowledge of the topography, offers here a measure of the limitations of the articulation representing this topography. [ibid: 90]

But there remains something strange about this line of thought. If the integration of the partial representations, such as the set of cross sections, were left without formal guidance then it would be clear why the partial representations could not articulate the surgeon’s knowledge. But neither would they determine the arrangement of organs as Polanyi has previously asserted.

The difficulty with interpreting this argument is that of balancing the claim that spatial configuration is both determined by what can be represented but remains ineffable and thus tacit rather than explicit.

I think that the clue to its interpretation is to realise that whether a symbol logically determines anything always, according to Polanyi, depends on a tacit element. This is supported by a different argument.

I may ride a bicycle and say nothing, or pick out my macintosh among twenty others and say nothing. Though I cannot say clearly how I ride a bicycle nor how I recognise my macintosh (for I don’t know it clearly), yet this will not prevent me from saying that I know how to ride a bicycle and how to recognise my macintosh. For I know that I know how to do such things, though I know the particulars of what I know only in an instrumental manner and am focally quite ignorant of them. [ibid: 88]

Polanyi suggests that the skill involved in the example of recognising a macintosh is akin to the practical skill of cycle riding. In both cases, the ‘knowledge-how’ depends on something which is not explicit: the details of the act of bike riding or raincoat recognition. Whilst one can recognise one’s own macintosh, in the example, one is ignorant of how. Thus how one does this is tacit.

If this argument is successful it is of general significance because it would carry over to the recognitional skill which underpins all medical classification such as diagnosis but also all linguistic labelling generally. Indeed, Polanyi makes this connection explicilty.

[I]n all applications of a formalism to experience there is an indeterminacy involved, which must be resolved by the observer on the ground of unspecified criteria. Now we may say further that the process of applying language to things is also necessarily unformalized: that it is inarticulate. Denotation, then, is an art, and whatever we say about things assumes our endorsement of our own skill in practising this art. [ibid: 81]

This connection between denotation and tacit recognitional skills appears to be the fundamental argument for the importance of tacit knowledge for explicit scientific accounts. Polanyi summarises the connection thus:

If, as it would seem, the meaning of all our utterances is determined to an important extent by a skilful act of our own – the act of knowing – then the acceptance of any of our own utterances as true involves our approval of our own skill. To affirm anything implies, then, to this extent an appraisal of our own art of knowing, and the establishment of truth becomes decisively dependent on a set of personal criteria of our own which cannot be formally defined.... [E]very where it is the inarticulate which has the last word, unspoken and yet decisive... [ibid: 70-71]

Although this connection looks to be of fundamental importance in linking the tacit dimension between recognising particulars and the use of language in general, Polanyi seems to overstate his case in a way which has significance for understanding tacit knowledge. He says of bicycle riding and macintosh recognition that we know how to do them despite being ‘focally quite ignorant of’ the particulars by which we know how to do them. That thought seems to justify his slogan that we know more than we can tell. But whilst that is a dramatic claim, it does suggest a dilemma for tacit knowledge.

How can tacit knowledge be both tacit and still knowledge? To count as knowledge, there must be something, some content, to be known. But if it cannot be said or told, what sense is there to the idea that there is anything known? If, on the other hand, the content can be articulated, in what sense is it tacit? The paradigmatic way to articulate knowledge is in language and that seems to mark the knowledge out as explicit.

I suggest that the best solution to that dilemma starts by contesting both Polanyi’s slogan and his account of recognition. Polanyi says that one is ignorant of how one recognises one’s own macintosh. But that is a surprising claim. It would be very surprising – and undermine the veracity of the claim – if one could say nothing about how one recognises something as one’s own coat. It is surely more plausible that one might say: “I recognise that it is my macintosh because of how it looks here! with the interplay of sleeve, shoulder and colour.” Or, “there is something about the shape of the lapel here! that is distinct.” One might be able to say this even if one could not recognise a separated sleeve, shoulder or paint colour sample as of the same type or describe the lapel when it is unseen or give a precise name to the colour.

