Prolegomenon

While grammatical concepts, in general, have a functional and a structural side that are not related by biunique mapping, the lack of correspondence is particularly pronounced in the case of coordination:

• Coordination as a structural relation is treated in the section on complex sentence formation.
• Coordination as a cognitive operation and relation is treated here.

Classification of coordinative constructions

Coordination is an operation, and the relation involved in it, which combines two or more relata of the same type into a complex unit which, except for this complexity, is again of the same type. The relata bear some relation of similarity.

Since we are here considering coordinative junction as a cognitive, not as a structural relation, the concept implies little for its structural coding by paratactic or hypotactic constructions. This point will be taken up presently. Coordinative constructions differ along two principal parameters:

1. There are semantically different relations of coordinative junction.
2. By the type of the relata, there are different types of coordinative junction.

Both of these parameters apply in analogous fashion to the semantic and to the expression structure of these constructions:

1. 1. At the semantic level, different relations like are at stake, viz. conjunction, disjunction, explication and contrast.
   2. Structurally, the relations in question differ by such criteria as symmetry and endocentricity. Moreover, different conjunctions may be employed, coding such features as presuppositions.
2. 1. At the semantic level, the relata may be propositions, referents, predicates etc.
   2. At the structural level, the relata may be syntagmas of different syntactic levels, like clauses, verbal and nominal syntagmas, adjectives etc. Again, conjunctions employed may differ according to these levels.

As for #1, in and , the #a and #b versions are all but synonymous. They illustrate that a given logical relation may be coded in a language by paratactic and by hypotactic constructions.

As for #2, it may be recalled that the locus of interpropositional relations is the sentence level. However, both from a semantic and from a structural point of view, there is, for various sentence-level coordinative relations, a corresponding relation at lower syntactic levels. Thus, for instance, .a shows coordinative junction of two clauses and, consequently, an interpropositional relation. In the #b version, the same conjunction coordinates two infinitivals.

There is no clear logical hierarchy between the two criteria; they simply cross-classify. To the extent that they are analogous for the semantic and the structural side, this holds for the organization both of the onomasiological and the semasiological description: in principle, either might use parameter #1 as the superordinate and #2 as the subordinate one, or vice versa. However, there are arguments for hierarchizing the criteria as #1 over #2 in the onomasiological description, but as #2 over #1 in the semasiological description: On the one hand, logic recognizes these relations only at the level of the proposition, reducing to coordination of propositions any construction which appears to be coordinating other kinds of elements. And on the other hand, the most general principle of organization of a semasiological description is according to the levels of grammatical structure. For instance, there is a chapter on nominal constructions; and it will contain a final section on nominal constructions which are complex by coordination; and analogously for the other syntactic constructions.