()\

I
,olt
I ) S'I'UDIES IN SOCIOLINGUISTICS
(
ilnt'litl Iililors:
N rkol:rs
('ottrlltnrl
Arlrrnl .lirwor ski
(
t t r r I i.l.l I I tt i vu: ilt,
lit'tt'rrlly I)rblislrcrl in tlrc Scrics:
Discourse and Practice
New Tools for Critical
Discourse Analysis
'litl|irr,q
ultttl'l'rtttlntt'nl: Rectmntendations,for Rreasf Cancer
ThgO Van LeeUWgn
Atlj r r vt tr t l'l'n t I nt( n I
l;clicir l). Ilobcrts
Lturguage in Time: The Rhytltm and Tempo of Spoken Interaction
Peter Aucr, Elizabcth Kuhlen, Frank Mller
Wltales, Candlelight, and Sruff Like That: General Extenders
in EngLislt Discourse
Maryann Overstreet
A Place to Stand: Politics and Persuasion in a Working-Class Bar
Julie Lindquist
S o c io lingu ist ic s Var iat io n : C ri it:al Il c.f la c t i t t t t.t
Edited by Carmen Fought
Prescribittg under Prc.ssure: l\tn'nl-l'lt\,.si(iutt
('tttn','t.t,ttt,,tt.:
and Antibiotics
'lanya
Stivcrs
()xt,'()t<t)
I)iscourseandPractice:NcwTitols.lor'('rititul I)i.vtttttr.srAttttlt,.si.
r/Nr\.r.r(sl\.r,r(riss
'l'llco
van Lccuwcn
'(xtx
'
fou1, S'?c'r,Ae
e1t
lfrt *&
f'
OXFORD
IINIVI'ISITY PRESS
()xlirtl
[]nivcrsity Press. Inc.. publishes works that fuflher
( )x l(nd t inivcrsity's ohjective of excellence
rr rcs(irch. scholarship. arrd educatron.
TO JeSSe,
orft,rrt Nerv
york
who has just experienced
Arrcklrrrrl Capc
'lirwn
Dar es Salaam lJong Kong Karachi
Kr;rtr t.urrpur Mcjrir! Mctbourne Mexico Ciry Nairobi
hiS fifSt day at SChOOI
New I)cllri Sharrghai fuipei
'foronto
With olliccs tl
^rrcnrinir
Ausuir []rrzil Chile Czcch Republic France Greece
(;u:rtcilrlir
llungary Italy Japan P0land Ponugal Singaporc
S(nrth K(rc Switzcrlard
'fhailaod
Turkey Llknine Vietnnr
('opyriglrt ()
200tt iry Oxford Universiry Press, Inc.
l'ublislrcrl by Oxford Universiry Press. lnc.
198 Mdison Avenue. New York, New York 10016
www.oup.cont
Oxlord is a reeistererl trademrk of Oxford University l)rcss
All rights rescrved. No pan of this publicarion may be reproduced,
stored in a retrieval system, or transmitted, io any fornr or by any mcans.
electronic. meclranical. photocopying, recording, or otherwise.
without the prior pernission of Oxford Universty
press.
l-ibmry of Conircss Catalogirrg in-Publication t)at
l,eeuwcrr.
'l'hco
van.
I)iscorrrsc and prrclicc : new lools fbr critical discourse anrlysis /'[1co van l,ccuwen.
p. cnr.-(Oxlirrd sludics in sociolinguistics)
ISRN 97lJ'O'19-.532330,6r 97lJ 0-19-532i]3 t-3 (pbk.)
l. (iitrcal rliscoursc rnrlysis. l.
'litlc
r,102.r.413 200r{
401 41 ik?? l(X)702:l(XX)
l'ilr'r l
"r
rlr, | ilr,,1',t.rtr.. r,l AnIr,,.,
,rrtrllr,,,r,,r
I)r5(
( )LrRSt ANt) PRACI tCI
irtttttigranls. ln otlrcr words, at home, one can easily single out the few deviant inrlivitl
uals wlro, unlikc "us," Australians, deserve the epithet "racist," and then turn thcn into
thc notorious
rcrsonifications
of prejudice and bigotry which "we" all (and especially
thc rnctlia) lovc to llrlc. Abroad, on the other hand, racism is much more pervasivc.
Anothcr grour ol-social actors who oppose or worry about immigrants and immi-
gration is lirrnrctl by
"us,"
the Australian people "as a whole." This group is more
sympathctically rfctted-lcss olien backgrounded, less often referred to generically,
and classificrl. il'at all, only as 'Australians." If they are activated, it is in relation
to mcntal
l)roccsscs,
such as being "bewildered" and "not understanding," "feeling
Lnrhlc to corc." and so on, rather than in relation to material and verbal proccsses, as
in llrc c rrsc ol thc "racists." And finally, they form a collective, which underlines their
surrosctl c()nscnsus about imrnigration issues.
'l
lrt: inrrnigrants themselves I have referred to as "them," and "they" are rela-
tivr'ly
'l.tcn
blckgrounded and often referred to generically, which helps to distance
tlrc r crulcr. lirln lhern. They are either assimilated or aggregated, and the aggregations
lrelr to rcrlcscnt thcm as a large "horde" about to invade "us," and as the object of
"rirrionul"
calculation, rather than as fellow human beings. They are also representcd
irbstllctly, and this, again, liequently involves the abstraction of their number. More
thrrn uny othcr category of'social actors, they are classified by their "ethnic origin,"
r'lirss, nrcc, lcvel ol'education, wealth, and so on--differences which are not made in
clrr(ion to "us," Australians. And immigrants from different ethnic origins are some-
lirrtcs lurnrcd logethcr in what I have called "associations," to create further categories
ol'rrrigrant. Il'thcy are activated, finally, it is almost always in relation to one activity,
tlrirt of'"inrrnigrating": in every other respect, they are acted upon by others.
'l'hc
govcrnrnent is rarely backgrounded or referred to generically, and it is often
individualiz.cd and nominated, that is, personified in the person of the prime minister.
It also transccnds classification and is always lunctionalized and playing a highly
activc role in relation to the immigrants. The social actors who form the exccutivc
arrr ol'the government, however, those who must actually "stop" the immigrants, are
supplessed: the article keeps the reality of "cutting back immigration" at a comfort-
able distance li'om the reader.
"Experts"
are represented in two ways. Either they are treated like elite persons
(highly activated, tunctionalized, individualized, nominated, and rirulated), or their
utterances are autonomized and/or collectivized, to imbue them with impersonal
authority and a sense of consensus among experts.
The writer of the article also refers to himself and to his readers. The latter ale
addressed directly, the former makes "the fcts" speak in his stead. It is not the author
himself, lbr instance, but "they" (i.e., "these dcvelopments") which "highlight the facr
that racism is seldom far below the surface." Whether or not the writer is also the social
actor who legitimizes the "fears" of "us," Australians, and "entitles" "us" (and Prime
Ministcr Hawke) to our f'eelings of pride, concern, etc., is not clear. Although this
lcgitimizing social actor plays an important role in the process of immigration, refer-
cncc [o hirn or her is always suppressed. Perhaps we are not too far fiom the truth as
wc rccognize here, through traces in the text itself, the highly active role of the media
in this social process, despite the careful stance of neutrality suggested by the way in
which most o[ the representation is attributed to sources other than the writer himself.
Representing Social Action
r
ln this chapter, originally written as a companion paper to the paper on wlriclr
chapter 2 is based, I continue the analysis ol'the
"Race Odyssey" tcxt, lbcusing, this
time, on social actions rather than on social actors, and asking thc qucstion "What arc
the ways in which social action can be representcd in English discourse?"
i. lntroduction
Even a short quote can bring out the critical relevance ofanalyzing thc rcprcscrttation
ol social action:
3.1 Many Australians, thc 1988 Fitzgerald Committee reported, werc "bcwildcrcd" by
the changing face of Australia today. They did not l'eel they understood or could
influence this change. They felt "bcsiegcd" by immigration.
Three kinds of social action are represented here: the action of "immigrating," the
reactions of "many Australians" to this action, and the "reporting" o1'these reactions
by "the 1988 Fitzgerald Committee." But they are notrepresentcd in the same way.
