contemporary
vw r,”’
icotemporary
SOCIETIES
Methods of Discovery Heuristics for the Social Sciences \Ietbeds nf Discevenj is organized around strategies or deef cuing arguments in order to find the best w ys to study social phenomena. Abbott helps social su nec’ students discover what questions to ask. This exciting book is not about the mechanics of doing social science research but about habits of thinking that ‘noble students to usc those mechanics in new ways, by ot un up with new ideas and combining them more effectively with old ones. \bhott organizes Metbeds of Discovery around general methodological moves and uses examples from throughout the social sciences to show how these moves can pe new line of thinking In each chapter, he cc vets several moves and their reverses (if these exist), discussing particular examples of each move as well as ts logical and theoretical structure. Often he goes on to propose applications of the move in a wide variety of empirical settings. The basic aim of Methods of Discovenj tc ffer readers a new wm y of thinking ibout directions for I cii resc arch and new ways to imagIne information relevant to their research problems. ANDREW ABBOTT is a professor in the Department of Ss’iologv at the Univenity ol (hicago He is the authot iost re ntly, of Chaos of Dzscip I i s and 1)rne Matters: On
The,
od M th ,d, both from the University of Chicago Press. The recipient of a Guggenheim Fellowship, Professor Abbott has written and lectured widely on
oc mpations and professions social science methods,
‘-oci il theory and the history of the ‘.bicago School of Sociology.
This series marks the coming of age of a generation and a discipline. It has been halt a century since the world’s leading sociologists engaged in a collective effort to make their thinking and research widely available and accessible. What has changed in the meantime? Just about everythingl Theoretical
fJIetl,icEIs c
way to relevance, and a provincial
transformations has been the cultural turn the recognition that meaning dynamics—codes, narratives, metaphors, values, and beliefs—remain central features of contemporary societies. In this series, the world’s leading sociologists show how these developments have transformed their areas of specialty. They do so by writing in a genre that has almost disappeared from the social sciences today—the essay. Well-written, clear-minded, and elegant, these brief compositions are major creative endeavors in their own right, even as they bring the ideas of the world’s most advanced thinkers into the world ofthelayreader. —Jeffrey Alexander, Series Editor
a
Discovery
1-leuristics for the Social Sciences
Pndrevv Abbott o .
H
5.
,.
03 (j
03
:ss o-t9—tuc-
—
,
,
90000
1 BI I BI1MIB hIB iuui hl I I 1 780393 978148
www.wwnorton
corn
;WW.N0RTON 51W YOlK -
•
LONDON
fl
JEFFPEV
C
ALEXANDEfl
4
iuIIh1hIiO1) S
I U
dL
-
*
U
i
10
0.
I I
iR
tit
iui Uw
i&4i
.)HIII1
\iiU IU
( .
t
‘±‘
-
‘---
____i.__..
--
-‘
-
0.
---.
t
--
,,fr
II C. L
-\
k \
k
£
tail
I
I
a 1.4.
1
I
lbS
lIs.,iuvh
114.
L
caaaa I
C
I
,1
Iii
I.
I
Itd.IathAi 1.44 lI’ L t,. I
xl
.t
I
Lb\
‘. ..
a
I.
ft
1
IdjflAr..
n4.a
iai..
Jail’
I.
1
U
al
a 1.1
Ll.aa:.c
%LIIaIlasia.V lilt)
(l
t
CONTENTS
AcKNowLEDGMENTs
To THE READER
xi
Explanation I. Explanation II, Methods III. Explanatory Programs
CHAPTER ONE
CHAPTEI Two
Basic Debates and Methodological Practices
I, Basic Debates H. Methods and Debates
tf 3
53
III, Cycles of Critique
60
IV. From Critique to Heuristic
75
CHAPTER THREE
Introduction to Heuristics
I. The idea of Heuristic II. The Routine Heuristics of Normal Science III. Topics and Commonplaces CHAPTER FouR
General Heuristics: Search and Argument
110
I. Search Heuristics
113
II. Argument Heuristics
120
H PfER
I ivii
General
Heuristics: Description and
Narration scip Ii
“I
1
SticS
146
fracal Heuristics
162
and
Positivum
Interpretivism
11
al s and N r Oitio I III. B havi rum (ultuca1isrn l\ Individualism and Emergenrism V Realism and Cc nstructaonism VI
fitextualiso
II
o
Cc
ii
nd Noncontcxtc alism
out
VIII
Conflict and Consensus IX, Transcendent and Situated Knowledge
C
i
a
i
It
tS
a
1 Ri
uP Puzsles
168 171 179 183 187 192 198 201 206 211
of It eas
213
11. Other People III, Literature
221
IV,
231
V \d
(ji
Si
138
Hcu s ro
in
(disi’ cia Six
I—Ic
I
‘iste U
nudity
I ussies
SsSIIY
N Rs IRF\(E
ACKNOWLEDGMENTh
137
226 234 242 249
255 259 265
has its roots in my book Chaos of Disc /1 ir While writing Chaos. I came to realize rha iii, LheOrl 01 IiOv social science developed also implied a VICW or flOzs SOCiãi can ence should be done. When Jeff Alexander ofibred TIC tt e portunity to put that vision into writing, 1 seized i’ Bu’ dx
Titis
LiTTLE TEXT
heady notion of fractal heuristics that came from Claos p Jiliiies quickly became part of a broader attempt to write sorro, thing my students needed desperately: an introduction to roe workings of imagination in social science, Undergraduate at graduate education in social science is so focused on ‘maki p a count (Lieberson 1985) that we gradually k11 the excitement our students bring to their work. I have, 1 hope, written 0 000i that will show how to protect that excitement, ifl no small Dirt by showing how practicing social scientists have themselves protected it throughout the last century. This book was written in the intervals of Leaching rid CeSIS almost completely on my experience as a teacher supervis ing papers, theses, and dissertations. First and foremost, then, thank my classroom and dissertation students at thc LJnvc sill of Chicago. They have taught me much: both drectiy, b, WhaL they’ve told me, and indirectly, by the struggles for which rep help was so often unavailing. More particularly, I thank my So ciological Inquiry and Thesis Proposal classes for having helped
AcKNo’
I 5DoM1\ [
ch’i iy thes ide is, I hope that this book will capture some hint ot the excitement of our conversations. I also thank uni— v i it cii en as at Sar )ugo, Harvard, Oxford, and Oslo, as well as a conference group at the Social Science History Associ ati in, II of whom heard and commented on Chapter One. b ok vas designed, and large parts of it drafted, during a stay at Nuffield College, Oxford, during h ik the Warden and Fellows of Nuffleld Hi cry ‘err 21 1 ut their support. One chapter (Chapter Six—typically, the first to he imagined was the ‘ast to be completed) was written dur i a or vis t, it 200 to th Department of Sociology at Fiarsard, hose kind hospitality I acknowledge. Erin York was assista it I o chased down various sources and ri r ax ,
time taken from the writing of her own splen did masters paper. Above all, I acknowledge the example of ( irge Pc lya, whose I ook Hou to 5oiie it first taught me the ri porrance of heuristic. Such success as this book may enjoy in
yes lrc i hi happy inspiration Its failures—no doubt con siderable—are mine. Chicago, November 2003
d
HE READ1ik f1
ti
M
resources
To
i’r IS A SURPRISING FACT that marty good stunents, wher they sic
down to rite course papers or bachelor’s theses or een doc toral dissertations, fear that they have nothing to say. They un derstand methods, They know about sources and data B it their own contribution seems to them obvious or rrivtal. It is little wonder students feel this way. Faculty stud lit’r arures and methods into their heads until there ia little to ni for anything else. Of course students think everything has al ready been said; they just read it last week. And of course they make theoretical arguments by blending ideas from here arid there like squirrels furnishing a nest with trash; they haven t learned to create theoretical arguments on their own. Small wonder they feel at a loss. What then does it take to have something to say I take tso things. The hrst is a puzzle, something about the social world that is odd, unusual, unexpected, or novel. The sccnd a clever idea that responds to or interprets or solves that puzzle Everything else—the methods, the literature, the description of data—is really just window dressing. The heart of good work is a puzzle and an idea. Although I shall talk about puzzles in the last chapter, fin is chiefly a hook about finding ideas. And while I talk ab ut social science methods and about research unrig those method.
tEADILi
i
ally about methods instead, it is i 1 i,ea/icit that animates methods. t vity annot ht taught As John Dewey put it, “(N]o y ii, ito idea can possibly he cot scyed as an idea from one )etSOfl 10 another (I 96 159), We can teach others the back ,r unds, the conditions, and the origins of an idea. But if we I
iii
cry is not
.
1t;
theinsus, e’t: I ittle more bluntly in Pride and Prqadrc: “We all love to in struct, though we can teaeli only what is not worth knowing.” Nonetheless, by teaching some basic tricks for produc ing ideas. I hope to give the reader tools for social scientific discosery, And I will show how these tools for invention and discovery relate to the methods that are—with good reason— thought to he our chief ways of prodLicing competent work. I have assumed undergraduate students as my basic audi ertec. But much of what I have to say can be useful to graduate students or even the occasional colleague.
Writing
this book
iia often reminded me of important things I had myself for— tter (and will forget again, no doubt) We are all on the same v, all tr ing to say niteresting things, all falling into bad fit ill strugghng to imagine the social world anew.
ci
I a e tried
sot added furhr citation if
s i
\
I
keep scholarly machinery to a minimum
ii general, whus the text mefers to an author and a
t
i
to
C. ci
type r
it
is not necessary to
c uniqit iv. AJ articles and books mentioned i
he P f rences at the end of the book. I
sc minimize I
idnotes as much as possible.
M E T HO D S
OF
Disco v E R
XPLANAT1 ON
I. 11.
ExPLANATIoN METHODS
A.
ETHN0GRAPHY
B.
HISTORICAL NARRATION
C.
STANDARD CAUSAL ANALYSIS
D, SMALL-N E. ill.
COMPARISON
FORMALIZATION
EXPLANATORY PROGRAMS
between rigor and imagination. What one proposes, the other evaluates. Every evaluation leads to new proposals, and so it goes, on and on. Many people think of social science less as a conversation than as a monologue. For them, it is a long speech that ends with a formal question, to which reality meekly answers yes or no like the plastic heroine of a Victorian novel, Yet no good re searcher believes in such monologues. Researchers know all about the continual interchange between intuition and method, just as they know about the endless teasing of reality as it evades them. Social science in practice is lass old-style romance
SCIENCE IS A CONVERSATION
than modern soap opera. The monologue version of social science is of course easier to describe. There are many excellent books about its machinery:
Sir
no
)s ut
I) s
0 lois
a questi m, I ow to design a study, how to ac quirt and analyze data, IIos to draw inferences. Indeed, many re v’g mi cd in on I p rti ula- ways of doing these 01 ‘s 11 them: ethnography, the ax inus methods, as we La se on lary data an’ lysis, historical mud comparative [Iv } s ow
in
to pn
“0
ihiids, and o on. All that is fine and good, But suci heok lorget the other vow the imaginative voice w hit surprise, and nos elt This discovery side of social um is more systeniatic rhin we think Social scientists use ,
mhits of imagination, mental moves they employ to hasten 0 a imbirs in chess these mental moves are I* •liw u ci on ars fur i ie punun esr loping. an rcalizing of possi h’ ii tim s Some sn general gambits imphcit in the nature of at ui to an I di set pi ion. a h Ic orhe s arise in conceptual issues that pei alr the disciplines All of these gambits work within , kind c n thod, I Ivy r iake up the heuristic of social sci cow, the mcars by wuicli so ial science discovers new ideas. as I said social reality often re met I en istic hecau Sist inc charms of methodology. As social scientists, we aim to n o scil ri ii teresti ig-- perhaps even true-—about social lite, Yet social realit often makes a stingy reply to even the
beCaSiSc examples I use in the book come trom mcielogy. But all of the s the social sciences are all mixed up together. riot SUDje amples will be socioiogkal. The social ScienCeS .1sare dology matters, theories, and a surprising amount of metho law K They are not organized into a clearly defined system their orientations from various historical accidents. Loosely the speaking, economics is organized by a theoretical cocicepi idea of choice under constraint), political science by an aspect h of social organization (power), anthropology by a mechu ast (erhnography), history by an aspect of temporality (the r Citf and sociology by a list of subject matters (inequality, the le ctiterw f u 5 th family, and so on). Thus. there is n sin an distinctions among disciplines. As a result, when one cu mel other discipline becomes too much of a bore, the othecs use else fun of it and steal its best ideas to put them to bcrter rsn, where. All of this flux means that a heuristics book can widely, as this one will.
,
cs SI U ‘nethi dok gical i ionologues returning tiny correla ions c en thu ugh challenged by the best of questionnaires, a iii sg it p1w sincled t ens us even though watched by not tlis of cattiest ethnography, returning boring stories even I y yeai of painstaking archival research. ho gi qe st n )Oe iii cal ty wa its a subtler wooing it wants rigor anti o this is a vsok ahou heuristic, a hook of aids to the social 1 ca e in a x iologist, many of the w ot i a ii
th aims, means and sumptions of social science research. 1 begin with explanatior because explanation s the purpose of social science. 1 turn in troduce some types of methods—some of the various ways i which social scientists have tried to be rigorous. I treat ttiesS
ThE FIRST TWO CHAPTFRS introduce
-
methods as concrete realizations of “explanatory piograms programs that carry out the different concepts of explanation introduced eamlier in the chapter Chapter Two turns to a more customary approach. charuu tetize methods in terms of set of conceptual issues—n’ He them, in fact. I flrst introduce these conceptual issues, then give the customary account of methods (1 skipped it in Cf a
\I
ho sri
Di
ii
cr Qi e) r hid says hat iethods arc best defined in terms of rhc nine issues. rhcn I leave the beaten path. I discuss the I t ]U( s that zidi method poses to the others and show that
tnt iues lead us it to an endless cycling through the methods Ihutli in theory md in practice), Moreover, the conee ual issues themselves urn out not to be fixed things; they lIa c an unsrab e, fractal character Not only do they differenti 1 n in m thod from an t ie thcy also differentiate internal stra icis mirhin each method-—and internal strands within the it stand I ids 0 Chapter One a d Two are the heavy lifting before the fun Ut lv u W7biI ti c r i, n Ufj1 of the hook is to stimulate imagination. it needs to present a clear sense of rigor as well. Orb rw se ye W( ii t be bie to tell the difference between m magi nation and foolishness. Recognizing that difference me ns getting se ure Sen e c I what explanation is, of why we seel explanations, and of what different kinds of explanations nd programs of explinatmon exist in social science. It also ic is I as i g solid grasy of more traditional ways of think ing about rigor, which are presented in Chapter Two, with its ra I I ass mcthodological debates in social science nd its end ess hms. a;1tiatel’. I will turn these isms from ri’ h c o yhal dehz is ito Inc heuristics.) Has ing set forth the basics of rigor in Chapters One and v I he i rut t i rapi anon Chapter Three discusses the general concept of licuris tic and sets forth the two simplest F or Sri sn tee es the adLliri e heuristic of normal science and the use of rommonpiace lists to generate new ideas. Chapter Four considers in letail ‘he general heuristic gambits that caicli for niportable no elti elsevhere and produce it by
transforming out cxiscin argunic irts Cliapre Ti c i oh heuristics of rime and space, tOe ileururic: chat Cliatie a. s describing or ensisi000g social reality s s to produce deas. Chapter Six examines the gambmn tha anse ou F 1 basic debates and rucrlmudological cjn cri of Cue rer To making a positivlst move widiri an iiterpre avi tradin ii, example. Finally, Chapter Sesen discusses uhe pc.bh: cc ci uating the ideas produced by lieuriscic It ask. how cc 90W good idea when we iCe one. and I have drawn examples from as far back as rh recently as 1999. Old Work is not juecessarilu bad york. Neo ten himseif is a good cxampte. Newton became r ie bleat n. rise nsedecas qIzesnu name in modern science by ve of the nature and origins of motion. He solved die prob cii n 1 simply assuming that ( motion exists nd motion b rends to peisist. By means of these assumptions eahy a ucar c of declaring victory, as sic would now put it) he v as zihe a uiar ums F 8 develop and systematize a general account of die re .
motion in the physical world. That is, by giving up on t e r question, he almost completely answered inc as;an questexu u following his example, we iearn that switching quest mm
powerful heuristic i950510. 1 sc ci c. )r e The very same move has occurred in sO a the great difficulties in tire work of Talccrr barsrms. the .aun nant American sociologist ot the mud—twejineili cci wiy
-
in explaining social change. Parsons held that soc:ai be1sasu: was governed by norms, which Wute themselves gcveei d ch were themselves governed f ez more betierai 1 values, wh values. In such c system, change tonid be conceived only .s local bieakdowmz, a problem event that had somehow wa’ed
Mt 111005
iii
F [)is (iS FRS
I
ExPLaNarLrn’
s riters handled the same prob— i i change by simply assuming that social nor it sital r 11; rat he’, it as the normal state or aftiitrs. With this assurnpcloii. the sarious historical sociolo pists who challenged Parsons were abh to develop much more simper
sing
x
us.
t
I
-
ourit
‘nyc
d
Sia
i c
1
Th)Vemflc
I mod
its,
of revolutions, and,
ity ii general. This was exactly
ye: hut neal s ioloLiists gave up on cx— plain iiip cflaflpe and si mph assumed it was happening all the time. Then all they had to do was figure out what is regular it d v iv it happen They diould have gone on to cx Icrnrsc L th’ pea much forgot about idll
I
Thus, urd work provides useful examples of heuristics just .mew v. orl does. This means that as I introduce the reader lie I tool kit )f heuristics in social science, I can simul
as
t
i
itt rn ci
has
duced
hr great lit ritage that that tool
it e
er’s be in, then
at
the beginning—with
xplanat ion. I
vei is’s
is an account that suffices it frees us to go on to the next pro0 1cm by bringing out current problem into a corn onse e world where it becomes immediately comprehensible. So soci biologists say they hae explained altruistic bcnax mr svcmi they show it to be merely an accidental result of selfish behav ior. They go no further because they think selfish behavior i self-evident;
it
needs no explanation
Third, we often sai we have an explanation of sornethi p when we have made a certain kind of argument aboru it: an ar gument that is simple, exclusive, perhaps elegant or even coun terintuitive. Thus, we may think Freudian psychology is hette than folk psychology because it is better worked aut, mc e complex, and more surprising. In this third sense, an account is an explanation because it takes a certain pleasing form, it somehow marries simplicity and complexity.
becausc
The first of these views—the pragmatic view that an explana non is an account that enables us to intervene—is the rsos familiar. Consider the explanation of germ-based disease. We think discovering a germ is explaining a disease because by dis covering the germ, we have discovered something that enables us to stop the disease. Note that this pragmatic approach to cx
in’s
to cxp a ocia life ‘II etc are three things Li iii hr i social scieiirist ay that a particular argument is an explanation. First, we sai something is an explanation when it
planation works best for phenomena that have somewhete a
ailows us to ulters ene in \vhaLe\ er it is we are explaining. For r ipli have cx lain ‘d the economy when we can manage
side action. It is this narrow neck—the necessity of a particular organism—that makes the germ-based diseases easier to fight
I
ac
V€
a n
a
Se.ond. looking
‘timla
ix d
p
e
t
hen we know how to eradicate
se av an account explains something when we stop
to itirther accounts of that something. An explanation
narrow neck of necessary causality: something absolutely neces sary to the phenomenon yet clearly defined and subject to out
than diseases ‘caused” by the interaction of millions of smal random events—cancer, heart disease, and arthritis, The move to the microceliular level in studying these diseases aims pre cisely to find a new realm where there
ma)
lie
a narrow neck—
LxPL i to end up in the natural sciences and ‘ii he na ic Will hi h tr r if w c ) irol for native ability? ihr co lugi Hajor: tor parental cue 1 )Urageflleflr for choice of i Ii h U us a 1 Jr orndLr die long—standing Ins or ai tin liii ri at the revolutio iary political parties of the Ii RL ii ii ury u al y 1 ad B r origins among artisans ir icr iii iii inn ii’i unsk lied tOis 1 laborers or agricultural at w iliac
,
.
,
‘.
laborers. ‘(/as this also true in areas where artisans were fes Was it true in Catholic as well as Protestant regions? east of risc Elbe? and so on. Finally, addition sometimes takes the form of aduing a new model or methodological wrinkle or theoretical twist. For aO science, tills might be taking a more careful look at the exact language that was used in interviews, to sec whether the order in which scientists said certain things re vealed new aspects of their assumptions. For a rational-choice
ethnographer
of
modeler, this might be trying four or five different forms o “game,” rather than just one or two, to understand a particular bargaining structure. For an SCA analyst, it might be putting exponential terms into the equation, to see whether certain in dependent variables had not only linear but also nonlinear effects. All of these—from simply adding data to adding a new di mension for analysis to adding a new methodological or theo retical wrinkle—ate basically minor, incremental additions. They are the tuck-pointing and reshingling and additionbuilding of normal science. They are the conservative strategy for social scientists, and it should Come as no surprise that graduate students—the most conservative of all social scientists (because they have the most at risk)—shouid be assiduous prac titioners of the additive heuristic. Libraries are filled with urn published doctoral dissertations that carry out such additive projects. Scholarly journals receive dozens of stabmusions nased on
them.
Such studies are profoundly useful. One brilliant contri bution does not fully establish a new argument. Adding new cases or variables or rules is always a useful first step in the full
ii
Ic
1
f de
1 1.
