previous day
all days

View: session overviewtalk overview

09:00-09:50 Session 11: Plenary Session IV
Location: New York
The moving target of language variation

ABSTRACT. Language variation can be described in spatial, temporal and social dimensions. The first one is the area of classical dialectology, the second of historical linguistics and the third of sociolinguistics. Social variation can often be understood as a more or less intentional expression of social group identity, whereas spatial variation is rather accidental and a confounder of spatial distance, which leads to less contact between communities of speakers and hence diversification. The linguistic expressions of variation are changing through time.

Social and spatial variation have been intertwined in a fascinating way in the history of the Dutch and Frisian language. Nearly half the speakers of Dutch in the Netherlands live in the areas that used to be Frisian-speaking in the Early Middle Ages. Dutch – as a fundamentally Franconian variety – and Frisian have existed side by side in Holland until the 17th century, and both have been in use in the present-day province of Fryslân – nowadays the core of the Frisian speaking regions – since at least the 15th century.

During the ages of contact, Frisian was mostly the less prestigious language of the two and over the centuries lost most of its area to Dutch. Some of the linguistic traces of Dutch were, however, shaped by its contact with Frisian. Frisian, in turn, absorbed enormous amounts of linguistic material from Dutch. Despite this mutual exchange of linguistic material, the two linguistic identities ‘Frisian’ and ‘Dutch’ remained intact. This lecture will illustrate how the cohabitation of these two languages developed, focusing on remarkable shifts in linguistic composition of the varieties and the linguistic identity of their speakers.

09:50-10:15Coffee Break
10:15-12:20 Session 12A: Acquisition
Location: Brussels
Norwegian linguistic profiles at the crossroads of oracy and literacy


The purpose of this talk is to systematize the language competencies of larger groups of Norwegian L1 users when taking into account both oral and written varieties. As we hope to unveil the situation is quite complex with possible ramifications for anyone who aims to investigate possible effects of bidialectalism in the Norwegian context, be it for example on scholastic or cognitive development, or both.

The following factors are essential: 1) There exit two official written standards of the language, Bokmål and Nynorsk, 2) Bokmål is more widely used than Nynorsk and dominates the linguistic landscape, with the consequence that users of the lesser used Nynorsk generally master Bokmål well whereas the converse does not hold, 3) the spoken variety sometimes referred to as Urban East Norwegian (UEN) is structurally speaking close to Bokmål whereas no single dialect has a similar status relative to Nynorsk, 4) children outside of Eastern Norway role play in UEN (or an approximation to it), 5) written dialect (or at least non-standard written practices) is widespread in private computer-mediated communication (CMC) outside of Eastern Norway, in particular among adolescents and young adults, 6) communities dominated by Nynorsk are mostly found in the rural parts of Western Norway and bordering areas in Eastern Norway, whereas local communities in Eastern, Central and Northern Norway are Bokmål dominated.  

As an example of the combined effects of the above a young user of Nynorsk from Western Norway will typically speak a dialect structurally different Bokmål, master both Nynorsk and Bokmål in writing, write dialect in CMC and have gone through a phase of playing in UEN. On the other hand, a young Bokmål user from the central parts of Eastern Norway will speak UEN, typically master only Bokmål and not Nynorsk, write Bokmål in CMC and will not have gone through a phase of role playing in a dialect different from his or her own. Contrasting with both of the former might be a young Bokmål user from Northern Norway who speaks a dialect notably different from UEN, writes in dialect in CMC, has gone through a phase of playing in UEN and who does not master Nynorsk particularly well.

These three linguistic profiles can be summarized as in table 1 where we label the first a bidialectal Nynorsk user, the second and monolectal Bokmål user and the third one a bidialectal Bokmål user.

Table 1: Three different linguistic profiles of Norwegian L1 users

                                                               Dialect (non-UEN)                    Bokmål                            Nynorsk

                                                        oral       written                     oral       written                     oral       written

1. Bidialectal Nynorsk                                                                                                       ?              

2. Monolectal Bokmål                                                                                                                              

3. Bidialectal Bokmål                                                                                                                      


We will argue that these linguistic profiles correspond to large groups of Norwegian L1 users, and we will also consider to what extent additional groups allowed by the typology also exist. We will furthermore discussthe importance of doing this exercise when exploring possible non-linguistic effects of language diversity.

Language shift in Norwegian schools


As noted by Trudgill (2002: 31) the Norwegian language situation is marked by an “enormous societal tolerance for linguistic diversity”, and, what is more, “linguistic diversity is officially recognized and protected”. Dialects are used by people in all layers of society and within all social domains, both public and private, and diglossia is rare (Røyneland 2009). In addition, there are two written norms of Norwegian, Nynorsk and Bokmål, both of which are used as primary written languages, while the well-recognised trends of migration and globalisation also make both English and minority languages present features in society. Thus, while making their own language choices, Norwegian youths must navigate through a highly varied linguistic landscape, with multiple linguistic practices.

This paper presents an ongoing doctoral project which explores various factors that may influence the language choices made by Norwegian 10th grade pupils. The study draws its empirical data from a digital questionnaire with 400 participants, and from focus group interviews with a selection of respondents. Of particular interest is the relation between the written languages Nynorsk and Bokmål. These are formally of equal status, but Nynorsk is the lesser used language, thus holding a weaker position in society. In school, pupils are expected to learn both languages, but from the 8th grade onwards, they may choose either Nynorsk or Bokmål as their primary written language. Previous research reveals traces of an ongoing language shift, where Nynorsk pupils shift to Bokmål either during lower or upper secondary school – particularly in the outskirts of the Nynorsk heartland (Øzerk & Todal 2013).

A first question, then, is whether (and to what extent) this trend is continuing. Secondly, if it is, we want to know why this is happening; which factors come into play when pupils choose one variety over the other? In trying to understand the complex web of motivating forces I draw on Fishman’s (1991) foundational work of language maintenance and shift, theories of language attitudes (Garrett 2010), language ideologies (Woolard 2008), and language and identity (Giles 1977). Keeping the multifaceted Norwegian linguistic situation in mind, the paper will particularly address the link between oral and written language choices, as well as the 10th grade pupils’ written language practices in formal (e.g. school texts) as well as informal (e.g. social network sites) contexts.



Garrett, P. (2010). Attitudes to Language. Cambridge University Press.

Fishman, J.A. (1991). Reversing language shift: Theory and practice of assistance to threatened languages. Clevedon: Multilingual Matters.

Røyneland, U. (2009). Dialects in Norway: catching up with the rest of Europe? IJSL 196/197, 7–31.

Giles, H.  (ed.) (1977). Language, Ethnicity, and Intergroup Relations, London: Academic Press.

Trudgill, P. (2002).  Sociolinguistic Variation and Change. Edinburgh University Press.

Woolard, K. (2008). Language and Identity Choice in Catalonia: The Interplay of Contrasting      Ideologies of Linguistic Authority. In Süselbeck et al. (eds.) Lengua, nación e identidad. Iberoamericana, 303–323.

Øzerk, K. & Todal, J. (2013). Written Language Shift among Norwegian Youth. International Electronic Journal of Elementary Education 5(3), 285–300.

Acquiring sociolinguistic competence: Austrian children and their standard-dialect repertoires


The German-speaking areas are characterized by considerable diatopic variation, determined by very different base dialects and by different relationships between the standard language and base dialects. The sociolinguistic settings have been conceptualized as a ‘diglossia’ (e.g., Switzerland), a ‘standard-dialect continuum’/‘diaglossia’ (Austria, Southern Germany) or as involving ‘dialect loss’ (Northern Germany) (Ammon 2003). For Bavarian-speaking Austria, a standard-dialect continuum has been observed on the community level, comprising, however, very diverse repertoires or patterns of usage on the individual level (Kaiser/Ender 2013).

The acquisition of variation has been a considerable research interest in other language areas, where 3-4-year-olds have been shown to use sociolinguistic variants of their L1, matching adults’ linguistic and stylistic constraints (e.g., Smith et al. 2007). Whereas standard-dialect variation in children has also been investigated in Switzerland (e.g., Häcki Buhofer et al. 1994) and Germany (e.g. Katerbow 2013), the question as to when and how this aspect of sociolinguistic competence is acquired by Austrian children is still pending.

In the present study, we collected speech data from 35 monolingual German-speaking children, aged between 3;4 and 6;4, in different interactional situations and with different interlocutors. Children were recorded twice while playing a game (‘Memory’) with other children and (1) a researcher speaking standard German and (2) a researcher speaking in the local dialect. They were also asked twice to re-tell stories to (1) a hand-puppet speaking standard German and (2) a hand-puppet speaking in the local dialect. Additionally, children were recorded while playing shop. These interactions were analysed for the proportions of standard, dialect, mixed, intermediate or ambivalent turns produced by the children. Thus, individual variational profiles could be traced which give insight into the children’s dialect-standard repertoires. As with adult speakers, patterns of variation can vary strongly between individuals. In the present sociolinguistic context, the most decisive factors seem to be diatopia (urban vs. rural area), age, and gender, which confirms the importance of these factors, which are known to constrain sociolinguistic variation in adults, also for the acquisition of sociolinguistic variation in children.



Ammon, U. (2003). Dialektschwund, Dialekt-Standard-Kontinuum, Diglossie: Drei Typen des Verhältnisses Dialekt – Standardvarietät im deutschen Sprachgebiet. In Androutsopoulos, J. K. & Ziegler, E. (eds.): „Standardfragen“, Soziolinguistische Perspektiven auf Sprachgeschichte, Sprachkontakt und Sprachvariation: 163-171. Frankfurt: Lang.

Häcki Buhofer, A. et al. (1994). Früher Hochspracherwerb in der Deutschen Schweiz: Der weitgehend ungesteuerte Erwerb durch sechs- bis achtjährige Deutschschweizer Kinder. In Burger, H. & Häcki Buhofer, A. (eds.): Spracherwerb im Spannungsfeld von Dialekt und Hochsprache: 147-198. Bern/Berlin: Peter Lang.

Kaiser, I. & Ender, A. (2013). Diglossia or dialect-standard continuum in speakers' awareness and usage. On the categorisation of lectal variation in Austria. In Pütz, M. et al. (eds.): Variation in Language and Language Use. Linguistic, Socio-Cultural and Cognitive Perspectives: 273-298, Frankfurt: Lang.

Katerbow, M. (2013) Spracherwerb und Sprachvariation: eine phonetisch-phonologische Analyse zum regionalen Erstspracherwerb im Moselfränkischen. Berlin: De Gruyter.

Smith, J. et al. (2007): "Mam, ma troosers is fa'in doon!" Community, caregiver and child in the acquisition of variation in Scottish dialect. Language Variation and Change 19/1: 63–99.

Finish your plate and clean up your language! A mixed-methods approach to standard and vernacular forms in child-directed control acts


This contribution studies variation in the way parents address their children, focusing on their choice for Colloquial Belgian Dutch (e.g. gij ‘you’) or Standard Belgian Dutch forms (e.g. jij ‘you’) of the second person pronoun. Particularly, we study parents’ style-shifts in control acts, i.e. “utterances designed to get someone else to do something” (Goodwin 2006: 517).

Linking the hyperstandardized linguistic situation in Flanders (Jaspers & Van Hoof 2013) with the Western-European ideal of democratic parenting (Pećnik 2007, see also Schaffer 1996), we expect to find the standard forms je/jij/jouw (Example 1) to be more typically connected to more indirect, softer control acts (the “best” language for the “best” type of parenting) than the vernacular forms ge/gij/uw (Example 2).

  1. Nu moeten wij jouw handje en jouw gezichtje een beetje wassen.
    (“now we have to wash your hands and your face a bit”)
  2. Zeg zeg zeg zeg zeg wa zijde [: zijt ge] daar nu weer aant [: aan het] doen?
    (“well well well well well what are you doing there now”)

This hypothesis is tested through a mixed method approach, where quantitative and qualitative analyses are used to study variant selection of ten Belgian Dutch parents. A total of 452 pronouns, isolated from a corpus of over 20 hours (>16,000 utterances) of self-recorded dinner table conversations (Blum-Kulka 1997), are subject to both regression trees and random forests (Baayen & Tagliamonte 2012) and to multimodal discourse analyses, scrutinizing the effect of parameters such as type of control act, repetition, mitigation and boosting and type of pronoun (Blum-Kulka 1997, Brumark 2010).

Results reveal how the choice for standard and vernacular is guided by an interplay of these features, nevertheless supporting our hypothesis that standard forms are reserved for “softer” control acts, additionally uncovering “irritation” as mediating factor for the use of the vernacular. 



Blum-Kulka, Shoshana. 1997. Dinner Talk: Cultural Patterns of Sociability and Socialization in Family Discourse. Mahwah, New Jerset: Lawrence Erlbaum.

Brumark, Åsa. 2010. Behaviour regulation at the family dinner table. The use of and response to direct and indirect behaviour regulation in ten Swedish families. Journal of Child Language 37: 1065-1088.

Goodwin, Marjorie Harness. 2006. Participation, affect, and trajectory in family directive/response sequences. Text & Talk - An Interdisciplinary Journal of Language, Discourse Communication Studies 26: 515-543.

Jaspers, Jürgen, and Sarah Van Hoof. 2013. Hyperstandardisation in Flanders. Extreme enregisterment and its aftermath. Pragmatics 23 (2): 331-359.

Pećnik, Nina. 2007. “Towards a vision of parenting in the best interest of the child”. In Mary Daly (ed.), Parenting in Contemporary Europe: a Positive Approach, 15-36. Strasbourg: Council of Europe Publications.

Tagliamonte, Sali, and Harald Baayen. 2012. Models, forests, and trees of York English: Was/were variation as a case study for statistical practice. Language Variation and Change 24: 135-178.

10:15-12:20 Session 12B: Dialect Levelling
Location: London
(t,d) Deletion and the Unsolved Problem of Morphological Effect in British English
PRESENTER: Carmen Ciancia


The sociolinguistic variable (t,d) – word-final (t,d) deletion in consonant clusters C(C)T/C(C)D – is  widely investigated in many US English dialects and, despite slight difference in constraint rankings, phonetic and morphological constraints were found to be largely uniform (Patrick, 1999). Conversely, in England, although little research on (t,d) has been conducted so far, conflicting results have been found: data from Manchester exhibit the usual “robust morphological effect” (Baranowski & Turton, 2016), with more deletion in monomorphemes (e.g. mist) than inflected forms (e.g. missed). However, data from York show that morphological class failed to reach statistical significance and monomorphemes disfavour deletion (Tagliamonte & Temple, 2005).

The lack of morphological effect found in York is one of the unsolved problems concerning (t,d) deletion. Guy (1991) argues that the probability of application of a variable deletion rule is conditioned by the morphological structure of a word, where deletion is most likely to occur in monomorphemes (e.g. mind), less likely in semi-weak verbs (e.g. left) and least likely in regular past tense verbs (e.g. called). It is also claimed (Santa Ana, 1996) that (t,d) is governed by a universal sonority hierarchy, such that less sonorous preceding segments (stops and fricatives) tend to favour the cluster reduction, whilst more sonorous preceding segments disfavour it.

This paper investigates (t,d) deletion in 36 East Anglian speakers from Colchester, Ipswich and Norwich stratified by class, sex and age. The data were gathered by means of sociolinguistic interviews, supplemented with reading passages and word lists. Elan was used to transcribe the interviews, whilst Praat was employed for critical cases. 4879 tokens (excluding those followed by /t,d/ stops and interdental fricatives) were coded and a mixed-effects Rbrul regression analysis was carried out, with speaker as a random effect. The linguistic factors examined include morphological class, preceding and following phonetic segment, voicing agreement, stress (on the consonant cluster and on the following segment) and style. These are marked as significant predictors, except for stress. None of the social factors is significant.

Results for morphological class match previous studies and fit the lexical phonology account. This constraint is statistically significant, even when preceding phonetic segment was included in the mixed-effects model: negative contractions (n 196) and monomorphemes (n 2204) favour deletion, whilst semi weak (n 687) and regular verbs (n 1792) disfavour it. Monomorphemes favour deletion even when negative contractions are excluded from the model. The word-final cluster reduction is also favoured when the following phonetic segment is a nasal, sibilant, stop or /l/, whilst a following glide, /r/, fricative, vowel and pause disfavour it. Findings for the preceding environment do not support the hypothesis that (t,d) is governed by a universal sonority hierarchy since nasals trigger the most deletion, followed by sibilants, then /l/ and finally obstruents and fricatives. (t,d) is also affected by style and voicing agreement, with heterovoiced tokens (e.g. bolt) favouring the reduction of final /t,d/ more than homovoiced (e.g. bold) tokens. This paper concludes by suggesting the emergence of morphological effect in British English, patterning with Manchester.


Baranowski, M. & Turton D. (2016). The morphological effect in British English T/D-deletion. Paper presented at the Linguistic Association of Great Britain, 7th September 2016, University of York.

Guy, G. R. (1991). Explanation in variable phonology: An exponential model of morphological constraints. Language Variation and Change 3, 1–22.

Patrick. P. L. (1999). Urban Jamaican Creole: Variation in the mesolect. (Varieties of English Around the World, G17.) Philadelphia & Amsterdam: John Benjamins.

Santa Ana, O. (1996). Sonority and syllable structure in Chicano English. Language Variation and Change 8(1): 63-89.

Tagliamonte, S., & Temple, R. (2005). New perspectives on an ol’ variable: (t,d) in British English. Language Variation and Change 17, 281–302.

Sociolinguistic variation in the Basque dialect of Garazi valley:preliminary data and results


The valley of Garazi is a rural area where most of its urban settlements are located in the French side of the Basque Country with the exception of Luzaide, which belongs to the Spanish state. This politico-administrative situation, together with the geographical distribution of population in these territories, have conditioned the area linguistically. The study of the Basque spoken in Garazi may be relevant for sociolinguistics due to the fact that several factors such as the importance of the political border between Luzaide and the rest of the towns, the loss of the transmission of the Basque language at home, the creation of the Basque standard “Batua”, the insertion of Basque in education and dialect leveling, are contributing to linguistic instability in the area, triggering changes in the way of speaking in terms of choice and use. However, very few studies have been carried out from the point of view of variationist sociolinguistics in Garazi valley and in the east of the Basque country, in general. We could only mention the project called Norantz (Oihartzabal, 2009) and the study of Santazilia in Luzaide (2009).

The aim of this work is to analyze the sociolinguistic situation of the Basque dialect used in the rural valley of Garazi both from the point of view of traditional and most recent variationist approaches (Labov 1966; Eckert 2012). Our main assumption is that the Basque variety of this valley is changing fast, and the specific hypotheses are that these developments are led by extralinguistic factors such as geography and socio-demography (age and gender). Regarding geography, the Basque variety from Luzaide is becoming different from the rest of the villages due to the political frontier. In terms of age, whereas oldest people’s way of speaking is close to the base dialect, young are having big influences of the standard, supralocal varieties and French or Spanish. Finally, regarding gender, older women seem to be more conservative and younger more innovative than men owing to the different situation that has experimented this gender in different generations.

Regarding methodology, we have decided to use the friends’ friend technique in our fieldwork in order to guarantee the obtainment of spontaneous speech. The interviews have been semi-guided in order to speak about similar subjects with everyone, but respecting the direction taken by each informant. Preliminary results are proving initial hypothesis, which a more large-scale study will have to confirm.



Eckert, Penelope. (2012). Three waves of variation study: The Emergence of meaning in the study of sociolinguistic variation. Annual Review of Anthropology 41, 87-100.

Labov, William. (1966). The Social Stratification of English in New York City. Washington D.C.: Center for Applied Linguistics.

Oyharçabal, Beñat. (2009). Norantz proiektua datu-basea. Available at www.norantz.org (last accessed 16 October 2018).

Santazilia, Ekaitz. (2009). “Luzaideko hizkuntz bariazioa”. FLV 41, 219-248.

Under Pressure: East Anglian English


Rapid shrinkage of the East Anglian dialect area in recent years has resulted in a classic “dialect death situation” (Trudgill, 1986:68). Despite this, Norfolk and the northernmost part of Suffolk, both linguistic sub-zones of East Anglia, show a resistance to the long-mid vowel mergers undergone in Early Modern English, whereby contrast between the long monophthongs /ɛː, oː/ and the diphthongs /ɛi, ɔu/ was lost (Wells, 1982). Trudgill notes the FACE merger reached completion across all of Suffolk in the 70s, while through lexical transfer tied to working class speakers and exposure to London varieties, the same is true for the corresponding GOAT vowel, but only in the south of the county (1978).

With the geographic scope of the GOAT merger uncertain, this paper therefore returns to the earlier assumption that linguistic diffusion between south and north Suffolk would be gradual (Trudgill, 1974). It examines how, why, and to what extent the northern dialect area of Suffolk has resisted the long-mid vowel merger, which resulted in the GOAT set, and has instead maintained contrast between [ʊu] and [ʌu] to differentiate, for example, ‘road’ and ‘rowed’. It assesses the nature of these mergers, their character vis-à-vis ongoing development through phonetic approximation, as well as their systemic motivations (external vs. internal). This paper will also explore these implications for both GOOSE and GOAT fronting in a dialect where the GOAT vowel remains variable, and the GOOSE vowel is recorded as fronted, as in [ʉ], in traditional forms of the dialect.

Data from 18 speakers native to Lowestoft (north Suffolk) are drawn from phonologically controlled reading passages, and the results of acoustic analysis of formant values (drawn from 25% and 50% points for diphthongs and monophthongs respectively) are normalised using the Labov ANAE method. Results are presented according to gender and across three age categories. Statistical analyses using Pillai scores highlight a change in progress where maintenance of the GOAT distinction is almost categorical for older speakers, gradient for middle speakers, and lost for younger speakers. GOOSE fronting is most advanced in younger speakers, while those speakers who maintain two distinct GOAT vowels are shown to treat each vowel differently with regard to GOAT fronting.



Trudgill, P. (1974). Linguistic Change and Diffusion: Description and Explanation in Sociolinguistic Dialect Geography. Language in Society, 3(2), 215–246.

Trudgill, P., & Foxcroft, T. (1978). On the Sociolinguistics of Vocalic Mergers: Transfer and Approximation in East Anglia. In P. Trudgill (Ed.), Sociolinguistic Patterns in British English (69–79). London: Edward Arnold.

Trudgill, P. (1986). Dialects in Contact. Oxford: Blackwell.

Wells, J. (1982). Accents of English II. Cambridge: Cambridge University Press.

Variation and change in four vowels of Achterhoeks


The Achterhoeks dialect, spoken in the eastern Netherlands, differs considerably from Standard Dutch (ABN) in its phonology and lexicon. Previous research focused on covering lexical differences (Schaars, 1987; Van Prooije, 2011), but less exists on its phonology. Additionally, there has been research into change in other Dutch dialects, such as Brabants (Swanenberg, 2009) and Limburgs (Hinskens, 1992), but not so much specifically considering Achterhoeks, especially in recent times.

This research focuses on the changes in the F1/F2 values of four Achterhoeks vowels between the years of 1979 and 2015. In order to compare them to their Standard Dutch counterparts, it was necessary to create lexical sets (titled PRAAT, KAAS, KIJK and HUIS) in which to describe the vowels in question. Achterhoeks speakers were recording translating sentences written in Standard Dutch into their dialect, as well as completing a picture task in order to gain a more spontaneous speech sample. The sentences used were designed to elicit dialectal pronunciations; they formed part of an earlier study by Leendert van Prooije in 1979, and so were reused in 2015 in order for the results to be comparable to the earlier study. The research focuses on speakers’ conscious representation of dialect in order to observe how what it means to speak in dialect differs from person to person; however, there was also the inclusion of a picture task in the 2015 research in order to gain a more spontaneous speech sample in addition to the data obtained through the sentence reading task.

Subtle changes in the Achterhoeks vowels were observed, suggesting a lack of stability in the use of the vowels in the dialect. Although the PRAAT and KIJK vowels showed little variation over time, the vowel in KAAS, historically realised as [e:] or [i:], appeared to be lowering towards the Standard Dutch [a:]. However, the most noticeable difference within the Achterhoek area, showing the most statistical significance, occurs with the pronunciation of the HUIS vowel when it appears after /r/, realised as either [u] or [y], with [y] acting as an intermediate form between traditional dialectal [u] and Standard [œy]. The overall picture supports the view that there is convergence on the standard variety within the Achterhoek, but that this is a gradual change which is affecting vowels differently. This research suggests that if Achterhoeks continues on its present course, it is likely that we will see the formation of a somewhat intermediate variety which displaces the older dialect.


