Analyzing Spoken Discourse-1 PDF
Analyzing Spoken Discourse-1 PDF
by
January 2002
Introduction
Emperor Pu Yi: Why are words are important for a gentleman, Mr. Johnston?
Mr. (Reginald) Johnston: Words are important for a gentleman, for if you cannot say what you mean,
you will never mean what you say.
(Bertolucci & Peploe: The Last Emperor)
Mr. Johnston’s reply to Emperor Pu Yi contains one truism and one fallacy with regards
to spoken discourse. The truism is that all utterances have a meaning, that is a
purpose or intention, nothing is said without a reason for saying it. In discourse
analysis these reasons are called functions. The fallacy is that the words in an
utterance, that is, the utterance’s linguistic form, determines its function.
The first part of this paper examines how function is often neither predictable nor
determined from form alone. Specifically how an utterance’s function often can only be
understood by reference to any or all of the following: situation, discourse structure, and
intonation, with an utterance’s linguistic form often being the least helpful in this
regard. The second part of the paper will analyze a transcript of spoken discourse to
provide an extended illustration of just how limited a tool form is for determining
function.
Part 1.
1.0 Form and Function
Form is concerned with syntactic structure up to the sentence level, i.e. the
arrangement of morphemes and words into the larger units of group, clause, and
finally, sentence. Form is also concerned with the syntagmatic relationship between
words within clauses and sentences. For example, “I’m taller than you” is different from
“You’re taller than I am”. Inverting “I” and ‘you’ around the comparative adjective changes
the propositional meaning of the sentence. Function however, is concerned with the
utterance’s purpose, i.e. what the utterance is meant to achieve. For example:
1
All “original examples” have been concocted for illustrative purposes.
The son uttered “You’re taller than I am” for the purpose of refusing to comply with a
command. This is a very different function of than that of:
A: Which of us is taller?
B: You’re taller than I am
(original example)
…where, “You’re taller than I am”, functions to provide information to a question. Nothing
about the form, that is the syntactic structure of the utterance itself, or the syntagmatic
relation between the words within it, allowed us to predict its function.
The above clauses have the form-classifications of, (I.) imperative and (II) interrogative,
but both could be assigned the functional classification of ‘directive’. The ‘directive’
function of the above stem from the verb ‘shut’ and whatever follows (‘the door’, ‘the
window’ ‘your mouth’ etc.). While example II above looks like an inquire that could be
paraphrased as “Are you willing to shut the door?”, this Can you equals Are you willing
paraphrase is faulty (Haegeman, 1983:83). For example, “Can you be quiet” does not
equal “Are you willing to be quiet”. Yet, in the case of II, the grammatical items (forms)
preceding ‘shut’ do have a purpose, namely, a ‘politeness function’. This is what Searle
(1975) calls an indirect speech act (Coulthard, 1985: 27). That is, an utterance with an
underlying base function performed indirectly by the performing of what, on the surface,
could be another speech act (function) form. In the case of II, a directive function is
indirectly performed by an interrogative form, which are often used for inquires
(‘questions’), rather than directly by an imperative.
While considerations such as ‘politeness’ may change the semantic label given the
utterance, from say ‘command’ to ‘request’, discourse labels are concerned with base
functions. Discourse-function labels like directive subsume semantic ones such as
2
All discourse function labels are taken from Francis & Hunston, 1992: pp. 128-133.
‘request’, ‘order’, ‘command’ etc. The above shows that when assigning discourse
function classifications, there are no form-equals-function maxims, such as “an
imperative always equals a directive”, or, “an interrogative always equals an inquire”.
1. Relation be relevant
Many utterances do not conform to these maxims, but retroactive application of these
maxims by the listener onto the utterance, allows him or her to interpret the function of
the speaker’s (potentially) ambiguous utterance. For example, if one roommate says to
another: “The music is a bit loud”, rather than focus solely on the utterance’s declarative
form, which seems like a mere statement of fact, the listener mentally rephrases the
utterance as if the speaker had fulfilled Grice’s maxims. The listener is able to interpret
the utterance’s declarative form as a directive, the speaker likely means: “Turn down the
music”. The utterance’s function is now understood, despite the speaker having violated
maxims of both “Quantity” and “Manner”. This constant effort to interpret each other’s
utterances is what Grice (1975) calls the “Co-operative Principle” (ibid.).
Consideration of situation is often vital for determining function. Sinclair et al. (1972)
lay out constative (i.e. what constitutes) rules for commands. For example, one rule is:
3
Note that ‘a)’ however is a purely form consideration.
