"Self-Repair" - from a careful look

Announcements of Self-Repair: “all i’m trying to say is, you’re under an illusion”
Stephanie Marx and John M. Swales

Introduction

When one speaks, “there is no going back and changing or restructuring our words as there is in writing” (Cook 1989: 115). However, although a previous utterance cannot be erased from history, further elaboration of it can often make it clearer or, indeed, alter the way it is interpreted. This is where self-repairs and self-reformulations enter the picture, providing speakers with the ability to change what he or she just said, at least to a certain extent. Typically, perhaps, we just try saying what we meant another time, but sometimes, we announce with an initial phrase that we are going to be trying again. These phrases, such as the one in the title, are the subject of this kibbitzer. Procedures

We began by brainstorming phrases that a speaker might use when he or she wanted to tell the interlocutors that an attempt to fix a speech mistake, clarify an idea, or rephrase an ambiguous utterance was coming up. These included items like “in other words”, “what I mean to say is”, which we then checked out on MICASE. We were careful to exclude comprehension check items like “you know what i mean?” While these are doubtless interesting, they are not announcements of self-repair. We also took note of Bruti’s findings (2004); although these were based on a corpus of biology textbooks, we were able to add to our list items like “namely”, “that is” and “e.g”. We also had a preliminary look at other-repairs, and found that announcements of these, such as “what you are trying to say is…” were really rare and could be ignored as a category. Quantitative Results

The total instances of each repair phrase found in MICASE are shown in the following table.

In other words	224
mean	50
trying to say	19
another way	18
that is to say	16
namely	15
i.e.	14
meant	11
what I’m saying is	7
clarify	4
rephrase	4
more specifically	2
misspoke	1

From the table it can be immediately seen that in other words is by far the most frequently used self-repair formula. This is followed by the verbs mean, say and put (it another way). Finally, we can note that the more lexicalized items such as clarify and rephrase are rare; indeed, paraphrase did not occur at all as an announced self-repair. Announced self repairs can be further broken down into those that are made immediately and those that are delayed, being brought up later in the conversation. However, only three instances of delayed self-repairs were found in the data (two using “meant” and one using “misspoke”), so separating them in the data seems a bit superfluous. Expansions or reductions?

The utterance following the announced self-repair could be longer than the first attempt, could be shorter, or could be about the same length. The following table and chart illustrate the totals calculated across all phrases in terms of expansions, reductions, and same-length reformulations. Expansions Reductions Other

197	119	43

Of all 359 instances of self-repair, expansions comprised about 55% of the instances. This percentage shows a not unexpected preference for people to follow announcements of self-repair with an elaboration of the original utterance, providing more rather than less information in the new statement. Reductions were employed 33% of the time, while in the remainder of the cases the repaired utterance was of approximately the same length as the original. We now give some examples: Expansions: 1. it is a call to personal and social transformation. in other words we can imagine and reconstruct various ways of creating sustainable relationships with each other, between partners between partners and children at different stages of our lives. 2. what clinicians call body angst, or i like this term better, bad body fever, alright, and i_ what i mean by that is a continuous internal dialogue with the self, about what’s wrong” 3. they’re called clades, and clades is another way of call- of saying monophyletic group. 4. you’re going from a sound here that’s labial i.e. involving articulators that are in the front of the articulatory apparatus 5. so when i said they were free i meant they were free relative to going out into the world and looking at each sensor value 6. uh and yet single women, uh that is to say w- uh, women who were never married or widows 7. we’ve got some kind of, thing that needs explanation here namely the fact that, every time we think of things we think of them in terms of time Reductions: 1 .men and women, have and they should have equal roles in both family care and work and public involvement, in other words both men and women should be, full human persons. 2. he’s throwing, kind of an, anchor off. notice that she bounces up and down. pulling the boat in, sways. no. kay so that’s what i mean by, boat-on-dry-land schtick 3. with anybody at least in a public situation i mean that, you sort of go out of your way to be polite and to not offend, um, and, and i suppose another way you can get at the same phenomenon is the ques- is in the notion of being honest. 4. the attitude that seems more modern to us than we would find in other thirteenth century thinkers that is to say, um, the, the, the utilitarian rule Similarities in length: 1. the labor movement which rose in the late nineteenth century in this country championed the middle-class white family ideal. in other words they tried to raise the male wage to the level, where the wife could become a full-time housewife. Conclusions

Announced self-repairs seems to be relatively infrequent—certainly when viewed against a corpus of 1.7 million words. Maybe most instructors prefer not to transparently admit that they haven’t immediately got their point across! And this reticence may also explain—at least in part—that in other words is by far the most frequent form of announced self-repair; like i.e. and namely, it seems a more neutral pre-paraphrase than something like the titular all I’m trying to say is..”. In the great majority of the formulations of announced self-repair, the expansions were more common than reductions; the one clear exception was mean. Why mean should be anomalous remains so far unclear. It is also a question for further study as to the relationship between version one and version two. Some times the direction seems non-technical→technical and sometimes the reverse. Other options are general→specific and abstract→example. Anybody like to follow up on our preliminary analysis? References

Bruti, S. (2004). Paraphrase types in the Pavia biology corpus: Some appositional constructions. In A. Partington, J. Morley, & L. Haarman (Eds.), Corpora and discourse (pp. 125-138). Bern: Peter Lang. Cook, G. (1989). Discourse. Oxford: Oxford University Press.