The "reporting" of the committee and the reactions of Australians are reprcse nted as
the doings and f'eclings ol'specific social actors, as observable and tangible occur-
rences. "lmmigration," on the othcr hand, is objectivated, reprcsented as a general-
izcd and intangible "phenomenon" rather than as an action by specific social actors
and as the equally intangible, yet inescapable quasi-natural process ol'"changc." At
the same time, it is also represcnted as affecting the Australian people, as "bewil-
dcring" and "besicging" them. Thus immigration itself remains an unexamined and
56 r)rs( ( )r lRSt AND PRACTtct
uncxalnir)ablc given, while the reactions to it are represented in all of their specifics,
as tlrough tlrcy should be our main focus of attention.
l{cprcscntational choices such as these play a key role in the "Race Odyssey"
tcxt. l'hcy firrm part clf a particular kind of racist discourse, a discourse based on
l'car-l'crr ol' loss of livelihood and loss of cultural identity, fear of the unknown
and unknowablc
"othcr." But they also have very precise grammatical and rhetori-
cul .culizutions:
(hc rcactions of Australians are represented through active verbs
1"rlrt:y lclt . "); thc actions of immigrants through nominalizations ("immigration");
tlrt. relrlls ol'tllc comnrittce are represented through specific speech act verbs ("the
l()llll lritzgcllld Curnrnittce reported..."); and the actions of immigrants are repre-
serrrr:rl tlrnrrrgh abstractions ("change") and metaphors ("besiege"). In this chapter,
I will hring thcsc two clcmcnts together. I will present a descriptive framework for
clitically analyzing nrodcs of representing social action, using critical, socioseman-
tic calcgorics such as "objcctivation," "naturalization," etc., but relating them to the
spccific grarnrnatical and rhetorical realizations which can help to identify them in
tt:xts. In short, I am skctching the outline of a sociological "grammar" of the repre-
scntation of social action-a systematic inventory of the ways in which action can be
represcntcd in English and their import in discourse.
2. Reactions
'l-he
govcrnment officials, experts, "concerncd citizens," immigrants, and others fea-
turcd in the "Race Odyssey" text are represented as involved, not only in actions, but
also in reactions. They have "legitimate fears" about immigration, feel "bewildered"
and "besieged," express "concerns," and so on, and the question of who is repre-
sented as reacting how to whom, or what, can be a revealing diagnostic for critical
discourse analysis. My corpus of "first day at school" texts contains many reactions.
Chilclrcn's stories, for instance, not only try to make children comply with the rules
ol'school but also, and above all, to like school, to feel happy in school. Texts aimed
at parents, similarly, show parents not only what to do, but also what to feel ("don't
worry il'your child cries," "enjoy the time you now have to yourself," and so on).
Sociological role theory recognizes this distinction. A role is not merely "a regu-
latory pattcrn fbr extcrnally visible actions," it also carries with it "the emotions and
atttu(fos th:rt hclong to these actions" (Berger, 1966:113). Role theorists see this as
rcsulting rn()rc or lcss directly fiom perflorming the actions and leave the question of
powcr ()ut ol'consitlcration. Berger cites the casc of the military officer: "With every
sulutc trivcn rrntl ucccptcrl, our man is fortified in his new bearing. . . . He not only acts
likc rrr r,llit'cr, lrc ltcls likc one" (1966: I l4). The children in "firstday" texts, on the
ollrt'r lr;rrrtl,:rrt'srrbjcctctl topower,andthetextsrellectanunderlyingconcernabout
wlrr'tlrr'r llrcy will rrrroliorrrrlly idcntily with their new role and with the institution
rn wlrt lr rl rs
rlrryt'rlorl.
le xts such as these do not attribute reactions equally to all
l)utr(
rl)iurts
('lilrlrt'n's
lrooks tlwcll on the reactions of children, not on those of
l)iu('nt\;ur,l
lr';r, lrt'rs
'li'xts
rrtklcssing parcnts dwell on the reactions of parents and
tlltlrllttt,ltol ott lltrrst ol lt';ttltt'ls,itntl sgtln.AsthepOwerOf SOCialaCtOfSdeCfeaSeS,
llt';rnrorlrt rl r'nrltvr'lt':t liorrs;rllrillulctl to lhctn inercasgs.
RnRlsr NTrNC socrAL ACTtoN 57
How arc rc:rcliorrs
llrrrrrrurticrrlly
rlistinct f}om actions? Halliday's transitivity
thcory (1967- l96l{, l9l'1.5) providcs sornc clear criteria: "mental processes" are, in
E,nglish, distinct lionr tlrc proccsscs that realizc actions ("material," "behavioral,"
and "vcrbal"
rloccsscs)
in {irur ways. First, unlike material, behavioral, and verbal
processcs, they cannot bc "probed" by means ofa "do" question (one can answer thc
qucstion
"What was hc doing?" with "Hc was washing up," but not with "He kncw
that she was coming"). Second, while material, behavioral, and verbal processcs
take the progressivc present, mental processes take the simple present (thus "l am
thinking" would be an action, "I think of you" a reaction; some rcactions can havc
behavioral manifestations). Third, the "senser" of a mental process (the participant
whose mental process it is, who thinks, I'ears, desires, etc.) must be human or, more
precisely, is treated as capable ofhuman mental processes by the vcry fact ofbeing
"senser" in a mental process: pets may be representcd as "scnsers," for instance.
Finally, the "phenomenon" (the object of the mental process, that which is thought,
fearcd, desired, etc.) can be realized by a clause as well as by a nominal group (one
can say "l knew he was coming" as well as "I knew him," lbr instance).
However, these criteria are not always suff icient for the identification of actions
and reactions in actual texts, because they are bound up with the grammar of the
clause, and l'ail to provide recognition criteria tbr actions and reactions realized
at other linguistic levcls, lbr instance, by elements o[ thc nominal group, as with
"unwanted" in example 3.2 and "tolerancc" in cxample 3.3, or across two clauses,
as in example 3.4:
3.2 Australiais...indangerofsadcllingitsclfwithalotofunwantedproblems.
3.3 Racial tolerance is wearing thin.
3.4 They were just a Iittle nervous. Would the teacher be strict?
Also, many reactions are not realized dynamically, by "rncntal process" clauses
such as "they l'eared...," but statically, by descriptive clauses such as "they were
afraid. . . " In other words, the grammatical catcgory "mental process" docs not l'ully
overlap with the sociosemantic category
"rcaction." As already discussod in the pro-
vious chaptcr, Halliday deals with this problom through his theory of'grammatical
'netaphor
(1985: ch. l0): thc sociosemantic concept "mental process" is said to be
realizcd literally (or "congrucntly") when it is realized by thc grammatical category
"mcntal process" and mel.aphorically when it is realized in other ways, for instance,
by a static descriptive clause. Thus, "I fear you" would be litcral, "I am afraid of
you" metaphorical. The two are difl'ercnt, of course, but I would prefer not to plivi-
lege one over the other as somehow more congrucnt with reality, or otherwise more
basic. They artl two dilferent ways ol'reprcscnting reactions, embodying two differ-
ent views of what reactions are, both cqually metaphorical or cqually literal: "Meta-
phors are not things to be seen bcyond. Onc can scc bcyontl thcrn only by using other
metaphors" (LakolTand Johnson, l9lt0: 239).
'fhc
point is to inve ntorize the dif'ferent
metaphors available for rcprcscnting rcaction, insofar as thcy arc critically nontrivial
and can be coupled with distinct lcxicogrammatical rcalizations.
Reactions cantte unspeciferl, through verbs likc "rcact" and "respond" and their
related nouns, adjectivos, and advcr-bs, or stecifted as particular tyrcs of'rcrcti()n,
I rt r I )L t
"
^.]l,
l'li^( |l( I
,rr,l r(, rr.r( ( \\lr( lr lyrt's ol rclction rrc, in a given discourse, attributcd to which
..r,r
r.rl ,r( t.r\ (:rn,;ulrrirr. lrt'rcvclrling. Somc discourses, forexample, in the field ol'
.rrlrt'tr.,nl', r('l)r('s( nt tlrc lrt:llrvior ol'thc "consumer" (the "voter," thc "audicncc,"
r't( );t.,
|lr(
(l.rrrrrrrrrtly rrolivirlctl hy alicctivc reactions, by desires, needs, and wants.
llrc rr';rr tr'n\ \\'lu( lr :ulvt'rliscls,
[)lanncrs
of'campaigns, etc., attribute to themsclves,
.r tlrt' orlrcr lr;rr,1. :rrr' rrrorc likcly to hc cognitive and rational: again, the greater the
rorvr't
,rl
,,o(
r:rl :r( t()rs, tlt: nlorc likcly it is that cognitive, rather than all'ective, r'eac-
Irorrs rvrll lrr' :rllr rlrult'tl l() lltctll.
ll,rllrtl.ry tlrstirrgrrislrcs lllrcc types ol'reaction, again on the basis of grarnmatical
r lrrr'lr,r lrr torrtlrsl tr,t col4ttitive and perceptiv mental processes,
ffictive
mentzrl
lr
( )( ( ,.,,( .,
r :rr t;rkc
"rnrposals"
as thcir ob.iect, that is, they can combine with pert'ec-
rr\ ( n, )rlunt(' t lrrtse s with
"to,"
as in example 3.5, and infinitive or impcrfbctive non-
IrrrrL r l;rurt s witlr thc ing participle, as in cxample 3.6 (cf. Halliday, I985: 23511'.):
{ ',
A rurrrlrt'r.ol critics want to see our intakc halved to 70,000 to 80,000.