References (Style: Heading 2):

Hinskens, F. (1992). Dialect levelling in Limburg. Structural and sociolinguistic aspects. PhD- thesis, University of Nijmegen

Schaars, L. (1987). Woordenboek van de Achterhoekse en Liemerse Dialecten. Doetinchem: Staring Instituut.

Swanenberg, J. (2009). Van Alterande Sorte. Brabants tussen Dialect en Standardtaal. Tilburg: Universiteit van Tilburg.

Van Prooije, L.A. (2011). [1984]. De Vakleu en et Vak (5th ed.). Netherlands: www.vakleu.nl

“The Cardiff accent will be gone … we’ll all sound the same”: dialect levelling in Cardiff English


This paper will present results from a research project investigating language change in Cardiff English (CE). Cardiff is Wales’ biggest city with a population of 357,000 and experiences the highest levels of in-migration in Wales (Markaki 2017: 3). However, there is a lack of recent sociolinguistic research in CE, besides Mees and colleagues’ real-time studies with data collection points between 1977 and 2011 (Mees and Collins 1999; Mees and Osorno 2015). Work in other urban centres in the UK has shown a reduction in the use of some traditionally local variants in favour of ‘supralocal’ forms amongst younger speakers, known as dialect levelling. The current paper will compare change across apparent time in two local CE features that have not previously been analysed in terms of generational differences:

  1. The (ing) variable: multisyllabic words ending -ing. Non-standard realisation /ɪn/ is well-documented in CE as well as other Englishes, but is rarely studied in relation to age as it is considered a stable sociolinguistic variable. This feature is not expected to show variation in apparent time, and has been analysed to provide current data on CE.
  2. The ‘Kerdiff A’: the BATH, TRAP, and PALM lexical sets show considerable inter- and intra-speaker variation from /a ~ æː/, with /ɛː/ as the broadest realisation. Fronter variants are stigmatised and produced as /eː/ in performative Cardiff accents (Coupland 1988).

Tokens have been extracted from a corpus consisting of transcribed speech from 20 interviews, comprising ten speakers in two age groups (ages 18-30 and 65+) with an equal gender split. Of the ten younger speakers, five attended Welsh-medium secondary schools. Variation in the two features is being analysed auditorily for (ing) and acoustically for the ‘Kerdiff A’.

Results from 2160 tokens of (ing) show that younger speakers use significantly less of the non-standard variant /ɪn/ than older speakers (χ2 = 42.92, p < 0.01). Furthermore, younger speakers educated in Welsh use significantly less /ɪn/ than English-educated speakers (χ2 = 74.23, p < 0.01). Analysis is ongoing for the ‘Kerdiff A’ but these initial results are unexpected and suggest a turn away from the traditional non-standard variant in younger speakers and those educated in Welsh. Both variables will be discussed in relation to Cardiff’s changing social context, and whether similar processes of dialect levelling can be seen here as with other cities in the UK.



Coupland, N. (1988) Dialect in use: sociolinguistic variation in Cardiff. Cardiff: University of Wales Press.

Markaki, Y. (2017) Migration flows and population trends in Wales. Available at: https://welshrefugeecouncil.org.uk/sites/default/files/msiw/pdf/MSiW%20Migration_Flows_Population_Trends.pdf [12/10/2018]

Mees, I. M. and Collins, B. (1999) Cardiff: a real-time study of glottalisation. In Foulkes, P. and Docherty, G. (eds.): Urban voices: accent studies in the British Isles: 185-202. London: Routledge.

Mees, I. and Osorno, C.H. (2015) Cardiff English: a real time study of stability and change between childhood and mid-adulthood. English Language Overseas Perspectives and Enquiries 12 (2): 53–77.

10:15-12:20 Session 12C: Syntactic Variation
Location: Madrid
A Sociolinguistic Study of Pronominal Expression in Mainland Colombian Spanish


This paper sheds light on the effects of the verb on Spanish subject pronoun expression (SPE). We explore eight predictors using 3,157 tokens from thirty residents of Medellín, Colombia. Interestingly, the overall pronominal rate (30%) is higher than those in other mainland communities (cf. Carvalho, Orozco, & Shin 2015; Lastra & Martín-Butragueño 2015; Otheguy & Zentella 2012, among others). The internal conditioning reveals person/number of the subject as the strongest predictor. Findings for verb class uncover that copulative, perception, and motion verbs, respectively, promote overt subjects. Although the favorable effects of copulative and perception verbs are consonant with previous findings (Enríquez 1984:240; Orozco 2015:24; Otheguy & Zentella 2012:164; Travis 2007:115; among others), the effects of motion, speech, and cognition verbs, respectively, are not consistent with findings in other communities. Inconsistencies in the effects of most verb classes compared across speech communities suggest that the effects of the verb on SPE may lack the overarching uniformity evidenced by other internal predictors (cf. Carvalho et al. 2015 and references therein; Torres-Cacoullos & Travis 2018). Thus, we tested verbs as standalone factors in multivariate regressions. The lexical effect of the verb provides more definite answers regarding how verbs condition SPE than the semantically-guided approaches used for four decades (cf. Bentivoglio 1980; Enríquez 1984). For instance, within copulative verbs ser ‘be’ favors overt subjects but estar ‘be’ favors null subjects. These findings imply that verb class, or rather classifications based on syntactic or semantic criteria, do not constitute the most accurate way to explore the effects of the verb on SPE (cf. Erker & Guy 2012; Posio 2015; among others). Moreover, the robust effect of age sets Medellín apart from most other monolingual Hispanic speech communities, as social factors do not consistently constitute strong SPE predictors (Carvalho et al. 2015:xv). Interestingly, the lower pronominal rate among younger speakers is consistent with findings in other monolingual Spanish varieties such as Peninsular (Prada 2015), Mexico City (Lastra & Martín-Butragueño 2015), Colombian (Orozco 2015), and Dominican Spanish (Alfaraz 2015). This finding appears to have cognitive and language acquisition implications, as younger speakers would be expected to have higher pronominal rates. In sum, this study contributes to expand our baseline knowledge of SPE in Latin American mainland speech communities.  



Bentivoglio, Paola. (1980). Why canto and not yo canto? The problem of first-person subject pronoun in spoken Venezuelan Spanish. Los Angeles: University of California M.A. Thesis.

Carvalho, Ana, Rafael Orozco, & Naomi Shin. (2015). Subject Pronoun Expression in Spanish: A cross-dialectal perspective. Washington DC: Georgetown UP.

Enríquez, Emilia. (1984). El pronombre personal sujeto en la lengua española hablada en Madrid. Madrid: Consejo Superior de Investigaciones Científicas.

Lastra, Yolanda & Pedro Martín-Butragueño. (2015). Subject Pronoun Expression in Mexican Spanish. In M. Carvalho et al., 41-59. Washington DC: Georgetown.  

Otheguy, Ricardo & Ana Celia Zentella. (2012). Spanish in New York: Language contact, dialectal leveling, and structural continuity. Oxford: Oxford UP.

Torres-Cacoullos, Rena & Catherine Travis. (2018). Bilingualism in the Community Code-switching and Grammars in Contact. Cambridge, UK.: Cambridge UP.

Genitive attributes and the von-construction in German: unravelling the threads


Similar to English (e.g. Rosenbach 2003, 2014, Szmrecsanyi 2010), modern German shows variation between two types of nominal attributes: The synthetic genitive phrase and the analytical prepositional phrase with von ‚of‘.


a.  ein  düsteres  Bild            von  der    Ineffizienz

      a    somber  impression  of  art.dat  inefficiency

     ‚a    somber impression        of         inefficiency‘


b.  ein  Bild              der      Inkonsequenz

     an  impression  art.gen    inconsistency

    ‚an impression      of        inconsistency‘

In addition, proper names resist the diachronic trend to postpose genitive attributes to a certain degree, exhibiting positional variation:


a.  eine  Komposition,  die      [Mannheims Stadtplan]        reflektiert

      a      composition   which         Mannheim.gen         street map reflects

     ‚a      composition   which       reflects Mannheim's        street map‘

b.     [dem  quadratischen  Grundriss  Mannheims]  nachempfunden

         art.dat  square layout       Mannheim.gen              imitated

        ‚imitating     Mannheim's   square layout‘

Up to now, the phenomenon has been studied for historical German (most recently Scott 2014), for a sample of proper names (Peschke 2014) and for possessive constructions (Lang 2018). Other studies have focused on single factors to find categorical (i.e. invariable) contexts in the sense of Rosenbach (2003) (e.g. Smith 2003). A comprehensive, rigorous analysis of all choice contexts, all possible factors and their interactions is, however, still lacking.

The present study seeks to close this gap. It is part of a new, corpus-based grammar of German that focuses on variation in the written standard, considering both intra- and extralinguistic factors. What sets it apart from earlier studies is its homogeneous database and the thorough statistical examination of all factors involved using a generalized linear model. The more formal data (mostly newspapers) will be complemented with informal Internet data (COW corpora). This paper attempts to show how the factors determining position and attribute type interact (e.g. animacy, length, semantic relation; time, region, register). It aims to explain the results from a functional perspective, showing that most of the variation can be explained by changing communicative needs.



Lang, K. (2018). Possession: Empirisch-funktionale Untersuchungen zu Genitivattribut und Präpositionalphrase mit "von". Munich: Iudicum.

Peschke, S. (2014). Merkels Politik vs. die Politik Merkels: Eine korpusbasierte Untersuchung zur Prä- und Poststellung von Eigennamen im Genitiv. In Debus, F.; Heuser, R.; Nübling, D. (eds.): Linguistik der Familiennamen: 233–248. Hildesheim: Olms.

Rosenbach, A. (2003). Aspects of iconicity and economy in the choice between the s-genitive and the of-genitive in English. In Rohdenburg, G.; Mondorf, B. (eds.): Determinants of Grammatical Variation in English: 379–411. Berlin, New York: De Gruyter.

Rosenbach, Anette (2014): English genitive variation – the state of the art. English Language and Linguistics 18 (2): 215–262.

Scott, A.K. (2014) The genitive case in Dutch and German: A study of morphosyntactic change in codified languages. Leiden, Boston: Brill.

Smith, G. (2003): On the distribution of the genitive attribute and its prepositional counterpart in modern standard German. University of Pennsylvania Working Papers in Linguistics 8 (1): 173–186.

Szmrecsanyi, B. (2010). The English genitive alternation in a cognitive sociolinguistics perspective. In Geeraerts, D.; Kristiansen, G.; Peirsman, Y. (eds.): Advances in Cognitive Sociolinguistics: 141–166. Berlin, New York: De Gruyter.

The Loss of the Preterite in Luxembourgish


With Fischer (2018) the loss of the preterite (Präteritumschwund) in German dialects has been thoroughly investigated. Fischer (2018) first describes the areal distribution of the remaining preterite forms and those in decline. She then reconstructs the process of the loss historically, and, finally, she names the expansion of the perfect forms (ich habe gemacht instead of the preterital form ich machte) as the main factor that caused the preterite to begin to decline and in some regions to disappear completely. However, the loss of the preterite in Luxembourgish (being originally a Moselfranconian and thus a German dialect) has not yet been investigated in detail. Krier (2015) describes the use and the loss of the preterite in Luxembourgish by using a corpus of free speech from the Chambre des députés (Chamber of Deputies). However, the exact areal distribution and a list of verbs with or without preterital forms has not yet been documented. This study focusses on this research gap.

The research questions of this study are the following:

What is the spatial distribution of the present perfect and preterite forms? Which dialect areas in Luxemburg still retain their preterital forms in addition to the perfect forms? Which areas have lost their preterite forms completely, if at all?

Are particular preterital forms that have been encoded with morphological information such as number and person more sensitive to being lost than others? Furthermore, are particular preterital forms more sensitive to being lost based on frequency, conjugation class, semantics, and function?

This study is based on survey data that were collected indirectly by means of a questionnaire at the beginning of 2019. With the questionnaire, verb paradigms for the present, preterite, and perfect tenses for 24 verbs from ca. 40 speakers of Luxembourgish dialects were collected.



Fischer, H. (2018) Präteritumschwund im Deutschen: Dokumentation und Erklärung eines Verdrängungsprozesses. Berlin/Boston: De Gruyter.

Krier, F. (2015): Der Präteritumschwund im Luxemburgischen: Mit einer Übersicht über das Verhalten des Konjunktivs II und einem Exkurs zur Präteritalgrenzregion. Zeitschrift für Dialektologie und Linguistik 82/1: 48–81.

The influence of social constraints on syntactic variation: Global and local patterns in the English dative alternation


The present paper explores the influence of social constraints on syntactic variation across eight regionally distinct varieties of English. To this end, the study focuses on patterns of variability in the English dative alternation, i.e. the variation between the ditransitive dative (e.g. Mary gives John a book) and the prepositional dative (e.g. Mary gives a book to John). Because the choice in syntactic variation is believed to be mainly driven by cognitive constraints – which have been explored extensively (see Bresnan & Hay 2008; Röthlisberger et al. 2017 among others) – , the influence of the speakers’ social background, e.g. their age or gender, has remained understudied.

The present study builds on previous work that shows the importance of social constraints on syntactic variation (e.g. Theijssen et al. 2011; Jenset et al. 2017) and extends the analysis to a large set of regionally distinct varieties of English. More specifically, 4,784 interchangeable dative variants were extracted from the spoken components of the International Corpus of English sampling data from eight varieties, namely Hong Kong, Indian, Jamaican, Philippine, Canadian, British, Irish, and New Zealand English. Variants were automatically coded for numerous constraints including the speaker’s gender, their age, education and occupation where information was available from the metadata accompanying each ICE-corpus. Next, conditional inference trees (Hothorn et al. 2006) and mixed-effects modeling (Pinheiro & Bates 2000) were used to explore internal patterns of variation in more detail.

Results largely corroborate previous research: conditional inference trees indicate that male speakers prefer the prepositional dative more than female speakers, particularly in spoken informal conversations in Indian and Jamaican English. Results of the mixed-effect model further highlight that age and gender influence variant choice significantly differently in the eight varieties, pointing to regional variation in the effect of social constraints. The present paper thus not only suggests that syntactic variants might carry social meaning, it also challenges the well-trodden notion that syntactic alternations cannot be considered sociolinguistic variables (see Lavandera 1978).



Bresnan, J. & J. Hay (2008): Gradient grammar: An effect of animacy on the syntax of give in New Zealand and American English. Lingua 118(2): 245–259.

Hothorn, T. et al. (2006): Unbiased recursive partitioning : A conditional inference framework. Journal of Computational and Graphical Statistics 15(3): 651–674.

Jenset, G. et al. (2017). Keeping the English dative alternation in the family: A quantitative corpus-based study of spoken data. Paper Presented at the 9th International Corpus Linguistics Conference, 27 July 2017, University of Birmingham. 4 October, 2017.

Lavandera, B. (1978): Where does the sociolinguistic variable stop? Language in Society 7(2): 171–183.

Pinheiro, J. C. & D. M. Bates (2000) Mixed-effects models in S and S-PLUS. New York: Springer.

Röthlisberger, M. et al. (2017): Cognitive indigenization effects in the English dative alternation. Cognitive Linguistics 28(4): 673–710.

Theijssen, D. et al. (2011): In a land far far away... A probabilistic account of the dative alternation in British, American, and Australian English. Manuscript.

Apparent time and real time variation in the Frisian verbal complex


The developments in Frisian verb clusters over the last decades have received a substantive amount of attention in linguistic studies: De Haan (1992, 1996), Ytsma (1995), Wolf (1996), Koeneman & Postma (2006), Versloot & Hoekstra (2016). Many of those are directed exclusively towards clusters of two verbs or towards one type of three-verb cluster. The current study examines both in an apparent time and real time framework. The data were gathered by means of an acceptability judgment task and a verb cluster elicitation task. Combining our results with earlier research, it is possible to give a broad picture of the state of affairs in the Frisian verbal complex: indeed, variation has continued to increase.

By means of a comparison between my findings and the distribution of different verb orders in the Syntactic Atlas of Dutch Dialects (Barbiers et al. 2006), I will argue that the developments in Frisian are going into the direction of regional clustering patterns rather than Standard Dutch patterns. This fits in nicely with what Heeringa & Hinskens (2014) found in their study of 86 Dutch and Belgian dialects: all dialects in the Netherlands converge to Standard Dutch, but in general dialects have converged towards each other.



Barbiers, S. et al. (2006). Dynamische Syntactische Atlas van de Nederlandse Dialecten (DynaSAND). Amsterdam: Meertens Instituut. URL: http://www.meertens.knaw.nl/sand/.

de Haan, Germen J. 1992. The verbal complex in Frisian. In Us Wurk. Tydskrift foar Frisistyk 41: 59–92.

de Haan, Germen J. 1996. Recent changes in the verbal complex of Frisian. In NOWELE Volume 28/29: A Frisian and Germanic Miscellany. Nielsen & Petersen (eds), 171–184.

Heeringa, W. & Hinskens, F. (2014). Convergence between dialect varieties and dialect groups in the Dutch language area. In Aggregating dialectology, typology, and register analysis; linguistic variation in text and speech. Szmrecsanyi & Wälchli (eds). Series: Linguae et Litterae: Publications of the School of Language and Literature, Freiburg Institute for Advanced Studies. De Gruyter, Berlin and Boston, pp. 26-52 and 452-453.

Hoekstra, E., & Versloot, A. (2016). Three-Verb Clusters in Interference Frisian: A Stochastic Model over Sequential Syntactic Input. In Language and Speech, 59(1), 43–58.

Koeneman, O. & Postma, E. (2006). Veranderingen in Friese werkwoordclusters. In Nederlandse Taalkunde 11(2), pp. 124-145.

Wolf, H. (1996). Wat en Hoe yn it Ynterferinsjefrysk. In Syngroan en Diagroan Undersyk nei Feroarings yn de Fryske Tiidwurdlike Einrige. MA thesis, University of Utrecht.

Ytsma, J. (1995). Frisian as First and Second Language. Sociolinguistic and Socio-psychological Aspects of the Acquisition of Frisian among Frisian and Dutch Primary School Children. Ph.D. dissertation, University of Tilburg. Leeuwarden: Fryske Akademy.

10:15-12:20 Session 12D: Sociophonetics
Location: Moscow
Towards a dialectology of tone in South Jutland


Most South Danish dialects, whether belonging to the Jutlandic or Island Danish dialect groups, mark the distinction between historically mono- and disyllabic words by means of tones. Historical monosyllabics have a falling tone, known as accent 1, and historical disyllabics have a two-topped dipping tone, known as accent 2 (Sørensen in press).

The phonemic status of tone in various South Jutlandic dialects is well-established (e.g. Bjerrum 1948, Nielsen 1959), but the only modern acoustic investigation of South Jutlandic tone is a sketch of the tonal system on the island of Als in an unpublished MA thesis (Maate 2004). Thus, most descriptions of the tonal movements of accent 1 and 2 words is based on the perception of dialectologists and interdialectal comparison of tone remains purely anecdotal (e.g. Andersen 1897: 65). Nielsen’s investigations of the dialect of Rømø (1959) and Bjerrum’s investigation of the dialect of Felsted (1948), however, do rely graphic representations of tonal movement drawn using a kymograph.

This paper takes the first steps towards a dialectology of tone in South Jutland based on acoustic evidence. The data for the analysis comes from a huge corpus of dialect recordings available through the Peter Skautrup Center for Jutlandic Dialect Research (Andersen 1981). The recordings were mostly made in the 1970s with dialect speakers born around the year 1900.

This paper analyzes tone as F0 movement across accent 1 and 2 words in the speech of dialect speakers from different parts of South Jutland. Generalizations on the tonal patters of the individual dialects and interdialectal comparison of tone is carried out using Generalized Additive Mixed Models (Sóskuthy 2017), which is an ideal statistical tool for analyzing dynamic speech data such as pitch movement.

This analysis gives a synchronic view of the dialects as they were spoken at an earlier stage. Given the rapid dialect leveling in Denmark, and the spread of Standard Danish, this work is an important part of the documentation of the traditional dialects. Perspectives will be drawn to how this analysis can be developed further by including modern recordings (such as those made by Monka & Hovmark 2016) in order to get a diachronic view on the dynamics involved in prosodic change.



Andersen, N. (1897): Den musikalske akcent i Østlesvigsk. Dania 4(1): 65-81

Andersen, T.A. (1981): Dialektbånd og databehandling. Ord & Sag 1: 11-18.

Bjerrum, M. (1948). Felstedmålets tonale Accenter. Aarhus: Humanistiske Studier

Maate, A. (2004). Tonemsystemet i Alsisk. Unpublished MA-thesis

Monka, M. & H. Hovmark (2016): Sprogbrug blandt unge i Bylderup anno 2015. Danske Talesprog 16: 73-112.

Nielsen, G. (1959). Musikalsk Accent i Rømømålet. Copenhagen: Udvalg for Folkemaals Publikationer.

Sóskuty, M. (2017). Generalised additive mixed models for dynamic analysis in linguistics: a practical introduction. [https://arxiv.org/abs/1703.05339]

Sørensen, V. (In press). Lyd og prosodi i de danske dialekter. In E. Hjort (ed.): Dansk Sproghistorie

‘Staccato’ rhythm is a marker in Stockholm

ABSTRACT. This study’s main finding is that rhythm is socially stratified and stylistically sensitive in Stockholm. Male speakers from the racialized working class have ‘staccato’ rhythm in casual speech and style-shift into less ‘staccato’ forms in more formal styles.

An epicenter of rapid social stratification, Stockholm is home to Europe’s ‘first’ multiethnolect. (Rinkeby Swedish, Kotsinas 1988). The variety is often described as ‘staccato’, but speech rhythm has never been examined in casual-speech production or in the context of style-shifting. Therefore, questions persist. Does rhythm stratify socially in the vernacular? If so, is it also stylistically sensitive, having reached marker status (Labov 2001)?

Thirty-six male Stockholmers, ages 24–43, participated in the study. All were either born in Sweden or arrived before six years of age. Seventeen self-identify as svensk (white) and 19 as invandrare (non-white) in the city’s racialized binary system (henceforth race, see Cornips & de Rooij 2013:130). They hail from a stratified sample of social classes, measured by means of a six-criteria index. I elicited three speech styles: casual, reading, and reading like a radio announcer (radio). The recordings were transcribed and phonetically time-aligned. Rhythm was operationalized by extracting [duration · mean dB · mean F0] from each vowel and calculating the normalized pairwise variability index of vowels (nPVIV, Low, Grabe, & Nolan 2000), resulting in 43 012 intervocalic contrastive units.

The results are visualized in Figure 1. Invandrare speakers have significantly lower alternation (more ‘staccato’) than svensk speakers in casual and reading. In radio, this difference is no longer statistically significant. Importantly, there is an interaction of race:class in casual, splitting the working class in two: invandrare have the lowest alternation (most ‘staccato’), and svensk have the highest (least ‘staccato’). In reading, the difference between invandrare and svensk is still statistically significant, but the class interaction has disappeared. As speech-style formality increases (casual to reading to radio), the slopes for class gets incrementally less steep.

My interpretation is that rhythmic alternation is socially stratified and that it has some degree of salience, causing speakers to adjust it with formality. I conclude that ‘staccato’ is a feature of multiethnolect and that it has reached at least marker status in Stockholm.



Cornips, L., & de Rooij, V.A. (2013). Selfing and othering through categories of race, place, and language among minority youths in Rotterdam, The Netherlands. Siemund, P., Gogolin, I., Schulz, M.E. & Davydova, J. (Ed.), Multilingualism and language diversity in urban areas: Acquisition, identities, space, education 1 (129–164). Amsterdam: John Benjamins

Kotsinas, U.B. (1988). Rinkebysvenska – en dialekt? [Rinkeby Swedish – a dialect?] In P. Linell, V. Adelswärd, T. Nilsson & P.A. Petersson (Eds.) Svenskans beskrivning, 16 (264–278). Linköping: Linköping University Press.

