Papers

  • Doing Sociophonetics with LAP Data

    This is a presentation on how I've developed audio preprocessing for noise reduction in archival recordings to improve their inteligibility for human annotators, as well as for automated processors.

  • The study of variation

    The study and formalization of intra-speaker variation within variationist sociolinguistics has followed a largely parallel history with generative phonology, always borrowing heavily from the generative theories of the day. More recently, structured probabilistic variation has become enshrined as a fact-to-be-explained by any theory of human sound systems in more mainstream phonology. This chapter outlines this parallel history of variation study from its origins in dialectology, the evolution of modern variationist sociolinguistics, and the development of more contemporary variation focused phonological theory, as well as critiques that have been posed over this history. The chapter reviews in considerable detail how the original notion of ‘variable rule’ was elaborated and complexified, and how variation is treated in constraint-based approaches. It concludes with a look towards the future of variation study that is incorporating more insights from psycholinguistics.

  • Crosslinguistic perceptions of /s/ among English, French, and German listeners

    This study reports the results of a crosslinguistic matched guise test examining /s/ and pitch variation in judgments of sexual orientation and nonnormative masculinity among English, French, and German listeners. Listeners responded to /s/ and pitch manipulations in native and other language stimuli (English, French, German, and Estonian). All listener groups rate higher pitch guises as more gay- and effeminate-sounding than lower pitch guises. However, only English listeners hear [s+] guises as more gay- and effeminate-sounding than [s] or [s−] guises for all stimuli languages. French and German listeners do not hear [s+] guises as more gay- or effeminate-sounding in any stimulus language, despite this feature’s presence in native speech production. English listener results show evidence of indexical transfer, when indexical knowledge is applied to the perception of unknown languages. French and German listener results show how the enregistered status of /s/ variation affects perception, despite crosslinguistic similarities in production.

  • Toward “English” Phonetics: Variability in the Pre-consonantal Voicing Effect Across English Dialects and Speakers

    Recent advances in access to spoken-language corpora and development of speech processing tools have made possible the performance of “large-scale” phonetic and sociolinguistic research. This study illustrates the usefulness of such a large-scale approach—using data from multiple corpora across a range of English dialects, collected, and analyzed with the SPADE project—to examine how the pre-consonantal Voicing Effect (longer vowels before voiced than voiceless obstruents, in e.g., bead vs. beat) is realized in spontaneous speech, and varies across dialects and individual speakers. Compared with previous reports of controlled laboratory speech, the Voicing Effect was found to be substantially smaller in spontaneous speech, but still influenced by the expected range of phonetic factors. Dialects of English differed substantially from each other in the size of the Voicing Effect, whilst individual speakers varied little relative to their particular dialect. This study demonstrates the value of large-scale phonetic research as a means of developing our understanding of the structure of speech variability, and illustrates how large-scale studies, such as those carried out within SPADE, can be applied to other questions in phonetic and sociolinguistic research.

  • Age vectors vs. axes of intraspeaker variation in vowel formants measured automatically from several English speech corpora.

    To test the hypothesis that intraspeaker variation in vowel formants is related to the direction of diachronic change, we compare the direction of change in apparent time with the axis of intraspeaker variation in F1 and F2 for vowel phonemes in several corpora of North American and Scottish English. These vowels were measured automatically with a scheme (tested on hand-measured vowels) that considers the frequency, bandwidth, and amplitude of the first three formants in reference to a prototype. In the corpus data, we find that the axis of intraspeaker variation is typically aligned vertically, presumably corresponding to the degree of jaw opening for individual tokens, but for the North American GOOSE vowel, the axis of intraspeaker variation is aligned with the (horizontal) axis of diachronic change for this vowel across North America. This may help to explain why fronting and unrounding of high back vowels are common shifts across languages.

  • Is phonetic target uniformity phonologically, or sociolinguistically grounded?

