“We want to describe how a human works”

An interview with Silke Hamann

1 December 2015

Silke Hamann is research-oriented lecture (NGO) at the Linguistics department at the UvA, where she studies the phonetics-phonology interface. She is currently working on loan words in Hong Kong Cantonese and Italian, at the restrictions on speech perception by amusics (with Jasmin Pfeifer), and at the phonetics and phonology of Bantu languages. She studied at the Goethe University Frankfurt and at the Free University of Berlin, and did her PhD at the University of Utrecht and the Zentrum für Allgemeine Sprachwissenschaft, Berlin.

“I started studying architecture in Weimar. I was good at arts and mathematics at school, and was fascinated by houses, flats; everything that people can live in since I was little. Quite soon I realized that this study was not really what I had expected. So I started German literature and English literature, because I always loved reading. While doing that, I had to do linguistics. Before that, I did not have a clue what linguistics was about. I found it really cool. Linguistics is so logical, it is all about finding a way to sort data, to describe it systematically, to form theories about it and test them, and I always have loved logical puzzles. So I added linguistics as my second major while studying at the Goethe University Frankfurt (the region where I originally come from). After my Zwischenprüfung – the former German equivalent of the BA – I went to the Free University of Berlin. At that time, I still did not know what to do afterwards. Towards the end of my study, one of my lectures asked me whether I would be interested in doing a PhD because they had a PhD position. I said ‘sure, if you think I am good enough for that, I will apply’. But since I did not have my degree yet, I was not accepted. However, I had made contact with a possible supervisor at the Zentrum für Allgemeine Sprachwissenschaft (ZAS) Berlin through that procedure, and a few months later I got the possibility to write an application for a position that was shared between the ZAS and the University of Utrecht. And I got it. I spent half of my time during my PhD in Utrecht, which I really enjoyed because there we were a large group of enthusiastic PhD students. With this position my family (all non-academics) finally felt assured with what I was doing and where it was going; they used to be a bit skeptical about my study choices. After my PhD I decided that I wanted to continue in academia. I did a couple of postdocs, going backwards and forwards between the Netherlands and Germany. First, I went back to the ZAS for a postdoc. Then I came to Utrecht with a VENI grant; I did not finish that, because I got a position as Junior Professor in Düsseldorf, which was a great experience but also a lot of work without having the chance of getting a permanent position after that (no tenure track). Now I am here at the UvA, and am really happy about it.”

[Answer on how it is to work together with her spouse, Paul Boersma] “Working together with Paul is both great and also tricky. It is perfect because we work on very similar things, so we can discuss a lot any time we want to. Of course we do not do that all the time. Actually we talk less about work the last years, as we are both involved in our separate projects. But still, whenever one of us has an idea, we can just talk about it. This might happen in the weekend, e.g. when we take a long walk at the beach, without the other one getting annoyed. However, in terms of career it was a bit problematic for me. We both work in the same area and we got together when I was finishing my PhD. So I was the inexperienced, young researcher, and I had to prove to myself and the community that I am not only the ‘wife of’. Luckily he was never my supervisor, so it was not that weird situation where a teacher and a pupil get together. For my dissertation, I did my own stuff, and then applied his theory, because I was really convinced by it (I still think that it is brilliant what he is doing, by the way). In this way we got into contact, and that is how we got together. We both love our work, and I think we influence each other a lot. In terms of positions, it is not easy to find work for the two of us in the same region, let alone at the same university, so we are very lucky to both work at the same department here at the UvA. Altogether I would say: it has its problematic parts, but the good parts far outweigh.”

“I came to phonetics and phonology quite late during my studies. I did a lot of syntax and liked it, but at some point I realized that the whole field depended too much on the ideas of one person: Whenever Chomsky came up with something new, almost everybody started following his new course. At that time I also got the chance to follow several courses on phonetics/phonology because I spent a year as an exchange student at the University of Manchester. There was not much on offer in phonetics and phonology at the Free University in Berlin. And sadly, this was and still is the situation at a lot of German universities. And by now it is turning like that at several Dutch universities, too. If I am correct, Paul is the only professor in Phonetics in the Netherlands by now, while there were five when I came to the Netherlands. Concerning phonology, the situation is slightly better, but not much. Here at the UvA, I could fill the gap that arose when Norval Smith – the only designated phonologist, retired, though my position is not permanent (yet). ”

“My key research area is the phonetics/phonology interface. That means I want to know how the measurable part of sounds (how we pronounce them, how they can be acoustically analyzed, etc.) interacts with the abstract knowledge of the sounds in our native language(s), the part that we cannot directly measure. At the moment, I am really interested in loan phonemes: new sounds coming into a language, like the English ‘g’ in Dutch. Older speakers might still pronounce loan words like ‘google’ with a ‘k’, but younger speakers, who have much more English input, know that it is a different category. You need input from English to be able to realize that there is another category you have to newly create in your native language. Together with an MA student, Alma de Jonge, I did an experiment in our phonetics lab where we looked at the sound ‘g’ in loan words referring to food. Pairs of participants had to create three-course menus with a given list of ingredients; in that way we got them to use the words we were interested in without them being aware of it. When you just give them listst to read they usually think too much about their pronunciation and say what they think is the right form but does not really correspond to what they would naturally do. We found that older speakers often pronounced mango and gorgonzola with the sound as in the beginning of Dutch geel, but that younger speakers use the English ‘g’ in these words. However, for the word spaghetti, both older and younger speakers use the sound in geel. Why do the young speakers use an originally Dutch sound for this but not for the other words? That is fascinating! It seems that a lot of other factors such as time of borrowing and frequency of the word, its orthography, the status of the language we borrow from, influence the adaptation of new sounds into our native language. And all of them have to be investigated in detail before we know what exactly is going on.”

