We study regional similarities and differences in language use on an anonymous mobile chat application in the German-speaking area. We use a neural network on 2.3 million online conversations to automatically learn representations of words and cities. These linguistic-use-based representations capture regional distinctions in a high-dimensional vector space that can be clustered and visualized to discover patterns in the data. We find that the resulting regional patterns are closely linked to the traditional division of German dialects, even though most of the conversations are written in standard German. The resulting maps correspond to traditional dialect divisions and language-external spatial structures, with a few notable exceptions that can be explained through external factors.

Our method also facilitates two qualitative analyses, allowing us to discover geographically-pertinent words for various regional levels, as well as creating regional group-specific style profiles based on various linguistic resources. The results of our study strongly suggest the existence of region-specific patterns of language use (“digital regiolects”) representing distinctive strategies of linguistic stylization in relation to linguistic resources and topics. As a methodological contribution, we show how linguistic theory can drive the application and direction of neural network-based representation learning, and how their judicious application provides the basis for qualitative analysis of large-scale data collections.

This article is an analysis of linguistic survey data representing German dialects in Switzerland in 1933/34 based on the so-called Wenker sentences. The data are impressionistic in terms of applied phonetic transcriptions, which were produced by non-specialists using the Latin alphabet. Due to the lack of pre-defined standardization, the phonetic transcriptions are very heterogeneous. From a technical perspective, this leads to very noisy data, which is why the validity of the Wenker data in general and the Swiss Wenker data in particular has been questioned. Using methods from computational linguistics, we compare, for the first time, Wenker data with linguistic data collected at virtually the same time by linguistics professionals. Direct comparison with a sample from the published atlas of German-speaking Switzerland (SDS) reveals that despite the noisiness of the data, they nevertheless provide reliable information, e.g., in terms of the spatial structuring of Swiss dialects. The study is thus a successful pilot for other corpus-based studies dealing with unstructured Wenker data in other regions.

This article extends previous analyses on (ay) and (aw) raising in Yami, an endangered Austronesian language spoken on Orchid Island, Taiwan. Two patterns of (ay) and (aw) variation co-exist in Yami—the more conservative/unraised variants and the innovative/raised forms—each of which indexes an authentic regional identity. Results from recent data suggest that Yami has undergone significant changes in just one generation as the raised variants are now preferred by younger islanders and have started permeating through non-raising areas. Traumatic shifts in the ecological context on Orchid Island help explain this sound change. Economically, rapid shift from fishing/farming to tourism trades promotes daily face-to-face communications among younger islanders, which has catalyzed the spread of nucleus raising to non-raising areas. Sociopsychologically, ongoing intergroup conflicts between islanders and mainland Taiwanese may further trigger the sound change, altering nucleus raising into a new communal system which highlights Yami identity in opposition to Others.

Finding the boundaries of linguistic variants and studying transitions between variants are key interests in classical linguistic geography. However, the definition of boundaries in areal linguistics is vague, and a quantitative characterization of transitions at the interface between dialectal variants is missing. We conceptualize these transitions as gradients, aiming to quantitatively account for the transition patterns which are traditionally only implicitly inferred from visualizations. Fitting of logistic functions in different spatial scopes (profiles as well as surfaces) is proposed as an approach to model the transition at the interface between the dominant usage areas of dialectal variants. Logistic functions can accommodate the breadth of boundary concepts, ranging from sharp isoglosses to transitions with different gradualities. The parameters of the fitted logistic models as well as supplementary measures then allow for the quantitative characterization and comparison of transitions across variables. To demonstrate the proposed methodology, we use Swiss German syntactic data on dialectal variables with a single transition zone.

We recorded older male speakers and younger female speakers of 86 local dialects of Dutch. Using these data, we analyze and visualize the influence of standard Dutch on apparent time changes in these dialects. Focusing for the most part on variation in the sound components, we test whether (I) dialect change is mainly the result of convergence to standard Dutch, (II) sound changes in two dialects which make them converge to standard Dutch also make them more similar, and (III) sound changes in two dialects which make them diverge from standard Dutch also make them less similar. We used three-dimensional (first hypothesis) and five-dimensional Levenshtein distance implementations (second and third hypothesis). These implementations are a novel step in dialectometry and in the study of ongoing processes of language change and their consequences for the dialect landscape. The findings corroborate all three hypotheses.

The use of both production and perceptual data has the potential to provide a more complete picture of linguistic phenomena than would otherwise be the case, including when exploring regional linguistic variation. Utilizing the social networking platform Twitter and an online survey, this paper reports on a descriptive analysis of the geographic distribution of a less-commonly used syntactic form of the Spanish verb gustar ‘to like, to please’, referred to as experiential gustar (e.g., cuando gustes ‘when you’d like’). The results from the analysis of 6,686 tweets together with the responses of 81 native Spanish-speaking participants in an online survey suggest that experiential gustar is produced and is perceived to be produced most often in Mexican Spanish, despite not being exclusive to that country. The paper contributes to the literature depicting the benefit of using both production and perceptual data in the study of dialectal variation, as well as to the literature documenting language variation in Spanish.

