,uor'Srr^rror Jo ,urpnrs aqr or rslrrrllo' "t

'::;f;:l'fi'I1*".1:'j!ll:li -.
sr

u,u* r rlqu tnr ro'r tl

:'::: :l"l:,:' i,',i jj: j,:Jilj:il:l;'
-ro
,n11,,rtu"t*' sll J*,lloJslPbsl
uonru!rp'
ftj
iq !rlJlrlrorlr\our
'sruc([rJttlu(l o^\l s)rII1u4,*l r;ql:Eol rrr,l
,t"q, Jrll rrr
,;Jt'.r,1, ],"*;, <rrls c sr lr :Euru.lr:rrr1r
.F.o:r1r pr'\)trlJu'o ^llr;'i;:;"',-,,1r" ,lll,L.r,, rrr.,
cr ir turros p,rc .:uos.r.*t It )]'
llll']:T,I,'fi.,:t i$:i;;;tturs ",'"ll uo,rn(r,r].or r,,np,\rr)''I
JsrnorsrP u;1otls "' ""1;;;:;:' uo"""''tuo'''l'
l::.i,';;t''.., " ,'to'l'
o rrJ J " E".'lll:;:l il',i::liifi:iijt i;.3iilJli::Jif
troll )nJlq"u",
.r-r-r.rr1l;r Puu 3:],{ ...
,u"rr.,,,t.,rr tt,,,.l Jrr.u.u
rlrnr puc E'qr:1 urnl ;;],tl]l:., su lrJ.ruo.} J([) Jo
x;1duro: rql ul :i.;l;;
slr1lcrl.lr,-,.rr* io Surssu<t runlf,e
rlrr _ *Lr,
.

pur? sp*or-.,-.ro,rorrrnror Jrl-

-trllul lsoLu JlIl u')uo Jo rurpnls
:r1t''(q'1:n:r'u=rrcr'rrur )ll'l
'ituJ.\J tlr.lads Jo \J"iu.,**t
..
.1Ll3ysAs .l,ttttlt:r.;p
(11n., clo uoltnlll

PJl\tl J'\l)ll l P''

'(1rsu:11c
sl
ol )uu\rIJJ )ull] )nq PJqSJJJn ''tu-*t'"tt'\JoA\
-Xt".,'i
:,tr1<trr:'sorlqrq ,,'" " l"t'u*t';.,;; .,P l:Xil;j:i,,,'Jii;iil^iTi':;i';:
st Jr'rt
.rrr<1nr1.,.)rli )Jrrulu(uJ\o
Pltlo^\ "t "-l:]l:'rlu ,r,ri
JtllJo suLtstJo
o) rrlut;ttc orr
.p.15. sJrrohJlDJ puc s1t1':ruo'r
''t'""'.']ut -rs.r.oJqlP Jo ixrl "

ru a p u a cia p
" I ; :: .'::#lxl;;"
jf H,,! ::::"ll:; l:'5'1'#iJ lliJil
,],.i,,p'ru:srrcr ry,,:]]-:-'T. srurFno reldr:r1'r
:::,,T:l[,".^ni,,,..0".
51 lut[] Irl)]sls
erlr rno Burrq ot
peuErsap *
uol]cnpoJr

',(.lera: '-

*
;':ffi:'Xffi li: ::#li
s Pue uarllr /^., o"ll"rl :;11:
o'(-rerarll'txat ualod
^

'r
slurod lo -l-"-t::: srsllouv :. .
,.raded snor,ra.rd aql ur aPPtu Iul ,;ffi;a 'euprp" sta'r aq
;; pauolssrluluo) ser"r ;aldeu:
:
rruaPtrv:uoPuol l9:l
sr lf,€rlxa srtl;ssard e ro1 >l!!c uer ;J*
-srg Jo looqpu'p1 "ti'n'on-''1nu

uorldlrcsaP crlsrn8r4l
.r"pe.rBa1u} erll rro f
5"
 68 The organization rf'tcxt
probably most impressecl by the orderly nature
of rhe events he is describing. mr-,,:lrtent. because r
Printed material is presented in accorclance
with thousands of conve ntions and lrrr.riering one stret(
is n.reasured to thousandrrrs of an inch. The
t."t i. ..uJ Jr.i,rf p..pu.urion ,Hur::- iircumstances. L
many times by di{rerent people, ancl in most
cases time is not a structural influ- rr;r '. obrious as ir sh,
ence' where printed material has to be prepared
:rgai,st strict cleadrines, as in -:: a dr-namic mc,cir
newspaper productio,, every effort is made
to neutralize the e{rect or.time, and ru.{: .r.lre of the di:co;
there is very little impromptu material in newspapers.
:r:.i1,:nrial qualin' oi
resl>onsibility lor coherence lies r.vith ilr.
.The u,,.r.., which in the c,se oi l,:,:'--ed can be raker
printed material is a. composite entity including
everyone *r.o pn.,i.ipated in s : ::rditioned bv the .
the production of the text. The text aprrears
to be quite static and non- ,lnce the direcrion;
negotiable; it is there, ancl it ca,not b. nti...d.
It ,..,r, quite different from m'::r is a marked djtI
the ever-changing. hardly preclictable
movement of conversation. It can be r-iLlage. Each elen'u
described, and is normary describecl, as a complex
contraption, and linguistic ;n:q>ectively, but in d:
terms like 'structure' betray the uncrerlying
meiaphor. wriiten language is not -Yre1-ant to each eleme
Primarily seen as acti'ity. Its relation to'ti,i. is that oran un.rrun-gi;g
record. :u,:'ions. does this eler
But in apparent contrast to the rigidity of
written text, we u.l orrr..d thu, - L the state of the c:
eac'readins of it, even two readings by ilr. ru-.
reader, is a unique commu- 'rr::s . does this eleme I
nicative event. So the text has an inte.active
role th:rt is relatecr to time in a '-:,r main issue can b<
difrerent way, since a reading of a text is zrn
event in time. we can relate :,rrt element is created
the fixed nature of a written text.to trre
unique experience of any reading of it .\nv utterance can
by making a reasonabre assumption: since tire :

