Sedivy et al. (1999) showed with eyetracking that adjectives are processed incrementally as they are encountered (heard) however looking specifically at English it is not clear how this is possible. As an example of the problem we can look at the phrase "the big square table". In this phrase we have the two adjectives and in hearing it listeners, as they are incrementally processing, would first search out the objects in the context (visual or otherwise) that are "big". But how is a listener assigning any sort of meaning to "big", a gradable adjective, that must have a scale when at the time of incorporation there is no available scale? Does this mean that each person/language/culture has some default threshold to be used in these moments of underspecification and if so how is that set? The listener then encounters the next adjective "square" which is not gradable (though more or less precision can be used) and now they have to update their prediction of the target object. But what does that update look like? Is it the intersection of the generically or absolutely "big" things and the generically or absolutely "square" things or is some set of generically or absolutely "big" things already picked and then the "non-square" things are weeded out? And finally when we get to the actual noun does this comparison class update process look similar to that of the nongradable adjective or is it in some way different across different parts of speech? These are the questions that I aim to answer through this ongoing project that I am working on in conjuction with Dr. Helena Aparicio.
Much of the discussion around thematic roles in a linguistic context has centered on verbs and how they assign these roles to thier different arguments. However any student of syntax knows that as soon as one tries to define or deliminate the roles from one another in concrete terms the distinctions become fuzzy. Dowty (1991) proposes an alternative to these broader thematic roles in the form of Proto-Agent and Proto-Patient roles which can be assigned by verbs in different combinations and which are what broader thematic roles had previously tried to capture. Reisinger et al. (2015) built on this work and showed that Proto-Roles can be understood and judged through online experimental platforms allowing for greater possiblities for experiments using these Proto-Roles. However throughout this work the focus has stayed primarily on predicate(verb)-argument(noun) pairs. There are many other parts of language that seem likely to be able to influence these Proto-Roles especially adjectives due to thier close relationship with nouns. The goals of this project are to find out whether adjectives can influence proto-role judgements and if they can how does that influence change depending on the semantics of the adjective. This is an ongoing project in collabaration with Dr. Marten van Schijndel.
In the Anatolian language of Hittite attributive adjectives are found both pre and postnominally. Previous work posited the prenominal position as the standard position for attributive adjectives with the postnominal position being only for exceptions and participles. Due to the number of exceptions included, that theory seemed ripe for reevaluation. Through analysis of diachronic sources of Hittite I reanalyze the prenominal position as a place for subsective adjectives and the postnominal position as a place for intersective adjectives which includes the so called "exceptions" as well as participles. This is hopefully just the first step towards fully analyzing the details of Hittite nominal phrase syntax.