Anaphora, logophora, indexicality: a syntactic perspective / 2 weeks

Sandhya Sundaresan

Chomsky's Binding Condition A states that an anaphor, like "himself", must be bound in its local domain by a c-commanding antecedent. This successfully generates sentences like: "John_i loves himself_i" and blocks others, like "John_i loves him_i". But the problem is that, as soon as we look beyond English, we see innumerable counterexamples to Binding Condition A:
1. In cases of "long-distance anaphora", the anaphor is bound across one or more clauses, thus violating not just syntactic locality but very often, also syntactic minimality,
2. in so-called "backward binding" structures, the anaphor c-commands its antecedent rather than the other way around (e.g. "Those nasty rumors about himself_i worried John_i terribly")
3. and in "logophoric" binding, the antecedent isn't even present in the same sentence as the logophor (E.g. "John_i was terribly worried. Were those nasty rumors about himself_i actually true?")

These issues are well-known but a satisfactory solution hasn't been proposed: the common strategy is simply to say that there are two types of elements: the "well-behaved" kind which obeys Binding Condition A and is derived by formal syntactic and semantic rules ("anaphor"), and the problematic kind which doesn't and is regulated outside of the grammar proper, by discourse-pragmatic principles ("logophor"). The problem is that these so-called "anaphors" and "logophors", in language after language, look the same on the surface; also, once we abstract away from the technicalities, we see that they also denote the same sort of object: a DP that lacks its own reference but must get it from somewhere else. And, perhaps most importantly, the "well-behaved"/"problematic" divide isn't as well-defined: "anaphors" are sensitive to discourse-pragmatic effects, and "logophors" are sensitive to structural rules.

In this course, we will look at crosslinguistic evidence from a variety of languages to challenge the traditional idea that anaphors and logophors are distinct elements. Instead, we will start with the idea that they are identical, but try to derive our "problematic" cases of long-distance, logophoric, and backward binding in a different way. The main idea I will try to convince you of is that the syntax must be enriched with discourse-pragmatic features: if we do this, we can explain the pragmatic aspects of anaphora/logophora while still retaining the structural restrictions imposed by the syntax. In the process of our investigations, we will also look closely at the interaction between anaphors/logophors and "shifted indexicals" (which have been argued by some to just be another type of anaphor). Although our investigations will be primarily syntactic, we will always be mindful of and be guided by the semantics of anaphora, logophora, and indexicality. In this respect, this course will complement Hazel Pearson's course: "Anaphora, logophora, indexicality: a semantic perspective".