Labov, W. (2001). Principles of linguistic change, Volume 2: Social factors. Oxford: Blackwell.

Low, E.L., Grabe, E. & Nolan, F. (2000). Quantitative Characterizations of Speech Rhythm: Syllable-Timing in Singapore English. Language and Speech, 43(4), 377–401.

Voicing and vowel quality in the German (multi-)ethnolect


We present data and results from a project which aims at a systematic sociophonetic and sociolinguistic analysis of the sound shape of (multi-)ethnic young speech in Germany. Towards the end of the last century, particular ways of speaking among certain groups of young speakers with ‘migrant background’ (and partly also their peers without such background) in several European cities became noted by researchers. While many discourse, lexical and grammatical phenomena of these so-called (multi-)ethnolects have been well investigated, there is still a lack of systematic quantitative research on the phonetic and phonological levels. Notable exceptions include publications by Jannedy & Weirich (e.g. 2014), Cheshire et al. (e.g. 2011) and van Meel et al. (e.g. 2016).

By applying quantitative acoustic and auditory phonetic analyses, we seek to establish how the pronunciation of a (multi-)ethnolectal core group differs from that of a (mono-ethnic) control group in different speech activities, ranging from spontaneous conversation to controlled read speech. To this end, 164 students aged between 13 and 21, studying at secondary schools in Stuttgart (Germany) and its surrounding region, were interviewed. Subjects were assigned to a core and a control group based on ratings by peers. Raters were asked to evaluate on a Likert scale whether the speaker sounds like they might speak other languages than German ‘at home’.

A core group comprising 24 (multi-)ethnolectal speakers and a control group of 10 speakers were then asked to read a list of sentences, perform a diapix/spot-the-difference task and take part in a semi-guided interview. The sentences on the reading list and the objects in the pictures used for the diapix task were specifically chosen to cover a range of variables that, according to literature and our own impression when listening to similar recordings made with multi-ethnic adolescents from Stuttgart (cf. Siegel 2018), might be characteristic of (multi-)ethnolectal ways of speaking.
We present results on two variables, voicing and vowel quality, and discuss the following questions (cf. Auer 2012):
Are there significant differences in voicing and vowel quality between the speakers categorized as coming from multilingual families and those who were not?
Are there differences that can be attributed to the speaker’s gender?
Are there differences due to different family languages or ethnic groups?
Which situational and interactional parameters favor the use of (multi-)ethnolectal features?



Auer, P. (2012) Ethnische Marker zwischen Varietät und Stil. In: Deppermann, A. (ed.): Das Deutsch der Migranten: 9–40. Berlin: de Gruyter.

Cheshire, J. et al. (2011): Contact, the feature pool and the speech community: The emergence of Multicultural London English. Journal of Sociolinguistics 15/2: 151–196.

Jannedy, S. & M. Weirich (2014): Sound change in an urban setting: Category instability of the palatal fricative in Berlin. Laboratory Phonology 5 (1): 91–122.

Meel, L. van et al. (2016): Co-variation and varieties in modern Dutch ethnolects. Lingua 172–173: 71–86.

Siegel, V. (2018) Multiethnolektale Syntax. Heidelberg: Winter.

Analyzing vowel variation of Dutch accents with Visible Vowels


Van der Harst (2011) measured f0, formants and vowel duration of the 15 full vowels of Dutch on the basis of word list data (monosyllabic words, before coda /s/ and /t/). The data set of speakers of standard Dutch is stratified by community (The Netherlands and Flanders), region (4x), sex (2x) and age (2x). They were selected according to dialectological and socio-geographic criteria via schools in medium-sized cities.

Using this data we compare the speakers’ accents by two measures. Using the first measure, speakers A and B are compared by calculating the average Euclidean distance between the vowels of speaker A and the corresponding vowels of speaker B. The second is Huckvale’s ACCDIST measure. This measure compares speakers A and B by correlating the mutual Euclidean distances between the vowels within speaker’s A vowel space with those within speaker’s B vowel space. The distance between the two speakers then is calculated as 1 minus the correlation.

Given n speakers, an n dimensional distance matrix is obtained for each of the two measures. The n dimensions are reduced to two dimensions by using multidimensional scaling. Distances between two regions X and Y are calculated as the average speaker distance between the speakers of region X and the speakers of region Y.

Initially, we included only F1 and F2 when measuring distances among speakers, and seven time points per vowel. We found a relative clear distinction between the Netherlandic and Flemish speakers, but some Netherlandic speakers showed up in the Flemish speaker group. Comparisons of regions revealed that those speakers are, not surprisingly, from the Netherlandic Limburg group, a southern variety of Dutch in the Netherlands, bordering at the Flemish Limburgian region. Furthermore, we found the Flemish Brabant region between the other Flemish regions and the Dutch regions which is in line with previous analyses, confirming that the accent in the core are of standardization in Flanders is more similar to the Netherlandic standard. Surprisingly, the accents of West-Flemish and Flemish Limburg standard Dutch are very close to each other, given the fact that these are the two outmost peripheral areas in Flanders with large dialect differences. However, when including F3 when measuring distances among speakers, the two regions become more distinct.

All the calculations and visualizations that we will show are done using Visible Vowels (www.visiblevowels.org) a web app and R-package for the analysis and visualization of acoustic vowel measurements: f0, formants and duration. Visible Vowels allows to convert and normalize vowel data and calculate some specific metrics. The app combines user friendliness with maximum functionality and flexibility, using a live plot view.

10:15-12:20 Session 12E: Panel Distributional Semantics
Location: New York 3
Enriching variationist analysis with distributional semantics methods


General Overview
The main methodological development in linguistics, including in variationist linguistics, of the past quarter-century has been the rise of corpus linguistics, with its concomitant commitment to the use of quantitative analysis methods. However, the increasing importance and availability of increasingly large corpora in variationist linguistics creates a need for appropriate methods for retrieving semantic information from corpora. This is because responsible variation analysis must take semantics into account: corpus-based research on grammatical alternation has demonstrated time and again that lexical and semantic effects are important constraints on grammatical variation. For example, we know that in the well-studied English dative alternation (give Tom some pizza versus give some pizza to Tom) semantic properties of the dative verb (transfer meaning versus communication meaning versus abstract meaning etc.) are significant predictors of dative choices. The trouble, then, is that as the field moves towards the analysis of increasingly large corpus resources, hand-coding data for semantic information of the type that influences the English dative alternation becomes ever less feasible. To deal with this problem, variation analysts are increasingly turning to (semi)automatic computational methods of distributional corpus semantics (e.g. collostructional analysis, semantic vector space modeling) to enrich the analysis of linguistic variation. This is a fairly “techy” subdiscipline in variation analysis, but one that is indispensable if we want to keep responsibly analyzing variation in big data resources.

Aims and objectives, key questions

Against this backdrop, the panel seeks to bring together specialists who use distributional semantic methods in variation analysis, to showcase best practice methodologies and to stimulate the exploration of new directions in the use distributional semantics in variationist linguistics. Questions that will guide the papers in the panel include the following:

  1. How can distributional semantics contribute to the semiautomated detection of lexical polysemy in corpus data, and more broadly, to the identification of lexical semantic structures like lexical fields or lexical relations?
  2. How can distributional semantics methods contribute to the (dia)lectometrical study of lexical variation?
  3. How can distributional methods help us better understand the determinants of grammatical variation?


  1. Dirk Speelman (KU Leuven) & Benedikt Szmrecsanyi (KU Leuven): Introduction
  2. Mariana Montes (KU Leuven); Distributional models and semantic variation across genres
  3. Stefan Grondelaers (Radboud Universiteit Nijmegen), Dirk Speelman (KU Leuven), and Paul van Gent (Radboud Universiteit Nijmegen): Getting a big data grip on language ideology change in the Low Countries: distributional semantics meets attitude research
  4. Christian Andersen (KU Leuven): Give him some space: semantic vector spaces and the dative alternation
  5. Haim Dubossarsky (Hebrew University of Jerusalem): What are word embeddings hiding up their sleeves?
  6. Stefano De Pascale (KU Leuven): Token-based distributional models and lexical lectometry
  7. Susanne Flach (University of Neuchâtel): Big data & small(er) questions: A distributional-semantic analysis of English reciprocals
  8. Weiwei Zhang (KU Leuven) and Kris Heylen (KU Leuven): From the cookbook of corpus-based lexical lectometry: a taste of Chinese
  9. Dirk Geeraerts (KU Leuven): Synopsis and discussion
Distributional models and semantic variation across genres


A distributional approach to semantics assumes that the meaning of an item can be derived from contextual features. For instance, we might characterize a word by the words it co-occurs with and weight their relationship with the strength of their association. This characterization can be represented as a vector to be compared to other vectors, leading to a vector space model (VSM). The spatial representation can be evaluated on different aspects, such as its general density and structure, both with a toolset of measures (see Speelman & Heylen, 2016) and through manual inspection of a graphical representation.

VSMs can be generated at both type and token level (see Heylen, Speelman & Geeraerts, 2012; Heylen et al., 2015). At the type level, two words are similar if they tend to co-occur with the same features. At the token level, two occurrences are similar if the words in their contexts tend to co-occur with the same features. This latter approach can represent the semantic structure of a type, illustrating how broadly spread or tightly grouped its senses are.

When applied to subcorpora of different registers, this methodology can contribute to variationist linguistics in two main aspects. On one side, applying the register classification to a token level VSM would show how the register distribution relates to that of the tokens (and thus the type’s senses) across vector space. On the other, a model built on a specific genre would illustrate the semantic structure of the type in that specific subcorpus, and could be compared to other registers.

For this case study we generate token-level VSMs to evaluate the distribution of tokens in different genres of an American English corpus. We would expect the token clouds (2D visualization) to exhibit a higher density in more technical genres, where the semantic variation of a type would be more restricted. The tokens used for this case study offer the advantage of being semantically annotated. Therefore, we can also assess whether the grouping in each genre matches the senses of the semantic annotation and how these are distributed in the different subcorpora.

The goals of this study are to evaluate the performance of VSM models identifying senses and illustrating the semantic structure of genres, and to report on the degree to which the meaning and genre subclouds match. This way, we present VSM models as a technique for the analysis of semasiological variation across registers.



Heylen, K., Speelman, D., & Geeraerts, D. (2012). Looking at word meaning. An interactive visualization of Semantic Vector Spaces for Dutch synsets. Proceedings of the EACL 2012 Joint Workshop of LINGVIS & UNCLH: 16-24.

Heylen, K., Wielfaert, T., Speelman, D., & Geeraerts, D. (2015). Monitoring polysemy: Word space models as a tool for large-scale lexical semantic analysis. Lingua 157: 153–172.

Speelman, D., & Heylen, K. (2016). From dialectometry to semantics. In Wieling, M., Bouma, G. & van Noord, G. (Eds.), From Semantics to Dialectometry: pp. 1–12. Groningen: University of Groningen.

From the Cookbook of Corpus-Based Lexical Lectometry: A Taste of Chinese


Lectometric approaches measure distances between language varieties (dialects, sociolects, registers etc.) by aggregating over observed differences in the realizations of a set of linguistic variables. In lexical lectometry, a variable consists of the alternative lexical expressions for one concept. In corpus-based lectometry, the observed realizations are culled from stratified corpora. Measuring semantically defined variables in corpora, and aggregating over them, poses specific methodological challenges that have been tackled in a number of studies (Heylen & Ruette 2013; Ruette et al. 2014; Ruette, Ehret & Szmrecsanyi 2016) with different statistical techniques, including Distributional Semantic Models. Yet so far, no general framework for corpus-based lexical lectometry has been formulated that systematically describes the issues and options in each step of the procedure so that it can be straightforwardly applied to new data and new languages, other than English (Ruette, Ehret & Szmrecsanyi 2016), Dutch (Geeraerts, Grondelaers & Speelman 1999) and Portuguese (Soares da Silva 2010).

This paper can be characterized as a twofold extension of the previous studies. First, it aims to establish a general framework for lexical lectometry research that considers most if not all options for different steps. Second, we want to go beyond the Indo-European languages by extending the framework on a typologically unrelated language, i.e. Chinese varieties.

For the general framework, we propose that a proper lexical lectometry research normally should involve the following steps: (1) compilation of a lectally stratified corpus; (2) sampling concepts as measuring points for lectometry; (3) identification of lexical expressions per concept; (4) disambiguation of lexical expressions in corpus data; (5) calculation of aggregated lexico-lectometric distances; (6) evaluation of measurement reliability and validity. For each step, we further provide possible options and caveats. For instance, step 2 and 3 can rely on existing concept-based lexical databases, like a synonym dictionary, or use corpus-driven keyword extraction and semantic vector space models. Step 4 can either make use of token-level distributional semantics models or rely on simpler n-gram language models.

To assess the portability of the general framework, both in practical and linguistic-typological terms, we perform a lexical lectometric analysis for varieties of Chinese based on data from large-scale corpora of Mainland Chinese, Taiwan Chinese and Singapore Chinese.



Geeraerts, D., Grondelaers, S. & D. Speelman. (1999). Convergentie en Divergentie in de Nederlandse Woordenschat. Amsterdam: Meertens Instituut.

Heylen, K., & Ruette, T. (2013). Degrees of semantic control in measuring aggregated lexical distances. In Borin, L. & Saxena, A. (eds.), Approaches to Measuring Linguistic Differences: 361-382. Berlin: Mouton de Gruyter.

Ruette, T., Geeraerts, D., Peirsman, Y., & Speelman, D. (2014). Semantic weighting mechanisms in scalable lexical sociolectometry. In Szmrecsanyi, B. & Wälchli, B. (eds.), Aggregating Dialectology, Typology, and Register Analysis: Linguistic Variation in Text and Speech: 205-230. Berlin: Mouton de Gruyter.

Ruette, T., Ehret, K., & Szmrecsanyi, B. (2016). A lectometric analysis of aggregated lexical variation in written Standard English with Semantic Vector Space models. International Journal of Corpus Linguistics, 21(1), 48-79.

Soares da Silva, A. (2010). Measuring and parameterizing lexical convergence and divergence between European and Brazilian Portuguese. In Geeraerts, D., et al. (eds.), Advances in Cognitive Sociolinguistics: 41-84. Berlin: Mouton de Gruyter.

What are word embeddings hiding up their sleeves?


Distributional models of word embedding (e.g., predictive models like word2vec or co-occurrence counts models like PPMI) have become prevalent in NLP studies and related research. The use of these models has also been established in usage-based linguistics, ranging from studies of polysemy (Schütze 1998; Heylen et al. 2015) and language variation (Jenset et al. 2018) to semantic change (Dubossarsky et al. 2015; Perek 2015).

In light of the popularity of these models within NLP and their diffusion beyond NLP, it is noteworthy that recent studies have reported deficiencies that make the word embeddings created by these models noisy (Hellrich & Hahn 2016; Antoniak & Mimno 2017; Dubossarsky, Grossman & Weinshall 2017; Karjus et al. 2018; Dubossarsky, Grossman & Weinshall 2018). These studies might seem of little relevance for linguists who want to use word-embedding as off-the-shelf models to study linguistic phenomena. However, contrary to this naïve impression, it has been shown that these deficiencies may drastically bias the analysis of the linguistic phenomena studied, and as a consequence, may lead researchers to unsound conclusions (Dubossarsky, Grossman & Weinshall 2017).

Our studies focus on the role word frequency and sampling have on the accuracy of word embeddings, and show how these two have far-reaching consequences for the study of semantic change (Dubossarsky, Grossman & Weinshall 2017) and polysemy research (Dubossarsky, Grossman & Weinshall 2018). In addition to both empirical and theoretical analyses, we propose a general method that allows the continued use of word embeddings by mitigating their deficiencies through carefully crafted control conditions.



Antoniak, M. & D. Mimno. 2017. Evaluating the Stability of Embedding-based Word Similarities. TACL 6, 107–119.

Dubossarsky, H., E. Grossman & D. Weinshall. 2017. Outta Control: Laws of Semantic Change and Inherent Biases in Word Representation Models. Proceedings of EMNLP, 1147–1156.

Dubossarsky, H., E. Grossman & D. Weinshall. To appear. Coming to Your Senses: on Controls and Evaluation Sets in Polysemy Research. Proceedings of EMNLP.

Dubossarsky, H., Y. Tsvetkov, C. Dyer & E. Grossman.2015. A bottom up approach to category mapping and meaning change. Proceedings of the NetWordS, 66-70.

Hellrich, J. & U. Hahn. 2016. Bad Company — Neighborhoods in Neural Embedding Spaces Considered Harmful. Proceedings of COLING-16, 2785–2796.

Heylen, K., T. Wielfaert, D. Speelman & D. Geeraerts. 2015. Monitoring polysemy: Word space models as a tool for large-scale lexical semantic analysis. Lingua 157: 153–172.

Jenset, G. B., J. Barðdal, L. Bruno, E. Le Mair, P. A. Kerkhof, S. Kleyner, L. Kulikov & R. Pooth. 2018. Continuous vector space models for variation and change in sparse, richly annotated Indo-European argument structure data. Presentation at SLE 2018, Tallinn.

Karjus, A., R. A. Blythe, S. Kirby & K. Smith. 2018. Two problems and solutions in evolutionary corpus-based language dynamics research. Presentation at SLE 2018, Tallinn.

Perek, F. 2015. Using distributional semantics to study syntactic productivity in diachrony: A case study. Linguistics 54(1): 149-188.

Schütze, H. 1998. Automatic Word Sense Discrimination. Computational Linguistics 24(1): 97–123.

Getting a big data based grip on language ideology change in the Low Countries: Distributional semantics meets attitude research


In this paper we introduce distributional analysis to facilitate the experimental investigation of language ideology change, a notoriously intangible process which is nevertheless essential to understanding why most European standard languages are increasingly becoming “mixed codes” of standard and non-standard forms.

A textbook case of the importance of language ideology is Belgian Standard Dutch (BSD), the norm for spoken Dutch which was forced onto the Flemish in a process of hyperstandardization (Van Hoof & Jaspers 2012). Although spoken BSD has never been a vital production reality, it continues to be ideologically dominant to the extent that it fuels controversy, and legitimizes the discrimination of Tussentaal, an increasingly vital colloquial variety which is still widely stigmatized. Following Kristiansen (2009), we have proposed a “competing ideology scenario” to account for the present-day language dynamics in Flanders, in which the conservative standard language ideology motors the use of BSD, whereas an emergent contra-ideology is sustaining the vitality of Tussentaal (Grondelaers et al. 2016).

There is some experimental evidence (Grondelaers & Speelman 2013; Rosseel 2017) that the conservative ideology attributes traditional prestige to the standard forms, whereas the contra-ideology motivates the dynamic prestige (pertaining to media cool and anti-norm challenge) of Tussentaal. No attempt has ever been made, however, to obtain an aggregate “perceptual map” of the Flemish repertoire, which reveals the simultaneous operation of the competing ideologies. The experiment reported in this paper was specifically designed in function of such a map.    

In a free response task, 211 respondents first returned three adjectives in reaction to the labels for five regional accent varieties, one ethnic accent variety (Moroccan-accented Dutch) and the two supra-regional varieties, viz. BSD and Tussentaal. Valence information (pertaining to the positive/negative character of the responses) and big data-based distributional analysis (to detect semantic similarity between the responses) were used to cluster adjectives into 11 positive and 11 negative evaluative dimensions. Correspondence analysis was subsequently employed to visualise the correlations between these evaluative dimensions and the investigated language labels.

Although free response techniques have been taken to the task for reproducing no more than the most conservative ideologies (Kristiansen 2009), our perceptual maps reveal a much richer stratification than the conservative “one variety good, all other varieties bad”-dichotomy we had anticipated: while BSD remains the uncontested (albeit virtual) prestige standard in Flanders, Tussentaal is increasingly accepted as a practical lingua franca. The fact that the Ghent-accent is clearly boosted by dynamic prestige features demonstrates that our free response data unveil the impact of both the conservative ideology and its modern counterpart. An apparent-time comparison between evaluations by younger and older participants, finally, undeniably demonstrates a value system in motion.     

10:15-12:20 Session 12F: Panel Frequency
Location: Paris
Frequency-grammar interaction in 100 year change of -<en> in bilingual context


This paper investigates one hundred years of language contact between Frisian and Dutch, focussing on the highly frequent word ending and suffix -/ən/. The data are taken from the bilingual sociolinguistic Boarnsterhim Corpus (Sloos et al. 2018; Sloos et al. subm.), which spans four generations Frisian-Dutch speakers.

Final -/ən/ can be pronounced in a variety of ways (1). In Standard Dutch, full realization (in higher registers) or nasal deletion (resulting in schwa) applies. In Frisian, these variants are also used, but a syllabic nasal or a nasalized schwa are more common and perceived as Frisian markers (Feitsma et al. 1987). The syllabic nasal is subject to place assimilation (Visser 1997). This nasalized schwa is optionally followed by a full nasal. In addition, zero realization occassionaly occurs. Typically, in the Dutch variety of Frisian speakers, all variants occur as well.

 (1)                 ən       ə       n m ŋ       ə̃       ə̃n       Æ

The pronunciation of -/ən/ changed across generations such that a complementary distribution developed. In the first generations, the distribution for the two language is largely similar. Full realization commonly occurs in both languages, but a syllabic nasal often occurs after /t d/. In the following generations, we observe divergence between the two languages. The syllabic nasal and full realization hardly occur in Dutch anymore and are gradually substituted by a nasalized schwa, often followed by a nasal. This nasalized schwa was simultaneously introduced in Frisian. However, in Frisian, a complementary distribution with the syllabic nasal developed. The context for the syllabic nasal is a following [-sibilant, -continuant] consonant; the nasalized schwa is the elsewhere allomorph.

The development in Dutch occurs clearly faster than the change in Frisian, but both are gradient. I will report on the interaction between grammar and frequency in the changes. I expect that the change in distribution correlates with lexical frequency such that higher frequent words change first; a pattern that is usually observed in reduction-driven language change (e.g. Bybee 2010, Philips 2006). Based on my previous research (Sloos 2013/2019; Sloos & van der Sijs (subm.)), I also expect that in different phonological contexts—viz. these of the preceding consonant—the rate of change differs.



Bybee, Joan. 2010. Language, usage and cognition. Cambridge: Cambridge University Press.

Feitsma, Tony, Els van der Geest, Frist van der Kuip & Irénke Meekma. (1987). Variations and development in West Frisian sandhi phenomena. International Journal of the Sociology of Language 64, 81—94.

Phillips, Betty. S. 2006. Word frequency and lexical diffusion. Hampshire/New York, Palgrave Macmillan.

Sloos, Marjoleine, Eduard Drenth & Wilbert Heeringa (2018). The Boarnsterhim Corpus: A Bilingual Frisian-Dutch Panel and Trend Study. In Proceedings of the 11th edition of the Language Resources and Evaluation Conference, 7-12 May 2018, Miyazaki (Japan).

Sloos, Marjoleine, Andrea Ariza García, Eduard Drenth, Wilbert Heeringa & Jeroen van de Weijer. Submitted. The Boarnsterhim Corpus: a bilingual Frisian-Dutch sociolinguistic spoken language database. Language Resources and Evaluation.

Sloos, Marjoleine & Nicoline van der Sijs. Subm. Dutch loanword integration in Indonesian: how lexical variation leads to changes in the grammar. Glossa.

Visser, Willem (1997). The syllable in Frisian. PhD dissertation, Vrije Universiteit Amsterdam.

Variation and the Lexicon


Knowing a language involves establishing mental representations for the forms and structures a speaker must manipulate and produce. Linguists hypothesize underlying representations of words and morphemes as essential components of the mental grammar.  But since such components cannot be observed directly, these hypotheses can only be tested indirectly, against observable phenomena.  Variable processes provide essential evidence for the nature of lexical representations.

This paper presents evidence bearing on several aspects of mental representations of lexical items, morphological structure, and lexical frequency.  Lexical representations are illuminated by how words are differentially treated by variable phonological processes.  For example, the internal morphology of words is revealed by phonological processes in several languages that treat segments that constitute discrete morphemes differently from those that do not.  In English, the final stops in missed, packed, representing the past tense suffix, undergo –t deletion less than those in mist, pact. Similarly, in Portuguese, final –s deletion occurs less often in plural senos ‘sines’ than in menos ‘less’.  The acquisition of these morphological representations is observable in child language, where the relevant words do not get the same treatment as in adult speech.  Variable processes also illuminate the underlying representation of exceptional lexical items; certain exceptional high-frequency words appear to undergo deletion processes far more than phonologically and morphologically comparable forms.  This suggests they have underlying lexical entries that encode the deletion process: English and has an allomorph an; Spanish entonces has an alternate entonce.  Selecting these allomorphs elevates the apparent rate of deletion in such words.

