Pragmatics2:3.

3I 1-323
InternationalPragmaticsAssociation

THE USES AND UTILITY OF IDEOLOGY: SOME R-EFLECTIONS

Michael Silverstein

Quelque grands quesoientlesavantages

dessignes,il fautconvenirqu'ilsont des
inconvdniens; et si nous leur devons
presquetouslesprogrds denotreintelli-
je lescroisaussila causedepres-
gence,
quetousses6carts.
A. DestuudeTiacy(1827[1801]:267)

Withoutwishing to commit the etymologicalfallacy in the understandingof a word's

meaning,I would like first to comment on the traditions of usage of the term
ideoloqv,a theme elegantly announced in Woolard's introductory discussion of
"issuesand approaches."
As is well known, it was Antoine Louis Claude Comte Destutt de Tiacy
(1754-1836)who invented the term, in that naturalizing move of the French
Enlightenmentrendition of l,ocke (or, to be sure, Condillocke) that sought to
understandhuman "nature." Ideology was proposed as that special branch of
zoologythat recognizesthe condition of humans,we animalswho have ideas as the
contentof what we should call our minds. Central here is the fact that any ideas
moredevelopedthan physiologicalsensationsare dependenton such ideas'being
clothedin signs,the organization of which by some systematicgrammar allows the
discursiveexpressionof a logical faculty of mind. Hence, for Destutt de Tiacy, there
is the generalscientificfield of ideologyproper, the scienceof ideas,of which the
subfieldof grammar studies the signiffing externalizations,as it were, in structured
systemsof articulated signs, and the subfield of logic the modes of rationality
orientedto truth and certitude of inferential states of mind (i.e., formation and
combinatorics of ideas).Such a sciencewould, for its propounder,also allow us to
diagnoseand understand "the causes of incertitude and flogical] error," thus
presumablyleading to an ameliorationof the human condition vis-d-visits natural
mentalfaculties.
It is particularly interesting,therefore,to see the fate of this term, proposed
as a formation parallel to any of the other "-ologies"of a systematicscientific
outlook. It has obviously become a word that now denotes a part or aspect of
Destuttde Tiacy's very object of investigation,and in many appearanceshas the
specifically "pejorative"use- to pick up on JaneHill's invocationof RaymondGeuss
(1981:12-22)- that presupposeswe know certain ideasto be dubious,in error, and
3I2 MichaelSilverstein

therefore suspector at least suspicious,in the manner of "mere" ideas as opposed

to material, historical,indeed factual "realities."l
For those of us who have no connotational- i.e., ideological! - problems with
consideringmental phenomena as historical and factual, if not directly material
realities,the concept of ideologyin our contemporaneoususagecan embrace the
terminological transition to a notion of a concreteobject of possiblestudy, while
making no judgment - at least in scientific and scholarly usage - about some
independent and absolute universe of Tiuth (with its capital T) and Validity
(positivistically,not positively,speaking)againstwhich ideologiesare measured. (In
this usage,it is no causefor concernthat scientificand scholarlydiscourse,too, is
"ideological,"not escapingfrom the universeof human mental activitiesas these
arise in conditionsof sociality.)It is thus with the sociologicaland anthropological
(or more generally"descriptive"[Geuss 1981: 4-12]and social scientific)concepts
of ideology that I should like to see us continueto be concerned,and certainly this
cluster of concepts- to re-term Woolard'sdifferentiationof 'features'of the term
ideolory - is central to all of the papersin the accompanyingsection,to which my
remarks are directed.
These different emphasesof social scientificusersof the concept-cluster-
whether by name or equivalent- form a kind of palimpsestor geologicalstratifica-
tion of connotationstaken from canonicalnineteenth-and twentieth-centurytexts.
Ideologyis an intensionalcharacteristic, predicableof a society,of a group or other
social formation abstractedfrom society,or even of individualswithin some defined
population, and can be understoodas "mental,"therefore,with the same problems

