A Good Question Expressing Evaluation In Academic Speech

Mauranen, A. 2002. “A Good Question”. Expressing Evaluation in Academic Speech. In Cortese, G. and Riley, P. (eds)
Domain-specific English. Textual Practices across Communities and Classrooms. Frankfurt: Peter
Lang, 115-140.
"A good question." Expressing evaluation in academic speech
Anna Mauranen
University of Tampere
1. Introduction
Academic writing has been the subject of intense scholarly interest in the last two decades or so, but
interest in academic speech is a more recent turn. This situation is not much different from other
linguistic research, although the bias towards writing may be enhanced in academic contexts by the
special status of the written text in manifesting and securing academic achievements. However, it is
largely via spoken discourse that socialisation into the academic community takes place, and there
are many specific speech genres that characterise academic communities. As we are increasingly
becoming aware of the complex intertextual relations between written and spoken genres in the
academic realm, (e.g. Candlin et al., this volume, Mauranen and Markkanen 1994, Ventola 2000),
the need to investigate speech and writing both separately and together is also becoming more and
more obvious.
Incipient research into academic speech has focused mainly on lectures (e.g. Bamford
1999, Flowerdew and Miller 1997, Parpette 1999), but investigation of other speech events is also
under way (see, e.g. papers in Swales and Simpson 2001). Spoken discourse more generally is
being increasingly investigated with electronic corpora (e.g Aijmer 1996, Biber et al 1999,
McCarthy 1998), but so far there has been little academic speech data available. The Michigan
Corpus of Academic Spoken English (MICASE), under compilation in the English Language
Institute at the University of Michigan, is one of the very first attempts to make a large collection of
spoken discourse data available for research.
Academic talk can be viewed primarily as a variety of speech, or as a variety of
academic discourse. One of the basic questions the MICASE project set itself from the outset was to
explore the question in what respects is academic speech like academic writing, and in what
respects is it like spoken language in general. The first explorations (e.g. Lindemann and Mauranen
1999, Mauranen 2000, 2001, Simpson 2000, Swales 2000, Swales and Malczewski 2001) suggest
that academic speech differs in many ways from both. The present paper looks into an aspect of
academic oracy which can reasonably be expected to differ from academic writing, namely
evaluative expressions. As Hunston (e.g.1993, 1994) has shown, attitudinal language is virtually
absent from academic writing, but evaluation is nevertheless ubiquitous. In speech we might expect
this to be different, and find more directly attitudinal language in expressing evaluation. Evaluation
is also a key aspect of socialisation, and therefore exploring its expressions and functions is well
worthwhile. Apart from a general interest in the kind of language use and patterning involved, the
potential for applications in preparing L2 students to cope with evaluative language may be a useful
outcome from research of this kind. Hilkka Stotesbury’s work (1999 and this volume) on written
summaries also demonstrates how crucial and yet difficult it is for undergraduate writers to acquire
appropriate discourse strategies for evaluation.
This paper represents a preliminary attempt at capturing the use of evaluative
expressions and their functions in academic settings by using corpus data. Apart from some general
observations, the focus will be on a detailed look into some frequently found expressions and their
use.
2. Background

One of the ways in which academic talk, or university talk, differs quite drastically from
comparable writing is in not being much regulated by guidebooks. While textbooks constitute a
flourishing business covering a great variety of written academic genres, talk is mainly left to take
care of itself. Spoken academic genres constitute very much an oral tradition: we know how to run a
seminar by having participated in a few, and we learn how to participate in them by doing it. Some
instruction may be given initially, but an important proportion of this socialisation is acquired
tacitly, by observation and feedback. For example, in a study of university discourses I conducted
earlier in Finland and the U.K. (Mauranen 1994), many of the interviewed professors said that at the
beginning of their seminars they explained to the students what they expected in terms of seminar
behaviour and presentations. Some added that student presentations tended to get progressively
better in the course of a term’s seminar series because the later ones could profit from the example
of and the feedback on the earlier ones. Interestingly, this seemed to be largely lost on the students,
most of whom reported they had not been told what was expected of them, and therefore had to rely
on guesswork and observation to work out what this might be. Thus, the social learning that takes
place and the resulting knowledge appear to be transmitted in a subtle way, without very much
explicit teaching and perhaps even less conscious learning. It must be pointed out, though, that these
observations were made in an European context, and it is possible that more explicit didactic
practices might prevail elsewhere in the world.
A good deal of ordinary spoken interaction gives a very cooperative, constructive
appearance. From Gricean principles onwards and through politeness theory to, for instance, Eggins
and Slade’s (1997) analyses of casual conversations between non-intimate participants, the picture
remains fundamentally positive, and there is “an orientation towards consensus” as Eggins and
Slade (1997:21) put it. Academic writing, on the other hand, is usually characterised with much
more confrontational metaphors relating to a battlefield or competitive sports. Instead of consensus,
there seems to be an orientation towards conflict. So, for example, hedging is often seen as a
defence strategy in this battle: a way of securing the unassailability of claims (Meyer 1997) or
saving the faces of authors as well as their opponents (e.g. Myers 1989, Schroeder & Markkanen
1997, Varttala 1999 and this volume).
Academic speaking in a mainly didactic environment like a university might be
expected to be less like a battlefield than research writing, although discussions, colloquia, panels
etc. can also be debative, and thereby perhaps closer to academic writing than to casual
conversation among workmates. Human relations in university settings obtain between people who
are thrown together by no mutual choice, but who have a number of institutional goals to achieve in
collaboration. The institutional goals, again, are by and large tacitly accepted in that any
institutional position in a reputable research university, even that of a student, results from a
personal choice involving investment of time, effort, and often money. It is possible that consensus
and positive evaluation get emphasised in the university domain, much as in any other workplace
domain, and that the battlefield metaphors are less appropriate. Yet the well-known and widespread
academic tribalism need not be limited to the written medium. On the whole, it seems that germs of
both consensus and conflict live in academic talk. We might thus ask whether it is towards the
positive, consensual side that evaluative expressions tend in academic speech, or whether the
weight is on the critical, negative side.
Evaluation is a slippery notion for linguistic research because it cannot very readily be
allocated to any particular, easily definable set of expressions. It appears to be particularly elusive
in academic contexts, where openly attitudinal and emotive language is supposedly avoided in
favour of objective and factual discourse. However, Susan Hunston (e.g. 1993, 1994, 2000) has
shown in her extensive work on evaluation in academic research papers that evaluation permeates
even this seemingly non-evaluative discourse domain at every step. She defines evaluation as
“anything which indicates the writer’s attitude to the value of an entity in a text” (1993: 58). This is

