Tacit knowledge second stab

My editor didn't like the first version of this entry for a collection on implicit knowledge. So here's the second go.

Introduction

The main promoter of the idea of tacit knowledge (or ‘tacit knowing’, as he preferred) in twentieth century philosophy was the chemist turned philosopher of science Michael Polanyi. His book The Tacit Dimension starts with the manifesto:

I shall reconsider human knowledge by starting from the fact that we can know more than we can tell. [Polanyi 1967b: 4]

But, as he is quick to acknowledge, it is not easy to say exactly what this means. It does, however, suggest a via negativa. What is tacit is what is not ‘tellable’ or explicit under a suitable understanding of that.

The manifesto accords with some of the phenomena that are sometimes called ‘tacit knowledge’ in the popular adoption of the label: recognising someone’s face, or a few hastily-drawn lines as a face; throwing and catching a ball; riding a bicycle; touch typing; skilfully playing the piano; reading a book or map; immediately diagnosing a patient; navigating the shoals of interpersonal relationships; understanding a language. These phenomena seem to involve normative, intentionally directed activities that express some form of knowledge but also something that cannot be fully, at least, put into words.

Two less everyday examples are worth setting out because they are often used as paradigmatic examples in more academic discussions of tacit knowledge.

The first relates to poultry farming, in which there is economic advantage in being able to determine the gender of chicks as soon as possible after they hatch. In the 1920s, Japanese scientists came upon a Chinese method by which this could be done based on subtle perceptual cues with a suitably held one day old chick [Martin 1994: 3]. It was, nevertheless, a method that required a great deal of skill developed through practice. After four to six weeks of instruction and practice, a newly qualified chick-sexer might be able to determine the sex of 200 chicks in 25 minutes with an accuracy of 95% rising with years of practice to 1,000 - 1,400 chicks per hour with an accuracy of 98% [Gellatly 1986: 4].

The second example is that of skilled Polynesian navigators who were found to be able to navigate small out-rigger canoes ‘across two or three hundred miles of open sea; and do so in almost any weather, and even when less than fully sober. How is it done?’ [ibid.: 5]. Investigation suggested that the skill took years to master and was context specific: they were only able to navigate the seas and in the wind conditions of their familiar part of the world and that they were unable to put into words on dry land.

In the first case, although the theory of chick sexing by genital differences was simple, the nature of the recognitional skill involved was elusive. According to R.D. Martin’s history, having read about the theory, Mr Kojima, a poultry farmer from Nagoya, reversed engineered the skill using first 60 day, then 30 day, then younger and younger chicks [Martin 1994: 4]. Dreyfus and Dreyfus’ 1986 account seems to imply that not even the chick-sexers could explain how they did it [Dreyfus and Dreyfus 1986: 197-8]. Their own knowledgeable ability was opaque even to them and hence doubly ‘untellable’. But this claim risks distorting the fact that it was necessary to make an immediate perceptual judgement as chick-manipulation changed the key visual cue. Chick-sexers also often learnt their judgements aided by diagrams later ‘outgrown’ once the ability was second nature [Martin 1994].

The very idea of tacit knowledge presents a challenge, however. Any account of it must show that it is both tacit and knowledge. But it is not easy to meet both conditions. Emphasising the tacit status, threatens the idea that there is anything known. Articulating a knowable content, that which is known by the possessor of tacit knowledge, risks making it explicit. What limits can be placed on what can be said still leaving something that can be known? I will suggest that an account can be developed that meets these dual condition influenced by Polanyi, Ryle and Wittgenstein. One consequence is that not all potential instances of tacit knowledge in philosophy – such as of the axioms of a theory of meaning for a language – will fit [Evans 1981].

Polanyi, tacit knowledge and the distinction between focal versus subsidiary awareness

Given, however, that the notion of tacit knowledge is only partly embedded in ordinary language and is so specifically associated with Polanyi, why not simply adopt his own account? There are two reasons. First, Polanyi is promiscuous with his use of the idea and so it is difficult to derive a clear account. As one Polanyian scholar concedes:

Now we are in a position better to understand why Polanyi’s notion of the tacit is unclear. Since he uses the term in different contexts to embrace all in human cognition that is not explicit, it is a comprehensive term that includes processes and contents from many ontological levels and different disciplines. Tacit processes and components include: (1) all the autonomic physiological processes that support at lower levels an act of knowing at any given moment; (2) plus the embodied skills, learned and perhaps innate, that contribute to one’s intentional actions (including cognitive actions); (3) plus all the psychological dynamics (integrations, judgments) involved in this knowing; (4) plus all the background presuppositions and beliefs that shape the prevailing personal framework of interpretation; (5) plus the immediately preceding thoughts and actions still active in working memory; (6) plus the network of sensitivities, moods, and feelings salient at the moment of analysis; (7) plus the influx of sensations, all vying for attention, relayed by the central and peripheral nervous systems, including sensory data from the body’s receptors; (8) plus the interests, goals, and expectations pushing thought and action toward certain satisfactions; (9) plus the words and grammar evoked as the transparent means to express and evaluate the meaning of this concatenation of influences. [Guilick 2016: 303]

The second reason is that Polanyi’s own account emphasises a phenomenological distinction between subsidiary and focal awareness which applies either to too narrow a range of cases or is merely implausibly generalised. To distinguish between focal and subsidiary awareness he uses the example of pointing to something with a finger

There is a fundamental difference between the way we attend to the pointing finger and its object. We attend to the finger by following its direction in order to look at the object. The object is then at the focus of our attention, whereas the finger is not seen focally, but as a pointer to the object. This directive, or vectorial way of attending to the pointing finger, I shall call our subsidiary awareness of the finger. [Polanyi 1967a: 301]

In attending from the finger to the object, the object is the focus of attention whilst the finger, though seen, is not attended to. Applying this notion more broadly, he gives the following example.

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]

This passage suggests that the skill involved in recognising a macintosh is akin to the practical skill of cycle riding (an important clue, to which I will return). In both cases, the ‘knowledge-how’ depends on something which is not explicit: the details of the act of bike riding or raincoat recognition. While one can recognise one’s own macintosh, in the example, one is ignorant, in some sense, of how. Thus, how one does this is tacit.

If this argument were successful it would be of general significance because it would also apply to all linguistic labelling and Polanyi explicitly makes this connection.

[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]

The wider argument for denotation being tacit seems to rest on an appeal to the clearer case of the recognition of particulars – such as a particular macintosh – which, Polanyi argues, depends on features of which one is focally ignorant. (The example concerns recognition of a particular macintosh as one’s own rather than merely that an object is a macintosh but nothing turns on this difference.)

However, to justify the wider argument, he needs to defend the general claim that explicit recognition of something as an instance of a type is always based on the implicit recognition of subsidiary properties of which one is focally ignorant. To recognise a feature (F, say) one must a) always recognise it in virtue of something else (subsidiary features G, H and I, eg.) of which b) one is focally ignorant. But it is not clear that either part of this claim is true.

Polanyi seems to assume that the question of how one recognises something to be of an instance of a particular kind – for example, a patch colour on a summer’s afternoon – always has an informative answer. To cover cases where it is not obvious what this is, he claims – implausibly – that one can have subsidiary awareness of one’s own ‘neural traces in the brain’ [Polanyi 1969: 147]. But whilst the question may sometimes have an informative answer, there is no reason to think that it always does at the level of the person. An excursion into optics and neurology may explain how some perceptual discrimination is possible without trading in (deploying rather than explaining) person level knowledge.

Even in cases where one recognises a particular as an F in virtue of its subsidiary properties G, H, I, and cannot give an independent account of those properties, it is not clear that one need be focally ignorant of them. It may be, instead, that the awareness one has of G, H, I is manifested in the recognition of something as an F or a particular F. One might say, I recognise that this is a macintosh (or my macintosh) because of how it looks here with the interplay of sleeve, shoulder and colour even if one could not recognise a separated sleeve, shoulder or paint colour sample as of the same type. Whilst it seems plausible that one might not be able to say in context-independent terms just what it is about the sleeve that distinguishes a macintosh from any other kind of raincoat (lacking the vocabulary of tailoring) that need not imply that one is focally ignorant of, or not attending to, just those features that make a difference. Recognition may depend on context-dependent or demonstrative elements, such as recognising shapes or colours for which one has no prior name. That suggests one has to be focally aware of them. (This notion of context-dependent or demonstrative elements will play a role shortly.)

For these reasons and contrary to Polanyian scholarship, it is better to sever an account of tacit knowledge from Polanyi’s preferred distinction between focal versus subsidiary awareness [contra Grene 1977].

A Polanyi-, Ryle- and Wittgenstein-inspired account of tacit knowledge

I suggested that the very idea of tacit knowledge presents a challenge. Tacit knowledge is both tacit and knowledge and it is hard to meet both conditions. There is a second strand in Polanyi’s work which helps address this, though abstracting away from his own account. At the start of his book Personal Knowledge he says:

I regard knowing as an active comprehension of things known, an action that requires skill. [Polanyi 1962: vii]

In combination with the first slogan – ‘I shall reconsider human knowledge by starting from the fact that we can know more than we can tell’ – and the idea of context-dependent demonstratives (above), this suggests a clue to tacit knowledge.

Taking tacit knowledge to be practical suggests one way in which it is ‘untellable’. It cannot be articulated except in context-dependent practical demonstrations and, further, only to those with ‘eyes to see’. It cannot be made explicit in words alone. Yet, at the same time, it has content: practical knowledge of how to do something.

In Tacit Knowledge, Neil Gascoigne and I set out an account of tacit knowledge as conceptually structured, context dependent, practical knowledge [Gascoigne and Thornton 2013]. We locate tacit knowledge at the level of the person, rather than the sub-personal, and within the ‘space of reasons’, the space of justifying and being able to justify what one says. [McDowell 1994: xiv, Sellars 1997: 76]. We motivate the account partly in response to ‘regress arguments’ offered by Gilbert Ryle and Ludwig Wittgenstein and defend the former against new intellectualist criticism [eg. Stanley and Williamson 2001]. But the support the regress arguments offer to a conception of tacit knowledge is nuanced.

