What do we actually do when we speak, or write? We produce soundsphonemes, morphemes—which amount to words, and these words are sometimes written down. We therefore replace objects, the world around us (to begin at the simplest level), with sounds (recognizable by others) or with written words. This replacing has long been studied and commented upon by linguists, and we know what is implied by notions such as "referent," "sign," "signified," and "signifier." Linguistics has also taught us that what seemed at first a simple enough operation is, in fact, so complex that we find ourselves faced with several awkward questions, the nature of meaning being one of them. It is this concept that I wish to discuss.
We have the thing, then, and we have the word. Meaning, here, is, or at least seems to be, no other than the mental "content" of the word as a phonetic or graphic unit; it is the reference. Thus, when I say /tābal/ or write table, the operation is accompanied by, or rather includes, the image (thing-presentation, Vorstellung) of a particular object. As we know, to speak supposes this operation of abstraction by which the word can replace the thing. Between word and thing, however (unless the speaker is psychotic), there is no relationship of resemblance or of identity; the sign is arbitrary—which is one more good reason why we can say that the thing is not the word and the word is not the thing. But, of course, while there is no identity between the two, there is a correspondence, a relationship: our linguistic activity does indeed function as if the word were precisely the thing. A replacement has occurred, a substi