I Th" historical
linguistin me wishesto note that thereis an interestingproblemhere in the
shift of meaningfrom the abstractfield-of-scientific-study senseto the concretesenseof (in the
singular) one of the objects-of-scientific-study. It would seem likely that the actual mechanism
involved a derived, adjectival usage,viz., iddoloqique,ideological,which would construe the
objects-of-studyby characterizingthem as such objects,namely,those studied by the field of
ideology.Thenceit is easyto seethe back-formationthat reidentifiesthe baseas the object-of-study,
sincethe -ique,-al formationsof adjectiveshavemoveddecisivelyin the directionof characterizing
objectson the basisof their own denotingterms,especiallyas theseshareformally identicalstems.
It is clear that contemporaneously with Destutt de Thacy'sintroductionof the term iddoloeiein
Paris,there was stimulatedtranslationcoinageof an equivalentin English,attestationsfrom 1796
and 1797appearing(reportingon the Frenchdiscussion) with the authorially-stipulatedsensesfor
ideolosv and ideoloeical (see O.E.D., s.w.). Apparently through a kind of delocutionary
quotation-translation of (pro-)Napoleonicusageca. 1813-1815, the derivationalset ideoloKv,
ideoloque,ideologist,ideologicalemergeswith a fiercelynegativeand mockingconnotation,leading
to a senseof unpractical,speculative, idealistsocialphilosophicalthoughtsand thinkers,whenceby
the 1830sand 1840s,the oppositionof (negativelyvalued)ideasvs. historicaland materialfactsis
establishedin English,especiallyideasassociable - accordingto one 1827citation of ideolory - "with
hot-brainedboysand crazedenthusiasts," that is, the negatively-valued
(mere)social-theoreticideas
of a group clearlyindexedas not that of the speakeror writer. Thus any ideologist,i.e., proponent
of the intendedlyscientificfield of ideolop, hasmerelyideologicalbeliefs,as opposedto ideasthat
correspondto material,historical,andfactualrealities.A possibleparallelshift in the noun-adjective
derivationalstructure may be observablein very contemporarytimes,in English close to home,
wherewe can note the shift of the erstwhileparadigms(noun) languase- (adj.) linguistic[= 'of
language'lvs. (noun) linguistics- (adj.phr.) of linquistics,as in the noun phraselinsuistictheory,
so that now the last phraseseemsto speakers(linguiststhemselves!) to mean'theory (asopposed
to descriptiveor other practice)of linguistics',not'...of langtage',and hencewinds up being an
"ideologicalnterm - in the newersense- of its disciplinarysocialorganization.
 Theusesandutilityof ideologt 3I3

as the conceptsof 'culture' or 'language'when these are consideredas "mental"

characteristics. Thus note that the sharednessof ideology,like that of culture ot
language,and the relationship of ideology to consciousness,are properties that
survivefrom Destutt de Tiacy's usage,which focusedupon the human condition as
ideational.Certainly,the culture concept and the concept of languagecentral to
severaltraditionsof anthropologyand of linguisticshave entailed their wrestling with
the natureof such sharedness, and the degreesof consciousness involved in such
culturaland linguistic phenomena, in many different ways. These constitute both
ontologicaland epistemologicalproblems for validity claims in the scientific manner
for our statementsabout "the" culture or "the" languagewe purport to describe,as
everyawarepractitioner knows,and they will not disappear,as is increasinglybeing
recognized (sometimesmore forcefully by external critics than by practitioners,who
then look like "ideological" apologists) by appeals to Cartesian certitude of
self-examining "intuition." Having such intuitions, and being able consciously to
formulateand communicate("share" in an active, agentive sense)them, may thus
be "ideological"to a degree greater than, and in ways more diverse than, many
studentsof culture and language in one scientistic mode or another would find
tractable. Indeed,if all cultural and linguisticphenomenaare essentiallyideological
"allthe way down,"this calls for a re-evaluationof how we might creep up on such
"material"or "objective"factuality as presentsitself in them.
Observehow this leads into the second aspectof the concept-cluster,the
social-situatedness of ideology. For if we are dealing with some ideational
phenomenon particularto or predicableof societies,groups, etc., like the classic
Romanticnotion of the world-view of "a people," "a nation," etc. that constitutes
their uniqueness,then insofar as ideology is characteristic of any sociocultural
phenomenon, it must inhere in what makes any social entity of whatever scope
cohereas that socialentity. We have only to be able to locate such a social entity,
to find the potential for a unique mental or ideological aspect associablewith its
members or participants.Once we yield to the distrustof the socially-locatable(as
opposedto the material and factual of early nineteenth-centurypejorative usage)
assomehowlessfactual than real facts, we see that the locatednessof ideology
takeson a kind of negativeconnotationas opposedto the locatednessof "culture,"
in both its technicaland its lay usages.Rather than seeingthat there is no such
thingasa socialfact without its ideologicalaspector component,many usersof the
concepthave simply yielded to the negativeconnotationsof a kind of charged
politicalrhetoric when it comes to analyzing the ideational aspect of social
formations. But the long-standingtextual basis for this bias is not hard to find,
revealed as it is even in the historical organization of dictionary attestationsin the
OED,
Thus,the locatedness appearsto thoseviewingideologyfrom acrossa dMde:
whetheron the perceivedlyopposite side of the scientisticstance at or in objective
'Tiuth" vs. the distortions or
mystifications in ideology (this is Woolard's third
member of the clusterof notionsin the word), or on the perceivedlyopposite side
of the de-mystiffing,de-bunking, committed championship of the victims of
dominant, powerful,or oppressivegroups or institutions,on whose behalf the
self-legitimization of the latter is declared to be ideological; these are the initial
intellectual
salvosof de-legitimation(Woolard's fourth feature of the concept). We
canseethat featuresthree and four share the stanceof otherness,the one being
 314 Michael Silverstein

epistemological, the other rather frankly political.Thesetwo stancescan, of course,