“Together with my PhD student Jasmin Pfeifer, I work on congenital amusia, which is a disorder that affects music perception. We came across that topic more by accident, but it turned out to be really interesting for linguistics, because amusia seems to have an influence on language perception. Initial anecdotal evidence showed that it is not just pitch differences used in language that might be impaired in amusic people, but also other parameters, e.g. vowel duration. Jasmin and I are looking at differences in the perception of vowel duration and stress at the moment. But what I find even more fascinating is how it is possible that, since amusics have problems in perceiving differences in pitch and duration, they do seem to be able to produce language that doesn’t seem different from the language by non-amusics. I am saying ‘seem to be able’, because that it something we have not measured yet, something we simply do not know yet: no one ever reported that amusics ‘sound funny’, but maybe they do produce things differently. That is something I really would like to look at in the future, so I am writing a grant proposal about this (on possible asymmetries in perception and production in general). I would expect some differences in production, but it would be even more interesting if it turned out that they actually pronounce everything like non-amusics, because then we would have to answer the question how one learns something that one cannot perfectly perceive!”

“I think language and music share quite a lot. In both fields, we deal with abstract categories and with restrictions on the combination of these categories. We abstract in language and in music, and how we do that, depends on what we are exposed to. I think this has to do with how humans work in general: we abstract away with whatever we perceive, we cannot store all the details. Another interesting parallel between music and language is the restrictions with respect to the output. In language you have restrictions on your articulatory system, and in music you have restrictions on what you can do with your vocal folds and fingers etc. in terms of instruments. Although there is also a big difference: even though almost all of us can appreciate music (apart from most amusics), we do not necessarily all sing or play an instrument, whereas we all are able to communicate in at least one language, the latter is not really optional.”

“I think formal theories are really important in linguistics. Having a theory that forces you to formalize all the details of all your assumptions is important, because it makes you think about these details, and it enables you to make testable predictions. That is what a lot of phoneticians forget; they just look at really small details. This is of course important, too, but I am more interested in the bigger picture, in a model that can formalize learnability and change of sounds. That is why I like Paul’s work that much, because he started with a theory where you had for the first time a linguistic formalization of the speech perception process. How perception works in terms of phonetic mapping onto phonological categories, using means that are familiar to phonologists. He started with Optimality Theory, and he is now turning to Neural Networks (together with Kateřina Chladkova, Klaas Seinhorst and Titia Benders). This is more attractive, because it is much more realistic than OT. I have been starting now with that because I think this is the way to go, to also convince the traditional phoneticians. They always thought that Optimality Theory is just a toy created by and for phonologists.”

“We also need a theory to makes sure that we are not just measuring without any underlying concepts or hypotheses. We cannot e.g. just measure and say that one sound is similar to another one, because we cannot quantify this similarity in a meaningful way, as perceptual similarity is language-specific, not a universal notion. It sometimes frustrates me that some phoneticians are unwilling to consider languages as whole systems. I am bashing the phoneticians now, but the other way around is also not perfect: a lot of phonologists are not interested in phonetic details. However, you have to account for phonetic details. If you think of language as a whole, and not just your little part that you are looking at, you need a model that maps phonetics onto phonology and the other way around. And with the work I am doing with Jasmin , we see that part of the perception process, sort of the basic part, is shared with music perception, which is where it seems to go wrong for amusics. I think this is a step we also have to include in our linguistic model: first there is prelinguistic perception, then linguistic perception, and then we go up into the structure to abstract categories.”

“I think it is extremely important that a linguistic model is biologically plausible, and that that is the way to go. What do we have with nice little models that might describe the data correctly, if they do not describe what the human listener and speaker is really doing? There are a lot of models around that account for speech perception that assume we can store every item that comes in, and that we calculate averages over these stored items. But our brain does not work like that, like a computer. For some reason, in how far a model describes what we as listeners do is still a criterion that has not been taken into account often enough. I really think that this is what it all should be about in our field: we want to describe how a human works. That is why humanities are also so important: to focus on the perspective of the human, not only on the measurable data. We have had all these discussions about the humanities, that they do not earn enough money, and that we have to get rid of all courses where we do not have ‘enough’ students. I find this really blind. Humanities will never earn enough money, if you look at it that way. But we need to teach people the ability to set up theories about the human brain and mind that create hypothesis that then can be tested. You cannot just look e.g. at the neurological part of the brain; you need the interpretation as well. In the last 20-30 years we saw this trend where everything turned to measuring, acoustics, brain activity; you name it. This is good, because we are finally finding out e.g. where things are happening in the brain. But that is only the first step, we still need the interpretation of that: what does it tell us about our mind. We cannot measure a category, for example. We cannot measure where the loan sound ‘g’ (as in English get) is in a Dutch brain. I think that is where the humanities come in. We should be careful not just to go for the physical, the measurable, part, and forget about the rest. That is why I keep going back to phonology, the abstract part of sounds that is concerned with the mind.”

by Gisela Govaart, December 2015

Published by  SMART Cognitive Science