The Structure of Multiple Cues to Stop Categorization and Its Implications for Sound Change

Total Page:16

File Type:pdf, Size:1020Kb

The Structure of Multiple Cues to Stop Categorization and Its Implications for Sound Change The structure of multiple cues to stop categorization and its implications for sound change Hye-Young Bang Department of Linguistics McGill University, Canada October 2017 A thesis submitted to McGill University in partial fulfillment for the requirements of the degree of Doctor of Philosophy ©Hye-Young Bang 2017 Abstract The central goal of this dissertation is to understand how multiple acoustic cues that signal phonetic contrasts are structured across linguistic and socio-linguistic factors, and what elements in these structures contribute to a long-term change in pronunciation norms in a speech community. The acoustic cues this dissertation focuses on are VOT and f0 which covary as cues to stop voicing categorization. The covariation of these cues is of particular interest because it is related to a cross-linguistically common type of sound change called ‘transphonologization’, where the relative importance of the cues that signal a set of phonetic categories changes over time. There are many questions that remain unresolved regarding acoustic cue covariation in speech production, including: (1) what is the mechanism and path of transphonolo- gization; (2) which aspects of the structure of cue covariation are language-independent, and which uniquely appear during sound change? This dissertation addresses these two questions in two studies, using large datasets from spoken corpora of three languages. Study 1 examines how transphonologization from VOT to f0 originates, how it prop- agates, and what consequences the change has for other aspects of the sound system of the language undergoing change. This question was examined using a dataset from a large apparent-time corpus of Seoul Korean (the NIKL corpus). The time-course of change in VOT and f0 was examined focusing on linguistic (word frequency and vowel height) and social (speakers’ age and gender) factors. The results showed a consistent trade-off between cues across talkers and contexts, and that the shift in cue primacy is more advanced in younger females’ speech and in the linguistic conditions where the VOT contrast is more reduced (high frequency words and before non-high vowels). These findings suggest that VOT reduction and f0 enhancement are spreading in parallel across speakers and the language in an adaptive manner, driven by a combination of production bias in one dimension (VOT) and adaptive expansion in another (f0). Study 2, building on the results of Study 1, sought to better describe how linguistic and social variables structure cue covariation across languages. It further aimed to identify aspects of cue covariation that are unique to languages undergoing change, by comparing Seoul Korean with two languages that are not undergoing transphonologization, English and German. The analysis examined corpus data from the three languages, focusing on two linguistic factors (word frequency and vowel height) one social factor (gender), and variation across individual speakers, and their effects on the relative use of VOT and f0 in signalling stop voicing categories. Regression models captured the effects of the linguistic and social factors, and cue use (or ‘weights’) for each individual was quantified using several methods: Cohen’s D, linear discriminant analyses, and support vector machines. For German and English, in high-frequency words and before non-high vowels, one cue is 1 consistently used less than the other cue, reducing total cue informativity for stop voicing contrasts. Interestingly, these are the same conditions where transphonologization in Seoul Korean is more advanced. The results for gender and individual speakers, on the other hand, showed consistent cue trade-offs in all languages. Taken together, the results from these two studies show that the cue trade-offs across linguistic conditions, observed only during the change in Seoul Korean, may have initiated to compensate for reduction in ‘cue informativity’ to avoid contrast merger, a pattern which is likely unique to transphonologization. On the other hand, the consistent cue trade-offs across speakers and gender in all three languages suggest that speakers have a goal of constant total cue informativity, while maintaining socially conditioned stylistic variation. The resulting structured variability across speakers may further serve as a catalyst to sound change, and change may be propagating by strengthening these existing structures. Broadly speaking, the findings provide evidence that variability in speech is not random but structured in multiple dimensions across contexts and talkers. This structure potentially reduces the variability that listeners must cope with in real-time processing and helps us understand how sound change occurs over generations. 