Type-shifting in formal semantics of natural languages (original) (raw)

In type theoretical approaches to formal semantics of natural languages, a lexical item in a sentence is assigned a type and using a number of composition rules and based on the syntactic structure of the sentence, the types compose to converge to a single type called t, the semantic type for sentences, together with a statement specifying the conditions under which the sentence is true. For example consider the example (1-a) below (1) a. Every linguist offended Fred. b. Fred offended every linguist. The syntactic structure of the sentence, its type assignments and the compositions are as follows(we omit the obvious truth conditions since it does not concern our discussion of type shifting): (2) t DP < et, t > < et, < et, t >> every < e, t > linguist < e, t > V < e, et > offended NP e N e Fred So the above type system correctly predicts that the example (1-a) is semantically interpretable. Consider the word every which has been assigned the type <et , ett>. We say the type <et , ett> is a locally correct type for the word every since this type assignment gives the expected result for a context, namely example (1-a). We call a type assignment, a globally correct assignment if the type assignment converges in all contexts containing the word. Whether there exist globally correct type assignments for lexical entries or not is a challenging question. There are at least four possibilities to be considered: (1) (Hypothesis A) Lexical items have globally correct type assignments. If the type assignment for a lexical item does not converge in a given interpretable context, then either the syntactic structure of the context needs to be changed or a new lexical entry needs to be defined. (2) (Hypothesis B) Some lexical items do not have globally correct type assignments but their locally correct assignments are systematically related.

Sign up for access to the world's latest research.

checkGet notified about relevant papers

checkSave papers to use in your research

checkJoin the discussion with peers

checkTrack your impact