Assistant Professor (with tenure)
University of Groningen, Department of Information Science
Martijn Wieling is an Assistant Professor at the University of Groningen, and Affiliated Research Scientist at Haskins Laboratories. His research focuses on investigating language variation and change quantitatively. He uses both large digital corpora of text and speech, as well as experimental approaches to assess differences in the movement of the tongue and lips during speech. Since April 2018, Martijn Wieling is the vice-chairman of De Jonge Akademie.
University of Groningen, Department of Information Science
University of Groningen, Department of Information Science
University of Tübingen, Department of Quantitative Linguistics
Ph.D. in Linguistics (cum laude)
University of Groningen, Faculty of Arts
Master (research) of Science in Behavioural and Cognitive Neurosciences (cum laude)
University of Groningen, Faculty of Science and Engineering
Master of Science in Computing Science
University of Groningen, Faculty of Science and Engineering
Bachelor of Science in Computing Science
University of Groningen, Faculty of Science and Engineering
In 2016, Wieling was selected as one of the 18 founding members of the Young Academy of Groningen for a period of five years. The Young Academy Groningen is a club for the University of Groningen’s most talented, enthusiastic and ambitious young researchers. Members come from all fields and disciplines and have a passion for science and an interest in matters concerning science policy, science and society, leadership and career development.
Wieling was selected as one of the youngest members of De Jonge Akademie (DJA) of the Royal Netherlands Academy of Arts and Sciences (KNAW) in April 2015 for a period of five years. In 2018, Wieling was elected as vice-chairman of De Jonge Akademie for a period of two years.The Young Academy is a dynamic and innovative group of 50 top young scientists and scholars with outspoken views about science and scholarship and the related policy. The Young Academy organises inspiring activities for various target groups focusing on interdisciplinarity, science policy, and the interface between science and society.
This four-year research grant was awarded to Wieling by the Netherlands Organisation for Scientific Research (NWO) for his project "Improving speech learning models and English pronunciation with articulography". Only 15.5% of the submitted project proposals were granted.
This one-year research grant was awarded to Wieling by the Netherlands Organisation for Scientific Research (NWO) for his project "Investigating language variation physically". Only 12% of the submitted project proposals were granted.
My research generally focuses on quantitatively investigating patterns in language variation and change. While I mostly investigate dialect variation, I also study the pronunciation of second language learners, congenitally blind speakers, or speakers with dysarthria (such as Parkinson's disease). Besides investigating pronunciation, I am als interested in identifying patterns in large, digitally available corpora of text. For example, in collaboration with the law department, we are currently investigating how we may predict court judgments on the basis of various legal and linguistic characteristics of the court transcripts. This project also emphasizes the interdisciplinary and collaborative nature of my research (see also projects, below).
For investigating patterns in speech, I generally take two approaches. The first approach is to analyze phonetically transcribed data using new quantitative, dialectometric techniques (see publications in Language, Journal of Phonetics, Annual Review of Linguistics, Language Dynamics and Change and PLOS ONE). The second approach is to track the movement of speakers' tongue and lips using (e.g.,) electromagnetic articulography (see publications in Journal of Phonetics, Journal of the Acoustical Society of America).
In terms of techniques, I am frequently using (and teaching courses about) generalized additive modeling, a flexible non-linear regression technique which can be used to model the influence of geography on dialect variation, but also to model time series data (such as collected using articulography, eye-tracking or EEG experiments). See this extensive tutorial (accepted in Journal of Phonetics) tutorial, but also publications in Language, Journal of Phonetics and PLOS ONE.
In this project we aim to automatically detect Dutch accents on the basis of data from the Sprekend Nederland data. Particularly, we are interested in identifying the acoustic and segmental characteristics of the different accents. This project is funded by the Centre for Digital Humanities of the University of Groningen.
In this project, we will study the effects of one language on another in our voices, by considering how recognizable Frisian phonological traits are in speakers' production of their first language, Frisian, as well as in their second language, Dutch. This project is funded by the University of Groningen (Data Science Projects 2017).
Law is everywhere: almost every human activity is regulated. Buying a sandwich, renting an apartment or going to a hospital, all these activities involve legal rules and consequences. For a stable and sustainable society it is essential that its laws are predictable. People need to know what a legal rule means and what likely outcome a potential court case will have. Authorities have tried to improve the law's predictability and transparency by publishing court judgments. For decades, summaries of judgments were published in written journals, which were not easily accessible for the public. Nowadays, courts publish their judgments online. For example, approximately 370,000 individual court judgments can be found on the website of the Dutch judiciary (www.rechtspraak.nl). Similarly, another 52,000 court judgments are published online by the European Court of Human Rights (http://hudoc.echr.coe.int). Each of these judgments contains a detailed and rich description of the facts, procedure, reasoning of the parties and outcome of the case. Of course, public availability of case law will help to improve predictability and transparency, but to analyse hundreds of thousands of legal documents, we need other approaches than the traditional and labour-intensive 'doctrinal analysis' (i.e. close reading of a single or a small number of judgments) conducted by legal researchers.
The goal of this PhD project is therefore to combine two distinctly separate disciplines, law and computational linguistics, in developing and evaluating quantitative, computational approaches to improve the predictability and transparency of the law. Techniques from computational linguistics would (for example) enable the automatic extraction, syntactic and semantic analysis of the judgment texts. The extracted features may be used in quantitative analyses identifying common patterns (see Wieling, 2012 for examples, albeit in a different field), which can subsequently be used to predict the outcome of a judgment. Such an approach would clearly be beneficial for the field of law. Surprisingly, a recent study (Vols & Jacobs, 2016) showed that between 2006 and 2016 fewer than 25 publications in Dutch legal journals were published involving statistics to analyse case law. While a quantitative approach to analysing case law is more prevalent in the US, it is primarily focused on specific American legal issues, and frequently contains serious methodological flaws (Epstein & King, 2002; Epstein & Martin, 2014).
In computational linguistics, specific characteristics of legal texts have been studied (see Francesconi, Montemagni et al. (eds.), 2010), but hardly any studies have attempted to use linguistic characteristics to predict judicial decisions. A very recent exception (also illustrating the timeliness of the project idea) by Aletras et al. (2016) reported an accuracy of 79% in predicting the judgments of the European Court of Human Rights. However, they focused only on a small sample (600 judgments) and used simplistic linguistic features (such as word frequency). The goal of this PhD project is therefore to take a more comprehensive, linguistically-oriented approach incorporating all available data, thereby developing a system which is able to detect common patterns in legal big data and use these to predict the outcome of a judgment.
This project is funded by the Young Academy Groningen.
In this IDEALAB-funded project, we investigate speech articulation in Parkinson's disease. Parkinsons disease is a degenerative neurological disorder that is characterised by a decay of motor function. Due to a loss of dopaminergic cells in the substantia nigra, both motor control as well as initiation deteriorate over time, which frequently leads to difficulties in speech producation, a phenomenom that is known as hypokinetic dysarthria. Previous acoustics studies have shown that pitch height and variation, articulation of both vowels and consonants and voice quality are often affected in hypokinetic dysarthria. So far however, the role of kinematics in hypokinetic dysarthria has not been given much attention. Yet, thanks to the recent development of electromagnetic articulography (EMA) it is now possible to track and measure the kinematic movements of the lips, the tongue and the jaw during speech production. This means that a more fine-grained analysis of the articulation difficulties in hypokinetic dysarthria can be performed. In this future study, the velocity, amplitude and coordination of kinematic gestures will be under investigation. Specifically, the temporal overlap of speech gestures will be studied as well as the location and rate of contractions within the vocal tract. In total, 30 Parkinson’s patients with hypokinetic dysarthria will be included. Another 30 participants will serve as healthy controls. In doing so, a detailed view of coordination in hypokinetic dysarthria will be obtained which will lead to a better understanding of the speech disorder. Moreover, it will shed new light on leading theories that have only approached speech from an acoustic viewpoint. Ultimately, the knowledge that will be obtained within this study can improve a early diagnosis of Parkinson’s disease and also dramatically improve speech therapy. In addition, it will also provide more general insight into the kinematics of speech.
In this project we aim to investigate how visualizing tongue movements using ultrasound may be used to help a learner to improve his or her pronunciation in a second language (L2). This project is funded by the Groningen University Fund and De Jonge Akademie.
In this project we investigate if the speech of congenitally blind speakers differs from that of sighted speakers, both from an articulatory as well as an acoustic perspective. Specifically, we investigate if automatic speech recognition performance differs between the two groups. This project was funded by VIVIS. The results of the project were presented at several conferences (e.g., see here: page 88), and are currently in the process of being written down.
In this project we are investigating articulatory differences between Dutch and German speakers' pronunciation of English versus those of native English speakers. Furthermore, we are assessing how visual feedback of the speech articulators may help improve non-native speakers' pronunciation of English. This project was funded by NWO (Veni grant). Results of the project can be found in several publications.
I have the following equipment available for research projects in my lab:
I am always looking for excellent PhD candidates or research assistants. If you are interested in the field of dialect or language variation, speech production research, or computational linguistics, please contact me.
This study focuses on an essential precondition for reproducibility in computational linguistics: the willingness of authors to share relevant source code and data. Ten years after Ted Pedersen's influential ``Last Words'' contribution in Computational Linguistics, we investigate to what extent researchers in computational linguistics are willing and able to share their data and code. We surveyed all 395 full papers presented at the 2011 and 2016 ACL Annual Meetings, and identified if links to data and code were provided. If working links were not provided, authors were requested to provide this information. While data was often available, code was shared less often. When working links to code or data were not provided in the paper, authors provided the code in about one third of cases. For a selection of ten papers, we attempted to reproduce the results using the provided data and code. We were able to reproduce the results approximately for half of the papers. For only a single paper we obtained the exact same results. Our findings show that even though the situation appears to have improved comparing 2016 to 2011, empiricism in computational linguistics still largely remains a matter of faith (Pedersen, 2008). Nevertheless, we are somewhat optimistic about the future. Ensuring reproducibility is not only important for the field as a whole, but also for individual researchers: below we show that the median citation count for studies with working links to the source code are higher.
We conduct the first experiment in the literature in which a novel is translated automatically and then post-edited by professional literary translators. Our case study is Warbreaker, a popular fantasy novel originally written in English, which we translate into Catalan. We translated one chapter of the novel (over 3,700 words, 330 sentences) with two data-driven approaches to Machine Translation (MT): phrase-based statistical MT (PBMT) and neural MT (NMT). Both systems are tailored to novels; they are trained on over 100 million words of fiction. In the post-editing experiment, six professional translators with previous experience in literary translation translate subsets of this chapter under three alternating conditions: from scratch (the norm in the novel translation industry), post-editing PBMT, and post-editing NMT. We record all the keystrokes, the time taken to translate each sentence, as well as the number of pauses and their duration. Based on these measurements, and using mixed-effects models, we study post-editing effort across its three commonly studied dimensions: temporal, technical and cognitive. We observe that both MT approaches result in increases in translation productivity: PBMT by 18%, and NMT by 36%. Post-editing also leads to reductions in the number of keystrokes: by 9% with PBMT, and by 23% with NMT. Finally, regarding cognitive effort, post-editing results in fewer (29% and 42% less with PBMT and NMT respectively) but longer pauses (14% and 25%).
When courts started publishing judgements big data analysis (i.e. large-scale statistical analysis and machine learning) within the legal domain became possible. By taking data from the European Court of Human Rights as an example, we investigate how Natural Language Processing tools can be used to analyse texts of the court proceedings in order to automatically predict (future) judicial decisions. With an average accuracy of 75% in predicting the violation of 9 articles of the European Convention on Human Rights our (relatively simple) approach highlights the potential of machine learning approaches in the legal domain.
In phonetics, many datasets are encountered which deal with dynamic data collected over time. Examples include diphthongal formant trajectories and articulator trajectories observed using electromagnetic articulography. Traditional approaches for analyzing this type of data generally aggregate data over a certain timespan, or only include measurements at a fixed time point (e.g., formant measurements at the midpoint of a vowel). In this paper, I discuss generalized additive modeling, a non-linear regression method which does not require aggregation or the pre-selection of a fixed time point. Instead, the method is able to identify general patterns over dynamically varying data, while simultaneously accounting for subject and item-related variability. An advantage of this approach is that patterns may be discovered which are hidden when data is aggregated or when a single time point is selected. A corresponding disadvantage is that these analyses are generally more time consuming and complex. This tutorial aims to overcome this disadvantage by providing a hands-on introduction to generalized additive modeling using articulatory trajectories from L1 and L2 speakers of English within the freely available R environment. All data and R code is made available to reproduce the analysis presented in this paper.
In this study, we investigate crosslinguistic patterns in the alternation between UM, a hesitation marker consisting of a neutral vowel followed by a final labial nasal, and UH, a hesitation marker consisting of a neutral vowel in an open syllable. Based on a quantitative analysis of a range of spoken and written corpora, we identify clear and consistent patterns of change in the use of these forms in various Germanic languages (English, Dutch, German, Norwegian, Danish, Faroese) and dialects (American English, British English), with the use of UM increasing over time relative to the use of UH. We also find that this pattern of change is generally led by women and more educated speakers. Finally, we propose a series of possible explanations for this surprising change in hesitation marker usage that is currently taking place across Germanic languages.
The present study uses electromagnetic articulography, by which the position of tongue and lips during speech is measured, for the study of dialect variation. By using generalized additive modeling to analyze the articulatory trajectories, we are able to reliably detect aggregate group differences, while simultaneously taking into account the individual variation of dozens of speakers. Our results show that two Dutch dialects show clear differences in their articulatory settings, with generally a more anterior tongue position in the dialect from Ubbergen in the southern half of the Netherlands than in the dialect of Ter Apel in the northern half of the Netherlands. A comparison with formant-based acoustic measurements further reveals that articulography is able to reveal interesting structural articulatory differences between dialects which are not visible when only focusing on the acoustic signal.
