Annu.Rev. Sociol. 1998.

24:345-70
Copyright? 1998 by AnnualReviews.All rights reserved

MeasuringMeaning Structures
John W.Mohr
Departmentof Sociology, University of California,SantaBarbara,California
93106-9430; e-mail: [email protected]

KEYWORDS:culture,meaning,network,structuralism,
institutions

ABSTRACT

The recent cultural turn in American sociology has inspired a number of

more scientifically oriented scholars to study the meanings that are embed-
ded within institutions, practices, and culturalartifacts. I focus here on re-
search that (a) emphasizes institutional (ratherthan individual) meanings,
(b) uses a structuralapproachto interpretation,and (c) employs formalalgo-
rithmsor quantitativeproceduresfor reducingthe complexity of meaningsto
simpler structuralprinciples. I discuss two core methodological issues-the
assessment of similaritiesand differencesbetween items in a culturalsystem
and the process by which structure-preservingsimplifications are found in
the data. I also highlight the importanceof two-mode analytic procedures
and I review some of the perceived benefits and criticisms of this style of re-
search.

INTRODUCTION
The cultural turn that has recently swept through much of American sociology
has meant that sociologists are ever more frequently focusing on the role of
symbols, meanings, texts, cultural frames, and cognitive schemas in their theo-
rizations of social processes and institutions. Although this resurgence of in-
terest in cultural phenomena is often associated with the shift towards more
humanistic and interpretative methodologies, an increasing number of quanti-
tatively oriented scholars have also begun to turn their attention to the study of
cultural meanings. In the process a new body of research has begun to emerge
in which social practices, classificatory distinctions, and cultural artifacts of
various sorts are being formally analyzed in order to reveal underlying struc-
tures of meaning.

345
0360-0572/98/0815-0345$08.00

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
346 MOHR

This work is scatteredacross many substantiveareas of research.It has no

coherentcenter or easily definable boundaries.There is, as yet, no clearly ar-
ticulated core methodology or statement of theoretical intent. As a conse-
quence, I adopta somewhatidiosyncraticset of inclusion rules for the material
covered in this review. My focus emphasizesstudiesthatmeasureinstitutional
(or cultural)ratherthan individualmeanings. I pay special attentionto those
projects that have (if only implicitly) adopted some variantof a structuralist
approachto interpretation.And I devote particularattentionto describingthe
ways in which formal analytic methodologies are being employed to reduce
complex collections of culturaldata to simpler, more easily intelligible struc-
tures of meaning.
I begin with a shortdiscussion of how the measurementof meaning devel-
oped in Americansociology up to and includingrecent attemptsto study insti-
tutional processes throughthe formal analysis of meaning structures.I then
highlight the main elements of a structuralapproachto meaning and describe
two methodological issues that are involved in conducting this type of re-
search-the measurementof similaritiesand differences, and the reductionof
complexity throughvarious types of formalanalyses. I focus next on one par-
ticularlypromisingclass of analyticalmethodsthatseems to me to hold out the
greatestpromise for futureprogress. I end with a brief discussion of some of
the advantagesand disadvantagesof formal approachesto interpretingmean-
ings.

MEANINGAND MEASUREMENTIN CONTEXT

Formalanalysis of meaning structuresis not new to the social sciences. Much
importantresearchhas been done by scholars in other disciplines, including
the pioneering work on semantic differentialtechniques by the psychologist
Charles Osgood and his colleagues (Snider & Osgood 1969). Cognitive psy-
chologists have carriedthis trajectoryforwardin a myriadof ways. Applica-
tions and extensions of this work have been pursuedby linguists, political sci-
entists, marketresearchers,and anthropologists.The lattergroupin particular,
and especially cognitive anthropologistssuch as Roy D'Andrade(D'Andrade
1995), have been especially instrumentalin developing the theory andmethod
of meaning analysis as a formal endeavor.
Formal analysis of meaning has also had a long-standinghome in at least
two areas of American sociology. Meaning measurementhas been a central
concern of sociologists who use survey methods to study opinions, attitudes,
and beliefs (Sudmanet al 1996). Contentanalysis of meanings in textual data
is a second areathathas been systematicallydeveloped since the early work of
BernardBerelson (Berelson 1952). However, with the exception of a few sub-
fields such as political sociology, where opinion researchcontinues to be im-

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
 MEANING 347
MEASURING

portant,both of these methodological projectshave become isolated from the

core researchtraditionsof sociological work. This is partly because the data
sources (opinion surveys and coded texts) have been difficult to acquire and
are generally limited to specialized areas of investigation. Many quantitative
sociologists also seem to believe thatthe dynamics of social structure(formal
organizations, social movements, processes of social mobility and status at-
tainment,and the like) can be measured,but that the more ephemeralaspects
of culturalmeanings cannot. Thus, the meaningful characterof social action
(and of institutionallife more generally) has appearedas part of the broader
theoreticalcontext which framesthese researchprogramsratherthanas some-
thing that is measureddirectly.
Recent work has begun to bridge the divide between culture and social
structure.The flourishing of symbolic interactionismand ethnomethodology
duringthe sixties contributedto an enduringand widespreadappreciationof
the socially constructed characterof the social world (Berger & Luckman
1967). Subsequenttheoretical statementsalmost universally emphasized the
dialectical or dualistic relationship between cultural meanings and social
structures (Bauman 1973, Sahlins 1976, Bourdieu 1977, Giddens 1984,
Swidler 1986, Sewell 1992). Slowly, these insights have come to be incorpo-
ratedinto the empiricalprojectsthatcharacterizecore regions of the discipline.
Culturewas first linked to the logic of the productionprocess in popularcul-
tureindustries(Peterson& Berger 1975, Peterson 1976). It was shown to be an
importantand measurable factor in the prediction of status attainmentout-
comes (DiMaggio 1982, DiMaggio & Mohr 1985, Mohr & DiMaggio 1995).
Institutionalistsdemonstratedthe significance of symbols and cultural pro-
cesses in the study of organizationsand their environments(Meyer & Rowan
1977, DiMaggio & Powell 1991). Social movement theorists showed the im-
portanceof cognitive frames(Snow et al 1986, Eyerman& Jamison 1991) and
culturalprocesses of identity formation(Damovsky et al 1995). Many other
examples could be pointed to.
However, the study of culturalphenomenais not the same as the analysis of
meaning. As Wendy Griswold has persistently complained, most of the re-
search continues to sidestep the problem of meaning analysis altogether.The
distributionof genres has been mapped,elements of culturalproductionhave
been counted, levels of culturalknowledge have been measured,increases of
organizationalhomogeneity have been demonstrated,but the meanings that
are constitutiveof these culturalphenomenahave largely been left aside. Gris-
wold has sought to develop methodological approachesthat incorporatesoci-
ology's empirical rigor and sophisticated understandingof social structure
while also taking the meanings embedded within literary and other cultural
texts as a critical element of the analysis (Griswold 1987a,b, 1992, 1993).
While she has been instrumentalin focusing attentionon this problem, Gris-

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
348 MOHR

wold's own work has largely been directedtowardthe study of literatureand,

althoughshe is rigorously empirical,the formalmeasurementof meaning has
not been her goal.

ACTORS,ACTIONS,AND OBJECTSOF ACTION:A

NEWER INSTITUTIONALISM
In this essay, I review many examples of sociological work that seeks to di-
rectly measureculturalmeanings. My focus will be largely methodological in
that I will organize the discussion aroundthe question of how this type of re-
search is conducted.But I begin with three recent examples that demonstrate
the potentialcontributionsof meaningmeasurementfor the empiricalinvesti-
gation of core sociological questions. What sets this work apartis the folding
togetherof culturalmeaningsand social structuresas primaryelements within
the same research design. Such an approachmuch more faithfully reflects
theorists' contentionthat social structuresand culturalstructuresare mutually
constitutive.Thus, in the following examples, meaningsaremeasuredin order
to show how social structuresare created.All three examples concernhistori-
cal changes in institutionalforms, a reflectioin of the fact thathistoricalsociol-
ogy has been an especially fertile terrainfor this style of researchbecause re-
searchers are constrained, by necessity, to analyzing meanings embedded
within texts (Franzosi& Mohr 1997).
CharlesTilly has been a long-standingleader in this style of research.His
recent work is especially instructive.Tilly (1997) seeks to explain the parlia-
mentarizationof British politics, that is, how the structureof political influ-
ence shifted from the local to the nationallevel while the natureof collective
claim-makingwas transformedfrom a reliance on spontaneousacts of (often
violent) protesttowardthe more formalizedgrievancemechanismsof political
parties and organized social movements. Though this might be viewed as a
problem of organizationalchange or of the formal transformationof political
structure,Tilly sees this as an occasion for culturalanalysis. He emphasizes
the need to understandhow the popularmeaning of politics changed during
these years, and especially how conceptionsof social rights shifted along with
ideas about how to make a collective claim in defense of one's rights. His
analysis relies on textual data, especially verbatim summariesof contempo-
rarynewspaperarticlesreportingon some 8,000 "contentiousgatherings"that
occurredbetween 1758 and 1834.
Tilly arguesthatshifting relationsbetween claim-makersand claimantsde-
fines the character of political institutions. For his analysis, Tilly divides
12,000 or so different contentious groups into 64 social categories (farmers,
friendly societies, workers, masters, constables, militia, parishioners,gentle-
men, local officials, and so on). He then looks for claim-makingrelationships

