Doing MICASE-based Investigations (II): Starting with a category or grammar point

Part II. Starting with a functional category
John M. Swales (October 2005)

Starting with a lexical item (such as we have seen with CONCERN) is relatively straightforward, at least initially, because one can be fairly sure of capturing all the tokens in the MICASE database. Starting with a functional category, in contrast, means searching the grammatical and pragmatic literature as well as racking one’s brains in order to come up with a list of possible realizations.

Making a start

The category I have chosen to exemplify possible procedures is that of making suggestions. Now, of course, what ‘counts as’ a suggestion presents itself an immediate problem. In fact, it is often difficult to separate suggestions from hedged or polite directives, especially when these are uttered by those in a position of power, such as instructors (Leicher, personal communication). As a way in, as it were, I have opted for a broad category that includes suggestions themselves, giving advice, and, a bit more dubiously, making recommendations and uttering indirect directives. The general term I use for these is “the suggest complex”. So the first task is to assemble a list of potential candidates for these roles. Here are mine (using a stable constructed example in each case):

1. I suggest you drop the course
2. My suggestion is/would be to drop the course.
3. I’d/I would drop the course
4. I advise dropping the course
5. My advice would be to drop the course
6. If I were you, I would drop the course
7. Why don’t you drop the course?
8. Why not drop the course?
9. You might wanna/want to drop the course
10. You could drop the course
11. I recommend dropping the course
12. My recommendation is to drop the course
13. How about dropping the course?
14. What about dropping the course?

So we have 14 possible realizations, some using nouns, some lexical verbs, some modals, and some fixed expressions such as “how about..?”. (Getting this list together probably took me about two hours.)

Concordancing and quantification

The next stage involves scrutinizing the concordance lines in order to establish which and how many fall within what I have called the suggest complex. This is easier with a program such as Wordsmith Tools that has a “delete” function, but it can be done via manual coding on the website. Either way, such a procedure is going to be a time-consuming task; in fact, so time-consuming that I have decided not to bother with #3 (“I’d/I would”) because there are more than 1150 examples. (This decision may come back to haunt me.) In addition, the procedure I have outlined is also not going to be an exact science. Sometimes, we don’t have enough context to make a firm judgment; at others, there is room for reasonable doubt. So, we do the best we can. The results that follow contain: a) the target structure; b) the total number of tokens; c) the number of tokens that fall within the “suggest complex”; and d) the percentage of c) in terms of b).

a) target item	b) total #	c)suggest-complex	d) percentage
you could	796	193	24%
what about	209	12	6%
why don’t you/we	141	120	85%
how about	106	39	37%
you might wanna	104	104	100%
suggest	96	46	48%
suggestion/s	65	4	6%
why not	57	15	26%
recommend	40	24	60%
recommendation/s	23	2	9%
advice	17	3	18%
advise	5	2	40%
if I were you	2	2	100%

This took me at least eight hours, including an hour for the write-up. So let’s hope the effort was worth it! To test this out, we will first see what might be concluded from the table and from the examples I jotted down as I went along.

Commentary on the table

a) If we focus on Column c), there seem to be three common ways of making suggestions in MICASE: you could, why don’t you/we, you might wanna/want to. However, the 24% percentage rate for you could shows that it is highly multifunctional. Most of the time, in fact, it operates to discuss possibilities or general options:

1) you could allow harvesting beyond a certain size if that’s somehow correlated with age class
2) you could actually call it a conditional
3) you could ask that question, I guess

In contrast, the other two common formulae are much more closely associated with the suggest complex—entirely so with you might wanna, largely so with why don’t you/we. In the latter case, the exceptions are typically rhetorical questions, as in:

4) why don’t we get the stability at just one?

b) There is a considerable difference in the percentages (as shown in Column d)) of “suggest” uses for the two prefabs how about…? and what about…?. The great majority of the what about…? consist—as might be expected—of instructor-initiated questions. Suggestions are rarer, and largely fall into one of these three syntactic frames:

5) what about the two thirty slot on Tuesday? (NP)
6) yeah but what about getting a double degree? (V-ing)
7) yeah what about if if we change the order of presentations (if-clause)

In fact, a similar set of syntactic possibilities seems to be available for the more frequent use of how about..? as a way of making a suggestion:

8) how about a show of hands…
9) how about figuring out initial Q, again?
10) okay. how about if we start with Erin, and just go around the room…

c) A next observation invokes negative evidence—noticing what isn’t there (much). For example, we can notice that the nouns suggestion, recommendation and advice are very rarely used to make suggestions (4, 2 and 3 tokens respectively). Here is the complete list for advice:

11) my advice to you would be to, uh, try and read the stuff before lecture..
12) one of the most important pieces of advice I could give you…is to try to..
13) pick your fights hm that’s my best advice.