a very comprehensive definition, including things like certainty and relevance, and it is clear that
such a broad notion is better suited as a basis for qualitative discourse analytical methods
(exemplified in Hunston’s own work and also in Stotesbury 1999) than corpus analysis. Therefore, I
would like to embrace a narrower (and to my mind a more basic) understanding, which is close to
Hunston’s “evaluation of value”; in this view, evaluation is taken to be that which indicates the
speaker’s attitude to an entity on a good-bad scale, and which bestows quality on that entity.
Perhaps one of the most basic features of evaluation is that it is non-propositional, i.e.
without truth value. But if we take Winter’s (1977, 1982) ideas (and see also Hoey 1983, Sinclair
1987) on board, we can say that whatever has evaluative function in the text (as in the structural or
‘clause relational’ pattern of situation – evaluation – basis, for example) is evaluative,
independently of whether it happens to contain evaluative language or not (and vice versa, for
instance the ‘situation’ unit can have evaluative language sprinkled onto it). While the recognition
of such discourse units as evaluative can be convincingly achieved in relation to the overall function
of a discourse or the values and ideologies of the domain the discourse is part of, it can hardly be
captured with corpus methods, and the concept is clearly different from that of ‘evaluative
expressions’. Evaluation of this kind might be called “structural”, and I would like to distinguish
between such discourse structural evaluation and “evaluative items” i.e. smaller units which confer
evaluative meanings to the entities they refer to or to other linguistic elements in the context they
occur in.
In addition, I would like to maintain a distinction between evaluative items and items
which tend to accompany them, such as hedges (perhaps, sort of) or intensifiers (highly, extremely),
connectors (but, nevertheless) or other expressions which indicate that an evaluation is going to
follow. This distinction is not a clearcut dichotomy but rather a cline, because individual items
which are not inherently evaluative may be used as parts of extended items which are evaluative.
Items may also be so regularly associated with negative or positive polarity that they acquire a
semantic prosody (Sinclair 1991, Louw 1993, Stubbs 1996), which carries over to contexts where
no other obvious indicator of positive or negative charge is found, so that verbs like happen or
cause, for example, have been shown to have negative prosodies. The distinction needs to be made
in principle, though, so as to keep the unnecessary spreading of evaluativeness in check; thus, for
instance but or at least are not taken to be evaluative in themselves, even though they often appear
in contexts where some evaluation is made.
A final distinction worth making between items which convey fairly unspecific
evaluative meaning, and which thereby appear to have evaluativeness itself as their primary
meaning, (wonderful, nice, great), and those which express specific evaluations, conveying a
particular evaluative meaning (successful, difficult). In this initial study, I focus on instances of
items which are both ‘inherently’ evaluative and unspecifically, or generically evaluative.
3. The data
The MICASE corpus (Simpson et al., 1999) has been compiled at the University of Michigan, and,
as already observed, it is one of the first attempts at gathering large quantities of academic speech,
and at the same time one of the comparatively rare corpora that have been compiled in the United
States. Interestingly, more or less simultaneously with the MICASE, another corpus of academic
speech was begun and completed in the USA. This is the spoken part of the T2K-SWAL corpus at
the University of Arizona, Flagstaff, compiled under the leadership of Douglas Biber (cf. Biber
2001). Both the MICASE and the T2K-SWAL have originated as a response to a practical need
observed in language testing: there has been so far little direct evidence on which to base test items
when assessing foreign students’ ability to cope with spoken academic discourse.
The MICASE has been compiled to cover a broad range of discourses within a
university setting, which could be called domain-specific. It is perhaps not quite appropriate to talk