VU
a no s
lirRoDucrioN 10 Heuiisiics
is alit and fitting that most of os begin oar reers with the additive heuristic, and it is not at al so 5115 nn ‘hat ii ianv u us icy leas c lit I U m c a n I hi tist is to impi ave on such nor C
U
it
-
inil 1
Runt niber Phli as definition: “The aim of hen ) ic t r cr1 Is id [U es d discovery and Ins cut ion s ss hat ss seek not just addition. How
iL
Lovent rut i
in
1 d
al
.
in
it n,
ii
SNL) fosl\to.\PaA U’
‘I erc s,
ri
is o
u t s fo
icvenrn n
tlii ig o a t adit on about inVentiOn, It s not Ii coal iii the srienccs but tatlier in the field of rhetoric. Ci I SC t or a cga ye word, tt label tricks of langtwgc or arutin cur P think or rhetoric as false or at least d ci vc Fl i ii cii u w itt. 0 1 di toric—people like isuciaLes ristot1e, Cicero. and Quiatilian—.--ss crc mainly con it. )Om
e”ned with tr litli ig people as knoss Iedeab1e speakers in pub ii Sc ii, 01 s i t lat C C K rts ii legal sen wigs. And so for .
a aood thing. both positise and creative. 11 l t I C Lii 551 i d t 55 of arguments was ceo tr I LU tile L lassical writers visi ni of rhetoric. (Ideally one could .1 i I u ‘s Ic ‘t taLkin ht i a tice speeches were \vt ttt ii .tlieid f t rue and rehearscd urtensivelu) Rhetoric text ‘u iks nut im rl began with section entitled un ui/lu, (Jnun— i Itcn. rliciork was
is
sailie toot i
i
in u-na
r’
o d
lie areek fo this was [mona, from the as See Clarke 1 b 3:., This section covered sy ‘a t ii a t p i r ci t arguments. The most ss lL to do so werc iniled U pics and included extremely i Ii im i es If tenu i d genus and
ie
ati i ss
/‘ ‘cud,.
‘
for arguments were cc lied commonplaces which were Ibmiliar notions, like th idea chat criminals did or did riot keep committi1g ch san’i: inilOr common beliefs that often came iu pairs one 01 e.Lcl side of an
species.” More concrete
sources
—.
argument.
Apprentice speakers learned huge lists of topics arid co r nionplaces and their subdivisions. lVfaStell in such lists was considered the foundation for effective argument. It is hardly surprising that in time there were complaints that oratory had become boring. What had been meant as a guide to in’senciig new ideas had become a machine producing endlessly familiat ones. e social scientisin have such rhetorical forms, 7 (
op’cs.
and commonplaces ourselves. The most famous—as farnuiar in high school students in America as the six parts of a classical speech were to similar students two millennia ago—is “com pare and contrast.” (It was on Aristotle’s and Cicero’s liStS, tOO.) “Pros and cons” is another enduring rhetorical fotn’i, also on most ancient lists, as it is in the repertoire of roost scholars cci day. Each of these rhetorical forms can be invoked in the heat of argument to provide a prefabricated layout ior a discussion. And each can sometimes become very mechanical. But the use of rhetorical forms and topics as means to in Vention suggests that there might he similar forms and topics for social science invention. These would be lists of topics that could be applied to any argument at any point to generate ness tIHflgs to Say. The idea is simple. You have a wied-and- rue list of
abstract categories or concepts, and when yoa find yourscl
running
out of ideas about
some aspect of social hfc,
y
ti go in
nis
M
the
list and
on must
hrt
Ds
hat it get some
IN1RODUCTION TO Fiio, ROtics
suggests to on. 11w problem is chat good lists of Categories or COflCCt5 to
1 IC to
050
u
times (that pl t ik t h too e ‘iousl and 0 c lists got very hot iug. 55 e are out going to be part iculailv worried about whether )ur lists art the right lists or the rue lists. it doesn’t matter 4 1 y I d r log i ly t p’stcmologically or wha ci. istie leas vo w of g aduatc school trying Ut) dci ide on th ‘right” abstract concepts and came to no con i
i
ists
n ancient
is
clusion
il ‘hat I i/Soil/il have thought about was which ist mi 1 re n svl. s tOil wc rc iigf t m on r su n pical lists —two classical and t vo modern—.that I myself ha c often found useful: Aris tOCles four causes, Kant’s list of categories, Kenneth Burkes Inc k ys of am tu n, a e Chants Mortis’s three modes of n s pa cs hr i ison “hose should be your at
.
lisr, Indeed, I ye mcd other lists from time to time. But Chest happeni to he the ones that have most often proved useful no ow They are al lists that hays recuircd in the works of ia y w it s sdc any different labels But let me reiterate mph
that
is is
twct
arifi h causs cy are right (although it would b 0 hard to come up with a concept of cause that didn’t nt Aristotle’s aflalyis one uay or an itherj Rather, it’s because icy s us 1 ‘I h y wIg u inakc o uck switches in our inteloct itta vi ems You Fm already htcn introduced to one ol tlieso lists. h the way: I used Morris’s modes of lan t
guage to Orgarure the tirsn chapter of this book.
‘
F&,ii Ca/1w 1 start with Aristotle’s font causes.
A.
Ar/ito//eS
its
a simple list:
mateual cause formal, or structurai. cause effective cause final cause When sse say, “The Republicans lost the election because they lost the women’s vote,” we invoke material Cause. In this case. something happens because of the social materials that went into makin or unmaking it Demography is pa excellence social science of material cause, It concerns numbers of peopic of varying types arid the ways in which those differing numbers shape social life. By contrast, we might say with Georg Simmel (1950) tfac all social groups with three members are inherently unbal anced, because two of the three always ally against th thicd (something those of us who were only children in two-parent homes know very well). Here we are saying something no about social material but about social structure. Is is the shapc of the triad that gives it its peculiar properties. This is i/c/t tiical
cause.
Aristotle’s nftk’tn’e cause is the most familiar of his four The effective cause of something is what brings it abous v hai fotces it to happen. So we say that a strike caused employer re taliation or that a newspaper caused a war. These are statements about a direct kind of forcing. By contrast, final cause refers to the aims of events \Vlien we say the cause of universities is the need for education, we u
INTRODIJCT1ONTO HEL1USTICS
atrributn, the existence of universities to their final cause (which toda is e often taIl thnctiuii, although thats not exactly sn 1 Ar iii ant) W u is c say h reason for pollu is s icc d cit air ye speal of final cause. 1. i p lee y t x t it Ject cc cause ne Ia v n as ii i nation I laiper political in ppus tic Os is liked to be their ?ay/nul cause. And the nunuers and distribution or those interests are the laws ‘naOe;.z/ cause. Every event has causes of all four finds. en Sti
inn
nothc r
a pis can
1 o
how usu
p
he ihur-cause list
they complain?) and whether that somebody, directly or rectly, maintains it because of this utilit). As you can see, the Aristotelian list is very useful. Time an again, you can come t p with something new b 1 swift ung 0 new type of cause from the one that you are implicitly usi Its also true that you can often come up with something ne by switching from one to another /ogin! corcept of cause. from
sii//cient cause (something sufficient to bring another thing about) to nertisay cause (something without which another thing cannot occur) and vice versa. But the Anstotelidrx hit probably more useful, which perhaps explains why it reappeac
w I •t’ to isa ( 1 sider uremploy iii I of ut ci 4 1 yme U in tcrins of its material. unenlyl y I Who art hey 5 ‘X7h r are they like? What finds of dualities do they share I )ues unemployment concern a
with so many dIfferent names and guises; cr alwaH be e in a tight spot to come up with a new attack on a problem
find
The Kantian categories, although much more abseract than Aristotle’s four causes, are also a useful list of topics. Kant thought there were some basic frameworks through which
I
or a transitory state for
at person
many
different kinds of
people’ This is to think of unemployment demographically. Or in an think at nemployn ent in terms Ut ts proximate, C//dC ani: H layoffs voik \\ ho decides who gets fired or I e icci
it
are
Ii
tin
1
urnc
lot
or ch os ng unemployment? drivir p lowered employment?
tives s
uneni ployment
terms of its firma!, ctrmtura/ properties. Couid it he the case chat unemployment is a general structural ciualiti of a certain production system and that i a
ii
s
in
erelv random torces dcjd who in particular is unemployed h ()i o can vieis ut cinploymet t jiimtionally, asking i
i
it
h
ci bee
neth p 1
t
ip usc
i
body (for example, Lot those remaining
fir some
is by lowcr i p wages t rarc ned wi ii unemployment if y can
B. Kaiiic Lzst of Categories
experience was filtered. There are twelve of these categorive, and they make another useful list of aspects of a proDlem
1
think about. Kant organized them under four basic headings quantity, quality, relation, and modality. In what follows, 1
give the categories comnlonsense meanings, not the formal philosophical ones Kant gave. Our aim is not to get Kant right but to make him useful for us. Quantity unity plurality totality
Mi i iii
Dis ovi,as
t
iNTRODUCTION TO 1uJRisTiCs
divisible or indivisible? Social class is a famous example hue L there a power elite, as C. Wright Mills thought? How umhc are elites and ruling classes? Are social classes unified wholes ot
ihi
loose units that fade continuously into one another? ihe Kantian quality categories are reality, negation, an limitarion These, too. suggest important ways to chan,e oar
kelatitu SuOstan’ c a
dUs.ilit\ d:pi
first conceptions of a research problem. The reality categoi raises the subtle but important question of eeii4matiuii, of mis
ikicic
ecipi’oLiL,
taking an abstraction mr a reality or—what u very common bad social science rhiniring—imagining that because we nas
1t)diLlt
name for something, ecpt
Ice
uf
it
ifl
is therefore real. Take the famous co
socialization, which is supposed to refer to all
tiic ticiiie
ing by which an infant and, later, a child becomes an adult I
ie Kant i
ty are
c nit y, plurality, and to lOt se ‘u it a number ol ssc nrial ways to rethink a rsiaaii cjiicsrion iiirv raises th issue ot rh i/flitS ot our iailS us: \Vhr ata they? \Vhv? How’ are thes’ unified What, it uXafl} la, is an ucuupation It’s ohs iuus what holds doctors tuatlier
ds
i
unit bat
1W )cWitOtS 1
wit
a liry at
c
u D cc
a
hat about ph sicians’ assistants? what s and scaitres s, Are these really units? ci thc concern k na jihu’, Ate there few or i att r h( v n n there arc 1 Could differ xi
icr
1
o for example, how many Does it ma u a difIyrcnr whether we r p xc a; t ii at cooks togeti r What about baby-sitters and Idar care ii urkets? Or social classes; how maui or them are t sea i
it
ii
tioiis
I
1 oudirv fdlse the problems of the in??! nW/ia of a subjecr. n i Lulibed xc h Ia’ Flaw would sic knuiv in xi- hat wa s is it
is by no means apparent that this word refers to anything orb r than the slim total of experiences a young human has, Put at other way, ii, isn’t clear what experience a young person has that could nor be said to be socializing that person for something or other. Nor is it apparent when socialization stops and life be gins. There is in fact absolutely nothing that is denoted spe cifically by this concept; it is simply a reification following from the (fallacious) functional argument that because people acquire skills, there must be some special process—different from the rest of life—that ‘trains” them. Thus, the reality ca egory invokes for us a crucial heuristic discipline, forcing us u ask whether the nouns we use in social science refer to real things. Negation, too, is a centrally important topic. I shall kiter discuss several heuristics based on negation; problensahzing the obvious, reversal, and the like. I shall also discuss the cen tral heuristic importance of making sure that your idea is capa
51
i
s
o
i
I’1TRODUCTION TO kiitURiSTICS
)iscs s
Pu or hctnL w -on. Xe should nvcr
to think about
toriet
ii icurist c cool Much of nor 1’aI c u c a ta ily rakes nit form of tth limits to general izata irs ‘xph ru g what so rologital positivists like to call Und what conditions is some argument mdi 101 ii pi ta ii P c s take effi t? These and a r t i c o A hnr’idred oilier quest ions all arise ironi thinking about linmira I n
i
it
a r
ii
t
So, tot i-xmmpln, wt might S id that many things that we ct invented at par i 11 a c ii ry st in Ii up it iditioti are in ndc r what co Ii ions do pcople invent tra w ii m s r ii 11 a nation iltion: \V h;n t her oar ouhooil i rhracenej? Whe
too
formed? Are there par iLUlUt kinds of people who are Ii i t t ) iv ‘r t t a Jitrons, Art they leaders of liPt social noscn rats? huh-n aristocrats Are there ways to differ traditio is All o these questions cuitiate iii cntcd and ‘real is
new!
‘
ri
e
is
liii we
t
y to set I routs
00
thit concept of invented
it
‘i’he Kantia i relational categories are even more important, I have fkni us hra aries in philosophy The first of them ai d the di is or ef the world into gien ci s u cc aeci lents s h nanc ar I rh properties of those things (acci iii na —
dents), In some f airs t social scic ore, the substance/accidents cane rory provides no useful basis for heuristics, When we say rtai m ape 1k r -xar pie, we know very well r r ron s the trbsranee ord the age is the properrt But iliac rio -
hat, sat, soci hoer L;, it is nor at all clear (no r ci icanori what the substance is and what I ill iii es Is so o It gv a iamt tbr cveiyhody with certain c It nt Ii Pu uid of dci; reis and training’ Then eduaai ion defines the sub-
ut
I a P
ni-st
It
is
liii
al values stance of sociology, and other things—people’s politic comi: types of employment, sociological ideas i-rid concepts—be a accidents. But I could just as easily dhne sociology define who hold certain kinds of jobs, in which case the jobs concepts substance, and political values, sociological ideas and kind ot and education itself become accidents. Note that this n of sub analysis begins to suggest char the whole distinctio , large stance and accidents is probably a mistake (as, indeed ing oo body of social theory believes). At the very least, reflect or seeing substance and accidents can help you change your way something. /depen The second of me relationai categoties is causality heuristic dence. Causal questions are obviously central to any ted list of causes. I woo t as we have seen in Aristotle’s celebra hack consider causality further here but simply refer the reader to that discussion pro The third relational category is reciprocit>. This, too, ns. Often vides a helpful way to rethink social scientific questio of ew we find ourselves in a cul-de-sac, trying to decide which educ things causes the other. We know that higher jevels of causes tion are associated with higher income, but which over which? Higher levels of education lead to higher income allows the course of life, but availability of higher income generations the transmission of educational advantage across educa There is a kind of reciprocity here between income and whose in tion that forces us to be much more specific about nvolsed come, whose educatIon, and what temporal orders- re er such The category of reciprocity reminds us to consid social dir chicken-and-egg models. Many, many systems in can h take this circular format of reciprocal causality. They
Ic i rinDs Ui Discos
I
s
e r
ui
a s
1
cxs
iN I’RODUC I iON lu H&TIC
vsrc’rns ci r sm ilize hcmselves, or they can ys ems tlia blow t (Loosely speaking, one arises e back th o 1 r io r negativc.) The reciproc i is us to hink e epl abc t suth systems. anrian Ca )r ci i x dality are possibility!
eXist em
c;n
DCXI
n cc
and
necessiry!contim
sihiliry reminds us ci it is easy to come up with social scicflce aryumcnrs that arc impossible and that, theretore, s c Iced to check our ideas constantly for possibility. This is particularly true because much social science is motivated by a desire to improve society, But certain kinds of improvements are IoicalI impossible It is impossible, for example, for eer one to be successful if being successful entails some form of suyeriL rity to others. Ar least it s impossible unless we de line 1 1 success as bein 1 absolutely idiosyncratic, Yet sc filled with argot nts that implicitly believe uccessful o v tr ust always reflect on the 0 t I ty in (01 strueti our arguments 1 e a xisrtr cc iai es iciest ions much like those of ins’. ‘I n iiauy types of social actors: dc andecl people the sam, ‘md so on. hieh of tiles ualiy have existcn t s groups rather than as sim i ple tvpc Inth ed, siiat does it mean to say’ have existence as groups 1 here are many famous examples of this set of heuris tic problcms It is easy, for exampiL, to talk about class. But do classes exisr And llrLt does it mean to say that classes exist? Are we talking about self-consciousness of ciass about coordi nated action about simple common experience? Or take occu parions, Are they simple categories of people? bodies of work? Organize assoclatior-is of workers liat does it mean to say 7
that a
occupation exiSts
Clearly ci
of contemporary social science
involve
n ost ta
i
u,
x’
gender nd tcL
p c f
ci In what sense? The hcuristic luesticrs by the category of existence are thus like c1iee of thc cc r 10 category. They lie in questiomn nouns we couirnooly 1 denote social groups and asking what kinds of ch rig chm
women a group?
ii
norms actually
label.
Finally’, the category ef necessity concinency raises n c heuristic questions bouc boss events relatc to one aOotInr in one sense, these are like the questions of rh Lrnitatioo fleunS tic: are certain relationships necessary, or are they conLingeru on other things (that is, limited)? But contingency- IS a ro uzh more complex phenomenon than mere limitation, ii. iPVite u to ask about the muiciple dependencies among social proee sc about the many paths that social processes can take, And ne L shy invites us to focus on necessary causality and its mu i canons, When half the young men of England, Fiance, rc Germany disappeared in the trenches ofQorld War 1 a gc me anon of young women couldn’t marry—because there was i one alive for them to marry. The resultant family structure ii i indeed the resultant larger social structures of employment an I shaped European society for generationa Like ‘or necessity pervades the social prucess A good Is ut
opportunity tingency,
heuristics will never omit
it.
The Kantian categories thus provide another useful tisi of heuristics. As with Aristotle’s four causes, ‘we can let the philosophers worry about the philosophical validity of this list is a useful checklist of things to thinx about. cs ii happens, Aristotle had a category list, too, which Cut up the world a little differently, Aristotle included two things that For us
it
H
4
n
ii
u
Dnr
ov
Kant iriade separate spae and nine, Both of these are them selve i setbi licur she reminders, lways ask yourself what the spatial n I emp ral scr ings t f ye ur problem are, How can tiny he eharied? idh aspects ci the Ti art necessary or bOth ‘
It flt
t( de te mine
Iariu s
whit Ii airs
n ‘our qu st or
ii
[05
1N’li{ODLCTION TO HEURiSTICS
iry
if
spate
1w pnahlcm) Are there regu (either social or geographical)
certain way, doing (or not doing) certa’t things (like fastening seat belts) agents—certain kinds of actors (It turned out p1ent ot older drivers were drunk on the road, be they were lea
anion—driving a
likely to get into accidents, possibly because they han more’ experience driving drunk and so were more skilled at it.)
C, Bi
i’
vi ig
‘at
1i Ke
rune—where people drink, how they get there, and hose they leave (This was Gusfield’s way of attacking the
‘,,
the modern setting brings us to the five keys of Iramiti in wt td th b the f IT tis literary critic Kenneth hi rhi i F is he ok A C ‘aimi ii il Al fl, action, actor, agent, scLtnig. purpose. We can use this At too as a heuristic aid to rethinking any ar ic ilar roblem Si cc tI is is a i mdc’ i list, I ian gis e a famous example. In us s lend d )ook TAr Culture Pub/u Piy,blemc Joseph Gus h.l I rec in ept ual ut d drunk driv rig. I Ic said (among many eti er di ni s h t ici ide it caused by drunk drivers are really a ‘a s ortat OI y robir rn, a problen of the iriJing, the locations where pen 1 ii drink The San Dreg i police had consulted Gusnell about a st dden ri e in accidents involving alcohol, He to
‘
d n
built Cur major hotels on vacant land near rIirerstatt highways all of tlmcm filled with bars and all of thieni i1iaccesSi) I y too i was pr tty likely that you were go-
i
it
t ra
If yr u
,
i( e ut n iobiic acidents involving alcohol, If peo A u se he pet irunk when riley can walk home (as in the pub in I ngland), they are nit ci kss likely to drive drunk 1K hit I F s ii liter m ne c ay analysis ot ale ohol based ac id ii s i i terms of Burke five keys ofdramatjsni: Are a (‘iclents be r understood as a matter of 5
question.)
agencj —vehicles and roads (If cars wouldn’t move unless seat belts were fastened around passengers. fatalities would be reduced.) puupose—why people decide to drive when, where, and how they do (Some people drive to get somewhere; others— young men, for example—drive to show off .
Ia Another excellent example of Burkean thinking is the intro mous paper of Lawrence Cohen and Marcus Felson that dui9a duced the so-called routine-activities theory of crime (I p s Prior theorists of crime had emphasized criminals (that is, and Felson noted char itive actors) as the key to crime. Cohen of crime takes three things: an actor (this had been the focus can prior research), a target, and an absence of guardians. We in think of an unguarded target as a certain kind of scene argoBurkean terms. The central thrust of Cohen and Felson’s after merit is that changes in scene caused the crime increase were 1960. More consumer goods were in the home they le), lighter in proportion to their value (and hence more portab people and the entry of women into the labor force meant fewer
I
s
\i. rum
Were
l
at
t )
)f pai
d lx
aloes o o ic
er nh ii
it
liars
he Ii
c
st
u
ir
is
13 irk
,
as
e 09 n ia
u
ai
Waft
p1 r ol
w
ii
‘l
9
es
II
Over
ci
iy
cii misc
i si
sr or
p
P’ )pcrr the cii thors ilL tLialiy goods n Sears, Roebuck cat
zen, o
\ve[1
as ii e percentages of
homes with p ci us taket rail d in 1960 and ith ‘r ip u I Iv uc jOLts Factors paralleled prity rir e tom 1)50 to I 9 5. Once ‘aised a si haL new theory in this case of cr
really , ust
is
INTKOOUC’iiON iO
it
p
n
Burki s list i po iii
Disc o ian
ii
us.
