Patrick Murphy: Linguistics

My research program can be described as applying psycholinguistic methodology (experimental methods and statistical analysis) to understanding dialectal features of Canadian English and Canadian French. I've studied three features of Canadian dialects:

1. Raising in Canadian English (raising of diphthongs /aj, aw/ to [ʌj, ʌw] before voiceless consonants)
2. Affrication in Canadian French (pronunciation of /t, d/ as [t͡s, d͡z] before a high front vowel)
3. Aspectual adjectives in Canadian English (sentences like "I'm finished my homework" or "I'm done my drink", which are grammatical for most speakers of Canadian English but not for most speakers of other dialects)

I've been interested in these dialectal features for their own sake but also to answer broader theoretical questions in linguistics. Here are three of the primary concepts that I've been interested in with my speech perception research:

• Contrast. Languages vary regarding which sounds they use to distinguish meaning (e.g., /t/ and /d/ in English), which sounds they consider equivalent for the purposes of meaning (e.g., different ways of pronouncing /t/ in English), and which sounds simply aren't used. I'm interested in how these phonological statuses affect perception, especially the ability of listeners to distinguish sounds.
• Phonotactics. Languages also vary regarding how they allow sounds to be arranged in words (e.g., /kn/ is a possible onset in German but not English). I'm interested in how listeners show awareness of these patterns to guide their perception (e.g., in cases of perceptual ambiguity).
• Dialect exposure. I'm interested in the perceptual adaptations (e.g., learning to distinguish more sounds) that people make when gaining exposure to another dialect, and whether the effect is similar to what happens when gaining exposure to a completely different language.

Dissertation

Committee: Philip Monahan, Jack Chambers, & Jessamyn Schertz

Dissertation abstract:

Listeners generally have a greater perceptual sensitivity to native contrasts compared to allophones (Whalen et al., 1997; Boomershine et al., 2008) or non-native contrasts (Goto, 1971; Sundara et al., 2006) in discrimination and other tasks. Recent research has emphasized the gradient nature of contrast, showing that many phonological relationships are intermediate or variable between contrast and allophony (Hall, 2009, 2013). This dissertation presents a series of experiments investigating the perception of what has been called marginal contrast or partial contrast using Canadian Raising as a testing ground.

Experiment 1 tests discrimination of raised and non-raised diphthongs ([ʌj]~[aj] and [ʌw]~[aw]) in different phonological environments, finding better discrimination in the contrastive environment where they can create different words than in the allophonic environment where they cannot, but only for one of the two diphthongs (/aj/ but not /aw/). This diphthong difference was ambiguous—it could be a property of the diphthongs themselves, or it could have been a result of the stimuli used, specifically the fact that [ʌj]~[aj] has more recognizable minimal pairs (e.g., writing/riding) than [ʌw]~[aw] (e.g., clouting/clouding). Experiments 2, 3, and 4 seek to clarify this partial contrast effect and diphthong difference, finding support for an inherent diphthong difference (using non-words in Experiment 2) and for an additional effect of the minimal pairs (Experiments 3 and 4). Experiments 1b, 1c, 2b, 3b, and 4b are semi-replications of these initial four experiments. They differ in lacking an additional experimental condition that was present in the original experiments, and in each case the original partial contrast effect fails to replicate, suggesting that partial contrast effects depend on quality/quantity of linguistic exposure. Finally, Experiment 5 tests discrimination of Canadian Raising diphthongs by Canadians and Americans, finding generally faster and more accurate discrimination by Canadians, with notable differences between Americans in regions that di er in how widespread Canadian Raising is.

Together, these experiments provide insight first and foremost into the effect of contrast—specifically partial contrast—on discrimination, as well as other topics such as cross-dialectal perception (and the effect of dialect stereotypes and dialect exposure on perception) and regional differences in the production of raising (and related phenomena) in Canada and the United States.

I got my first taste of working with French dialects (and doing field-work) when I recorded 10 speakers of Acadian French (the traditional dialect of Atlantic Canada, distinct from Quebec French) at Université Sainte-Anne, a francophone university in southwest Nova Scotia. Interesting findings for me included the rhotics, which varied between the apical tap [ɾ] and the English-like retroflex [ɹ] (rather than the more standard uvular [ʁ]).

Praat (editing recordings and making stimuli), jsPsych, OpenSesame, or Testable.org (designing and running the experiment), R/RStudio (analyzing results), LaTeX or RMarkdown (writing the paper or making the poster/presentation). For reading experiments I've used SR Research Experiment Builder and Ibex farm (online). I like Zim Desktop Wiki for organizing notes.