Velar nasal plus in the north of (ing)land George Bailey University of Manchester @grbails UKLVC11 - 31st August 2017
1. Introduction Velar nasal plus Historical origin The life cycle 2. Methodology 3. Results Unstressed (ing) Stressed (ng) 4. Conclusion Summary 2
Velar nasal plus (Wells 1982: 365) Presence of post-nasal /g/ in varieties spoken in the North West and West • Midlands of England Liverpool (Knowles 1973); West Wirral (Newbrook 1999); • Manchester (Bailey 2015; Schleef et al. 2015); Cheshire (Watts 2005); Birmingham (Thorne 2003); Cannock (Heath 1980); the Black Country (Mathisen 1999; Asprey 2015) Well-attested in dialectological literature but the nature of its variation is • relatively understudied ! Even has its own emoji: • Envelope of variation can be split into two distinct environments: • [ ɪ n] [ ɪ ŋ ] [ ɪ ŋ g] (ing) e.g. runn ing, wait ing (ng) [V ŋ ] [V ŋ g] e.g. ki ng, si ng er 3
Velar nasal plus (Wells 1982: 365) Presence of post-nasal /g/ in varieties spoken in the North West and West • Midlands of England Gi* z-score Liverpool (Knowles 1973); West Wirral (Newbrook 1999); • 2 Manchester (Bailey 2015; Schleef et al. 2015); Cheshire (Watts 1 2005); Birmingham (Thorne 2003); Cannock (Heath 1980); the 0 -1 Black Country (Mathisen 1999; Asprey 2015) 1960s 2015-17 (Orton et al. 1978) (based on data from MacKenzie et al. 2017)
1. Introduction Velar nasal plus Historical origin The life cycle 2. Methodology 3. Results Unstressed (ing) Stressed (ng) 4. Conclusion Summary 5
Historical origin • Origins of (ing) and (ng) variation closely intertwined • (ing) originates from two Old English suffixes: present participle -inde and verbal noun form - ynge/-inge (Visser 1966) • Reduction (and later deletion) of the final vowels -> simplification of the consonant clusters leading to nasal place contrast (alveolar vs. velar) -> conflation of two forms • Simplification of the / ŋ g/ cluster never ran to completion in the North West of England, leading to surface variability between [ ŋ ] and [ ŋ g] that still exists today • Diachronic evidence suggests that the rule deleting post-nasal /g/ evolved in a very systematic way, following the ‘life cycle of phonological processes’ (Bermúdez-Otero 2011) 6
1. Introduction Velar nasal plus Historical origin The life cycle 2. Methodology 3. Results Unstressed (ing) Stressed (ng) 4. Conclusion Summary 7
The life cycle of phonological processes (Bermúdez-Otero & Trousdale 2012) Phonology split into three ‘cycles’ • Phonological processes begin as post-lexical • rules before climbing into more embedded domains over time 1. P HRASE - LEVEL : rule can see the whole phrase (i.e. across word boundaries) Time e.g. Jon Snow is the King in the North ] 8
The life cycle of phonological processes (Bermúdez-Otero & Trousdale 2012) Phonology split into three ‘cycles’ • Phonological processes begin as post-lexical • rules before climbing into more embedded domains over time 1. P HRASE - LEVEL : rule can see the whole phrase (i.e. across word boundaries) Time 2. W ORD - LEVEL : rule can only see the word itself e.g. Jon Snow is the King in the North e.g. Morrissey is a talented singer from Manchester 9
The life cycle of phonological processes (Bermúdez-Otero & Trousdale 2012) Phonology split into three ‘cycles’ • Phonological processes begin as post-lexical • rules before climbing into more embedded domains over time 1. P HRASE - LEVEL : rule can see the whole phrase (i.e. across word boundaries) Time 2. W ORD - LEVEL : rule can only see the word itself 3. S TEM - LEVEL : rule can only see the stem e.g. Jon Snow is the King in the North e.g. Morrissey is a talented sing er from Manchester 10
The life cycle: synchronic predictions Synchronic implication under a cyclic framework: • words where the /g/ is eligible for deletion (i.e. in coda position) in more • cycles -> more chances for /g/-deletion to apply -> higher probability of surface [g]-absence /t,d/-deletion (Guy 1991) and /l/-darkening (Turton 2014, 2017) have been • analysed under similar frameworks Higher probability of deletion finger singer sing it sing || sing tunes Phonological computation _V _#V _#|| _#C /f ɪ ŋ .g ə / /s ɪ ŋ g/ /s ɪ ŋ g/ /s ɪ ŋ g/ /s ɪ ŋ g/ Stem-level /f ɪ ŋ .g ə / /s ɪ ŋ .g ə / /s ɪ ŋ g/ /s ɪ ŋ g/ /s ɪ ŋ g/ Word-level /f ɪ ŋ .g ə / /s ɪ ŋ .g ə / /s ɪ ŋ .g ɪ t/ /s ɪ ŋ g/ /s ɪ ŋ g.t ʃ u ː nz/ Phrase-level Chances to apply: 0 1 2 3 11