be combined,as in traditions of a politically-committedscientisticananlysisof
"ideological"formationsthat som€versionsof political economydeclarethemselves
to be.
Observehow the specializedsensesWoolard points out are on the order of
Putnamianstereotypeslocatable in specific groups of users in historically located
traditions of the division of linguistic labor of the very term ideology and its
derivatives.These all seem to have emergedout of the very same,ideologicallyin-
formed processesof linguistic change that start in the indexical penumbra of
denotational usage of linguistic forms-in-context,is re-valorizedin the realm of
ideologically-grounded apprehension,and returns to usagein terms of transformed
indexicalvalue. Such explainsalso the charming sociolinguisticfact that Woolard
brought up in oral presentationthat there are two phonologicalforms, [ay]deology
and [i]deology, probably indexicallylocating a speaker in different communities of
discourseand probablyemphasizingdifferentstereotypeaspectsof the concept-clus-
ter in actual usage, e.g., cultural anthropologesevs. social anthropologeseor
sociologese,among others. Like the classiclabovian example of the woman who
infusedthe stratificationindexicalvalue of the alternativepronunciationof yleys]eg
vs. v[ahz]eswith denotationalcontent (the first denotinglessvaluable objectsthan
the second),thoseof us who pronounceand listento thesealternativesprobably do
behave like native speakers ought to in respect of the term ideolog.vand its
derivatives,in which the role of authorizing,canonicaltexts read and re-read over
two centuriesinfuseseven our phoneticusagewith now one, now another dominant
conceptualstrand,despite ourselves.
Tirrning to the other papers in my assignedtask, I want to emphasize two
general themes concerningthe ideologicalaspectof the casesdiscussed,and then
comment upon the utility of the concept of ideology- as opposedto the concepts
of culture and language themselves- when dealing with the facts of language.
Thesetwo themescomprisefirst, the mediatingpositionof ideologyin what can only
be called a dialecticprocessat many planesof abstractionand many distinct orders
of socialityand of socialprocess.The secondtheme is how, exactly,to understand
the "sitedness"of ideology in a number of frameworksrelevant to the analysisof
language:where - to spatializemetaphoricallyat least- to look for ideolory's many
manifestationsin social life, and the implicationsof the answersto this for social
scientificpractices.The papershere have rich material relevantto these concerns.

The necessityof ideology in the dialectic of indexicality

Once we recognize that the "realities" of meaningful social practices emerge from
the experience of indexical semiotic processes,we should resign ourselves to
enjoying the fact that it's indexicalityall the way down. That being the case,we
should seethat meaningfulness is a dialecticproperty of socialsemiotics(other than
such aspects of denotational languageas are justifiably referred to systemsof
Saussurean-Bloomfieldian-Chomskian'sense'-generating morphosyntax).And the rub
is, the only way to break into such dialecticalsystemsis with the inherently ironic
concept of ideology.Ideology,in other words, is defined only within a discourseof
interpretation or construalof inherentlydialecticindexicalprocesses, as for example
 Theusesandutiliryof ideologt 315

the processes of making or achievingtext (entextualization)by using language and

othersignmodalitieswhether at the denotationalplane or the plane of interactional
textuality(thoughof course,for languagein particular, both planes of textuality are
alwaysin play).'
Now any indexicalprocess,wherein signspoint to a presupposedcontext in
whichthey occur (i.e., have occurred) or to an entailed potential context in which
theyoccur(i.e.,will have occurred),dependson some metapragmaticfunction to
achievea measureof determinacy.It turns out that the crucial position of ideologies
of semiosis is in constitutingsucha mediatingmetapragmatics,giving parties an idea
of determinatecontextualizationfor indexicals,presupposableas shared according
to interestedpositionsor perspectivesthat follow upon some social fact like group
membership, condition in society,achievedcommonalty of interests, etc. Ideology
construes indexicality.In so doing ideologyinevitablybiasesits metapragmatic"take"
so as to createanother potential order of effective indexicality that bears what we
canappreciatesometimesas a truly ironic relation to the first. lrt us term these,
asit were,hypotheticallypre-ideologicaland decidedly post-ideologicalindexicalities
respectively the first-order and second-orderaspectsof indexicality.3(Of course,
giventhe dialectic nature of the semiotic process,we should understand these
relativeordersto be generalizableto nth and n+lst.) One kind of approach to
invokingideologyin the analysisof social signsis to start with the observation that
everysystemor modality of social signs is infused with indexicality, that therefore
suchindexicalityis caught up in a dialectic process mediated by ideological
formations,and that therefore there is no possible absolutelypre-ideological, i.e.,
zero-order, socialsemiotic- neither a purely'sense'-drivendenotationalsystemfor

a
- It is,of course,essential
to keeptheseplanesof textualitydistinctone from another,sincethe
fint answers the question,"What hasbeenArillhavebeensaidein-and-bysome use of signs?"and
thesecond"WhathasbeenAvillhavebeen done in-and-bysome use of signs?'The first is a model
of denotational(referential and modalized predicational) coherence over some span of
event-duration, frequentlyexpressed in termsof propositionalor'informationalncontent;the second
is a modelof social-actionalcoherence,frequently expressedin terms of social acts in some
framework for description.It must be seen,however,that both of thesekinds of text are models of
gradient, interpersonal, indexicalty-consummated achievementsin the processualrealtime of using
signs,the decontextualizable,genred appearance of which to the user is an important,
ideologically-inforrned perspectivalreality that makes indexical presupposition seem to gel as
structureautonomousof realtime contextualization(available, for example, for armchair
niaosociologizing of linguisticpragmatists andphilosophers).
And notehowsuchstructurebecomes
vMdwhenconcretized in a text-artifactthat can perpetuallybe reanimated- e.g.,by reading a
printedarrayasa text - in a new entextualizingevent.