There are some recognitional judgements that do not depend on subsidiary elements (as recognising a macintosh by its sleeves or lapels). One might recognise a colour as the same as another in a variety of ways. One might judge the colour of a sample to be the same colour as a previously encountered one or as the very same (indistinguishable) shade as another adjacent sample. In such cases, the judgement of sameness of colour or sameness of shade might not be made in virtue of anything else (such as a label) but rather a direct judgement of sameness of colour or shade. Still, that is not to say that we know more than we can tell. We can express what we know by pointing to the samples and saying that they are the same colour or shade. In so doing, we express what we know.

The same goes for practical knowledge-how. If one knows how, and is able, to ride a bicycle then one knows that the pedalling should be like this! and the brakes applied progressively like this! It is not that one is ignorant of how to ride even if one does not pay attention to it. If one does not pay attention to how one is riding that does not mean that one cannot put one’s ability into situation-dependent words, but rather that one does not. (That one cannot without paying attention is trivial.)

The connection between practical knowledge and tacit knowledge on this account suggests the need for practical teaching to convey the relevant knowledge. Such tacit knowledge inherits the well known feature of bodily skills that learning them in general requires practice. The same is true for recognitional skills even though they do not require bodily dexterity. Here is one reason for that.

I have discussed recognising a particular macintosh as one’s own and also of recognising a colour. But there are differences. The former is literally a matter of re-cognising the very same entity. But recognising that something has the same colour as something else is not a matter re-cognising the same entity [Travis 2006: 189-93]. Colours are not entities but properties (whatever exactly the preferred scientific account of colour). To recognise something as having the same colour as something else is to judge it to be relevantly similar with respect to colour. In some contexts that might be judging that the same broad colour word applies to it, for example, ‘red’. In others, it might turn on being of indistinguishable shade to the human eye in standard lighting conditions. In clinical cases it might be within limits fixed by medical significance (is the blood sufficiently oxygenated?). To learn to ‘recognise’ or judge colour in such circumstances is an open ended skill requiring practice like other skills.

That however is the general nature of medical classification. It turns on applying diagnostic labels because of relevant similarities between individual cases and hence on context-dependent recognitional judgements. There is thus some continuity between bodily skills, such as the ability to perform a particular surgical technique, and recognitional skills.

Context-dependent recognitional judgements are a way of putting what one knows into words. But, unlike paradigmatic explicit knowledge they are not context-independent or general claims. Rather, they depend on the presence of what is judged. They are tacit or silent insofar as they cannot be put into words outside particular environments or contexts. Thus they contrast with the general and context-independent claims of much of science. Polanyi’s achievement was to draw attention to their importance, nevertheless, for science.

Polanyi, M. (1962) Personal Knowledge, Chicago: University of Chicago Press
Polanyi, M. (1967) The Tacit Dimension, Chicago: University of Chicago Press
Travis, C. (2006) Thought’s Footing, Oxford: Oxford University Press

Re/Mis-reading Bolton and Hill

As a quick reminder for when I can again get access to my books, now packed up in crates for a room move:-

Marking a student essay at the end of last week, I realised that I think of Derek Bolton’s and Jonathan Hill’s attempt to locate meaning in nature as having two key claims.

i) ‘the meaning (information) that regulates action is encoded in the brain’ [Bolton and Hill 2004: 86].

ii) Nature contains not just the patterns exemplified in billiard ball causation but also distinct patterns exemplified in intentional causation.

In my Essential Philosophy of Psychiatry I summarised the second point thus:

The explanatory power of everyday intentional or ‘folk psychological’ explanation derives from the causal power of reasons. But the historical division between reasons and causes voiced, for example, by Karl Jaspers puts this claim under threat. If, instead, one distinguishes between intentional and non-intentional causation, folk psychological explanation ceases to be exceptional and in need of special philosophical explanation and becomes instead a particular example of a more general phenomenon.

