Skip to ContentSkip to Navigation
About usWorking with usGood reasons to work with usTestimonialsTales of Talent

Martijn Wieling: computational linguist

On entering Martijn Wieling’s office, he apologizes for the fact that it’s difficult to shut the door. A large part of his office is filled with a huge case and a very unusual-looking device with sensors protruding from it. It turns out to be an articulograph, a device for monitoring people’s tongue and lip movements when they talk. Such knowledge could prove very useful when learning a foreign language, for example.

Martijn Wieling, fotograaf: Gerhard Taatgen

From accountant to IT specialist

Wielings enthusiasm for articulography came as a surprise to him. ‘I’d wanted to be an accountant ever since year 4 of secondary school. That’s why I started an on-the-job training scheme in accountancy after I left pre-university education. But despite my original enthusiasm, I soon realized that accountancy is rather like a straitjacket in which you are confined by rules with very little room for manoeuvre. After I decided to stop, I spent a few months working in horticulture until I’d worked out what I wanted to do next. I came across a degree programme in Computer Science which seemed to combine theory with practice. It was a good choice.’ Wieling obtained a Bachelor’s degree in Computer Science and was admitted to the new Research Master’s in Behavioural and Cognitive Neurosciences, a programme that he took alongside the regular Master’s degree programme in Computer Science. ‘I wanted to get into research, but enjoyed doing other things outside my studies as well. I did various student assistantships and had a seat on the University Council for a year.’

Exam training

‘Something else I really enjoyed and spent a long time doing was coordinating exam training courses. I started as a teacher for these courses in Leiden, but then it occurred to me that Groningen should offer these as well. So together with the UOCG [now ESI, Educational Support and Innovation], I introduced the concept in Groningen. We started with 60 pupils, but the courses proved so successful that we now help some 700 schoolchildren every year. I coordinated this for many years, together with a colleague from the UOCG. I only stopped as coordinator in the final year of my PhD.’

Electronic learning environment and student performance

After completing the Research Master’s, Wieling was offered a PhD position by the UOCG, with Professor Adriaan Hofman as his supervisor. He carried out research into the effect of electronic learning environments (such as Blackboard) on student performance. ‘We demonstrated that students benefit from being able to watch video lectures. But I missed the more computational side of research in this project.’

Research into dialect variation

So when a PhD position for a computational study of dialect variation became available in the Department of Humanities Computing, Wieling immediately was interested. ‘This position allowed me to devise new quantitative methods for studying dialect variation. While you can certainly map the areas where the word ‘car’ is pronounced in a particular way, these boundaries are word-specific. Studying a larger number of different words gives you a better idea of the possible dialect areas. This is much more objective.’ Wieling’s dissertation was a technical amalgamation of two fields: dialectology and sociolinguistics. Sociolinguistics focuses on speech variation in relation to social differences, such as the differences between men and women or between young people and the elderly, whereas dialectology examines the link between speech variation and geography. The method that Wieling used made it possible to determine both the influence of various social factors and the influence of geography on dialect variation. ‘I think it’s very important to share methods, which is why I published details of this particular statistical method on my website. In addition, I discuss the method in the short statistics courses I’m sometimes asked to give to linguists at other universities, including Cambridge, Montréal and Tübingen.’

Attaching sensors

‘I was keen to continue in research after I’d completed my PhD. Thanks to a Rubicon grant, I was able to conduct research in Tübingen, Germany, where Professor Baayen had purchased an articulograph. This device is used to monitor the position of the tongue during speech and thereby allowed us to investigate dialect differences at a new level. We took the device to a number of schools, where we studied the way pupils spoke in their local dialect. We asked all of them to pronounce a long list of words, while the device monitored the movement of their tongue and lips. Attaching the sensors was not always easy. You have to stick them onto the lips and tongue and that can be quite tricky. For example, some of the children didn’t follow our instructions to keep swallowing during the experiment, so the excess saliva in their mouths loosened the sensors, causing them to detach before all words were pronounced. No, attaching those sensors wasn’t one of my finest hours ,’ says Wieling with a smile on his face. ‘But we collected a lot of valuable data. We now know that the tongue position of the northern speakers is really different from that of the southern speakers, irrespective of whether people are speaking dialect or standard Dutch.’

How do you pronounce a foreign language?

Wieling was recently awarded a prestigious Veni grant for new research using the articulograph. ‘We are going to use the device to register the way that people pronounce a foreign language. How do Dutch people pronounce English, for example? We have asked people from the Netherlands, Germany and Great Britain to take part. The device is portable so it can be setup in different locations. Well, the case weighs almost thirty kilos so it’s not that easy. But it has been to London’, says Wieling with obvious satisfaction. ‘One of the things we observed is that Dutch people’s tongues are further back in their mouth when they pronounce ‘th’ in English than those of native English people. The next step is to see whether (and to what extent) visualizing the tongue and lip movements can improve the way people pronounce foreign languages. It could be very useful in education, but may also have the potential to help people with speech impediments.’

Making data publicly available

Wieling was recently made a member of the Young Academy, a platform of talented young researchers. ‘The main advantage for me is that it will allow me to share my fascination for science with a much wider audience. I have occasionally been interviewed in the media about my research, but I’ll be able to reach a lot more people now. In addition, I think talking about and influencing scientific policy is incredibly important. For example, I believe that research data from publicly funded research should be made publicly available. Society has helped to pay for this research, so society should have free access to the information as well. It also must be made clear how the researchers obtained their results, so you should be able to precisely reproduce what researchers have done in their analyses. To me transparency is very important in research.’

More information

Last modified:18 August 2015 11.09 a.m.
printView this page in: Nederlands