Wliyi And one can also
m
i
ility,
ersi of rue famous old \X hat? Qiiere, 7 en How \X h
aliotlier s
\\ ho see in
it
of Aristo ember that the utility of all of these lists es icss in their nm city than ii their heuristic power Re port rs ist the who’ what-w hi a ist to remind themselves to touch a1 1 the bases We ire more interested in using lists tt) re n inc us that ou 5 thcor often foe its excessively on one or an atiler ispeci of wh t We st u ly. \V1 cn we iced to think anew, it s usually a p tic stioi ut figuring a it what aspect of our analy siS u old La ‘hang 1 ‘ode cc a 1 w Pen. iii
1)
)i
ii
I
ft
1
lx
se
tio
is ht
Pr
gi ati
i
re ii
U
(
‘ci
harles
Ti ‘u
Ic i lents and things to which they refer, tel ii ions lv tn•eeri symbolic statements
don in which they are made What is rad ii (.hapt r )nt is its noting that mans ‘Clleafu5 be ies e’ rhat prag manic approaches to explana .
iS
is
sysrc
in I ii nu xi cm a 1 Oil ny 1
?/L
p M rris’s three aspects of symbolic si ma i ic and )ragm itic This list was of u, eel ii Lb prer One Syntactic relations are relations n U i ci ts h lie a ter 5 mantic relations are rela
A final topics list sy ter s. yi at t 001st
en
Un/ti
ii
iairiy strong echo
f.
tii ient
Hi’.exisro,n
hi.
tion are the only “real” ones. I used the Morris trici to start a thinking about explanation more broadly than is cusiomaic That is, I used the Morris argument heuristically. It can of course be used in other contexts. There is no rheces sary reason, for example, to chink that it applies only to bolic systems. You could think about the S) ntax of markets th semaneics ut r C\ls 1 th tile connections between groups n tile market and tence outside it. And you could go on to think about \vlidi am tots in n’iarkets are doing (saying) and what the actions th pragmatic context) of those market assertions are One way 0’ stating Marx’s analysis of work is to say that there was a FUji’ damental error in the belief of liberal economic theory iii tlic
(internal
market
relationships) over against
separability of the syntax of markets (that is, the wage teL tionship) and the semantics of the social groups in those mar kets (workers and capitalists as they were outside the market1. Liberal theory said these things could be separated’ ±V1 rv showed, in endless empirical detail, that they could not. Maybe this is far-fetched, but seeing market telations as related to so eLI relations outside production in the same way linguisLic syntax is related to meaning and reference makes the tradi X’e can think of tional analysis of work suddenly look alive. 1 new questions to ask, Wmi
stir
MORIUs LIST, I come to the end of my own curient
set of topical lists. Social scientists use many sucH lists tiltougn their careers. I have often used knowledge, feeling, action (from Plato, Aristotle, Kant, and any number of others) as a use ful commonplace list. Many of us have used various lists of Se cial functions—-Talcott Parsons’s adaptation goal attainment
1,ITRODUCTIOL’i TO BEUSIST1cS
MFTIIuDa 01 Disc OVFfl
uitegcauon
and
pattern maintenantL.
fbr example. Most of us
aso us’- the discipLinei from time to time as a commonplace La 11 ions’ liii ‘Whatwouldananthro polognt savt Sometimes there s no tstcr ay to come up with a in dea than’ vi under ban ‘,rnlebcxl) from a difierent dis I c I t I utyo.u’i a Ihisisparticularlysobe taust as I iaottd in the precedin 6 chapter. academic disciplines a t orjaniad around ditkrvnt dimensions of diffirence. ‘Ii ed vi vi t isethscndTanyotheriuts.But in dosing m• discussion of topics and commonplace lists, 1 sant to underscore ns o cautions First, do not reif’ these lists. I u .oneofthem,we ‘I’ Li dun t need to aSSLtiiie. thur turret tness or truth. They are siffipl) userul lists or reminders o things to think about, re n t I n u a,c stuck l)ont worry about their re .thq t r truth. Second. dont oeruse them. Classical rhetoric died because .
•
-
d
I
n
i
t tcsarieattindei Soevcrythirigfrom
cendeiloms to rib qes to pun. gristle ‘was turned into ground beer. Don’t use these lists as some kind of comprehensive sys 11 pt ‘at ryar tea ‘hquesticnsthrough.Just use them vihen ou et stucL Use them to stimulate your thinking When you find that stimulation, turn to working out d l tc iew artu aent ) i t run through every Last i c iris it list tor ever> last idea and then try to put eserything together. Youil an er get anyvi here P anl r w at hcunstigoes ton.way Youare ía: better nfl naiLing one major leap and then viorkn*g out all the details and subparts of that leap than you are tying to v ri tt r a1 riacrieap aidsukleap thatcouldbe -
•
I
109
taken. Take the time to s ork out the details of a majot hcuris tic move. As we shall see in the next chapter, most brilliaa at-
ticks and books are built on one panicidar hon. Thc. aathor made a big move, then spent a lot of time workzn ota the details.
ULAItCH A Si) ARt UMEN1’
ur F o Chapter 1 (i\ER i, HEuRIsTICS
Search and Argument
I Sca
I-Icr
H
AN
A
)RRU\\ 1L
A
lAk NC
B. B
11
I
L
/
N
11
H
Fr
Y
Mriiio
H C Sil s
NC
S
is aa C
I
I
\L(
I oar i FMATJZ1\H 1vlim
(
s
iu ,iit
FElL
Oh\ mr
S
Assr
MPTIHN
ic(l[L 511 .11
n
Ii c
Nt
DLSC Ru
general iules for coming up
\s itli iicw ideas. 1 shall illustrate these heuristics with a variety of CX ip Cs ra n on s& vera] cliseij incs. I he examples are illus ftC 15 C
a
Cu i
not dcfniitivc ihc ieader should not get the idea that a ai x ipi lius tate one a ci only one heuristic, In fact,
I end c p rc usi
ig
h b (Urpi et
Li IL iiii
S
11 us
a
long r i )1
C’u tses i i
exampies. Just as there are several ways to vu r i il cd, there rna be several ways to in
orne ,
d n tl
IL t
:1 ii ct ual 11
uves of ari given article or hook. SC Inc iplcs lat ere peruAed as clever ftc r they were ritten. Such papers are Curi i s ci scien es. fhe economist Ronald
hr ned 9, B
economics until the Bfius çCoase won the Nonel Prize o 1991.) The aiithropologisr Fredrik Barth’s ALvlii; o FaiN! Ura flI_OtiOn was puhished in 1960 but ditto become a classic ‘aration until much later Ludwik Fleck s pioneering book on scientific thought styles lay fallow from its publication in 1iN until it was i’cpopularized by Kuhn in the 1960s and finall translated into English in 1979 That people took so long to tecognize the clean itv of these works perhaps cells us something important bout the natur of creativiti. Much of ir has to do with how one’s ideas fit w tim others’ current beliefs. Ciecuicity is ,e1a;ioni. Coast’s woN. went unappreciated until the rest of the economics eomrnurut came around to the broad conception of economic thinkmng that Coase took for granted. Fleck’s book was completely ignored until Kuhn’s Srnct,’ire iif Scliuiriftr Rrco!uriu,ts prcpared people for it. Often a mainstream cannot see nev ideas as crc
l.v Ra
‘sic
lleUKuncs
ii
on the nature of the firm wi pub ur lid ot mc a touchstone of modern
ative. Often it cannot SeC rhen at all. This tells us about an important limitation on the practicL auttience. If you of heuristic. Yhu can easily be too radical for an aim to have an impact. you have to adjust your heuristic gam bits to your audience——whether it is a bunch of college friends, a seminar, or a subdiseipline. Note also that the cyclical rela tions among methods and the fractal character of social scien tific debates mean that it is quite possible to be too radical to one group while being insufficiently radical Lor’another. Practi tioners of SCA might find Fleck’s view of the conditionality of facts so radical as to he irrelevant, while contemporary sociolo
gists of science would find him tame. This rule—be different but not rca dfferenr—cakes us bacH to some earlier themes. As I said in Chaptet One, the ai zi of
)iscos I RI is
SEARCH AND ARGUMENT HEURISTia
sa something interesting—perhaps een true—about social reality. 1 e hale some conventional ways to do that ‘a hich Wc call mcthods. The rule to he different but .ot too d,Lfereiic rmcnds us that Each methodological commu ia cts owr it how far i too tar It changes from time 50 . 1 c J
SOILI1CE
to
I i y sou si i y age remember well vi ci v s d ed r rultiple regression in c hi be d r nanoseconds by eleven v wed for i action’ we always repooled riances No suc care exists today ‘I here are. howeer, neuer rules about what’s OK. 0 heuristics in this book wcll sometimes take you clean lb out of ‘a hatever standard world you’re currentl in. That’s the ufl (it it, us tar I m cunerned, But on should be advised lit oi a e too c utside the usual methodological communi icre arc things thu ak strange noises in the so i•it if I haPs w i rethodological communities b a at t i c i cc cvi s you won’t have to deal w hose tI i s cgular I s s if yr u don’t want to, ii this chap i ci the next I discuss general heuristics. L nlike those in Chapter Six, these do not derive directly from tiie fractal debates 01 Chapter Two, Thet are tested ways of broadening ‘aliat you are doing, ‘a’as to come up with new idcs, ncs methods, or new data, ways to get unstuck. Re niber that the, are not spec in all aimed at any particular w aspe I i search p ess I hey are useful at van1 d ‘[lit
u. s w k ids of c, i
t
arch h
al h irista gambits in this istics the simplest form of
general heuristic. They involve seeking out new data, methour, and ideas. They are the first step beyond the additive heuristics of the preceding chapter. ‘The second are argument heuristics. These are ways to play with or pose arguments n order to c e ate openings for ideas Like search heuristics, argLtrnent hernia tics are general strategies for producing new ideas. But rathe than helping you look outside your pioblem or data or way thinking, argument heuristics help you look within bendirLl what you have into new shapes and new uses. I. SEARCH
-
HELRISTICs
Search heuristics are ways of getting new ideas troir elsewhere, When you use search gambits as heuristics, you are betting that someone else has already thought seriously about your problem or something like it and that you can borrow that thinking. The central search heuristic is analogy, it could be at analogy about data: “the marriages I am studying are really like negotiations in business.” Or it could be an analogy about a problem: ‘the problem of trying to explain why unions fail is just like the problem of trying to explain why X-ray machires fail.” Note that in the second case, we aren’t saying that unions are like X-ray machines, only that the process of failure has a certain logic to it in any circumstance. A specialized but important search heuristic is the borrow ing of methods. Borrowing usually involves analogy but goes beyond it to invoke riot only some ideas but also a whole appa ratus of analysis. It can be quite general or narrowly specific Let us now look at these two search heuristics in detail w th some famous examples.
S
I
I I
1
U ct
SSARCkI AND ARGO MSNT HeuRisTics
ccii
Id/
S
‘he hi
is
of the general sti 15 iuki hg au analogy so ung that an X IS really a G. I ri s yc Cc yon w c pet ring Y. That would have nc.xt (hoc f i is etc usine the n/Jiiiis heuristic.) Examples ii ci
1
mans
was’s mo t
important
lieu
cial s iencc. Applying tarn hd ( cc models to cxj lain stare formation in teudal as r k n nal gy etween fudal kings and rn r irhoiia1 actors Appi yi up ecological models to ho ii I srI ii H irg ss p1 d em t citucs in 1 95, and Haniiais i d Freeman applied thetis to organizations in 19— 1 y
iii
in
ii
creat
s
-
n
aHab
y bcrw en I iman societies and biolog ical s sic nis Apply ing economic models to family planning H
iki
p a i
ean m tkirp n ando ‘y betw cii people having children and I ix i p 1 nnburgcrs [flew
U
ft
mciv seem like far-tethJ cunalogies, but they were m Vc (ft sidcr t c conc inc analogy. Gary Becker,
greatcst ipostle of this analogy. began his career with what as z t i it ii sly ist inishiiig book, I Ic Economiac Dir urns ‘u/i ci. Stippose, Becker said, we think ahoLit racial dis H nat on as I as cii y an c ‘on mit phenomenon We can estimatc a price of discrinuuiation by the thiloiving method: IC
arc t.h
lì curl
wages pa d in southern textile mills that efll by all-u lute labor forces with wages paid in mills em ploy T nixed or all black labor foices The diffcience will be the 11cc thc bc tots ow tier is willing to pac for his discrimina vc
mci
tion
cis
Ii lip
i he were h iyii p it like a stnt if clothes. We can then all ihit app crams k m ic roeconomlcs to hear on that price,
aiian zing
oh
rp
te with labor demand and supply; 5 fluctua h rr c c-oft hetw en spcnding one’s money on dis how
it
11
crimination versus spending it on ocher things cnew capital for the plant, for example), and so on. Becker’s analogy must have seemed shattering at the time. Indeed, nobody outside the eco nomics profession paid a lot of attention to The Economics of Dii crimiilatio,i. But the analogy was powerful, arid when Becker began to analyze more mainstream topics, like family-planning decisions, his work began to be regarded as truly revolutionary. Analogy is fundamentally different from addition. It means truly changing the terms of analysis, not simply adding some thing to them. It has a risk to it: there will be naysayers. At the same time, it can be very productive. Many analogies take the form of Becker’s, which begins with the theory and method and moves toward the data. The Becker claim was really “You may think that phenomenon X cannot be analyzed with my theory/method T, but in fact you’re wrong: it can be.” It is equally common for people to start from the data and use analogy to find new theories and methods. That was the source of the ecology analogies men tioned above. Park and Burgess looked at the raw complexity of the city of Chicago and asked whether the city looked like something that someone else had already come to understand. The answer was that it looked like the thing biologists call an ecology. So one way to understand it was simply to borrow the language and some of the analytic machinery thought up by biologists to analyze complex natural systems: the city is an ecology. Ditto for Hannan and Freeman, with their ecological approach to organizations. Organizational fields, too, can be seen as ecologies. Looking for analogies from the data end is the more cornmon experience for students. Suppose you are interested in the
aLARCH AND AxGuiuoiT
y i ties are uverned I c usual line of analysis treats this problem quite traditionally, as a question of understanding p1itics voting, councils, bureaucracies. But it is perfectly possible to treat city government completely as a problem of ecorlonuc’s: economies of favors, economies of patronage and polities, economies of location, In this analogy, city politics becomes simply an economs’, and you can apply to favors, pa tonage, and decision-making all the machinery of economics: and demand tradc-o Is budget constraints, elasticity,
c
in. You may Oifl e
o
u
d
t
ia
lot
the final paper using ci whatever surface rhetoric
t
writing
bu
rowed arguments and ideas u I o id I it s t ia y rem puzzling when you think urely t t a flu a! t us as problems of power, au ty arid iufiucr c A tI i r ripie makes dear, one of the to! aspects of analog is it most often the ideas you bor )W w:ll be quite well wt deed out, When you forage in other you
i
hr
n
)
disciplines and subdisciplines, you will find the intellectual supplies plentiful and is eli kepr, ripe for the taking. Analogies don’t always
ode, not even the Ones that make it
i
into print. In two essays, the famous sociologist Talcott Parsons once gas e an analysis or power and force in economic terms tqlda,b), He treated power as a medium of exchange, exactly ii mont s lie treated o cc is the ‘gold” backing up the “rnonc’v ‘) syst i H r flu cted on the uses of embodied capital”)
p
m
lineal growth (exactly analo—
.
1
Ml this hinged on a simple, di— b v cen pm e inc money, pa a yn hat analogy are brilliant but
t )HOIT r
r
t
1 i t
bizarr
v
I hey
c 1
Iliant because they make us
Hau uscrics
think about power iii a completely nosy way They tie bizaiie because Parsons never med the analogy co quescuxi mc din bution of power to iodiduals. “fec this is rh base LOpC ,r politics—who gets what where, hew, and ishsci—tnaugo that of economics (other than iyiarxian economics) This exam pie teaches another useful lesson: in analogy, somethiiig cats traliy important can be lost—in adehcion to the SoiTiethiny gained—unless ise are yen, careful.
Note char analogy is not simply a matter of goiny o orh’t disciplines and other bodies of knowledge. It is fi ‘st tin most having the ability to break out of the standard barnes Vu put around phenomena Having this ability means seei i example, that there is a close similarity between schools, p
fo
ons, and mental hospitals (David Rothman The Discoi cry of nv Asylam); that bodily fluids like mucus and semen cross bound
aries in the sante way unclassifiable objects do IMary Dougdn Pui’ii and Danger); that everyday interaction can be trnued drama (Erving Goffman, The Presentation
of
,in,F
iii
LiJl). Obviously, it is crucial to know when and how an anaiog works; after all, many people besides Goffmai; have seen life as drama, not least among them William Shakespeare. Often the key to an analogy is not having it but being wiling to work out the details, which
exactly what Goffman did, To cultivate analogy, you must do two things. First iou have to be willing to make rash connections. Tlus willingness is itself a character trait, and you will need to get a sense of iS
whether you are more or less analogical than others, If you ha e too little analogical power, you need to cultivate it, if t
much, you may need to restrain it. But to use analogy ode tively, you must have not only the character but also the means
-
Saioo-i A\D jRGui 051 Hacacoics
)drinons 01 D1%( m tRY
I IS
must read broadly in social science and beyond. The more y Jo I ye to dia on, th better. That is why many great social a I ti n es tiways reading outside their r sornc oil high school or r ys u giiit I iou
and putting them
nw uses, tit’s also one of the reasons whi many yreat social scientists began life as histo rians, phi sicists, chemical enyinecrs, literary critics, and even nun
o ceo
it
ifs
yti
ii
i’s,)
th ori
iot
a
it is
t
i
of analogies are generally n And analogy often pro igument which must then
those meriiods people Liavc invented for atiaIyairp G .5cm times these are quite general borrowings. Most ol r e statist ca tools in SCA were borrowed in toto Ito n biology and econometrics (which got most of them from lit It y i i an case). uorrelational analysis, multiple regression mpcritrlcim a ‘
and quasi-experimental design, hypothesis rosLxng———ncarh all were deeiopd to analyze crops and raids and rucuhiLao genetks. Other techniques ‘arrie from cisevhcre ‘ilic riot tional methods used by social scientists to analyze boy lot
1 y thos vi c usc t i 1 mdv iii tatting point fr at be carefulli elaborated and critically worked out on its own. but the overall fcr is that many an intluential paper has its 1 simple nalogy that is carefiully worked out. t ui a dod i te evident in famous titles of i kg is t ivas U 01101 1 f Livors, vocabularies of lik ich i a s of kn v edge.’ a i i so on, each one of which yr.” ‘jx 1 ii daunts the andogv invols ed Analogy is the queen of heuristics.
things take to happen (how long unnl a certain hi id i’ Ia gets passed, how long until a given compan folds) were dcvcl oped to investigate the failure of industrial de iCCs and the sur vival of sick patients. Ar the other end of th social sciences, much of anthropoiog’, particularly SilieC (diitFuru Gairri. S mous methodological essay Thick Description,’ has boiio a I
13 b
ing myself. In the early 1980s, I realized that one could rlun of occupational careers—one of the most basic things to be ex plained in all of sociology—as simple sequences of events.
cli
ice
1
ci
o
a
ihrt
‘her
‘
mean for
t
lri tug analogy. That force
method, Students generally I borro v i T1ie feel that th y are learning the methods a of ttns or that tick! and that their faculty supervisors will expect them to use the local methods. Certainly in methods ourses that’s true enough. BLIt dot the more general course pa
i
r ow
Its
ii
i.
a steal
,
ii
I i.e ainly for
arch
t
six
i
t
hing
ptrs
and professional work,
do
can ho put simply “if only I could make an analogi between X and G. 1 could use all ut
all
horro
n
relati
1 gene heavily from the textual-analysis methods developed b ations of literary critics. Often, however, the borrowings are more specinc ctnh CsL ach borrou” on contested analogies. I am responsible for Ofle 5
they were simple sequences, ame could appi “sequencing” methods to them, and I had heard about the hen computer algorithms Just titan being developed ho compLitet scientists, cryptographers. arid biologists to compare dies, ran’
reasoned that
1 i
sack code systems, and comb
protein
y not apply ( h databases. 7
these to social data? This idea proved quite powerful and spawned a in i industry. But I had lust something important in die na1o’.
I 20
‘I
\1i TuniC, 01 Dm OiiR\
set
,tAkCH AND ARC LMLNT rhwRisTiCS
a b o c gy md or ipuur science were not gcner a p i uLii d mc r ) i as areers are generated in time. Surek dir earli stages of a arecr are more important in some seflsO than th later ones because they can dominate where one nd ut ) I I I mowtd id ut thing with that im ci
o
S
mc
am I
auci ii u
stiC,
its veal side s well as its strong on ou ing was nsequentix nor a complete
mu
I
anal p. bi d
wi n
e all os a wide command ducij hots. Jr is by freeing oncu I from the cm veiiuionai association of ccrtain objects of anal VsiS wi rIm certain kinds of methods that one opens oneself to [lii ri Ii possibilities of horn using, But 1 r ring oneself means I Si 5
i2 I
ents marriages by getting cranky adolescents out of the house lowering unemployment by keeping millions of young people out of the labor market, providing a maximally supportive cii vironment in which young people can experiment with erotu. and emotional relationships, and that you know a lot
more
so on. You
will suddealy nod
about the educational purposes of
a result of this reflection. More important, now ,ou can see the crucial questions about the en/irarzona1 purposes ut college in a way that you couldn’t before you thought about all college
as
t ci
the noneducational aspects of college. You have pioblemat zed the obvious. A second argument heuristic is to make a rcs’ersal. Since everyone assumes universities educate students assume they
as cit k u uv edge, d se or distant, acci fuhl s ught f oti er methods and means of inalvsis AnaloLizers and horros ers unust always be reading and itarn1n
prevent education. List all the ways college life suppresses edo cation: scheduling boring classes, providing differing individu als with uniform, uncustomized learning. There are dozens of’ ways—the nucleus of a good, contentious paper. Reversals arc
>t
.lem
m
ci
it
u
iu
i
t
a
Ii
not necessarily reversals
it
Argumi ut heuristics are Ways 01 turning old and familiar argo nuents hiro new and creative Ones Search heuristics look elselie in id Ar t ment heunis i ‘5 work with the ideas one lit
ia.
ua
u n ook unfamiliar and strange. ‘1 he hst a ulInc ru heuristic is to problemarize the obs ious. Hr example. po ibieniarize the ohs ious notion that college is ihouit em mug thin s Sup ose the prposc 01’ college isn’t ed t
VI is i i phi t I t 1 H ‘ccl is there any tea son a I y cull c inpht lie cxi ecu’ to has e ally purpose? Think of all dir alternat\ e reasons (other than education) tor the exis tence of elk ‘es, and make
i decert case
for each: saving par-
of
truisms,
however,
although that
always a useful place to start. You can also just reverse phrases and ideas. I look at my bookshelf and see a copy of Edward Launaann and David Knoke’s book The Organizational Stare. As I know well, the book tells how state actors (bureaucracies boards, legislatures) are embedded in and surrounded by net works of organizations that seek to influence policies in vaniour ways. But suppose I turned the title around and made stare the adjective and organization the noun: Statist Organization( .m). What would such a book be about? Perhaps the ways in whicl organizations take on the properties of states—monopoly 0 force? e 1 ll, not real force, but perhaps economic force? ho reaucracy taxation? How can an organization be said to ham
Ii
1
i
1)
(
(YR
state
Nt w l Lens-— se that I haVe a topic. It it ii t tt e hid off na thc r worh ri-lath oship does en ii
ci
(I
ii
1 Id
i
e
p
(it 5dHzaliOfl
N
at
es
iii
‘iiat
c ipl
.
