Presentation mode and the number of speakers: Influences on structural priming in 4-year-olds

Research Areas: 

Structural priming is considered as one possible implicit learning mechanism in language acquisition. This project focusses on the influences on structural priming in preschool children. For this, we study whether the presentation mode (a video vs. a live presentation) and the number of speakers (one vs. two persons) have different effects on the production of passives in describing pictures.

Methods and Research Questions: 

The aim of the project is to find answers to the following question:
Are 4-year-old children sensitive to the number of speakers talking to them live or in videotaped action sequences and do rely on social references when being primed structurally?

1. One or two person speaking? Many imitation studies focus on the type of models imitated by children, for example mothers vs. strangers or peer vs. adults. Assuming that language development is based on imitation as well as priming effects (reported as enhancing grammar learning) we investigated, whether children are sensitive to the number of speakers talking to them.

2. Video or live? Young children are able to learn different skills from watching television. However, they perform better when they are taught by a person providing social references. While this effect of presentation mode has also been shown for the acquisition of nouns and verbs in language development (e.g. Krcmar, Grela, & Lin, 2007; Roseberry, Hirsh-Pasek, Golinkoff, & Parish-Morris, 2009), the impact of the presentation mode on the acquisition of syntactic structures has been barely studied.

Structural priming has been discussed as enhancing implicit grammar learning (Chang, Dell, Bock, & Griffin, 2000; Savage, Lieven, Theakston, & Tomasello, 2006). Recently, several studies have shown that structural priming is possible in three- and four-year-old children (e.g. Shimpi, Gámez, Huttenlocher, & Vasilyeva, 2007).
Using a priming paradigm, we investigated to what extent 4-year-olds rely on the social cues of one person vs. two persons speaking to them when being primed with passives. Furthermore, we compared the efficiency of structural priming with passive constructions in a video-based to a live presentation mode. We trained the children with six different actions combined with passive constructions each.

The children were asked to describe two picture sets, with eight pictures each. Additionally there was one training picture in the beginning. The posed question was “What is happening to_______?” to focus the child’s attention on the object of the sentence. The actions on the pictures showed the handpuppets using in the training. We counted the number of passives in the description of the pictures. Then we compared the number of passives in the pre- and posttest. We coded the children’s utterances as “passive” when they were built as follows:
(AGENT) (be) (PATIENT) VERB(transitive participle).


1. One or two persons speaking?
Whereas the median score on passives in the group with one speaker in the live presentation increased from Md = 0 to Md = 1, it increased from Md = 0 to Md = 2 in the two speaker condition. The Wilcoxon Signed Rank Test revealed the trend that only the children in the two speaker condition produced more passives after the training than before              (Z = -1.79, p = .073 , r = .28).
This study indicates that children are sensitive to the number of speakers using a specific structure in a priming situation. In a learning situation, two speakers presenting one particular grammatical structure might also entrench this structure more than only one speaker using this structure.
In the video condition, no statistical effect of the number of speakers was observeable.
We believe not only simply automatic priming processes were effective within the repetition of the passive in the live condition but also a direct adaption to the social situation. Maybe the structure was better perceivable with two persons using it or applied higher pragmatic pressure on the child. It is also possible, that the children interpreted the repetition of the second speaker as an approval for correctness. This was not possible for the children in the video condition.

2. Video or live?
Comparing the video to the live presentation in the one speaker condition the Mann-Whithney-U Test revealed no significant difference between the conditions at the first testing (U = 214.5, Z = -.463, p = .643) whereas at the second testing a significant difference between the conditions was detectable  (U = 148, Z = -2.327, p = .02). This was also similar in the two speaker condition in which the Mann-Whithney-U Test revealed a trend in the second testing (first testing: U = 202, Z = -.525, p = .600, second testing: U = 147,5, Z = -1.927, p = .054). The results from these analyses suggested that the effect was more pronounced with live than with video-mediated primes providing evidence that in this study structural priming, too, benefitted from social interaction.