4
A Rank scale is taxonomy from the smallest unit(s) to the largest. In grammar (form) these are the
morpheme, word, group, clause, and sentence (see Brazil, 1995:29). Smaller units comprise the larger ones.
least two turns involving at least two people. According to Sinclair and Coulthard, the
exchange structure has at least two elements, Initiation and Response. For example:
The ‘why’ above is not really a response, in that it does not answer the Initiation
(question) above. The “why” is labeled by Sinclair and Coulthard as R/I (an element that
may be a 2nd initiation, or a response to the 2nd initiation, following an I), and is an
optional element (indicated by brackets). The exchange above can now be labeled: I
(R/I)(R/I) R (Coulthard, 1985:136).
Nothing about the above R: “Can’t, I’ve got exams to mark” required the other participant to
say “That’s too bad”. “That’s too bad” is an option labeled as Follow-up (F). A Response does
not predict or require a Follow-up unlike an I or R/I, which predict and require a
Response. The structure of exchanges can then be described as I (R/I) R (F) (ibid.).
7.0 Moves
A switch between an element (I, R/I, R or F) requires a move. In section 6.0, the
Initiation was made using an eliciting move, but in the following:
I: I’m tired
R: Me too.
(original example)
There are also two types of eliciting moves and two types of informing moves:
Each type of move can only occur once in an exchange and in only one, therefore
predictable, order: e1, i1 or e2, i2 (ibid: 75). This idea of one type of element or move
predicting the subsequent move explains how discourse structure itself, allows people
discern a single function from the variety of forms that can realize it.
For example:
(original example)
An e2 Initiation predicts a polar (Y/N) response, yet the responses in form appear to be
irrelevant to the elicitation, i.e., none of the responses contain ‘yes’ or ‘no’. However, the
discourse constraints demand a yes no answer, so the ‘initiator’ knows that a Y/N
polarity assertion is being made by the ‘respondent’ regardless of the form realizing it.
Employing Grice’s maxims, the ‘initiator’ will attempt to interpret how the ‘respondent’s’
response has realized a ‘Yes’ or ‘No’ (all Responses above would be ‘no’).
8.0 Intonation as a function determiner
Brazil (in Coulthard 1992: 37) points out that intonation provides moment-to-moment
functional and emotional value to utterances. Due to space constraints, this section will
only show two examples of how intonation can act as a function determiner.
Prominent Syllables: Making syllables or words in a tone unit prominent can change
the function of the same form. //CAN you SWIM a LENGTH john// is a question
because CAN is prominent. Whereas //can you SWIM a LENGTH john// is more likely
a command because “ …‘can’ is non-prominent and non-performing; it would make no difference if
“would”, “could” or “will” were substituted into what is essentially a politeness formula.” (Coulthard,
1985: 131) (see also section 2.0).
Pitch-Key Sequence: Utterances are rarely said in monotone, rather the tone and pitch
within or between utterances often changes, and these changes in pitch convey the
utterance’s intended function. A pitch difference, called the key, can be labeled high,
mid, or low relative to the pitch preceding it (Coulthard, 1992:42). Pitch can set up
constraints/expectations of what should follow in response to an utterance. Take key-
tone termination, the function value of the key of the last prominent syllable in a tone
unit.
Fig.1
IRitating you say VERy irritating
Doctor: //VERy // Patient: // //
(ibid.:46)
The doctor’s utterance was declarative in form, but due to the intonation was correctly
interpreted by the patient as a polarity elicitation, with the response “very irritating”
serving as a ‘yes’ (ibid.).
A mid-termination is used to elicit expected confirmation, and therefore is not a polarity
elicitation (Coulthard, 1992: 46).
Fig.2
Its DRY
Doctor: // skin// Isn’t it// Patient: //mmm//
The patient only needs to, and is only expected to, confirm the doctor’s elicitation. Note
also that the key of the patient's response concords with the doctor’s elicitation. This is
because a high key is used as a contrastive, so the high tone (Fig.1) “very irritating”
means ‘yes’ not ‘no’. Whereas the mid-key (Fig.2) “mmm” is neutral and additive,
meaning “That’s correct” (Ibid.).
Interestingly, a low termination does not constrain a response on the listener, unlike
high or mid-termination, and is similar to follow-ups (see section 6.0) in this regard.
Utterances that do not require responses are often boundary exchange markers
(indicated by ///), marking the end of one mini-topic and the shift to another
(Coulthard, 1992:49). For example:
By using a low-termination the doctor is saying, “enough about your skin for now, lets
move on…”. The patient knew from the low-termination that the doctor was not
initiating or eliciting a response, and therefore does not contribute one, despite the fact
that the form of the doctor’s utterance (“very dry skin”) could easily have been an elicit
given a different intonation.