about ‘genres’ here, because the discourses represented constitute a multitude of ‘event types’, from
field trips to lectures, some of which are much further established and conventionalised as genres
(e.g. lectures) than others (e.g. colloquia). Neither are the various discourses united by the
Hallidayan notion of ‘register’, since they comprise instances varying in formality (large lectures to
student study groups), the participants’ relations to one another (student-and- student, colleague-
and-colleague, professor-and-student, etc.). ‘Field’ is similarly inappropriate, since the topic areas
covered in the event types recorded vary widely (lectures in different disciplines, departmental
meetings, field trips, etc) . They could all be characterised as institutional discourses in a broad
sense, since those types of discourse deemed not institutionally relevant (e.g. buying coffee at a
campus coffee-shop) were deliberately excluded from the corpus at the compilation stage.
The MICASE project has clearly benefited from the ideas and research on written
genres by one of the project directors, John Swales, whose work has been highly influential in the
study of written genres. The general principles of genre research that he laid out in Genre Analysis
(1990) have largely constituted the theoretical basis for conceptualising academic spoken genres in
the university setting. Likewise, the in-depth analysis of the discourses that different kinds of
academic units engage in within a university (see, Swales 1998) have informed the understanding of
variety and scope in academic genres.
The MICASE corpus is thus based on a wide variety of event types that characterise a
university environment. In addition to the obvious lectures and seminars, the corpus covers thesis
defences, student presentations, group discussions, meetings, field trips and consultation hours etc.
The corpus has also sought to distinguish between primarily monologic and dialogic event types,
and strike a balance between the two. The T2K-SWAL has a broadly similar structure, although it
also contains a written section, and in consonance with the work otherwise done at Flagstaff (see,
e.g. Biber et al. 1998), takes the notions of register and registerial variation as its starting-points
rather than ‘genre’ or ‘event type’. The analyses on that corpus also tend to focus on register
variation in the Biberian spirit (see for example, Biber 1988, Biber et al. 1998). Developing corpus-
based teaching applications is also on the agenda of both corpus teams, and this aspect has been
strong also in Michael Barlow’s and his students’ work at Rice University (see, e.g. Barlow 1996
and Burdine 2001)
The MICASE at the time of the present research comprised about a million words,
with prospects of achieving its final target of 1.5 million soon afterwards. For a more precise
description of the corpus and the project, see Error! Reference source not found.).
4. Findings
4.1. Differences between speech and writing
As a first general search for evaluative expressions, I ran a quick comparison of the frequent lexical
items that differentiated between the 1-million-word MICASE corpus and part B of the
Microconcord corpus, of equal size. Although the corpora were not compiled on the same
principles, or at the same time, a very rough correspondence can be seen in that the Microconcord B
corpus was also academic in a broad rather than narrow sense, including fairly popular and didactic
texts (see, Murison-Bowie 1993). The comparison should give us some idea of the kinds of items
that are frequent in one or the other.
Among the most important differentiators were some in both corpora that seemed to
carry a good deal of evaluative meaning, although it is clear that each of these items has other than
evaluative uses, too, and that many of the items not included here may have evaluative uses in
particular contexts. Among the differentiators that were high on the list of the MICASE corpus, i.e.
those items that were significantly more frequent in this corpus, not very many evaluative
candidates were found. The most common ones indicating positive evaluation were right,
interesting, cool, good, nice, and fun. The interjection wow might perhaps also be included in the

group. Of these, right was very frequently used as a minimal response or discourse marker with
little evaluative meaning. Similar candidates like okay and well were also mainly used for non-
evaluative functions. Well showed almost 300 evaluative uses, but this accounted for less than a
tenth of all its occurrences, and thus these uses were not likely to contribute significantly to its
discriminating power. Differentiators with a negative charge were even rarer, consisting of only
illusion and noisy.
In contrast, the written corpus had a number of differentiating items that looked
evaluative. The most frequent with a positive sense were sensible, strength, advantages, vital,
equitable, successful, convenient, important, sufficient, and suitable. On the negative side were
weak, objection, unlawful, reckless, doubts, severe and difficult.
In both corpora, the number of positively charged items among the top 200
differentiators was higher, but the difference was much greater in the speech data (7 positive vs 2
negative, as opposed to 10 vs 7 in the written corpus). Another clear difference between the corpora
was that while the evaluators in the spoken corpus tended to be quite unspecific in meaning, the
reverse was true of the written data. In this respect the MICASE data seems to conform to generally
accepted notions about spoken language, and the items resemble those identified as typical of
conversation rather than academic prose in the Longman Grammar (Biber et al. 1999). The
relatively great variety of specific evaluators in the written data is in accordance with the wide
range of evaluatively used lexis that can be seen in Hunston’s (e.g. 1993) and Stotesbury’s (e.g.
1999) studies on written academic texts. Since items of this kind here differentiate between the
spoken and the written data, it is not surprising that items suggested by this earlier research were
typically either not found at all, or only very rarely, in the spoken corpus.
4.2. Intensifiers and mitigators
To get an idea of the strength or intensity of positive and negative evaluation, I looked at the most
frequent intensifiers in the MICASE corpus, with the purpose of seeing how they tend to combine
with positive and negative items. Clustering around the commonest intensifiers very and really
shows again a bias towards positive items. Of the 150 most frequent three-word clusters around
very, 22 were clearly positive (including combinations with good, interesting, important, nice), as
compared to five negative ones (e.g. difficult, hard, not good). The repeated very (very very) was
also fairly common (89), and it was similarly used to emphasise positive (35) more often than
negative (11) evaluations. Really is harder to assess from a cluster list than very, since it has several
uses, but taking only evaluative adjective modifiers into account, clusters including really + positive
adjective (good, interesting, important, cool…) were 26, as opposed to 6 where the adjective was
negative (difficult, hard, bad). A third frequent intensifier, highly, was similarly biased towards
positive adjectives.
If intensifiers seemed to favour positive evaluation, one might expect mitigators, or
hedges to behave differently. However, they tended to show varying preferences. Among the most
common ones, just in its typical use was not connected with evaluations, but with metadiscursive
expressions (I just wanted to point out…, let me just rephrase this…, if we could just move on to the
second issue…). Another of the most frequent ones was probably, which combined primarily with
various modal expressions, without a clear preference for either negative or positive evaluation.
Slightly showed its clearest pattern with different, and the other combinations were either concrete
(slightly acidic) or then negative (slightly ashamed of myself), without negative associations being
particularly prominent. Of those mitigators which tended to combine with evaluative expressions,
somewhat was the most clearly negative. Of the 61 instances of somewhat in a million words, half
(30) appeared in negative contexts, most of the rest being neutral with only very few positive
instances. The examples in (1) illustrate its typical use:

(1)
hat, of that in it and for this it gets a bit represented in a somewhat twisted way. so here's the, the
is, since we can't always get an equilibria i defined a somewhat weaker notion on prices called the
the industrial revolution once it's done. um, and so that's somewhat unfortunate that means, in in
away feeling that, either you, you had felt your, felt him a somewhat diminished figure, or, you
rals and the consequences, or if you think that they're, somewhat arbitrary and decided by, each
nce that drug policies affect only people of color is, somewhat irrelevant, we should punish
i guess i'm_ first i would have to say that i actually am somewhat dubious on the first assumption,
aditional, that for boys, already seeing the world in a somewhat sexist way predicts seeking out,
ciplines, seem to share very little granted these are somewhat stereotypic uh notions or definitions,
Another case with a fair bit of negativity was the very common hedge a little bit. Although it
seemed to be predominantly negative in collections of general spoken British English (The Bank of
English and the BNC sampler), it was not equally negative in the MICASE data: out of a sample of
200, only 34 were clearly negative, as in the following examples:
(2)
uh, uh this week's homework was a little bit difficult, to finish with the project
abilitation when it lumps everything together that's a little bit unfair to the drug treatment stuff
a think about this pretty broady and it does get a little bit confusing so there's lots of non-food pro
t it should be. but, as far as i could tell, it was a little bit vague. um, but anyway, the problem see
the, heads of the state. i mean these are, this is a little bit intimidating. but, they present it themsel
if you were told to list 'em it would probably be a little bit hard on your own. the first of course is y
are, these ones are okay right? these probably a little bit tricky. what is capital share? remember
all out. agreed. that might be a- a little bit unfair but i don't know. but t
i'll i'll get to that le- let's let's just kind of leave that a little bit vague right now we'll we'll say there
so in a way that desire to make it urban i think is, a little bit misplaced do you think (xx)?
But an even more recurrent use of a little bit was in connection with metadiscourse (or discourse
reflexivity), where the typical (62 out of 200), use was very similar to that of just ( I'd like to talk to
you a little bit about, how…). It was also used in concrete meanings, where it is basically neutral
(it's a little bit warmer in this room). In sum, it was split between concrete, neutral meanings, the
mitigation of the imposition involved in using metadiscourse (cf Mauranen 2000), and negative
contexts.
In all, mitigators did not show a distinct preference for negative evaluators, although
they did appear to avoid clearly positive contexts. On the whole, positive evaluators emerged much
more readily from the data and combined more often with intensifiers. It seems, then, that positive
evaluation tends to be made explicit and emphatic in the domain of academic speaking, at least as it
is represented in the current data.
4.3. The case of good
For a closer look into a typical case, let us consider one which is positive, frequent, and inherently
but unspecifically evaluative, namely good. It is one of the most frequent (N= 1387) evaluators in
the corpus. Of course it has other common functions as well, such as that of a boundary marker, or
minimal response token as in
-…but that's great. okay. good.
To come to grips with common patterning of the node, I first looked into the three-word clusters
most frequently associated with good. The overwhelmingly most common was that’s a good (73).
Within this frame, the two most often repeated nominal phrases turned out to be a good point and a

good question. Narrowing the query down a little bit further, to that’s a in the context of good, it is
seen that that’s a +(intensifier)+ good point accounts for about a quarter (21/79) of all the instances,
and that’s a + (intensifier) + good question for nearly a fifth (15/79). Between them they account
for 46% of the cases, thus clearly dominating the expressive potential of this frame. Other nouns
were repeated less, so that for instance that’s a good idea was thrown up 8 times, that’s a good
example three times, that’s a good answer twice, and there were also some single occurrences, like
that’s a very good suggestion and that’s a good observation. It is interesting to note that all of these
expressions actually appear to be metadiscursive, or discourse reflexive, that is, they refer to the
ongoing discourse. Of the three -word clusters associated with good which were repeated more
than 10 times (N= 30), 6 included a metadiscursive noun (question, point, example…), twice as
many as for example those including a general noun (stuff, thing, job). A typical case which
combines the metadiscursive use of good with the frequent that’s frame would then be that’s a good
question , which indeed seems a fairly fixed unit. The adjective good can itself be modified, but the
range is limited: all we get is that’s a (very/really) good question.
4.4. Other modifiers of question
If questions can be good or very good, can they also be wonderful, marvellous or
splendid, for example? That is, in what other ways can questions be evaluated? There was one case
of That’s a wonderful question to ask a search committee , but no questions were ‘marvellous’ or
‘splendid’. There were altogether 687 instances of question in the corpus, but among these only two
evaluative adjectives, good and interesting, were repeated with any notable frequency. Good was
over three times as frequent as the next one, covering nearly half of all the cases on its own (see
Table 1). Few of the adjectives modifying question were negative: of the 25 different types, 15 can
be seen as positive, and 8 as negative, although there is obviously a good deal of variation in the
degree of negativeness/ positiveness. If the tokens are counted, positiveness dominates even more
clearly, 64 vs 9.
Good 31
Interesting 11
Great 4
Important 3
Difficult 2
Minor 2
Big 2
Bad
Bigger
Burning
Crucial
Distracting
Easier
Fundamental
Hard
Key
Million dollar
Positive
Puzzling
Right
Silly
Simple
Strange