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
 MEASURINGMEANING 349

among these groups by analyzing his data on contentious gatherings to see

what actors made what kinds of claims (attacking,donkeying, petitioning, ar-
resting, applauding,addressing,and other similar efforts) against what other
actors.A blockmodel analysis, divided by time periods, shows how the mean-
ing and structureof political action changed dramaticallyduringthese years.
DiMaggio & Mullen (1993) provide a second example. Their study con-
cerns how American communities went about celebrating National Music
Week in 1924. Like Tilly, DiMaggio & Mullen are interestedin the way power
and influence are organizedand, also like Tilly, they see the social structureof
communitypolitics as being built up out of shifting systems of meaning. The
events studied here are not contentious gatherings,however, but formally or-
ganized communityrituals.They collected dataon 833 events occurringin 419
differentcommunities,focusing on the actors,actions, and objects of action in
each event. In this case, the actorswere the event participants(clubs, churches,
ethnic associations, employee groups, and the like). The catalogue of actions
included the various types of musical events that were planned-those that
tended to reinforce status group boundaries(e.g. religious or classical music
concerts) and those thathad a more inclusive focus (band and patrioticmusic,
group sings, and so forth). The objects of action were the types of audiences
that were encouraged to participate:Were they general admission events or
events thatwere staged in such a way as to assemble people who were defined
as membersof specific associations, congregations,or as residentsof particu-
lar institutions?
Once again, it is the relationsbetween these threesets of elements-partici-
pants, performances, and audiences-that DiMaggio & Mullen use to con-
struct a measure of the shifting institutional logics of community political
structure.What they discover are four relatively distinct modes of organizing
communityritualsthat they then link to the foundationsof political authority
in each community.They identify these as (a) "ritualsof ratification"thattend
to reaffirmtraditionalarrangementsof the social order;(b) "ritualsof commu-
nitas"that constitutecommunitymembersas individualswho are unmediated
by private associations but, nonetheless, members of a collective unity; (c)
"ritualsof civic unity"thatconstitutepeople as individualconsumersandpar-
ticipants in a local economy that all have a vested interest in supporting;and
(d) "ritualsof incorporation"that identify community members as workers,
membersof ethnic groups, and other diverse collectivities that must be some-
how integratedinto a whole.
The thirdexample is a study by Mohr& Guerra-Pearson(1998) of commu-
nity social welfare agencies in New York City during the Progressive Era.
Here the question is how particularkinds of organizationalforms come to be
institutionalized. Like much work in the organizational ecology tradition,
Mohr & Guerra-Pearsonfocus on the ways in which differenttypes of organi-

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
350 MOHR

zations compete for resourceniches by makingjurisdictionalclaims over spe-

cific regions of institutionalspace. Unlike work in this tradition,however, the
focus is explicitly shifted to the interpretationsthat organizationalactors pro-
pose about the meaning of institutionalactivities. As before, three sorts of
foundational elements are studied. Using written descriptions of organiza-
tional activities for some 600 organizations,Mohr& Guerra-Pearsoncollected
informationat fourdifferenttimes aboutstatuscategoriesthatwere used to de-
scribe relief recipients (men, women, boys, girls, children, sailors, travelers,
the working, able-bodied,and so on); classes of social problems(criminality,
delinquency, disability, immorality,and the like); and technologies of organ-
izational action (such as general relief, employment assistance, vocational
training,and character-building).
In this study, it is the structureof the organizationalenvironmentthat is
mapped.Relationsbetween organizationsare measuredaccordingto the simi-
laritiesof theirclaims aboutthe region of institutionalspace in which they seek
to operate.Similarityamong claims is definedby the extentto which organiza-
tions apply the same technologies to the same statusidentities,afflicted by the
same types of social problems. Differences of interpretationarise when alter-
native combinationsare invoked (as when differenttechnologies arepurported
to be effective for treatingthe same classes of problems). Multidimensional
scaling enables the organizations'niche locations to be mappedout on the ba-
sis of the claims thatwere made abouthow poverty problemsshould be inter-
preted and addressed.Using this approach,Mohr & Guerra-Pearsondemon-
stratehow settlementhouses waged a (losing) battlewith social work bureauc-
racies and othermore rationalizedorganizationalforms in theirbid to become
the primaryinstitutionfor delivering community social welfare services dur-
ing these years.
I have begun with these examples for several reasons. First, all three em-
ploy structuralmethodsfor measuringmeanings.In Tilly's case, the meanings
concern collective actors' sense of public rights and their conceptions about
the appropriatemeans for securingthose rights.DiMaggio & Mullen studythe
meanings of community and sets of emergent norms about how to symboli-
cally consecrate the social order.Mohr & Guerra-Pearsonanalyze organiza-
tional interpretationsof community social problems and contesting claims
about appropriatesolutions to those problems. In each case, the relevant sys-
tems of meanings are studied empiricallythroughan analysis of the relation-
ships thatactorsimpose on variousprimitiveinstitutionalelements (actors,ac-
tions and objects of action). Finally, all of these studies make explicit linkages
between the systems of meaningsthatareinvestigatedandthe social structures
in which they areembedded.Indeed,the point of each projectis to demonstrate
the ways in which enduringsocial institutionsare explicitly constructedout of
a complex process of negotiation and contestationover culturalmeanings.

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
 MEASURINGMEANING 351

THE STRUCTURALAPPROACHTO INTERPRETATION

The key methodological maneuverin these projectscan be tracedto the struc-
turalmethodof interpretation, generallyassociatedwith the semiotic, structural-
ist, and poststructuralistintellectual projectsthat thrived in Europebeginning
in the 1960s. Although diverse in both methods and goals, most of this work is
eitherin the traditionof or in responseto the work of the Frenchlinguist, Ferdi-
nand de Saussure [(1916) 1959]. Structuralism(in this Saussuriantradition)
was founded on the argumentthat meaning was constitutedthroughthe sys-
tematic distinctionsthat differentiatewords (or sounds, or signs) from one an-
other.A numberof analyticprinciplesfollow from this. Contentis seen as be-
ing fundamentallyarbitrary.Patternsof differenceswithin a broadersystem of
culturalobjects become the focus of analysis. The interpretationof meaningis
seen to be connectedto the analysis of the system of relationsthat link cultural
objects. As developed most famously in the writings of the anthropologist
ClaudeLevi-Strauss(1963), much of the work of interpretationcomes to be di-
rected towardthe identificationof underlying(deep) structuralprinciplesthat
serve to organize the largercomplexity of relationalpatterns.
Thereis an enormousliteraturethatdiscusses, evaluates,andcritiquesthese
intellectual projects. Useful introductionsare provided by Hawkes (1977),
Pettit(1977), Kurzweil(1980), Wuthnow(1987), andCaws (1988). Moreover,
as the discussion in Caws (1988), D'Andrade (1995), and Emirbayer(1998)
make clear,the developmentsof the Saussurianstructuraltraditionparalleleda
broadvariety of other intellectualmovements thatwere also shifting towarda
relationalmode of analysis duringthe same period. The researchthat I review
here is (for the most part)only loosely coupledto these literaturesand I will not
spend the time to develop these connections in any detail. I will simply point
out thatmost of the work thatI discuss in this review adopts(if only implicitly)
some type of structuralapproachto the problemof interpretingmeaning.
As an example of this mode of analysis, consider Karen Cerulo's (1988,
1995) researchon national anthems. Cerulo shows that national anthemscan
be analyzedas culturalmeaning systems and thatthe structuresthat are identi-
fied in the process can be used for comparativeanalysis. She accomplishesthis
by treatingmusic as a series of symbolic codes: "melodiccodes, phrasecodes,
harmoniccodes, form codes, dynamic codes, rhythmiccodes, and orchestral
codes." Musical notes are the elements that, when combined, generate these
code systems. Hence, it is the patternof relationsbetween the notes thatis rele-
vant;the musical notes themselves lack any specific intrinsicvalue, ratherthey
"derive their meaning according to their placement within a larger system"
(1988, p. 319). Througha complex and ingenious series of measurementsde-
voted to assessing the relationalpatternsof the notes, Cerulomaps out the mu-
sical structureof over 150 national anthems. The "meaning"of the music is

