Because of the great diversities of both semantic fields and contexts it is very useful to distinguish three different dimensions of meaning, which in varying ways relate to both the semantic domains of individual words as well as to the conditioning contexts in which such words occur. They are most easily described in terms of a series of contrasts: (1) situational vs. behavioral meanings; (2) linguistic vs. extralinguistic meanings; and (3) intraorganismic vs. extraorganismic meanings (Lounsbury, 1955).

The contrast between situational and behavioral meanings involves a broad field of investigation, for this distinction includes both the stimulus-bearing parts of the context and the responses to it. For example,a particular blaze may constitute the stimulus for the verbal response fire. At the same time, the screaming of fire in a crowded building may be the stimulus for panic. Thus in describing the meaning of fire we must include not only the stimulus situation involving rapid oxidation, but also the behavioral response to such a symbol. It might be argued that the reactions of the crowd are directed toward the referent of fire, not the symbol. On the other hand, announcing to a group that there is a large, destructive conflagration in the building will not produce the same types of reactions as the symbol fire, even though the former may be regarded as in some measure a valid lexical substitute for the latter.

The differences between the linguistic and extralinguistic aspects of meaning are often overlooked,since in general we tend to think only of the extralinguistic elements. That is to say, in speaking of such words as boy, dog, tree, and hill, we tend to consider these words only in relation to certain extralinguistic referents–the objects in question. When we do so, we are simply concentrating our attention on the extralinguistic distributions of these symbols. In other words, we are matching the occurrence of such symbols with the existence of certain types of extralinguistic objects in the environment. But these same words also have linguistic distributions. For example, in general they occur in a certain position in a sentence which we describe as occupied by a noun. Three of these words, dog, tree, and hill, may also occur in a position normally occupied by a verb, e.g. they will dog his footsteps, the dogs can tree the coon, and he will hill up the corn.

There are some linguistic symbols which have virtually no relevant extralinguistic distribution. That is to say, their occurrence must be treated almost exclusively in terms of linguistic distributions. For example, the English particle to, as it occurs before infinitives, e.g. to go, to run, to kill, can scarcely bc "dcfincd" in tcrms of some extralinguistic contexts. Rather, its distribution must be described in terms of other words. The same situation exists, of course, for many linguistic forms, e.g. -ly ( quickly, friendly), -th (width , depth), and de- (detach, deceive ). In many instances grammatical forms combine both linguistic and extralinguistic elements of meaning, as in such categories as number, mode,person, size, and shape, while grammatical case (e.g. nominative, accusative, dative, etc.) involves primarily the linguistic relationships between symbols.