rules part 1 CAN WE KNOW RULES?
Wittgenstein on Certainty
Alec McHoul
Murdoch University
CONTENTS
Abbreviations
Preface
Part One: Are Rules Knowable?
References
ABBREVIATIONS
The following abbreviations are used throughout to refer to works by Wittgenstein:
BB The Blue and Brown Books (Oxford: Blackwell, 1969). 2nd edn. Followed by page numbers.
CV Culture and Value (Vermischte Bemerkungen) (Oxford: Blackwell, 1980). 1st English edn. Followed by page numbers.
LC Wittgenstein's Lectures: Cambridge 1930-32 (From the notes of John King and Desmond Lee) (Oxford: Blackwell, 1980). 1st edn. Followed by page numbers.
LE Paul Engelmann. Letters from Ludwig Wittgenstein with a Memoir (Oxford: Blackwell, 1967). 1st edn. Followed by page numbers.
LSD 'Language of Sense Data and Private Experience', (notes taken by Rush Rhees of Wittgenstein's lectures, 1936), Philosophical Investigations, 7, 1984, pp1-45 & 101-140. Followed by page numbers.
OC On Certainty (Über Gewissheit) (Oxford: Blackwell, 1974). 1st edn, corrected. Followed by section numbers.
PI Philosophical Investigations (Philosophische Untersuchungen) (Oxford: Blackwell, 1968). 2nd edn. Followed by section numbers for Part I and page numbers for Part II.
PR Philosophical Remarks (Philosophische Bemerkungen) (Oxford: Blackwell, 1975). Followed by page numbers.
RFM Remarks on the Foundations of Mathematics (Oxford: Blackwell, 1978). 3rd edn. Followed by page numbers.
TLP Tractatus Logico-Philosophicus (London: Routledge & Kegan Paul, 1961). Followed by section numbers.
Z Zettel (Oxford: Blackwell, 1967). Followed by section numbers.
The following abbreviations are used in Part Two only:
B&H G.P. Baker & P.M.S. Hacker, Wittgenstein: Rules, Grammar and Necessity (Oxford: Blackwell, 1985). Followed by page numbers.
L&R Norman Malcolm, 'Wittgenstein on Language and Rules', Philosophy, 64, 247, 1989, pp5-28. Followed by page numbers.
PREFACE
The problem of rules
Throughout this text, I am concerned with the problem of rules, and the ways in which it has been raised, treated, dismissed and neglected as a crucial problem in the social sciences, the humanities and cultural studies. If I tried to state, from the outset, the 'nature' of the problem — let alone the 'nature' of rule itself — this would unduly circumscribe the details of the investigations. And it would also unnecessarily restrict the ways in which you read them. To write anything is inevitably to exclude — no matter how much I might wish not to close off possibilities. Still, I do propose to offer some prefatory tales of caution about several common approaches to the problem of rules. I will refer to four approaches to, or discourses on, rule. The four are distinct, though we shall see that they overlap at several points.
1. The first holds that any discussion of rules necessarily complies with the aptly named 'standard rule-account' of social, cultural or linguistic action (O'Keefe 1979). On this standard account, rule is a cognitive phenomenon, like a mental signpost, directly informing the action in question. It is something located in the hidden personal grottoes of actors' minds. In Derrida's (1978) terms, the standard account considers rules to be a sort of 'Department of Internal Affairs'; or in Austin's (1975), a 'backstage artiste'. It begins with the unquestionable assumption that any text about 'rule' (about 'rule-governed' or 'rule-generated' behaviour) has to be of this kind.
Yet the 'standard rule-account' can be set amid a plurality of accounts. It can be thought of instead, against the grain of its own self-understanding, as a merely one discourse among others on the problem — as one which could never be said, by definition, to exhaust every possible formulation of the problem. It is nevertheless an account which, among others, will necessarily be repeated (albeit critically) in the space of this text.
2. A second approach, sometimes overlapping the standard cognitive account, takes it that any discussion of rules must propose the existence of a pre-given and relatively stable 'natural' structure. This structure has a variety of names. 'Grammar' is one; as is 'structure'. Another common one is 'the normative social order' (Wilson 1971; Filmer et al 1972). Here 'normative' refers to the idea that social actions are rule-governed in a single, once-and-for-all way. Hence it strongly suggests that the term 'rule' should be used in a prescriptive or stipulative way. This approach is common in structural functionalism (SF), the dominant paradigm of Western sociology. SF pictures a world in which rules emerge as apparently 'natural' features of an underlying morality, a shared social 'consensus' (Durkheim 1964; Parsons 1951). For me, SF offers simply another story about rule, among others. And my interest, especially in the first of these two investigations, is not to adjudicate between the stories. It is in how the various stories are told.
3. In a third approach, the term 'rule' has a slightly more qualitative ring to it. It refers to the substantive, propositional knowledge which a social or linguistic group supposedly holds and which allows them to act or speak. Transformational generative grammarians (Chomsky 1980) as well as some British philosophers (Moore 1959) have held such a propositional account of common sense. Effectively, this is a close cousin of the standard (cognitive) and normative (SF) accounts, for it considers rules to be 'deep' propositional forms of knowledge which supposedly lay the ultimate ground for the 'surface' bits of language and action observed in actual sentences and in everyday life. There is a definite tendency, here, to think of the deep rules as causes of the observed phenomena and, from time to time, theorists speculate on how these might be 'wired in' to fixed attributes of the human central nervous system. Again my initial interest is in how such discourses on rule are possible in the first place. I don't want to contribute the drawing of battle lines between them.
4. There is a further approach to rule, represented by ethnomethodology. Ethnomethodologists tend to reject the entire nexus of standard, normative and propositional accounts of rule (Wilson 1971; Zimmerman and Pollner 1971; Pollner 1987). They have shown instead how 'making reference to rules' and suchlike are quite routine everyday practices with no special privilege. Further, they argue that the methodical accomplishment of everyday affairs — such as offering accounts of rules — can be a topic for empirical investigation in its own right (Garfinkel 1967; Coulter 1979a; Psathas 1979). This prevents rules from becoming a special means by which social scientists can gain what they think of as a unique access to the causes of other people's social behaviour. So, to put it crudely, in ethnomethodology, the question of analytic rules is displaced in favour of an empirical analysis of 'rule-talk'.
In what follows, I want as far as possible to try to run counter to (or at least differently from) all four of these central 'theories' of rule. But still: so dominant is their effect upon anyone writing about rule today that I confess to being sceptical of complete severance. Any possible new position there may be is unclear to me. If it exists, it lurks within my own discourse like a kind of blind spot. And if it is new, it is only so in context and combination.
Firstly, contrary to the standard account and its almost dogmatic certainty about what rules actually are, I want to use the term 'rule' in an unremittingly secondary way, to cover whatever has already been held to ground social, textual or cultural orders of things. As well as being secondary and derivative, this use is also a highly interrogative rather than a prescriptive one. For the very question of whether there actually are such grounds (and if there are, whether they are describable, analysable or even utterable) underpins my investigation. The standard account, we remember, prefers to consider 'the grounds' as consisting of mental phenomena — mental states, for example. And so too does its propositional cousin in large part. I do not wish, here, to replace that notion directly by a less mentalistic and more materialistic account. Though if pressed, I would admit to favouring a materialist track. Instead, I want to ask questions about the status of the phenomenon of rule in the various analytic discourses, including — but not especially in order to diminish — mentalistic ones.
Moving on to the second, normative, approach: it could be said that I too start out from a distinctly normative position. This is true to the extent that everyone has to start somewhere. And a good enough place seems to be the widely-held and relatively innocuous presupposition that there is a social order (whose grounds may or may not be knowable), along with its attendant assumption that the task of the social sciences is to ask how that order is possible. My position is normative, however, only insofar as it remains reasonably sure that social, linguistic or cultural action cannot consist of just anything. There are, I think, limits — an inside and an outside, though not necessarily a centre — but their location (their locus) remains to be plotted.
Necessarily, then, I inhabit some common ground with several sociological, linguistic and cultural studies discourses. Along with them, and their varying accounts of the matter, I assume that there is something which renders possible the orders of things we investigate. This 'something' might be called 'conditions' in one discipline, 'determinations' in another, 'structure' in another and so on. I remain uncertain as to what it should be called, but refer to its features in as 'rules'. I do so in a piecemeal and perhaps inconsistent way, without definite commitments to specific positions (idealist, materialist, behaviourist, Cartesian...) on how the ensemble of these features might or might not be arranged. The best term I have at present, though it is still grossly inadequate and derivative, is: 'conditions of possibility' (Foucault 1979).
Within the loose parameters given so far, and only if pressed, I would have to say that what I am calling 'rules' are anything which might have been held as constituent conditions of possibility of a language, a society or a culture. Different disciplines have different collecting terms for referring to this domain in general: cognitive structure, the social order, social structure, tacit knowledge, common culture, stock-of-knowledge-at-hand, grammar, competence, methodicalness, paradigm, Weltanschauung, logic-in-use, bases of social conduct, means of production, mythic structure, performance rules, use(s), praxis and so on. The list multiplies almost daily as new disciplinary centres are effected and named. To be sure, I want to shift the terrain, but only with a gentle nudge and ask: just what is this production of centres? And the very fact of asking this question and not others shows how inevitably embroiled I am in the already-available discourses on rule.
But, after all, isn't this merely the same as asking the ethnomethodological question about rule all over again? Let us see if there are differences. Ethnomethodology's own shift of focus, when it asks 'how do rule-accounts work?' rather than 'what are the rules?', certainly displaces the question of rule. But, having done so, it then turns to a functionally similar question which is no longer just about rules but about the 'methodic bases' of talk/action — including among other things talk-about-rules. But, at the end of the day, are these 'methodic bases' any different in principle from the other disciplinary centres mentioned above? In his seminal statement about ethnomethodology's conceptual shift over the problem of rules, Garfinkel (1967, p33) insists that
All procedures whereby logical and methodological properties of the practices and results of inquiries are assessed in their general characteristics by rule are of interest as phenomena for ethnomethodological study but not otherwise.
In other words, ethnomethodology is not so much in the business of doing as of studying, for example, rule-attribution. Nevertheless, it is clearly in the business of doing the attribution of methodicalness to everyday actions and that too might just as well be referred to as 'rule'. Indeed much ethnomethodological work (especially its conversational analysis branch) has directly glossed this 'methodical work' in terms of rules (Sacks 1974) and their more general 'systematics' (Sacks, Schegloff and Jefferson 1974). Therefore, ethnomethodology, too, creates a 'whatever-it-is' which apparently grounds the orderliness of everyday activities. So it, too, might be investigated in terms of how it produces a specific disciplinary centre. But we should remember that any such inspection of ethnomethodology always risks being a repetition of ethnomethodological techniques. Critique of standard, normative, propositional and ethnomethodological discourses on rule can never situate themselves completely outside those discourses. They all make their appearance in this investigation. They are repeated here. But that repetition has a distinct modality of its own. That modality is signalled by one kind of interrogation which none of our companion disciplines routinely make as part and parcel of their inquiries: are rules objects of knowledge? Or, can we know rules at all? That is: when descriptions of the nature or the operation of rules are routinely called for in sociology or linguistics, for example, this always assumes their status as objects of analytic knowledge to be unquestionable. The investigation ahead, quite simply, questions that.
The use of On Certainty
And with regard to that question, Wittgenstein's position on rules has always seemed equivocal to me; a source of puzzlement and a field ripe for investigation. At times he appears very sceptical about the possibility of knowing rules at all. Yet, at other times, rules appear to be less problematic for him. At these points they appear absolutely fundamental to questions of linguistic conduct and also quite open to direct inquiry about their nature and operation. In the Investigations, for example, Wittgenstein writes that such fundamental rules do exist: '..but they do not form a system, and only experienced people can apply them right. Unlike calculation rules' (PI p227). Then, in the later text, On Certainty, even these more straightforward rules of calculation are put on the list of matters for critical inspection (OC 26-29). On this view, questions of rule — of applying a rule, of learning and following a rule and of conduct-in-accord with a rule — all seem to be still-open matters for linguists, cultural theorists and social scientists. And several commentators, for example Winch (1963), have read Wittgenstein accordingly. But there remains for me a sneaking suspicion that On Certainty could be read differently — in a way which potentially jeopardises the very expressibility and knowability of rules.
Part One of this investigation sets out to track that suspicion: it asks whether or not we should be sceptical about rules. Part Two takes up the other possible reading of Wittgenstein: that all rules are quite ordinary affairs — like the Highway Code or the rules of chess — and that to entertain any scepticism about them, even as one of two possibilties, is one of philosophy's many misleading temptations.
So the investigation in Part One of this book is first and foremost a reading of On Certainty; or perhaps a meta-critique for the reading is by no means direct. It asks how it is possible to read Wittgenstein on the question of rules and their knowability/expressibility in social science. Hence the object of the reading is not so much 'the text itself'; it asks what happens when we submit the text to a specific test: does it work for or against the knowability of rules? This is not to over-privilege On Certainty, to establish it as conclusive arbiter in the matter. On the contrary, it is simply to acknowledge that, in most other cases, our question is answered with a very firm, yet disturbingly silent, affirmative. (Wittgenstein's own Remarks on the Foundations of Mathematics, Part VI is one exception and it also informs Part One.) On Certainty, even if it can't answer the knowability question, at least promises to open up the space of that question. In terms of form, opening this interrogative space means inserting wedges between the paragraphs of Wittgenstein's text.
So while the investigation in Part One works upon selected passages of On Certainty, it is not in any usual sense of the word a commentary. (In fact the problems of commentary and the description of one text by another are among those which the investigation prizes out of OC.) It uses the raw material of On Certainty to generate its own product. To that extent selections have been made. All passage from about OC 300 are treated. This is basically the fourth part of OC (according to Anscombe and von Wright's preface) — give or take the odd paragraph — comprising the work Wittgenstein did in the last few months of his life. This fourth part constitutes, I believe, the text's most original contribution to questions of rule and knowability. The paragraphs before 300 are much more reminiscent of other Wittgensteinian work on rule. They are not unique to On Certainty, and so I have been much less methodical in my treatment of them. They set the scene, establish a certain vocabulary and then drop out of view.
What is at stake here? Simply this: if rules are inexpressible, a good deal of work in the social sciences and the humanities fails by its own criterion that it should be able to discover rules, their mode of operation or their local discursive equivalents. Put another way: if rules are not open to description, those disciplines which claim to express them must actually be doing something quite different. But what?
Faced with such an impasse, it would be very tempting to lapse into nihilism. But the real problem, as I see it, is to discover how such nihilism may be rethought through an inspection of the minute steps by which the spectre of inexpressibility arrives to block the way. Essentially, then, mine is an arguably positive reading of On Certainty; one which tries to reach into the particular and almost infinitely proliferable consequences of a general problem. Wittgenstein succinctly expressed this problem in the notes now called Culture and Value:
The limit of language is shown by its being impossible to describe the fact which corresponds to (is the translation of) a sentence, without simply repeating the sentence. (CV p10)
Perhaps what is inexpressible (what I find mysterious and am not able to express) is the background against which whatever I could express has its meaning. (CV p16)
That is: in any description 'of the world', is there anything outside of, beyond, or without citation or repetition? Can the language-game ever be treated — as empirical analysts demand, almost insisting thereby on a medical model of treatment? Or can it only be quoted/repeated? Can the rule be inspected from 'outside' itself or only followed from 'within' its confines? Is there any difference-from-the-social which social science absolutely requires before it can produce utterances about (both 'around-and-about' and 'with respect to') the social? What is fundamentally in question here is the very possibility of description itself.
Turning again to matters of form: in Part One, each selected paragraph from On Certainty is referenced by its original number as a brief heading for each set of remarks. Some readers may wish to inspect the relevant paragraphs in the interstices between my remarks. Others may wish to read the remarks in the interstices between Wittgenstein's paragraphs. I cannot not stipulate in advance which will turn out to be figure and which ground. The remarks can be read in either of these ways as leading to or following from the relevant paragraphs of On Certainty; or they can be read in relation to each other. Where I think one set of my remarks inevitably follows from the previous (but nevertheless results from a fresh paragraph of On Certainty) the first set is followed by a dash to mark continuity.
Some remarks appear in double parentheses. These subsections branch out from the text. They are a kind of footnote or marginal comment but I have included them inside the 'main' text for simplicity of access. At the conclusion of a parenthetical section, the sense of the text immediately prior to it is picked up, or else a fresh remark is started.
Apart from the above notes — which do no more than invite the reader to have a copy of On Certainty to hand — I leave the use of the text to the reader as far as possible.
Origins and destinations
These investigations represent more than ten years work. They began as discussions with Ian Hunter in 1979 at Griffith University concerning problems in discourse theory and discourse analysis. At that time Ian and I conducted an honours seminar in the School of Humanities called 'Oral Discourse'. Things appeared to have reached something of a head with the problem of the status of rule and its possible distinction from the discursive effect that I call the 'rule-expression' (cf. the remarks on and following OC 476). Ian and I were asking: how can we make claims to have 'uttered the rule' separately from having merely repeated it in order to produce the 'rule-expression'? It was at this point that Ian raised the question that pervades this investigation: are rules objects of knowledge? It seemed to me that the entirety of social scientific work on rules depended upon a positive answer to that question. Not least of my worries was that I began to see how my own work in conversational and textual analysis was deeply under threat from a negative answer.
The studies in Part One were begun in note form as early as 1979 at Griffith University and continued on a smaller scale in 1980 at the University of Wollongong. Since 1981, I have picked up the threads again at James Cook and Murdoch Universities. Comments by John Frow and Horst Ruthrof and Bob Hodge have been invaluable in these later stages. An earlier and much rougher draft, including some unfortunate textual errors (all my own doing), appeared in the 'Monographs, Working Papers and Prepublications Series' of the Toronto Semiotic Circle (Volume 1, 1986). I am deeply grateful to Clive Thomson at Queen's University for his help with this version.
Part Two of this investigation ('Are rules social?') is effectively a coda to Part One. It reinspects the question of the knowability of rules in the light of recent debates between leading Wittgenstein scholars about rule. It is also a critique of the scepticist impulse which drives Part One, turning instead to the possibility that 'nothing is hidden' (Malcolm 1986), that rules are knowable as the most ordinary things in the world. Perhaps rules are knowable because they are intrinsically social. But some readers of Wittgenstein have contested even this, suggesting a non-social sub-set of rules (Baker & Hacker 1985). Part Two was written while I was Honorary Research Fellow in the Department of Sociology at the University of Manchester in 1989 and it was circulated there as a 'Manchester Sociology Occasional Paper'. I am very grateful to Rod Watson, John Lee, Wes Sharrock and David Hatch for the time and effort they put into helping me rethink the question of rule.
Fremantle, WA
September 1990
PART ONE:
ARE RULES KNOWABLE?
We must begin with fragments...
26-29
Here we are reminded that, whatever we might say or write about rules, there is always a question of their indeterminacy and their defeasibility.
43-49
That there are rules is enough. (But contrast the description of language-games in OC 56 and 82.) We could say instead, 'The rules can be displayed but not formally written, shown but not said'. And so: we can't simply go ahead and give or tell the use, or the rule. We can show how it's done — and so teach it. We can teach it by practice. Compare the position of the Tractatus: we can't go ahead and give or tell the form of the proposition, but we can show it. From the Tractatus to On Certainty Wittgenstein's position on the limits of language remains effectively the same. There are, in effect, two related versions of 'whereof we cannot speak...'. What I want to show is that the later text doesn't work in quite the same way; that it does not involve exactly the same sorts of embargoes on certain forms of discourse as those to be found in the Tractatus.
From 'We could say instead...' to '...teach it by practice': this is only one reading — one which Hunter favours, I think, when he asks whether rules can be objects of knowledge. To begin with, and as a kind of heuristic only, I would like to show, or rather to read OC to see whether it can be read as showing, that rules can be written but also that the writing of them is primarily an activity, a language-game pursued under the compulsion of another rule. I want to know whether Wittgenstein can be rescued for anthropology in general. But it may turn out not to be possible.
The ominous question, for everyone from Chomskian grammarians, Parsonian functionalists, Malinowski and Lévi-Strauss, through to transcendental phenomenologists, logicians and conversational analysts, is that rules may not be objects of knowledge.
Why do I want to preserve the idea? Because I have no idea what work I could do on the world without it. For now, the alternative seems to be nihilism — perhaps of the kind implied by Silverman's chapter on Method/Rule in Reading Castaneda (1975).
Hunter wants to find a way of reading Wittgenstein which removes 'the myth of the knowability of rules' — but he also wants to find some way of working without it. 'That rules can be known' — this is such a ground of action for me that I get a kind of vertigo imagining what could be done on any analytic front without its support. The only therapy for that vertigo is that OC (for example, 375) seems to abstain from judgments about grounds. 'The standard has no grounds!' (PI 482). If OC also says that rules cannot be objects of knowledge, then one must choose between readings. It would be stranger for me to imagine a text or a reading without contradictions.
61-65
Here the connection which PI makes between rule and meaning is repeated. This immediately widens the scope of our investigation. The question of rule slides into the question of meaning just as it slides into the question of objectivity. It is no more than a point of departure for questioning a whole form of life — the social sciences.
86-87
Between forms of life, grounds may shift. An 'hypothesis' in one becomes a 'ground' in another. But can any actual investigation ever occur between forms of life (for instance, between science and common sense)?
88
Could there be a common universal set of 'foundation' propositions for all forms of life — not pertaining to a particular but to all forms? Given 86-87, doesn't this depend on who 'we' are and whether 'inquiry' is one of our language-games?
93-95
'My picture of the world...' — and aren't there a number of these? (See OC 422 on 'Weltanschauung'.) For a picture of the world could never be uniquely mine. The grounds for this picture are like rules (but...).
95-97
Within a form of life there may also be a switch of grounds; an exchange of grounding for empirical propositions and vice versa. This would have to be so with a mythology (einer Art Mythologie).
99
What is 'hard rock' here? Again: temporarily hard rock, only relatively hard rock or what?
102-103
I cannot describe the system of my convictions: but they do have a system. They are so anchored that I cannot touch them.
105
The 'element' (Lebenselement) — Garfinkel is reputed to refer to this (Tyler 1974) as though he were the fish that 'discovered' the water; as though all the rest were blind to something absolutely essential.
110
'An ungrounded way of acting': I should not know how to take this.
111
The general rule of multiplying cannot be given — rules for particular multiplications can. We might then have to say: there are distinct kinds of rule — ground rules and empirical rules. Perhaps ethnomethodology and conversational analysis deal only with the latter and that's why, as Mehan and Wood (1975) say, no one has ever done any empirical work on the more general 'background expectancies'. When Hunter objects to the idea that rules can be objects of knowledge, does he mean this to apply to both 'levels' of things we might call rules? In defence, we should have to say that the former cannot be known (described) while the latter can. (I am still looking for escape clauses.) —
116
These 'levels' might be: that which stands fast (Rules1) and that upon which I act right here and now (Rules2). (Deep and surface structure?) And then the problem is whether or not the former may be formulated discursively. Is an alternative way of working, then, one which is entirely confined to 'surface structure'? This at least I can (say I) know. With the other: it simply stands fast for me.
121
Can one say this? So far: no, not quite. But we can say: 'Where there is no possibility of doubting, we are dealing with something which is not properly called 'knowledge'. This destroys the idea that 'background assumptions', 'tacit knowledge' (Polanyi) etc., are describable phenomena. It may also be why terms like 'problematic', 'paradigm', 'Weltanschauung' etc., can be neither adequately defined nor empirically investigated.
What of the case in anthropology; say, where what a tribe could not doubt, I should have to? For instance, that chickens' entrails can foretell the future (the Azande). Here, I may doubt and so I may properly speak of knowing this. The Azande, on the other hand, cannot doubt. It would not make sense in their form of life to doubt this. Does this mean that I have a privileged position; is there something I know (about them) which they cannot? And presumably vice versa? (This is why Castaneda might be able to know Don Juan's way of knowledge but why Don Juan can only display it.) Wittgenstein's formulation would seem to allow for this. Hunter might consider it a problem of 'privileged knowledge'. But is it? That is, even at the level of Rules1 (that which stands fast), there may be possibilities of access and description where forms of life themselves show no, or little, family resemblance. Isn't the case with Chomsky's claims somewhat different from this? This difference shows that there is a problem in Hunter's question; and the problem is that it seems to require a sort of universal unspeakability. Moreover, isn't the case of, say, Wieder's (1974) description of the 'convict code' different again? But how? And aren't there a number of these differences within, say, all the things we call 'ethnomethodology', 'text grammar', 'narratology' or 'structural myth analysis'?
Which position would best be called 'relativism' here: mine, fuzzy as it is, or Hunter's?
126
And this system could be described — like a system of knowledge, unlike the system of certainties which ground our doubts and our knowledge (see OC 115 and 102-3). These are different systems-with-rules: Rules2 and Rules1 respectively. (See the remarks on OC 116, above.)
131
The grounds of our judging are complex convictions which underpin how we experience. They form a system, a culture and have rules. But those rules are not objects of (our) knowledge in the form of life which they underpin. Believing he can say something about these, Garfinkel (1967) tries to make his own form of life 'anthropologically strange'. Wittgenstein appears to be ruling this tactic 'off limits'. But presumably the same problem would not occur with 'empirical' rules (Rules2) such as those described by Harvey Sacks.
134
Object constancy: 'Experience always proves the law of object constancy'. Well, this is true in forms of life where experience is governed by such a law (which is of the Rules1 kind). I put something down — say, a book — and turn away. I turn back and it has gone. In my form of life something other than the book's 'merely disappearing (of its own accord)' has occurred. But such a weird explanation might be perfectly all right in a form of life which does not count object constancy among its Rules1 (its system of assumption). It just happens that this is close to the last explanation we should give. You can't find your car keys even though you 'definitely' left them on.... When they turn up elsewhere, you are neither disturbed nor do you consider them to have moved by themselves. There could be a culture, again, where this consideration is expected and expectable. We 'don't operate this way' but it has nothing to do with our empirical rules of conduct. These come (analytically) later — though they may be learned first.
135
With these things: the 'disturbing' does sometimes happen. We can reach the limits of 'reasonable explanation'. Pollner gives some examples in 'Mundane Reasoning' (1974). I think what happens in these cases is that the Rules1 (for example, that what has happened will happen again or that objects remain constant) do enter into our reasoning — though in a peculiar way. Perhaps we 'become aware' of them even though it would make no sense to doubt them. Otherwise — what are Wittgenstein and Moore doing here? How are they able to chew over the nature and status of these things? For example, how is Wittgenstein able to say they form a system if they are not accessible to knowledge and description? Wittgenstein is once more pushing at the limits of language. But there is nothing to say this doesn't go on in 'ordinary' settings, outside the work of philosophers. Here, throwing away the ladder is not a possibility. Here, even to possess a ladder is to accord a kind of privilege to oneself.
136
Doesn't the peculiar logical role of Rule1-type propositions in some way distinguish them from other sorts of propositions, empirical proposition, for example ('Moses led the Israelites' etc.)? Is it correct to call Rule1-types 'empirical'? For they seem to underpin our decisions as to, for example, what might and what might not count as an empirical proposition. However we cannot perform this operation on them, in turn.
139
'Our rules leave loop-holes open, and the practice has to speak for itself'. This shows there is something about engaging in a practice which is more than the following of rules — a supplement perhaps. (Therefore we should not conflate rule and practice.) But it would be misleading to think that this supplement consisted of 'rules for following rules'.Practice = rule + application;
but the question of 'application' is problematic. (We must be thinking of Rules2. What would be a loop-hole in 'The earth has existed long before my birth'? In passages like this we have to consider Wittgenstein to be discussing rules as of the Rules2 type.) The question of application has to do with what Wittgenstein calls 'examples'.
140
The use of the word 'rule' in 'Rules1' is, in a certain sense, inappropriate. (And yet it is this kind of rule which I think Hunter is addressing when he speaks of unspeakability.) —
143
What the child swallows down in this example is some aspect of the complex and mutually supportive total system of convictions underpinning and anchoring the form of life she is learning to practise — and still we want to call these 'rules' (= Rules1), perhaps wrongly. Wittgenstein is closer to the mark in OC 116 when he refers instead to what 'stands fast'. In 141-144, he uses the term 'believing'. But isn't believing (like knowing and doubting) a set of practices which is also itself underpinned by 'what stands fast'? Needham (1972) raises the same question and shows that it can't be answered satisfactorily. The question of 'belief' is open to investigation. I should have reservations about uncritically substituting 'what I believe' for 'what stands fast for me' (in Wittgenstein's sense). There are practices of believing but not practices of what stands fast, and so on.
150
The point of not-doubting (Nichtzweifeln): on one side of the point lies what stands fast (Rules1), on the other practices (Rules2 + applications). If my bifurcated reading of the question of rules in OC works, then Nichtzweifeln is its fulcrum. A mechanism like this ought to help us decide the status of, say, Chomsky's rules, among others. This piece of meta-theoretical equipment looks initially as if it can do various sorts of work. We may not be able to address or describe the unspeakable — but we may, paradoxically, be able to decide what does and does not count as such when we meet it in practice. There seems, at the very least, to be a way of asking whether or not some account of rules or practices can be given.
152
But 'what stands fast' can also be discovered. That which is not-doubtable is at least open to this. Can its systematicity be described though? It seems that many working with the problem of 'culture' have tried to do this. And instead of a description we get a word for the totality: 'Weltanschauung', 'world vision', 'Weltbild' (OC 162), 'Zeitgeist'; 'totality' itself is an example, one also used by Lukács. Other examples are: 'the complex and mutually supportive total system of convictions underpinning and anchoring the form of life' which I used above; and Wittgenstein's 'substratum' (OC 162). We seem to be able to endlessly proliferate these synonyms but can we do investigations of the phenomenon they address? Is it an object of knowledge? Now Hunter may be right. Nevertheless: there do seem to be some things we can actually say in this province: Wittgenstein says them.
In 152: how does this discovery come about? Certainly not by the 'method ... of enquiry' referred to in OC 151. Wittgenstein has already said that this is not to be arrived at by an investigation (OC 138). Yet, 'I can discover...': and this seems to be what cultural studies has been singularly unable to do.
154
This is how I feel about Hunter's doubt about rules as objects of knowledge. Does this tell us something about forms of life which only have the minutest differences? This difference seems to radically affect our respective modes of investigation. Until now: rules being objects of knowledge has lain beyond the point of Nichtzweifeln for me. Now I come across a form of inquiry which doubts this. What should I do? If I begin to doubt, shouldn't I need grounds for doubting — something which stands fast and makes the doubt a reasonable one? And why shouldn't there be a form of inquiry which doubts this new ground? And so on. Either this or one says: 'No! The problem is to stop this recursion. The question of rules has to be cleared up — brought to a halt.' What to believe here?
163
With the final question: No, you are not 'to say' it — this is simply how we operate. But we can imagine things otherwise. Yet what we imagine will not be scientific procedure as it's usually conducted. So here: we may actually be saying something important about scientific procedure and how it's usually conducted, after all. And this could be vital, say, for an anthropologist investigating a form of ethnoscience where such a proposition (or another like it) does not happen to stand fast.
Elliot (1974) used this very proposition, for example, to address the relation between common sense and scientific procedure. There are forms of inquiry where such a 'discovery' is significant — where it has currency. And these forms of inquiry have acknowledged the difficulty raised in OC 166. (Remembering the difficulties also raised in the above remarks on OC 143.) Thus we might say instead: '...the groundlessness of our grounds'. (Glauben here might better be translated as 'trusting'.)
167
Some more synonyms: 'matter of course foundation' (selbstverständliche Grundlage), 'that which goes unmentioned'. These all stand in for 'Rules1'. And OC itself appears as a complex mention of these.
So, disregarding problems of 'privileged access' and so on, we still have to ask 'unmentionable for whom?' And the answer looks like, in the particular case, Lavoisier and, in the general case, those who would operate smoothly and unproblematically with such a foundation.
And (despite the reservations expressed above) it's here that Althusser's term 'problematic' comes to have a certain usefulness. Like Nichtzweifeln it seems to establish a point or fulcrum at which 'what stands fast' (Psf) is divided from 'what may become a problem of investigation' (Pi). And here, for instance, it would be useful to distinguish the problematic of some natural sciences from that of some philosophies of science where there is a definite chain of relations:
This shows how we can get quite 'expressible' accounts of rules. But a shift of forms of life is involved. The relativity of these shifts is displayed by a further considerations. Anderson (1972), for example, uses such chains as the one above to construct a philosophy of science. There is an inescapable problem in this though, for the description can subsequently be seen, with a very slight shift of the Psf/Pi division, as part of the very phenomenon it describes. It's this sort of shift that Wittgenstein addresses re. 'mythology' in OC 95-99. And these shifts happen. And they do not necessarily carry along with them problems of 'privileged access'. (The problems are manifold enough as it is.)
All accounts will have this feature: that, working from some matter-of-course foundation other than that which underpinned the production of the account, the account will be readable as a part of that which it sought to describe. For 'what is part of what' will shift with the foundations upon which such judgments are made. So I do not intend to say that we can 'break free' of this. Furthermore, it would be with such an assumption (that there can be pure description ungrounded in any form of life) that we should encounter the problem of self-ascribed privilege. This appears to me as the source of such troubles and not the fact that one set of grounds can be used to describe practices (etc.) performed on the basis of another set. We should be clear about this distinction.
((The psychologist measures his best student's IQ. It gets calculated as, say, 110. The psychologist can doubt that this is actually the student's IQ, but that his tests actually measure IQ is, for him, beyond doubt; and likewise that IQ is actually this or that. A cultural anthropologist, on the other hand, may raise a whole volume of questions about such matters. But to regard his objections as privileged is nonsense. For the anthropologist, IQ tests are simply part of the phenomenon they attempt to describe — and, again for the anthropologist, that phenomenon is not IQ but the social relations of tester and tested. Here it's not a question of which is the truth but rather one of which form of life is practised. However, when the psychologist insists that his tests purely describe a certain phenomenon 'in the world' regardless of his discipline's assumptions, grounds, etc., then we must acknowledge a certain arrogance on his part. But the same would be true if the anthropologist made a similar claim about her version of IQ tests. The question is: whether they could enter into dialogue. And it's here that Feyerabend's (1975) notion of incommensurability gets a foothold. Many (but not all) 'critiques' and 'replies' are not exchanges within the same language-game. They talk or write past one another. A lot of what passes for critique in the social sciences is not 'about' the work critiqued but is, rather, the working through (repetition) of some other position. The Psf/Pi chain is so tight here that one position's analytic resources will be topics of investigation for a host of others. Because most social sciences both undertake investigations of human conduct and are instances of that conduct, a good deal of theoretical and methodological work becomes indistinguishable from empirical investigation. Malinowski is as open for inquiry as Trobriand Islanders.))
185
Still using the (problematic) terms 'Rules1' and 'Rules2': it is, in this example, noteworthy that Wittgenstein talks of 'doubting our whole system of evidence' — a Rules1 phenomenon. Previously: doubt itself involved practices anchored by such systems. Here: it seems as if doubt is possible — but we should imagine someone doing this only when working with some alternative 'system of evidence'. And this seems to be easier than doubting some empirical item given by a system of evidence (a Rules2 phenomenon). We might more sympathetic towards a flat-earther who completely rejects our system than to a serious historian whose only departure from normal historical evidence was that he doubted the existence of Napoleon. The former is 'anthropologically interesting' while the latter is simply suffering from dementia. The usual view (OC 186) seems more reasonable; but is it?
189
For Wittgenstein, description (of rules?) seems possible. Explanation would be a means of conducting ourselves with a system of Rules1 (and it would involve the pragmatic following of the corresponding Rules2). But when it comes down to the possibility of a multiplicity of such sets of Rules1 (systems), we are no longer engaged in explaining as such. But perhaps we can describe (systems of?) rules. This does not mean that our descriptive language is ungrounded in such a system. It would simply mean that form of life (f1) could not explain form of life (f2) — yet it could describe it. And it seems reasonable to hold now that a description of a set of Rules2 (for example, in conversation analysis) is fairly unproblematic — so long as its status as a description is kept in mind and we don't inflate it into an explanation. But Rules1?
191
Here we are going around in a circle because what counts as 'reality' or 'the facts' is given by the form of life which 'decides' whether one of its propositions is or is not in accord with these. This is where justification comes to an end (192).
195
Mistakes are made once we accept a form of life. They are not made about the 'givens' which anchor a form of life we have accepted. Hence the Einsteinian cannot say the Newtonian has made a 'mistake' in his calculation of a time/velocity equation. All she can say is that different equations were used and that they (or their paradigms) form the very substance and foundation of the Newtonian universe. Here there are no questions of mistakes. Though perhaps there are questions of what Berger (1966) calls 'alternations' (changes of conceptual foundations).
((I wrongly accuse Althusser of making a 'mistake' about the nature of reading and so forth (McHoul 1980).))
204
I tend to speak of a form of life as if it were a way of seeing; and of the games in that form of life as if they were — by contrast — ways of acting. But the form of life is only the family of related games. It is our acting with the empirical Rules2 of certain language-games that shows our adherence to the 'deep' Rules1 of a form of life. And again this goes to show the peculiarity of calling the latter 'rules'. If anything, it is they that are founded on the (more knowable) Rules2 of the language-games and their application.
To say that another culture 'sees' the world in a different way is only to say that it engages in different practices. We can't begin by describing the 'seeing'. Or indeed, pace Althusser, with the not-seeing. Castaneda's lunacies at least contain this important lesson — 'If people did not sometimes do silly things, nothing intelligent would ever get done' (CV 50). Looking for Rules1, for what lies at the foundations of a culture, is an absurdity as a 'pure' exercise. Describing cultural praxis is the possible (doable) version of this. (Though I am inclined to believe that this too is a repetition.) Why do we feel that there's something deeper? Something more general but something which we can't address? (For example, background expectancies, tacit knowledge and so on; a systematically describable Weltanschauung.)
((It is no more than a useful heuristic — pace Plato and Durkheim — to think of this foundation as consisting of objects.))
209
And now Wittgenstein refers to a 'picture' founding my practice. So it is again very tempting to think of a way of seeing (Bild/Sehen). 'Foundation', 'mythology', 'grounds', 'picture': all of these must be metaphorical. And so must be 'Rules1' and 'the systematicity of our system of rules of action'. Yet there is something binding and firm here. And given 204, it seems to be a dense texture woven by our relatively regular actions in a form of life (another metaphor). But what gives it its strength? Not a single structure or 'forma'; not a single thread running through the whole 'matrix'; but the complex overlapping, the wefting and warping of our actions themselves (PI 67). Trying to act as if the earth only existed for the last 100 years is like trying to weave a thread against the weft and warp of a form of life. But we can imagine a whole fabric woven this alternative way — only, for now, all the other supporting threads are missing.
((And here it's also simple and dangerous to resort to concepts like social structure, linguistic structure and so on. Yes: it's 'there' plain as day but the structure-as-such has no independent existence. Yet talking this way still has its uses — for example, when it's the structure that we want to change. Or rather: when we want to talk about changing it — for change-in-practice will mean coming to terms again with the complexities of the weave.))
210
To change a society is to change its matter-of-course foundation. In which case, the 'much' that 'seems to be fixed' has to be shunted back into use (from whence it came). But the possibility of doing this already presupposes the existence of a form of life where it is already in play. And that form of life is identical with what we would establish by change. The discovery of new ways of speaking and writing is already revolutionary.
((It seems to be a fundamental error to conceive language and social structure as entirely different sorts of objects. It is not as if the relation between them were one of base and superstructure and it's not as if speech, writing and other forms of language were constitutive of social structure. And so on. These are useful metaphors. The hardest thing to grasp is their identity.))
211
The pointless exercise is to seek competing synonyms for the 'it' of this passage. And descriptions of it are but wayward descriptions of practical action. Another synonym: 'scaffolding' (Gerüst) — and so it becomes very easy to slip into the reificatory metaphor of structures and frameworks.
215
And we also see that the notion of 'clear applications' is problematic.
216
The peculiar idea that truth is scarce and requires privileged access. It is hard to apply this idea to a familiar form of life. But where scarcity and privilege are associated with the practice of writing we almost want to laugh at the idea. And aren't there certain sorts of writing that we treat in the same way?
((Zimmerman (1974): in the welfare agency, officials found it laughable that a client should believe himself to be proving his identity by furnishing just any scrap of paper whatsoever with his name written on it. In an episode of the 'Goon Show': two characters muse over the advantages of a having a watch rather than a piece of paper with the time written on it.))
223
The social character of being crazy: the mentally ill are so because they are so labelled (according to a certain theory) — but this is only one stage. They are so labelled because they doubt (and so forth) where reasonable persons do not. And their doubt involves certain language-games which reasonable persons do not (know how to) play.
224
The past is no more 'fixed' than the present or the future. It's just that there too we have our preferred constructions (our favoured narratives). And this is where we fall back. It underpins the difference (in this form of life) between 'I have forgotten' and 'It didn't happen'.
229
And vice versa? Even if this is so: there is no sharp demarcation between one and the other. It's not as if either one were primary and the other dependent. Sometimes it suits our (theoretic) purposes to construct such relations between separate objects which only became separate in the first place by virtue of the 'application' of the same theories. And reflexive theories of 'mutual dependence' are just as troublesome for the same reasons. The division is as arbitrary that expressed by the phrase 'a proceeding and the rest of our proceedings'. Such divisions can be made anywhere at all.
230
Applied to our own investigation: how do we decide whether rules do or do not constitute knowledge? What do we do with a statement 'I know the rule...'? Although there are occasions when its use seems quite clear (such as when a doubt is raised yet it seems almost impossible for anyone not to know it), in any wider analytic sense, it seems perfectly useless.
Yet we can imagine forms of life that valued such statements. And then we think of all the other bizarre things that people have made collections of, and how certain items became valued, rare, treasured. (Aspects of the Theory of Syntax.)
233
And having this feeling of saying something funny (Komisch) also seems to go along with saying 'I know the rule...'. There are practical circumstances where this is in order. But outside these, the peculiarity is glaring — the peculiarity of announcing the obvious.
Lectures where I announce Harvey Sacks' rule that 'one person talks at a time' get this reaction — hence Sacks appears in the Private Eye Book of Pseuds. But then why are the Euclidian axioms a discovery? Who would act differently? (And the same is true in the case of Sacks.) In both cases: by knowing the rule we can locate regularities in all sorts of other, related cases. And these may be highly counter-intuitive. (This is presumably a valuable intellectual commodity — see remarks on OC 216.)
'Imparting a picture of the world' has something to do with this. Cultural studies seems to have set itself exactly this task.
234
And this (233) is perhaps what impels Alfred Schutz (1972) to make a catalogue of the things which 'stand fast'. And perhaps there is something too in his 'epoché' of the natural attitude' (1962, p229), a systematic refraining from doubt. Except this unduly exact specification of a system seems to be troublesome when it includes anthropological generalisations (Rules1).
238
Another synonym for Rules1: 'fundamental attitudes' (Grundanschauungen). But isn't it rather these which underpin our having attitudes about something?
Where these are contradicted: 'I should simply have to put up with it' (as with a disease). The interlocutor is (just) speaking from a different terrain. This we can see at a glance. But to describe the form of life which such a detractor inhabits is another question. So what is it that we do when we go looking for what people know and use in order to see something (like this) 'at a glance'? The sorts of rules we want here may be very different from the sort Wittgenstein calls 'fundamental attitudes'.
239
This goes to show that 'beliefs' is a poor synonym to add to our list. Knowing and believing seem to be different practices. The Catholic histologist knows that the wafer is not the flesh..., but he may still believe it. And both of these predicates have something to do with the foundations of the form(s) of life he inhabits. But those foundations cannot then be called something he knows or something he believes.
240
A concept like 'parent' rests on its intermeshing with all sorts of supportive concepts — 'sexual relations', 'care-giving' and so forth. And the concepts support is derived from these threads. If 'parent' happens to be the one we are interested in, it seems to be primary and all the others subordinate. A moment later, if we shift focus, these dependencies could easily be reversed. But we are still on the same familiar territory. The same experience is being used. Another form of life would be one where, say, the concept of 'parent' were supported by quite different ones — 'cloud patterns' or 'star constellations', for example. (I take this material from a lecture by Val Presley.)
And, of course, this has nothing especially to do with that other language-game - proof. Though we can forge a connection.
241
We can see life as the continual confirmation of our hypotheses about the world — but this is only a way of thinking and perhaps an overly scientistic one. Someone else might say: 'God continually reveals himself in his creations'. And another still: 'The evidences determine the pattern and the pattern the evidences. Each is used to elaborate the other.' The 'hypotheses' version is one Chomskians like to use with respect to how a grammar is learned. On this version, when my daughter was very young, she held the hypothesis that there is a general rule of negation:
S(neg) --> NP, VP, not
And, according to the theory, her hypothesis should tend to be disconfirmed as she listens to others form negative utterances in the regular way.
But is this what happened with (her) language learning? Why, for instance, did she hold this hypothesis for nearly two months? And why didn't (or couldn't) she operate 'verbally' with hypotheses (in her interaction with others) at the time? What would the game of hypothesising look like when a baby carried it out? (As opposed to, say, an archaeologist?)
There may be rules for hypothesising. Yet it seems bizarre to think of the acquisition and application of rules as akin to hypothesising (and that hypothesising always being successful!) We have inverted the picture in so thinking. Coulter (1979a) makes a similar point, referring to transparent vs. opaque concepts.
The problem is not confined to cognitivists. Behaviourists think of their subjects as learning S-R 'rules' by having hypotheses confirmed and disconfirmed....
We can play games where rules and hypothesising are akin — like when we think of rules for a series of numbers — or in the wonderfully reflexive whist game where a 'judge' decides or makes up 'whist rules' which the players don't initially know so that, as the judge corrects their play, the rules gradually 'dawn' on them. And this is how some natural science operates. But here we have to think of rules as akin to natural phenomena, properties of the world. Some forms of life are anchored in this way, while some are not.
248
Rules which lie at (as) the foundations of knowledge (Rules1) — this seems to be what we're looking for. But the terms of the quest are badly expressed. For these foundations are simply part of a whole system of (for example) concepts — 'the whole house'. And this is where we ought to be looking. But how, given that the systematicity of the whole evades us? It evades us because we are thinking of it as being like a foundation. In fact, a description of Rules2 would be good enough, for it is here that we begin to describe people's actions. And it has to be 'good enough', because anything 'deeper' that we might look for only exists as the product of our confusion (about language).
((We want to grasp some thing — so we reify what it is that people do.))
249
The false picture: that doubt is the absence of certainty. While in fact: in order to doubt, plenty of things must be already in place — as certainties. But the stipulation of what's in place seems to get us nowhere. (Moore, Chomsky and others have been misled by this picture.)
253
And this shows that 'belief' can be used in two different ways.
254
So the crucial (critical) term becomes 'behaves'. And here we can begin to think of rules.
255
His actions are 'like doubting' but they are performed on 'beliefs' (cf. OC 253) which are not founded but founding.
256
Thus our work with 'anthropological' rules could become historical.
257
And (like the reformed half-wit) we are similarly distressed when we try to reconstruct exactly how we moved from one system of concepts to another.
((Kuhn and his ilk may well be in the business of trying to do description here, as is most of psychiatry. But are the 'how and why' of it even items of knowledge? When we 'shift terrains', which 'neutral' terrain can supply the language with which to address the shift? 'A kind of persuasion' (OC 262).))
