100% found this document useful (2 votes)
2K viewsConversational Routines in English - Convention and Creativity PDF
Uploaded by
Choi PhilipCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
100% found this document useful (2 votes)
2K viewsConversational Routines in English - Convention and Creativity PDF
Uploaded by
Choi PhilipCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
/ 134
Conversational Routines in English
vitor Studies in language and linguistics
General editors: GEOFFREY LEECH & MICK SHORT
Lancaster University
Already published:
A Dictionary of Stylistics
‘The Communicative Competence of Young Children.
SUSAN H. FOSTER
Linguistic Parisi
GEORGE THOMAS
‘Women, Men and Language
Second edition
JENNIFER COATES
Lexical Ambiguity in Poetry
SOON PENG SU
Understanding Metaphor in Literature
English Spelling and the Computer
ROGER MEFTON
Conversational Routines in English: Convention and Creativity,
KARIN AMER
CONVERSATIONAL ROUTINES
IN ENGLISH:
Convention and Creativity
KARIN AIJMER
Longman
London & New York ‘Addisom Wesley Longman Limited
Edinburgh Gate
Harlow, Essex CM20 2]
and Associated Comproiesthroughout the world
Published inthe Ute Stats of America
‘by Addison Wesley Longman, New York
(© Addison Wesley Longman Limited 1996
‘The right of Karin Aijmer tobe identified as the author of
this Work has been asserted by her in accordance with the
‘Copyright, Designs and Patents Act 1988.
Allright reserved na part of his publication maybe
fad, sored in aretieval syst, oF wansraed
inany form or by any means electronic, mechanical
Photicopying recording or ctherwise without either the
Prior wen permission the Publishers or licence
rmiing restated copying in the United Kingdom issued
bythe Copyright Licnaing Agency Li
90 Tatenkarn Court Read London WIP HE
First published 1996,
ISBN 0582 08212-9 Cs
ISBN 0582 08211-0 Ppr
British Library Cataloguing-in-Publieation Data
‘Accatalogue record for this book is
available from the British Library
jng-n-Pablcation Data
‘Ames Karin
‘Conversations routines in English convention and creativity /
avin Aer
{em — (Gries in anguage ad linguistic)
Indies biblographia reference and nde
ISBN 0-582-O861120 (paper) — BBN 0-582-08212-9 cas)
{-Englsh language-Spoken Engl. 2. English anguast—
Discoue analy 3 English anguoge—Prenk aalis
4.Computationat linguists. 3.Comveration 6 Semitic
{hte Series Stales in language std lngustcs Landon,
England
PEIOI SAT? 1956
‘a4 s6-17540
cP
Set #n9/1 RPO gu ad
Produced by Longman Sng hers Pe
Paedngoppre
fete 25%
To Celia, Kristofer, Paula and Lovisa Contents
Preface
Acknowledgements
CHAPTER ONE: Introduction
Wl
12
13
4
15
16
17
18
19
110
un
12
1B
‘Aim and scope of the present study
Material and method
Frequency of conversational routines in
spoken language
Psychological aspects of conversational
routines
‘Conversational routines and ritualization
Lexicalization, grammaticalization and
idiomatization
Conversational routines and meaning
Criteria of fixedness
1.8.1 Repetitive phrases and pragmatic
idioms
1.8.2 Prosodic fixedness
‘The processing of conversational routines
Routines and discourse
Conversational routines and grammatical
analysis
1111 Grammatical deficiency
1.11.2 Syntactic integration and position
‘A model for describing the structural
flexibility of conversational routines
‘The pragmatic function of conversational
routines
1.13.1 Conversational routines and
illocutionary force
1.132 Indirect speech acts,
10
n
2
13
“4
15
7
18
18
9
a
m4
4
4
vii CONTENTS
Charrer Two:
1.133. Conventionalization of inditect
speech acts
1.14 The pragmatics of conversational routines
1.14.1 Conversational routines and frames
1.14.2 Factors of speech-act frames
1.15 Conversational routines and language
teaching
Thanking
21 Introduction
22. Thank youlthanks as anillocutionary force
indicating device
23. Thanking and politeness
24 Strategies of thanking,
25. Gratitude expressions
2.6 Continuation patterns
27 The grammatical analysis of gratitude
expressions:
28 Prosody and fixedness
29 Distribution of thanking over different texts
2.10. Thank you/ thanks as stems
2.10.1 Expanded forms of thanking.
2102 Thanking and intensification
2.10.3. Prosody and intensification
2.104. Patterns of compound thanks
2.11 The functions of gratitude expressions
2.111 Thanking and ritualization
212. Thanking asa discourse marker
2.12.1 Thanking as a closing signal in
adjacency triplets
2.122 Thanking in proposal-acceptance
sequences
2.123. Thanking in telephone closings
2.124 Thanking in different turn positions
2.13. The pragmatics of thanking.
2.13.1. The constraints caused by the
object of gratitude
2.14 Frames for thanking
2.141 Variation in standard situations
2:18 Conclusion
(CHarren Taree: Apologies
3.1 Introduction
32 Defining apologies
33 Apologizing strategies
25
26
26
7
28
33
35
35
39
39
41
41
2
45
v
8
SI
51
52
56
58
6
2
78
BL
(Carrer Four
34
35
36
37
38
39
310
an
32
33
314
35
41
42
43
44
45
46
47
CONTENTS
The form of apologizing a
Continuation patterns 87
‘The grammatical analysis of apology
expressions 88
Apologies and prosody 88
Distribution of apologies over different texts 89
Collocational fixedness and flexibility 31
39.1 Fully expanded apology expressions 91
39.2 Apologizing and intensification 93
3.93 Prosodic devices emphasizing the
politeness expressed by the
apology 94
394 Compound apologies 94
Apologies and function 7
Retrospective and anticipatory apologies 98.
3.11.1. Disarming apologies 100
3.1.1.1 Disarmers and corrections 102
311.2 Disarmers as requests for
repetition 102
3.11.13 Disarmers in dispreferred
responses 103
‘Thestructural function of apologies 106
3.121 Apologies in telephone openings 106
3.12.2 Apologies intelephoneclasings 106
‘The type of offence 108
3.13.1 Talk offences 109
3.13.2 Time offences ua
3.133. Space offences m5
3.13.4. Offences involving social behaviour 115
3.135 Offences involving inconvenience 116
Apologies and pragmatic frames m8
3.14.1 Frames for standard situations 9
Conclusion 121
Requests and offers 124
Introduction 14
‘The speech act assignment mechanism and
indirect speech acts 124
Indirect specch acts and pragmatic principles 126
Indirect speech acts and implicature 127
Pragmatic ambiguity 128
Defining requests 129
Requestive strategies 130
‘A taxonomy of requests 131
48 ‘CONTENTS
(CHAPTER PVE:
48.1 Requestives, advisories and offers
482 Explicit and implicit indirect requests
49° Requests and politeness
49.1 Assertive and tentative indirect
requests
492 Requests and style
4.10 Continuation patterns
411 The grammatical analysis of requestive
routines
4.12. Describing request expressions
4.12.1 Prosodic modification
4122. Requests and discourse type
4.13. Indirect requests and speech act stems
4.14 Types of stem
4.14.1, Mitigated indirect requests in the
form of declarative sentences
4.142 Want and need statements
4.143. Mitigated indirect requests in the
form of interrogative sentences
4143.1 Canfeould you
41432. Willfoould you
41433 Requests in the form of
permission questions
4.15. Lexical mitigating devices
4151 Please
4152 Just
4.16. Internal and extemal modifiers
4.161 Requests and extemal modifiers
4162. Combinations of modifiers
4.17 Referential strategies
4.18 Requests and pragmatic conventions
4.18.1 Requests and the situation
4.18.2 Frames fr requests
4.19 Imperatives
4.19.1, Imperatives and politeness
4191.1 Do-+ imperative
419.12 You + imperative
4.20 Patterns expressing offers
421 Conclusion
Discourse markers as conversational routines
51 Introduction
52 Coherence and discourse markers
53 Discourse markers characterized
54 The metalinguistic function
14
136
139
40
141
ww
M4
45
M5
146
Ww
148
49
154
156
157
159
161
163
166
169
170
170
178
175
7
178
180
182
183
187
189
195
200
200
201
203
References
Index
58
59
5.10
Sn
512
CONTENTS
Relevance theory and communication
55.1 The interpretation of discourse
markers in relevance theary
‘The approach to discourse markers in
this work
“The linguistic properties of discourse
markers
57.1 ‘The discourse marker slot’
57.2 Prosodic fixednoss
5.7.3 Positional fixedness
574 The grammatical analysis of
discourse markers as stems
Contextual properties of discourse markers
5.8.1 Discourse markers.as deictic
‘pointers’ referring backwards
and forwards in the discourse