    In this paper, I investigate to what degree phonetic uniformity in diachronic vowels shifts can be accounted for in terms of a shared phonetic implementation rule of phonological features [6, 10], versus a shared social evaluation of the phonetic realizations [19]. I take a particular focus on the parallel fronting and subsequent retraction of the GOOSE, GOAT and MOUTH vowels, as well as the raising of the preconsonantal FACE and pre-voiceless PRICE vowels in Philadelphia, drawing data from the Philadelphia Neighborhood Corpus [15]. Using generalized additive models [21] I fit models for these vowels accounting for gender, date of birth, educational attainment, and vowel duration using tensor product smooths. Looking at the correlation of the byspeaker random intercepts, back vowel fronting appears to be highly correlated, thus likely phonologically grounded, while FACE and PRICE raising is not, thus likely socially grounded.

  • Using the Tolerance Principle to predict phonological change

    Language acquisition is a well-established avenue for language change (Labov, 2007). Given the theoretical importance of language acquisition to language change, it is all the more important to formulate clear theories of transmission-based change. In this paper, we provide a simulation method designed to test the plausibility of different possible transmission-based changes, using the Tolerance Principle (Yang, 2016) to determine precise points at which different possible changes may become plausible for children acquiring language. We apply this method to a case study of a complex change currently in progress: the allophonic restructuring of /æ/ in Philadelphia English. Using this model, we are able to evaluate several competing explanations of the ongoing change and determine that the allophonic restructuring of /æ/ in Philadelphia English is mostly likely the result of children acquiring language from mixed dialect input, consisting of approximately 40\% input from speakers with a nasal /æ/ split. We show that applying our simulation to a phonological change allows us to make precise quantitative predications about the progress of this change. Moreover, it forces us to reassess intuitively plausible hypotheses about language change, such as grammatical simplification, in a quantitative and independently motivated framework of acquisition.

  • The Role of Phonology in Phonetic Change

    This article reviews the role phonology plays in phonetic changes. After first establishing what kinds of changes qualify as phonetic changes for the purposes of discussion, and laying out the theoretical outlook that is adopted here, I review the most obvious cases in which phonology plays a role in phonetic change. These include (a) the way phonological contrast can lead to phonetic dispersion, (b) the way phonological natural classes can define a set of segments to undergo a parallel phonetic shift, and (c) how phonological biases may lead to instances of underphonologization. Throughout, I discuss alternative approaches to these phenomena.

  • Generations, lifespans, and the zeitgeist

    This paper is equal parts methodological recommendation and an empirical investigation of the time dimensions of linguistic change. It is increasingly common in the sociolinguistic literature for researchers to utilize speech data that was collected over the course of many decades. These kinds of datasets contain three different time dimensions that researchers can utilize to investigate language change: (i) the speakers' dates of birth, (ii) the speakers' ages at the time of the recording, and (iii) the date of the recording. Proper investigation of all three time dimensions is crucial for a theoretical understanding of the dynamics of language change. I recommend utilizing two-dimensional tensor product smooths, fit over speakers' date of birth and the year of the recording, to analyze the contribution of these three time dimensions to linguistic changes. I apply this method to five language changes, based on data drawn from the Philadelphia Neighborhood Corpus. I find relatively weak evidence for lifespan effects in these changes, robust generational effects, and in one case, evidence of a zeitgeist effect.

  • Response to Berkson, Davis, \& Strickler, ‘What does incipient /ay/-raising look like?’

    Berkson, Davis, and Strickler (2017) provide an invaluable report on incipient /ay/-raising in Fort Wayne, Indiana. Their data suggest that /ay/-raising conditioned strictly by phonetic voice-lessness is a possible early stage in the development of /ay/-raising. This raises a particularly vexing question of why /ay/-raising has gone on to be conditioned by phonological voicing in all North American varieties for which its interaction with /t, d/ flapping has been examined. It suggests that the process of phonologization reorganizes the distribution of phonetic variants, rather than simply discretizing phonetic precursors.

  • The early influence of phonology on a phonetic change

    The conventional wisdom regarding the diachronic process whereby phonetic phenomena become phonologized appears to be the ‘error accumulation’ model, so called by Baker, Archangeli, and Mielke (2011). Under this model, biases in the phonetic context result in production or perception errors, which are misapprehended by listeners as target productions, and over time accumulate into new target productions. In this article, I explore the predictions of the hypocorrection model for one phonetic change (prevoiceless /ay/-raising) in detail. I argue that properties of the phonetic context underpredict and mischaracterize the contextual conditioning on this phonetic change. Rather, it appears that categorical, phonological conditioning is present from the very onset of this change.