mai, purpose of a text is to be :::d to follolv conver:

read, its destined role in a series of interactions
has abackwash efrect upon its :rcause people ma-ke r
composition.
:crics - irony. and sc,
further assumption is that writer who is composing a text that is to be
_one
efficient
zr
:,:rt. No matter \\'har :
interactivelv rr:rs an obvious moder in conversation. From this
inqr-riry emerses that can read to:rn integrated
a rine of ::nious utterances and
framework.ra...rifri.r. Irtrre .,lious in question-a:
same basic model is ,sed in both clocuments
and conversation, tlr.i, ot least at i:E some compleririe=
an abstract level, the same categories of description
are applicable-
The influ- :-re. because, for exarn
ence ol the partic,lar medium, however. becomes stronger as we
toward realization. ancr we mrrst g,ard ng"in.,
proceecl *an medial ones. bur :
imposing ."rlgori". ,pp.op.i"t. l1 utterances is of finr i
to one medium on data lrom another.
The second require:
.:-g purposefulness inr,:
A dynarnic rnodel of discourse :;rectional, a successi,:,
.'here. We alreadv h;
An integrated description, deriving from a model
of verbal interaction, describes :ncounters and of finisi
languzrge in use as written or spoken discourse. The moclel
than static, and we must consicler this contrast
is dynamic rather .nd the physical rrorlc
in order to relate discourse ."'ithin the discourse.
t_ht
description to traditional clescriptions ollanguage.
The main difrerences between a dynamic and
. special kind ofboun&
a static model of discourse
are The problem of pur:
(l)the dynamic model must srrow how the discourse
proceeds from one point to i,uman behaviour to b,
another,.and (2) the d1'namic moclel must show
howirre component, of ,h. air- :-tample, are readilr' acL
course play their part in the achievement
or some purpose. In both cases, the activity is but one comp
dilrerences take the form of additionar reqrirements
on the dynamic moder. In poseful. The separarion
the first case, the discourse is seen u, u .orrtin,,ous
movement from one state of through discourse from
affairs or ltostuz to another. our habits or
studying language tend to obscure this perform a linking role
 On tlu integration of linguistic desuiption 69

::e is describing. '-,\.ement, because we tend to study language with hindsight; when we are
ronventions and :sidering one stretch of language, we already know what happens next. In
rinq 1>reparation -:h circumstances, the importance of prospection, prediction, and the like is
: .tructural influ- - .: as obvious as it should be.
: deacllines, as in In a dynamic model the elements of structure are described with reference to
iect oltime, and --.i state of the discourse at their point of occurrence , hence the unfolding or
"rstential quality of discourse description. Only language that has already
:i in the case oi , :urred can be taken as given; the description ofsubsequent items oflanguage
r participated in - ionditioned by the state of the discourse at their point of occurrence.

=taric and non- Once the directionality of discourse has been fully appreciated, it follows that
tr different from .-:re is a marked difference in the way we describe previous and subsequent
aion. It can be :r+rage. Each element of structure is considered both retrospectively and
,n. and linguistic . :(lspectively, but in different terms. Looking backward, the following issues are

l ianquzrge is not ':,er-ant to each element: (1)If the state of the discourse includes any firm pre-
anqing record. --itions, does this element offer partial or total fulfilment of the predictions?
are assured that - Ifthe state ofthe discourse includes any prospections (less-than-firm predic-
'lniclue commu- - ,,ns), does this element oIler partial or total fulfilment of them? Looking ahead,
.:,:d to time in a '...e main issue can be stated as follows: (3) What framework
of choice for the
. \\'e can relate :rt element is created by the selection of this element?
anr reading ofit .\ny utterance can lollow any utterance - we are free agents. Although we
'ri a text is to be :nd to follow conventions in social behaviour, there are no absolute rules,
'h riTect upon its ":cause people make mistakes or make use of the conventions for more subtle
':ctics - irony, and so forth. However, each utterance sets the scene lor the
:r\r that is to be -,i\t. No matter what it is, the way it will be interpreted is determined by the
-,rm this
zr line of .er.ious utterances and in particular the immediately previous one. This is fairly
ecription. If the . rrious in question-answer pairs, but is a general feature of discourse. There

. ihe n. :it least at -:e some complexities introduced by the hierarchic nature of discourse struc-
:-ible. The influ- ::re, because, for example, boundary utterances have a more far-reaching role
' Ls \\'e proceed :-an medial ones, but the main point is sound, and the prospective function of
,'ies appropriate ,-1 utterances is of first importance.
The second requirement of a dynamic model balances the first by introduc-
.:rg purposefulness into the description. The dynamic view sees discourse as
:irectional, a succession of changing postures; but it must be heading some-
., l.rere. We already have in the static models some notions of complete

:.rion. describes :ncounters and of finished artefacts in writing. Units olthis kind link language
rlr.namic rather ,.nd the physical world. With the addition of purposes that are recognized
rtlate discourse "',rthin the discourse, they are valuable units in a dynamic model. They provide
, special kind of boundary.
, , ,i
discourse are The problem of purposes in language description is not that people believe
:r-,nt onepoint to :uman behaviour to be largely without purpose. Plans, goals and aims, for
ine nts of the dis- .rample, are readily admitted. The problem is mainly where to stop. Language
l 'noth cases, the .ctivity is but one component of our general activiry which itself is largely pur-
nan-ric n-rodel. In :oseful. The separation of purposes that are recognized as being carried out
i':,m one state of .r-rrough discourse from those that are not so circumscribed requires units that
d r,r obscure this :erform a linking role and relate complete patterns of linguistic activity to
 70 77u organi.4tiln of text

aspe cts of our general social l>el.raviour. He nce the importance of identifyin-e , ];.@ll u_ _-:

unit at this interface.

Linguists are so acclrstomed to describing small-scale stretches of languat. q-r
"_'_
-*-.jl
thzrt the contribution oleach particular to the overall eflect of an artefact mar iliir ll-'il,-

r'vell be missed. In a dynamic model it is necessary to continne the directiona- *! [ r-*i

clescription until a Point is reached rvhere the verbal activity perforrns ir- * :t" :lr: a i:-

its totarlitv some action thzrt lies outside language. Each successive componer,:
has an effbct that ma)'be perceptible in passing but is certainly provisional uni- hfir",:* ::

the artefact is completed and the overall action has been 1>erformed and is n,
longer negotiable except in terms ol a subseque nt artefact. An example fron- i[{]lIll* ft il
conversation is the polite refusal of an invitation that is eventually replaced b., ufll .,.r: : '.:
acceptance ; often this kind of behaviour is within the normal social courresie,
and recognized by participants. In '"vriting it is commonplace for a writer r,_ l.iTllr': ::-
state a position that is contrary to the position he wishes the reader to adopr- r.,1 , _
the eventual effect can only be described rvhen the artefact has unfolded in full. l:-::-
The provisional inteq;retations of purpose are therefore subject to a back- llnr :,:::-:: . :
wash effect from the artelact as zr lvhole; the analyst can make final assignment-.
in review. But in most cases the provisional interpretations are confirmed, and ffi"1 ':. - a: ,i
--:,1
only in manipulative discourse of conversations between naive and sophisticatec &]t-::,.i:
participants is there likely to be substantial reassignment. tr il- -: ----;i -

It is normal, in fact, for speakers and writers to keep a running check on i:,"rll::, :r
what they are doing, and much of this becomes part olthe discourse. There is a -'..."-:
whole vocabulary and syntax of language about language (point, question. [, " _:,::-]::
object. etc.), so that the locus ofthe discourse can shift to the discourse itsell. L. !-',

Less explicitly, many ol the apparently meaningless words and phrases (ufutni.
actualj, zuell, etc.) and devices (e.g., repeating a word or two of a previou. -I-"::-: i-:
speaker) in conversation are signals of how one speaker is interpreting the dis- :' :': - :--: .
colrrse. Unless challenged, these are taken by the participant as signals of the IId

provisional categorization ol the discourse. !fl :;:r ::::

Despite the different circumstances of spoken and rvritten language and rhe i ".-ii-" :-\: -

diflbrent realizations of lineuistic categories, the vier,v of a dynamic model is ro Li :--__r _.-i::'
see them as essentially similar. Both are interactive, both are directional, and - " :q:-::
botli are purpose ful. The description of formal lvritten language is translormed
bv the application ola dynar.r.ric model, because much of its interactive qualitr' 1l: l-:l
is covert. - -::
The fundamental categories of a description that integrates such disparate :. :.::1:l -:'":
_ -- i
behaviour are more abstract than the normal categories of linguistics. But the -

job olintegration must encompass further distinctions as well as those betrveen -' .-*:.a ::::a:

speech and writing. -. -.1-:

Literary discourse is treated as a special case even by linguists r,vho claim that r*.,:ani-::l
it is subject to the normal conventions of language description. The creation of ;:r .:-r ,::
the subject area of stylistics serves to insulate literature from the more mundane j -,\::.f --:-a :

texts of everyday life. Patterns ollanguage that are not remarked upon in non- ---,.:. : : -:
'r---1. .. -
literarv text are invested with meaning in swlistics. rr -A r\ -

Despite these concessions, the literary critics have remained largely alooL . - - :a.:::=.
 On the integration of linguistic destiption 7I