260
And in this matter, Wittgenstein differs fundamentally from Moore (and most philosophers). But who can say who is right? What techniques could be used to choose between them? When Austin (1962) ploughs piecemeal through Ayer's peculiar way of writing, what is it that compels me to side with Austin?: only my prejudice towards the idea that expressions should be confined to their 'normal' linguistic usage; my literalism. But I can see that Ayer would have it otherwise. These are but (marginally) different forms of life. It's the convictions I already have that I use in deciding between them.
((It might be 'healthier' to give these up from time to time. But can I? And this is why some people find it hard to accept Derrida's (1977) deconstruction of Searle's notion of the literal.))
262
Persuasion: transactions whereby we become 'of the persuasion that...'; our form of life changes; new games can now be played and so we act according to different rules. But it's not quite as if 'being of the persuasion that...' were itself constituted by 'deep' rules (Rules1).
264
The example here is necessarily one involving an extremely vast cultural or anthropological displacement.
267
Austin, Ayer, phenomenology: all three (and several others) come to mind.
272
This could be rewritten as: 'I am a member of some collectivity which unproblematically accepts it as beyond doubt and acts accordingly.'
We act according to rules. And we act according to that which we unproblematically accept (grounds, foundations, etc.). But this does not mean that these grounds are rules. (Are they rulelike?) Though it's easy to conflate the two, especially when — in our search for rules — we decide that we'll try to discover what it is that people act according to. If we do this, we will always be open to the charge that what we find will not necessarily constitute a rule.
((This is another methodological problem for linguistic and cultural studies.))
274
Interdependence (zusammenhängender): this does not mean that we can specify its general systematicity. The whole fabric never seems to be (all) that clear.
Consider the following from Garfinkel (1967, p123) where he's trying to specify how certain things hang together for us with respect to sexual identity. Doing this (like Wittgenstein in OC 264) requires an examination of cases where membership of the collectivity for whom it hangs together has become problematic:
For normals, the presence in the environment of sexed objects has the feature of 'a natural matter of fact'. This naturalness carries along with it, as a constituent part of its meaning, the sense of it being right and correct, i.e., morally proper that it be that way. Because it is a natural matter of fact, for the members of our society there are only natural males and natural females. The good society for the member is composed only of persons who are either one sex or the other. Hence the bona fide member of the society, within what he subscribes to as well as what he expects others to subscribe to as committed beliefs about 'natural matters of fact' regarding distributions of sexed persons in the society, finds the claims of the sciences like zoology, biology and psychiatry strange. These sciences argue that decisions about sexuality are problematic matters. The normal finds it strange and difficult to lend credence to 'scientific' distributions of both male and female characteristics among persons, or a procedure for deciding sexuality which adds up lists of male and female characteristics and takes the excess as the criterion of the member's sex, or the practice of using the first three years of training to decide sexuality, or the provision for the presence in the familiar society of males who have vaginas and females who have penises.
So certain things hang together for what Garfinkel calls 'the normal' and certain other things for the scientist and still further things for intersexed persons.... But how does this difference get seen and noticed? That is, how is the difference practically constituted? Only when these forms of life are brought up against one another — say, in popular scientific magazines, at cocktail parties or in Garfinkel's text — could that happen. And here: it's a not matter of persons actually comparing the grounds of their versions of the case but of practices, behaviours, actions, like those mentioned in Garfinkel's last sentence. So it's not as if the grounds were composed of rules, 'in themselves'. Rather the question of rules arises when we consider the (here: contrastive) practices of different collectivities. Grounds only appear to be 'manifested' in such cases of gross differentiation. But everyday rules of action appear not to need contrastivity. And this underpins the differences between a great deal of earlier and later ethnomethodology.
279
Hanging together (hängt zusammen): it would seem to be impossible to stipulate the 'how' of this hanging together. Certain things hang together.
Not 'learns' (aufnehmen/lernen) rather one swallows it down with the things that are learnt (OC 143).
And here we are tempted to invoke or manufacture categories like 'the unconscious'.
280
But what he is in no position to doubt is not (necessarily) what he has seen and heard. He learns to see and hear (etc.) in such and such a way — and here there are rules to be acquired. But what, as a consequence, turns out not to be doubtable is neither learned, nor is it a matter of acquiring rules (as such).
((It is much more like the acquisition of a picture consistent with what is learned and with the rules of procedure.))
282
'I cannot say...': and this is because the propositions in question are not opinions. These are what we rely on in order to engage in games like 'expressing an opinion'. And where the opinions differ, these will not.
Also: might such a thing perhaps be taught (gelehrt) but not learned (lernt) (OC 279) as such? Do we have to be taught grounds which differ from those which every reasonable person has? — while we simply pick up (aufnehmen) these latter?
((Do our first set come free while the rest are only got at considerable effort (as is the case with Castaneda)?))
286
'...by far [for us]': so social science continually seeks to replace what people commonsensically know and use ('ideology', 'bad faith'...). For 'us' (social scientists): we are continually in the position of having to compare our system of knowledge with theirs, even though ours supposed to be a system of knowledge about theirs. But, by this stage, it's always much easier to make judgments (Lukács' 'implied class consciousness', Comte's 'religion' of sociology). Social science exits today almost totally exists in order to make common sense into its poor relation. It is a literary criticism; a tool for producing discursive hierarchies. Wittgenstein's 'poorer', read ironically, points to social science's chauvinism; and so to its lost relativity (Vico); its failure to attempt to engage in description of its proper object.
Sacks say something like: 'how can we find out what it is people know and use when it's already decided that what they know and use is wrong?'
287
Something gives a form of life shape, relative stability, predictability (invariance)... but this 'something' is not a set of general laws on which particular actions are based. But can we then say: 'There are no rules of action'? No: the problem is one of confusing (empirical) rules of action with the grounds (foundation, picture, mythology, fabric) of a form of life. The discovery of the former is nothing special; an empirical matter.
288-291
Know and believe: Wittgenstein wants to display the peculiarity of Moore saying that he 'knows' these things. But at the same time, he wants to reserve the expression 'I know...' for occasions of its normal usage (OC 260). And Moore's seems to be an absolutely ordinary usage, given that he wants (somehow) to say these things. Yet: where could they be said? And what would be the use of saying them? This is their peculiarity. So why should 'I believe...' be so different (Needham 1972)? Isn't believing a practice or set of practices which is also anchored in a form of life?
((Wittgenstein appears to be asking, however, what is this like for me and not for the entire membership (OC 281).))
292
The Einsteinian cannot give the lie to Newtonian mechanics. The same is true of the social scientist and the collectivities she studies (and of which she may also be a member).
300
'Level' (stufe): a 'correction' of our whole way of seeing things differs from a correction made within it.
What I (probably wrongly) call 'Rules1' and 'Rules2' are distinct levels or stages. (For instance: to 'correct' the view that all persons are either one sex or the other vs. to correct the view that so-and-so is one sex and not the other.)
302
'Perhaps we are wrong' about some of the 'assumptions' we hold most dear about the world: this step, associated with both scepticism and phenomenology, has to be nonsensical. Because: what makes something 'right' is not that it is an objective property of the world but rather that our form of life conducts itself on the basis of the world being like that. To 'be wrong' here is not a matter of imagining the world to be actually subject to entirely different principles 'in the facts' — it is to propose an alternative form of life. In all this: what the world is 'really' like remains untouched and untouchable. There is a deeply seated irony involved in the scepticist position which appears to require assumptions like: 'The world may actually be x, y and z, but no one has yet seen it this way.'
Talk about 'reality' does not given the lie to our conventions — at most it proposes alternative ones. And then there's the practical problem of recruitment and conversion. (A case in point occurs during scientific revolutions.)
304
There can be no mistake about all people being male or female (in a certain form of life) but only about whether some person is (fe)male. (The levels of difference and mistake.)
How the world is for (some) 'us' cannot be a mistake.
Sociology's ironic version of common sense is not merely arrogant (in the rather technical sense I want to reserve for this term), it is also and more importantly confused about language and what can and what cannot be called a 'mistake', 'false', 'wrong' or 'ideological'). What is not scientific does not necessarily stand condemned. And here: 'a step like the one taken in relativity theory' (OC 305) is most definitely needed (Is that step a moral one? Exactly?)
Examples occur frequently in the sociology of knowledge and social psychology even though one may think of these disciplines as archetypal terrains which already realise this relativistic step. But the following sorts of language are still used there:
The sociologist moves in the common world of men, close to what most of them would call real. The categories he employs in his analyses are only refinements of the categories by which other men live — power, class, status, race, ethnicity. As a result, there is a deceptive simplicity and obviousness about some sociological investigations. One reads them, nods at the familiar scene, remarks that one has heard all this before and don't people have better things to do than waste their time on truisms — until one is suddenly brought up against an insight that radically questions everything one had previously assumed about this familiar scene. This is the point at which one begins to sense the excitement of sociology. (Berger 1966, p33)
Until one can propose a dependable technology of assessment, which surpasses lay judgments, one can neither challenge nor refine the commonsense concepts and propositions out of which social psychological analysis has arisen. (Scott and Scott 1979, p99)
The examples could be multiplied endlessly. We could cite Berger's (1966, p83) discovery of 'the hard rock of human stupidity', or the 'stupidity, superstition and the congenital bestiality that jurists share with the bulk of their fellow citizens' (p85). But a science which forgets its forbears is doomed. And which science's forbear is not common sense? And how many sciences also propose to investigate 'the common world of men'?
We can either investigate that world or we can construct a distinctive scientific community: but both?
((Sociology does involve a sort of relativity — but it is like classical Newtonian relativity. For example: 'For the lawyer the essential thing to understand is how the law looks upon a certain type of criminal. For the sociologist it is equally important to see how the criminal looks at the law' (Berger 1966, p41). But this extends no further. For where one of the parties represents science and the other non-science, the same does not hold. Here: what is correct and what counts as a mistake for sociology is quite clear, quite absolute. And I obviously don't want to say that sociology is mistaken in working this way. It would be better to say merely that sociology has no special prerogative for undertaking investigations of society.))
307
'How the words are used' ... 'grounds for my way of going on': the use can have its description (and this is what Wittgenstein himself is offering), but not so the grounds.
Language-games can be described, as can their rules — but the general fabric of which they form a part cannot. As though the family resemblance between them could be stated as a set of (more) general rules. (Family resemblance works as much by difference as by similitude or identity.)
308
The empirical vs. the undoubtable which underpins it: the former enters into our reasoning — it is here that we conduct our reasoning. The latter stands fast for us; it does not enter into our reasoning; but without it our reasoning would not be like this. There are rules of, for example, reasoning — empirical rules — and there are also un-reason-about-able grounds. When we consider the problem of rules and their application, these levels should not be confused. (For instance, the two uses of 'methodology'.)
((Yet at times, and with time, the empirical and the undoubtable will merge or even change places.))
309
This depends on whether we are still allying 'rule' with the grounds of a form of life (picture, mythology, Rules1) or with occasioned, situated and practical maxims of conduct (Rules2). Remember: rules are the sort of things that can have mistakes in their application (OC 26-29) while there can be no mistakes in our 'whole way of seeing'. Rules in the strict sense and their applications appear to be empirical matters.
But if, by 'rule', we intend that of which we are certain, that which stands fast for us, then rules are definitely not objects of knowledge. (Dare I say: 'for us'?) But still on this extreme view, rules have their applications (= practices). And applications are empirical matters. So we should not want to think in this connection of Rules2 for the application of Rules1. Even the mentioning possibility seems corrupt or fishy.
Either way then: we can speak of a merging of rule with the empirical; though our notion of merging will vary according to the position we take.
OC 309 represents the fundamental question in my debate with Hunter. Are we to ally rules with grounds or are we to see them as separate but somehow connected levels? Is the question of rule an empirical question? What can we say for and against these readings? First: most of what I have written so far has rendered any discussion of 'Rules1' peculiar. What Wittgenstein calls variously 'foundations', 'matter-of-course foundation', 'grounds', a 'picture of the world', a 'mythology', 'hard rock' — these seem to be different from rules. They seem to lie deeper. (The metaphor of depth is compelling in this context.) Second: would anybody want to call 'that the earth existed 100 years ago' or 'that I have two hands' rules? But third: if these lie 'deeper' than rules, the empirical, the observable seems also to be underpinned by rules. At this point: rules do appear to 'lie deeper'. In the comments on OC 272, I suggested that we may act according to, say, a picture of the world and also to the rules that it furnishes, but that this does not make them equivalent.
((I want to argue for a three-step 'equation': foundations, rules and action (the empirical). This seems much too easy, but the picture does present itself.))
And according to this picture, the really unspeakable thing comes first - this may be the sphere in which Wittgenstein rules out knowledge as such and prefers to refer to certainties. For ethnomethodology, this first domain is that of 'background expectancies' and 'tacit knowledge'. Mehan and Wood (1975, pp98-116) attribute these to a generalised model of man (the 'reality constructor') and unfortunately, at least for the present reading, refer to 'social knowledge' and 'interpretive procedures'. (These are roughly the Schutzian anthropological universals and Garfinkel/Cicourel on 'normal forms', 'reciprocity' and 'et cetera'.)
The awfulness of all this surfaces when Mehan and Wood produce a sense-data narrative for its explication (1975, pp102-106). Like Moore and Ayer, they are thereby cornered and forced into a peculiar form of expression: 'A man driving to work sees a fuzzy object.... Before the man can begin driving ... leave his bed ... put on his clothes ... drink his coffee, he must assume that objects are what they appear to be on the surface...' and so on.
Wittgenstein seems to point to this: all of this can be the case, but attributing it specifically to the domain of knowledge is ill-suited to the case. (What would it be like to doubt here?) And so it's no surprise when Mehan and Wood report that 'no studies describing the use of social knowledge [sic] and interpretive procedures in actual settings have been conducted' while also noting that this is no domain for 'free invention' (p106).
((Interpretive?: surely this is one hidey hole where that term never applies? There's always something that's not open to interpretation. And to say 'these are the grounds of our interpreting' is equally meaningless — for they're also grounds of our everything else. Nose-picking procedures? Promising procedures? Yes: but they do not constitute the rules of promising or nose-picking. These are quite different.))
No studies have been performed — what could they look like? But the description of rules (as such) is a different matter.
The discovery of the medium (or element) in which we 'swim' is beyond us — but not so the rules of 'swimming'. Garfinkel's (1967) breach studies try to show that (if not quite how) the social world is founded on such and such — but these studies too are practices conducted in that world and they have their rules. They do not 'reach beyond' practice. What could reach 'beyond practice' here? 'Theoretical practice' — but....
For us, 'see[ing] general features' (Mehan and Wood 1975, p107) is like asking for the rules of the rules, the abstract systematicity of the system. The early parts of the Blue and Brown Books speak of the peculiarity of this.
Are we confined to the particular case? Well, what's so terrible about that, except that the whole of (social) science runs against us? Here: the social world and the world of social science actually look like mirror images, but in the sense that nothing could be quite so oppositional.
Without speaking of sharp boundaries (OC 319): rule and (the) empirical (proposition) merge into one another. And so do rule and ground(ing proposition).
((Now I have this picture. The ground of this text is relativism.))
310
The crux of the matter is in our doing, not in the expression of grounds. (As if this latter were 'pure' theory. 'Pure' theory would be the ground-free stating of grounds!! What I have in mind instead is a rule-governed description of rules.)
312
We want to say 'the grounds are arbitrary'. But who could extricate themselves from the grounds in order to actually see this? What connects up — we are inextricably connected up in. (This methodological aberration has its earliest concrete expression in Bacon's attempt to remove 'natural interpretations' and its extends through to Husserl, Ayer and others.)
315
If this ('Is there a table...') were questionable, we should (already) live in a different world. Where the idealist raises these 'questions': he is not speaking to us or as one of us. He is not 'doing' our world. As if that were possible not to do it! It's so easy to accord privilege to this sort of writing; just so as to follow the narrative — to see where it leads. But, of course, it doesn't even start out. We are tricked into a form of representation (here and elsewhere). Idealism has only the outward form of a search for something.
Because it goes wrong over the question of grounds, it cannot 'learn how to...'.
317
Choice: compare Mehan and Wood's mythical 'free invention'. Not just anything will do here: but not that we can state precisely what will. (The indeterminacy of grounds.)
Something remembered, in a slovenly way, from Sacks: not just any word can get used as an address-term. But is there therefore a finite list of those which can? Still: on finding occasions where whatever happen to get used as address-terms are in fact used this way, we can begin to see and describe some of the ways in which addressing gets done.
And, albeit with a shift to 'doubting', 'believing', and 'being certain', isn't that just what Wittgenstein is doing here?.
318
Does Althusser characterise methods by seeking their limits: by seeking those questions which don't arise for them? And aren't these sorts of methodological propositions also propositions within a method? 'Yes', we want to say, 'but a different one'. So how far can we take this differentiation? Doesn't it too come to an end? And at the end, there simply can't be any more methodological propositions, only propositions within a method.
It's here that Hunter's question about rules begins to make sense to me. And it's only our seeming ability to proliferate methods infinitely that prevents me from going along with it.
((That is: what does the 'end' look like here? Is it something (the only thing) universal but whereof we cannot speak? The descriptor is always ultimately part of the thing it would describe — but only in that we can always find a perspective from which anything is part of any other thing.))
The lonely hour of the last instance never comes.
319
For 'logic' read...
321
Again: there is a 'blockage' against 'pure' theory (OC 310).
((And does practice 'get in the way'? In order to say such a thing we should have to imagine nothing whatsoever, including being silent. Nihilism is also a form of practice.))
Each transformation here already supposes a transformation of forms of life. (Being precedes....) And this is what we mean by 'blockage'. So when we say '...can theoretically...': this only reminds us that we can imagine other forms of life. In theorising, we are always struggling to establish a form of life (in and through language). But to the extent that we still 'make sense', the umbilical is never quite severed. So the seeming simplicity of proliferation is never quite that unproblematic (cf. OC 318).
And here we are tempted to say: 'Something binds the discourses of a natural language', but in getting thus far we have already ruled out the possibility of saying what this is.
Language or languages? The consequences of the choice are enormous. But the choice is only 'theoretical': can we choose between pictures of the world like this? (Say in imagination?) Or do we only retrain? Everything points to 'language': but then so many doors are closed. (Were they ever open? or were they just grammatical errors?)
Why does Wittgenstein want to say this ('any empirical...')? In what sort of world could he say it? What is preventing him?
To imagine a language may indeed mean to imagine a form of life (PI 19): but it's precisely imagining that we have to steer clear of if we cannot get clear about it. (As if imagining took place unconstrained by language.)
(Possibly aside from imagining) there are certain grounds which we have to share, which we cannot treat as empirical propositions, if we are to make sense. Some norms of description just can't be displaced. What is even relatively hard (OC 99) is practically unshakeable.
And what kind of investigation would this leave us with? None. By way of investigable topics, there are certain things we can do nothing with. (Elsewhere, for instance as resources for investigation: we can do nothing without them.)
In this sense: grounds are best left lie, as a rule.
Yet: doesn't it seem flippant to say, 'Some bizarre collectivities will always try to play language-games where these are doubted'? Won't they after all have their rules for successful participation?
((On Wittgenstein not being able to say what he wants to say: there's a constraint here on the use of the words 'theoretically' or 'in theory'. And still we can point to practical instances, empirical instances, if one prefers, where an empirical proposition has become so seminal that it acts as one postulate among others for an entire mythology, for a form of life which collectivities have lived and are living. A case in point: 'Marx met the Paris proletariat in 1848'.
So the constraint on 'theoretically' is this: it is still nonsense to speculate about certain empirical propositions becoming 'deep' postulates. This seems to involve an imaginary proliferation of possible forms of life. There's a grammatical relation between the uses of the words 'theory' and 'possibility'. At the same time, only a limited number of forms of life are available and all discussion of 'possibilities' must be done in the language of one of them. (This world speaks all possible worlds.) This constraint brings us very close to the central contradiction of Wittgenstein's text: the use of (linguistic) relativism in the pursuit of limits (of language).
The proliferation of games: everything comes to an end.))
The available means constrain all references to (and therefore the construction of) means which are not yet available. (We are both the products and producers of our own history. We make our own history but not under...).
This famous passage from Marx (The Eighteenth Brumaire, 1852; in Feuer 1959, p320):
Men make their own history, but they do not make it just as they please; they do not make it under circumstances chosen by themselves, but under circumstances directly encountered, given, and transmitted from the past.
Marx then goes on to say how, at times of revolutionary crisis, the revolution takes on the tokens and vestiges of previous victories. What is often forgotten, when this gloomy passage is quoted, is that Marx then goes further still by saying that, for effectual change to occur, the seemingly inevitable temptation of relying on past terms must be overcome.
And the Wittgenstein's position in some passages of OC (317, 321), on the problem of speaking anew or speaking creatively, is almost identical to Marx's. The means available to us are those given by the form of life we occupy and its language-games. To construct a new form of life we require new grounds - but grounds are not that easily shifted. They can't simply be conjured up and put to work just like that, by fiat. (For example, in the case of the struggle for a discourse which dissolves the problems of 'theory' and 'practice': would that be a practical or a theoretical struggle — and if we can't decide, would the struggle itself not exemplify already the very thing it struggled for?)
What is significant is Marx's next metaphor in the same passage:
In like manner [re. translation of new crises into old terms] a beginner who has learnt a new language always translates it back into his mother tongue, but he has assimilated the spirit of the new language and can freely express himself in it only when he finds his way in it without recalling the old and forgets his native tongue in the use of the new....
Presumably we can do this when there is a community, a practical working model, a form of life, already using a language other than our own. But the problem of forging an entirely new form of life, of being entirely revolutionary, is that we never have a working model available. The mode of the future is never given to us. There is something deeply seated and embedded — as well as unpredictable - about a language. The revolutionary is always in danger of believing that she can speak from outside it — just like that, again, by fiat; imagining and then persuading: all via the given means of the community she would reconstruct. How, then, do we change society? (That is to say all of the following: how do we construct a different form of community? How do we imagine a new form of life? What are the limits of a language and how are they overcome, if they can be?)
Language and society do not 'come together' simply for peculiar varieties of 'sociolinguistic' research or for speech-act theory, etc. They are always already indissoluble. In fact, it is only our current and bizarre compartmentalisation of intellectual practice that tries to sever them, thus calling for a further suture. The 'relation' of language and society looks likes an interdisciplinary topic, but this is only because things got badly disconnected a long time ago. (For a cyclical narrative, return to 'This passage from Marx...' above.)
Wittgenstein (despite some readings of the Investigations) is not giving us a picture of language as a tool-kit for free creation. It's 'not as if we chose...' (OC 317).
An example — what is a 'thought experiment' (Einstein)?; if not the production of the world as the logical empirical outcome of a few (novel?) 'givens'? Why, then, do we construct a monumental technology (particle accelerators, etc.) to 'prove' that the world is in fact like this? Why, for a particular collectivity, does it only seem certain once all this has been done? (When Einstein was asked where his laboratory was, he took out his notebook.) Is it because grounds/postulates which do not ostensibly support an array of empirical propositions cannot be called 'certainties' any more than unsupported empirical propositions can be called 'knowledges'? A 'real' world is one where both postulates and empirical propositions are relatively stably in harness; tethered together; bound. And the Einstein case is only one (very tangible) index of the massive work required for even minor differences in a form of life to occur. What is it that's involved in such a bond/tethering? Surely this: rules. Learning relativistic physics means learning how to proceed in a world which is such and such. Rules of procedure bind together the un-reason-about-able grounds and the production of relevant 'reasonable' propositions. Otherwise any proposition could be a proposition made on any grounds. What creates and constrains this relation we call 'rules'. Without rules, grounds and empirical propositions 'float free'. (And we can we even imagine a world like this?)
Creation (relativism) and constraint (limits) are questions of rule. Investigations of rule are also meaningless where the investigator does not both hold the grounds and observe (preserve) the evidence. I do not say 'know' the grounds. Therefore, 'knowing' the evidence: yes. But 'knowing' the rule? And here we see, for instance, Garfinkel's crucial distinction between substantive and procedural knowledge (which is based on Ryle). And also: the bankruptcy of 'sign/referent' theories.
((How can there be a form of life whose language can even address questions of (the relation of) ground, rule and empirical actions? Where do we write or speak from here? In order to do so, the notion of 'theoretically' in OC 321 has to be activated. In addressing such matters it's not as if we were unconstrained by the very constraints we seek to address. And 'forging' a way of speaking or writing is just as difficult here as elsewhere. But this is what we must take on by virtue of the very project of asking 'what readings of OC are there which respectively do and do not display rules to be objects of knowledge?' Of course, at least one of those readings renders the question which generated it meaningless! It cuts off 'that sort of talk'. And this may be why our allegiances fall in a direction opposed to that reading — at least so far.))
324
For example somebody who believed, in 1200 AD, that the Earth rotated around the Sun. Or: somebody who believed in 1905 that the elliptical orbit of Mercury might fractionally rotate or that velocity has an absolute upper limit (OC 336).
All we are saying here is that we can imagine people who are 'out of step'. Unlike in the case of the Tractatus: there's nothing sacrosanct about science here. But there is, in another sense. In that sense in which science provides certain unshakeable conventions for us and in the sense in which it is neither here nor there to indicate the 'arbitrariness' or 'hollowness' (OC 312) of those conventions (grounds and rules). Because then we could ask: 'arbitrary with respect to what?'; whereas on the present reading we can say 'non-arbitrary with respect to just about everything which connects up with those grounds'. There is very little which is more peculiar than saying 'such and such is just a convention'. For one who would say this, the standard must have a further ground. And this is nonsense.
326
The question here is not 'who?' but 'what?'
328
And this too is a convention. (Yet) nothing could be firmer, more certain.
332
The difference between denying on empirical evidence and denying because 'no one can speak otherwise': these are very different sorts of 'appeal'. Moore seems to confuse them though; as do Mehan and Wood when they appear slightly surprised that no empirical studies of the grounds of a form of life have ever been undertaken.
((If what they claim about Crowle's (1971) thesis is true, Crowle must be a Martian. See Mehan and Wood (1975, p106, fn9).))
336
Merkmal = sign, characteristic. Diachronically, it makes sense to think about shifts in the grounding of a form of life. Wittgenstein and his 'never having been on the moon' would be a case in point. For me, in the 1980s, this has come to be a bit more like the empirical case of 'never having been in the village of X' (OC 332). It is not something which we (here and now) deny because it has to be denied necessarily. Here is an interface where grounds are shifting diachronically (but because of practice).
Synchronically, it seems to make less sense: here we are always closer to the possibility of ein objektives Merkmal. That is, those who maintain a belief in, say, the scientifically absurd (creation in Genesis, that a person once had only a single parent...) do so 'despite', 'in the face of' or 'along with' their own beliefs. The Christian only allows so many special cases. And here the rules of divinity (tradition/convention) show what is to be marked as a bona fide special case and what is not. For instance: we couldn't imagine these beliefs forming a life practice. (A Christian wedding where jars of water, some loaves and some fishes were brought along, everyone quite naturally expecting their transformation? Or: virgins routinely giving birth?)
337
To take on trust: in this case we should have to be able to imagine what it would be like for the trust to be betrayed. Could the scientist do this? For instance, imagine a betrayal of his trust in the fact of his workbench being solid. Taking for granted entails a situation where trust is not in point. Trust is connected with betting (a 'pretty safe bet'). And could we say that we all engage in the wager that the sun will rise tomorrow? — a wager which will always come off? No: this is absurd. But with the case of, say, religious beliefs — isn't this case more like trusting? Or betting?
((And Alfred Schutz treats all the grounds that we adhere to as being like wagers — hence his talk of the man in the everyday world's 'suspension of doubt'. 'I trust and I trust that you trust that...'. For phenomenologists, the world can always be different from how it seems — it's just that they take 'seeming' to cover everything, and this really is absurd. We could also say these things about phenomenalists, Ayer for example.))
338
The people we imagine here are like Ayer and Schutz: 'we could doubt but...'. As if this were a voluntary activity. But: what would ground the activity of doubting, and what rules of procedure could we use? The amazing thing is that people (like Husserl) have actually tried to answer these questions instead of seeing the absurdities which they point to by virtue of their being uttered (and so dissolving them).
Popper: 'We can't conclusively prove that...'. But starting from where we do, all of the grounding propositions are unfalsifiable — including the falsifiability criterion itself! Because of this: the form of life, 'science', simply does not routinely involve procedures for their falsification. Even to imagine such procedures is unscientific.
((Likewise: to imagine procedures for the falsification of the immaculate conception would be unchristian.))
339
Having 'doubts' here cannot actually affect our everyday conduct. Whatever: we must still conduct ourselves by the same rules as everyone else.
341
Metaphors: shunted on to a siding (OC 210); hinges on which things turn OC 655). Raising questions is always based on something unquestionable. What is it that Husserl, etc., seem to achieve by trying to ask questions of the latter, the unquestionables? The most we could do here in all seriousness is to display — give testimony to — a certain constraint.
((Husserl's notion of 'transcendence': in theory, the postulates of a form of existence get turned into empirical propositions. But what can now be said to ground this activity of transformation? In order to do this (and not as a result of doing it) an entirely foreign form of life must be invented; as if we could do such a thing at will. Because idealism simply assumes transcendence, it is not the means by which we can develop strategies for transcendence. Those problems are always already overcome for the idealist.))
In phenomenology: the hinges are seen as 'just part of the door'. And then: nothing can turn.
342
In der Tat: what cultural studies has lacked is a strong material base. As if (by definition) we were dealing with the immaterial here and so must reject materialist theses.
((Yet we 'allow' idealist accounts of science.))
But all the time the materiality of the act, of practice, of technique and of rules has been available. Have we been daunted by the spectre of behaviourism?
Where cultural studies has adopted materialist positions it has never identified this aspect of materiality exactly. (Brecht came closest.) Cultural production has consequently been seen as immaterial production whose proper understanding lies in an investigation of 'actual' material production. In fact the material production of cultural commodities in cultural practice is also open to us. Aesthetics has nothing to do with 'mind', (Goethe, OC 402) and everything to do with practice. (Likewise: semantics, OC 347.)
Hunter: 'The unmaking of the totalist-structuralist mode involves showing that "the relation of discourse to its object involves the relation to its social function". It is not the Saussurian problematic that begins an account of this relation, but the negative problematics of the later Wittgenstein.' (From his lecture notes.)
We should still be able to say where something is unmade from and, consequently, some quite positive things about matters such as: rule, practice, production, discourse.... Otherwise the idealist/phenomenalist trap of voluntaristically unmaking/transcending yawns. That is, we assume that what we would undo already has exogenous means available for its undoing.
343
The problem is not one of exhaustion (comprehensiveness) but of constraint (limits).
345
Orders of questions: the absurdity, in phenomenology, of asking 'What do you see?' and 'What do you see (is given) in this act of perception?' In the same question which presupposes my seeing, that seeing is also 'called into question'. There are just some things that we cannot do with language. (Which does not imply that what we can do is finite.)
347-350
The expression in philosophy attempts to be contextless. The problem lies in the game of 'imagining a pure utterance'. In this game we turn to certain marks or to possible mental resonances in order to puzzle ourselves about the meaning of the expression. In most other games nothing quite like this occurs. There is an emptiness (to the expression) as if it were to language what charades is to 'actual' practice.
When an object is investigated in philosophy, it is as if one said of it, 'Imagine what goes on with this...' but also required an answer in the abstract - a sort of general synthesis of all possible uses.
((Imagine the equivalent in mathematics. The philosopher asks 'What is the meaning of this equation?' while everyone else might, for example, find its solution or use it to calculate.))
351
- Have these words a meaning?- Yes, they are for X, but also for T and Z, and perhaps for a few other things besides.'
(Where X, T and Z... are practices and where the meaning in each case is a question of the application of procedural rules for doing X, T or Z.... But to use the term 'rule' in an attempt to gloss something X, T and Z... have in common and to call this the 'meaning' is absurd.)
Rule is no more a matter (exclusively) of generalisation than it is (exclusively) of differentiation (cf. Shwayder 1965). In the social sciences, rules are recursive and nonlinear and, as Gleick (1988, p24) puts it: 'Nonlinearity means that the act of playing the game has a way of changing the rules'.
353
On Certainty begins by pointing to the peculiarity of saying 'I know...' if this is followed by something certain. From at least OC 288, Wittgenstein also tends to refer to this as something known. At 288 he asks: 'I know all this? I believe it', but then goes on to refer to a 'body of knowledge' (Wissenskörper). At 353, he asks whether or not it is legitimate to say the forester 'knows that that's a tree...' etc. And if it is permissible, then it's also correct to distinguish between a sort of knowledge where we may doubt and a sort where doubt is nonsensical.
((Perhaps it's this: there are still some situations where the particular game demands that we speak of knowing something quite certain — situations quite unlike the one Moore finds himself in. For instance: in cases of confusion (OC 349), where something's not quite clear or in psychopathological circumstances (OC 355). In 353, there appears to be a third person situation where it's also reasonable to speak of knowing. And if this is correct, how much more so in the comparative anthropological case? 'They (tribe X) all know that a certain poison may predict the future of men'. The method here seems to be to try the sentence on for size, as it were. But what procedure do we then use in order to decide whether it fits? And is there an embargo on describing the procedure's rules?))
355
Here it may be wrong to say 'I believe...', thus implying possibilities of doubt and testing. (This is how I should want to deal with cases of belief — but not with matters of certainty.) Yet at 288 it is considered correct to say 'I believe that the earth existed...', etc.
Isn't there something peculiar here? Are the cases of 'that's a chair' and 'the earth existed...' quite so different? Is there, say, a difference between empirical and non-empirical certainties in the offing? Surely not. What is relevant here is the question of 'non-doubting behaviour' and its relation to everyday practice. (One difference might be my degree of certainty in each case.)
357
What could 'the certainty that is still struggling' be? Perhaps it could only occur when some entirely new form of life is in the offing.
358
Throughout: I have accepted this way of regarding certainty. It should be worth asking what is bad about this expression.
((Nb: '(a) form of life'.))
359
Instinctual? Or institutional?
360
Again, 'KNOW' appears to be in order. Also: 'belief'. But it's still not as if we could do anything with this. And therefore they seem not to be in order. For —
367
When 'know' and 'believe' are made analogous, the consequence is that mistakes have to be explicitly ruled out. Whereas: there is no question of a 'mistake' (as such) here.
370
And (crucially) can't we ask 'How do I know...?' either doubtingly (as in scepticism) or in the further language-game of social inquiry?
((Or indeed both — as in phenomenological inquiry?))
Is there really a barrier here to constructing a mode of empirical investigation out of these 'negative problematics'? In one case 'How do I know...?' is asked as a move in the same game in order to bring in doubt, to destroy, disrupt. In the other case, we make an investigation of grounds and rules... as Wittgenstein himself may be doing. For example, doesn't Wittgenstein almost answer 'How do I know...?' by saying 'By virtue of the absence of doubt'? Nb: 'the essence of the language-game'.
371
There are certain things which anchor our ability to engage in games. But they are more settled than rules of procedure and it would also be peculiar to say we know them. In every case: we do not engage in doubting behaviour with them.
And doesn't it seem equally peculiar to say that rules, too, are matters of certainty?
((Applying the inversion of the method of trying the expression on for size — remarks on OC 353 above.))
372
But there is no implication here that we should not make an investigation of the practices (games) which our being certain about something anchors. We are certain about particular things to do with hands and we also engage in games which use expressions to do with hands. But in investigating these games-with-rules we do not expect this to involve us in statements about our certainties. Descriptions of rules, like rules themselves, can always be revised. Not so certainties.
374
(Cf. OC 44) So: we don't learn to do X by learning the rules for X. We seem to learn, as it were, by example. But does this mean that there are no rules?
((What methodological principle states, for instance, that formulations of rules should not be counter-intuitive? It seems entirely reasonable that they could well be: for this seems to be in line with the strong thesis that they cannot be objects of knowledge.))
That is, do we simply take the obvious and the certain 'on board' and immediately proceed to competent practice? Suppose I knew that, for tribe T, ancestors return as trees, the wind governs human fertility, bathing in a certain river cleanses the miscreant of evil-doing, and all the rest of these obvious matters. (That is, suppose I was taught this.) Could I then pass as a competent tribal member? The experience of anthropological fieldworkers indicates otherwise. Many report gradually picking up the skills and procedures of everyday tribal life and only then being able to reflect on the 'abstract' constituents of their mythology. Membership — competence in a form of life — seems to require something to connect its founding mythology with its observable practice. A term we might use here would be 'orientationals' (tacit maxims, rules).
375
It seems consequently pointless to talk of an arbitrary, hollow or false system of 'belief' (repression, ideology, etc.). If we could do this, we should have to count even a reliance on (our form of) arithmetic among the examples.
378
Here I am trying for a picture of the sort of ambiguity involved in 'knowledge is based on acknowledgement'. On the one hand, it seems to have to do with a resemblance between the language-games with 'know' and 're-cognise'. On the other hand, it seems to have to do with 'sociological' problems of recognition (as in 'I recognise your authority', 'He was recognised throughout the world as a genius'). It's only Wittgenstein's critical work on mental processes and mental phenomena generally (for example, at OC 47) that makes me tend to the second reading. So much in Wittgenstein connects up with this reading.
Yet the first reading is so much more like philosophy. Western philosophy appears continually to want to pose and answer problems for a sort of standardised individual, but an individual nonetheless. When it addresses knowledge, it asks 'how can one know?' — and likewise for the whole gamut of philosophical topics. The standpoint is invariably one of individual cognition, reasoning, or perception and so on, depending on the position taken. What is so peculiar about Wittgenstein is that he begins, as it were, with plurality, with a collective, and locates philosophy's problems in terms of its trying to answer collective questions in terms of individual units (objects, minds, etc.). The problems of philosophy then become, for the most part, (disguised) sociological questions. The crucially collective notions of 'form of life' and 'language-game' speak to this. Problems of knowledge, language, meaning, rule, order, certainty, grounds, evidence, criteria and so forth are all problems of (crudely) social location. It seems to be thoroughly misleading to try to handle these questions by asking what they look like for an individual. Perhaps one of the tricks language plays is to have us think that subjectivity precedes intersubjectivity. And even the latter term, when taken alone, seems to refer to a 'collection of units' with the stress more on 'unit' than 'collection'. Wittgenstein's work on knowledge, certainty and rule is fundamentally connected with a more sociological question of intersubjectivity (rather than, say, 'other minds'), of what makes social order possible. This is why we cannot yet be specific about a reading of 378. It leads off in many directions (PI 525).
((I would hold this, despite Janik and Toulmin's (1973) reading of Wittgenstein as a 'moral individualist'.)) —
388
A case in point: Moore's questions have nothing to do with proof of the existence of 'external' things ('how do we know that...?', 'how can we be absolutely certain that...?'): they have to do with what can be said where and the form of life that both determines that saying and is produced by it. The problem, as Moore poses it, might only arise for an individual sceptic. Moore tries to solve the problem by pointing to community, to common sense. But his version of 'common sense knowledge' renders it functionally similar to individual cognition. All Wittgenstein is doing here is to push Moore's 'solution' closer to concerns about community. And so the 'solution' becomes a dissolution: for the problem itself cannot exist in a collective social space. That is, what would it be like here to engage in a 'proof of the existence of external things'; to continue to conduct oneself as a sceptical individual over every matter and so on?
389
Here things are not shown by personal assurance or personal experience.
((Wittgenstein on pain: the discursively useful limiting instance where individual sensations can be spoken of — thereby showing questions of meaning, knowledge and so forth to be mistaken (or perhaps metaphorical) when dealt with in the same way.))
Whether or not something is known, acknowledged or recognised as a tree is, in one sense, a matter of community. Whether Moore does or does not know it therefore has to do with his membership. Reflexively, his membership has to do with what grounds are accepted, what rules are followed and (ultimately) what actions he is seen to be engaging in. And none of this has anything to do with the individual person called G.E. Moore.
391
And what isn't an 'ordinary' case? Miming? Children playing? And here there may well be no door involved. So doesn't this simply tell us a little about how we use the word 'ordinary'? Doesn't it also show the word's highly relativistic use? For, in the ordinary case of mime, it would be impossible for someone to really believe that there was a door. Then there's children and what's ordinary to the world of play, drama, fantasy, art and so on.
((One very peculiar use of 'ordinary' is where the ordinary (actual) and the theoretic are divided. For instance, where 'ordinary' language addresses the world, the film, the event and so on, and where theoretic language too addresses the same things, though in some supposedly different way, as if it were immaculate language freed from life practices.
Hunter (1978, p56) refers to 'theory's own self-understanding as "dis-embodied" knowledge'.))
392
Certain kinds of doubt are not possible in this community. Others are (apparently) possible but are not useful for anything here. And there is also the case, which Wittgenstein omits, of perfectly ordinary, reasonable doubting.
On the contrary: the possibility of the language-game (if we think of it in terms of social order) depends on plenty of things not being doubted. Yet some, but not all, of these things are able to be doubted.
'Contradiction': because some doubts are possible (even if they accomplish nothing) we can actually generate propositions which contradict, jeopardise or doubt the very grounds (postulates or axioms) which made it possible for them, along with other propositions, to be uttered in the first place.
((There is a superb moment in his Rules when Durkheim (1964/1895, pxliii) says that to call social facts 'objective' is not to locate them in some category of reality but to take a certain mental attitude towards them! Really: a doubt like this has no place in Durkheim's text. It could have no use in the practice of Durkheimian sociology.))
395
If this 'comes out' in the speaking and the acting, then one method in the behavioural sciences would be to track back from the speech and action to knowledge. This has been a fairly implicit methodology in work which has made inquiries about rules. But is it possible to do it? Isn't an attention to someone's speech and action which is oriented to the discovery of what the speaker knows as peculiar as an utterance of 'I know...'? That is, don't we routinely listen for other reasons?
((And isn't part of what Wittgenstein is saying: 'we routinely leave these matters unstated; nothing is to be accomplished by stating them'?))
The 'reasons' for uttering (and writing) must be closely related to those for hearing (and reading).
396
A disjuncture between knowing and saying one knows arises from the embargo (constraint, limit) on the latter. This is perhaps Wittgenstein's second version of 'whereof we cannot speak'. But here: I (or indeed Wittgenstein) can say that he knows.
397-398
In saying 'I know that the earth...', it is very likely (but by no means certain) that one will also think about the earth...etc. Perhaps, for the individual speaker, what is so strange about saying these things is that one has to think about things one has always known but never previously thought of.
((This may be unrelated to the problem of why it is peculiar to utter these things publicly.))
401
'...of the form of...': in one place these may be (or act as) empirical propositions; in another, though of this form, they may act as part of the very foundation of language.
From the Investigations: the surface form of a proposition and its employment in a specific location seem to be unconnected matters. Why does pragmatics or linguistics (semantics) take this 'connection' as a central task? Is semantics such an uphill battle because it is actually trying to forge a connection and not simply describe one which is already made? Take the following well-know example:
- John is easy to please.
- John is eager to please.
Why should someone even debate their 'similarity'? Can we get significant information about 'native speakers' by describing two things they wouldn't confuse (on pain of having their membership, sanity, etc., revoked) and the reasons for their not confusing them?
If someone runs for a bus, can we get information about him by asking why he didn't boil an egg instead? Perhaps: but there are infinitely many things someone is not doing at any point and only a few (if any) will he be pertinently not doing. For example, in cases of abstaining. (This idea is from Harvey Sacks.) Such investigations (except in cases of pertinent absence) seem thoroughly absurd. Yet it is a crucial bit of methodology in linguistics. As if, in saying 'John is easy...', we are actually refraining from saying 'John is eager...'.
402
Hypotheses are necessarily non-permanent. They may guide particular utterances and activities for a while. But they too will be underlain by, say, the grounds of a particular scientific discourse. Hypotheses (by comparison with fundamental propositions) can hardly constitute foundations. The opposite is true in most circumstances. (There are parallels here with Lakatos's categories for describing what can and what cannot be queried within a scientific research program.)
403
Here we might as well say: 'It is the truth'; being fully aware that we are using the word 'truth' more or less interchangeably with 'unmoving foundation of a community's language-games'. And also being aware that this is not how Moore uses it.
For relativism: absolute truths are to be found everywhere. They are by no means a scarce commodity. That the earth is round is an absolute truth and so is that the earth is flat. These can both be part of the unmoving foundation of different language-games. Whether the communities engaging in those games can communicate with one another is a separate matter. And doubtless, if they could, they would want to debate what counted as truth and what not. Deciding on and sanctioning truths seems to be just one language-game and not one that occurs particularly often.
404
Einstellung: is this properly rendered as 'attitude' — with all its psychological and personalist connections? No, this must be something always already held collectively. It is something like: 'the position the membership take with respect to...'.
Something's being certain is partly conferred. (As when we think of language as a 'convention' — we are uninclined to think of the convention actually being drawn up like a charter. Likewise we never think of the practical task of conferring a certainty! But 'after the fact', it can sometimes be useful to deal with it in this way.)
((What language would we use to draw up a language? What grounds could we rely on in conferring the title of 'certainty' on some proposition? None. But these things did not come about as a natural order of things. Something at least resembling a drawing up of conventions and a settling of grounds must have taken place. (The problem is most clearly addressed by Vico; and Lévi-Strauss has the same problem with the natural/cultural eidos of the incest taboo.) If this 'settlement' once happened quite rapidly: now we hardly notice its occurrence at all. All of this must be a metaphor or analogue, but I use it to push further away from thinking of language as a reflection of the (natural) world and certainty as a property inherent in worldly states of affairs. The metaphor shifts attention away from an objectivist (Durkheimian) universe, the place of language in it and persons' knowledge of it. (If one leg moves, so must the rest of the table.) It shifts attention towards a universe composed of human activities, of action and the deed. Convention, conferral and other deeds of covenant are merely parts of the metaphor but can be taken quite seriously.))
406
Here there is a tendency (especially in philosophy) to confuse things which are fundamental to (or at the foundation of) quite ordinary procedures with 'fundamental' (metaphysical) revelations about the world.
Wouldn't this be different again in sociology where, for instance, it might be decided that the former would constitute 'fundamental' analytic insights into the social world? This would be one of the tacit conventions of ensuing investigations. We might then get excited about quite mundane matters and discourse 'about' them. (The passage beginning 'Nothing could be more remarkable...' on page 4 of CV expresses this exactly.) But the excitement would be very different from that of the philosopher who thought that she had discovered in the same things a route to metaphysical certainty.
The difference is this: the philosopher might use these utterances to aid her in making discoveries about the world (as it were, straight off), while the sociologist's interest in them would be connected with discoveries about how some collectivity went about constructing accounts of events in the world. Compare the following: 'It is certain that X is the case'; 'A collectivity, C, is certain that X is the case'. And is the second an empirical discovery about a non-empirical matter? Doesn't most of what Wittgenstein says change its focus if we change our interest from a philosophical one to one which gears into making studies of what goes on in ordinary life? Doesn't Wittgenstein practically recommend such a shift of interest? In and as philosophy though? No: for the discourse of philosophical collectivities could well be investigated by the form of study Wittgenstein is indicating without the result itself being philosophy. Far from it.
407
When we refer to 'commonsense knowledge' we may be using the word 'knowledge' commonsensically. That is, in pointing to some fundamental certainty of a collectivity, members might describe their relation to it as one of 'knowing'. This seems fine. But not when we consider the different use that 'knowledge' has in philosophical language-games. (Also: statements of the 'I know...' type.)
