Gilquin, Gaëtanelle
[UCL]
The human mind is a black box and access to conceptual structures is necessarily indirect (Nuyts 1993). One such indirect means of access is language. Because it explicitly encodes and transmits conceptual information, linguistic behaviour is often considered a “relatively privileged” source of information on conceptualisation (Pederson & Nuyts 1997). This view implies that one deals with the disparities, regularly emphasised in the literature, between language comprehension and language production (see e.g. Dieser 2006), two aspects which can be reconciled by invoking the distinction between “input” and “output”. More problematic, however, is the lack of consistency within language production itself. In this presentation, two types of language production will be investigated and compared, namely corpus data and elicitation data. More precisely, it will be shown that the two types of data provide different answers to the question of cognitive salience.
The corpus data were extracted from Switchboard, a collection of transcribed telephone conversations in American English. For the elicitation data, a test was designed were American informants had to produce, for a series of linguistic stimuli, the first sentence they could think of. All the data were analysed according to the frequency with which the different senses of a given word were produced, this frequency being arguably a measure of cognitive salience. The analysis reveals marked divergences between the two measures. To give but one example, the most common use of the verb 'give' in the corpus is the delexical one (e.g. 'give a smile', 'give a look'), while the sense that is most frequently elicited in the experiment is that of “handing” (e.g. 'give somebody a cookie'). This confirms the discrepancy between corpus data and elicitation data brought to light by some recent studies, cf. Roland & Jurafsky (2002), Nordquist (2004) or Shortall (in preparation).
While these results may seem to question the validity of language production as a window to the mind (if different types of language production offer different pictures, what can be said about conceptualisation?), it will be argued that the two types of evidence tap into different things and provide complementary views of the mind.


Bibliographic reference |
Gilquin, Gaëtanelle. Language production: a window to the mind?.22nd Scandinavian Conference of Linguistics (Aalborg, du 19/06/2006 au 22/06/2006). |
Permanent URL |
http://hdl.handle.net/2078.1/112530 |