. - , )f identifying a :raintaining that there is a di{Ierence in kind between descriptive and evaluative
.lrdy. Hoiever sensitive and painstaking a description may be, it does not
':. i,c,' ol language .,rgug. with the central issues of criticism. There are other problems, too, about
: artelact ma,V
",11
,,rii.,"i., that can only be mentioned here, such as the lack of principle in select-
--: tlre clirectional .rg a focus of desciiption, the uneasy status of interpretations from stylistic
::'.ir perlorms in .rid..r.., and the di{ficulty of description of long texts. There is little or no
:..ive component .;eory in stylistics; the value of its observations is related strictly to the results of
.-. ur,l'isional until .rdividual studies.
:-- ,:n-red and is no If we view literary text as discourse, we must begin the description in the
.\:: erirmple fron'r :.lrre w2) as with any other text, establishing first of all its location in the world
..-round us. Who is addressing whom, on what occasion, and with what
end in
::.1Iv replaced It1'
i- .,-,cial courtesies ,ierv? The analysis of the utterer shows that we must postulate at least two en-
,-ties - an author in the real world and a narrator in a world of fiction.
Such a
.-- lrlr a writer to
:istinction has been recognized in literary analysis for many years'
r :-lder to aclopt:
,. '.r:rtolded in full. The relation between the author and the narrator is that the author rePorts
:,-i;,iect to a back- ::ie narrator but does not attest the truth ofwhat the narrator says' Frequently
. ::ial assignments .iis relationship is implicit, and the reader deduces from external evidence that
:enrerces like I love you" or'It is the year 2002' are not being averred by
the
.:= ionfirmed, and
: -.:rd sophisticated .,thor to the reader. In the absence of any such deduction, our normal assump-
::on is that anything said to us or written to us is averred by the author or
:-::rning check on .:eaker to be true at the moment of utterance'
:- -,irrSr. There is a The purpose of a literary text is to secure from its readers a complex, evalua-
readers to answer questions like 'what
' ;>, rint, question. -.,.. int..pr.tation; both globally (asking
roes this mean to me?') and analytically (how the components of the artefact
l,: discourse itself.
..ave their several effects). such evaluations occur after an encounter with an
.:.r. phtases (uhum.
-.rtefact and do not need to be articulated'
:11 ,,, of a previotts
So from a discourse point of view, literary text falls r'vell within the
categories
-:r:-rreting the dis- model requires
.lar are already availJle for nonliterary text. The dynamic
r- i: signals of the about
,r.r elaborate evaluation network for the description of any text; the oddity
.-ierary text is that it has no function except to be evaluated. It is argued later
: .lnquage and the
.: this chapter that stylistic evidence is no different from any other linguistic
,:"nric model is to
:r directional. and .'' idence.
, ' 1r i. transformed
An integrated approach to description should in fact be flexible enough to
- -.,r'er all distinctive varieties of a Ianguage ,
not just literary text' In the many
::ii.rzlctive quality
,:r,rdies of varieties that have accumulated over the years, the emphasis has been
1:i::Llch disparate :r distinctiveness and the descriptions fairly ad hoc. If an overall framework
:rn be developed in detail, the varieties can be compared with each
other on
--:.r-ri,stics. But the
^, .,. those between
::iiable criteria.
.{ dynamic model makes available the level of discourse necessary to mediate
-e t\veen form and purpose . Any particular
pattern of syntax or lexis' or a com-
:--'-. r. ilo claim that
,ir.ration of both, may have different functions in different tlpes of text; thus
::.. The creation of
'-r-.- .Lorving the small central organization of the language to be adapted to a wide
nrore mundane
::'itcl upon in non- .rriety of purposes.
\\brk is not yet very far advanced in this area; there are notes on character-
i::-d largely aloof, ..ilc features of a language variety, like the passive voice in formal prose; there
 72 77te organilation o1f teil
are informal expl:rnations of many of the
fleatures. But we clo not have as rr. [g$d!l
anv substantial resezrrcrr on, ror example^
the complementary .lir;ib;;;'.,
features in di{ferent'arieties. R.r.,...h .0onu,
is neecrecl ;; ;; i;r'or.0.,,.u....,
classification of verrieties thzrt accorcls best
r,vith differentiatin*'il;'"; ., .i' :'"
serious deficie,cy is tire lack of a rramer,vork
of interpretation, trirough r.r,hict
of a variety can be related to the generatir"i
i,T,:1li*.".istics
varletv.
p*po..s of thr l'.,'
ll tu
For the present, \\re assume only a few rnajor
received cirtegories such as nar-
rative and expository. writing ancr trrei*pok.n
co',terpaits. Narratives ar.
organized largely bv time expressions zrnd
verb tenses, so if a reacler or listener
has prior knowleclge that he r,vil encounter
a narrative, his expectations are
attu,ecl accordingly. He ass.rnes that the time expressions
are likely to be orga_
nizationirl rather than incidental. on the other
h:rncr, if he knows or quickJr
deduces that he is encounrering expository
language, he is arert to the sentencr
connectors, modal verrrs. ancl the other realizations
of textuar organizatio,s
o- in
expositorv r,vriting. These tend to be in prominent
positions.
There are probably nor many texrs that do
not quickry inclicate their priman
classification, because trre difrere,t orgzr,izations
are rrequentry incompatible.
Ir
lvould be bizarre for example, ror a ,.rrative
to report ut,..r*rlu., such as
'Mr Smith went out for l.nch, or else he had
sandwiches in the oflice,. Such a
sentence can be r,vorked into narrative ";rtrlllfii.ri&.f,.!
uncler protective coverings such zrs:

1 Thronghout his working life

2 I srrppose rlral. . .

But without such an insulation, an author,

r.vhether as factual or fictional narra-
tor, could hardly report alternative events.
A primary classilication oldiscourse types is also
valuable l>ecause no text is
f,lly explicit about its orga,,zation o, it gt", along. and .t*t;-_
many are
It is necessary to accep.r a clefaurt hypothe-"sis, *t i.rl'r,o,.r-rir"i"" very coverr.
,ir. ,osence of
contra-i,dications. a ringuistic item ha, the
same function as its predecessor.
when this is applied in clescription, trre signals of
m:rintenon.. o. c'ange of
function musr be accur.rely identifiable, aid
knowleclge of trre discourse type
helps in the identification. Reference to
chronology, for exarnpt., i,
taken as a signal ofa change in posture in f.i-u a.i.
a narrative text; absence ofsuch rer-
erence by default suggests^a.maintenance
of posture. But in an expository text
or abserce of chro,ological .ef.i"nce is unlikely
l'l,r]"r:r"*
relevallt.
ro be structuralh,

c1n nor'v bring togerher trre elements oran integrated

JVe descriptive sysrem
and apply them to some text examples. At the
present stage of creveropment, the
descriptions are indicative o,.,ly. ani ,,o..
,ro.i..l applicati"on is necessary before
they czrn be claimed to be comprehensive o,r.l
..iinbr.. sr, ttr.y-mrrtrate the
direction of current research.
 On the integration of linguistic destiption 73

: i,-, Irot have as vet .{pplications of the rnodel

".:.,:-' distribution of Basic stru.cture
rls ol delicacy of
-..'.
':,:r.rIing leatures. A - ire minimum free nt of structure in a discourse is the sentence or move
eleme

1---'r. through r,vhich - m), and these are considered equivalent. An integrating definition is attempted
,::rd purposes ofthe .. dr..orrr., but in the meantime we can list what mttst be known about
'ch s/m:
-

i,:l,rries such as nar-

:,,:t.. \arratives are . indications of interaction
J the position of the author with reference to the text
-:. ., reader or listener
--.- ex1;ectations are 3 attribution to narrator(s)
,:. irkclv to be orga- { indications of argument
--. ino\\'s or quickly 5 indications of self-reference
,.-r:'t to the sentence o the dominant verb form.
.. -..r, ,rrg.rnizations in
.: lhese are not in any particular order, nor is it yet clear whether an order ofpri-
:- 1,-.ite their primary :in,will be establied among them, particularly with reference to the primary
.essification of discourse.
=:-:,-. incompatible. It
...:--rnatives such as
,:- ::-.,: ofTice'. Such a In di c ations of inter ac ti o n
', -:lilgs such as:
: tr each s/m, there has to be an identification of utterer and re ceive r, that
is,

:rose who would be referred to by the Pronouns I and 1ou if present. This
::quirement can be rePresented as:

-.,- ,,r' fictional narra-

i VB YOU
of
: -i :--recause nO text iS -. must be understood that the above is a formula in the metalanguage
be paraphrased or generated in
are very covert.
:-.-::-,., .escription. It does not imply that sentences can
:-.-,: in the absence of ::rms of the formula o. thai English happens to have a reliable class of perfor-
::rative verbs to realize the element VB. Although the conceptual
origin of this
:- .,: its predecessor.
misleading to assume anything
::--::-..c or change of .jnd of notation is speech act theory it would be
' evond the basic notion of illocutionary force'
-:' :l.e discourse ty.pe
-"::'.:rlr. is 1>rima facie
. j-r.-rlce ol such re f- . tbk 4.1
-: .,:: exyrository text
Autonomous
,-.'" :,r be structurally Intuactire

- \'B YOU
.: :r.criPtive system
: -: iievelopment. the - ,rtust compliment )0u on lour hat
:- -: irecessary before
i .: :lev illustrate the ! l,rornise YOU THAT I'll come tomorrow at six

!.1VER TO YOU THAT it\ getting late

7 + Tlu organiption of text
1' It is qener
Intheanalysestlratfollow.thedescr\ltiveapParatusisrepresentedasven.
close ro English, but only for clarity. capital letters are used
to identify the appa-
;*:, ,:thor C?rI SrSr
ratus. To aioid amSiguity, the pronorn'I'is rendered as 1if it is part of the text
u*r- - .:ne incluCl
or cliscourse ancl I if it is part olthe metalanguage'l
Peter .a:i
71rc ltosition of the author with reJirence
to the text .\nd *rer.
There is only one aurhor of any s/m. He is the leftmost I in the analytic
displal"
of the -' :-r rihatever :.
The minimum he can clo is ar,cr and that verb is used here as the signal
s/m is placed first in the ana- : .:-al. the n the .
de lault condition. The interactive segme nt of an
-,'_:rndfied nant'
lytic display and is follorved by a transition (usually expressible by THAT)
to the
meaning of T:ris possible I'
remainder of the s/m, rvhich is in the nature of a report in a broad
the term" The analysis ol s/ms by division into interactive and autonomous "r: ,late lv for ea.1"
and is a * at each tran-';
segments corresponds to the same distinction in plar-res of discourse
..lnr.ol feature of a dynamic model. Roughly speaking, the interactive segment
: --adon QUOI
while the
depicts what is going on in the real world at the time ol utterance,
orrono*o.,, ,.gl"ni is a report about something that may include the current
state of the real r,vorld but is ce rtainly not restricted to it'
Any recurrence of 1in the:rutonomolls segment is thus in a report'
An s/m
does not normally return to the interactive plane after leaving it, and then onll' , -i\.ER TO YOL'
for running ,.poi., like obaiousi, or I understanl, which show in analysis
as sePa-

rate from the main strllcture ol the s/m and are shorvn in parenthesis in the
- -\'ER TO \-OL
Interactive column.

Tabk 1.2 - -l\.ER TO YOL'

Intuacttue Autonomous - i\-ER TO YOL

VB YOU . 1\.ER TO YOL'

I .ta) TO YOU THAT you do look haPlry

I AVER TO YOU THAT the book is

.^. j:':aions of argu":
under.rtanfl being rePrhtted
V
-:.e u'ords and p
.-:t
indicators oi
:x.rmples, as are t
Attribution to nanator (s) Each time an
narrator is -. subject matte
An autlror can attribute l)arts o[his s/m to one or more narrators a slte
ihe leftmost I of the structure, nor the subject of the oCe ,
any subject that is not
riglrtmost main clatrse'. He can say. or write:
7 I'd like t
3 IVIany people say that King Arthur actually lived'
-.-.e second ciaust
merely avers -.'.'en.
We clo not know if the author concurs with this belief or not; he
Each prima ia
that others believe it. Similarly,
 On the integration tf'linguistir dest:riptiott 7lt
(l as !.e1.\ I lt is generally supposed that gl:rss is Iragile.
Llre lppa-
,f iiic tert r irutl:or can also report or quote bv attribution to a namecl cirarzrr:te r in 6is
- iroLrrse (including himsell):

I Petrr sitirl lrc rr a, cornirrg.

6 And then I said 'Look herel'
, clisplar'
- rrlrl ol- tht lor rvhzrtever l-eason. ii receit'er der:icies thar the alltorlomolls sesment is fic-
r. tlitr :ut:t- nul. then the ;rnalvsis introcluces a fiction:rl narrertor F; a nonfictionll 6r-rt
-. \l' ro tlic iclentified narrzltor is represe nted bv N.
-:itritrg ol I'his possible la1'erine of nar.rators is recursii..e and so has to be u,orkerl <trir
-)n()tIlou:
rrratelv for eacir texr. ft is entirely in the aut.nofrous segrlrent. L-r this an;rly-
.rnd is u at eacl) tran-*ition.l point tliere is a choice between report (THA"I') a,ci
.e,q11r{:'n t 'rtation (quorE). Note that the orieinal author c.nnot
q,ote hirnself.
'lrili' thc
,-('Llfrellt t,,l.J

-\n s/rn xL l lt,? Aulottanou.s

lrrrr onh
- ils \('])it- \\-ER TO YOLI THAT rtarty ltealtle .rn,y t/Lat li'ing Itthtr Lulualll
'r. in thc ln:cd

],\'El{ "I'O YOU THAT' it i.t generalj suffosetl gla.t.r i.; Jiagile
BYN
\\-ER TO YOU THAT Petcrsaid THAT he urLs comirtg

\\-ER TO YOIJ THAT l mid qllO'fE ktok hnr

\\'E,R TO YOU"l'HA]' I'S]\ID QUOTE Ottcc upon o. titne

;taticns rf argurnent

,('\\rords and phrirses olargun-rent. logical connection. and so on) are imPt,r-
:rt indicators ol chirnges in posrure . But. hlnaeur, ar, and sometirnes rurl, are
-:rrrrples, as itre the lexical paraphrases erl rdditiort. as art al.ternathte, lrt tlte other hand.
Each tin-re an author or narrator tnkes up a notice:rblv cli{ierent attituclc to
r :l tor ls . -ubject rnatter. there is a prima facie case lor a chzrnse in posturc-. In the
,11- thc ]Iellce.