3 In my 1979piee, on "l:nguage structure and linguistic ideologr" I reference Robert K

Merton'sdifferentiationof manifestand latent functions of social practices(see Silverstein 1979:
204),
especiallyasthis alignsideologically-informed metapragmatic ntakes'(there calledfunction,)
onindexicalfunction(therecalledfunctionr)with Merton'smanifestdimension,and muchof actual
indexical
effectiveness- especiallyin the way of entailmentsof indexicaloccurrence- with Merton's
btent dimension.Given the dialectic of ideological engagementwith indexicality, however, the
conceptof orden of indexicalitymediatedDyideologically-informedmetapragmaticsis a much more
uscfulformulation,since in certain respectsideologicalformations depend on facts of signs
ftnctioningindexically(at whateverorder) just as a higherorder of indexicalfunctioningdepends
onthepresupposition of ideologicalformationsfunctioningmetapragmatically.
316 MichaelSilverstein

the referential-and-predicationalexpressionsof any language,nor a totalizing system

of purely "symbolic"values for any culture. Certainly the confluence of deconstruc-
tive and Manrist political-economicapproaches,insofar as coherent,amount to an
argument to this effect. But it is not my purpose here to elaborate on these.
Rather, I want to point out that ideologies as invokable schemata of
explanation/ interpretation of the meaningful flow of indexicalsgelled into text-like
chunks, are generally thought of as relatively perduring with respect to the
indexicals-in-context that they construe. And we recognize such schemata
characteristicallyby the way that they constituterationalizing,systematizing,indeed
most importantly naturalizing schematathat "explain"the indexicalvalue of signsin
terms of some order(s) of phenomena stipulatively presupposableby - hence, in
context, autonomous of - the indexical phenomena to be understood. Such
naturalizingschemata,in the way of "JustSo Stories"about the direct, transparent,
indeed sometimes iconic indexicalityof the phenomena at issue, in their most
elaborated forms, go to the very heart of issuesof what human nature is all about
in a universe that has certain absolute characteristicswith respect to the field of
indexicalsemiosisat issue.
We can think of any of the demographic dimensionsof identity summoned
up indexicallyin the use of language:men vs. women, generationn vs. generation
m within a kinship or even age-systemorder, ethnicities Ao B, or C, various
class-orientations or -identities within a structure of stratification, etc. Any
reasoningthat interpretsa presuppositional relationship- "Such-and-suchesuseform
'...',while - is potentiallyan ideologicalone rationalizingthe indexical
so-and-sos..."
value of the forms in terms of schemataof social differentiation and classification
that are independentof the usagesat issue.And we shouldnote how systematicand
elaborated such rationalizations in pre- or non-semiotic frameworks can be,
ultimately grounded in a cosmologicallyor cosmogonicallytotalizing vision. At their
most elaborate, ideological construals/constructionsof indexicality - the ultimate
extensionalfacts, note - turn the indexicalsinto seeminglynatural indexical icons -
emblems with consubstantial intensional content or stuff with respect to their
original contextual surround, little detachable design elements that are like
ready-made texts deployable for social-interactiveeffect and which preserve their
intensional characteristicsas a seeminglyautonomous"code" of meanings.a (We
have only to think of the almost literal evaluationof words - such as v[ahz]es- as
$10-items,comparedwith their $.10 denotationalequivalentsin other socio-econo-
mic classregisters.)
Thus also note that many of our theorized sociolinguisticconstructsare
inherently suffusedwith the ideologicaldimensionof the semiotic processesthat
realize them. For example, the concept of register,as a minimally binary paradigm
of "sayingthe same thing" indexicallyappropriate to two distinct contexts of usage,

a Is this not the essenc€

of totemismthat hasbeencompletelymissedby structural-functionalists
and structuralistsalike? Here note the similarityto phenomenaof fetishization,of reification,and
of commodification,as they have been termed in their respectiveliteraturesinformed by Marx,
Durkheim, Lukdcs,et al. I sketchthis generalization of the phenomenato the wider semioticrealm,
fully in keepingwith and developedfrom scatteredinsightsof the late RolandBarthes(seeespecially
1968;1983),as I havedetailedin two asyet unpublishedpapersat other AAA meetingsin 1984and
1988,on 'The qvalue'of objectuallanguagen and nDe-voiceof authority.'
 lhe usesand utitityof ideologt 317