2 Résumé Cette thèse a comme objectif principal d’éclaircir (a) comment sont structurés les multi- ples indices acoustiques qui signalent les contrastes phonétiques, une structure qui reflète des facteurs linguistiques et sociolinguistiques, et (b) quels éléments de ces structures contribuent aux changements à long terme des normes de prononciation dans une com- munauté linguistique. Les indices acoustiques sur lesquels se focalise cette thèse sont le délai d’établissement du voisement (DEV) et la fréquence fondamentale (f0), qui sont des indices covariantes permettant de classer les plosives. La covariation de ces indices est d’un intérêt particulier car elle est liée à un type de changement de son commun nommé la ‘transphonologisation’, où l’importance relative des indices acoustiques change au cours du temps. Il y a beaucoup de questions non résolues en ce qui concerne la covariation des indices acoustiques dans la production de la parole, y compris: (1) quels sont le mécanisme et le chemin de transphonologisation; et (2) quels aspects structurels de cette covariation des indices ne sont pas uniques à une langue donnée et quels aspects structurels ne sont présents que lors des changements de son? Cette thèse aborde ces deux questions dans le cadre de deux études à grande échelle, pour lesquelles les données ont été extraites de corpus parlés de trois langues. La première étude examine comment a lieu la transphonologisation du DEV à la f0, comment ce phénomène se propage et quelles sont les conséquences du changement sur d’autres aspects du système phonologique de la langue. Cette question a été abordée grâce à des données du changement en temps apparent provenant d’un corpus de coréen de Séoul (le corpus NIKL). La chronologie du changement de DEV et de f0 a été examinée en se concentrant sur les facteurs linguistiques (la fréquence des mots et la hauteur des voyelles) et sociaux (l’âge et le sexe des locuteurs). Les résultats ont démontré d’abord qu’il existe un compromis entre les indices acoustique selon le locuteur et le contexte et ensuite que le changement de primauté des indices est plus avancé dans le parler des jeunes femmes et dans les conditions linguistiques où le contraste du DEV est réduit (dans les mots à haute fréquence et devant les voyelles non hautes). Ces résultats suggèrent que la réduction du DEV et le renforcement de la f0 se propagent en parallèle de manière adaptative entre les locuteurs et la langue, le tout motivé par un biais de production dans une dimension (le DEV) et l’expansion adaptative dans une autre (la f0). La deuxième étude, s’appuyant sur les résultats de la première, a cherché à mieux décrire comment les facteurs sociaux et linguistiques sont impliqués dans la structure de la covariation des indices dans les langues. En outre, visant à identifier les aspects de la covariation des indices qui sont propres aux langues en changement, la deuxième étude offre une comparaison entre le coréen de Séoul et deux langues qui ne sont pas 3 en cours de transphonologisation, soit l’anglais et l’allemand. L’analyse a examiné des données de corpus pour les trois langues, se focalisant principalement sur deux facteurs linguistiques (la fréquence des mots et la hauteur des voyelles), sur un facteur social (le sexe) et sur la variation entre les locuteurs individuels, ainsi que les effets de cette variation individuelle sur l’utilisation relative du DEV et de la f0 dans la signalisation des groupes de plosives. Nos modèles de régression ont estimé l’importance des facteurs linguistiques et sociaux. De plus, l’emploi des indices acoustiques (ce que l’on nomme également le «poids») pour chaque locuteur a été quantifié de plusieurs façons: le D de Cohen, des analyses de discriminantes linéaires et des machines à vecteurs de support. En allemand et en anglais, un indice acoustique est systématiquement moins utilisé dans les mots à haute fréquence et devant les voyelles non hautes, réduisant ainsi la quantité d’information offerte par cet indice pour ce qui est des contrastes entre les plosives. De plus, ce sont les mêmes conditions où la transphonologisation en coréen de Séoul est la plus avancée. Les résultats pour les sexes et pour les individus, par contre, ont démontré qu’il existe des corrélations systématiques dans toutes les langues. Pris dans son ensemble, les résultats de ces deux études démontrent que le compromis d’importance relative des indices acoustiques associé aux contexte linguistique, observé uniquement en coréen de Séoul Coréen où l’on retrouve la transphonologisation, aurait pu s’instaurer pour contrer la réduction de « l’informativité » des indices pour éviter la perte de contraste, une tendence qui est probablement unique aux cas de transphonologi- sation. Cependant, les changements de l’importance des indices acoustiques selon le sexe et l’individu sont systématiques dans les trois langues, ce qui suggère que les locuteurs ont un but partagé par rapport à la quantité d’information linguistique qu’ils souhait- ent transmettre tout en produisant des variantes stylistiques socialement conditionnées. La variabilité structurée entre les locuteurs peut également servir de catalyseur pour la transphonologisation et ce changement peut se propager en renforçant ces structures ex- istantes. De façon générale, les résultats démontrent que la variabilité linguistique n’est pas aléatoire, mais qu’elle est plutôt structurée en plusieurs dimensions selon le contexte et le locuteur.