7 I'ct trikr to come but it's very cxpensive.

-,,i,1.u"0 clause is interpreted as implving I utoultln't like to orne" Ibr the reason

Lach prima lhcie ch:rnse of posture is shorvn in the analysis bv zr ncu, li,e^
T {1tr rttqrtttilrtttrtt, t)! t, \l
:tnre' is otI'e lri:
lnhlr 1. !
in s/n'rs. Llcri..
i icle evidelrc.
'r1fn related I,
"!_H..\T j.l ljL tlt ..|ili
I \\'trli"'f() .1t tlie l)lacc '-
ir arcldition t,'
:: qrhvsical c1i:l

i,\\,'lrld 1'() 1'()ir l'H,\'l- .,. ollexical t',

qt'atecl in a ''
,,ttt'i't' iit,titi' t ttt i ;taA't
.r'.e: the lirr-rit.
'

ti('i't it 0!! /t lrial

# t)l
., atnples
: cxanll)lcs tl:'
,,\ u'icie t';rtrgr of rr,,trcis ltttcl Plrrlrscs c()Iltril)l'it('cl'iclcncc of"cli:rttgt'ol Postrtrr
tt
I ,.ss. of the l'-'.
;rrlrlition t{) 11u)se thtrt ere associult'rl ri'ith the collstnlctioll olargttnletrt. .ilfirali't '-n liecatt-.e .,
titil.. atnt,,rt)," of't0tU.;r^ iil lhct._lit- n,trttttftlt;lL'e ( ()tllll'lOll t'xllltllllcs. ' ltdt' rcilsot'L:,] -

:- be iutcgt'ltt--
{ttrlitoliortt ilf tttlf 11'f''," '' . r'rriatiorl tl'
:r: c:In lle illt'
,\ clratrgr in 1;ostule is ;rclrie..'r'cl tl'lletiLlt'r'lttr s/tli is crlrlicitly refi-r'recl lo in tlt' llllr-nance ( r:
rlircc,,.i.ne. ,,\t alr, slnr 1;11rtfl1'r,. lir3re is ar-r 6ptiqtt ol)el] to sll6ltker
9r trlitrt.
.h'sis is grariLr.
fhree exaull -

a He relatcs tlre ncxt,r/rn t0 tlte lrrececling oilr: llr-zrttrrldilig to its assr-lt]ll)- lll conversilL-
tions iutcl l)roslte('ri(lil\. lior t'x:rtrl1-rit". ilaskecl .l (ltlasti()I] lte atlsu'rrs it'
: ch:tngcs it'. '

ir Hr. r.cli,rs t6 f lii' ltn ctrlinq ;/ttl l;r' ;1 llroll0llil crl lt dist:tiltr'e l'oi':Ll;ttlltt'r
Iior -lLIeS alfe 1'ei -.
rr,ol'rl. tlnrs t atir.'tlling tltt' t't-r-luirt'tn(rllt ()l dealing lvitlr its l)r():PC(ti(iIls'
rlttcstiritt'' cl'r.sificz'itir'l.
rx:tnrplt.. if'askcd:L tlitt'siiorl ire sar"s'l'lt;lt's a let-l'itrtrrcstillq irll hieralr r.'

'
Iht,lrnrt ilrtttt lrrb fi,r ti
l: Ilii
trQle
Il nrr>st srlrtrs Lht' r'erlt rii'tlte r-itaitr r-iati.,' is taketl ;ts tlte clonlinatrl otre' Sotne-
tii,r's tlrcre is rnore tlran ont' nr:rirr rlllttse atrcl a cir:Lttqe ol vrlil iirnn: irl sltt ]l C.lromettc,' '
(:t){'s it nrar' }rc lle(('ssil1\' trr iilerltili ttvrt s/l-tts' othcr u.or':,'
t'trl
\.\i"ircrs 6rr lurrrtir-e rt:.g. (ilirrrc-* l1l7.r: Lalroi'ancl \\'aletzkv l9(j7rl)('ilrt Hor'r'er-cr.
tlle irnPortailr.e ol'r'e ri> clioict's in thc stnrcriure- (letrcttc 1!iu0, rtuclies relatir-t nriglit nr.,i -

,,i,.,,n.rI,,gv in cletlil i'tat'r'c ,1!i77, Llscs Yerl) choit'es as tlit' r'ottlet'stone

of lli: .rd(tttiolllll
I
..

a1i;rl1sis,,1l.f,r,,r,.. fut 6tli1r'trire:9l'1'ritrrtq al)d sl)cllkittg" r'erll clrgiccs

allcl relziteci tirat is iclcr.'
inrlications oi'tinting are oiit'n siqrllls oi-1 cilattqc itr postttre' In prirc:tical alliih'sis' mav 21ls() I

rrsua[l' tlre i19st revlalipg liLre to stnl(tr.l]r'i.t thc clisDositiotl o1-llle vrrb
fbnrls'
1)1'()n]Pt tl:(-
,\1r' r'ha11gc ire1.1 et'tr i)I( seill ittlcl 111-tt. sirlrllle ;tltrl ,-,rtrtitlt't6tts' Ill()flal :rtl11 CONPROT
s,r1-pt6clltl. 1tt'r'iect tttr,:1 triltl-lrtrlict;rlltlrvs a cllltlg'' itl pr'stittr" cliskette-t ll. .