SeARCH AND ARGUMENt HEURISTiCS
n 1
gotten thete—to citi Hie waves (if lecent layoffs and e clear tha for many people atI citizenship of a kind, with V
s a e o gut ation but also rights in ki tds ot OrganizationS ha e citizens 1t i i toty have there been such
i
F loss does the i lea o employees rights grow p A o s ddt n 1 are Fit cleus ( t Lizzie, Note, too, 0 I h is e drifted ti’om teve ‘sal to anaIog: the new title forced th rg
ilizata us?
e
‘
—
ad ai es n se 5 w oh y of science, among other acids. 5 A ou eos see, making an assumption is often a prelude F itt sri g Yot usuall make ni assumption in order to sim pIll) 01 to transIae. A ii d tif rt in rgument Is ‘u istic is reconeeptualization, 0
sn
ti
iii
ii
that
s
lege is not really about sexuality at all hut about bragging rights People date not because they are interested in intima -y but in order to prove something to people other thaji those they are dating. Therefore, dating should be catcgorizd with other forms of bragging. Who knows if such an argumeut true, but it suggests an interesting way of rethinking a familia phenomenon. Let us now consider these argument heuristics in more de tail, using examples.
he ‘d
c tis is ip to the ss odd of organ— zat’oia. But tiic Stdtti ig 1 ioait was a ;imple grammatical cv rs t t ri 1 rt found the nuci us of the idea. third argcinint heuristic is to make an assumption—usu t H s c what it iets you fhe most familia iii? a rjsf on r of hese r sh assumptions in social science is to assume that some :t r d( ii’ ar ‘r no sal’ th t assu ript on buys you all the methods or mit roeL onomics and gme theory. Ut also has a con ra v tk P bet S trio s’s tel brited assumption that all rationoiit bust he hounded in some way.) But you can as on r r ot r hines ‘Vi u can assume, for example, that beiuse tiot human activities are conducted through lan— ‘a ge laisgu ge I old t se key to til socu 1 explanation. One must therefore analyze t in any situation, This assumption led ii
pose we reconceptualize college dating. Perhaps dating in coi
hat von thought ‘aas I) is really Is or even P Sup-
A. Problem.atizing the Odious Is there something everyone thinks is obviously true? A usetu heuristic is to attack it systematically. Much of the time this gets nowhere; people are often right. But a substantial amoul of the time, well-accepted and carefully tested ideas are pro foundly wrong. They turn out to have been not carefully tested -
at all. Perhaps the most famous recent example of this heuristw is lime on thi Cros.c by Robert Fogel and Stanley Engermars Fogel and Engerman attacked several widely accepted “facts’ (1) southern slavery was dying as an economic system imrrie diately before the Civil Wdr, (2) slave agriculture was eco nomically inefficient (and, consequently, defense of it waS economically irrational), and (3) the southern economy as a whole was actually retarded by the existence of slavety. ogel and Engerman rejected all of those propositions, which had been mainstays of the scholarly literature for many years when they wrote their book. In the process of that rejection, they demonstrated dozens of counterintuitive results: the money
Li
o a
u c
I tv
or
in
Div
Rs
ScH
g abo as highe t than what it would c b i ha they (c i Ira shaiecroppcrs (1974:1:239, 1 no n an I rgu pIdntatir us I ad black nianagemenr 1:212, I : and so on Fot,cI aud Eniermans twovolume work aasai a alot upoit publication and fox man y years thereafter. F gel ax d Engernxan wcte quite clear abou t pioblematizing i obv o i 1 i a t ti ry devo e nan pages to expl aining how v s A J it nxx i s if ‘lay that uas so etroneous became standard They also reveal I : appendix A) that they were not roe hat prohlematizers or these ohs jous facts and point to tnioi dii aa difhc ohs sOt h a heuristic sometimes faces. uo
1
i
i
-
i
C
Fiscncr’s Tii Due/i aruvn,g ix r s deflated by his book is the Hotion hit pen P who live in inc arc i ‘to ‘t isola ted—have rcw er I riends and acqailiran ces—ttran people in smal l towns or anal serri nra. I his belier is a staple of pop psychology and e i if n uch sc it us scholarly work. Fischer went out and sim i k e v 1 xc ii r utn out tlia the tru sm was wrong, alti ig [k ii iy ti u sf15, it c ‘taint P a grain of truth in iliac thc A in LI ot people orhanines know are somewhat different from rhoe rural people know. They are more likely to b nonkin. But tins turns out to he because urbanites are more likely I c y CiliA fcOplc. pcoplc lookinn for ness oppo rtunities and I ax I s .1 i it is I cople Ii) arc more likely to have ii tw rks it I if n m ku a e likel o live ii cities /r other rea -i Ayain, prohlernatizing the obvious led to au exciting in— \ cstiAatlori one tIiii hallcngcd old truisms and raised new tics titi n s lx
/
i
xooip
is
iauue
p ii e ma y a
ii
-
‘
-
,todt ox tiles i
act
id
i
t
it
xi r s
need to taLc on so i innumental a project ii oc r s avery r the city. The world is lit-
AHO ARGutLNT
HEURIs1iCS
L
tered with obvious 1dm
that are wrong Newspapers aria ia azines, with their strong interest in astonishiog their readea. arc fine sources of unsupported pieces of ommoii sense .oo sider the beliefs that members of generation X hold cci ‘ax attitudes or that the 1950s were particularly staid oi that Americans are losing their belief in God or that the family P falling apart as a social institution None of these lus muha truth in fr, but all are standard fare in public niscourse. Social science is full of such hollow truismS, tiaü, Takt thr common belief that social change is happening faster than cvu before. it is not even clear what this means, much less that t it in any way true, yet it is a devour assumption of dozens of arti
cles and papers. Or to consider something more controversial, take tile idea that departures horn equality in human 5Stci’fls need to he explained. This is a universal assumption of nearig all social scientific wtiting on inequality. We maicc thi
sumption every time we write articles on the causes of inquah ity across genders, races, classes, and so no. If inequality i, these areas doesn’t need to be explained we don’t need to scare articles about it. Now, we might want to get rid of inequalit} for moral or political reasons, but why should we think it need some special explanation? That is, why should we think it iS unusual? We normally explain things that are unusual states of affairs, as I noted in Chapter One. Yet inequality, far from be ing unusual, seems to be nearly universal n human systems. Ir something is universal we have to think very differently about its causes than we would if it were sonic special state of affaits Or you can simply take something as a problem that no o ye else has treated as such. Vl’ten Bruno Latout and Stcve i’ool 19 ( ) 9 gar did an ethnography of hId in a scientific laboratory a
D
SEAm
tiivias
It’ i’calized that
Ii id taken life in the i u and uiiprohlematic. Tummy the weapons of t made it suddenic ne a ii s i mgc i tug die obvious gross 5 out of the habit of always t
er
we
tug things that are said oi taken for granted It’s like a t cm uniling in the background on your computer. Every arctlmL or, very ceneral izat ion. es ry background assumption that ou run into, should be scanned with this simple check: Is thaL rrallr
rue? Could I get somc where by regarding this as a problem rather [hail as something taken for granted? The must extrc me s ersion ot this scan is simply turning such arguments t
OP
their heads, That
it
n
13
is
the heLiristiL of reversal,
to
which I now
\ lazn a Rn
Ar other of the central argument heuristics is to make a rever
sal Sometimes this is simply a grammatical reversal, I was once ashed to ss rite a paper for a special journal issue on the subject or boundarie 3 Boundaries and boundary crossing had become I sin inabit so I was bored with the idea ‘Boundaries, ii tl hugs, of boundaries of things, of boundaries of t m self in tile shower one day Suddenly, the s v id I had the phrase “thi ‘igs of boundaries.” V 1 1 an? I puzzled over ‘t (after I got out of the I
V
1
e n
11
i )i p
a
real crisc
1 Se spell
I undiries, ‘11 t
I
ii
V 1
1
Maybe social things nuch of my lifli studying)
edges comt first, then the t us by h ng a border with place A iu us hont ci ith lace F and yet an— and tht n hooked them up to make
ii AND ARGUMENT
Haui is in s
something continuous and all of a sudden there was an insid and an outside, and we called the inside a nation. The resulting papel- —titled “Things of J3uw’imniea. ci’ course—--grew out of that simple reversal, I made up die glut then tried to think of phenomena that fir it. Ufrer, CveL5d not such a simple grammatical move but rather reVerSal some standard theory. rvrnong the most famous exrnp1es 01’ this is [-Toward becker’s paper “Becoming a Marihuana fJser based on ethnography among marijuana users at a time when marijuana use was much less common than it is today. becuer cer started from the standard view of “deviant behavior tain people have propensities to do deviant things. In such a view, people take up pot smoking because of son’iething cnarac ‘:
that
terological, a motivation to be deviant. Becker turned that idea on its head: “[I)nstead of deviant motives leading no deviant around; the deviant DehijViot 10 behavior, it is the other way
time produces the deviant motivation” (1962:42). Becker’s ar gument was that people had to learn to think of the loss control and other physiological symptoms of getting high aS ot
pleasant experiences, rather than confusing or frightening ones. Hence, behavior came first and motivation—sometimes--’ afterward. This is precisely the reverse of’ our standard assunap
-
tion about human behavior. That reversal opened up zones investigation and possibilities of interpretation to Becker that
ot
had been closed to others. Note that it is
not
necessarily clear, without talking to the
authors who use thus trick, whether the data fotced it on the n “things of boundaries’ or it came to them in a flash, like idea. But the best reversal papers combine data and interpre my
tation in a way that seems magical. Mark Gtanovctret
Mt
—
‘Stren iios ci tt’r ii
n.
•tak ‘lies tells its reversal right in the title, Gra ss as tuicrested !fl what makes inrerpersotial Coflnec is ‘ii In am, chub is had drawi sociograms, 0 ii
i iagr ut s r u
i
‘on tt s
ri t in g
II
Ir
‘stARt H SND ARGUMeNT 11itiSIRZ
1005 ‘II Dist,o tas
n pi mt lsf ip
pc p
with huts between the us between pm pl —-connections
e as pc ints and onnc C t a
CXCII huge of money, or what It w s lwax s 1 osely assumed that di nse soclograms_ Oframs Hi wbi ii lUtist Ot Soifleont 5Ot ficetions are also his as
i in ix
H IC Cti( H
\ n
I
It’
so
C Ni it t i ) IS’ C H Set tI 115
k Wii. ‘ii
tIn e
i
ran us ttt r
null degrct of i
ji
Ittasr if at
y
iotn
the
art
ed w
s
Strong type
that if
eni nection iii a group
as well
is
sum,
e wit
iii
a eli
of net-
tin nk about that has several we
of
hnkc aaroa
gaps he tSVf en tI Cr5 I IC ft ft icjUt tiCS (50-tCi led xvea ties) actuall 3 do r a u il thc cotictin because they WLrC bridges between tin ne e dl c n tctit is fell rapidly if they were taken away. tt iii,
“
“
.
‘,
nt
r
filL
disappeared, it didn’t
r su h niatti r, bet ause the two individuals involved were prob bly c n set e I dirt ugh scv I c tiser eople as well. ‘r
Cr, nt
‘s empirical data ins nivec hnding employment. I iu ie a mr ma rise iaeople Granovetter studied usually it ux d ji hs through sonic sect ndha id connection—a weak tic- —r
In
n
Vt Li
uh r is
t
ian hiouj4 ai
s as
friend Many
i
nmedia e friend, The key
sour distant friend’s tint ic’s
to
em-
sister, not your
best
havt had the r xperience of this kind e dc ntal” ob ot tact, A id w all itink of it as unusual, In f ut a C ‘a 0 ettt’n S (licoritital argument shows, it’s the cornii ui ‘xp nit im, \ irhin oc r elk 1 lie, ill the people we know im p1
us c the ann jul i lb
of
marion ss t lan
Iccause
are tied to t e sat it ptoj IC wi ate Er is tliiougls their ft iends ‘rn/sit/c the J cjne that lieu i itt rrnat on comes iii ii
t wy
vll fa Anothir example is Paul DiMaggio and Qaltei bc built on a mous papet “The Iron Cage Revisited,” which wa fcQ direct challenge to the Hannan and Freeman papei I ‘tCfltiO a) earliet (the one that borrowed ecology to study organizatloii va The central question of the Hannan and Freeman paper answei why are there so many types of organizations? Tiseir
gio and was that ecological forces produced differences. DiMag asked, Powell simply turned that question on its head They iri why do all organizations look alike? Obviously, on the emp extent at cal side, the two pairs of authois were looking to some s ha: different aspects of organizations But the fact remain different they used their different questions to make very and things out of what they did see in common. DiMaggio or Powell argued that only at the beginning of their lives were ntia ganizations subject to the ecological pressures tbr differe they were tion that Hannan and Freei’nan had seen. Afterward, pushed toward each other by forces of “isomorphism.’ gio My interest here is not with the content of the DiMag heuris and Powell paper but with the now familiar namre of its paper on tic gambit. The paper turns the argument of another that its head, seeks a way to allow both to be right tb zation ecological differentiation comes early in the lives of organi l theo and isomorphism comes late), and then lays our a genera u ith ex retical argument about isomorphism and illustrates it arid amples. Reduced to its barest form, it’s just like Becker certain Granovetter: “They’ve told you that X is true, but under ons, conditions X is false. Let me tell you about those conditi the This is the siniple reversal heuristic, and it produced—in —three Becker, Granovetter, and DiMaggio and Powell papers of the most widely cited works in modern sociology. saying
SEARCH AND ARGUMENT HEURISTICS
v np
y liii I or
\t
olves n ak
be data it a oi ‘31/h tio a ph1 sicist turned sociologist, noticed tflat thea ire some inobiliG systems in sx bali holes, rather 0 v it n t v ( 1) No a an become side t f larva d u nil the corrat resident resigns. Then cii v S uV 11 ai I to b c pit dci t This merely t he liok’ —ci (Ocant y—rnu\ to some other place. Then H) i i) ii se to ill I i lace lea iH a I ole somewhere else. I. nm ills the v ones chain gets to th edge of the ss stem. aid 5 onhod elirers acadenne administration from outside to ii
Ii
HH
clv
k
as li_i
(
pm t
ia
tF
oiorl er vox. in icl
‘3 Lod
I
si a
a vsicrsal
i
i s If i abolished, ending sotem. /,e/ have initiative.
I )le opei
and n body can move cx v here he or she Lllooses; rhc possibilities are dictated by niL L 1 s ti at xisr ss he i ai indivie ial wai is to move. 0 I iw hat hr i e w u a wh )le class of occupations like thu (tootball coaches collece presidents, Protestant clergy, ii
e
it
I(
or
t
a SvSte vs
ii ss
ip.
bil
ts
ii
h L
ass
Hi
li
rys al
e
P
not a., hi’!
liiJi
is
ti is
s
Inch
it
Sf
ii
mmli larger class of mo was liasels true (tiniversity departit a was i
cdi al
iffs). This insight turned
mobility on its head; it said that constraints e iv He 10110 trot t in mc bility ti an cuber the choices or the tact r C th Sc I Yny to itiove. Iills res ersal like so many things, had it roots in analogy. is
0 e
aw Ir
iat
,
in
he a n icom uctors there are electron are mote ot less negatis e electrons, absences that
i o
w i
In.
Ia
on
with posiuve rather than the pliy>i ist already knew about a
haigc So \\ hi te in v bali lit les Ia ed an
it
un
mist rol
.
Perhaps the
1.31
suggestion to make a reversal in the thinking about pecpic; mo bility simply worked its way out through his subconscious. Like so mans of my examples, the idea of vacancy chains i.. an example of several kinds of heuristics comiog together. One of these is reversal—making holes more important than people. Another is analogy—between mobility systems and crystalline solids. The third is borrowing methods, for ‘lnte turned his
insight into empirical analysis by invoking a general class of probability methods (Markov models) well known (as of 1970) by physicists but unfamiliar to most sociologists. C. 11aking an Assumption Making an assumption—usually a simplifying assumption— can be a powerful heuristic. As I noted above, a simplifying as sumption is often a step toward borrowing, usually from a discipline that analyzes simpler or more tractable systems. Thus, by assuming that “value” was a conservable substance
like energy, economists were able to borrow the mathematical tools of statistical thermodynamics whole cloth (Mirowski 1989) There are other reasons for making an assumption, besides adapting someone else’s methods. Assumptions make for tract ability; they make systems easier to think through In formal demography, for example, it turns out to be useful to disregard men. As far as formal demographers are concerned, all men do is impregnate women; there are always plenty of men around to do that. It is the women who have the initiative; their age specific fertility behavior determines the size and shape of a population. So demographers generally start from investiga
SL4,RCH ND ARGUMCIn HEuins’iiC:
tlOHN
0f
iie sex.
that x omen can de termine their oss n fertility getting pregnant if. and only if tlRy C 0, popuLLions
o
01
t
assuming
ii ti igui
Ciwetn suth ractability as Ii art ddibeiat I chust i, and background as Sunipnuns, u lutli are merely implicit. All forms of analysis has e iriplii assumptions, it is always a useful exercise to re let t nud stuon 1 t se asst ption But 1 am here con c F 1 us asst ptions wh ch are designed sump
us
.
up a 1 iation anal s An exeJlrnt example of st cli an assumption comes in ihan aid Dunans iu’ie;i Ouiipaiii led! . 1i’/fC;uh, already 5 tO ()
L it
iptei
e as a assic cxamplc of SCA work. 1 ii ze t dependent variable of the espo I lit s C lent iol start by si tidying the wa•’ it is altected b 1 iidpendent variables like father’s job status, respondent s education, and respondent’s rst job When we vii it 1 C(jUdtl to c t u iare I c e effects one thing we ast F a at s s tictul the arrows describing what itft’ct vhar it lie nio Ii i——is thc same for every case. This translates into tOe assumption that every case follows the same stor, Oh i sly is a it cal ss lnption Otis Dudley Dun 1. 1 ill c )yi fist t vim did the study, knew this erfec 1 wcli I ideal ed mod I order was fathers job status and tdther’ 5 education taken together lead to respondent’s edu cai ion, whit Ii leads to respondents brst-job status, which leads u It nt’ i rentF statu )bviously, many cases will r 1 i e p’ Mr i back t school after start us mcii idiers i ay makc deliberate status sacrihces to IRH
t
‘
,
)i
guarantee tHeir sons’ educations; and so on. but c maulac tOe radical assumption that th sequence was eserywhere thc Sctl,c.
Duncan was able to apply path analytic leglession ard eals some powerful guesses about the ralative npor.u these iorces in shaping men’s lives, Th etual reLticai.sI p were of course weaker than they seemed becauSe they ware u a
ditional on an assumption known to be aeruncoLs n degree. But the power of the assumption was great, oh ci suIts, even though conditional, ware worth dir puce 3 strong assucnption—iike the Duncan assumption— An creates the possibility of reversal. Although Duncan w cu ‘welt aware of his radical assumption, many of his foliowtrs n sight of it Obviously, a useful heuristic gambit is to hallcn, such a foundational but forgotten assumption. Petcr Abcl (1987) and I did exactly that with the Duncan assumption i sisting that we investigate the order of events in careers. result: a variety of new concepts of career as well as new me ii” ods for analyzing narrative models for social lifc, Another body of inquiry that was built on question no standard assumption is the bounded-rationality literatuin noted above. Starting in the early 195 Os, the economist bert Simon challenged the idea that all ecouomic actcurs ue rational. In his book Models of Man, Simon argued that ration ality was bounded—because there are costs to the unformatu i one needs to be rational, because the problems invohed mat Dc too difficult to solve, and so on. He proposed that people “sat
isticed” Ifrom satiij plus sal/ice); they make decisions by sctt11 minimal thresholds for success and then search for actions iiiy until they hnd one that beats the threshold. Later researcf ci have elaborated on this idea in dozens of ssays.
