In my dissertation I examined the interaction of the phonological grammar and lexicon through the lens of Lexical Conservatism (Steriade, 1997). Lexical Conservatism is a theory about what kinds of lexical and phonological factors influence speakers’ decision about how to pronounce a novel word. The hypothesis of Lexical Conservatism states that speakers avoid creating novel allomorphs, and instead preferentially recruit other stem allomorphs from elsewhere in a morphosyntactic paradigm to resolve marked structures created by affixation. I probed these questions using three experiments on English stress placement (chapter 2), and two experiments on Spanish mid-vowel diphthongization (chapter 3).

I found that English words which have phonologically advantageous Remote Bases are produced on a free-response task with right-shifted stress more often than nouns which do not. This effect only holds for individual participants who know both the Local and Remote Bases. The lexical qualities of the Remote Base such as semantic similarity and frequency, as well as classical phonological factors such as the Stress-to-Weight Principle and the avoidance of long lapses, also impact the stress placement in the novel form. I also find that the role of the Remote Base can be influenced on the level of the individual trial by priming, indicating that the Remote Base is actively recruited from the lexicon by the phonology, implying a rich and dynamic interaction unfolding over time and across grammatical domains.

I modeled this grammar-lexicon interaction in a Maximum Entropy Harmonic Grammar framework, proposing a new theory where each Base in the lexicon gets to exert a pull on the novel derivative, which is cross-cut by markedness considerations (chapter 4). This model incorporates information about the lexical status of the Remote Base into a contemporary constraint-based phonological framework by treating the lexicon as prior on the accessibility of different words to the grammar (chapter 5).

A rich body of psycholinguistic research demonstrates that the phonotactic well-formedness of a monomorphemic (non)word has an effect on the way that word is processed when perceived, and also impacts the cognitive effort involved its production in speech. This study examines the effect of phonological markedness across a word boundary on processing, looking at prosodically close English Adjective-Noun sequences in a well-formedness judgment task, a speeded production paradigm, and an accompanying speech-error analysis. I found that different types of phonotactic markedness are distinct in their strength of impact on processing in speech production, although only a subset of the constraints tested show a significant impact of markedness compared to unmarked items, and only one of them shows this effect in the expected inhibitory direction. These results raise interesting questions about the relationship between the neural-level language-production system and the symbolic-level phonological grammar and how the two interact in speech production.