![]() ![]() In order to make sense of this correlation we must postulate that all words that have this constellation of properties form a natural class, i.e. Thus, we observe a formal correlation between two seemingly independent factors: a) the ability to occur in environment (1b), and b) the ability to have an -ed affix. But what happens if we try to add an -ed affix to the words in (2b)? See for yourself! (3) The roots of these words, too, can occur in the same slot in (1b). Notice too that the words in (2a) are complex, formed with a root and the affix -ed (you can apply the criteria we developed in the previous chapter to justify this). The words in (2a), then, form a distributional equivalence class, which also includes bowed. The words in (2a), for instance, can be inserted there, but the ones in (2b) cannot. If we remove from it the boldfaced word, we get a frame (1b), or context, into which we can insert any words we can come up with. ![]() Distributional criteria are not used alone but in conjunction with morphological criteria, for instance, to justify the existence of a category of words or a part of speech in a language. So fundamental was this technique to the methods developed by American linguists during the first half of the 20th century that it gave name to their whole school of linguistics: Distributionalism. In order to justify a classification of words into grammatical parts of speech, then, we need to develop grammatical criteria, in the sense of criteria for classification that are not tied to vague, anecdotal, or limited semantic notions.Ī well-established test of a word’s category comes from its distribution, that is, the context or set of contexts where a word can (or cannot) occur. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |