John Nerbonne

University of Groningen, Alfa-informatica, Faculty Member

Followers

381

Following

Co-authors

Public Views

Interests

Uploads

Papers by John Nerbonne

Poster: Catch project Scratch: Script Analysis Tools for Cultural Heritage. Statistics on Queries and Line Matching

A Central Asian Language Survey

Language Dynamics and Change, 2016

We have documented language varieties (either Turkic or Indo-European) spoken in 23 test sites by... more We have documented language varieties (either Turkic or Indo-European) spoken in 23 test sites by 88 informants belonging to the major ethnic groups of Kyrgyzstan, Tajikistan and Uzbekistan (Karakalpaks, Kazakhs, Kyrgyz, Tajiks, Uzbeks, Yaghnobis). The recorded linguistic material concerns 176 words of the extended Swadesh list and will be made publically available with the publication of this paper. Phonological diversity is measured by the Levenshtein distance and displayed as a consensus bootstrap tree and as multidimensional scaling plots. Linguistic contact is measured as the number of borrowings, from one linguistic family into the other, according to a precision/recall analysis further validated by expert judgment. Concerning Turkic languages, the results of our sample do not support Kazakh and Karakalpak as distinct languages and indicate the existence of several separate Karakalpak varieties. Kyrgyz and Uzbek, on the other hand, appear quite homogeneous. Among the Indo-Iran...

Download

Determinants of Dialectal Variation

Linguistic Distances Workshop at the joint conference of International Committee on Computational Linguistics and the Association for Computational Linguistics

The Association for Computational Linguistics, 2006

Data Analysis, Machine Learning and Applications. Proceedings of the 31st Annual Conference ofthe Gesellschaft für Klassifikation e.V., Albert-Ludwigs Universität Freiburg, March 7-9, 2007

The use of registered names, trademarks, etc. in this publication does not imply, even in the abs... more

Download

Toward a dialectological yardstick*

Journal of Quantitative Linguistics, Aug 1, 2007

Dialectometry measures the differences between dialects in ways which may involve many independen... more Dialectometry measures the differences between dialects in ways which may involve many independently varying parameters which must be specified in combination in order to arrive at measures of difference. The existence of many parameters of measurement and possible interaction introduces the problem of how to choose parameter values and combinations of them intelligently. This paper proceeds from the assumption that dialectology proper must reveal geographic coherence in language variation in order to propose a yardstick with which to compare measurements made using various parameter settings, and it presents some results of its application.

Download

The handbook of dialectology

Wiley-Blackwell eBooks, 2018

Application of phylogenetic methods on dialect pronunciation data

Measuring Dialect Differences

We measure varietal dierences in general, and dierences with respect to standard languages in par... more We measure varietal dierences in general, and dierences with respect to standard languages in particular (dialectality, in Herrgen/Schmidt's sense) in order to systematize observations about dialect dierences, to make sense of exceptions, and to enable measurements based on randomly selected material, thus obviating issues of potential bias. Finally, measurements allow the characterization of abstract relations among language varieties. We illustrate some issues with simple techniques for categorical data introduced by Séguy and rened by Goebl, viz., issues concerning frequency, irrelevant variation, and competing forms. We proceed to measuring pronunciation dierences, focusing on dierences in the pronunciation of the same words in dierent varieties. Caution is needed to isolate pronunciation dierences from dierences in inectional morphology, sandhi, and intonation. We characterize the dierence between sound segments and develop a measure of the dierence between the sequences of those segments in words, including insertions, deletions, and swaps (epenthesis, elision and metathesis). Automating measurement techniques exposes the issue of validation, which lay largely unexamined in earlier dialectology. We propose to validate measurements based on the degree to which they correlate with dialect speakers' judgments of dierence, justied by the presumed function of linguistic variation, that of signaling provenance.

Download

Low Saxon Dialects across Borders — Niedersächsische Dialekte über Grenzen hinweg . Edited by Alexandra N. Lenz, Charlotte Gooskens & Siemon Reker

Diachronica, 2012

Download

Validating and using the PMI-based Levenshtein distance as a measure of foreign accent strength