Ryle’s regress argument is summarised pithily thus:

If a deed, to be intelligent, has to be guided by the consideration of a regulative proposition, the gap between that consideration and the practical application of the regulation has to be bridged by some go-between process which cannot by the pre-supposed definition itself be an exercise of intelligence and cannot, by definition, be the resultant deed. This go-between application- process has somehow to marry observance of a contemplated maxim with the enforcement of behaviour. So it has to unite in itself the allegedly incompatible properties of being kith to theory and kin to practice, else it could not be the applying of the one in the other. For, unlike theory, it must be able to influence action, and, unlike impulses, it must be amenable to regulative propositions. Consistency requires, therefore, that this schizophrenic broker must again be subdivided into one bit which contemplates but does not execute, one which executes but does not contemplate and a third which reconciles these irreconcilables. And so on for ever. [Ryle 1945: 2]

There has been a recent flurry of literature on the precise nature of this argument and thus whether it is successful [eg Stanley and Williamson 2001; Noë 2005; for detailed assessment see Gascoigne and Thornton 2013: 51-79]. But it involves something like the following regress:

Suppose all know-how can be articulated in words as a piece of knowledge-that: grasping some proposition that p. Grasping the proposition that p is itself something one can do successfully or unsuccessfully, so it is also a piece of know-how. So, on the theory in question, it will involve grasping another proposition, call this q. But grasping the proposition that q is itself something one can do successfully or unsuccessfully, so it is also a piece of know-how. So, on the theory in question, it will involve grasping another proposition, call this r.... etc

If the first step of the reductio is designed to ‘articulate’ or represent a piece of otherwise merely tacit knowledge at the heart of recognition, it will lead to a regress. Ryle himself suggests that it can be used to undermine what he calls an ‘intellectualist legend’ which attempts to explain practical knowing-how though a deeper, theoretical form of knowing-that based on grasping a proposition. His counter argument is that, since grasping a proposition can be done well or badly, the only way to avoid the vicious regress is to grant that intelligence can accrue to, and be manifested in, practical knowledge directly, without further theoretical explanation or underpinning. Knowing-how is more basic than knowing-that.

What is the relationship between Ryle’s argument and the clues extracted (above) from Polanyi to the nature of tacit knowledge? On the one hand, it offers some nuanced support. Polanyi’s claim that we know more than we can tell is one clue. Given that Ryle argues that knowing-how cannot be explained through knowing-that or grasp of a proposition – both paradigmatic of what can be put into words – if, following Polanyi’s second clue, tacit knowledge is equated with practical knowledge then Ryle’s argument suggests limits to the way of putting it into words. It cannot be explained, at least, in knowing-that terms.

But, on the other hand, the idea that practical knowledge can express intelligence directly – without needing to inherit it from grasp of a proposition – suggests the following thought which runs counter to the claim that we know more than we can tell. (And thus it puts Polanyi’s two clues in tension.) Consider the following piece of practical knowledge: the ability to recognise a raincoat as a macintosh and thus denote it ‘macintosh’. Why is the denoting of raincoats as ‘macintoshes’ not what articulating or expressing this piece of recognitional knowledge amounts to, thus discharging any tacit element? In this particular example, because it involves linguistic denotation, why is that not putting all the relevant knowledge in play into words (calling this coat a ‘macintosh’, eg.)? Why assume, as Polanyi does, that there is always a further, though tacit, answer as to how one recognises that something is an instance of a general kind?

Another regress argument, this time from Wittgenstein, suggests an answer.

Wittgenstein considers teaching a pupil to continue a mathematical series correctly who ‘judged by the usual criteria… has mastered the series of natural numbers’ and is then taught to continue series of the form 0, n, 2n, 3n, etc. for arbitrary values of n

Then we get the pupil to continue one series (say “+ 2”) beyond 1000 – and he writes 1000, 1004, 1008, 1012. We say to him, “Look what you’re doing!” – He doesn’t understand. We say, “You should have added two: look how you began the series!” – He answers, “Yes, isn’t it right? I thought that was how I had to do it.” —– Or suppose he pointed to the series and said, “But I did go on in the same way”. – It would now be no use to say, “But can’t you see…?” – and go over the old explanations and examples for him again. In such a case, we might perhaps say: this person finds it natural, once given our explanations, to understand our order as we would understand the order “Add 2 up to 1000, 4 up to 2000, 6 up to 3000, and so on”. This case would have similarities to that in which it comes naturally to a person to react to the gesture of pointing with the hand by looking in the direction from fingertip to wrist, rather than from wrist to fingertip. [Wittgenstein 2009 §185]

The pupil reacts in a divergent way (from us) to explanations of how to continue. He appears to have acted under a divergent interpretation of the expression of the rule. But if to be successful an explanation of a rule has to be correctly interpreted then any expression of the interpretation might be subject to divergent understandings whether explicitly stated in a symbolism or by examples.

“But how can a rule teach me what I have to do at this point? After all, whatever I do can, on some interpretation, be made compatible with the rule.” – No, that’s not what one should say. Rather, this: every interpretation hangs in the air together with what it interprets, and cannot give it any support. Interpretations by themselves do not determine meaning. [ibid §198]

Since any interpretation has to be specified in some way, if understanding depends on an interpretation, no specification would be sufficient in itself. Hence the label ‘regress of interpretations’.

In an early work – Changing Order - on the role of tacit knowledge in science, the sociologist Harry Collins assumes that the regress argument shows that the correct way to continue, the correct notion of sameness, cannot be specified or said [Collins 1985].

[S]ince in spite of this we all know the correct way to go on, there must be something more to a rule than its specifiability’ [ibid: 14].

Thus there is a genuine gap between what is said in an explanation of a rule and what is understood when all goes well and something is needed to plug the gap so as to explain the otherwise mysterious ability to go on correctly.

The extra element is social entrenchment such that the sense of sameness in continuing in the same way is relative to a community. It underpins the ‘mysterious abilities that enable us to know when to continue ‘2,4,6,8’ with ’10,12,14,16’ and when with ‘who do we appreciate?’’ [ibid: 22, italics added].

Collins calls this ‘tacit knowledge’.

[T]he member of a social group who has the ability to continue the sequence ‘2,4,6,8’ with ‘10,12,14,16’ as a matter of course, without even thinking about it, also possesses something that the stranger to our culture and the newborn do not. This is sometimes referred to as ‘social skill’ but we can call it tacit knowledge without doing too much violence to the term. [ibid: 56 italics added]

According to Collins tacit knowledge cannot be put into words, is invisible in its transmission and capricious [ibid: 129]. This suggests a picture of rule following in which explanations are always inadequate but are patched by invisible and capricious tacit knowledge: the picture that Wittgenstein puts into the mouth of his interlocutor in order to reject it.

“But do you really explain to the other person what you yourself understand? Don’t you leave it to him to guess the essential thing? You give him examples a 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 mean” would amount to: “various interpretations of my explanation come to his mind, and he picks one of them”. So in this case he could ask; and I could and would answer him. [Wittgenstein 2009: §210]

There is, however, a different way to connect the regress argument to tacit knowledge. The moral of the regress of interpretations according to John McDowell is that:

We learn that it is disastrous to suppose there is always a conceptual gap between an expression of a rule and performances that are up for assessment according to whether or not they conform to the rule, a gap that is made vivid by saying the expression of the rule stands there… We must not acquiesce in the idea that an expression of a rule, considered in itself, does not sort behaviour into performances that follow the rule and performances that do not. [McDowell 2009: 100-1]

This, according to McDowell, is the moral of §201 of the Philosophical Investigations: ‘there is a way of grasping a rule which is not an interpretation, but which, from case to case of application, is exhibited in what we call “following the rule” and “going against it”’ [Wittgenstein 2009: §201]. For those who are party to the relevant practices, there is no gap between the expression of the rule and being told which way to go. A signpost, for example, not merely a signpost under an interpretation, points the way. On this account, the regress of interpretations is stopped before it can start. Hence, the expression of a rule is explicit in its directions.

But it is only explicit for those with the right eyes to see, or ears to hear, an explanation, for those, in Stanley Cavell’s phrase, who share the ‘whirl or organism’ of our form of life [Cavell 1969: 52]. This in turn grounds out in the applications of rules, which are context-dependent, conceptually structured and practical. The regress of interpretations argument shows that any informative explanation of what following a rule correctly, going on in the same way, comprises soon gives out. Any general account of the relation of rules or conceptual generalities to particular judgements falls prey to the regress of interpretations. What can be made explicit via the art of denotation – in Polanyi’s phrase – rests on the tacit, not because of a distinction between focal and subsidiary awareness but because going on in each specific way simply is accord with a rule or concept, is also context dependent and, by linking conceptual generalities to particular cases in judgements, is practical.

Because the relation of the tacit and explicit is subtle, it is worth giving a worked example. It is possible to make a béchamel sauce by following a recipe: to measure ingredients and to cook at a specific level for a predetermined time. But it is also possible to make one ‘bi t’rack o’ t’ee’ (by the rack of the eye) as it is said in Yorkshire dialect. One judges a rough equivalence of butter and flour, sautés the resultant ‘rue’ until it looks right (so as to cook the flour) and then adds milk gradually and intermittently while stirring continually, watching the consistency fall to almost that of the added milk and then stiffen again. Gradually, the rate of returning to stiffness slows. Were one to add too much milk it would never thicken again. Thus the end point is fixed by a judgement that the rate of stiffening has slowed sufficiently that a final addition of a particular amount of milk will, combined with the residual heat energy of the sauce, yield the right consistency by the time it is served. Cooking thus is an expression of a conceptual understanding of the effects of dilution, heating, stirring, thickening against the jeopardy of adding too much milk. But the expression of the concepts comprises demonstratives of the form ‘that degree of stiffness’ and ‘that quantity of additional milk’. It involves the right kind of manipulation of the sauce mixture with a spoon and the right degree of heat applied so that it bubbles thus. Although it is possible to make a béchamel sauce following an explicit recipe, the cook who makes it ‘by eye’ does so employing a form of conceptually structured, context dependent, practical knowledge, or tacit knowledge, on my account. Further, its expression and articulation in a practical demonstration depends for its success on the ability of the audience to see in the particularity of a few cases the generality of the rules involved. They must have ‘eyes to see’. They must share the whirl of organism.