Researchers in dialectometry have begun to explore measurements based on fundamentally quantitative metrics, often sourced from dialect corpora, as an alternative to the traditional signals derived from dialect atlases. This change of data type amplifies an existing issue in the classical paradigm, namely that locations may vary in coverage and that this affects the distance measurements: pairs involving a location with lower coverage suffer from greater noise and therefore imprecision. We propose a method for increasing robustness using generalized additive modeling, a statistical technique that allows leveraging the spatial arrangement of the data. The technique is applied to data from the British English dialect corpus FRED; the results are evaluated regarding their interpretability and according to several quantitative metrics. We conclude that data availability is an influential covariate in corpus-based dialectometry and beyond, and recommend that researchers be aware of this issue and of methods to alleviate it.

This article explores two fundamental dimensions in sociolinguistics: the dynamics of linguistic variation and change in international languages and the exploitation of data proceeding from significant countries. These issues will be addressed through examination of a particular syntactic feature and a possible change in progress: the occurrence of null direct objects in Spanish. It is shown that for Spanish, a widely used international language, social factors have not been decisive in explaining the distribution of the phenomenon under investigation. This study shows that while direct object omission is not conditioned by typical social variables such as sex, age, and gender, it is unevenly spread throughout the Spanish-speaking world: Mexico and the continental Caribbean use it more than other countries, such as Spain or Chile. Besides the relevance of geography, some semantic, discourse, and contextual factors are shown as determinant for the direct object omission. Finally, this paper reflects on methodology, specifically the use of a macroregional sociolinguistic method for data analysis as well as the advantages and shortcomings of a specific data collection technique that capitalizes on technological tools with global reach: the internet survey in an international scenario.

This paper explores the cultural and historical forces that created variation in terms for “cemetery”, including links between language and material culture, using cemetery terms found within two Linguistic Atlas data sets to demonstrate how colonial influence, cultural changes, and physical locations contribute to linguistic variation. Speakers’ lexical choices in the 1930s still show the effects of the religious and social climates of the seventeenth and eighteenth centuries. Northern and southern colonial trends were still influencing regional language use several hundred years later. Furthermore, for the LANE data we find that the physical location of historic cemeteries has an effect on speakers’ use of specific lexical items.

This paper takes a quantitative perspective on data from the project Syntax hessischer Dialekte (SyHD), covering dialects in the German state of Hesse, an area with rich dialectal variation. Many previous dialectometric analyses abstracted away from intralocal variation (e.g., by only counting the most frequent variant at a location). In contrast, we do justice to intralocal variation by taking into account local frequency relations. The study shows that the border between Low German and Central German—one of the most important isoglosses in German dialectology—is not relevant for syntactic phenomena. At the same time, a comparison with character n-grams (a global measure of string similarity) reveals that the traditionally assumed dialect areas, primarily defined according to phonological developments, are still present in the twenty-first century data. Different from previous studies, our results are obtained from a uniform data base. Therefore, the differences between syntax and phonology cannot be due to variation in sampling, elicitation method, or time of elicitation.

In the early 2000s, the SADS, an extensive linguistic atlas project, surveyed more than three thousand individuals across German-speaking Switzerland on over two hundred linguistic variants, capturing the morphosyntactic variation in Swiss German. In this paper, we applied TESS, a Bayesian clustering method from evolutionary biology to the SADS to infer population structure, building on parallels between biology and linguistics that have recently been illustrated theoretically and explored experimentally. We tested three clustering models with different spatial assumptions: a nonspatial model, a spatial trend model with a spatial gradient, and a spatial full-trend model with both a spatial gradient and spatial-autocorrelation. Results reveal five distinct morphosyntactic populations, four of which correspond to traditional Swiss German dialect regions and one of which corresponds to a base population. Moreover, the spatial trend model outperforms the nonspatial model, suggesting a gradual transition of morphosyntax and supporting the idea of a Swiss German dialect continuum.

This contribution focuses on the use of geben ‘give’ as a put verb in Upper German dialects in Austria from a historical and a recent perspective. On the basis of comprehensive historical and contemporary data from German varieties and Slavic languages our analyses provide evidence for the central hypothesis that this phenomenon traces back to language contact with Czech as already suggested by various scholars in the 19th century. This assumption is also supported by the fact that Czech dát ‘give’ in put function has been accounted for since the Old Czech period as well as by its high frequency in both formal and informal Czech written texts. Moreover, our data analyses show that geben ‘give’ as a put verb has been and is still areally distributed along and spreading from the contact area of Czech and Upper German varieties.