In this study, we use and validate an adaptation of the Levenshtein distance (Levenshtein, 1965) ... more In this study, we use and validate an adaptation of the Levenshtein distance (Levenshtein, 1965) to determine pronunciation distances between accented speech. The Levenshtein distance measures the minimum number of insertions, deletions and substitutions to transform one (phonetic) string into the other. Our adaptation (Wieling et al., 2009) uses automatically obtained, sensitive linguistically sound (Wieling et al., 2012) segment distances (based on how frequent sound segments correspond in substitutions). We apply this adapted (i.e. PMI-based) Levenshtein distance algorithm to data from the Speech Accent Archive (Weinberger & Kunath, 2011). The Speech Accent archive contains more than 1000 transcribed speech samples in English from people with various language backgrounds and is digitally available at http://accent.gmu.edu. Each speaker reads the same paragraph of 69 words in English. By calculating the average PMI-based Levenshtein distance between all native American English speakers in the dataset (115 in total) and each individual speaker, we obtain a computational measure of pronunciation distance between a speaker's pronunciation and 'average' American English speech. To assess how well our automatic method matched human judgments of how native-like a certain accent sounds, we conducted a perception experiment. On the basis of the native-likeness judgments of more than 1100 native American English participants we determined the average native-likeness for 286 samples of native and non-native English speech. As the correlation between the average native-likeness judgments and the (log-transformed) PMIbased Levenshtein distances was r =-0.8, we conclude that our automatic procedure is a valid method to measure foreignness of transcribed speech (Wieling et al., forthcoming). Given the wealth of data, it is interesting to investigate the accent differences. For example, Figure 1 shows the visualization of the pairwise pronunciation distances between some of the world's English accents.

Download

Dialectvariatie, dialectafstanden, en dialectindelingen: Een formele analyse van uitspraakverschillen

Download

A corpus of regional Dutch speech

P.I.E. Peter Lang eBooks, 2013

We present in this paper a phonetically transcribed corpus of regional Dutch speech from the Neth... more We present in this paper a phonetically transcribed corpus of regional Dutch speech from the Netherlands and Belgian Flanders and some example analyses using it. The corpus consists not only of 200 common words, but also of 200 nonsense words as these were pronounced by radio announcers from regional radio stations. The announcers regularly speak on radio programs with an explicitly regional remit, and they agreed to our recording them using regional speech but also using the standard language (standard Netherlandic Dutch in the Netherlands and standard Belgian Dutch in Flanders), which they also spoke professionally. The corpus is publicly available for analyses concerning, e.g., the relative differences between Dutch and Belgian regional speech, the relative similarity of regional speech to standard speech in the two countries, and the probity of models such as Auer's & Hinskens' (1996) "cone model", within which regional speech should be found. Nerbonne, Van Ommen, Wieling & Gooskens (to appear) have indeed used the corpus material to test whether regional speech adheres to the predictions of the cone model.

Download

Language and Space: Theories and Methods

We measure varietal di erences in general, and di erences with respect to standard languages in p... more We measure varietal di erences in general, and di erences with respect to standard languages in particular ( dialectality , in Herrgen/Schmidt&#39;s sense) in order to systematize observations about dialect di erences, to make sense of exceptions, and to enable measurements based on randomly selected material, thus obviating issues of potential bias. Finally, measurements allow the characterization of abstract relations among language varieties. We illustrate some issues with simple techniques for categorical data introduced by Seguy and re ned by Goebl, viz., issues concerning frequency, irrelevant variation, and competing forms. We proceed to measuring pronunciation di erences, focusing on di erences in the pronunciation of the same words in di erent varieties. Caution is needed to isolate pronunciation di erences from di erences in in ectional morphology, sandhi, and intonation. We characterize the di erence between sound segments and develop a measure of the di erence between the sequences of those segments in words, including insertions, deletions, and swaps (epenthesis, elision and metathesis). Automating measurement techniques exposes the issue of validation, which lay largely unexamined in earlier dialectology. We propose to validate measurements based on the degree to which they correlate with dialect speakers&#39; judgments of di erence, justi ed by the presumed function of linguistic variation, that of signaling provenance.

Geographic distributions of linguistic variationreflect dynamics of differentiation

De Gruyter eBooks, Dec 19, 2007

Download

Learning computational grammars

This paper reports on the LEARNING COMPUTATIONAL GRAMMARS (LCG) project, a postdoc network devote... more This paper reports on the LEARNING COMPUTATIONAL GRAMMARS (LCG) project, a postdoc network devoted to studying the application of machine learning techniques to grammars suitable for computational use. We were interested in a more systematic survey to understand the relevance of many factors to the success of learning, esp. the availability of annotated data, the kind of dependencies in the data, and the availability of knowledge bases (grammars). We focused on syntax, esp. noun phrase (NP) syntax.

Download

Learning computational grammars using inductive logic programming (ILP)

Computing and Language Variation: International Journal of History and Computing - Volume 2

Edinburgh University Press eBooks, Aug 15, 2009

Abstract Computing and Language Variation explores dialects and social differences in language co... more Abstract Computing and Language Variation explores dialects and social differences in language computationally, examining topics such as how (and how much) linguistic differences impede intelligibility, how national borders accelerate and direct change, how opinion and hearsay shape perceptions of language differences, the role of intonation (melody), the differences between variation in pronunciation and vocabulary, and techniques for recognizing structure in larger collections of linguistic data. The ...

Computational Phonology: International Workshop held with RANLP-2007

To what extent are surnames words? Comparing the geographic patterns of surname and dialect variation in the Netherlands

HAL (Le Centre pour la Communication Scientifique Directe), 2006

John Nerbonne

Uploads

Papers by John Nerbonne

Log In