Recommended publications
  • Phonetic and Phonological Systems Analysis (PPSA) User Notes For
    Phonetic and Phonological Systems Analysis (PPSA) User Notes for English Systems Sally Bates* and Jocelynne Watson** *University of St Mark and St John **Queen Margaret University, Edinburgh “A fully comprehensive analysis is not required for every child. A systematic, principled analysis is, however, necessary in all cases since it forms an integral part of the clinical decision-making process.” Bates & Watson (2012, p 105) Sally Bates & Jocelynne Watson (Authors) QMU & UCP Marjon © Phonetic and Phonological Systems Analysis (PPSA) is licensed under a Creative Commons Attribution-Non-Commercial-NoDerivs 3.0 Unported License. Table of Contents Phonetic and Phonological Systems Analysis (PPSA) Introduction 3 PPSA Child 1 Completed PPSA 5 Using the PPSA (Page 1) Singleton Consonants and Word Structure 8 PI (Phonetic Inventory) 8 Target 8 Correct Realisation 10 Errored Realisation and Deletion 12 Other Errors 15 Using the PPSA (Page 2) Consonant Clusters 16 WI Clusters 17 WM Clusters 18 WF Clusters 19 Using the PPSA (Page 3) Vowels 20 Using the PPSA (Page 2) Error Pattern Summary 24 Child 1 Interpretation 25 Child 5 Data Sample 27 Child 5 Completed PPSA 29 Child 5 Interpretation 32 Advantages of the PPSA – why we like this approach 34 What the PPSA doesn’t do 35 References 37 Key points of the Creative Commons License operating with this PPSA Resource 37 N.B. We recommend that the reader has a blank copy of the 3 page PPSA to follow as they go through this guide. This is also available as a free download (PPSA Charting and Summary Form) under the same creative commons license conditions.
    [Show full text]
  • Lenition and Optimality Theory
    Lenition and Optimality Theory (Proceedings of LSRL XXIV, Feb. 1994) Haike Jacobs French Department Nijmegen University/Free University Amsterdam 1. Introduction Since Kiparsky (1968) generative historical phonology has relied primarily on the following means in accounting for sound change: rule addition, rule simplification, rule loss and rule reordering. Given that the phonological rule as such no longer exists in the recently proposed framework of Optimality theory (cf. Prince and Smolensky 1993), the question arises how sound change can be accounted for in this theory. Prince and Smolensky (1993) contains mainly applications of the theory to non-segmental phonology (that is, stress and syllable structure), whereas the segmental phonology is only briefly (the segmental inventory of Yidiny) touched upon. In this paper, we will present and discuss an account of consonantal weakening processes within the framework of Optimality theory. We will concentrate on lenition in the historical phonology of French, but take into account synchronic allophonic lenition processes as well. This paper purports to demonstrate not only that segmental phonology can straightforwardly be dealt with in optimality theory, but, moreover, that an optimality-based account of lenition is not thwarted by the drawbacks of previous proposals. This paper is organized as follows. In section 2, we will present the main facts of lenition in the historical phonology of French. After that, section 3 briefly discusses previous analyses of this phenomenon, mainly concentrating on their problematic aspects. Section 4 considers the possibilities of accounting for lenition in Optimality theory. Finally, in section 5 the main results of the present paper are summarized.
