Ordinary language is all right.
One could divide humanity into two classes:
those who master a metaphor, and those who hold by a formula.
Those with a bent for both are too few, they do not comprise a class.
I like it when people in movies are complimented on their cooking and they say, 'it was nothing, just burned a few eggs'.
When I burn things I try to burn them evenly on all sides, if they have sides.
—Run aground again, on §87. Wittgenstein talks about explanations of what (or who) one means; about 'doubts' about the words used in the explanation ('aber was meinst du…?'). But how are those doubts?
He proposes a somewhat counter-intuitive sequence of terms open to doubt, given that the example is about 'Moses': 'Israelites', 'Egyptians', sure: but then (eventually) 'red', 'dark', 'sweet'? At the very least, the direction there, and the character of the latter terms, suggests a drive toward something like a tractarian way of resolving doubts: if these words (i.e. these words' meanings) are unclear, there will be nothing to do but find the framework in which the most fundamental attributions of reality or unreality can be made, and the meanings of words like these reconstructed.
Were that possible, would it be a removal of (something we could call) doubt? Wittgenstein dismisses skepticism quite late in the Tractatus, actually in the last set of remarks (under 6.5) before the final one, 7. And after the (other?) remarks on ethics (and the world). Skepticism is obvious nonsense (6.51) when it doubts where nothing can be questioned (because nothing can be said); this is a region in which 'unsere Lebensprobleme' live untouched by any possible answers to scientific questions (6.52), and thus show their character as (a curious sort of) disappearing problems (6.521). This little analysis of skepticism is juxtaposed to a second assertion (6.53) about philosophy being something that is done (the first was at 4.112: 'not a doctrine, rather an activity': the logical clarification of thoughts effected by providing elucidations): it describes an unsatisfying, but correct, method of doing philosophy with another person. Or 'to'. And, juxtaposed to this, a rare (lone?) remark (6.54) about the import of the other remarks and the book as a whole (save for the tricky 5.6's), which simultaneously claims a way in which they elucidate while implying that they somehow fail to observe the correct method in philosophy: something which readers will discover 'eventually', 'am Ende'. (Will they then be the ones who, per the preface, are afforded pleasure by understanding the book?)
When, on this picture, is philosophy done? Over with? If skepticism is 'obvious nonsense' ('when…'), what would be the source of a wish to 'say something metaphysical' (like? 'the world exists'? 'there are objects'?)? Not just a failure to understand the logic of our language (4.003)? (Not if 'Alle Sätze unserer Umgangssprache sind tatsächlich, so wie sie sind, logisch vollkommen geordnet' [5.5563]?)
There are two points here I don't understand well at all. Why is the picture one of doing philosophy with, to, for, on, another—who wishes to 'say something metaphysical'? And in what sense is the activity of philosophy completable (for a person), or, say, guaranteed to be (in principle)? (So that the solution promised to the 'Lebensprobleme' is one that can be found?)
—In contrast to the uses to which language-games are put in the early sections of the Investigations, the context of use for many of Wittgenstein's examples drawn from our language (the place where, I think, issues of doubt and removal of doubt are being addressed) is not so often made clear.
The comparison between rules and signposts in §85 can seem unnecessary somehow, perhaps like the succession of comparisons in §§11–12: handles, then tools. Why not just proceed to the language-game? At the very least, the comparison serves to reinforce the sense of a rule 'standing there', or lying there, like a thing, so that in §86 when rules are embodied in things, the most immediate doubt is likely to be something like: how can this thing tell me what to do?
If a rule tells us what to do, and we embody a rule telling us how to apply that rule in a thing which will thus seem to challenge us with the question, how can this thing tell you what to do?, then maybe we are getting to the point of a paradox: if applying a rule is something we do, and we do things by applying rules, then how do we manage to do anything at all??
With table in the left hand and schema in the right, the need to 'do' something with each is made concrete. (And so the sense of impossibility comes in where?)