The dual challenge for an account of tacit knowledge can be met in this way as follows. It is tacit because it cannot be codified in context-independent and general terms – it is in that sense ‘untellable’ – while at the same time it is knowledge because it is articulated in the application of general concepts to the particular case, grounded in practice, which is a form of know-how. One consequence of this account is that explicit knowledge always rests on a body of tacit knowledge. Against that background – the art of denotation – some things can be made explicit.

An ontology rather than epistemology of tacit knowledge?

Taking tacit knowledge to be person-level and conceptual (as well as practical and context specific) is not the only possible approach, however.

In his Tacit and Explicit Knowledge, Collins also contrasts tacit knowledge with what is explicit [Collins 2010]. But the characterisation of the explicit is not as what can be expressed linguistically – as Polanyi and I assume – but rather with what Collins calls ‘strings’ [Collins 2010: 57]. Strings are ‘bits of stuff inscribed with patterns: they might be bits of air with patterns of sound waves, or bits of paper with writing, or bits of the seashore with marks made by waves, or patterns of mould, or almost anything’...’ [ibid: 9] The motivation for this is to avoid the ‘freight of inherent meaning that makes the notions of signs, symbols and icons so complicated’ [ibid: 9]. One repeated worry is that strings (and hence signs as kinds of strings) do not have an essential meaning: ‘strings are without meaning… a book is a physical thing, not a meaningful thing’ [ibid: 34]. So rather than presupposing meaning or mental content, Collins stresses the physical nature of strings: ‘a string is just a physical object and it is immediately clear that whether it has any effect and what kind of effect this might be is entirely a matter of what happens to it’ [ibid: 9].

But since strings are just ‘bits of stuff inscribed with patterns’ their elaboration, transformation, mechanization and explanation are all counted by Collins as instances of what is explicit. Hence the focus of the book turns out not to be the nature of the knowledge that rational subjects possess but the worldly objects of knowledge: the processes or tasks or worldly patterns themselves.

One example of this is the way that Collins, as I have, attempts to demystify tacit knowledge by connecting it to what is practical:

We are just like complicated cats, dogs, trees, and sieves… Sometimes we can do things better than cats, dogs, trees and sieves can do them, and sometimes worse. A sieve is generally better at sorting stones than a human (as a fridge is better at chilling water), a tree is certainly better at growing leaves, dogs are better at being affected by strings of smells, and cats are better at hunting small animals… if we were to stop talking and just get on with things — that is, if the tacit was not made mysterious by its tension with the explicit — there would be no puzzle at all about the body, per se. That teaching humans to accomplish even mimeomorphic actions is a complicated business, involving personal contact, says nothing about the nature of the knowledge, per se. (ibid: 104-5)

The final claim reiterates the idea that what matters to Collins – ‘knowledge, per se’ – does not depend on how humans think of or conceptualise tasks, something in the realm of sense, but rather on the mechanical process, in the realm of reference. In my example above, Collins would not distinguish between the béchamel sauce maker who follows a recipe and the one who uses uncodified judgement if, by chance, they undertake the same process. The process, not the knowledge, the way things are understood by a subject, matters. For Collins, the tacit status of the knowledge of a subject can be undermined through a kind of action at a distance if the process they follow is codified elsewhere [Thornton 2013]. The passage does not explicitly equate tacit knowledge with the behaviour of cats, dogs, trees and sieves but it does suggest that thinking of sieves sorting stones is a good way to demystify human tacit knowledge. Whatever the independent merits of an ontology of strings, however, it is far from an account of knowledge itself. It is pitched at the level of reference rather than sense and ontology rather than epistemology.

Polanyi was right to stress the importance of a form of knowledge at the level of the person, expressive of human minds, tied to context and practical and presupposed by, but exceeding, linguistic codification. The challenge is to articulate it. Construing tacit knowledge as conceptually structured, context dependent, practical knowledge is one way to meet that challenge.

References

Cavell, S. (1969) Must we mean what we say?, Cambridge: Cambridge University Press
Collins, H. (1985). Changing Order: Replication and Induction in Scientific Practice, London: Sage
Collins, H. (2010) Tacit and Explicit Knowledge, Chicago: University of Chicago Press
Dreyfus, H.L. and Dreyfus, S.E. (1986) Mind Over Machine, New Youk: Macmillan
Evans, G. (1981) ‘Semantic Theory and Tacit Knowledge’. In Holtzman, S.H. and Leitch, C.M. (eds.). Wittgenstein: To Follow a Rule, London: Routledge and Kegan Paul
Gascoigne, N. and Thornton, T. (2013) Tacit Knowledge, Durham: Acumen
Gellatny, A. (ed) (1986) The Skilful Mind, Milton Keynes: Open University Press
Grene, M. (1977) ‘Tacit knowing: grounds for a revolution in philosophy’ Journal of the British Society for Phenomenology, 8: 164-71
Guilick, W. (2016) ‘Relating Polanyi’s Tacit Dimension to Social Epistemology: Three Recent Interpretations’ Social Epistemology, 30: 297-325
McDowell, J. (2009) The Engaged Intellect, Cambridge, MA: Harvard University Press
McDowell, J. (1994) Mind and World, Cambridge, MA: Harvard University Press
Martin, R.D. (1994) The Specialist Chick Sexer, Melbourne: Bernal
Noë, A. (2005) ‘Against intellectualism’ Analysis 65: 278-90
Polanyi, M. (1962) Personal Knowledge, Chicago: University of Chicago Press
Polanyi, M. (1967a) ‘Sense-giving and sense-reading’ Philosophy 42: 301-325
Polanyi, M. (1967b) The Tacit Dimension, Chicago: University of Chicago Press
Polanyi, M. (1969) Knowing and Being, Chicago: University of Chicago Press
Ryle, G. (1949) The Concept of Mind, London: Hutchinson.
Ryle, G. (1945) ‘Knowing How and Knowing That’ Proceedings of the Aristotelian Society, 46: 1-16
Sellars, W. (1997). Empiricism and the Philosophy of Mind. Intro. by R. Rorty and study guide by R. Brandom. Cambridge, MA: Harvard University Press.
Stanley, J. and Williamson, T. (2001) ‘Knowing how’ The Journal of Philosophy 97: 411-444.
Thornton, T. (2013) ‘Tacit knowledge and its antonyms’ Philosophia Scientiae 17: 93-106
Wittgenstein, L. (2009) Philosophical Investigations, Oxford: Blackwell

Everything is made of peas!

I was trying to explain to some nursing students the way philosophy works. I realise I had an uneasy mix of pride and shame in my discipline. Philosophy typically goes like this.

To start with, most non-philosophers assume that some things are obviously made of peas (such as piles of peas on a plate) and some things less obviously (such as mushy peas). There are some publicly disputed cases (guacamole) and some hard cases (split peas, which are really nuts). But the folk metaphysics of peas seems calm.

Then some bright spark philosopher offers an argument to the effect that everything is made of peas. His/her argument starts with the obvious cases (peas), gives a method of construing some other cases as made of peas under a slightly wider understanding of ‘pea’ and shows that Wittgensteinian arguments from the 1970s that nothing other than peas are made of peas were themselves question-begging. A new pea-metaphysics prevails in which the bright spark’s PhD students expand the pea ontology to some further interesting cases using S5 modal logic that few others can grasp but nod anyway.

There are jobs to be had in peas.

But then other somewhat envious technically minded philosophers show that some of pea constructions are implausible under better understandings of the key concepts on which they depend. Others suggest non-pea alternative constructions. And then a John McDowell highlights to everyone’s surprise that the very motive to assume that peas must be at the heart of things depends on a blindingly obvious but somehow missed merely optional assumption driven by a misreading of the C17 scientific revolution. There’s much embarrassment all round and everyone drops peas from the tripos. And strangely, the 1970s Wittgensteinian pea-eschewing arguments make a comeback.

This usually takes about 20 years.

Tacit knowledge

Introduction

The main promoter of the idea of tacit knowledge (or ‘tacit knowing’, as he preferred) in twentieth century philosophy was the chemist turned philosopher of science Michael Polanyi. His book The Tacit Dimension starts with the manifesto:

I shall reconsider human knowledge by starting from the fact that we can know more than we can tell. [Polanyi 1967b: 4]

But, as he is quick to acknowledge, it is not easy to say exactly what this means. It does, however, suggest a via negativa. What is tacit is what is not ‘tellable’ or explicit under a suitable understanding of that.

The manifesto accords with some of the phenomena that are sometimes called ‘tacit knowledge’ in the popular adoption of the label: recognising someone’s face, or a few hastily-drawn lines as a face; throwing and catching a ball; riding a bicycle; touch typing; skilfully playing the piano; reading a book or map; immediately diagnosing a patient; navigating the shoals of interpersonal relationships; understanding a language. These phenomena seem to involve normative, intentionally directed activities that express some form of knowledge but also something that cannot be fully, at least, put into words.

Two less everyday examples are worth setting out because they are often used as paradigmatic examples in more academic discussions of tacit knowledge.

The first relates to poultry farming, in which there is economic advantage in being able to determine the gender of chicks as soon as possible after they hatch. In the 1920s, Japanese scientists came upon a Chinese method by which this could be done based on subtle perceptual cues with a suitably held one day old chick [Martin 1994: 3]. It was, nevertheless, a method that required a great deal of skill developed through practice. After four to six weeks of instruction and practice, a newly qualified chick-sexer might be able to determine the sex of 200 chicks in 25 minutes with an accuracy of 95% rising with years of practice to 1,000 - 1,400 chicks per hour with an accuracy of 98% [Gellatly 1986: 4].

The second example is that of skilled Polynesian navigators who were found to be able to navigate small out-rigger canoes ‘across two or three hundred miles of open sea; and do so in almost any weather, and even when less than fully sober. How is it done?’ [ibid.: 5]. Investigation suggested that the skill took years to master and was context specific: they were only able to navigate the seas and in the wind conditions of their familiar part of the world and that they were unable to put into words on dry land.

In the first case, although the theory of chick sexing by genital differences was simple, the nature of the recognitional skill involved was elusive. According to R.D. Martin’s history, having read about the theory, Mr Kojima, a poultry farmer from Nagoya, reversed engineered the skill using first 60 day, then 30 day, then younger and younger chicks [Martin 1994: 4]. Dreyfus and Dreyfus’ 1986 account seems to imply that not even the chick-sexers could explain how they did it [Dreyfus and Dreyfus 1986: 197-8]. Their own knowledgeable ability was opaque even to them and hence doubly ‘untellable’. But this claim risks distorting the fact that it was necessary to make an immediate perceptual judgement as chick-manipulation changed the key visual cue. Chick-sexers also often learnt their judgements aided by diagrams later ‘outgrown’ once the ability was second nature [Martin 1994].