{ r, lr,l:rr1, K:rlc likcd bcing the last name on thc register.
li rr r'l)tr\,( rrrt'ntlrl
rroccsscs
are distinct in being ablc to take another kind of nonfi-
rurr r r,nsrrrrttion, {llc "accusative cum infinitivo," as in 3.7, or its "progfessive" ver-
,.r.r.
;r\ rr i l{ (cl'. Matthicssen, 1992: 108):
l. / | lrcrrd lrirn walk away.
'I X I llr:lrrrl hirl walkins away.
('ol'nitivt:
rncntal pl'ocesses can take "propositions" as their object,
"that" clauses (in
rvlut lr "tlrirt" rnay bc clided), as in 3.9 (cl'. Halliday, 1985: 235fL):
t
()
llo
lthc
prirnc ministcrl thinks our current intake is about right.
liirrrrlly, ccrtain key perceptive verbs, such as "sce" and "perceive," and certain atl'ec-
tive vclbs, such as "l'eel" and "fear," can act as cognitive processes and take "proposi-
tiorrs" as the il obicct, as in 3.10:
l. I0 l'hcy did not f'eel they understood ol could influence this changc.
'I
lris introcluces a ditbrent interpretation of the cognitive process, and a dilfbrent
glourrtling firr the knowledge expressed, lending it a pcrceptive flavor, in which "scc-
ing" bccomes like "knowing," or an aff'ective flavor, in which the diflbrcnce bctwecn
"l'ccling" and "thinking" is diminished, so that "f'eelings" can suppolt "propositions."
Othcr dimcnsions of the representation of reactions, such as the dillbrcnce bclwce n
"l l'car..." and "I am afl'aid..." (between what, below, will be called "activltion"
and "descriptivization"), can also apply to actions and will bc discusscrl in lrtc.
sections.
Applying this liamework to an analysis of the "Race Odysscy" tcxl, wc c:ul
obselve, first of all, that thc text contains I | 3 representations of acti()ns (73
rcrcenl )
and 42 representations ofreactions (27 percent): reactions clcarly play r sigrrilit'rurl
RTPRISiNTINC SOCIAL ACTION 59
rolc in tllc wly irrrrrriglation is rcprcsentcd hcrc. The vast majority of these reactions
is attributccl to "Arstralians" (-52 percent) and their government (21 percent) and to
those who rrc "conccrncd about" or "resentful of'' immigrants in other countries ( I 9
pcrccnt).
'l'he
reactions ol'immigrants, on the other hand, are not represented. Immi-
grants only provoke reactions. Their thoughts and feelings and observations are not
considered relevant. The tcxt is squarely written from the point of view of those who
have, or aspire to have, the power to regulate the "intake" of migrants, to "cut" it,
"halve" it, and so on-even though the text also tells us that 40 percent of Australians
were either born ovcrscas or havc at least one parent who was.
There is a distinct pattern also in thc distribution of the different types of reac-
tions. Anti-immigration reactions in countries other than Australia are for the most
part atfbctive, cmotional, unreasonable. Anti-immigration reactions within Austra-
lia, on the other hand, are as often cognitive ("believe," "consider," "think," ctc.):rs
all'cctivc (although scnc of the cognitive reactions have an affective "llavor"): thc
"legitimate l'cars" of Australians are not portrayed as irrational f cars, but ns r((t\()tt
able l'ears, emotions checkcd and held in balance by rational considcrltions, lhis
in contrast to tlrc crnotions of the true "racists" in other countrics. "llcsclltrttcnl"
is reserved lbr Bruce Roxton. the racist the Australian media love to halc, tts tlrt:
exccption to the rule that Australians are not racist and have only rcasonablc untl
undelstandablc
"conccrns" about immigration. The Australian
Sovernmcnl's
rcac-
tions, finally, are lirr thc most part cognitive, and the affective reactions attributcd t()
it are positivc ("hopc, "bc disposed to"), this despite the fact that the article criticiz.cs
the governmcnt's immigration policy and in keeping with the general pattcrn in thc
media that "cmotivcncss" dccrcases as status and power increase.
3. Material and Semiotic Action
Social action can be interprered as mctterial or semiotic, as "doin8" or as "meaning,"
in other words, as action which has, at least potentially, a material purpose or ell'cct
or as action which does not. Examples 3. I I and 3.12, for instance, represent the same
stage in the "first day" scenario, the child's rebellion against entering sclrool for thc
hlst time, but where 3.1I "materializes" it,3.l2 "semioticizes" it:
3.1 I Darrcn lcsistcd thc teachcr's attempts to settle him.
3.I 2 "We don't want to sit down." said Masnus. "We want to go outsidc."
'l'he
same contrast can be observed in these two versions of a speech act by
Nclson Mandcla, both liom the Times (7 February 1994): in 3.13, it is material-
iz.cd, rcprcscntcd as action ("attacks"), in 3.14 semioticized, represented as meaning
( "tlcscribcd"):
l. l l'l'hc ruling National Party has urged Mr. Mandela to end his attacks on President de
Klcrk.
.l l4 Mr.MandeladescribedthePresidentasaweaklingwhohadnotraisedawordofpro-
lcst against right-wing attacks on ANC offices and carcd nothing for black lives.
()0
I)I5('O(JRSI AND PRACTICI
Again, thc point is not that one is metaphorical and the other literal, but that the same action
is rcprcscntcd in two different ways, each conveying a distinct attitude toward, and interpre-
tation ol, the action; the one equating the power ofwords and deeds, the otherdistinguishing
bctween the semiotic act of "describing" and the material act of "attacking ANC offices."
The distinction between material and semiotic action can be related to specific gram-
matical rcaliz-ations (see Halliday, 1985: ch. 5). "Verbal processes" occupy an intermediate
position betwccn material and mental processes, between actions and reactions. On the
onc lranrl, thcy rescmble doings (one can answer the question "What did she do?" with
"Shc
is saying hcr prayers"). On thc other hand, they resemble cognitive mental processes,
irr tlrat tlrcy can takc a "proposition"
as their object, which is not the case with material
l)r(xrcsscs
(onc cannot say "Mr. Mandela attacked that the president was a weakling").
Matcrial actions can bc transactive or nontransactive. The former involve two par-
ticipants, thc "actor," the "one who does the deed," and the "goal," "the one to which
the proccss is extended" (Halliday, 1985: 102-5). The goal, according to Halliday, must
be a "thing," that is, "a phenomenon of our experience, including, of course, our inner
experience or imagination-some entity (person, creature, object, institution or abstrac-
tion); or some process (action, event, quality, state or relation)" (1985: 108). Nontrans-
active actions involve only one participant, the "actor," who in thc case of "behavioral
processes," must be human. This distinction is not just a neutral, grammatical one. It
distinguishes also between actions which have an effect on others, or on the world, and
actions which do not. In "first day" texts, lbr instance, children's actions are mostly
nontransactive: children are rarely represented as having an effect on the world. They
may play, run around, whine, cry, draw, or sing, but who they are running to or playing
with, what their whining and crying achieves, or what it is they draw or sing about, is
not represented, at least not in "first day" texts aimed at adults. Children's actions arc
portraycd here as mere behavior, without content, purpose, or effect, and this is not so
in thc case of teachers, whose actions are almost invariably represented as affecting thc
childre n in their care. The actions of migrants in the "Race Odyssey" text also tend to
bc nontransactive: they "immigrate," "anive," constitute a "mere trickle" or "an influx,"
rnrl so orr. Clcarly, the ability to "transact" requires a certain poweE and the greater that
powcr, thc grcater the range of"goals" that may be alTected by an actor's actions.
llallidly's description of the "goal" of a material process is very broad, and it
rnrry bc cliticllly rclcvant to distinguish bctween transactions with things and trans-
Irclions witlr pcoplc. Grammatically, "material processes" may treat both as "things."