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
352 MOHR

thus representedby these structuralmappingswhich, as Cerulo demonstrates,

closely correspondto the sorts of genre distinctions(e.g. between a marchand
a hymn) that musicologists employ to differentiatethese styles of music. See
Timothy Dowd (1992) for a similar approachto analyzing Americanpopular
songs.
Cerulo (1995) proceeds to show that several basic social processes very ef-
fectively predict the type of musical composition that a specific country is
likely to adopt.In particular,she shows thatposition within the world political
economic system, relationshipto neighboringcountries,to a specific colonial
power, and to a cohortof othernations, all influence the characterof the musi-
cal compositionthatis chosen for a nationalanthem.She uses a similartype of
analysis to map out the basic structuralcharacteristicsof national flags (by
looking at how color, pattern,and images are arrayedrelationallywithin the
flag). Her analysis demonstratesthatthe same factors influencedthe selection
of these sorts of national symbols as well.
Cerulo's work illustratesthe core principlesof the type of structuralanaly-
sis that is the focus of this review: (a) basic elements within a culturalsystem
are identified, (b) the patternof relationsbetween these elements is recorded,
(c) a structuralorganizationis identified by applying a pattern-preservingset
of reductiveprinciplesto the system of relations,and(d) the resultingstructure
(which now can be used as a representationfor the meaning embeddedin the
culturalsystem) is reconnectedto the institutionalcontext thatis being investi-
gated. To begin such an analysis, threeproblemsneed to be solved. First,basic
elements of the culturalsystem have to be identified. Second, a relevant sys-
tem of relationsmust be identified. Third,the patternof relationsmust be re-
corded. These are not trivial tasks.
Initial decisions regardingwhat will count as the culturalelements whose
relationalarrangementwill be analyzed is critical to determiningthe signifi-
cance of the meaningstructuresthatcan be uncovered.Because Cerulowas in-
terested in identifying aesthetic structures,she turnedto expert literaturesin
graphic design and music theory for guidance in selecting relevant elements
for inclusion. This illustrationis generalizable.Every structuralanalysis must
begin with the identificationof a relevantset of culturalitems, and these items
are never simply available in an immediate fashion. An informed culturalor
institutionaltheory is necessary to enable us to notice and to be able to make
relevantdistinctionsbetween the constituentelements.
The same cautionarynote applies to the selection of relationalmeasures.
Elements of a cultural system can be similar or different in any number of
ways; the trick is to identify those relationsthat matter.Some of the most im-
portantcriticismsof the Frenchstructuralistswere thatthey frequentlygot lost
in their own overly formalisticmodels of structure.This is what Philip Pettit
describedas the dangersof mere "pattern-picking"(Pettit 1977, p. 41). When

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
 MEASURING
MEANING 353

one reads, for example, Clifford Geertz's (1973) criticisms of Levi-Strauss,it

is hardnot to be persuadedby his complaintsthatthe structuralistswere wrong
to focus solely on the culturalphenomenain isolation from what we might de-
scribe as the institutionalcontexts of culture.This was manifestedin theirten-
dency to treatculturalcodes as closed systems. For the structuralists,the logic
of culturewas all too often presumedto lie within the culturalsystem itself (or,
as in the case of Levi-Strauss,within some intrinsicorderingpropertiesof the
humanmind).
Thereis a solution to this dilemma,however;relationsof elements within a
cultural system should be determinedon the basis of how the elements are
linked to the practicaldemandsof the institutionalsystem of which they are a
part. This is an argumentthat goes underthe general heading of practicethe-
ory. Classic statements of this perspective include Bourdieu (1977), Geertz
(1973), and Giddens (1984). Useful summaries and commentaries can be
found in Ortner(1994) and Friedland& Alford (1991). The argumentis that
any culturalsystem is structuredas an embodimentof the range of activities,
social conflicts, and moral dilemmasthat individualsare compelled to engage
with as they go about negotiating the sorts of everyday events that confront
them in their lives. This insight has direct implications for the measuringof
meaning structures.It suggests that when we think about identifying a set of
culturalitems, asking how they are related to one another,and assessing the
type of structuralmodel thatmight be relevant,it is importantto begin with the
question of what type of practicalutility such a culturalsystem plays within a
concrete institutionalsetting. Ideally, relationsbetween the culturalelements
should be assessed by looking at how actors, organizations, or institutions
make practicaluse of the culturaldistinctionsbeing investigated.

MEASURING RELATIONS
Once the relevant elements and relationalcontexts of a culturalsystem have
been identified, measurementcan begin. The first task is to compareeach ele-
ment in the culturalsystem with every otherelement in terms of the identified
relationship.Fourtypes of measurementstrategiescan be distinguished.Simi-
larities and differences among culturalitems can be assessed accordingto (a)
subjectivejudgements, (b) common attributes,(c) relationsto others, and (d)
structural-functionalprofiles.
Subjective Similarity
Two culturalitems can be said to be similarto one anotherto the extent thatin-
dividuals make a cognitive judgement that they are of "the same sort." Be-
cause they often arefaced with the need of collecting datain differentlanguage
situations or from nonliterate respondents, cognitive anthropologists have

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
354 MOHR

been especially innovative in devising a great many alternativemethods for

collecting these kinds of data(see Hays 1976, Weller & Romney 1988, D'An-
drade1995).Forexample,respondentsareaskedtojudge the similarityof pairsof
culturalitems directly, or they are asked to select which two out of three items
are most similarto one another.A thirdmethodis to ask the respondentto sort
cardsreferringto culturalitemsintopiles of thingsthat"gotogether."Scorescan
then be aggregatedacross a numberof respondentsto derive a set of average
measuresof inter-itemsimilarities.RichardShwederused these techniquesto
investigatehow individualsin differentculturesunderstandthe concept of the
person(Shweder& Bourne 1991). Forexample, he used pile sorttechniquesto
assess how individuals in India organize their understandingof 81 different
personality traits. Included here were terms translatedas "crooked,""lazy,"
"obstinate,""brutallyfrank,""loyal," "harmless,""fickle," "contemplative,"
and so on. The resulting 81 x 81 matrix was used by Shweder & Bourne to
demonstratethat the respondents' culturalunderstandingof these terms was
structuredaroundgeneralizeddimensions of power and social desirability.
Attribute Similarity
Items within a culturalsystem can also be comparedon the basis of the sets of
selected attributesthat they share. This type of technique can also be used in
collecting interview data. Csikszentmihalyi& Rochberg-Halton(1981) em-
ployed this strategy in their study of the meaning of household possessions
among Chicago families. A list of items was first elicited by asking respon-
dents "Whatarethe things in your home thatare special for you?"Interviewers
then asked aboutthe meanings that were attachedto each of these items. This
resultedin a listing of 1694 objects (collapsed for analysis into 41 types). The
respondent's meaningful associations were classified into 37 categories of
value, such as, the object is special because it is a memento,orbecause it is part
of a collection, or because it belonged to a specific relative, or because it re-
flects a specific achievement, etc. The 41 types of culturalitems were then
comparedto one anotherin terms of the types of valuations associated with
each. The same generalmethodwas used by Mohr& Guerra-Pearson(1998) to
assess the discursive similarityof community social welfare organizationsin
the article discussed earlier.In this case organizationswere comparedto one
anotheron the basis of the similarity of the claims that they made about the
types of solution technologies that were seen to be appropriatefor classes of
persons and types of problems.
Relational Similarity
Similarityamong items within a culturalsystem can also be assessed by look-
ing for the presence or absence of varioustypes of social relationshipsthatlink
cultural objects together. Thomas Schweizer (1993) also studied household

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
 MEASURINGMEANING 355

possessions (in an urbanneighborhoodin FrenchPolynesia, a groupof hunter-

gatherersin Zaire, and a peasant community in Java). In Schweizer's study,
however, categories of possessions were seen to be similar or different from
one anotherdepending upon which individuals in the community possessed
the same sets of items. So, for example, on the basis of who owned what in
Papeete, a refrigeratorwas more like a radio and a bicycle was more like a
kerosene stove. Thus, Schweizer interpretsthe statusmeaningof variousmate-
rial possessions by understandingwhat members of the community possess
what types of goods. Also importantare measuresof similaritybased on rela-
tionships that directly connect the culturalitems (or elements) to one another.
Boorman & Levitt (1983) used this method to study how meanings were or-
ganized in federalbankruptcylaw. They coded the internalcitationreferences
of the 55 subsections of Article 9 of the Uniform Commercial Code. Thus,
ratherthan asking individualsto sort culturalitems that went together, Boor-
man & Levitt asked which subsections of legal code were "substitutable"in
the sense that they were similarly referencedby other subsections of the legal
code. This is the same type of measurethatwas used by Tilly (1997) in his par-
liamentarizationarticleto measuresimilaritiesamong categories of collective
actors. Those groups that had similar profiles of relationships to all other
groups were linked together in the blockmodel analysis as being structurally
equivalent.
Similarity of Structural Function
In some situationsthe measurementof similarities and differences has to in-
corporatemore complex information.To assess the structuralmeaning of cul-
tural items embeddedwithin sequences and narratives,it is necessary to pre-
serve more than pairwise informationon similarities. It is also necessary to
know how items are located within the broadercontext of the narrativeor se-
quential orderingof events (Abbott 1983, 1990). For example, most contem-
poraryapproachesto content analysis requirethat words be comparedon the
basis of their syntactic or narrativefunction within the text. How one should
gatherthis data is very much underdebate (see Roberts 1989, Franzosi 1989,
1990, Carley 1993, Carley & Kaufer 1993). In general, however, most strate-
gies call for the preservationof two types of information-the semantic func-
tion of a word within a field of meaning (e.g. within a sentence) and the spe-
cific relationshipsthat occur within that field to other (functionally defined)
words. For example, Roberto Franzosi proposes that similarities of words
within a text be recordedin such a way that their functions within a localized
semantic grammarare preserved.This is an importantextension (and formali-
zation) of Tilly's methods. In his study of contentious gatheringsin Italy be-
tween 1919 and 1922, Franzosi(1997) coded newspaperstories so thatvarious
categories of agents (laborunionists, owners, brownshirts,police, and various

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
356 MOHR

others) are comparedto one anotheraccordingto the extent to which they ful-
filled the same institutionalfunction (instigatorsof violence, recipientsof vio-
lence, and so on). To get at this kind of informationit is necessary to record
who does what to whom, and in what sequence.