408
Yet: what could 'slipping up' be like outside philosophy? — at least with respect to 'I know that...'.
409
The point here is: 'I know that that's a fact' is not something which indicates the speaker's knowledge. The stress, as it where, is elsewhere. This, unlike knowledge (in the form of practical rules, clues-to-action), simply could not let someone down. For this not to be my foot means that something fundamental about my world is no longer assured. If this goes — anything goes! Our knowledge, I want to say, is somehow built 'on top of' this sort of foundation.
410
This would tend to imply that some putative bit of knowledge about that system could have no value for us insofar as it could not be located within that system.
411
Of course, 'assume' is the wrong word here. For: assuming is something we do (an action or a thought) on the basis of the very things Wittgenstein here calls 'assumptions' (Die Annahme). Assuming is not something belonging to the foundations of our system of language-games, but that the earth has existed... does belong here (cf. Dilman 1973).
412
And the explanation would, of course, be one of the uses of the expression which the person was supposed not to be able to imagine. Where we legitimately say 'I know that this is my hand', we are (for example) pointing to our knowledge of the use of the word 'hand' — to the rules of its use. Knowledge and rules seem to be tightly bound to one another but not so tightly to the things Wittgenstein wants to say we're certain of.
414
This is the sort of passage which governs my reading of the others: 'What we have here is a foundation for all my action. But it seems to me that it is wrongly expressed by the words "I know".' So: we have to divide foundations of action from rules of conduct. It is the former which are not strictly objects of knowledge. And here it would simply be pretentious to say something like 'Well, in the ordinary course of their lives, of course people don't know about these things, but the special attitude of the philosopher, the phenomenologist or the sociologist (etc.) renders them objects of professional knowledge'. The same arrogance is not involved in descriptions of the rules which are in force on some occasion.
415
'[C]ertain propositions seem to underlie all questions and all thinking'. Underlie: but this underlying is not a knowledge we have, like a deep stratum. The underlying, as it were, is not part of the business of knowing, questioning, thinking, inquiring, etc. Rather it is somehow separate; it 'lies apart'. Its relation to 'knowing' is not direct (as it would be, say, in a formulation involving 'deep' and 'surface' structures of knowledge).
So, for Wittgenstein, founding propositions are equally unavailable to everyday knowledge and to scholastic (professional) knowledge. Not that there's any great distinction between the two, for Wittgenstein, except insofar as they partly constitute different collections of games (practices). Though, again, it is quite evident which of the two Wittgenstein prefers.
The importance of this for our inquiry is: in the social sciences there is a favoured device which requires one to say that, for everyday (non-scientific) activity, there is a stratum deeper than that empirically available to the actor and that this 'unknown' stratum is responsible, in turn, for his or her observable conduct. While it is unknown to the actor, it continues to do its work; it is available to the privileged 'sight' of the investigator and constitutes an object of specialist knowledge.
Wittgenstein, here, displays several problems with the device. Firstly: any form of life will have its foundational propositions which one cannot properly be said to 'know'. Second: this will be the case equally for professional investigations of this domain. That is, investigations will neither be describing features of everyday knowledge nor will they, thereby, produce special professional knowledges. What is underlying here is literally unknowable.
'The foundations of knowledge' seem to be an anthropological rather than a philosophical topic. It has to do with forms of life; with 'tribal' mythologies. Wittgenstein shows the peculiarity of claiming philosophical relevance for these (so called) 'knowledges' (re. proof of an external world and the rest). '[I]sn't the use of the word "know" as a pre-eminently philosophical word altogether wrong?' (OC 415). But he also seems to cut loose from their moorings anthropological inquiries into the 'same' matter. The topic seems to properly belong to this domain but we seem prevented from proceeding with it, just as much as we are in philosophy (though for different reasons). Something prevents claims being made to the effect that these grounds can become objects of any 'sort' of knowledge.
Is the anthropology indicated here, then, merely a form of practice itself — one which might save itself from theoreticism by having a strong argument against the denigration of 'mere practice'? If, on the other hand, theoreticism aims at a pure investigation of the foundations of knowledge, it constitutes itself, on this view, as a practice with an unknowable object. (Brecht had similar objections to certain varieties of Marxist theory.)
((Our inquiry here is not one into the grounds and conditions of knowing per se but rather one which asks how these figure in any form of life and what their relation is to the question of rule and the question of action or practice.))
416
The 'test' here is: if this can be doubted then so can anything and everything else. And doesn't the example here look like just any empirical proposition?
417
Knowledge endures or persists separately from other 'immediate data': images, memories (see OC 419), etc. Knowledge, in this respect, is a matter unascertainable from the contents of consciousness.
A phenomenological investigation of the contents and processes of experience will not inform us about (social) knowledge. Why might we think it will?
418
This is nonsense if 'is' is taken to be the sign of identity.
A possible re-write: understanding is accomplished only within the confines of a collectivity's (already given) version of the world. These confines involve, metaphorically, a screening out of other possible versions. In these other versions, I would lack understanding, not being trained in them. In part, I accomplish what understandings I can by remaining blind to the others. And this sounds better than, say, Schutz's 'epoché' of the natural attitude'; the willful suspension of scepticism and the rest of the baggage.
All we point to here is the irrevocable location of all practice within a form of life, its grounds and its rules. When we say 'not an object of knowledge' we point to the mistaken belief that the 'irrevocable location' can be cancelled and knowledge produced in the resulting 'neutral space'. Of course, the possibility of this space and its 'pure' (scientific) knowledge are myths underpinning certain scholastic practices; all irrevocably located, anchored in their 'realities' (to use Mehan and Wood's cumbersome term). Like other myths, this one is not self-evidently true. Rather it secures what can pass as truth in certain forms of scholarship.
419
In these cases, we are dealing with something that lies beyond the question of 'known truths'. What is to pass as knowledge and as truth rests on the unquestionability of these propositions.
420-421
This sort of disjuncture is like a kind of culture shock. The world appears differently — even with respect to once familiar objects, and so on. The rules of the game, and which games are to be played, become unclear. Re-learning and un-learning come on to the agenda. But where we make transitions from one form of practice to another, doesn't the transition (as a practical matter) make sense on both sides? (The inhabitants of both banks are familiar with the bridge.) So isn't the sudden inexplicable transition that Wittgenstein describes in this passage an impossible one in any practical sense? Doesn't it run counter to everything in the world of routine events? (We think: it would have to be a practical joke, a mock conspiracy, a drama and so on. See Fowles's novel The Magus (1975).)
It is nevertheless interesting to note that the stakes involved in such (putative) transitions involve our sanity. Though I don't want to say that the insane have necessarily made transitions of this kind. The view would be too close to the romantic position of Laing, Cooper et al. All I am indicating here is the perimeter of a form of life.
What is the use of such (erstwhile) transgressions for describing the limits of language and community? Are they any more than an illustrative device; a step saying 'Look how absurd this is!'; or are questions of the transgression of limits and transference between forms of life serious propositions? What could be accorded the status of an occasion on which 'things such as I don't dream of at present ... happen'?
422
'Like pragmatism': this may literally be pragmatism, the doctrine whereby any utterance is estimated solely by reference to its practical bearing on human interests. Except it is 'like' pragmatism insofar as the scope of terms like 'estimate', 'practical' and 'interests' are extended beyond their narrower uses in (traditional) pragmatism.
'Here': when 'I am trying to say something...' or when we try to imagine the transgressions and transferences of 420-421? In the first case: the Weltanschauung is that of a conventional taxonomy of philosophical terms. In the second: it is what Wittgenstein has been trying to address all along as a mythology, as the very conditions of our knowing and so on (the system of convictions, Rules1).
'Kind of': for there's something analytically unsatisfactory about 'Weltanschauung'. Like 'pragmatism', it seems to have a possible application here — but not one that would accord with its uses in any mainstream philosophical discourses. There would have to be a fractionally different use — a different discourse — and hence a transference just as impossible or unlikely as that in 420-421.
This gives one account of Wittgenstein's reticence in these passages (and elsewhere) — 'It ... seems so' (418), 'like a madman' (420), 'I am trying to say...' (422), '...like pragmatism' (422), '...a kind of Weltanschauung' (422). Not only is the imprecise expression sometimes exactly what we need; sometimes it is all we can give. By contrast, the formulation in the Investigations makes it sound as if there were a choice (PI 71).
Wittgenstein's tentativeness may be partly due to what many investigations encounter as a double-bind: what we would express is 'the conditions of expression', including (ipso facto) the conditions of our investigative expression. The choice would seem to be between the idea of an ultimately ungrounded form of expression, on the one hand, and silence, on the other. But is there even this choice? Here we seem condemned to practice: with no comfortable pedagogic hierarchies of theory and practice, event and commentary, and so on. What is thereby denied the investigator is a certain (dramatic) irony that has long been the hallmark of Western pedagogy: knowing something which all 'on stage' do not (or rather knowing they 'really know' something while they don't know they know it). This last parenthetical expression captures the irony of retaining two senses of 'know': natural unreflected folk knowledge vs. the clear analytic vision which tames the 'primary', 'savage' kind by means of categorisation and the promise of a full inspection.
If we are to insist on the pretence of analytic or theoretic 'distance', then we get tangled in our own webs, spun from double-uses of 'know' and their conflation.
423
What is 'fishy' here is precisely the relation of Moore's expression to the double-bind sketched above. His use of 'know' forces us to separate philosophical from 'ordinary' communities. If 'know' is to be used here, it has to be allowed a special sense. But what is the warrant for that sense? Only this: the imposition of philosophy's (self-understood) authority over some 'ordinary' discourse where that hierarchy is itself a product of philosophical discourse. In locating such 'knowledge' as this as a condition of our expression (any expression), Moore denies the very possibility of his own utterance. (For: it, in turn, would have to be conditionless, groundless.)
Wittgenstein is not insisting on a version like this: Moore's expression is fishy vis-à-vis some ordinary expression; since ordinary expressions are the authorised standard, Moore's expression is deviant, deficient.... Rather, Wittgenstein is entirely dubious of the ordinary vs. non-ordinary binary. That is, he admits to the meaningfulness of Moore's expression 'in particular circumstances'. So we may be tempted to say 'Yes, there's only one ordinary use, in particular circumstances, in daily life...' but then to add, 'Then there's Moore's use, in philosophy, which is ahistorical or a-situational...'. But in fact, this list of qualifiers does not rip Moore's expression right away from particular circumstances and into some other space of commentary. Rather it furnishes just another possible set of particular circumstances for the expression. (Here they will have to do with particular philosophical traditions and institutions.) 'Outside these circumstances' does not mean outside any whatever.
So Wittgenstein does not display unbounded reverence for the 'ordinary' as against its other: rather he merely refuses to accept the misleading and ironic view of certain sorts of discourse as extra-discursive, extra-ordinary.
The dubiety of 'ought it to?': this acknowledges the potential trap of repeating or reproducing Moore's separation between philosophy and common sense. If the separation is accepted, Moore's expression looks fishy vis-à-vis ordinary uses. But this is not the point — for the problem with the expression is that it requires such a division in the first place. If the separation is not accepted, we should also have to suspend certain of our problems with Moore. For example, we should now have to acknowledge how his expression makes perfect sense in certain ('philosophical', 'exemplificatory', etc.) circumstances; where these descriptors (suggesting the extraordinariness of the discourse) would be taken critically, as part of Moore's own, uninvestigated, self-understanding of his enterprise.
There is something fishy and also something perfectly understandable here. Ironically, going along with Moore down the divided paths of philosophy and common sense makes him look all the more fishy. (The same is true of Austin's objections to Ayer; the discourse of sense-datum theory existed, for all its self-ascribed privilege.)
Most importantly for us, in 422 and 423, is the reminder that what ultimately grounds (any) discourse cannot properly be called 'knowledge'. (For wouldn't knowledge itself be an effect of those grounds?) And given our denial of a folk vs. professional knowledge hierarchy, we should have to say that the ultimate grounds of discourse (and therefore knowledge) cannot be known.
((The still unresolved question is whether this is coterminous with work on 'rule'. The problem has certainly been constructed this way on enough occasions, here and elsewhere.))
((One possible way of formulating the problem of accounts, descriptions or ethnographies of events in one form of life (say Azande witchcraft) by means of events in another (say, correspondence in American Anthropologist) would be to remember that cultures engage in continual borrowings, emulations and parodies of one another (Sharrock 1974). Here, then, the ethnography would be a re-write or borrowing of events in one discourse according to the requirements of another. In these sorts of cases we can refer, perhaps, to analogic or metaphoric relations between two 'texts': event-as-text and ethnography-as-text (cf. Clifford and Marcus 1986). Occasionally we might find multiple stages occurring, with intervening (and also analogous) texts produced: tape-recordings, films, transcriptions, observation notes, interviews, gossip, etc. We can see ethnographic forms (indeed all social science descriptions of events in the 'social world') as analogues of what they describe: of the same order of events, attentive to and productive of, for example, the same temporal and spatial sequences, the same detail, etc., as what they describe. Here: what's retained on one side of the analogue and what's not will vary — just as I may have a graph with axes analogous to some other graph (both logarithmic for example) or with a curve analogous to it, or both. The formulation of accounts and events as analogues of other accounts and events would thus have to acknowledge Mary Hesse's (1963) further distinction between positive, negative and neutral analogies. In such a way the myth of 'pure description' makes way for analogue, parody, bricolage, plays of difference and identity. In short: writing in Derrida's (1976 1978) sense. Thus we might legitimately ask what structural anthropology has learned about the doability of structural anthropology from having investigated certain South American Indian uses of natural products. (Discourses reproduce themselves, not other discourses.) In asking such a question, however, it might be worth recalling how that asking itself would be analogic. (Here 'analogue' is suggested as but one relation of discourse to discourse that might occur.)))
424
The uses of 'know' here all look post-initial; they don't appear to be able to occupy the opening turn of any possible conversation. And some look like instances where a speaker has been 'put on the spot' by a previous speaker's turn. They also look like defences, excuses, apologies and so on. Yet Moore's use ask for a different context so that they can be used like initial utterances. But what could follow from them?
((Seeing what could follow involves the production of a text like On Certainty.))
To include '...it's not surmise' is to defend the utterance. Doesn't it seem peculiar to use a defence when no accusation has been, or could be, made?
425
Does the empirical givenness of the tree have anything to do with the truth of the assertion about the tree? Yes: but only if we are using a very specific means of legitimating the truth of an expression.
(a) There are several things which we don't doubt in this form of life. (b) The form of life also involves techniques for distinguishing between true and false propositions. There is nothing to necessitate isomorphism between (a) and (b) as techniques within a form of life.
What is odd is that a particular technique of distinguishing true from false, which we call 'empirical verification', should be cited here ('disclosed', 'confound').
What are 'the circumstances that give this sentence meaning'? I think again of defending an issue. This is like in OC 591. These would be the sorts of circumstances where we might want to find out whether we were dealing with the weak case, 'I believe...', or with the stronger case, 'I know...'. But with some propositions, there could be no question of this kind; even 'I know' would be too weak. And if these fall: so does everything else — all the knowledges and beliefs which hang together with them.
So what would it be like to be 'fallible' about this? Say, where the speaker was lying, being deliberately misleading, joking, acting in a play? (All the cases that Austin and Searle dismiss as non-serious or 'etiolated' uses.)
This is problematic for, according to 425, there would have to be propositions such that:
(a) 'I know that...' could be uttered of them, but only in certain circumstances;
(b) 'I only believe that...' could be uttered of them, but only at the risk of being completely misleading;
(c) I cannot be mistaken about them; yet
(d) I am not necessarily 'infallible' (unfehlbar) about them.
((Can we even say that a person is infallible?))
Given the sense in which Wittgenstein talks of certainties being defeated (that is, their turning out to be empirically false), we can only assume that this is what is meant by 'infallible'. Certainties, then, would suffer from the possibility of empirical disqualification.
And doesn't this itself seem to involve something like a complete revision of everything else we hold to be the case (about the empirical world)? Here, we should probably want to consider the fallibility of the empirical evidence before considering ourselves to be 'fallible about' the matter in hand (if it were a matter of satisfying conditions a, b and c above). OC 429 is like this.
Is this a case: Castaneda holding the proposition 'Men cannot fly without the aid of special apparatus'?
((One important correction emerges from 425. Namely: those versions of relativism (relativist idealism) which have held the grounds of a culture or a form of life to be akin to beliefs, acts of faith or arbitrary agreements (because it is possible to cite examples of their relative inconsistency across cultures) are quite mistaken. Mehan and Wood (1975) address this same problem as 'the fragility of realities'. To make such claims is to assume a privileged language, severed from any form of life and purely describing cultures and their (relative) (in)consistencies. Consequently, there could be no practical correlative of such a 'discovery' (whatever Berger says about 'transcendence' and 'ecstasy'). What grounds a culture is so firm that even the (usual) 'harder' re-writes like 'corpus of knowledge' are inadequate. The idea of 'fragility' is nonsense; dangerous nonsense.
An example of this: relativist idealism argues (for instance) that propositions like 'Men and women are each suited for exclusive sets of tasks' can be shown to be arbitrary beliefs (and so forth) by virtue of numerous cross-cultural counter-examples. (Margaret Mead's Samoan work and its supposed influence on the rise of U.S. feminism in the 1960s would be a case in point.) Does the utterance itself dissolve the problems generated by such a hard and fast 'belief' (?) in any given culture? (Instead we could use the term 'assumption', 'certainty', 'taken-for-granted', 'expectation', 'prejudice' and so forth according to speakers and locations.) If not: what would be the practical correlative? In what domain could knowledge of this kind be used? Struggles for social and political change remain unaffected by relativistic talk of the supposed 'fragility' of current dominant forms; for those struggles cannot be carried on in the supposedly ahistorical domain that these relativisms require in order to think of cultural grounds as (only) 'beliefs'. All struggles are struggles within practice and within history. Talk of 'fragility' may be suitable as a rallying cry or recruitment slogan. (At recruitment time, the war is always portrayed as easy and victory as close to hand.) But no actual strategies or tactics have been formulated on the basis of propaganda.))
427
Here we'd want to say: he could not conduct himself this way without knowing (?) such and such. In order to act this way he would have to know (?) — what prevents the subject of our inquiries from saying 'I know...' surely prevents us from saying 'He knows...'. Problems of knowledge have nothing to do with problems of mental operations.
From this (472): it seems as if there is an important connection between rules of conduct and 'the thing we are concerned with' (which may or may not be appropriately formulated with 'I know...').
In order to (correctly) see such an exhibition, we should need to be fellow-members. But what does that entail?
428
If we consider the primordiality of practice (including linguistic practice) over and above the 'literal' meaning of what is said — then we could ask what sort of a move uttering a sentence with 'I believe...' is. And the incongruous answer would be: it's an attempted move in a game which, in order to be played at all, requires the utterer to hold what she says she believes in a much firmer way than 'I believe...' expresses in this game.
429
In 425 there's the idea of something (a tree) being 'disclosed as something different' or 'confounding me'. And this appears to be connected with my 'fallibility about' the most ordinary and certain things of the world. There's much in this that looks like empirical disconfirmation.
By 429, however, these most ordinary and certain things in the world appear to 'dominate' experience itself: as if experience could not change them without the prior acquisition of, or training in, an entirely different form of life.
Training would be an example of 'previous experience' and the events or practices of training could not constitute grounds for propositions like 'I have ten toes'. Rather these propositions (?) themselves appear to be grounds for everything else (including, for instance, my relative certainty about previous experience).
Here I want to question the term 'proposition'; for there are troubles with it in this context insofar as it refers to a linguistic item, perhaps even an utterance. Hence I really ought to refer to my certainty over the matter of having ten toes (should it be called into question, etc.) rather than my certainty about the proposition 'I have ten toes'. But: if grounds do not exist in the form of utterances, how do they exist? Where would they be located? (All references to mental equivalence are out of the question here.) Again: why should the question have to be asked when everything speaks for our (unspoken) acceptance of grounds. I use 'for' to mean both 'in favour of' and 'in place of'.
Note: the unspoken acceptance, in many cultures (except perhaps sports commentary), of an prohibition on stating the obvious.
For an expression to make sense, it should speak for (tie in with) everything that is obvious in the game where it is used, but it should never speak of anything so obvious.
Is Moore's use of expressions like 'I know I have ten toes' actually a breach of some 'taboo'?
((Wittgenstein's textual strategy might be this: there are certain limits to language (etc.) which cannot be formulated in language. They can be seen when I continually hit up against them with Moore-like utterances — utterances close to the edge. This is like a magnetic field whose limits become visible only by virtue of the fact that certain objects can't penetrate; or gravity, whose qualities cannot be described but whose effects are overwhelmingly visible and measurable. 'What is gravity?' is almost a cosmological question (parallel to the questions of ethics and aesthetics). 'What does gravity do?', however, is no cause for wonder or debate. The 'rules of its existence' are unavailable while the 'rules of its conduct' are only too plain. Cultural studies, like natural science, should perhaps confine itself to questions it can answer.))
430
The enormity of the cultural disjunctures which must be envisaged before Moore's expressions can apply: talking to Martians! And isn't there a sort of craving after alienating disjunctures ('in order to find out about ourselves')? We would like to think: it would all come clear if we had access to a totally alien presence. However: do all of our certainties have such scope as to pertain to humankind in general? What about being certain that the Earth is (roughly) spherical? And where we encounter people who do not hold fast to this, we talk about their beliefs — as opposed to, say, our knowledge — about the shape of the Earth. Both formulations have their problems. As with social collectivities, so presumably with languages.
431
First paragraph: 'I show this knowledge...' — so I must have it, but there is something troublesome about saying 'I know...'. And (re. the notes to OC 427): how peculiar that I can talk about it (as knowledge) when it comes to writing a passage in a philosophy book (or in On Certainty) but that something prevents me from saying 'I know...'.
It may be the case that I show this knowledge: but if 'I know...' cannot be uttered of it, this is as far as it must go. (a) Something is shown in what I say (OC 431) — (b) so what would be shown in an utterance about the 'something' that is shown in what I say?
((Only peremptory formulations like 'theory and practice', 'philosophy and everyday life', 'sociology and the social world', 'resource and topic' can save us.))
At the end of the day, what is shown in the second case, (b), would be the same as was shown in the first, (a). Thereby the second question becomes absolutely superfluous.
Second paragraph: certainly, we might gather that the person is sure of his ground, knows his way around and so on. But only if some special investigation were being carried out. That someone should be familiar with the terrain on which they live is one of the most ordinary things in the world. Even to 'gather' this assumes the air of a doubt or a check. And the assurance, 'I know...' would also assume that such a doubt had been or could be raised. All of this usually passes without much consternation however.
Even 'the conclusion that I knew' is not compelled: as if there were some kind of constant calculus going on (or some regular appraisal of the world in constant motion) which we could (if we wanted, where there was a doubt) pay attention to or leave running on its own. (Here: cultural competence would be like those digital watches where you press a button to get the read-out if you want it but where the seconds keep ticking over anyway, invisibly, if that button is not pressed.)
We do not wait upon the order of the rule; but neither do we summon it to meet the circumstances.
432
And this term 'knowing' is not quite correct here: for it seems to be divorced from other sorts of evidence and from any particular circumstances. And still we want to ask: what is it that's going on here (as it were) 'behind' my conduct and which we continually want to gloss with 'I know...' and 'my knowing...'?
We want to put a name to it: yet it facilitates the very practices of naming; so we get a problem like that of (a) and (b) in the previous set of remarks (OC 431).
433
There is an important difference between, on the one hand, 'I know...' and the circumstances of its use and, on the other, claims about human knowledge.
Doesn't Moore's use of these propositions (those that 'lie apart from the route' we routinely travel — OC 88) tend to assume that they are contextless.
((What criticism of Moore is being made here exactly? Earlier (OC 26-105) Wittgenstein relates the Moore propositions to the 'system of convictions', the 'mythology', the (relatively) hard rock which grounds our talk and action. And the trouble, in those passages, is Moore's use of 'I know...' and his claim that an attention to these propositions can lead to 'absolutely certain knowledge' (Austin 1962), to the fundamentals of inquiry. Wittgenstein does not seem to deny the importance of these 'propositions' but claims instead that, insofar as they act as grounds for all we do and say, they have (a) an overwhelming ordinariness and (b) problems associated with their expression and its contexts. In terms of everyday practice these 'convictions' are both trivial and crucial. There appears, therefore, to be a double warning. The first warns us against taking them too seriously, vis-à-vis their centrality for philosophical debates about, for instance, absolutely certain knowledge or about the existence of an external world. The second warns us against taking them too lightly, vis-à-vis their importance as anchors for, for example, linguistic activity. Either way: there is a further warning about possible uses these expressions may have and the extreme care that must be taken here concerning what can and what cannot be said either with or about them. Wittgenstein appears to use them, then, as an occasion for displaying (or pointing to) the limits of language. Here, as in the Investigations, Wittgenstein's dissolution of philosophical problems takes the form of showing how problems and questions only arise once the limits are illegitimately transgressed — that is, once mistakes are made about the possible uses of words. In OC, Wittgenstein looks through many of the possible uses of 'I know...' and shows how Moore could not possibly be using his sentences like that. The contradiction is like the one which arises in OC 468. In 433, again, legitimate uses of 'I know...' are the re-stress, the assurance, the defence. Already Wittgenstein has shown the fundamental absurdity of doubting in these cases. But here (433 and as far back as about 300) there seems to be the pretence, 'supposing someone however absurdly did doubt' (for instance, in the case of philosophical idealism). How, then, could his doubts be answered? For if they are absurd, there is surely an answer to them. (The parallel with solipsism is an interesting one in this connection.) If one use of 'I know...' is as an assurance against these doubts — this would be how we should actually reply. And this seems to be what Moore does in 'Proof of an External World' (1959). But still, even in the case of such a reply, the utterance and its circumstances would not reveal anything of the philosophical importance attributed to the propositions by Moore.))
The problem that remains, on this reading, is: Wittgenstein seems at least to concur with Moore that there is something fundamental here (and that the problems lie in the expressions that can arise when we attempt to say just what). One position arising from this might be that there is a universal apparatus or corpus underpinning all action and speech but, by virtue of its very existence as an underpinning, it cannot be addressed in its own right. Indeed all the attempts to synonymise or express the grounds of expression in our present investigation would suffer from the same in-built deficiency. The core of the problem is this: we start with something 'that sounds like pragmatism' (OC 422) and so giving up on expression at a certain point can sound like mysticism. That is, there is at least one investigation (of the expression of the grounds of expression) whose procedural requirements would always remain unsatisfiable. Or: there is something about language which prevents us from ever discovering its mechanism and that 'something' is at least part of the mechanism. Furthermore, we can show how speakers' conduct continually orients to this limit, this unspeakability — while this, itself, cannot be directly addressed. A generalised logical grammar ticks on in practice but remains unavailable to theory (for there is no practice-less theory ideally distanced from the grammar).
If such a reading of OC can be made, it would echo many traditional Russellian readings of the Tractatus as 'pre-relativistic'. The parallel or analogue could be shown this way: according to the Tractatus doctrine of saying (aussagen) and showing (zeigen) and its picture theory of the proposition, a state of affairs is asserted by a proposition by virtue of its having the same structure as the proposition. Proposition and state, word and object, and so on, are isomorphically structured. This 'common structure' cannot itself be asserted but is nevertheless shown (zeign) in the symbolic expression. When philosophers attempt to address the question of the structure of a proposition, they are prevented from doing so. At best, their statements will only show or display that structure. The structure itself remains something 'whereof we cannot speak'.
So too, in OC, we find an analogous problem. Here too the conditions of expression cannot be expressed. Thus:
Conditions of expression :: Structure of a proposition
The rest of the analogy follows.
In a sense, Wittgenstein's ideas about the topography of linguistic structures or conditions have changed between the Tractatus and OC, yet that terrain, whatever its topography, remains impenetrable. The candidates for what is to count as structures or conditions have changed, but their unaddressability remains constant. In passages like OC 427 and 431, the doctrine of saying and showing seems to be alive and well, if reformulated.
We are being warned off a certain sort of investigation of language which happens to coincide with almost everything that's been done so far. A vacuum — silence — nihilism — no way of working: only these negative possibilities seem to present themselves.
Only the most absurd relativist would be satisfied to leave things at that. Work could only be done in the form of neutralish correctives on linguists, philosophers, etc. No positive thesis could be advanced as a counter-mode of investigation. So while the linguists and philosophers were upbraided and downgraded, all our work could consist of would be the practice of reprimand. Whatever we thought of their work, ours would be dependent on their work getting done; a kind of eternal opposition. (Cf. contemporary semiology's utter dependence on realist discourse as a point for the application of critique.)
I would like to show in my reading that, if part of the argument involves showing how speakers' conduct continually orients to the limits of language, then such analytic work might be an end in itself.
From OC 427: '... his conduct exhibits the thing we are concerned with'. And from 431: 'I show this knowledge day in, day out by my actions and also in what I say.'
And so a form of analysis might take its lesson from this but instead of the investigable object being 'the thing we are concerned with' or 'this knowledge', and instead of stalking it through 'surface' investigations of conduct, action or speech, the order of things could be reversed. Working from the problem of limits and the impossibility of forging ideal and perfect formulations of general conditions (that is to say, working from reasonably definite notions of what can't be done), empirical investigations of naturally occurring conduct, speech and action might be undertaken — with the particular location itself constituting the confines of our generalisations.
((Compare The Blue and Brown Books: 'the contemptuous attitude toward the particular case'; 'the craving for generality' (BB 18).))
Work on the 'surface' business of the management of discourse, utterance-by-utterance, does not appear to suffer from problems of generalisability.
One immediate topic might be: given that ideal-perfect formulations of language can't be made, how do (for example) conversationalists formulate what they're saying, what they've said and what they're about to say? (See Heritage and Watson 1979; 1980.) For example, how do they summarise their own speech events, in Garfinkel and Sacks' (1970) sense of 'formulate'?
((Much of the above can be found in Garfinkel's (1967, pp4-7) work on 'indexical' and 'objective' expressions — if we use it as a basis for reflecting on objective expressions about the problem of expression and their possible non-existence.))
It would be quite wrong though to suppose that an investigation of the above kind would 'solve' the unspeakability problem. A further problem from (early) ethnomethodology displays this. Conversational analysis, it should be mentioned, may not suffer from exactly this problem though it (and other 'sociolinguistic' investigations) will certainly suffer from a parallel one.
The problem: early ethnomethodologists held that the social order emerged out of methodic practices engaged in by persons in their everyday affairs. They held further that the coherence, planfulness and regularity of those practices was displayed with, in, and as those very practices as they unfold on any actual occasion. They took as their object the discovery and description of the above state of affairs, as a continuously empirical answer to the question of how the social order is possible. They referred in this case to the 'reflexive' or 'incarnate' character of the accountability (methodicalness) of everyday practice. Assuredly, they held, every discursive event entails a display of its own methodic character ('formulating' being a very explicit case in point). But: once this was established in principle and some practical investigation had to be carried out, ethnomethodologists became uncertain as to whether the methodicalness they described, as a property of the occasions under investigation, was a feature of the occasions' accounts or instead a feature of the investigative account or description. Was the domain or the model generating the methodicalness? According to the criterion of reflexivity or incarnateness, the description given by an ethnomethodologist would itself necessarily have to display its own 'proper' methodicalness. And for all practical purposes, the two levels of methodicalness were indistinguishable. (Though this could be solved by assuming universality and invariance.) In this respect ethnomethodologists have by and large been dubious about referring to'rule' exactly — but we can at least see how what we're calling 'rule' would suffer analogous problems — problems which, up to now, have only been avoided by assuming that pure descriptions of speech practices can be given; where 'pure' stands for 'uncontaminated' by any features of that which is (ostensibly) only described. The problem in ethnomethodology consists in no more and no less than this (and this would also be the case with analogues of the problem): to say that methodicalness is a feature of actual occasions is only to say that whatever the analyst locates as methodic will be called an immanent property of those scenes. The immanence, that is, is ascribed.
434
What is the 'but' at the end of this paragraph? Perhaps this: experience may show us what sorts of behaviour are exhibited by someone's knowing something (for example, that a man can find his way about a certain house; that his judgment can be trusted... and so on) but, experience will only teach us that people know something if we are prepared to equate their knowing and their behaving in certain ways.
Of actors 'knowing' social rules, Garfinkel says:
That is a very interesting notion of 'know'. The 'know' here has to do not with what one might have in mind in some secret place. It is not a case of your having to calm a respondent or seduce him in order for him really to tell you. Then you would be illuminated on what he had been hiding all along. Instead, 'know' consists in a structure of activity. That is what the 'know' consists of. It is not that the member has it somewhere in the nervous traces or that he has it according to a theory of personal action. (Garfinkel in Hill and Crittenden 1968, p47)
How did Moore find out, or learn, that everybody knows (with certainty) that there is at present a living human body... and the rest of his propositions? (Assuming, of course, that the relation of 'all' people to these propositions is such that they can be said to know them.) Did he find this out through experience? Or did he experience the world in such a way that any person acting in the world would have to be acting in this relation to those propositions? If so: it would then be easy to see why Moore, with his philosophical training and so on, would want to formulate that relation in terms of knowledge. Then, because of certain underpinnings of action and because of certain trainings, we can see how Moore would claim to have discovered a class of absolutely certain propositions (some of which, like 'Here is one hand, and here is another', have a special role in the proof of an external world). But 'behind' this claim there are always the practices and the trainings. So Moore is right in every respect but one (and a crucial one): what makes the fundamental fundamental is not its status as knowledge, where one's conception of knowledge is something like holding a store of substantive propositions, but rather its status as practice (allowing that one might - though Moore doesn't — have practice as one's conception of knowledge).
A reading of the above kind assumes we can distinguish between versions of knowledge as (a) a sort of mental storehouse plus the store's contents and (b) a term collecting various sorts of (overt) practice. But can we? And, even if a strict separation could be made, what would persuade us against the idea that each is some kind of regular transform of the other?
If we bifurcate knowledge into (a) mind and (b) practice: firstly, neither seems to take us intrinsically further that the other and, secondly, we have no distinct evidence of their incommensurability. Is any sort of choice available here? Perhaps we only 'feel' that knowledge has been traditionally dealt with as a mental store of substantives and should like to give an alternative account a chance to display its cogency. 'Perhaps it is, after all then, a matter of faith whether we want to give a materialistic or an idealistic account of knowledge.' Yes: but which is of these accounts could ask this very question? The other seems to embargo questions of that kind, of choice. That is, one of the accounts would want to include the argument that practice and training (rather than choice) is 'behind' the giving of that very account itself. So to ask the question, 'Perhaps it is, after all then...', assumes one of its possible answers to be less legitimate than the other. It fails to give equal opportunity to the contenders. By the time we get around to formulating this 'choice', one of the alternatives it proposes is already preferable.
435
These problems about the nature of knowledge may be in no way connected with the use of the word 'know'. Is a concentration on the use of the word a means of un-bewitching ourselves?
Philosophy is the attempt to be rid of a particular kind of puzzlement. This 'philosophic' puzzlement is one of the intellect and not one of instinct. Philosophic puzzles are irrelevant to our every-day life. They are puzzles of language. Instinctively we use language rightly; but to the intellect this use is a puzzle. (LC p1)
436
Claims to certainty: here we seem to be making a mistake about the relations between our uttering something (which, say, no one whatsoever would disagree with) and the state of the world ('beyond' our language and its uses?). But what is the mistake?
Is it just the assumption that things (in the physical world, for instance) unconnected with such an insignificant act as someone uttering something should have to be bound by what goes on in language?
Or is does the mistake lie in assuming that there is a separation in the first place between utterance and object? Removing such a mistake: even the most remote non-discursive entity as God would be a product of linguistic practice and that practice would render certain statements incapable of falsehood, by virtue of consisting (in part) of procedures for the demarcation of truth and falsehood. For: on this view, what else could be marshalled as counter-evidence apart, that is, from further utterances?
Either way (positivism or relativism), claims to certainty, even for truisms like Moore's, are mistaken. For the positivist: the empirical world is pre-linguistically available for scrutiny and no claims to certainty can be made until a scrupulous empirical investigation has been carried out. For the relativist: the world can never be pre-linguistically available so that claims to certainty will not be pure and simple representations of how the world is but always embedded rather in certain trainings and practices. OC 436 can be taken in both ways. In the first case: we feel humbled beneath the immensity of whatever lies beyond social, linguistic or cultural order. In the second case: we see perhaps the peculiarities of accepting the first view; its idea of 'the real' comes to seem heavily ironic. The practices of reading as deathly serious and as undercuttingly ironical — two practices? Or two 'ways of knowing' the world?
437
Everything so far in On Certainty would lead us into saying that this can have no consequences (except under very peculiar circumstances — such as where someone expresses a doubt where doubt is hardly possible; as it were, when he speaks for something which nothing speaks for and everything against).
Using the arguments of the Investigations: where something routinely has no possible consequences, one is not able to go on — and 'going on' would be co-terminous with the practice of understanding. These propositions (claims to certainty) do not seem to have a place in routine discourse when tied up with the sorts of things Moore claims to be certain about. That is, we do not go around claiming the impossibility of the falsity of truisms! Hatch (1978) discusses the sorts of empirical discursive devices that arise out of this basic conversational embargo on stating the obvious. Sacks in his lectures (for example, 1971, September 10) also talks of the near-total restrictions on speaking of something one knows one's interlocutor to know already.
But in many of the cases where we have an analytic interest in this phenomenon, it turns out that the thing which should not be told (because the teller already knows the other knows it) should not be told because the teller has already told it some time earlier — and this is why the other happens to know it. Thus — unless I have an atrocious memory — I do not tell you twice in the space of five minutes that I just changed my 'phone number. But this case of the other already knowing is very different from the case with something which is not told because everyone knows it. 'I changed my 'phone number' second time around is different from 'the earth existed before my birth' any time around. (Though both can be done as jokes, we should remember — or as, say, mock displays of lunacy.)
In this respect, there is something at least interesting (as Wittgenstein says in OC 437) about Moore's propositions insofar as nothing does, nor can it, (usually) follow from them. In an actual case we might get 'So what' and its variants — deprecations of the previous speaker's competence. Hence these truisms present an interesting paradox for cultural studies: namely that knowing them is a condition of competence in a culture, yet speaking them is a breach of that competence.
Cultural studies (semiotics and so on) seems to reserve a special place for the crucially known and the equally crucially unexpressed. But this is only a paradox within a particular view of language. In other respects, not speaking or writing is a perfectly definite and positive discursive practice. It is not that silence is the opposite of speech or writing; rather it is a form of it. On an extremely instrumentalist view: 'passing over in silence' is a perfectly legitimate speech act along with promising, denying, accusing and the rest.
Does this dissolve the problem? Hardly: for we now have to say that there are a class of things such that knowing them (if 'knowing' is the correct term) is crucial to competence and such that passing them over in silence is the relevant means of addressing them. (Just as psychoanalysis is the relevant means of addressing certain 'unspoken' desires.)
Returning, however, to the formulations in OC 431 and the Garfinkel passage (remarks on 434): this 'knowing' that is so critical to competence is not a separate matter from practice. Having established that, in this case, the (universally?) competent practice is to pass the things in question over in silence: we cannot now say that something is first known and only then treated this way. Rather this 'knowing' is that treatment, that practice. Simply this, then: there are matters which are crucially passed over in silence as a condition of competence (while others are optionally treated this way). We can talk of someone 'knowing' these things and (hence) fulfilling but a single condition of competence when we observe them acting in this way. (Instead of 'knowing' being in quotation marks, it would be better represented under erasure: knowing.)
This single condition of competence has become crucial in anthropology, linguistics, cultural studies and so on insofar as almost all formulations of rules in these fields would be treated this way. If there is a domain of rules, we can be certain that they are of this very kind. (This is why the Moore propositions will do for now as placebos for them.)
The very speaking of them would either be (a) a breaking of them or (b) impossible or (c) really another instance of following them. Otherwise: to speak or write these rules would require a language in accord with some other, entirely different (new, alien) rules. If, then, the rules of utterance are universal, we, as the utterers of those rules, must be able to place ourselves outside the universe. To believe we can do this must be in the category of a false belief.
If, on the other hand, the rules we want to express are not universal, we must have some alternative society (sub-universe) in which that expression can take place — an alternative membership. The relevant problem here is: can this be done by choice — as an act of volition? How could such a counter-practice be established if, prior to its establishment, it were necessary to find out and enumerate one by one the features of the practice it opposed? (Remembering that this counter-practice is also established precisely in order to do that very job.)
If these sorts of limits are actually working in relation to the linguistic, cultural and social sciences, how would we account for their current means of coping within these constraints? The answer to this question would seem to inaugurate a possible field of study. But what would be involved here?
((Are rules objects of knowledge? Yes, insofar as they are 'objects' of certain discursive practices. Does this mean they can be addressed? Yes, but silence is their correct form of address. Do we risk anything more than a charge of incompetence when we attempt to utter rules? Yes, we risk either the charge of holding the mistaken belief that analysts can locate themselves beyond the universe they address (transcendentalism) or the charge of holding the mistaken belief that linguistic, cultural, social, historical, political (etc.) alternatives can be voluntarily constructed (idealism). Where is the warrant for the use of 'mistaken belief'? In the very conception of rule that you start out with.))
Only let's cut out the transcendental twaddle when the whole thing is as plain as a sock on the jaw. (LE p11)
438
Yes: but the grounds might well already have been given. In most (ordinary) cases, we will not actually give them there and then.
((For example, where do we have to present our credentials? Think of the relation here to 'trust'.))
With many things of this kind: could we actually even point to some (prior) occasion in which they were given?
439
What carries the weight of conviction is this 'everyone takes it for granted' — it remains unspoken.
((Silence and power — Gramsci.))
The double sense of 'suppose': society, language and culture may be enforceable as what is supposed to happen.
The OED on 'suppose': Assume as a hypothesis, imagine. In this sense, it is easy to make idealistic formulations of the problem of rule. But there is this case where 'suppose' acts as a reminder of the very opposite — the fixedness (on any occasion) of correct procedure:
This ambiguity catches the very ambiguity of rule itself: we feel that there's a natural fluidity here; that rules can be imagined, 'kicked around' like hypotheses. And treating them this way certainly adds facility to our investigations. But: rules never operate this way 'in practice' (as if what we thought we were doing were not 'in practice'). Analytic work on rule always requires this (relativistic) indulgence: as though the theorist could try these things on for size while the subjects of her investigations were doomed to something more fixed and rigid.Bill: That wz a jo:ke people, (.) Bill: That wz- [ Ellen: Yeh. (.) Bill: That- En yer spoze tuh smi:le (Jefferson 1979, p94)
440
There is every difference between an investigation which merely includes the statement: 'There is something universal here...' and one which begins 'There is something universal here and this is...' and which then goes on to (attempt to) spell the matter out.
But yes, there is a continual requirement of grounds for claims to knowledge, at least insofar as they may be demanded if someone is not convinced. In a sense: they only become visible by a kind of default. How then should we convince someone that we had grounds for knowing the grounds of certain forms of conduct?
We fool ourselves here that there are unfolding levels or layers (grounds of grounds, rules of rules, etc.). But could this be the case? (And isn't it soothing that it is not the case?) An illustration: in a famous book on French verbs, the grammar begins this way:
Definition du verbeOn appelle verbes les mots qui expriment qu'une personne ou une chose existe, est dans tel ou tel état, fait telle ou telle action. (Verbs are those words which express that a person or thing exists, is in such and such a state or performs such and such an action.) (L'art de conjuguer, p4)
When defining certain sorts of verb by saying that they show something exists, we do not say that the word 'exists' in this last expression is part of some meta-verb (because for example, it encompasses a whole class of other verbs including itself-when-not-used-meta-verbally). It is no different from that which we use in the expression 'The table exists'.
((Hence a number of mistakes in conceptions of scientific truth:
Given some language in which we can write statements, its meta-language will be the collection of statements about that language (rather than in it). For example, if we talk in French about sentences in English, then French is being used as a meta-language for English. Any meta-language will, in turn, have its own meta-languages; for example, someone might be talking in Spanish about us talking in French about the English language. This hierarchy can continue indefinitely. There exists an infinite hierarchy of logical languages, the subject of each being the one below it. Logical statements in a particular language cannot be called true or false without stepping outside that language to employ a meta-language. (Barrow 1988, pp260-261).))
441
'I know...' used (a) as an assurance or re-assurance and (b) to convince someone: both uses can be available; but some circumstances can cut off one or the other. For example, at law there is no room for (mere) assurance. With 'I know that's a hand': are we dealing with one or other of these uses? That would depend — but who would need to be (re-)assured or convinced that that's a hand? No one. So the force of the assurance as it were passes to the speaker in a case like this. There may conceivably be a circumstance where we need to be able to find whether someone is operating with(in) the same system as us and if he did this we could be (re-)assured in this respect. (This seems to be almost the only sort of circumstance Wittgenstein could be considering here.) What is most significant here is that the doubt does not get directed at the question of whether or not it's a hand but at the question of who's operating with(in) what system. If we doubted the business about the hand here (that is, if we tried to make certain uses of 'I know it's a hand') we would have to doubt everything else, which is not possible. And to have to say 'I know that's a hand' must imply an earlier doubt. Where something in the scene does not quite fit — and someone uttering 'I know it's a hand' would be such a case — we seem to have to search for weak links in the arrangement: the occasion, the speaker, the utterance and so on. But it is only in the last resort that the whole system (as in OC 141-2) might be a candidate weak link. When we're forced to select this, something fundamental in our lives has already broken down. Better to say that there is something dubious about a speaker's adherence to our system. But, with 'hand' and 'the existence of the earth' and so on, this is (practically) to specify the speaker's insanity. Still: if, as Wittgenstein says, we know the circumstances and if the utterance is okay in this respect, then this must be 'an assurance that the person...is normal in this respect'. For the gelling of utterance and circumstance is one of our means of recognising normality. (But these seem to be very big ifs; and notice that Wittgenstein is not tempted to say just what the appropriate circumstances might be. What comes to mind here?)
The world I inhabit is such that it is inconceivable that persons should not know that this is a hand when one is held up. But it's almost as inconceivable that anyone should find an occasion to say out loud, 'I know that that's a hand'. The propositions that ground our form of life contain among them certain embargoes on what's to be said and what's not.
442
If we imagine we know, what's wrong with saying 'I imagine I know'? Simply this: one who imagined they knew could (or would) only say 'I know'.
Moore (in the Defence) would say that I cannot know p to be true if p is in fact false. But this seems only to show that Moore uses 'know' in a very peculiar way. In fact this proposition can only be taken as a sample case (for, say, illustrative purposes) of Moore's strange use.
((How often, in philosophy, is this what we're actually doing: inventing strange uses and then seeing if certain stories can be told by consistently exploiting those uses?))
To imagine I knew something (which I turned out not to know) would entail my also imagining a system of grounds in which I could know it. Could I do this all at once or does imagining one knows take a particular training, as does knowing? (For example, Tolkien records that he knew it was full-moon for a certain battle in The Lord of the Rings. He knew, given such and such. Both the knowledge and the grounds were imagined. Here, imagination is more like a shift in the key than a shift in the musical notation. The same rules apply — but they are transposed. Or: a spectral shift in physics.)
443
Here we have a formula (or mnemonic) for translating statements with 'know' so that we're unlikely to think they involve some mental equivalent of what is known. That is to say: here we prevent knowing from becoming a mental process by translating statements with 'know' into statements about circumstances, events, etc. The probability of a mistake in the knowing is similarly given.