Variation also provides tests to distinguish apparently homophonous morphemes.  The English suffix –ing occurs in both nouns (a building) and verbs (I’m talking), but its variant realization –in occurs much more frequently in verbs, and priming effects occur more strongly within-category than across-categories, suggesting that two distinct morphemes may actually be involved.

Do grammar and lexical frequencies meet?


In the early days of formal phonological theory (introduced in and developed out of Chomsky & Halle 1968), the locus of variation –in so far as variation was at all an issue- was conceived to reside in the rule machinery. Dialect variation was accounted for through rule ordering: two rules which are differently ordered constitute either two different (categorical) grammars or inherent (quantitative) variation.

The theory has gone through several major developments. Today phonology is conceived as an autonomous module of the grammar which is itself, in turn, organized modularly, with interfaces with the lexicon, morphology and syntax. Each module and interface is the object of (ever evolving) subtheories.

Over the last decades a new paradigm has developed, which is sometimes referred to with the umbrella notion ‘cognitivist’. In connection with speech, sound structure and sound change, this includes Usage-based phonology (Bybee 2001) and Exemplar Theory (Pierrehumbert 2016). In this paradigm, grammatical knowledge is claimed to emerge bottom-up and structure is not given a priori. "The cognitive and psychological processes and principles that govern language are not specific to language, but are in general the same as those that govern other aspects of human cognitive and social behavior” (Bybee 2001:17).

In this approach, language use requires little, if any computation. In principle everything is stored redundantly: every realization of every item is stored in anecdotal memory in bundles of maximally concrete articulatory, acoustic, grammatical, semantic and pragmatic information concerning the single occurrences ('tokens' or 'exemplars'). All occurrences are organized in exemplar clouds, around a prototype; the lexicon is hence a cosmos of multi-dimensional networks. Token frequency (i.e. frequency of usage) and type frequency (distributional frequency) are the stem cells of grammar and they have been claimed (and sometimes demonstrated) to play a role in processes of language change and thus in variation.

In this talk, some of the main differences and similarities between both paradigms will be sketched in broad outline. Subsequently, some recent hybrid models will be briefly introduced. A hybrid model will be presented which builds on insights into the life cycle of sound change (Kiparsky 1995; Bermúdez-Otero 2015), a theory from which predictions can be drawn regarding the place of lexical effects in sound change. Some of these will be discussed and empirically tested on data concerning instances of both historical and ongoing sound change in specific varieties of Dutch.



Bermúdez-Otero, Ricardo (2015). Amphichronic explanation and the life cycle of phonological processes. In: P. Honeybone & J. Salmons (eds), The Oxford Handbook of Historical Phonology. Oxford: OUP, 374-399.

Bybee, Joan (2001). Phonology and Language Use. Cambridge: CUP

Chomsky, Noam & Morris Halle (1968). The Sound Pattern of English. NYC: Harper & Row

Kiparsky, Paul (1995). The phonological basis of sound change. In: J. Goldsmith, (ed.), The Handbook of Phonological Theory. Cambridge, MA: Blackwell, 640--670

Pierrehumbert, Janet (2016). Phonological representation: Beyond abstract versus episodic. Annual Review of Linguistics 2, 33–52

Type frequency and morphological neighbourhood determine competitiveness


The competitiveness of a pattern depends on its frequency. If there is competition between two patterns, the one that is more frequent is expected to win out against the pattern that is less frequent. Consider competition between irregular simple pasts in /æ/ and in /ʌ/. These are competing, which is observed from errors, dialectal variation, and language change. For example, flang changed to flung, and rang is under pressure from rung (Bybee & Slobin 1982). This suggests that /ʌ/ has a competitive edge on /æ/. However, a type frequency count of simple pasts in these vowels does not convincingly bear this out. We observe that /ʌ/ occurs seventeen times,  whereas /æ/ occurs thirteen times.  Furthermore, the average raw token frequencies of the pasts in /æ/ are higher than those in /ʌ/, contrary to our expectations.

The above approach does not take the morphological neighbourhood into account. Morphological neighbourhood is a concept based on relevant analogies. The morphological neighbourhood for a given item consists of the set of those items which are closely similar to it, both formally and semantically. For example, fruitful belongs to the morphological neighbourhood of fruit. Strictly speaking, fruit also shares an analogy with four, but this analogy is not relevant to morphology: fruit and four are too distant in form and meaningto consider the effect of this analogy. In this respect, analogy resembles gravity: if the distance is too large or the object too small, we may neglect its effects. To determine a significant morphological neighbourhood, we only consider items close in form and meaning.

Considering the simple past, the only form close in meaning is the past participle, since it also involves past tense meaning and since a past tense form of a given paradigm is similar in form to the participle of that same paradigm. The upshot of this is that the participle belongs to the neighbourhood of the simple past and vice versa. When examining competition between the two vowels mentioned, we must take the morphological neighbourhood into account. The reason for this is that primary activation of, say, flung (simple past) will yield a secondary activation not only of all simple pasts in /ʌ/ (like stung): it will also yield a secondary activation of participles in /ʌ/ (rung, done, and so on). This boosts the competitiveness of the vowel /ʌ/, explaining the observation by Bybee & Slobin (1982). So, type frequency may be viewed as a proxy for what psycholinguists refer to as primary activation, and morphological neighbourhood (items target through relevant analogies) may be viewed as a proxy for secondary activation.

We will show that morphological neighbourhood  also allows us to clarify the following empirical puzzles:

  • Why is the suffix -en of irregular participles atrophied in English but flourishing in Dutch?
  • Why is the suffix -t/-d of irregular participles flourishing in English but atrophied in Dutch?
  • Why do irregular pasts and participles in English end more in -t/-d than in other endings?
  • Why do irregular pasts and participles in -t/-d show more conflation than other endings?
  • Why is superconflation (homophony of past, participle and infinitive) always in -t/-d?
12:20-14:00Lunch Break
12:20-14:00 Session 13: Posters Presentations
Location: Central Hall
“Man said ‘one two five’”: Stance, Style & Pronominal Variation in London Adolescent Speech


In recent years, a wealth of variationist research has documented the emergence of a new ‘multiethnolect’ in London – what researchers have termed ‘Multicultural London English’ (MLE: Cheshire et al., 2008). This research has overwhelmingly depicted MLE as the emerging working-class vernacular. However, it remains unclear to what extent MLE can truly be defined as a homogenous dialectal variety or whether it should be considered more of a youth style (cf. Kerswill, 2013).

To investigate this issue, I focus on one feature often considered typical of MLE: the first-person pronoun man as in: “man’s hungry”. Existing research on the feature has examined the semantic and syntactic properties of man (Cheshire, 2013; Hall, 2017), however less is known about the stylistic possibilities of man and its sociolinguistic distribution.   

Drawing on interviews and self-recordings collected during a year-long (2016-2017) sociolinguistic ethnography in an East-London youth group, I examine the social and linguistic distribution of 49 tokens of pronominal man. Distributional analyses show that not only have the semantic referential values of man developed (1st person à 3rd person, cf. Cheshire 2013) but it is only used by speakers who self-identify as part of the ‘Gully’ – a membership category that is characterised by a more ‘urban’ identity. By subjecting the data to close interactional analyses, I show that, as a 3rd person pronoun, man has developed a unique rhetorical function for these speakers. Specifically, when used to address the 3rd person subject, man ‘Others’ the interlocutor, thereby excluding that person from the ingroup, thus building solidarity amongst those who use man – the Gully. I then go on to draw on social media data collected during the period of ethnographic research to situate this feature and the corresponding ‘gully’ identity within the broader sociolinguistic context of London.  

This analysis therefore not only contributes to the debate on the nature of multiethnolects (Svendsen & Quist, 2010) but also joins a growing call for variationist analyses which integrate perspectives beyond the paradigm in examining variable forms (e.g., Snell, 2010). 



Cheshire, Jenny. 2013. Grammaticalisation in social context: The emergence of a new English pronoun. Journal of Sociolinguistics. 17(5):608-633.

Cheshire, Jenny; Sue Fox; Paul Kerswill & Eivind Torgersen. 2008. Ethnicity, friendship network and social practices as the motor of dialect change: linguistic innovation in London. In Ulrich Ammon & Klaus J. Mattheier (eds), Sociolinguistica: International Yearbook of European Sociolinguistics, vol. 22, pp. 1-23. Tübingen: Max Niemeyer Verlag.

Hall, David. 2017. The impersonal gets personal: A new pronoun in Multicultural London English. Generative Linguistics in the Old World. 14-18 March. Leiden University, The Netherlands.

Kerswill, Paul. 2013. Identity, ethnicity and place: The construction of youth language in London. In: Auer, Peter, Martin Hilpert, Anja Stukenbrock and Benedikt Szmrecsanyi (eds.) Space in language and linguistics, pp. 128-164. Göttingen: Walter de Gruyter.

Snell, Julia. 2010. From sociolinguistic variation to socially strategic stylisation. Journal of Sociolinguistics. 14(5): 618-644.

Svendsen, Bente A. & Pia Quist. 2010. Introduction. In Pia Quist and Bente A. Svendsen (eds.) Multilingual Urban Practices in Urban Scandinavia, pp. xiii-xxiii. Bristol: Multilingual Matters.

Let's ask the people: Folk perceptions of foreig-accented Icelandic


Language variation in Iceland has, for a long time, been characterised by minimal regional and socially non-stigmatised phonological variation. This well-established situation has lately been disrupted with the number of non-native speakers growing rapidly in the past couple of decades, now amounting to 10% of the country’s population, and quantitive research showing that perception of foreign-accented Icelandic is subject to linguistic stereotyping based on speaker’s country of origin (Bade 2018). Those results call for further investigation of people’s perceptions of and associations with different variants of foreign-accented Icelandic as well as for explanations of the motivation behind attitudes towards non-native Icelandic.

Folk linguistics and perceptual dialectology have proven to be valuable research fields that have provided diverse insights into how variation in language is perceived, what ideas and stereotypes exist towards the variation and, most importantly, why people react to language variation in the way they do (see the work of Preston, e.g. 1996; 1989).

By combining folk linguistic approaches with methods from both language attitude research, i.e. verbal guises (cf. e.g. Huygens & Vaughan 1983; Lindemann 2003), and the social sciences, i.e. grounded theoretical data analysis (cf. Strauss & Corbin 1994), „a multidimensional approach to what are ultimately folk questions [are expected to] help build a more complete and accurate picture of the regard for language use and variety“ (Niedzielski & Preston 2010:96) within the Icelandic speech community. With folk, i.e. laymen’s, perceptions and ideas about language variation and foreign accents in Icelandic at the centre of this research, focus-group interviews were chosen as primary source of data collection.

Five focus-group discussions with six to eight informants each were conducted. Apart from general discussions on language variation, informants were asked to listen to six audio recordings, i.e. verbal guises, as well as to perform a brainstorming and mapping task, followed by detailed discussion on attitudes towards the speakers and the individual motivation behind those attitudes. In this paper, we will report on first results of this study.



Bade, S. (2018). Evaluations of foreign accents in a purist speech community: the case of Iceland. In: Villena-Ponsoda, J.-A. et al (eds.): Language Variation. European Perspectives VII. Amsterdam: John Benjamins, SILV. Forthcoming.

Huygens, I. and G. Vaughan (1983). Language attitudes, ethnicity, and social class in New Zealand. Journal of multilingual and multicultural development 4: 207-223.

Lindemann, S. (2003). Koreans, Chinese or Indians?Attitudes and ideologies about non-native English speakers in the United States. Journal of Sociolinguistics 7(3): 348-364.

Niedzielksi, N. and D. Preston (2010). Folk Linguistics. Berlin; New York: Mouton de Gruyter.

Preston, D. (1996). Whaddayaknow?: The Modes of Folk Linguistic Awareness. Language Awareness 5(1): 40-74.

Preston, D. (1989). Perceptual Dialectology. Dordrecht: Foris.

Strauss, A. and J. Corbin (1994). Grounded Theory Methodology: An overview. In Denzin, Norman K. and Yvonna S. Lincoln (eds.). Handbook of Qualitative Research (pp. 273-285). Thousand Oaks: Sage.

Stereotypes, language and society: different prejudices about four accents of Italian and possible changes over time
PRESENTER: Camilla Masullo


This work investigates the existence of linguistic stereotypes in Italy regarding four different accents: two Italian accents, one from Northern Italy and one from Southern Italy, and two foreign accents, a British and a Senegalese accent. We also wonder whether a linguistic stereotype can change over time, with a specific analysis focused on the Southern Italian accent.

Method and data

We chose the verbal-guise method to indirectly extrapolate judgements about the four accents.

We first recorded the accents selecting four speakers, one for each accent, all male and between the age of 25 and 50 years, to avoid the influence of these social variables in our corpus and to reduce as much as possible voice quality’s mismatches. The speakers were recorded while reading the same Italian text from a neutral weather forecast in order to eliminate textual information influence. Thus, the substantial difference between the recordings was about accent.

Than we created a questionnaire of four sections, one for every accent, used both for quantitative and qualitative analysis. Firstly, respondents listened to the recordings, and then they answered to ten questions on the pleasantness of the voice and the supposed economic, social and cultural status of the speaker. A final section elicited some sociolinguistic information about the respondents (age, sex and origin). We collected 273 questionnaires later reduced to 166 in order to create a homogenous sample: thus, our corpus consisted in questionnaires from Northern Italian respondents, aged between 18 and 35. We used the software IBM SPSS 21 to run monovariate and bivariate analysis on which chi-square and Cramer’s V were calculated.

The qualitative analysis was led on a smaller selection of data. We chose questionnaires completed by Northern Italian respondents aged over 50 from the total sample (for a total of 18), in order to compare their opinion with the previous generation.


Data from quantitative analysis outlined a prejudice scale of the four accents: first position is occupied by the British accent; then come the two Italian accents, without significant distinctions between them, followed by the Senegalese one. The British accent is characterized by an admiration prejudice: in the 61% of cases, respondents attributed a high economic and cultural status to this speaker. Conversely, the 55% of respondents considered the Senegalese speaker friendly but with a low cultural and economic status. This result shows the existence of a paternalistic stereotype associated to the Senegalese or, even more generally, to the African speakers of Italian.

Data from the qualitative analysis outlined an ongoing change about the Southern Italian accent’s prejudice: if there is no socio-economic difference between the two Italian accents for the actual generation, a negative stereotype is associated with the Southern accent by the older generation. This demonstrates that linguistic prejudices may shift over time due to social changes and migrations in the language community, and that in Italy we may be dealing with a change in progress.

Diatopic variation of polar questions in two areas of the Italian peninsula
PRESENTER: Alice Betti


Aim: This work aims to investigate the intonation of polar questions in two areas of Italy, Ripacandida (PZ) and Siena (SI), by following the path of other scholars (e.g. Marotta & Sardelli 2009, Interlandi 2004) in mapping socioprosodic variation within Italian regional varieties. We also want to verify a possible relationship between segmental and suprasegmental variation (e.g. Albano Leoni, 2009), concerning the production of polar questions.

Data: Our corpus consists of 5h 33’ of recordings with 8 Italian native speakers, balanced by sex and by location; all speakers performed a sentence reading task (Corpus Frasi Fisse, CFF), and a series of map-task dialogues. The audio files have been cataloged according to the following parameters: geographical area of origin, phase of elicitation and informant. All these data form the oral archive CIGReS (https://github.com/AliceBetti/CIGReS) which consists of 410 polar questions (160 from the CFF, and the remaining 250 from the maptask dialogues) equally distributed between the two Italian varieties under analysis.

Main results: The polar questions of CIGReS have been analyzed both qualitatively and quantitatively. We used PRAAT plugin_momel-intsint (version 2016-03-29 Hirst, 2007) to observe the difference in F0 curves in the two Italian varieties: for instance, Sienese speakers show a more ascent F0 terminal contour than speakers from Ripacandida.

Fig. 1 The intonational profiles of a Ripacandida (left) and a Siena (right) speaker in the sentence reading task.

Quantitative analysis considered 5 continuous variables: duration of the stressed vowels, F0, intensity, F1, and F2 as calculated on the mean point of each stressed vowels; the statistical analysis was performed in IBM SPSS 20. The results showed that there is a correlation between segmental and suprasegmental variation, in fact there was a greater variation of the vowel space, between the two geographical areas, during the production of global interrogatives but not during the declarative statements, which are not diatopically marked from a suprasegmental point of view.

Conclusion:  Our analysis has proved that a diatopic difference between two varieties of Italian exists at both a segmental and a supra-segmental level, in particular for F0 terminal contour and variation in the vowel space, especially in the fronted vowels.



Albano Leoni, F. (2009). Dei suoni e dei sensi, Il Mulino: Bologna.

Hirst, D. (2007). A Praat plugin for Momel and INTSINT with improved algorithms for modelling and coding intonation. In Trouvain, J. (ed.) Proceedings of the XVIth International Conference of Phonetic Sciences, Saarbrücken, 6-10 agosto 2007, Saarbrücken: Universität des Saarlandes, 1233-1236.

Interlandi, G.M. (2004). L’intonazione delle interrogative polari nell’italiano parlato a Torino: tra varietà regionale e nuova koinè, PhD Thesis, Università di Pavia.

Marotta, G., Sardelli, E. (2009). Prosodiatopia: parametri prosodici per un modello di riconoscimento diatopico. In Ferrari, G., Benatti, R., Mosca, M. (a cura di) Linguistica e modelli tecnologici di ricerca, Atti del XL Congresso Internazionale di Studi della Società di Linguistica Italiana (SLI). Vercelli, 21-23 settembre 2006, Roma: Bulzoni, 411-435.

Language variation in university college students: A real time study


The purpose of this research is to identify and quantify language variation in real time in a community of students in the university college “Giasone del Maino” in Pavia (Italy). We chose this sociolinguistic setting firstly for the heterogeneity of the student population: as students come from different regions of Italy,  the college setting provides an optimal situation of language contact between standard Italian and Italian regional varieties. Moreover, the peculiarity of collegial life allows us to consider the student body as one community of practice (Eckert and McConnell-Ginet:1992; Wenger:1998), which, in our hypothesis, would accelerate the processes of diachronic variation.

Data were collected through interviews aimed at eliciting three distinctive speech styles at different level of control over the speech production: spontaneous speech, through a short self-presentation of the speaker and some general questions; semi-spontaneous speech, through different linguistic tasks; and controlled speech, through the reading of a word list. The whole corpus has been designed to investigate different phonemes related to both vocalism and consonantism of Italian. The word list reading task includes 49 tokens of real words. To study diachronic variation, the data were collected at different points throughout the academic year.

For this poster-presentation, a case-study of 6 speakers (3 male and 3 female) of two different linguistic areas (3 from northern Italy and 3 from southern Italy) have been analyzed. Data were collected at the freshmens’ arrival in college (Freshmen T1) and at 11 months after arrival (Freshmen T2). A first analysis was performed on dental affricates in the word list reading (144 tokens). As demonstrated by previous analysis (see Meluzzi 2016), these phonemes present three sonority degrees (voiced, voiceless and mixed) with a high degree of variation across the Italian regional varieties. A comparison of the data relating to the sonority degree of Freshmen T1 shows a prevalence of voiced outcomes within southern students (32.4% voiceless, 17.6% mixed and 50% voiced) with respect to the northern students (52.8% voiceless, 30.6% mixed and 16.7% voiced) [p=.012]. After only 11 months, it is possible to see in Freshmen T2 an important reduction of the voiced dental affricates of the students of the south (50% voiceless, 13.9% mixed and 36.1% voiced) which are close to the outcomes of the students of the north (58.3% voiceless, 19.4% mixed, and 22.2% voiced) [p>.05]. In the Freshmen T2 data,  the difference of north-south origin is no longer statistically significant.

These first results, although partial, suggest the presence of a process of accommodation towards the northern Italian regional variety by southern speakers. This phenomenon can be justified both by the perception of a greater prestige of the northern Italian regional variety and by the influence of the peculiarity of the community of practice.

dat de vogel aan kwam vliegen / dat de vogel kwam aangevlogen: Functional and Variative Dimensions of Dutch komen ‘come’ + Motion Verb


The Dutch motion verb komen ‘come’ can combine with either a past participle or infinitive to specify simultaneous aspects of a single motion event as in (1):

  1. De agent kwam de straat ingefietst.ptcp/fietsen.inf

‘The police officer came cycling into the street’.

The past participle variant is losing ground in Dutch (e.g. Dal 1954; Vogel 2005); however, the corpus data in Beliën (2016) suggests that it still has a strong footing in contemporary Dutch. Speakers can make use of either the past participle or the infinitive variant; however, there are slight regional, country-specific preferences for Standard Dutch (Haeseryn et al. 1997) and idiolectal variation (Cornips 2002). Given that speakers can use both variants in parallel and that there is a difference in the morpho-syntactic form of the variants, this raises the question as to a semantic differentiation between the two (e.g. Beliën 2016; Schäfer forthcoming). For example, Beliën (2016: 30; italics in original) suggests that while “both variants describe an unfolding, unidirectional motion event towards a contextually construable vantage point, the variant with the past participle highlights the end of a process, while the infinitive variant does not”. This distinction remains to be systematically tested.

In this contribution, we will present the results of a questionnaire with sentence completion and rating tasks, the purpose of which was to answer these questions:

  • How are the variants horizontally distributed in standard-near speech levels of spoken Modern Dutch?
  • Is the observed variation associated with extra-linguistic variables such as speakers’ age, gender, and level of education?
  • What role do semantic restraints play in steering the variation? Specifically, is the choice of the form contingent upon lexical aspect (=Aktionsart) of the motion verb? Is the formal choice dependent on the distinction between internal (manner) and external (path) movement?
  • If a motion verb expresses external movement, is the separable verb particle aan ‘towards’ required?



Beliën, M. (2016): Exploring Semantic Differences in Syntactic Variation: Dutch komen ‘come’ with a Past Participle or an Infinitive. In Bannink, A. & Honselaar, W. (ed.): From Variation to Iconicity: Festschrift for Olga Fischer on the Occasion of her 65th Birthday, 17–32. Amsterdam: Pegasus.

Cornips, L. (2002): Een vreemde eend in het rijtje. Over het aspectueel hulpwerkwoord ‘komen’. In In verband met jan Luif: 29 Variaties op een thema door vrienden en collega’s bij het afscheid van Jan Luif. URL: http://cf.hum.uva.nl/poldernederlands/backup_luif/luif/cornips.html [last accessed: 25.09.2018].

Dal, I. (1954): Indifferenzformen in der deutschen Syntax. Betrachtungen zur Fügung ich kam gegangen. Norsk Tidsskrift for Sprogvidenskap Bind XVII: 489–497.

Haeseryn, W. et al. (ed.) (1997): Algemene Nederlandse Spraakkunst. Groningen: Martinus Nijhoff.

Schäfer, L. (forthcoming): “kommen” und Bewegungsverb in westgermanischen Varietäten. In Speyer, A. & Hertel, J. (ed.): Syntax aus Saarbrücker Sicht 3. Beiträge der SaRDiS-Tagung zur Dialektsyntax. Stuttgart: Steiner.

Vogel, P. (2005): Neue Überlegungen zu den Fügungen des Typs sie kamen gelaufen. Zeitschrift für Germanistische Linguistik 33(1): 57–77.

Presenting the Nordic Word Order Database


In this talk, we present the Nordic Word Order Database (NWD), focusing on the rationale behind it and the methods used in data elicitation and data analysis. NWD is an online database with a user-friendly search interface, hosted by Tekstlab at the University of Oslo. It contains elicited production data from speakers of all the Scandinavian languages, including several different dialects. So far, 7 fieldworks have been conducted, and data from altogether around 250 participants (age: 16-60) have been collected (appr. 50000 sentences in total).