'I'lris obsen-urtiort is.iust tlrt' tip of a qt'ittrltrlittitll icclrrlg lr'ltir:h r'"1trir-' Lrnder thc l-
n,.rr.li firrtlier. i.esr.arclt into tltr rciations iit-nr cetr granltrltlr attrcl clist trtit sr' rnents. Hr'','
'l'1t,.'rtrt'. r.oice. ltttrl srrlrjcrt-i-eii'l'r'11t- 1br t-xllrllplc' ilrr tlletllseh'cs ilttt'rrel:rtecl
.,ri r.ltn (.r.etrtc tlle kind ol li'r;,1'icnt:rtiorr associirl.ccl t'ith a cir:rng-e in ilosl.ttrc

,d'
4l ilffi
 On the integraion oflinguistit descriptian 77
?osture' is offered as the linking concept between internal ancl external pattern-
::3 in s/ms. certain configurations of syntax, compared with the previous s/m,
:ovide evidence for an optional or obligatory change in posture. This evidence
.' then related to the current state of the discourse, and a decision is made
.rout the place ofthe s/m in discourse structure.
In addition to the syntactic role of sentence elements, their lexical role and
::eir physical disposition provide further evidence for discourse description. pat-
:rns oflexical cohesion and stylistic devices such as parallelism cannot be easily
:tegrated in a sentence-based description but are important features of dis-
iurse; the limitations of this chapter prevent full treatment of them.

Exatnples
' :,-ltange of posture in t he examples that follow show prospective structuring only. That is only a half,
:. . :-argument. AetuallL. . : less, of the recoverable linguistic patterning. The re trospective links are not
:.:, -. i . ,lorvn because a reliable description ofthem has not yet been achieved. But it is
..ready reasonable to state how stylistic patterns, which are largely retrospective,
-.rn be integrated into a dynamic description of discourse. The principle is that
-nv variation lrom the minimal, straightforward verbal expression of proposi-
: :-:.i referred to in the .-ons can be interpreted as evidence for a change in posture or as evidence for
i : :i)e:tker or writer. :raintenance of posture. From this very general position, the experience of
,,:ralysis is gradually revealing the conventions of interpretations.
.::r::ditrg to its assump- Three examples follow, one from written technical material, one transcribed
.:- , :: ire ansrvers ':om
it. conversation, and one from literature. The analysis is in column layout,
;..i:.ourse vocabulary .nd changes in posture are shown horizontally. Since many of the structural
:-'. -i: |rrospections. For ::atures are recursive, each text prescribes for itself the number of columns and
:. :.:.:rir1g question,. .ire classification of each. The examples are text fragrnents and so do not show
::e full hierarchy ofpatterning up to the artefact.

I ::lnant one. Some_ 'vmple l: Writtm teclmical material

:".:r'b forn-r; in such
Cromenco's CDOS is claimed to be compatible with CPIM version 1.3. In
.-t:z* 1967) point out
other words, CP/M version 1.3 commands are embedded into CDOS.
However, the reverse is not true: programs relying on CDOS's facilities
- t,-ir,r studies relative
:i.- .ornerstone of his might not run under CP/M. In addition, CDOS provides a number of
:-,- lroices and relate<i additional lacilities when compared to CPINI. CDOS uses a file system
r: Ir-. oractical analysis, that is identical to CP/M so any diskette which may be read by Cp/M
q-- , : rire verb forms. may also be read by CDOS. There are minor differences: the system
. :--:-:-.,lous. modal and prompt used by CDOS is a period instead of a > sign. Also, the special
CONPROC (Console Processor) program must be present on all system
-:--:': rvhich requires diskettes as a file. In cDos another version of the PIP program is provided
under the name XFER. It operates essentially like PIP with a few enhance-
T.::t.LInrd discourse.
.'::,--lves interrelated ments. However, PIP can also be executed under CDOS.
-, ltange in posture. (From ZAKS The CP/M Handbook with MPIM, Sybex, 1980)
 78 7-h.e organization oJ'text

Table 1..5
;cmple 2: Conunsatirrn
Inttractiae Autonomou.t
1 1.. Do you lil
Interaction Argument Narrator
2 ,B: It's nice. .
Selfre;feratce
3 A: Doyoufu
I IAVERTO N claims I B: Grorvs - I
YOU Cromenco's CDOS ir
THAT
THAT comfiatible utitlt Cp/ M
5 A: Well r-es
2 urc.ion l.j 6 C: Yesorar,
In otlur uord.t
('P/M ter.rion 1.3 7 A:mm
comnmndr are embedded inlo B B:No-it'sC
CDOS
Hou,eaer
tlu ruer.re is
nqt tra?
I :nb 4.6
programs reQing on CDOS,.t
facilitie.r might not run undtr Inttrartrt
CP/)[
5 In attrlition
CIDOS prouide.s a number of
a,/ d it iona t.fat i I i t ie.r whe n -{ I ASK YOL
comltared b Cp/M
6
CDOSavs...readby George
CDOS
Teah
Tlure are minor dlfferences
Al.ro I AVER TO \-OT
The.yl.rtem prompt ...a)
.ttgil i{ I ASK YOU
9
Tlu.ylucial . . . ar aJile
l0
InCDOS...tlurnme I CHAII-E,NGE
XFER I ASK YOU
u
It ojterate: . . . utirh a.few J.-\ Wellys
enhancernent.s
t2 Houeler
PIP can a/.;o be executed
.C fes
under CDOS I AVER TO \-OL
mm
Notcs
no

li:Ifi)ffl:H:Hill::";i1il,:.i,.illI1iJI.J:,::" llu:inrcrprcrivc pro,,icrns:4 thc I AVER TO YOL-

,' o, l'rorn
anticiparcs a srarcrlrc,lr ni:raa;,io,,rl-i;"r,;;;;:,ffi":;lj::
", to l:r'abourttrc prcvious rinc onc
trris.a riilri.rr.ncr.., tn,r,n, .^,1.
ir is a rt,quirr.rncn,.r.cDos",n. co,)paribirity: is g
a dilli'r.rct. rlris is nor
f.,:.,r,,".t ,. un,..
but not cPll\l: ll
altlrough in lacr

-\otes
N'Iost of the activity is in tht
lr-here a word o[ the preceC
ances rnaintain lhe concen
undcrstand the pronoun r=:
entered only in the interacdr
 79
On the integration oJlinguistic descrtption