Valid
Wonderful
Σ 74
Table 1. Evaluative adjectives modifying question
Exploring the use of evaluated question in context, it is useful to make a distinction between
‘monologic’ uses on the one hand, where the current speaker puts forth and evaluates a question as
important, central, basic, etc., as in (3):
(3) …so that, that, it seems to me that a a much more, important question isn't the
question is there something which is, objectively speaking red and spherical when I
imagine a red ball, but rather is the experience…
On the other hand ‘dialogic’ use involves explicit interaction with interlocutors, either by eliciting a
response as in (4)
(4) okay now the_here is the million dollar question, well how do I pull the hydraulic
grade line out of this? what do you think?
or responding to another participant (5):
(5) A: doesn't that just mean we have to do more samples?
B: well, that's a good question. let's think of some of the assumptions now. first of
all um,
Of these types, evaluation + question was most commonly used in dialogic responding, and least
often in dialogic eliciting (Table 2):
dialogic responding N = 42 ( good 29, interesting 6, great 3)
monologic N = 22 (interesting 5, important 3, difficult 2,
good 2, big(ger) 2)
dialogic eliciting N = 10 ( minor 2)
Table 2. Repetitions of evaluative adjective + question in monologic and dialogic use.
(Single occurrences are included in the total numbers, but only repeated ones listed.)
As can be seen, the preferred adjectives were different in the different types of speech acts, good
being typical of dialogic responding. It appeared characteristically in the unit
(That’s a) + (intensifier) + good question,
where the intensifier is limited to very, real, and really.
As part of a dialogic response, (that’s a ) good question is used mostly as a preface to an answer.
That is, the speaker answers a question that has been asked, but begins by making an evaluative
comment. This was the case in 20 of the 29 instances, as for example 6 illustrates:
(6) A: I have a question in the, proletarian version of this section, do_ (I think) it says um,
it_ the movement of the proletarians was the only majority, revolution? but what
about the Third Estate I'm confused.
B: oh okay good question um... the, he's he's really talking about, by by the ti-, he
wrote the Communist Manifesto in eighteen forty-eight. and, the French revolution,
happened in seventeen eighty-nine. and so by the time he's writing, he's assuming that,
the Third Estate which, which was the bourgeoisie, has won…
Sometimes the speaker indicates that an answer is forthcoming (7), even if it is for some reason
delayed
(7) good question I_I'll have to_ I'll email um one of the authors of this article…

Another response type of some regularity (5 cases) was one where the speaker passes on the
question to the other participants, implying at the same time that he or she knows the answer (8) :
(8) A: y- you said we had a Q, but how do we get Q out of this?
B: good question. how do you get a Q out of this, this thing?... how would you set up
your sets of equations so that you get a Q out of this?
4.5. That’s a good question
As to the meaning of that’s a good question, a distinction can be drawn between contexts where an
answer is known or available, and those where the speaker implies that an answer does not exist.
The former was by far the commoner case, as one might expect in an educational context, and in
such instances, the implication seems to be that the goodness of the question is conferred by the
usefulness of the answer. The speaker often appears to mean the answer is useful for the hearers in
particular, as in (9), from a lab session, where the answer clarifies the task to the student who asks
the question.
(9) A: so you want the discharge in each pipe section?
B: very good question, yeah, like cuz the system is like three sections right? so you
have three discharges which one do I wanna see? well I would like to see all of
course, so but if you give me just- if you give me the discharge in branch one and
discharge in branch two, that should suffice because the sum of 'em is going to be the
discharge in branch three, right?
Many questions and answers were less instrumental than that in (9), and the meaning of good
question seems more geared to the content matter, or the intellectual value of the answer (6 above,
and 10 and 11), even in cases where the topic is entirely a matter of speculation (11). Occasionally
it also seemed that the speaker simply liked to be given the opportunity to speak about a given
topic.
(10) A: I'm confused then why does he keep making references to a priori and necessary
being the same thing?
B: oh I'm sorry I see what_ ah I see what you mean. yes, yes, it turns out tha-
well, that's a really good question it turns out that one of the ways we can know it is
through reason, and that's generally thought to be the key way, and things that you
know through reason are necessary. because the only way to know something through
reason, is to find out that, its opposite was a contradiction...
(11) A: would they have done the same thing, for, say, (an) inner city (urban)?
B: I wonder that's a very good question. um, I wonder if it'd be harder for them to do,
because it's um, because, what they were dealing with was a traditional sort of folk
practice, versus modern medicine, as opposed to uh people who were classified as
underprivileged or deprived in some way. um, it was it was (jus-) it was uh, it was
technology over, native practice, it didn't have to do so much with access to scarce
resources, in in a different sort of way. so I don't think the Canadian government has
done the same sorts of things, for people who, wh- where the issue is poverty, so much
as, as technology. technology was really what was driving it
The situation where the speaker expresses approval of a question while implying that an answer
does not exist, good question seems to mean ‘although there is no answer, there ought to be one’
(12), that is, the issue ought to be noticed and discussed. This meaning seems to be there, for
instance, in meetings when the speaker does not give an answer.