The database contains information about six different syntactic macro-phenomena: (1) Subject shift: the placement of pronominal and NP subjects with respect to sentence adverbials; (2) Object shift: the placement of pronominal and NP objects with respect to sentence adverbials; (3) Long object shift: the placement of light pronominal objects with respect to NP subjects; (4) Particle shift: the placement of verb particles with respect to pronominal and NP objects (and subjects); (5) Embedded V2: the placement of finite verbs with respect to sentence adverbs in different types of embedded clauses; (6) Non-V2 in questions and in contexts of certain adverbs. For all of the phenomena, at least one of the Scandinavian languages has developed a word order pattern that is categorically different from the other languages. There is also a relatively well-established description of the diachronic development of the relevant variable. In addition, the phenomena are known to be infrequent in natural speech. The database therefore provides an important addition to corpora – it gives a new means to investigate both inter- and intra-individual variation.

Two production experiments are used for data collection, and they include largely the same sentences in all languages. Both build on a simple design whereby a participant reads a sentence presented on a screen aloud at a given cue. The participant is then provided with a new start of the sentence, and is asked to transform the sentence. Five different transformations are used, in order to avoid priming: (a) inversion: The baker bought not the book yesterday à Yesterday… (bought {the baker} not {the baker} a book (tests phenomena 1, 3 and 5); (b) tense change and inversion (phenomena 1, 2, 3 and 4); (c) passive to active (phenomena 2 and 4); (d) main to embedded clause (phenomenon 5) and (e) embedded to main clause (phenomenon 6). The experiments are recorded, and the sound files are segmented and coded for word order, and now accessible through NWD.

The results indicate that both syntactic variation and prosodic/intonational properties elicited in the experiment match the patterns found in spoken language corpora. Moreover, the results reveal considerable variation both within and between speakers, but also a surprising stability within the individual languages: syntactic phenomena that show a variable pattern in one language has a fully categorical distribution in a closely related language. The database can thus be used some core questions about linguistic variation, e.g.: How much variation in form within/across languages, dialects/sociolects and individuals remains once we keep the meaning constant?

Wagwan, fam?!: Lectal focusing in grime music


The last few years have seen an increase in sociolinguistic studies exploring language variation and change in hip-hop-related media contents (Cutler & Røyneland 2015, Stæhr & Madsen 2015). Despite its relevance for young multicultural Britain, however, Grime – a hyper-local MC-led musical genre that originated in the streets of East London in the early 2000s – has so far only scarcely found its way into sociolinguistic discourse (Adams 2018, Drummond 2018). This paper reports on the Grime Project, an ongoing research project which explores a 2.5 million word corpus of over 3000 grime songs. The UDE Corpus of Grime (UCG) is a fully annotated and POS-tagged corpus containing metadata on extralinguistic information about 137 grime artists, including their demographic background from the early days of this musical genre until today.

A key analytical tool discussed in this paper is a modified version of the LFI (lectal focusing in interaction) metric, a measure developed by Sharma and Rampton (2015) to track how much an individual speaker strategically “[...] shifts toward one or another style in one interaction [...]” (Sharma & Rampton 2015: 3). Preliminary findings suggest that the language used by grime artists is, quite surprisingly, located on the more standard end of the spectrum, with strategic leaps towards a more vernacular style to establish street credibility.

Overall, the project has three distinct aims:
(a) to examine how both identity and street credibility are constructed by use of varying degrees of vernacularity and thus to explore the co-construction of authenticity, identity and youth-street styles particularly in the context of a largely DIY music production scheme,  
(b) to assess the extent to which the genre’s linguistic repertoire draws on Multicultural London English (Cheshire et al 2011) and Multicultural Urban British English (Drummond 2017), and
(c) to propose a reproducible analytical framework for exploring MC-led musical genres.

Adams, Z. (2018): ‘I don’t know why man’s calling me family all of a sudden’: Address and reference terms in grime music. Language & Communication 60: 11–27.
Cheshire, J. et al. (2011): Contact, the feature pool and the speech community: The emergence of Multicultural London English. Journal of Sociolinguistics 15(2): 151–196.
Cutler, C. & U. Røyneland (2015). Where the fuck am I from? Hip-hop youth and the (re)negotiation of language and identity in Norway and the US. In: Nortier, J. & B. A. Svendsen (eds.): Language, Youth and Identity in the 21st Century. Linguistic Practices across Urban Spaces: 139-163.
Drummond, R. (2017): (Mis)interpreting urban youth language: white kids sounding black?. Journal of Youth Studies 20(5): 640-660.
Drummond, R. (2018): Maybe it’s a grime [t]ing: TH-stopping among urban British youth. Language in Society 47: 171–196.
Sharma, D. & B. Rampton. (2015): Lectal Focusing in Interaction: A New Methodology for the Study of Style Variation. Journal of English Linguistics 43(1): 3-35.
Stæhr, A. & L. M. Madsen. (2015): Standard language in urban rap – Social media, linguistic practice and ethnographic context. Language & Communication 40: 67-81.

Morphological variation, standardization and language change – The case of Luxembourgish


Luxembourgish, being one of the youngest and smallest languages in Europe, is particularly interesting for the study of language variation and change because of its high degree of variation on all linguistic levels. One of the main reasons for this is the characteristic of Luxemburgish, being primarily a spoken language, whereas the processes of standardization and institutionalization are still under development.

In my paper, I will report on a study combining variational and perceptual linguistic approaches to tackle questions concerning language change and standardization. It is embedded in a larger research project that focuses on morphological variation in spoken and written Luxemburgish on the one hand and speakers' perceptions of this variation on the other hand.

Although Luxembourgish linguistics is a very young field of research and the focus in recent years has been more on describing the language system than on dealing with variational linguistic topics, these have not been completely neglected (s. Conrard 2017). In terms of morphological variation however, only the morphophonology of the past participle (Gilles 2011) and some aspects of plural formation (Dammel/Kürschner/Nübling 2010) have been investigated to some extent. Thus, an empirical in-depth survey of morphological variation patterns is pending.

The paper analyses the morphological variation of selected phenomena in spoken Luxembourgish to determine the factors controlling the variation of inflection patterns, for example the past participle (kaf vs. kaaft 'bought'), certain frequent adverbs (domat vs. domadder 'with it'), and the superlative (dat schéinst vs. schéinsten vs. schéinstent Haus 'the most beautiful house'). Data for the study has been gathered using a crowd sourcing mobile application for spoken language data (“Schnëssen[1]”). The sample contains language material from approximately 1000 speakers covering a wide demographics in terms of age group, social background, and geographical region. The analysis of this corpus will shed light on different structuring factors of morphological variation, e.g., variation patterns, spatial and social distribution of the variants, and the relation between language variation and change.

Strikingly, the largest part of the morphological variation patterns found in the data cannot be linked to regional variation. Rather, there seems to be a correlation with the speakers' age, other social parameters, or language-internal factors. Moreover, in combination with a perceptional linguistic approach it becomes clear that variation seems to be an integral part of the speakers' individual norm.



Conrard, F. (2017) Variation durch Sprachkontakt. Frankfurt/New York: Peter Lang. (Luxemburg-Studien / Études luxembourgeoises)

Dammel, A. & S. Kürschner & D. Nübling (2010): Pluralallomorphie in zehn germanischen Sprachen. Konvergenzen und Divergenzen in Ausdrucksverfahren und Konditionierung. In Dammel, A. (ed.): Kontrastive germanistische Linguistik. Teilbd. 2. Hildesheim [u. a.]: Olms.

Gilles, P. (2011) Morphophonologie des Partizip II im Luxemburgischen. In Gilles, P. & M. Wagner (ed.): Linguistische und soziolinguistische Bausteine des Luxemburgischen: 51-82. Frankfurt a.M. [et al]: Peter Lang. (Mikroglottika. Minority language studies : 4)

[1] The mobile application "Schnëssen" allows collecting spoken language data using image descriptions, translation and reading tasks.

Linguistic features across political borders: the case of Basque dialects in France and Spain


More than a century after the first systematic studies on dialectology and language variation, and despite “the European tradition of organizing linguistics according to national language philologies” (Harnisch, 2010: 276), it is almost a truism that significant linguistic boundaries do not necessarily converge with political borders (Veny, 1992; Thomas, 1999, Viaut & Pailhé, 2010).

In this paper we explore the limited influence of the political border between France and Spain on the Basque language, i.e. a minority language spoken on both the north (or French) and south (or Spanish) sides of the border. On the basis of old texts and modern records, we show how permeable has the border been during the last three centuries. More precisely, we analyze the geographical distribution of some phonological innovations that have been commonly used among Basque dialectologists to establish dialectal and sub-dialectal boundaries: the mid vowel raising ei / V, the insertion of the glide [ʝ] and the uvular [β] after /i/ and /u/, the palatalization of /n/ and /l/, and the neutralization [s] → [ś] before stops.

In particular, we discuss the spread of these innovations along the coast of Gipuzkoa and Lapurdi, two neighbouring seaside provinces pertaining to Spain and France respectively, where different linguistic varieties have been described (Bonaparte, 1863; Zuazo, 1998). Furthermore, we try to show that, in some cases, the bordering varieties converge against inland areas. This reveals that innovations did not always spread from more populated towns, and focuses our attention on the communities connected to the sea life.

On the whole, this case study provides an opportunity to consider various issues related to variation and geolinguistics, such as: the influence and/or permeability of the political borders, the spread and models of innovations, the influence of the linguistic communities.



Bonaparte, L.L. (1863). Carte des sept provinces basques montrant la délimitation actuelle de l'euscara et sa division en dialectes, sous-dialectes et variétés. London: Stanford’s Geographical Establishment.

Harnisch, R., (2010). Divergence of linguistic varieties in a language space. In Auer, P. & J.E. Schmidt (eds.). Language and Space. An International Handbook of Linguistic Variation. Theories and Methods: 275-294. Berlin: De Gruyter.

Thomas, P.L. (1999). Frontières linguistiques, frontières politiques. Histoire Épistémologie Langage XXI:1: 63-82.

Veny, J., (1992). Fronteras y áreas dialectales. In Euskaltzaindia (ed.). Actas del Congreso Internacional de Dialectología (Iker-7): 197-245. Bilbao: Royal Academy of the Basque Country.

Viaut, A. & Pailhé, J. (2010). Langue et espace. Pessac: Maison des sciences de l’homme d’Aquitaine.

Zuazo, K. (1998). Euskalkiak, gaur. Fontes Linguae Vasconum 78: 191-234.

Auxiliary Selection in Yiddish Dialects


Standard Yiddish (StY) has two perfect auxiliaries zayn ‘be’ and hobn ‘have’ (1) inherited by Middle High German (MHG). While in MHD the opposition was originally controlled by aspect (imperfective vs. perfective) it is no longer stable in modern German (cf. Eroms & Dal 2014: 141–146, §69). There is almost nothing known about the situation in Yiddish.

According to Jacobs (2005: 70) the Northeastern Yiddish (NEY) dialects show a "generalized use of hobn as the sole AUX for past tense formation" (2). Such a development is known from Swedish, Norwegian Dialects and Standard English (Harbert 2007: 304).

(1) StY: ix hob gešribn ‘I wrote’, ix bin gəzesn ‘I sat’  
(2) NEY: ix hob gəšribn ‘I wrote’, ix hob gəzesn ‘I sat’ 
                                    (Jacobs 2005: 70)

In my talk I want to present data on this phenomenon from the questionnaires of the "Language and Culture Archive of Ashkenazic Jewry" and compare the results with what is known about auxiliary selection in closely related varieties like German and Dutch (Gillmann 2015). The data will also be reviewed under Sorace’s (2000) "Auxiliary Selection Hierarchy".



Eroms, Hans-Werner und Dal, Ingerid (2014): Kurze deutsche Syntax auf historischer Grundlage. Berlin: de Gruyter, 141–146, §69. 
Harbert, Wayne (2007): The Germanic Languages. (Cambridge language surveys). Cambridge: Cambridge University Press.
Jacobs, Neil G. (2005): Yiddish. A linguistic introduction. Cambridge: Cambridge University Press.
Melitta, Gillmann (2015) Auxiliary Selection in Closely Related Languages. The Case of German and Dutch. In: Malte Rosemeyer und Rolf Kailuweit (Hrsg.): Auxiliary Selection Revisited. Gradience and Gradualness. Berlin/New York: De Gruyter, 333-358.
Sorace, Antonella (2000). Gradients in auxiliary selection with intransitive verbs, in: Language 76, 859-890.

Large-scale corpus-based analysis of regional varieties of Danish


As long as it has been physically possible, the Danish dialect research centers in Aarhus and Copenhagen have been recording dialect speakers throughout the country in order to preserve their language and culture for future generations (Andersen 1981). During a particularly active period of recording between 1971-1976 alone, more than 500 hours of recordings were made covering more than 500 different parishes. In the time since, the research centers have focused mostly on developing dictionaries of dialect-specific vocabulary (Jysk Ordbog and Ømålsordbogen), for which the recordings have only been used very sparingly. Instead, these have relied on older written sources and questionnaires.

As part of two related research projects on regional phonetic and phonological variation in Denmark, we have been gathering the recordings of Danish dialects and updating the existing metadata into a modern standardized format. Taken together, the corpus is unique in its sheer scope and breadth of coverage, but in spite of that, it has never been used for any kind of systematic research. However, it is crucial that such systematic research be done, as the bulk of Danish dialects are either extinct or moribund (Kristiansen 1998, but see also Monka & Hovmark 2016).

In this poster, we discuss challenges and opportunities of working with a corpus of this size. Aside from the sheer scope of the corpus, challenges are posed by the varying and generally subpar quality of the recordings, as well as the lack of systematic transcriptions. That being said, the scope of the corpus also provides unique opportunities to identify and map patterns of variation at a very high level of detail, e.g. by using modern computational methods in dialectometry (e.g. Prokic & Nerbonne 2008, 2013). Using the data, we will be exploring variation in syllable structure, consonant realization, intonation and tonal phenomena in regional variants of Danish. This poster showcases opportunities provided by the corpus and methodology by presenting preliminary research on microvariation in prosodic features and in the realization of laryngeal contrasts in stop consonants (expanding on Puggaard 2018). This will show how large-scale corpus-based studies using fine-grained acoustic analysis and statistics can be used to illuminate aspects of Danish regional variation which have previously rarely been subject to empirical study.



Andersen, T.A. (1981): Dialektbånd og databehandling. Ord & Sag 1: 11-18.

Kristiansen, T. (1998): The role of standard ideology in the disappearance of the traditional Danish dialects. Folia Linguistica 32 1-2: 115-129.

Monka, M. & H. Hovmark (2016): Sprogbrug blandt unge i Bylderup anno 2015. Danske Talesprog 16: 73-112.

Prokić, J. & J. Nerbonne (2008): Recognising groups among dialects. International Journal of Humanities and Arts Computing 2 1-2: 153-172.

Prokić, J. & J. Nerbonne (2013): Analyzing dialects biologically. In Fangerau, H. et al. (eds.): Classification and evolution in biology, linguistics and the history of science. Concepts, methods, visualization: 147-161. Stuttgart: Franz Steiner.

Puggaard, R. (2018): Realizations of /t/ in Jutlandic dialects of Danish. Linguistica Lettica 26: 368-395.

ZIta: a corpus of dental affricates variation in spoken Italian


Aim: The ZIta project will provide a first phonetic corpus of Italian dental affricates, in order to allow a fine-grained analysis of these phonemes. In world’s languages, dental affricates are rarely and marked phonemes, especially the voiced /dz/. Moreover, within the Italian dialects and regional varieties, these phonemes show a high degree of variation for what it concerns voicing, duration and place of articulation.

Methods: ZIta include dental affricates as realized in real Italian words, and in both controlled and semi-spontaneous speech by Italian L1 speakers of different sociolinguistic backgrounds. Each speaker will read a word-list and perform a map-task with a second speaker of the same sex and place of origin. Each task includes 54 tokens of real words containing a dental affricate equally distributed in four phonological contexts, that is word-initial (#C-), post-sonorant (SCV), and intervocalic singleton (VCV) and geminate (VCCV). All the tokens are annotated in PRAAT, by also distinguishing three sonority degrees (voiced, voiceless and mixed). The corpus will be constantly updated with new data and get freely accessible through GitHub: on this platform scholars will find audio samples in .wav format, transcriptions, metadata, PRAAT scripts, and referenced publications.

First results: For this presentation, a case-study of 4 male speakers of two different linguistic areas (Sicily and Lombardy) have been analyzed in the wordlist task, for a total amount of 216 tokens. The considered variables were sonority degree, consonant duration, and place of articulation as measured on the central point of the fricative segment.

A significant correlation has been found between the voicing of the dental affricate, the phonological context and the origin of the speaker: the two speakers from Sicily tend to produce more mixed affricates in the post-sonorant context (SCV, 66,7%), whereas Northern speakers present the mixed variant in word-initial position (#C-, 40,9%); moreover, the geminate intervocalic realization is more likely to be realized as voiced by Southern speakers (32,5%) and as voiceless by Northern speakers (39,4%). The chi-square analysis shows that these results are statically significant with a Cramer’s V of around 0,3, thus indicating a moderate effect of the correlation.

As for the consonant duration and the center of gravity, a significant correlation has been found only for voiced affricates in word-initial contexts, with the Northern speakers showing a more advanced place of articulation of the fricative segment (Mean difference = 904,17 Hz, t(22)=2,398, p=,025, Cohen’s d=1,02) and with slightly longer realization of the voiced dental affricate (Mean difference = 36,137 msec, t(22)=3,210, p=,004, Cohen’s d=1,3).

Conclusions: The ZIta project will built the first phonetic corpus dedicated to dental affricates as realized in different varieties of regional Italian. The corpus will help highlighting social structured variation of peculiar phones in a sociophonetic perspective.ù+

14:00-15:40 Session 14A: Lexical Change
Location: Brussels
Changes in the Historical Lexicon of Bernese Swiss German


Published between 1962 and 1997, the Atlas of German-speaking Switzerland (SDS) contains more than 1500 linguistic maps depicting dialect phenomena (cf. Baumgartner et al. 1962-1997). Specifically, it also contains data from 1944 on the greater area of Bern where NORMs were mainly taken into account. Since then, only very specific factors of the Bernese Swiss German varieties have been examined (e.g. Hodler 1969 on Bernese German syntax, Marti 1976 on Bernese German grammar more generally, and Siebenhaar 2000 on socially determined variants in the city of Bern), but the dialect has not been examined in its entirety.

I have collected new data for Bern and its greater area using select variables already documented in the SDS, allowing for a direct comparison of linguistic variables collected at different points in time, but with similar elicitation methods. To this end, 20 localities in the greater area of Bern were surveyed, taking into account three different generations (=a younger, middle-aged, and older generation) as well as one professional group (=farmers) for each locality. The aim of this study is to investigate which linguistic variables have undergone change, and then to determine/provide an explanation for the observed changes, or inversely, for the observed linguistic stability.  The data collection includes a range of lexical items, which could be considered today as archaic forms/relict forms even though they were documented as very stable variables in the SDS. Very often, the oldest group of speakers still use these items, the middle-generation of speakers has a passive knowledge of them, and the youngest generation of speakers either has no or scant knowledge of them, even if these lexical items designate concepts, which are still present and in use today as the different lexical variants of the word ‘cream’ in examples 1–3 show:

  1. Nidle (historical Bernese form)
  2. Rahm (modern Bernese form)
  3. Sahne (Standard-German form)

As the questionnaire of the study also includes motivation- and identity-related questions, I will in this talk not only focus on the newly emerged changes in the lexicon of Bernese Swiss German, but also look for the speakers' motivation to rename a concept such as cream in the example above. Furthermore, I will also show that the archaic forms should not be considered as lost but to play an important role when speakers of Bernese Swiss German consolidate their linguistic identity. Finally, I will discuss whether these lexical changes have to be considered as an expansion of the lexical field of a certain item or as a semantic shift and I will discuss whether there exist tendencies to choose a variant from other Swiss German dialects or from Standard German when renaming an archaic form.



Baumgartner Heinrich, Hotzenköcherle Rudolf (1962-2003). Sprachatlas der deutschen

Schweiz. Bern, Basel: Francke Verlag

Hodler, Werner(1969). Berndeutsche Syntax.Bern: Francke Verlag,

Marti, Werner (1985). Berndeutsch-Grammatik für die heutige Mundart zwischen Thun und Jura.Bern: A. Francke

Siebenhaar Beat, Stäheli Fredy, Ris Roland (2000). Stadtberndeutsch : Sprachporträts aus der Stadt Bern. Murten: Licorne-Verlag

A New Perspective on Subject Pronoun Expression: The Effects of Lexical Idiosyncrasy


This study explores verb-related predictors on subject pronoun expression (SPE) using 13,350 tokens from four speech communities: (a) Barranquilla, Colombia, (b) Medellín, Colombia, (c) New York City Colombians, and (d) Xalapa, Mexico. Although verb semantics has been found to be a robust SPE predictor, it has been explored using several different predictor/factor configurations (Carvalho, Orozco & Shin 2015; Otheguy & Zentella 2012; Torres-Cacoullos & Travis 2018; among others). Our multivariate regression results for lexical content and verb type largely concur with those of previous studies, confirming that verb semantics significantly conditions SPE. Nevertheless, these results fail to augment our collective knowledge. Thus, we analyze the lexical effect of the verb—a recently proposed alternative to verb semantics (Orozco 2018)—by testing verbs as random effects factors in further multivariate regression analyses. These results uncover opposite statistically significant tendencies between verbs in the same semantic category in all four communities. For example, in Barranquilla recordar ‘remember’ favors overt subjects but acordarse ‘remember’ has the opposite effect; in NYC ir  ‘go’ favors overt subjects but salir ‘leave’ favors null subjects; and in Xalapa and Medellín, respectively, ser ‘be’ favors overt subjects but estar ‘be’ exerts the opposite tendency. These results provide a more detailed account of how the verb conditions SPE. They show that grouping verbs according to semantic criteria in exploring how they condition SPE fails to uncover important differences. Moreover, cross-dialectal comparisons reveal that, despite tener ‘have’ being the most frequent verb in SPE contexts, the lexical effects of the verb lack the cross-dialectal consistency exhibited by all other internal SPE predictors (cf. Carvalho, Orozco & Shin 2015:xv; Torres-Cacoullos & Travis 2018). For instance, estar  ‘be’ favors overt subjects in Barranquilla, has a neutral effect in New York, and favors null subjects in Xalapa and Medellín. Concurrently, hacer ‘make, do’ promotes overt subjects in NYC and has a neutral effect in Barranquilla but favors null subjects in both Xalapa and Medellín. Therefore, these findings set the verb apart from all other linguistic SPE predictors. They suggest that the apparent differences in how the verb conditions SPE across different speech communities may be caused by the effects of lexical idiosyncrasy. That is, the lexical effects of the verb in a given speech community differ from those elsewhere; thus, challenging the premise that the internal conditioning on SPE is largely similar cross-dialectally. This analysis expands our analytical scope, as it improves the accountability of our findings on SPE. Further, this study, by offering a new perspective on the lexical effect of the verb, contributes to opening exciting research paths.  



Carvalho, Ana, Rafael Orozco, & Naomi Shin. (2015). Subject Pronoun Expression in Spanish: A cross-dialectal perspective. Washington DC: Georgetown UP.

Otheguy, Ricardo & Ana Celia Zentella. (2012). Spanish in New York: Language contact, dialectal leveling, and structural continuity. Oxford: Oxford UP.

Torres-Cacoullos, Rena & Catherine Travis. (2018). Bilingualism in the Community Code-switching and Grammars in Contact. Cambridge, UK.: Cambridge UP.

Lexis-oriented sociolinguistics: Methodological, theoretical, and analytical foundations


This paper introduces a methodological and theoretical framework that allows for analysing lexical variation and change. Although, traditionally, lexical variation is thought of as a throwback to dialect geography, recent studies have showcased the rich potential of investigating lexis in order to further our understanding of language in its social context (e.g. Sandow and Robinson 2018). By synthesising findings from a growing body of socio-lexical studies as well as presenting new data, we outline a proposal for a lexis-oriented sociolinguistic research programme.

