What breed is the Trojan Horse?

Comments on Lesley Milroy, "Variation as an Interactional Resource"

Penelope Eckert, Stanford University and Institute for Research on Learning

Lesley's examination of variation between /t/ and glottal stop explores an extremely promising line of research for variation. By moving to the level of conversational structure, this analysis moves out of the segmental arena and into the social, while sticking close to the structurally accountable ground. I'm also in very deep agreement about the possibility that tags are acting as a "trojan horse" for introducing glottal stop into pre-pause position. What I will explore in my comments is the nature of this horse. Specifically, I will explore an alternative interpretation of the interaction among glottal stop, tags, and pre-pause position; and ultimately I will raise some issues about the relation between gender and the linguistic behavior that this interaction produces.

To introduce this, I will begin with a picky point. There is, to my mind, some difficulty with assigning the glottal stop the status of "turn-delimitative signal" to the extent that one thinks of it as a signal ". . . that a speaker is prepared to yield the floor." (Milroy p. 7). I would be willing to accept that the glottal stop does delimit the end of a turn in some way, but not as a speaker-change signal, since by the time the last segment in the last word has been pronounced, it's already too late for any hearer to act on such a signal. This is a trivial remark on its own, I realize, but I want to turn attention away from the focus on the turn-taking aspects of pauses, and towards other things that are relevant to turn-final position.

Speakers are generally ready to relinquish the floor when they feel that they have completed the verbal task at hand*formulated the question, executed the command, delivered the compliment, made their point, etc. There are plenty of signals (e.g. Duncan 1972 and Local et al 1986) that are given well before the speaker has finished the turn, giving other participants the opportunity to prepare to take the floor. I would be inclined to think that such things as prepausal glottal stop have another turn-delimiting function, specifically to signal something about what the speaker has actually accomplished once the turn is over. Prepausal position is an important site for emphasis of various sorts, and the details of this case suggest to me that something of this sort is indeed going on here.

My own data on preadolescents and early adolescents in California show a similar use of glottal stop in prepause position, and it clearly signals "attitude" (in the sense "she's got an attitude."). Thus it isn't uncommon to hear the final segment in _fuck it_ replaced by a glottal stop. Glottal stop also tends to accompany other emphatic resources, such as assigning secondary stress to final unstressed syllables and the raising of schwa to [I]. All of these co-occurred, for example, in an example I have of a girl trying to convince her friends to wait for her, saying, "Wait - lemme get my jacket!" While I don't claim, or even believe, that glottaling in my data is anything like glottaling in Lesley's data, I can't help wondering if there isn't a similar or at least related interactional dynamic. This is particularly plausible to my mind, given the role of tags.

Lesley emphasizes that there is a very high rate of prepausal glottal use in tags, but there isn't enough information to figure out what the rate is like in other words. If pre-pausal glottal stop occurs overwhelmingly in tags, as it appears, then one might consider that the tag is as crucial as, or even more crucial than, prepausal position. The key to this question will lie in the nature of the tag uses that display glottal stops. And this ties in as well with the issue of gender. It has been well established (e.g. Cameron et al 1988, Holmes 1984) that gender differences in the use of tags lie not so much in the raw number (or rate) of tags used, as in the kind of use. Men have been found to make greater modal use of tags, while women have been found to make greater affective use of them. Tags can be used to request propositional confirmation, to elicit interaction, to seek agreement, to threaten, etc. Any coincidence of the use of glottal stop with some particular use or set of uses of tags would point to the nature of the interactive value of glottal stop.

Not only are the young women using more glottal stop than the rest of the population, but there is a significant difference in the use of glottal stop among the young women between the working and middle class. One might expect, if the use of glottal stop is spreading from medial to prepausal position, that there would be a constant relation across the population between the frequency of use of glottal stop in the two positions. Such a constant relation exists for all segments of the population except the younger working class women. If one combines both the medial and the prepausal occurrences of glottal stop, the young middle class women lead the young working class women by a significant margin (26% vs 15%) - a margin only somewhat smaller than the margin in medial uses only. (The same middle class lead occurs among the young men.) Given the significance of the numbers in both positions, one can't pass this off as noise. And one can't say that the use of glottal stop is simply moving from medial to prepausal position among women in general. Clearly the working class and the middle class women are doing something quite different with glottal stop, and the question is whether they've simply developed diverging rules, or whether they're actually doing something different interactionally. I'd be inclined to guess that interactional style is a good candidate for explanation of this difference, and that the use of tags is closely tied in with this style.

This possibility is also indicated by K's use of tags and glottal stop with her brother on the one hand and her friend on the other. Lesley notes that the percentage of glottal stop is far greater in the interaction with the friend than with the brother. My first intuition would be that K has a very different relationship with her brother and her friend, and - at least in this case - quite different kinds of interaction. But in addition, as Lesley points out, the difference appears to fall out from the difference in the rate of use of tags in the two interactions - most particularly the use of "and that." An examination of the function of "and that" may well point to the interactional crux of the matter. Another small indication of the relation between the use of glottal stop and the nature of the interaction - perhaps"attitude" - is in Lesley's note that two out of the four pre-pausal glottals produced by males occurs in "isn't it". I'm led to wonder if these are instances of the non-standard aggressive form [In I] documented by Cheshire (1981).

The line of explanation that Lesley explores in this paper is an exciting step towards filling the amazing void in the study of variation in use. As is my wont, I would like to push to use conversational structure not only as a source of explanation in itself, but as a platform for the exploration of meaning.


D. Cameron, F. McAlinden and K. O'Leary. 1988. Lakoff in context: The social and linguistic function of tag questions, Women in Their Speech Communities: New Perspectives on Language and Sex. Vol. ed. by J. Coates and D. Cameron, 74-93. London and New York: Longman.

J. Cheshire. 1981. Variation in the use of ain't in an urban British English dialect. Language in Society 10.365-81.

S. Duncan. 1972. Some signals and rules for taking speaking turns in conversations. Journal of personality and social psychology 23.283-93.

J. Holmes. 1984. Hedging your bets and sitting on the fence: Some evidence for hedges as support structures. Te Reo 27.47-62.

J. K. Local, J. Kelly and W. H. G. Wells. 1986. Towards a phonology of conversation: turn-taking in Tyneside English. Journal of Linguistics 22.411-37.