1. Introduction Velar nasal plus Historical origin The life cycle 2. Methodology 3. Results Unstressed (ing) Stressed (ng) 4. Conclusion Summary 12
Methodology Quantitative approach using twenty-four • sociolinguistic interviews conducted with North Western speakers ‣ two speakers recorded in 1971 for a real- time component Blackburn Stratified by age and sex (all ‘working class’ • Manchester speakers) Interviews typically one hour long, followed by a • reading passage and word list Transcribed and force-aligned using the FAVE • suite (Rosenfelder et al. 2011) All tokens coded by hand for [g]-presence • Mixed-effects logistic regression using lme4 in • R, with random intercepts of speaker and word • 3760 tokens of (ing) ~ 1459 tokens of (ng) The Linguistic Atlas of England - Orton et al. 1978
1. Introduction Velar nasal plus Historical origin The life cycle 2. Methodology 3. Results Unstressed (ing) Stressed (ng) 4. Conclusion Summary 14
Overview Unstressed (ing) • [ ɪ ŋ g] almost completely 100% absent in conversational data (0.7%) • Even the plain velar nasal [ ɪ ŋ ] 75% is rare (11.9%) Proportion of tokens Variant • Rates of alveolar -in are high ɪ n even in contexts (and for 50% ɪŋ social groups) that usually ɪŋɡ disfavour this variant 25% • weak age-grading pattern, and only for female speakers 0% • no effect of part of speech BegleyJ BethS BruceG ChrisT ConnorL DaveJ FeliciaD FrankE GloriaJ GraceG GrahamR HarryG JimmyC LillyR MaryB MikeM MollyF TanyaC TheaS WadeT WandaJ WendyJ WillowA (cf. Tagliamonte 2004 in York) speaker 15
Grammatical category Unstressed (ing) BegleyJ BethS BruceG ChrisT ConnorL DaveJ 100% 75% 50% 25% 0% FeliciaD FrankE GloriaJ GraceG GrahamR HarryG 100% 75% 50% 25% N Rate of -in 0% 40 JimmyC LillyR MaryB MikeM MollyF TanyaC 80 100% 120 75% 50% 25% 0% n e b v u r TheaS WadeT WandaJ WendyJ WillowA i e o t v c n e 100% j d a 75% 50% 25% 0% n e b n e b n e b n e b n e b v v v v v u r u r u r u r u r i e i e i e i e i e o t o t o t o t o t v v v v v c c c c c n n n n n e e e e e j j j j j d d d d d a a a a a Grammataical category
Grammatical category Unstressed (ing) Unstressed (ing) Surprising given that the • effect is strong both in York the US (Labov 2001) and even elsewhere in the UK (e.g. York - Tagliamonte 2004) Absence of part of • speech conditioning also attested in nearby community of Wilmslow (Watts 2005) SED data from the Linguistic Atlas of England - Orton et al. 1978
Style Unstressed (ing) 100% Rates of velar nasal plus increase for • the reading passage, but only slightly; predominantly used in word list 75% Proportion of tokens Could this reflect something other than • Variant prestige (e.g. speech rate or prosody)? ɪ n 50% ɪŋ Suggestions that [ ɪ ŋ g] is seen as ‘less • ɪŋɡ socially attractive’ than [ ɪ ŋ ] anyway (Schleef et al. 2015) 25% over-articulate and associated with • an “unenergetic, uptight attitude 0% towards life” (p. 207) conversation reading word list passage Style 18
1. Introduction Velar nasal plus Historical origin The life cycle 2. Methodology 3. Results Unstressed (ing) Stressed (ng) 4. Conclusion Summary 19
Results Stressed (ng) 100% Highly variable in • conversational data, unlike (ing) 75% Proportion of tokens No main effects of • age, sex, part of Variant ŋ 50% speech, or lexical ŋɡ frequency But strongly • 25% conditioned by morphophonological factors 0% BegleyJ BethS BruceG ChrisT ConnorL DaveJ FeliciaD FrankE GloriaJ GraceG GrahamR HarryG JimmyC LillyR MaryB MikeM MollyF TanyaC TheaS WadeT WandaJ WendyJ WillowA speaker 20
Life cycle’s predictions Morphophonological effects 100% Prediction: correlation between surface • rate of application and the number of cyclic levels in which the rule had chance to apply 75% Turns out to be the strongest predictor • N Rate of / ɡ /-deletion of [g]-presence 150 200 one chance : 19% deletion • 50% 250 300 ‣ ( SINGER -type tokens) 350 two chances : 46% deletion • 25% ‣ ( SING # V -type tokens) three chances : 67% deletion • 0% ‣ ( SING # C -type tokens) one two three Number of cyclic domains in which / ɡ /-deletion can apply ‣ ( SING #||-type tokens) 21
Recommend
More recommend