The very idea of tacit knowledge presents a challenge, however: it has to be tacit and it has to be knowledge. But it is not easy to meet both conditions. Emphasising the tacit status, threatens the idea that there is something known. Articulating a knowable content, that which is known by the possessor of tacit knowledge, risks making it explicit. What limits can be placed on what can be said still leaving something that can be known?

Given his central role in discussing tacit knowledge, I will turn first to Polanyi, though as one Polanyian scholar concedes:

Now we are in a position better to understand why Polanyi’s notion of the tacit is unclear. Since he uses the term in different contexts to embrace all in human cognition that is not explicit, it is a comprehensive term that includes processes and contents from many ontological levels and different disciplines. Tacit processes and components include: (1) all the autonomic physiological processes that support at lower levels an act of knowing at any given moment; (2) plus the embodied skills, learned and perhaps innate, that contribute to one’s intentional actions (including cognitive actions); (3) plus all the psychological dynamics (integrations, judgments) involved in this knowing; (4) plus all the background presuppositions and beliefs that shape the prevailing personal framework of interpretation; (5) plus the immediately preceding thoughts and actions still active in working memory; (6) plus the network of sensitivities, moods, and feelings salient at the moment of analysis; (7) plus the influx of sensations, all vying for attention, relayed by the central and peripheral nervous systems, including sensory data from the body’s receptors; (8) plus the interests, goals, and expectations pushing thought and action toward certain satisfactions; (9) plus the words and grammar evoked as the transparent means to express and evaluate the meaning of this concatenation of influences. [Guilick 2016: 303]

My aim is to suggest a coherent person-level notion of tacit knowledge influenced by Polanyi rather than to attempt to make sense of his wider encyclopaedic ambition. One consequence is that not all potential instances of tacit knowledge in philosophy – such as of the axioms of a theory of meaning for a language – will fit [Evans 1981].

In Personal Knowledge, Polanyi examines how what can be said or, more broadly, articulated both leaves room for, and depends on, something outside what can be articulated [Polanyi 1962]. There are two key arguments which stress the tacit element. (I will return to a third suggestion that helps with the knowledge status later.)

One depends on limits on the kind of representation available to summarise explicit knowledge in science, thus indicating a space 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.

Polanyi on the limits of 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 possesses 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, which here stands for tacit, because it cannot be captured in a representation. Polanyi goes on to argue that even if human bodies were identical and 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’ [ibid: 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 is surprising. Polanyi concedes that the set of cross-sectional representations, presumably with further information about their inter-relations, ‘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 the maps alone would not be an articulation of the skilled surgeon’s knowledge. But then neither would they fully represent the arrangement of bodily organs. With the further information, however, why would this not count as an explicit articulation of the surgeon’s knowledge and thus not tacit?

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, eg. cross-sections, were left without formal guidance then it would be clear why they could not articulate the surgeon’s knowledge. But neither would they determine the arrangement of organs as Polanyi has previously asserted.

I think that the clue to its Polanyi’s argument is his claim that whether a symbol determines anything always depends on a tacit element. This is supported by a different argument concerning the distinction between focal attention and subsidiary awareness.

Recognition and the distinction between focal and subsidiary awareness

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 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. While one can recognise one’s own macintosh, in the example, one is ignorant, in some sense, of how. Thus, how one does this is tacit.

If this argument were successful it would be of general significance because it would also apply to all linguistic labelling. Polanyi makes this connection.

[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.

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]verywhere it is the inarticulate which has the last word, unspoken and yet decisive... [ibid: 70-71]

The argument for denotation being tacit seems to rest on an appeal to the clearer case of the recognition of particulars – such as a particular macintosh – which, Polanyi argues, depends on features of which one is focally ignorant. (The example concerns recognition of a particular macintosh as one’s own rather than merely that an object is a macintosh but nothing turns on this difference.)

To justify his claim about denotation he needs to defend the general claim that explicit recognition of something as an instance of a type is based on the implicit recognition of subsidiary properties of which one is focally ignorant. To recognise a feature (F, say) one must a) always recognise it in virtue of something else (subsidiary features G, H and I, eg.) of which b) one is focally ignorant. But it is not clear that either part of this claim is true.

Elsewhere, Polanyi uses the example of pointing to something with a finger to distinguish between focal and subsidiary awareness.

There is a fundamental difference between the way we attend to the pointing finger and its object. We attend to the finger by following its direction in order to look at the object. The object is then at the focus of our attention, whereas the finger is not seen focally, but as a pointer to the object. This directive, or vectorial way of attending to the pointing finger, I shall call our subsidiary awareness of the finger. [Polanyi 1967a: 301]

In attending from the finger to the object, the object is the focus of attention whilst the finger, though seen, is not attended to. The finger is not invisible. It could itself be the object of focal attention were it attended to. This suggests that the first part of the general claim that Polanyi needs itself faces an objection based on a regress. The recognition of an instance of a type depends on subsidiary awareness of something else which could have been the object of focal awareness and thus would have depended on subsidiary awareness of something else.

This is a potential rather than a vicious regress but, nevertheless, suggests something implausible about Polanyi’s general claim. It does not seem reasonable to think that it is always the case that recognition of the instantiation of a kind, for example a colour, depends on subsidiary awareness of something else. Polanyi seems to assume that the question of how one recognises something to be of an instance of a particular kind always has an informative answer and to cover cases where it is not obvious what this is, the defensive move he makes is to claim – implausibly – subsidiary awareness of ‘neural traces in the brain’ [Polanyi 1969: 147]. But whilst it sometimes may have an informative answer, there is no reason to think that it always has. This puts the first part of Polanyi’s claim – that to recognise a feature one must a) always recognise it in virtue of something else – under pressure. What of the second aspect: that one must be focally ignorant of the subsidiary features?

Even in cases where one recognises a particular as an F in virtue of its subsidiary properties G, H, I, and cannot give an independent account of those properties, it is not clear that one need be focally ignorant of them. It may be, instead, that the awareness one has of G, H, I is manifested in the recognition of something as an F or a particular F. One might say, I recognise that this is a macintosh (or my macintosh) because of how it looks here with the interplay of sleeve, shoulder and colour even if one could not recognise a separated sleeve, shoulder or paint colour sample as of the same type. Whilst it seems plausible that one might not be able to say in context-independent terms just what it is about the sleeve that distinguishes a macintosh from any other kind of raincoat (lacking the vocabulary of tailoring) that need not imply that one is focally ignorant of, or not attending to, just those features that make a difference. Recognition may depend on context-dependent or demonstrative elements, such as recognising shapes or colours for which one has no prior name. That suggests one has to be focally aware of them.

While it is true that one is sometimes focally unaware of features that underpin one’s recognitional abilities and one cannot always say in general term on what features one’s recognition depends these facts do not support the general claim that (focal) recognition of one feature always depends on merely subsidiary awareness of something else. And if not, Polanyi has not offered a general reason to hold that recognition is a tacit skill by his own lights.

Contrary to Polanyian scholarship, it is thus better to sever an account of tacit knowledge from the focal versus subsidiary awareness distinction [contra Grene 1977].

A narrower Polanyi-inspired account of tacit knowledge

I suggested earlier that the very idea of tacit knowledge presents a challenge. Tacit knowledge is both tacit and knowledge and it is hard to meet both conditions. There is a second strand in Polanyi’s work which helps address this, though abstracting away from Polanyi’s own account. At the start of his book Personal Knowledge he says:

I regard knowing as an active comprehension of things known, an action that requires skill. [Polanyi 1962: vii]

In combination with the first slogan, this suggests a clue to tacit knowledge. First, it cannot be made explicit and, second, it is connected to action, the practical knowledge of a skilled agent. The latter suggests how tacit knowledge has content: as practical knowledge of how to do something. Taking tacit knowledge to be practical suggests one way in which it is ‘untellable’. It cannot be articulated except in context-dependent practical demonstrations. It cannot be made explicit in words alone.

In Tacit Knowledge, Neil Gascoigne and I set out an account of tacit knowledge as conceptually structured, context dependent, practical knowledge [Gascoigne and Thornton 2013]. We locate tacit knowledge at the level of the person, rather than the sub-personal, and within the ‘space of reasons’, the space of justifying and being able to justify what one says. [McDowell 1994: xiv, Sellars 1997: 76].

We motivate the account partly in response to ‘regress arguments’ offered by Gilbert Ryle and Ludwig Wittgenstein and defend the former against new intellectualist criticism [eg. Stanley and Williamson 2001]. But the support the regress arguments offer is nuanced. To take the latter, Wittgenstein considers teaching a pupil to continue a mathematical series correctly who ‘judged by the usual criteria… has mastered the series of natural numbers’ and is then taught to continue series of the form 0, n, 2n, 3n, etc. for arbitrary values of n

Then we get the pupil to continue one series (say “+ 2”) beyond 1000 – and he writes 1000, 1004, 1008, 1012. We say to him, “Look what you’re doing!” – He doesn’t understand. We say, “You should have added two: look how you began the series!” – He answers, “Yes, isn’t it right? I thought that was how I had to do it.” —– Or suppose he pointed to the series and said, “But I did go on in the same way”. – It would now be no use to say, “But can’t you see…?” – and go over the old explanations and examples for him again. In such a case, we might perhaps say: this person finds it natural, once given our explanations, to understand our order as we would understand the order “Add 2 up to 1000, 4 up to 2000, 6 up to 3000, and so on”. This case would have similarities to that in which it comes naturally to a person to react to the gesture of pointing with the hand by looking in the direction from fingertip to wrist, rather than from wrist to fingertip. [Wittgenstein 2009 §185]

The pupil reacts in a divergent way (from us) to explanations of how to continue. He appears to have acted under a divergent interpretation of the expression of the rule. But if to be successful an explanation of a rule has to be correctly interpreted then any expression of the interpretation might be subject to divergent understandings whether explicitly stated in a symbolism or by examples.

