Interaction data in second language output (Dictogloss) tasks:

a new light on the identification, classification and value

of Language Related Episodes (LREs)

Alan Fortune, King's College London, and Dilys Thorp

Our paper builds on recent studies of second language learners engaged in different types of form-focused collaborative output task.  These studies include several articles which have appeared in Language Awareness journal, e.g. Swain and Kowal (1994), Storch (1997, 1998), and which contribute to the general literature on the role of output in SLA and how learners ënotice the holeí (Doughty and Williams, 1998) - gaps in their interlanguage systems - when pushed to produce L2 data.

In our study, like the earlier ones, we identify and classify Language Related Episodes as units for measuring the amount and type of form-focused interaction which takes place  We go further, however, in evaluating this analytical framework.  In applying it to our L2 data, we found more problems than previous studies have recognised.  Swain and Kowal identified grammatical, meaning-based and orthographic episodes.  Storch employed a slightly different framework.  We shall

offer an elaboration and refinement of their typologies.

Our study involves mixed nationality EFL students co-constructing Dictogloss (Wajnryb, 1991) texts.  Our data consists of transcriptions of the conversations of student triads engaged in the production of their texts.

In the paper we shall

(1) discuss difficulties of identifying and classifying LREs and  then present an amended typology.
(2) discuss the relative usefulness of quantitative and qualitative approaches to the data
and
(3) present data extracts for analysis and discussion.