Here: to know something and to have grounds for knowing it become questions of practice, and of training in that practice, respectively.
Already Wittgenstein has made it clear that to ask for the ground of the grounds is to ask the impossibile. The grounds are groundless. If having grounds for knowing something is a matter of training in certain practices, then to ask for what underpins those practices (rules, regularities, methods) may be the same sort of demand. But is this the same thing as saying that those practices' overt displays of orderliness cannot be described? The systematicity of this 'element' (medium) seems to befall us — yet it must be abundantly evident (if the doubts about it mentioned in OC would bring down the whole system). In this respect we should perhaps prefer to say that every display of it always already constitutes a description of it with no further (analytic) ado. What then are these analytic re-descriptions that we once thought to be write-ups of the rules? Well, many ordinary activities get formulated on the scene of those activities, as we have seen (Heritage and Watson 1979, 1980). That is, we often not only say x but also add a comment on what it is we are doing in saying x. We say: 'Well it was only a joke', 'Are you trying to tell me I'm not wanted?' and so forth. But these sayings about we were doing previously must remain sayings nonetheless, and subject to their rules.
Pulling together the Tractatus and OC, Brand (1979, pp16-17) outs the matter well: 'I cannot place myself outside language and speak about it, that is, about how it describes the world. One cannot speak about what can be said. One can only say it.' An extreme relativist would say language and the world are unconnected. A less extreme one would say they are so intimately connected that we cannot get beyond either in order to describe how they are connected.
A similar problem exists with 'on the scene' (members') formulations of their talk — for they do not bring to any final closure the sense of the occasion. It goes on being produced (even) during such formulations. In this sense it parallels professional or analytic discourse about discourse. At some point we simply have to acknowledge self-evidence.
444-445
In the first case (444) we add the grounds on which our knowledge is based in order that those we address may decide its reliability (or otherwise). And the grounds in this case have to do with the reliability of particular activities: looking something up in a timetable and so on. In the second case (445) we look in vain for a statement of grounds because there are no trusted practices which we can turn to. 'I just looked again at them' (see OC 430) adds nothing whatsoever except perhaps a doubt as to my own normality. And this is only because the very reliability of our checking procedures is itself grounded in the stability of matters like 'I have two hands'.
Now we want to ask: is it the fact that these propositions stand fast or is it the operation of certain reliable practices that grounds our action and knowledge? This does not really seem to be a question at all: for the two are identical. There really does seem to be something arbitrary (and to do with a rather unimportant linguistic convention) about whether we talk of this in terms of such-and-such-is-the-case (nominal) or such-and-such-is-done (verbal). There is a point where the complex dialectic between them ceases — a point where they become (for practical purposes) identical.
This is (almost) analogous with the wave/particle separation. Positive analogies are:
- The structures are all-pervasive though minute;
- A choice of two competing 'forms' can be made but this becomes relevant only at a grosser level;
- The question of their empirical observability is irrelevant; for they are always amply manifest at grosser levels (such that the phenomena at these levels could not be as they are without these grounds being such and such).
However (applying this inversely): does the nonexpressibility of grounds mean we are in the presence of something fundamental? Surely not: for I can be equally stumped for similar additions to certain sorts of (mere) belief. 'I believe the dead are watching us': where I should have to add 'I don't know why, I just have a feeling about these things'. This is why I think systematic grounds are often called 'beliefs' (as in 'the beliefs of this tribe about fertility'). We often call them 'beliefs' because they are different and distant from the things we happen to cling fast to. In our form of life they would be beliefs, matters of faith or superstition. And, as we saw just now, systematic grounds (when expressed in the form of propositions) have this process-grammatical similarity in common with beliefs: they are difficult to follow with indications of their likelihood. I still don't think we should continue to exploit this (dangerous) confusion.
((For example, Brand (1979, pp7-8) compounds the problem by referring to 'incontrovertible belief' — this may exist but we can easily imagine things to be otherwise. Is this also true of 'I have two hands'? I should like to reserve the term 'belief' for what we take on faith. The incontrovertibles, by contrast, we should never consider to be matters of mere faith. Imagine: 'I have two hands. I don't know why, it's just a basic faith I have always had'. Here we should want to know what someone was doing making a faith out of something so ordinary. Our faiths are usually matters which require a leap into the unknown (rather than the incredible). We can doubt our faith, but this too must come to an end. Yet where it comes to an end is not a matter of faith but of certainty. We may be unsure of someone's beliefs, but to imagine for a moment that he would have trouble with 'I have two hands' is preposterous. If there is a connection: beliefs are sometimes called personal convictions. Even if this is a contradiction in terms, it is nevertheless instructive. Beliefs are (properly) personal matters. To have beliefs is to inhabit a world where there are such things as individual persons. And for this reason (that is, because not all world systems are alike) we can imagine someone being entirely without beliefs. Peter Ustinov (1978, p21) writes of his grandmother's propensity for bursting into tears over the young Ustinov when telling various stories from the Bible such as the crucifixion. Of his father he says:
No doubt my father was subject to the same treatment, perhaps to an even more passionate degree, for his mother was younger then, with a more objective and even colder fear, which may explain the fact that he was the most irreligious man I ever knew, not in the sense of blasphemy or of agnosticism; he merely completely ignored the whole business and never seemed to feel the need to accept it or reject it, or even fear it as a superstition. It didn't exist.
Belief is (or can be) of this sort. But could the same be said of matters like my having two hands? Imagine: 'I used to believe that but it soon lost its fascination for me'.))
446
This kind of certainty is nothing remotely like a personal conviction — it is nothing that I hold in some secret grotto. As Garfinkel says (see comments on OC 434), it consists in the structure of an activity.
There seems (again) nothing to choose between 'the game rests on this' and 'this is presupposed (vorausgesetzt) in the game'.
Graphically: it is there before we set out.
447
And what is there holding together these two propositions? Without them a system would collapse. 'A system is, so to speak, a world' (PR 152).
But here: I cannot speak about the system which is built up through the overlapping of these and other propositions. I can only speak with(in) it. I should not know where to turn for an alternative system with which to address it. For the system itself provides the notions of given and alternative, among others.
((Sometimes we think these question are abstruse but in fact what's being addressed here is the impossibility of addressing the world without remaining part of it. And now it almost seems not to make sense to address what is after all an impossibility. The only warrant is that it has the use of pointing out an illicit step which has been taken in almost every area of social science. While the sort of contentment Wittgenstein draws on in, say, OC 471 may be philosophically respectable, it's hard to see how the social sciences would face it. If any discovery and elucidation of features of the grounding of human activity is ruled out here, how can the social sciences recover? What alternative analytic practice is there?))
449
Social scientific work has always held (at least implicitly) that it could elucidate this foundation and also (sometimes) the training by which it is imparted. Recently it has been discovering that such elucidations themselves involve a transgression of this foundation. One position which has survived this re-questioning advances the view that professional analytic work need not be subject to the same foundational limits as the work (activity) which it attempts to describe. This, too, can now be shown to be methodologically troublesome. (I omit any criticism of its moral consequences.)
Here Wittgenstein is again showing the fly out of the fly-bottle. And this is okay in philosophy. But in the social sciences, the fly-bottle looks a much more comfortable proposition. For, once outside its confines, once the question of describing (or otherwise accounting for) human activity is made problematic, social science enterprises face revisions which could render them almost unrecognisable. (Sociology, for instance, would be able to make coherent and 'primary' investigations of social change with respect to its own methodology.)
450
With the construction 'perhaps that is a...', one would be trying, if this were the only form used, to train someone to doubt everything. Could anything be learned from this? No: there simply could be no practical consequences.
Practice and certainty: these are so directly coupled as to be almost synonymous. Our engagement in any activity whatsoever, including the act of refraining, shows our certainty about something.
If Wittgenstein is right about the dependence even of doubt on the existence of a foundation of certitude, then there is always something which anchors our activity. The doubt which doubted the existence of rules could not be a doubt.
Here we might want to say: when we address or formulate a rule or rules, we are simply acting under the compulsion of a rule (perhaps the same one, perhaps one which we don't address); but this doesn't prevent us from producing rule-descriptions. And would we be right in saying this? For: in order to make this claim, we should need to be able to distinguish, a priori, rules-of-description from rules-for-description (roughly: analysts' and members' rules). But what would be the grounds for such a distinction? (One rule for science, another for life?)
Nothing whatsoever supports the idea of rules-for-distinguishing-kinds-of-rule. There can be no such mise en abyme if, that is, we are able to consider rules as features of the grounds of action. At some point in the regress, we would lose sight of practice (and hence also of rules). If explanation does come to an end somewhere (a phase shift), then there is a finality about the stratum of rule itself, and our problem of (in)scrutability is not yet solved. There is something terribly naive about the idea that we can just go ahead and describe, speak or write about a rule.
Are we then going to say, of something of this kind, that we (can) know it?
451
But here too we should require, for this rule to make sense, that a reasonable doubt had previously been raised. Say, for instance, that a movie set had been constructed in a place where there was a (single) actual tree. This tree becomes part of the set along with all the imitation ones. Only in these extraordinary circumstances (a) could a reasonable, grounded, doubt be made and (b) could, therefore, such a thing as this be sensibly uttered.
In 451: when one 'make[s] the meaning more definite', that very action destroys the indefiniteness of Moore's game. It destroys, too, his perplexity over these sorts of utterance. For here, 'making more definite' means showing the familiar paths that lead off in every direction (PI 525) from the proposition in question. But, comparatively, what leads off from 'That is a tree'?
452
Whether it is reasonable for something to be beyond doubt is not open to personal decision. It is through and through a public matter (for the language-community in question). Perhaps we look for a rule by which such matters could be decided — and this is sometimes how the word 'rule' is used. Yet such a rule could not exist; for membership of this community means that such decisions never get made. If we must use the word 'decision': what is beyond doubt is always already decided (for us).
((The operation of a rule can have multiple outcomes. But here: the 'outcome' is always singular.))
The putative 'rule' in 452 is a rule-before-rules. Here we can have no clearer instance of the fact that there is no rule for deciding between rules. No rule over-rules the rules.
453
Law too is practice and must rely on something. Here, too, the judging has to come to an end somewhere. The learned judge is only as wise as anyone else when it comes to the question of where a doubt is reasonable or unreasonable. In English law, one is proven guilty 'beyond reasonable doubt', not 'beyond doubt'. We could, that is, always find some 'carping' doubt. But what is a reasonable one is given by the whole character of the world we inhabit. 'This doubt is reasonable'. Why? 'Because the way the world is for us, things could reasonably have been otherwise'. For example, the 'murderer' could have been in another town that day but he could not have been on Mars.
'Reasonable doubt' ... 'reasonable person': there is no difference here.
What can be doubted is given to us through the fixity of the undoubtable. This supplies the criteria of reasonableness.
Again, 'Here no reasonable person would doubt' would most likely be said when someone had for some reason proposed an unreasonable doubt. This is what the 'Here' refers to: where someone has doubted or looked like doubting something absolutely fundamental; something which would have to be undoubtable in order for anything whatsoever to get done — including the expression of a doubt.
454
For there is a connection between logic and reasonableness. The logic of a language and what the language-community 'finds' reasonable are very closely related. There is a relation between 'One can legitimately say...' and 'It is reasonable to hold...'. Wittgenstein, in these matters, always seems to ask whether or not something can legitimately be said (for instance, in 453). That something can be said seems to be related to considering certain sorts of circumstance (the ones in which it might be said) to be ones which would occur in this (present) form of life.
Expressing x and x being reasonable or unreasonable are related. Doubt is only one case in point. There are restrictions on the expression of the unreasonable. The reply 'that's unreasonable', for example, involves a very harsh downgrading of the prior utterance.
455
In 2 x 2 = 4, we're not involved in working something out. But where does the recognition come to an end? Obviously, with 2426 x 18 = 43668, we do have to work through to find an answer. The fact is not immediately recognisable. But where, between the two 'multiplications', does one procedure (recognition) give way to another (calculation)? (Think of our training: of learning tables up to, say, 12 x 12.)
'But if you work these two out using a calculator, you must take similar steps in each case.'
Yes, but what a peculiar thing to use a calculator to 'work out' the first of them. We might, for instance, be checking to see whether the calculator were functioning properly. So even here, the two 'multiplications' are behaviourally different (different as matters of social practice — this shows that the position adopted here is not behaviourist). There is every difference between the things we might check and those we cannot check because they are fundamental resources for (for example) the very procedure of checking itself. In this case, it is probably correct to say that we 'recognise' them. But in another sense, they are so familiar that even this is hardly in question.
((Recognising is a little like calculating here. Just as we don't calculate 2 x 2 = ..., so we don't strictly recognise it. Its familiarity is beyond both.))
So: if recognising is not quite right here, what does the language-game rest on?
456
Past a certain point, doubt makes no sense — for doubting is an activity and the very grounds of that activity cannot be subject to doubt. Because the activity of doubting is (merely) a linguistic activity makes no difference. To think we can doubt everything is to think of language as a form of direct commentary on the world yet distanced from it; whereas language is actually an inseparable part of our worldly activity.
Certain sorts of doubt are fantasies. We have to be able to imagine another world before we can imagine them.
((For sure: these problems can actually arise, as in the case of Pollner's (1975, 1987) 'reality disjunctures', but only where some actual collectivity materially construct another world, as it were. Here anthropology is only useful because it highlights what is always already under our noses. It actually has us recognise 2 x 2 = 4 and the rest. Philosophy, too, 'is in fact the synopsis of trivialities' (LC p26).))
But everything that's accomplished (including the comparison of cultures) is accomplished within a culture — as a repetition of its own techniques. And so to accomplish anything we have to accept some other things as perfectly fixed.
((Of course, whether things are 'really' fixed this way is not a reasonable inquiry: for the acceptance of them supplies what might be called 'our sense of the real'.))
There is no not doing this; and most of the time there's no debate about it, no selection between competing cultures. At this point cultural relativity ceases to be an issue. (It takes us here but can't be used in investigations of, or actions in, a single culture — which is to say, all actions.) In its place we can begin to talk of self-evidence. Now, we can say that 2 x 2 = 4 is self-evident, and so on.
So this is just the wrong sort of question: 'How is self-evidence accomplished?'; even though good deal of energy in social science has been dissipated in the search for an answer to it.
The question of 'cultural production' is much more immediate (visible) than this conundrum. Human life is essentially practical: looking for the structures of practice is like looking for something more essential. This is perhaps what Marx had in mind in his Eleventh Thesis.
At the question of material practice our spade actually begins to turn.
((In Marx: is material practice what is meant by 'the real foundations' or is there supposed to be something more fundamental still? There do exist readings of Marx where some unstated base is supposed to lie deeper.))
457
'...resides in the nature'? Perhaps it is better to say that a number of concepts interlock here: certainty, rule, knowledge, language-game and so on. The residential locations in this case are mutual.
Syllogism might be such a language-game; though a highly restricted one to be sure. We step through from premiss to conclusion, in this case, without trouble. All whales are mammals and mammals are warm-blooded; so all whales are warm-blooded. This is a cogent move in the game: we are certain of the procedure we have undertaken. All looks well. We feel confident that having followed the rule for the production of syllogisms, we can now claim to know that if such-and-such, given such-and-such else, then such-and-such more. These fibres are interwoven. (And this would all be true even prior to the specification of the 'rules' of syllogistic reason by Aristotle.)
And here too things come to an end. We cannot ask: how is it we know that if ... and ... then ...? We cannot specify the procedure we undertook. We do not get beyond our certainty that this is the case. Likewise, in experience, we are not aware of something going on (like a computation) in us as we move from premiss to conclusion. Even where logicians can stipulate such moves, where ethnomethodologists can reconstruct rules, where social psychologists can deliver cognitive maps and so on, nothing persuades us that these representations actually occur (except our subscription to the conventional procedures of these — theoretical — language-games). They — the theories — furnish instances of the insurpassability of rule, certainty, game, etc., even where, on the face of it, they appear to address topics like 'analysis' or 'description' of this or these.
They are not described or analysed here. How could they be?: for they are resources essential to the production of descriptions and analyses (for example). Here we never attain pure topics for description (that is, ones which do not 'contaminate' the resources used for that very description).
Any official theory, of whatever political colour, is, for ethogenics, a problem, not a resource! For all we know, each reality, as constituted, is an aspect of a composite social world and is a truth. (Rosser and Harré 1976, p173)
This is an example of a methodological criterion which can never be ideally satisfied.
When we earnestly inquire 'But how do we know...?', the perfectly ordinary shrug (or 'We just do') is not an answer which shirks the problem. It is an answer (even though it does not point to a process). And if the answer is: 'Intuitively!', are we to accept it too? Perhaps: providing we don't then get a discipline which attempts to plot the anatomy of intuition.
Most of the time the termini of human language and action seem to be terribly remote or deeply buried (Chomsky, for instance, or Freud). In Wittgenstein we are led into considering the possibility that our haste towards this 'core' passes over, in the very early 'surface' stages, the phenomena we seek.
458
Doubt arises when the certainties we unspokenly rely on do not 'come together' in the correct way. But here we do not doubt the grounds. Rather we ask what could have gone wrong; and we only select an answer from those things which indeed could have gone wrong. (A gasket has blown — but not: the whole engine has mysteriously turned to paper.) The certainties we rely on themselves inform us (negatively) as to what this class of possibilties is.
Thus, if I lose my watch and search the house for it and if this search includes looking in my watch's usual place on the dresser and if I then later find the watch to have been in that very spot all along, I then have a number of things I can doubt and a number I cannot. For instance I can doubt whether I looked properly in the first place; but not the fact that my watch is unable to move of its own accord. What separates these? One doubt is quite all right; the other, if entertained, throws everything else into disunity, and this 'everything' is not an overstatement. The whole of our collective life would be threatened.
There may be societies, however, where one's senses are always held to hold fast but where, on the other hand, animation can be ascribed to objects like stones (which we hold to be inanimate).
'...objects like stones which we hold to be inanimate': the very language seems to baulk at this sort of expression. It becomes 'poetic'.
(('...Yea plants, yea stones detest, / And love...'. (John Donne 1967, p32 — from 'A Nocturnall Upon S. Lucies Day'.)))
459
There is a point to which absurd, merely frivolous or obsessive doubts cannot be taken: like checking each apple. Being absurd (and the rest) requires the same grounds as more routine sorts of doubt.
At what point do we begin to ascribe mental illness? How far into a radical disjuncture with our normal not-doubting-points must one go before this sort of step is taken? And isn't this step simply a way of re-inscribing normal procedure in the face of some potential break? 'Ascribing mental illness': nothing is more normal(ising).
So in this extreme case — which is also a highly practical instance — the 'psychological terms [like 'mental illness'] merely distract from the thing that really matters'. Jeff Coulter (1973, 1979a, 1983) works through this matter in ways which are enormously significant for social science (cf. McHoul 1988).
Here, the 'thing that really matters' is the practical organisation of the world.
I also want to take up the question of belief (both the religious and surmising varieties) here because, as we have seen, it is very easy to talk of the grounds of that practical organisation as 'beliefs'. And this seems quite wrong. Coulter (1979b) also works through beliefs as features of practical organisation (as against the mental constants which putatively ground it).
In this respect we can show religious belief and surmise to be (a) related and (b) practical accomplishments rather than the grounds of practice. We might do this not out of of any intrinsic interest in these topics — an interest peculiar to the sociology of religion, speech-act theory applied to acts of believing, etc. — but because some social theory wrongly holds (a) that the fabric of social structure relies on collectively shared beliefs (or on 'betting' on how the world is, in the surmise-version) and (b) that these beliefs are mental processes or entities.
In these respects we might look at the sorts of procedures involved with the term 'belief'. Extrapolating from Wittgenstein's division between religious and surmising beliefs, we could set up the following:
- a Belief that — propositional belief
- b Belief in — moral/normative belief
(And Coulter makes a similar distinction.)
With a, belief is always weak next to, say, knowledge claims. We say: 'I believe we're 50 miles from Brisbane'. And the driver says, 'No, we left over an hour ago'. So we revise our belief, knowing the sorts of speeds we have travelled at. We don't say, 'I know we're 50 miles out...' in the first instance. This procedure with 'believe' is almost reserved for cases like those in 458, cases where some quite reasonable doubt is involved.
We can see this clearly in the case of ironical replies to questions. In the following case: Pye saying he 'believes' something beyond any possible doubt (namely, that he just said something in particular) is an ironic affront to the competence or even the sanity of his interlocutor:
'Sark.' 'Yes, sir,' said the man in the little quayside hut. 'A return fare. Six shillings.' 'A single my friend,' said Mr Harold Pye. The man in the little hut looked up and frowned at the unfamiliar face. 'Did you say a "SINGLE", sir?' 'I believe so.' (Peake 1972, p7)((This passage offers a great deal of material for conversation analysis (McHoul 1987). For instance one interesting feature of it is the negotiation of a certain sort of (almost) blame. The question is: which is most peculiar; that some 'unfamiliar face' should want a single ticket to the tiny and exclusive island of Sark; or that the person in question should have made a mistake in his request and have any possible correction there may be in the offing initiated by the ticket clerk?))
With b however, we are inclined to say that belief of this sort is relatively stronger. It seems to have a different relation to knowledge and certainty. For instance, we tend not to get parallels with the a-cases, like:
- I believe in the Lord Jesus Christ!
- He never existed.
- Oh, I guess you're right
(except, again, as jokes and the like).
But this b-variety of believing is limited. Most of the time, belief entails the view that what is believed is open to doubt. (Although it may have to be actively protected from it.) It is merely believed as against being known or certain. Coulter says that 'I believe...' is a downgraded claim compared with 'I know...'. Also, in the third person, we say, 'John said he'd been to Mars and Joe believed it!' to show Joe to be a fool.
It is not impossible to imagine a conventional language-practice where all religious (or b) belief was treated in the same way. Under this convention it would be peculiar for theists to make claims that they believed X, Y and Z and expect other people to see that it was therefore something strong and fixed. In this convention, saying you believed X, Y and Z would presuppose the weakness of your claims; it would be taken as evidence of the downgrading of X, Y and Z not as evidence of their likelihood of being true.
((In both cases, 'believe' has an apparent relation with the 'personal': a-cases with a proffered opinion, up for grabs; and b-cases with a personal conviction. Perhaps the family resemblance rests in the possibility that b-cases are made in order either to identify and consolidate a religious group from non-members (as with the Credo) or, relatedly, to confront non-members with claims to strong and fixed propositions which they might previously have thought of as nonsense. Then: a- and b-cases would both relate to this matter of downgrading.))
a- and b-cases of belief, contrary to the possibility raised in 459, appear to be much more closely related to doubt than to certainty. a-cases are plainly uncertain expressions. The b-cases are doubt-productive by comparison with certainties. To speak of a certain matter as one of 'belief' itself raises a doubt.
Either way: it would be peculiar to talk of the end point of all our doubting (459) as 'belief'!
((This way of treating belief (cf. Needham 1972) foreshadows a lot of trouble for phenomenology. An example is Husserl's concept of doxa, a character of all believing, not only believing in simple positive certainty. This short phrase is anathema to the above analysis of belief. How can there be an act of consciousness common to each and every case of believing? How can all believings be conflated? Our relation to 'simple positive certainty' is not one of believing... and so on. These sorts of problems make a good deal of trouble for those who characterise Wittgenstein as 'a phenomenologist purely and simply' (Brand 1979, pxxv). Many symptoms in Brand, such as his odd distinction between 'interpretation' and 'Wittgenstein-immanent text' (pxxi) detract from Wittgenstein's strongly anti-phenomenological position in passages like OC 459 — a position which is consistent from the Tractatus and before, right up to the last paragraph of On Certainty.))
Another example which the above version of believing could immediately show to be troublesome follows:
Man lives in the meanings he is able to discern. He extends himself into that which he finds coherent and is at home there. These meanings can be of many kinds and sorts. Men believe in the reality of these meanings whenever they perceive them — unless some intellectual myth in which they also come to believe denies reality to some of them. (Polanyi and Prosch 1975, p66)
Perceiving a meaning? The reality of meanings? Meanings with reality denied them? Belief in the reality of meanings? I can imagine Polanyi and Prosch being 'at home' in this. But is it right?
A final example of a problem which becomes clear on the Coulter-Wittgenstein account of belief: how can this request even be made?:
Elaborate on the observation that caution or certainty about a belief reflects belief about that belief. Discuss the support appropriate to such second-order beliefs. (Quine and Ullian 1970, p93)
460
Do points at which we are no longer prepared to doubt (shown by the superfluousness or absurdity of 'This is a hand...') actually furnish us with important instances of 'stopping points'? That is: are these limits sites at which we might raise the question of rules (as such) and their knowability? There does seem to be something absolutely basic being rapped against here: but is it the phenomenon of rule? If not, a good deal of my investigation is in jeopardy. (And we should then perhaps have to turn to the 'obvious' work on rule in PI 139-242 and RFM 303-353.)
A reason for raising this point here is that the example in OC 460 seems to deal more with a single (candidate) rule than with the phenomenon of rule generally. The instance would be a rule we have already encountered, one which penetrates a good deal of, but by no means all, discursive sites and which can be roughly glossed as 'the embargo against the statement of the obvious' (cf. Hatch 1978). Is there any good reason for us to suppose that this particular rule over-rules?
((To show Moore's propositions to have that rare (rule-displaying?) location (or non-location) in everyday discourse would be a very suitable way to read Wittgenstein. In this case: while Moore considered his own strategy as one in which the 'unarguable' truths of 'common sense' were stated, he would be now shown instead as one who actually proceeded much more closely to Garfinkel. That is: supposedly locating candidate 'transgressions' of 'common sense' (breach studies) in the attempt to display 'what we have known all along' (but were afraid to ask). ))
Here we see the centrality of Wittgenstein mentioning the rare but imaginable cases where (even) these (Moore) propositions come to be uttered. That is, we might hear them in terms of something having gone awfully wrong, or as a rare instance, a special case. Whichever: we are still left with the problem of the relation of utterances and their location as opposed to (what we hoped to extract here) a possible solution to it.
So in a fundamental sense, all 'empirical' attempts at locating the rule or ground face a crisis. And Garfinkel's search for the methodic bases of everyday interaction would be just one case. This is because transgressive instances merely look like the negation (difference) which displays the existence of the rule; yet all that happens in these instances is that we move under another rule's aegis. How else should we make sense of Moore or Garfinkel (or of Lévi-Strauss, Chomsky and so on)? Even if a rule could be given, the system of all our rules could not.
Stating the rule depends upon an unruled discursive space whose construction is the unique prerogative of the theorist (rule-descriptor).
There is a warning or therapy going on here in the way Wittgenstein addresses the problem of rule. He does not ask: (a) 'How do we locate and describe rules of...?', but (b) 'What are the troubles with proceeding as if we could locate and describe rules of...?'. Yet addressing problem (b) says something (negative) about problem (a) and (again negatively) about the concept 'rule' as such.
Is this a nihilistic step? Hardly: because it shows precisely the futility of the conventional step of unproblematically severing the 'theoretic' from the 'pre-theoretic', the primary from the secondary and so forth. When that distinction is problematised, a critique of 'the language of theory' becomes a contribution to the sort of grounded discovery-work which the charge of 'nihilism' denies. What would it actually be like for there to be 'nothing we could say'?
Work on the Moore-propositions can constitute work on the general problem of the knowability of rules and, hence, the general problem of following rules:- eventually.
461
For Moore there is a sense in which reflecting on the utterly commonplace ought to allow us to elaborate the relation between our knowledge and the world (objects of knowledge). For Russell, there is a parallel sense in which doubting the commonplace ought to reveal this same relation: the grounding of our knowledge. Both of these projects begin with the assumption already in place that grounds can be elaborated.
But Wittgenstein seems to insist that there is just a point where we routinely stop — certain moves do not enter into the language-game. And is this a boundary, a limit, even a foundation as Moore and Russell would (perhaps) have it? Does the discovery that 'we can go no further' (in this form of life) have consequences for classical nomothetic studies of language, culture, etc.?
Running up against these boundaries seems to be a definite practice we can undertake (in philosophy and sometimes also outside it). But the way Wittgenstein's poses the problem continually stresses that there is no practice 'beyond' them — that language becomes necessarily absent past a certain point. Purely and simply elaborating rules could then not be one of the language practices that anyone ever has engaged or could engage in. All erstwhile elaborations turn out to be only further repetitions of the (ruled) discursive practices routinely engaged in by the language community in which the elaborator was trained.
At the same time: limits also seem capable of being displayed.
This is the case with the imaginary situation in 461. And the question it raises is this: how often do we routinely bump up against these limits, outside philosophy? Or, when we do: how are the utterances in question to be taken? What is their use; in what language-game? The game of giving information seems almost certainly ruled out — except in the most astonishing circumstances. Cases of insanity and science fiction come to mind — instances in which there is the pretence or conviction that the world is radically different in some way. Why should philosophy want to be added to the list?
Either an utterance is nonsense or it is a legitimate move in some peculiar game.
'...how can he be certain?': and, equally, how can we be certain of a claim to elaborate a rule? For: such a rule would underpin our certainty (be a basic constituent of our certainty). What, then, would underpin the discourse in which the rule was elaborated?
((Here I am not trying to invite an infinite regress of any kind (cf. Z 693). I want to answer: 'nothing would do this' — the utterance is nonsense or else simply part of some other game. (The games lie on a plane, not on infinitely regressive levels.) The elaboration of general rules of grammar, action, etc., presupposes gameless utterances, practice-less speech, pure theory: knowledge of the world but not in the world.))
David Silverman (1975) says that such a procedure (the elaboration of the grounds of the social order) forgets its own community — its own origination in social and discursive practices. The category of forgetfulness is redundant though, in one sense, for the procedure, the pure rule description, is an impossible one to completely realise. There is always a remainder. Silverman wants theory continually to 're-member' its origination. But the remainder is always a reminder. In fact Silverman's problem seems to be that some ways of working in social science preclude actual researchers from being aware of this impossibility. And this is another possible sense of 'forgetfulness'. But how could Silverman's text (which relies on the identity of language and community) ever be perplexed by what particular human subjects were or were not aware of? This matter of 'awareness' would be a minor accompaniment to the working of language.
Sometimes in Silverman there is a definite sense that 'overcoming forgetfulness' would entail being continually prepared to state the grounds of the social origination of (a particular) theoretical discourse. And it seems that this is the very thing which the limits of language (the boundaries of community) prevent.
Silverman learns one thing from Wittgenstein: that we cannot say 'how things are' with language. He does not learn that this also means that we cannot say 'how things are with language'.
Why do we go on trying such things in social science, linguistics, cultural studies? — as if there would eventually be some sort of breakthrough and the rules (or whatever) would be laid out before us like a perfect calculus? And am I trying to put an end to these activities? I think not. I hope not. I am not even, like Silverman, trying to have them include an 'awareness' of themselves as activities along with others. Remembrance would only be a repetition of the activity already being carried on — it could not be a special private screening of 'what's happening':
I give myself an exhibition of something only in the same way as I give one to other people. I can display my good memory to someone else and also to myself. I can subject myself to an examination.' (Z 665-6)
Perhaps I want to point to some other way of working which does not assume that the grounds of language and action are utterable. Wittgenstein seems to be pointing here to a dissolution of a central problem in social science: and it is difficult to imagine a new problem to take its place. The place, itself, is odd.
462
Moore doesn't mention a fact which is known only to him because he is concerned to show that common sense consists of a series of absolutely certain propositions which are (a) known by all and (b) known to be known by all. In this way it is assumed that they may be relied upon as the basis for all sorts of other human events, language practices, communicative exchanges and so on. They may also, and incidentally, be relied upon in the production of the sorts of things which only Moore might know, in common with (say) a relatively small coterie. The difference between these kinds of fact or knowledge is identical with the difference between the statements in OC 444 and OC 445: 'The train leaves at two o'clock' and 'I have two hands'.
How close is Wittgenstein himself to this position? Perhaps quite close. He seems also to hold that there is a texture or network (web) of certainty — a bedrock of propositions — a mythology which sustains what we do and what we know in a form of life. He refers (in 461) to a 'background' (cf. Garfinkel 1967). In uttering a 'background proposition', the background for that proposition itself is missing. Once more, I rpeat, we do not therefore begin a regress. The difference between Moore and Wittgenstein is that Moore held there to be some importance for philosophy in (the uttering of) these background propositions whereas Wittgenstein seems to want to leave them as they are — or to put them 'back to work' in the everyday stream of events. Further, Wittgenstein seems to want to say that Moore collecting them specially for philosophy is nonsense — they could not be sensibly uttered; their 'sleeping' (being in the background) being essential for the sense other utterances. Nothing could act for them as they act for ('genuine') utterances. So, when we actually say 'I have two hands', this is either nonsense or it requires some extremely peculiar or esoteric 'background' in order to make it sensible. Either way: no rule gets uttered. Possibly: some other rule is (simply) followed.
I don't think Moore is correct about the propositional character of the grounds of culture or common sense. And I remain unsure as to what Wittgenstein wants to say about this matter. But this is a matter for another investigation entirely (McHoul 1981). The point is: the propositions addressed by Moore and Wittgenstein may stand for — they may even be metaphors for — whatever it is we conceive to be the grounds of social and cultural order. (A suitable translation is needed for each different research paradigm operating on its specific different terrains and bases). And it emerges here that when analysts claim to be uttering the rules of some discursive practice, they are actually following a rule of, or repeating, some (other) discursive practice. A whole domain of meta-theoretical work on language is ruled out (vis-à-vis its own self-understanding as 'representing the rules' — for example, of representation).
463
The practice of joking, for instance, might be going on when a proposition like 'That is a tree' is actually uttered — a proposition which Moore might cite a fundamental certainty, say, or use as a step in the proof of the external world. And these are all practices ranked together: joking, citing, proving. None, as it were, may claim another as its sub-province. At least: such a 'claim' would be false. All practices are nomic, but none utters its (or another's) rule(s).
((It is tempting here to think that only the uttering 'gets in the way' of the expression of rules! Or to put it another way, to think that the (for example, physical) fact may be a regularity we hold to and that the proposition - the uttering of the fact — relies on there being certain other rule-facts that we hold to. Then we should get something like: (a) '"That is a tree" is (merely) an utterance along with others and is rule-governed but that that (pointing) is a tree is, or is like, a general rule'. This sort of objection has in fact been put to me by logicians. The cat, they argue, has fur and paws, while 'the cat' does not and so on. It would be strange to have to deny such a distinction. But it is a distinction which can only be made through a certain discursive practice — namely, separating use and mention in linguistic philosophy, quoting and so on. And this practice does not seem to operate particularly well vis-à-vis the question of rule. So it does not seem to touch the question of the knowability of rules. It would come out something like: (b) 'There is some knowledge we have of rules which (on the model of knowing the cat vs. knowing "the cat") is outside or prior to the attempt to utter the name of, or proposition for, that rule'. But then: what would the rule be like? Isn't it better to say that not only can rules not be uttered, they also cannot be known? A lot of social science seems to rely on something like (b) above but then to add a clause like: '...but our discipline's meta-language can capture the rule itself, unlike the object-language in which the rule operates'. Just as (b) relies on the simpler (a), so this new clause relies on an utterance like: '"This is a tree" and that that (pointing) is a tree are different — in respect of their rules — for the people we study, but we, as analysts, can take them as either the same or different in this respect just as we please'. That is, the analyst accords himself the privilege of the option of a rule-free discourse for the special (time-out) occasions when he is revealing some rule or other.))
In the above, the relation between proposition and state-of-affairs stands in for, is a metaphor for, the relation between rule-utterance and rule. How else would it be possible to proceed? Is this how OC proceeds?
464
An utterance can be out of place either because it is routinely used elsewhere ('Hi' used as anything else but an opener) or because it is routinely left unsaid ('I know you are so-and-so'). We rely on both of these to 'get things going'. What connects them is their utter familiarity. They are overwhelmingly expectable and preferred. This absence — unlike the absence of many things — is non-trivial for our social existence. Sacks (n.d.) gives 'He didn't even say Hello' as a case in point. 'Then half-way through he mentioned that he knew me to be so-and-so...', said of a friend, might be a similar exclamation about an unusual event.
What do we do with persons who routinely do these sorts of (irregular) things in conversation? For example, how often could you get away with saying 'Good morning' out of place before being incarcerated?
In this case: have we pointed to something we rely on so fundamentally that we can say that we have a rule of conversational procedure? Is the utterance 'If using a greeting in conversation, use it at the start' a conversational rule? For instance: how would this be used? To whom? It seems that this utterance is not entirely separable from the one which started us thinking along these lines: 'I knew all along that you were so-and-so'. And one did perhaps know this all along — but could this possibly be? Likewise: when do we know about the rules for performing greetings? (Cf. PI, p59 — marginal remark (b).) So finally: is it knowing that is going on in the case of the duration of a rule across some extended event? Perhaps not but: '...it seems to me that I have known' (OC 466). And this is perhaps one way of reconstructing what happened (for myself or for, say, a conversationalist I have been listening to on a tape). Where a rule appears to be operating, it seems quite natural to say that the participants knew it. But this might just be another way of saying that it was operating and not a specific attribution of knowledge. Then: 'They knew' would mislead our investigation.
465
Here it is the fact that the sentence is uttered out of context, not the kind of sentence it is, which makes it peculiar. Why does Wittgenstein substitute 'They know nowadays that...' for 'I know that that's...'? Perhaps because even a useful sentence, one which does not declare something absolutely obvious, can be meaningless depending on where and how it is used; so that 'meaninglessness' is not some unique feature of the Moore-propositions. It is simply the case that they are likely to have a very reduced range of sensible contexts over and against, say, 'They know nowadays that there are...'. So, perfectly ordinary sorts of constraints are operating vis-à-vis the uselessness of Moore's 'background' propositions. There seems here to be a spectrum of sentences — some which we have no trouble inventing contexts for ('Good morning'), some which are a little more restricted ('You must call a doctor') and some which we should have to work hard on before a suitable setting emerged ('I know that that is a tree'). Somewhere on this continuum lies the point of not-doubting. Doubt and certainty have their 'operationalisations' in this way — this is how they emerge discursively.
If 'I know that that's a tree' is ploughed back into OC 465, we find that we shouldn't speak of someone suddenly uttering it as we might of someone uttering 'They know nowadays...' out of context. That is, we couldn't perhaps imagine this 'something else' that she was thinking of in the interim. There is even a little trouble connecting the 'tree' sentence with something as peculiar as a trance. In both of these cases we are still looking for a plausible use of the sentence.
With '...speaking without understanding what he's saying': it's essential that Wittgenstein use a rare context — a trance. For presumably this is no routine occurrence. Understanding does not work this way: like a (mental) monitoring of our own (material) language-use. Understanding an utterance involves using further (competent) utterances. It involves correct moves.
In a play by Oscar Wilde, a child asks his father 'Do you understand everything you say, father?' The father replies that he does but only when he listens very carefully. It is perhaps because understanding does not operate this way that we find the remark funny.
((Why do we think 'I can't really have meant that' when we catch ourselves saying something quite appropriate to the circumstances but not exactly the utterance we (retrospectively) wanted? Say we said 'blue' where 'green' was correct. This too is a lapse but not like the one above — with the trance. It is not a lapse of understanding or competence but a lapse into error, say. And sometimes we do monitor ourselves here. We do say: 'Hey did I just say "blue"? It ought to have been "green". How stupid.' But we do not (?) say: 'Did I just say "They know nowadays about species x"? I realise it was irrelevant. I thought this was a conversation about insects. How stupid when it's in fact about skiing.'))
(('Someone who, dreaming, says "I am dreaming", even if he speaks audibly in doing so, is no more right than if he said in his dream "It is raining", while it was in fact raining. Even if his dream were actually connected with the noise of the rain' (OC 676). And supposing someone asleep in the corner says just the right thing at just the right time in a conversation (like 'Yeh, I'll have some tea') and this is actually part of his dream? Or: a young child who can't calculate happens to say 'three' when someone asks for the sum of two and one? Here I am comparing being right about a state of affairs and making the correct or legitimate move in a language-game. They are related games — not isolated events. 'I'll have some tea' is only correct when followed by further appropriate moves. For instance, a positive response to, 'So you are awake then?' We don't just perform correctly for a while and then lapse. Competence is not like this.))
466
One of its (competence's) effects is that, on reflection, we do seem to have known something all along (cf. Chomsky).
Competence resides in the proper application of the rule — but this does not make the rule an object of knowledge. Competence concerns conduct or practice but not knowledge; or rather, competence concerns knowledge only insofar as we consider knowledge as conduct or practice. In following the rule, and in applying it correctly, it is wrong to speak of something we have known all along separately from the following.
As ever, 'rule' points to practice, to discourse. It points to the familiarity of practice, to repetition. In 'doing over and over' there is not a point at which 'the rule of doing' is (I want to say 'directly') addressed. Rather, the '...over and over' supplies the rule.
For instance, mistakes in the application of a rule do not come to light by virtue of our comparing the action we witness with something like a (write-out of a) rule. Rather: 'Practice in the use of the rule ... shows what is a mistake in its employment' (OC 29).
'Rule' does not refer to something we consult. Human conduct does not work by a process of matching; for example matching up the act and the rule for it.
Social science almost invariably gets going, however, by assuming that conduct is like this — split into act and rule, into event and knowledge (or, sometimes, into practice and theory). And, once going, it assigns itself the task of re-cementing or stitching up the gap thus created; by, for instance, 'writing out the rules'.
If the rule and the practice are inseparable, then social science's grounding and its principal task are undermined. (I say this not to remove a way of working but to find another.)
How strange that in daily life we show someone how, while in social science we are continually trying to say or write 'the how' (Psathas 1980). 'What is "learning a rule"? — This ... And what is pointed to here is something indeterminate' (OC 28). Unbestimmt: indefinite, vague, not fixed, not certain, not open to ascertainment or definition. And this is just how our rule-talk is. Why do we proceed as if there could be a science, a way of knowing these rules?
467
In philosophy of a certain kind — say in 'doubting and reflecting' — it looks as though we have 'elaborate[d] the relation of knowing' (Brand 1979, p7).
((Its relation to what? — for OC 90 is full of parody, something for which Brand seems to have amnesia. The idealism which OC 90 parodies, Brand swallows. Was there ever a Wittgenstein who was on the side of a phrase like: 'So that the fact is taken into my consciousness'?))
As if we got anywhere with these claims about knowing. It is just that there are discourses (in philosophy principally) which understand themselves as elaborations of this kind. Anywhere else they emerge as insane, as repetitious babblings. Even within philosophy they sometimes have to be excused.
468
Here we search for some 'normal circumstances' for the utterance. And those circumstances incline us towards the idea that the utterance is doing something other than relaying information. When we want to hear it informationally, we get stuck. The utterance, the practice and the circumstances must come together in a certain way, or (sometimes) in certain ways, unless this sort of 'sticking' is to occur. But are we 'consulting' a rule for the conjuncture? No: we seem to be able to see that something has gone wrong — but we should be hard pressed to give a full account of what generally happens, of what happens when things go right. (And it's not because our science isn't quite up to that yet but will be one day.)
If ... one wanted to give something like a rule here, then it would contain the expression "in normal circumstances". And we recognise normal circumstances but cannot precisely describe them. At most, we can describe a range of abnormal ones. (OC 27)
Here (468) Wittgenstein points to a peculiarity of the Moore propositions by citing one such abnormal case. What goes wrong is that an 'abnormal' conversational sequence is attempted: if we are given that Ui can only be heard as one of a range of acts Ai, Aii or Aiii, then a problem arises when an elicitation of the kind of act it was produces another utterance, Uii, which states the act to be none of Ai-Aiii, but in fact some fourth, Aiv. The simplest way through this contradiction is: the way Ui gets heard has not undergone a change. A single use cannot alter public usage. Uii is simply wrong: that is not how Ui was meant. The speaker did not know the meaning of Ui; he did not know what he was saying. There is not at present a community which uses Ui this way.
But such a definite range of uses (Ai-Aiii) is never given by an explicit rule. Nevertheless, we can definitely tell — as in this case — when something goes wrong. But that 'telling' does not entail matching our knowledge of a rule with our hearing of the utterance. Upon inspecting something going wrong, it seems (466) that 'the rule for the unusual case' was known all along. At these points in our discursive practices (when things go wrong) we could quite easily give other accounts. (Cf. standard 'forms' of wrongly argued syllogisms.) Our training inclines us to consider them 'rule breaches' and to account for them by giving the rule for the normal, 'unbreached', case. A story of this kind might simply be a case of following a further rule. And what is the rule for accounts of breakdowns as rule-breaches? - This. (Would an example be a Durkheimian analysis of anomie?)
On this version of things: what would be the 'deregulated' discourse that gave, but did not follow, a rule?
The central methodological question of the social sciences is: how are rules to be assigned to practice? The question assumes that a rule-practice relation 'governs' human action. But does it? Isn't this very relation a local discursive product? If so: we need to entirely re-work questions of discourse, production and so on; and that re-working would have to end the proliferation of (co-)relational or matching mechanisms in social science. (And I do not mean just statistical correlation.)
At this point, the discursive, as it were, is flattened. It becomes uniform and planar. There is no longer a great drop (or leap) from its level to that of the real, the economy, and so on. We can no longer proceed by seeking to match discourse, culture or speech (etc.) with something which is given pre-discursively (the real, the economic or the world, etc.).
469
With 'I wish you luck' the range of acts or applications (Ai... An) is certainly a lot wider than with 'that's a tree' (468) or 'They know nowadays...' (465). With some expressions the range is so large that they could arise almost anywhere (like 'That's how it is'). Perhaps there is a spectrum from the almost unuseable ('That's a tree') to the almost infinitely useable ('That's how it is'). But an expression's having meaning (or otherwise) on any occasion would not have anything to do with its position on this continuum.
470
If there were doubt about my name, how many other language-games would be ruled out which are now routinely played, and so routinely that things, everything, would need to be vastly reorganised before they could come to a halt?
We have the expression, 'That's so, or else my name's not...'. Here: the thing has been doubted; it is subject to debate. But the party defending it is able to display her conviction by comparison with an indubitable keystone like 'My name is...'. If this is then doubted, almost everything else falls. Surely enough, that one's name is... is hardly the sort of thing that one could establish (festellen) or rank among some list of truths. This is not how our reliance on personal names works. With this (and the rest) already in place, the games of 'establishing beyond doubt' and 'considering as true and false' can proceed. Without it, they would not get off the ground. If 'establish' (festellen) is at all appropriate in this case, it is in this way: that my name is... is always already established. Its establishment is not something I can do. There is a point at which we can no longer tinker with language. —
471
That point is Wittgenstein's 'beginning' — the 'always already'. It is here that our language and action are grounded and (although we are tempted to try) it cannot be penetrated by investigative languages for they too are identically grounded. 'Going further back' cannot be attained — or entertained. Wittgenstein, perhaps, wants a critical therapy which warns us against, and eventually prevents us from, attempting the unattainable and needlessly frustrating ourselves in the process. For the therapy, this beginning would need to be obvious and accountable. But can this very beginning itself be accounted for? Or only shown? — by beginnings.
'Going further back': seeking an account of the rule is an instance of a transgression-which-cannot-be-performed.