582 Discourse markers and person deixis,
Functional properties of discourse markers
59.1 Global and local discourse markers
59.11 The pragmatic functions
expressed by local
discourse markers
59.12 Functions of global
discourse markers
‘Combinations in the discourse marker slot
Discourse markers and cognitive frames
Condlusion
208
209
21
a
212
216
216
a7
218
218
20
221
21
226
231
232
235
246 Preface
The aim of he prevent work is to sty the range ad frequency of
tional routines in the London-Lund Corpus of Spoken English
‘The numberof spoken texts which have been colleced and transcribed
make up almost haifa million words. The original recordings on which
the texts are based were made mainly during the 1960s and 1970s at
University College, London, as a part of the project Survey of English
Usage (SEU).
‘The samples represent various text types such as face-to-face conversa-
tions, radio discussions and telephone conversations, described in more
howdy
Finally, the phrase is fixed in the sense that ¢he idiomatic function (= a
greeting) is associated directly and unambiguously with the linguistic
form. A high degree of fixedness is also indicated by the fact that, in a
2
INTRODUCTION
discourse perspective, how do you do must be followed by another greet-
ing, usually a copy of the same phrase (A: How do you do? ~B: Hew do you
o?). Gee section 1.10.)
‘According tothe same tests, the greeting how are you isles fixed. It can
either be interpreted literally or have an idiomatized pragmatic function
The idiomatic meaning is illustrated by the following exchange:
How are you?
Fine how are you?
However, the phrase can be understood literally asa question fits asked
by one’s doctor and answered by a specification of one’s health (Makai
1972:175). Grammatically its also les fied. The form of the verb can be
changed (How haze you been?) the subject can be the third person (ow is
he?) The phrase can be embedded in another structure (He wondered how
1 yas) and it can be expanded (How are you today?).2
‘The majority of routines are like how are you, which is fixed according
to some criteria but not others. Thank yu is another example of how each
phrase needs to be described on the bass of several differen¢ criteria. The
phrase cannot be passivized (“you are thanked), and the verb cannot be
replaced by a (near-synonym (‘appreciate you). However, thank you can be
modified (thant you very much), embedded (I wish fo thank you), etc in
‘ways which will be further discussed in Chapter2.
Haggo and Kuiper (1983: 541) account for the grammatical and seman-
tic restrictions on formulas and their potential variability ina finite state
grammar without loops (see section 1.12). Another model is provided by
lexicalized sentence stems, an alternative I explore in section 1.12.
181 Repetitive phrases and pragmatic idioms
Conversational routines can also be fixed because they always appear a
ceriain form. Manes and Wolfson’s work on compliments (1981) is of
special interest here. To begin with, Manes and Wolfson noticed that there
isan almost infinite number of (indirect) realizations of a compliment.
However, some variants are never chosen although they are grammatical
‘and would be functionally appropriate. Manes and Wolfson found, for
‘example, that as many as 85 per cen (out of 686 compliments) fell into one
of only three syntactic patterns such as I (really) like NP (Manes and
Wolfson 1981: 123) Furthermore, just two verbs (like and love) occurred in
8% per cent ofall the compliments containing verbs expressing a positive
‘evaluation (Manes and Wolfson 1961: 1279, Of the compliments with an
adjective, 229 per cent contained nice (Manes and Wolfson 1981: 117) (See
Figure 12)
Manes and Wolfson concluded that compliments were formulas
(routines), just as much as greetings or thanks, because speakers used
only a restricted set of expressions, As a result, compliments are
a (CONVERSATIONAL. ROUTINES IN ENGLISH
ie fm [B] on
r) 2 ‘ta ie] ne
PRO is realy (a) ADV NP
You Via! vealy ADJ NP
NOUV (NPY ealiy) ADV
Yu have (al eally) ADU NP
(Wit) AD NPP
ADIN
fant ABU NP
Other patorns|
ee zemnpo
Figure 12 _The distribution of syntactic pattern in compliments (after Manes and
Wolfson 1981: 121)
constrained as o patter productivity just. as much as thanks or apologies.