3
i\li I 11005 01 1 )1s Os ThY
Mik up an I denying major assumptions thus constitute s another basic heuristic in the social sciences. Both moves pro— dose challenging and surprising results I R ipi ‘Pm ìç A final arurnent heuristic is reconceptualization By this, I mean taking a familiar or takn—Ldr-granted phenome non and t cat i S I t weic an example of somethin g quite differ s Te t it v t s a of X but t Y a even better, Z. gave I in the precedug chapter the famous example of Joseph (jusfields reconceptualizat ion f drunk-driving aced ci ts a a setting or location problem (too many people have 1) Iriv t rd r t d uik s ci il places) rather than an actor probleiix (coo many people are unable to control their cars be— cause of alcohol intake— the co Ic”pr implicit in the phrase d unk-cli v n ‘1 But automobile accidents had already pro P ci a far iou e a riple of rect neeptLi’ilization by a non—social ientist Prior n the writings of Ralph Nader, it was thou ght that high speed caused’ accidents. Nader s book Unsafe at Any Specd tecoriceptualized injuries from automobile accidents; they ve c a driver agent) pn blem I ut a car (material) problem. Gu’ficld then lat r reconceptuahzecl accidents invo lving alco hol as not a driver (agent) problem but as a locatio n (place) problem (Fhus, both of these are based on moves in the Burke five key Lu ( C saptei ihr e Somcri nes recunccptuali anon s almost forced on one by data. in Pie I 9Os, some eriminologsu ooriceel that rates of iiiotot. dc theft fell radically in states with compu lsory— h 1 nc a $ (Mayl e (1 irke ane Eliot 1989). They saw a P Ic explaratior for this if they reconeeptualized motorcy— ‘,
ScRCH A4D AROL
eNT 1-iCt
f14s
thit tand, latu, niatt .mnoc crime) as driven by j cit nity: it was an opportunistic rather than a planned ac so I a lichuer allO compuisotyhehmnet state, xi you haven suddenly decide to steal a motorcycle, the ) ce ,ill °l yoc at once for the helmet violation ann tlicu Haute ciii ou are a thier. The fact that motorci dc tucK fal
dc
compulsory helmet laws makes xmmediatc Sf SC 5’ Pet stop thinking of the crime as planned ad start rninkmg u as opportuniStiC But the nbtiOfl of opporruirlstle ci me dial views ul lenged long-standifli “criminal personalit Hence, the reconceptualization was a radical 5 tie Reconceptuallzadbon is always easier when one s with the lists of topics or commonplaces I mennoned in P preceding chapter. A seasoned social scientist always ke”p these kinds of lists in mind. He or she is always rethinkin things of interest. Is my case really X or really Yl can 1 sa\ nina something new by recasting the whole fraineworP which I siew my problem?
SFARC1i AND
ARGUMCNT HEURISTICS ace
the
simplest of
general heuristics. Analogy and borrowing, the major search heuristics, open to our use distant areas of investigation and thinking that aren’t normally part of our repertoire. But s 1 noted, one can take advantage of these other’ ‘areas only n unr iS aware of them in the first place. That’s what makes ‘nsacsablc
reading and broad taste crucial to a good social scientist, Them provide the basis on which search heuristics wcrk ArgulilLin heuristics, by contrast, make changes id what we ali’ead: at hand. problernatizing the ob\ ions., making reversals, inakuiy assumptions, and rconceptualtziflg_t5e are all ways o
M
Fiuis
LEst
15 HIS
In r we al ady have and making it into something new nd strange t nlike analogy and borrowing, they aren’t sr adtli rI km wiedge But they aren’t cp nheiu on ieadin or 1 dt th or know ledge either, They are simply a Lb pL’Ricnt inauci 01 j’ra tic ol hasing the h Ihit of doing them Nun, too rh it ni iking ass imptions differs from the other chrct rr imrnt heuristics, The other three are guaranteed to no[i,c l’hey e\pIiCirIv change or challenge some LaUS( publi thin By ()nttast rnakin a r big ass imption is often something ‘
ChapLer Eivp (ENEkAL Hiisi
DscRrnoi’
I.
is
with Becker s assumptions about fimily-planning “rationality,”
whit h stayed rontroversial for a lot g time It is probably the case that a good hr uristic assumption is a radical one—one that ,ets noticed, lieware of assumptions that are mere Conveniences,
NAATiUi’ti
DESCRIPTIVE REt RIST1CS
A.
(curious ol but his r a her followers are not That arrainly ss is the case with Dunran rsSuinption cibout uni was certainly nc the case torni areri’ sun nrncs, a ho igi’ an ant hot
AM)
CHANGiNG CONTEXT
B. Ci-jANGING LEVELS C. SETTING CONDITIONS:
11.
LLMPINO AND SPLit tINU
NARRAYIVE HEURISTiCS A STOPPING AND PLTTING iN MOTiON ENCY B. TAKING AND LEAVING CONTING C. ANALYZING LATENT FUNCTIONS
D. ANALYZING COUNTERFACTUALS chapter were largely Coil Ti-IF GENERAL HEURISTICS of the last and our general conceptions a” cerned with the methods sse use 1 will focus on how we ac objects of study. In this chapter, the
y as something in the world tually imagine our object of stud indeed, one could think o”’ both at a moment and over time. time. In the discussion o” these as the heuristics of space and mentioned the importance ii topics lists in Chapter Three, I lists) of space and time. both Kant’s and Aristotle’s category suggesting some pa’. This chapter recognizes that importance, useful in recasting our tie ular heuristic moves that have proved social space and of its lI )W conceptions of reality’s layout in through social time. I3
1J15CRieY1ON AND NARRATOx
)
s I iprion
C
a
i
F
ii
i
a suniptioiis I
i
i
V
u
in
gi Ic
social
orocess Every de d challenging those
av o r i cc ncw deas I ii r i descrt v on alsv.vs i is a foreywund and a back— 1 around. a tocl area anda conteXt. So is hen we srudy industrial nrms, for example. ii take the economic conditions they face at
15%
cuss is
as context. s’e also think of the s i rkers who isurk in them as pair Of their context. and is e consider the local politics and schuols n the toss its where they arc located as part of their con a ss C rudy tumid di liamiLs, howes er we take the r s in which the tIn ii ncmhers work as context, c SC ouls and ndibhb hood n which tamily mem
text tad us a,
There S n
I
r
ntext t
s
t icr t
lctc
n
mple
it
ays an et’fe ru Iii
.
if
ii it
reason
to
make some-
pair of the thcal area. The ont i ious But in order to i
it we St
, we makc 1 id
some
gr( und Challenging these
d and others I.
i
d c
ic
it
aiiv dest ription also li
c
a “leve1, in the
sense
that
thete arc things we imagine that are bigger than our object of study, things of
hich it
is a part
and that possibly determine
hings that are smaller titan it, things it contains and in turn detcrmiiies. An important heuristic move is to change this . to decide that nxa be rh determining action i les ci ut amds 5 itt. and
I tie
[alt
the s
it
c
at a
dillercnr les ci than o
e thought
it
did. Consider
school 1 here is a king history of re de ide wl he the determining action
if success in ii’
is
He ORiSliCS
at taletit) Wi i takes place within individuals (differences and values), or withu families (differences in family resources for teaching) Ii school systems (differences in school resources determining level of this literature, the explicit question is the causality eve ywherC Finally, a description doesn’t necessarily apply description, to sa it Perhaps we want to limit the range of a this rafle applies in some places but not others Changing one that raises iru application is another important heuristic, for cxampia portant and novel questions for analysis. Suppose, in various public media— we argue—as is commonly stated Americans are frightem that illegitimacy rates among African up this queshoit ingly high. An obvious heuristic for opening description might appiy. to analysis is to ask where else that and so on (ii among whites? Hispanics? the highly educated? throughout the pop turns out that illegitimacy rates are rising
ulation as a whole.) the question o More generally, condition-setting concerns Chow big” the phenomenon of interest is. Qe might be study for example. hue ing the rise of professions in modern history, occupations are ‘sot perhaps the rise and structuring of expert but are part of a really phenomena that happen in isolation formalizing all sort much larger movement regularizing and accounting), la ins of behaviors: investment (formalized in the creation of the codification of laws), and even music tin he study tag even-tempered scale). in that case, we really should (This was Mac a broader phenomenon called rationalization. eher’s argument.) 1
1
j 5
( J
)J
C on
I
ci l)ism
cRunioN Nfl NARRAIIOIS D u 1
ei
p sc tention and wha ii
ci
level rearrangi i 15
what
is iii our
focus of at
outside.
For cx mpie, sup o you are stLidying why students choose to go to iarticular colleges \ on gather material from college ss hooks promoron 1 in iterial Web sites, sod so on. You ,
. But on don’t 5 smdc studs ins’ inftrescs nd se cli parteoi bud m ucli Sm lents -em to apply to a strange variety of schools, i in x of four-i cai colleges and u iis•crsities, urban and rural, tamou 5 and not C irnous Moreox er, students seem to ins ly in ins i ci ftei once s between schools. How shout changing the sontext* (‘mild it he that applying to and cc isideri ig olkge ate rcalhc, at first, about staking out a po
Ce sjx
id
i >
tic
sition as a kind of person cit home t in high school or among a fri ndshi 1 p ouj ? I hat is, the in ntext of the lecision is not Si imhy tik St Lident in the abstract hot the student as someone ss o s t viiig o t dl his patent that he is sensible or her Inc mis that she is cfari up or Ii is school that he can run wi di the be t, ml so n. Fins c rnt xi nat ows and changes as decision time dmss s ieat and family economic and practical realities Be t e in ii ssu is on o ontsxt. We change our dunking about college apphieatiois he asking whether we have if r cht 0 e. u ti c p ohlcm A splendid example ot context changing is Atlie Iiochs ha pp 1 art Hot I sd lie s hook puts together ci hi 7 iii
‘
normally coiisieiercd cparaLe. e icu nions and work. Traditionally, studies of n ork treated emo 5 tional life as a conrrxi for understanding what goes on c1 th workplace. There had been a substantial litciature on the to formal organization of the workplace” and whether n helpeu Cc
two realms of investigation
at
(Jiaiiginp Co itCXt is a poss ertu] heuristic bemuse it brings to w h s e at ±Cdl s c x&rt or it rearranges the gi thet rh n. a cc wc )flncct sot nil things. I at i not thinking here so much if i h ides di ii the out yr letern ines what happens. (I’ll con soler Put next,) I am n ore concettied with simply rearranging rilillip
iI1uiusrics
hindered the organization. This literature saw friendship, pci sonal rivalry, and so on as a part of the Context of thC thinmi structure of the organization, but no one had thought about emotions as part of the foregiound, as part of work inseir. Hochschild’s decision to make emotions the foreground led bet to the concept of “emotton work,” work that involves changin mind (Inc’s own feelings in order to produce a “proper state in others” (1983:7). It also led her to a remarkable stuny of the lives and experiences of people who do such emotion work (flight attendants, bill collectors, and others), which rerrans one of the most interesting pieces of sociology of the last quat ter century. Bringing emotion to the foreground was a brilliaru idea. Changing context isa particularly powerful heuristic tool because contexts are usually established by largely conventional rules within disciplines and disciplinary subcommunities. In many ways, undergraduates are better positioned to change thc contexts of their problems than faculty members ace, because they don’t know the conventional contexts assigned ii tfe literature. It 15 always worthwhile to think about chai ging the context. Are there parts of your phenomenon that you a ‘e treating as background that could become foreground, or s ice versa?
Mr i in ms
i2
h
or
L’ihP1u’ a”i)
1)i os I RI
LriJ
social phenomenon—-work, say, or start thinking about it. i w s ia e i eve! at whi Ii r. T±k the example of tuies When sve ask what cities look k c—-h a he ix shaped, wh t kinds of people live where in xse think about
sonw
first ira ii nation is to think at the level of iTv dual city S we look at who doesn’t like whom and U r xs ho doesn’t win t ) Ps tiex t horn and who moves where, ss ir. i. an I is hi hr. c look at tfansportatioii structures, land val P
m I so in
cci,
o s, n Ii i
tit
-
iii
StE CS
jI
.n iglit
ad!
he that
t
ie Structure of cities is mainly
dc cermi ii d b son’ larger flier omenon, th natioHal or global t cc i, ldr example This is the theme of Saskia Jr r it
sassen s 1ij C/s/cd Cip, which holds that the structure of cer tain “primate i ‘ies” Thea Yr. rk I ondon Tbkyo—is deter n ii d It ir ia u c- as cent rs for producer se’ices (law, ix nking, insurancr. and othc r services businesses need) in the global tcotio n ‘I his centrality generates a de 1 ci 1 01 It 0 in kinds U emok vees, a ho in turn have certain C OUflt
‘,
P tids of incomes and tastes, which in turn generate a lot of
055 thu
in rnai Lets
at
d kinds
it
employment. That is, the
ptodur. cr sc’ivk es industry thri ses on c mcenrration, which in it in I tate wh cc p d leer-sc cx ices employees lise and what kinds of retail operations and setvkes must be locally available. 131 mp!aation, Sassen s argument (the dominance of the inter I
ii
i’
1
cv si n of labor) ould be extended to other types of
in a glob ii teoiiomv, Ii this argument, explaiucn
(JutS
a
i
in iy
he
i
hr.. ni
o
i
k me nistaku Tb
“lower-level’ phenomenon real phenomenon of interest
1 1 irger and cia lowc r-level one driven by the part
NAt
A1h5,
LdTieS
C )Uise I it plai s in th larger one. The same acgume it can Of ucive iPx made in reverse. In a famous article on “The Cuniui precie1 tore of Local Urban Culture,” Gerald SLctdes -argud time the opposite of Sassen. Ant city. lie said, acquires over Wa cain political habits acid rigidities. These will be in rthifly f ur unique, and they will overdetermire the fate or all Sorts o with ban change: political, cultural, even industrial. Chicag al rn its relatively cohesive elite closely tied to an aging politic mu v chine, is qU1LC different from multi-elite New Ydrk and -
i -u open and freexvheeling Los Angeles. To see a sflgle paetea d city poliocs is to look at too genetal a level, Thor. only Aou re one not see particular cities as determineci by global structu politics one should also riot oclieve in general patterns of city 0 but only in a genera! process (aging) thai produces unique p each city, What matters is uut that one or the other of these argu works ments is right or wrong hut rather that both of these further have become cecebrated and fruitful foundations for research studies of Lirban life. Both led to extensive bodies or levels. because both ins oke the important heuristic of changing t Perhaps the most extraordinary example of sucf contex “glob changing in recent social science—the grandfather of all seucf alization arguments—is Fernand Braudel s monumental of the The Aiedderranican, Braudel argued that the ‘events” hoc Mediterranean in the sixteenth century were just so much of dVefltS sam and jetsam on the surface of toe sea. The nature leve was dictated b rvhaL he called utiJuncture, a middle chan les historical reality that included flue tuauiorla in prices, pments ni uiaval practice ama in trade patterns, and develo war. But puss er, in types of governments, and in torins of
terils
ill
Yet
l•1
DESLKIIXILON i’o NAaRATIOI HEUjuSTICS
it ‘a )t
)ISCOVLR
beneath everytli ng lila a stead lug foundation, was “struc are, the nor ha iging and leteri hr my basis of Mediterranean i ii he ci vironnient—geography, seas, irc b g S r it boundaries, ii iatu..— but also i icluded foundational islam human prar tues ci e nat ire of towns, nornadism and “transhu re ‘ular long-distance rn gration and return), types of ii a “
ships a id other s ch tI ings F r Braudel structure was the m sr impo tacit about it or bLimidred pagcs worth), conjuncture âme ucond (about fise hundred pages, but then conjunctural di ng thai ge I, 50 tb y ro P mc e space) Events——the stuff of the las three hundred pages of n st Ii srr rics— take oul Biaudrl s two volumes. The crc ctara and conjunctural con texts krr rmcne them. B ru id s la k abounds n interesting heuristics. His u side-dov n map of Africa sh ‘a ng ‘how the great Sahara dcsert don mates the sea 1,1972 1: 169) is a spectacular exam ple of r sersal Bu his most ea raordinaR effect was to give tral g ieracot s of I v I-raising arguments, from the w rid-sysr rn r ieorv of the lfOs to the globalization theory of the i )90s, c\Il of these mesult from Braudel s radical chang ing of level, his insistencc that grand conjuncture above all
us
t
ii cs
s s
ic vent hch w.
the 1
(.Stthir C Oilt/I.
(011
1 1101J11fl),
a/I]
S/2ll!Ifl2g
cIa ions is a matter I’ deciding where a particular de a pt cii a plies P it anorhc r way, it is a matter of deciding ft spli some social phenomena apart or lump them ‘a ethr Sec hog
to
01
‘ethe T ins a iothiet way to cli oh of svhar Sasseri did in The Global o a t at tl e book cit ‘a a distinction between the ,
and Lois great producer-services cities—New York, lokyo, tion ‘was don—and all other urban places. Of course, the distinc Lei overdrawn, Many other cities partook of this or- that ch1ara ‘le istic of the global triumvirate. But precisely what made drawio book powerful and attractive heuristically was that d Sassen c such a tight line around the phenomenon allowe allowed dci write about an extreme version of it. This in turn depth tba to explore the phenomenon of globalization at a larger class o might not have been allowed had she analyzed a push an ar cities. Making a strong distinction allowed her to gument to the limit distcnctio One could, by contrast, choose nor ro make a pheriome but to lump things together as instances or a single this in social non. Among the most celebrated examples of s. Elias took science is Norbert Elias’s The Civilizing Proces manners, dozens of subjects that used to be separate—table on-—and as nose blowing, spitting, bedroom behavior, and so ” which ne sembled them into an image of private “civilization, ion of then even more audaciously connected to the format , cons i modern states. All of these things together, he argued l’s ‘structure tuted a grand “civilizing process.” Like Braude here rha Elias’s civilizing process was a huge conception. But ining idea was not Braudel’s of changing our idea of the determ we had level of a system but rather an argument that things g and the thought utterly separate—-the history of nose blowin one largc history of the absolute state—were in fact part of process. oLis 01 Again, there is flO need for the student to he so audaci together grandiose. But it is often a useful heuristic to lump such a things that others have left separate. Merely to propos
DESCRIPTION AND NARRATiON Ha RtSTc$ se )S I it
iV t i it us )i I
I (
not and whc thee w whether we view processes dynamically or
lum un to mwr us to raise a hundred untrue sting questions id
mr
sSiWs
inv stigatiun
isno the sameas spit Ii rs put ci it t h tower 1 w I is th determining tine. Rather, it s t iption applies os er ularitu u .tsS ti P a st r iunge thaii w had thought fhu has been the over for the last twenty ott n Is a Wo r t whc n n example ( yntloa l:psteins \Vomni iii Lau makes the C a
yr
H,
His
a
oar
ce
n
scn
v
trs’tr
already nany hue Its about lawyers, hat whiie them lawyers who 1 ilI pent rahzamions n th ni door apply ii 051 bee u hr c rd ‘r of the day in ‘plittng I s r a odds. ci hi.i a,td rcual studc, geud r studies, and so on. and has NulL liuss CVei’, hat it is quite gemicual heuristic Carlin’s research, o m ii p ohr ci t o dr with hi ii made precisely the same claim about ii 7 /x L
n
V
se
are
to
many
move
a
act
iii
s
ru
smal
,
ers—
ii. Nsiii Ucs
Jetome
solo practices— that they were quite different that Epstei made about women lawyers. othet laws
iawyels
from
vist
riptu c x
1 list
:5051
heuristics
I \ iriat
c e ts Pints’s
ic
RisIU
s
v
it
o
rivilia
s
prop(
se tIiarges to tile way reality is de
olve ehauigiiig the way we LiSt In this Sense, to think ib ut social life as it is process’ is as nuch a narrative lieu us
its
is
mose
ip
nuivibee of separate a ci st r (TI is u id rscoies an important C k d n ii t mattLr whit we ia/i tile WdS in s hiLh we Pu c nec ideas I ‘Fh mc are fi)ur ideas iu’t m ung as si
.
desk c
ptl\
c
Ofle
It
WCi’v es
it
new’
n
h 1
n mratioi
entrr
ur
n—the 1920s seem closer inspection—looked at in isolatio
pretations be extremely diverse. Conversely, many static inter Consider cond come quite different when seen dynamically. managers of tuons in high-tech industries today, The senior tess statically these companies view the situation more or and stock within the narrow time frame of quarterly returns experience their market value. But the workers themselves their longer, dynamic work within -
time
the
Depending to
freeze
or
on
our
research
interest,
we
frame
are
careers.
of
going
to
want
time
flow.
heuristic involves contingency Many so based or They models disregard cial about in sev results kinds the belief that the thc t specifically interested i i and that if eral well disregard the the pathways, we might details example that determine them. A people’s lives. A long literature
A second
narrative
are
same
we
can
of
come
aren
con
as
impor
‘
I
life indeed sometimes that’s the best way to understand social characteriza much of history works this way. Grand-narrative le, most histo tions come apart on close inspection. For examp jazz Age, bm uc ries of America speak of the l920s as the
ways
generate
i ant nartat or
seem crazy
contingencies.
point:
ieuristics to disCuss.
fas’
science
narratives
a
focus on contingency. that has been An obvious first move is to take something sely, to mIt viewed statically and put it unto motion or, conver it Static something that has been seen narratively and make who are usual, there is no great issue of faith here. To those one), it may cinated by the processual nature of social life J m ic But to treat freeze framing as a legitimate heurist
fhe first two involve the thinking about a problem:
ot’
rather
tingencies
of
this
comes
from
the
on
interesting
DiiSCR1PIiON
1 InIV ui I)is US ticS
11
Id that ncgati c life events— sickness, hm ICJS efrient c nempiovment. and o on—could lead to various lit W len several writers pro dis ess he S irpi c is it ma r age and so on otto Ii e tics it sit i seoul LI haS C t Iii Sanlu eftect, sonit thing that turns our to (iLl l us tI cot nymley— distress came only b no or I s tt h
alit
I
ur itd ou to be irrelevant. ega ye hc mctines conringenev matter hei th n we think. 5 centrally )o lie P at hand so ietii cs contingency i arc n Whit v ic chat model, mentioned ot i n i articr. Is an eXalilpie Ut a completely contingent model, at least he
t c
it
cc
presence ot such over m ans that we arc working of t C m cci I h g ics ci ut mali Cs. White s model is ultimately a ctt the ss rung vs hich clii large system has dominance over o ut vu
‘a’
ih
nic
ci
uJivil lOis
ireI s. Fl
uls es latent functions. Latent ar uiiplannc or largely UflflotiCe d results of social vhich, I wevc t nit out to be impor it t ii
A uhrd rarr,cct\ e it
heuristic i in
that these latent functions bein mu ceo portent ii ran acki uwledged functions ‘hen I ii n th (>1 V US I used an implicit exam C lu (it latent tunctions: th alternative purposes ot college.