Very often a form seems to have the potential to realize a number of different functions,
for this reason form, in itself, is not a reliable predictor of function. While the choice of
one form over another to realize a given function may be made for reasons of
‘appropriateness’ or ‘politeness’ -- changing nuance and semantic classification—these
type of form choices do not alter an utterance’s underlying function. Yet the same form
uttered in a different situational context, position within discourse, or with different
intonation give the form different functional values, making these considerations much
more reliable determiners of function than considerations of form.
PART 2.
1YS: It’s not fair. Sarah got a new bike when she was fourteen.
2F: Do you keep a running tally of everything we buy for Sarah?
3YS: Well, yeah
4M: Alexis, we can’t afford that—oh, a Christmas card from Geordy?
5F: So?
6M: Well I finally took them of the list! We haven’t laid eyes on them since we had
Alexis.
7F: Jill, it’s a card not a dead fish.
8M: I know, but if I send them a card now, it’s going to look like we’re only sending
them because they sent us one.
9YS: Sarah got new boots?
10M: Stop keeping score.
11F: Those are nice, they must be warm
(older sister enters kitchen, sits down to eat)
12YS: It’s alive, It’s alive!
13OS: How come we never go to church for Christmas anymore?
14YS: (inaudible)
15F: Well, you know, in the last few years—you girls, um—haven’t, ah—shown much
enthusiasm.
5
Although the following is an excerpt of a longer stretch of dialogue, for clarity and reader convenience,
the lines have been numbered 1-22.
16OS: So do you guys believe?
17M: In God?
18OS: (nods)
19F: Your mother was raised Catholic, and I—um—wasn’t so we…
20M: You know, you don’t have to attend church—not going doesn’t mean that um…
21F: Exactly, I mean you don’t have to go to believe in God
22YS: Do we really have to talk about religion? It’s Christmas.
•Line 2, being interrogative in form and because the interrogative begins with “Do you”,
we can see from the form it is a neutral proposal, seeking a yes/no response without
indicating the polarity of the expected answer.
•Line 6 includes two declaratives, and it is hard to imagine how the same forms could
be anything other than informative in function.
•Line 15, the declarative clause, along with the marker “Well” and the starter “You
know”, combine to make this line clearly an informative.
• Line 19, “Your mother was raised Catholic” by itself, if given a high termination could
be an inquire, however, while incomplete and interrupted, the continuation “and I
um- wasn’t so we…” would make the entire line an informative in any discourse.
•Line 20, again incomplete and interrupted, has two declarative clauses, which together
would make an informative in any discourse (however, note that here “You know”
acts as a starter).
Line 1, has two declarative forms which look like an observation and an
informative. However, “It’s not fair” ends in a high termination, which gives it the
function of “an invitation to adjudicate” (see: Cauldwell & Allen, 1999:47). “Sarah
and ‘she’ are given prominence in “SARah got a new bike when SHE was eleven”,
making the (perceived) contrast between the parents’ generosity towards the OS and
the lack of generosity towards the YS salient. The entire line is indirectly an inquire,
which could be paraphrased as “Why can’t I have a bicycle too?”. The (M)other
begins to reply to this inquire in line 4.
Line 3, the “Well” is a marker, because it is its own tone unit. “Yeah” is an
informative due to it being an answer to the neutral proposal of line 2, however, in
a different discourse “Yeah” could have perhaps been a confirm or a concur or
engage.
In line 4, the first half is an informative response to line 1. The second half, “oh, a
Christmas card from GEORdy” is more of an inquire than an informative. We
know this from situation (i.e. Geordy is rather persona non grata with my Mother),
which is why she placed both prominence and a ‘contrastive’ high key on “Geordy”.
(M)other is really asking “Why has Geordy send us a card?”.
In line 5, (F)ather does not see the problem so does not understand the relevance of
the question. “So” therefore is an inquire, meaning: “What’s the problem?”. We
know this only from the relation of ‘so’ to line 4.
Line 9 is declarative in form, and as YS has just watched my mother wrap the
boots, line 9 on the surface looks like an observation. However, due to the
“questioning” (high termination) intonation it seems to be an inquire. Given
knowledge of the sibling rivalry situation, line 9 is really asking “Why does OS get
new boots (and not me)?”.
Line 10 is imperative in form and would be a directive in many discourses. Yet this
directive function is most clearly indicated by intonation and discourse position.