FINDING STRUCTURE
Having collected informationregardingthe relations of similarity (or differ-
ence) among a set of items within a culturalor institutionalsystem, the next
task is to find structure-preservingsimplifications that may allow the com-
plexity of the system to be more easily understood.Ideally, one hopes to iden-
tify some deeper,simpler,structurallogic-that is, a principleor set of princi-
ples that account for the arrangementof partswithin the culturalsystem (Ort-
ner 1994). The relationsbetween culturalitems are the key to such an investi-
gation. The analyticaltask is to discover how these relationsare relatedto one
another.Structuralistmethodsare gearedtowardthe identificationof transfor-
mationsthatallow the relationsamong the relationsto be reducedto more eas-
ily understandableand or visible patterns. I discuss four general classes of
methods for accomplishingthis: (a) multidimensionalscaling and clustering,
(b) networkanalysis, (c) Boolean algebra,and (d) sequence analysis. Each of
these approachesmakes use of the relationalinformationin differentways in
orderto highlight various qualities of the meaning structure.
MultidimensionalScaling and Clustering
Multidimensionalscaling analysis (MDS) is one of the oldest and most widely
used methodsfor mappingout the relationalsystem of differencesin the meas-
urementof meanings. Simply stated,an MDS analysis readsin a squarematrix
of similarities or differences (actually half of a squarematrix is usually used
because the inputdataareoften symmetrical)andproducesa transformationof
the data that seeks to locate all of the objects in a common (two- or more di-
mensional) space in such a way that the similarities in the input matrix are
transformedinto Euclideandistances. An MDS space would thus representa
series of objects in such a way that(generallyspeaking)if two items are similar
to one another(in the input matrix)then they are located near one anotherin
the space. If they are dissimilar,they are located far apart.When analyzed in
two (or three)dimensions,the items can thenbe easily plotted in such a way as
to visually convey the relationalstructurein which they are embedded.
Clustering methods have also been used for studying cultural meanings.
Thereare a greatvariety of clusteringmethods,but all seek to group items to-
gether according to some algorithmicprinciple for deciding which are most
alike given the input matrix of item by item similarities. One of the primary
differences between clustering methods and MDS is that the latter locates

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
 MEASURINGMEANING 357

items in a way thattakes into accountthe totalityof all relationalsimilaritiesin

each iteration,whereas clustering methods tend to connect items to clusters
that are deemed (by some criteria)to be most similar in a localized (or pair-
wise) sense. The two methods are often used together(as when the coordinates
of an MDS space are submittedto a cluster analysis in orderto more clearly
designate subregions within the space). The Sage books on these topics are
both excellent and still quite useful (Kruskal & Wish 1978, Aldenderfer &
Blashfield 1984).
These techniques have been widely used by cognitive anthropologistsfor
studying meaning systems. As D'Andrade (1995) explains, the 1970s were a
time when basic theories about cultureas knowledge systems were connected
to formalmethods of analysis. It was duringthese years that many of the clas-
sic applicationsof these techniques to understandingthe structureof cultural
meaning systems were pioneered. For example, Burton & Kirk (1977, 1979)
used MDS to model gender assumptionsin Maasai culture. They asked their
respondents to think about the personality characteristicsof various social
identitiesdifferentiatedby age andgender(small boy, olderboy, warrior,adult
male, small girl, older girl, and young adult woman). Using triad tests, they
collected informationon similaritiesamong 13 personalitytraits(disobedient,
brave, frank, hardworking, socially competent, playful, lazy, respectful,
fickle, suspicious, clever, skittish, and stingy).
Theirfindings showed thatsmall boys were expected to be playful, success-
ful, brave, respectful, clever, and disobedient,but as they aged, they were ex-
pected to also become hardworkingandto no longerbe disobedient.Once they
became warriors,however, expectations shifted so that they were expected to
be realistic and, once again, disobedient,while stinginess was frowned upon.
Burton& Kirk relate these shifts in culturalbeliefs to changes in males' roles
within the community.For example, as warriors,young men are often on raid-
ing partiesaway fromhome where they areexpected to use theirown initiative
and, when appropriate,to disobey the commandsthathad been given them by
the elders in the village. They arealso expected to work togetheras a group,re-
lying on one anotherfor survival,makingstinginess a significantliability. The
analyses show a very differentpatternfor women. Young adultwomen are ex-
pected to be lazy and stingy and they are expected not to be successful, clever,
or brave. Kirk & Burtonnote that the shift in normativeexpectationsfor men
and women that occurs after circumcision is striking. For men, the change
marksa time of ascendantauthorityand autonomy.Forwomen it marksa clear
accelerationof negative stereotypingand formaldisempowerment.They con-
clude that a primaryfunction of circumcision rituals in Maasai culture is to
dramatizeand institutionalizedifferences in power between men and women.
Numerous other examples from this period of anthropologicalresearchcould
be pointed to. Anotherclassic example is the study by Roy D'Andradeand his

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
358 MOHR

colleagues in which they used MDS to model the structureof meaningsregard-

ing disease categories among English and Spanish speaking students(D'An-
dradeet al 1972).
Sociologists have traditionallyused MDS techniques to study social net-
works or social organizationalstructures,such as in Laumann & Knoke's
(1987) study of the structureof policy domains. More recently, sociologists
have begun employing MDS analyses to map out culturalmeaning structures
in a fashion thatis more reminiscentof the work of anthropologists.For exam-
ple, Ennis (1992) and Cappell & Guterbock(1992) used MDS techniques to
model the structureof American sociological specialties. Ventresca & Yin
(1997) used MDS to analyze the meaning of social categories in a broad his-
torical sample of nationalcensus surveys. Mohr (1998) uses MDS techniques
to analyze gendered assumptions in the structureof poverty classifications
used by ProgressiveEra social reformers.
Network Analysis
The one subfield within Americansociology thathas always been thoroughly
structuralist(in the relational sense) is social network analysis. Typically in
this researchtradition,individualsarethe nodes, andvarioustypes of social re-
lationshipsthat link individualsto one another(friendship,kinship, social ex-
changes of various sorts) constitutethe "ties"out of which a patternof social
organizationis constructed.Over the years, networkscholarshave developed
an enormousarrayof conceptualand methodologicaltools for thinkingabout
the structuralpropertiesof social networks. The methods can be divided into
two general types-those emphasizing connectivity and those focusing on
structuralequivalence (Burt 1978, Wasserman& Faust 1994). The formerare
largely concernedwith how the relationshipsbetween individuals in a social
networkare mediatedby the structureof ties thatdirectlyconnects them. Con-
cepts like networkcentrality,between-ness, and adjacencyare especially im-
portanthere. The latterapproachfocuses on ways in which individualsstandin
structurallyequivalentpositions because they share common patternsof rela-
tionships to all others in the network. The key concept here is the idea of a
structuralrole (White et al 1976). While most of the empiricalwork has been
directedto understandingthe propertiesof social networks,these methodscan
also be extremelyuseful in the study of culturalmeaning structures.
KathleenCarleyhas been one of the pioneers in applyingnetworkmethod-
ologies to the study of culturalphenomena,especially in her work on textual
analysis (Carley 1986, 1993, Carley & Kaufer 1993). Carley uses network
analysis techniques to map out the structureof meanings within narratives.
Like Cerulo, Carley then proceeds to use these structuralrepresentationsof
meanings to compare culturalphenomena. For example, Carley (1994) uses
these techniquesto map the structureof meaning in 30 science-fiction novels