I have not read it, but the advertising for an anthropology text (Fox 1976) reads: 'In Encounter with Anthropology Professor Fox attacks the prevailing orthodoxy in anthropological circles that human culture is both arbitrary and learned.' What would it be to give an account of the absolute (non-arbitrary, non-learned) grounds of (a) culture? What would be the cultural location of this account? What would it be for there to be knowledge of the grounds of 'human culture' which was separate from those grounds: a text which merely related the grounds but remained ungrounded by them? What would be the grounds of this text's language? Obviously, they could not be 'human': so what?
None of the answers can be given. And yet, how easy to say that the 'prevailing orthodoxy' of a currently unpopular discipline is all that is being challenged! How easy to slip into (the) 'further back'. A couple of words will do it.
What is more surprising is that analysts convince themselves that they have succeeded in these ventures. The literature is strewn with social, linguistic and cultural grails.
Here, even stating what one wants to criticise is troublesome. Working through particular cases seems to be the only viable way of proceeding with a critique of 'attempts' to utter the grounds. And fairly obviously, no critical working-through can simply utter, let alone undermine, the grounds of the discourses it addresses. If it could, there would be no need for the critique, for the 'further back' (471) would already have been reached. Critique must (simply?) repeat the discourse it critiques; in this case, any discourse which offers explicit grounds. It repeats a discourse because it cannot be just another pseudo-attempt to recount the grounds of discourse and it repeats precisely this discourse because it cannot be some further discourse which has this one as its (pure) knowledge-object.
So how does the critical repetition differ from its object, the 'original' (non-critical) repetition? I will try to give an answer to this below, even though it remains sketchy and experimental.
Summarily, at least this has been pointed to: questioning the knowability of rules undermines a number of fundamental ways of proceeding in social science. Once the unknowability of rules is accepted, the discursive ceases to be 'accountable' (in terms of, for example, the relations of discourse and the real, language and the world, text and context etc.). The level of the discourse is 'levelled' to the point where accounting for it becomes another repetition of certain discursive practices. (This idea of 'for' assumes a meta-level which is ruled out.) If this is inevitable, critique of those social science practices whose self-understanding involves 'recounting the rules' must itself constitute discursive repetition. Briefly: 'Are rules objects of knowledge?' furnishes the further question: 'What separates critical from non-critical repetitions?' (We should not, then, expect an answer to this last question cast as a general rule.)
The question needs a fuller study, but perhaps one mode of attack would be as follows. The (original) non-critical repetition emerges from a training in which it is assumed that a discourse can represent, say, real economic relations, actual rules of grammar... and so on. There is a necessary reliance, in the training (and in the repetitions it furnishes), on the unity of addressing-discourse and addressed-object. Non-critical repetition is pseudo-unified repetition.
Critical repetition, on the other hand, works against that training. The unity of the discourse's self-understanding is reformulated as fragmentary (of necessity: for a discourse's knowledge-object cannot be given pre-discursively). The discourse appears here in terms of separate, discontinuous and contradictory repetitions. The original training itself is refused and is seen instead as a means of patching over (suturing) those fragmentations and contradictions into a pseudo-unified and discursively representable knowledge-object. This training is deleted, but replaced with another — one in which the option of transparent representation, for instance, is no longer present.
Non-critical repetition understands itself as representing a knowledge-object. Critical repetition understands itself as refusing representation of this kind and seeks, in repeating the discourse, to display such a pseudo-unity as fragmentary and contradictory. Critical repetition understands itself as repetition.
Instead of meta-theoretical utterances (supposedly) about, say, psychoanalysis, we should find a repetition of the discourse of psychoanalysis which showed its contradictions: a critique of psychoanalysis doing several incommensurate repetitions of psychoanalysis. The discontinuities of the discourse cannot be shown by representation or meta-theory (which are pseudo-unified repetitions) but by repeating the fragments.
Some of the 'assumptions' working here are:
1 That language does not operate representationally (by correspondence with the world) but by repetition (iteration);
2 That the knowable, the utterable and the repeatable are to be entirely equated with one another;
3 That the grounds of language are not just not representable but also not repeatable (knowable, utterable);
4 That non-critical training reverses the above and assumes a representational unity for language and its knowledge-object and also for language and its grounds;
5 That the discursive training can be 'unreversed' (overcome, replaced) by a training based on 1-3 above.
472
Learning what is and what is not to be investigated: a discursive training always takes some things (we could say) for granted. But those are always the grounds of the discourse itself. What may be investigated is whatever those grounds lead to; it consists of moves in those games. What may not be investigated are the grounds themselves — the principles of the investigative discourse.
In learning the discourse, we learn what is to be investigated. We do not simultaneously learn what is not; for we do not learn the principles of separating what is from what is not to be investigated. We 'take on' the principles of the discourse. So how much less likely are we to learn the moves by which those principles are doubted? It seems to me that training in a discourse does not even train us to state its discursive principles, let alone doubt them.
((Wittgenstein seems to be sustaining a massive metaphor here. The Moore propositions concerning supposedly indisputable features of the physical world are used to illustrate (or as analogues for) the whole question of what can and what cannot enter into dispute and what (on the other hand) always remains 'behind' to anchor games like disputing. If there is a Wittgenstein of the picture theory and a Wittgenstein of the 'use' slogan, perhaps there is also a third who saw that use, too, must have its grounds but not in, say, representation. Again, the doctrine that does seem to be resurrected from the Tractatus is saying and showing. In the Tractatus, a proposition rested on representational grounds. A state of affairs, that is, was held to be asserted by a proposition by virtue of having a structure identical with that of the proposition. This 'common structure' was the representational grounding of the proposition. It was not itself susceptible to being spoken of (said: aussagen) but only to being shown (zeign) in the symbolic expression. The expression could say..., but only show its own grounds. In OC this is also held to be the case, except that the proposition's representational grounds have been replaced by a doctrine which envisages those grounds as the use the proposition has in a language and in a form of life. After PI, and in OC, it's only what is to count as a proposition's grounds which has changed — those grounds remain unsayable.))
473
When we learn x as the norm, we do not also learn that x is the norm. Logically, this could not be part of x.
That x is the norm might perhaps be, then, one of the 'alterations' we undertake? But are these alterations ever this drastic? A garment may be altered but at some (which?) point in this process the alterations may be so great that it becomes another garment altogether. The equivalent 'test' in language would be whether we could continue to use the sorts of utterances that such drastic alterations would generate. Would they constitute moves in any game at all? This is the extent of the 'alteration' we can do. Is giving the grounds of the discourse (saying that x is the norm and describing x) within these limits? I think Wittgenstein wants a negative answer here and that this is why he would have trouble with the Moore propositions.
For analogous (or identical?) reasons, the rules of grammar, behaviour, economy, etc., may not be given as such. And here talk of 'rule' is (along with 'representation', etc.) one of the possible variations of the grounds of a discourse. I want to say, 'Whatever the nature of these grounds (rules, representations...), they do not enter, as such, into the practice of the discourse'. The only way the cultural sciences have survived an enormous (even final) crisis has been for them to cling fast to the myth that they are not cultural practices themselves (or that they are also something else).
474
The use, the worth of the game on some occasion: this tends to be immediately obvious. It works or it does not. And so: it may or may not be played again. Mostly, it's hard to find failures here — with entire games. But all of these 'matters of practice' or experience work on the basis of game-grounds. So they are not identical with them (cf. OC 429). Empiricism falls: as the attempt to find the grounds in empirical practice. Rationalism falls: as the attempt simply to specify the grounds (of knowledge and action) or by deriving principles of knowledge from an inspection of what's known. —
475
And the danger is always that we may turn to a kind of behaviourism in their stead. What kind? In the Investigations the question arises: 'Aren't you at bottom really saying that everything except human behaviour is a fiction?' (PI 307). And this picture does seem to have taken command here (OC 475). Human behaviour: yes, but this does not mean the shallow conception of behaviour associated with empiricist behaviourism; for behaviour is always grounded (it is never 'just any' behaviour), though not by the grounds given by a rationalist model.
((Using Ryle's (1949, p46) example: we could distinguish the chess player's move from that of a drunkard who just happens to put the piece in the right place. And this distinction does not require a 'ghostly process' as the mark of its difference. The question is: is the behaviour grounded in the rules of chess?; does the move 'articulate' with other moves in the way the rules provide for? Behaviourism simpliciter could not make such a distinction and rationalism makes the distinction too mystical a one.))
Having cleared both the empiricist and rationalist versions, Wittgenstein still wants a separation between (discursive) practice and its grounds. This separation does not repeat a similar one in the discourse(s) of empiricism/rationalism.
Even the raising of the problem of grounds seems to be, however, an 'alteration' (OC 473). —
476
Learning the practice comes first. And this may occur without the question of its grounds arising. Questions about the grounds emerge later as another practice. Wittgenstein wants perhaps to remove the 'special' emblem which is reserved for questions of grounds. This emblem seems to imply that there was no work to be done on the question of practice but that, on the contrary, this was all 'sewn up'.
The separation of practice from its grounds is a little bit like the separation of the empirical from the logical. It is even more like the separation of the empirical proposition from the rule. A proposition may act sometimes as an empirical proposition and sometimes as an expression of a rule:
The justification of the proposition 25 x 25 = 625 is, naturally, that if anyone has been trained in such-and-such a way, then under normal circumstances he gets 625 as the result of multiplying 25 by 25. But the arithmetical proposition does not assert that. It is so to speak an empirical proposition hardened into a rule. It stipulates that the rule has been followed only when that is the result of the multiplication. It is thus withdrawn from being checked by experience, but now serves as a paradigm for judging experience. (RFM 325)
Because a proposition may do this, we cannot make a general or universal separation between 'empirical-type' and 'rule-type' expressions. This is why '...it [is] impossible to hold a firm frontier between the 'logical' and the 'empirical' (Manser 1980, p19). But only in this general sense. In any particular case, what is to count as the rule and what as the empirical proposition is plain. In particular circumstances, we should be able to say how the proposition was acting. The passage Manser could have been considering is OC 96 where the 'hardening' of empirical propositions is also mentioned. But at OC 97, Wittgenstein goes on to add:
The mythology may change back into a state of flux, the river-bed of thoughts may shift. But I distinguish between the movement of the waters on the river-bed and the shift of the bed itself; though there is not a sharp division of one from the other.
In any actual case, I must make this distinction.
A mistake I have been making so far is that I have been taking problems with the knowability of rules as leading to the impossibility of giving expressions of rules. But now it seems that there is a difference between the empirical proposition and the expression of the rule. So there can be such expressions! A rule may have many expressions.
Does this mean that 'following a rule' is indefinable? No. I can surely define it in countless ways. Only definitions are no use to me in these considerations. (RFM 321)
If I now see a rule in the sequence [of numbers] that is given me — can that simply consist in, for example, my seeing an algebraic expression before me? Must it not belong to a language? (RFM 329)
My earlier mistake was this: I thought that candidate rule-expressions must in fact be something else and so concluded that 'rules can't be expressed'. My mistake attributed the wrong mistake to social science. Its actual mistake lies in treating the expression of the rule (of which there may be many) as the rule. What social science wants is a rule that will stand before it. And it has let rule-expressions (of a certain sort) stand in for what it has wanted to do with rules. This 'metaphorical' way of working has always been taken quite literally by social science.
But aren't we guided by the rule? And how can it guide us, when its expression can after all be interpreted by us both thus and otherwise? I.e., when after all various regularities correspond to it. Well, we are inclined to say that an expression of the rule guides us, i.e., we are inclined to use this metaphor. (RFM 347)
'The expression of the rule guiding us' is like the confusion between the meaning of an utterance and its mental accompaniment. In so many things, the metaphors (or stand-ins) are useful and quite endemic to our ordinary (and our scientific) discourses.
What social science seems to want is an expression of the rule which purely represents (we might say, which is) the rule. And a certain (correspondence) theory of representation can provide this. What it wants is an expression which stands clear of the route usually travelled by linguistic expressions; a rule-expression which is not in a language but which is allowed (perhaps magically) to stand out beyond language. (For example, Chomsky's re-write rules, tree diagrams and so on. The magic? Perhaps the rationalists hope this will be explicated one day by neurologists.)
Someone writes up a sequence of numbers. At length I say: 'Now I understand it; I must always...' And this is the expression of a rule. But, only within a language! (RFM 329)
The exclamation mark might gloss 'how else could it be?'
The expression of the rule also has its uses. Using the expression involves certain techniques.
And so we move well away from the dominant idea that grounds are some kind of 'deep structure'.
For the ground keeps on giving us the illusory image of a greater depth, and when we seek to reach this, we keep on finding ourselves on the old level.
The difficult thing here is not to dig down to the ground; no, it is to recognise the ground that lies before us as the ground. (RFM 33)
When our empirical investigations lead us to the expression of a rule, we tend to ask: how does this 'rule' now get used on one occasion or another? What this actually amounts to is the question: what are some other possible expressions of the rule?
Realism holds that states of affairs and also rules are objects of knowledge. This is how expressions of them act. 'Not empiricism and yet realism in philosophy, that is the hardest thing' (RFM 325). Expressions of rules and empirical expressions are doubtlessly different. Realism treats the logical and the empirical separately. Why then does it treat the relation of states of affairs to empirical propositions in the way it treats the relation of rules to rule-expressions? Realism requires not only a uniform world, but also a single way for language to work.
Expressing a rule is different from an empirical proposition. This much is sure; but it does not make the former more 'special'. We are not taken any 'deeper'. 'The old level' (RFM 333) is never left behind.
Rules as such are not expressed, if by this we mean 'contained in rule-expressions'. They (rules) are not expressions. Claiming we know them is no different from claiming unmediated knowledge of other aspects of 'what is, of being.
Our relation to rules is not one of knowing, but one of following.
Previously I have been dubious about the exact significance of Wittgenstein's remarks on certainty vis-à-vis the Moore propositions for an investigation of rule. But now the matter seems to be a little more solid.
'But at every step I know absolutely what I have to do; what the rule demands of me.' The rule, as I conceive it. I don't reason. The picture of the rule makes it clear how the picture of the series is to be continued.'But I know at every step what I have to do. I see it quite clear before me. It may be boring, but there is no doubt what I have to do.'
Whence this certainty? But why do I ask that question? Is it not enough that this certainty exists? What for should I look for a source of it? (RFM 350)
There is quite simply something most peculiar in asking for a description of the rule. When it's done we certainly make mistakes about the product. (For example, the fetish surrounding E = MC2 or the transformation rules in Chomsky, as representations of how the world or language are.) This peculiarity has to do with our imagining that we are somehow going deeper into the matter by giving a rule-description — as if the fundamental could be penetrated; whereas it is its being a fundament which even allows us to give these descriptions:
Following according to the rule is FUNDAMENTAL to our language-game. It characterises what we call description. (RFM 330)
(The mistake is like that of someone under the compulsion of the Tractatus view of language: as if, when we went deeper and deeper into language, we could cross the barrier between names and simple objects and so directly grasp a suddenly languageless world!) Sometimes it looks or perhaps feels as if the spade has gone deeper upon the very stroke at which it turns and, so, can go no deeper. (We feel we are pushing the knife in further, but it turns out that the blade was merely bending under the pressure.)
OC 476: we can, and do, engage in talk about, in questions of, and in expressions of the rules for, and the grounds underpinning, our ordinary (linguistic) activities.
That is, we can (in philosophy but also elsewhere) ask ourselves a question corresponding to 'Is there such a thing as a unicorn?' only with respect to some quite ordinary object. In doing so: we might end up with an expression about our ability to distinguish separate objects, say, rather than working with an unbroken continuum. And here we think, perhaps, we have hit upon something held fundamentally by everyone (in the culture). We might talk of this as a rule-expression and we might begin to list rules of object-separability, object-constancy and so on (Mehan and Wood 1975, pp66-70 and passim).
In a certain sense (only), both Moore and the ethnomethodologists are working this way.
When Wittgenstein raises the question of 'how to set about satisfying oneself of the existence of unicorns', he is pointing to the ordinariness of the sorts of 'investigations' we could do here — or in the case of the social sciences. That is: there are also techniques and practices of this sort of investigation (for example, into rules and grounds). How do we learn this method?
((Here I am emphatically not trying to rejuvenate the old chestnut about the infinite regression of investigations of rule. Rather, I am trying to show that the expression of the rule is first and foremost an expression; an element within particular discourses and subject to their rules. It is not merely the expression of such a rule. And it is certainly not identical with the rule. Discursive rules have to do with techniques; practices which we have learnt — via the application of techniques. I do not think it is possible to go 'deeper'. So how could I be invoking a regress where, as the custom has it, we are supposed to go deeper and deeper all the time?))
477
What is the search for rules all about? (A kind of will to thwart the possible?) It is as if we were looking for something absolutely certain which we could say, without fear of contradiction, is at the root of all language and action, and so on. (And even if the search is necessarily fruitless, we should still have to account for the persistence, historically, of such a 'will'.) But why have we been so tempted as to think not only that this is how language and action work (namely, by having grounds which can be specified) but also that these grounds should concern a knowledge of objects? (Is the will a will to ultimate control — control of the very roots of being?)
((Perhaps our idea that statements about the existence of objects are fundamental stems from their grammatical simplicity; that is, from the extension of a certain practice.))
478
Again: sometimes we say that it is our knowledge of the existence of objects which is fundamental and sometimes we say it is our believing this. Against this, I want to say that a certain practice gets going — there is something we do. And after a while there's something else we do which is just another activity and also a questioning about what it was we were doing with that first practice. Knowing X and believing Z...: these could both be practices in the first sense. So why is it that we have tended to pick on these as answers to the 'secondary' question? —
479
Propositions like Moore's are answers of this kind; 'I know that there exists a living human body...' etc. If I have to have known this during the whole period since I learned the term 'body' — then it is very peculiar that I should have so few occasions for making the proposition do any work. It is peculiar that 'I know that there exists...' is not a move in many of the games which we called 'first practices' above (notes on OC 478). Perhaps this sort of knowledge not only tends to come later — say when we perform certain (religious or) philosophical practices; the knowledge that there are physical objects may not come at all. Similarly: 'I am a human being' — this can be true of me, even though I don't know it (yet).
We sometimes want to say: the young child could not get by in the world without knowledge that there are physical objects. But is this true? Couldn't just about anybody get by without it? What would a specific case of 'not knowing that there are physical objects' look like? Without that sheer difference the point has no significance. I am inclined to think of Molière's Monsieur Jourdain who spoke prose without knowing it.
480
Here the quibble is perhaps over a certain expression: 'that a tree exists' ('Proof of an External World', (Moore 1959)). Is it that the child can be said to know this expression when the practice in question is learning the use of the word 'tree' through another expression: 'Lovely tree!'? Why should the inference be drawn that because one expression is uttered (to the child) the other is known (by the child)? On what grounds, I would like to say, is that inference drawn?
We appear to be able to learn, and later to use, 'tree' without the grounds of these practices consisting in the knowledge of objects or the knowledge of further expressions concerning those objects.
And this is only one type of ground-expression — the type using knowledge of objects.
((Only in the translation does 'that a tree exists' appear in quotation marks. This makes a difference.))
481
There is a certain 'distance' from which we normally encounter our own ordinary expressions — a distance provided, metaphorically, by the use that we make of them. But our mistake (in linguistics, say) is to think that things can somehow be made clearer by getting 'closer' to them. (This compares with RFM 333: 'the illusory image of greater depth'.) The case is actually like this: we have something in focus (say with a telescope or camera lens) — if we move the focus back it gets blurred. We rightly feel we are at too great a distance from the object. But we are equally off focus if we move the lens closer. Either side of a small margin and we lose sight of the usual workings of our expressions.
In philosophy, scepticism may be one product of overfocussing in this way.
Parallel problems in the social sciences are harder to categorise and much more widespread. One way of categorising them would be to refer to 'inspectionism' — a belief that one only needs to inspect the right materials in the correct manner in order to see social, linguistic, economic, etc., fundamentals, of either a micro- or macroscopic kind.
With our ordinary expressions and actions, we can get some peculiar results in this way. How much more so with 'I know that that's a tree'?; where the fact that it is uttered in the first place is, itself, peculiar and produces an overfocus on something which routinely could just be let pass.
482
Overfocussing on 'know' can lead to metaphysics of various sorts. Why is this? (See notes on OC 477.) 'The pursuit of the incorrigible', 'a ... hankering for something absolutely certain' (Austin 1962, p104).
What is not clear is why 'the incorrigible' and the 'absolutely certain' must be conceived in terms of knowledge, especially if knowledge is viewed as something indwelling, something like thinking.
Austin (1962, pp104-5): even Ayer and the empiricists are interested in things which 'go on in us' (seeing, perceiving and, ultimately, knowing).
For Wittgenstein: matters of knowledge are characteristically open to doubt: so what is involved in saying that one knows that that's a tree? (Cf. the doctrine on negation in LC 22: 'A proposition therefore is any expression which can be significantly negated'.)
How could the foundations of a (public) practice consist of something which went on in an individual — even if it went on within every individual?
Here OC leads us to the conclusion that human activity could never be satisfactorily explored psychologically; that there is nothing mental at the heart of material practice (cf. Coulter 1979a). (And Heidegger also seems to have believed this.) This does not mean that there is currently a sociology which is up to the task. Everything indicates otherwise.
483
Note that, in all these cases, a situation has been contrived such that there can be a doubt: because of partial blindness; because it is perfectly feasible that a person should not be home at a particular time; because there can be mistakes about the dating of parts of buildings and so on.
Knowing is like any other practice: sometimes it operates and sometimes it does not. This is true generally and also for particular cases of knowing x and y. It is clearly not as if it were like a background tone by which all other notes could be measured; a sort of constant and fundamental companion of all our language, thought and action. But this is how we think of it when we think of it as the medium in which rules are couched.
In anthropology: 'a corpus of knowledge' is often ascribed to cultures and collectivities as the basis of 'observed behaviour'. On the one hand, the empirical activity; on the other, the knowledge on which it is founded. This seems to be quite mistaken. (And not simply because there is a confusion between theoretical model and empirical domain.)
((For example, what would it be for an Azande to doubt not the outcome of the oracle's prediction but, more fundamentally, that the oracle can predict at all? What would it be for a modern European to engage in doubt over whether or not the Earth exists?))
Against the model in which knowledge is considered to be the medium of the rule, too, then: rules are not objects of (in) knowledge, just as on another model they are not (contained in) rule-expressions.
484
Knowing and saying how we know: in most (genuine) cases, there is no problem. Where our knowing something comes into question on some quite ordinary occasion, we not only affirm the fact that we know but we tend to say, additionally, how this comes about. There is a doubt, perhaps even a challenge, here.
In the social sciences we often work on the same commonsense pattern of action. Having made the mistake of attributing the fundamentals of social action to the realm of knowledge ('knowing the rule', etc.) we then ask how we know — as if there had been a genuine doubt.
((Hence: 'making the routine problematic', 'systematic doubt', 'the epoché', etc.))
Knowledge (categorisation) -> 'Doubt' (method) -> How known? (research question)....
The mistake is in the first move; the rest follows as a sort of reflex (a tradition).
Epistemology and ontology: it is not just one of these which is troublesome (so that the other gets preferred), but rather the pair as a whole and also the division it represents. Knowledge and the real: it is the tendency towards the construction of dependencies between them which is troublesome and not the direction of any particular dependency.
485
One possible confusion is this: with 'I know...' we may think there is no doubt involved because the person who says it may be asserting as much; that is, that for her there is no question of doubt. But we forget that her very assertion emerges in a situation where there is the possibility of doubt.
There is another trace of subjective individualism in this: we look at what might be going on for (inside?) the one speaker, but not at what's taking place in the situation (context, form of life, etc.).
So: with 'I know' there is a doubt even if it is not the speaker's doubt; with 'I believe' there is a doubt and it may well be the speaker's (where belief is not, say, the theist's — fully confident — belief).
((In children's arguments we find affirmations like 'It is so'. And here there is a doubt but not the speaker's. For many cases 'X is the case' and 'I know X is the case' can be freely exchanged for one another. Thus from Coulter (1979b, p180):
A: He could not have committed suicide.B(to C): See what I mean? He still believes it was an accident.
A: I don't believe anything of the sort; it was an accident.))
Knowing and believing can be taken as quite routine practices, involved with certain discursive techniques such as the ascription and avowal of beliefs, the contestation of knowledge and fact and so on. What could we say about these in general?
((Perhaps that the family of practices with 'belief' tends to be weaker with respect to doubt and certainty. It also seems to be much more definitely associated (for its doubt or certainty) with the commitments of the speaker. When we appear to be quite happy to let something pass as someone's belief (whatever it is), we shall nevertheless take it that he is trying to give it a place in a more public arena if he later tries to claim it as knowledge. This is why the difference may be crucial in court cases.))
Someone going through a list of propositions and checking them as known or (only) believed: should we say that he is 'checking on the certainty' (Sicherheit) of these? Perhaps.
Something else that this kind of work with (material) belief-practices would block would be essentialist moves towards a characterisation of all knowing or all believing. One example we have already mentioned is Husserl's concept of doxa. Apparently, for Husserl and phenomenology, our relation with certainty is, strangely, one of belief. And more strangely still, it can be suspended! There is a massive manipulation here of the usual uses of 'belief'.
486-490
Distinguishing between games of knowing and believing (something): some fundamentals of our language-practices do get described in certain ways. We can speak and write about our names being such-and-such; we can say that the language we're using has specific features; we can mention the existence of physical objects and the rest. These can additionally enter into the games with 'belief' and 'know'. And this can then be confusing — as are certain other peculiar collocations of language-games.
But this is what we have with 'I know...' and even 'I believe my name is...'. For our relation to these things is neither one of knowing or one of merely believing.
((Test: knowing implies the possibility of doubt. (Merely) believing strongly suggests it. Can x be doubted — ever?))
Rather, these things (the existence of physical objects and so on) must be in place, I want to say, even before games like 'knowing' and 'believing' can take place (OC 121). They are practices along with the rest. Here I agree with Coulter. If there were a menu of items available for semiotic or anthropological analysis, knowing and believing would be on it. By this I mean that we shouldn't want to locate either as the 'fundamental medium' in which rules or grounds exist. Rather, they too will have their rules and grounds, along with conversing, gossiping, dancing, promising, marrying and so on.
Being practices, knowing and believing are neither the medium of grounds nor are they, in their turn, registered in any mental medium.
There is nothing fundamental about knowing, believing or the mental. The philosophers, the theologians and the psychologists are on the wrong track, if it's the grounds of human action that we're after.
This does not mean (as Coulter may — I think wrongly — be read as saying) that there is only the ascription and avowal of knowledge and belief and not other things that go under the titles of 'knowing' and 'believing'; or that no 'inner' realm of the mental actually exists. On the contrary, there is nothing wrong with the idea of a mental life and we may even hold there to be one that can be 'explored'. But there are no bogies here — no ghosts in machines. 'The mental' has its form of existence but it does not operate this way — hidden, as it were. What OC points to instead is that, whatever the mind may be, it will tell us absolutely nothing about the grounds of practical actions.
Garfinkel (1963) is wrong in the sense that there is plenty to be found 'in the mind' ('under the skull' in his metaphor) — a whole baggage of accompaniments. The mistake is to confuse them with something fundamental, something which 'stands fast' in all human practice. —
491-492
So, we can get unduly caught up in the contrast between (practices of) knowing and believing — as if there were circumstances where one or the other had to be the case. What we don't see is that it might be neither and that we are mistaking the comparative strengths of knowing and weaknesses of believing for a much more fundamental distinction which can perhaps be glossed as 'being certain vs. surmising' (OC 491).
The much more fundamental distinction is (practically) between two worlds, or perhaps 'realities'. One we are so familiar with that things simply can't be otherwise; the other we are asked to imagine by persons who should like to fill out that 'otherwise' in extraordinary ways (Russell, Castaneda...). As if the defeat of something on which immunity from doubt had been conferred could be treated like a (mere) change of beliefs.
What is this verb: 'to seem to be immune from doubt'?
This seeming might only be sustained in, say, (very good) science fiction stories.
'Would I react as I do when a belief has proved to be false?' Certainly not; but what happens when an anthropologist claims to suddenly 'see' what it is for a particular group of people to hold that the Earth is flat? Can she do this? If she can: is it just one of her beliefs that has fallen? — or must she now reconsider all her judgments? In an extreme case like Schneebaum's Keep the River on Your Right (1972) where a New York artist becomes a stone-age tribesman for some years, we must surely say that it was more than his beliefs which proved to have alternatives. Or could someone just 'become' a cannibal and sustain that life-practice? So in these descriptions, are the grounds laid bare? Perhaps: but why should we accord particular privilege to these rule-accounts? Often we want to do so because we think we 'see' cultural rules by a comparative method rather than by 'direct' rule-descriptions. (Anthropology as a sort of litmus.) But the differences on which the comparisons are based, we have to remember, are the products of quite readily available discursive practices. One such practice involves the notion that the anthropologist 'sees' how the Earth can be flat by a sort of 'mental click', and that the 'click' can be represented textually so as to repeat that effect 'in' the reader. And now we are confusing rules with mental states and textual practices with representations.
It is not, as it were, the 'content' of what stands fast for us that is important. (Anthropology, science fiction and Castaneda can play with these notions endlessly according to their various house-rules.) What is important is that there is immunity from doubt over particular matters and that problems arise from confusing them with the rule-descriptions produced by the very discursive practices which they underpin. The cross-cultural material is beside the point here. To do it algebraically: culture I has rules a, b and c...; culture II has rules x, y and z.... Rule a underpins practices ai — an; rule b, practices bi — bn and so forth. In this case, a rule-description of rule a, in discursive practice, say, yi, is a product of rule y and not a representation of rule a.
In speaking of a, it speaks from y.
'Speaking from' precedes (provides the conditions for) 'speaking of'. 'Speaking from' anchors our practice as community practice; it anchors what Garfinkel would call 'shared understandings'. 'Speaking from' approximates to Garfinkel's 'how something was said' and 'speaking of' his 'what was spoken of' (Garfinkel 1967, pp28-30). —
493-496
What is it to 'speak from' (for this seems to bring us at least into the general area of rules)? Three strategies become available here: to see 'speaking from' as the recognition of authorities (OC 493); to see it as a matter of holding to proposition-like rule-descriptions (OC 494); to see it as (like) 'the way the game has always been played' — as a matter of custom and convention (OC 495-6).
All three might count and all are metaphorical. Attempting to say 'what sort of a thing' we are treating here (a more generic activity than providing rule-descriptions) is inevitably metaphorical, as is the use of the terms 'rule' and 'ground'.
((The proposition is not an empirical proposition. 'It has rather the character of a rule'.))
When we refer to rules (of social action, say), we might be unnecessarily caught up in the metaphor — taking it literally — in the way that some peoples of the South Pacific hold fast to a geocentric view of the universe for the practical purposes of navigating.
The irony of social science: its membership of a world where (something like) rules are inevitable and also unobtainable. Result: a number of disciplines with a 'missing' central phenomenon. (And that is something we might well welcome.)
The propositions of social science are meaningless because, to take them as representations rules, we should have to imagine doubting things which, if doubted, would signify the end of all our acting and judging. We should have to remove the very thing which we would describe.
((Schutz's solution is a solution by fiat. He says, in effect: go ahead and do just that.))
And so the propositions would have the ground removed from under them — they could mean nothing. The comparison with the propositions of logic according to the Tractatus is, once more, quite clear.
((Compare rules with the form of the proposition.))
For the propositions of social science not to be meaningless, they would need to rest on some 'rule' which over-ruled propositions like that in OC 494, 'I cannot doubt this proposition without giving up all judgment' (in the matter of doubt-immune propositions). And could we imagine such over-ruling? -
497
Doubting comes to an end somewhere. The mistake is in attempting to explore beyond that point or even to locate it discursively. But 'giving the rule' is such an investigation. This is why Wittgenstein's work on certain philosophers' demands that one doubt the indubitable (cf. the law of identity) is relevant to an investigation of 'rule' in social science.
The impossibilities, one should like to say, are immensely similar. (Another research metaphor?)
One holds out against a certain temptation in both cases ('not being shaken', 'not being wrong'). And, in doing so, something 'definitional' is upheld — the objects of description (in social science) or of doubt (in scepticism) — whose absence would need to be presupposed before the very practices of describing (this) and doubting (this) could get off the ground.
In both cases it means resisting what is becoming (in philosophy and social science) a regular step. It means stopping sooner than is becoming usual rather than continuing beyond that point.
The appeals to 'memory deception', 'being misled', 'failing to satisfy oneself' and the rest are appeals to (empirical) practices which don't 'translate' into the realm of rules (for they are ruled practices themselves). So it's not a question of accurate vs. inaccurate rule-descriptions. It's not as if one were merely prevented from precision here; as if one day we should eventually discover precise enough tools for the 'correct' description of rules. I am not accusing social science rule-descriptions of being the wrong ones.
((An analogy would be right vs. wrong pictures of unicorns which, depending on your view, can either be infinitely pictured or not pictured at all, though for different reasons.))
Rather: there is something more problematic about the whole business of giving rule-descriptions than social science has acknowledged. For it now seems that the attempt to do so (and also to hold them to be representations of rules) presupposes, we should say, an analytic place-of-commentary from which rules are absent.
That is: social science would speak of rules, without speaking from rules (see notes on 491-492 above).
Language about (= representing) grounds would be groundless language. So: either we have to say that what looks like language on the topic of grounds in social science is not that; or we have to say that the language is groundless. The second possibility could hardly be entertained. So what is that language?
We might adhere to a picture in which social scientific language generates its own cultural objects (the ego, the subconscious, anomie, deep structure, the signifier...) which do not precede their analytic discourse — which are in fact products of those discourses.
((Is this anti-representationalist position an idealist position? Insofar as it resists a division between 'the real' and 'discourses of the real': yes. Actually, the discourse/reality duality is dismantled here, whereas a (linguistic) idealism would pose the existence of transcendental objects formed out of thought or talk. For me: social and cultural objects could never have that transcendental character. And in another sense, all that is being done is to re-call certain discursive objects to the labour of their production — to their constitutive practices. Innocuously then: if the (social) world is (created out of) human practice, no exemption need be granted to theoretic objects — even where they would 'address' (or represent) those very processes of production. The assumption that they can represent in this way is identical with the assumption that social scientific language can 'match' social objects.))
It may be that we actually have to abandon any empirical investigation of discursive production in general and confine ourselves to showing what is done on specific occasions — but as a critique of general rule-descriptions. (Critique replaces second-order description?) And where would such a critique speak from? The discourses themselves.
498
Bedrock (Fundamente): to raise doubts here contravenes the regularities of a common technique. To repudiate or castigate the contravention is, then, a familially related technique that can arise. But to claim knowledge of the technique is not so related. It, itself, contravenes — it is incorrect (unrichtig).
It is not possible for me to see contravention at a single glance simply because 'I know' the technique or the rule. My relation to the rule (and hence to empirical cases of its contravention) is not one of knowledge.
Ethnomethodology — Garfinkel's breach studies, for example — has assumed that raising troubles at bedrock will allow us access to an 'everyday corpus of knowledge'. The position in Wittgenstein appears to run counter to this picture. Though by no means unique in this respect, ethnomethodology hangs on a presuppositional step which is — at the very least — open to investigation in its own right. That is, the methodic or rational sensitivity which members show when faced with breaches, contraventions, troubles, doubts etc., may not, in fact, be translated into the proposition 'members know the rule(s)' but only into (say) 'members follow the rule(s)'.
More radically than in ethnomethodology, practice and technique precede knowledge (or knowledge-effects).
If the rule generates (for instance) knowledge-effects, how can the rule then be a knowledge-effect? This does not deny that rule-expressions are of this kind.
Specific techniques of representation (for example, bits of language) get continually reproduced. Their sedimentation is institutional and historical. That institutional reproduction determines certain knowledge-effects. Then the mistake arises when we try to take those knowledge-effects and show how they are grounded by, for instance, a specific relation to transcendental objects; when this is not how they developed at all. If this is the picture we use, then the only conclusion that could be drawn would be that knowledge is not 'grounded' (in this sense) at all.
'Bedrock' always tempts us to think of some sort of concrete bond between our knowledge(-effects) and a given, independent ontological ground. But this picture itself only emerges in certain techniques.
Epistemology (considered as the search for relations between 'knowledge' and 'the real') is one discourse among others — and one which can lapse.
Saying that we know the rule, then, is a symptom of this mistake we make in trying to 'trace back' epistemic effects to their grounds. The grammar of 'to know' is taken wrongly, and so is that of 'rule'.
If we can ever say that rules ground our knowledge it is not because they do a job of matching (anything to anything).
'Extant and reproducible techniques of representation produce knowledge-effects': this is a uni-directional process. Even when it looks like we have reversed the direction and traced 'knowledge' back to its source, we are actually re-invoking the former process.
499
'"This is a hand" — and so we can proceed to something more general'. It looks as if we are taking steps which bring us closer to what are called 'grounds'. In fact we are no closer than we are with 'This is a hand' on its own. For: we can't go 'back' from here to what grounds the proposition. What can we do? Only repeat a certain technique.
500-501
How can social science get itself off the ground by proposing that 'knowing the rule' can be examined when almost all expressions of the form 'I know...', and some rule-expressions, have almost no application? How can this take place without a hair being turned?
'It looks like something we should not ordinarily (be able to) say'. And the response to this, on the part of social science, is often given in terms of a science vs. ideology distinction, or one of its variants. I want to say, then, that a specific sort of discursive technique gets controversies like this (science/ideology) going. They emerge as forms of defence.
If we return to why the defence is (per)formed in the first place, we can see that social science might well resist the step of contravening the logical grammar of 'know' and 'rule'. In these notes I want to find out how social science would look in such an instance.
One thing, at least, is sure here: that social science practice could in no way be taken to be beyond or outside social practice. The meta-linguistic step could not be taken in any form.
'Logic cannot be described'. Perhaps, then, social science would have to forgo all forms of (what Schutz calls) 'second order' descriptions. How do we arrive at what traditional methodologists would see as this very bleak (or blank) picture of their routine ways? By 'look[ing] at the practice of language'.
The practice of language as it is routinely undertaken can certainly tell us something about philosophical and epistemological peculiarities (for example, about claims to know the rule). But this does not mean we can also describe the (principles behind) practice of language.
Ethnomethodology begs us to (dis)solve social science problems by turning to 'naturally occurring' language practice. The mistake is then to treat the new turn as identical with a second order description of language practice. That would only return (some of) the social science problems which the step was supposed to obviate. In this respect, ethnomethodology goes 'too far'.
So what could 'look[ing] at the practice of language' mean if not the (re)formation of a kind of second order analytic description? This is the hard thing to answer. For everything points to ways of giving up social science analysis but not to the means of its transformation.
'Look at' (ansehen) might be a perfectly ordinary activity and nothing to do with analytic description at all.
((And aren't there, after all, seemingly innocent or pastoral — 'pre-scientific' — techniques in Wittgenstein? As if the problems of philosophy would vanish once we could make do with pre-philosophy, with 'ordinariness'; a sort of collective natural rationality which can only be gone through but never described. Practice? So then a discourse which described 'the practice of language' (like ethnomethodology) could clear up our troubles?))
502-504
'I know...' is always the effect of certain circumstances or conditions. It is not at the centre of things. Although we may want to think of it as the grounds what we do and say, the hard thing is to see 'I know...' as just another effect.
And when I write 'certain circumstances', this does not have to do with any interior inspection I might make (OC 502); nor does it (always) have to do with any 'objective' inspection I might make (OC 503). The 'circumstances' have to do with evidence of a different sort.
'Evidence' (Zeugnis, Evidenz) here has to do, more strictly, with testimony — with the discourse of 'all the others' (OC 503), of 'other people' (OC 502).
((It is a problem in its own right that we think of 'evidence' not as a formation of discourse but as a transparent representation of the real. When 'all the evidence speaks...', this is a socio-cultural event. When someone goes 'to find the facts', we forget that 'facts' are elements of a language.))
With 'knowing the pain', what can the others say? But then this doesn't mean (as we tend to think) that only I can know it. It's simply that 'knowing' isn't actually produced as an effect in these cases — and that's why it means nothing to say 'I know...' (504).
Knowing is not private, yet it is not public-because-reality-is-like-that. It is simply public. And that is the sense of 'evidence' here. The real does not transcend this discursive plane. —
505
So if anyone should think that this is the concept of 'nature' as used in the Tractatus, they couldn't be more wrong.
((So is this a form of idealism which says that 'nature's' own evidences are public discursive techniques? No: for no one is saying that 'mind' produces anything at all specifically; in fact that is specifically denied. Mind is one site among others where discursive techniques may have their effects. And those techniques are anything but private.))
It is better to say that one knows by favour of nature — for this specifically counteracts the idealism of saying that nature is a function of what one knows - except that our concept of the natural then has to be worked on. So much of our training points us toward an equally untenable realism once idealism has lapsed. The point is to supersede the dichotomy itself.
506-509
Against Russell — not everything can be doubted at once. The expression of the doubt is itself a technique of the language and I rely upon it in order to get started on my doubting. Without it, even my doubts couldn't get underway. The practice of doubting would not make sense; it would not be a technique in this community.
Asking 'what can I rely on?' (OC 508) always already performs the technique it asks about. (Cf. 'I should like to inquire about the use of the verb "to inquire"'.)
Try not relying — that is the hard thing. But still this does not mean we have access (outside those techniques which we rely on) to the 'how' of the relying getting done.
((Again, Psathas (1980) begs us to move from 'knowing how' to 'knowing the how', as if this involved something like a change of attitude.))
'Trusting' (509), 'relying' (508): all the time our description tries to cover a meta-technique. Social science discourses have been founded on the strength of the candidates they have put forward here. And some people then want to know why Wittgenstein is not specific in these instances. 'Trusting something': how vague! But that anything at all can be said here is close to miraculous, for it is very close to being self-annulling. Wittgenstein is trying to show us how we might stop doing something. 'Something? What?' Would 'second-order analytic description' be a possible answer — given that it seems to be just another repetition of the relying or trusting that it would (ideally) describe? OC 509 is also a case in point. And to take 'trust something' any further would only compound its trouble.
Here the vague concept is not only exactly what we need — it is all we can use (if we are trying to say that the meta-discursive step can't be taken).
510
'But from his utterance (Äusserung) "I know..." it does not follow that he knows it' (OC 13).
'Or again "I believe..." is an "expression" (Äusserung), but not "I know..."' (OC 180).
'If I say "Of course I know that that's a towel" I am making an utterance (Äusserung)' (OC 510).
These three uses of Äusserung in OC seem not to be consistent. But what they all point to is the mistake (in philosophy and elsewhere) of looking for something behind the uttering or the expressing, the (almost physical) act of saying 'I know...'.
((The peculiarity of OC 180 arises because the putative 'something' is the subjective grounding (subjektive Wahrheit) of claims to knowledge and the model of expressing (externalising) a mental condition or state is the one that is being treated. See OC 179. Äusser = outer, external.))
The immediacy of 'I know...' stems from its place in a course of events, in an institution, and not from its emergence from (commonly held) substantive grounds or from 'private' mental grounds. 'I know...' has no necessary privilege (vis-à-vis the connection of the proposition following it with the real) over and above other sorts of 'prefaces'. If it seems markedly stronger than 'I guess...' or 'I wonder...', it is because the situation calls for(th) a strong preface. If anything might count as a ground, it is the institutional embedding of claims to knowledge. But that's precisely the sort of thing that philosophy has not wanted to think of as a ground. Somehow, for it, the institutional is but an effect of conjunctions of the supposedly more firm domains of the mental and the material (characteristic candidates for bedrock in philosophy).
What Wittgenstein considers in OC is that those candidate grounds might actually be seen as, themselves, effects of discursive institutions.
But that does not mean that speaking subjects are mindful of the underpinnings of their utterances as such underpinnings. 'I have no thought of verification.... I don't think of past or future'. Because: then the mental act might just as easily stand in for the discursive institution. To think this was the space for the utterance would then be identical with this actually being its space.
An utterance's space is not open to doubt. It is fundamental.
One of its effects is that we might make a separation between the knowing or speaking subject and the known or spoken-of object. This is one way in which a reconstruction of knowledge claims might be made. To repeat, it is only one way. By contrast we tend to think that all matters concerning knowledge are the province of epistemology — when in fact that discourse traditionally deals with knowledge only as a subject-object relation; so that 'different' epistemologies differ only in respect of what sort of relation they prefer. A non-epistemological theory of knowledge would want to remove subject-object relations from its centre. Instead: knowledge could be treated as practice, technique and institution; as knowledge-effect; as a concrete form of repetition.
511
Any sureness is an effect of the fact that this is what is always done unquestionably; that this is a repetition of specifically this....
Sureness emerges from a training (in the use of objects, words, and thoughts...).
And the mistake is to look for the grounds among the relations of the effects of the grounds. From this way of proceeding we can always get concepts which are more or less convincing. But because the philosopher can rest easy for a while (with empiricism, rationalism, materialism, idealism and the rest as theories of knowledge and its grounds) does this mean that that (common) way of proceeding is necessary?
512-519
In order to sustain his argument over his certainty about fundamental matters (that the Earth is round, that I know my name...), Moore implicitly relies on the possibility of our knowing otherwise. He relies, that is, on the possibility of 'something really unheard-of' occurring so that he might, as it were, inspect the state of knowledge of someone experiencing this. (Castaneda's books are imaginings of the same order).
Wittgenstein's is not an argument against these counter-fundamentals happening. It is an argument against Moore's (possible) conclusions from them (or their regular non-occurrence).
To give up some things is to give up everything: because in order to give them up the whole system of which they form a part must be given up.
And that system is not just mine. It does not have its roots at an end-point called 'my state of knowledge'. It is not giving up a 'state of knowledge' (in the individual) which is the difficult thing for Wittgenstein here.
The trouble with this and other philosophical examples is that when philosophy wants to find out about thought, it inspects what I might be thinking or what's going on in, or as, my thinking. When it wants to find out about knowledge, it inspects what I might know or what takes place 'as' I know something. When there's a doubt, it imagines what it is for someone to doubt something. Either this, or there is a process of ratiocination which supposedly leads to knowledge, thought... etc., as essences. (From time to time these 'methods' are combined.) Moore's preference is for the first method: an inspection of the individual's 'state of knowledge'.
And for Wittgenstein, this is the source of his troubles and their solutions. What Moore claims about knowledge is the effect of a certain (empiricist) discourse.
((Note how Moore takes great pains, in the 'Defence', to give his exact position with respect to classical empiricism.))
For Moore, the troubles that might occur were houses suddenly to turn to steam (and so on) would be troubles with, so to speak, the individual's 'cognitive interior'. The lapse of knowledge and the raising of doubts are events which go on inside someone and it is on this scene that Moore documents such 'events'.
Could changes in 'knowledge' ever be the result of epiphanies? No more could knowledge be grounded in mental revelation.
Giving up, or else making, fundamental judgments, neither could be just 'for me' — and inspecting what goes on in me at such 'moments' will be irrelevant to the case. At best it will be ancillary, an accompaniment. OC 518 is crucial here.
'...a language-game ... consists in the recurrent procedures of the game in time' (OC 519). What could be more clear a statement of Wittgenstein's adherence to the notion that it is the repetition of a technique which 'lies behind' language and action? So: '...it seems impossible to say in any individual case that such-and-such must be beyond doubt if there is to be a language-game' (OC 519).
And Moore's mistake is to conceive the game with 'I know...' (the game of knowledge-claiming) as an individual case or cases; whereas the existence of the whole enterprise depends on its alterity (its being-for-others) and its repeatability (its regularity).
