Most test items attempted resolved in agreement with LRE resolution Table 23 (partially reproduced)

Post-test items corrected in agreement with LRE resolution

Items attempted

Items corrected in

agreement with

LRE resolution

Items corrected in agreement,

as a proportion of items attempted

Mean items per participant

Group (n = 30) 249 182 73.1% 6.1

One-to-one (n = 15) 160 124 77.5% 8.3

170

Of the test items attempted by participants, most of these – between 71% and 78%, depending on the mode – were resolved in agreement with the LRE resolution, that is, the correction made on the post-test was the same as the correction proposed during the PE task. This may constitute evidence that there is an association between LREs and learning, with new knowledge constructed (or existing knowledge consolidated) in the LRE surfacing again on the isomorphic task.

In the following extract, for example, Ofelia initiated an LRE regarding the greeting Hi and its informality, and with scaffolding from her teacher in the form of prompting and the provision of an L1 equivalent, was able to provide the correction Dear:

Ofelia First this Hi Teacher Hm

Ofelia Is like a bit informal

Teacher OK, what do you think would be better?... Ofelia I really don’t know how to make it better but,

Teacher Hm, if you write a letter or an email, usually, how do you begin?... Is there an expression in English like, a bit like estimado [dear]

Ofelia Ah like, Dear

Teacher Yeah, exactly, so you could change that for Dear Ofelia I wasn’t sure if it was too personal or not, I mean

Teacher Yeah, you can use Dear for, for er… yeah for a formal email, a formal letter, that’s fine

Ofelia OK…

In the post-test, Ofelia corrected “Hi” by writing “Dear”. This correction therefore appeared to relate to knowledge constructed in the episode, in which there was evidence that, while Ofelia had previously been aware of the item “Dear”, she had not been fully aware of its usage. In this way, it may be interpreted that the test was able to confirm that learning had occurred, and that, in Vygotskian terms, a concept that had previously been spontaneous had moved towards the scientific.

Also of interest were test items resolved in agreement with LRE resolution when the LRE had itself been incorrectly resolved. Georgina and Gema, for example, settled on the incorrect form “Dear Mister” as an alternative to “Hi”:

Georgina OK, well this is formal no?

Gema Yes is formal because it’s a student and he is, er, Andy, he is writing to, to Georgina To a university no? So this “Hi”… first, this is not right

171 Georgina I think Dear

Gema OK yes, Dear, Dear Mrs or Dear Mister

Tellingly, both participants then reproduced the incorrect form “Dear Mister” in the post-test. As in La Pierre’s (1994) study, it appeared that participants had retained the incorrect knowledge that had been constructed the previous week. Swain (1998) argues that retention of incorrect knowledge may be a greater indicator of learning associated with languaging than the reproduction of correct knowledge, since learners may have known the correct forms prior to the episode. As noted above, in the absence of a pre-test it is not always possible to determine whether episodes are the site of construction of new knowledge or the consolidation of existing knowledge, and so the presence of learning in an undesired direction is evidence that learning has in fact occurred.

Similarly, individual learner Isabella verbalised two episodes in quick succession, both of which were resolved incorrectly: firstly, she removed the “S” from “Mrs”, and then changed the comma after “Hi Mrs Horowitz” to a colon:

Isabella Yes, well, it seems to me that is very very, too er formal, er informal, this email, to a University is not the right, how to say, I think the right way to say to a person that her, erm Andy, I do not know this person, is very informal I think… grammar… vocabulary… OK, so… “Mrs Horowitz”, this is not correct, is without S, Mister, and the comma I think is not correct, it need two points…

Later in the same task, she chose the incorrect “I wait for your reply” as a way to sign off the letter:

Isabella: and “bye for now and see you soon”, of course no, in this letter is very informal maybe, please answer me soon, I wait for your er, er, reply, yes….

She then made these same three erroneous “corrections” on the post test. These responses were therefore marked as resolved in agreement with LRE resolution, although the resolutions were themselves incorrect. Again, the test responses may be interpreted as evidence that the LREs had been the site of construction or consolidation of knowledge, albeit erroneous.

No significant differences were found between modes in test items resolved in agreement with LRE resolution, mirroring findings from Nassaji & Tian (2010), who found that although pairs demonstrated greater accuracy than individuals when

172

completing cloze and text editing tasks seeded with phrasal verbs, there were no significant differences in learning gains. In this study, as in theirs, learning appeared to occur regardless of mode.

5.4.2.4 Few test items attempted resolved in disagreement with LRE