howeverdefined. Note that the "sameness"of denotationalvalue, built into this

concept,is an ideologically-drivenmirage from the structural linguisticpoint of view;
thatthereare indexically-inflectedparadigmsof "equivalence,"from the perspective
of usersof the denotationalforms, is what makes this denotational"equivalence"
factual within some sociolinguistic order. No examples of literally equivalent
registershave,to my knowledge,appearedin the literature, yet it is a construct that
nicelycapturesan important way that native speakersof languagescomprehend
variationand variability in their linguistic forms, one that allows us to capture
somethingof what it means to have a linguistic community whose ideology of
"speakingthe same language" must obtain in spite of obvious processes that
dialecticallybalance dialectal and superposedvariability in usage (the terms are
Gumperz's1968:383-4) in their constant reciprocal interconvertibility in different
ordersof indexicality.sYet the imagination of indexically-functioningvariability as
organizedby registers - seeing indexicality through the lens of alternatives of
denotationalusage - presents the social context that metaphorically surrounds
languagethrough the idiom of discussingthe appropriatenessof alternatives of
formulatingone'sdenotationalmessagein this or that way. As Don Kulick observes
in hispaper,"languageideologiesseemnever to be about languageso much as they
areaboutother things, such as gender and expressionand being civilized," all, we
mightadd,featuresof language'scontext of occurrence.
So that people have ideologiesof languageis a necessaryentailment of the
factthat language,likeany socialsemiotic,is indexicalin its most essentialmodality.
And that irony is the essential trope lurking always in ideologically informed
contemplation of language,whether by "them" or by "us," is a consequenceof the
actualdialectic manner in which ideology engageswith pragmatic fact through
metapragmatic function, in a kind of spiral figurement up the orders of indexicality.
All of our accompanyingpapers demonstratethis.
Irvine's paper takes up the distinction between deference-and-demeanor
indexicals and honorific registers.The latter always involves a strong ideological
component expressedin a metapragmaticsof appropriatenessto addressee-focused
conten(presupposition)at the first order of indexicality that ironically operates
ratherdifferently,generally speaker-focused,at the second.6Moreover, as Irvine

5
Notethat the intuition of "sayingthe samething" in the ideologically-informed capacityof
native
speaker mntemplatingdialectal/superposed indexicalvariation is a very different one from
anything
a structurallinguistmightwant to happenupon for purposesof settingup morphosyntactic
andhence'sense'equivalences in the realm of morphosyntacticpgg[ggg relationships.A
praphrase in thismnceptualization hasnothingto do with an indexicalparadigmof equivalence-in--
context,
ofcourse, contextualization andentextualizationbeing,the theorygoes,entirelyindependent
'performance" - like paraphraserelationships- giving
characteristicsdistinct from characteristics
cvidenc€for the structureof linguistic(sc.,grammatical)ncompetence.' (Small child watchingthe
4 Empror paradeby: "Or are they?")

6
Thur,notethe caseof Javanese deferencehonorificationlaid out in Silverstein1979:216-27,
andin greaterdetail in Errington 1988. The multiple dimensionsof indexicality involved,
speaker-to-addressee,
speaker-to-referent, speakerestimationof Agent-to-DativeNP-denotatum
deference,
contrast
with the linguisticideologyof addressee-focused
unidimensional, linear nfineness'
oflanguageaspartof the appropriatebehavioralenvelopeor ego-centered, addressee-focused bath
in whichoneshouldstrive to immerseone's interlocutor,the more 'fine" the interlocutor (in a
 318 MichaelSilverstein

points out, in some languages these honorific registers organize the lexicon of
nominal-headinggrammatical forms insofar as the categoriesof honorific/neutral/
pejorative are expressedsystematicallythrough the normal paradigms of inflection
and agreement.In this sensethey are "grammaticalized" in somelanguagesto a very
great degree, as in Yao and ChiBemba, while in others there are simply paradigms
of lexical register-alternantsin sometimesone-to-one(Zulu), sometimesmany-to-
-one denotational "equivalence"(Javanese).
By contrast, Irvine notes that the Wolof linguistic ideology, centers on the
binary contrast of articulate griots and inarticulate nobles (the latter forming a kind
of verbal Lumpenkonigstum, one can imagine) as speakerswith immanent social
characteristicsindexically presented.As Irvine has discussedin an earlier paper on
greetings (Irvine 7974), the ideological alignment of these social identities relative
to discursiveinteraction looks somethinglike the following: Griot : Noble :: active,
moving : inert, stationary:: low status: high status:: speaker: addressee:: fluent,
loud, rapid, intoned speech: dysfluent,soft,slow,monotonespeech:: first pair-part
role (soliciter) : secondpair-part role (responder).Doing somethingto index a shift
in these expected alignments is the obviously speaker-centered possibility for
effectively troping, but not trampling upon theseexpectations.Thus, speakingmore
elegantly in the griot manner will not a nobler person make, obviating a higher-or-
der indexical effect. Irvine's paper thus nicely problematizesthe contrastsbetween
ideologically-supportedindexical systemsof honorification or of avoidance-respect
(Zulu hlonipha), with their addresseefocus, their degrees of enregisterment and
even grammaticalization, and the characteristicsof various kinds of text-bound,
discursive indexicals the ideologies engaging which seem to operate distinctly.
Highly-ideologized,enregistered,grammaticalized(or at least lexicalized)honorific
and avoidance-respectindexicalsare imperialistic;they underlie an expansivetropic
potential in what seems to be a great deal of strategic "metaphorical switching"
(Blom & Gumperz (1972:425) with effectivesecond-orderindexicality.
In Hill's discussionof the reflective discourseof nostalgia among Mexicano
(Nahuatl) speakersin the Malinche area,we see the topic of ideological concern,
not unexpectedly, in the addressee-focusedhonorification indexes, prototypically
conjured up in images of set interpersonal routines like greetings.One of the most
interesting facts here is the irony of ideologically-informedusage, located as it is
with respect to a past viewed with highly-developedlonging for days of mutual
"respect," and positionally-definedself-certitudein discursiveinteraction. (It is an
irony so palpable that at least one commentator, Dofra Fidencia, catchesall of this
in her own counter-nostalgicwit.) Indeed, the very irony pointed up by Hill (citing

