Perceptual factors license phonological contrasts in Chamorro Richard Bibbs ◦ Department of Linguistics e-mail: [email protected]

RESEARCH QUESTION: What governs a wider range of vocalic contrast before laryngeal in Chamorro?

1. Chamorro distribution 3. Evidence for perceptually motivated licensing 4. Against an exceptional coda hypothesis

๏ Mid in Chamorro only occur in closed stressed syllables Proposal: Laryngeals permit a wider range of vocalic contrast ๏ Mid vowels in stressed open syllables are not just the result of [4,5,14] due to the persistence of vowel formant information intervocalic consonants being syllabified as codas [10]; here’s why: High vowels Mid vowels Gemination [gúː.pu] ‘fly’ [tém.mu] ‘knee’ ๏ Vowel formants persist through the laryngeal, providing longer vowel ๏ The C of certain –CV suffixes geminate when a word has a closed [úː.luʔ] ‘worm’ [mét.gut] ‘strong’ steady state and transition information as perceptual evidence [líː.ʔiʔ] ‘see’ [pók.puk] ‘bump’ ๏ realized as creaky word-medially stressed syllable, and a word-final open syllable [4,5,14] [gék.pu] [gik.pók.ku] ‘my flyer’ ๏ Stress shift triggers alternations, both raising and lowering: [tém.mu] [tim.móɲ.ɲa] ‘his/her knee’ [mét.gut] [mit.gót.ɲa] ‘stronger’ ๏ Gemination does not trigger for forms with a stressed [pók.puk] [puk.pók.ɲa] ‘his/her bump’ before an intervocalic [tém.mu] [tim.móɲ.ɲa] ‘his/her knee’ [téː.ʔuk] [bé.ʔi] [be.ʔíː.ɲa] ‘his/her bandage’ ๏ Also, raising of mid vowels in nativized loans: F2 ‘thick’ [bó.ʔu] [bo.ʔúː.hu] ‘my bubble’ • Notice that stressed [hóː.dzu] < Spanish [óʝo] ‘hole’ F1 Penultimate lengthening mid vowels remain… [béː.lu] < Spanish [βélo] ‘veil’ ๏ Vowels in penultimate stressed open syllables are lengthened [4,5,14] [túː.giʔ] (109ms) ‘write’ [mét.gut] (55ms) ‘strong’ Front Central Back ๏ Vowel formants do not persist through oral consonants: (5 tokens) (4 tokens) • Chamorro vowel High i u [díː.suʔ] (98ms) ‘squat’ [pók.puk] (45ms) ‘bump’ inventory [4,5,14]: Mid e o (7 tokens) (7 tokens) Low a ɑ ๏ Mid vowels before intervocalic consonants are lengthened, indicating an open syllable, i.e. no coda assignment [túː.giʔ] ‘write’ [bóː.han] (116ms) ‘hand-fan’ [dóː.ʔak] (148ms) ‘cataract’ 2. Exceptionality of mid vowels before laryngeals (8 tokens) (7 tokens) F2 [téː.ʔuk] (112ms) ‘thick’ [bóː.ʔan] (116ms) ‘froth’ ๏ Some mid vowels exceptionally occur in stressed open syllables F1 (3 tokens) (4 tokens) in the native vocabulary ๏ High vowels in stressed open syllables become a major puzzle if this Mid before laryngeal consonant Mid before hypothesis is adopted ๏ Distinctiveness of contrasts captured through constraints referencing [bóː.ʔan] ‘froth’ [góː.fis] ‘lungs’ perceptual distance between formants in positional inventories [7,11] CONCLUSION: [téː.ʔuk] ‘thick’ [póː.tu] ‘rice-cake’ ๏ Assign F1 levels relative to vowel height: i = 1 , ɪ = 2 , e = 3 [7] ๏ Reference to phonetic cue information allows a cohesive account of [déː.ha] ‘see’ [éː.tsuŋ] ‘crooked’ ➢ Mindist:F1:2 – Assign a violation if distance between F1 levels is ≤ 2 patterned exceptionality in Chamorro ๏ An observation: before laryngeals, mid vowels are more common ➢ NoMerge – Assign a violation for every pair of merged vowels ๏ Other approaches, such as licensing by cue [13], may be equally ➢ Periph – Assign a violation for every non-peripheral vowel ([o e]) than expected [5,12]; is this just chance? effective, but still maintain integration of perceptual factors within Perceived F1 contrasts better signaled with longer formants ๏ A chi-squared test for significance can be conducted on bisyllabic ๏ the phonological system Formant length scales perceptual distance between formants: native roots from the Revised Chamorro-English dictionary [12] ๏ ๏ A purely phonological account of exceptionality is possible, a la [8], laryngeals multiply by 1.5, oral consonants multiply by 1 but not as effective for this case due to a lack of evidence for a Mid vowel High vowel Total unique diachronic pathway to explain lexical categorization [3] Intervocalic 29 48 77 íːxʔ ~ éːyʔ MINDIST:F1:2 NOMERGE PERIPH ๏ Neither is there evidence of loan word influence conditioning a laryngeal (19) (58) → íːxʔ ~ éːyʔ (3) * separate stratum that exceptional forms occupy, a la [9] Intervocalic 107 368 475 íːx,yʔ *! (117) (358) th oral éː ʔ *! * References: [1] Bessell. (1992). Proceedings of the 28 Meeting of the x,y Si Yu‘us ma‘åsi‘ǃ Chicago Linguistic Society. [2] Bird. (2011). International Journal of Total 136 416 552 American Linguistics, 77(2), 159-184. [3] Blust. (2009). Pacific Linguistics. [4] Acknowledgements: many thanks to my Chung. (1983). Language, 59(1) [5] Chung. (In Progress). University of íːxCoral ~ éːyCoral MINDIST:F1:2 NOMERGE PERIPH patient and helpful Chamorro consultants, California, Santa Cruz. [6] Crosswhite. (1998). Phonology, 15(3), 281-316. ๏ X-squared = 7.38, df = 1, p-value < 0.01 – significant, not chance! íː C ~ éː C (2) ! * [7] Flemming. (1995). University of California, Los Angeles. [8] Gouskova, M. x y * and to Ryan Bennett, Sandy Chung, Junko Ito, (2012). Natural Language and Linguistic Theory 30(1), 79-133. [9] Ito, J., & Amanda Rysling, Grant McGuire, and all Mester, A. (2008). Lexical Classes in Phonology. [10] Khan. (1976). ๏ How might the patterned exceptionality [15] of mid vowels before → íːx,yC * members of the 290 research seminar for Massachusetts Institute of Technology. [11] Padgett. (2003). Natural laryngeals be explained? éː C * *! Language & Linguistic Theory. 21(1), 39-87 [12] Revised Chamorro-English x,y their support and insight Dictionary. [13] Steriade, D. (1997). Manuscript. [14] Topping & Dungca. (1973). University of Hawaii Press. [15] Zuraw, K. (2000). UCLA.