Itshould be remembered, however, that even if the patterns which can be
established by quantitative methods are of interest, they have different
properties from other routines
At this point we need to make a distinction between different types of
routine, According to Stenstrém, structures can be analysed both on the
‘grammatical and pragmatic level (ee Stenstroms 1990: 172F multi-level
analysis of so-called discourse markers). Coulmas (1979: 240), on the
‘other hand, discusses the possibility of classifying routines into repetitive
(recurrent) phrases and pragmatic idioms. “Repetitive phrases’ (such as
really like NP) do not have special syntactic chatacteristics, ie. they
are not restricted as to pattern productivity and they are not transforma-
tionally defective. Pragmatic idioms’ are characterized by typical
idiom features, such as a special meaning, fixed form and grammatical
irregularities.
Finally it should be noted that there is a fuzzy boundary between dif-
ferent types of phrase. As Lamb points out (1992, unpublished), there are
a great many phrases which are only weakly or slightly idiomatized, Le
they keep ther literal meaning. Such transparent phrases are, for example,
the probiem is .., what Fmean i .., let me tell you something.
18.2 Prosodic fixedness
Prosodically, conversational routines correspond to tone units. Tone unit
boundaries and hesitation pauses within the tone unit therefore provide
additional clues about the degree of fixedness of a conversational routine
(cee Altenberg and Beg-Olofsson 1990: 18; Peters 1983: 10). There is a fixed
set of tones to choose from. As Kuiper and Haggo point out, a fixed pro-
sody makes ‘storage and retrieval easier, as. fling system using cards of
“
INTRODUCTION
Afferent colour in each section is easier to use than a single colour system’
(0984: 228),
Keller (1981: 1061) regarded “a drop in the intonational contour" as a
sore reliable criterion for ‘gambits' (conversational routines with a dis
course-organizing function) than syntactic independence or postion in
the uterance. Moreover, a characteristic intonation pattern can be chosen
in stereotypic or standard situations illustrated by the tone on sorry when
cone bumps into someone (see Ladd 1978 ~ "stylized intonation’). Finally,
it should be stressed that although intonation is always to some extent
fixed, a conversational routine may have several prosodic patterns each
‘with complicated conditions of use. Sory i, for example, used with arise
(fall-rise) tone ifthe speaker asks the hearer to repeat Something, but with
a fall tone to signal sincere regret.
A fixed prosody also means that the stress cannot be moved from its
normal position to any other element of the phrase (see section 1.1). There
are, however, degrees of prosodic fixedness, and changes in the prosodic
patter can be made for certain purposes (thank you ery moch-> Hak you
ery much)
1.9. The processing of conversational routines
On a general level, the analysis of routines and idioms depends on
whether language performance is regarded as governed by grammatical
rules or as the ‘putting-together’ of ready-made or partly ready-made
cognitive units stored in the long-term memory. Inhis connection itis
fruitful to think of the ‘language text’ as being produced and interpreted
according to two different principles.
‘According to the so-called ‘open choice principle’ discussed by Sinclair
(1987:379), the language text isthe result of a number of ‘slot-by-slot’
choices made by the language user on different levels. A language text
organized according to the ‘idiom principle’ imposes heavier constraints
on what is said” since the principle presupposes that language users have
at their disposal a number of lexicalized or semi-lexicalized phrases
which are stored in the long-term memory (see Sinclair 1987: 320). Such,
phrases can be interpreted directly by the hearer without the need of
semantic rules operating on the syntactic structure in a compositional
fashion (see Morgan and Horn’s notion ‘short-circuited implicature’; see
section 4.4). The idiom principle is compatible with a tendency to eco~
rnomy in Janguage (see Leech 1983: 67) since the speaker can easily and
rapiclly get access to the lexicalized or idiomatized chunks which are
stored in the long-term memory. Looking things up in the long-term
memory is quick and provides the speaker with extra time to plan ahead.
