hello chris , 

here are relevant publications :

http://eric.ed.gov/?id=ED037721

http://www.tandfonline.com/doi/abs/10.1207/S1532690XCI2103_02#.VhiLUeyqqko

http://www.erudit.org/revue/TTR/1995/v8/n1/037201ar.html

http://eric.ed.gov/?id=ED037721

Thanks Rahimi. I was really looking for studies which combine the think-aloud technique with learner corpora. This may prove difficult as learner corpora and introspective methods belong to different zeitgeists. However, your articles are potentially very relevant  to my research and will be read in due course.

pleasure chris , 

good luck with your study

a.

I know different publications about EA but not combined with Corpora,but you can see the publications of Hilary Nisy she has PhD in this field :)

I haven't heard of any, but have you looked at some of the very early work (1960s) on error analysis. They often catalogued errors with bits and pieces of real data (not quite corpora but the idea is similar). It is also possible that there were 'follow-up) studies of the (erroneous) forms that might have used psycholinguistic research methods. 

Excellent point, Martin. As prisoners of our time, we too often tend to ignore language research carried out in the pre-corpus era.

I suggest you to have a look at these:

Lado, R. (1957). Linguistics across cultures: Applied linguistics for language teachers. University of Michigan Press

Sinclair, J. 'The automatic analysis of corpora', in Svartvik, J. (ed.) Directions in Corpus Linguistics (Proceedings of Nobel Symposium 82). Berlin: Mouton de Gruyter. 1992.

They are dealing with the Corpus Based Approach and Contrastive Analysis Hypothesis- as Martin mentioned they are old but still great.

Hi Gaabor,

Sounds interesting. Do these papers use psycholinguistic research methods such as the think-aloud technique? I want to use learner corpora to find out  which errors students make with "some and any", and psycholinguistic methods to get inside their heads as it were, and try to find out why they make them.

Well, they use more like error analysis and predictions about mistakes, involving describing the languages, comparing them and predicting learning difficulties, plus some behavioristic approaches like "learning is a question of habit formation, and this could be reinforced or hindered by existing habits".

The basic concept was that the difficulty in mastering certain structures in a second language depends on the difference between the learners' mother language and the language they were trying to learn.

They also analyzed juxtapositions of language differences. But more or less the basis for these theories were more "grammar" and cognition oriented than psycholinguistic (sadly we are seriously lacking of empirical data from this field.)

Hi Chris,

You are right, to a certain extent. We tend to ignore research done in SLA some decades ago, to follow the theoretical trends of the moment. Yet, regarding Error Analysis and Contrastive Analysis, the fact that both rely, probably too much on behaviourism, seem to fail to provide more specific information not on the "what" but rather on the "how" and "why" of SLA acquisition and especially development. Neverhteless, I find your attempt interesting, since yõu intend to combine psycholinguistic methods, as I understand, to find out the possible reasons of why such "errors" happen. One of the issues that you should have in mind (allow me the piece of advice) is that the language learning process is not linear and, for that reason, variable. Sometimes highly variable. Hope this has contributed to you research. This is indeed a very important point, despite the decades that separate the initial phase of error analysis and contrastive analysis.

Hi Custodio,

I agree entirely that language learning is a non-linear process. Cognitive methods of grammar such as emergent grammar, pattern grammar and constructivist grammar all recognize this. A  Complex Systems approach to language acquisition may well help us to understand this process by showing the overlapping networks involved in language learning and the tendency for progress to involve peaks and troughs and returns to previous states before finally stabilizing. At the very least, it provides some useful metaphors for understanding the processes involved and some potentially interesting heuristics for future research. In this respect, I would suggest that you read Complex Systems and Applied Linguistics by Larsen-Freeman and Cameron (Oxford University Press). Chapters 5, 6 and 7 are particularly relevant, but you would probably need to read the earlier chapters to better grasp the issues involved. Of course you may already have read this book; it's hard to write on Research Gate without teaching someone's grandmother to suck eggs!