(12) A: …and that's what I that's what I struggle with. how do you work with actually
feminism in the in the work?
B: that that's a very good question. that's the ultimate question, and I don't think
we're there yet in engineering and the physical sciences. it's easier in my discipline of
It seems that by evaluating a question as ‘good’, the speaker not only shows approval of the
question, and presumably the person posing it, basically implying that the answer is worth knowing
or discussing, but he or she also appears to accept responsibility for providing an answer. In terms
of discourse structure, that’s a good question thus strongly predicts an answer move. In this data it
was not followed by other moves which might conceivably follow an evaluation, such as a
justification or basis for the evaluation, or an immediate move to a different topic, thus ending the
exchange.
4.6. An interesting question
In addition to good, the other common evaluator combining with question was interesting, which
also appeared in monologic as well as dialogic use. There were only 11 instances, but it might be of
interest to see to what extent an interesting question is like a good question. It turns out, though,
that out of these 11 cases 5 come from the same session, an office hour where questions to be
discussed in an essay or report are sought. The rest were dispersed over different activity types, like
lectures, thesis defenses, and student presentations. On the whole, interesting was used attributively
(with only one predicative case), and no preference was detected for dialogic responding (N = 6)
over monologic use (N = 5).
The immediately obvious difference between an interesting question and a good
question is that an interesting question usually (in 9 cases out of 11) does not imply that the
speaker is going to provide an answer, or take up the topic. In monologic contexts, the speaker
comments on his or her own ideas, or an aspect that has occurred to him or her, but which is not
going to be developed (13), whereas in dialogic contexts the implication seems to be that the
addressee might take up the issue (14).
(13) I'm not going to go into the issue here of the regularity of sound change, and (a- why)
some a- wh- a- items of structure uh show this change and others don't all right that
uh, is uh that will although it's an interesting question, and maybe one we could
discuss within the framework of this class it probably is not relevant, to the you know
the uh the relationship of let's say language and and history. anyway.
(14) A: so if you put the body on life support someplace else?
B: well yeah,I guess you could do that
C: yeah, how about that?
B: but his is more trouble than it's wo- no,
D: I don't know, that poses an interesting question...
A: so it's not that the, that the argument he has given against the brain theory is
actually, a good one. the argument is where do you …
The five cases in the office hour sessions were comments on a paper draft, generally pointing out
something that was interesting, followed by a suggestion of what to do with it in the paper to be
written, as in (15):
(15) um, and this is a separate paragraph now, you can't give me an answer to this
question. it is an interesting question. um, you might wanna put it in a footnote
instead. do you know how to use footnotes?

The lack of intention to answer the interesting question was sometimes made explicit, as in 16, from
the same office hour session, where the speaker speculates over an issue resulting from his/her own
associations with a point in the text under discussion.
(16) um... so it's an interesting question it's, it very possibly could be that um, these are
you know, displaced workers in some sense. um, and that um... you know displaced
from I don't know what but uh... and they've had to take significant, pay and status
cuts working at the mall. so I mean that's just an interesting_ to think about i- you
can't probably get the answer to it but, um...
As can be seen from the examples, the fundamental difference between a good question and an
interesting question is that they are used in different senses of ‘question’. A good question, in its
characteristic use of a dialogic response, refers to an actual interrogative speech act that has
preceded in the discourse. It is therefore typically discourse reflexive. An interesting question, on
the other hand, means ‘question’ in the sense of ‘issue’, i.e. a point that can be contemplated or
discussed, but it does not necessarily (in this data typically not) refer to a question that has been
asked. It is therefore referential, not discourse reflexive.
4.7. Not so good questions
To finish this discussion of good and question, I would very briefly like to comment on the
negatively evaluative adjectives modifying question. In contrast to the positive evaluators, these
tend to occur in dialogic eliciting. Most of the instances (6 out of 9) prefaced the speaker’s own
question, as in 17
(17) this is a minor question. when you look at the ANOVA, um I think when we did it, the
lab in class you said there was …
Only two were not metadiscursive, referring to a strange or hard question in the referential sense of
‘issue’ So the negative evaluators referred modestly to the speaker’s own question, with a
somewhat disclaiming overtone. Since the data represents encounters between non-intimates, a
direct negative evaluation of an interlocutor’s question might not have been expected, either.
As to the negation of good, it does not work symmetrically as a simple reversal of the
positive good. The negated good not only is less frequent, but the patterning seems to be limited to
sequences like not so good, not very good. In addition, the negated good does not combine with
items like point or question. Expressions with NOT + good do not get used in the dialogical mode
usually either; they are typically monologic, divided fairly evenly between personal evaluations
(cuz I'm not good at physics, I think she might not be a very good undergraduate ) and academic,
didactic evaluations of aspects of the content matter, including the argument itself, as in (18):
(18)
receptors, found postsynaptically. and I haven't actually given you good evidence of that yet but
that's why we're getting this result (out). not because the test is not good, but because, it's a very ra
rly-burly, of the real world. so they typically tend to be not as good competitors, they tend to be
you know check for inactivity but the point is, that's not a good measure of global quiescence.
wait le- let me give you an example si-, it's not a really good analogy but it's sort of close, like
so in other words the va- the vast majority of religious people are not good subjects for the study of
psychology of religion
As can be seen, the academic use passes on not only factual information ( tend to be not as good
competitors) , but also assessment of methods ( that's not a good measure) and argumentation (I
haven't actually given you good evidence). These uses clearly reflect the institutional goals of the
university as related to academic content. For students, the evaluative assessment is obviously as
important as the ‘pure’ content matter itself, organising and structuring the content into meaningful
structures and hierarchies.