Lmc
lfOiCCLc n mat be the
IL
1aeht
‘fe is
1
i
not
case
tot educatic it
)ur
i
C)
hut for reducing unemploy Cc nit of the full-time labor
the ostensible function, un accnt hnictn ir ibis particular example of 1 n rent JIL I c 1 i s kid by probk matizing the r ens lhu lUrk lion of an instilution and looking for latent 1 nu. There arc mant othur things to problematize about In ii or
education
1.
NnacieviiUs Haissice
college, of course: common beliefs about who goc. hem, v people actually do there all day (raculty iju Out teach alt d.i example) and so on. But looking for latent tutiCeiGlis I’ 01
Ia 0
a useful heuristic My final narrative IteutiStic is the coantcrracrjal. kiict Li have happened if, Some disciplines arc perucuhcrly vs it set for counterfactual analysis. Economics has a particular advantag here, because of its ability to “Impute” prices to ucipeiced rhinn .
.
to havL by estimating the costs of the ocher things people forego r the unpriced ones. But counterfactuals are aho widely used history. For example, thit implicit counrerfactual itt A. J P. T lors Oi’igins ojthc Second Udiuld \l’di’ (dtscussed in Chapter Unef that if Hitler had not invaded the Soviet Union and gratuitously gotten aua declared war on the United States, he might have ent cu with most of his gains up to that point. The counterargum -
that, however, is that he got the earlier gains by making precisely Someone those kinds of audacious moves, but on a smaller scale. earlier bold svho knew when to stop could neer have made the his “mis take moves that got him to the point where he made Taylut s tuesis Thus we see that the core of the argument anour political system lies in the nature of Hitler’s personality’ and the in German that allowed his personality such comprehensive sway is’here tile argopolicy. By thinking counterfactually, we see merits hinges are.
more d Let us now examine these narrative gambits with tailed examples. A. Stopping and Puttin,ç in Motion If your pies The first narrative heuristic involves history itself. it d nanhicl it ent analytic strategy is static, how about making
DOiUtOuN
static As with so many heuriatics the question isn’t whc thor the social world is in fact I istot ic il or not Y iu don t have to be Max Weher to know the
it’s
dy’n.rmit, iiow about making
it
n v r o r a c it t it Bt t or ir imcs it’s useful to attcnd to that history rnd somerilies it isni. The more tamiliar move is 1mm tatlc to dynamic. Wlien fixed, a slice of (it r W( mu r to a ness town or school, it seems ye irs do w know which ta ing hi i c ( i IC
rNi)
NAk(Rrc15O HiR5LS
I I
I
a ociety itt t se out of northeri Burma, visiting nearly every in eacniy utiob, aflh lie Most of Iìis field liOteS SCCEC destroyed y, his tewS eLtA it my wrote his reat book nom memor and what published materials he could bud, impi ed ii tile Leach s central point was that the stability In his fiaai ctiSit’ classic community studies svas a mirage.
cally blunt
prose, he
The generation
wrote:
of Brililli
amlsropuioisLr
UI
A
iich I
belief in th rrcievaii ef lihr;. organ aiion. QI’iat 1. eall for the understanding of social is irrelevar meant by these arguments is not that history bus PioresIct that it is too difficult to pat 00 prpea rr pho1dcrs who is one of the molt stucc
parts arc hi urging and which stable, Theories we adopted at tirst seem silly o ice we understa id that this or that part of the slice was ni fmr in rapid motion, 1 hus you might wonder why up in a new location. 0 St nra ave t’ st i i i I Yr i m ghr It velop a story about problems between the store
has
and the mali, pr blenis Wit hi comper tors aid so on. But then store’s new loca 1 a new des elopment spi mgs up around the i ealizt bar ‘n fact your flivorite store I yr u s ay s ddcnl 1-ad bcn it arc 1 n the riginal ma] 1 only temporarily, while be static ts new quarters were bring built. What seemed to turns out t has e oceil in motion, hut because you first got to
is also in equilibrium analysis in British social anthropology analysis. f ci ads ocatc of the use or’ history in anthropological hrwe tie 1 h has-’IiOt yet explaused host rh 1,ICufl5jlEciILt5 two positions can be resolved (I 61:282—i’
ts temporary location, you didn’t see that. the c ntral ditficul its of assessing any social situa Urn ion at .i. sinyk moment is precisely out inability to see the snapshot merely as part of a moi ie reel. This point is made the n ost influential works of sub u w ri ig a c r cv i i one by i hiop dogy: PPz ica/ Syurmi oJ Highland Burma ii d r Edmund Leach. 1,eachr set out to do ‘a functionalist study of a r it
hr s w
i
ssu ethnographic slice of life. He ‘ a the 1 ii hs intc the work when the Second World i w u tone Shortly afterward, Leach flu ii a ‘n War itt r the next five years drifting in and ed the army and
o’irmun’tx was onli a Ic v r i
sIngle
“
proudly proclaimed its
Evaiis—PrircliarJ,
u
Otters the best mm e Leach was right about “history,’ of course. long-rein change possible is to put Ones data in motton, to see made that mo rather than simple equilibrium. But having He prese s—U Leach himself made a quite peculiar reverse symbolic SVSLCITiS Cr equilibrium by saying chat the ritual sso soJeries ir burn a. the Kachin act as rj thete were oqu1ibrLins stabilit nut u r Their cultural system draws on a language of thing. fne arithi it that language to do “historical, changing tribes symbols for pologists’ mistake, then, lies in taking the rh fliuVe into liiithe reality. Curiously’ enough, Leach made (This wading into the yr re tory and then took it hack again. he corn non arso ‘a and then hurrying back ft shore seems to
1 1)
fltt)i spoloists. Marshal flukes nt
ml
s
dc
the he-u tiStC
an
es’eLts isas
tI
i’,C
it
f Histrirj
is he he irudc. Leach Up Vi ith the issue of
really there,
‘
hether
it
did or
r equilibrium. in some ways, a t off by trying to keep
is
art
Ia
I
C
1
‘SeC
mi
it
i.
rIte dow or
Iri lliatt(’
boil
I Sahi ifl s induential [Lnds
Ii the saffle ino e.) vi v wh
dii
is bather
DCSCRIPTLON AND INAI{IiAYLON HuiRisTics
)5i1i
wha -r il imio utility of invoking
Stan as
es
well
was testify ing
—
as
to
ignoring the passage
ed
r
hut ii usr
wit
stop
tI s motion
put our problem that is already there.
i o
I\ pilli s a vatit to do rhi is hen our interpretation of some en m we L y the narrative in is h r p c mi part Lu s Id 1 t ti It w th igs s e actually know. i ewelleni e\ainple 01 iii s is th magniticent historical ihioTji//,it; Tin Pmnmi.J L.in/ t Error by Em et Om aripns wu
I
hi
Ilij
idur
y ma
r
ti
l hr In
ourse of modern 01 S I ither France had been the last representatives of the h
itt
thit
pcrevcd jmi tly in thi i r role as heies They are noticed in history Path rn no I iat
beliefs (the highest Ca— ii SOS, I a p t ccii I isted ompletely —until death -—--attat thur ceremony of “heretication”) and for the
ft
resul lad
i&u t
mcii
diem Be provoking these cru itral role in establishing the 1 [eyed a r’ s r in arini duo, rh southern lie most pan or is nit is now Fi aiu u. That is the usual story of the nut.
tel
I
economy and residence, about ramily and marriage. AFOUL sheep and migration Suddenly, Catharism hsnmes 1iu atthing strange and perplexing but something deeply consul hensible in the context of the culture at the time. Hisrori ii us becomes ethnography in this book; long-dead nistorical re-cowls rise to a living, daily culture. rhis practice of stopping the clock s
give
tir S
0)
But the inquisition that rooted out th he:es kept Je-tail notes. And Le Roy LaIrie realized that oiie could re-an ttie 0)quisitorial tecords nor so much as evidence about ( thai u ocr se but a, evidence about the community as e whole. b ii
iiss ft s th
p. a
‘its i
t soppressu
[south -rti I
Cl 1(0.
an iniportant
uc, cnn
of historical Inquiry. ihe maya of stopping the clock IS central to all forms of equilibnoir. analysis. Thus in many branches of economics, a marker may he far from stable at any moment, but by analyzing its behavioi n it is important
not only in
areas
equilibrium, Uric can specify the direction of the forces plaI1tng upon it. So. too, in certain forms of game theory. Eve-n ex tended games—games that take place as repeated mlays o ur time—can sometimes be reduced to a strategic form in which the answer is given at once, no matter how ifle repeated pun s might get there in practice. Stopping the clock essentially enables you to attend to inure things in the present, It allows you to broaden the context possibly to change levels. That is, it can be a gatesimy—uke- so many of these moves—to other heuristic moses. (e one-n think reality is fundamentally historical. But it is still useful to imagine it, from time to Lime, as frozen for a moment. I best can be big moments to be sure. When Braudel is justify irig ins avnept
of scruc’curc-’
112
J/iu
It/Jf.a’a,ili.ii;J,
he write’s
ur
aria
‘iti moos Ot
I
Discos FOS
Hi t thtsc’ ts o I un rtd yeats, I t50—1650, should demands I in i a ol ci ía unit, at least in s me respects, clearly some plaiiarion (1 972:2 89) I lo can two centuries be a 1 pretending that they are, Veil, they call t, hut h i tumcnt c III 9 ii )U Sc e to sonit in porta it insights
an I carlnh ( ft lfi C ( ontinycili y also produces an important heuristic. One can cii ratc fliany ness s icis s of a theory ot a regularity by arguing tInt it ‘s ‘ontinge it on so se hi g Conversely, one can some 1 disregarding contin in s sri cluce cx raord’ni my results b id ed one I me standard moves made in ency UI ian r i 4 13, La
fr1
,
formal and quao to atis e work. As ‘in oxan pit of taking cc 1itinyenc seriously, consider licha I Pio ‘c and Charles Sabcl s Iryrument in The Second Indui m,l I’ ii i th t here as nothing foreordained about mass ro lii ti i. A ca rdit g o their irgurnc’nt, modern economic grow th might have been sustained by small, flexible produc— I here was no absolute need for assembly lines and iteichat ye hi parts. Pior and Sabel’s controversial argument has i cured at e iormc us mass of research on those areas of tht v ri ( ot liv estein (set tie iy and northern Italy, for cx— mm il ss he e complex webs of flexible producers have indeed
tiomi u
IirS,
ccl. A numbt t 01 interesting consequences followed from hc b oP I cast, the book suggested investigating the web ik s ihc inra tor strut tures educational systems, and credit idustia areas of “flexible 4,”ir, the I’m It n ci cO s spc t ia izatc n. ‘I hat is, time I ook had direct consequences
mmvii
old for industrial policy Sccond, it suggested te’thinking the tif Was the role of artisanal nam rats e 01 in iustriaio’atiom
HFaJRUS 1C5 DeSCRIPTION AND NARRKHON
made to appeat? What wetc labor as tangential as it had been trial divide’ for the labor move the consequences of the “indus fication’ created rnoru powcr ment? On the one hand, “massi created conditions that mmdc ful employers. On the other, it the history of moder WutIt union recruitment easier. Suddenly, looked ditfe rent. ents implying chat pechaps By contrast, there are also argum an even we think. Making contingency isn’t as important as Normal Accidents, Charles Pein stronger argument in his hook work out a relativel systemarc row suggested that one could , such as nuclear-plant ac theory for rare and contingent events like. The book opens wltn cideiits, ship collisions, and the t of the Three Mile Island thrilling, utterly contingent accoun what kinds of factors allow accident in 1979. Perrow then asks probability—to cioffin contingency—in the sense of random two. The first factot is cl c nate systems. He comes up with systems have lots of feedback complexity of a system; complex multiple functions, possibly in loops and lots of parts serving The second factor is the cu ambiguous or unmeasurable ways. d systems are strongly’ time pling of the system; tightly couple sequences of action in thm dependent, svith many invariant successfully operating. Perrow’s and, typically, only one way of nts—that is, “sysrernatmcalbi basic theory is that normal accide most common in comple produced” contingent events—are coupled systems. He thereby systems that are also tightly contingency. achiees something of a theory of is always important in Considering the role of contingency heuristic moves of either invok thinking about social life. The out contingency can of temi ing more contingency or ruling Suppose you are writing a burst open an intractable problem.
kuSS1L DCSCRLPTUS aD iARRAIiua t1.t
I
S I. L
ii
Li.
it
1 3 U pr fessiot like pharmacy, radiogra pi ii n ci 1 and it looks as pin aiRS i ursini. You ead books or each one ft fcssional taking Ut ic I iC ri y e 1 t II ig mni c ru turns A tn sune time, they seem to vi m ire 1 it. w di ( thu para 1 ofessions or I ti h ti h s iso v I i tL mcdii nc itU 11. 1 hat seems a ratllcr tiaccid, dull design 1 w
pu es ionali. n o i e risc hu there is lots of u a i qx L thin a way to regard all of thsc little fights SqUi[)i hug iot as icr I olei t I u t is s srcma ii .By ic\ up rhc competi held you cn see them as con or’ all at Once in a rompeti us vacancy chains, rescni a lmirid set of resources As iii Ql3ite s mtingency by seci p it as the outcome of a n o n strac tin ed s stem of on1petiLion ‘iou base mos ed to a new icd yin dl IL ZC en crc ia of ompetition uithin which liv I h ui dcrswod ss stemarically,
se fights c C U i
it
,cl
in
il
)fl
at U
t
Iii
lat
or
cc
ia
cc
cc into a
I gone out of fashion
SCV
half ccnturt Functional arguments are Ii y s ml to hc i cnply elab rated versions i e i inal-c ioice aryumcnts function equals purpose equals Somc t IUS ti cy are purely logical i e in v u fl
i tIe
v
lath i1uii
I in uL
fl.
lid!).
irc as ii
ss nih, ss c a[ 11 1
nit L
U
the
ate
ci
oust
au
st
classu
‘irnpetati e lunction
ii
tnins
I
ui)
ri
alsvavs
it
Rust
argument
that
wcui for a society to
identify the social
structures
‘P1 ci tSom ti ies such arguments arc sImols lariunaluzations for moral arguments about how so t Ia (iigaulizt cic c The dci )aa user i Unc 10 (Inal isfl 15 deep and complex, but I itini iiiiei
t 10
ii
I)
e ViAL
1111 S
csted in functional auguinc’ its merely as heuristics. Of-
e and nacerug ten we look at a social institution or struccu abs inns pupoic o theory of it based on what seem to us the do mar tflcic src functions that t serves. But it may well be Oi OtI1\ i5 k. forces keeping it in place richer putpaSiet LILCSC mc ices V flcctrng on latent functions can take us to we can then analyze as we
Set
fit.
s is Kid ard E An example of latent functional analys its fairly strong po wards’s book Contested Teriain. Divested of with a distinct p0 nt litical overlay (Edwards was a radical that the usual hiscor ci view), the book’s basic argument is got h all wr my employment relations in the United Sratm efficienct filu\’efOcflt The traditional argument was that an rh ear y yet rs of th dominated American labor relations for management, WItflILS twentieth cenwry This was “scientific on the piece mates, stringent work rods and so StOP watches, n in the film disc/na world of work skesvered by’ Charlie Chapli scientific manapemer t limes. On the traditional interpretation ion; rh attempt in was driven by the engineering profess grown out of rationalizad ronahze labor on the shop floor had management was then production itself. In this story, scientific war by thc ‘human rc a replaced in the 1930s arid after the much broader focus ufl dons” school of management, with its capitalism, and similar workers Ides and happiness, welfare
policies although nbc humar What Edwards pointed out sas that , gentler form of relations school looked like a kinder ous expansIon Of management, in fact a concealed an enorm vas fly extended limes bureaucratic rules and regulations that —quite persuasiVely control over workers’ lives. He argued of n’eanagement was t that the teal purpose of both schools
Dssc IPTION
I)cocFin
I lie sc nate, ostensible functions may I vc e ii p reilciet v ad aking care of workers,” respec th dy But tile lrent tunet ion w is the same in both cases: contn 1 or chc. s urkers cur 001 purposes, whether that control was t d sc iplint (I
C
abo I rc c
is that the seacch for latent r 1 till )t Cr t it is u niport in social tortes. I have several tiIie mentioned the latent fUnctiofl of college in safeguarding emplOyment opportunities tor the larger labor force. This may see n an cid nay to think but in the years when American a
I
t
c
ml. lit pci
t
rkers organized labor indus r al w s c r i ate I St’(ii ‘ly O f o d d iy attempt to crc ate the kind of combined cippri ntiieship-scliooluig system that trains much of the labor fofte In (rermanv. Such a system would in fact, have threat )bs Colleges unction of controlling unem o n n ci c I 1 ib
b nc c iripotta t bin you tE ink So latent fu ctiuncl inalysis S always a useful strategy. You ma\ Eu rn up nor In ny at a! 1. But you may turn up important
1 yn
t
lay
tiungs indeed.
r
i
ill
F m Ny. 1 xc ish to consider countertactuals One of the most userul narratue heuristics is i bat ij We are used to practicing ci is in our own lix es, as in II 1 hadn’t gone hiking in Europe mc dii hax c met and gotten married.” Fwm I c s ix met v c i I .c r P p ijit of vcv f coursc, there are hundreds, thousands nia) be men tells ol thousands, or people to whom one oul :1 have been successfully married, Our lives might have differed in anious ways, but most of them wouldn’t have been xci cc r u n i 1 in cci ms, It ‘s say of ur ultimate type of
AND NARRcEiO HIuiusTic
aneial situation, 1 employment. our hr
UI
the socoeconociQ
ta
yecieral itaeresc in in tus of our children. There is thus little . aitlic ugh th c vesrigating counterfactual might have bens met may be considerable personal inteiest in impiur. vicaily Often, however, counterfacruars are m melmaiii xeaL 5 \XJould there have been something like a fksc heroine the major n out Adult Hitler? Would Chicago have had become a rnajoi tail hub of the Midwest if St. Louis dif& rent ii Would American history look fundamencaily noticed that cln watchman in the Watergate complex haetnc latch was taped on a basement door in 19PP uctive. We often Posing counrerfactuals can be very prod g our case agains do it merely for the purpose of improvin what aid h p them—that is, to improve the argument for s and i%7Os’— pen. But sometimes—particularly in the 196u rately developed counter-factual analysis nas been an elabo mode of analysis. ctual heurist One of the most brilliant uses of counterfa mi: 6cou,in, Fewas Robert Fogel’s Railroads and zIeieri ca/, Em;zu railroaus were eirnn gel problematized the obvious “fact” that ld have liappene. to American economic growth. What wou ad? Obviously. thor he wondered, if there had been no railro of the actual would have been a lot of canals. But what to a considerable nomic consequences? As Fogel pointed out, ut; () percent ni extent the railroads were git’e;i their role flat-o and state gox cit their total capitalization came from federal merits
as
gifts.
Indeed, Fogel s
mtroductoty
chapter is kil I
liar facts and 51fl’ with such information, a brilliant use of fami t of us accept as a pie economic theory to demolish what mos t r u is rn.
DESCRIP HON
I
igu
lu I t C
1
log
cc ncludes that “no single Ic euom yrowrh during the ninu HO i m id 1 ak n overwhelmii. die i al 0 the economy” e lies w h it di I pages contain i it ii Ii ii i pi nary markets hi at d us oi P e American iron P nu d a r hr t I wit, and so on. Ion A i 1 i p u i ot only something it
etc
v
s
r
r o 1
r s 1
i
1
e i
1 C i
it
a c s if
a e
1
t
t
1 a a
i
(
i
va I
e
I
y
)
i
ill
v
r
I C
jx
bu t
I
u
iore important, teaching
r an
a
r
ceived and accepted
av thi te ity and energy fogel has iIC lisp a eve his .areer, But coun i alv iys usefu Reti rn to the example of lit Liii ippy wth any one of hun t k y n ders anding marriage and 1 -l ta lcd ly ii in of dating and house— u i the aige b it icr that shape ‘pools” of inc ogeth r barriers like college atten ieo u w an put the issue of partner s veryt ne knows the number of plausi 1 tble n thi ma I et declines rapidly at cer i out for t ample at the end of one’s xx 1 y us 11 s makes us think about an hoi w uk p Ic o about finding and Ii 1 ii ig as college 1 (That t e lat one m 11 facilitating the mar t r r t err i ura1 experiment of sorts c i ih i p pulatuon doesn’t go to
4NL) NARRATiON
I Lu ins tic.
the population tends to college (And note that that portion of marry earlier!) spins us out. via Thus, starting with a simple counteriactual unei leaps, into a wide variety of a number of other heuristic marriage p4ttctlls TIn esting hypotheses about marriage and It drives us LO prob is the utility of counterfaetual analysis
of new ways in whiel lematize the obvious and suggests dozens to think.
us to the end of this survey of general preceding one hay heuristic gambits. This chapter and the producing new ideas. I discussed a wide variety of ways of will again—that these should underscore—as 1 have before and not about the true heuristics should not be reified. They are ideas. They should he and the untrue but about finding new things. taken as aids to reflection, not as fixed many of these exam They are also very powerful. Although heuristic move, you should ples have involved more than one work out the results o use them one at a time and carefully you iHto deep water very each one. Otherwise, they can get
S BRING 1 COCNTERFACICA
quickly. final 01 even the inos These general heuristics are not the science. That honor powerful set of heuristic tools for social on the basic debates of goes to the fractal heuristics founded Chapter Two. 1 now turn to them.