  • Variation and Change in the Use of Hesitation Markers in Germanic Languages

    In this study, we investigate crosslinguistic patterns in the alternation between um, a hesitation marker consisting of a neutral vowel followed by a final labial nasal, and uh, a hesitation marker consisting of a neutral vowel in an open syllable. Based on a quantitative analysis of a range of spoken and written corpora, we identify clear and consistent patterns of change in the use of these forms in various Germanic languages (English, Dutch, German, Norwegian, Danish, Faroese) and dialects (American English, British English), with the use of um increasing over time relative to the use of uh. We also find that this pattern of change is generally led by women and more educated speakers. Finally, we propose a series of possible explanations for this surprising change in hesitation marker usage that is currently taking place across Germanic languages.

  • Filled Pause Choice as a Sociolinguistic Variable

    In this paper, I argue that filled pause selection (um/uh) is a sociolinguistic variable, conditioned by both internal and external factors. There appears to be a language change in progress towards selecting um more often than uh. In all respects, the (UHM) variable appears to pattern quantiatively just like all other sociolinguistic variables which have been examined, even though the locus of (UHM) variation would seem to be firmly in the speech planning domain. Combined with the quantitative systematicity of sociolinguistic variables across the full range of linguistic modules, I argue that the locus of variation may not be in the grammar, but rather constitutes a separate domain of knowledge, perhaps what Preston (2004) called the “sociocultural selection device.”

  • An Evaluation of Sociolinguistic Elicitation Methods

    This study investigates the effects of different elicitation methods on the speech of a single speaker of San Francisco English who is participating in a systematic set of vocalic sound changes known as the California Vowel Shift [6]. We contrast data obtained from classic sociolinguistic interview methods with data from self-recordings, as well as data from various methods for eliciting spontaneous speech that are typically used in laboratory settings.

  • I’m done my homework—Case assignment in a stative passive

    We present an analysis of an understudied construction found in Philadelphian and Canadian English, and also in certain Vermont varieties. In this construction, the participle of certain verbs can appear along with a form of the verb be and a DP complement, producing strings like I’m done my homework , I’m finished my fries , and (in Vermont) I’m started the project . We show that the participle in the construction is an adjectival passive, not a perfect construction. We further argue that the internal argument DP in the construction is receiving Case from the adjectival head a , similar to what happens in all English dialects with the adjective worth , and that the internal argument is interpreted via a mechanism of complement coercion. The microparametric variation we find across English dialects with respect to the availability of this construction is accounted for by variation in the selectional restrictions on the a head.

  • One hundred years of sound change in Philadelphia: Linear Incrementaion, Reversal, and Reanalysis

    The study of sound change in progress in Philadelphia has been facilitated by the application of forced alignment and automatic vowel measurement to a large corpus of neighborhood studies, including 379 speakers with dates of birth from 1888 to 1991. Two of the sound changes active in the 1970s show a linear pattern of incrementation in succeeding decades. The fronting of back upgliding vowels /aw/ and /ow/ shows a reversal in the direction of change, beginning with those born after 1940. The study also finds a general withdrawal from two salient features of local phonology, tense /æh/ and /oh/, led by those with higher education. Younger speakers with higher education have also reorganized the traditional Philadelphia tense/lax split of short-a to form a nasal system with tensing before all and only nasal consonants. The development of the Philadelphia vowel system can be understood in the geographic context of neighboring dialects. Features in common with North and North Midland dialects have accelerated in use while features in common with South Midland and Southern dialects have been reversed in favor of Northern patterns. The microevolution of a linguistic system can be seen here as subject to phonological generalizations but driven by social evaluation as features rise in level of salience for members of the speech community.

  • Phonological Rule Change: The Constant Rate Effect

    The detailed quantitative study of language change, as found in studies such as Labov (1994) and Kroch (1989), has raised two central questions for linguistic theory. The first is an issue in the theory of language change itself, namely: do changes in different components of the grammar progress in the same way? The second question addresses the relationship between the study of change and the development of synchronic linguistic theory: can quantitative, diachronic data help to choose between alternative analyses of synchronic facts? This paper addresses both of these questions with the case study of the loss of word-final stop fortition (frequently termed "devoicing") in the history of German, and concludes that the answer to both questions above is "yes".

