Hi all,

I am currently writing my essay question on the IRF Model and it's really interesting. I recorded a one to one class of a high school student who has some great vocabulary but very poor grammar.

Anyway it was great to record the class and I finally took about 17 minutes of the recording time and transcribed it, giving me about 211 lines of exchanges. I have broken up the exchanges into their appropiate boxes (at least I think I have).

My tutor has given me some great advice about looking at things from my angle and I am looking forward to 'spotting' some interesting exchanges where I can take them out and really break them down. I have added a lot of 'extra' actions too which will also be a major point in my paper. I like the way I can pretty much do anything with the model as long as I can back it up!

Seems a bit of talk about it on the CELS mailing list so it's a good idea to put it up here and get some other opinions.

Looking forward to your comments,

David Moroney (Korea)

I did the same question as you as well. I had created a different kind of lesson, one in which it was an entirely student centered lesson. Once I had transcribed the lesson, and gone through the SC model, I realized it wasn't going to fit at all. But on p.6 of the original SC book, it specifically stated that the model was not designed for TBL student centered type lessons. The IRF model really doesn't explain either that the 'I' can be the student initiating the exchange.

The model isn't as rigid as you may think but what I ended up doing was the following;

I examined my transcript and fit into the IRF model what I could. There wasn't much.

Then what I did, was look at things that the model didn't allow for, such as embedding. I made that part of my paper.
I also examined the parts where there were more than one conversation happening at the same time.
Following that I decided, which gave me many sleepless nights, was to create a model and 'complement it' with the SC model.

I came up with a total of additional 6 acts, I redefined the exchanges and explained why I had done so. I came up with a different formula, not just IRF, but
I (R) (R-F)n (F) which means the center RF could be a number of exchanges between students students or students and the teacher. This also meant that I needed to redefine the way in which the data was presented. I also showed that the teacher feedback wasn't static and I inserted a floating header in which the teacher's questions or answers, could be anywhere.

Then I showed that the IRF model was just a series of display questions, (you'll need to bring in Long and Sato, or Coulthard for that) and that the student centered lesson with the transcript as it was eliminated display questions and made the classroom a more authentic learning experience for the students. Referential questions were about 98% of the lesson. Embedding was also allowed with this model as well as simultaneous conversations. The students could easily be the initiators and the could also be the R and the F, with the teacher coming in only for the purposes of scaffolding or clarifications.

I made sure that I stated that it was a complemetary model and the SC model was still the standard, because in reality, many classrooms still follow this model. (or the model is still a fit to many classrooms).

Mark (Japan)

After I had transcribed the lesson I realised that I don't offer a lot of feedback to the students, instead I usually ask another question after I say 'Hmm' or 'Ok' which is really interesting. Usually it was another question to keep the flow of the conversation.

I like your new formula Mark. I will have to look at mine more and see if there is a different formula for my class as well. Maybe [IR(n)]F where the n respresents the infinite number of times a conversation can take place before a teacher gives feedback, meaning that asking either display or referential questions were being asked and answered for long periods of time (in reality about 6 or 7 turns). Of course I could be way off but I am learing as I go along.

The student did pause a lot in class so I did mark those with a silent stress ^ and a pause denoting the amount of time in seconds that particular pause lasted ^[Pause 5].

I noticed too that my student and I used a lot of action in the class to get our point across so I will label them [Action] or something like that as the IRF model only has non-linguistic features in the form of directive (d) and react (rea). There was a lot of nodding heads in agreement and some laughter too, all which will be noted too.

It's truely amazing the amount of time one can spend studying a 12 minute conversation!

Regards,

David