What is very striking about Kripke's skeptic is an assumption he makes which, once it is brought to light, ought to make us suspicious of his whole way of proceeding; this assumption both is essential to the argument and is itself unargued for. I mean the assumption that if there are semantic facts they will have to be reducible to facts specified non-semantically: for the skeptic is in effect demanding an answer to the question ‘what does meaning/reference consist in?’ which does not just help itself to the notions of meaning and reference. Thus the candidate answers Kripke considers all attempt to say what constitutes a given semantic fact without simply using semantic concepts directly (or indeed indirectly) – viz. actual application, states of consciousness, dispositions to use. The kind of reply that is being implicitly judged illicit is one that simply uses semantic concepts, as follows: what it consists in to mean/refer to addition by ‘+’ is for the speaker to mean/refer to addition by ‘+’ – this is the sort of ‘fact’ that meaning consists in. Remember that Kripke's skeptic is out to show that semantic discourse is not fact-stating; then his implicit assumption is that semantic discourse cannot be regarded as factstating just as it stands. The skeptic is assuming that unless semantic facts can be captured in non-semantic terms they are not really facts; but why should this assumption be thought compulsory? So the question for Kripke is why we cannot give the truth conditions of ‘he means addition by “+”’ simply by re-using that sentence, frankly admitting that no other specification of truth conditions is available – precisely because semantic statements cannot be reduced to non-semantic ones.