  • Redevelopment of a Morphological Class

    Coronal stop deletion (or‚`TD Deletion‚`) is the paradigm sociolinguistic variable. It was first described in African American English (Labov et al., 1968) as a rule whereby word final /Ct/ and /Cd/ clusters simplify by deleting the coronal stop. It has since been found in many dialects and varieties of English. Aside from the very regular phonological and phonetic factors which condition whether TD Deletion applies, morphological structure also appears to have an effect. The three morphological categories of primary interest are (i) monomorphemes\vphantom{\{}\}, (ii) regular past tense verbs and (iii) semiweak past tense verbs. In almost every dialect studied, the order of morphological classes from least favoring deletion to most favoring deletion is as given in (1). (1) monomorphemes {\textgreater} semiweak {\textgreater} regular past tense In this paper, I will be focusing on the difference between semiweak and regular past tense. I will pursue a revised version of the analysis in Guy \& Boyd (1990), casting it in terms of Competing Grammars and Distributed Morphology. Specifically, I will propose that the rate of phonological TD Deletion is the same for the regular past and the semiweak. What leads to higher TD Absence in the semiweak verbs is variable morphological absence of /t/, i.e., there is a competing morphological analysis where the past tense of keep is simply "kep", instead of "kept".

  • Cross-derivational feeding is epiphenomenal

    Baković (2005) proposes that patterns of sufficiently-similar segment avoidance are the result of interacting agreement and antigemination constraints, a pattern known as cross-derivational feeding (CDF). The bleeding interactions between epenthesis and assimilation which prevent adjacent sufficiently-similar segments in English are shown to follow, however, from extragrammatical considerations. Several case studies provide evidence against the major predictions of CDF.

  • The Spread of Raising : Opacity , Lexicalization , and Diffusion

    The centralization of the low upgliding diphthong (typically called Canadian Raising, here just Raising), is frequently cited as an example of phonological opacity. Conditioned by a following voiceless segment, Raising continues to apply when an underlying unstressed /t/ is flapped on the surface. Dialects which have both Raising and Flapping, then, maintain the distinction between "writer" and "rider" in the quality of the vowel, rather than the voicing of the stop. Exceptions to the simplest formulation of Raising have been reported on in the past. Underapplication of Raising in pre-voiceless environments can possibly be accounted for by prosodic structure (Chambers, 1973, 1989; Jensen, 2000; Vance, 1987). However, a few reports from the Inland North (Vance, 1987; Dailey-O'Cain, 1997) and Canada (Hall, 2005) suggest that the regularity of Raising's conditioning has deteriorated, allowing raised nuclei before underlyingly voiced segments. The distribution of these raised variants is unpredictable within a speaker's phonology, but stable for given words, suggesting that Raising has lexicalized, and is undergoing diffusion to new environments. This paper focuses on the phonological status of Raising in Philadelphia. Raising was identified as an incipient sound change in progress in the LCV study of the 1970s, and has been revisited for study in connection with its masculine association (Labov, 2001; Conn, 2005; Wagner, 2007). After examining data from 12 boys, ages 14 through 19, it appears that Raising has lexicalized here as well. [{\textasciicircum}y] frequently appears before underlyingly voiced stops, as well as before nasals, but not in a phonologically predictable manner. Certain words seem to be selected for consistent overapplication however. "Spider" and "cider" are lexical items with raised nuclei for which there is broad agreement between speakers. However, there are also a number of lexical items which show more interspeaker variation, such as "tiny", produced variably as [tayni] or [t{\textasciicircum}yni]. Importantly, across all of the data, the effect of the lexical item on overapplication of Raising is stronger and more significant than the effect of surrounding phonological environment.

  • The Spread of Raising: Opacity, lexicalization, and diffusion

    Canadian Raising is typically described as the centralization of the nucleus of /ay/ before voiceless segments. However some recent studies in areas affected by Raising have shown that the current conditioning factors are not as regular as reported previously (Vance, 1987; Dailey-O’Cain, 1997; Hall, 2005). This paper explores the status of Raising in Philadelphia. Examining data from 12 boys, ages 14 to 19, it appears that Raising has lexicalized here as well. While Raising occurs before a number of voiced stops and nasals, the words which experience Raising most regularly suggest that it has spread due to its opaque applications.

  • .bib