Sinclair stresses the fact that the two models of language are incompat-
Ible with each other: There is no shading of one into another; the switching,
1s CONVERSATIONAL ROUTINES IN ENGLISH
from one model to the other will be sharp’ (1987: 324). On the other hand,
the principles or processes for interpreting sentences do not represent
an either/or choice in discourse interpretation, but operate in a parallel
fashion:
‘We should also recognize thatthe choice between understanding a combination
om the bass of is components and understanding a lexcalized formation as a
Unit isnot anelier-or choice at al since these two proceses may both operate,
in parallel, Its very likely tht such parallel operation is usual fF com:
binations that areas yet only slightly leaned,
(Cam 1982, unpublished, p. 1)
Parallel processing implies thatthe hearer has @ choice between recog
rizing a combination of words asa lexicalized unit or analysing it on the
basis of the grammar and the lexicon, and that switches from one prin-
ciple to another can take place on-line. In Lamb's model there would be
two possible processing routes (see Figure 1.3).
Routines
Recognition of Processing by means
texcaizes unit, tthe grammar and
the lesicon
Figure 1.3 Parallel processing of conversational routines (strongly and weakly
‘idiomatic phrases)
Even when a lexeme has been ‘strongly’ idiomatized, its components
can still operate as such fo provide connotations based on the meanings
they have when occurring in isolation (cf. frankly speaking, to tell yor the
‘ruth. If we take it that there are parallel ways of processing idioms, we
«an also explain that the semantic information associated with the com-
ponent parts of the idiom can be recovered by means of ‘recycling’ (cf
Haggo and Kuiper 1983: 549 ~ ‘semantic reanalysis’; Kjellmer 1991: 123 —
‘manipulation’. An example is when How do you do? is responded to by
How do Ido what? for husmorous effect (Makkai 1972: 173).
This view of language processing and interpretation means that many
phenomena, which traditionally belong to grammar, may also be analysed
as routines and patterns. Makkai advocates the view that much of gram-
‘mar ough to be studied in an idiomatic perspective (1992, unpublished).
Here are some examples:
+ uses of words which are collocationally restricted
(now you come to mention it)
6
INTRODUCTION
+ uses of words cooccurring with certain complements
(hark you for coming, sorry to trouble you)
+ words which are ordered in a special way
(perhaps you would like to come)
‘+ combinations of words which repeatedly occur together (indirect
requests ike cm you, would you, etc)
There are many problems with studying utterarces as routines or larger
routinized patterns.” Even if one believes thatthe bulk of language is to
some extent routinized, its clear that speakers may differ with regard 0
what patterns they have stored and use. Moreover, itis difficult to for-
rmulate pragmatic rules or frames for indirect requests which are suffi
ently constrained to explain how the extralinguistic situation is matched
with lexical forms,
1.10 Routines and discourse
Minimally, a routine is a single word (eg, thanks, sorry), but most routines
consist of several words. Is clear that speakers can store quite long
phrases in their long-term memories (eg. the routine can ! keep you for a
‘moment and tell you something which can preface a message). A rostine
need not even be a syntactic constituent as is shown by routine phrases,
such as the fact is. cam you ... . Some routines are discontinuous with a
slot for lexical insertion: as far as / .../is concerned.
Routines do not only occur alone, but they can be repeated and juxta-
‘posed. The structures which emerge are composed of single building-
blocks, stored! as units in the speaker's long-term memory, which appear
in the order that they are generated. The combination of routines can be
‘more ot lese fixed:
) B Aldvelye
thank* you !very ‘much in-ded#
(9.2.568-9)
In general, the length and complexity ofthe gratitude phrase depend
on the social situation and the amount of politeness it requires (see
‘Chapter 2). Besides, we find sequences of routines. The smallest ofthe for-
‘mulaic discourse structures can be described as aft adjacency pair
‘Richards and Schmidt 1983a: 128; Schegloff and Sacks 1973). Adjacency
pairs are sequences of utterances produced by two successive speakers
such that the ‘first pair part’ constrains the other speaker to provide a
response in the ‘second pair part. An exact copy of the frst part is char-
acteristic of pairs which are completely formulaic such as greetings (Hello
= Hello). In discourse structures which are less tightly constrained, there
isa choice between different strategies in the second part ot the second
part may be optional. Compare, for example, the following adjacency
v CONVERSATIONAL ROUTINES IN ENGLISH
pairs consisting of a request and different second pair parts (from
Richards and Schmidt 1983a: 129)
Request Grant A: Can you mail these for me please?