“But how can a rule teach me what I have to do at this point? After all, whatever I do can, on some interpretation, be made compatible with the rule.” – No, that’s not what one should say. Rather, this: every interpretation hangs in the air together with what it interprets, and cannot give it any support. Interpretations by themselves do not determine meaning. [ibid §198]

Since any interpretation has to be specified in some way, if understanding depends on an interpretation, no specification would be sufficient in itself. Hence the label ‘regress of interpretations’.

In an early work – Changing Order - on the role of tacit knowledge in science, the sociologist Harry Collins assumes that the regress argument shows that the correct way to continue, the correct notion of sameness, cannot be specified or said [Collins 1985].

[S]ince in spite of this we all know the correct way to go on, there must be something more to a rule than its specifiability’ [ibid: 14].

Thus there is a genuine gap between what is said in an explanation of a rule and what is understood when all goes well and something is needed to plug the gap so as to explain the otherwise mysterious ability to go on correctly.

The extra element is social entrenchment such that the sense of sameness in continuing in the same way is relative to a community. It underpins the ‘mysterious abilities that enable us to know when to continue ‘2,4,6,8’ with ’10,12,14,16’ and when with ‘who do we appreciate?’’ [ibid: 22, italics added].

Collins calls this ‘tacit knowledge’.

[T]he member of a social group who has the ability to continue the sequence ‘2,4,6,8’ with ‘10,12,14,16’ as a matter of course, without even thinking about it, also possesses something that the stranger to our culture and the newborn do not. This is sometimes referred to as ‘social skill’ but we can call it tacit knowledge without doing too much violence to the term. [ibid: 56 italics added]

According to Collins tacit knowledge cannot be put into words, is invisible in its transmission and capricious [ibid: 129]. This suggests a picture of rule following in which explanations are always inadequate but are patched by invisible and capricious tacit knowledge: the picture that Wittgenstein puts into the mouth of his interlocutor in order to reject it.

“But do you really explain to the other person what you yourself understand? Don’t you leave it to him to guess the essential thing? You give him examples a 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 mean” would amount to: “various interpretations of my explanation come to his mind, and he picks one of them”. So in this case he could ask; and I could and would answer him. [Wittgenstein 2009: §210]

There is, however, a different way to connect the regress argument to tacit knowledge. The moral of the regress of interpretations according to John McDowell is that:

We learn that it is disastrous to suppose there is always a conceptual gap between an expression of a rule and performances that are up for assessment according to whether or not they conform to the rule, a gap that is made vivid by saying the expression of the rule stands there… We must not acquiesce in the idea that an expression of a rule, considered in itself, does not sort behaviour into performances that follow the rule and performances that do not. [McDowell 2009: 100-1]

This, according to McDowell, is the moral of §201 of the Philosophical Investigations: ‘there is a way of grasping a rule which is not an interpretation, but which, from case to case of application, is exhibited in what we call “following the rule” and “going against it”’ [Wittgenstein 2009: §201]. For those who are party to the relevant practices, there is no gap between the expression of the rule and being told which way to go. A signpost, for example, not merely a signpost under an interpretation, points the way. On this account, the regress of interpretations is stopped before it can start. Hence, the expression of a rule is explicit in its directions.

But it is only explicit for those with the right eyes to see or ears to hear an explanation, for those, in Stanley Cavell’s phrase, who share the ‘whirl or organism’ of our form of life [Cavell 1969: 52]. This in turn grounds out in the applications of rules, which are context-dependent, conceptually structured and practical. The regress of interpretations argument shows that any informative explanation of what following a rule correctly, going on in the same way, comprises soon gives out. Any general account of the relation of rules or conceptual generalities to particular judgements falls prey to the regress of interpretations. What can be made explicit via the art of denotation – in Polanyi’s phrase – rests on the tacit, not because of a distinction between focal and subsidiary awareness but because going on in each specific way simply is accord with a rule or concept, is also context dependent and, by linking conceptual generalities to particular cases in judgements, is practical.

An ontology rather than epistemology of tacit knowledge?

Taking tacit knowledge to be person-level and conceptual (as well as practical and context specific) is not the only possible approach, however.

In his later Tacit and Explicit Knowledge, Collins also contrasts tacit knowledge with what is explicit [Collins 2010]. But the characterisation of the explicit is not as what can be expressed linguistically – as Polanyi and I assume – but rather with what Collins calls ‘strings’ [Collins 2010: 57]. Strings are ‘bits of stuff inscribed with patterns: they might be bits of air with patterns of sound waves, or bits of paper with writing, or bits of the seashore with marks made by waves, or patterns of mould, or almost anything’...’ [ibid: 9] The motivation for this is to avoid the ‘freight of inherent meaning that makes the notions of signs, symbols and icons so complicated’ [ibid: 9]. One repeated worry is that strings (and hence signs as kinds of strings) do not have an essential meaning: ‘strings are without meaning… a book is a physical thing, not a meaningful thing’ [ibid: 34]. So rather than presupposing meaning or mental content, Collins stresses the physical nature of strings: ‘a string is just a physical object and it is immediately clear that whether it has any effect and what kind of effect this might be is entirely a matter of what happens to it’ [ibid: 9].

But since strings are just ‘bits of stuff inscribed with patterns’ their elaboration, transformation, mechanization and explanation are all counted by Collins as instances of what is explicit. Hence the focus of the book turns out not to be the nature of the knowledge that rational subjects possess but the worldly objects of knowledge: the processes or tasks or worldly patterns themselves.

One example of this is the way that Collins, as I have, attempts to demystify tacit knowledge by connecting it to what is practical:

We are just like complicated cats, dogs, trees, and sieves… Sometimes we can do things better than cats, dogs, trees and sieves can do them, and sometimes worse. A sieve is generally better at sorting stones than a human (as a fridge is better at chilling water), a tree is certainly better at growing leaves, dogs are better at being affected by strings of smells, and cats are better at hunting small animals… if we were to stop talking and just get on with things — that is, if the tacit was not made mysterious by its tension with the explicit — there would be no puzzle at all about the body, per se. That teaching humans to accomplish even mimeomorphic actions is a complicated business, involving personal contact, says nothing about the nature of the knowledge, per se. (ibid: 104-5)

The final claim reiterates the idea that what matters to Collins – ‘knowledge, per se’ – does not depend on how humans think of or conceptualise tasks, something in the realm of sense, but rather on the mechanical process, in the realm of reference. The passage does not explicitly equate tacit knowledge with the behaviour of cats, dogs, trees and sieves but it does suggest that thinking of sieves sorting stones is a good way to demystify human tacit knowledge. Whatever the independent merits of an ontology of strings, however, it is far from an account of knowledge itself. It is pitched at the level of reference rather than sense and ontology rather than epistemology.

Polanyi was right to stress the importance of a form of knowledge at the level of the person, expressive of human minds, tied to context and practical and presupposed by, but exceeding, linguistic codification. The challenge is to articulate it.

References

Cavell, S. (1969) Must we mean what we say?, Cambridge: Cambridge University Press

Collins, H. (1985). Changing Order: Replication and Induction in Scientific Practice, London: Sage

Collins, H. (2010) Tacit and Explicit Knowledge, Chicago: University of Chicago Press

Dreyfus, H.L. and Dreyfus, S.E. (1986) Mind Over Machine, New Youk: Macmillan

Evans, G. (1981) ‘Semantic Theory and Tacit Knowledge’. In Holtzman, S.H. and Leitch, C.M. (eds.). Wittgenstein: To Follow a Rule, London: Routledge and Kegan Paul

Gascoigne, N. and Thornton, T. (2013) Tacit Knowledge, Durham: Acumen

Gellatny, A. (ed) (1986) The Skilful Mind, Milton Keynes: Open University Press

Grene, M. (1977) ‘Tacit knowing: grounds for a revolution in philosophy’ Journal of the British Society for Phenomenology, 8: 164-71

Guilick, W. (2016) ‘Relating Polanyi’s Tacit Dimension to Social Epistemology: Three Recent Interpretations’ Social Epistemology, 30: 297-325

McDowell, J. (2009) The Engaged Intellect, Cambridge, MA: Harvard University Press

McDowell, J. (1994) Mind and World, Cambridge, MA: Harvard University Press

Martin, R.D. (1994) The Specialist Chick Sexer, Melbourne: Bernal

Polanyi, M. (1962) Personal Knowledge, Chicago: University of Chicago Press

Polanyi, M. (1967a) ‘Sense-giving and sense-reading’ Philosophy 42: 301-325

Polanyi, M. (1967b) The Tacit Dimension, Chicago: University of Chicago Press

Polanyi, M. (1969) Knowing and Being, Chicago: University of Chicago Press

Ryle, G. (1949) The Concept of Mind, London: Hutchinson.

Ryle, G. (1945) ‘Knowing How and Knowing That’ Proceedings of the Aristotelian Society, 46: 1-16

Sellars, W. (1997). Empiricism and the Philosophy of Mind. Intro. by R. Rorty and study guide by R. Brandom. Cambridge, MA: Harvard University Press.

Stanley, J. and Williamson, T. (2001) ‘Knowing how’ The Journal of Philosophy 97: 411-444.

Wittgenstein, L. (2009) Philosophical Investigations, Oxford: Blackwell

Saying and showing in the later Wittgenstein

The echo of the Tractarian saying versus showing distinction in the later rule following considerations

2.172 A picture cannot, however, depict its pictorial form: it displays it…

4.121 Propositions cannot represent logical form: it is mirrored in them. What finds its reflection in language, language cannot represent. What expresses itself in language, we cannot express by means of language. Propositions show the logical form of reality. They display it. [Wittgenstein 1961]

For what we thereby show is that there is a way of grasping a rule which is not an interpretation, but which, from case to case of application, is exhibited in what we call “following the rule” and “going against it”. [Wittgenstein 2009 §201]

Introduction

The paradigmatic role of the saying-showing distinction in the Tractatus Logico-Philosophicus concerns the representational powers of pictures and propositions. The account of representation apparently articulated cannot, somewhat paradoxically, be put into words but something relevant can be shown. On ‘resolute’ interpretations, however, the distinction itself is part of what has to be jettisoned when the insight offered in the book is finally understood, an aspect of the ‘ladder’ to be thrown down. On a traditional reading, the distinction is what permits something substantial still to be conveyed even when the ladder is thrown down. On an elucidatory reading, such as McGinn’s, the distinction survives as an aspect of the continuity between the earlier and later Wittgenstein though McGinn offers as an example of that continuity only epistemic matters from On Certainty which is only of indirect connection to representation itself.