('r'iticrrlly,
it is irnportant to distinguish between actions which affect people and
;rt'liorrs wlriclr allcct other kinds of "things." I will refer to the former as interactive
;rrrf fo tlrt' lirtlcr rs instntrnental transactions. In the case of interaction, the action
rs rt'lt lt'tl lo lry nrclns oia verb which can only take a human goal, as with "hug"
rr \.15 :rrrtl "tlt:ny cntly" in 3.16; if the goals of such verbs are not human, thcy
t;rrr rrsu:rlly lrt' irrtr:rrrctcd as metonymical displacements (as when someone "kisses
llrt'I'rorrrrrl" ort wlit'l s()rnc()nc else has stood), instrumentalizations of social actors
(r'1' rr ( \l)r('ssr()rrs likc "thc bullet killed him"), or projections of human social
llr:r{
tr{ r.., orrto tlu'lrt'll:rviol ol'lrninrals, plants, and even inanimate matter:
I l
"
l\'1,I v l\.ttr' t:It to lr[l ;rrrl lrrrgl',ul hcr.
IIr, I'r.r'trlr'('l ,\.,t,rn(l(.\r('ilt .ll:rvt'ltt.t'nrl|tnicdentrvtoelegantrestaurants.
RTPRTSENTINC SOCIAL ACTION
ln ittstrumrntul transactions, the goal may be either human or nonhuman. In
othcr worcls, instrunrcntal transactions represcnt people as interchangeablc with
ob.iects, lirr instance, through verbs like "use," "transport," "destroy," "carry," etc.
Not surprisingly, instrumentalizations are common in texts which are to some degrcc
bureaucratized (as with "make use of other children" in 3.17) and less common, lir.
instance, in stories (cf. 3.18), whcre the transaction might, literally and figurarivcly.
be rcpresented as interactive, dialogic:
3.17 Make use of other children to help him get dressed or usc scissors.
3.18 "Susan," called Miss Laurie, "show Marv Kate the doll's housc and all the othcr
things."
The same distinctions apply to semiotic action. Semiotic action can be transactive (as
in "she addressed him") or not (as in "she spoke fbr an hour"). Again, the actions of
lower-status actors are more often represented as nontransactive:
3. I 9 When she woke up the children were singing.
The instrumentalization of scmiotic transactions is rcalized through verbs ol-
"exchange" and "transport," e.g., "give," "offer," "receivc," "providc," "supply," "con-
vey," "put across," usually together with some kind of specification of the kind of
speech act involved (e.g., "information" in 3.20) and possibly also of the contcnr
conveyed by the semiotic action (e.g., "about new entrants," also in 3.20):
3.20 More than half of our samole schools received information about new cntrants lhrr
the preschools.
But semiotic action can involve an additional dimension, the very dimension which
makes it semiotic: it can convey meanings. When semiotic action s behavioraliz.ed,
this dimension is not represented, and semiotic action is treated as similar to othor
forms ofaction, divested ofits ability to reach beyond the here and now ofthe corn-
munication situation, its ability to represent the "then" and the "there," to remember
lhe past and imagine rhe future.
When semiotic action is not behavioralized, the meanings conveyed by
the semiotic action are also represented, resulting in embedded representation,
representatiot-within-the-representation. This embedding can take a number of
forms. It can take the form of the quote, in which case it includes not only rhe
meanings conveyed by represented social actors but also their wordings. In texts
like "Race
Odyssey," this is typically reserved for high-status actors (3.21) or used
to enhancc thc credibility of an embedded representation. When there is a choice
bctwccn quoting or not qu()ting, quoting may imply something like "l could not
havc sairl it bcttcr rnyscll'" rr "No lrther comment or interpretation is needed" as,
rcrllrps,
in f .22:
.1.21
"l
lrrrx.thrrt :rswt'lrorr,"lrt.
ll,rirrrcMinistcrllawkcl saidrecently,"thatwemaybe
Irblt' lo look lrt lrirlrrr lt.vt.ls ol irrrrnignttion."
62 l)ts( ( )t rf<st ANt) PRACTtCt
1.22 "N:rtive Virncouverites will be made to feel like strancers in their own citv. . .." wrote
onc rcador ol-Thc Province newspaper.
Many "first tlay" texts arc procedural, telling parents or teachers what to do on
thc "first day," and whcn such texts include quotes, therc is often an unspoken sug-
gcstion that thc rcprcscntcd wordings should bc adhered to if the semiotic action is to
be ef fectivc, rs in this
riccc
of advice to parents:
3.23 Yet by bcing asked specifics such as "Did you make a new friend?" or "What was
thc vcry best thing you did today?" your child may find it easy to launch into a vivid
tlcscription of their day.
lictulitiott does not include the wording and is realized by reported speech:
t.24
'l'hc
mayor of Kawaguchi has'Joked" that with so many dark-skinned foreigners in
town, Japanese are having trouble seeing them at night.
'l'ho
content ol'a semiotic action may also be conveyed in a more abbreviated
lirrrrr. lry specilying either the nature of the signifier (form specification), as with "a
nrotlicrnr of inlbrmation" in3.25, orthatof thesignified (topic specffication),aswith
"ilrorl
lllc first day" in3.26:
.1.2-5 livcry school requested a modicum of information from the parents.
_J.26 l)arcnts should make a point oftalking about the first day.
'Iirric
srccif ication is typically realized by a "circumstance of matter," a phrase with
"lrout," "concerning,"
or some similar preposition (Halliday, 1985: 142), such as
"ahout the first day" in example 3.26. Form specification is realized by some term
dcnoting a kind of speech act (e.9., "ioke," "story," "lesson," "nonsense") or commu-
nicative act using some other semiotic mode (e.g., "song," "drawing," "diagram").
When not accompanied by a quote, rendition, or topic specification, it is in fact close
to behavioralization and no longer represents the meanings conveyed by the semi-
otic action. This often occurs in the representation of semiotic acts by lower-status
social actors, lbr instance, in the representations ofchildren's talk in "first day" texts
addlessed to adults.
Applying the distinctions introduced in this section to the "Race Odyssey" text
shows that the key categories of social actor featured in the text are represented as
involved in different types of action. Migrants are only twice represented as involved
in semiotic action (the Pakistanis "tell lies under the name of Allah," thc Sikhs
"press for the right to have Mounties in turbans in Canada"); 94 percent of thcir
actions are material and almost all of these are nontransactive (86 percent): migrants
just "anive," "immigrate," "trickle in," and so on. Yet, as already mentioned, thcsc
actions are nevertheless also represented as having an effect on Australians an<l
on "conccrned citizens" elsewherc, albeit in an objectivated way: what affccts thcsc
"concerned citizens" is not the migrants themselves, as identifiable and spccific
actors, but "migration," a vague and intangible "phenomenon
rir lrir \l N r rN(; s()clAL ACTTON 63
'I'hc
Atslrlrlilul govr:rruncnl, loo. is nrrrslly n:rrcscntcd as involved in material
:rction (l{7
rcrccnt),
rrrrtl its lctions alc
rlctlorrriurnlly
instrumental: it "takes in,"
"prognrrns," "lralvcs," "culs," "slol)s," ctc. lts rclation t() nrigrants is represented, not
as intoraction with pcoplc, but as a calculated, mcchanical operation upon people.
"We," Australians, and the "racists" in other countries are represented as involved
in matcrial and semiotic action, though somewhat more in the former (69 percent
material and 3l percent semiotic in the casc of Australians;60 percent material and
40 percent semiotic in the casc of the "racists" in other countries). The actions of
the Frcnch, Japanese, Peruvian, and Canadian "racists," however, are not instrumen-
tal like those of the Australian government. Thcy are actual interactions. "Racists"
in other countries "attack," "deny entry," "insult," and so on. The actions of "us,"
Australians, on the other hand, are more abstract and rarely interactive. Sometimes,
nontransactive (they "sit," "read," "dcbate"), at other times more or less instrumental
(they "avoid problems"), they always take place at the level of "public opinion," of
surveys and opinion polls, ofreading newspapers and contributing to "debates"; this
is in contrast to the actions of "racists" elsewhere, which are represented as actual
acts of racism.
Two categories of actor involved in the "debates" are represented as only contrib-
uting semiotic actions: the "experts" whose findings are rendered by the writer, and
the writer himself', who "italicizes," "highlights," and "calls into question" and who
reprcsents his own words as "facts" which teach us a "lesson." The only reference to
him as a person occurs in the byline ("argues David Jenkins"), but the nature of his
speech acts is reprcsented throughout: despite his careful adherence to journalistic
practices such as attributing anything that could be interpreted as opinion to named
"sources," the article is not a neutral "report," but an "argument" and a "lesson."