The second move is suggested in passages such as (if I had their book I am sure there would be a better passage):

The point is in brief that if the goal is explanation and prediction of intentional behaviour (complex organism-environment interactions), then the methodological assumption has to be that the agent is regulated by information about the environment, that is, by intentional states, either mental, or encoded in the brain, or both. It is, on the other hand, perfectly possible to do away with intentional concepts, but then only non-intentional behaviour can be predicted, for example, physical movements of the body. [ibid: 73]

So here’s the question for those with tidy minds. Are Bolton and Hill ‘bald naturalists’ in McDowell’s phrase or not? The first claim suggests that they are on the assumption that the states that what does the encoding can be described in bald naturalistic terms. That is, on the assumption that it is a reductionist statement. But of course, Bolton and Hill argue explicitly that it is not intended to be that. In my summary I said:

It is not intended as an attempt to reduce meaning or content to non-intentional terms. Such projects, according to Bolton and Hill, presuppose the fallacious distinction between reasons and causes and attempt to reconcile them by showing that one side is really an instance of the other.
Nor is the slogan a claim on behalf of the idea of Fodor’s language of thought hypothesis (discussed in the first section of this chapter). The authors reject representationalist or cognitivist explanations of mental content by invoking the later Wittgenstein’s attack on theories based upon static inner states or syntax. They present the following challenge.
[T]he point is that we can ask of any proposed model of encoding information: ‘In what sense is there anything semantic here?’ [ibid: 71]
Following Wittgenstein, they argue that what makes something meaningful is that it plays a role in the guidance of action.
[W]hat you have to add to, or have instead of, symbol manipulation, in order to achieve intentionality is the regulation of action, that is, meaningful (goal-directed, plastic) interactions with the environment. [ibid: 70]
Brain states do not encode meaning by embodying internal syntactic or sentence-like structures but by guiding action in accordance with norms.

But I took it that the encoding thesis was needed to do some work in making the application of intentional causation to humans less mysterious. I took it to be like that partly because, following the text, I approached the claims in the other order, so that the encoding thesis is used to explain or shed light on the otherwise dark talk of intentional causation when applied to people, a task the authors do indeed seem to accept. I summarised this idea thus:

How is it possible that reasons are a species of intentional causality? Is the information deployed in other intentional-causal sciences, such as biology, of the same order and explanatory of the meaning that is found in mental health care? The recent philosophical problem is not so much to determine whether reasons are causes but how it is possible that reasons are causes. The solution that Bolton and Hill propose is that the brain encodes meaning.

Having set up the problem this way, I wasn’t impressed by their deployment of Wittgenstein, was worried about the metaphor of encoding and was struck by the fact that it was brains that did this. All that suggested an implicit form of reductionism, albeit one not attempted in any detail, which contrasted with the idea that meaning or content is a whole person level phenomenon, not something for brains. Talk of brains guiding action was just a confusion etc etc.

But I now wonder whether the better way to read / use / teach Bolton and Hill’s idea is to ignore the encoding thesis and concentrate instead on the idea of intentional causation. If so then they should not be filed under ‘(unwilling) bald naturalism’ but something quirkier.

The basic naturalising move would then be not an encoding thesis to place meaning in something describable in more naturalistic terms: brains and brain states. It would be instead the claim of the ubiquity of intentional causation.

Now if there is to be no gap between this idea when applied to the dances of bees and to folk psychological accounts of people – a gap otherwise needing to be filled by the encoding thesis – this will place not just a thin notion of information out there in the world (on my previous reading the encoding thesis codes between full blown human semantic intentionality/intensionality and the thin information supposedly carried by brains and bees alike). Nor will that information be – as I would prefer – an abstraction from human semantic intentionality/intensionality: a facon de parler, a convenient short-hand. It will instead be a world full of semantic meaning, more McDowellian than McDowell. Pan psychism.

OK it is obviously not what the authors would intend. But I will at least try to re-read their claims in this way when I get the book back.

Bolton, D. and Hill, J. (1996; second edition 2003) Mind Meaning and Mental Disorder, Oxford: Oxford University Press