In this chapter, I discuss an analogy between the saying-showing distinction as it seems to apply in the Tractatus – whether or not it survives as genuine insight – and the rule following considerations in the Philosophical Investigations. But even in this latter context, difficulties remain.

On one approach, the rule following considerations highlight the gap between what can be set out in any explanation of a rule and what – philosophically – one took to be conveyed. Examples of deviant interpretations serve as the equivalent of ringers in arguments for epistemological scepticism. And hence so much the worse for ‘saying’. One response is a radical form of scepticism – Kripke’s meaning scepticism – in which there is less to rule following than previously thought. Another (eg Brandom) is to fill the gap between what is said and what is understood by invoking a layer of norms underpinning explicit rules implicit in practice. This suggests a role for the saying-showing distinction with norms merely shown in implicit practices underpinning what can be said, explicitly. Neither approach, however, fits what Wittgenstein himself says.

A third response is to deny that there is any necessary gap between an explanation and what is understood. The regress of interpretations only threatens on the assumption that the expression of rules ‘stand there’ like a normatively inert signpost and offer normative guidance only under an interpretation. This response undermines the suggestive argument for a distinction between saying and showing but, nevertheless, one remains at the heart of representation. The content of a rule, such as one governing application of a word, grounds out in particular judgements. For example, there is no further explanation of the fact that ‘blue’ applies to light and dark blue but not to green than that this is our natural practice. It depends, in Stanley Cavell’s fine phrase, on our whirl of organism. The content of our concepts is thus given by our practice. This is the aspect of the ‘logic’ of representation that has to be shown in the later Wittgenstein.

Saying and showing in the Tractatus, very briefly

On first appearance, the Tractatus Logico-Philosophicus concerns the representational powers of thought and language. Drawing inspiration from the practice in French courts to depict road-traffic accidents using models, the Tractatus seems to explore the idea that thought and language latches onto the world through a modelling, or more generally, picturing relation.

In the proposition a world is as it were put together experimentally. (As when in the law-court in Paris a motor-car accident is represented by means of dolls, etc.) [Wittgenstein 1979: 7]

Like the courtroom model, elements in language stand for elements in the world and their relations as set out in language – or in thoughts expressible in language – mirror those of the worldly correlates. Such a general approach – massively refined and extended throughout the body of the book – promises to shed light on such phenomena as that thoughts and utterances can be false as well as true, they can depict situations – putative facts – that do not obtain. Language is also generative: grasp of finite words and a grammar enables a speaker to understand an infinite number of novel sentences and, on this picture, thus know what would be the case if they were true.

But one consequence of the account – assuming for the moment that it is, as it first seems, an account – is that it best fits the relation of statements, thoughts or propositions to contingent facts. Logical statements such as tautologies and contradictions have to be accommodated as limit cases of the structures that enable the expression of contingent claims but resulting, strictly, in nonsense. Given the account of what is expressed by statements setting out contingent claims, tautologies and contradictions do not say anything, though their logical form is shown.

One consequence of this is that the very articulation of the conditions of possibility of the representational powers of language and thought, which seem to express the necessary structural isomorphism of language and world, runs counter to the account of sense that is, apparently, designed to deal with contingent claims. If the account of language apparently set out in the Tractatus were true, one consequence of that would be that the expression of that account in the Tractatus would be violate the bounds of sense and would be meaningless. As though to accommodate just that dawning realisation, the penultimate remark runs:

6.54 My propositions serve as elucidations in the following way: anyone who understands me eventually recognizes them as nonsensical, when he has used them—as steps—to climb up beyond them. (He must, so to speak, throw away the ladder after he has climbed up it.) He must transcend these propositions, and then he will see the world aright. [Wittgenstein 1961]

In fact, however, the idea that aspects of what makes the representational powers of language and thought possible cannot be stated runs throughout the work. For example:

2.172 A picture cannot, however, depict its pictorial form: it displays it.

2.173 A picture represents its subject from a position outside it. (Its standpoint is its representational form.) That is why a picture represents its subject correctly or incorrectly.

2.174 A picture cannot, however, place itself outside its representational form.

2.18 What any picture, of whatever form, must have in common with reality, in order to be able to depict it—correctly or incorrectly—in any way at all, is logical form, i.e. the form of reality.

4.121 Propositions cannot represent logical form: it is mirrored in them. What finds its reflection in language, language cannot represent. What expresses itself in language, we cannot express by means of language. Propositions show the logical form of reality. They display it. [Wittgenstein 1961]

In other words, a distinction between saying and showing runs throughout the seeming account of representation. Taking pictures as models, the one thing a picture cannot depict, the Tractatus suggests, is its own means of depiction. To do that, it seems, would be for it to pick itself up by its own bootstraps.

This brief account has been hedged about with ‘seeming’s and ‘apparent’s because of an exegetical controversy inspired by the difficulty just outlined and which can again be very briefly summarised. On a traditional approach, the conflict between the attempt to set out, in book form, an account of how representation is possible and the realisation that such an attempt undermines itself is uneasy and forces Wittgenstein to add to §6.54 the final comment:

7 What we cannot speak about we must pass over in silence. [ibid]

But on the traditional interpretation, while silence may be the proper conclusion of the exploration of representation, something really has been learnt along the way about how language and thought hook onto the world, something that pertains to the picturing relation misguidedly apparently described. Although strictly nonsensical, the account of the picturing relation gestures at a kind of metaphysical truth. Of this approach, Frank Ramsey memorably said that what cannot be said cannot be said and cannot be whistled either. Such an approach helps explain Wittgenstein’s later return to the same questions with a different style, at least, of philosophy.

More recently, a ‘resolute’ reading has pressed the idea of continuity between the earlier and later Wittgenstein, the latter being explicit that nonsensical utterances are simply utterances that lack sense: no use has been given to some particular combination of symbols [eg Conant 2000, Diamond 1991]. There is no paradoxical nonsensical sense. Nonsense cannot be used to encode any truths, effable or ineffable. Reading this view back into the Tractatus, it denies any notion of the communication of truths about the forms of representation via ‘substantial’ nonsense. The Tractatus serves as a kind of reductio ad absurdum of a philosophical venture. The correct conclusion from §6.54 is that, aside from itself and a few other framing propositions, the whole of the book collapses into nothingness. On this view, the distinction between saying and showing is not an insight but part of the house of cards to be collapsed.

More recently still, Marie McGinn has defended a third ‘elucidatory’ view [McGinn 1999 and 2001]. Very roughly, she attempts to preserve much of the content of the Tractatus while also accepting that nonsense really is nonsense. While it is a mistake to attempt to offer an account of how language and thought can represent the world as though from a perspective outside language and thought, the Tractatus can be used to shed light on representation from within. Hence, she argues that the saying-showing distinction can be defended as an aspect of how we need to think about representation from within it. Further, she argues that this distinction is an aspect of continuity with the later Wittgenstein though the example she offers concerns hinge propositions in On Certainty rather than claims more directly about the nature of representation.

In the rest of this chapter I will explore a reason to think that the saying showing distinction is a feature of the later Wittgenstein’s discussion of rules in the Philosophical Investigations.

The prompting problem in the Philosophical Investigations

To set out the role of the saying versus showing distinction in the later Wittgenstein, I will start with a familiar scenario from the discussion of rule following. Wittgenstein has an extended discussion of teaching a pupil to continue a mathematical series correctly. At one point, he imagines a pupil who ‘judged by the usual criteria… has mastered the series of natural numbers’ and is then taught to continue series of the form 0, n, 2n, 3n, etc. for arbitrary values of n

Then we get the pupil to continue one series (say “+ 2”) beyond 1000 – and he writes 1000, 1004, 1008, 1012. We say to him, “Look what you’re doing!” – He doesn’t understand. We say, “You should have added two: look how you began the series!” – He answers, “Yes, isn’t it right? I thought that was how I had to do it.” —– Or suppose he pointed to the series and said, “But I did go on in the same way”. – It would now be no use to say, “But can’t you see…?” – and go over the old explanations and examples for him again. In such a case, we might perhaps say: this person finds it natural, once given our explanations, to understand our order as we would understand the order “Add 2 up to 1000, 4 up to 2000, 6 up to 3000, and so on”. This case would have similarities to that in which it comes naturally to a person to react to the gesture of pointing with the hand by looking in the direction from fingertip to wrist, rather than from wrist to fingertip. [Wittgenstein 2009 §185]

In this example, the pupil reacts in a divergent way (as far as we are concerned) to the explanations offered of how to continue and Wittgenstein suggests one possible way of making sense of this: that the pupil understands the explanation in a systematically divergent way. In a word he uses a little later, the pupil appears to have acted under a divergent interpretation of the expression of the rule.

The example may seem to present a problem only of third person epistemology: of conveying or understanding from an explanation of a rule how to go on. But a problem also attaches to the understanding of the person doing the explaining.

“But I already knew, at the time when I gave the order, that he should write 1002 after 1000.” – Certainly; and you may even say you meant it then; only you shouldn’t let yourself be misled by the grammar of the words “know” and “mean”. For you don’t mean that you thought of the step from 1000 to 1002 at that time – and even if you did think of this step, still, you didn’t think of other ones. [ibid §187]

The more fundamental issue is not just the epistemology but rather in what understanding the series can consist, given that a proper understanding has to rule out divergences of the sort postulated. But if, to be successful, an explanation of a rule has to be correctly interpreted then any expression of the interpretation might be subject to divergent understandings. It does not seem to matter whether the explanation is explicitly stated in a symbolism or by examples such as the first few numbers of the series.

“But how can a rule teach me what I have to do at this point? After all, whatever I do can, on some interpretation, be made compatible with the rule.” – No, that’s not what one should say. Rather, this: every interpretation hangs in the air together with what it interprets, and cannot give it any support. Interpretations by themselves do not determine meaning. [Wittgenstein 2009 §198]

Any interpretation has to be specified and the original dilemma is still in play for this. Earlier in the Investigations, Wittgenstein introduces a yet simpler idea: signposts.