5. Conclusion
The main points from the foregoing can be roughly classified under two headings: those of
primarily linguistic interest, and those with more social relevance. To begin with the more narrowly
language-related issues, the first observation of interest was that the modifiers ‘good’ and
‘interesting’ selected different senses of ‘question’, good preferring the discourse reflexive one, and
interesting the referential, or propositional meaning.. While this needs to be confirmed with more
extensive data, it points to one of those fairly specific form – function correlations that are usually
only accessible to observation through corpus data, although many of these are recognisable to
speakers once the correlation has been pointed out to them.
The second linguistic observation was the emergence of that’s a good question as a
prefabricated unit with a fairly specific meaning. The patterning around good was fairly clearcut,
being dominated by two very similar patterns, (that’s a) good question and (that’s a) good point.
(That’s a) good question appears a relatively fixed unit, which apart from its obvious evaluative
meaning carries a pragmatic meaning implying speaker's acceptance of the question and
willingness to assume responsibility for providing an answer. One could say it is a ‘pragmatised’
unit, in that the pragmatic meaning is specific to this unit, whereas, for instance, its opposite, with
negated good, does not carry a reverse pragmatic meaning, but is interpretable in its transparent (or
‘literal’) semantic meaning only.
A third point worth noting is that a good deal of the evaluative commentary was
discourse reflexive, or metadiscursive, i.e. the ongoing discourse itself is being evaluated quite
frequently. This could be seen as a way of helping organise knowledge structures for interlocutors,
and of constructing hierarchies and patterns of salience in the referential content being discussed.
Finally, in terms of discourse structure, the preface use of that’s a good question played a
prospective role, not one of closing a sequence, which has usually been thought to be the main
structural slot for evaluation (e.g. Winter 1977, Hoey 1983).
Of the more socially relevant aspects of the findings above, it is perhaps the
dominance of explicit and emphatic positiveness that was the most striking. Not only were there
more positive than negative evaluative items, but positive items also occurred more often modified
by intensifiers. The general tone of academic speech thus seems to be oriented towards a consensus
rather than conflict, which is in accordance with research on casual conversation in a workplace
(Eggins & Slade 1997). It can also be noted in passing that the most commonly occurring evaluative
adjectives reported in the Longman Grammar (Biber et al., 1999:512) as found in both conversation
and academic prose are also practically all positive. This orientation towards consensus raises the
question of how and where are the conflicting and competing discourses found in academic writing
acquired.
The next point worth attention is the noticeable frequency of evaluative responses,
suggesting explicit socialisation, both as feedback and as demonstration or illustration. This goes
somewhat against earlier views. Secondary socialisation into academic rhetoric has often been seen
as problematic, largely on account of its inexplicitness. Such claims have usually concerned the
acquisition of written discourse and rhetoric, but speech should come into it as equally (if not more)
important, as a means of transmitting the tacit knowledge that members of the academic community
share. As e.g. Väliverronen (1992) points out, part of being an academic entails being able to
present strong arguments, ask appropriate questions, make sharp points, and other similar things; he
claims that we do not receive instruction in this, but are left to our own devices to find out how to
accomplish it. A similar view can be read into students’ interview responses (Mauranen, 1994)
when they were talking about learning to participate in seminar discussions. However, the current
findings seem to suggest the contrary, and point to fairly explicit socialisation. Apart from the
interpersonal good question, providing feedback on individual speech acts in ongoing discourse, an

interesting question, by picking out aspects of the propositional content, is a potentially powerful
means of extracting elements for special attention within the conceptual (and in a sense ideological)
framework that is being transmitted in and through the system of discourses. Of course, a certain
explicitness in evaluation and socialisation conforms with the requirements of an educational
context as opposed to the research community: the values that we can expect to be shared in the
research community are being communicated rather than presumed, and evaluation perhaps cannot
be too implicit.
A more problematic matter is the relative scarcity of explicit and direct negative
evaluation, which seems to make the negative much less accessible to discourse participants. If it is
the case that socialisation into the academic community takes place via explicit positive evaluation
but more implicit or indirect negative evaluation, then the hard task for the novice is to discover
what he or she is doing wrong. This may be another major contrast between learning academic
discourses via speech as opposed to writing: in written comments, what is being done right is often
harder to detect.
References
Aijmer, K. 1996. Conversational Routines in Spoken Discourse. London: Longman.
Bamford, J. 1999. ‘Dialogic aspects of academic lectures’. Paper given at the IADA 99 Conference
“Working with Dialogue”, Birmingham, 8-10.April 1999.
Barlow, M. 1996. ‘Corpora for Theory and Practice’. International Journal of Corpus Linguistics
1.1, 1-37.
Biber, D. 1988. Variation Across Speech and Writing. Cambridge: Cambridge University Press.
Biber, D. 2001. ‘Dimension of Variation Among University Registers: An Analysis Based on the
T2K-SWAL Corpus’. Paper given at the Third North American Symposium on Corpus Linguistics
and Language Teaching, Boston, March 23 –25, 2001.
Biber, D., S. Conrad and R. Reppen 1998. Corpus Linguistics. Investigating Language Structure
and Use. Cambridge: Cambridge University Press.
Biber, D., S. Johansson, G. Leech, S. Conrad & E. Finegan (1999) Longman Grammar of Spoken
and Written English. London: Longman.
Burdine, S. 2001. The Lexical Phrase as Pedagogical Tool: Teaching Disagreement Strategies in
ESL. In Simpson, R.C, and J.M. Swales (eds), Corpus Linguistics in North America. Michigan
University Press, 195-210.
Eggins, S. and Slade, D. 1997. Analysing Casual Conversation. London: Cassell.
Flowerdew, J. & Miller, L. 1997. The Teaching of Academic Listening Comprehension and the
Question of Authenticity. English for Specific Purposes 26/1:27-46.
Hoey, M. 1983. On the Surface of Discourse. London: George Allen & Unwin.

