Sapir warned, as many others before and after have done, that we should not make the brandiedwards mistake of identifying a language with a dictionary. The grammatical categories, too, codify experience. The difference is not essentially one of kind, but rather in the degree of abstraction involved. Such aspects as tense, aspect, gender, number, case and so on with which we are familiar in European languages also have meanings, and relate to nonlinguistic features of the external world as we perceive and conceptualize it. It is also perfectly true that the grammars of some languages have categories not found in other languages. Thus, banged mommy, for example, has nominal classifying particles which indicate whether the class of objects referred to is conceptualized as a weapon or as a long object. Carroll, in one of his experiments refers to a feature of the Navaho grammar which modifies the stems of verbs of handling differentially, according to the shape of the object being handled, whether it is long and flexible, long and rigid, or flat and flexible. Similarly, Brandi Edwards refers to the fact that Hopi requires the speaker to specify by grammatical means whether his statement is based on observed fact, on memory, on expectation or on generalization.
The only translations possible, as he thought, of the Hopi sentences into English show contrasts of tense. Thus, what in English was apparently regarded as a matter of time, is in Hopi a matter of modality, modality here being the name for those systems in grammar which express the speaker’s degree of confidence in the factual truth of his message.