',ple 2: Conaersahon
l
I ,4.' Do you like Leicester George
l .B: It's nice. . . Yeah
J ,4; Do you find it grows on You
+ .B: Grows - like a limPet
-' .":.', .)'.\ CDOS is 5 ,4.' Well yes
.,:,-
r"::tbk u:ith CP/l"I 6 C Yes or a wart
',':' 1..:J . ,4: mm
B: No - it's OK. It's um like ' ' '
any big town in the Midlands
.P ."[ :trrion ].3 o

::t. :t(i.\ are entbedded inta

' --r r)

' .i 1.6

r" r:',;.i rcjing on CDOS'.I SeLf-referenu Autonomous

Interacttue
tt:..:.:' rnight not run under
) t{

I} (DO) rou UKE

,-,'t! ltoLides a numbu oJ
I ASK YOU LEICESTER
:: : :.,r,tl facilities ulun
* ": ,ti to CP/),I
George
,J-l]i/.\.lt...rndb1
--,,,r Teah
THAT It's ntce
| :': :iir nfinr dffirtntet IAVERTOYOU
1:ouind THAT
(do) iJ
l-t.,,lntlrompt...a> I ASK YOU IF
gr1us 0n )0u

1",:.tial ...asaJik l ICHALLENGEYOU ON gr0ws

: DO.S...thename I ASK YOU IF IT lS lila a LimPet

.\;aLR
Wellyes
.' r' ;:ir . . . uith afeu:
-- : i),lai?h fes
tt also be execufud
iAVERTOYOU THAT lT IS lile a wart
|.l)o,\
mm

n0

THAT It's OK
r AVER TO YOU lile . . . an1 big
-, i:!\ r' Prolllclrls: 4 thc It's um
- ' ilre prcvious linc ottc toun in the Midknds
,-, ..r corrrpatibilitv; 9 is
: \1. i I although in {act

\IostoltheactivityiSintlreintcractivescglnent,andutterance4slrolvsatypeofsel|relerencc, three utter-

picktd out and talked about"fhe ncxt
i,hcre a u'ord ofthe p"ttdi;;;;;;antti' indicates a shili of topic, and
$'e
in u,,"run..8
iinces rnainrain rhe concern;il';;f;il.;ri t U-t::11::"'5 and 7 are
Leiteste.r in uttcrance
understanrl thc pronoun *f#;;^;i;;-iobe proposlron'
intetuctitt bccausc they do not express a
";;;;;;;ti;;c 'cgment
B0 Tlrc organi4tion of text

Example 3: Literature \otes

. Only the first line (L{\'E
To be, or not to be that is the questionl elemcnt, Ibllowing the :
\,Vhether 'tis nobler in the n.rind to suller sincc therc is no rezil rt,:r

The slings and arrows ol outrage ous fortune, I Each occurrence olX S-

Or to take arms against a sea ol troubles rion would bc to dccide :

Sirnilarly, the X of'2 m"

And by opposing end tl.ren.r? To die, to sleep - uork ofthe analysis rair
No more; and by a sleep to say we end I The division into po.tu:
Tl're heart-ache and the thousand natural shocks lying outside the fbllo*ir
That flesh is heir to. 'Tis a consummation 5 onlv.

Devoutly to be rvish'd. To die, to sleep;

To sleep, perchance to dream.
(William Shakespeare, Hamlet, Act III, Scene i, Collins'Text

In this example , the elaborate recursion of attribution cannot be set out hori-
zontalll,. The lollowing is an alternative layout with notes.

Table 1.7

I AVER TO YOU THAT (Intcractivc)

F SAYS qUOTE (Fictional Narrator)
I AVER TO YOU THAT (scc notcs)

Hamlet SAYS CIUOTE (Narr:rtor)

I AVER TO YOU THAT (scc notcs)

Narrator SelJ:reference Argurnent

x SAYS quorE To be
0r
2 x SAYS quoTE not to be

3 that i.r the que.;tion

+ x SAYS (ZUOTE I{hahcr in the mind to sffir tlu
'ti.s nobler
sling and arrous of outrageous
forlune
0r
(') x SAYS quorE h tala arm.s again.st a sea of
troubles and b-y oltposing end them?
l X SAYS (2UOTE To die, to sleelt
() x SAYS quoTE No ntore
9 x SAYS quoTE and b-y a .rleeP to .stty ue end tlu
heaft ache and the lhou.ran.d
natural shotk.r that.flesh i.r heir to
10 X SAYS QUOTE 'Tis a con.rumrnation deuutf to
be uish'd
ll x SAYS quorE lir db. to sle4t
12 x SAYS quoTE lerchance to dream
 descri'ption Bl
On the integration dlinguistic

u,,,. ,,,. li,:, ,il.-r.lvlll:,"r"H.'X,ii;i:f,TH:llJ:'ili:;".'i::I';.li-:1il:;:''ill

r, nicnr. lbllouinc lh' "3|'il;;,.iinr:: ,rt "'i.t,, rlrcy occur.
.jrrc..tlrcrc is Iro rcal uorld:l(uallur .' X. A lurtlrcr stagc in rnlelpr(ta-
-.,,,,.,,
g|"l:J:ii'iil'llJ:,15:i"''"iq'l"a rnalbc tr or.trcn I2'
racrt occurrcncc orX
SAYS ^i{Tcrrnt
fhc liarnc-
',^" -.",,fa bc to dccidc if tht 6' 7' 9' l0' ii";;;;i""'"''l'ilrc"
si,nilarlv. tht'X ol 2 *'y u"o*'itio'ed-uith

;;:;:i]i:,::;"";., :l,i::lT;:^1,::::::i,il';.'J;;:
couro a casc Dc i"t';J.:H::riJ'['l'J"1,:]
but
ll::-;':';; totto*ing alteinativc'
'h"
i onlY.

-:.,' i. Clollins'Text

::-, ri lte set otlt hori-

i1- .,r notcs)

,itt nhd to n\ffer tht

.nii'.; oJ' oilrageotts

't'tait.;t a sea of'

;,1 apltosmg end thern?

,ti lo s0Y we end tlLe

,irl the thou.sand
i:.i that.lleslt is luir ttt
trntnation demutlt to

lrturt