This paper reports on the type and frequency of pronunciation patterns among a group of Hul'q'umi'num' (Central Salish) adult language learners. With its rich phonological and morphological systems, Hul'q'umi'num' contains many consonants and consonantal sequences that are unfamiliar for learners whose first language is English. Analysis of a set of four pronunciation tests reveals that learners' pronunciation is affected by test, speaker, word, number of unfamiliar sounds within the word, and specific phonological make-up of these sounds. These findings lay the foundation for more in-depth studies targeting each of these factors, to understand their implications for a) additional language pronunciation in Hul'q'umi'num', b) models of second language pronunciation more generally, and c) pedagogical approaches to Hul'q'umi'num' pronunciation.

No abstract

The reliability of pausing as a correlate to prosodic, syntactic and discourse units has been debated in commonly-studied languages (outlined, for instance, in Krivokapic 2007). Preliminary research in Nxaʔamxčín (Interior Salish) (Caldecott & Czaykowska-Higgins 2012) has shown that pausing could be a more reliable acoustic correlate of these structures than pitch, in line with previous research indicating that Salish languages do not exhibit a strong reliance on pitch to mark information structure (Caldecott 2017; Caldecott & Czaykowska-Higgins 2012; Davis 2012; Koch 2008, 2011). The current study further tests this hypothesis by examining the occurrence and duration of pauses in a 9.5-minute spontaneous narrative by a fluent speaker of St’át’imcets (Interior Salish). Pause distribution with respect to syntactic and prosodic boundaries is analysed. Preliminary analysis suggests that St’át’imcets follows previous research on spontaneous narratives in some respects but not others. Very long pauses (<5.5 s) mark major thematic shifts (Oliveira 2000) and most pauses (70%) occur clause-finally (Henderson, Goldman-Eisler & Skarbek 1966). Most clause-medial pauses occurred between a determiner and noun, following Goldman-Eisler (1968), but contra Gee & Grosjean (1983). A t-test indicates that clause-final vs. clause medial pauses are not significantly different in duration (p>.05) (contra Goldman-Eisler 1972).