In the individual case, all sorts of peculiarities can arise, or can be imagined. But the regularities they rely upon cannot be open to the same treatment.
((Ethnomethodology's reading of Wittgenstein as a fellow radical situationalist is put into some jeopardy here.))
520-521
What can be known is through-and-through a public matter. It is not 'ascertained' by checking on who avows a matter of knowledge. What gives knowledge its guarantee is not something going on in someone (Moore) and, indeed, avowals are not in fact representations of such inner states, processes, etc. 'I know...' cannot cancel 'It cannot be known'.
But often this is precisely the step taken in social science; as it were a dialectic of self-evidence. It operates in the defence of all meta-linguistic claims when faced with (what has come to be known as) 'the problem of relativism'. The discourse which claims (avows) itself to be an incarnation (complete adequation) of the real claims to speak for (of?) all other discourse. This is undermined by showing that the grounds or rules of discursive operation (which it claims to represent) are not knowable but rather (for instance) that they provide the very possibility of knowledge. And then what is claimed — for example, with the rules for taking turns at conversation? Simply this: 'Here they are! We do know them!', followed by a reading-off of some rule-expressions.
This is identical with 'I know that the earth existed before my birth'. Just as Wittgenstein is not disputing the matter or the earth's existence, so I am not trying to make any sort of point about whether or not it's those (as against some other) rules which operate 'in the world'. The problem is not one about the earth or about turn-taking. And so even Sacksian proficiency in finding myriad cases 'orienting to the rule' will be like finding ancient fossils or having Russell assured by one older than him, and the rest of the empirical evidence. The point instead is: this is not the 'method' (so to speak) by which what counts as knowledge does count. (Even if it is sometimes a cause, it is not a ground — OC 74, 130. The notion of 'cause' is practically a component of that empirical method itself.) Instead: we must turn to repetition; to 'the recurrent procedures of the game in time (wiederholten Spielhandlungen in der Zeit)' (OC 519).
And no amount of avowal and assurance that something is so will lead to an acceptance that the claimant 'knows the rule'. Rather we need to ask: is this the correct technique?; is this how the practice is normally carried out? — including the practice of avowing knowledge.
((My tendency is also to think of belief-avowals and -ascriptions like this, rather than as contrasting with the case of knowing. Does 'I believe' cancel 'It cannot be believed'?))
It is as if, in denying Moore's claim to know such-and-such a thing, we were accusing him of speaking falsely — as if we knew what he 'really' knew. But no such thing is being said. Moore might well know this; but to treat 'knowledge' as raw material for exteriorisation, as something only the knowing subject can attest to by checking on an inner state: this is what has gone wrong.
A good deal of our public world could simply fall away if Moore's means of arguing could routinely hold up; if this always passed.
Moore's is only a single example of an attempted meta-practice — but the trouble with it can be generalised thus: the way in which a meta-practice proceeds cuts off the very sort of world which it would represent. By avowing its difference from that world, it is always already a (mock) transgression of it. But it accounts only for a world without the transgression.
What this claim to know evidently knows nothing of is the procedure of claiming to know. While its repetition of that practice may look all right, it is in fact a (mock) transgression; but, because there can be no 'transgressions of a practice', it is simply another practice — and not a claim to know.
522-523
In these cases, we are quite prepared to give some sort of primacy to 'mastering', to 'training' and to the repetition of 'technique'. So 'practice' seems to be able to stand alone. It can come first and have 'knowing the meaning' come later as an effect.
((Perhaps this is why Wittgenstein, say in the Investigations, keeps returning to the simple language-games of training and instruction; that is, as a mnemonic for the primacy of practice. Contrast: as an example of inquiry not knowing where to come to a halt, Chomsky asking for the cause of the practice (formal competence).))
'Mastery of natural language' is a product of the techniques of natural language. It is something it does or has done (lassen — to have done, to bring about). There is no further to go.
Attaching the name of its colour to an object: this can only occur where the practices of using colour words are already in the bag. Then the practice runs smoothly, for the most part. Yet here no one thinks, 'ah, to know the meaning of this word I must consult my inner resources'. So why do we do it in other cases? Perhaps we are (later) trained to accept a sort of weird picture of our training? Certainly this is so if some theories of education are looked at (both cognitivist and behaviourist).
What is it that relocates the phenomena of meaning, knowledge, understanding (and the others treated in the Investigations) internally to us? How do we acquire this picture (Weltanschauung? Ideology?) of our own acquisitions of meanings, knowledges and so on?
((Perhaps the conflation which that 'and so on' implies is itself an effect of that picture providing a seemingly common place for a number of actually quite diverse phenomena.))
On this picture of things, an active subject can be said to lose 'control' over his own knowledge, meanings, understandings. But 'control' was itself (but) an effect of that wrong picture. It wants to retain a sort of helmsman for passages through 'inner space'. But, paradoxically, it is equally damaging for that picture to take on an overly-radical subjectivism (solipsism). For then we are construed as controlling but not 'surmis[ing] the meaning of our words'.
So this picture can shore itself up against both radically materialistic and radically subjectivistic attacks. It has the force of persuasion — of 'common sense'. The grounds lie within and can be inspected or reflected — but each interior remains similar within a degree of tolerance. Thus arise dustbin explanatory categories: collective consciousness, common culture and so on.
Wittgenstein points to the abandonment of that picture. 'Surmise' is not prevented by the device of the common interior; it is prevented by an identical exterior — a ubiquitous exterior. If there is an agent at all it is language. And language goes on 'in' us only in the same sense that it goes on 'in' texts, 'in' police stations and so forth.
Practices of language (language-games) have as their effects certain 'knowledges', including conceptions of the subject and of interior-grounding. We might not be surprised at the rejection of the idea of surmising meanings, because the particular conception of an interior ground which emerges as an effect of our discourse obviates such a notion (solipsism, for example). But our failure to be surprised is for the wrong reason; for it is the domain of public, material practice which actually does the obviating here.
The picture Wittgenstein wants to dislodge sees practice as an epiphenomenon of a shared set of (mental) grounds which can be the subject of an unambiguous and representational commentary — which can, that is, be unmediatedly translated into propositional form. 'Practice' is, on this picture, just what we do when we put the deeper merchandise into motion — when we exteriorise it, when we show we have it 'all there'. (Intelligence testing is but a single drastic product of the picture.) The picture is an idealist form of the tendency to 'go too far' along the path of explanation. Nothing lies behind the practice. Attestations of 'grounds' (mental or otherwise) are its epiphenomena — if anything — if, that is, it's correct to talk of a centre at all.
As Coulter (1979a) shows all too well, 'mind' itself is the effect of certain discursive practices. (The problem with Coulter's text is that it then tries simply to describe those techniques. To do that it would have to be possible to recognise pure objects — mind, mental phenomena like 'motives' and so on — separately from their productional techniques.) But it is the near-ubiquity of the picture as a mode of explanation and reflection (!) upon human affairs that gets us to the position where 'doubts at bedrock' can arise. Because it holds the grounds to be 'in us', it views doubting as a form of self-doubting. 'Do I really know?', 'What can I really be sure of?' and so on: followed by an inspection of the relevant interior recess. And that sort of 'self-doubt' always looks very healthy — it immediately invokes a certain scientific nobility. It has the nod of the most popular and the most esoteric of philosophical methods. So doubting turns out to look not only permitted but practically obligatory. Whereas, in fact, from the very start, the whole business is in jeopardy. These matters are not the unique properties of the reflecting subject, in the first place, for that subject simply to take up and subject to doubt. They don't dwell in any arena where that sort of activity goes on. They inhabit a different environment where doubting has (must have) come to a stop. If there are 'doubts' here, the very production and intelligibility of self-doubting questions would be lost. (We think our subjectivities should be free when in fact it is the picture which includes subjectivity that entraps us. Instead, we should be free of the subjectivist picture of ourselves.)
I would say that doubting here is like trying to pull oneself up by the bootstraps — except that someone will then come forward and perform what their picture pictures as that very trick. Only with a shift of the whole terrain could they see that it's not unequivocally that trick that's being performed at all.
524
OC 524 explicates the phenomena which, in a form of life, 'go without saying'. Our sureness over these is shown, but not said — it is expressed by the absence of its needing to be said. And conversely, any instance of its being said raises a doubt. Hence scepticism confounds of one of the regular techniques we use in language. It is based on an over-literal attention to everyday language; for in everyday language 'I am sure...' only arises as a rebuff to contradiction, either raised by another person or imagined as possible by oneself. It does not mean that the one who says it is sure in the usual sense. The sureness, here, has the edge off it.
These are two 'practice[s] of the language'.
Wittgenstein, here, gives us instances of those practices, but not (the rule for) the practice. There is not a meta-description of the technique or rule — merely a reminder of what we know about 'I am sure...'.
Mostly the technique produces a particular sort of knowledge-effect: the 'immediate inclination to check'. The 'mistake' in philosophy is to think of the technique as producing another effect — certainty. The translation runs:
Utterance Philosophical re-write 1 'p' 1' 'It is the case that p' 2 'I am sure that p' 2' 'The speaker is sure that it is the case that p'1' is possibly a mistake in itself. It assumes only that the 'constative' aspect of p is in point. But 2' is in effect a reversal of 2 because it strengthens the case for p while 2 weakens it. 524 reminds us that we could easily 'retrain' ourselves to hear 2 as 1 and also in the usual way we do now, that is as expressing a slight doubt. What Wittgenstein refrains from, in his explication, is a description (like that above perhaps) of firstly the technique, then the knowledge-effects, then the means of matching the two. That sort of work is ruled out by the very conception in which the practice is held to produce the knowledge-effect. For that conception prohibits an inspection of the object, of the effect separately from the language practice, the technique. To inspect that object is to repeat the technique. What passes (on an alternative conception) as innocent scientific inspection of the practice, or the object or their relation, is actually repetition of the very practice which is being apparently inspected. Inspection is repetition.
Ideology (in science and philosophy) involves passing off inspection as something other than repetition. It might be passed off — in realism, for example — as a pure means of addressing 'natural' objects. But it can also be passed off as a critique of 'another' discourse (when it is, in this case, a repetition of the critiquing discourse). This happens in critical phenomenology for example (Habermas, Ricoeur, Telos...).
Wittgenstein resists an inspection of the practices with 'I am sure...'. Instead we are merely reminded of those practices. And this is why the sentence 'What is important is whether they go with a difference in the practice of language' (524) performs so much work, yet looks so idle. It looks as if we were sure of that all along. Wittgenstein, here, pushes (what might properly be called) 'analysis' to its limit.
Beyond the reminder — the invocation of our training in a language practice — and not discretely separate from it, lies the inspection which poses as other-than-repetition (which poses, then, as pure representation of the practice, as the rules for the technique).
((Note how Castaneda pushes Don Juan for the 'how to do it' to be uttered in so many words. Don Juan remains characteristically silent. To continue to practice the description of everyday techniques that phenomenological sociology demands, these matters have to be written off as 'philosophy':
[Ethnomethodologists] point out ... that this is a problem for philosophers rather than sociologists interested in analysing the everyday world. Ethnomethodologists are concerned to describe how members can be seen to be achieving their everyday world; a world both they and anyone else can recognise. Of course, they accept that an ethnomethodological study can be made of ethnomethodological studies; and that a further ethnomethodological study can be made of the ethnomethodological study of ethnomethodological studies and so on. They merely point out that it is not [in] their particular interest to continue indefinitely along this chain. Moreover, the production of any account anywhere along this chain requires it to be accomplished by the use of members' common-sense methods of practical reasoning. Their interest is in these methods. (Cuff and Payne 1979, p153)
Here the title 'neo-positivism' is not misplaced; for, although phenomenological sociology realises that technique precedes social objects, the 'technique' is then posed as a new 'natural object' for positivist inspection. The initial realisation, logically, precludes such a move. Some have taken this as an indication that a step back from 'phenomenological realisation' should be taken. Our dilemma is how to step forward from here.))
((Compare: the relative silences of Don Juan and of Garfinkel after Studies (1967). The problem is knowing where to come to an end; not relentlessly pursuing 'the spirit of exhaustive inquiry' (Castaneda 1970, p11). ))
525-526
In these cases, we are imagining someone who is 'at odds' with the rule that we usually follow with, say, colour-names (Farbnamen). And, as with certain forms of chauvinism and xenophobia, we imagine that the differences from the following of the rule are somehow blanks, gaps, absences or omissions. They are other than what the rule routinely generates. So we do not see them as actions generated by another rule — which they are. They are just some other practices - yet we always want to say they are paler forms of the 'right' practice. This is like 'schools' of linguistics arguing over what is to count as syntactic structure; when in fact the 'arguing' is an effect of the same techniques that produce the versions of syntactic structure in the respective schools. The schools generate different objects altogether.
'I am sure that...' is not a misplaced form of declaring one's sureness; it does not do that. It is not an awkward form of following that rule; but rather an instance of following another rule — the rule for giving a slight doubt. Cultural difference, expressions in mid-training, colour blindness — these are routine and recognisable techniques, each with their rules. They are not fractured forms of cultural homogeneity, full training and standard colour vision. These objects do not escape being objects of knowledge determined by particular discursive practices.
The 'distance' from the culture we (wrongly) feel these instances carry around with them is like the strange-ification that phenomenological sociology attempts. And in this case too we are dealing simply with the following of other rules and not the location of a scientific discourse beyond rule (beyond rule in order to describe rule).
When we ask for the rule, what we get is the following of the (or another) rule. Epoché is such a case; it does not lead us to discover the essence of everyday consciousness, it produces certain knowledge-effects collected as 'the essence of everyday consciousness'. And this is (merely) an object in its (phenomenology's) domain. It looks like a second order description or analysis of, or about, some natural phenomenon (consciousness) — but it in fact invites us, when we look closer, to refrain from (attempts at) second order description. Instead of asking the — scientifically 'noble' — question 'do you want to find out about everyday consciousness?', we should ask the question 'do you want to repeat the discursive practices of phenomenology?' And the choice is a question of linguistic strategy; not an 'epistemological' choice over matters of investigative subject and investigated object.
To ask for the rule, and expect a form of words, is like asking to see oneself blink (through the 'whole blink', with both eyes) when looking in a mirror.
527-528
Here we have a technique which (on some readings) contains a paradox: mastering the use of a word does not lead us to being able to say anything about our 'sureness' about what its 'referent' is called. Yet — when challenged — these are the very sorts of things we (seem to) invoke as a defence. We then say we are sure; we say we know it; we point to our status as native speakers.
This perhaps says most about a technique for learning (perhaps a second) language by interrogating its users. The technique looks for referents for each word; and there is something quite natural to this way of asking, for demanding the referent can be a routine community practice, stemming out of a particular conception of what language is and does. Joining in this game — say as informants — we become entitled to make statements about our sureness over the referent, our knowledge of the language and the rest. All the time, we are in the grip of a technique of the language.
Chomsky asks 'what is this thing "knowledge of a language"?' — and he is in the grip of a technique where the concepts of knowing, sureness, nativeness and reference come together as a routine part of it, the technique. Quite rapidly we are down the path of that technique (we have to be to read Chomsky) — repeating it ourselves, in and as our reading — and it looks as if we can derive cast-iron propositions about language and what knowledge of it is. We seem to have arrived at being.
This 'being', this 'is' is the 'is' of literalist scholarship. It is the 'is' of 'If he says "I am sure that's red" then he is sure'. To doubt that 'is' is not to 'take the edge off' such a sureness, to turn it into a poor relation of sureness. It is simply to separate community members' uses of one word (a colour word) from their uses of others, like 'I know...', 'I'm sure' and so on. The latter techniques are not about the former — they are other techniques and produce quite different effects. They have, for instance, to do with the language-game of challenging.
We continually press these techniques too hard by expecting them to be access points to the rules of other, more 'humble', techniques — like the practices with names and colour-names. We are obsessed with wanting to find representations for rules which flow counter to the rules for giving representations. This would be like a clock which stood outside the flow of time. It could not tell us anything. We think our instruments measure certain phenomena — clocks measure time and representations measure rules. But in fact the instruments are governed by their objects.
'I know it is; German is my mother tongue' is not so much an expression about mastery of techniques as an expression governed by it.
((In this sense, mastery itself is an effect of the repetition of the techniques and has nothing to do with control over them by an autonomous subject. 'Mastery' is like a process which happens to us.))
529-532
'I am sure...' and 'I know...'; these expressions have nothing to do with the 'state' (531) of the utterer; nor with her/his 'being sure' or 'having the knowledge'. In fact they can show an absence of competence on the part of the speaker or his/her interlocutor.
One who says — without it being a return statement to an interlocutor whose competence is doubtful — 'I know...' risks (strongly) having this reflect on their own competence. Children speak this way, as do foreigners — often in order to impress upon one their growing appreciation of particular cultural practices (the ones with colour for example). This is like a literal expression of one's knowledge whereas, in fact, competent members tend to use 'I know...', 'I'm sure...' much more metaphorically — as polite ways of addressing the other's grasp of the cultural techniques in question. Moore makes this mistake: he believes statements like 'I know...' can be — or are routinely — used quite literally. And don't some of the language-games of philosophy work that way? Moore's discourse gains its effect from having a move in one language-game stand in for that of another — though both sound the same: 'I know...'. This is like mistaking the question 'Do you feel like eating a piece of cake right now?' (spoken, for example, by a dietitian) for the invitation which has the same form.
But in deflecting attention from what someone knows to what practices someone repeats — in deflecting attention from 'knowing the language' to 'how we talk', for example — aren't we simply identifying another problem and not solving the initial one? Chomsky and other rationalists seem to think this is the case. And so they continue to ask ontological questions about our 'knowledge' (for example, of the language) despite the emergence of quite separate disciplines dealing with discourse praxiology. I want to see this ontological insistence, this will to being, as a further instance of Moore-ish literalism over 'I know...' and its variants. Chomsky writes:
...we could [still] search for explanatory theories that are concerned with the state of knowledge attained by someone who knows a language and the basis in the human genetic constitution for the acquisition of such knowledge, recognising that even full success will still not have answered all questions, for example, the question 'how we talk', just as an account of our knowledge would still leave open the question of 'how we act'. (Chomsky 1980, p80; my italics)
Chomsky's representation of the field opens up two positions. One turns to the question of 'how we talk' (praxiology), asking that we take 'communication' rather than 'genetics' as 'the essence of language'. The other position is that repeated by Chomsky. He doesn't say that 'how do we talk?' cannot be answered — though perhaps he thinks it's peripheral — but he does say there are specific and exclusive questions to be answered which can go on being answered without affecting one another. One question — the one Chomsky wants to answer — turns to the genetic basis of language and the other to what is actually performed by that biological equipment. So it is easy to see that, for Chomsky, the praxiological position is always an effect of the geneticist-rationalist one. Only the latter discourse is active in the cited text. Praxiological techniques are not repeated here.
The geneticist discourse holds itself to be 'other' than a social technique. This is its self-understanding. It poses the existence of tacit, background 'knowledge' and, only once this is in place, can the relatively more transparent sphere of action and talk emerge. Its own position as discourse is rendered irrelevant. It imagines itself as the pure wind of knowledge (seele, soul) blowing around the depths and hollows of pre-discursive 'knowledge'. It is precisely the knowledge of 'knowledge'.
From this position, Chomsky's geneticist discourse can delete the praxiological discourse it couples with the signature 'J.F.M. Hunter' and assume its rightful place as the 'motor' of medico-neurologico-genetic research into language structure — the 'real work'. (Though of course the analytic techniques of generative transformational grammar require none of this in order to work. The two are quite distinct.)
By virtue of its very constitution, and especially by virtue of the fact that that constitution depends on the separability of geneticist discourse from all questions of discursive action, what the discourse can't see is that praxiology might speak from elsewhere. (That is, that geneticism's narration of praxiology is not a repetition of praxiology but of just more geneticism.)
What geneticism could not account for would be a discourse which held (or ran): 'No, something called "knowledge" does not underpin action and something called "genetic constitution" does not underpin talk; instead, talk and action are all there are, and talk about "genetic constitution" is but and instance'. (And this also shows that the term 'talk' is limited — a geneticist and logocentric gloss on 'discourse'.)
This cannot be said within geneticism, as a repetition of geneticism. It is for this reason that geneticism and praxiology can never enter into 'dialogue' but must always speak past one another. There is no point on Chomskian terrain where the 'how-we-talk' position (discourse analysis, praxiology) could be demolished because it could not be constructed there, on its own terms, in the first place.
For discourse analysis or praxiology, grounded discourses, like geneticism and praxiology, cannot give the grounds of discourse. The 'hidden rules' which appear to emerge from them are always actually rule-expressional effects. The 'hidden rules' are always actually 'surface forms' (making a nonsense of that distinction) generated by the very unspeakable rules they are mistakenly thought to address. Chomsky's position — thereby — becomes self-annulling.
For the sake of consistency then, praxiology would have to abandon general theory — the search for (praxiological) rules of discourse. In this respect, Chomsky is right when he politely intimates that it could provide no challenge to a rationalist geneticist theory of language in the form of an alternative general theory. Instead it tries to point a way outside general theory. 'Minimalism' might be an appropriate label here.
532-533
On the contrary, to use 'I know...', as Moore does, is to repeat a technique of a particular philosophical discourse. The technique permits the utterance to be placed alongside utterances which use terms like 'mind' and 'state' without anyone who is also repeating this discourse thinking anything strange is happening. The repetition of 'I know...' has nothing to do with the more usual ones. This is what Wittgenstein glosses by 'out of context' (ausser dem Zusammenhange) (533).
Imagine: 'Just then, my state was like this: "I know that that's a tree"'. Hasn't something strange occurred here?.
Usually we don't collocate 'know' and 'state' in any such way; something is wrong with identifying 'I know...' as the expression of a state (OC 588). Instead, for each of the peculiar 'I know...' utterances:
I can imagine circumstances that turn it into a move in one of our language-games, and by that it loses everything that is philosophically astonishing. (OC 622)
What Wittgenstein wants to say about Moore's (and other philosophers') uses of 'I know...' expressions, I want to say about social science's uses of expressions about rule. That is, they are misguided because expressions like 'The rule for ... is ...' have their place as moves in various language-games. They are moves in those games, not utterances standing beyond the material practice of the games — just as surely as 'reading the riot act' is a feature of civil disturbance. The temptation in philosophy and social science is to take these expressions ('I know...', 'The rule is...') as investigable — as expressions — for their secrets about human knowledge and conduct.
The moves in philosophy and social science also relate to one another insofar as the latter are mistakenly held to lead us to (or to stand for) knowledge of the rules — professional scientific knowledge, that is, as opposed to mere 'cognising' — Chomsky's term for 'naive' knowledge! On the contrary, professionally or otherwise (practically or theoretically, we might say), we only ever)know the rules via their effects. Hatch (1980, p47) alludes to this with the term 'notionality'. 'Notionality' glosses the idea that rules or devices are simply not available to full analytic specification and that any necessarily partial specification 'of' them is, in effect, another instance of a rule's being followed. In this sense, we can only talk of the rule as is effects — by following it — by repeating just that technique. Such discursive effects include those theoretical texts which try to do an exhaustive description — a terminal or closed description.
To speak of the rule is (but) the effect of a(nother) rule.
534
Regarding this pleonasm (which is 'just what I want to counter the first sentence with'), there is a certain identity of language-mastery and competent practice. Yet it is not as if 'two things' were being matched.
One tends to think that Wittgenstein's injunction against 'special' philosophical uses of terms — and his continued insistence on ordinary language being in place just as it stands (TLP 5.5563) — both seem to point to a philosophical practice in which the philosopher firstly 'listens' to ordinary language and then this somehow unveils its own 'perfect logical order'. The key to the problems of philosophy (on this reading of 'ordinary language' in Wittgenstein) would be to have everyday utterances instruct the philosopher.
But there is another reading; one where Wittgenstein criticises Moore for following exactly the above sorts of policy; for trying to find out how the 'I know...' utterances can tell us something in the direction of 'problems of knowledge'. This reading would then arrive at a (philosophical?) strategy which consisted entirely of 'unadorned' ordinary techniques, which merely followed the already available rules and so forth. It would usher a minimalist participation — a very quiet intervention.
The philosopher (?) — like the child who is trained or training — 'must be able to do certain things'. That is, she must be able to perform the techniques whose repetition (re)produces quite local and practically available institutions. (So perhaps the term 'philosopher' becomes redundant.) Then an emancipatory interest would require those repetitions to be critical repetitions. Althusser held philosophy to be class-struggle in the domain of theory. Instead it might be held that philosophy can be institutional struggle in the domain of (theoretical) practice.
535
The language-game in which the child asks what a plant is called and the language-game in which she can answer these questions too: this reminds us that the crucial concepts in the family of knowledge and rules are 'training', 'institutions' (for example, natural history), 'practice', 'technique' and 'repetition'. The last of these is the upshot of the lesson in the Investigations that to know such-and-such is to be able to go on in a certain way.
((Is this behaviourism, since knowledge is no longer associated with a kind of ratiocination, and since 'mind' appears to have dropped out of the picture? Even within the unnecessary realism of the anti-behaviourist discourses on 'mind', we can say: no one has suggested at all that that particular location of, or site for, techniques' repetitions be abandoned. See the remarks on OC 522-3.))
The mistake in many educational studies is to assume that the reply 'That is called a...' is a token of the knowledge which the child has. Instead that practice of expression is, for all practical purposes, the 'knowing' — to use an extremely clumsy expression. Here too (in studies of training) the mistake is to think of the utterance as the shadow-form of the 'real' knowledge. (Many careers have been built on claims to state what it 'really' is a child knows when she utters such-and-such. These claims are clearly fraudulent, bogus.) Once again, such a belief sends us on the track of correspondences between language and language-transcendent entities of various kinds.
In all of these instances, concrete material practice is always treated as the mere trace of something more durable, more certain.
To illustrate this: imagine a language-game in which some people speak to each other in a common tongue — but very softly and quietly. Another person tries to understand them by finding regularities in the muffled whispers. Many analysts proceed as if our natural language were always very like this game, and they were the ones interpreting.
Or again: the Germanic letter-sign, the rune, gains its name from a term meaning 'whisper' (Borges 1974). It is as if something stood in the way of the language speaking clearly. Runes were carved into wood or rock surfaces. It is as if the barrier to clear speech were one of trying to hear the very rock itself. For the Norse peoples and the Germanic tribes, there was a definite magic about the runes — the magic was in the technique which some learned people had of getting words out of a stone. Before literacy was massively common, parishioners stood in awe of the priest who could, it seemed, utter the powerful words of God by standing close to some special sheets of paper (Hoyles 1977; Burke and Porter 1987).
Many still proceed as if our language were like this all the time — as if it needed a continual supplement. Some seek an interpretation of utterances, some a grammar. Some are looking for the meaning. All are looking for what underlies discursive practice. They want to add that to discursive practice. And this is like a superstition — a cult of magic. It is like the paranoia of the one who hears language as other people's unintelligible whispers; or like the secret and magic runes whose key may be found only by a wise scholar or a gifted priest. Some, guided by the superstition, have sought the hidden speech in people's minds — the lingua mentalis. Some believe it to be locked within it even before the child speaks — even before it is born. Some others have sought the key to the whispers in a relation between themselves and the objects to be found lying around the whispering places (the con-text). And there are still other favoured ways of conducting the (re)search. It is a second search — very closely modelled on the first, the search for the soul. (Cf. Derrida 1976; Ulmer 1985).
One way of doing this, which is sometimes wrongly attributed to Wittgenstein (Geertz 1973), is that we should 'really listen' to the whispers 'as the participants themselves would'. But that, too, seems to involve a secret behind or within the whispers. (Is another culture a secret?)
Another strategy might be to ask how we could desist from following these 'ways'. This would mean showing that they were always already 're-runs' of the techniques they (on their own self-understandings) would ideally penetrate and, further, that the re-running of those techniques does not seem to pose any undue problems of comprehension for members.
I want to say: actually there is nothing mysterious here at all, once we realise who 'we' (analysts) are; for our position is no different from that of any other language-users.
536
This is like the philosopher who has not yet got the concept means at all.
537
'Having the concept' is to be master of the technique of 'calling'. This must precede the technique of referring to one's knowledge of what a thing is called.
538
Any training, for example in the technique of 'calling', is not instant; the process takes its time. Only our concern with mental states as the fundament of knowledge makes us want to ask: 'At the given moment, was that state (knowing what X is called) in place?' It is not the state we should inspect but the training.
539-541
The behaviourists may be right about the first stages of learning in which one 'react[s] in such-and-such a way' (538). Certainly dogs and so on appear to be capable of such 'learning'. But there is a confusion (not about dogs but about language) when knowledge is ascribed to them.
The behaviourist account falls down when it tries to describe what happens when the training is in place. It confuses training in a technique with the (later) repetition of that technique. It concludes that knowledge must have been 'there' all along from the first moment of training. But the rationalist account is even worse — it considers knowing to be a state, but one which must have been 'in place' all along. The claim to know is a claim that one's training is pretty much complete. Someone whose training is complete can, if necessary, claim to know what a certain thing is and go on doing this indefinitely. Of the dog and the novice, such a thing cannot be said.
To claim to know what X is called, one must be able to extend the concept to cover anything. Using the concept to cover anything is not the simple quantitative extension of the limited case of successfully calling X 'X'. That description would cover what a dog learns and explains why dogs, strictly, 'know' nothing.
((It is in this respect that Chomsky's formulation of a grammar remains attractive: the infinite re-application of a finite set of rules. Yet there is no necessary reason why this formulation has to be accompanied by the idea that universals must be 'in place' prior to training. Equally, it by no means necessitates an analytic practice which proceeds to describe the rules. Transformational grammar might continue oblivious of a 'theory of linguistic universals' and vice versa. The two things seem to involve very different techniques and are not related to each other as process and product. In this disjuncture might lie a critique of Chomskian practice which doesn't have to encamp itself in behaviourism or empiricism.))
542-543
A position we could try would be as follows. What lies (literally) prior to the 'knowing' is an 'ability' which is perhaps connected with Ryle's (1949) notion of a 'predisposition' — and the knowing is founded (without pleonasm, OC 534) on that ability if we are going to say that it is founded at all. And that is simply because the 'knowing' is rightly ascribed (in that language-game) once the ability is no longer 'limited', once the concept is 'in place'. (And none of this has necessarily to be read as 'generalising' and the like.)
A concept is a unit of 'culture'. It takes over after a certain point. This is how it is that language speaks through and produces subjects (and not vice versa).
The language-game of ascribing and claiming knowledge is an instance of how, if language can be said to refer to anything at all, then it is to other language (cf. Silverman and Torode 1980).
...is it by chance that ... the meaning of meaning ... is infinite implication, the indefinite referral of signifier to signifier? (Derrida 1978, p25)
But in saying this we are only pointing to the way in which the network (Foucault: quadrillage) of cultural techniques is built up by language-games flowing over, into, through and past one another. That is, the instance of ascribing knowledge, simply because it recalls the practices of training, does not give it the status of meta-language. The level of the extra-discursive (including the level of the interpretation of discourses), insofar as it exists at all, exists as the effect of quite specific discourses. (Cf. Foucault 1980, pp197-198) —
544
As a consequence, the ascription and avowal of knowledge has its place and its setting. Because the technique could be deployed at any moment and one could truthfully say that one knows the colour of the sky, or one's own name, does not mean the technique will be deployed. And this is no more and no less the case than that one could — at any moment — remove one's clothing; that one has the physical ability to do so. If it weren't likely to be mistaken, I should say that the continual presence of a possibility of knowledge-statements - on which Moore relies so heavily — is something like a physical (as opposed to, say, a cultural) possibility. And in this respect, it is no different from other techniques like promising, calculating or questioning.
((Could we imagine a school of philosophy which, instead of blowing knowledge-statements out of proportion, did the same thing with, say, promises?))
545
At very early stages — but for quite a long time — some children can only operate with a single colour word, say, 'green'. When they match it up correctly with a green object, those around — if they don't know the child is operating in a restricted way — can be very impressed. The assumption they make is that this use of 'green' is like their own. What they wrongly give the child credit for is having learned a rule. One of the stages of training can involve the repetition of rule-like operations when no rule as such has yet been learned. (And this, of course, leads to the 'wrong' rules getting tried out: 'gooses', 'readed', 'thinked', etc. — all the instances of which Chomsky makes too much.)
((As a child, sitting on a bus with my mother, another bus drew alongside. I said 'Crossville' and other passengers seemed amazed that, at the age of about three, I could read the name in gold on the side of the green double-decker. Of course I read nothing. That was just the word I used for green buses; often wrongly. Just because the right concept wasn't in place, it does not mean that no concept was in place.))
Saying 'A child knows which colour is meant by the word "blue"' does not apply to a single instance; it points to an articulation between the child and the entire concept of colour. The initial discovery that such an articulation exists might be accompanied by 'I know...'. But such uses are 'countables' (Sudnow 1967); after one or two, or three or four perhaps, it ceases to be news.
What is already known ought not to be spoken of in so many words — this is perhaps how cultures get going. If Moore is pointing to anything, then it is the tautology that cultural principles are universal to cultural members. But nothing is to be gained from uttering the unutterable; one then wants to know simply which culture is doing the uttering this time.
In philosophy (of a certain kind) there still exists a naivety about discursive techniques and their limits. There is still a sense — as with Moore — that 'the' language can go anywhere, do anything.
The objection is sometimes raised that it makes no sense to talk ... about alternative conceptual schemes. Our present conceptual scheme is the only one we can envisage. We are, as it were, trapped inside it; we can never get outside to ask whether ours might be worse or better. All we can do is examine it from the inside, seeing how the various concepts function and which beliefs rest upon which. It would follow that to ask Kant's sort of question about what is required for any possible conceptual scheme — or for any possible sensible experience, whether like ours or not — is quite hopeless. For it supposes that we can get outside our own conceptual scheme so as to compare it with whatever others there may be, and to discover what features they all have in common.
But this is mistaken. Without throwing off our conceptual scheme we can perfectly well imagine what alternatives would be like; [below] I shall try to describe how experience might be for someone in a timeless world, and his conceptual scheme could hardly be the same as ours. (Walker 1978, p17)
And this is like: for all you say about the impossibility of describing general cultural or linguistic rules, here is a perfectly good one..., with this followed by an expression like 'S --> NP^VP' or a 'systematics' for the exchange of turns at talk (Sacks et al 1974). Isn't this sort of defence rather like Moore's since it is offered as a counter-instance? As though an item of language were like a physical proof.
'[Below] ... I shall try to describe ... and his conceptual scheme could hardly be the same as ours'. 'Describe' appears to have been sundered from 'conceptual scheme' But why don't we make the obvious connection: that 'his conceptual scheme' is an effect of the description?
546
A culture is not what it states but what it passes over in silence — not its reason but what it partitions off as folly — not itself but its other — not the totality of its contents but what is confined beyond its limits, ever unpresent(ed): 'whereof we cannot speak'.
Hence the distinction between saying 'I know...' (Ich weiss) and what everybody knows (kennt); and the difficulty of their conjunction — a conjunction which Moore wants to bring off as the most ordinary thing in the world: 'I know this is my hand'. Before 'I know...' can be said, there has to be a disjuncture between the saying and what everybody knows. Otherwise we are trying to utter nonmeaning.
((In OC, Wittgenstein has to scramble around the multitudinous paths of language in order to find examples of such disjunctures — where, for example, not the colour but the tones are in question.))
This realm, so dear to the philosophers — mind — could have as its origin (its conceptual origin) a metaphor — like so much else in language (cf. Rousseau and Condillac). The metaphor would be this: mind is the hidden realm, the dark side, the 'department of internal affairs' (Derrida 1978, p36). It stands for whatever is unavailable — it has its own Language prior to language (lingua mentalis) and it is this Language we must understand in order to understand language, its ways and its rules (grammar). As if beyond language and reasoning there were another — both like and unlike it. As though language were but one continent, one space which could be spoken from another. Hence the mystique about 'stepping beyond' (ekstasis), the hyperbole.
At the very moment of the attempt, the discourse freezes — it paralyses itself. Although it appears to go on (and express the grammar, the economy or whatever), nothing further is actually understood. Naturally: beyond the limits there is no progress to be made.
547-548
From the technique of trainings we can say we know such-and-such but (equally from the same techniques) we only utter such things as and when specific conjunctures arise. Whence our 'knowledge' — what we can correctly be said to know. So to use 'knowledge' to gloss the 'grammar' of a culture (the ensemble of techniques) is to speak — to use the most charitable term — metaphorically. Learning, having the concept, knowing, using — these things seem to hang together in a way which empties the mystery from a question like 'What are the grounds of our knowledge?' All we can say here, perhaps, is that the techniques do hang together. This happens. The mystery is removed by the (crude) direction: look and see what it is proper for us to say we know — and when it is proper for us to say it. This might at least prevent us attaching the tag 'knowledge' to the wrong sort of thing.
This is identical with the mistake about assuming that a knowledge lies behind the child's use of 'red' and 'blue' (in OC 547). Quite clearly, in this particular case, the knowledge still lies ahead.
((The etymology of 'infant': one who cannot speak:
infancy: ...want of distinct utterance...infantilism: ...an utterance or trait worthy of an infant...
[L. infans, infantis-- in-, not, fans, pr.p. of fari, to speak; cf. Gr. phanai]
(Chambers Twentieth Century Dictionary 1977, p670)))
549
In court, we do not get accused of perjury if it later turns out that our account was wrong (for example, there were three robbers, not four) if we gave our evidence in good faith but were just mistaken — really being sure at the time that we were right. The defence counsel may even ask us if we are sure about our details. Being sure about something can change while, presumably, its having been true cannot. (Of course this is not to say anything about our methods for establishing truth). One who says 'I know that there is a chair there' when there is no chair is not necessarily lying, even if the proposition itself is false.
And this has to do with which language practices are repeated where; for questions of the truth and falsehood of propositions do not exhaust the totality of questions about language even if they loom large in such practices as attributing mistakes, errors, slips, omissions and the like.
((One effect of the unbalancedness of philosophy: the compulsion to find the truth of every matter when plainly the features T/F do not apply in every case. Particularly they do not apply in cases where 'everyone knows' — where this membership re-members itself by passing over the matter in silence. Moore wants to get some mileage out of attaching the tag 'true' to some of these matters — a special kind of truth-beyond-doubt. And doesn't this 'specialness' itself seem to indicate the peculiarity of this step? Another symptom of this peculiarity: one might just as well, and equally arbitrarily, settle for the label 'false' in these cases. Being sure about something and its being true not only do not always apply co-terminously, they may not even legitimately apply to the same sets of utterances.))
550-551
The question (a) 'How does he know?' is not identical with the question (b) 'how is knowledge possible?'. The first may be answered — and it is routinely asked. The answer must turn upon a technique — 'This is how he knows' — and here, as in Wittgenstein, 'this' points to something like the repetition of a technique.
((In philosophy, at least in the West, there appears to be no embargo upon extending the first question (a) to cover the second (b), or upon answering the second by an investigation of the first. What goes (wrongly) without saying is the idea that the possibility of knowledge is always identical with the possibility of knowledge-for-a-subject. Where one is absent, an agent may always be legitimately supplied in this metaphysical game.))
So: to complete the tautology, it appears as if we are being deceptive, fraudulent — as if we are not doing philosophy when we point to techniques and trainings and when we refuse to answer the 'philosophical' question about the fundament of knowledge. It is as though philosophical issues had been evaded and, in their place, some sort of empirical social psychology had promised an empirical answer ('This is how the training went...').
But I want to say that the mistake in philosophy — the 'idea' that not only does knowledge have an origin, a centre, but that tit can be revealed — is identical with the mistake in social science about the representation of rules. The two are examples (in the sense of samples rather than exemplars) of a fundamental crisis — the crisis that erupted when 'language invaded the universal problematic' (Derrida 1978, p208). This was
...the moment when, in the absence of a center or origin, everything became discourse — provided we can agree on this word — that is to say, a system in which the central signified, the original or transcendental signified, is never absolutely present outside a system of differences. The absence of the transcendental signified extends the domain and the play of signification infinitely.
At this moment, all forms of (loosely) interpretation which seek to name, to represent a(n essentially absent) centre are marked as self-annuling projects; for the naming, the representing is no more than a further repetition. The movement toward an absent centre may be a movement in any direction — except that the movement of the signification supplies the centre which is its (supposed) end, its telos.
Some reassuring foundation for knowledge, some grammar, some rules of social organisation: these are samples of the transcendental signified whose connection with any possible form of signification is in crisis.
((This crisis is marked — as one of its instances — in the passage from the Tractatus to the Investigations and, more specifically, in the realisation that the general form of the proposition cannot be given.))
In 'philosophy', markers of the crisis can be traced in a range of texts — in Wittgenstein, Heidegger and Nietzsche for example. But in social science, that 'tradition' is all but absent. (For example: once semiology lost its empiricist techniques and the materiality of the signifier began to be stressed, it no longer remained as a fashionable form of doing cultural studies.)
Social science appears to ransack philosophy for possible empiricist outlets. Wittgenstein (via the texts of Austin) gets transformed into speech-act analysis, Husserl-Heidegger into ethnographies of the eidos of everyday events, and so forth. And, insofar as it too has affirmed the presence of a centre, rationalism has also been appropriated as a source of 'social science practice'. The name 'analysis' belongs properly here. 'Analysis' seeks the reduction of the empirically given to its (eidetic) constituents. The analytic imperative of social science paradoxically retains the nomothetic platonism of Husserl and, perhaps equally ironically, the ideographic aristotelianism of Descartes. There is a sense in which 'social facts' are convenient analytic fictions for Durkheim and an equal sense in which Weber is the theorist of an objective social structure.
The rationalism of empiricism and the empiricism of rationalism which the social sciences so unswervingly produce over and over: these remainders are displayed in their relatively harmonious shuffling movement from living presence to eidetic structure and back. (Marx, the calculator of the internal structural motions of capital, is also the ethnographer of the English working-class of the 19th century.)
We should give the simple title 'empiricism' to this entire movement where, for example, that name names the naming of nomos (the law).
Because we pick our examples from social science (law, rule, structure, grammar), there is no necessity to stop here. One might add: God, man, nature, the body. Or: the final particle, the first moment, the genetic code and so forth.
Another project might be to trace the 'history' of these concepts — the formation and interweaving at specific junctures of the techniques they in fact gloss. And the hardest thing would be to do this without doing philosophy (or history of ideas, or sociology of knowledge, etc.). Foucault (1976, 1979) points to the sort of project I have in mind.
Now we can see why pointing to a technique with 'This is how this could be known' is philosophically (etc.) unexciting. (We can 'just see Russell's face'.) But that only tells us something about philosophy and not about the practice of pointing to a technique. That practice just gets repeated (in philosophy and elsewhere).
To understand is, indeed, to go on — repetition, iteration. The text or the signifier is understood when it is done over, re-played. And the understanding cannot be accessed separately; for example, by seeing whether or not the reading (considered, say, as a mental event or series of them) matched some putative fixed point (such as, meaning, structure and so forth). The re-playing can't be matched, it can only be done. If the 'understanding', the object, the product, were to be matched with its productional technique (the point of some ethnomethodology; for instance where the object 'queue' is matched against the pure technique of 'queuing') it would have to be possible for the object (meaning, sense, understanding, reading, etc., in particular uses of these terms) to be recognisable to the putative meta-language in the absence of its technique. The object would have to be available as pure self-presence. And it is this which does not seem to be possible.
((In some moments of social science, the object is an 'empirical' one like a queue or a reading; in other moments it is 'theoretical' like a myth-structure, the unconscious, etc. Naturally, the locus of the 'technique' will be correspondingly shifted across these cases — but the argument still holds. For instance, it would not affect the argument were we to shift from the object 'a deep structural representation' to the object 'neuro-genetic basis of language'.))
Until very recently in these remarks, we have proceeded as if the centre (nominally 'rule(s)', because of our acknowledgement of social science) were unspeakable or unwritable because it governed speaking and writing themselves and so remained eternally hidden. Now we have to face the harder option — that the centre is (was, has been, will be...) absent.
((The expression 'is' is always wrong here — for by a conventional move we press it into the service of presence. This makes absence look like a mere form of presence. Parallel to this is the ambiguity of 'nothing is certain' where we may want to mark the absence of all certitude or the presence of some thing ('nothing') wherein certainty surely resides.))
In the Tractatus, the 'whereof we cannot speak' always presented itself for a certain form of representation. We tended to think of showing (repetition) as a (philosophically unsatisfactory) version of saying. In the Investigations, the 'rules of the game' — the code — are always just around the corner, being shown, indexed, pointed to. Their outlines are traced via an investigation of their limits. ('There are no words for the shape of this but, here, feel...'.) This investigation of certainty now seems to point elsewhere. (And here a definite telos is to be avoided at all costs). Specifically: it points to the cutting away of terra firma in the matter of language. The machine which set the flight in motion turns out to be an effect of the motion of flying — and there is a distinct vertigo.
Is this a new scepticism — or, worse still, do the constructs 'technique', 'practice', 'discourse' and so on merely crystalise to take up the place of the absent centre? What are the texts of Wittgenstein doing with the concepts 'language-game', 'form of life' and so on? It would be very strange if they were rekindling a system, a systematic philosophy — of any kind, with all its differences. What we have to discover is how they point us away from philosophy — which has always been 'systematic' philosophy. Clearly, they do not indicate an empiricist social science — a typology of behavioural acts. Equally clearly, they do not indicate silence, nihilism. So...
If something is shaken here it is not the certainty vacated by the sceptics and 'reinstated' by Moore — it is something of the order of the 'origin' of the concept of certainty and its repositioning as the effect of the repetition of particular techniques. And not just 'certainty' — or at least, 'certainty' is worked through only as an example.
What is presaged here is a sort of micro-history (historia, knowing) of the concept. But this too is wrong — the wrong formulation of these investigations; for it simply raises a new centre and another empiricism.
((Feyerabend's move 'beyond empiricism': if only it were that easy; a sort of voluntary act like giving up smoking.))
How is it possible to conceive of a philosophy and also a social science which does not begin with an originating agent — be it consciousness, economy, man, etc., and which does not end with the problem of describing or representing that agent?
((Perhaps the most immediate problem is: how to stop investigations which do proceed that way.))
552
Here is an instance of the trouble. No one is going to say that such and such is known. It simply isn't said. That technique is never repeated, though some like it are; for example, the technique for dispelling others' doubts. This signifier never comes into play. The philosophical inquiry here is like the inquiry 'And what happens when the umpire takes a turn to bat?'
But even though it is like this inquiry, it still has a certain force. We don't want to say it's just nonsense. And this is because philosophy trains us to scan beyond the range of the signifier itself for the grounding of these matters. A proposition's being true or false is a case in point.
And suppose we then say — okay, it is true. You do know about the chair and all. That's the effect of the technique of truth-inquiries: the proposition 'You do know...'. This can be accepted. But then we have to say that that's just more technique. It doesn't prevent the illicit move 'Do I know I am conscious...?' and its like. This is a separate move, another signification and the two go on independently. The second is not an answer simply because the first was not a question.
((This would hold, too, even if the second move registered the result 'false'.))
It is simply that ascribing or avowing knowledge and registering truth are separate practices. But at the level of the signifier (which is all that can concern these investigations) the first utterance just does not get going. The clause 'if one doesn't say it' rules it out — there is no signifier here.
((Yet, 'I am conscious' can be a signifier, just as 'blb, blb, blb' can be.))