In §86, the schemata for reading a table different ways are 'introduced', one of them 'attached' to such a table. Nowhere is it said that the training in how to read tables, drawing those left-to-right arrows, has been abolished, or anything of the sort (diversified?). The schemata, rules for rules (specifically, for the table), are not situated within the training–institutionalized rules–praxis scheme in use earlier in the Investigations. But the language suggestive of their being fully operationalized, 'instruments for the use of the language' just as the tables are, exerts a great deal of pressure on the reader to think of the table and its schema as being exactly on the same level, as if the Chart Operator were standing there with a table in his left hand and a schema in his right, the thing in his right hand to be used so as to use the thing in his left hand, with that relationship forcing him to cast the thing in his right hand as if it could not be used without another of the things somehow like one of the ones he's holding, in a third hand.
In his concluding questions Wittgenstein refers to the potential further rules explaining the schema, explaining how to use it. To whom, by whom? With the giving of orders or the instruction in the language far from the scene all these two paragraphs later, no one may come to mind. Nor what would need to be explained. But, following the pattern Wittgenstein often uses when entertaining the need of or use for an explanation, we might say that the introduction of different ways of reading tables itself provides a context in which explanations may be needed, and given—'to prevent misunderstandings'. (And this may be against a context in which pupils were earlier trained in the normal way of reading a table, the way in which they draw a series of arrows with their fingers.) And there, with established ways for reading a table provided, an explanation could consist in, say, pointing to one schema rather than the other. So that attaching a schema to a table is itself an explanation of how the table is to be read. And in the same way, the people giving and in need of explanations can be imagined. The boss, to the new employee. The manager, in light of new table-reading regulations recently passed. The friendly co-worker, correcting an innocent mistake liable to be punished by termination. (And how many of these, if they needed some kind of further explanation, would not quickly be explained by resort to arrow-drawing with the finger?)
The language-game introduced in §86, a variant of the builders' game of §2 in which a table or chart specifying the correspondences between words and stones is used by B in carrying out orders, is the first since §64. It seems to signal the imminent closure of §§65–88, and thus of the work of the whole first part of the Investigations (before its runs of remarks on logic and philosophy).
Before that, the remarks can be divided into sequences clustered around the games explicitly introduced as such, and detailed, in them. In §§2–36, there are five such games: the builders' game of §2 (A calls the names of the several kinds of building stones, and B brings him the stones), its expansion in §8 (with the number, color, and place of stones now playing a part in complex orders from A, who also points and employs samples of colors), the expansion of §8's game in §15 (with names affixed to the tools now used in building), the evidently related game of §21 (in which A requests, and B gives reports about, the number, color, or shape of stones at a certain place), and in §27, 'a language-game in its own right' absent from §2 and §8, asking and explaining (by pointing) what a thing's name is.
In §§37–64, after early references to most of those initial games (to §2 in §37, §8 and §27 in §38, and §15 in §41), the first new game is one devised to fit an account of names and being taken from Plato, as the game in §2 was proposed to fit the passage on names taken from Augustine in §1. In this game (§48) sentences formed from sequences of 'words' 'R', 'G', 'W', 'B' serve to represent combinations of colored squares arranged on a surface in a certain pattern. The next two games each involve orders concerning objects composed of parts, in which each game has a variant in which the objects have names, and a variant in which only their parts have names: there is the initial such game (§60), and a similar one played with a table (§62). The final game (§64) is a variant on the one from §48 in which single names will now signify squares composed of two colored rectangles.
§48's game is important. A number of questions about truth and correspondence and existence, refracted through the criticism of the account from Plato's Theaetetus, this itself a refractory criticism of parts of the Tractatus, are addressed in relation to it. But it's also important for its role in the gradual exposition of ideas proper to the Investigations. The use of a table, or chart ('Tabelle'), in a game is first introduced in §53a–b, and immediately (§53c–§54) employed as a way of understanding roles that rules might be said to have in language-games. (This is just the second point the idea of a rule for a language-game has come up: and in §31, the first, it seems as if the idea appeared in order to entertain the plausibility of denying the need for it: 'one can also imagine someone's having learnt the game without ever learning or formulating rules'.)
