The shift in focus yields a kind of turn which can be labelled 'pragmatic'4 (understanding the term as a referring to giving pride of place to the practical5): the turn from studying language as a system of signifiers associated with their respective signifieds to studying it as a tool for interaction.
In this way, Quine's original idea that in order to understand what meaning is we should study linguistic behavior (especially within the setting of radical translation) slowly mutates into the idea that the truly important thing is the behavior itself – if studying it brings us also the understanding of the concept of meaning, 7 very well; if not, the worse for the concept of meaning and we should simply throw it by the board.
The problem of understanding the role of rules within human linguistic conduct, then, can be portrayed as that of steering among the Skylla of regularism, claiming that a rule is by its nature explicit (we have already seen that this leads to a vicious circle) and the Charybda of regulism, claiming that rule-governed behavior is nothing more than regular behavior (which would erase any difference between a stone's following the law of gravitation by falling and a person's following the rule of traffic by stopping at a red light)8.
This is harmless unless we fall into the trap of understanding this as a picture of a real relation of denoting between an expression and a concept.
Hans Kamp has suggested to me that one of the ways to express this is the following: „A theory of the implementation of memory presupposes a theory of understanding of meaning“.
DRS's; (B) the inferential properties of DRS's are in some sense more explicit than those of English sentences (the properties can be somehow read off from the DRS's themselves); and (C) (3) corresponds to Every farmer who owns a donkey beats it. This yields an understanding of the nature of the praxis of semantic analysis which may differ dramatically from the commonsense view.
However, it has tremendous consequences for grasping the possibilities and limitations of drawing philosophical consequences from such a semantic analysis; and by corollary also for understanding the nature of semantic analysis itself.
All these results have greatly advanced us in our understanding of the nature and structures of our language; however, I think that to become really operative, they must be placed within the framework of a more sophisticated theory of language; a theory which would not rest on some naive picture of expressions as signs which we use to label exhibits of the world-museum, or to externalize our thoughts.
The truth of this is sufficiently manifest from the single circumstance, that the philosophers of the schools accept as a maxim that there is nothing in the understanding which was not previously in the senses, in which however it is certain that the ideas of God and of the soul have never been; and it appears to me that they who make use of their imagination to comprehend these ideas do exactly the some thing as if, in order to hear sounds or smell odors, they strove to avail themselves of their eyes; unless indeed that there is this difference, that the sense of sight does not afford us an inferior assurance to those of smell or hearing; in place of which, neither our imagination nor our senses can give us assurance of anything unless our understanding intervene.
One of those situations is connected with the way we treat infinity; and it is just this situation that is crucial for our understanding of the nature of meaning.