When Wittgenstein locates the difficulty as that of knowing where to stop, one heuristic could be to sniff for an upcoming (transcendental) signified and to proceed with caution.
So the mistake, in this case, is assuming that the effect of one technique can stand as centre for another. And, of course, it can: insofar as we are going to retain the centring of practices. In fact only the effects of other techniques can stand thus. But they always emerge as non-centres, as supplements, to use Derrida's term. So the mistake is — technically, not condescendingly — understandable (repeatable).
With a glance towards decentring, the problem could be dis-solved. And that, too, is a kind of therapy.
One problem that supplementarity raises is the reading of OC in which the step into context is read as a step towards providing a new centre for discourse. It looks as if '"I know I am conscious" can be a signifier' should carry the rider 'in certain circumstances' or 'given the right context'; so that we scan beyond the signifier to a new possibility of matching an object with a technique; a new possibility of analysis. Strictly, the location of meaning in the concept of context is a supplement — an addition to the discourse. Its retention is like the retention of truth, sign and the like. It is almost our only way to proceed. 'Context' ought, now, to remain parenthetical — a transitional term — at best a metaphorical displacement for empirical specificity within the indefinite play of signifiers. Its own day has already come as a 'fallen centre' (Derrida 1977). But no one is yet pretending that there is any pure discourse, shaken free of its (false) centricity. (Quine's 'web' of knowledge?)
((Why does 'context' appear more innocuous than 'man', 'structure' or 'consciousness'? Why does it pass almost unnoticed as a central concept? To find that this is what it is, all at once, has a kind of menace that only gradually seeps out — like, for example, in The Castle.)) —
553
Something queer about 'I know...' in the wrong context: all along this has been an item of concern and investigation. Perhaps what is troublesome is not the absence of an answer but the mode of posing the question. That is, perhaps we ought to try to work around the relativistic centre which can be called 'context' (or 'occasion', 'situation', 'place', 'circumstances' and so on).
'Context' is paradoxical. At once it seems to free everyday language and events from fixed essences (the unconscious, a calculus or grammar) and, at the same time, it constitutes an essence or centre within (relativistic) social science discourses.
Hence, in the sharp division between the 'everyday' and the 'scientific' (or its equivalents) which is a necessary product of these discourses, the former can emerge as decentred, loose and 'contextually oriented', at the expense of shifting the entire burden of centricity on to the latter, 'scientific' discourse.
'Context' has this equivocal position in phenomenological and relativist-ethnographic disciplines. Here, at the moment of empirical analysis, 'context' is held to be other than an essence pre-existing actors, rules and other constituents of 'scenes'. But at the moment of inspection of its own discourse, phenomenological sociology finds 'context' to be a concept with very high priority. 'Context-sensitivity' must then be located, transcribed, analysed, rescued from everyday events. Its centrality as an analytic concept is overwhelming.
Likewise, for empirical analysis, 'indexicality' points to the absence of an essence. But the position of 'indexicality' within, say, the practices of training ethnomethodologists, is identical with 'function' or 'equilibrium' in traditional American sociologies.
This incoherence in the discourse generates a variety of 'queer' moments very like the instances of 'I know...' analysed in OC. For instance, the notion that context-sensitivity is co-existent with context-independence in conversational analysis (Sacks n.d.; Schenkein 1978) makes the expression of conversational rules a legitimate practice. But it, thereby, generates (heavily disguised) tautologies like 'One person talks at a time', 'Speaker turns recur' and so on. (This is endemic to some ethnomethodological work.) The position of such rule-expressions at the level of the 'object' discourse and their position at the level of the 'analytic' discourse become disjoined. In one place they are peculiar, mad or funny (for instance, Sacks and others appearing in the Book of Pseuds) in another they are official wisdom.
And now we are saying 'In one place ... in another place...'. We might equally say 'in the context of everyday speech', or 'in the context of scientific exchange'. But what this glosses is not so much context as discourse. We want to say 'The expression arises differently in two different places' and instead we could say 'There are two distinct expressions, each repeated within a separate discourse. It's a matter of some minor empirical interest (phonological interest perhaps) that they have the surface form "I know..." in common'.
The expressions 'special occasion', 'when there is some need' and also 'unjustified and presumptuous' and 'perfectly justified and everyday' can be read as glossing this problem. There seems to be no particular reason for examining the question of truth — once again.
((I want to ask: what is the reason for the constant occurrence of examples of self-annulment in social science discourses? How is it that they seem forced, compelled to specify central concepts which have systematically ambiguous positions, in both the domain being studied and in the scientific model conducting the investigation? Quite obviously, if the meta-linguistic step cannot be taken — and it appears that it cannot — then that ambiguity must have some other explication. It is equally clear that the ambiguity and self-annulment would be replicated in any social science which turned upon the concept of a hidden centre; and that it could clearly not be replicated in a decentred social science. It is, of course, entirely possible that the myth of the meta-language arises as a possible way of deferring the problem of self-paralysis. Where a central and damning contradiction exists, one can continue in the face of it by relegating one half of the contradiction to another 'level' or 'dimension' and specify that these dimensions shall not be adjacent or overlapping, or translatable — except perhaps under special conditions.))
554
A language-game is not a context but a discourse. Its site, that is, its social 'co-ordinates' are irrelevant. 'In the book', 'in my head', 'in the kitchen' (and so on): these distinctions are relevant in the case of assigning context-labels, but redundant in the case of deciding which discourse.
Hence the expression '...as soon as I say this sentence outside its context' has to be read as deflected, as partially metaphorical. The 'context' (or rather, the language-game) lets us know that this is a legitimate reading. ('In its language-game' vs. 'outside its context'.)
No one thinks to ask: just what is this language-game of Wittgenstein's, the one wherein — and not whereof — he speaks? Is the lame excuse of climbing up and only then discarding the ladder — which implies a certain empiricist temporality — available here? Who plays the language-game of language-games? Not even a god itself could tell me how to do that.
Perhaps we can locate here Wittgenstein's particular variant of self-annulment - the custom of the language-game is our reference point, that which tells us how we may and may not speak, so what would it be that authorised a discourse 'about' the custom of the language-game? Either: theory has its extra-discursive position of privilege, or: it is the custom which authorises discourses 'about' itself and so theory becomes an instance of practice, but, like it, cannot have access to any hidden totality which authorises the 'play'.
In every case of self-annulment in the human sciences, we can begin to discern a double appearance of language. In some discourses only (that is, what follows is not as rule but an instance, though I suspect a very widespread one), logos (discourse) is held to generate two distinct continents of human existence — one natural (physis) and the other cultural (nomos, law, custom). Hence
L ::= p/n
The first methodological move is then to decide upon further constituent features for the initial set of constituent features (p/n). Physis proves troublesome as, in the absence of nomos and hence logos, it cannot be expressed:
L ::= p/n
p ::= ----------
The social sciences, here, tend to turn away from the natural — although various infantile (pre-utterance?) forms like exchange theory believed p to be expressible and so to be the crucial feature (at this level) in social science. Sociobiology and human ethology continue the fiction that natural man can be spoken. (Of course, in the natural sciences, physis is uttered; the question of logos is entirely effaced so that nature can seem to write itself into the journals. Some Neanderthal social science held that the social world could be treated analogously — hence the positivisms.)
To continue: the problem of the centre arises in the further specification of nomos. All previous forms of social science have written a centre (man, God, rule, structure...) under this heading; usually along with other theoretical objects acting as supports. We can represent these by dummy entries
L ::= p/n
p ::= ----------
n ::= x^y^z^ ....
where '/' shows division while '^' shows a choice of one or more than one. For example, in any instance:
n -> x z or
n -> z or
n -> xyz
These instances cover the majority of social science discourses. The trouble that arises is that, in the very critique of discourses which believed p to be expressible or believed n to be identically expressible with p, there arose the nomocentric social sciences (all anti-positivist forms). But almost simultaneously with this move, as a necessary correlative of the critique of the expressibility of physis, nomos is compelled to take on a new constituent: logos itself. Now logos makes its double appearance:
L ::= p/n
p ::= ----------
n ::= x^y^z^L ....
Now we can derive all sorts of double-appearance phenomena:
L -> L or
n -> x^y^z^p/n ....
and other troublesome derivations (tautologies, infinite spirals). The obvious procedure has been to distinguish between L1 and L2, the discourses of the investigator and the investigated respectively. Clearly this cannot work without the ascription of a privileged form beyond the entire calculus — one from beyond logos, physis and nomos. Transcendence is, then, one move to be made along with the meta-linguistic step. Another move is to scrap all but nomos and to have it as the unspeakable presence which is the fixed point in the play of utterance and actions. By definition, nomos, on this view, is pre-discursive. Naturally, some have turned to biology again here. And this seems to exhaust the steps that can be taken (along with the fall-back of empiricism — enforced methodological naivety, a sort of community solipsism) in the face of the crisis of logos' double appearance.
((Is logos another centre? In this crisis the answer is always Yes and No. Yes because, as one feature of nomos it acts as the centre for certain social science discourses (like semiology). No because it is always already the condition of existence of what counts as a centre or totality. As its condition of existence, it cannot also be a centre or totality.))
Possibly, then, not even a decentring of the human sciences will lead us out of the crisis — for it alone does not cancel or conceal logos' necessary double value (in either logonomy, the law of language, or nomology, the language — or science — of law). What is set up is a double-shuttle — a perpetual motion machine where no question is ever resolved but is always deferred. The beginning is lost, forgotten or just absent — as is the end. No investigation is ever freshly opened — it has always already been conducted before and none is ever completed and closed up. With the double-appearance of logos, infinitely great or small methodologies can be concocted — and there is always more to be said; for every methodology must implicitly or explicitly allow for a recircuiting at the point(s) of logos.
Social science discourse — at the point of, but not beyond, the crisis — is in turmoil. But it is pointed towards (if this is the correct word) a position where origins and ends are shunned as final repositories of explanation; a position which 'affirms play and tries to pass beyond man and humanism ... the reassuring foundation, the origin and the end of play' (Derrida 1978, p292). In this case, it appears that nomos itself could fall, to be as mythic as physis; that the rule is seen and written neither as an available nor as a hidden foundation but as an absent one. With the disappearance of nomos, both appearances of logos become one again — there is then only discursive practice, including theoretic practice. In this case the division between theory and practice falls and social 'science' rejoins social 'life' — as if anything but the flimsiest of screens ever separated them.
555
The attribution of how one's knowing something comes about is, itself, a language-game. (Or else there are games of attributing 'how one knows' to reason, to experience, to intuition, to a position in the 'web' of knowledge and so on.) In this sense, discourse precedes experience, reason, and so on. Discourse produces those categories. The concepts of 'being in a position' and of 'general consent' mark the priority of discourse (technique). Being in a position to know is being in a position to have a certain discourse 'play itself out' through one. It is to have the consent of the discourse which masters one.
((Wittgenstein often writes as if language were a set of available techniques ready, as it were, for the taking — waiting for the volition of the autonomous user — their master — to arrive on the scene and take them up, use them for some extra-discursively given end. There is this voluntarist element in Wittgenstein.))
'People will agree...'; but that agreeing is not done outside the language-game (of attributing knowledge to, say, experience). The agreeing is itself a discursive technique which has been learned. Agreement doesn't secure the language-game, like a contract with universal signatories. Agreement itself has to be already in place.
We don't hesitate to agree in these circumstances. The agreement, almost, happens to us with matters like 'I know that water boils...'. And it happens right away. It is almost not actually agreement, because the alternative — disagreement — is never presented. With language-games one doesn't choose to agree; and neither does one choose to take part in one rather than another. Often, it is only retrospectively that we discover 'which game', that we find a sense in what we have said or written.
'The writer's thought does not control his language from without; the writer is himself a kind of new idiom constructing itself' (Merleau-Ponty 1964, p9).
'My own words take me by surprise and teach me what I think' (Merleau-Ponty 1952, p97).
In the absence of a centre for our discourses to speak/write (and so rightly position our knowledge once-for-all), it is we who are spoken/written. That is how we know.
The mistake is to situate the epistemic as the object which, above all other, has to have its origins accounted for. (As if, by doing this, we would have documented the mechanics behind the well-spring.) The epistemic is, certainly, only another effect; so why do we imagine its particular foundation to be more fundamental than that of other 'objects'? Knowledge-effects somehow take us in. They have us read too much power into the force of their signification. Epistemology carries its own, tautological guarantees. By persuading us to read everything as a case of knowledge, or as the derivative of knowledge, epistemology persuades us of its own universality and privilege as the knowledge of knowledges. Nothing seems more natural, under this compulsion, than to take the 'theory of knowledge' as the most basic position which an investigation might reach. The epistemic positions itself as the origin of discourse.
And yet the epistemic is always an effect of an institutional discourse (for example, empiricist discourse). Again: the double appearance of logos — and a fundamental contradiction in the practice and theory of epistemology.
((In any case, isn't it peculiar that we seek the grounds of knowledge in a theory of knowledge itself and not in a theory which ought, rightly, to make no explicit use of the concept it would preferably generate as its outcome? And don't we proceed this way precisely because we are already able to operate (with) the concept 'knowledge', wield it in a certain way; but yet we remain unable to find that that operation, that wielding, is the very answer we seek to the 'problem of knowledge'?))
556-557
To lie, to mistakenly say something which is untrue, to be confident (although wrong) in one's assumptions or beliefs, to know something to be false... (and a longer list could be given): each is a distinct practice.
Does this mean that the concept 'being in a position' is peculiar to language-games with 'know'? Surely not. One may be in a position or a situation (Lage) to do all sorts of things (if not believe — quite).
Here Wittgenstein wants to situate what will pass for knowledge as a matter of community membership, contract and agreement (and not, say, as a matter of correspondence with transcendental or ideal objects). The concept of position (Lage) appears to speak quite plainly to that matter. It is the member-position of the speaker that, on this theory, decides whether the utterance is warrantable as knowledge.
And here it seems as if, in order to get this innocuous concept of 'position' off the ground, we have to install certain props for it: membership, contract, agreement, speaker. And then some distinct autonomy must be granted them in their own right. So they end up as instances of the very ideal objects we wanted to avoid. The step into culture, practice, custom, etc., going by the name of 'conventionalism' or 'relativism', is itself a general theory — a centred and closed theory — of a kind. And its most notable aspect as such is its concept of the autonomous — though often collective — agent.
((Perhaps we should remember that although the term 'agent' once simply meant one who acts, it has come to mean on who now, acts on behalf of a higher power. In this modern sense, the autonomous agent is self-contradictory.))
558-559
'The foundations of our language-game', 'it is not based on grounds': now here is a decision — a rupture in our reading of On Certainty and Wittgenstein as a whole. Water boiling, the sun rising, my name being such-and-such, this being my hand: we are tempted to think — like Moore — that we are enumerating the propositions that act like anchors for our knowledge. Certain propositions — one wants to say, 'pre-empirical' propositions — must lie there and must be absolutely unequivocal; this layer must be put down on the shifting grounds of.... Of what, precisely? For prior to our discourse there is nothing. Perhaps: our being is not founded after all upon an elusive presence but upon an absence. With such a founding we can both refer to 'the foundation of our language-game' and, simultaneously say that 'it is not based on grounds'. Now the concept of a hidden centre appears nonsensical. The centre of language is inexpressible not because it is (logically or otherwise) obscured or eclipsed but because it is absent — it is absence. Sense (or meaning) arises out of absence, simply by virtue of its difference from absence. The words find their meaning in the gaps and spaces they leave. That meaning is always unpredictable (Unvorhersehbares), not given by pre-discursive experience, reason, intuition, etc. —
560
It is not rationalist epistemology or empiricist epistemology (and so on) which are misplaced — such that we continually play one off against the next in the strange belief that we are making a choice — it is epistemology itself which it is necessary to scrap, to do without. It is the question of the grounds of knowledge (whether as foundation or as web, or as any metaphorical allusion) which needs our attention — not the answer to that question. The maze of calculations leads us to a point where the question appears as a next and quite ordinary step. But the pathway is taking us nowhere — for our attention ought to be turned instead to the language-game.
So instead of asking for the grounds of knowledge we shall want to know the means of (re)production of language-games — their determinate conditions. Yet that investigation is not an empirical or descriptivist one, and the 'conditions' in question are not matters of empirical context (which would have to pre-exist the game — for nothing pre-exists the game). The language-game is always already in place. Our work (or play) on its (re)production is not a search for origins or pre-existents (or indeed for its end, its telos).
Work on (or play with) knowledge can only be work on the language-game's reproduction, that is, work in that reproduction, at the site of the language-game itself, in its conditions and institutional repetitions. It cannot be work (an investigation) of the second order. Hunter writes:
The question of the 'grounds of knowledge' will thus be re-posed as a completely new problem: the problem of the conditions of reproduction of technical and discursive practices. It has already been indicated that this is not simply a move from epistemology to sociology. The means which problematise epistemological descriptions in fact rule out all second-order descriptions of discursive practices — including those characteristic of the sociology of knowledge — by locating critique on the terrain of the actual reproduction of the technical and discursive practices. Critique thus has a new field and a new object. (Hunter 1980, p13)
And even where it looks as if one has left the terrain of the language-game itself and moved up to some other descriptive level, still it is the language-game which is reproduced.
I want to say: we have to work, instead, 'consciously' on the 'terrain of the actual reproduction' of discursive practice. But 'consciously' is not correct.
((An example of ideology at work: Gouldner — to name one text-corpus — wanted to restore sociological problems to the domain of social problems, to avoid, thereby, a crisis. But by what means were sociological problems ever divorced from social problems, science from life? That there ever was such a moment is doubtful — yet it is a myth which gets both theoreticism and its opposite going in sociology.))
((Surprisingly it is Freud who reminds us that everything begins with reproduction and that the question 'reproduction of what?' is pointless. This is the crucial shift — in psychoanalysis — from a topographical metaphor to a dynamic one (Freud 1955, p611). In speaking of a dynamics of language we indicate a flow which '...is there — like our life' (OC 559), not to be represented but lived.))
What is being asked is that social science's congenital genitive be replaced by a perennially absent locative — that it cease attempts to speak of the language-game and simply speak in it. And this is not a matter of choice — the mythical 'choice' of methodologies.
561-562
The concept of 'knowledge' is a concept — a piece in the language-game. To deal with the question of knowledge would be, for example, to train someone in particular moves. (Not to describe particular moves.) And this means that knowledge cannot pre-exist the language-game. It is not behind our speaking and writing. In an extension and corruption of Merleau-Ponty: I speak (not necessarily aloud, of course) in order to find out what I know. This is like checking the score, seeing how the game is going — which of course one does as one plays. Knowledge is like 'the score' — it is something which is an effect of the game but also a part, a 'piece' of it. Looking for how what we know underpins our practice is like saying 'The score is so-and-so, but the game is not yet begun'.
Perhaps it is important not only to imagine a language without our concept of 'knowledge' (562) but also to bring about such a language. The point may not be to describe language-games (how could it?) but to change them. And, given the impossibility of the former, the latter may even be the easier move. How do critical interventions get made in the necessary absence of the meta-linguistic step, of second-order descriptions?
One thing is at least clear here: a (pseudo-)descriptive repetition could not be a critical one and a mode of ceasing that ubiquitous form in social science would be a necessary prerequisite to critique. Then: how to intervene in — how to repeat — the discourses of social science in such a way as to show the contradictions of descriptivism?
((For example: Silverman and Torode's (1980) technique of describing what it is the discourse of, say, Althusser says and then running this against a description of what it is that discourse does, is massively self-paralysing — especially where it is description they want to de-authorise. Such a way of working would have to be ruled out. And this paragraph is an example of it.))
563
As with 'knowledge', so 'proof' — and other forms of reason — and also 'experience' have their place in the language-game, in particular techniques. For instance: in the technique of reassurance, in stating a source of reliance. And we are misled by this. We think: one says 'I know' — how does one know? — well, one says 'You can rely on this — here's the proof' or 'You can rely on this, I just saw such-and-such occur'; and this order(ing) of discourse points to a hierarchy with rational, experiential (etc.) grounds at the base, knowledge at the middle level, still somewhat hidden, and language at the surface. What we are forgetting is that a few elementary techniques of the language-game have been carried out. The pyramid itself (our feudal conception of knowledge) is the effect of a further operation ostensibly upon those other operations of the language-game. Another technique — a supplement — gives this pyramidal structure to language-knowledge-grounds.
So the technique of substantiation (Begründung) comes to an end — though it is eternally repeated up to that end point. What is this end that one comes to when doubt and (hence) the supplementation of grounds (Begründung) comes to an end? Beyond the end there do not exist more — secret — grounds awaiting some new philosopher or philosophy. The end is the location of the second-order in the first-order; of the (linguistic) description in language. What one realises when one reaches the end, the limits, is that one is eternally in(side).
564
'I know...' works as a technique in the language-game — and like other techniques it has its own features. One of these seems to be that it works in order to gloss the work of certain other techniques. It 'pulls together' the means of finding out certain things, here: counting, measuring and so forth. This seems to be its 'natural' place: for when there is no problem with these means of discovery, 'I know...' can be dropped.
((Is this a second-order description? It tries to be: but isn't there a sense in which we have to find out about the language-game so as not to make mistakes in it? Wouldn't it be part of training to see how 'I know...' might apply and how it gets misapplied in pseudo-descriptivism? This would seem to be part of the therapy of doing critique in the language-game.))
Perhaps it is this glossing technique with 'I know...' that is behind 'knowledge' coming to be the signifier of all signifieds in epistemology. It seems — normally, naturally — to cover all manner of practices, and then we have to think: well, mustn't there be something in this? something substantial behind this routine gloss? 'Know' seems to cover all these things; mustn't there be some hidden, transcendent, process called 'knowing' or an entity called 'knowledge' which lies at the centre of our practice? Then we are off on the scent; the domain of practice/technique seems to have been left behind in the search for something deeper. But that search (the various epistemologies) only gets going — we forget — by confounding something relatively arbitrary about one technique (the one with 'I know...') and taking it as the index of something essential.
((An analogue of the trouble is the way in which functionalist linguistic analysis finds a 'semantics' in arbitrary matters of form or style:
Our analyses suggest ... that lexical items, linguistic forms and linguistic processes carry specific meanings.... (Fowler, Hodge, Kress and Trew 1975, p186)
It is only when we acknowledge the meaning carried by the items themselves, that linguistic form can be demonstrated to be a realization of social (and other) meaning.... In Saussure's system a term has meaning by virtue of its opposition to other terms, though in itself the term is without content. In our theory terms in systems have meaning by opposition, but they also have content in their own right. (Fowler et al, pp188-9)
For instance, the way Halliday (1971) in his analysis of The Inheritors locates the passive consciousness of a tribe in the fact that they are never spoken of (narrated) as agents or as attributing agency in active sentence forms. Here there is an enormous pun which sustains the analysis. The 'agent' of grammar stands for human agency. The 'active' (or the 'passive') modality stands for a form of action or consciousness; a 'mood' perhaps — for the extensions of the pun are endless. A similar set of puns can be got out of concepts like 'subject', 'object', 'verb(al)', 'genitive' and the like. (See especially Coward and Ellis (1977) for this use of 'subject'.) ))
The arbitrary in signification can always have these techniques (which are self-understood as 'secondary') playing over and upon it, in order to rescue it for certain orders which are its very own effects.
((Moore finding so much in 'I know this is my hand' and Halliday finding so much in 'The stick grew shorter at both ends' (where Lok is watching a bow being fired at him) — both of them, I want to say, are making the same mistake. Finding a relation between what is known and what is uttered and finding a relation between social structure and linguistic structure are the same error. To find a relation, they have by definition to be theorised as, at some moment, separate entities (plus a means of connection). To find a relation, their inseparability has to be denied, violated. But on what grounds? What both could not tolerate would be the notion that 'social structure' or 'knowledge', in the form of 'social structural effects' (like class) or 'knowledge effects' (like innateness) are effects of discourse, where discourse is not describable 'in advance' by the categories of common sense, sociology, descriptive linguistics and so forth.))
The technique of matching is almost ubiquitous in social science of: the grammatical with the social, the technique with the object, the idea with the substance and so on. And in every case the measure — the thing against which one matches — appears like a tablet of stone. Still we go on — trying to refine the same techniques when others can be sought which do not require a 'correspondence' (in the widest possible sense).
((I cannot forget the plea of Aboriginal lawyer, Pat O'Shane, at a joint conference of Australian historians, anthropologists and sociologists. She asked that, above all, these social sciences end their schemes of 'classification, de-classification and re-classification'. Here it looks as if we just do not know how not to measure.))
565-566
What is forgotten when, in the face of the practice of language-game #2, we say 'the child knows that that's called "a slab"'? We forget precisely this: that the step with 'I know...' or 'the child knows...' is another technique (of the language-game) — an extension (erweitert) of it — a supplementary move. And somehow, we think, that supplement must have been possible all along and not just tacked on afterwards. We think: in using that name, the child must have known it was called.... So this knowledge must have been in place. Here: we mistake the operation of expanding or supplementing (erweitern) for the operation of proving something — the knowledge — to be (erweisen).
The business of knowing is not a privileged operation which — by virtue of its own constitution — must cover all (or be in place during all) operations. That a construct like this (ubiquitous 'knowledge') operates in various discourses only gives us information about those discourses and their self-understandings as general theories.
((Berger and Luckmann (1967): 'whatever passes as knowledge' — for there to be a world where such passing goes on, sociology already (prior to that world) has to have set in place a very definite concept of knowledge. This is not a move — even if it sounds like one — in which the categorisations of 'the practitioners' are merely allowed to shine through, unmediatedly. Those moves are always constrained (by the discourses of sociology of knowledge) to be knowledges. And that is certainly a constraint.))
Wittgenstein's objection in 565-566 is that we tend — wrongly — to think of knowledge as an omnipresent centre of discursive practice. The reason why we then experience difficulty in saying anything about that knowledge is that the concept itself — 'knowledge' — is just part of this supplementary technique for glossing centres. Our concept of 'knowledge', used in philosophy, constructs the centre, and yet we somehow manage to think that it should be describable, that (metaphorically) we should be able to 'have knowledge' of it.
Accordingly, epistemology can become a maze of infinite regression, so that we wonder how to proceed in the labyrinth. Instead, it may be better to start without a concept of ubiquitous knowledge. Then, the philosopher cannot even begin by considering his task in terms of knowledge of the knowledge of, for example, what something is called.
Whatever the philosopher's task is, or becomes, it cannot involve superordinate ordering.
567-568
The question of 567 is wrongly put because the use of my name (for example by myself) or my dealings with boiling water that, quite routinely, get glossed by 'I know...' or other attributions of 'knowledge' do not mean that I have this or that or the other kind of knowledge (somewhere, in my head, perhaps?). The question, we might say, connects up two techniques of the language-game wrongly and puts us on a path where we end up by describing an essence — knowledge.
((The 'theory of learning' asks itself an analogously peculiar question: how does the knowledge get in place? From our language-game we set up all sorts of imaginary concrete worlds that the game might gloss, if it happened to work in a certain way — for example, literally. It is as if our language-game were a conversation and philosophy and other sorts of inquiry were the conversation's 'touched off' topics — those which arise out of an incidental word occurring in the course of talk.))
((A student once said: it seems to me as if, to do philosophy, you take a word and talk about it for a while and then you stop. Then you take another word and talk about it and so on.))
The verb 'to know' must have preceded the noun 'knowledge'. And this might be a metaphor for the procedure of practice. For using the name must precede using a knowledge-ascription ('He knows the name').
One mistake — like that made in (or as) sociolinguistics — is to take this precedence of practice as an index of just another object for empirical description. Yet it does not point to a shift of objects but to the absence of objects of description.
That the ascription of my 'knowing the name' comes only after my using it time and again does not mean that we have to move from a description of the knowledge to a description of the use. This is only partly clear in Telling How Texts Talk (McHoul 1982).
569
Neither 'inner' nor 'outer' experience (Erlebnis/Erfahrung) secures (our) knowledge. Knowledge has nothing to do with the fact that we are experiencing subjects.
570
I am called (ich heisse) many things other than my name, and some of them I don't know (568) in any sense. I remain untrained in those applications.
571
In English there is a routine response to a doubt — 'It is true or my name's not...'. If it then turns out that this is not my name, it is not just me who is not believed (assuming I'm not being deliberately deceitful). Something would be 'wrong' with the whole technique of persons being called such-and-such, of their having names. And what is it for that to be wrong? Utterly wrong?
To find something wrong with a technique would have to involve having access to something other than its repetition. For instance, it might involve having a list of the technique's features and pointing to an inconsistency among them. Or it might involve showing the technique to be logically incompatible with others that routinely surround it. These sorts of steps point beyond repetition. To what?
And aren't peculiar steps like this involved in all attempts to carry out, for instance, an empirical investigation of our naming techniques (Sacks and Schegloff 1979; Sacks 1972)?
We like to think of 'indifferent' descriptions and overtly political critiques as distinct academic exercises. Hence: the fact/value distinction and its variants. But an 'indifferent' (Althusser: 'innocent') description of our naming techniques is just as bizarre as a critique of them — perhaps more so.
There are some things we do not step beyond (in the form of either a critical or a descriptivist transcendence) because they 'orient', one might say, 'all our steps'. If these are gone — or not part and parcel of our analysis — then we have stepped precisely nowhere. And by 'gone' I mean absent as techniques we are carrying through — repeating.
((Earlier I might have said 'these are grounds'. Now there simply seems no need to say this at all. Rules or grounds (which hold identical positions in social science and philosophical discourses respectively), if taken to be transcendent, yet describable, centres of practice or 'knowledge', are chimerical. In another sense, of course, 'rule' might just be a stand-in for 'repetition' itself and perhaps (then) confused with some 'basis of repetition' or 'essence of repetition'. (Cf. Garfinkel and Sacks (1970) on 'reproducibility'.) But by this stage the line of practice has already been clearly drawn.))
572
'Knowing I can't be wrong' means: we have this particular technique of naming and not some other kind; for example, the allotment of different sized and coloured tattoos on the forehead.
It is possible for people to be wrong in applying of the technique of naming to themselves. And this has its effect on other applications. For instance, it might set in motion techniques of (re)training or incarceration. These things are 'naturally' tied up; for no one could neatly separate failures of discursive technique from failures of legal or customary obligation. To do so would entail a neat cleft between logos and nomos themselves. But how — we want to ask — could laws be given outside language and how could language be lawless? Here it's not a question of specifying the exact relation between essences (law and language) but of noting the identity of certain practices with themselves.
There seems to be no move beyond 'This is how things go amongst us'.
573-575
'Something running against the evidence for my name being...': this is the event (we might say) that would test our knowledge. And Moore's account tacitly rests on such an event. For Moore, there could be, even if it is only the barest possibility, something to indicate the contrary of my name being.... This is how he is able to give these matters the title 'certain knowledge'. Naturally, this event is not one which our practices could sustain — perhaps even in imagination. The test cannot arise. The difference (my name being other) which would make a declaration of knowledge meaningful is itself eternally absent.
So this is no epistemic centre; for it is not a kind of knowledge at all. The 'I know...' must be other than an avowal or ascription of knowledge. For instance (575) it might indicate reliability — but for the assurance to be called for appears strange.
However and wherever this 'I know...' happens to be used, it is the use that is in question; not something behind the utterance. '[T]he usefulness of this sign must emerge from experience' — this points to the utterance coming first and finding 'what was said' coming later.
The discourse (logos) produces the 'use' of the 'I know...' and, afterwards, we find out what that use was. It is not that the 'use of the discourse' is something we — as prediscursively constituted subjects — have under our control. Rather, following upon the repetition of the discourse, we find out how it was we were produced in it.
It is Moore's preference to find out from the repetition something about our 'knowledge'.
((Peculiarly, Moore would attribute such a preference to 'common sense'. But surely common sense, while it might include most genuine empirical propositions as items of 'knowledge', would not want to honour such a platitude as 'I know my name is...' with that title. Even empirical propositions are expected to have a definite 'newsworthiness' to them before they can be thought of as 'knowledge'. For common sense: what is categorically not knowledge is what everyone knows. There has to be a not-knowing before knowledge can be in place. In this respect knowledge-signs are not particularly privileged in any respect. They are subject to the general (Saussurian) requirement of oppositionality. To know and to not know have to go together. It is in the essentialising of 'knowledge' that Moore privileges it above other signs. And that is a move of traditional epistemology — not of common sense.))
((An instance: that someone might well not know my name is not only feasible but a routine occurrence. Yet: that I might not know it is absurd — that is how the language-game goes with names. Nevertheless, my name being... is the proposition in question in these two different techniques of the language-game. And so: they can be confused — run together to produce, for example, jokes:
Mulla: Did you see me come into your shop?Craftsman: Yes
Mulla: Have you ever seen me before?
Craftsman: No
Mulla: Then how did you know it is me?
(Shah 1975, p207)
In this case (which is a little different from the case with personal names), one simply says 'That's how the personal pronoun gets used. There was no question of my "knowing" or "not knowing" here at all'.)) —
576-578
What is actually in question here is a training. And one can — though barely — imagine a training so deprived that such basic matters were missing from it ('Teach me!' — 578). But if that were missing, then so would any training that could lead to the moment of questioning my name being.... The steps along the way would have to be missing for the exact step in question to be missing. ('Only the fourth storey is missing' — of a ten story building.) One untrained in the language-game with personal names would not be trained in the questioning of personal names.
And the use is the repetition of the technique in which one is trained. It is absurd to ask, 'How do I know I am not mistaken about that?' (576), because it is not a question of being correct or making mistakes but only a question of this: was one trained in the technique or not?
((To repeat: I am not arguing for anti-cognitivism, nor behaviourism, for example, but for anti-epistemology.))
If a basic discursive technique like person-naming is not in place, there has been a very peculiar lapse indeed. One might almost say that there has been a slip in the discourse itself — some unevenness in its reproduction of itself — a jerkiness in its self-inscription, a slip of the pen.
((It is here that social science goes searching for its wild children — to see what happens when bizarre circumstances mean that a specific individual is missing a very basic technique. One then happily forgets that it is just another technique which sends us scurrying for basics, for 'absolutely certain truths' which we assume to be hidden in the confines of some personal subject or other — as if difference itself might almost become visible! 'The influence of "the social" on "the subject"' is not a legitimate phrase (énonciation) within sociological discourse and it does not refer to a topic (énoncé) for sociological investigation.))
The absence of a 'basic knowledge' from the practice of person-naming is paralleled only by the absence of some founding practice more basic than it. That is, it is not as if what we gloss by 'knowing the name' had as its sub-technique 'refusing to entertain arguments to the contrary' (577). This gloss does not gloss a practice at all in this case — though it may do so in others, perfectly innocuously. Our practices no more stand upon refusal of their 'opposites' than upon knowledge of their definiteness. None of that matters — it only points to a supplementary practice; to something we might do as well.
((The ambiguity of 'as well' is significant here — something may do as well as another thing, and thereby either (a) replace it or (b) add to it. 'Supplement'/'supplant'.))
The 'higher authority' (578) might be a 'higher author'; for it would somehow have to be able to hierarchise practices which we (everyone!) simply repeat. That is in the nature of 'knowing the truth'. This author's particular (and hidden) supplement would indeed have to be a fundamental or originary text undergirding our practice.
'But then my eyes would have to be opened'. This shows quite clearly that the path of investigation which requires a 'higher authority' who 'truly knows' leads us nowhere other than into mystification.
579
So here is a use of 'certainty' (Sicherheit, 'safety') which is innocuous in respect of 'centring devices' and 'grounding principles' and the other paraphernalia of general theory. One uses the language-game with people's names (with the names one 'knows') with safety — the greatest possible safety.
But this is not to remark upon a general feature of discourse, nor upon 'inner states'. This 'safety' indexes only the irreducible routineness of practices. It points to their iteration.
Contrary to some versions of social science (Mehan and Wood 1975), the fact that discursive practices are iterated, rather than founded, does not make them 'fragile'. It is only a traditional metaphysical notion of grounding which makes anything else (anything 'less' in the common idiom of metaphysics) look a poor relation; a relation moreover who moves in and declares the foundations overthrown — just like that!
Iteration by no means indicates fragility or infinite license. If critique sets out to undermine the foundation of (social) practice — if this is its intervention — then there is no move for it to make. The task may be more difficult still; for the iterative may be far less fragile than the founded: a more difficult target. For critique will — it seems from here — not be able to do anything other than serve the iteration it would ideally alter, or obliterate. The critique, always 'condemned' to the sphere of practice, will have its work cut out even making a critical intervention into practice — a critical iteration. This displays the immutability — the definite non-fragility - of everyday practice.
Relativism does not obliterate 'certainty' with a sweep of the pen! That would be to give it too much power. But neither does it necessarily speak on behalf of a an apolitical idealism. The mixed fortunes of relativism — its chronic under- and over-valorisation in the political sphere — has meant that it has never been seriously explored for its value in critical practice.
'Certainty' (in its various forms) becomes a matter of technique. As such, the reassuring fiction, some as-yet-undiscovered but fixed centre to all our practice, moves off the scene. And with this move come and go different versions of what might pass as critique. And if nothing else, it is possible to see how a new project for the social sciences is (whatever it is) co-terminous with a project for social critique.
580-581
'I know...' might be wrong in at least two ways: either because I do not, whatever I may think, or because I do not and I am trying to deceive. In some cases, only the first use is possible; in others, only the second; in still others, both.
With the Moore propositions, however, neither could seem to be the case. (Imagine the case of being deceitful about 'I know that that is a tree'! Or, indeed, of someone finding that, after all, they had simply been wrong all along!)
Moore's 'I know...' appears to excludes being 'wrong' (falsch). In other words, his propositions with 'I know...' have no routine use in any of the various techniques of false-speaking (falsches Spiel?). They can enter into various kinds of practice but not into sharp practice.
Someone not being able to help themselves and repeating 'I know...' over and again even though wrongly: well this might happen — but it would be bizarre if it were one of Moore's 'I know...' utterances specifically, and not, for instance, a genuine empirical proposition.
The child learns 'I know...' as a technique in the language-game — and one such technique is with a Moore proposition. It arises when, for some odd, perhaps philosophical, reason one gets challenged about a matter of conventional certainty.
Is there actually a difference between someone not being able to help themselves from wrongly using 'I know...' (if we can imagine it) and someone not being able to help themselves from using it aright? In the first case, we should say they were under the compulsion of an iterative rule which is utterly foreign to us. In the second case, they are simply under the compulsion of a rule. When something really bizarre happens (like imagining someone saying 'I know that my name is...' wrongly) — we immediately want to say: here is someone who is simply beyond the rule, beyond all rule(s).
((We might say: this subject is spoken or written by another discourse.))
((Is there a sociology of deviance or only a non-sociology of deviants? Think also of the myths of the genius creator and the madman — the 'wound and the bow' motif in literary criticism.))
For empiricism, the subject looks like a unique collection of particular marks, if we freeze it in history:
And when some peculiar pattern emerges, empiricism gives up and says: this subject is just like that. It is a law unto itself. What empiricism fails to see is that this 'subject' is actually composed like this:
That is, of flows of discourse (chains of signification) composing a dynamic lattice — a network, a grille, a mesh of forces. Subjects arise at (or as) specific conjunctures of those forces and are in the process of flux.
MacCabe writes, for instance, of 'the crowd of voices that compose us. Voices that bear witness to the incompatible discourses that have traversed our flesh' (1978, p133). Again: '...one is confronted with the problem of understanding the individual as a set of overlapping and contradictory practices which produce a plurality of contradictory subjects' (p153). MacCabe reminds us of 'Benjamin's characterisation of a word and a human life as the same...' (pp155-6). In parallel, Derrida (or Jabes) continually repeats this tale of the word-subject: a 'race born of the book' (Derrida 1978, p64), 'he who writes and is written' (p65). Here 'words choose the poet' (p65) so that 'the art of the writer consists in little by little making words interest themselves in his book' (p65). There is, then, no doubt about the answer to the question 'Do we formulate speech [speech?] or does it fashion us?' (p69).
The 'bizarre' arises where a discursive formation which does not routinely traverse (what we can call) a subject-space happens to be found in that place. In this case, the 'bizarre' and the 'routine' are not structurally separable — they are matters of temporality. Effectively it is time which separates Moore's 'I know...' from its routine uses.
582-583
It looks as Moore's propositions are trying to cover what it is that lies behind the repetition of certain techniques. They appear to arise when the 'underneath' of practice is in question. Sometimes we can see how 'I know...' inserts itself into practice (582); but with the Moore-types it is harder. They want to question practice as if from the outside; to absent themselves from practice. But 'what's beneath practice?' is like 'what's beneath this stone?' where the 'stone' in question is just a visible part of the Earth's mantle. These 'I know'-types abscond from practice; and the only way they can make sense is to reinstate them — hence the translation procedures in OC 583. Here, for example, what looks like a deep 'knowledge' behind a mere surface practice of naming turns out simply to have been a case of naming simpliciter, or perhaps of teaching, imparting the name.
'I know that p' (for the cases in hand) is transformable to 'p' just as much as the 'instruction', 'Let it be known that p' translates as 'that p'. (Compare this with the supplementary '...is true' in LC 50.)
((We could imagine a language identical with English except that it allowed sentences to begin with 'that'. Speakers of this language would use such a form where we use 'I know'. For example, they would use it in official proclamations and in responses to challenges or contradiction and so on. These people, most likely, would not develop metaphysical epistemologies.))
The point of these remarks is to show that 'I know that...' cannot point to the interior of practice. Practice is centreless and what appear to be declarations of knowledge-at-the-centre turn out to be routine 'surface' instances of particular practices. There is no other kind.
Sometimes, 'I know...' signals a separate technique from the simple utterance of the proposition minus 'I know...'. But this is a case of different techniques coming into play and not of the absence of technique. —
584 & 586
And so the translation ofI know + the obvious
into a form of expression we already use (like 'be familiar with') will not always come off. Certain varieties of 'know' have their — relatively unique — place(s). But this arbitrary discursive arrangement need not (indeed must not) lead us to the wrong conclusion about such utterances. For example, to the conclusion that there is a particular insight to be got from them; that they have a unique and privileged position among discursive techniques; that they speak to a 'knowledge' which embraces the totality of discourse.
In all sorts of circumstances, our being acquainted with something (our 'knowing' or 'being able to say' what it is) can be legitimately expressed; and expressed using 'I know...', and 'I know what that is'. Necessarily our acquaintance with things is not expressed for every possible case or on every possible occasion!
So it is not as if 'our acquaintance' were an interior object which persisted always but happened to 'find expression' only now and then, at the command of some entity called 'the occasion'. (Here 'occasion' is as troublesome as 'context' — it seems to allow the language-game to speak beyond itself.) Acquaintance, knowing what something is: an inspection of the bare bones of discourse leads us to imagine analysable objects of this kind. But the techniques of the discourse, in practice (which 'I can say what that is' comes closer to than the alternatives), point away from such the existence of such objects. They point (not to the 'occasionedness' of but) to the 'discursively generated' nature of acquaintance-effects. They point to knowledge as an ability to go on to say something: to repeat a discourse of a certain kind.
We ask: 'What's behind the utterance "I know what that is"?' — and the form of the proposition in (the) question pushes us in the direction of an object — the knowledge, the being acquainted. Then the question becomes 'what is it to know or to be acquainted?' All along this analytic chain we could have stopped and taken the proposition as a move; a move which is re-inserted by the form 'I can say...'; then we should have looked for the other techniques around it. And we might have asked ourselves, if there were still any doubt: is this a 'place' where acquaintance gets avowed? But that would probably be all.
We always seem to want to push our techniques further than they can (naturally?) go — we want to find in them other, hidden, techniques. As though there were a kind of metaphoric relation that could go on between techniques. (When, in fact, metaphor is just another instance of technique.) Finding metaphor where none exists is a kind of paranoia — a pretence that one alone can tell what others are speaking about. And naturally that is specious. Being the only one who can tell is like saying that you alone can follow the rule. And that would be a misapplication of the expression 'follow the rule'. ('When you say "I know...", you are representing your knowledge.' Yes, but only if 'representing my knowledge' is a synonym for saying 'I know...'. And then, of course, we have gone no further at all.)
Sometimes our misapplications hang together quite coherently. Sometimes they congeal into quite solid chunks of metaphysics.
((After Kuhn's metaphysics, social science can rest content with an account of its competing forms of metaphysics; an account which produces them as 'pre-normal science' — awaiting forever the dominant paradigm. 'Pre-normal science' is, of course, itself a centre for the play of metaphysics to move around. Religions can form just as well around continually absent messiahs as around continually present ones. The point would be to remove both or any forms of that constraint. In this respect, the struggle for a method in social science is identical with the struggle for emancipatory discourse.))
585 & 587
I know + the obviousmay well work differently from, or identically with, simply stating the obvious alone. But only a mistaken belief in the possibility of second-order descriptions of knowledge could lead one to take the former as an empirical proposition describing 'my knowledge'. (And the latter presumably as an ontological proposition?)
It is as if one had said, '"I know..." cannot work in such a way as to represent my knowledge' and then received the reply, 'Of course it can, listen, "I know that that's a tree"', with this last proposition uttered very succinctly and deliberately — the way one might talk to foreigners (if a native) or natives (if a foreigner). In this case, one is offered something a bit like an empirical proof where none can exist.
((Likewise, ethnomethodologists might say: 'Well, for all you say, here is a perfectly good naturally-occurring instance where someone reports a conversation they just overheard, listen...'. (Then a tape is played.) 'Now try and tell me there can't be second-order descriptions. Members do it all the time!' Naturally, that's not an empirical demonstration either; at least of the matter in question.))
This is like a certain moment in the debate between Habermas and Gadamer over the concept 'misunderstanding'. For Habermas, consensus or rule is like a form of understanding that can exist and be negated by misunderstanding. There is always the possibility of the absence of the consensus or rule. But Gadamer speaks differently. For him consensus or rule cannot be negated. It undergirds the possibility of any (empirical) understanding being negated. It is like the Wittgensteinian point where doubting comes to an end.
Is it not, in fact, the case that every misunderstanding presupposes a 'deep common accord' (tiefes Einverständnis)? (Gadamer 1980, p131)
Habermas' reply contains several empirical examples of misunderstandings — of 'corruptions' of discourse. And obviously, there is no way that that technique could ever show the absence of Gadamer's tiefes Einverständnis. But consequently:
Is it not, rather, the case that something like a 'supporting consensus' precedes all misunderstanding? We can agree on the answer, which is to be given in the affirmative, but not on how to define this preceding consensus. (Habermas in Bleicher 1980, p203)
Indeed, for Habermas, any description of that consensus is 'the result of pseudo-communication' (Bleicher 1980, p204, stress added). And this, by contrast with the empirical cases, does constitute a reply to Gadamer. It is a demonstration which displays the peculiarity of Gadamer's search for an expression of the consensus of our expressions; and as such, it establishes Gadamer's work on consensus as an instance of pseudo-communication. We cannot agree on how to define the consensus because it would itself always undergird the descriptive practices, the representational techniques, which tried to do so. So: as a project for social science, the tiefes Einverständnis (the reassuring foundation, of whatsoever variety) might as well be abandoned. It must be abandoned. (Here Gadamer's concept begins to look like Moore's 'certainty'.)
((I want to say: here is something so absolutely constraining upon our discourse that social science ought to consider it absent; whereas, of course, 'something so absolutely constraining' is exactly what all but the most relativist discourses in social science have hankered after for so long. I do not want to say: give it up as an object of analysis because it's not there. I want to say: give it up because it's everywhere. Here. It doesn't have social being because it is social being.))
