(Sektion 199: Philosophische Untersuchungen von Ludwig Wittgenstein).
In Section 199 of the Philosophical Investigations, Wittgenstein shifts his careful consideration of rule-following (see also Philosophische Untersuchungen 143) and expresses the grammatical ties between the following of a rule and the possibility of a private system of rules. The notion of a rule, in the Investigations is all tied up with the possibility of a private language or private rule, as commentators have never been lax to remark (see, for example, Saul Kripke on Wittgenstein).
Wittgenstein's notion of a private language is not that of a language which is physically private or isolated, rather it is of a logically private language, that is, a language (or rule) that is private to a solitary individual or event of necessity. This can be contrasted to an isolated language, such as that of an individual stranded on a desert island for the remainder of their life -- they might develop a language and system of personal communication, but, presumably, this system (and its rules for success and defeat) could be conveyed to a team of rescuers if the occasion arose. A private language, on the other hand, is one that cannot be conveyed to entities that aren't 'privy' to its workings -- this 'not' is of logical compulsion, and a private language is necessary closed in upon itself.
In Section 199, Wittgenstein denies the possibility of a private rule or a private language -- he notes that following a rule is a custom. A rule can be followed only insofar as there is a customary application of the rule, only insofar as there is a custom for following this rule. This helps explain the differences between a rule and its interpretation. Prior to this passage, Wittgenstein cleverly cleverly described how flexible the act of interpretation is -- any activity can be interpreted as conforming to the practical rules of just about any other line of activity -- a rule can be interpreted in a variety of ways. (For instance, Saul Kripke's quus function is the same function as our normal usage of plus (+) except that if either of the two arguments being computed when quusing are greater than 56, than the result is always 5 and not the sum (using plus) of the two arguments). When we do arithmetic or balance our bankbooks, our reasons for asserting that we are plussing according to normal rules rather than quussing are, Wittgenstein has shown, not logical rules. They are practical, customary, cultural, etc..
There is no formal reason (or grammar) behind which we can hide from Kripke's quus function. Yet, there are customary rules, practical institutions, and cultural grounds that we can refer to in asserting that we are doing this rather than that. How, for example, do we know that some person is playing chess and not football? (After all, their wood-pushing could be interpreted according to the rules of football, and systematically so, with enough explaining and time on our hands.) The answer: -- is there a use in the practice of football for four-inch pieces of wood and an eight-by-eight grid upon which they are placed? Does playing football have a use in practices that involve using four-inch pieces of wood? Not these four-inch pieces of wood and this eight-by-eight grid. How do we know that we aren't quusing? Because nobody quuses: there is no custom for quusing, there is no rule there, where people don't follow it: an invisible signpost demands no obedience.
199. Is what we call "following a rule"* something that only one man, only once in his life, could do?--And this is of course a note on the grammar of the expression "following a rule".
It cannot be that a rule was followed on only one occasion by only one person. It cannot be that a report was made, or an order given and understood etc., on only one occasion.--Following a rule, making a report, giving an order, playing a game of chess are all customs (uses, institutions).
Understanding a proposition means understanding a language. Understanding a language means having command of a technique.
* - Anscombe gives "obeying a rule" for "einer Regel folgen" in her standard translation, but I think that this is too cumbersome. One might also ask, "In what sense can a rule by obeyed?" Orders are obeyed, military commanders are obeyed, but rules (as customs and uses) aren't. At least we don't speak of them as such; not in normal usage. Rules are, rather, followed.