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
 MEASURING
MEANING 359

written duringdifferentperiods of history. She codes informationon ways in

which the authorsdescribe the features,actions, and other characters'percep-
tions of robots.By studyingthe relationshipsbetween the featuresthatwere at-
tributedto robots by the novels' authorsas a network of concepts, Carley is
able to representchanges in the meaning structureof how robots were por-
trayedover time. Priorto 1950 they were describedas dangerousnonmetallic
humanoids composed of batteries, electron tubes, and human parts that in-
spired fear, anger, hatred,and pity in others. By the 1980s they were instead
being describedas clever, loyal, curious,and capableof sarcasm,being embar-
rassed, making love, programmingcomputers,and dreaming.
A very different example can be found in Hammond's (1972) analysis of
the ruinsof Mayanarchitecture.He employed networkmodels of connectivity
(between plazas) as a way of studyinghow public space was used. On the basis
of these models, Hammondwas able to deduce which plaza served as the cen-
tral marketplaceand to postulatethat the Mayan ballcourtgame was, initially
at least, a sportreserved for an elite audience.
Structuralequivalence approachesto networkanalysis have also be used to
understandculturalphenomena.Mohr(1994) uses this methodto mapthe struc-
ture of moral discourse within which poverty relief agencies differentiated
amongvariousclasses of genderedstatusidentities(mothers,soldiers,widows,
workingmen, the blind, andvariousothers)duringthe ProgressiveEra.Here, it
is the status identitiesthemselves that are the items within the culturalsystem.
Relations among identities were measuredby assessing the profile of relief
practices that were applied to each (58 differentpractices were investigated,
e.g. some classes of the poor were given money by the state, otherswere placed
in the poorhouse, some were given counseling by the church, others were
given job training by private agencies, and so on). A block-model analysis
showed that it was possible to identify differentrole positions within the sys-
tem of moral discourse that closely correspondedto Theda Skocpol's (1992)
argumentsaboutAmericanwelfare politics duringthese years. The discourse
role analysis was also shown to be an effective predictorof the likelihood that
any given status category would be describedwith morally coded terms.
Other examples can also be pointed to. Peter Bearman(1993) uses struc-
tural equivalence to show how changes in the social organizationof elites in
16th-centuryEnglandled to the emergence of abstractreligious and constitu-
tionalist rhetorics. Anheier & Gerhards(1991a) use structuralequivalence
measuresof the role structureof Germanwritersto explain the developmentof
culturalmyths aboutwriters-the myth of the writeras a "notoriousloner,"as
"apoorpoet,"as "themisunderstoodgenius,"andmemberof a "romanticliter-
ary circle" are all shown to correspondto the social organizationof German
writers. Finally, PatrickDoreian (1987) uses structuralequivalence measures
to reanalyze Hammond's data on Mayan architecture.

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
360 MOHR

Boolean Algebra
Boolean algebra is an increasingly popular method for formally analyzing
qualitativedata.CharlesRagin (1987) has been instrumentalin bringingthese
methods to the attentionof sociologists in recent years. This approach(also
known as qualitativecomparativeanalysis, or QCA) begins with a set number
of cases that vary in terms of the presence or absence of some outcome (or
product).Cases are groupedaccordingto theirprofile of features.Boolean al-
gebrais then appliedto identify logically irreducibleandnonredundantcombi-
nations of features that are associated with certain outcomes. As in network
methods,the methodpresumesa finite numberof cases within a system (rather
than a sample from a population),and the focus is on qualities (presence and
absence ratherthan measuredquantities)of the features.Unlike the methods
discussed above, the usual goal of a Boolean analysis is to ascertainwhich fea-
tures of a set of cases have a causal relationshipwith some other feature (or
outcome) that is being explained. In this sense, QCA is similar to logistic re-
gression analysis, althoughas Ragin explains (1995), there are variety of rea-
sons why one might preferusing QCA, includingthe fact thatit can be applied
with many fewer cases.
Because of the emphasis on hypothesis-testing, QCA has generally not
been appliedto the analysis of meaningstructuresin the same manneras MDS
or networktechniques.Thereare some exceptions.Degenne & Lebeaux(1996)
use Boolean techniquesto analyzethe structureof belief systems.Using French
survey datathey show a causal orderingof religious beliefs and practices (re-
spondentswho pray regularlyalso go to churchregularlyor they "believe in
paradise,purgatory,hell," both of which imply thattheirchildrenwill be given
a religious education, and so on). More generally, however, applications of
Boolean analysistendto be very much in the spiritof the work describedin this
review, and to resonate especially well with those works in which meanings
are seen to be a constituentelement of institutionalpatterns.Forexample, John
Foran(1997) uses QCA to analyze the role that five factors(dependentdevel-
opment, economic crisis, repressive/personaliststate, world system opportu-
nity, andpolitical culturesof resistance)have played in the success, failure,or
quiescence of a dozen recentrevolutionarysituations.Forandemonstratesthat
all factors must be present for success, and that the last three of these factors,
including a particularculturalorientation,are especially salient.
Sequence Analysis
Sequence analysis is another qualitatively oriented methodology that is in-
creasingly being employed by sociologists. These methods are used for find-
ing reduced form patternsin the sequencing of events throughtime. Andrew
Abbott (1988, 1992) has been instrumentalin bringing these methods into

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
 MEASURINGMEANING 361

mainstreamsociology. This is anothermethodological approachin which ele-

ments within a system are comparedon the basis of their similaritiesto one an-
other.These kinds of methods can also be used for identifying certaintypes of
culturalmeaning structures.
Abbott & Forrest(1986) used these methods to understandculturaldiffu-
sion among communitiesof English MorrisDancers.By studyingthe sequenc-
ing of dance steps, Abbott& Forrestwere able to identify common dancetradi-
tions that they used to analyze the diffusion of cultural forms among nine-
teenth-centuryruralEnglish communities.Othermethodshave also been sug-
gested for analyzing narrativestructures.Peter Abell's (1987) work empha-
sizes the comparisonof sequences in multiple narratives.David Heise (1988,
1989) has developed Event StructureAnalysis (ESA) and a computer pro-
gram,ETHNO,for its implementationwhich has been appliedby LarryGriffin
(1993) to studyingthe unfolding of historicalevent logics. Here again, a struc-
tural representationis created of a particularinstitutional(meaning) system
that can then be used as a formalbasis for comparativeanalysis.

MEASURINGTHE DUALITY OF SOCIALAND

CULTURALSTRUCTURES
I startedthis review by noting thattheoristshave emphasizedthe dualityof so-
cial and culturalstructuresand I suggested thatone importantreasonfor meas-
uring meaning is that doing so allows us to take this theoreticalmandateseri-
ously. The threeexamples of institutionalanalysiswith which I began were se-
lected to illustratehow the measurementof meaningcan be used to analyze the
culturalconstructionof social structures.It should be apparent,however, that
we also need to attendto the ways in which social structuresproduce cultural
meaning systems. Indeed, my earlierdiscussion of practicetheory would sug-
gest that such an approachis critical for culturalanalysis. My focus until now
has been to suggest some ways in which meaning structurescan be measured.
In this section, I discuss several methods that can be employed to focus atten-
tion on the duality that inheres between culturaland social structures.These
arethe class of methodsthatare describedas two-mode dataanalyticstrategies
because they simultaneously order both columns and rows of a data matrix.
Several methods are available.

Correspondence Analysis
Probablythe most familiarexample of this kind of approachis correspondence
analysis (Weller & Romney 1990). Like MDS, correspondenceanalysis can be
used to representa set of culturalitems in a dimensionalspace, therebyallow-
ing theirunderlyingstructuralpatternsto be representedvisually. At the same
time, correspondenceanalysis enables the relational foundations of the cul-

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
362 MOHR

turalsystem to be treatedas a set of (social-structural)elements thatare them-

selves ordered in terms of their relations to the cultural domain. Thus, this
methodology allows both the social and the culturaldimensions to be plotted
within the same measurementspace.
As an early proponent of practice theory, Pierre Bourdieu was naturally
drawntowardthis methodology and he has used it extensively in his research.
In Distinction (1984), Bourdieu uses data on the culturaltastes of different
class fractionsto identify the class based logic of culturalgoods. Thus, Bour-
dieu shows thattaste in music, film, painting,recreationalactivities, and other
cultural domains is organized according to the structureof social classes in
France, and using correspondenceanalysis, he is able to simultaneouslyplot
the various social locations (e.g. the determinantsof class structure)according
to theirrelative location within the space of culturaldistinctions.These analy-
ses have been centralto Bourdieu's theoreticalgoal of measuringthe relation-
ship between social and culturaldomains, and he used them to demonstrate
that, for example, class locations are orderedaccordingto two salient dimen-
sions-the total volume and relative composition of cultural and economic
capital.
Because of its ability to representthe dualityof mutuallyconstitutedstruc-
tures, correspondenceanalysis is beginning to be used more frequently by
American sociologists as well (althoughother approacheshave also begun to
be developed, such as the proposalof Borgatti& Everett 1997 for using MDS
to analyze two-mode data).Ann Mische (1998), for example, uses correspon-
dence analysis to analyze how youth movements in Brazil were structuredby
the types of political beliefs which they held at the same time thatthe political
discourse itself was structuredby the groupswho were constitutingit.
Lattice Analysis
Galois lattices are anotherapproachthat is explicitly orientedtowardthe rep-
resentation of the dual ordering of rows and columns. Whereas correspon-
dence analysis can be viewed as a two-mode extension of MDS techniques,
lattice analysis provides a dual mappingof rows and columns based on Boo-
lean set theory.
Mohr& Duquenne(1997) use lattices to analyze the changing institutional
logics of poverty relief duringthe Progressive Era. Here the duality between
culturalmeanings and social practices is operationalizedexplicitly. Reform-
ers' classifications of the poor (distressed, destitute, fallen, deserving, home-
less, indigent, misfortunate,needy, poor, stranger,and worthy) are shown to
be embeddedwithin a hierarchicallyorderedmeaning system by the structur-
ing of organizationalpractices(giving advice, giving food, giving money, pay-
ing a person to chop wood, placing a relief applicantin an asylum, and so on).