4. Objectivation and Descriptivization
Actions and reactions can be activated, represenfed dynamically, or deactivated,
represented statically, as though they were entities or qualities rather than dynamic
processes.
Whcn activated, the actions or reactions arc grammatically realized in the verbal
group of a non-embedded clause. When deactivated, they are realized in other ways,
to be discussed below. From the point of view of a theory of representation, however,
they remain represcntations of actions and reactions.
Objectivated actions arc realized by nominalizations or process nouns which
cithcr linction as suh.ject or object of the clause (as with "migration from traditional
sourcc corrrlrics" in 3.27) or firrm part ol'a prepositional phrase (as with "immigra-
tiorr" in
'l.2tl):
1.27 MiSrrrtiorlirnrtrrditional sourcccountricslikeltalyandGreecehasdriedup.
I .)ll
'l
lr('y t':rnrrot voit't: lt:gitirnatc lcars about irnmigration without being branded racist.
(
)lr jct liv:rliorr t'rrr ;rlso lrc rcalizctl rnctonyrnically, by various kinds of displacement,
lor irrstrrnct'. lty tt tttttrtrttlii.ttlitn,llrc subslitution of the tirne associated with an action
()r (.r( f f ()n lor tlrc ircliort or rcaction itself (as with "about the day ahead" in 3.29);by
.slttttittli..tttiutt, tlrc substitution of a place associated with an action or reaction lbr the
rut t ion or rclrctiorr itscll' (as with "school" in 3.30); or by various forms of'prolepsls,
Ior instrrrcc, tlrc suhstitution of the product of an action or reaction fbr the action or
rcrclion itscll'(ls wilh
"no parting tears" in 3.3 l):
.i.29
'l'hc
lirnrily can talk calmly and happily about the day ahead.
.1..10 School isjust about to begin.
.l-3 | I sw no parting tears.
'l-his
docs not cxhaust the possibilities. Another form of displacement, common
in "psychological" fiction, is the displacement of reactions onto the weather, as in thc
lollowing quote fiom a detective novel by Georges Simenon (1979: I I l), where the
first clause objectivates the reaction of the hero, and the second the hero himself', by
substituting "the rain" for the hero, as "senser" ofa (descriptivized) mental process:
3.32 lt was still raining the following day. The rain was sofi, cheelless and hopcless, like
a widow's tears.
When an action or reaction is objectivated, the representation downgrades it in order
to givc priority to something else. In procedural texts, for example, priority might
be givcn to secuencing, as in 3.33, or to signaling whether an action or reaction is
optional or obligatory (or something in between: "helpful," "advisablc," "dcsirable,"
ctc.) as in 3.34:
l.l3 l'reparation firr thc first day at school should start early.
3.34 Mccting the teacher is also important.
ln lcxls wlriclr.juxtaposc scveral representations of the samc practice, priority might
bc givcrr to rnrxlality:
1. l5 lt worrlrl sccr)l lhll il is logistically possible to include mothers in the classroom.
Vt'ry licttrcrrtly, ltowcvcr, obicctivation scrvos to add purposes and/or legitirnutions
f() llrc rcl)r'(rscnlalion. Ilctc is an cxarrtplc lioln Illich's Deschooling Society (1973: 5l):
1 {(r Alit'l:rtior wlrs u dircct conse(lucncc ol'wage labour which deplived man of the
orxrrtrlrity l() ('r(:tlc and recrcate. Now young people are pre-alienated by schools
tlrrl isol:rlc tltcnr whilc thcy pretcnd to be both producers and consumers ol'their
owrr LrrowlerlLlc, which is conceived of as a commodity put on the markct in
st llrtll
'llrt.
lrsr t l:rrrst'ol tlris cxccrpt contains two objectivations: "alienation" and "wagc
l;rlrour
"
l lrlrt olrt'r'livirtions allow the two actions ("working fbr a wage" and "alien-
;rtilrll') to lx' lilrkt'tl lo crcll othcr by a causal pfocess ("was a direct conscqucnce
ol
")
llrrs. tlrt' nt'l';rlive connotation of'"alienation" can transf'er to "working filr a
RI PRI STNTINC SOCIAL AC IIO\ 65
wagc" anil, by rrrcans ol'an irnplicit comparison, also to schooling, Illich's main topic.
In othcr wortls, verbs dcnoting logical relations can link objectivated actions to
each other in order to evaluate and thereby legitimize or delegitimize them. In texts
like Illich's Deschooling Society, legitimation and delegitimation are in fact the
ovcffiding concern, and lepresenting the social actions involved in "schooling" takes
a back seat. However, legitimation and delegitimation cannot stand on their own.
They must be related to a representation of the actions and reactions which thcy
legitimate or dclegitimate, however reduced, generalized, and abstract this represen-
tation may be.
Objectivation also allows social actions to be classified, labeled. The objecti-
vatcd action or reaction then premodifies another objectivation which abstracts an
aspect or quality liom that action, for instance, its "ritual" or "fbrmal" or "strategic"
nature:
3.37 Entry procedures are largely a malter for the head teacher to decide.
3.38 Local education authorities vary in their admission policies.
3.39 Childrenlearnmostof whattheirteacherspretendtoteachthemfrom...merepar-
ticipation in the ritual of school.
This, too, can play a role in the realization of legitimation. To call an action or set of
actions a "ritual," lbr instance, delegitimizes itlthem, bringing negative connotations
of "emptiness" and "hollowness" to bear on the action of "schooling."
Actions and reactions can also be descriptivized,fhat is, they can be represented
as more or less permanent qualities of social actors (as, for example, with "smiling"
in example 3.40 and "specially trained" in example 3.41) or of other elements of the
represented practice, e .g., the props required for its enactment, as wilh "her favorite
vidco" in 3.42:
3.40 A snriling teacher met them at the door.
3.41 A spccially trained squad of teachers will go into homes to show parents how to
prcparc their children fbr formal education.
.1.42 Aquictwayofcclcbratingherfirstdayatschool mightbeaviewingofherfavorite
vidco.
Ileactions arc of'te n dcscrintivized in this wav:
3.43 ls...this nation... irlpcrvious to racist sentiment?
3.44 Wc arc cnli(lctl to bc proud. . . about our immigration program.
Descriptivizrrtirrn c:ur lx' crrlizctl by cpithets, as in 3.40, 3.41, and 3.42, or by the
attributcs ol'rlt:sc'rirtivc r'lrnrscs rs, lirr instance, in 3.43 and3.44.
In thc
"lt:cc (
)rlysscy" tt'xt, only -53 pcrccnt of actions and reactions are acti-
vatccl. Iltt llrc iun()unl ol tlt':rclivirlion is nol constant throughout the text. In the first
twclvc lilrt:s, tlrt'st't lron wlriclr tlt:;rls with tllc bchaviorof the "racists" in othercoun-
trics, ucliv:rtion is nrut lr rrrort' lrt'trre rrl (74 pcrcent). The actions and reactions of
llrcsc
"r'it'isls"
;rrt' r:rrt'ly rlt lrt livlrlctl: irr lil perccnt ofcases, they are placed squarely
66 I)Is(]OIJRSI AND PRACTICE
in tlrc lirrcground of'the text. The actions of the writer of the article and of the various
cxl)crfs wlrosc rcports and surveys are quoted are also predominantly activated.
"Wc," Australians, are activated in 54 percent of cases. Objectivation occurs
nrostly in rclalion to actions and reactions that could be interpreted negatively: "fear,"
"disharrnony," "tlislocalion," and so on. Descriptivizations, which represent actions
rur rerctions its rn()rc or less permanent qualities of social actors, occur most often
in relation to rclativcly positive reactions (see 3.43 and 3.44 above, for instance).
Activation is tlrus used to foreground the rational and nonracist nature ofAustralians'
conccrns about inrmigration ("believe," "think," "feel") and to background their basis
in ltclings ol'insecurity and fear.
'f 'lrc
rrctions and reactions of the Australian government are activated in 44 per-
t'cnr ol clscs, rnainly in relation to semiotic actions ("say") and reactions ("think,"
"bclicvc," "lropc"). The government's material actions, on the other hand, the actions
wlritlr nratcrially affect migrants ("program," "cut," "stop," etc.) are objectivated,
lrrrt'kiroundcd. As in the case of "us," Australians, descriptivization is used in rela-
liorr lo
xrsitivc
rcactions ("is confident," "is disposed to high migration").
