Prevelar mergers

In some parts of the country, "egg" often rhymes with "vague," and some people say "bag" the same way, too.  

In the Northwest, this merger (collapsing of sound categories) affects three front vowel sounds /æ, ɛ, e/ before voiced velar consonants /ɡ, ŋ/ ("hard-G" and "NG").  In sociolinguistic interviews with Seattleites, young adults were less likely than their parents or grandparents to pronounce "tag, bag, lag" like "beg."  In another study, I recorded people in a formal lab setting, where many pronounced the sounds as different -- like the vowels in "bat, bet, bait."  In a perception study, young adults categorized all three vowel sounds as merged, but older people did not.  These patterns suggest that /æ, ɛ, e/ before /ɡ/ have become more merged over the generations, but merger is not as acceptable in formal settings, and young people may be avoiding it.    

Dr. Freeman studied this merger as a graduate student as part of the Pacific Northwest English Project, and she's been keeping it going as a side project. Future phases may investigate prevelar pronunciations in Oklahoma as part of the Oklahoma English Project.

Recent Papers:

Older Papers & Presentations:

Related work:


Student Corner

Data and materials:

  • Dataset (A): 20 audio-recorded sociolinguistic interviews with White Seattleites in two age groups (half young adult, half middle-age), evenly split by gender, some with family, 2-3 hours each, conducted in homes or libraries. Included tasks: conversation/interview about family history in area, minimal pairs, reading passage, 300-word list in carrier phrase.
  • Dataset (B1): Button-box responses from 20 Northwesterners (WA, OR, ID) in three age groups (younger, middle-age, older adults) who heard synthetic /b_/ syllables with front vowels and indicated which /b_d/ or /b_g/ word each could be.
  • Dataset (B2): Audio-recordings of each participant from B1 reading a 40-word list (3 times each) for each English vowel and front prevelars. 17 speakers are also in Dataset C.
  • Dataset (C): Stereo audio-recordings of adult Northwesterners in age- and gender-matched pairs completing six collaborative tasks in sound booth. Overlap of 17 speakers from Dataset B. See Dr. Freeman for details.
  • Time-aligned Praat TextGrids of word lists, passages, and two collaborative tasks with hand-corrected boundaries for at least all measured vowels
  • Measurements (F1, F2, F3, f0, duration) of plain, prevelar, and prenasal vowels
  • Audio of synthetic front-vowel /bV/ syllables (B1)
  • Spreadsheets of responses and response times (B1)
  • TextGrids with stance annotations for two collaborative tasks (C)
  • Participant demographics (all); social network worksheets and interview notes (A)
  • Interview scripts, experiment instructions, prompt slides, passages, lists, consent, flyers, email scripts, IRB forms, etc.
  • R scripts for plots and stats (all); PsychoPy experiment and SynthWorks files (B1); stance annotation protocols (C)

Project ideas:

  • Compare pronunciations between Northwest and Oklahoma English Project
  • Examine prevelar mergers or prenasal vowel trajectories in Oklahoma
  • Play Northwestern prevelars to Northwesterners to see if they can tell near-merged minimal pairs apart
  • Play Northwestern prevelars to Oklahomans: Can they tell them apart? Are they confusing? How do they react/what are their first impressions of the speakers?
  • Play Oklahoman prelaterals to Northwesterners to see how they perceive/react 
  • Consider prelateral mergers in Northwesterners