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
 MEASURING
MEANING 363

The use of lattices demonstrateshow it was that the culturaland practicallo-

gics were mutually constitutive. In other words, the orderingof the poverty
classifications is shown to be determinedby the orderingof the relief activities
and vice versa. By mapping the changes in this dualistic structurethrough
time, Mohrand Duquennedemonstratehow the institutionallogic of the nine-
teenth centurypoorhouse system was replacedby a far more progressive sys-
tem founded in the rhetoricand practice of social work professionals.
The paperby Thomas Schweizer (1993) discussed earlierprovides another
example. Recall that Schweizer analyzes the relationshipthat inheresbetween
the status logic of materialpossessions and the social rankingof individuals.
He shows thatthe meaning of possessions can be interpretedby understanding
what membersof the communitypossess what types of goods. Simultaneously
(and dually), Schweizer assesses the rankingof individual communitymem-
bers in the social orderby observing what materialgoods they possess. The
structuralduality of these two orders(the individualmembersof the commu-
nity and the materialpossessions thatthey hold) consists in the fact thatthe or-
deringof one is simultaneouslydependentupon the orderingof the other.Vin-
cent Duquenne(1995) provides an additionalexample of the use of lattices to
analyze status ordersand materialwealth.
Hierarchical Classification Models
A final example of two-mode analytictechniquesis the hierarchicalclassifica-
tion model developed by Paul de Boeck and SeymourRosenberg,which is im-
plementedin the softwareprogramHICLAS(Rosenberget al 1995). This is an
interativealgorithmthat accomplishes the same goal as a lattice analysis by
employing set theoreticalprinciples to cluster items in the rows (objects) and
the columns (attributes) of a two-way two-mode binary matrix such that
equivalentitems are classed together,the classes of objects are hierarchically
ordered,the classes of attributesarehierarchicallyordered,andthe two hierar-
chical ordersare relatdto one another.
This technique has been employed to measure several types of cultural
meaning systems. Anheier & Gerhards(199 lb) use this methodto understand
how famous writers (Ernest Hemingway, William Shakespeare,and others)
were organizedinto a field of literaryinfluence by comparingwhich contem-
porary German writers cite them as having been inspirationalin their own
work. The results allow Anheier & Gerhardsto describe the dual mappingof
blocks of contemporarywritersonto a clusteringof literaryinfluences in much
the same way as Bourdieu's analysis maps social classes and culturaltastes in
the same measurement space. Rosenberg (1989) uses HICLAS to analyze
meaning structuresin Thomas Wolfe's autobiographicalnovel, Look Home-
ward,Angel. In this case the set structureof various attributesof individualsis
dually mappedonto the set structureof characterswithin the novel. Rosenberg

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
364 MOHR

shows with this methodthatmembersof Wolfe's family are organizedin clear

subsets of attributes(his brotherFred is a subset of his sister Mabel, who is a
subset of his mother).Rosenbergalso demonstratesthatWolfe's descriptionof
his own identityat differentages is hierarchicallynested, and thatWolfe's de-
scriptionof his self shifts throughlife to being more and less like variousmem-
bers of his family.

ADVANTAGES AND DISADVANTAGESOF

FORMALANALYSIS

Precisely because the measurementof meanings lies at the boundariesof the

more scientistic and the more humanisticapproachesto analyzing social phe-
nomena, it is guaranteedto createcontroversy.NormanDenzin (1991), for ex-
ample, complainsthatscientifically orientedstudies of culturehave a tendency
to reify andreducethe intrinsiccomplexities of culturalmeanings.Thereis no
doubtthathe is rightaboutthis. All measurementprojectsend up reifying real-
ity to some extent. Moreover,it is clear that there are othermore hermeneutic
and post-structuralistapproachesto the problemof interpretationthatproceed
along very differentpathwaysandhave a greatdeal to offer. It would be a mis-
take to presumethatthe type of structuralistmodels of meaningthathave been
reviewed here are in any sense exhaustive or even that they should be ceded
some sort of empiricalprimacy.
Measuringmeaning structuresdoes have its benefits, however. As Robert
Merton(1957) suggested in his classic essay regardingthe advantagesof em-
piricalresearch,the use of formalmethodsbringsa pressurefor the clarification
of one's concepts and may lead to the discovery of anomalous findings that
call theoreticalassumptionsinto question.Moreover,as Hage & Harary(1983,
p. 9) arguedin their attemptto persuadeanthropologistsof the advantagesof
graphtheory, the shift towardquantitativerelationalmethodologies also has a
numberof more specific advantages.While my focus in this review has been
somewhat different,I think Hage & Harary'slist is applicablehere as well. In
somewhatmore generalizedform, these advantagesare the following:
1. Structuralmodels are iconic. They look like what they represent.Whether
one uses MDS techniques to generate a picture of the similarityrelation-
ships between culturalitems or one employs a lattice to map out the hierar-
chical dualityof culturalitems and actions, these techniquesallow us to see
the patternsof difference out of which meanings are constructed.Once we
see these patterns,we can begin to understandthem.
2. Structuralmodels provideus with a rich conceptualvocabularyfor thinking
about meanings. Simplicity, complexity, centrality,duality, permutations,
and transformationsare ideas that have a naturalfoundationin structural

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
 MEASURINGMEANING 365

models. As we tryto grapplewith understandingphenomenasuch as cultural

boundaries,culturalidentities,culturaldifferentiation,and culturalnarrativ-
ity, I suspectthatthese kindsof structuralconceptswill be of muchuse to us.
3. Structuralmodels provideus with the tools for quantifyingaspects of mean-
ing structures.As I have tried to indicate here, we can quantify meanings
without reducing them to some artificially linear metric. Instead we can
build structuralmodels that simply representrelationshipsof similarityand
difference that are embeddedwithin practices. Once we have found a way
of measuringthese kinds of phenomena,however, we (and others) are in a
position to be able to replicateour interpretationsandto subjectour ideas to
variouskinds of formaltests to help assess the validity of ourinterpretation.

CONCLUSION
While I will be pleased if qualitativelyorientedscholars are persuadedby the
merits of these argumentsand the utility of these techniques,my primarygoal
has been to convince sociologists who alreadyuse quantitativemethodsto em-
brace the measurementof meaning and to incorporatethese measures into a
more balanced approachto social researchthat recognizes the duality of cul-
tural and social structures.In this regard I fully agree with HarrisonWhite
when he argues that "interpretiveapproachesare central to achieving a next
level of adequacy in social data"(1997:57-58). I also agree with Jepperson
and Swidler's (1994) contentionthat cultureis no more intrinsicallydifficult
to measurethan other social phenomenaand that the greatestimpedimentsto
the formal analysis of cultureare conceptualratherthan methodological.
I have focused here on one very basic level of conceptualization.I have ar-
gued thatculturalmeanings arebuilt up out of structuresof difference andthat
by attendingto the patternsof relationsthatlink items within a culturalsystem,
we can use formalmethodologiesto measureand analyzemeaningstructures.I
have also suggested that a critical componenetof any such endeavoris deter-
mining the linkage between meanings andpracticesand thatrelationsbetween
the formershould be determinedby their embeddednesswithin the latter,and
vice versa. Beyond this, I have said very little here aboutwhat purposesthese
measurementsshouldbe put to or how culturalanalysis shouldproceed. I have
done this quite intentionallybecause I believe that these are general purpose
methods that can be employed to accomplish a great variety of things. This
should not be takento indicate thata theoreticalcontext for theiruse is unnec-
essary or undesirable.On the contrary,to use these methodswithout a theoreti-
cal goal in mind is unlikely to yield much that is of sociological interest.
In this review I have provided a sampling of the types of researchprojects
that are beginning to appearin the sociological literaturethat seek to analyze
the structureof institutionalmeanings.As I suggested earlier,thereis as yet no

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
366 MOHR

coherentbody of literatureor collection of scholarsidentifiedwith this project.

Thereis, however, tremendousenthusiasmandopportunityfor researchof this
sort.Before closing, it seems appropriateto offer a small list of suggestions for
scholars who have an interest in this style of research.In summaryform, my
advice is to first,get the text, second, find the use, andthird,mapthe meaning.
Get the Text
Meanings are complex things. Any attemptto model them or subject them to
formalanalysis invariablyinvolves a gross simplificationof the culturalmate-
rial. As a consequence, it behooves us to postpone this simplificationprocess
as long as is feasible. If we can wait until the thirdstep-the stage at which we
are consciously trying to reduce the data accordingto some structuralprinci-
ples-we have much greatercontrol over the way in which we simplify the
culturalmaterial.In practicalterms, this means that we should always try to
gather data as unobtrusivelyas possible. It is best to avoid imposing an arbi-
trarycoding scheme on the data if it can be avoided. Instead,we should try to
find a way to get the entire"text"into a computer(or to come as close as possi-
ble to this ideal).
In this regardit is probablyworthpointing out thatwe arejust now entering
what must surely be the golden age of textual analysis. What sets this moment
in history apartis the incredible proliferationof on-line and on-disk textual
materials. Previously, scholars who were interested in doing some form of
contentanalysis were compelled to spend huge amountsof time readyingtheir
texts for analysis.Now one can easily sit at one's desk andmore or less instan-
taneously summonup a fantasticarrayof culturaltexts in electronicform. This
includes softwareeditions of many contemporarynovels (currentlybeing mar-
keted for use in laptop computers)for about the same price as a hardbound
book, newspaperand magazine articles (publishedon the Interet), annualre-
portsof corporations(availableon CD-Rom), full transcriptsof free-formcon-
versations and other types of social exchanges (occurring in virtual "chat-
rooms" and user interestgroups on the Interet), to name but a few.
Find the Use
As I arguedearlier,the most critical componentof the meaning measurement
process is the identificationof the systems of relationsthatwill be used to gen-
erate the assessment of similaritiesand differences among culturalitems. Be-
cause there are invariablyany numberof ways in which these relationsmight
be assessed, it is essential thata theoreticallyinformedmetricbe appliedat this
stage of the analysis. The best rule of thumb in this situationis to locate and
evaluate the relevant domain of practicalactivity in which the identified sys-
tem of culturalmeanings is embedded. Differences in practice produce (and
are producedby) differences in meaning. Thereforethe goal of an empirical