B. Sure
~ Put off B: Sure, but won't have time today.
=Challenge _B:_ Why do you always ask me to rasil them for you?
~ Refusal Bi: Sorry, but I won't be near the Post Office,
Discourse patterns (which can be regarded as a special kind of adjacency
pairs) may occasionally consist of three pars, al of which are routines,
‘A: ‘Would you pas the milk”
% ‘Were’
‘Ac “Thanks”
(atter Gottman 1975 65)
“The first two moves of the conversational exchange are a request by A and
the granting of the request by B. The third move (which is optional)
acknowledges the response by B and closes the conversational exchange.
Routines are frequent in certain discourse types which impose
constraints on the speaker’s short-term memory stich as auctions (see
Kuiper and Haggo 1984), sports commentaries (ee Kuiper and Austin
1990), elgious ceremonies (see Paul 1989). Another example is telephone
calls where routines are found in the opening and the closing ofthe call
(see Hopper 1989; Schegloff and Sacks 1973). This will be illustrated in the
following chapters.
1.11 Conversational routines and grammatical analysis
‘To what extent can conversational routines be described grammatically?
In the following sections \sball discuss if and what syntactic information
must be specified for conversational routines.
11L1 Grammatical deficiency
According to Bahns etal. (1986: 695), formulas do not show any syntactic
peculiarities. Many conversational routines can be analysed as sentences
am grateful for fyour help/, '™m sorry, how do you do}, subordinate clauses
@s far as Teno, i may sy <0), noun-verb structures (I mean, you know,
nnon-finite clauses (to fell you the truth, frankly speaking), prepositional
phrases (infact), etc. Tere are also grammatically defective routines such
as Sentence fragments, independent subordinate clauses (if you cou ep
‘me), irregular wh-questions (how about coffee), non-sentences (next slide
please), and elliptical constructions of different kinds (ant a drink
Routinized sentence fragments can be of different kinds. The short
phrases sorry, beg your pardon oF thank you are analysed as sentence
8
INTRODUCTION
fragments because they lack the subject (or both subject and copula).
Since the missing elements are known, the phrases must be analysed as
the result of situational ellipsis (Quirk et al. 1985: 11.38).
Even if routines have a constituent structure, they are not necessarily
{generated by the grammar. To give an example, you know and I men are
8 fixed combination of noun and verb which must be analysed as dis-
course markers (see Stenstréim 1990), Outside this group, however, noun-
verb structures are an open-ended category: In comment clauses such as
he thought, he std, etc, the verb can, for example, appear with an adverbial
(he suid immediatly), and itcan be exchanged for other synonymous verbs
he reflected, he supposed, he mentioned.
‘As pointed out above, conversational routines are usually multi-word
units. One-word routines are difficult to describe in grammatical terms,
‘and many of them would belong to Fillmore et al.'s class of extra-gratn-
‘matical idioms’ (Fillmore et al. 1988: 505). In thanks, for example, itis not
clear what elements have been omitted as thanks could be expanded in
‘more than one way. Both I ote you my thanks andl I give yow thanks are
possible sources for thanks (Quirk et al. 1985: 12.34). Alternatively, we
Could argue that the syntactic properties of the phrase should be
deduced from its syntactic subrcategorization. However, this is not very
helpful since thanks can be sub-categorized in two different weys. The
phrase must be a verb when its followed by an intensifying adverb (eg.
nk very much) but it can also be a noun which is premodified (many
thanks). We conclude therefore that thanks is grammatically rather like
hello which is not amenable to syntactic analysis.” Discourse-organizing
routines such aswell, now which function as ‘openers’ are also difficult to
analyse grammatically. They can be analysed neither as elliptical struc-
tures nor as special parts of speech.