Lclst ofien activated, and hence most backgrounded, are the actions of the
nri'.r'rrnls llrcmsclvcs. They are activated in only 20 percent of cases, mostly in thc
f irsr rwt:lvc lincs, that is, in their interactions with "racists" in othercountrics, where
tlrcy
"tr:ll
lics,"
"prcss
for the right to have Mounties in turbans," "bring bulging
wrrllcts," ctc. A l'cw of thc deactivations descriptivize their actions, represent their
:rt'liolrs lrs nlorc or lcss pcrmanent characteristics-negative ones: they "are users
ol wt:ll:rrc," "havc high unemployment rates." But most of the deactivations are
olrjccliv:rtions of'that onc crucial action, rcpresented in all of its generality: "immi-
gnrtion."
5. Deagentialization
Actions and reactions can be agentialized, represented as brought about by human
agency, or deagentialized, represented as brought about in other ways, impervious
to human agency-through natural forces, unconscious processcs, and so on. I will
distinguish three types of deagentialization: eventuation, existentalization, and
naturalization.
In the case of eventuation, an action or reaction is represented as an event, as
something that just "happens," without the involvement of human agency. The ques-
tion "by whom?" cannot relevantly be asked in connection to it. One of the ways in
which this may be realized is through a material proccss denoting involuntary action,
as in the case of "lose" in this examole:
3.45 Parents lose key role in policy making.
Many such processes have an inherently passive sense (e.g., "undergo," "cxpcricncc,"
"suf1'er"), even though, grammatically, they are active:
3.46 They receivcd a sudden cold-shoulder from neighbours and co-workcrs.
Rt PRt st N l rNc socr^l Aclo\ 67
Anothcr possible realization combines objcctivation with generalized proccsses suclr
as "happen" or occur":
3.47 Such palicipation does not seem to happen at senior school.
3.48 Signs of fatigue occurred regardless of preschool experience.
Many of the things we might ordinarily think of as events can also be representcd ls
actions, and vice versa. Compare, for instance, the following two examples;
3.49 Cohenfoundthat,atthebcginningofthetrial,thesubjects'eyeswerefixitedatthc
center of the screen. When a small object was presented at thc left or thc right ol tl)('
screen, subjects moved their eyes to read this digit.
3.50 When an object begins to move across the field of view, the eyes will begin [o n)(]v('
after it after a time interval of about 0. I 5 seconds.
Example 3.49, an extract from a scientific paper, represents human action ("r srnall
object was presented") and their actor ("Cohen"). Example 3.50, an extract ll'()rn rn
introductory textbook, represents an event (the "object moves") and removcs lhc
traces of the human experimenter/observer without whose actions, in thc cnd, no
representation of natural events is possible.
Conversely, human social practices can serve as a model for representing naturll
events. Susan Sontag (1979: 64) has pointed out how the representation ofcanccr rs
frequently modeled on military practices:
Cancer cells do not simply multiply, they are "invasive."... They "colonize" liom thc
original tumor to far sites in the body, first sctting up tiny outposts ("metastases").
And, as Freud (1975
[901]:
221) has shown, what most of us think of as chancc
accidents can also be represented as brought about by human agency:
[Bungled actions] prove to be govemed by an intention and achieve their aims
with a certainty which cannot in general be credited to our conscious voluntary
movements.
The English language provides resources for representing proccsses as eithcr
actions or events. But it cannot dctermine which phenomena should be intelpreted
as actions and which as events. The same occurrences can readily be represented
either as actions or as events, either as voluntary ("the subjects moved their cyes")
or as involuntary ("the eye moves"). Both options are, in principle, always open.
Contextual, not ontological, factors dccide which is taken up in any given instance.
This is why I use "transformational" terms in both cases, fbr agentialization and
lirr eventuation.
In the case of existentialization, an action or rcaction is represented as somcthing
thal "simply exists." The action or reaction itsell'is objectivated and fills the slot of
thc "cxistcnt" (thc entity predicated to exist) in "existential clauses," clauses which
rrsscrt lhc cxistcncc of somcthing and frequently begin with "there" ("there is...,"
l)r'|J(
( )l rli"l ANI) I'RA( TICE
"lhclcexisls..";
llllliday, 1985:130),asin3.5 l.Altcrnatively,existentializationis
lcrrf izctl irr olr.jccf ivalcd ways, e.9., through nominalizations like "existence" 3.52:
-1.51
'l'hcy
bclicvcd that the immigration program existed for the benefit of politicians,
btrrcaucrats, and the ethnic minorities.
.1.52 l'he vcry existencc of obligatory schools divides any society in two realms.
ln thc crsc of naturaLizatior?, an acton or reaction is represented as a natural pro-
ccss by mcans of abstract material processes, such as "vary," "expend," "develop,"
ctc., which link actions and reactions to specific interpretations of material pro-
ccsscs-to discourses of rise and fall, ebb and flood; of birth and death, growth
and decay; ol'change and development and evolution; of fusion and disintegration,
expansion and contraction:
3.53 Tolcrance is wearing thin.
3.54 All that is changing.
Again, objectivated naturalizations are common. The following example com-
bincs existcntialization ("a change is coming") with an objectivated naturalization
("changcs"):
3.5-5 With thcse changes is coming a change in community attitudes.
Naturalization is the most common form of deagentialization in the "Race
Odysscy" tcxt, and thc writer uses it to give his predictions an air of inevitability:
racisrn is 'tlcvcloping," "dcbatcs" arc "building," and, the most frequent naturaliza-
tion, "changcs" arc "coming."
(r. (
icrrcr;rliziltion iln(J Abstraction
I)if li'r'crrt rt'rrcst:rrlirlions rn:'y
4eneralir
actions and reactions to different degrees.
WIrrrt in l
-5(r is
lcrrclllizctl
("rrrilk timc"), for instancc, is in 3.57 broken down into
scve rirl nrort'srt'cilic irctions ("gutlrcring on thc rug," "naming
a color," "fetching the
rri lk" )
i\lr Mrlktirt'istrcrrlt'tl :rslsrccificcvcnl.insomcclassesandpassesalmostunnoticed
rr ollrcr s
t 5 /
'l
lr' wlrolt' t lrrss is grrlhcrcrl on thc rug. The teacher names a color and children who
.rrt rvr':rrrrrr', rt lt'lr'lr llrt:ir lttilk.
llrkrn' tlrt' trvo t'r:rrrrrlt's loictlrcr,Onc could construct a taxonomy (table 3.1). But
rl rvorhl lr' :r lirrorrorrry ol ir
rcculial
kind. In most taxonomical diagrams, the terms
;rt tlrr. l,rvcr rr,rrlt's rvorrLl lr:rvc lo bc rcatl as subordinate: "gathering on the rug," for
cr:rrrrrlr',rvoultl lr:rvt tolx'rt'url lrs"r kintl of'rnilktime."Here,however,"milktime"
nrilk time
RTPRESENTINC SOCIAL ACTION 69
gathenng on thc rug n naming a color n fetching milk
rllE 3.1. Temporal Composition Taxonomy
generalizes a sequence (this is indicated by the carets in the diagram). The diagrarn
can also be seen as a meronymical or "composition" taxonomy in the sense of Mar-
tin ct al. (1988: 149) exccpt that "composition" is here temporal rather than spatial.
"Gathering on the rug," "naming a color," and "fctching the milk" are parts of (mero-
nyms of) "milk time." Such temporal composition taxonomics can bring out how
thc micro-actions which would make up "fetching thc milk" (gctting up liom the
rug, walking to the table where the milk is displayed, etc.) constitute "actions," how
actions constitute action sequenccs or "episcldes," and how sequences of episodes
constitute "practices" (cf. Barthes, 1977: l00fT.). But it is difficult to provide linguis-
tic criteria fbr recognizing the level of gencrality of isolatcd actions such as "gathcr-
ing on the rug." Generalizations become apparent only in an analysis of the semantic
relations between difl'erent representations of the same actions and reactions within
one text (in which case, it is possible to reconstruct the action taxonomies inhcrent in
that text) or in comparing texts, as I have done informally in this section. Generaliza-
tion is neveltheless an important issue in critical discourse analysis, as texts which
are mainly concerned with legitimizing or delegitimizing actions and reactions tend
to move high up on thc generalization scale, including only the names of episodes or
of whole social practices.