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
 MEASURINGMEANING 367

analysis should be to assess how the various cultural elements are differen-
tially implicated in alternativeforms of practice.
Map the Meaning
Not all of the examples of culturalinterpretationthat I've describedin this re-
view use structuralmodels as a way of representingculturalmeanings. Clearly
situations exist wherein this kind of formal approachsimply isn't feasible or
desirable.Nonetheless, partof what I've triedto suggest here is thatthe project
of finding meaning structuresis not really differentfromthe projectof measur-
ing meaning structures,or rather,thatthere is a difference in the degree of for-
mality ratherthan in the type of endeavor. But, having said that much, I do
thinkthatwe can profitablymake greateruse of formalmethodsand especially
the sorts of two-mode modeling techniques that I discussed earlier.
Finally, the techniquesthatI have referredto are readily available and usu-
ally quite simple to use. Scaling and correspondenceanalysis programsare in-
cluded in most major statistical packages. Boolean algebras and sequence
analysis programsare easily obtainableand gearedto run on most microcom-
puters. Network analysis packages are likewise widely available and, nowa-
days at least, quite user-friendlyand intelligible.
Of course, as anyone who has employed these methods will readily testify,
there is nothing simple or determinateabout the interpretivework that one
must performafterhaving completeda structuralanalysis. In a sense, once one
has mappedthe meaning structureof a given institutionalpractice,one is in the
same position as a scholarwho has just emergedfrom an intensive field study,
chock full of ideas and images. Any visual representationof a meaning struc-
tureis still largely a Rorschachtest upon which one must seek to projectan in-
terpretation.But, to the extent thatone has carefullythoughtaboutthe linkage
between cultureandpracticeandhow it informsone's dataselection, measures
of differentiation,and structuralmodeling, so too one should have little diffi-
culty in interpretingthe meaning structuresthat have been identified.
ACKNOWLEDGMENTS
I would like to thankAnn Swidler for originally inviting me to write up a ver-
sion of this review for the ASA CultureSection Conference on Meaning and
Measurement,held at George Mason University in August 1995. Bill Bielby,
Paul DiMaggio, Mustafa Emirbayer,Noah Friedkin, Roger Friedland,John
Martin, Pep Rodriguez, Bruce Straits, John Sutton, and an anonymous re-
viewer provided valuable comments and advice.

Visit the Annual Reviews home page at

http://www.AnnualReviews.org.

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
368 MOHR

Literature Cited
Abbott A. 1983. Sequences of social events: leges: the social and conceptual structure
concepts and methods for the analysis of of sociology specialties. Am. Sociol. Rev.
order in social processes. Hist. Methods 57:266-73
16(4):129-47 Carley K. 1986. An approachfor relating so-
Abbott A. 1988. Transcendinggeneral linear cial structure to cognitive structure. J.
reality. Sociol. Theory6:169-86 Math. Soc. 12:137-89
Abbott A. 1990. Primeron sequence methods. Carley KM. 1994. Extractingculture through
Org. Sci. 1:373-92 textual analysis. Poetics 22:291-312
Abbott A. 1992. From causes to events: notes Carley KM. 1993. Coding choices for textual
on narrative positivism. Sociol. Methods analysis: a comparisonof content analysis
Res. 20(4):428-55 and map analysis. Sociol. Methodol. 23:
Abbott A, Forrest J. 1986. Optimal matching 75-126
methods for historical sequences. J. Inter- Carley KM, Kaufer DS. 1993. Semantic con-
discip. Hist. 26(3):471-94 nectivity: an approachfor analyzing sym-
Abell P. 1987. The Syntax of Social Life: The bols in semantic networks. Comm. Theory
Theory and Method of ComparativeNar- 3:183-213
ratives. Oxford: Clarendon Caws P. 1988. Structuralism: The Art of the
AldenderferMS, Blashfield RK. 1984. Cluster Intelligible. Atlantic Highlands, NJ: Hu-
Analysis. Beverly Hills, CA: Sage manities Press Int.
Anheier H, GerhardsJ. 1991a. Literarymyths Cerulo K. 1988. Analyzing culturalproducts:
and social structure. Soc. Forces 69: a new method of measurement. Soc. Sci.
811-30 Res. 17:317-52
Anheier H, GerhardsJ. 1991b. The acknowl- Cerulo K. 1995. Identity Designs. The Sights
edgment of literary influence: a structural and Sounds of a Nation. New Brunswick,
analysis of a Germanliterarynetwork. So- NJ: RutgersUniv. Press
ciol. Forum 6(1):137-56 Csikszentmihalyi M, Rochberg-Halton E.
Bauman Z. 1973. Culture as Praxis. London: 1981. The Meaning of Things: Domestic
Routledge and Kegan Paul Symbols and the Self Cambridge, UK:
Bearman P. 1993. Relations into Rhetorics. CambridgeUniv. Press
New Brunswick, NJ: Rutgers Univ. Press D'Andrade RG. 1995. The Development of
Berelson B. 1952. Content Analysis in Com- Cognitive Anthropology.Cambridge,UK:
munication Research. Glencoe, IL: Free CambridgeUniv. Press
Press D'Andrade RG, QuinnN, Nerlove S, Romney
Berger PL, Luckmann T. 1967. The Social AK. 1972. Categories of disease in Am-
Constructionof Reality. New York: Dou- erican-English and Mexican-Spanish. In
bleday MultidimensionalScaling, Vol. 2, ed. AK
Boorman SA, Levitt PR. 1983. Blockmodel- Romney, R Shepard, SB Nerlove. New
ing complex statutes:mapping techniques York: Seminar
based on combinatorial optimization for DarnovskyM, EpsteinB, Flacks R. 1995. Cul-
analyzing economic legislation and its tural Politics and Social Movements.
stress points over time. Econ. Letters 13: Philadelphia:Temple Univ. Press
1-9 Degenne A, Lebeaux M-O. 1996. Boolean
Borgatti SP, Everett MG. 1997. Network analysis of questionnaire data. Soc. Net-
analysis of 2-mode data. Soc. Networks works 18:231-45
19:243-69 Denzin NK. 1991. Empiricist cultural studies
BourdieuP. 1977. Outlineof a TheoryofPrac- in America: a deconstructive reading.
tice. Cambridge, UK: Cambridge Univ. Curr.Persp. Social Theory 11:17-39
Press de Saussure F. [1916] 1959. Course in Gen-
Bourdieu P. 1984. Distinction: A Social Cri- eral Linguistics.New York:McGraw-Hill
tique of the Judgement of Taste. Cam- DiMaggio PJ. 1992. Cultural capital and
bridge, MA: HarvardUniv. Press school success: the impact of status-
Burt RS. 1978. Cohesion versus structural cultureparticipationon the grades of U.S.
equivalence as a basis for network sub- high school students.Am. Sociol. Rev. 47:
groups. Sociol. MethodsRes. 7:189-212 189-201
Burton M, Kirk L. 1979. Sex differences in DiMaggio PJ, Mohr J. 1985. Culturalcapital,
Maasai cognition of personality and social educational attainmentand marital selec-
identity. Am. Anthro. 81:841-73 tion. Amer.J. Sociol. 90(6):1231-61
CappellCL, GuterbockTM. 1992. Visible col- DiMaggio PJ, Mullen A. 1993. Organizing

