The realization rules that convert the propositions and relations into language have been characterized in several ways. Bloom's (1970) characterization is a transformational grammar that has a directly semantic base structure. That is, instead of formulating a base structure in terms of abstract linguistic entities, and then invoking some sort of "semantic interpreter" (as in Chomsky, 1965), the deep structure in Bloom's model takes the form of an underlying conceptualization. This leads to problems since it implies that the conceptualization is linguistic to begin with.
Schlesinger (1971) presents a set of realization rules that are not influenced by transformational linguistics. He looks for rules that could convert conceptualizations to sentences. He offers a formal treatment for only the simpler steps of converting "intention- markers" (as he calls the underlying conceptualization) to sentences. An example will suffice to give an idea of the outlines of his model. The following one is taken from Brown's (1973, p. 112) discussion of Schlesinger's model. Consider the sentence "John catches the red ball." The intention-marker for this sentence would include the following conceptions and relations:
red | is attribute of | ball |
(the (red ball)) | is object of | catches |
John | is agent of | (catches (the (red ball))) |
The intention-marker is operated upon by the following realization rules:
Rule R1 may be expanded to read: When is an attribute of b, then a precedes b and a is an adjective, b a noun, and the combination itself a noun. Rule R2 would read: When a is an object of b, then a follows b and a is a noun, b a verb, and the resultant combination itself a verb. Rule R3 reads: When a is an agent of b, then a precedes b, and a is a noun, b a verb, and the result is a sentence.
This example of some rewrite rules provides a rough idea of some of the things the child has to learn. What "rules" the child actually uses is not known. As the child acquires more advanced languaging abilities, the mysteries increase. Rather than get further involved in the unknown, we will move on to how the child begins to make finer distinctions in his speech-to "modulate" the meanings of Stage I with grammatical morphemes.