Spontaneous Speech Data Compared to Elicitation Data: The Test Effects

Leonie Cornips

Meertens Instituut
In this talk, I will first address the reliability (validity) of written questionnaires that are used to collect dialect data in a large geographical area. Second, I will focus on the discrepancies between spontaneous speech and oral elicitation data in Heerlen Dutch. Heerlen is situated in the province of Limburg in the southeast of The Netherlands, near the Belgian and German border and it is the result of a process of abrupt language shift in the beginning of this century with the local dialect as the source and standard Dutch as the target language. More specifically, I will address the question to what extent spontaneous speech and oral elicitation data are consistent with linguistic and social distribution regarding the infinitival complementizers om (standard Dutch variant) and voor (local dialect variant) and dative objects (dialect variants):
    1. ...moet ik terugkomen om dat half jaar ( ) af te maken (12: Bert)
      'I have to go back to finish that half year'
    2. ...je komt hier voor te studeren (12: Bert)
      'You will come here to study'
The data in Heerlen Dutch are collected by means of (i) recording spontaneous speech (33,5 hours) and (ii) administration of oral elicitation tests. Crucially, the variants of the complementizer om and voor occur even while one speaker is maintaining the same level of speech style (individual-level) in spontaneous speech. For example, (1a) and (1b) were uttered by one speaker ('Bert') on the same occasion.
Regarding the elicitation data, the social variables do not display the same distribution as the spontaneous speech data. I will discuss that these patterns are primarily due to distinct test effects.
Back to the programme
Syntactic Atlas of the Dutch Dialects (SAND)