While in this case the line was uttered with a low-termination, given a high
termination and if found in an exchange like the following:
Line 12 is a joke version of a greeting (i.e. Good morning). This can only be
understood by reference to situation. Namely, it is late in the morning, OS being a
teenager sleeps in whenever possible, and line 12 was delivered in mock imitation of
the line from the original Frankenstein film, said by Dr. Frankenstein when he re-
animated the dead and dormant monster. Only those with knowledge of the
situation and cultural (film) reference can understand line 12 as a somewhat
disparaging greeting.
While line 13 is both interrogative in form and an inquire in function, OS did not
respond to YS’s greeting, and chose instead to direct an new (I)nitation at the
parents instead. Line 13 therefore implicitly functions as a reject as well. By
deliberately not responding to the greeting, OS was telling YS to (being polite here)
“Get lost”
Line 17 is a return and is a R/I eliciting move of a clarification exchange, the line’s
position relative to lines 15 and 16, indicates it is a return.
In line 21, we know the “Exactly” functions as a react, because “Exactly” is its own
tone unit. The remainder of line 21 is a reformulate, because it paraphrases (and
completes) line 20. In a different position this line would have been an informative.
Line 22 in its interrogative form, looks like an inquire. However, because the
“REALLY” was prominent (rather than “DO”) it seems to function as a kind of
protest. YS is criticizing the “appropriateness” of the entire exchange from lines 13
to 21 and this line could also be seen as a type of implicit directive to drop the
subject. “Do we REALLY have to talk about religion” ends in a low termination, so
YS is attempting to effect a boundary exchange. “It’s Christmas” is an observation
said for humorous effect and to lighten the mood.
Even if another discourse analyst were to dispute the function labels assigned in
section 12.0, the fact remains that the functions will need to be determined by
reference to considerations other than form. If Garfinkel, Grice and Searle [see sections
3.0 & 4.0] are correct in that function, while intended, is really never said but only
interpreted, perhaps such disputes are inevitable. Even if one were to disagree that the
functions of the utterances in section 11.0 were indicated by their forms, as suggested
in this paper—and this may only be a reflection of this writer’s lack of imagination-- this
would only increase the already great preponderance of functions not predictable from
form found in the transcript. If nothing else, the transcript in section 10.0 shows that,
in informal conversations at least, more often than not, function is not predictable from
form. There is no reason to believe that this transcript is in any way unique in this
regard.
However, if students are to become competent, avoid sounding stilted and not develop
what could be called “fossilized comprehension”, students need to be weaned from this
‘linguistic conservatism’.
Secondly, if students are to be exposed to the different functions a form can realize (and
vise versa), it is important that these are taught in the order that provides students the
greatest “surrender value” (see Willis, 1990: 42). That is, the functions that a form most
frequently realizes need to be taught first. This may not be the function(s) presented by
traditional (i.e. pre-corpus linguistics) grammars.
Thirdly, this also has implications for the type of dialogues to be included into
textbooks. Most textbook dialogues usually are chosen (or scripted) either to illustrate a
specific grammar point via repetition (e.g. a dialogue of endless ‘tag-questions’), or are
(or resemble) formal one-on-one interviews following a Q&A, Q&A pattern, with direct
questions being given direct answers. Perhaps textbook writers should consider
including more dialogues like that in section 10.0 of this paper, especially since
“interview” type dialogues (and in class activities) represent a rather specialized and
limited type of discourse. It may also not be enough having a form’s many functions
individually illustrated in separate dialogues. Perhaps more dialogues, in which a single
form is performing two (or more) functions (or where a single function is realized by
more than one form) would be more effective in increasing learner tolerance of the
inconstant relationship between form and function.
A final tentative conclusion is: since there are very few, arguably no, one-to-one
form/function relationships, this inconstant relationship between form and function
suggests that perhaps forms can never be comprehensively presented and explained, as
attempted in various pedagogical grammars. It seems more likely that forms will only be
fully learnt, in all their various function-manifestations, by experiencing them in
context and in use.
REFERENCES
Coulthard & Brazil. (1992) “Exchange Structure”, In M. Coulthard (ed.) Advances in Spoken Discourse
Analysis. London: Routledge. p.p. 50-78
Francis & Hunston. (1992) “Analysing everyday conversation”,In M. Coulthard (ed.) Advances in Spoken
Discourse Analysis. London: Routledge. p.p. 123-161
Haegeman, L. (1983) “’Will you’ and ‘are you willing’: a faulty paraphrases” ELT Journal 37/1.
Sinclair & Coulthard. (1992) “Towards an analysis of discourse”, In M. Coulthard (ed.) Advances in
Spoken Discourse Analysis. London: Routledge. p.p. 1-34
Sharwood Smith, M. (1994) Second Language Learning: Theoretical Foundations. New York: Longman