In this study we investigate the effect of age of acquisition (AoA) on grammatical processing in second language learners as measured by event-related brain potentials (ERPs). We compare a traditional analysis involving the calculation of averages across a certain time window of the ERP waveform, analyzed with categorical groups (early vs. late), with a generalized additive modeling analysis, which allows us to take into account the full range of variability in both AoA and time. Sixty-six Slavic advanced learners of German listened to German sentences with correct and incorrect use of non-finite verbs and grammatical gender agreement. We show that the ERP signal depends on the AoA of the learner, as well as on the regularity of the structure under investigation. For gender agreement, a gradual change in processing strategies can be shown that varies by AoA, with younger learners showing a P600 and older learners showing a posterior negativity. For verb agreement, all learners show a P600 effect, irrespective of AoA. Based on their behavioral responses in an offline grammaticality judgment task, we argue that the late learners resort to computationally less efficient processing strategies when confronted with (lexically determined) syntactic constructions different from the L1. In addition, this study highlights the insights the explicit focus on the time course of the ERP signal in our analysis framework can offer compared to the traditional analysis.
Dialectometry applies computational and statistical analyses within dialectology, making work more easily replicable and understandable. This survey article first reviews the field briefly in order to focus on developments in the past five years. Dialectometry no longer focuses exclusively on aggregate analyses, but rather deploys various techniques to identify representative and distinctive features with respect to areal classifications. Analyses proceeding explicitly from geostatistical techniques have just begun. The exclusive focus on geography as explanation for variation has given way to analyses combining geographical, linguistic, and social factors underlying language variation. Dialectometry has likewise ventured into diachronic studies and has also contributed theoretically to comparative dialectology and the study of dialect diffusion. Although the bulk of research involves lexis and phonology, morphosyntax is receiving increasing attention. Finally, new data sources and new (online) analytical software are expanding dialectometry's remit and its accessibility.
This study uses a generalized additive mixed-effects regression model to predict lexical differences in Tuscan dialects with respect to standard Italian. We used lexical information for 170 concepts used by 2,060 speakers in 213 locations in Tuscany. In our model, geographical position was found to be an important predictor, with locations more distant from Florence having lexical forms more likely to differ from standard Italian. In addition, the geographical pattern varied significantly for low- versus high-frequency concepts and older versus younger speakers. Younger speakers generally used variants more likely to match the standard language. Several other factors emerged as significant. Male speakers as well as farmers were more likely to use lexical forms different from standard Italian. In contrast, higher-educated speakers used lexical forms more likely to match the standard. The model also indicates that lexical variants used in smaller communities are more likely to differ from standard Italian. The impact of community size, however, varied from concept to concept. For a majority of concepts, lexical variants used in smaller communities are more likely to differ from the standard Italian form. For a minority of concepts, however, lexical variants used in larger communities are more likely to differ from standard Italian. Similarly, the effect of the other community- and speaker-related predictors varied per concept. These results clearly show that the model succeeds in teasing apart different forces influencing the dialect landscape and helps us to shed light on the complex interaction between the standard Italian language and the Tuscan dialectal varieties. In addition, this study illustrates the potential of generalized additive mixed-effects regression modeling applied to dialect data.
I frequently teach (invited) statistics courses for linguists focusing on mixed-effects regression and generalized additive modeling. The first technique, mixed-effects regression modeling, is the method of choice when analyzing linguistic data where subjects have provided data for multiple items (e.g., reaction times). The second technique, generalized additive (mixed) modeling (GAMs) is important as it allows to model complex non-linear relationships between predictors and the dependent variable (e.g., time-series data such as EEG data). I've been invited to teach these courses at (e.g.,) Cambridge, Montréal and Toulouse. Slides (which are regularly updated) of these courses can be found here. If you are interested in this type of statistics course (generally ranging from two to five days), you are welcome to contact me.
We enjoy demonstrating how we collect data on tongue and lip movement during speech. If you'd like a demonstration at your school or event, please contact me. Below you can see an impression of my team at the Experiment Event for children organized by De Jonge Akademie, the NS, the Spoorwegmuseum, and Quest Junior.
Through a project grant of De Jonge Akademie, I was able to create a comic about my research (designed and drawn by Lorenzo Milito and Ruggero Montalto). You can view the comic below or download it here for free. Please contact me if you would like to receive a printed copy of the Dutch version of the comic (as long as supplies last).
(Note that especially the English news coverage in 2014 got many details wrong, see this Language Log post.)