Our analysis of socio-lexical variation is carried out from semasiological (one lexical form, multiple meanings) and onomasiological (one meaning, multiple lexical forms) perspectives in data collected from Sheffield and Cornwall, U.K. We use a combination of innovative elicitation methods that allow for managing socio-lexical variation effectively (cf. Lavandera 1978). These include spot-the-difference tasks, naming tasks, and experimental sociolinguistic interviews. In accounting for the methodological prerequisites we focus on the three points in more detail. Firstly, we consider what kind of data are required. For example, ‘what is a lexical sociolinguistic variable?’ from both onomasiological and semasiological perspectives and, if there are no absolute synonyms, how do we conceptualise lexical items as ‘variant’ forms of a ‘variable’? Secondly, we turn our attention to how best to collect such data in speech styles of varying formality. Lastly, we consider the data collected using these methodologies.

Our data show that patterns of lexical usage do mirror those of phonological and morpho-syntactic variables. However, we also observe that lexis provides unique insights into sociolinguistic variation which are both new and nuanced. Semasiological variation, e.g. gay ‘happy’, gay ‘homosexual’, or gay ‘lame’, appears to be largely conditioned by traditional sociolinguistic parameters such as age and social class and is highly susceptible to lifespan change. Specifically, we see that across the lifespan some speakers can reject not just particular variants, but the entire variable, e.g. by not using gay in any of its potential meanings. Onomasiological variation, e.g. tourist, holiday-maker, tourist, or emmet, despite exhibiting some these correlational patterns too, is highly amenable to strategic stylisation as social identity work and is, therefore, best explained by a language ideology framework. Using an identity questionnaire, we can quantify strength of local identity and show that identity is a stronger predictor of onomasiological usage than social class or gender.

We outline a framework, with supporting data, which can be employed in order to develop a more holistic understanding of how and why speakers vary in their usage at multiple levels of language. We provide evidence from two case studies which serve as proof of concept for a new lexis-oriented research programme which we suggest satisfies the lexis shaped gap in the variationist enterprise.



Lavandera, Beatriz. (1978). Where does the sociolinguistic variable stop? Language in Society, 7 (2), 171—82.

Sandow, Rhys. and Robinson, Justyna. (2018). ‘Doing Cornishness’ in the English periphery. In Natalie Braber and Sandra Jansen (eds.), Sociolinguistics in England, 333—361. London: Palgrave.

It’s all English to me: Examining attitudes towards verb conversions in British and Australian English


English speakers are renowned for their prescriptive tendencies and an increasing body of research has been dedicated to understanding speakers’ bugbears (see, e.g., Cameron, 2012; Curzan, 2014; Tieken-Boon van Ostade, 2018). Nevertheless, English-speaker attitudes towards verb conversions such as to hospitalise and to burglarise constitute an intriguing yet hitherto under-researched linguistic phenomenon. Which variant is considered to be part of the standard variety – to go to hospital or to be hospitalised; to burgle or to burglarise – depends greatly on the English variety under investigation (Ebner, in press). Some features, such as to hospitalise, are classified as Americanisms by British English speakers; however, the same variant can be perceived somewhat differently in Australian English – whose speakers are characterised by their extreme prescriptive tendencies, especially towards supposed American English (Severin & Burridge, in press).

In this paper, we illustrate how semantic differentiation of the variant to hospitalise has taken place in Australian English. Responses from British and Australian informants via an online questionnaire demonstrate that this particular feature has acquired a new meaning which contributes to more favourable perceptions of to hospitalise in Australian English compared to British English. The same phenomenon can be seen with to burglarise, albeit to a lesser extent, suggesting the existence of a further change in practice.

Making use of quantitative and qualitative data we collected from British and Australian participants, we further assess attitudes towards these allegedly nonstandard verb formations, examining the distribution of the attitudinal data according to social variables such as age and gender. Through this research, we demonstrate that linguistic phenomena stigmatised by one group can be openly accepted by others; as a consequence, we strengthen the claim that prescriptivism is a complex social phenomenon and not simply borne out of ignorance.



Cameron, D. (2012). Verbal Hygiene. 2nd edn. London: Routledge.

Curzan, A. (2014). Fixing English: Prescriptivism and Language History. Cambridge: Cambridge University Press.

Ebner, C. (in press). “Good guys” vs “bad guys”: Constructing linguistic identities on the basis of usage problems. Values and Multiplicity: Identity and fluidity in Prescriptivism and Descriptivism. Multilingual Matters.

Severin, A & Burridge, K. (in press). What do “Little Aussie Sticklers” value most? Values and Multiplicity: Identity and fluidity in Prescriptivism and Descriptivism. Multilingual Matters.

Tieken-Boon van Ostade, I. (2018). English Usage Guides. Oxford: Oxford University Press.

14:00-15:40 Session 14B: Standard Language
Location: London
Comparing standard language change dynamics in Flanders and Italy: style-shifting in Flemish and Italian commercials
PRESENTER: Jacoba Waumans


Several evolutions in late-modern society have challenged the idea of homogeneity within European languages. These dynamics have led to the emergence and spread of new supra-regional varieties competing with traditional standard languages in different formal contexts (Kristiansen & Coupland 2011; Ghyselen et al. 2016). Recently, the increased use of these varieties has been linked to dynamic prestige, mainly attributed to their omnipresence in mass media (Grondelaers et al. 2016), but generally little is known about their social meanings.

This paper focuses on two highly comparable language areas, Flanders (Belgium) and Italy, where intermediate varieties in the diaglossic standard-dialect stratification are believed to be undergoing a bottom-up process of informal standardization since the latter half of the twentieth century. These varieties are italiano neostandard in Italy (Cerruti et al. 2017) and tussentaal in Flanders (Grondelaers et al. 2016). We want to gain further insights into the role of dynamic prestige in the diffusion of these varieties by exploring their social values through a quantitative analysis of inter-speaker style-shifts in recent spoken commercials in both areas. By comparing similar data from the two language areas, we get a better understanding of the status of informal varieties in Europe.

We turn to advertising as a data source to analyze the prestige of language varieties. As marketeers stylistically exploit linguistic elements to build a brand identity based on positive values such as authenticity and reliability, commercials have been shown to be highly suitable to investigate social meanings (Van Gijsel et al. 2008).

We collected a corpus of Flemish and Italian commercials, broadcast on national radio and television (the Italian data between 2015 and 2019, the Flemish in March 2018). A colloquiality index was computed for each spot element, expressing how strongly tussentaal and italiano neostandard features are represented. Next, regression analyses were used to determine which aspects of the ads (e.g. type of spot element or product) were linked to increased usage of the colloquial features. For both languages, non-standard features are more frequent in situations where people interact, which confirms their supra-regional suitability in informal contexts. In Flanders, higher colloquiality indexes in trendy telecom commercials suggest a dynamic prestige. Nevertheless, our results contradict a presumed weakening of the “standard language ideology” in terms of authority and objectivity, as both standard varieties prevail in informative spot elements, and Italian children’s commercials normatively avoid neo-standard features.



Cerruti, M. et al. (eds.). (2017). Towards a new standard: Theoretical and empirical studies on the restandardization of Italian. Berlin / Boston: De Gruyter.

Ghyselen, A. et al. (2016). Studying standard language dynamics in Europe: Advances, issues & perspectives. Taal en Tongval 68(2): 75–91.

Grondelaers, S. et al. (2016). Destandardization is not destandardization: Revising standardness criteria in order to revisit standard language typologies in the Low Countries. Taal en Tongval 68(2): 119–149.

Kristiansen, T. & N. Coupland (eds.) (2011). Standard languages and language standards in a changing Europe. Oslo: Novus.

Van Gijsel, S. et al. (2008): Style shifting in commercials. Journal of Pragmatics 40(2): 205–226.

Ex situ focusing in the Cypriot Greek koine: on non-convergence to the syntax of the standard variety


In both Standard and Cypriot Greek in situ focusing is achieved through phonological prominence; as regards ex situ focusing, in Standard Greek there is movement of the focused constituent to a preverbal position while in Cypriot Greek ex situ focusing necessarily involves clefting (Tsiplakou, 2017):

 (1)    to                     ˈspiro                                        ˈvlepo

          the.SG.ACC   Spyros.SG.ACC           see.PRES.1SG

          ‘I am looking at SPYROS.’                                         (Standard Greek)


(2)     en      ton                     ˈspiron                   pu   θoˈro

          be      the.SG.ACC      Spyros.SG.ACC    that see.PRES.1SG

          ‘It’s Spyros that I am looking at.’                                (Cypriot Greek)

In recent research it has been argued that the Cypriot Greek koine displays convergence to Standard Greek in core areas of its phonology, morphology and syntax (Tsiplakou, 2014); however, ex situ focusing is one area of the syntax of the koine where convergence to the standard variety has not taken place; for instance, in a corpus from sociolinguistic interviews in the Cypriot Greek koine (Tsiplakou et al., 2016) there were no instances of focus movement. Interestingly, in production which is intended to be Standard Greek or, indeed, is otherwise in Standard Greek, e.g. in formal oral or written texts, clefts show up couched in Standard Greek morphology:

(3)       ˈine      eˈsis                 pu        ˈkanete                 ti                    ðiafoˈra

            be        you.PL.NOM that      make.PRES.2PL  the.SG.ACC difference.SG.ACC

            “It’s you that make the difference.”

This paper presents and discusses results from a quantitative survey which aimed to test whether clefting is treated by Cypriot Greek speakers as a syntactic structure that is common to Cypriot and Standard Greek, which may explain the prevalence of clefting in the otherwise structurally mixed Cypriot Greek koine. The questionnaire tested for the acceptability of focus clefts in Standard Greek (clefted adverbials, clefted subjects, clefted direct objects and clefted indirect objects). On the whole, the data indicate a strong preference for clefts as a syntactic focusing strategy, as their acceptance rate was significantly different from chance. The results point to imperfect acquisition of focus movement, the corresponding ex situ focusing structure of the standard variety, and this even at end state. An explanation is suggested in terms of the Interface Hypothesis, according to which acquisition of phenomena which pertain to an interface (e.g. syntax-semantics, syntax-pragmatics / syntax-discourse) is hard to achieve; focusing pertains to the syntax-discourse interface and acquisition involves not only acquiring the syntactic structure but also mapping it onto the relevant information structure, e.g. assigning it the properties of contrastivity, exhaustivity, etc.  (cf. Sorace, 2011; Tsiplakou, 2017).



Sorace, A. 2011. Pinning down the concept of “interface” in bilingualism. Linguistic Approaches to Bilingualism 1(1): 1-33.

Tsiplakou, S. 2017. Imperfect acquisition of a related variety? Residual clefting and what it reveals about (gradient) bilectalism. Frontiers in Communication 2. 10.3389/fcomm.2017.00017

Tsiplakou, S. 2014. How ‘mixed’ is a mixed system? The case of the Cypriot Greek koine. Linguistic Variation. Special Issue: Three Factors and Beyond, Vol. 1: Socio-Syntax and Language Acquisition: 161-178.

Tsiplakou, S., Armosti, S. & Evripidou, D.  2016. Coherence ‘in the mix’? Coherence in the face of language shift in Cypriot Greek. Lingua 172-173: 10-25.

Destandardisation in a standardising context? Reflexively used pronouns in Danish


With regard to third person singular pronouns with genitive function, standard Danish (and Scandinavian languages in general) distinguishes between pronouns co-referring with the gram­matical subject of their clause (i.e. pronouns used reflexively), and pronouns not co-referring with the subject. When the pronoun is co-referential with the subject, the possessive pronoun form sin (reflexive form) is chosen, when it is not, a (non-reflexive) genitive form of a personal pronoun is chosen (i.e. the Danish equivalents of his, her, ones and its).

In spoken Danish (as well as in varieties of Swedish and Norwegian), however, the use of re­flexive forms varies considerably, and non-reflexive forms are often uses in reflexive contexts:

han  havde    hundredhalvtreds                myggestik                                                            på           hans       venstre   side                  af            kroppen (std. Danish: sin)
he                     had                                        hundred and fifty                                              mosquito bites     on           his                    left                                                         side                         of            the-body

The norms regarding reflexively used pronouns are explicitly taught in the educational system and deviation from it is diligently corrected by many Danish teachers, at least as far as written lan­guage is concerned. When it comes to the spoken language, there has probably always been a great deal of variation, and none of the traditional dialects of Danish have norms with regard to reflex­ively used pronouns which are exactly like the prescribed norm of standard Danish. First and fore­most, the dialects of the western part of the country (the Jutland peninsula) have a less widespread use of reflexive forms (Jul Nielsen 1986).

The paper presents the results of a large scale survey of reflexively used pronouns in contempo­rary spoken Danish. The data material comprises recordings with 261 different speakers recorded in the period 1978-2010 (a large proportion of these have been recorded twice within this period) from four localities in Denmark. All reflexively used pronouns have been analysed using mixed effects models in order to uncover linguistic, social and geo-spatial factors constraining the use of reflexive vs. non-reflexive forms.

The results indicate that there is an ongoing change in the eastern part of Denmark (most pro­nounced in the capital Copenhagen) in which non-reflexive forms are gaining ground. The change is thus in the direction away from the prescribed standard. The paper argues that the development may be seen as a relaxation of the standard norm and a revalorization of formerly low status speech. This may explain why the standardization which in most other respects characterises the Danish speech community (e.g. Maegaard et al. 2013) seems to have stopped in Jutland with respect to re­flexive pronouns, as the traditional language use here resembles a radical version of the modern Copenhagen language use.



Jul Nielsen, B. (1986): Om pronominet sin i jysk. Danske Folkemål 28: 41-100.

Maegaard, M.; T. J. Jensen; T. Kristiansen & J. N. Jørgensen (2013): Diffusion of language change: accommodation to a moving target. Journal of Sociolinguistics 17 1: 3-36.

Competing norms of Standard pronunciation: Evidence from -ig-variation in Austria


From a sociolinguistics perspective, the term “standard” can be defined as “standard usage” (Gebrauchsstandard) meaning a speech form, which is used by highly educated speakers in formal speech settings like reading aloud (Kleiner 2014).This talk deals with the “standard usage” concerning unstressed <-ig> in Austria, e.g., König ‘king’. Although German pronunciation dictionaries propose the realization as [ɪç], studies like Kleiner (2010), Hildenbrandt / Moosmüller (2015) and Thévenanz (2018) show that <-ig> pronunciation differs depending on the speakers’ regional background, the phonetic context as well as the setting. Especially in southern Germany and Austria predominantly [ɪk]-variants occur. Thus, the plosive realization of <-ig> is accepted as being norm compliant in Austria. Although there are some studies on <-ig>-pronunciation, a comparative and nation-wide examination focusing on the “standard usage” of Austrian laypeople of different generations is still missing.   

The study presented in this talk aims to fill this gap in research. The data were collected within the Special Research Programme “German in Austria. Variation – Contact – Perception” and come from thirteen villages in Austria spread over different dialect regions (Central and South Bavarian, Alemannic, transition zones). In total, 1550 <-ig>-realizations from 52 speakers (26 woman) of two generations (18–35 and 65+) representing “standard usage” in four different settings (reading a text, reading a word list, naming pictures, translations of dialect sentences to standard German) were transcribed phonetically. Thereafter, the following (socio-)linguistic factors were examined: phonetic context, part of speech, setting, gender, age and regional background.

The results show that in Austrian “standard usage” <-ig> predominantly is pronounced like [ɪk] (91%). The investigation of the phonetic context and part of speech reveal that most fricatives occur in lexemes containing the morpheme -keit (e.g., Tätigkeit ‘activity’) (36.5%) and in final position in cardinals (e.g., zwanzig ‘twenty’) (14.7%). Furthermore, in non-written speech (picture naming, translating) more fricatives occur than during reading aloud (13.3% vs. 5.1%). In addition, a gender-effect could be proven, since women (10.9%) use more fricatives than men do (5.3%). In contrast, age does not significantly influence <-ig>-variation. Concerning regional background, more fricatives occur in Central Bavarian including the South/Central transition zone and Alemannic than in South Bavarian and the Alemannic/Bavarian transition zone.



Hildenbrandt, T. & Moosmüller, S. (2015): The pronunciation of -ig in three varieties of Austria. In Torgersen, E., Hårstad, S., Mæhlum, B. & Røyneland, U. (eds.): Language variation – European perspectives V. Selected papers from the seventh International Conference on Language Variation in Europe (ICLaVE 7), Trondheim, June 2013: 111–128. Amsterdam, Philadelphia: John Benjamins Publishing Company.

Kleiner, S. (2010): Zur Aussprache von nebentonigem -ig im deutschen Gebrauchsstandard. In: Zeitschrift für Dialektologie und Linguistik 77 (3): 259–303.

Kleiner, S. (2014): Die Kodifizierung der deutschen Standardaussprache im Spiegel der faktischen Variabilität des Gebrauchsstandards. In: Plewina, A. & Witt, A. (eds.): Sprachverfall? Dynamik – Wandel – Variation: 273–298. Berlin, Boston: Walter de Gruyter.

Thévenanz, C. (2018): Standardsprechsprache im Spannungsfeld zwischen Norm und Variation – Nebentoniges -ig in Österreich. Masterthesis. Universität Wien.

14:00-15:40 Session 14C: Urban Vernaculars
Location: Madrid
Benim: A new first-person pronoun in vernacular Swedish


The Turkish loanword benim has become a first-person (1.pers) ‘honorific’ pronoun in Stockholm’s multiethnic urban argot. The word has a self-aggrandizing indexicality, which can be intensified with a pronominal illeism (e.g., 1b).

Knowledge about this innovation is sparse. Kotsinas and Doggelito (2004) define benim as I/me (jag/mig). In an earlier analysis, I identified benim as especially frequent in Swedish slang (Young 2018:181), but a more thorough account of the pronoun’s pragmatic and social function is lacking.

The data comes from hip hop songs and YouTube vlogs from 2012 to 2019. Over 120 occurrences of benim were examined. Subject (1a, 1b) and object forms (2) are common. Possessive forms (3) are few.

(1a)       benim  kommer  göra  shorba

               I          come     to make mess

              ‘I will make a mess.’

(1b)     benim,    han     e      boss

               I          he      is      boss

              ‘I'm the boss.’

(2)          du      har    benim     i      din    mun

              You     have     me        in    your    mouth

             ‘You gossip about me regularly.’

(3)         benims  nia,   den    e      ej  latch

                My     niner     it      is    not  nice

              ‘My niner is mean.’

The use of benim resembles the 1.pers use of man in Multicultural London English (Cheshire 2013) and the Japanese 1.pers masculine honorific pronoun ore, the use of which elevates the speaker and deprecates the interlocutor (other-deprecatory, Miyazaki 2004).

Interestingly, illeisms as in (1b) occur only with a copula. In other words, I found no constructions like (4) in the corpus. Such ‘double subjects’, however, are not particularly marked in Swedish because left-dislocated topicalization like in (3) and (5) are highly typical (Johannessen 2014). Therefore, I suggest that the historical development of the illeism in (1b) could have been encouraged by the frequent occurrence of constructions like (3) and (5).


(4)       *benim,   hon   kommer   göra   shorba

                  I        she     come      to make     mess

              ‘I will make a mess.’

(5)       Johan,     han    är      bra     komisk                  (Johannessen 2014:1

               Johan     he     is      rather    funny

              ‘Johan is rather funny.’



Cheshire, J. (2013). Grammaticalisation in social context: The emergence of a new English pronoun. Journal of Sociolinguistics, 17(5), 608–633.

Johannessen, J.B. (2014). Left dislocation in main and subordinate clauses. Nordic Atlas of Language Structures 1, 8–15. http://www.tekstlab.uio.no/nals/system/archives/98/original/NALS-left-dislocation.pdf.

Kotsinas, U.B. & Doggelito, D. (2004). Förortsslang. Stockholm: Norstedts.

Miyazaki, A. (2004). Japanese junior high school girls’ and boys’ first-person pronoun use and their social world. I S. Okamoto och J.S. Shibamoto Smith (Red.) Japanese language, gender, and ideology: Cultural models and real people, 256–274. New York: Oxford University Press.

Young, N. (2018). ‘Copycats, ja dom shouf’ - Using hip hop to compare lexical replications in Danish and Swedish multiethnolects, University of Pennsylvania Working Papers in Linguistics 24(2), 174–184. https://repository.upenn.edu/pwpl/vol24/iss2/20/.

Multiethnolects as contact languages


In North-west Europe, new varieties of established languages have emerged in the past 40 years following rapid migration. Arrivals from other countries acquire local languages with varying levels of fluency, while their children, acquire something like ‘native’ proficiency. It is these (young) people’s speech that is in focus. To varying degrees, the new varieties differ structurally from the established varieties, and the obvious question arises as to the mechanisms behind these changes (Cheshire, Nortier & Adger 2015). A well-established term for these varieties is multiethnolect, but some researchers prefer such terms as urban contact dialect (Wiese 2017) in order not to prioritise ethnicity over other components, such as class.

I take an eclectic contact linguistics approach. I will evaluate the extent to which multiethnolects are allied to creoles and to koines, that is, whether there are parallel processes involved in the formation of multiethnolects and either of these other types, especially in terms of restructuring and simplification. In addition, I address the evidence that has been adduced both for and against the presence of intergenerational transmission, in Labov’s (2007) sense (Huenlich 2016). Trudgill’s deterministic theory of new-dialect formation (Trudgill 2004) is a relevant model, because it assumes that frequency and intensity of contact between speakers override social factors, especially ‘identity’: since a reasonable amount is known about the linguistic input to many of the multiethnolects, including the numbers of speakers involved, it makes sense to apply the determinism model in an attempt to account for the appearance of certain features and not of others. Useful in combination with Trudgill’s model is Mufwene’s earlier Founder Effect (Mufwene 1996), which argues that a founding population of speakers has a strongly disproportionate effect on the developing dialect compared to later arrivals. This is a plausible position to take for multiethnolects. However, this model does not account for the role of language learning in dialect formation, a factor which needs to be incorporated, at least in creole studies (Winford 2017).

The paper is based on primary data from large-scale projects on multiethnolects in Scandinavia, Germany and the United Kingdom. My conclusion is that multiethnolects share properties with other contact languages, but because of their particular patterns of development, essentially involving rapid language shift followed immediately by a process akin to koineisation, they have unique properties.



Cheshire, J., Nortier, J and Adger, D. (2015) Emerging Multiethnolects in Europe. QMUL Occasional Papers in Linguistics. QMUL, London.

Huenlich, D.W.K. (2016). The roots of ‘multiethnolects’: Effects of migration on the lexicon and speech of German-speaking school children. PhD, University of Texas, Austin.

Labov, W. (2007). Transmission and diffusion, Language 83 (2): 344–387.

Mufwene, S. (1996). The Founder Principle in creole genesis. Diachronica 13:1, 83‐134.

Trudgill, Peter (2004). New-dialect formation: the inevitability of colonial Englishes. Edinburgh University Press.

Winford, D. (2017). The Ecology of Language and the New Englishes: toward an integrative framework. In Filppula et al., (eds.) Changing English: Global and Local Perspectives, pp. 25-56. Mouton De Gruyter.

How pre-adolescents use ethnolectal features in urban areas: A case study of German-speaking Switzerland


Features of ethnolectal Swiss-German have often been associated and discussed in connection with adolescents with a migrant background (see e.g. Tissot et al. 2011). While these features have also been used by non-migrant adolescents for stylistic and indexical purposes (Auer 2002, Schmid 2017), little is known about their usage and the social meaning attached to them within the youngest speech group, i.e. pre-adolescent children. In this study, we investigate the extent to which such ethnolectal features (see, for instance, 1-2) have spread to the spoken vernacular of pre-adolescent Swiss-German children below the age of 12.

  1. Pragmatic expressions, e.g. Altä? (as a pragmatic marker)
  2. Syntactic expressions, e.g. omission of prepositions, articles, pronouns or auxiliaries, as in Chani bleistift? <Can I pen?> for Chani en bleistift ha? <Can I have a pen?>

The present study seeks to empirically investigate earlier anecdotal claims about the usage of these features by tapping into spoken data collected through the diapix task (Baker & Hazan 2011). Ethnolectal and other youth linguistic features are defined and extracted from the data on the basis of Auer’s (2002) list of seven ethnolectal features (see also Tissot et al. 2011) and on the basis of perceptual data provided by the caregivers, peers and the children themselves. The participants are all primary school children aged between 7 and 10 who live in the city of Winterthur, an urban area in Switzerland with roughly 100,000 inhabitants.