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
 MEASURING MEANING 369

communities: models of collective action work for the sociology of culture. Sociol.
in national music week, 1924. Presentedat Methodol. 17:1-35
Am. Sociol. Assoc. Meet., 88th, Miami Griswold W. 1987b. The fabricationof mean-
Doreian P. 1987. Equivalence in a social net- ing: literary interpretationin the United
work. J. Math. Soc. 13:243-82 States, GreatBritain, and the West Indies.
Dowd TJ. 1992. The musical structureand so- Am. J. Sociol. 92(5):1077-1117
cial context of numberone songs, 1955 to Griswold W. 1992. The writing on the mud
1988: an exploratoryanalysis. In Vocabu- wall: Nigerian novels and the imaginary
laries of Public Life: Empirical Essays in village. Am. Sociol. Rev. 57:709-24
Symbolic Structure,ed. R Wuthnow.New Griswold W. 1993. Recent moves in the soci-
York. Routledge ology of literature.Annu. Rev. Sociol. 19:
DuquenneV. 1995. Models of possessions and 455-67
lattice analysis. Soc. Sci. Info. 34(2): Hage P, HararyF. 1983. StructuralModels in
253-67 Anthropology. Cambridge, UK: Cam-
Duquenne V, Mohr JW, Le Pape A. 1998. bridge Univ. Press
Comparisonof dual orderingsin time. Soc. Hammond NDC. 1972. The planning of a
Sci. Info. 37:(2):227-53 Maya ceremonial center. Scientific Am.
EmirbayerM. 1997. Manifesto for a relational 226:83-91
sociology. Amer. J. Sociol. 103(2): Hawkes T. 1977. Structuralismand Semiotics.
281-317 Berkeley, CA: Univ. Calif. Press
Ennis JG. 1992. The social organizationof so- Hays TE. 1976. An empirical method for the
ciological knowledge: modeling the inter- identification of covert categories in eth-
section of specialties. Am. Sociol. Rev. 57: nobiology. Am. Ethnol. 3(3)489-507
259-65 Heise D. 1988. Computeranalysis of cultural
Eyerman R, Jamison A. 1991. Social Move- structures. Soc. Sci. Computer Rev. 6:
ments: A Cognitive Approach. University 183-96
Park,PA: Penn. State Univ. Press Heise D. 1989. Modeling event structures.J.
ForanJ. 1997. The futureof revolutions at the Math. Soc. 14:139-69
fin-de-siecle. Third World Q. 18(5): KirkL, BurtonM. 1977. Meaningand context:
791-820. a study of contextual shifts in meaning of
Franzosi R. 1989. From words to numbers:a Maasai personality descriptors. Am. Eth-
generalized and linguistics-based coding nol. 4:734-61
procedure for collecting textual data. So- KruskalJB, Wish M. 1978. Multidimensional
ciol. Methodol. 19:263-98 Scaling. Beverly Hills, CA: Sage
Franzosi R. 1990. Computer-assistedcoding Kurzweil E. 1980. The Age of Structuralism:
of textual data. Sociol. Meth. Res. 19(2): Levi-Strauss to Foucault. New York: Co-
225-57 lumbia Univ. Press
Franzosi R. 1997. Mobilization and counter- Laumann EO, Knoke D. 1987. Organization
mobilization processes: from the "Red State: Social Change in National Policy
Years" (1919-20) to the "Black Years" Domains. Madison: Univ. Wisc. Press
(1921-22) in Italy. Theory Soc. 26(2-3): Levi-Strauss C. 1963. Structural Anthropol-
275-304 ogy. New York: Basic Books
FranzosiR, MohrJW. 1997. New directionsin Merton RK, ed. 1957. The bearing of empiri-
formalizationand historical analysis. The- cal researchon sociological theory. In So-
ory Soc. 26(2-3):133-60 cial Theoryand Social Structure.Glencoe,
FriedlandR, Alford RR. 1991. Bringing soci- IL: Free Press
ety back in: symbols, practices and institu- Meyer J, Rowan B. 1977. Institutionalizedor-
tional contradictions. In The New Institu- ganizations: formal structureas myth and
tionalism in Organizational Analysis, ed. ceremony. Amer.J. Sociol. 83:340-63
WW Powell, P DiMaggio, pp. 232-63. Mische A. 1998. Projects, identities. and so-
Chicago: Univ. Chicago Press cial networks: Brazilian youth mobiliza-
Geertz C. 1973. The Interpretation of Cul- tion and the making of civic culture. Pre-
tures. New York: Basic Books sented at Am. Sociol. Assoc. Meet., 93,
Giddens A. 1984. The Constitutionof Society: San Francisco
Outline of a Theoryof Structuration.Ber- MohrJW. 1994. Soldiers, mothers,trampsand
keley: Univ. Calif. Press others: discourse roles in the 1907 New
Griffin LJ. 1993. Narrative, event-structure York City Charity Directory. Poetics 22:
analysis, and causal interpretationin his- 327-57
torical sociology. Am. J. Sociol. 98: Mohr JW. 1998. The classificatory logics of
1094-1133 state welfare systems: towards a formal
Griswold W. 1987a. A methodological frame- analysis. In Public Rights, Public Rules:

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions
370 MOHR

Constituting Citizens in the WorldPolity DeSoete. Teaneck, NJ: World Scientific

and National Policy, ed. C McNeely. NY: Sahlins M. 1976. Culture and Practical Rea-
Garland.In press son. Chicago: Univ. Chicago Press
Mohr JW, DiMaggio P. 1995. The intergen- Schweizer T. 1993. The dual orderingof peo-
erational transmission of cultural capital. ple and possessions. Curr. Anthro. 34:
Res. Soc. Strat. Mob. 14:169-200 469-83
Mohr JW, Duquenne V. 1997. The duality of Sewell WH Jr. 1992. A theory of structure:du-
cultureand practice:poverty relief in New ality, agency and transformation.Amer. J.
York City, 1888-1917. TheorySoc. 26(2- Sociol. 98 (Jul):1-29
3):305-56 Shweder RA, Bourne EJ. 1991. Does the con-
MohrJW, Guerra-PearsonF. 1998. The differ- cept of the person vary cross-culturally?In
entiation of institutional space: organiza- Thinking Through Cultures. Expeditions
tional forms in the New York social wel- in Cultural Psychology, ed. RA Shweder,
fare sector, 1888-1917. In Bending the pp. 113-55. Cambridge, MA: Harvard
Bars of the Iron Cage, ed. W Powell, D Univ. Press
Jones. Chicago: Univ. Chicago Press. In Skocpol T. 1992. Protecting Soldiers and
press Mothers. The Political Origins of Social
OrtnerS. 1994. Theory in anthropologysince Policy in the United States. Cambridge,
the sixties. In Culture/Power/History:A MA: HarvardUniv. Press.
Reader in ContemporarySocial Theory, SniderJG, Osgood CE. 1969. SemanticDiffer-
ed. NB Dirks, G Eley, SB Ortner, pp. ential Technique:A Sourcebook. Chicago:
372-411. Princeton, NJ: Princeton Univ. Aldine
Press Snow DA, Rochford EB Jr, Worden SK, Ben-
PetersonRA. 1976. The productionof culture. ford RD. 1986. Frame alignment pro-
Am. Behav. Sci. 19:669-83 cesses, micromobilization,and movement
Peterson RA, Berger DG. 1975. Cycles in participation.Am. Sociol. Rev. 51:464-81
symbol production: the case of popular Sudman S, Bradbur NM, Schwarz N. 1996.
music. Am. Sociol. Rev. 40:158-73 ThinkingAbout Answers. TheApplication
Pettit P. 1977. The Concept of Structuralism. of Cognitive Processes to Survey Method-
A Critical Analysis. Berkeley: Univ. Calif. ology. San Francisco:Jossey-Bass
Press Swidler A. 1986. Culture in action: symbols
Powell WW, DiMaggio PJ. 1991. TheNew In- and strategies.Am.Sociol. Rev. 51:273-86
stitutionalism in OrganizationalAnalysis. Tilly C. 1997. Parliamentarizationof popular
Chicago: Univ. Chicago Press contention in Great Britain, 1758-1834.
Ragin C. 1987. The Comparative Method. TheorySoc. 26(2-3):245-73
Moving Beyond Qualitativeand Quantita- Ventresca M, Yin X. 1997. Representing
tive Strategies. Berkeley: Univ. Calif. economy and society: comparative trends
Press in what states count, 1800-1970. Pre-
Ragin C. 1995. Using qualitativecomparative sented at Am. Sociol. Assoc. Meet., 92nd,
analysis to study configurations. In Toronto
Computer-AidedQualitative Data Analy- WassermanS, Faust K. 1994. Social Network
sis, ed. U Kelle, pp. 177-89. Thousand Analysis. MethodsandApplications. Cam-
Oaks, CA: Sage bridge, UK: CambridgeUniv. Press
Roberts CW. 1989. Other than counting Weller SC, Romney AK. 1988. Systematic
words: a linguistic approach to content Data Collection. NewburyPark,CA: Sage
analysis. Soc. Forces 68(1):147-77 Weller SC, Romney AK. 1990. Metric Scal-
Rosenberg S. 1989. A study of personality in ing. Correspondence Analysis. Newbury
literaryautobiography:an analysis of Tho- Park,CA: Sage
mas Wolfe's Look Homeward Angel. J. White HC, Boorman SA, Breiger RL. 1976.
Pers. Soc. Psy. 56:416-30 Social structurefrom multiple networks. I:
Rosenberg S, Mechelen I, De Boeck P. 1995. Blockmodels of roles and positions. Amer.
A hierarchical classes model: theory and J. Sociol. 81:730-80
method with applications in psychology Wuthnow R. 1987. Meaning and Moral Or-
and psychopathology. In Clustering and der. Explorations in Cultural Analysis.
Classification, ed. P Arabie, L Hubert, G Berkeley: Univ. Calif. Press

This content downloaded on Thu, 24 Jan 2013 21:05:24 PM

All use subject to JSTOR Terms and Conditions