At such a point, it is unlikely that there could be any illusions about the special or second-order nature of 'I know that that's a tree' in respect of 'that is a tree'. And this is also shown by the cases of their practical interchangeability.
588
'In the first sentence a person is mentioned...' (587). And so, according to a certain training we have, we get swept along by the formppn + verb + [proposition]
into a meditation which asks: what is the relation between the person referenced by the personal pronoun (the 'subject') and the proposition? According to the same training, the likely contender is always 'a certain state'. Given a reference to someone's relation to a proposition, a legitimate next move is always a reference to that person's 'knowledge' or 'state'. But that move is not so clearly available for propositions framed with 'that is a...'.
The question of deciding between 'I know that's...' and 'that is a...' is a question of what moves might come next. And one move among many is a reference to inner states. But this does not means that 'I know that's...' implies, glosses or represents some actual inner state. It is not 'about' an inner state even if, in retrospect, we have a tendency to analyse it that way.
In a decision between 'I know that's...' as against 'that is a...': we could not checked our decision by looking at the two propositions themselves. It is specifically in the utterances surrounding them (in 'replies' to them, perhaps) that they differ. And in that difference may lie any difference in 'meaning' they are said to have. (Semantics is always process-semantics.)
We could imagine a language-game where
ppn + verb + [proposition]
led invariably to an interest in the technique it 'indexes'. Then, 'I know that that's a tree' might be replied to by 'Go on then!' And indeed, this is closer to how 'I can say what that is' (586) works. In the same language-game, too, 'I know that that's a zebra' could only be uttered by persons — but they might think that that was because persons are the agents of the techniques and not because they have inner states.
If we think of the notion of 'uttered by persons': according to 588, the zoo notice would not count but, presumably, the text On Certainty itself would. And this is not terribly satisfactory. All sorts of intermediate cases, very like the zoo notice can be collected. For instance, a sampler hanging on a wall which reads:
I KNOW THE LORD MY GOD
IS A JEALOUS GOD
Perhaps instead of 'uttered by persons' or 'uttered by texts', we could say that particular discursive techniques favour or routinely use 'I know...' while others favour 'that is a...'. (Otherwise, we are going to have concepts like 'texts [not] written by persons' and other troublesome matters.) On this picture of things, the site at which the technique worked (the text of a book, or of a conversation) would not necessarily be relevant.
In On Certainty there is lingering 'instrumentalism'. The idea that texts are uttered and controlled by subjects is almost but not quite dispelled; as if subjects had to hand a range of possible instruments which could be selected in order to fulfil pre-given purposes; as if those purposes were given outside language, outside any text, and the texts were then selected to match them; as if the texts were 'uttered by' text-autonomous subjects. The trace of a free agent still exists — even in OC — as a possibility beyond language. Some of the character of OC and of Wittgenstein's corpus in general changes if we refuse to think of subjects as uttering texts and come to consider them as being uttered by and through texts. The subject 'becomes' (as) an effect of the interleaving of textual lattices.
So when 'I know...', as it were, is uttered, this means that a particular fragment of subject-formation is in process. It has its meaning by specifically excluding other subject-formations. 'That is a...' can work differently from this. But it is not the presence of an autonomous agent of selection that makes 'I know...' meaningful. If there is a coincidence (in large part) of 'I know...' with a specifically human site of utterance — that is simply, for want of a better expression, an 'empirical' noticing. It is like noticing that
PULL TO STOP TRAIN
PENALTY FOR IMPROPER USE $50
tends to occur at certain sites only.
589-590
Whatever kind of 'state' (Zustand) is in question here, it cannot be a subjective state. For 'there is no subjective sureness that I know something. The certainty is subjective but not the knowledge' (OC 245). For someone to learn to recognise that they know something, they need to be trained to recognise a public event or co-occurrence just like any other in many respects. Not that this technique of recognising that we know something is much repeated in this form of life!
Do we usually speak of 'a state, where what is said is...'? Perhaps it is more usual to talk of the site of an utterance.
And do we recognise our knowing — our production in various epistemological discourses — in any form at all, as a 'state' or otherwise? It might be said that to make this kind of recognition there would have to be a second order of knowing — a knowing that we know.
Knowledge-effects are the products of various discursive techniques — so knowledge of knowledge-effects would have to be the product of a technique of techniques. It is difficult to see how this could come about for, once again, we should have to be in a position to recognise the object, the knowledge-effect, without the technique that generated it. Naturally, in performing such a recognition, we should be simply repeating the very technique in question and so reproducing the very object that would be our supposedly pure topic — our alienated object of recognition.
From here: it seems, in principle, impossible to address the question of one's possession of knowledge — for that knowledge is never held independently for the purposes of description, critique or analysis.
There could be a story which ran: a natural object — a rock — could be collected and described in terms of its composition (= analysis) or its production. But a cultural object, insofar as it has a production, has it equally at the very moment of description or analysis. So to point to the production of cultural objects does not mean that it is the 'production and not the object' that ought to come under empirical scrutiny. To point to the production is merely to indicate the kind of repetition which must be present at the moment of inspection.
It is not as if one had to choose between the knowledge-effect and its technique. The two are always bound together as one. So what would it be to believe that the 'possession of knowledge' could be a separate object for discussion?
591-592
'I know...' in all of these cases works to lay a certain emphasis — either anticipating contradiction or in the face of contradiction. This is the let-out clause for 'I know...'. That clause arises time and again in OC. It is almost as if this 'I know' were transformed into a kind of punctuation — say a mark of stress like italicisation or raising the voice. And: one which might be left out or substituted (592). So, even where Wittgenstein gives it its 'natural' place, there is an attempt to efface the 'I know' wherever possible. Cf. the translation procedures discussed in OC 584-6 and at other points.
After Moore's extravagant claims, the 'I know' has become almost unrespectable, so that OC's project develops into seeking ways around it.
((This is almost analogous to Nietzsche's attack on Descartes' 'I think'.))
And this is why it is necessary to criticise notions like the 'possession of knowledge' (590): for Wittgenstein's attempts to efface the 'I know' in various ways (translation to 'I can say', substitution by punctuation, and so on) indicate a position where second-order knowledge-descriptions become impossible, whatever form they might take.
((This, of course, does not mean that 'I know...' has to be erased from usage; it just means that we should not get too enthusiastic about it as a means of access to 'absolutely certain knowledge' or any of its variants.))
But the contradiction of this text is that the question of knowledge (its grounds and the means of their description) is still one that the text asks. It interrogates 'knowledge' at the same time as that very interrogation (along with the absence of 'answers') always displays the impossibility of this way of proceeding. What the text says or asks is 'What are the grounds of knowledge?' and what it shows is that epistemological questions cannot be asked or answered in any satisfactory way. Of course, it is possible that this failure can only be shown by instantiation. That would be perfectly consistent with the primacy the text allocates to practice and repetition.
((Analogously: one might show a pupil how a musical theme could not be inverted and played alongside itself harmoniously — where that involved the playing of a particularly bad piece of music.))
If this is so, it would not be the first time that Wittgenstein had displayed the futility of particular moves in philosophy by making them in a certain way. This is one of a number of coincidences between OC and the Tractatus, consistent with the retention of a form of the saying/showing doctrine. Eventually, then, the second-order work with 'I know...' in OC has to be seen for what it is and discarded.
On Certainty may be an instance of the critical repetition that it has led us to seek as a method in the social and human sciences, replacing the specification of the rules of discursive repetition.
What kind of iteration of Moore's discourse is OC? While 'Proof of an External World' and so on have their successes, OC fails by the same standard. It does not stand conclusively counter to Moore's 'I know'. It reworks the 'I know' to the point where it begins to fail; the point where the contradictions which generate that discourse become visible. OC's success lies in that 'failure'.
593
The game with negation is not the formal inversion of the game without it. 'I don't know' does not stand to 'It is not' as 'I know' stands to 'It is'.
((And of course it is well known that the presence of formal-grammatical negatives does not necessarily act as inversion in everyday language. 'You're going to Brisbane' and 'You're not going to Brisbane' can act identically as interrogatives. If the former is interrogative and the latter declarative, they can act as identical (negative) instructions and so forth.))
Is 'I don't know...' a confession? Sometimes it will act that way. But does it give access to a hidden realm of (absent) knowledge? — in a way that 'I know' apparently does not? That is: does it formally negate what some philosophers have mistakenly considered 'I know...' to assert?
Instead of this, perhaps we ought not to be pushed by the non-substitutability of 'I don't know...' and 'It is not...' into the single alternative paraphrased above. Instead, perhaps 'I don't know...' could simply lead us to investigate some other techniques — for instance that of confession (cf. Foucault 1981).
In OC, the 'I know' forms can usually be completed by the addition of highly nominal forms: '...that that's a tree' and so forth. And so it becomes a form of revelation to connect these with techniques or practices. The routine mistakes we make about language when its techniques involve nominal forms should be well-known. And they are relatively well known in social science, often leading to substitutions. Investigations of 'the police' become investigations of 'policing'; 'myth' becomes 'mythic production', and so forth. All sorts of substitutions have arisen at the theoretic and analytic stages of social science practice: 'Reality' -> 'reality work'; 'Language' -> 'utterance'; 'Truth' -> 'truth claiming'. In these cases, we try to eliminate an investigable object and replace it by an equally investigable practice.
But in so doing, we are repeating exactly the same mistake about language. The nominal form no more stands for an object than the verbal form stands for a practice. Both are (parts of) practices in their own right. Social science's mistake about 'ordinary language philosophy' is to think that it indicates a new method — the investigation of practices (and their description as 'verbals'). But what it actually leads to, at least in part, is a cancellation of all forms of second-order description. What it says about 'I know that that's a tree' also goes for 'I know how to exchange turns at talk' or 'I know how to read texts'.
The 'how' is a false passageway to modified and legitimate second-order descriptions (of social practices as opposed to social objects). We think: because we can't refer to objects, we can refer instead to procedure. (Sometimes we may model this switch on recent developments in physics and its conception of matter.) And we do not realise that procedure has just become another object for description — absolutely along with the rest. Methodologically, nothing has changed.
So when we say that 'I don't know...' might lead us, not to suppositions about a hidden realm of knowledge, but to investigations of techniques like the confession — this must always mean that it will lead us to the repetition of techniques (like confession).
The question of knowledge cannot be addressed either by mystical repositories (internal or otherwise) or by second-order descriptions of practice. It can only be addressed by practice itself. In fact, both of the alternatives constitute practices but not ones which 'address' (and this word is misleading - 'redress' might be better) the 'problem of knowledge'.
'I don't know...' has its particular uses in that practice (it crops up from time to time).
594-595
Couldn't we imagine a world in which language-games were such that (the equivalent of) 'My name is not...' followed by one's name would be just ungrammatical? — as ungrammatical as 'Name not my ... is'?
It is almost an arbitrary matter that the massively obvious can be 'disputed' among us. Or, to put it another way, that the form of the dispute (disputation techniques) can be carried on in respect of it. (But what is 'in respect of' here?)
Of course, the sceptic has the field to himself here — for the routinely obvious cannot be well defended in disputes, by the usual techniques of proof and mustering evidence. So many hollow victories can be won in arguments about that really being a tree or such and such really being my name. For 'making connexions with innumerable things' which make (?) these matters certain is not a routine method of argument. (It is literally circumstantial.) All it can do is show that a number of things hang together — are trapped by the same lines of force. Co-occurrence can often be thrown out of court as merely circumstantial.
This is how objections like that in 595 can be made. The form of the proposition, the form of this dispute, conspire (one might say) in the production of such monsters. Our natural language (like our number system), if we work hard enough on it, can throw up all sorts of 'anomalies'. Hofstadter (1980) provides a veritable catalogue of them. And this is how it is possible to speak in contradiction to the wor(l)d one speaks (in).
Utterance — person — world: this is not a hierarchy of levels with 'effects' running up and down it. (The heads are pointedly missing from the arrows.) A contradiction in one is a contradiction in the other. Where something is 'out of place' (like 'My name is not...') — it does not mean that one of the 'levels' is out of kilter with some other — it means that one is absurdly proposing that everything is other than we have taken (?) it to be — that every practice isn't like it is but otherwise.
((And that is how 'in contradiction to the world' should be taken.))
Once the empiricism which initially separates utterance, person and world (and subsequently attempts to rejoin them by finding empirical connections) is scandalous, so is the scepticism which plays upon 'disparities' between the 'levels'.
We want to say: 'Your name is not...? How could you be that uncertain?' As if there were a scale, a matter of degrees here. Instead we should have to say 'Your name is not...? Well, that's just beyond the bounds of possible uncertainties.'
596
And the reason this sequence:- My name is NN
- Can you be mistaken?
- Noor this one:
- My name is NN
- Can you be mistaken?
- Yescan have meaning is that they can (albeit barely, especially the former) be moves in the language-game. Yes: moves — but moves analogous to sacrificing one's queen at the first available opportunity without an exchange. The sense lies in this being a possible practice. Here the game of chess reaches its limits as a sensible game. The strategy of 'I'm unsure whether or not my name is NN', however, like that of queen-sacrifice leaves what we might call 'the rules' intact. These instances do not show us the rules any more than 'What do you mean, you had a flat tyre?' (Garfinkel 1967, p42) does.
((All Garfinkel's breach studies do is to elaborate (repeat) extraordinary strategies — they do not give us analytic access to 'rules'. We cannot be sufficiently disturbed into anomie, the absence of rule — no matter how vulgar the attempts to negate regular strategies.))
The queen-sacrifice, we want to say, is purposeless, given the ways that chess operates. (Do the rules stipulate that the players must try to beat each other?)
597
'I don't think so' could just be a 'polite' substitute for 'No (but I do think you must be very strange to ask)'. It could also be a reaction to a picture which the second turn ('Can you be mistaken?') produces. Namely this one: that when some normally quite routine matter is questioned, the question must be pointing to a describable state of things. The picture could be modelled on one from natural science. For instance: the question 'why doesn't the moon just drift off into space?' (to give a very simplistic example) could be replied to by a whole gravitational physics. (But the questions must come to an end somewhere.) And we think, perhaps, that to question a matter like my name being such-and-such would lead to a whole social 'physics'.
This is how scepticism and positivism are intertwined. 'I don't think so' is the answer which opens the possibility of a positivism. It asks to be completed by '...but go on, see if you could think how it might be that my name was not...'. Then there might be a theory about apparent and real names or some other nonsense.
In sociology, 'Is this how they are acting?' routinely gets the reply, 'No it's only how it appears; what they're actually doing is...'.
((To make charges of idealism against positions which attempt to eliminate general theories is to be able to think appearance and reality as centres which are always already in place (as 'realities' to use another discourse). The way around the charge is to show appearance and reality to be the product of an interrogative method ('Can you be sure...?') which is misplaced. A typical 'front' for that way of working is the 'reasonable', 'open minded' but, above all, 'scientific' expression, 'I think so, but...'.))
((A contradiction in science: science ought to be 'ever open', the questions ought never to be closed — according to a very common conception:
...no subject which claims to be scientific can ever reveal the ultimate truths. The acceptance of ultimate truths would mean that the scientific enterprise had come to an end because scientific inquiry requires, in principle, the scrutiny and questioning of any fact. Science must be open-ended. Open-endedness does not mean, however, that solutions and explanations are no good at all because they do not represent the ultimate truth. (Cuff and Payne 1979, p10)
But also: unless something is treated as closed, unless the questions allowably come to an end somewhere, no science get practised. Both centred (absolute) and decentred (relativist) positions could, then, be positivisms.))
598
Again: the routine thing that happens when a doubt is raised in the matter of what my name is, is that one feels one could conduct some kind of inquiry or lead the doubter through a process of reasoning — beginning with something she will accept and making sure that all the steps meet with her approval — ending in the acceptance of my name being such-and-such. One 'describe[s] the case'. But what could this 'case' (Fall) be? For instance what would there be that the doubter could accept as a first premiss of the argument towards my name being what it is? If, that is, she is not prepared to accept the very thing the description leads us to?
There is certainly, we might say, the 'routine' case where a record has gone wrong; a document missing; an early abandonment by parents followed by a discovery of the certification much later. Perhaps there are other contexts too.
But in cases where there is no such available narrative — what sort could we possibly invent? Is there any narrative which is not a variant upon the simple denial of a mistake? Are we, for instance, any the wiser if someone says 'I am called this because that is how things are with the practice of naming amongst us'?
But Wittgenstein, while he sees these sorts of stories filling in for the statement of denial (and vice versa), wants to retain 'a very important distinction' between them. (At least this is how I read the rather puzzling last sentence of 598.) What could that be?
It might be this: that the existence of a reasoned account treats the previous question ('Can't you be mistaken?') as, in some way, valid. It, as it were, seals approval upon the question as a viable one. It turns the question into a technique of the language-game. (Going, perhaps, too far, it gives the utterance its meaning retrospectively. See the last sentence of 596. Here it is both question and answer together which have meaning.) The denial, on the other hand, and equally in accord with OC 596, gives a sense (Sinn) to the question but it turns it into having been a quite different sort of move from that generated by a reasoned account. It makes it quite plain (it shows without saying) that people do not usually doubt such matters. Perhaps we could say it puts both the speaker and his utterance 'in their place', while the account-type reply opens up a weird (non-)place for the question.
Within the constraint that there must be a reply (which may or may not operate), it is perhaps strategically important in such cases to say immediately, 'No, there could never be a mistake here'.
The sort of account to which the question might lead is identical with social science accounts of why a certain practice is like it is. It is significant here that 599 uses an example from chemistry — because here there are all sorts of acceptable accounts that chemistry teachers use all the time. Science language-games consist of moves of this kind (among others). —
599
But the chemistry teacher's reply to the question 'Can we be sure water boils at 100deg.C?' consists of another set of the propositions of chemistry — to do perhaps with the physical chemistry of thermometers and other instruments. What the inquisitive student gets is the 'connexions with innumerable things which make [the proposition] certain' (594).
The difference between scientific and other discourses is, perhaps, only that in science such recitations are routine, acceptable, a regular constituent of inquiry.
The scientist, by giving the surrounding parts of the discourse, does not give the grounds of the proposition. Modelling itself on science procedure, social science investigation has tended to produce similar accounts at similar points of dispute and questioning. And from time to time — wrongly — the accounts have been given the status of rules or grounds of social action. Philosophy has worked this way too.
((Sometimes we want to pile the propositions of a discourse end on end — so that one rests on another more fundamental than it. At other times, the discursive operations do not get carried out this way. They are, perhaps, more horizontal in their relative arrangement.))
- Are you sure that's your name?
- Is there water in the ocean?
- No: the ocean is water
- And similarly the 'you' you're addressing is the one who bears that nameWe put propositions side-by-side with one another — we are not in a position to cite grounds. The 'misunderstanding of the nature (Wesen) of our language-games' is the expectation that among the repetitions of their techniques will be one which addresses the grounds of the game; a sort of repetition-free repetition.
600
Is there any question of doubting a discourse or is it a matter of whether or not one repeats those techniques and not others? And if one repeats, one has no choice other than to 'trust'. If one does not, then the question of trust cannot be entered into.
((Feyerabend, for instance, wants to include certain techniques under the rubric of 'physics' — that doesn't mean that even he distrusts physics, except in a very loose sense of the verb.))
Some things we have to take without (much) evidence; specifically those which are used in the generation of evidence. This does not mean that we take them on faith — only that that's how these things are carried out here, at the moment.
((Garfinkel exhorts us to make the everyday world 'anthropologically strange'; to make the most routine occurrence problematic by seeing it 'for another first time'. This is asking us to go beyond what cannot be gone beyond. It is a search for an analytic repetition of practice which moves itself outside all (other) repetition. Its difference is the difference. And Garfinkel, too, treats the stability of the social world as a matter of, or analogous to, trust (Garfinkel 1963). Then the analytic question is: what are the grounds for that trust? By this point, we have made too many mistakes to turn back.))
Would the production give us the grounds? Does an account of the production make us rest easy about the question of grounds? Isn't the account of that production ('I have heard, seen and read various things') just another text, another reading, another repetition which might be just as much a candidate for interrogation with respect to grounds? Whatever is, is produced — but that necessarily includes accounts of production.
In all of this we return to (we have never left) the 'level' of the sheer practising of techniques — the practising which must go unaccounted or, alternatively, which is continuously and ceaselessly accounted insofar as it is repeated at all. Garfinkel is right — practices are indeed their own accounts. That is sheer tautology. So what could there be to make social science reports about?
((Am I distrusting social science? No I am mistrusting it. What else could you do with a range of techniques which existed in order to make inscriptions of the grounds of practice?))
601
The propositions of 601 bulk large in the production of this text (OC and my remarks on it) — they do not ground it. They are propositions of the text.
Instead of gazing at the expression or at its accompanying feeling, we are implored instead to 'always think ... of the practice'. In this alternative case, there remains, I believe, the possibility of 'misunderstanding ... the nature of our language-games' (599) perpetrated by the routines of inspecting either the expression or the feeling in the search for meaning. The repetition of the mistake lies in this: 'thinking of the practice' constitutes a variety of second-order operation upon the practice — ideally, at least. (One might as well write 'describing the practice'.) That is, the practice and the meaning are narrated as separate elements — just like the expression and the meaning or the feeling and the meaning. Those separate elements are then to be shown to be connected (again) in various ways.
((Of course this self-understanding cannot be right — 'Thought is a symbolic process. It does not matter a damn where it takes place, provided the symbolic process happens' (LC 25).))
And so we may arrive at a 'Wittgensteinian' methodology for the social sciences which involves a refurbished matching procedure. Instead of 'meaning' and 'practice' we might get variants like 'social meanings' and 'discourse' — with the point of the operation being to find the former embedded in samples of the latter. In this way, 'critical linguistics' establishes its art of interpretation:
Interpretation is the process of recovering the social meanings expressed in discourse by analysing the linguistic structures in the light of their interactional and wider contexts. (Fowler et al 1979, p196)
All this might be directly in accord with 'thinking of the practice' — but it merely repeats the matching procedures of the essentialism and mentalism it would replace. 'Thinking the practice' is a symbolic process — it is a practice; and in this case the practice still hangs around a matching technique. (What is this irrepressible will to match?)
The point to realise is that the (performance of the) practice is the meaning. Hence all matching procedures in social science — all forms of second-order description — must be effaced.
Here is the central contradiction again: OC points out a position where the only viable way of dealing with the question of meaning is via practice (Praxis). But we fall short of realising that method if we construct practice as just another metaphysical concept. What is not opened up, then, is the critical question of whether practice can be described, or whether such descriptions are always already an instance of a certain technique being practised. Yet much else in OC points to just such a position — to the absence or impossibility of meta-descriptions. It points to the absurdity of our 'describing the meaning' independently of the practice and 'describing the practice' without 'incurring' the meaning, in order to see how well matched the independently achieved 'meaning' and 'practice' are.
Yet 'always thinking of the practice' now re-opens the possibility of just such an impossibility.
Along with the meaning/practice operations (as the mode of operation of an empirical social science) goes another supposed Wittgensteinianism, namely that language-games are objectively available tool-boxes from which individual users make their selections. In this version, as we have seen, the autonomous subject (which OC itself jeopardises on more than a single occasion) is alive and well. For instance, Fowler and Kress hold that '..social structure provides the resources, individuals mediate their realisation' (in Fowler et al 1979, p196). The concept of 'resource' is very troublesome here. It indicates an extra-linguistic subject capable of deploying an extra-linguistic rationality in the extra-linguistic act of speaking or writing from among linguistic 'forms'. At the same time, OC puts the cap on the extra-discursive. Such a domain, it argues, could never be (linguistically or discursively) attainable. That is, it could not be attainable at all. Second-order description, a meta-language, the free selection of alternatives and the autonomous subject-of-selection collapse together as a mutually supporting 'house of cards'. One cannot be left standing in the absence of (any of) the others.
The mistake is 'contemplating the expression itself' (logocentrism) or 'contemplating the frame of mind' (humanism). And the mistake lies as much in the 'contemplating' as in the terms which follow it. So 'contemplating the practice' carries forward the mistake to a new domain; the very mistake that OC elsewhere rules out.
If there were a proposition which guided my investigation, it would be this:
There is always the danger of wanting to find an expression's meaning by contemplating the expression itself, and the frame of mind in which one uses it, instead of [by engaging in the practice]. (OC 601)
602
With physics: if this is indeed the discourse one is repeating (in say reading the textbooks of physics — OC 600), then it is hardly a case of saying one believes the discourse or that one knows the discourse to be true. One either repeats the discourse or one does not.
In the repetition of the discourse of physics 'I believe in physics' and 'I know that physics is true' are not moves. They may be moves in some other game. For in physics, the question of truth will have its own means of establishment (for instance a set of physical propositions might be found to be isomorphic with an already existing set of mathematical ones — or vice versa). That would make a nonsense out of declaring physics as a whole to be true or false 'within physics'. It would constitute a category error.
'I know physics is true' might be important in positivistic social science or philosophy. In emulating physics discourse, positivism might be able to show that a set of its propositions were isomorphic with a set of accepted physical ones; accepted because 'I know physics to be true'. (One might say that if pressed.)
And all that has happened here is that one has found some analogies between discursive operations.
((The case with 'belief' probably cannot be treated this way at all however. To reply to a detractor that one (merely) believes in physics would be tantamount to conceding the case. It is difficult to see clearly the use of this expression either within or without the discourse of physics.))
Even if all languages implicitly contained their own meta-languages, this would be evidence of the impossibility of the meta-language. In such a case, it becomes impossible to speak of discourse D1 within discourse D1 and equally impossible to speak of it within discourse D2. Each attempt merely constitutes an additional move within each discourse (as I think Tarski (1941) realised though he is not always read this way). This would be how the notions of repetition and supplementation are intertwined.
Language has no answer to its own question: 'What lies beyond language?'
603
What separates the 'physics' of 602 and the 'ethno-physics' described here? Are they separate discourses? What are the statuses of 'I know...' (or 'I believe...') declarations in respect of this physics?
((And of course it is possible that 602 was also 'about' this other physics.))
I should want to say that the case is exactly the same; to treat the 'I know...' within ethno-physics as identical with the previous mistake. Further: this latter is the mistake Moore makes.
Do members know the propositions of ethno-physics? Well: that is one way of speaking about something which so routinely informs our activities. But this is just a substitution:
x + know => inform + y
It is not the same as accepting 'I know physics to be true' as a move in the game. The collection of 'experiences' catalogued in 603 — which has its analogue in physical 'experiments' with falling bodies plus 'experiments' with air resistance plus... — is not likely to be summarised by that expression. Perhaps it might be 'summarised' by the confidence with which one puts the ink bottle down on the table.
In each case, our investigation leads us to a practice (which is also sometimes an expression).
What it is that I order 'my own activities in accord with' (the rule) is not expressed by 'I know that physics is true' or some variant of this. Of course, I could teach somebody to act this way also. I could train her or him. And the uttering of 'I know...' would not be part of that process either.
604-605
When ordeals by fire were used in reaching verdicts there was nothing at all absurd about using such a technique. That was the technique.
((Relying on a particular boiling point may not be quite the same. The statement of a physicist (or anyone else for that matter) would be the sole technique for reaching a verdict. Its very self-evidence would make it a step along the way to a verdict — one technique among many. It might be part of an exchange of evidence — a trial. And it is this form of trial — and not the statement alone — which replaces the ordeal. In the ordeal (and this is what makes it so gruesome for us) the utterance 'Witches do not burn' is never questioned because it is never raised — there is no exchange for it to be raised in.))
Can we imagine another society outraged by our practice of sending people to prison on the basis of a measurement like a boiling point? It is these 'what ifs' that show how locked-in to certain practices we are. But the 'what ifs' (and perhaps, too, some of the empirical findings of anthropology) seem to assume that we can step out of those very practices. They expect something which they show to be impossible. They have it that we are separate and separable from those practices whereas our very being is intimately connected with their repetition.
Only someone other than us could treat the propositions of our (ethno_) physics otherwise. Only someone other than us could 'treat' them at all. And what would be the practices which would produce that someone? I cannot imagine.
((Of course, the sceptic imagines that he himself is this someone. And (how) does that come about?))
606
One of my judgments or actions is the judgment that 'to my mind someone else has been wrong'. So how could such a judgment, as it were, threaten to topple all others?
To judge something is to repeat a certain technique in respect of it. To be unsure in one's judgment is another or, perhaps, a different form of that same repetition.
Again: we always seem to want to find a hierarchy — a set of levels which one can move down through and so get a 'deeper' explanation — perhaps even reach an ultimate level of explanation.
((The Tractatus is a seminal work in this connection. But many other texts work in an analogously atomistic way. Yet they do not see the folly of that.))
Exactly the same trouble emanates from holistic endeavours at explanation as from reductionist forms: both discursive formations produce (often indistinguishable) hierarchisations. Their practice, we should say, generates those levels and their relations. (Sometimes 'practice' or 'social practice' is one of those levels which both feeds and feeds off others.)
An instance of this notion that our judgments and actions may always rest on some more deeply embedded level follows, albeit an absurd one:
...there is almost no leakage from one level to a distant level. That is why people can have intuitive understandings of other people without necessarily understanding the quark model, the structure of nuclei, the nature of electron orbits, the chemical bond, the structure of proteins, the organelles in a cell, the methods of intercellular communication, the physiology of the various organs of the human body, or the complex interactions among organs. All that person needs is a chunked model of how the highest level acts; and as we all know, such models are very realistic and successful. (Hofstadter 1980, pp305-6)
Naturally, the measure of their success will be how they stand up to 'scientific testing' against 'real' models thrown up by the previous levels! This is the consequence of taking (note!) personal action as but the highest level of a number of others (no matter how autonomous). A similar error, this time in thinking with analogies between 'levels': 'For millennia Nature has slowly been fashioning first our minds, and then, within them, an imitation of her workings that we have come to call the laws of Nature' (Barrow 1988, p1).
Even if we do not always think of a certain physics (exactly) underlying the most important concepts of the social sciences (like 'action', 'behaviour' and so forth) — one which exists but which gets exponentially more complex at each successive level and therefore is to remain untouched — we nevertheless want to continually probe them as levels and to explain things in terms of their relation to other levels (above or below — for there is a determinism in both strategies).
Perhaps it is just this: social science cannot handle the question of practice alone as a de-hierarchised, separated 'level' of phenomena. (For example, Raymond Williams' (1973) attempt to remove the hierarchy of base and superstructure in marxist theory leads to a retention and modification of it.) Here, even the word 'level' is quite wrong. Even the cautious quotation marks drag back the ups and downs of social scientific analysis.
Social conduct, it seems, has to rest upon the unconscious, the economy, and so forth; it has to emerge from specific contexts or stem the gap between culture and nature; it has to.... And the list could continue, because the only way of treating practice is as (and in) the repetition of certain practices. As soon as explanation (or, indeed, understanding) is called for, so is an analytic meta-language and the rest. Then social science is in familiar territory once more. Social science has to be able to break out of this geological metaphor — one which creates the separation between theory and practice (autonomous levels with 'almost no leakage from one level to a distant level' (Hofstadter 1980, p305)) — a separation which has subsequently, as part of the very task of social science, to be re-welded.
The mistake is identical with the mistake in asking for an answer to the question of what grounds my judgment. Social science makes that mistake so routinely — it is virtually part of its Weltanschauung at this time. But that mistake 'is no ground for any unsureness in my judgment, or my actions' (OC 606). For social science as it stands, it is not a mistake at all. That such a layered method is in place may be, in fact, one of the symptoms that social science uses in diagnoses of its own sureness about many matters.
Where something (like hierarchisation) reaches this almost universal level (so to speak), it becomes unable to express difference. Its own meaning is jeopardised. It becomes indifferent, to use Barthes' pun (1981, p119). Many social science techniques involve attributing significance only where activity on one level has its effect, isomorphism or analogue on another. This universal difference positions those social science techniques as 'outside the levels' and so as meaningless in their own terms.
The movement for disciplinary change comes, perhaps, not when there is a scandal of such proportions that it cannot be covered up, but when there is just staleness or saturation. For example, the constancy of French structural linguistics from Saussure to theories of the 'materiality of the signifier'.
((Are we addressing a definite change or just a change of style again? Is 'practice' not just becoming a new bottom point? Is it not (in a different metaphor) just another centre?))
607
We want to say that what is beyond doubt just stands; it does not only stand when it stands beside a statement saying that it is indeed beyond doubt. Moore's propositions do not prevent these things from standing, nor do they 'make us realise' that they have stood all along unknown to us, nor, that is, do they facilitate their standing. These matters just hold for us.
((And they hold so that we can, for example, imagine something; one case of this being where we imagine that something else holds in their place. Science fiction texts 'about' other worlds are products of the textual practices of this one.))
Moore's propositions are like the labels one sees in some homes which have pre-school children living there. Pinned to the door is the label
DOOR
and so forth. His propositions might (we could think) be used in training one who did not know the slightest thing about us. But the propositions themselves would also be matters of mystery for such a person.
((The anthropologist, having read his grammar of the tribal language, makes it known to the chief that he now needs to know the fundamental principles of the tribe's life. The chief says something completely strange to the researcher, like 'The light on the river never darkens'. Only after living with the tribe for many years can the anthropologist find out what it was he was being told then.))
Trying to go beyond undoubtables merely standing, Moore tries to give a use to these practically useless propositions. —
608
Now: how can we say that the propositions are 'practically useless' when we are prepared to call them 'good grounds' (if we are prepared to give anything this tag)?
Well, use comes to an end also; practice comes to an end — and to want to express the propositions which give that use its 'significance' (that is, by being distinct from that use) is inevitably useless. It is not part of any discursive practice which we repeat — at least in this form of life.
So what is it to be 'guided in our actions'? Is the guiding, as it were, part of the repertoire of our actions or beyond it? If we are guided — then we think we can produce the guide itself. We have the techniques of social science at our disposal for producing such guides. What is so hard is to see the 'guides' as, perhaps, 'guide-expressions', as literally the products of social science techniques. The guide-expressions, too, are effects of guidance.
((We live in a complex structure for which the plans have been lost — say. And some of our time is devoted to trying to reconstruct the 'original' plans by observing the structure as it now stands. It is like trying to reconstruct the recipe from the cake, or the score from the performance. But part of the complex structure — in this case — is a whole history of techniques for drawing up plans themselves. Our plans emerge as part of the structure which they are supposed (autonomously) to represent. And this is not a case of levels or recursion or meta-languages.))
For all this, we are inclined to say, the structure of our practice has its grounds; just as the proposition has its general form. But to give the form devoid of propositional 'content' (to compound the metaphor), is another matter entirely.
How do we know the propositions of physics constitute a good ground? We continually show this in (or as) our practical activity.
The importance of ethnomethodology is that it is fundamentally correct in realising this position for practical action — but it is equally fundamentally mistaken in assuming that this is a new object for social science investigation (and of a basically empirical type). Practice effaces all other possible objects — but it does not re-fill the space left by such an effacement (and so cannot be treated in the same way as the 'centres' it replaces — or shifts). The 'methodological' question is not how to fill that space at all but how to continue with it unfilled and to move on to a new terrain altogether.
609
Here we come upon another 'what if' (see remarks on 604-605) which displays how locked into our everyday action we are (in any form of life). So what is it to (apparently) 'talk about' a form of life and the relations between forms of life? What is it for language-games to combat each other? How could it be that they could have common ground over (or on) which to battle?
What happens here is that one discourse's repetition works on another discourse but remains a repetition of a single discourse. (Just as the model 'enemy' soldiers which a General places on his camp table are part of his military equipment.) The same or a similar thing happens in the other 'camp'. No discourses actually come into conflict or combat — except in this spectral sense. There are simply posturings using self-generated models of the 'other' discourse. (A butterfly's colouring which makes it look like an owl to its prey are part of it being that butterfly in that system of eco-relations.)
From such techniques (which we are calling 'spectral') emerge discourse-specific hierarchies 'of' discourses — of the kind discussed above in the remarks on OC 606 — usually with each discourse either omitting its own self-representation from the stack altogether or else placing it at the top. This is perhaps what is behind the notion of using a language-game as a base to combat another.
What should we want to say then of empirical instances of discursive colonialism — where, for instance, a tribal people's 'religion' (is that what it is?) is displaced by another. Or where someone is trained to accept certain techniques as opposed to others which were previously accepted 'without a word'?
In this case we are making the mistake of confusing the discourse with the sites of its repetition. A clear distinction should be made here. But, I still want to say, discourses do compete for 'the same' sites — various physicses, for example, for the position of dominance in Western science, and so forth. And this would seem to be the proper business of social science.
((And it is often such competitions for what seem to be the same sites that get used as first examples for training novitiate sociologists. There is something characteristic in this, if not exactly something fundamental.))
How — by what practice — might we come to examine instances of discursive combat for a particular site? The only answer would seem to be to, as it were, take arms in the struggle. And that would be the work of social science. It would not be merely preliminary to 'real' analytic work.
((But is it not peculiar that a kind of relativism should lead us to such a position?))
In this hypothetical example, the 'participant observer' might try to 'bracket' his physics and step up along with the others to consult the oracle. But, for us, social science work might involve repeating the oracular discourse in the face of physics — say in its sacred sites like the laboratory, the journal Nature, the secondary school science curriculum and so forth. What the strategic value of repeating it in the annals of philosophy of science is (cf. Feyerabend 1979) is another matter again.
But this is how we might learn something of the relations between a discourse and its routine sites and of the (possible) interfaces between discourses. At least such a move would entail repeating the discourses in question and not indulging in the belief that we can address them unmediatedly.
610-611
How could we judge which of two discourses was right or wrong? This is the same order of trouble as the question of whether we are right or wrong in combatting the 'other's' discourse. And that order of trouble is this: judgments as to what might be right or wrong, of whatever kind, are techniques of the language-game. For such a technique to engender judgments as to the wrongness of other techniques of the same language-game would simply be for that last technique to be eliminated altogether. (And perhaps such a collapse could seriously endanger the entire discourse?)
There seems to be a built-in defence which language-games have — by controlling judgments of right and wrong, they need not fail to judge themselves favourably. And this means that other discourses are 'natural' enemies; they are 'naturally' wrong.
The slogans we have in such cases consist of the discourse's own defence mechanism at work. (This is almost a Darwinian idea.) Where, for example, it is 'axiomatic' that national boundaries are sacred, we get: 'The good of the nation made it necessary that I...'. Where a certain economic order it unquestionable: 'To act otherwise would have been to jeopardise the existence of free-enterprise'. And so forth.
Of course, these things really (wirklich), practically happen. Discourses come into contention and struggle. But even this struggle is not strictly a mutuality for the contending discourses. It is not as if there were some natural discourse to which both 'subscribed' and which stated the grounds of contention. Discourses may be struggling with others which do not 'know' it — in a certain respect. Struggles between discourses always have their practice in struggles within discourses.
((Dr Johnson's remark about two people shouting between houses on a narrow street: 'they are arguing form different premises and they will never agree' leaves open the question: will they ever actually argue? A further relevant anecdote — Mark Olsen at the 1982 conference of the Sociological Association of Australia and New Zealand told of the leader of the New Zealand Boilermaker's Union who was approached by the employers' representative and asked if his union could appoint a lawyer to represent their case in order that negotiations might be speeded up. The union leader replied: 'Well, if you really want to understand our position, why don't I get one of my boilermakers to put your side of the case?' See Olsen (1982).))
All this makes the business of critique quite problematic. In fact, many of the debates 'between' social science paradigms cannot be said to 'degenerate' into, for they always already are, insults, slogans, charges of myopia and the rest. One discourse obviously cannot convince another of the reasonableness of its grounds until the second has already accepted those very grounds in order to take part in the techniques that the first discourse uses for convincing.
There are no good Newtonian grounds for accepting Einsteinian relativity. In this sense, a switch of grounds may be virtually arbitrary. (Who switches? Members? Fields?) This is not a question of which discourse best represents 'real world' phenomena; nor is it a question of the competing interest of collectivities. It is a question of the specificity of the operations of the respective discourses for producing distinct knowledge-effects. In all of this: nothing is being said of 'real' objects, nor of the relations between them and how they are 'known' by various disciplines. That is, there can be a treatment of material operations without ontological or epistemological stipulations.
This is the very form of a decentred social science that these notes have been attempting to indicate. But what it means is that, in the absence of appeals to 'real' universal objects outside, beyond, but common to different discourses, and in the absence of appeals to 'interested' epistemological claims to their most adequate representation, there is no longer a space in which discourses can come up for comparison. Marxism's material operations for the production of 'ideology critiques' and functionalism's material operations for the production of 'legitimations' are not objects of comparison. A material operation is either carried out or it is not.
Between discourses there are only re-trainings. So what is the nature of critique?
612
The critique of a discourse (the combat) is always within the discourse itself. Or else, it must take the form of a training in another — this is persuasion (Überredung) or conversion (Bekehrung).
We think that, because a subject is formed from a number of intertwining discourses (see the remarks above on OC 580-581) and so might — from another perspective — be said to range over that collection of discourses, then discourses themselves have such capacities. We confuse discourses with the subjects they produce.
If the subject can do this — then it can be done. Most certainly: but this does not mean that when a subject is the site of one discourse's repetition, it somehow transfers everything across with it when removing to another discursive domain.
Critique cannot get going by (a) finding a particular discourse at fault for inadequately, incorrectly or partially representing the 'real' objects of the world as they stand independently of discursive operations; (b) finding a discourse at fault by pointing out the interests of its users (as a subject, a class, etc.) and their production of epistemological categories to 'fit' those interests rather than the 'real' objects. (I am thinking of Habermas, but also Althusser.)
This is because the order of discourse provides the material operations for the production of 'objects' and of (interested) 'subjects'.
Critique, on the other hand, might get going by (a) finding strategies for the critical repetition of a certain discourse; repeating it, going through its techniques in such a way that its contradictions are shown; (b) finding strategies for re-training. And these two moves might be identical.
613
Schutz (1962, pp14-15) confidently separates what members of the social world know and what they can find out by knowing which specialist agencies to turn to. (For example, not knowing about carburettors but knowing about car repair shops.) And all through this part of OC (599-613), where science has been in question, a similar distinction has been lurking. (See also Foucault (1977, pp41-43) on sociétés de discours which fall into two types — a relatively esoteric and a relatively publicly diffuse kind.)
'Is it wrong for me to be guided in my actions by the propositions of physics?' (608) — well, surely, the follower of Schutz might say, it is not as if you know these things first off. You yourself say that you only 'believe' you know (600). You are quite dependent upon the agencies you do know first off for this sort of knowledge, the physicists that is. Perhaps that is why you would be 'astonished' by water freezing on the gas stove and hand the matter 'to physicists to judge' (613).
And it is precisely this space between levels of knowledge that Feyerabend (1979) exploits in his defence of witchcraft, astrology and the rest of anti-science. It is as if this 'secondarity' of knowledge meant the possibility of control — of the establishment of a priesthood and a religion. (Cf. Foucault (1977, pp23-28) on the subject of the commentary and the 'décalage entre texte premier et texte second' (p26).)
Wittgenstein opens up the same space by making his personal knowledge (of his old friend so-and-so) a 'still greater certainty' than the matter of water boiling rather than freezing on the gas-ring. All this is, of course, in accord with the use of 'certainty' to cover a personal, subjective relation to knowledge (OC 174, 245, 404) — perhaps it is this: the 'attitude' of the subject to its knowledge.
This, of course (if we accept these distinctions at all), is going to have to mean that someone else might equally be more certain of the water boiling than of the matter of a person being his old friend so-and-so. That is, there is not much sociological interest in the matter.
But working back from this point: we might also want to question the very notion of direct and secondary knowledge and the gap it opens up (to be filled with something like 'an attitude to one's knowledge').
The discourses of physics (I like to think — perhaps wrongly) go on at quite some remove from me. Their operations are only partially determinative of this personal subject. And doubtless I should be much more shocked by a person turning out not to be my old friend than by the freezing water on the stove. But the distance of that discourse from a particular subject seems to make very little difference here. It leaves the practice of physics untouched.
On the other hand, what is this direct or primary knowledge I am supposed to have? (Within 'empiricism', Austin's (1962) treatment of direct and indirect perception is seminal here.) Is there some personal discourse which supplies direct knowledge? Is this how my certainty about my old friend is grounded?
All that is at stake here is: which discourse is repeated and which is not? But to be repeated it must already be in place. If its iteration takes place 'in' a certain subject — that does not mean that this knowledge is more 'direct' than the propositions of physics which do not happen to have this iterative site.
If 'certainty' or 'directness' (and now these seem to boil down to the same thing) is only a matter of something tacked on by the subject after the fact, I should still want to ask: what is the discourse that furnishes the technique — the 'tacking on'? Subjects do not, presumably, know where to go attaching their allegiance to 'greater certainties' and lesser ones because of some private language. Or: are one's 'attitudes to one's knowledge' outside the confines of language and practice?
614
Yes, here it is 'know' — the question of subjective relations to that knowledge has been dropped, and so one may talk of the 'foundation of all judging'.
What I want to move against here is the possible notion that the 'knowing' is something personal. In these paragraphs (especially 606 and 613-614), one might think of knowledge a personal. But where the name which has been used for so long suddenly fails (and so on, for the other examples) this is a public failure and means that a whole network of social relations and discourses would have to have been rent in some important fashion.
It is not just 'from us' that 'the foundation of all judging' would be taken away in these instances. For that foundation is not (just) 'mine'. If anything, I am its effect and not vice versa. '[T]o imagine a language is to imagine a form of life' (PI 19) and when gaping anomalies seem to be occurring in the fabric of that language, it is actually the case that there is another form of life that is coming into being (or being imagined). If it is bizarre 'for me' — that is because how I am being produced is itself changing. Personal uncertainty — one might say — is the effect of my position in a whole network of discursive relations. —
615
The network is indexed by utterances like 'things behave kindly'. Those statements point to a certain stability of the forces, a fixity which can be made to look fragile by any kind of reflection. 'What if the sun did not rise again?'
Certainly there are 'adjustments' that can be and are made here — but they are incursions into our whole form of life. The strategies and techniques for making them would properly be called a micro-politics. That politics would work on our language-games.
To get some new practice off the ground (and so displace another) is so difficult to achieve because it tenses the net locally and reverberates through its entire warp and weft. The rending of a single intersection of strands threatens potential chaos for the whole fabric. —
616
The 'fabric', the 'facts', things 'behaving thus and thus' — these are the stable (but not necessarily benign) vectors of social life, whose movements we feel directly. That fabric's actions upon us are us — just as it consists of our actions upon it. Society is an us/it.
((A fractured Einstein: space tells matter how to move, matter tells space how to warp. The physical universe is one space/matter element.))
And so it might look as if there is a struggle between 'me' and 'the facts' in some given case. 'Me' in the saddle and the 'facts' bucking. But that struggle, that tension, is co-extensive with the social forces which constitute the fabric. Scepticism, for instance, is a position with its own vectors — even though its self-understanding is that it lies outside the confines of social forces. The discourses of Marxism, for another instance, have their locales also, as do those discourses which argue for the critique or the violent overthrow of the entire social fabric. The techniques at these sites are thoroughly familiar. Their positions are known. They are on the map.
((This is not a functionalist, equilibrist or consensus apologetics for the burying of struggles. Rather it begs those struggles to see their loci and speaks to a strategy of a certain kind of minimalism.))
The problem