A rule stands there like a signpost. – Does the signpost leave no doubt about the way I have to go? Does it show which direction I am to take when I have passed it, whether along the road or the footpath or cross-country? But where does it say which way I am to follow it; whether in the direction of its finger or (for example) in the opposite one? – And if there were not a single signpost, but a sequence of signposts or chalk marks on the ground a is there only one way of interpreting them?... [Wittgenstein 2009 §85]

If one understands which way a signpost points then one knows it points in the direction of the ‘finger’ not the other way. But one might have taken it to point the other way. The signpost does not take a viewer by the throat as Achilles suggests that logic will take the Tortoise by the throat [Carroll 1905]. So it can seem that the signpost points only under an interpretation. But in what would grasp of such an interpretation consist? In this case, the most obvious candidate would be a mental image of a signpost with an indication of which way it points. How? By yet another mental signpost, perhaps?

This case makes the problem clear. In this key respect (which way it points), the inner signpost is no different from the outer one. If the outer signpost needs an interpretation then so will the inner one. And that threatens a vicious infinite regress, stopping short of completing which will leave the subject with no understanding of which way any of the – inner or outer – signposts point. It will leave no understanding short of impossibly completing the infinite series of sign-posts because for any sign-post all one knows is that if the next higher order sign post points left then so does this one and if it points right then so does this one, but as yet, one does not know which way that higher order sign points. For it, an equivalent conditional applies with respect to the yet higher signpost.

I suggest that the apparent gap in the example of §185 and thereabouts between what can be put into words by the teacher and how the pupil subsequently acts suggests one place to locate a saying-showing distinction in the later Wittgenstein. But, as I will argue, this is a mistake.

Different responses to the prompting problem in the Philosophical Investigations

Whatever their subtler differences, some broad distinctions can be drawn in responses to the prompting problem along a particular spectrum.

One type of response is to concede the apparent sceptical implication of Wittgenstein’s hypothetical case and to conclude that less is communicated or understood in an explanation of a rule than is typically, pre-philosophically, taken to be the case. In different ways, Saul Kripke and the earlier Crispin Wright take this line. According to Kripke, the correct conclusion to draw from what looks to be the starting point for an epistemological scepticism concerning the idea of shared grasp of rules is not that knowledgeable access to those facts is impossible but, rather, that there are no such facts. According to Wright, the applications made supposedly in accord with rules in fact shapes those rules [Wright 1986]. Rules are plastic to ongoing use. On such a view, the apparent gap between what is said in explanations of rules and what is understood is solved by radical pruning of what is understood. The idea that a rule is something that can be understood but which also determines ongoing use, sets a determinate standard for subsequent applications, is simply given up.

There can be no such thing as meaning anything by any word. Each new application we make is a leap in the dark; any present intention could be interpreted so as to accord with anything we may choose to do. So there can be neither accord, nor conflict. [Kripke 1982: 55]

Such a response does not support a saying-showing distinction, however. While it suggests that there may be disturbing limits to what turns out to be sayable, there is insufficient resource in what is shown as well.

A second type of response to Wittgenstein’s discussion is much more promising. It is to conclude that there are limits to what can be made explicit in an explanation of a rule but that the potential gap between what can be explained and what can be understood is plugged by something implicit. Robert Brandom takes this view. He argues that the regress of interpretations suggested by Wittgenstein highlights the role that applying any such interpretation correctly in situ must play.

No sequence of interpretations can eliminate the need to apply the final rules, and this is always itself subject to normative assessment. Applied incorrectly, any interpretation misleads. The rule says how to do one thing correctly only on the assumption that one can do something else correctly, namely apply the rule… Absent such a practical way of grasping norms, no sense can be made of the distinction between correct and incorrect performance-of the difference between acting according to the norm and acting against it. Norms would then be unintelligible. [Brandom 1994: 21]

Hence, according to Brandom, Wittgenstein’s argument shows the reliance of explicit statements of rules on something implicit: their application in practice.

Norms that are explicit in the form of rules presuppose norms implicit in practices…Norms explicit as rules presuppose norms implicit in practices because a rule specifying how something is correctly done (how a word ought to be used, how a piano ought to be tuned) must be applied to particular circumstances, and applying a rule in particular circumstances is itself essentially something that can be done correctly or incorrectly. A rule, principle or command has normative significance for performances only in the context of practices determining how it is correctly applied. [ibid: 62]

In support of this, he cites a passage that serves as a kind of diagnosis and conclusion of Wittgenstein’s discussion.

This was our paradox: no course of action could be determined by a rule, because every course of action can be brought into accord with the rule. The answer was: if every course of action can be brought into accord with the rule, then it can also be brought into conflict with it. And so there would be neither accord nor conflict here. That there is a misunderstanding here is shown by the mere fact that in this chain of reasoning we place one interpretation behind another, as if each one contented us at least for a moment, until we thought of yet another lying behind it. For what we thereby show is that there is a way of grasping a rule which is not an interpretation, but which, from case to case of application, is exhibited in what we call “following the rule” and “going against it”. [Wittgenstein 2009 §201]

The appeal to norms implicit in practices halts the regress below the level of what can be explicitly stated. As McDowell critically summarises, Brandom presents a picture: ‘in which, below the level at which speech can be described by saying which concepts are made explicit in it, it can be described as subject to norms specifiable in terms that do not require an understanding of the idea of explicitly expressing this or that concept’ [McDowell 2009: 97]. Brandom’s project – in Making It Explicit and elsewhere – is to give an account of this implicit level.

It must be possible to make sense of a notion of norms implicit in practice-which participants in the practice are bound by, and can acknowledge being bound by-without appeal to any explicit rules or capacities on the part of those participants to understand and apply such rules. [Brandom 1994: 26]

This is a challenge because, on Brandom’s view, the examples of divergent ways of going on exemplified by the pupil learning a mathematical series show that there is more to following a rule than exemplifying just any regularity. Hence ‘Such practices must be construed both as not having to involve explicit rules and as distinct from mere regularities’ [ibid: 29].

Brandom is not the only interpreter of Wittgenstein to suggest a gap between what can be said explicitly and what is implicitly understood. The sociologist Harry Collins suggests that the gap is plugged by a form of tacit knowledge [Collins 1985]. Charles Travis draws a distinction from the philosophy of thought and language between singular thoughts such as object-dependent thoughts and non-singular thoughts such as descriptive thoughts to suggests a gap between prior explanations of a rule and the object-dependent or singular understanding of it one might have when faced with a potential instance.

Some thoughts are only available to us given suitable acquaintance with our environment… 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... [Travis 2006: 130]

The issue it raises is how prior understandings can require novel ones. [ibid: 131]

Travis plugs this gap between two sorts of thought with ‘parochial sensibilities’.

In speaking to our fellows we speak to those expected to share (enough of) our parochial sensibilities, and to apply these in understanding what we say (in the ways they do). [ibid: 138]

Construing Wittgenstein’s discussion of rule following as suggesting a gap between what can be explicitly expressed in explanations and what is implicitly understood in rule-governed practice might suggest a role for a distinction between saying and showing in the later Wittgenstein.

The problem with that assumption, however, is that the arguments to the effect that the content of a rule cannot be said also counts against what can be shown. If the regress of interpretations counts against linguistic explanations of rules it also counts against practical explanations. Any finite stretch of practice could be subject to an infinite number of diverging interpretations as the example of deviant pupil suggests.

More generally, any reading of Wittgenstein that postulates, and then attempts to plug, a gap between explanation and understanding threatens the normativity of meaning since the same gap that seems to open up between teacher and pupil would also apply to intra-personally between an understanding at one moment and its application the future.

“But do you really explain to the other person what you yourself understand? Don’t you leave it to him to guess the essential thing? You give him examples a 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 mean” would amount to: “various interpretations of my explanation come to his mind, and he picks one of them”. So in this case he could ask; and I could and would answer him. [Wittgenstein 2009 §210]

This passage provides a route to the third response to the prompting problem which is to deny that there is a gap between explanation and understanding. A clear example is John McDowell’s paper ‘How not to read Philosophical Investigations: Brandom’s Wittgenstein’ quoted above [McDowell 2009].

McDowell offers an alternative view of the role of regress of interpretations to Brandom’s claim that it is the first stage of an argument to locate normativity implicitly in practice, below the level of what is explicit. McDowell highlights Wittgenstein’s use of the example of a signpost pointing in a particular direction as an example of a rule or the explicit expression of a rule: ‘A rule stands there like a signpost’ [Wittgenstein 2009 §85]. McDowell suggests that the danger of this image is that ‘“stands there” suggests a conception according to which the rule, or its expression, considered in itself, is normatively inert’ [McDowell 2009: 100]. Once that has been assumed, it can seem that what tells people which way to go is not the signpost itself but the signpost under an interpretation. But this initiates a regress because whatever serves as an expression of the interpretation of the signpost will also seem to stand there ‘like a signpost’ and be similarly normatively inert.

The moral of the regress of interpretations according to McDowell is that:

We learn that it is disastrous to suppose there is always a conceptual gap between an expression of a rule and performances that are up for assessment according to whether or not they conform to the rule, a gap that is made vivid by saying the expression of the rule stands there… We must not acquiesce in the idea that an expression of a rule, considered in itself, does not sort behaviour into performances that follow the rule and performances that do not. [ibid: 100-1]

This, according to McDowell, is the moral of §201: ‘there is a way of grasping a rule which is not an interpretation, but which, from case to case of application, is exhibited in what we call “following the rule” and “going against it”’. For those who are party to the relevant practices, there is no gap between the expression of the rule and being told which way to go. The signpost, not merely the signpost under an interpretation, points the way. On this account, the regress of interpretations is stopped before it can start. But it is stopped at the level of what is explicit. For signposts, the relevant rule or norm is the indicated direction in which one should go not some lower level norm expressible independently of that explicit rule.

This third approach to Wittgenstein’s discussion of rule following undermines one reason to think that the saying-showing distinction might apply to the later Wittgenstein: the assumption that the regress of interpretations, and the example of deviant responses to explanations of rules, places limits on the what can be made explicit or said and thus has instead to be shown. Despite initial appearances, the role of showing to complement saying is not, however, located in the supposed gap between explanation and understanding. On a proper reading of Wittgenstein, there is no such gap. In the final section, I will suggest where the distinction might lie.

