Using syntactic priming to increase robust learning

Project title	Training oral production in learning second language grammar
PI	De Jong (post-doc, now faculty)
Co-PIs	Perfetti, DeKeyser (faculty)
Others with > 160 hours	John laPlante
Study start date	March 2007
Study end date	September 2007
Learnlab	French
Number of participants	30
Total Participant Hours	75
Datashop	Log data are available.
Current status	Transcription and coding are in progress. (May 2009)

Abstract

The transfer of comprehension skills to production skills in second language learners seems to be, in many cases, neither spontaneous nor complete. Although there has been a considerable amount of discussion about the link between receptive and productive language skills in both the psycholinguistic and the applied linguistics literature, the question of how to bridge the gap between comprehension and production skills has not received much attention yet outside the narrow area of speech. This project investigates whether the proposed syntactic priming task can support transfer of comprehension skills to production skills with respect to two grammatical structures, thus supporting fluency in production. In this task, each instance of production is immediately preceded by an instance of comprehension. Because comprehension activates the required structural representations (e.g., knowledge of syntactic frames or verb forms; syntactic priming), the activated representations become more accessible for subsequent production, which will facilitate processing and increase fluency. Thus, new procedural knowledge for production may be formed (refinement), or existing knowledge may be strengthened.

Glossary

Comprehension skills

Skills in listening and reading

Production skills

Skills in speaking and writing

Syntactic priming

The phenomenon that speakers tend to use syntactic structures they have recently processed; also known as structural priming

Research questions

• Research question 1: Can errors in oral production be prevented by activating (i.e., increasing the availability of) the correct structure by aural input?
• Research question 2: Does oral syntactic priming lead to learning, i.e. a long-term increase in the availability of grammar knowledge for production, resulting in increased accuracy?
• Research question 3: Does oral syntactic priming lead to learning, i.e. a long-term increase in the availability of grammar knowledge for production, resulting in increased fluency?
• Research question 4: Does the increase in production skills transfer to semi-spontaneous speech?

Background and significance

The transfer from comprehension skills to production skills in second language (L2) learners seems to be, in many cases, neither spontaneous nor complete. Although there has been a considerable amount of discussion about the link between receptive and productive language skills in both the psycholinguistic and the applied linguistics literature, the question of how to bridge the gap between comprehension and production skills has not received much attention yet outside the narrow area of speech. This project investigates whether the proposed learning task can support transfer of comprehension skills to production skills with respect to grammar, thus supporting fluency in production.

Syntactic priming refers to the phenomenon that speakers tend to use syntactic structures they have recently processed. An explanation in terms of Levelt et al.’s (1999) model of language processing is that syntactic information can be pre-activated, for instance by listening to a sentence containing that syntactic structure. In subsequent production, it will be easier for the grammatical encoder to produce syntactically correct output because it can benefit from activation of the relevant syntactic knowledge components. McDonough (2006) argued that syntactic priming can support L2 acquisition, when learners have a choice between a simple and an advanced form, or between a non-targetlike form and a more appropriate form. When syntactic priming occurs, the structure is temporarily more available, and the learner is more likely to produce the more advanced or appropriate form in subsequent utterances. In our view, syntactic priming enables students to process advanced and appropriate forms more often than in traditional activities. This way knowledge components and procedures in the grammatical encoder are strengthened more effectively.

Dependent variables

Comprehension

• Acuracy (2-choice forced-choice)
• Reaction time

Production

• Accuracy (correct choice of word order; correct choice of mood)
• Response duration
• Number of pauses
• Length of pauses

Semi-spontaneous production

• Accuracy (correct choice of mood)
• Number of pauses
• Length of pauses
• Phonation/time ratio
• Location of pauses (before verb – elsewhere)

• Near transfer, immediate: The immediate posttest is the same task as the training tasks; however, items are presented in separate blocks of comprehension items and production items. The content of the items is different, that is, the vocabulary if different from the training tasks.
• Near transfer, retention: The delayed posttest is the same as the immediate posttest. It will be administered three to five days after the training and immediate posttests.
• Far transfer, retention: The delayed posttest is followed by a semi-spontaneous speech in which the target structure is elicited. This test is included only in the study on the conditional mood.

Independent variables

• Training: presence/absence of syntactic priming (availability)
• Time: immediate vs. delayed posttest (long-term retention)
• Task: controlled sentence-level production vs. extended semi-spontaneous speech (conditional only; transfer)

Hypotheses

It is expected that the proposed training—in which production of a sentence immediately follows comprehension of a sentence with the same structure—will increase accuracy in production during training. Fluency would also increase in that there would be fewer and shorter pauses, and response duration would be shorter. A more traditional training program in which comprehension and production are practiced in blocked tasks will also increase accuracy and processing speed in production, but to a lesser extent.

If the syntactic priming training indeed increases accuracy during training, it is expected that accuracy and fluency on the immediate and delayed production post-tests will be higher in the syntactic priming condition (long-term retention). The location of pauses may indicate that students in the syntactic priming condition pause less before the grammatical structure to be learned (pronoun order, or choice of mood) than in the non-syntactic priming condition.

It is hypothesized that students in the syntactic priming condition will achieve higher accuracy and fluency in the semi-spontaneous production posttest than the students in the non-syntactic priming condition (transfer).

Findings

Data was collected from March through September 2007. Data transcription has not started yet.

Specifically for this study, software has been developed that can record audio over the internet, so that, in the future, similar studies can be part of an on-line course. Although this delayed the project, we have gained valuable experience in programming such software and in using a Flash Media Server to capture audio. This tool will be made available to other projects at PSLC.

Explanation

This study is part of the Refinement and Fluency cluster. The studies in this cluster concern the design and organization of instructional activities to facilitate the acquisition, refinement, and fluent control of critical knowledge components. The overall hypothesis is that instruction that systematically reflects the complex features of targeted knowledge in relation to the learner’s existing knowledge leads to more robust learning than instruction that does not.

This study addressed the core issues of fluency from basics, scheduling of practice, and transfer. Students’ earlier knowledge of one specific grammatical construction is used to promote its transfer from comprehension to production skills by syntactic priming. This ‘syntactic priming’ can take place from comprehension to production because there is at least a partial overlap between the knowledge components and/or procedures involved in the skills of comprehension and production. Such priming through comprehension may enable students to use certain grammatical structure in production which they could not use correctly otherwise, or just not as often. Knowledge components and procedures related to the grammatical structure will be strengthened by syntactic priming, so that they are more readily available. This in turn will lead to higher accuracy and fluency.

This particular way of scheduling comprehension and production tasks is expected to lead to higher gains in accuracy compared to scheduling that does not involve syntactic priming. Fluency is expected to increase because a basic skill—one particular grammar point—has been practiced. Transfer is expected to take place from comprehension to production skills.

Annotated bibliography

-

Descendents

-

Further information

After a delay due to the unavailability of qualified transcribers, transcription of the data of both studies (pronouns, conditional) is currently underway (December 2008). Expected date of completion: March 2009.

Screen shots

Screen shot 1: Pronoun order listening task.

Screen shot 2: Pronoun order speaking task.

Screen shot 3: Conditional verbs listening task.

Screen shot 4: Conditional verbs speaking task.