Preliminary analysis of the data indicates that use of ethnolectal features is indeed not restricted to adolescents but is already common among children younger than 12. The most frequently attested ethnolectal features are change of genus of nouns and omission of definite articles. The paper will discuss these findings against the backdrop of language external factors, such as mobility of the speaker, number of and contact with speakers of migrant backgrounds, age and gender of speaker, and will explore the influence of these factors on inter-speaker differences. In doing so, our paper contributes to a growing body of work charting children’s acquisition of language variation (De Vogelaer & Katerbow 2017).



Auer, P. (2002). ‚Türkenslang’. Ein jugendsprachlicher Ethnolekt des Deutschen und seine Transformationen. In Häcki Buhofer, A. (ed.): Spracherwerb und Lebensalter: 255-264. Tübingen: Francke.

Baker, R. & V. Hazan (2011): DiapixUK. Task materials for the elicitation of multiple spontaneous speech dialogs. Behavior research methods 43(3): 761–770.

De Vogelaer, G. & M. Katerbow (2017) Acquiring sociolinguistic variation. Amsterdam: John Benjamins.

Schmid, S. (2017): Differenzierungsprozesse im Sprachgebrauch von Jugendlichen in der Deutschschweiz. Zur sozialen Interpretation von ethnolektalen Sprechweisen in Schweizer Medien. Bulletin Suisse de Linguistique Appliquée (Vals-Asla) Nspécial(t. 1): 105–116.

 Tissot, F. et al. (2011): Ethnolektales Schweizerdeutsch. Sozio-phonetische und morphosyntaktische Merkmale sowie ihre dynamische Verwendung in ethnolektalen Sprechweisen. In Glaser, E., Schmidt, J.E. & N. Frey (eds.). Dynamik des Dialekts. Wandel und Variation: 319-344. Stuttgart: Steiner.

Phonetic features of (multi-)ethnic urban vernaculars in German-speaking Switzerland


Since the turn of the millennium, the phenomenon of so-called (multi-)ethnolects has been observed in different cities of German-speaking Switzerland (Tissot et al. 2011). This kind of speaking differs significantly from traditional Swiss German dialects. For the time being, however, a sociophonetic investigation of this phenomenon is lacking. This has already been done quite extensively in other (Western) European cities such as in the project about Multicultural London English (Cheshire et al. 2011).

We recorded 53 pupils in two different schools, which are either located in a very multicultural neighborhood in the city of Zurich or in a more monocultural one. To compare the varieties and speech styles of adolescents with and without migration background in the city of Zurich, various selected phonetic features are examined. We conducted individual interviews asking the subjects about their language biographies and let them describe a picture. In addition, the informants played a game of ‘spot the difference’ in pairs, we let them read out loud phonetically rich sentences and finally conducted group interviews. While the adolescents were naïve to the exact purpose of the recording during the first three speech styles (individual interview, ‘spot the difference’, read speech), they were explicitly told what the recordings are all about for the group interview because we are interested in their attitudes towards (multi‑)ethnolects.

In a third school, other adolescents were asked to evaluate to what extent the recorded pupils sounded mono- or multiethnic by means of a short ‘screening’ experiment (Bodén 2010). The realizations of different phonetic features are analyzed statistically. For example, Swiss German dialects are traditionally lacking voiced obstruents and speakers make use of a fortis-lenis distinction (Fleischer & Schmid 2006); instead, ethnolectal Zurich German is assumed to use voicing to distinguish obstruent pairs (Schmid 2012). Furthermore, we hypothesize that suprasegmental features (in particular speech rhythm) differ between the two groups of adolescents.

The research gap filled with our project is not only relevant for the fields of language variation, sociolinguistics, and phonetics, but is also of great social relevance. (Multi-)ethnic urban vernaculars or rather the speakers thereof are often associated with negative stereotypes, which we will investigate by means of two perception experiments at a later stage.



Bodén, P. (2010). Pronunciation in Swedish multiethnolects. In Quist, P. & Svendsen, B. (eds.): Multilingual Urban Scandinavia: 65–78. Bristol: Multilingual Matters.

Cheshire, J. et al. (2011): Ethnicity, friendship network and social practices as the motor of dialect change. Sociolinguistica 22: 1–23.

Fleischer, J. & Schmid, S. (2006): Zurich German. Journal of the International Phonetic Association 36: 243–255.

Schmid, S. (2012). Segmental features of Swiss German ethnolects. In Calamai, S. et al. (eds.): Sociophonetics, at the crossroads of speech variation, processing and communication: 69–72. Pisa: Scuola Normale Superiore.

Tissot, F. et al. (2011). Ethnolektales Schweizerdeutsch. In Glaser, E. et al. (eds.): Dynamik des Dialekts: Wandel und Variation: 319–344. Stuttgart: Steiner Verlag.

14:00-15:40 Session 14D: Sociophonetics
Location: Moscow
A Phonetic analysis of the which~witch merger in Edinburgh, Scotland


The which~witch merger is nearly complete in most English varieties. It has been complete in Southern Standard British English (SSBE) since the 19th century (MacMahon 1999), and is barely maintained in other varieties (Wells 1982; Trudgill 2004). However, the merger is still ongoing in many parts of Scotland (Brato 2007), including Edinburgh (Schüzler 2010; Rierson 2013). We consider it with respect to social class, and investigate whether it is phonetically abrupt or gradual.

Two studies of /ʍ/~/w/ merger in Edinburgh find Middle Class speakers participating in the merger (Schützler 2010; Dickson 2016). Schützler (2010) attributes this to accommodation to SSBE norms. Dickson (2016) finds Working Class men are also leading in the merger, which seems contradictory. To explore these results further, we compare speakers from two Edinburgh neighbourhoods which are local icons for social class membership: Morningside, (Middle Class) and Leith (Working Class) (Esling, 1978; Johnston, 1984).

Data are drawn from the Edinburgh Speaks corpus, a collection of interviews conducted for student projects at the University of Edinburgh. Interviews with 9 speakers (4 Leith, 5 Morningside) recorded in 2014 were coded for whether etymological /ʍ/ was realised as [ʍ] or [w], and [ʍ] tokens for duration of voicelessness.

The result of the categorical coding is striking: every Leither has less [ʍ] than every Morningsider. Neighbourhood has a significant effect in a mixed effects logistic regression (β = 2.4, z = 4.6). A neighbourhood effect is marginal for proportional [ʍ] voicelessness (β = 0.18, t = 1.7) , and there is no significant effect of speakers’ overall rate of merger (β = -0.07, z = -0.28). These results suggest that the /ʍ/~/w/ merger is a phonologically abrupt change. The direction of the neighborhood effect is counter-evidence to Schützler’s (2010) hypothesis regarding accommodation to SSBE norms. The fact that it is more advanced in Leith, where there is less Anglo-English contact than Morningside suggests this is at least in part an endogenously motivated change.



Brato, Thorsten. (2007) Accent variation in adolescents in Aberdeen: First results for (hw) and (th). Proceedings of ICPhS XVI. url: http://www.icphs2007.de/conference/Papers/1420/1420.pdf

Dickson, Victoria. (2016) The sound of Social Mobility: Investigating “New Middle Class” Speech in Edinburgh English. M.Phil. thesis, University of Oxford, UK.

Esling, John H. (1978) The identification of features of voice quality in social groups. Journal of the International Phonetic Association 7:18-23.

Johnston, Paul. (1984) Variation in the Standard Scottish English of Morningside. English World-Wide 4(2):133–185.

MacMahon, Michael K.C. (1999) Phonology. In Suzanne Romaine (Ed.), The Cambridge History of the English Language pp. 373-535. Cambridge: Cambridge University Press.

Reiersen, Øystein. (2013) Edinburgh Jockney? A socio-phonological study of accent variation and change in Edinburgh English. Masters thesis, University of Bergen.

Schützler, Ole. (2010) Variable Scottish English consonants: The cases of /ʍ/ and non-prevocalic /r/. Research in Language 8:5–21.

Trudgill, Peter. (2004) New-dialect formation: The inevitability of colonial Englishes. Edinburgh: Edinburgh University Press.

Wells, J. 1982. Accents of English. Cambridge: Cambridge University Press.



This paper addresses the methodological issue of reconstructing diachronic change in patterns of variation, by way of a case study of North Germanic preaspiration. As argued, among others, by Pétur Helgason (2002), variable preaspiration was characteristic of Proto-Norse and has been inherited in many present-day varieties. In some languages, notably Icelandic, it has stabilized as a phonological phenomenon, but in others it remains variable. The patterns of this variation, unsurprisingly, also differ between varieties (e.g. van Dommelen et al. 2011). In this paper I address the diachronic implications of this variation.

I propose to treat the rise of differences in patterns of variation as an instance of phonetic change without phonological change (e.g. Fruehwald 2017). I present an acoustic study of preaspiration in two varieties of Norwegian: Western (Rogaland) and Northern. Rogaland Norwegian is commonly described as a preaspirating variety (e.g. Oftedal 1947), whereas for Northern Norwegian there are only a few scattered reports of preaspiration, but no detailed. Based on the acoustic study, I suggest that preaspiration may have been under-reported in the literature. Further, I use Bayesian regression modelling with Stan to quantify the attested patterns of variation. Model comparison and posterior predictive checks show robust differences in the structure of variation between Western and Northern Norwegian. I argue that a diachronic interpretation of differences in the structure of variation is possible, and that it can provide important clues for historical enquiry.

In the particular case of North Germanic, I focus of the effects of (phonological) vowel length and (phonetic) vowel duration on preaspiration duration. I show that despite their differences, the attested variation patterns demonstrate a shortening of preaspiration after (phonologically) long vowels. I argue that this shortening provides a precursor to the otherwise somewhat mysterious sound change found in several North Germanic varieties, whereby aspirated stops are lenited to voiceless unaspirated stops intervocalically (cf. Steblin-Kamenskij 1974). Specifically, I argue that this mechanism explains why lenition of aspirated stops is neutralizing in varieties where unaspirated stops are categorically voiceless (e.g. Icelandic), but non-neutralizing when the unaspirated stop series allows (variable) voicing (e.g. Helgeland Norwegian). This, in turn, has important implications for the reconstruction of the age of preaspiration (supporting Pétur Helgason’s dating of it to Proto-Norse) and for its relationship to phenomena such as ‘West Jutland stød’ (e.g. Ringgaard 1960; Kortlandt 2003).



Fruehwald, J. (2017): The role of phonology in phonetic change. Annual Review of Linguistics 3(1): 25–42.

Kortlandt, F. (2003): Glottalization, preaspiration and gemination in English and Scandinavian. Amsterdamer Beiträge zur älteren Germanistik 58: 5–10.

Oftedal, M. (1947): Jærske okklusivar. Norsk tidsskrift for sprogvidenskap 14: 229–235.

Pétur Helgason (2002): Preaspiration in the Nordic languages. Stockholm: Stockholm University.

Ringgaard, K. (1960): Vestjysk stød. Aarhus: Universitetsforlaget.

Steblin-Kamenskij, M. (1974): The Scandinavian consonant shift. Arkiv för nordisk filologi 89: 1–29.

Van Dommelen, W., Holm S. & Koreman, J. (2011): Dialectal feature imitation in Norwegian. In Proceedings of ICPhS XVII, Hong Kong: 599–602.

Fundamental Frequency Range in Welsh-English bilingual speech: An investigation of cross-linguistic differences and sociolinguistic variation


Recent work on fundamental frequency range (FFR) in Welsh-English bilingual speech in north west Wales (where the majority of the population speak Welsh) has reported significant cross-linguistic differences between the two languages for female speakers but not for male speakers (Ordin & Mennen 2017). This complements the results of work on segmental variation in north Wales which also found that women were more likely to differentiate between realisations of /l/ in Welsh and English (Morris 2017). It is not known, however, the extent to which FFR varies both within and between the two languages in different areas of north Wales (particularly in areas where Welsh is not spoken by the majority) and in the speech of those who have acquired Welsh through immersion education.

The current study therefore aims to examine both areal variation and the influence of other social factors on FFR in three areas of north Wales. Specifically, I address the following research questions:

  1. To what extent do Welsh-English bilinguals from north Wales have distinct FFR in their two languages?
  2. Are there differences between western, central, and eastern areas of north Wales and to what extent can this be accounted for by the social history of the Welsh language in these areas?
  3. To what extent do social factors influence FFR both within and between Welsh and English? Particularly, do speaker sex and home language influence variation?

Data were collected from 48 Welsh-English bilinguals aged 16–18. The sample was stratified equally by area (western/central/eastern), speaker sex (male/female), and home language (Welsh/English) in order to examine the extent to which FFR is influenced by extra-linguistic factors. Participants were asked to read The North Wind and The Sun in both languages. The recorded reading passages were then segmented into intonational phrases and f0 level and span were analysed acoustically as dimensions of FFR (cf. Mennen et al. 2012).

The results of the ongoing data analysis will be discussed with reference to dialectal variation and the influence of language background on bilinguals’ speech production. Finally, I will discuss the results in the context of previous perceptual research which shows an association of certain segmental and suprasegmental features with a particularly Welsh-influenced English accent (e.g. Williams et al. 1996).



Mennen, I. et al. 2012. Cross-language differences in fundamental frequency range: A comparison of English and German. The Journal of the Acoustical Society of America 131(3): 2249–2260.

Morris, J. 2017. Sociophonetic variation in a long-term language contact situation: /l/-darkening in Welsh-English bilingual speech. Journal of Sociolinguistics 21(2): 183–207.

Ordin, M. & I. Mennen. 2017. Cross-linguistic differences in bilinguals' fundamental frequency ranges. Journal of Speech, Language, and Hearing Research 60(6): 1493–1506.

Williams, A. et al. 1996. Perceptual dialectology, folklinguistics, and regional stereotypes: Teachers’ perceptions of variation in Welsh English. Multilingua 15(2): 171–199.

14:00-15:40 Session 14E: Panel Distributional Semantics
Location: New York 3
Token-based distributional models and lexical lectometry


Type-based distributional semantics as embodied in vector space models has proven to be a successful method for the retrieval of near-synonyms in large corpora. These words have then been used as variants of lexical sociolinguistic variables in lectometric studies (see Ruette, Geeraerts, Peirsman, & Speelman [2014] ). However, a limitation of type-based VSMs is that all senses of a word are lumped together into one vector representation, making it harder to control for polysemy and subtle contextual distinctions. In addition, operating at the lemma level, these type-based VSMs are not able to pick out the relevant corpus occurrences that are the input for the lectometric calculations.


Our paper reports on methodological research aiming at better semantic control in the lectometric use of VSMs. Several methods will be compared w.r.t. their value for lectometric studies of lexical variation. First, staying at the level of type-based VSMs, we measure the degree of semantic similarity in a cluster of potential near-synonyms, and use that degree as the weighting measure for the frequency of the lexical variants in the variable. The other solutions build on token-based VSMs to disambiguate senses of lexical variants (Heylen, Speelman & Geeraerts 2012). This type of VSMs identifies different meaning/usage tokens of a word in a corpus that are represented as token clouds in a multidimensional space, with token clusters revealing the senses of the word. By superimposing the token clouds of the lexical items, one can distinguish which meanings are shared by near-synonyms and determine the ‘semantic envelope of variation’. The remainder of the study will show how the calculation of the overlapping area can be carried out, using a set of cluster overlap indices evaluated in Speelman & Heylen (2015).


The fine-tuning of VSM-based lectometry targeted here contributes to the scaling up of lexical variationist research, by providing methods for dealing with corpora whose size exceeds manual analysis. At the same time, token-based VSMs comply with the need of detailed analysis by allowing the possibility of zooming in on the behavior of individual tokens in order to determine more subtle contextual distinctions.



Heylen, K., Speelman, D., & Geeraerts, D. (2012). Looking at word meaning. An interactive visualization of             Semantic Vector Spaces for Dutch synsets. In M. Butt, S. Carpendale, G. Penn, J. Prokic, & M. Cysouw             (Eds.), Proceedings of the EACL-2012 joint workshop of LINGVIS & UNCLH: Visualization of Language             Patters and Uncovering Language History from Multilingual Resources 6–24. Avignon, France:             Association for Computational Linguistics.

Ruette, T., Peirsman, Y., Speelman, D., Geeraerts, D. (2014). Semantic weighting mechanisms in scalable lexical             sociolectometry. In Szmrecsanyi B., Waelchli B. (Eds.) Aggregating Dialectology, Typology, and Register.             Analysis Linguistic Variation in Text and Speech. Berlin: de Gruyter, 205–230.

Speelman, D. & Heylen, K. (2017). From dialectometry to semantics. In Wieling, M., Kroon, M., Van Noord, G.,             & Bouma, G. (Eds.). From Semantics to Dialectometry. Festschrift in honor of John Nerbonne. UK: College             Publications. 325–334.

Synopsis and discussion

ABSTRACT. There is no abstract in the word document.

14:00-15:40 Session 14F: Panel Frequency
Location: Paris
Speech perception through a phonological filter: neurophysiological evidence


The Mismatch Negativity (MMN) is an automatic neural response that occurs when an auditory regularity is disrupted. The measure has been used extensively in speech perception research, and shows sensitivity to a range of factors from low-level auditory properties to high-level phonological properties like phonemic versus allophonic status. Eulitz & Lahiri (2004) reported a directional asymmetry of the MMN: a sequence of [o] disrupted by [ø] resulted in a larger MMNs in German speakers than vice versa. They presented this as evidence for coronal underspecification in the representation of [ø]. Elaborating on this finding within the Featurally Underspecified Lexicon framework (FUL; Lahiri & Reetz, 2010), I will present a line of research aiming to expand our understanding of the relation between phonological representations and their neural reflections in the MMN.

In the French vowels [u], [o], [y] and [ø], underspecification is assumed in the front and mid vowels, while back and high vowels have the features [dor] and [high], according to FUL. Analogous to Eulitz & Lahiri (2004), MMN responses from French speakers to these vowels showed directional asymmetries: larger MMNs for changes from back to front vowels than vice versa, suggesting underspecification of front vowels, and larger MMNs for changes from high to mid vowels, suggesting underspecification of mid vowels. However, very different response patterns were found in another study with French speakers, contrasting the vowels [i], [e], [ɛ] and [a]. Rather than larger MMNs for a change from the peripheral vowels [i] and [a] towards the mid vowels [e] (and possibly [ɛ]) than vice versa, an overall effect of direction was found with larger MMNs for changes to a higher vowel, and a gradient effect of distance in the vowel space. These findings indicate that either these vowels have phonological representations that are different than proposed by FUL, or that the MMN in this paradigm also reflected other properties than the phonological representations. To elucidate the role of acoustic stimulus properties, an MMN study with Dutch speakers was conducted to compare responses to the [o]-[ø] contrast with non-speech analogues. The most complex non-speech analogues, matching part of the spectral envelope of the vowels, were found to elicit the same directional asymmetry as the vowels, in contrast to more simple analogues. This suggests there might be a special role for certain spectral shapes in the mapping of the acoustic signal onto phonological categories.

Lexical storage of reduced word pronunciation variants: evidence from frequency effects


In many languages, words show substantial variation in how they are pronounced in informal situations. For instance, French ministre is mostly pronounced as /ministr/ in formal speech, but if often reduced to /minist/ and /minis/ in informal speech. In this talk, I will discuss the question how listeners understand reduced word pronunciation variants. I will argue, on the basis of two series of word recognition experiments in French, that many word pronunciation variants may be stored in our mental lexicon. Recognition of these variants may then involve the activation of their lexical representations.

In both series of experiments, we tested native listeners and Dutch learners of French. In the first series [1], participants had to classify auditorily presented words as “real words” or “pseudo words”. Among the real words and pseudo words, participants heard real words with a schwa in the first syllable. This schwa was either present (e.g. /pəluz/ for pelouse ‘lawn’) or absent (e.g. /pluz/ for pelouse ‘lawn’). After this lexical decision experiment, participants rated for every schwa word in the experiment how often they thought that the reduced variant without schwa occurs compared to the full variant (e.g. /pluz/ versus /pəluz/). The results showed that listeners recognized word pronunciation variants more quickly the more often they think that these occur. Importantly, the learners’ reaction times were better predicted by their own variant ratings than by the natives’ ratings. Listeners’ own exposure thus appears to determine their processing speed of word pronunciation variants.

In the second series of experiments [2], we investigated whether the recognition of a reduced word pronunciation variant is more hindered by a low frequency of occurrence or by a larger deviance from the full variant. We conducted a cross-modal priming experiment in which participants heard prime sentences containing a prime word. Right after the prime word, participants saw a printed target word, for which they performed lexical decision. The experimental target words ended in an obstruent-liquid-schwa cluster (e.g. ministre). The prime word was either unrelated to the target word or was the same word as the target word and was pronounced in full (e.g. /ministr/), in a low-frequency, mildly reduced variant (e.g. /minisr/), or in a frequent, but highly reduced variant (e.g. /minis/). Most importantly, the results showed that participants reacted more quickly when the prime word was the frequent, but highly reduced variant of the target word than when it was the low-frequency, mildly reduced variant.

These results suggest that frequency of occurrence plays an important role in the recognition of reduced word pronunciation variants and consequently that these frequencies are mentally stored. I will argue that this result can most naturally be explained in an account assuming that the mental lexicon contains representations for reduced word pronunciation variants with their frequencies of occurrence.

[1] S. Brand & M. Ernestus (2018). Listeners’ processing of a given reduced word pronunciation variant directly reflects their exposure to this variant: evidence from native listeners and learners of French. Quarterly Journal of Experimental Psychology 71, 1240-1259.

[2] S. Brand & M. Ernestus (submitted). Understanding reduced words: the relevance of reduction degree and frequency of occurrence.

15:40-16:15Coffee Break
16:15-17:30 Session 15B: Adolescents
Location: London
BATH and TRAP Variation amongst Early Adolescents in West Cornwall


It has been argued that the urban turn in sociolinguistics means that rural varieties, such as those found in Cornwall and the wider South-West of England, have been notably under-researched (Britain 2012). In this paper, I will show that rural adolescents have not simply succumbed to the effects of standardization and, just like their urban counterparts, are innovative in their language use. I do this through an analysis of the vowels in the BATH and TRAP lexical sets in the speech of forty-two schoolchildren from west Cornwall. Data were collected from two structured elicitation tasks (map tasks and word lists), eliciting 850 BATH tokens and 1,400 TRAP tokens. I examine how these variables are stratified according to the macro-social categories of social class and gender, and consider local orientation, quantified using a questionnaire, as a variable in order to provide insight into the micro-level issues surrounding social class.

In west Cornwall, the BATH vowel is traditionally /æː/, while the TRAP vowel is slightly raised (as in RP) and variably lengthened to /æː/. Acoustic analysis of the BATH vowel indicates that, despite the expected shift towards the RP variant /ɑ:/, an innovative shortened form of the traditional west Cornish English variant is emerging amongst certain groups of early adolescents in west Cornwall. In tandem, the traditional long TRAP vowel in Cornwall is shortening to a more RP-like variant. This shift may be explained by the social meaning of ‘long <a>’ in the South-West, as lengthened TRAP has been shown to be perceptually linked to South-Western rural identities, particularly those connected to the to the concept of the uneducated and unsophisticated ‘farmer’ (Montgomery & Moore 2018). Variants linked to rurality face particular stigma due to prevalent stereotypes of rural areas as, “backward, conservative, boring, dangerous, threatening, ‘uncultured’ and uneducated” (Britain 2017: 174). Therefore, this paper posits that speakers participating in this shift may be maintaining a sense of ‘localness’ in their fronting of the BATH vowel, but avoiding the stigmatized ‘farmer’ associations of the lengthened variant by keeping their BATH vowels short.

In addition to providing much-needed insights into the status of two traditional features of west Cornish English, this paper also demonstrates how different acoustic elements of a variable may carry subtly different meanings, and how speakers may use these creatively to project desired identity traits.



Britain, D. (2017). Which way to look?: Perspectives on “Urban” and “Rural” in dialectology. In Montgomery, C., & Moore. E (eds.): Language and A Sense of Place: 171–187. Cambridge: Cambridge University Press.