metric ultimatelycosmologically centeredthroughthe King or Sultanin socialstructure).Ironically,

the more 'finely" one can speak,the more one showsself- at a secondorder of indexicalityinformed
by, and depending on, the ideological understandingof the practice - to be deservingof
ideologically-informed'finenessnof speechand behavior from others! Janet Morford, of The
University of Chicago(p.c.) reports variousspeakersof Frenchobservingthat their families of
orientation were so cultured and polite, they would not use reciprocalT (of the famousBrown &
Gilman [1960]set, Tl{) under any circumstanc€s, evenin the domesticgroup.Note here againthe
semnd-order indexical value, focused on 'culturednessnof speaker,in what has been obviously
emergingin certaincirclesasa registerof honorificationbespeaking positionalidentityof speakers,
not necessarilyaddressees.
 Theusesand utilityof ideologt 3I9

severalexemplaryworks by Kenneth Hill and herself, q.v.) is like that of l-abov's

(1972)now famous"lower middle class"speakersof New York City English showing
maximal"linguisticinsecurity"before, i.e., ideologicalallegianceto, the standard
registerthey seem basically (or basilectally) not to use. Around the Malinche the
ideologicaldiscourseof nostalgia for the pure, unadulterated Mexicano of mutual
respectseemsto peak among those whose verbal protocols show maximal mixing,
i.e.,interference/code-switching, between Castellano (Spanish) and Mexicano. As
Hill notes,"the discourseof 'nostalgia' [being] in fact a pragmatic claim about the
present,using'pastness'as a naturalizingideologicalstrategy,"there is room even
withinthe communityfor seeingthe incoherenceof the first-order and second-order
indexicaleffectsso generated,the first aligned with the valued "pure" formulaics of
honorification,compadrazgo, in a seemingmemorializationsof a time when everyone
knewhis or her proper place (and hence the other person's too), the second with
contemporaneous social stratification and ego-focusedmobility, people construing
their local, traditionally Nahuatl-speakingcommunity within a nation-state order
dominatedby Spanish.And we can see that the speech community exists in the
intersectionof different sociopolitical interests around the far-from-neutral
ideological portrait of the old daysas a canonicalethical-moraltext and enregistered
modelfor current behavior. Yet such interests articulate, debate, and contest the
imagesin terms seeminglyunderstandableone to another.
Contestationis, indeed, the theme of Kulick's paper on Gapun (Papua New
Guinea) speech genres of kros-making vs. oratory. The first, invective- and
obscenity-filled dialogicoutburstshurled between two houses,are negativelyvalued
andrationalizedas the product of essentiallyimmature personalitieswho do not
keeptheir heds,their autonomousselves,while the second,well-turned monologic
thoughtssharedat a public meeting, are positively valued and rationalized as the
productsof personswho have achievedsove,i.e., "sawy" and exemplary Christian
and civic ideals oriented to the common weal.' At the ironic, second-order of
indexicality, this becomes an indirect index of being an adult female, prone to
burningkroses,so it is thought, or being an adult male, able to burnish oratory in
orderlyinteractionthrough serial monologue, as in the ritual politics of the men's
house.And yet, exactlyparallel to the real irony of male sociopolitical life among
the Merina long ago pointed out by Ochs (Keenan 1974), men can strategically
orchestrate or authorizea lcrosin which their wivesor other female relatives are the
talkingfteds,thusplayinga potentially important role in issuesof contestedinterests.
Kroskrity'sdiscussionof Arizona Tewa showsthat that ideology of language
usecenterson te'e hi:li 'kiva speech',which seemsto be conceptualizedboth as a
genreandasa registerwith respectto servingas the informing centerpiece of what
outsiders seeas "linguisticconservatism"or "purism" more etically. Thus, insofar as
identityis manifestedthrough modalities of using language,all of the demographic
andattitudinalcharacteristicsof people are potentially aligned as metaphors one of

' would
It seemfrom Kulick's paper that the people of Gapun considerthe boJ to be a
nonologicgenre,in a way,andthe oratoricalspeechto be a dialogicone, in termsof their ideolory
of groundingthe valueof thesegenresas functionaleventsindexingsomethingof the personality
claracteristiaof the speakers.This provides another, subtle form of irony especially in the
chuacteristically
femalegenre,the lzos.
320 MichaelSilverstein

another by the conceptualizedproximity to the the kiva, in practice or in entitle-

ment, they can summonup in linguisticusage.Functionallysimilar to the hegemony
of standardregisterin linguisticcommunitieslike thoseof modern nation-states,that
is, dominating at the top-and-center of the conceptual array of "enregistered"
indexical forms, the forms of kiva register become potential superposedvariables
in the repertoire as a function of position and strategyof users.And we see thereby
the interesting irony that a woman announcing a yard sale in the manner of the
traditional male public crier, using enregisteredforms close to, though not at,
kiva-"standard."is neverthelesscommendedfor the effort.

