Abstract
In this paper we report on an ongoing multiinstitution effort to encode inferential patterns associated with adjective modification in English. We focus here on a subset of intensional adjectives typically referred to as “non-subjective” predicates. This class includes adjectives such as alleged, supposed, so-called, and related modally subordinating predicates. We discuss the initial results of corpus-based investigations to discriminate the patterns of inference associated with these adjectives. Based on these studies, we have created an initial annotation specification that we are using to create a corpus of adjective related inferences in English.