    [Show full text]
  • Phonological Processes
    Phonological Processes Phonological processes are patterns of articulation that are developmentally appropriate in children learning to speak up until the ages listed below. PHONOLOGICAL PROCESS DESCRIPTION AGE ACQUIRED Initial Consonant Deletion Omitting first consonant (hat → at) Consonant Cluster Deletion Omitting both consonants of a consonant cluster (stop → op) 2 yrs. Reduplication Repeating syllables (water → wawa) Final Consonant Deletion Omitting a singleton consonant at the end of a word (nose → no) Unstressed Syllable Deletion Omitting a weak syllable (banana → nana) 3 yrs. Affrication Substituting an affricate for a nonaffricate (sheep → cheep) Stopping /f/ Substituting a stop for /f/ (fish → tish) Assimilation Changing a phoneme so it takes on a characteristic of another sound (bed → beb, yellow → lellow) 3 - 4 yrs. Velar Fronting Substituting a front sound for a back sound (cat → tat, gum → dum) Backing Substituting a back sound for a front sound (tap → cap) 4 - 5 yrs. Deaffrication Substituting an affricate with a continuant or stop (chip → sip) 4 yrs. Consonant Cluster Reduction (without /s/) Omitting one or more consonants in a sequence of consonants (grape → gape) Depalatalization of Final Singles Substituting a nonpalatal for a palatal sound at the end of a word (dish → dit) 4 - 6 yrs. Stopping of /s/ Substituting a stop sound for /s/ (sap → tap) 3 ½ - 5 yrs. Depalatalization of Initial Singles Substituting a nonpalatal for a palatal sound at the beginning of a word (shy → ty) Consonant Cluster Reduction (with /s/) Omitting one or more consonants in a sequence of consonants (step → tep) Alveolarization Substituting an alveolar for a nonalveolar sound (chew → too) 5 yrs.
    [Show full text]
  • The Phonology, Phonetics, and Diachrony of Sturtevant's
    Indo-European Linguistics 7 (2019) 241–307 brill.com/ieul The phonology, phonetics, and diachrony of Sturtevant’s Law Anthony D. Yates University of California, Los Angeles [email protected] Abstract This paper presents a systematic reassessment of Sturtevant’s Law (Sturtevant 1932), which governs the differing outcomes of Proto-Indo-European voiced and voice- less obstruents in Hittite (Anatolian). I argue that Sturtevant’s Law was a con- ditioned pre-Hittite sound change whereby (i) contrastively voiceless word-medial obstruents regularly underwent gemination (cf. Melchert 1994), but gemination was blocked for stops in pre-stop position; and (ii) the inherited [±voice] contrast was then lost, replaced by the [±long] opposition observed in Hittite (cf. Blevins 2004). I pro- vide empirical and typological support for this novel restriction, which is shown not only to account straightforwardly for data that is problematic under previous analy- ses, but also to be phonetically motivated, a natural consequence of the poorly cued durational contrast between voiceless and voiced stops in pre-stop environments. I develop an optimality-theoretic analysis of this gemination pattern in pre-Hittite, and discuss how this grammar gave rise to synchronic Hittite via “transphonologization” (Hyman 1976, 2013). Finally, it is argued that this analysis supports deriving the Hittite stop system from the Proto-Indo-European system as traditionally reconstructed with an opposition between voiceless, voiced, and breathy voiced stops (contra Kloekhorst 2016, Jäntti 2017). Keywords Hittite – Indo-European – diachronic phonology – language change – phonological typology © anthony d. yates, 2019 | doi:10.1163/22125892-00701006 This is an open access article distributed under the terms of the CC-BY-NCDownloaded4.0 License.
