ABSTRACT
The central goal of this work has been to study the interpretation and biases of ellipsis sentences, and to explore the role of parallelism between DPs in the processing of ellipsis sentences. Parallelism between DPs and other elements of unelided conjoined structures was originally found by other researchers (e.g., Frazier et al., 1984, 2000; Henstra, 1996) to produce a facilitation effect in the processing of unambiguous sentences. In this book, many types of parallelism between DPs have been identified and shown to affect processing, supporting the Parallelism Hypothesis, but not to determine the basic structural hypothesis, as claimed by the special purpose Alignment Hypothesis. The structural effects found for Gapping sentences, for example, confirm that ellipsis sentences should be parsed using normal processing mechanisms that include structure-based processing principles as well as parallelism, contra the Alignment Hypothesis. The effective types of parallelism have included similarities between DPs in grammatical and extra-grammatical features as well as prosodic similarities which favored different focus structures. Such factors have biased the interpretation of a wide range of ambiguous ellipsis types, including Gapping, Bare Argument Ellipsis, Comparative Ellipsis, Replacives, and Stripping sentences. This list includes ellipsis types which contain the conjunction and as well as ones that do not, showing that parallelism effects are not limited to conjoined structures.