Keeping the sites of ideologr in sight

I want very briefly to point to the differencesin explicitnessof the ideological as

such phenomena mediate social life, and also to the special position of certain
institutional sites of social practice as both object and modality of ideologization.
Again, the papers here demonstrate something of the commonalty as well as
diversity of ideologicalmanifestations.
A first important distinction is the degree to which, and the textualized
manner in which, ideologiesmanifestthemselvesin and as phenomenaldiscourses
about socialpracticessuchas usinglanguageitself.Suchmetadiscourses, even in the
partial, though mutually-reinforcingform that any of Hill's Mexicano consultants
articulated, frequently present themselvesas nomicallEcalibrated metapragmatics
- a metapragmaticsstipulating law-like regularities,independent of epistemological
concernwith time, place,or other eventivecontingenciesof the pragmaticpractice
that is its object of discursivefocus.(This nomic calibrationcan be contrastedwith
metapragmatic calibrations of a reportive or reflexive type [see Silverstein 7992:
48-53], which roughly deal with the realms, respectively,of "those" ald "these"
pragmatic phenomena. Myth, liturgy, and generalized historical or etiological
explanation of social phenomena tend to be in nomic calibration with their
denotationalobjects-of-exposition. )
Such a metadiscourseis semioticallysituatedto advantagein being a mode
of rationalizingexplanation,representingde-contextualizable characteristicsas the
basisof how indexicalsignsinstancetvpesof meaningfulness. Thus, any accountsof
how characteristicsof individuals recruited to the roles in semiotic events are
indexedby certain linguisticforms, have this characteristic:women vs. men engage
in, or are skilled at, distinctgenresof discourse- extensionalobservation- because,
women are such-and-suchand men are so-and-so--intensionalization in another
etiologicalschemagiving the "essence"of the socialcategorli these are the valued
vs. devalued forms of languagebecausethey are in keepingwith tradition or the way
of the ancestorsor not - the Malinowskian"chartermyth" about indexically-pregnant
forms; etc. When people have elaboratemetalinguisticideologiesof this metadiscu-
rsively-evidenced sort, this is one place for the socialscientistto start in figuring out
the nature of the indexicaldialecticthat is the socioculturalobject of investigation.
Sometimes,the metadiscourse emergesin descriptiveimagesof linguisticand
other semiotic usagein particular contexts,genred,eventively-situated stereotypes
of usage that form a kind of canon of ideological values. Hence Kroskrity's
kiva-style,the contextualizeduse of which links this context,through an ideological
 The uses and utility of ideologt 32I

imagingof it, to any other context in a kind of effectively intertextual - really,

interdiscursive - ideology that measuresany contextualizeduse of languageby the
degreeto which the kiva-talk can be seen as immanent in it. The Mexicano scenes
animatedby speakersas typical of days gone by, also seem to have some of this
canonicalcharacter as textualized Gestalten that are not so much constructed
ideologically by a discoursof intensionalrules or principles,as exhibited - "meta"ani-
mated- by a trope of re-presentationthat functionsas characterizingrepresentation.
Notein manyof the papershow such textualizedimagescombine with rationalizing
metadiscourses of principle in a complex narrated and expository genre.
Which brings me to point out that ritual sites - Arizona Tewa kivas, Gapun
meetinghouses,Zulu royal courts, etc. - and sites of understood ritualization of
usagc- prototypicscenesof honorification of a Mexicano compadre,for example -
are very frequently the descriptive exemplification in terms of which ideologies
articulatesemiotic value for indexicals and their determinate, textualizable
grounding in a universeof meanings.This is not by chance,of course,since in such
contexts of use,usageis in fact such that we can find tractable interactional texts.
Not onlysocialscientists,of course, but any concerned and interested parties can
appreciate the value-settingnature of ritual.
Thuswe seealsothat ritual is a site of tremendousideologicalpower, in that
- asDurkheimiantradition stipulates- it constitutesan autonomousmetatext for
socialpractices the indexicaldimensionsof which are implicated in the tropic figure-
mentsof ritual. Ritual is semioticallyself-grounding,in that its very enacted form,
inwhateversign modalities,constitutesits own reflexively-calibratedmetapragmatics,
at thesametime as this enactedform is constitutedout of tropes - metasigns- that
presuppose the existenceof nontropic, i.€., literal and absolute "reality." This
presupposition of the "literal"code - Barthes(1968:89-90)callsit, generalizing,the
denotational code- by an axiallyself-groundingentextualizationmay look non-"real
andfactual"to the outsider, but such suspicionis, of course, irrelevant in the
semioticuniverseof indexicality. We know of no ritual that does not get its
effectiveness in this manner, from the most mundane "performative" usage of
4 denotational language, to the most elaborate,scripturally-ordained liturgy. Ritual is
self-grounding as indexical-iconwithin its figured universe, and as such makes a
strong,thoughsemioticallyimplicit casefor what can only be called an ideological
orderwithin a culture. The site of institutionalized ritual and ritualization, then,
provides an essential placewhere societiesand socialgroupsin effect articulatethe
ideological.To figureout how a ritual is performatively efficacious,then, in its tropic
modalities, is frequentlythe key to being able to articulate an ideology seemingly
effectivein socialpractice. Further, as we see in the accompanyingpapers, such
sitesarethe foci of metadiscursively-evidenced ideologicalformations about social
life moregenerally,being powerful institutional forms in two senses,both in their
ordaining/(re)vitalizing effectivenesswith respect to the "denotational" codes they
regimentnomically,and as genred eventive images that are close to ideological
consciousness when rationalitycontemplatesitself.s