    [Show full text]
  • Himalayan Linguistics Segmental and Suprasegmental Features of Brokpa
    Himalayan Linguistics Segmental and suprasegmental features of Brokpa Pema Wangdi James Cook University ABSTRACT This paper analyzes segmental and suprasegmental features of Brokpa, a Trans-Himalayan (Tibeto- Burman) language belonging to the Central Bodish (Tibetic) subgroup. Segmental phonology includes segments of speech including consonants and vowels and how they make up syllables. Suprasegmental features include register tone system and stress. We examine how syllable weight or moraicity plays a determining role in the placement of stress, a major criterion for phonological word in Brokpa; we also look at some other evidence for phonological words in this language. We argue that synchronic segmental and suprasegmental features of Brokpa provide evidence in favour of a number of innovative processes in this archaic Bodish language. We conclude that Brokpa, a language historically rich in consonant clusters with a simple vowel system and a relatively simple prosodic system, is losing its consonant clusters and developing additional complexities including lexical tones. KEYWORDS Parallelism in drift, pitch harmony, register tone, stress, suprasegmental This is a contribution from Himalayan Linguistics, Vol. 19(1): 393-422 ISSN 1544-7502 © 2020. All rights reserved. This Portable Document Format (PDF) file may not be altered in any way. Tables of contents, abstracts, and submission guidelines are available at escholarship.org/uc/himalayanlinguistics Himalayan Linguistics, Vol. 19(1). © Himalayan Linguistics 2020 ISSN 1544-7502 Segmental and suprasegmental features of Brokpa Pema Wangdi James Cook University 1 Introduction Brokpa, a Central Bodish language, has a complicated phonological system. This paper aims at analysing its segmental and suprasegmental features. We begin with a brief background information and basic typological features of Brokpa in §1.
    [Show full text]
  • L Vocalisation As a Natural Phenomenon
    View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by University of Essex Research Repository L Vocalisation as a Natural Phenomenon Wyn Johnson and David Britain Essex University [email protected] [email protected] 1. Introduction The sound /l/ is generally characterised in the literature as a coronal lateral approximant. This standard description holds that the sounds involves contact between the tip of the tongue and the alveolar ridge, but instead of the air being blocked at the sides of the tongue, it is also allowed to pass down the sides. In many (but not all) dialects of English /l/ has two allophones – clear /l/ ([l]), roughly as described, and dark, or velarised, /l/ ([…]) involving a secondary articulation – the retraction of the back of the tongue towards the velum. In dialects which exhibit this allophony, the clear /l/ occurs in syllable onsets and the dark /l/ in syllable rhymes (leaf [li˘f] vs. feel [fi˘…] and table [te˘b…]). The focus of this paper is the phenomenon of l-vocalisation, that is to say the vocalisation of dark /l/ in syllable rhymes 1. feel [fi˘w] table [te˘bu] but leaf [li˘f] 1 This process is widespread in the varieties of English spoken in the South-Eastern part of Britain (Bower 1973; Hardcastle & Barry 1989; Hudson and Holloway 1977; Meuter 2002, Przedlacka 2001; Spero 1996; Tollfree 1999, Trudgill 1986; Wells 1982) (indeed, it appears to be categorical in some varieties there) and which extends to many other dialects including American English (Ash 1982; Hubbell 1950; Pederson 2001); Australian English (Borowsky 2001, Borowsky and Horvath 1997, Horvath and Horvath 1997, 2001, 2002), New Zealand English (Bauer 1986, 1994; Horvath and Horvath 2001, 2002) and Falkland Island English (Sudbury 2001).
    [Show full text]
  • The American Intrusive L
    THE AMERICAN INTRUSIVE L BRYAN GICK University of British Columbia The well-known sandhi phenomenon known as intrusive r has been one of the longest-standing problems in English phonology. Recent work has brought to light a uniquely American contribution to this discussion: the intrusive l (as in draw[l]ing for drawing and bra[l] is for bra is in southern Pennsylvania, compared to draw[r]ing and bra[r] is, respectively, in British Received Pronunciation [RP]). In both instances of intrusion, a historically unattested liquid consonant (r or l) intervenes in the hiatus between a morpheme-final nonhigh vowel and a following vowel, either across or within words. Not surprisingly, this process interacts crucially with the well- known cases of /r/-vocalization (e.g., Kurath and McDavid 1961; Labov 1966; Labov, Yaeger, and Steiner 1972; Fowler 1986) and /l/-vocalization (e.g., Ash 1982a, 1982b), which have been identified as important markers of sociolinguistic stratification in New York City, Philadelphia, and else- where. However, previous discussion of the intrusive l (Gick 1999) has focused primarily on its phonological implications, with almost no attempt to describe its geographic, dialectal, and sociolinguistic context. This study marks such an attempt. In particular, it argues that the intrusive l is an instance of phonological change in progress. Descriptively, the intrusive l parallels the intrusive r in many respects. Intrusive r may be viewed simplistically as the extension by analogy of a historically attested final /r/ to words historically ending in a vowel (gener- ally this applies only to the set of non-glide-final vowels: /@, a, O/).