Moreover, whole sentences can be syntactically indetertainate, as illus-
trated by Haggo and Kuipers (1983: 5434) example there yow are (a formota
used when giving somebody something). First, they say this sentence is
probably nota transformation of you are there Secondly, we could hay say
that you is the subject of arc. Thirdly, te status of there is difficult to detor-
rmine.!® Now ifthe phrase isnot simply a transformational variant ofthe
ppkrase you ae ther, this could be taken to mean that it should be analysed
as a grammatical construction, 2. ‘a symbolic model that pairs syntactic
conditions with semantic and pragmatic conditions’ (Lakoff 1987: 471; cf
Fillmore et al. 1988). Asa grammatical construction the phrase can base dif
ferent syntactic, semantic and pragmatic properties from the corresponding,
simple sentences (see also the discussion of stems in section 1.1),
LdL2 Syntactic integration and position
Conversational routines expressing thanks and apologies constitute inde-
pendent conversational moves (see Stenstrim 1990: 147). On the other
8 (CONVERSATIONAL ROUTINES IN ENGLISH
hand, routines with a discourse-organizing function (called discourse
‘markers in Chapter 5) are partly integrated in another grammatical struc-
ture,
Following Andersson (1975) they can also be called speech-act adver-
bials because they have adverbial function and “because they modify the
~ speech act of saying something rather than what is said’ (Andersson 1975:
27;my translation). The adverbial constitutes a prosodic unit marked by
a tone unit boundary. It has initial (or final) position in the wtterance of
which itis a part.
What is interesting about Andersson's analysis is that speech-act
adverbials are characterized not only functionally and prosodically but
also grammatically. In Figure 1.4, a distinction is made between speech-
act adverbials, which have pragmatic function, and other elements
(adverbs, clauses, et), which are integrated in the proposition and havea
propositional function.
The speech-act adverbial (= a) occupies a separate node in the tree dia-
‘gram, indicating that its pragmatic scope extends over the whole structure
although itis not part of the grammatical core. Elements which have a
propositional function (= b) must be part of the propositional structure,
‘which is interpreted compositionally by the semantic rules.
There are several arguments for distinguishing between two syntactic
analyses. First, there is a contrast between speech-act adverbials and
adverbs not referring to the speech act which is clear in pairs such as:
(a) Frankly (speaking), he behaved well
(b) He behaved frankly
Also, whole clauses as adverbials can have two interpretations reflected
in different structural analyses. To give an example, if] may say o has
pragmatic function and is not part of a conditional sentence. The clause
contrasts with jFclauses which state the condition under which the main
clause is true (fit rans, I carry an umbrella).
s s
s
ADVERBIAL
|
Ne ve NP ve Aoverstat
(@) pragmatic tnetion (©) propesitonaltesetion
Figure 14 The constituent structure of (a) speech-actadverbials and (b) of other
adverbials i final position
(After Andersson 1975: 29)
20
INTRODUCTION
Secondly, the analysis proposed in Figure 1.4 can explain the ambi-
guity of the following (constructed) example from Sinclair and Renouf
(1988: 152)
‘Tobe fair, Tom divided the sweets evenly.
In the pragmatic interpretation, the person who is fair is the person speak
ing or writing, and the phrase is a judgement on Tom. In this case, the
phrase is attached pragmatically as a speech-act adverbial or a disjunct)
to the following utterance. Inthe other interpretation, fo could be para~
phrased as in order to’, and the person whois fair i Tom. The infinitive
{sa purpose clause and belongs tothe proposition.
In the wellknown syntactic analysis of adverbials by Quirk etal (1985
8.121), adverbials which are only loosely integrated inthe syntactic struc-
tureare called disjuncts. Their independent status is shown by various
grammatical operations. For example, they ae distinc from other syntac-
Hc classes of adverbials (adjuncts, subjuncts) because they cannot be
rade the focus of aceft sentence (it is rankly speaking that he dd it} or be
the basis of alternative interzogation (“wus it frankly speaking or strangely
enough that he didi).
1.12 A model for describing the structural flexibility of
conversational routines
Can conversational routines be accounted for in generative grammar?