Gcneralization can be seen as a form ofabstraction; they abstract away from the
more specific micro-actions that make up actions. Othcr forms of abstraction abstract
qualities from actions or reactions. Example 3.58, for instance, abstracts away from
the substance of what parents actually do when they "interact" with teachers, sug-
gesting perhaps that it is not important what hey do, so long as they are socn to
"interact with," "relate to," "be involved with" the school. This kind of absrraction
I will refer ro as distilLation:
3.58 Your interaction with teachers throughout school life can have a very positive effect
on your child's attitudes.
Example 3.59, again takcn from Illich (1971: 338), abstracts away from whar reach-
ers do in order to concentrate on their agentive power:
3..59 The teacher cxcrciscs a kind of power over their person which is much less Iim-
ited by conslitution:rl and consuetudinal rcstrictions than the powerwielded by the
gualdians ol' tf lrcr social cnclavcs.
Theso distillaliors nol orrly higlrlight sonlc aspecl of an action at the cxpense of
otlrcrs, lhcy ulso rt::rlizt'l)rtrlx)s(:ri lrrrtl lcgitirnations: purposcs through the kinds of
rualitics lrighliglrtt'rl (rlrt.
rrrrxrst's
ol'soliciting thc complicity of parenl.s with thc
7O r)r\(
( )l rRsl AND PRAclcE
school systcnr and o1'controlling children, fbr instance) and legitimations ancl dclc-
gitirrurtrons llrrough the evaluative associations which, in thc given context, may
cling t() thc tcn)ls that establish the refelence to the action or reaction. "Involvenrcnt,"
"intcrrction," and so on have positive connotations in discourses ol'counseling and
thcrapy, wlrilc powcr may evoke negative connotations, certainly in the context ol'
llliclr's /)cslooling Society, which argues lbr returning some power to the clients
of cducatiorrrrl ploltssionals. Since the same qualities can be distilled from a hetero-
gencous varicty ol'social practices, distillation also allows practices to be compared
and classified along the dimensions of cuality highlighted by the distillation. Fields
of'socirrl
rlirctice
in which thc same kinds of purposes and the same kinds of valucs
rrul tlisvrrlultions obtain can thus be demarcatcd. And such fields are scrved by the
rnstilrtions whose (theorctical) practices ofdistillation elaborate thesc purposes and
vrrlrcs: intlivitlual psychologies in the case of "interaction," "involvement," etc.;criti-
c:rl socioloqics in thc casc of"'power," "control," etc. Such discourses teach us to see
tlrt' turrlitics in a hctclogcnoous variety ol'social practices and supply the legitima-
lions rntl
l)ur.lxlscs
that support thcsc practiccs-or the delegitimations that allow
Ilrr'il t'r'ilirttc.
Sorrc ol'tllc lnotlcls ol'rcprcscnling actions and reactions that we have alrcady
rlisr'rrsst'rl rrlso irvolvc lrbstraction.
'l'lrc
lirllowing combination of objectivation
;rrl t'vt'llturrlion. liu inslirnce . higlrliglrts tllc
"ccromonial" quality ol'the activity ol'
":rsst'rrrlrl'
irr tlrt' lr:rll":
t (l)
l lris (('r( rnor)y is tlt'sirlncrl to lr:lr thc child ltcl she now bclongs to thc school.
'llrt'rt:
is;rlso i nrt:lulrngrrislic lirrrrr ol'tlrslillatron: thc distillation of eithcr thc signi-
fit:, rs in .l-(r|, ol tlrc sigrrilictl, rs in .1.(2:
-1.61
'l'hc
rrotiorr ol'rllytirnc rcprcscnts a lna.jor discontinuity in the cxperience ol'rnost
children starting school.
3.62 The telm physical education refers to that period of physical activity when the chil-
dren use spccific apparatus and movements to promote muscular coordination and
agility.
Distillations in the "Race Odyssey" tcxt include "backlash," "l'reeze out,"
"bcsiege," "disharrnony"-all distilling the negative aspect of actions the prccisc
nature of which we can fbr the most part only gucss.
7. Overdetermlnaton
In chapter 2, I characterized the symbolic representation ol'socil ilclors rs ovcr
determination-as a form of representation in which the replescntcrl social rrt'tors
can rel'er to actual social actors in more than one social practicc.'l'lrrs, lllc kirri of
the lairy tale can stand fbr the father, the company director, or tlrc
rolrlic:rl
lt'irtle r,
and the tcam ol'"prof'essional heroes" in wcstcrns (Wrighl, l()75) lirr tlrt' tcrrrrr ol
<klctors,thctcarnol'scientists,thetcamol'Dresi<lcntiirl irirlt's.:rrrtl soorr'l lrisrlt'firritior
lil t,ll \l N lrN(; s(x,tAt ACTtoN 71
of symbolization cn bc arplictl lrlso to sotirrl rrcliolt: lllc slaying of'the dragon in
the fairy tale can stand lirr ovcrcornirrg tlrc
()r:rlipul
conllict, passing the entrance
examination, winning thc clcclion, in slurrl, lirr any trial which achieves the goal of
a hero's quest. And the killing ol'thc cncnry in thc westcrn can stand for overcom-
ing, violently or nonviolently, any thleat to a society or group, whether disease
or disaster. rebellion or crimc. or any c()nlpetition. whether from rival scientists
or political opponents. Such storics, set in a fantasy world, a mythical past, or an
imagined future, are openly fictional prccisely to allow a multiplicity of references,
each one as valid as any other. As Wright says of the western: "Myths present d
model of social action based on mythical interpretation of the past" (1975: 188;
emphasis added).
Symbolization can be local or extended, that is, it can extend over all or part of
a representation. Allegories, symbolizations extending over the whole of a text, as
in f'airy tales or westerns, are at the lcast localized end of the continuum. Metaphori-
cally represented actions, such as "steer" in 3.62 and "build up" in 3.63, are at thc
local end of the continuum. Note that they difler from the naturalizations and abstrac-
tions I discussed earlier: the verbs are concrete material processes which can takc
only human agents; if used with nonhuman agents, they humanize the natural world,
rather than naturalizing the human social world:
3.3 It helps to see yourself as the teacher's partner who can support her ef'forts to stecr
your child through the primary skills of learning how to leam.
3.4 She turned thc pages with an expert build-up ofanticipation.
Like distillations, such metaphors highlight a quality of action rather than reprcscnf-
ing the action itself. But they do so by means of a concrete image. Also likc distil-
lations, they may introduce purposes ('steer' in 3.63, for instance, introduces thc
purpose of "control") and legitimations (e.g., "build," with its positive connotati()n,
in 3.64), and as a result of their overdetermining potential, thcy can create covcrl
classifications of action along the dimensions of the quality or qualities highlightcd.
But, unlike thc classifications brought about by distillation, these classifications
remain covert. In 3.63, "steering" remains a metaphor, opening up a potential fi)r
classification, whereas, say, in Habermas's theory of communicative action (1984),
"steering" becomes a major technical term ("steering mechanisms") applicablc to
social practices in general.
The title of the "Race Odyssey" text is such a symbolization. The archetypal
tale of a heroic and perilous journey back home, back to the cradle of one's idcntity.
symboliz.cs Australians' cuest for "racial" identity and connects it to noble qucsts ()f'
this kind generally.
The second fbrm of overdctcrmination is inverson.In "The Story of Asdiwal,"
Lvi-Stlauss (1967) described the role of inversion in a myth of the Tsimshian
Indians, a people from the Pacific coast of Canada. The economic activities (fish-
ing, hunting, etc.) represented in this myth accurately describe Tsimshian practices
but, says Lvi-Strauss, an anthropologist who would seek to use the myth as evi-
tlcncc firr ctlrnographic description would be bound to make errors, for other Tsim-
slrirrn
rrlcliccs
alc invcrtcd in this story. It was, for example, Tsimshian custom
',
Il r rrlr',t /\t.I) I'liA( |l( I
l,r rr',,nrr r ri, nr)v(' to tlrt' village of their husbands after marriage ("patrilocal
nr.ur,rt,( ). lrut rn
"llrt'Slotyol-Asdiwal,"theoppositeoccurs:thehero,Asdiwal,
nr.\'r'.. r(, rlr rrll:r1't'ol lris wilc aftcr marriagc ("uxorilocal maniage"). Myths,
s:rys l.t'r,r lirr;rrr.r { l')(r/. I l), "tkr not give accurate picture[s] of the reality of
nllrvr'lrlr'. lrrrt r,,ort ol cortttlt:rlttint wlrich seems sometimes to be in harmony
wtllt tlt(' rc:rlrty, rrrrtl \()nr('lln('s lo
t:rrt
l}om it." The reason for inversion of this
kintl lrt's rr tlr(' l(',trtrnrlrrtl'. luttt'liotr ol'tltc nlyth. "The Story of Asdiwal" seeks
Io lclirlirrrrzt' tlr('( nst()rrr ol
r:rlliloc:rl
ltrrrriagc: thc hero's deviant actions lead to
:r crisis wlriclr rs rrol rt'solvt'tl unlil otlcl is rcstorcd and the hero lives again in the
villlrlc ol llts lirtlrcr:
"('xtr ( nr('
Irirsrtr(lrrs
lrre irn:rgirted in order to show that they are
trtlt:tt:rlrlc"
(lritl.: 10).