8
I hau.elaborated
someof the themesaboutthe powerof ritual and ritualizationin Silverstein
l9Z:passim,andin an asyet unpublishedpiece,"Metaforcesof power in traditional oratory,nms.
lS1. S€€alsonowthe importantpieceon "poeticsand performance"by Bauman& Briggs(1990),
322 Michael Silverstein

But the casesreported in the accompanyingshow the sitednessof ideology

in this way very clearly,with no further comment necessary.

The utility of the ideologr concept

Finally, I wish to make an observation on the very fact of existence of the

accompanyingpapers,namely, that they all use an analyticconcept of ideolory in
approachingtheir cultural data, which itself shouldbe a sign.Perhapsmerely - and
'the C-word'
unfortunately,if so - a postmodernistsignof why ,culture"is becoming
(though its adjectival usage,in cultural, ought to be fairly neutral [cf. fn. 1]). But
perhaps also because as linguistic anthropologistsdealing with the semiotic
complexitiesof language-use-in-context (thus being expansivein scholarlycharge),
we are faced first-off with indexical facts, facts of observed/experiencedsocial
practices, the systematicityof which is our central problem: are they systematic?if
so,how? with respect to whic& institutional forms? (re)aligning whosevalues in the
semioticallyseamlessdialecticof mind-in-history?
The total cultural fact - and I include here primordially the total linguistic
fact - consistsof answersto all thesequestions,organizedin terms of a claim of how
certain local practiceslocal knowledge is 'cultural'. As [r Maitre Saussurehimself
observed,theoreticalconceptsare only of real power when they dictate their own
methdology.Ideolory of linguisticand culturalphenomenais sucha concept,I think,
epistemologicallyuseful in making the phenomenal"realities"of both participants
and observers(these are roles, not individuals,of course!) experience-nearand
tractable.
The splendidpapersaccompanyingon which I have been askedto comment
abundantlyand richly make this clear.

References

Barthes,Roland (1968)Elententsof semiolog (Tlansl.,A lavers & C. Smith). New York: Hill &
Wang.

Barthes,Roland (1983) Thefashion system(Tfansl.,M. Ward & R. Howard). New York: Hill &
Wang.

Bauman,Richard & CharlesL. Briggs(1990)'Poeticsand performanc€as critical perspectives

on
languageand social life."Annual Reviewof Anthropologt19:59-88.

Blom,Jan-Petter& JohnJ. Gumperz(1972)"Socialmeaningin linguisticstructure:Code-switching

in Norway." In J.J. Gumperz & D. Hymes (eds.),Directionsin sociolinguistics:The ethnographyof
communicalrbn. New York: Holt, Rinehart& Winston,p.407-34.

reviewingmuc{ recent literature.

 The usesand utility of ideologt 323

Brown,Roger& Albert Gilman (1960)"The pronounsof power and solidarity."In T A Sebeok

Cambridge,Mass.:MIT Press,p.253-76.
(ed.),Sry/ein language.

Desturrde Tbary,Anroine, Comte (1827) [180U Eldmensd'idiologie; 1, Iddologieproprenrentdite.

Paris:[lvi.

Errington,J. Joseph(1988) Stntctureand stylein Javanese:A semioticview of linguistic etiquette.

Philadelphia:Universityof Pennsylvania
Press.

Raymond( 1981)Theideaof a citical theory:Habermasand the FrankfurtSchool.Cambridge:

Geuss,
Cambddge
UniversityPress.

Gumperz,JohnJ. (1%8) "Linguistics.III, The speechcommunity." In D. Sills (ed.),Intemational

of the social sciences.New York: Macmillan & Free Press,vol. 9: 381-86.
enqvlopedia

pattems.Philadelphia:University of PennsylvaniaPress.
labov, William (1972)Sociolinguistic

Michael(1979)"languagestructureand linguisticideolory."In P.Clyneet al. (eds.),The

Silverstein,
elements: on linguisticunitsand levels.Chicago:ChicagoLinguistic Society,p. 193-247.
A parasession

Michael(1981) "Metaforcesof powerin traditionaloratory.' l,ectureto Yale University

Silverstein,
Department of Anthropolog. [ms. in author'spossessionl

Michael(1984)'The lalue' of objectuallanguage."

Silverstein, Paperdeliveredat a symposiumof
theAmericanAnthropologicalAssociation's83rd annualmeeting.[ms. in author'spossession]

Michael( 1988)"De-voiceof authority.'Paperdeliveredat a symposiumof the American

Silverstein,
Anthropological
Association's87th annualmeeting.[ms.in author'spossession]

Michael(l9Z) "Metapragmaticdiscourseand metapragmaticfunction."In J. A Lucy

Silverstein,
langtage.Cambridge:CambridgeUniversityPress,p. 33-58.
(eA.),Refleive