The introduction of the table in §53 is prompted by questions in §51 about 'correspondence' in the language-game from §48. The questions are posed on the level of the words-to-squares connection (and, interestingly, the first one connects that 'mere' connection to our language, our color words), concerning which a distinction is drawn between the teaching of the language and the practice ('Praxis') of the language (the word 'Praxis' is consistently used as part of this distinction in its few occurrences in the book). And here, ideas about what this connection, this correspondence of names to things (squares), could consist in are characteristically put forward in questions (which are not answered in this section, nor before corrective comments in §51b's closely investigated 'details' and §52's mouse parable): could it be that the person describing complexes always says 'R' when there is a red square, etc.? But what if he says 'R' mistakenly, when it's a black square (and what's the criterion for this being a mistake?)? Or does 'R''s signifying a red square consist in the users of the language always having red squares come before their minds when they use the sign 'R'?
I'm not going to examine it right now, but I think if you read these questions as skeptical (or anti-skeptical), Wittgenstein's suggested answers will seem unsatisfying. Essentially, he shifts the burden over to ideas about how the language is taught, how rules for its correct use might be instituted, or how instruments embodying its correct use might come to be employed as part of the use of the language. Then, in making several quick remarks about 'the kinds of case where we say that a game is played according to a particular rule' (§54), he almost seems to take back these suggestions by including the possibility that for a particular game, no rule is employed in teaching nor in play—one learns the game by watching how others play it—but 'we say that it is played according to such-and-such rules because an observer can read these rules off from the way the game is played', with the distinction between mistakes and correct play shown, says Wittgenstein, by 'characteristic signs of it in the players' behavior', recognizable as such even without knowing the players' language.
(It's instructive to compare that series of suggestions to the note of dissatisfaction voiced in the initial question in §51 about what correspondence consisted in: 'In describing language-game (48), I said that the words "R", "B", etc. corresponded to the colors of the squares.… After all, the explanation in (48) merely set up a connection between those signs and certain words of our language (color names)'. How could that be enough to get doubt going—when it is the initial stipulation in the description of the game that is being doubted?)
The same sort of dynamic plays out, I think, in the introduction of the language-game in §86. This game is like the builders' game in which B fetches stones per A's written orders, except that now a table is used which shows the shapes of the stones. The table is 'a rule which [B] follows in carrying out orders' (cf. §53). Wittgenstein notes, as is customary by this point, the sort of training in this language-game which its 'speakers' will receive: 'one learns to look up the picture in the chart by being trained, and part of this training may well consist in the pupil's learning to pass with his finger horizontally from left to right; and so, as it were, to draw a series of horizontal lines on the chart'.
Then, also as is customary, there are questions posed in light of the game that's been described. Supposing two ways of reading a chart were introduced, and one such way was 'added' to the chart (the translators use 'added' for 'beigefügt': I think 'attached' would be better, imagining a cardboard chart with a photocopied xerox of a 'left to right' arrow schema paper-clipped to it—the kind of instructions you get when you buy a brand new chart, or start your first day at your new job as Chart Operator Level I), as 'a rule for its use': then, three questions are asked and not answered: Can't we imagine further rules to explain this second one? Was the first chart incomplete without the schema of arrows? And are others without theirs?
It's possible to read answers off these questions by taking them to be asked rhetorically ('yes, pointlessly', 'not necessarily', 'no reason why'), but if we take them as posed suggestively, instead, to something like a skeptic, the sort of skeptic drawn around these parts to the idea of excluding (all imaginable, possible) doubt, removing uncertainty about what to do in a particular case, then won't the mere idea of there being two different schemata for using the chart seem to leave open the threat of skepticism rather than closing off anything like it?
What I want to say, I guess, though I can't see how to arrange to say it clearly, is that now that a language-game is being introduced to provide (Wittgenstein's special, sometimes limited, local form of) clarity about the idea of a rule for the use of language, at a point at which the addition of rules to already-described language-games to answer certain skeptically-inflected queries has already been part of Wittgenstein's ongoing response to skepticism, once he does again add rules to a game, the result is unprecedentedly ambiguous: does it satisfy a skeptic or not? And part of the precedent to this has to be, I think, the way in which earlier uses of language-games for critical purposes could often seem much more definitive (but that impression is probably partial: cf. §64).
§86 is the closest yet to a repetition of the way of operating with language attributed to §1's shopkeeper.
In a way, 'a rule stands there' is the whole content of §85: that phrase alone achieves enough distance from another—'a rule tells us what to do'—for us to be able to think both that we have our part to play, and that the rule says nothing—nothing without us.