2. Summary of the chapter
Ford and Bresnan’s study combines methods from corpus linguistics and psycholinguistics in order to analyse the dative alternation. Dative constructions can assume two forms in English: the double object dative as in He showed the woman the ticket, and the prepositional dative as in He showed the ticket to the woman. The syntactic order is changed as the theme (the ticket) and the recipient (the woman) switch places. Ford and Bresnan examined which factors influence the choice of the dative form.
Corpus analysis
Drawing on the Switchboard Corpus of American English Telephone Conversations, they created a database of 2,349 datives. With the help of mixed effects modelling and the free statistical software environment R, they determined fixed effects, which either increase or decrease the likelihood of a prepositional dative.
The following table provides an overview of the effects.
pronominality of the recipient |
pronoun (e.g. me) vs. nonpronoun (e.g. my kids) |
pronominality of the theme |
pronoun vs. nonpronoun (see above) |
definiteness of the recipient |
definite (e.g. my kids) vs. indefinite (e.g. some people) |
definiteness of the theme |
definite vs. indefinite (see above) |
animacy of the recipient |
animate (e.g. my kids) vs. inanimate (e.g. the poorer school districts) |
number of the theme |
singular (e.g. this car) vs. plural (e.g. our cars) |
previous dative |
no previous dative vs. previous double object dative vs. previous prepositional dative |
log length difference |
the logarithm of the length of the recipient (counted in words) minus the logarithm of the length of the theme (counted in words) |
verb sense |
Bresnan and Ford differentiated not only between verbs, but also between verb senses of one verb: e.g. give in the transfer sense vs. give in the abstract sense |
Table 1: Factors influencing the dative choice
Bresnan and Ford included the influence of each factor in a model formula and used it to calculate the ‘corpus probabilities’ of the prepositional dative for all sentences. A detailed explanation of the model formula can be found in chapter 16 on pages 298-99. The question arising from those calculations was: Can humans make such accurate predictions, too?
Psycholinguistic methods
In order to analyse human intuitions about language, Bresnan and Ford conducted three psycholinguistic experiments. They investigated the participants’ judgements, their online processing, and sentences the participants produce themselves. As artificial items might have caused various problems, the scholars used items from the Switchboard Corpus again. About 40 students from the USA and Australia participated in each experiment.
Bresnan and Ford created a questionnaire with 30 items from the Switchboard Corpus. All of them end with a choice between the double object and the prepositional dative. The participants had to rate the naturalness of these alternatives by distributing 100 points between them, e.g. 50 – 50 if they considered both equally natural. The results showed that the participants were sensitive to the variables influencing the dative form.
Participants were presented with sentences containing prepositional datives one word at a time, while the reaction times to the word to in the prepositional dative were measured. If people reacted fast, it was assumed that they had anticipated the next word. The experiment showed that the participants were sensitive to several predictors.
Participants were given sentences from the Switchboard Corpus containing the beginning of the dative construction (i.e. all elements up to the verb) and had to complete the sentence. Bresnan and Ford used this method because the results of the first two experiments indicated that Australians prefer the prepositional dative more than US-Americans. As no comparable datasets existed, this method presented a way of comparing the two varieties. The results proved that Australians do indeed use the prepositional dative more often.
Convergent evidence
Using convergent evidence from corpus linguistics and different psycholinguistic methods requires far more effort, knowledge and resources than the use of a single method. However, a multi-method approach takes into account many factors is thus able to capture speakers’ intuitions and mirror complex linguistic reality.
Created with the Personal Edition of HelpNDoc: Create iPhone web-based documentation