AS we have seen, conversational routines may be invariable and/or
grammatically irregular; they can have idiomatic meanings and be
bound to the situation. The majority, however, are partly productive
and are not grammatically peculiar. Exactly how one describes con-
versational routines and other formulas seems to be a consequence of
‘one's views on lexicalization and the boundary between the lexicon and
grammar. Fillmore et al. describe the current grammatical scene as
follows:
‘Current formal medels of grammar take a severe view of the distinction
between knowing and figuring out: they assign as much work as possible to the
‘computing 6: figuring out part of knowing how to use language, and they
attempt to keep a! 2 minimum those aspects of linguistic competence that have
tobe represented as stored or known,
(illmore etal. 1988: 502)
In generative grammar a distinction is made between a grammatical rule
component generating sentences and the lexicon representing those
aspects of linguistic competence which are stored or known. It is possible,
however, that generative grammar is, in a way, ‘too perfect to account for
‘conversational routines. This leaves us with a number of possible models.
for describing conversational routines. In so-called Construction
a (CONVERSATIONAL ROUTINES IN ENGLISH
Grammar (see Fillmore etal 1988), conversational routines would pre-
sumably be grammatical constructions specifying syntactic, lexical,
semantic and pragmatic information.
Recently, one has also seen different proposals for enriching the
lexicon in order to account for structures which do not belong clearly
either in the lexicon or in the grammar (see Jackendoff's 1975 discussion
of full-entry theories using redundancy rules versus impoverished entries
‘and Haggo and Kuiper 1983: 5399)
1m the theoretical framework of the present study, which is based on
Frawley and Syder (1988), the form of conversational routines isa fairly
abstract sentence stem, consisting ofa ‘core’ with possible extensions. The
following lexemes ae linked by family resemblance andl related by a stem
containing sorry as its core:
)_Y'msorry to keep you waiting,
I'm so sorry to have kept you waiting.
Mr X is sorry to keep you waiting allthis time.
(Pawley and Syder 1983: 210)
‘The stem is a collocational frame” with slots for grammatical categories.
INP be-TENSE sorry to keep-TENSE you waiting
(Pawley and Syder 1983: 210)
Clearly, the stem is different from a productive grammatical rule as it
contains (lexical) elements which are not inserted by means of lexical
rules. The capitalized elements represent slots into which, for example, a
noun phrase or a tense morpheme may be inserted,
Lexicalized sentence stems are placed by Pawley in the “ideal lexico-
sgraphers’ lexicon’ (1986: 114). On a deeper level they are cognitive units
which provide an economical way of storing linguistic variants which
show some similarity, and they can be used ready-made in speech pro-
duction. The availability and flexibility of the stem may explain why
Iexicalized sentence stems are so frequent in spoken language; the
number of lexicalized sentence stems in English ‘probably amounts, at
Jest, to several hundreds of thousands’ (Pawley and Syder 1983: 213; see
also section 1.3)
From a formal grammatical point of view, a dictionary entry for a
lexicalized sentence stem is a ‘mini-grammar’ (Pawley and Syder 1983:
216). The stem has the capacity to generate a number of ‘possible routine
phrases’ and can be exploited by speakers for ‘creative’ departures
(Fillmore 1984; 128). Variants are formed from the stem by substituting,
elements by ‘inflection’, modifying existing structures, combining ele-
ments into sets, embedding structures in larger discourse units and by
deleting optional elements.
Grammatical constructions or lexicalized sentence stems can be com-
pared with a finite state grammar without recursive properties (Kuiper
2
INTRODUCTION
and Haggo 1984: 221)" The grammar runs through a number of states
‘with an initial and terminal state as endpoints, producing words at each
transition, assigning to them a certain probability. Since the model repre-
sents the simplest type of generative grammar, it has little processing
‘capacity (requiring only one slot in the short-term memory).
‘One can notice, finally, that although the lexicalized stems discussed by
Pawley and Syder (1983) have the form of sentences, itis obvious that
stems can also consist of clauses or phrases with fixed and variable parts
and a specification of the elements which can fill an empty slot inthe stern.
ADVERB Sorry to
very
rightfully
terribly
sxufully
truly, ete
Discourse-organizing routines (speech-act adverbials) consist of
