The inclusion of such data as these in the study of language has followed a long period in which linguists concentrated on segmenting 'enate' sentences (sentences with identical surface structures) and classifying the resulting parts:
| The | cook | used | cornmeal. |
| The | people | ate | mush. |
| The | children | ate | mush. |
To insist that the principal things to be learned in a language are its 'patterns' is one thing, but this word may be interpreted with the same latitude as 'habit' (see above). To define 'pattern' enately, as 'a sentence or phrase with all of the content words removed' (Brown, 1967, p. xviii) is unnecessarily narrow. In this sense, the sentences in the frame above would all represent the same pattern, which could be represented somewhat as follows:
| Article | Personal Noun | Transitive Verb | Noun |
and the five phrases about the cook using cornmeal would be regarded as representing five different and presumably unrelated patterns. To define 'pattern' in this way encourages the writer of materials to ignore the extremely productive agnate relationships, such as connect the five sentences about cornmeal. But this book is not the place for detailed discussion of either of these matters.
13