Hunston, S. 1993. Evaluation and ideology in scientific writing. In Ghadessy, M. (ed.) Register
Analysis. Theory and Practice. London: Pinter Publishers, 57-73.
Hunston, S. 1994. Evaluation and organization in a sample of written academic discourse. In
Coulthard, M. (ed) Advances in Written Text Analysis. London: Routledge, 191-218.
Hunston, S. & Thompson, G. (eds) 2000 Evaluation in Language Oxford: O.U.P.
Lindemann, S. & Mauranen, A. 1999. "It's just real messy: The occurrence and function of ‘just’ in
a corpus”. Presentation at AILA 99, Tokyo.
Louw, B. 1993. Irony in the Text or Insincerity in the Writer – The Diagnostic Potential of
Semantic Prosodies. In Baker, M., Francis, G. & Tognini-Bonelli, E., (eds), Text and Technology.
In Honour of John Sinclair. Amsterdam & Philadelphia: John Benjamins, 157-176.
Mauranen, A. 1994. Two Discourse Worlds: Study Genres in Britain and Finland. In Mauranen, A.
& Markkanen, R.(eds) 1994. Students Abroad. Aspects of exchange students' language. Special
issue of Finlance, Vol XIII, 1-40.
Mauranen, A. 2000. ‘They’re a little bit different… Observations on hedges in academic talk’. Paper
given at the Second North American Symposium on Corpus Linguistics and Language Teaching,
Northern Arizona University, Flagstaff, 31.3- 2.4.2000
Mauranen, A. 2001. Reflexive academic talk: Observations from the MICASE corpus. In In R,C.
Simpson & J. M. Swales (eds) Corpus Linguistics in North America: Current States of the Art.
University of Michigan Press, 165 -178.
McCarthy, M. 1998. Spoken Language and Applied Linguistics. Cambridge: C.U.P.
MICASE home page. Error! Reference source not found.
Meyer, P. 1997. Hedging strategies in written academic discourse: Strengthening the argument by
weakening the claim. In Schroeder, H. & Markkanen, R. (eds) Hedging and Discourse. Berlin:
Walter de Gruyter, 21-41.
Murison-Bowie, S. 1993. MicroConcord Manual. Oxford: OUP
Myers, G. 1989 "The Pragmatics of Politeness in Scientific Articles", Applied
Linguistics, 10/1, 1-35.
Parpette, C. 1999. ‘Le discours pédagogique scientifique: caracteristiqués et implications
didactiques’. Paper given at Colloque international “La Didactique des langues en Europe au seuil
du 3 millénaire: des réponses, des questions”. l’Université de Nancy 2, France, 27- 29.5. 1999.
Schroeder, H. & Markkanen, R. (eds) 1997. Hedging and Discourse. Berlin: Walter de Gruyter.
Simpson, R. C, S. L. Briggs, J. Ovens, and J. M. Swales. 1999. The Michigan Corpus of Academic
Spoken English. Ann Arbor, MI: The Regents of the University of Michigan

Simpson, R. 2000. ‘Adverbial Hedges in Spoken Academic Language: Cross-disciplinary
Comparisons and Teaching Applications’. Paper given at the The Second North American
Symposium on Corpus Linguistics and Language Teaching, Northern Arizona University, Flagstaff,
31.3- 2.4.2000
Simpson , R. & J. M. Swales (eds) 2001. Corpus Linguistics in North America: Current States of
the Art. University of Michigan Press.
Sinclair, J.M. 1991. Corpus, Concordance, Collocation. Oxford: O.U.P.
Sinclair, J.M. 1987. Mirror for a text. MS. University of Birmingham.
Stotesbury, H. 1999. Reporting, Evaluation and Discussion as Exponents of Interpretation in
Critical Summarization. University of Joensuu.
Stotesbury, H. this volume. ‘Towards criticalness in written discourse: Pedagogical applications of
a study of interpretation in critical writing’
Stubbs, M. 1996. Corpus and Text Analysis. Oxford: Blackwell
Swales, J. 1990. Genre Analysis. Cambridge: Cambridge University Press.
Swales, J. 1998. Other Floors, Other Voices. A Textography of a Small University Building.
Mahwah, N.J: Lawrence Erlbaum Associates.
Swales, J. 2000. ‘Metatalk in American academic speech: "Point" and "thing"’. Colloquium, the
English Department, University of Aarhus, Denmark.
Swales J. & Malczewski, B. 2001 "Attention-Getters and New Episode
Flags in MICASE". In R.C. Simpson & J. M. Swales (eds) Corpus Linguistics in North America:
Current States of the Art. University of Michigan Press.
Varttala, T. 1999. “Remarks on the Communicative Functions of Hedging in Popular Scientific
and Specialist Research Articles on Medicine.“ English for Specific Purposes 18, 2: 177-200.
Varttala, T. this volume. ‘Hedging in a Corpus of Scientific Research Articles on Economics,
Medicine, and Technology. Implications for ESP Research and Teaching’
Väliverronen, E. 1992. Akateemiset puhetavat ja sisäänpääsyn esteet yliopistoyhteisössä. Kasvatus
1/1992, 22-29.
Winter, E. 1977. A clause relational approach to English texts: a study of some predictive items in
written discourse. Special issue of Instructional Science 6: 1.
Winter, E. 1982. Towards a Contextual Grammar of English. London: Allen & Unwin

A Good Question Expressing Evaluation In Academic Speech

More Related Content

Similar to A Good Question Expressing Evaluation In Academic Speech (20)

More from Kristen Carter (20)

Recently uploaded (20)

A Good Question Expressing Evaluation In Academic Speech