    [Show full text]
  • An Analysis of I-Umlaut in Old English
    An Analysis of I-Umlaut in Old English Meizi Piao (Seoul National University) Meizi Piao. 2012. An Analysis of I-Umlaut in Old English. SNU Working Papers in English Linguistics and Language X, XX-XX Lass (1994) calls the period from Proto-Germanic to historical Old English ‘The Age of Harmony’. Among the harmony processes in this period, i-umlaut has been considered as ‘one of the most far-reaching and important sound changes’ (Hogg 1992, Lass 1994) or as ‘one of the least controversial sound changes’ (Colman 2005). This paper tries to analyze i-umlaut in Old English within the framework of the Autosegmental theory and the Optimality theory, and explain how suffix i or j in the unstressed syllable cause the stem vowels in the stressed syllable to be fronted or raised. (Seoul National University) Keywords: I-umlaut, Old English, Autosegmental theory, vowel harmony Optimality theory 1. Introduction Old English is an early form of the English language that was spoken and written by the Anglo-Saxons and their descendants in the area now known as England between at least the mid-5th century to the mid-12th century. It is a West Germanic language closely related to Old Frisian. During the period of Old English, one of the most important phonological processes is umlaut, which especially affects vowels, and become the reason for the superficially irregular and unrelated Modern English phenomenon. I-Umlaut is the conditioned sound change that the vowel either moves directly forward in the mouth [u>y, o>e, A>&] or forward and up [A>&>e].
    [Show full text]
  • A Study of Some English Phonological Rules and Their Processes
    A study of Some English Phonological Rules and Their Processes Idha Nurhamidah Fakultas Bahasa, Unissula Semarang Abstract This article discusses some English phonological rules and their phonological process. The rules dealt with in this article include aspiration, nasalization, vowel lengthening and regressive assimilation with fixed spelling—just to limit; despite the fact that there are more phonological rules. Each of the rules under study is described in terms of generative phonology followed data analysis to justify the application of the rules in line with the underlying forms. More significantly, the study will be able to give linguistic evidence that English has rules in their systems of sounds that are rule-governing and generative in nature Key words: phonological rules and processes, aspiration, nasalization, vowel lengthening, regressive assimilation, rule-governed INTRODUCTION Despite its high importance in the learning of English as a foreign language (EFL) in Indonesia, pronunciation is least paid attention to. This is understandable since EFL teachers assume that there are a lot of dictionaries equipped with phonemic or phonetic transcription for each word entry to which students can easily refer with respect to the pronunciation of a particular word. However, it is also true that the acquisition of a foreign language is much facilitated by pronunciation drill since the production of speech sounds can only be made perfect through drills. Pronunciation drills, in other words, are very important to train the speech organs for the production of at either the level of individual sounds or that of words, or even that of phrases and clauses, along with word stress and intonation patterns.