By the way, I am using the term error analysis in a broad sense to incorporate all studies, from whatever school of linguistic thought, that attempt to throw light on learners' errors. This definition would certainly include psycholinguistic studies, which, I quite agree, are preferable to methods which are based on behaviourism.

Complex Systems and Applied Linguistics by Larsen-Freeman and Cameron is indeed a nice book, I can recommend it too

hi chris ,

My papers on error analysis could offer some ideas and findings.

I was just looking back at some recent answers, and it struck me that your focus seems to be on errors. I think it is also true that learners are quite often correct as well as wrong with the same forms, and that, in order to understand their developing interlanguage, it is important to investigate both. Quite what kind of tests do you have in mind?

Hi Martin,

You are right. I will need to investigate correct as well as incorrect uses of my target forms-some and any. I have focused on errors so far in this thread, as the main purpose of combining learner corpora research with psycholinguisitc methods is to discover what students get wrong with these forms and why, with a view to  developing a new pedagogical approach to this area: the final part of my PhD will be a recommendation of how to write sample teaching materials for some and any. 

However, because of the possibilities that learners may avoid difficult uses, it is not sufficient to focus on what students get wrong. It is also necessary to  search for correct uses in learner corpora to determine which uses the learners do not need help with, and  to compare learner corpora with native speaker corpora to discover which uses if any students tend to avoid.

As you say, correct use of a form can also provide insights into learners' developing interlanguage. However, as I will not be carrying out longitudinal research, I will not get a complete idea of how learners' ability to use some and any develops over time.

  The main part of my research involves the use of reference corpora to discover the semantic and pragmatic meanings and collocational and colligational behaviour of the two forms.  After that, I intend to investigate learner corpora to see what help learners need with these forms. The results of the learner corpora research will then lead to some pyscholinguistic tests which will attempt to throw some light on the causes of learners' diffculty with these forms. For example, learners will be given dialogues/ written texts/contextualized sentences in which they have to insert some or any .

I will use introspective and/or retrospective think-aloud protocols to discover what is going through students' minds when they select the form. Introspective think-aloud protocols ask students to state what they are thinking while they are doing the task. The problem with them is that the act of verbalizing their thoughts may disrupt task performance, thus bringing into question the validity of the results: they may perhaps give a reliable measure of what students are thinking at this moment, but what they are thinking may not be a completely faithful reflection of how they normally choose between some and any.

Retrospective think-aloud protocols ask students to explain their thought processes after the task. Although the knowledge that they will be interviewed afterwards might affect the way some students approach the task,, they are likely to be less disruptive of task performance than introspective protocols. However, there are reasons to doubt both the reliability and the validity of the data obtained: the reliability is brought into question as students may forget some things that went throught their mind during the task and thus provide incomplete data; the validity of the data may be affected as learners may offer a posteriori reasoning which is not a true reflection of what they were thinking while they were doing the task.

Clearly, neither introspective nor retrospective think-aloud techniques are perfect!  If you can think of any other ways of discovering how learners choose between some and any and why they make certain errors, then let me know.

In any case, the research using learner corpora and psycholinguistic research will probably have to take place over a relatively short time period-perhaps three or four days at the most- and will therefore only provide, at best, snapshots of students' problems with these forms at a specific moment in time, rather than a true reflection of their interlanguage development. It may be possible to speculate about interlanguage development by comparing the performance of students at different proficiency levels but I am not sure if  I want to include such a speculative approach in my PhD.

In some ways, your study reminds me of Green, P. & Hecht, K. (1992). Implicit and explicit grammar: An empirical study. Applied Linguistics, 13, 168-184, but with a narrower range of target forms and a different approach to getting at explicit learner knowledge. I expect you have read it but just in case…

Hi again Martin,

No, I hadn't read it. Thanks for the pointer.