A lexical-interpretive approach to passives
Since the synonymy of active-passive pairs is not explicitly accounted for by a rule in the syntactic or semantic component, a grammar that captures the active-passive relation must include a rule that does so. In the theory to be sketched here, the synonymy of actives and passives can be specified in terms of the semantic equivalence of predicates and the semantic functions associated with them. Without going into detail, we may consider that any predicate (i.e., verb, adjective, gerundive nominal—and with exceptions, derived nominal) governs a particular set of semantic functions which is determined according to the conceptual structure of the semantic class of the predicate.8 Thus a predicate of motion carry, as in John carried the chair from the study to the kitchen, governs the semantic functions “agent” (the agency responsible for the motion), “theme” (the object moved), “source” (the location the theme is moved from), and “goal” (the location the theme is moved to). In the example, John is agent; chair is theme; study is source; and kitchen is goal. In the passive version of the example, The chair was carried from the study to the kitchen by John, the NP’s perform the same semantic functions, even though the form of the predicate has changed from active to passive and the syntactic positions of the NP’s have been re-ordered.
From this discussion it should be clear that the two sentences can be interpreted as synonymous if the two predicates are semantically equivalent. This poses no problem for active-passive pairs because the equivalence of active and passive predicates has always been assumed. Thus a rule of semantic interpretation which marks active and passive sentences as synonymous could be stated as:
(1)
This rule would have to apply after those interpreting quantifiers, negation, and coreference.
Using rule 1 for determining the semantic equivalence of sentences, it becomes possible to approach the near-paraphrase relationship between the following:
(2)
Examples of this sort were mentioned in Chomsky 1965 as sentences which are obviously related in meaning but which could not be transformationally related. We already know that buy and sell are predicates of motion, and that the NP’s in each sentence have the same semantic function—i.e., John is goal; Jack is source; the book is theme. Therefore the meaning difference between the two sentences is determined by the semantic fine structure of the two verbs. The predicates buy and sell, aside from belonging to the same semantic class, refer to the same event, a financial transaction. They differ in terms of the interpretation of agency, which seems to be a matter of focus.
In 2a, John is agent; in 2b, Jack. Yet each predicate entails the other. Since such entailments are part of the semantic representation of these predicates, this information is available to rules of semantic interpretation like 1. If the definition of semantic equivalence of predicates accounts for such entailment relations, then 1 should be able to account for the near-paraphrase relation between 2a–b without resorting to unmotivated syntactic analyses.
Since the theory being developed here does not include a single phrase marker from which passive and active sentences are derived, the relationship between active and passive predicates must be captured in the lexicon. For this reason, it is necessary to make a few comments on the structure of the lexicon and derivational morphology.