FRACThL Haciusrics
Chapter Sax F1 ‘iti.u llEtRIsrrIcs
I Pu
ii, Ax.sc 111, hi
I\
ININS ID
‘NI)
N
VII (
EL t°l Ii VNF
taR STION
SlissI AND
SlistI AND
1 icaui.x FISM L
CONS CRL(1IO\ISM
)N’ lxi A1 155
V
Nl) NONCON[LXTDALISM
101 1. so CoNs i RAINI
(ii ( I AM) (ONSENs1
\\S lADEN F AM) I It 511 D KNOLLiit F
IX. FR
\X/
i
IN
I t 1)111551 AND (N Lit R Si lSi\l
V R’
\ 111
DsL)
,1\r,’v
ii
i
ol heuristics. The ciddjtje rules or creating minor variations iii
o l.ti
simpl(t are
LEN rhrce general types
he sceond are lists ii p iwric topics and common noide u ns. nuns that i e cai Ise as stimuli t) point us in new directio e— The third -i hr general hr titistu S ot Chapters Four and Fiv by ma are more self -consuor s des ices tot producing ness ideas ticular pr far op arguments ltsctipt ons, and narratises in par ay
ths (tap er, 1 t sfe up a lout th type of heuristic, one that teat debates ut uh social sciences that I dis .u iscs ifl tilL u ( hapter ‘Iwo Jr maC use of a particular quality of ussed 1 chap thest dchates, one th it I r orid btiefly at the end of that mething that tht it tru tal narui e A tr ic ml is simply Sc it r
IC.
get to it. A famou irc looks the same no matter how close we whose fronds is a little ferc tal is the woodland fern, each of e ferns made up 01 tiniti made of leaves that are actually littl
ferns, and so on. pter T o are fractals ii The great debates 1 discussed in Cha ortant debates no matter the sense that they seem to be imp ch we rake them up. Takr what the level of investigation at whi constructionists. Realists the famous opposition of realists and and repeatable. Con think social reality is a real thing, fixed k the actors and mean s structionists don’t. Consrrucrionist thin along, by playing witis ings of social life are made up as we go past repertoires. Realists don’t. tty clear idea of who thc Now most sociologists have a pre nists are. Survey analysts arc realists are and who the constructio historical sociologists to be usually thought to be realists and olars are usually realists; soci constructionists. Stratification sch nists-—and so on. But suppose ologists of science are constructio nce and isolate them somehos. we take some sociologists of scie argue internally over precisely Sure enough, they will start to that structionism. Some will argue this issue of realism and con d dge produced by a certain kin science is a given type of knowle is issue is how that knowledge of social structure; the big you es. Others will argue that shaped by larger social structur itself is until you understand cannot understand what science build language that scientists use to the actual flow of the daily gets rationalized in textbooks. the scientific knowledge that over will fall into violent debates That is, the two groups even ism versus constructionism precisely the issue of real strong con ipline regards them all as though the rest of the disc ed in the fact, exactly what happen structionists. (This is, in
Mr ii
)LV, )i
15 [1 V
1)is
FRACTAL HJ1GR1STICS
when the field had a kind of I think more things are socialh onstructed than you do” con r thc L n led ip s ith the Wh( I field retending, somewhat inn ooslv. niar it difin believe in the reality of anything at
tionist questions were asked: is a series of harassing acm ni event or many? Do closed-form survey questions necessariI Cu erce respondents to follow a certain pattern? When is a que tion to be considered “suggestive”? These were all 1. ft sail debates that the sociologists of science were at leave i i air
Jo rake an aampit on rhe or het side of the discipline, the soot logy of tn iw was tor a I ny mc one f the strong realist holds ni sociolog. Ci tine statistics have a long history in
1980s, but they’ were located in a community chat cite disc pline widely regarded as realist. As this example shows, the central social scientific debacn of Chapter Two are fractal in nature. No matter how large u how small the community of social scientists we consider, must of these issues will be debated within it, even ii we chink tILa, the community already represents one extreme or the Other on
uc cig
Ann rica
ot
so
pobl
oct
in the I
life, and
1)5()
emed inure obviously n il titan at attest But ni the I 950s, there emerged within ri is hilil 5 r ‘al sc breracur a oii rrucrionjst critique. This la aJing riieo,v’ trjed tli tifc Was something more to be coming c mi nJ titan simply doing the act; you had to i
few (Vents n
letaincd, held, charLed, convicted, and sentenced. Many people slinped away at each step along the way, yet only a ii 0 lid vo i re ill hecoi e a labeled criminal. The labelers insisted rh it rite long-observed inverse correlation be t C IL1L ii,
ia si tus at d 1111 nalit happened because lower claSs tttierirlers were mtire likely to make it through the long p occ s t ice cc s f om art to oiiviction Criminality was not a simple. teal flier hut a complex, constructed one. t
er
s(
In e was als
sintilat but smaller debate within the f1///) na/is! group of criminologists. These realists ft o ii P ar ecaus of the unreliability of arrest statis tics Clinago’s crime rate rose 53 percent in one year (1962), canwhi e,
I
k
ii
h d ot changed hut reporting piocedrires had. So a vo ifcrousiotip argued that arrest statis i s in it ily cc st t t cite1 d cii inc shotild be measured h mire s (ii Vl( rim5, not by counts ot offenders. And in set
a
r i p
or
n
n vi
w ha
it
in iza umi sri eys, loz.cns of realist 1 construc-
-
the issue By itself, that is just a curious fact. But this cu irus fact means that we can use the basic debates as heuristic too Wherever we find ourselves with respect to the compiev arrangement of forms of knowledge that is social science, xc can always use these fractal heuristics to produce new questim and new problems. re on A simple example of this comes from the literatu anxiety’ and stress. How are we to explain stress? Who suffers investi most? ‘‘hat can stop it or mediate it? The literature w gating these questions from the 1960s through the 1980s an outsider strongly positivist. But what is most noticeable to positivis. reading the stress literature is that whenever the develop searchers came up against a blank wall, they would open new re narratives and reinterpretations of data that would re lookm search vistas for them. Thus, the original literatu events) and only at the correlation between stressors (unusual proved to Dc distress (unhappiness). When those correlations defined a weak, researchers scatted to think about “coping,”
ii
sti
I
L
n non on
it
1
)
(ISA
the pad (that is it the narrative)
distress, Diffbrem es in coping skills and re ount for hit weak simpl orrelation between listress: beitec copcrs would suffer less from a t
umber of stressors than weaker copecs would When C( ping variables proved weak in their turn, analysts started asking even more subtle inrerpreti e questions, such as
jajt ssflat point does heavy drinking change from a coping
into a symptom?
strategy
(Kessler, Price, and Wortman
l95 5nd. Now the answer to this last question has been the sLibjeCt
7
of numerous famous novels (for example, Fitzgerald’s
Btaah/u/
is
DamizJ) arid filnis (fur example, Tik Daji
iua’
of
XV/,n jiji Rued, which show well that there is nothing like an objectu e answer to it. But JUSt thinking about it gave the strc5S positivists something new to do. They weren’t stuck any longer in their cuhde-sac with the lousy correlations. They had ness questions to investigate. That is ss liar I mean
hen 1 say that the main importance of
ne Ira tal debates may not be as organizing principles of the
hut rather as heuristics for the disciplines. Indeed, I r ropose that the great debates had their first exis—
o
iou
tt
tat
v ew
w
Is
at
n
s and bet ame gei teral, organizing principles
hole disciplines and methods only because so e it, believing so nany substantive things, cs. On this a gurnent it is their wide— that lea heorists of various disci— ill )( ally d : osas into what appear
i
i
1
i
I
I
i i
v c
ii
1
1St i
gut
i
i
taL
ELti’S
sCOiLR\
g I b tes It e I m just pulling one of ks everst. Ii direction of causation, and still credil I
I n n it sure whether this
die lilac histocnJ its an interestitig pssik day in the
arguru
i t
holds
arid
uus is
not
.saluate n-UCi
.lc5
(
cii Ocl
scielicC,
In
sunilna.
debases or I Lhc picas central
sji.d wit
mt
OSc’Uc I widcl (if i nplicidy) used iU a fleurStmC 1 In ts glint. tl alt open up new questions and possibilities. of the uthurs 1 haS u C mi11 as common a heuristjc as any ca etuicO. wo othur hiticS, Lne can he greacis
themselves
Like
uken Cu bt ocher heuristics. they dtould not be p:obn nature o things. çTnas has been the
the with
ufu
ti
a.
itcaC.i
USCIU 55 it flu them as great debates. But treated well, fle ar whute Lti,L1mO, part of your heuristic armamentalium. yuud to the nine basic de 1 shall organize this chapter according nebate. I silali cisc bates discussed in Chapter Two. For eaLh one an be used few examples. trying o shoss how each flat die curfeut deflhl matter what the method, no matter ss examples fur horn hnmces tions or the research. 1 do not give traciitoflS. That wuuid h for each of my five methodological methods x 9 debates, aiw CuLl 90 exampleS (2 choiceS x 5 than I viart cc pci iookifl don’t want to read them all any more enough pusscbiiinies a. giv for them all But I shall try to offer heuristics And I shall ciii you a sense of the tichness of ftactal the grain: deep nterpretiViw phasize moves that went against who cry our ndividuahsri who turn positivist. emergentiSts to seltct papers that have 1 ac and so on. As bef( re, 1 have tried social science. aithougl S a stiong influence on subsequent av lo favorite ecenc work some cases, I’ve been seduced by bewildeling liversity of me advance fdr the almost
gize in
cc tsed
fractal heuristics but that is part of the point; v a ewtldeting vat Lety of throughout the social sciences in
amples,
i
OS
duat
1 fu
III )L) 01
11 \‘IN\i AN 0
s
C
lu
IRACTAL HEUiiS1CS
Dt.( lvi 05
1\iERPREI\ Isal tat i bate is b
wee i
posit vism
and inter
I sho ik m asure social reality tiarmafiv and thinking ou can’t and shouldn’t. in fact, isv i find vansples of posirvist and interpretive moves s c
n
iv
tI
ir
kin
y in
a i
s C I h gi ci t ‘ad t or I ie two no engaged in y an icsa t dial nue So in etllno5raph\, sometimes our irn hyre counted bowling pLilse is tO 0Unt Lhings tds cilliam F. T in omen ne our impulse is to S t A H U lois into von son interpretis detail (as M itchc 11 Duneier does .n disc ussing police busts in iaku jfk). In SCA-type analy in ‘is i sand positivism ate too numerous to count, is, s the other way as I just i at tic ii t. ii iv outed in my disLussion of the stress literature. A partu LLiaily elegant example ul an interpretivist move in I Saral Fenstermaker Berk’s ha d huh I. s intlu nital article tin models for thi household division of labor I 1d8c hand and Beth are attempting to evaluate the “new” the household as ‘a produc— s ith ts booty 1011 c c 1 iii employ an ext c ssely elaborate positivist de Sit i F 101 sign: a two-stage least squares operationalization of a structural
qu don model rn a Pita set on the allocation of household hc t i I ends up i an interpretive discussion B is count th ddnitiois of “sharing’ atid substitution” in house hold tasks \oting the complex differences between hus s tlic ousehold division of labor v i. c I c s a bha ges i wli h ta. ks the xvii doe affi.’ct which tasks the hisband does but not SlOe versa>, the authots point out that hus th other family members in 1 ito jot sily bai I cci to a ci
respondents ii production, and they provide quotations from “metal tUP lustraring three different models for this “sharing’: definiiciuiss of port,” “assistance,” antI ‘supeiwised help.” Thes subsumi on o if sharing have different implications for the Ion rise projecL husbands’ effort for the u ives’ effort and hence Beck ‘gnu Boric of analyzing the famil as a production ssmm. of the exa c leave the reader wondering about the question In short f trade-off between husbands’ and wives’ housework. are run, he ter all the rules are followed and all the regressions form of relntOr way out of a quantitative dilemma takes the story preting a variable by anchoring it in a more complicated blind alley with more ambiguous meanings. Thus, a positivist
is
escaped via an interpretive move.
in a It is equally important to note moves coward positivism One ex place like historical analysis, where we least expect it. ample is the paper by V. 0. Key on critical elections (I95) in one of the single most influential papers in political science the the twentieth century. Key’s paper removes elections from one-by-one tell-a-story approach that had been common before his time. By analyzing detailed counts of votes in particulan constituencies over many national elections in a row, he showed that in certain elections there were sudden realignments tflat
then persisted for three or four elections thereafter. Key’s move might be seen as a form of temporal lumping; his argument was that the event” of realignment was often bigger and rnorc enduring than it seemed. But it is important to see that Key made his claim stand by taking a distinctly positivist turn ii a literature that was until then given mostly to historical, diacuc sive analysis. It was by getting analytic that Key made his mark.
1i iHor,s oc
tiniIo, torn Li n vie recent move OCCUrS in John Mohr’s hr Iliant analysis ot images of mvcrtv it n neteenth-cencury w Yorl (\.[ ii arid ugu one .99), Mohr sought to un “ovi r ew Yorkers images md c m epts of poor and needy leoplc. Rachet ida i thilow rh sormal srrateg of critical an tIySi of tcxt a iout )()Vtfts, lit kd hc o h ial descriptions of H ci s n lozetis or •New Yirk St ual service agcncies into crimp iter. -ft then a ials zed di nc olin ia[ descriptions by say ins. that two types of need) people were ‘close, in the eyes of f th y appear d in the sam descriptions to N w Y rhe gethet. ()ricc lie had cak ulared the closeness’ of all possible par of cv ‘e he void use Lli’Stt ritig and scaling methods that
a
,
n e’ ci rca into It srers and pkt ures. As a result, he product d an tsroundingiy cur prehensive picture of poverty is it was ens isioncj by the er a encies that dealt with it. The xc s as a a it tH positivi. t en, but it revealed aspects of o
u ‘Ii
ci su
nineteentl -cc ntury theories ot pus crty that had nes er yet been :onceiv d Ii short,
hnd that i In posit is hr ‘mnrerpi erA is choice can be mad b) tity kind of analyst at any point Often, as in the uses m ntioned herr, the moves have their most decisive effect Vt
when thc) go against the expected direr hon. 1 e don’t expect o intet nit ‘C in wt n SCA and we don r expect positivist niovcs iii historical and cultural analysis Therefore, the results e mon dazzling. in tact, either move S ml nbA ii a y n thod a any point. And at toy eVel! One can Can ly envision moving on to the next 1ev 1 if let au and en ik ny e th r novc with respect to any
of such Inns es
I lieu could
I
see in
dl
ti
p es aLe ni t al elections. it is clear that one fir mor posicis ist than Key in evaluating the tres at
£RttCThL hc hiSTaS
I)tst en c
arid indec Lion of whether critical elections really do exist. just mar. One there is a large literature since Key that has done election resaics can also imagine making the move to counting did, but thi and to examining longer runs of elections, as Ke of the results insisting on a more interpretive form of analysis original papet Key’s own rendition of the phenomenon, in the ies the phenome is completely demographic. He simply identif t to lntelprt. non in the voting patterns but makes no attemp of new party organ it. Was it a result of new party ideologies? izationi of legal changes in registration? Was it a downstream a new result of the changed immigration laws of 1.924? of soro voting coalition in subgroups? of a change of heart by iately major subgroup? The possibilities are many and immed did encourage a large interpretive, historical literature, which had ur in fact emerge to try to explain the phenomenon Key covered. pre In short, nor only is the pairing of positivist with inrer it aism twist a heuristic pairing useful across all methods, hut trul) a applies at any level in those methods. This pairing is one way fractal heuristic. If your current thinking is blocked, to move ahead is to use it to sidestep the blockage and open UI) new research problems and opportunities
II.
ANALYSiS ANt) NARRATION
of analysis -sod Like positivism and interpretivism, the pairing a fractal heuristic. narration is used throughout social science as h as a story, sum Sometimes we riced to follow a story throug compare the bits, but times we treed to break it into bits and level, the switch between no matter what the method or the and often used. narration and analysis is always available -
i-a
In
die
t F)
FIIAC1AL HLURIa’iiCS
ins
nrst instance, this switch can be seen as simply the
in motion, discussed in tin i e e c h t r. F e\arn in one of the most influen a cu c I rs I i der icratu t (dsce West and D& turn erman argued that gender is not a set of traits, nor 1 vailaule. nor a rote, bLit the product of social doings of some sort” I hs I 2(g That is gender a performance, a process of a cc c r at d nv g certain symbols in cer La conic s wit i the tent pointing to oneself as gen dered. l he insistence that gender is nor a fixed thing but an unguin a pc rfhrmancc challenges gender research whatever the is] d ira n h twans hrgerting about si i tnt olc, a id ss i ung how topic mark gen p her distiuctiom, us cr time. In SCA analysis like the Berk and beth pupel Just mcntioned, it means ins esrigating trade—offs in ii i ss ut in t tad i thai r sun ing that there are stable hi c of ci at d w mc in, rid nclrrative heuristic of freezing
or setting
,
hut the anah us narration h
untie
move is often not just a
matter or settIng n motion or stopping hut can be a specific ith espec t i i paticular urrent method. in ethnogra fliO tr x i p1 s s up dii of iii last twenty years has tow i ci mu tfloft narrati ten poral approaches. The new cthnogtaphv embeds its local events in larger narratives ,
ul
ulture
ss
xb ot
intact as in tin work of Marshal] Sahlins on Cap H wa i [195 1), d hopiiig capitalism (as in the iael iraw y ii he ill ing in American factories
[19 9J, globalization (as in Janet Salart’s work on young girls in I-long Kong factories [1951)). or some other large-scale his u al pr s I n in at hroj ci gical linguistics, which re 1 c it r m ink al id ai ytic machinery than the rest
1
of anthropoiugy, the rsoe toward mote OiVC methods tu.. been marked. Indeed, the drift to narration iS So stl’Olsy thu ethnography is ripe for an antihistorical move——perhaps hasi
on an ahistorical theory like rational die ice perhaps based ji renewed insistence on local ethnographic validity. The constant tug of war between narrates e rid oa1yii.. moves is even more evident if we consider not a piunicc ii methodological tradition but a genetal field of research Studies of social class in modern societies are a good example. The great classic of rnid—twenneth—centur social—class studies ‘a as W. Lloyd Warner’s immense “Yankee City’ scudy of Newburyport, Massachusetts, a kind of industrial-strength ethnogapl done by dozens of workers who talked to hundreds of peple and evaluated their class status based on their language, furni ture, place of residence, and manl other things (Warner et al. 1963). The social-class concept Warner used in this work wis highly analytic and static (Warner, Meeker. and Eells 1949). Not surprisingly, it ‘aas widely attacked by historians. While
some of the attacks were, predictably, based on the simple “putting into motion” heuristic (that is, “Warner got it wrong because he took a snapshtt when he should have watched tI” e movie”), the most damaging was Stephan Thernstrom s highly analytic study Poverty and Progress. Thernstrom traced individu als through manuscript census records, counted noses, and showed that there was far more class mobility than Warner had suspected.
From our point ot view. Thernstrom made a narrative move in that he looked at th life histories of individuals rather than simply talking to all of the residents of Newburyport at point. He made it in a very analytic way, howevei in that se
Fsi
M rjc>ii, ot DIStOVERS
d ci
ii ii
als,
u
cia.
dc tailed histories of individu i h c icduced tlitir lives to coded sequences of the C ss V y h Id over ti ne. Ifos narrative S S ci y L c’ople
n cci vi w
t t
01 seek
move with u naf cc a en o it a, t stiongly with the con n mpor ItI nios by BIau and Dun an’s airea ly discussed Amer ‘.ien Or zrpetioI ruJrtIm Thes students of mobility_and indeed ch whole tradition thee stood ‘n——conceived of the “riarrac is of mobility as a jump from the static class status of a fathec to the Statit class status of a Son. The move was ana I ci it r early ii le c s, assuin, ng away most of the lifetime stand ig, most of the change in the f ::npai ‘an nd (as xse have seen in an t tin t rc e tiler di ipter ill if ci t r us-in lividual variation in the “nar mi me pattern of ci uses All of this in order to make dramatic i
r
C
fat it
‘
cia
s
,
àfldi\’tk
compar!sons.
ho make so man\ ‘analytic moves—moveS away from narra
sounds is omcisoine, of course, hut it is important to real izc I at a litcma mc has to makc such choices in order to move
tint
—
1h
litcrature dcliberately as1 V a nit pa rs of the history in order to get at oth is F x cxampi hc enormousl influe itial paper of Robert ilodge, P ml Siel a id Peter Rossi on the history” of occupa tional prestige iii the United States establishes that the occupa tional prestim’ ratings are stable over time (1966), a crucial i c
dog cal mc bl 1
o
‘
elenlenc in the StrUt tural s iew taken by the Blau and Duncan u o nid most laccr sociological study of mobility. But the I )cg e iid R si parer accomplished this by assuming hat cli c V c c i Ii irgm in th natu e of occupations themSc v b nut ‘t U 15 i id 196 ‘Cot assumption was necessary, .