Although there are more examples of the verb suggest to make suggestions (46 in all), these are in fact under 20% of the total occurrences of the lemma. Overall, the data is pretty clear; these nouns and verbs are not used to make suggestions, but to either report them, as in:

14) also we suggested to her, that she look to local businesses..
15) Pat was suggesting a much more scientific mhm writing format

Or to draw tentative conclusions:

16) a constant friction suggests something closer to a solid than to a liquid

Pragmatic differences

With the frequency data nailed down, we can now turn to possible pragmatic differences. Consider first this pair:

A) why don’t you/you might wanna:

Look at these examples of why don’t you:

17) why don’t you grab a chair and join us?
18) why don’t you look it up, find out?
19) why don’t you zip through this introduction?
20) why don’t we take a five minute break?

As these examples show, this form of suggestion-making tends to be used when what is being suggested is simple, straightforward and not particularly onerous. As Adolphs notes in her PhD thesis, …”the immediacy of the suggestions using this prefab indicates that ‘short-term’ obvious problems are addressed rather than…long-term ‘troubles’” (p. 242).

Now consider the alternate you might wanna formula:

21) you might wanna talk about this difference…
22) and since you missed some stuff you might wanna do all the homework
23) you might wanna check with your T-A about this
24) uh I think you might want to probe a little deeper into this

As we can see, these tend to be somewhat more “imposing” and time-demanding. In consequence, a more polite form is adopted. The might wanna is already hedged, so a further hedge, such as in 24), is not that common. But here is an exceptional example (my emphases):

25) you might just wanna like again sort of draw out some conclusions from this a little bit.

B) why not/how about:

In a 2002 paper Svenja Adolphs examines these two prefabs and concludes:

“We find, for example, that ‘why not’ is mostly used to address some wider issue in the context of a discussion, with the aim of complaining or lamenting. ‘How about’, on the other hand, is used in suggestions directed towards an identified problem related to other participants in the conversation” (2002: 58).

Her data is restricted to relatively few examples and is based on British talk. So, how does this interesting distinction pan out in MICASE? Certainly, the 15 relevant examples of ‘why not’ take place in largely lecture contexts and are not related to individual problems of the interlocutors, as in:

26) so if you can have the cliffs along the beach and they are, undergoing,
this slow gradual change, why not organisms as well?…so somebody had
come up with this, this idea, and you’ve probably all heard of Lamarck,..

In fact, as many as seven of the 15 are strawman arguments setting up a later denial of the ‘why not’ proposition, as in:

27) …one could ask why not just a dump a lot of this stuff in your engines,
…um uh they will work, but one has to be careful because they work,
ultimately, by corroding the metal…

In contrast, the 39 ’how about" suggestions tend to occur in more interactive contexts, and when they do occur in lectures they tend, like examples 8 and 10, to be making suggestions about classroom procedures. In the more discussive contexts, they rarely deal—as we might expect—with personal issues and problems, but with suggestions about word choices, about calculations, or about other rather small-scale issues:

28) how about improving or how about um, enhancing, protecting
29) how about figuring out initial Q again.
30) how about if you couple it with glucose?

Distribution

A final issue that we can explore is the distribution of suggestions. If we just focus on the two formulae in the original list that are most predictive of suggestions (why don’t you/you might wanna), then there are eleven speech events where they occur more than 0.3 times per thousand words (using Wordsmith’s plot feature). These consist of :

• 3 office hours
• 2 advising sessions
• 2 small lectures
• 1 lab
• 1 meeting
• 1 seminar
• 1 service encounter

As this list shows, most of these fall towards the highly interactive end of the MICASE continuum.

Conclusions and implications

We have now reached the preliminary end of the story—a story that has taken some 20 hours to construct and to be revised following suggestions from the MICASE team. I have taken a fairly broad view of the suggest complex, but it would seem that the great majority of tokens have the overall function of trying to change a situation, often one raised in the immediately preceding discourse, and hopefully for the better. In the MICASE data, there are a limited number of formulae for doing this, and these rarely employ lexical items like suggest or advice that we might have expected. We have also seen that don’t you and you might wanna tend to be used in different contexts of imposition, and that why not and why don’t you also function in very different ways. All these are points that can usefully be brought to the attention of those who are attempting to acculturate to the spoken discourse in a US university.

Reference

Adolphs, S. 2002. Genre and spoken discourse: Probabilities and predictions,
Nottingham Linguistic Circular, 17: 47-60.