    [Show full text]
  • Toward a Unified Theory of Chain Shifting
    OUP UNCORRECTED PROOF – FIRST-PROOF, 04/29/12, NEWGEN !"#$%&' () TOWARD A UNIFIED THEORY OF CHAIN SHIFTING *#'+, -. ./,0/, !. I"#$%&'(#)%" Chain shifts play a major role in understanding the phonological history of English—from the prehistoric chain shift of Grimm’s Law that separates the Germanic languages from the rest of Indo-European, to the Great Vowel Shift that is often taken to define the boundary between Middle English and Early Modern English, to the ongoing chain shifts in Present-Day English that are used to estab- lish the geographical boundaries between dialect regions. A chain shift may be defined as a set of phonetic changes affecting a group of phonemes so that as one phoneme moves in phonetic space, another phoneme moves toward the phonetic position the first is abandoning; a third may move toward the original position of the second, and (perhaps) so on. Martinet (12(3) introduced the argument that chain shifts are caused by a need for phonemes to maintain margins of security between each other—so if a phoneme has more phonetic space on one side of it than on others, random phonetic variation will cause it to move toward the free space but not back toward the margins of security of phonemes on the other side. Labov (3414: chapter 5) gives a lucid exposition of the cognitive and phonetic argu- ments underlying this account of chain shifting. Despite their value as a descriptive device for the history of English, however, the ontological status of chain shifts themselves is a matter of some doubt. Is a 559_Nevalainen_Ch58.indd9_Nevalainen_Ch58.indd
    [Show full text]
  • Some /L/S Are Darker Than Others: Accounting for Variation in English /L/ with Ultrasound Tongue Imaging
    University of Pennsylvania Working Papers in Linguistics Volume 20 Issue 2 Selected Papers from NWAV 42 Article 21 10-2014 Some /l/s are darker than others: Accounting for variation in English /l/ with ultrasound tongue imaging Danielle Turton Follow this and additional works at: https://repository.upenn.edu/pwpl Recommended Citation Turton, Danielle (2014) "Some /l/s are darker than others: Accounting for variation in English /l/ with ultrasound tongue imaging," University of Pennsylvania Working Papers in Linguistics: Vol. 20 : Iss. 2 , Article 21. Available at: https://repository.upenn.edu/pwpl/vol20/iss2/21 This paper is posted at ScholarlyCommons. https://repository.upenn.edu/pwpl/vol20/iss2/21 For more information, please contact [email protected]. Some /l/s are darker than others: Accounting for variation in English /l/ with ultrasound tongue imaging Abstract The phenomenon of /l/-darkening has been a subject of linguistic interest due to the remarkable amount of contextual variation it displays. Although it is generally stated that the light variant occurs in onsets (e.g. leap) and the dark variant in codas (e.g. peel), many studies report variation in different morphosyntactic environments. Beyond this variation in morphosyntactic conditioning, different dialects of English have been reported as showing highly variable distributions. These descriptions include a claimed lack of dis- tinction in the North of England, a three-way distinction between light, dark and vocalised /l/ in the South-East, and a gradient continuum of darkness in American English. This paper presents ultrasound tongue imaging data collected to test dialectal and contextual descriptions of /l/ in English, providing hitherto absent instrumental evidence for different distributions.
    [Show full text]
  • Chapter X Positional Factors in Lenition and Fortition
    Chapter X Positional factors in Lenition and Fortition 1. Introduction This chapter sets out to identify the bearing that the linear position of a segment may have on its lenition or fortition.1 We take for granted the defi- nition of lenition that has been provided in chapter XXX (Szigetvári): Put- ting aside stress-related lenition (see chapter XXX de Lacy-Bye), the refer- ence to the position of a segment in the linear string is another way of identifying syllabic causality. Something is a lenition iff the effect ob- served is triggered by the specific syllabic status of the segment at hand. The melodic environment thus is irrelevant, and no melodic prime is trans- mitted between segments. Lenition thereby contrasts with the other family of processes that is found in phonology, i.e. adjacency effects. Adjacency may be defined physically (e.g. palatalisation of a consonant by a following vowel) or in more abstract terms (e.g. vowel harmony): in any event, as- similations will transmit a melodic prime from one segment to another, and only a melodically defined subset of items will qualify as a trigger. Posi- tional factors, on the other hand, are unheard of in assimilatory processes: there is no palatalisation that demands, say, "palatalise velars before front vowels, but only in word-initial position". Based on an empirical record that we have tried to make as cross- linguistically relevant as possible, the purpose of this chapter is to establish appropriate empirical generalisations. These are then designed to serve as the input to theories of lenition: here are the challenges, this is what any theory needs to be able to explain.
    [Show full text]