(iorrrrulrrblc invr:r'siorrs trccrrr in "lirsl tluy" tcxts. In the children's story I quoted
in chaptcr 2, l littlc girt trrkcs r lion to school.'l'hc lion is not only an inverted partici-
pant, hc cngagcs in invcrtctl rctions, sucl) rs "swishing his tail" threateningly during
"r'cgistcr tirnc."'l-his crcatcs an incidcnt that disturbs the ordcrly unfolding of the
"rcgistcr" episodc:
3.(r5
'l'hc
tcacher stopped calling the registcr when she saw the little girl and the lion, and
all the othcr children stared at the lion, wondering what the teachcr was going to say.
The tcacher said to the little girl: "You know you are not allowed to bring pets to
school." The lion besan to swish his tail-swish!-swash!
When such deviant actions occur, they cause remedial actions or episodes to be
inserted into the representation. Unrepentant deviants must be corrected in a punitive
cpisode. Or an attempt must be made "to show that what admittedly appeared to be
a thrcatening expression is really a meaningless event, or an unintentional act, or a
jokc not meant to be taken seriously, or an unavoidable 'understandable' product of
cxtcnuating circumstances" (Goffman, 1974:.223). The teacher training texts in my
"first ilay" corpus provide many instances of this:
3.6(r
'l'hc
skilllul tcachcr can save thc new entrant's face by showing hersell'to be on his
sitlc: "l Ic's not rcally crying, that'sjust a noise he's making."
I{crncdial actions rnry tlcvclop into full-scale social practices and become central in
tcxls rlclrling with dcviant bchavior. Newspaper reports on truancy, for instance, often refer
to tnruncy itscll'orrly hy rrrcans of'a handful ofobjectivations ("truancy") and abstractions
("thc
rroblcrrr")
lnrl conccntratc on tho social actions which have been institutionalized
(or luc
rlarrrrcrl)
to counlcrrct it, although there are, ofcourse, also reports which focus on
llrt: r lt'v i;rrl lx'lr:rviol itscll, lor instancc, by means of short interviews with truant children.
'l'lrc
s;rrnc is lnr(' ()f nrc(li:r rcproscntations of many other social problems.
l)t'vrnt ;rtlions irlw:rys prcscnt a threat to the orderly unfolding of social prac-
I rt t's.
'l
l rr'y r t'vt':rl l lrc contradictions that underlie these practices, contradictions such
rrs llrt' rrrrt'trr;rl tlistr ilxrtion ol-thc rewards of education among diflerent social groups.
(
)rr tlrr. otlrt'r lrirrrrl, rcrrcscntation of deviant action is also attractive, for it allows the
v r( ;r rr )us tr rrrrsl'.r't'ssion ol'social norms. Hence, it forms the staple larc of many kinds
rl tcrrs n('s,\. ( r irnc, Iiclion, gossip, and so on.
Social
action
REPRTSENTINC SOCIAI
^CTION
7 I
'l'lrc
actions ol'Bruce Roxton, thc racist Australians love to hate, are inversrons
ol'"corrcct" social practices. He "damages" and "distorts" and is therefore rightl'ully
"rcsented" by Australians-but his deviant actions are extensively covered by thc
Australian media, year after year.
B. Conclusion
By way of conclusion, I provide two summaries. Thble 3.2 summarizes my answcr
to the qucstion with which I started and gives an overview of the principal ways
Unslecified
f
Cognitive
ReactionJ
f
Affective
L Perceptive
,fAl:cnti;rliz:rlion
livcnltt:ttiott
-4
1
LDcagcntialization
)
l')xistt:rrtiirliz:rlion
I
L Ntlurlrlrz:tlrott
.
( it'ttet rtltzrtl tott
fAbstraction-
-)
I
I l)rstillrrtion
L
Cloncrcl iz:r I ion

Interactive

Matcrial__-
Transactivc
I H
L
Instrumental
I rl L Nontransactive
Action
---) | |
I l- tlehavioral
Form specificarion
rsemioric
i_J I
I
'l
I
Topic specification
I
LNonbehavioral
-)
t
l-
Rcndirion
I
L
Quotation

Activation
)
Objcctivatitrn
L Dcacrivarion __)
L f )cscript ivizrrti,,n
.[Silrglc
dclt'l ntirrt rort
1 f Strrtl,,,ltz.rlr,,rr
LOvcrtlclcr lrrr:rl rorr j
I 1,,u,-,.,,,,'
TABLE 3.2. Social Action Nt'lworL
DISCOURSE AND PRACTICT
in which actions and reactions can be represented in English discourse. As before,
square brackets stand for either-or choices (e.g., deactivation can take the form of
objectivation or descriptivization), curly brackets fbr simultaneous choices (e.g.,
reactions can be activated and agentialized, or deactivated and agentialized). I will
also summarize what this descriptive framework has allowed me to observe about the
"Race Odysscy" text.
Both the actions and the reactions of the "racists in other countries" are repre-
sented. Theil reactions to immigrants and immigration are portrayed as emotive and
negative, and their actions are transactive, as actual interactions with immigrants-
material interactions, such as "attack," as well as behavioralized semiotic interac-
tions, such as "insult." For the most part, these actions and reactions are activated,
hencc prioritized in the rcprcscntation.
Australians are also representcd as involved in actions and reactions. But Aus-
tralian reactions are more oftcn portrayed as cognitive, considered, and reasonable,
and whcn they do tend toward the affective and the negativc, they are mostly objec-
tivatcd. The actions of Australians also differ from those of immigration critics in
other countries: they tend not to involve interactions with immigrants. Australians
are represcnted as people who stand on the sidelines and monitor the events as they
unlbld, but not as directly participating in them.
As fbr the Australian government, both its actions and its reactions are repre-
sented, but the former are in the majority. Government action is either semiotic and
activated, or material, instrumentalized, and objectivated: what the government says
takes center stage; what it does, its instrumentalizcd transactions with immigrants,
remains in the background.
The immigrants thcmselves are never represented as reacting, and their actions
arc mostly nontransactive and objectivated. Only thc action of "immigrating" itself
is rcpresented straightforwardly. Other immigrant actions are veiled behind distilla-
tions, such as "freezing out" and "besieging," or behind naturalizations which repre-
sent their actions as "change."
Thus, the "Race Odyssey" text allows us to glimpse racist actions in countries
other than Australia, but blocks us from access to what happens in Australia itself.
Australian "criticism" of immigration takes place at the altogether different level of a
reasonable "debate." But it is a debate which keeps a safe distance from what actually
goes on in Australia and which refers to actual interactions between Australians and
immigrants to Australia only in vague, objectivated, generalized, and abstract ways.
And as far as the immigrants themsclves are conccrned, their voices are never
heard.
Time in D iscou rse
I
ln this chapter. I describe the semiotic resources of English discourse for represent-
ing the timing of social practices. My analysis draws on my corpus of "first day at
school" texts as well as on an unpublished study of the representation of time in stalT
newslctters, in-house magazines, and other written intcrnal communications from
Rank Xerox (collected in 1995) and in six articles from thc Employment Gazette
(July and September 1993).
1 . lntroduction
In his rcmarkable book about time, the sociologist Norbert Elias stressed that time,
and thc way we think and talk about it, is a product of the actvity of timing-the
activity ofmeasuring one kind ofactivity or evcnt sequence against anothcr kind of
activity or event sequence (1992:43):
The reifying character of the substantival form "time" disguises the instrumental
character ofthe activity oftiming. It obscures the fact that thc activity oftiming, e.g.,
by looking at one's watch, has the function ofrelating to each other the positions of
cvents in the successivc ordcr of two or more change continuums.
In this chaptc I focus on the resources which the English language provides
fbr representing this activity. As before, this is not an end in itself, not a labor of
classification for its own sake. Sociologists (and also musicologists; see Van
Leeuwen, 1999) have drawn attention to the correspondcnces between the timing of
4