Current project: Variflex

Previous projects

May 1997 I started in a joint project between Utrecht and Antwerp University, in which human language learning is simulated with the help of an artificial learner.

My PhD. thesis focuses on the development of finiteness. To gather more insight in this process, I try to reveal the small steps that take children towards the final adult stage.

A data-driven model of language acquisition: Computational and psycholinguistic investigations

The aim of the project is the development of a computational psycholinguistic model of morphosyntactic aspects of language acquisition. This includes a psycholinguistic investigation of the acquisition of morphosyntax, and more specifically the acquisition of the morphological and distributional reflexes of the feature 'finite' in Dutch. A computer model of these linguistic phenomena will be implemented in which the principles of similarity based reasoning will be represented.

This project is funded by the "Flemish - Dutch Committee for Dutch Language and Culture" (VNC) under the auspies of the Belgian National Science Foundation (NFWO) and the Dutch Science Foundation (NWO).

Elma Blom, Frank Wijnen. University of Utrecht, UiL-OTS
Paul van Geert, Jan Koster, Gerard Bol. University of Groningen
Masja Kempen, Steven Gillis, Walter Daelemans, Georges de Schutter. University of Antwerp

Here are two examples of the movies we have shown to Dutch and English children to elicit verb forms. You can view them with, for instance, Quicktime.

This movie is about a little dirty pig that is taking a bath. She/he is washing her/himself until all the dirt is gone.

This movie expresses some more drama; it is about a boy who is playing on the street. When a car approaches, he has to run away.

The acquisition of finiteness

Considering the complexity of grammatical knowledge, children learn grammar with remarkable speed. A 3-year-old child has already knowledge of most syntactic properties of the target language. One of the questions I am interested in is how children achieve this. Take verb movement, which is a structural means to mark finiteness. In Dutch main clauses, the finite verb is moved from sentence-final to sentence initial position (a trace of this movement operation can be observed in particle verbs: Ik wil hem vanavond opbellen/Ik bel hem op ...). For children learning Dutch, verb movement is not obvious. A closer look at the the input data of Dutch children shows that there is, for instance, only marginal lexical overlap between unmoved non-finite verbs and moved finite verbs. Dutch children have to overcome this "poverty of stimulus" effect. Is the input too poor to find out that finite and non-finite verbs belong to one category? I do not think so. Children have to learn that finite verbs and non-finite verbs belong to one categorie and that the syntactic distinction between the two is due to finiteness features. Inflectional morphology provides a cue for this generalization.

Infinitives and bare stems

English root infinitives contain bare stems whereas root infinitives in Dutch, German and French contain true infinitives. The asymmetry goes further: English root infinitives contain more stative predicates than root infinitives in Dutch, German and French, they are less often modally used and contain more often wh-words. How to explain these cross-linguistic differences? First of all, within Dutch, German and French there is an asymmetry between children's root infinitives and finite sentences: the latter are more often modal, more often stative and contain wh-words. Secondly, children tend to truncate words and use stripped forms. Children in all languages children do this, but only in English this results in finite sentences that cannot be distinguished from root infinitives.

Stative predicates, cognitive immaturity and input

Patterns in child language are not necessarily due to deficits in their language knowledge. One of the most difficult tasks of a child language researcher is to disentangle the complex of factors that influences child language. There are phenomena in early child language for which different explanations can be given and are given. The absence of stative root infinitives in Dutch child language is such a phenomenon. Dutch children use an abundant number of stative predicates in their early finite clauses, whereas their root infinitives contain hardly any stative predicates. There are a number of reasons to believe that this asymmetry reflects cognitive immaturity of children between 2 and 3 years old, more specifically the absence of a theory of mind, and reflects distributions in the input.

These are three examples of issues that I discussed in my thesis, From Root Infinitive to Finite Sentence: the acquisition of verbal inflections and auxiliaries, This is the covertext of the book. The dissertation is available January 2003 in the form of a book (LOT-dissertation series). The electronic version will be downloadable from the site of the Utrecht University Library.

Supervision: Frank Wijnen, Henriette de Swart, Paul van Geert
External supervision: Nina Hyams