Michelle Suijkerbuijk

• Master of Arts
• PhD Candidate at Radboud University

We present a corpus of 8400 Dutch sentence pairs, intended primarily for the grammatical evaluation of language models. Each pair consists of a grammatical sentence and a minimally different ungrammatical sentence. The corpus covers 84 paradigms, classified into 22 syntactic phenomena. Ten sentence pairs of each paradigm were created by hand, while...

Neural language models (NLMs) have frequently been tested on their ability to model grammatical phenomena in a "human-like'' way, but their behavior is often not compared to actual human data, making it unclear whether their behavior is actually human-like. An example of this can be seen with long distance filler-gap dependencies and their island c...

We present a corpus of Dutch 8400 sentence pairs intended for the grammatical evaluation of language models. Each pair has a grammatical sentence and a minimally different ungrammatical sentence. The corpus covers 84 paradigms, classified into 22 syntactic phenomena. Nine sentences of each paradigm were rated for acceptability by at least 30 partic...

We present a corpus of 8400 Dutch sentence pairs, intended primarily for the grammatical evaluation of language models. Each pair consists of a grammatical sentence and a minimally different ungrammatical sentence. The corpus covers 84 paradigms, classified into 22 syntactic phenomena. Ten sentence pairs of each paradigm were created by hand, while...

Argument doubling, also known as (contrastive) left-dislocation, is common in spoken Dutch, but it is unclear exactly what triggers it. Earlier proposals in the literature showed that the construction is not used for marking contrast, and suggested it is used for marking shifted topics instead. However, the results from a Spoken Dutch Corpus study...

The current study investigates whether a Long Short-Term Memory (LSTM) network can learn the wh-island constraint in Dutch in a way comparable to human native speakers. After establishing with an acceptability judgement task that native speakers demonstrate a clear sensitivity to wh-island violations, the LSTM network was tested on the same sentenc...

The current study investigates whether a Long Short-Term Memory (LSTM) network can learn the wh-island constraint in Dutch in a way comparable to human native speakers. After establishing with an acceptability judgement task that native speakers demonstrate a clear sensitivity to wh-island violations, the LSTM network was tested on the same sentenc...

While kinship relations in Dutch are usually introduced by a possessive determiner, Twitter users have recently been observed to use a definite article in that position. To learn more about the characteristics of this construction, we performed an exploratory investigation of the definite article possession construction with Dutch kinship terms on...

In an online production experiment, we investigated the effect of sentence position on the preference for either a nominative or object form of an object pronoun restricted by a relative clause in Dutch. Results show a significant preference for the nominative form of the restricted object pronoun in sentence-initial position as it was chosen in 95...