Often – and especially in theories under the heading ‘semantic holism’ – we explain the meaning of a statement by another statement: ‘Pete is a cat’ means ‘Pete is a feline mammal’.  And we may make use of this second statement as delivering criteria in judging the correctness of applying the first statement (‘No, this isn’t a cat at all, since this isn’t a mammal, but some Martian penguin vaguely looking like a cat.’). Thus lin­guistic descriptions made in some language occur in our semantic rules for that very language. The problem of the regress starts with the simple question, how this second description is rule-governed itself. If another rule is introduced we need another one, and so on. So Wittgenstein, in his discussion of rule-following, considers any theory like semantic holism as inappro­priate. Any interpretation would need an interpretation itself, and therefore leaves us hanging in the air.^[1]  On the other hand, according to Wittgenstein, any behaviour could be made fitting any rule if we just re-interpret the rule, since the rules aren't de­termined enough. For example: If you re-interpret ‘mammal’ as ‘mammal seen before 2007, and fridge ever after’ you are free to come up with strange statements about cats in 2007.

I will consider two solutions to the regress problem, which are also two interpreta­tions of Wittgenstein: (a) the recourse on some capacity (Colin McGinn´s Wittgenstein^[2]) and (b) the recourse on the praxis of rule-following (Saul Kripke´s Wittgenstein^[3]).

(ad a)

If one appeals to a capacity of rule-following the regress of rule-following vanishes: We speak the way we speak, since we have a corresponding capacity. This seems to explain nothing. That we swim because we have the capacity to swim, stated thus, does not explain anything (how do we do it?). The same holds for language. Secondly: Inasmuch as the regress should be avoided the capacity to follow a rule must not be rule-governed itself. Semantic decisionism (being the opposite of orientation on rules), however, vio­lates our intuition that we have reasons to employ this expression and not another. Wittgenstein says himself that when we follow a rule, we don't choose.^[4]  And understanding ‘capacity’ as employing a causal mechanism (like our capacity to stand upright) is in danger of giving away on the normativity of language usage: our choice of word can be wrong causality just operates.

(ad b)

In Kripke´s solution, that which determines the correct application of an expression is the common practise of a language community. No rule but ‘praxis’ speaks for itself. Rule-following is specified by a behaviour that accords to the rule. We fit in, tune in to our linguistic community. Convergence within the community constitutes the rule in the first place.

But is this an alternative to (a) or to semantic holism? Instead of a subjective ca­pacity we now have a social habit. But how do I take part in this habit? I seem to need the capacity to participate in the doings of my speech community. Since there are – ex hypothesi – no rules laid down to which I adjust I have to tune in to my community. The notion of ‘praxis’ is by itself not clearer than that of ‘capacity’ or that of ‘rule-following’. Even if the conformity within the language community is part of securing the proper application of the semantic rules, it cannot substitute for explicit regula­tions. A mere praxis is not better than any other praxis. That people talk this way today about cats does not make it the right way to talk so. In the face of rule scepticism we might even say: How can we distinguish between wrong applications and a change of rule? If we orient ourselves to the majority of the speakers, how is it possible to criticise the majority for an incorrect employment of an expression? Why can’t Sue be the only left competent user of ‘cat’? Shouldn’t we, at least, require that the majority adheres to its own standards (i.e. rules) of usage so far? The ma­jority might be the institution to change rules, but once it has laid down a rule we should be able to criticise its behaviour with reference to these stipulations. We seem to lose the normativity of rule-following if the majority is in its praxis always right. And the praxis of a community is as re-interpretable as any behaviour. So (b) also leaves the scepticism undefeated.

These are reasons enough to see whether we don't do better in solving the regress problem within semantic holism. The main problem of semantic holism is our intuition of a regress. Justification, further on, is holistic: We give reasons without going back to some ultimate ‘given’ or ‘incorrigible’ facts, we refer to other reasons instead, even if this means going in circles, given that the circles are wide enough. Reasons (statements held to be true) support other reasons/statements, but we can always put the last reason given to test.

Nevertheless, we can solve the regress as follows: Because of the holistic procedure of justification (and therefore of meaning something) we are allowed to keep asking for further reasons in principle, but in doing so we employ a meta-rule of sufficient foundations:

If there is no founded/reasoned doubt, there is no need for further foundation/argumentation.

We employ the semantic rules in some situation and try to conform to the habits of our community. If someone asks us why we do so, we explain our usage by reference to the fulfilment of the criteria of use (i.e. give reasons by citing a second description like ‘feline mammal’). This duty is part of the normativity of meaning. But if in respect to the fulfilment of some criteria in such an argument after several steps there is no longer reasoned doubt (i.e. no foundation for belief in their non-fulfilment), why should  we proceed in founding our claims? Our argument now is (relative to all claims founded in that debate) sufficient. Relative to our knowledge of this state of the argument and our knowledge of the rule of meaning it is the optimal logical procedure to evaluate the usage as ‘correct’. All reasons we have now speak in favour of this evaluation. This is neither an act of decisionism nor an act of some capacity, but the application of our rule-following procedures which can be taken up again in principle and has been interrupted only at a suf­ficiently clear point. So the meta-rule seems to be the lesser evil in comparison to the consequences of (a) and (b). We have not overcome the principled problem of the regress, but we can see that it is harmless if we employ our rule of sufficient founda­tion. The regress problem has our intuition of foundation as its driving force. But with respect to our intuitions the meta-rule as a principle that an argument just has to be sufficiently clear seems equally strong. Our pre-understanding of rule-following, therefore, doesn't decide the matter. And all other reasons in our comparison of (a),(b) and se­mantic holism speak in its favour.

So we can use semantic holism to explain meaning by rule-following. There is nothing wrong with stating the meaning of a sentence by giving another sentence.