Britain, D. (2012). Countering the urbanist agenda in variationist sociolinguistics: dialect contact, demographic change and the rural-urban dichotomy. In Hansen, S., Schwarz, C., Stoeckle, P. & Streck, T. (eds.): Dialectological and Folk Dialectological Concepts of Space: 12–30. Berlin: Mouton de Gruyter.

Montgomery, C. & Moore, E. (2018). Evaluating S(c)illy voices: The effects of salience, stereotypes, and co-present language variables on real-time reactions to regional speech. Language 94(3). 629–661.

“Your muscles are [weːkɪn] like harder”: Stylistic variation and identity shifts in a Scouse secondary school classroom


In this paper we discuss the use of Liverpool English, or “Scouse”, in educational contexts. We consider the results of a study which includes audio recordings of lessons from several secondary schools in the Liverpool area as well as semi-structured interviews with students and teachers. Drawing upon the premise that speakers are not carriers of a dialect but agents in constructing social reality through deploying resources they have available in contexts specified to conditions of use (Snell, 2013; Blommaert, 2005), we explore how adolescents use and perceive Scouse in the classroom for purposes of self-construction and differentiation.

We focus on 14 hours of audio recordings of Year 9 (age 13-14) English lessons in an all-girls Catholic school on the Wirral in Merseyside. The classroom data includes examples of students speaking to each other, speaking directly to teacher, speaking aloud in front of the class, and reading pre-prepared passages aloud. To this end, we explore the effects of stylistic variation. We particularly focus on the NURSE~SQUARE merger that is now increasingly observable in varieties elsewhere in the wider Merseyside area (see Newbrook 1999; West 2015)

Analysis of the data illustrates that realisations of NURSE tend to have a fronted vowel similar to [eː]. This aligns with the wider tendency across in the north-west of England where NURSE and SQUARE have merged. In parts of Merseyside, particularly Liverpool, this merger is in the direction of a fronted variant (Watson and Clark 2013) and so is not altogether unexpected. However, students’ NURSE variants also display F2 variation which correlates with speech style, highlighting a pattern where the highest proportion of merged, fronted vowels occurred when students spoke to each other. As the style of speech becomes more formal, NURSE tokens display a lower F2 which approaches (but is never as low as) RP [ɜː], indicating an overall fronting of NURSE in Merseyside English. Students’ audiences similarly affect the fronting of NURSE as they tended to realise a more fronted “Scouse” variant when speaking to other students, even when presenting before the whole class, but a more backed variant when speaking directly to the teacher, despite the teachers also being from Merseyside.

Drawing upon the quantitative results as well as the interview data, we argue that although students perceive the classroom as an abstract formal setting, it is specifically the English teacher as a figure to which their speech style seems to respond. Students temporarily construct themselves as conforming to a perceived correctness when addressing the teacher but they reclaim their Scouse identity and its associated social values when their perception of audience diversifies. The use of NURSE and its variants, thus, function as a resource for subtle but focused identity shifts despite the reported relaxed attitude towards Scouse at the specific school. This practice reflects an association between standardness and educational attainment that seems to be valued in the specific school.



Today’s youths are avid users of social media and computer-mediated communication (CMC). In their informal digital writings, they use a language variant called ‘digi-talk’. Many parents fear that digi-talk harms youths’ literacy skills or formal writings, such as at school (Spooren 2009). To determine if such worries are at all justified, we conducted two large-scale studies with Dutch youths in secondary and tertiary education.

The first study measured youths’ (N = 338) social media use through extensive surveys. In the second study (N = 408), half of the participants were primed with social media, specifically WhatsApp, while the other half performed a non-CMC related control task. All participants wrote school texts: essays in the survey study, stories in the experimental study. The 746 school writings were first analysed in terms of the higher-level features of lexical richness, syntactic complexity, writing productivity, and formality: relations with CMC use were discovered in the survey study, especially for lower educated youths (Verheijen, Spooren, & Van Kemenade, submitted), but no direct impact of WhatsApp use was found in the experimental study (Verheijen & Spooren, subm.). The current research presents follow-up analyses, focusing on orthographic surface features, because fears of language deterioration are often fixed on spelling. The 746 texts were again analysed, this time for three kinds of orthographic ‘deviations’: textisms (orthographically unconventional words), non-standard orthographic details (punctuation, capitalisation, spacing, diacritics), and misspellings. We calculated the relative frequency of these features to the total number of words per school text.

Quite surprisingly, fewer spelling errors were found in the school writings of (a) youths who were primed with WhatsApp immediately before writing a story in the experiment, than youths in the control groups, especially for adolescents (evidence of a direct impact!), as well as (b) youths who reported owning smartphones in the survey, than youths who owned old-fashioned or no mobile phones. Yet more textisms occurred in the essays of youths who reported using predictive and corrective dictionaries in CMC, than in those of youths who did not. Dutch youths’ CMC and smartphone use were thus positively related to their orthographic performance in school writings, in terms of fewer spelling errors, but their use of auto-correction and auto-completion were negatively related, evident from more textisms. This suggests that the informal language variant of digi-talk as used by many youths in their digital writings is not dangerous to the orthography of their school writings, as long as youths formulate their own words and sentences rather than passively rely on word predictors and correctors.



Spooren, W. (2009). Bezorgde ouders? De relatie tussen chat en schrijfkwaliteit. In W. Spooren, M. Onrust, & J. Sanders (Eds.), Studies in Taalbeheersing 3 (pp. 331–342). Assen: Van Gorcum.

Verheijen, L., & W. Spooren (submitted). The impact of WhatsApp on Dutch youths’ school writing.

Verheijen, L., W. Spooren, & A. van Kemenade (submitted). Relationships between Dutch youths’ social media use and school writing.

16:15-17:30 Session 15C: Lifespan Change
Location: Madrid
Linguistic malleability after university: Investigating the effect of post-university trajectories on linguistic practices


This project explores individuals’ linguistic malleability across emergent adulthood, generally considered to be the years 18-25 (Arnett 2000). To date, panel research has focused on linguistic maturation effects in children (Van Hofwegen and Wolfram 2010), the influence of academic choices at the cusp of high school and university (Wagner 2012), or the roles of speakers’ socio-economic trajectories between middle adulthood and older age (Sankoff and Wagner 2011). These analyses suggest that speakers demonstrate sensitivity to the linguistic market, producing more standardized variants in life-phases and labor contexts that are subject to normative practices (Wagner 2012; inter alia). Though the life-phase of emergent adulthood has been studied by sociologists, there is a dearth of research on the ways in which individuals who attend university exhibit change their linguistic patterns as they transition to the post-academic job market.

Our research expands on previous work in two ways. First, by exploring six individuals who were recorded in their 2nd/3rd year of undergraduate studies at the University of Newcastle (i.e. between 19 and 21 years old) and again 3-5 years later, we expand the window of analysis by adding an underexplored age-bracket to the investigation of post-adolescent variability. Second, while most previous panel analyses tend to focus on single variables or on variables that are phonetically related, our research follows Sankoff and Blondeau’s (2007) call to comparatively explore malleability in the use of different types of variables across the life-span: stable variation in the realisation of (ing); changes in progress in the realisation of (t), the (FACE) vowel, and the quotative system. We use mixed-effects regressions to model the linguistic and social factors that correlate with intra-speaker malleability.

We argue that speakers’ post-university trajectories represent the outcome of the “continuous modification and reconstruction of … linguistic identit(ies) over the course of the lifespan” (Dickson & Hall-Lew 2017: 249). In aggregate, our research provides one more piece in the objective to map the factors that influence individuals’ linguistic malleability across the post-adolescent life-span (Bowie and Yaeger-Dror 2015).



Arnett, J. (2000). Emerging adulthood: A theory of development from the late teens through the twenties. American Psychologist 55(5): 469-480.

Bowie, D. & M. Yaeger-Dror. (2015). Phonological change in real time. In P. Honeybone & J. Salmons (eds.): The Oxford Handbook of Historical Phonology. 603-618. Oxford: Oxford University Press.

Dickson, V. & L. Hall-Lew. (2017). Class, gender, and rhoticity: The social stratification of non-prevocalic /r/ in Edinburgh speech. Journal of English Linguistics 45(3): 229-259.

Sankoff, Gillian & Hélène Blondeau. (2007). Language change across the lifespan: /r/ in Montreal French. Language 83(3): 560-588.

Sankoff, G. & S. Evans Wagner. (2011). Age grading in the Montréal French future tense. Language Variation and Change 23(3): 275-313.

Wagner, S. Evans. (2012). Real-time evidence for age grad(ing) in late adolescence. Language Variation and Change 24(2): 179-202.

Van Hofwegen, J. & W. Wolfram. (2010). Coming of age in African American English: A longitudinal study. Journal of Sociolinguistics 14: 27-52.

Lifespan developments in spoken and written data: A case study of Margaret Atwood’s restrictive subject relativizers


Recent work shows that one of the main assumptions of the apparent time construct – that speakers’ idiolects remain stable after adolescence – does not always hold up. Some speakers participate in ongoing language change (Sankoff, 2004; Sankoff & Blondeau, 2007), while others undergo retrograde change as they age (Wagner & Sankoff, 2011). Understanding how individual grammars make these adjustments requires real time data.

Since real time studies are time consuming and expensive, the present study investigates what we can learn about linguistic malleability by considering written data in addition to spoken data. The focus is on one individual, Canadian writer Margaret Atwood, and on restrictive subject relativizers – specifically, the variation between who and that with animate antecedents (as in “the woman who/that writes), which is stable and age-graded (D’Arcy & Tagliamonte, 2010), and the variation between which and that with inanimate antecedents (as in “the paper which/that investigates relativizers”), which is undergoing change (Jankowski, 2013)

The data come from two sources: The CBC Digital Archives, a collection of 20 television and radio interviews with Atwood (recorded 1967-2000) and the 16 novels Atwood published to date (published 1969-2016). The data were coded for a variety of social and linguistic factors and analyzed using comparative sociolinguistic methods and mixed effects modelling.

Results show that both spoken and written data offer crucial insights into lifespan developments. The spoken data show that Atwood has remained stable throughout her lifetime, consistently favouring standard variants. This is unsurprising given Atwood’s professional stature, but it contrasts markedly with Toronto, where Atwood has spent most of her life, demonstrating that individuals can and do deviate from communty-level patterns. However, the written data also show that Atwood is participating in the change from which to that. This may be due to editorial influence, a possibility currently under investigation. The written data further shows that Atwood uses relativizers for character development in her historical novel Alias Grace, which has markedly different patterns from her other novels. This suggests that relativizer choice may not be as covert as previously assumed and that lifespan change intersects with stylistic variation. Together, these findings indicate that researchers can gain valuable insights into lifespan developents by taking multiple modes of communication into account.



D’Arcy, A., & S. A. Tagliamonte (2010). Prestige, accommodation, and the legacy of relative who. Language in Society 39: 383–410.

Jankowski, B. L. (2013). A variationist approach to cross-register language variation and change. Unpublished Ph.D. dissertation. Toronto, ON: University of Toronto.

Sankoff, G. (2004). Adolescents, young adults and the critical period: Two case studies from “Seven Up.” In Fought, C. (ed.), Sociolinguistic variation: Critical reflections: 121–139. Oxford, UK: Oxford University Press.

Sankoff, G., & H. Blondeau (2007). Language change across the lifespan: /r/ in Montreal French. Language 83: 560–588.

Wagner, S. E., & G. Sankoff (2011). Age grading in the Montréal French inflected future. Language Variation and Change 23: 275–313.

Six informants recorded in 1967, 1996 and 2018: Different factors influencing individual language variation, change and stability


In 1967–68, Nordberg recorded 85 individuals (born 1876–1951) for the first large-scale sociolinguistic investigation of language variation and change in the Nordic countries, and in 1996, I made the recordings for the project Continuity and Change in Present-Day Swedish: Eskilstuna Revisited, directed by Nordberg. The new study combined the merits of a panel study, with 13 re-recorded informants (born 1913–1950), and a trend study, with 72 new informants. All informants were natives of Eskilstuna. As a result, both individual and generational language change over a period of nearly 30 years could be studied.

Seven morphological and morphophonological variables were analyzed using quantitative variationist methods. The quantitative approach was combined with qualitative explanations, specifically concerning the individuals in the panel study, along with the concepts of integration and social mobility. Each variable has two distinct variants, one standard form, which agrees with the written form, and one traditionally used in spoken language in Eskilstuna. The non-standard forms of these variables are not unique for Eskilstuna; they are or have been more or less common in colloquial speech over larger or smaller areas in Central Sweden.

The expectation was that all the variables were in the process of rapid change from the regional dialect towards the spoken standard; however, the rate of change at the level of the community had been low. The changes in the seven variables over the 30 years in the direction of standard speech was manifested both as individual change and as generational change, but different comparisons between the panel and the trend study demonstrated that a trend study is the most reliable way of investigating real-time change. The panel speakers were not representative of the Eskilstuna population in 1996; on average they used more local variants than the trend speakers, as they also did in 1967. The results from the panel study also demonstrated that idiolectal change is strongest before age 50, but there were also informants older than 50 who had changed, either in the direction of more standard or more local speech.

Now I have returned to Eskilstuna once again to record as many panel speakers as possible. I found six of the 13 panel speakers from 1996 (born 1928–1950), and recorded at least 50 minutes with each informant. It is unusual to have this kind of material with recordings from three points of time and 50 years between the first and the third recordings. I know much about the six informants after recording them in 1996 and 2018, listening to the recordings from 1967, and in addition, asking them different questions for the integration index in 1996. In this paper, I will describe and discuss how the individual variation, change and stability is influenced by a combination of different factors.

16:15-17:30 Session 15D: Crowd Sourcing
Location: Moscow
Sociopragmatic factors in gender assignment in reference to female persons in Luxembourgish


While a correlation between gender and sex usually applies to appellatives and anthroponyms, one can use feminine as well as neuter forms when referring to female persons in Luxembourgish. Thus, the „natural gender principle“ is violated. Feminine and neuter gender are assigned following a complex, partially variable system. In addition, due to the still ongoing standardization processes, Luxembourgish is generally characterized by variation on many different linguistic levels.

While the fundamental parameters for feminine and neuter gender assignment are well-known, there are many variable contexts whose assignment parameters are more difficult to identify. In fact, two categories of reference contexts can be distinguished: The first category includes mostly invariable contexts such as female first names – morphologically treated as neuter in Luxembourgish –  and, accordingly, taking neuter targets, or controllers such as last names and/or titles (e.g. Madame ,Ms.’) triggering feminine gender on their targets. In contrast, the second category includes controllers, which can trigger either feminine or neuter. More precisely, such gender assignment conflicts mainly arise when referring to female persons with more complex names (e.g. first name + last name). Moreover, a few special cases (e.g. Schwëster ‘sister’) belong to this second category. First investigations on this specific topic (Döhmer 2016, Nübling 2015, Nübling et al. 2013) have suggested that gender assignment primarily depends on sociopragmatic factors (such as age, respect, etc.) in cases of variation. Therefore, the aim of this study is to further analyze the exact sociopragmatic factors playing a role in gender assignment considering the different reference contexts.

Thus, the study analyzes elicited data from an online questionnaire (2700 participants) as well as audio recordings collected by means of the Luxembourgish survey application Schnëssen. On the one hand, the results confirm the decisive role of factors such as age of the referent and familiarity (vs. distance): Neuter gender assignment is more likely when referring to young persons; older referents rather trigger feminine gender. On the other hand, the results show that the speaker’s age influences gender assignment: Younger speakers assign neuter gender more often than older speakers. Additionally, some results indicate that female speakers tend to assign neuter gender more often than male speakers. The role of this sociopragmatic factor (sex of the speaker) is to be investigated in more detail. Finally, some results reveal a language change in Luxembourgish. A very striking example is the pronominalization of the feminine appellative Schwëster ‘sister’, which underwent a substantial language change from feminine to neuter gender assignment.



Döhmer, C. (2016): Formenbestand und strukturelle Asymmetrien der Personalpronomen im Luxemburgischen. ZDL Beihefte: 13-38.

Nübling, D. (2015): Between feminine and neuter, between semantic and pragmatic gender assignment: Hybrid names in German dialects and in Luxembourgish. In: Fleischer, J. et al. (eds.): Agreement from a Diachronic Perspective: 235-265. Berlin/Boston: de Gruyter Mouton.

Nübling, D. et al. (2013): Dat Anna und s Eva – Neutrale Frauenrufnamen in deutschen Dialekten und im Luxemburgischen zwischen pragmatischer und semantischer Genuszuweisung. ZDL 80 (2): 152-196.

Regional Distribution of Contemporary Lithuanian Dialects


At the beginning of the 20thcentury, like in many other European countries, a plan to document the spoken language systematically was also brought up in Lithuania. In the 1950’s, the necessary data for the Lietuvių kalbos atlasas (LKA) 'Lithuanian Language Atlas' was collected. This linguistic atlas was published in the three volumes Lexika 'lexicon' 1977, Fonetika 'phonetics' 1982 and Morfologija 'morphology' 1991, containing dialectal data for 800 localities visualised in 376 dialect maps (cf. Mikulėninėet al.  2014). As the aforementioned volumes were published, dialectologists very often viewed language variation mono-dimensionally. But since the attempts to create a linguistic atlas of Lithuanian, dialectology has undergone a paradigm shift, and today, dialectology has become a dynamic, sociolinguistically and perceptually oriented field of research.

Contemporary projects are methodologically based on a new paradigm where crowd-sourcing methods are used to help collect large amounts of data in a short time. These new data corpora are then compared with the data corpora collected during the traditional dialectological period, which makes language variation and change relatively easy to identify and document. Given the existence of the LKA, the Lithuanian situation is also appropriate for comparison.

In this talk, we want to focus on the results of a pilot study conducted in Lithuania in 2018, for which a questionnaire was drafted in order to document language variation and investigate change. The questionnaire aimed to survey eleven variables from the LKA to allow for a first comparison of data.  The questionnaire was accessible online for a total of one day during which we collected 119 completed questionnaires. The results make clear that the two major dialect regions (Aukštaitija 'Upper Lithuania' and Žemaitija 'Lower Lithuania'are still visible; however, their original area of distribution, as shown in the dialect classification that emerged out of the data of the LKA, has changed considerably. Furthermore, the results show that the linguistic homogeneity, as documented for the majority of all variables in the LKA, has been lost in most dialect regions, basically through speaker mobility. In this talk, we also want to focus on the recently discovered heterogeneity and present new dialect maps, which give us a clearer picture of the linguistic situation in Lithuania today.



Mikulėnienė, Danguolė et al.(2014). XXI a. pradžios lietuių tarmės: geolingvistinis it sociolingvistinis tyrimas. Vilnius: Briedis.

16:15-17:30 Session 15E: Language Contact
Location: New York 3
Language Variation and Language Contact – Analyses on PUT verbs in Austria


The presentation will discuss a variation phenomenon that might have been supported by language contact, namely the semantic development of German geben ‘give’ towards a posture verb, namely towards a PUT and even TAKE verb. As such, German geben ‘give’ can be used in constructions that refer to object movement with an inanimate target (cf. (1)) or an inanimate source (cf. (2)). While tun ‘do’ is prevalent as a PUT/TAKE verb in most of the dialects and regiolects of German, the areal distribution, the varietal status of the posture verb geben ‘give’ as well as its historical ‘source’ is contested. At least in the VWB-NEW, geben is considered to be one of the posture verbs even in the written standard of Austria (A) (cf. (1.a)). The presentation will empirically investigate the (linguistic and sociolinguistic) competition of the two posture verbs in the Austrian language area. The empirical basis of the discussion are ‘old’ dialect data on the one hand, and experimental data on present-day German on the other hand (Bowerman et al. 2004). As will be shown, the success of geben as a new PUT verb in Austrian German is mainly due to German-Slavic language contact.

(1) Target-oriented object movement

a. “[…] eine Parkscheibe hinter die Windschutzscheibe geben […]” (VWB-NEW, geben)

Literal: […] to give a parking disc behind the windshield

‘to put a parking disc behind the windshield’

b. “Der eine Teppich ist nicht so groß, den könnte ich in den Backofen tun.“ (http://www.parents.at/forum/showthread.php?t=675914)

Literal: The one carpet is not that big, I could do it into the oven

‘One of the carpets is not very big, I could put it in the oven


(2) Source-oriented object movement

a. “Teig aus der Schüssel geben“ (http://tortenzwerg.at/7-days-of-love-kekse-mit-glasinneren/)

Literal: give the dough out of the bowl

take the dough out of the bowl

b. “sie soll die 'Pratzen' aus der Tasche tun.“ (http://sbgv1.orf.at/stories/505948)

Literal: she should do the ‘claws’ out of the bag

‘she should take her claws out of the bag.



Bowerman, Melissa, Marianne Gullberg, Asifa Majid & Bhuvana Narasimhan. 2004. Put project: the cross-linguistic encoding of placement events. In Asifa Majid (ed.), Field Manual Volume 9, 10-24. Nijmegen: Max Planck Institute for Psycholinguistics.

VWB-NEW = Variantenwörterbuch des Deutschen. 2016 (‘German Dictionary of Lexical Variants. The standard language in Austria, Switzerland, Germany, Liechtenstein, Luxembourg, Eastbelgium, South Tyrol and Rumania, Namibia and Mennonites‘ settlements.) Ed. by Ulrich Ammon, Hans Bickel and Alexandra N. Lenz. Berlin/Boston: Walter de Gruyter.

Do the contact languages influence the distribution of prepositions in Estonian dialects?


In Finno-Ugric languages, an adpositional phrase is generally built up of a complement noun followed by an adposition – a postposition. However, in Estonian, alongside with other westernmost Finno-Ugric languages, an adpositional phrase has two possible structures: the adposition can either precede or follow the complement noun, i.e. both prepositions and postpositions are used. (Tauli 1966; Palmeos 1982; Ehala 1995; Grünthal 2003; Janda et al. 2014). The origin of prepositional phrase structure in Finnic languages shows, however, little evidence of being contact-induced. However, B. Heine and T. Kuteva (2005:4–50) have shown that grammatical constructions may acquire a higher frequency of use as a result of equivalent structures used in a contact language. The effect of contact languages on the Estonian adpositional system is thus likely, as the linguistic landscape of Estonia has been remarkably diverse for thousands of years. The main local contact languages of Estonian-speaking areas are from three branches of Indo-European languages – Baltic, Germanic and Slavic, and from the western branch of the Finno-Ugric languages – Votic, Finnish, Ingrian, and Livonian.  

The current paper provides new findings on the topic of contact-induced change by comparing the distribution of prepositions in Estonian dialects with the respective contact languages. The purpose is to determine whether the usage frequency of prepositions is higher in areas mainly in contact with prepositional Indo-European languages. The topic is approached from a corpus-based, frequency-driven viewpoint. I use frequency data obtained from the Corpus of Estonian Dialects (http://www.murre.ut.ee/murdekorpus). The geographical distribution of the prepositions is observed and, additionally, statistical tests are applied.

The results show that the usage frequencies of prepositions differ throughout the area: there occurs a gradual decrease in the density of prepositions from the northeast to the southwest of Estonia. As the strongest influences with prepositional languages are by the southern border and the northwest coast of Estonia, the final results do not support the prediction that contacts influence the frequencies of prepositions. It may thus be concluded that, at least in the case of one Finnic language, the results give support for the general view that the adpositional system is not easily influenced by contact languages.



Ehala, M. (1995): Explaining the bipositional head-complement order in adpositional systems. In Palek, B., F. Dane, O. Fuijmura & J. V. Neustupný (eds), Proceedings of LP’94: Item Order in Natural Laguages, Prague, August 16–18, 1994, Proceedings of the Conference: 325–339. Prague: Charles University Press.

Grünthal, R. (2003) Finnic Adpositions and Cases in Change. Helsinki: Finno-Ugrian Society.

Janda, L. A., L. Antonsen & B. A. B. Baal. (2014): A radial category profiling analysis of North Sámi ambipositions. High Desert Linguistics Society Proceedings 10: 91–102.

H., Bernd & T. Kuteva. (2005) Language Contact and Grammatical Change. Cambridge: Cambridge University Press.

Palmeos, P (1982) Eesti keele grammatika: II osa, neljas vihik. Tartu: Tartu Riiklik Ülikool.

Tauli, V. (1966) The Structural Tendencies in Uralic Languages. The Hague: Mouton.