Saying and showing in the Philosophical Investigations

As described earlier, on the account of language and world seemingly set out in the Tractatus, elements in language – names – stand for elements in the world and the relations as set out in language – or in thoughts expressible in language – mirror those of the worldly correlates in virtue of a shared pictorial or logical form. Names stand in, or go proxy, for objects of some form. Philosophical Investigations §§1-64 consider the possibility of establishing word-thing links through ostensive definitions. Ostensive definitions are, however, like any explanations of linguistic rules in that they could be misinterpreted. Only in the context of a great deal of stage setting is the content of the definition determined. This point is not made to undermine the pre-philosophical role of ostensive teaching using samples. Any explanation of meaning might be misunderstood but that does not undermine such explanation. Ostensive definition, however, cannot provide the foundation for language apparently presupposed in the Tractatus (modulo the earlier comments about its proper interpretation) because it would presuppose further meaning-related necessary stage setting.

What role then do samples have in ostensive definition? Wittgenstein suggests that the connection between a word and a sample is made within language. Samples play a role within the realm of representation rather than crossing between representation and what is represented.

What about the colour samples that A shows to B: are they part of the language? Well, it is as you please. They do not belong to spoken language; yet when I say to someone, “Pronounce the word ‘the’”, you will also count the second “‘the’” as part of the sentence. Yet it has a role just like that of a colour sample…; that is, it is a sample of what the other is meant to say. It is most natural, and causes least confusion, if we count the samples as tools of the language. [Wittgenstein 2009 §16]

[W]e explain that “sepia” means the colour of the standard sepia… This sample is an instrument of the language, by means of which we make colour statements. In this game, it is not something that is represented, but is a means of representation. [ibid §50]

Suppose someone asks what colour sepia is and, selecting a paint sample card, one says ‘This is sepia’. Such a comment is meta-linguistic. It is about the system of representation. The sample is not just an instance of sepia but serves – is being used here – as a representation of other instances of sepia in the world. Both the word and the sample serve as devices of representation: both have to be applied according to rules. To apply the sample, one compares it with, for example, particular flowers and judge of those that look sufficiently similar in the right lighting that they are ‘this colour’, perhaps indicating the sample. Hence one can also say that they are ‘sepia’, using that word.

In a different context, for example a quiz, one might say of an unlabelled paint card that it is sepia and that would be a judgement about that part of the world, about that instance of colour, not the system of representation. In this second context, the paint card is not being used as a sample and counts instead as a part of the represented world.

Returning now to the saying-showing distinction. Suppose that one did want to say to what colour the word ‘sepia’ applies in such a way as to cross the divide between representation and represented. One might say that sepia is a reddish-brown. But such an act of saying only gets one so far because one is still relying on a prior grasp of the meaning, the worldly application, of those words. One might hold up a sample of sepia. But if so, again, that presupposes grasp of how to use a colour sample to represent colours. For example, one shouldn’t hold it up in the dark or strongly coloured light, or compare it with objects by taste or size. How then can we articulate the connection between words that represent and things represented? All attempts at saying stay on the side of the system of representation when what was wanted was to bridge the gap between representation and represented. That they stay on the side of representation/representing is indicated by the fact that one needs to know how to use them in accord with a rule. By contrast, represented instances merely sit there and ‘take it’.

The way the gap is bridged is shown not said in the application of words to things in utterances that do not express rules. It lies in the unfolding practice of either directly applying a sample to objects such as flowers and judging that that is the same colour, that is: sepia, or by just saying of things by a reckoning of the eye that they are sepia. Such comments are not about the system of representation but about the world. But they show the method of representation in the ongoing practice of application.

The fact that there is something to be shown is made vivid by Wittgenstein’s examples of pupils who get rules wrong despite everyday explanations, such as the pupil who continues a series that a teacher had tried to indicate via an ordinary explanation of the plus 2 rule, 1000, 1004, 1008 etc. Their ongoing practice would show a different system of representation, whether or not it is coherent to us. They may agree with all our explanations, but they go on a different way. That helps illustrate that something substantial is being shown by our going on as we do, in the contrast with them.

Despite the widespread significance attached by commentators to the deviant pupil, there is a complication here. As Jonathan Lear argues, it is at least open to question what sense we can make of this example and others like it, such as the tribe who sell wood by area rather than volume [Lear 1982, 1984, 1986]. They may not actually be alternatives available for us.

But these counterfactuals cannot for us express real possibilities; for the notion of people being other minded is not something on which we can get any grasp. The possibility of there being persons who are minded in any way at all is the possibility of their being minded as we are. [Lear 1982: 386]

To the extent to which the brief sketches Wittgenstein offers might, under greater critical scrutiny, collapse into nonsense then they do not offer a substantial contrast and if that contrast is necessary to suggest the role of showing, so much the worse for showing. But other, more mundane, examples suggest themselves.

In the Blue Book, Wittgenstein considers the case of light and dark shades of blue.

There is something remarkable about saying that we use the word “strain” for both mental and physical strain because there is a similarity between them. Should you say we use the word “blue” both for light blue and dark blue because there is a similarity between them? If you were asked “Why do you call this ‘blue’ also?”, you would say “Because this is blue, too”. One might suggest that the explanation is that in this case you call “blue” what is in common between the two colours, and that, if you called “strain” what was in common between the two experiences of strain, it would have been wrong to say “I called them both ‘strain’ because they had a certain similarity”, but that you would have had to say “I used the word ‘strain’ in both cases because there is a strain present in both”. Now what should we answer to the question “What do light blue and dark blue have in common?”? At first sight the answer seems obvious: “They are both shades of blue”. But this is really a tautology. So let us ask “What do these colours I am pointing to have in common?” (Suppose one is light blue, the other dark blue.) The answer to this really ought to be “I don’t know what game you are playing”. And it depends upon this game whether I should say they had anything in common, and what I should say they had in common. [Wittgenstein 1958: 133-4]

Such a case is a good example of a view ascribed to the later Wittgenstein: the ‘autonomy of grammar’ exemplified in passages such as:

We have a colour system as we have a number system. Do the systems reside in our nature or in the nature of things? How are we to put it? - Not in the nature of numbers or colours. [Wittgenstein 1981 §357]

Just as samples are best thought of as part of the system of representation rather parts of the represented world so systems of concepts are not susceptible of a kind of realist explanation that their distinctions are dictated by distinctions brutely present in the world. And hence there is a role for an anthropological perspective on representational practices to chart the distinctions thereon [cf Lear 1986]. These distinctions are shown by the ongoing practices of speakers and thinkers in making the judgements they do, applying the rules they follow. They are, in Stanley Cavell’s famous description an aspect of the ‘the whirl of organism Wittgenstein calls ‘forms of life’’:

We learn and teach words in certain contexts, and then we are expected, and expect others, to be able to project them into further contexts. Nothing insures that this projection will take place (in particular, not the grasping of universals nor the grasping of book of rules), just as nothing insures that we will make, and understand, the same projections. That on the whole we do is a matter of our sharing routes of interest and feeling, senses 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’. Human speech and activity, sanity and community, rest upon nothing more, but nothing less, than this. It is a vision as simple as it is difficult, and as difficult as it is (and because it is) terrifying. [Cavell 1969: 52]

The worldly distinctions that correspond to a system of representation is shown by what we do.

This, however, suggests a nuanced role for the saying-showing distinction in the later Wittgenstein only partially mirroring its apparent role (on a traditional reading) in the Tractatus. The relation between language and the world is shown in ongoing practice which shows, to an anthropological perspective, the relation between a system of representation and what is represented. Any attempt to say in what that relation consists remains within the system of representation, forging connections from word to other words or words to samples, which are themselves best thought of as linguistic entities (because their use as samples presupposes rules of representation). But this is not to say, as some interpretations of Wittgenstein claim, that there is any gap between an explanation of the rules that comprise the representational system and what is understood that showing fills. The saying-showing distinction is distinction of perspectives rather than an attempt to communicate in mysterious way strictly ineffable insights.

References

Brandom, R. (1994) Making It Explicit, Cambridge, MA: Harvard University Press

Carroll, L. (1905) ‘What the Tortoise Said to Achilles’, Mind 4: 289-90

Cavell, S. (1969) Must we mean what we say, Cambridge: Cambridge University Press

Collins, H. (1985). Changing Order: Replication and Induction in Scientific Practice, London: Sage

Conant, J. (2000) ‘Elucidation and Nonsense in Frege and Early Wittgenstein’. In Crary, A. and Read, R. (eds.) The New Wittgenstein, London: Routledge

Diamond, C. (1991) ‘Throwing Away the Ladder: How to Read the Tractatus’, in The Realistic Spirit, Cambridge, Mass.: MIT Press

Kripke, S. (1982) Wittgenstein on rules and private language,Oxford: Blackwell

Lear, J. (1982) “Leaving the World Alone”, Journal of Philosophy79, 382–403

Lear, J. (1984) ‘The disappearing “we”’ Proceedings of the Aristotelian Society supplementary volume 58: 219-58

Lear, J. (1986) ‘Transcendental anthropology’ in Pettit, P. and McDowell, J. (eds) (1986) Subject Thought and Context, Oxford: Clarendon Press pp2367-98

McDowell, J. (2009b) The Engaged Intellect, Cambridge, MA: Harvard University Press, 2009b

McGinn, M. (1999) ‘Between Metaphysics and Nonsense: Elucidation in Wittgenstein’s Tractatus’, Philosophical Quarterly 49: 491–513

McGinn, M. (2001) ‘Saying and Showing and the Continuity of Wittgenstein’s Thought’, The Harvard Review of Philosophy 9: 24-36

Travis, C. (2006). Thought’s Footing, Oxford: Oxford University Press

Wittgenstein, L. (1958) The Blue and Brown Books, Oxford: Blackwell

Wittgenstein, L. (1961) Tractatus Logico-Philosophicus, trans. D. Pears and B. McGuinness London: Routledge and Kegan Paul

Wittgenstein, L. (1979) Notebooks 1914–1916, ed. G. H. von Wright and G. E. M. Anscombe, trans. G. E. M. Anscombe, 2nd edn, Oxford: Blackwell

Wittgenstein, L. (1981) Zettel, Oxford: Blackwell

Wittgenstein, L. ([1953] 2009) Philosophical Investigations, Oxford: Blackwell

Wright, C. (1986) ‘Rule Following, Meaning and Constructivism’, In Meaning and Interpretation, C. Travis (ed.), 271–97 Oxford: Blackwell