The use of loanwords is generally attributed to a social feature, like social prestige, and to semantic features, like the need to fill a lexical gap. However, few studies take into account variation in the use of loanwords within a speech community, and directly compare the frequency of loanwords from more than one source language. This paper contributes to research on lexical borrowing by comparing the distribution of loanwords from three different source languages in two large databases of dialect data. We take an onomasiological perspective, which allows us to gauge the frequency of borrowed lexical items vis-à-vis alternative expressions. Using Generalized Additive Mixed Modeling, we show that the usage of loanwords can only be explained by taking into account the interaction between semantics and geographical diffusion. Our analysis confirms that the patterns that occur almost exclusively reflect changes in socio-cultural history.

This project uses GIS mapping to analyze spatial trends in spoken language, testing how features identified as part of the “Southern dialect” by the Atlas of North American English (ANAE; Labov et al., 2006) are used in the Digital Archive of Southern Speech (DASS; Kretzschmar et al., 2013). We analyze vowel mergers, diphthongization, monophthongization, fronting, and several consonantal features. Rather than drawing isoglosses, we use local spatial autocorrelation analysis to reveal subregional patterning in the data. We present a series of maps illustrating the realization of Southern speech features as enumerated by ANAE. We find little evidence for ANAE’s Inland South region based on acoustics, and while some areas surveyed in DASS align well with the portrayal of Southern speech presented by ANAE, others do not.

This paper revisits the question of national varieties in German, a concept that still has not found solid empirical proof, at least not from a quantitative-etic perspective. The question raised is how data collected in the course of the international research project, Variantengrammatik des Standarddeutschen (Handbook of Grammatical Variation in Standard German), and the application of spatial statistical methods to geographical variant distribution patterns can contribute to this discussion. With the aid of Factor Analysis and spatially modified clustering of variant distribution patterns it is shown that—at least from a quantitative/corpus linguistic point of view—there is no evidence for assuming nation-specific varieties. Importantly, this result does not imply the existence of a homogenous German standard language; rather, it is argued that geographical variation does not self-evidently legitimize the assumption of standard varieties.

This article presents the results of an interdisciplinary project that explores street name changes in Leipzig, a city in eastern Germany, over the past one-hundred years. Our analysis focuses on the ways in which semantic choices in the streetscape are recruited to canonize traces of the national past that are “supportive of the hegemonic socio-political order” (Azaryahu, 1997:480). We triangulate results from variationist sociolinguistics, Linguistic Landscape (LL) studies and geographical analysis to visualize waves of street (re)naming during a century of political turmoil. Drawing on historical archival data allows us to interpret spatial and temporal patterns of odonymic choices as the public embodiment of subsequent political state ideologies. The analysis provides quantitative and longitudinal support to Scollon and Scollon’s (2003) claim that the indexing of officially sanctioned identity and ideology as well as the appropriation of human space are performed by and in turn index state-hegemonic politics of memory.

This paper studies the perception of regional variation in Russian among young Russian nonlinguists in Moscow, Perm, and Novosibirsk. I explore the labels used in 55 perceptual maps and categorize them in order to investigate the perceived character of regional variation among young Russians, including their explanations for regional variation. The data analysis shows that claims about regional variation are based on the informants’ assessments of variation in Russian, but also on assessments of domains that they perceive as related to regional variation: style and accent, as well as extralinguistic features such as geography and climate. Based on this analysis, I argue that the line between regional language variation and other variations can be conceived of as fuzzy.

This perceptual-dialectology study investigates how listener-related social factors impact the geographical recognition of regional accents. In contrast to much prior research on English, our focus is on Dutch, which lends itself well to our study, allowing for notable regional accents within a well-defined standard language. Using a map-based recognition task in which 1,578 listeners placed forty representative speakers on a map based on fragments of their speech, we investigated the regional biases in accent recognition and the extent to which each listener’s awareness of these depends on their familiarity and proximity. Education, geographical knowledge, and distance to listeners’ own regions significantly predicted their accent-recognition accuracy. Moreover, we found a curvilinear age effect, which we interpret in terms of age-related changes in geographical and social mobility. We show how these effects in our design lead to meaningful accent-recognition patterns in groups of listeners.

By means of the first comprehensive apparent-time study of Austria’s traditional dialects, this paper explores the use of adnominal syntactic constructions of expressing the semantic relation of possession. The article focuses on both the geographical variation and the interplay of syntax, semantics, and pragmatics. The analyses are based on data from direct recordings of 162 speakers from forty villages and on written questionnaire data from 103 of these speakers from thirty-seven villages. The analyses reveal clear geographical patterns for those constructions in which the possessor phrase precedes the possessum phrase within the entire construction. We propose to focus on the discursive-pragmatic properties of the possessor phrase to explain the fact that each of the observed dialects allows the possessor to precede the possessum. We provide evidence that referential anchoring, combined with the concept of accessibility, is the key to explaining the syntactic order within the used constructions.