‘
oum
‘.
if
.
wc u r
t i
liink U a peopic were
rating
the pres
.EIAi 1iLuRe rim
things throughout the period. BLC in fact tOe er lbcu identities of occupations like secretary and bookkeep cha! d. ignoring that chaoge—ar almost completely in char periu 1 to be paid. Oni b least for a while—was the price that had can you opea nd ‘r assuming away some parts of a narrative tige of
the
same
parts to analysis. sis/narration Ce Before leaving the icactal heuristic of analy that move with the should consider some examples of studies already highly an grain rather than against it studies that are move to become esen more so or e further narracis e studies that are already narrative but mak agaInst the gram moves. The reader should nor tnine thac the only possibility. ens itselt by niov An example of narrative ana1sis that deep level iS Goraii Them ing to an even more complex narrative tal and the Rise m born’s influential paper on ‘The Rule of Capi of the classic nar Democracy” Therborn s paper considers one by comparing rative problems—the rise of democracy— modern democracies. capsule form) the histories of two dozen al Oi’ig.ni / [us argument starts where Barrington Mooras Soci the tiotion that the Dictatorship and Democraci’ leaves off, with ingent process, ior tlic rise of democracy is a complex and cont constellation of variables, as t result of a single variable or work of Seymour Ma peared to be in the mLcch more analytic ts that prior narrative tin Lipsur and others But Therborn insis essentially related to analyses have left our another narrative in or threat of foreign that of democratizarun: participation or its threat was central war. He makes a strong case that war er and au s to spread access to pow in forcing bourgeois state populations. He thus thority more broadly throughouc their
alytic
but make
a decisive
It a ly a c m a thcr vei i u v Ii i i du vasiio I c Ii r n i r to look. s ii s i 5 1I die liistc c
o)
.
vhar
FRAC [AT HEURiSTic.S
.1
i
c )mplex historical cc complex (Note that his a s ii le vat iable of war but s differen wars played in
ison
I rat spl bu s ‘he othet direction) i An Vi’ll mon x w ta iss cnd tIc Theory of Price s John Moth s “Rat o I decade, until ot ccc or i Movemi nts, a papi r t sa 1 rot its kernel a theory that Robert Lucas and others tasl oc e transformed our s iew of governn i t inte vention in the econ omv, Much, an in onomist makes i t onglx analytic move in a tradition of research that is alreaux highly analytic; not only are economic actors rational maximizers at the first level, he says, nut thin in iact act the way tefloni 6 would. The paper starts with a pureR rorinal analysis of an economy in which produc ers are predicting the prices thin will be able to get for their goods in turure rime intervals. Ti specifically attacks Herbert iniori’s Iivporlwsis of ‘bounded rationality” (discussed in C
11151 IC
‘
Chapter Four; sec Simon It
is
suflietiflies
cu.)numies leads
982.>:
arcied chat the assumption of rationality in ci
theories inconsistent \ith, or inadequate to
expl.un. uhserucd phenomena. especialir ihanges over rime. Our h pochesis is based on ex.a Civ the opposite point of view: it at it’
it iiaflUi u union u models do i\[urh
1961
not assume
enuligli rational
16; cmi hasi idded
T i ‘I there a re a substan argunient I ce I firms expectations and e b ciii and predictable di I ice vould have been able to sin lii hcILi\ iot of rl c ii 1 s cu an a tal ng advantage of it. On the ,itc i tirm ii
general economic assumption that people arc ational, sonic n
would therefore have done that (if it were possible), and if fore we are safe in assuming that prices as they turret tly e n reveal all such predictions about the future including se c speculative ones. For if secret speculative reward exist di i someone has taken advantage of it and hence removed the pos sibility fiom the market.i Muth’s argument was later used i attack Keynesian management of the economy Since govn ment fiscal policy was a matter of public recoid, it was argue speculators would take advantage of any difference betweer government-supported prices and “real market” prices ii iii process canceling the effects of government intervention. Our interest here is less in the policy implications of Muth s ,
celebrated article than in its seemingly extremist insistence that an already absolutely analytic literature become even more ana1 tic. Effectively, the Much paper assumed that at least at the level of expectations, firms (as a group) were as good ci predicting the future as were economists. As Muth himstt pointed out, this was quite close to “stating that the marginal
revenue product of economics is zero” (1961:316). Not and’ were economists analytic, but they also might as well assume that the firms they studied were as analytic as they. This ex traordinary assumption produced two or three decades of citing research before the rarionai-expectation hypothesis WaS finally deserted for newer, more exciting ideas. Thus the analysis/narration debate also functions as a fractal distinction, We should note, however, that the ordei
/p wh J’i differe one takes narrative or analytic turns makes a big Taking a narrative turn after an analytic one does not get you t the same place as taking an analytic turn after a narrative o se
10
4(1.
5
ii
I’RACTAL hEO RiS’i iC.
D
sons to different people with respect to different theoctes. in— example shows that the order in which you invoice tmacta heuristics has a big impact on where you end up
c ornes win the story of my own men hoi owin of opti mabman hng methods ftorn biology, e 1 had de to i ad in dia jic I ur I d I ts dorrowuig bacaus k’J that r ss as important to think about tha full scqucnces of ‘c N r tIn r than jus a it psiara instance ot cm U( pits epation. I hat is, I made a iiarratis a turn hrst, i lovitient and hc tuli uqn ‘net of Hinleone s work life as ud cc an n ut us tat n ci
oc
S
this
vi cSNI) CULTURALISM vioiusii 1 Bi III. s
Lii 0 w as analytic, I rca ized that 1 could i cxi anpoi tam omp ra ira rs C y am ioying the sequencing-Lompanson al The got irhrns t 1 iat s tra usc ci to toit pait siands of DNA. distances’ ber een careers, and I aig rirhms wool I create a.id ha a Lssit LCIT., n. ng thc usaal am, of panemn search
i’necho Is. Bc cot ipa is ii i u. hen wa srtid v otkei s v make the an tc turn rirsc, we inevitably think of individual episodes of pat tu ilar w rha P winy c itpl w 1 (0 (10 partiL ular things at a an moment This in turn leads to thinking in terms of labor m i bats wit re th e yorker-jul units are riansacted. If we than wake a ‘iarrat1se mose and start to ask abotit the chang ulat labot market, we are seeing am m I joy, i ti e dirrecnt sat of rhmy,s
than
arc
visible using
the
methods
I
hut rather cOntinuotEs set of transactions. The questions of interest aren’t Ic c i wr h t ther the historical develop IN tL H hir Ott 01 a geitetal labor market. changes in likelihood of limy, s it types of individuals ii hi my, hi r s ii ,, ( la c,
ha ci
v
in
We
lii d, an I so ha m
OIi
mit it
He
s ti
w
is
is
i
L
ha
i
cc t
ti
ious sat of people
niterestnlg It is not r gilt set an I one the wrong. Rather, they’re sum s hut lot diffterit rcan I ii pot at t q
ft
em
of
ju
t
15
r
With the heuristic involving behaviorism and cuhurabsan we move away from debates about forms of analysis to the heuris tics drawing on differences in how we think about the ontology of social life—the elements and processes that we imagine make up the world. In this first case, the issue is whether we In cus on social structure or on culture, on observable behavior or on meaning. One of the best examples of this heuristic I have aireadt given: Howard Becker’s magnificent paper on marijuana use used this as an example of making a reversal in Chapter Four, The reversal Becker made involved just this heuristic Rather than assuming that attitudes precede behavior, as is more oc less standard, Becker argued that behavior produces attitudes, He was playing with our sense of the relation between behavior and meaning. A useful way to see the fractal character of this contrast is cc look at two influential papers, both in a single methodologic’sl tradition (SCA), one of which takes a behaviorist turn and Ofle a cultural turn. We normally think of the SCA tradition of methods as largely behaviorist, unconcerned with the meanings of things, hut even within that framework it is possible to move in either direction. As it happens, both of these papers consider the application of economic ideas to family life. Ir one that application is part of the hypothesis, while in the other it is something to bc explained.
FaACTkI. rlLuimTIes
Mt LFIODS Oi I)IScOVERY
a n ove toward behavior. George Faikas’s “Education, Husband and \agc Rates and rh Division of Libor between a o re if rh lust papers to look directly at the family k or of-]ibor ‘ueuon with itt nig modern data, Not sur— of social pr snity. it has been very influential. It s a modcl cxc 11cm d ira at d eflec ive analysis and, perhaps eses to which the U IC it p wiant, with leir alternative hypoth test three basic auth ii i ci e .ju ii artet t ion. karl as aims to economist’s theories about the household division of labor: the old w i tat viiw ii t oupi s seek o maximize total househ e utility and henct a Iji it their division of labot to the relativ house v of husband and wbe to make money outside the tb itt in 1 theory ri it middle- and upper—class hc h IL husbands and WIVeS are more lik ly to accept women’s work CI( nce, witi
and the “relatise resources’ argument that ri ic s n chin ath n (iI# available wages outside the u a ‘v home) drivi rite dii ision of labor. dhat n behavic inst about the paper is its insistence that we rt ii des bout the household division of labor cxai ii e ii
JJtSIJL
r se ill
home,
hut retrial perform inCc’. Hence, the dependent variables are the ds re is ifhs in mal is srk outside the I ome and the husban on households I irtid Ii viii it housess oiL Most earlier isork wa h rse 1 it e hi ogra hic or interview-based research that gave ltss at ention to behavior th i to attitudes. Indeed, it was slear from it cat 11cr reseat ch that those attitudes took the It r Ia iv resou es and subcultural by pothe— s. \ hat was not known ivas is he tlic r behavior did as well. I )td upn i and middle-class IloLiselIolds just talk a good line, ri car I c it Li va easy to suspect that couples might o
it t
i
i
ii d
-
t
talk a more Lgalitarian line than they actually lived, s r hap al pened, Farkas found that the relative-resources (edanu,x differences) theory did badly, subculture (class dirferences best, but the wage-rare (ecological) theory could noc be rulul out. As often happens, the big results were surprises; chat rh presence of children played a central role in dererminin the division of labor, and that division of labor changed radically over a family’s life cycle. For us, the important matter here is that by insisting an predicting behavior, not attitudes, Farkas made a distincti bc haviorist move within a tradition generally regarded as already quite behaviorist. It was a matter of doing what we airc’ady do. but doing it better. We can see the contrasting move—wlklch ry is more surprising—in Ron Lesthaeghe’s widely cited “Centu of Demographic and Cultural Change in Western Europe. Lesthaeghe’s paper advances our understanding of changes in demographic behavior, but it does so by moving toward c Lilt nrc. The There are two heuristic moves involved in the paper. larger. first is locating demographic change within something descrig This move of lumping things together is one of the uence of tive heuristics of Chapter Five. An important conseq he employs Lesthaeghe’s choice of the lumping heuristic is that lumping: factor a quantitative technique aimed specifically at regression analysis. As opposed to SCA’s much more common of differ techniques, which are designed to separate the effects whether cerran ent variables, factor analysis specifically asks larger phenom variables cannot be lumped together as part of starts looking. ena. (It is important to realize that once one
tIJCTA
M
I ii
D
o
D st
ic there are fo mal, mathematical methods for many heurist techniques reach far more is ov s St itisti al cod mathematical rho ot rn us- --or a course on sociolog I ua Ily at a plane ic al statisbcs- m p1 t make yot think.) is less interesting for or (c r puiposCs If srhacghes paper culture. this is clear from i s lula pi p lu cc for its move toscard t
uc ip ni
ip
so
ite
icc s:
/ ri til cv it Ii e is it es ci e pelt of a broader emancipation i\lorc e i ic ll th demographic regulatory mecha I nisms. iphc Id by the accompanying communal or family au to the principle of t tot iv an I cxc hane pattcri s, give way dot
Ljisi’Ii
)‘ [‘tY
1 oi e, hc tel
not
‘i
behavioral cue. ncl he manages to use quanticec v cc
i
es to do it! nic u 1 i, chat wcthiu a pirieulae taa,ni iu of aschJ 1 \Ve see, th that is widely undecscood as sctougly behaviui sc I i nil in SttO1as eitht ditectton. Farka’s iTme ward behavior that can be measured Leerfia ghc a is to -vaid u individualism ha. Cd Lit 1 lea cultural construct cche rise of 1
sibie to
move in
sured’ only as an implicit commonality amolsg cs.iStii measured variables. Once again, we see that a ommirmeti one level to one or the other side of a fracrac neuristic due did translate into a comm:tment at the flCXC ievl. Al] toads arc- al ways opeti.
allowing an c xtension of the do
mic tationalit to ti e pht-noinenon of reproduc I Ia purpose of this exercise is to explore the extent tion niiptiality can be to v hi h itrIent changes in fertility and v e ved as ii hI stations of cultural dimension that had a! the demographic transition in y cmcrgcd at the ime
IV. INDIvIDLALISaI AND LMERGLNEIISM
F or pe çLst hat ghe 1 9S 1:41)
and forces. Philosophically, emergentism has found itself the embattled position. Every reader of Durkheim s ‘u1c1de hr ow that the author spends many (probably roe many) pages etcaisni. r ividu i d fending his emergentist views and attacking 1
main of
ci us
t the In making this move, Lesthaep te moved ver much agains in train of demography as a social science Demography is s. Its nanv w ‘s m 0 the i sost behaviorist of the social science entral variables arc ra es of Ii. Lt unmistakably explicit behav -
of iors. birch marriage, death and migration. The apparatus lite-tahlc a salysi through which rates of these four behav age and marriage it ten piod ii:e estimates of populations social science. Yet ti cturc s, is one of the glories of fotmal to Lestheeglue s whole enrerplist ii this influential article is ,
isake
is
see ccc mo taphi
hange as a part of a cultural shift,
The debate over individuals and emergeucs ms I ecu out oc mc most enduring lii soca1 science. Methooologicai individnaeiss are forever insisting that only individuals are real Yet most I us are closet emergencists with working beliefs in SoCial gluaps
Yet this pairing, coo, can be a fractal heurist c Emergei nat literatures invoke itidividuahst tneories and iCe ‘versa. Ofi CCII see this in any methods tradition, In ethnography, foi example. the dominant tradition is etfinography of groups, from Mai nowski onward. Yet there is an ecjually old teadicion of incluvid ual study or hf lust ory, beginning mu W. 1. Lii mas an I Florian Znaniecki s dye-volume series on Life Fe/it/i Pejicj,i;
FItACTAI htLRiSfia,
51t inoos or Discos ms
I Ivi
ries and is loch was largeli built on life histo I anthropology has lifmhisn iv docuincuts Th hisu ri al turn f i md viduals, as in Sahlins’s t c 5 1 f sun a locus r it orical A. r ( ta m C iok i i rh 11 iwamian Islands Hist des between anal ysi 1 s 01 1 urse seesasveci br m my desa ry. \‘ithmn great man biogrephical history and corporate histo tis 0 bus els of analysis are often pirricular historical isorks, thu comnpletel\ triter iss used. actal duality in action is to Apai i a pro I wa ri sec ci is I is major nethod tradition, i m pa e s wthi a
i
t a
s in so is a It ris li. m Am mg the most famous book ive lence isrr rhc lasr half century’ is The Lugh ‘f Colltct to show why or iii \lannur olson. Jr. Olson s basic aim is activities; he starts uopie )oin groups and participate in group misc. he wants to question ft om a resolutely individualist pit use of the benefits they t a t n t peopir joi s pioup beca tI a H mote that pro ps often provide benefits for t sir i r ribute or not. all h i is mb r. whether the members cont they are called, \Vhei it comes to these collective goods, as tive to take rhine who can cut assay ss tab it have every incen do so are the ilium is ithout contributing anything. (‘I hose who larized the con Etc e ii dee s 0) son s was nit analysis that popu is the I u sit Is ta tt s- mf ftc t iding.”) But if this 1 ide collective cxpiain why p ups that prov is i v s lflswt r to this question was inge— et i ist’ Olson p )c I incentis es—various Co is ins okimip Wlhtt lie called selective ii r
‘
•
contribute (giving ways the t.roup has of taretin&r those is ho t (giving them pun thrum posit is e tess ards) and those who don’ there were further problems (who was to
islinmeimnA. Of cOUtsC, s c mm c a time ia
‘I cove imentivesr and so on), but the
t 1 ctive aCtiOO uo book ignited a debate on the nature of colle ed in the classr toe continues to this day. Al of this was argu senraniuns ul mal style of economics, using fairly simple repre all of it scam ted supply, demand, contribution, and so on. And iduais. in the traditional manner, with isolated indiv logist Hath An the same time Olson was writing, the socio r direction. Whue son White was moving in precisely the othe ly time ce”erse employed similarly formal methods to ask near similar mineresis question: not how is it that individuals with define individuals get together in groups but rather should we positrons in all of as similar when they are located in similar individual interests their social groups For Olson, similarity of aim of coliaborac came first, and location in groups (svich the nd. For White, it ing on producing collective goods) came seco in groups came was exactly the other way around. Location g similar (in in first, and we could understand people as bein of social location terests or in anything else) if their patterns were similar. ctural Equiva Francois Lorrain and Harrison White s ‘Srru s not from the no lence of Individuals in Social Networks” start rather, from the hon that there are individuals and groups but, relations between notion that there are individuals and types of papers, many them. As is often the case with such original exposition char levels of complexity were included in this complexity and have since been forgotten. But hidden in the of category thco couched in the impenetrable mathematics study of networks: was a concept that would revolutionize the ely speaking, struc the concept of structural equivalence. Loos actors all of whose turally equivalent actors are defined as those network ties are the same:
IRACTAL 1-icUitisTiCs
Mt i w,Ds (ii I)isi
6
vrav
s structurally chuisalent to b if a relates to I ( in e\acrty rita sane ways as b does ?rom obh ci then, a and h ics of lit h ,ie o P a structure 51) it )i utable. lndted, in L u:iy ajws clam, riley arc substit its and b. (Lorrain [cc m ni to ideritif t Sc a is I hat xsoMs, a
a
ci
at
suci
aid \\
asr
)
hit
I.
3
elaborate the \yhitc and hi cullabort to 5 in I lol O\ ers would into a Corflpre inutpr at sit mi oral equivalelkc. making it structures. tr undcrstandin roics and social I ci sty ii de come hrst; in Similarity became network similarity. Relations div’dcab sCC( 1k individual cots in t a, Li cli i, ss 5cC that toves toward ate possible despite the usual associ C ptions or emerfcni ones methodological individualism. The ctit H of iotmilitation with re rk analysis is cxt tamely instructive in this 1 ic his 01 tthose opposed to at 1 Tha “intl s idualist network analysts of networks Whitc- -j m s Coleman, for cxc rnpie) conceived lity” in nethugeli n ret ms I cliques and iseasured ‘ccntra y called vh t â t ie emergcntisrs like X’hne (usuall Vt P structural equiva suructuralists ii t us Inc aturc) focused on widely cited paper lence [ he strucruralist Ronald Burt wrote a Not sur he as a I the two against each other (1983). i v it lence won. orisingly gis en Burt’s allegiancc, structural equiva eventually devel— But th indi id ualiscs went metrily on and has ig a lo c I network ties was a kind of i d P i one t at and James or it dividuals, Baptised by Pieire Bourdieu r sac; notion has become Ci leman with the name social apital, this now virtually a or of tht great grt kth oncept I the 1990s, analyses of field after u iidar I variable in traditional SCA-type ,
pared down the eLton held Meanwhile, the struci uralists have ns that drove Whre s rare logic of multiple types of relatio rk” concepts of ma original work and are developing “netwo incentive theories of tra kets that invoke many of the classical wrote about games to ditional microeconomics. Peter Abell structural concept u networks’ (1990), bringing together the concepts of gamc networks and the relatively individualistic theory.
y takiüg new turns So this fractal heuristic, too, is steadil as it drives the research within the old turns, and so on. Just in more routine social ac frontier, so also is it available for us ualism or emergentisici ence. Making a move toward individ a pohlem or findiny always available as a means of rethinking a new line of investigation. V.
REALISM AND C0NSTRLcTI0N1sM
s probably the The interplay of realism and constructionism spent the last thirty most familiar of these debates. We have applied to nearly years hearing the phrase ‘social construction” class, nationality, everything in the social world: race, gender, edge—whate er. ethnicity, aesthetic judgment, scientific know1 is relatively Because of this huge amount of material, there . Readers re no little need to illustrate constructionist moves are some particu doubt already familiar with them. But there and it is useful cc larly interesting versions of construcrionism, because tney hava look at those. Among the most exciting, research, are the such a big potential impact on social science very census figures analyses of social statistics showing that our ately questionable embody dozens of arnitrary and often deliber ieres, Simon Szteter coding decisions. Writers like Alain Desros
1LACiAi H u1s1iu
sh iwn ft exam ftc that the occu priulal rarista s ci e ise to measure achievement were quite id in re m ott-mt that they change d vi p a it a aiim acily horn census to census, \X here do we count family
a
U
Ia u,
ts H
aVC
usins and uncles who help o a in .asiorciili in thur telativc’s businesses? i/hat do we do w th hi s i r p iii ml ra f nvisihl work”—easual work for ii urinal c eonomV of C leaning, e bud care. lawn ri t he nit is
V
-
Wic’
at
I
lit
h 5
nets
an I
‘
I s o i What do w in the of illegal labor—paid ork in the drug rndustrn for example. ‘cX hat do we do when the in of tO Ott ipation doesn r change hut its status and its emd i ci e OSiriOn of secretary 1 tpna1 rn bytes ii,, as ap n
mow
-
th e tl yec is of th tnentieth cetiturl One has only to p 1st thse qut dons to realize that what the openings eflltti to he methodoloiical problems are all c I maje usc a cli I rat it ions, Fry in iy out how our labels for the labels are oct upatira iS caine rico existence and how loosely 11 tell us exitemely impor related to tie r air res un lerr earl tant things ihoot the labor market, nderstatiding the creation tells is more about how o house vi if t ic cii is aicgo Icork was strut tui cd in the nineteenth century than could a
ii
oat ii ud cc ea 101 die cc inns categoric s as unproblematic. It is also ci oe that once you make the first Constructionist all the type. >1 lobs that got mixed into au yo dig ncI out ol the category of hookkeepeinyoti then face the im -
I what that con uructed world looks like as a “construetedh real one. in the on hay a g it r to generate a firm “his au ot o( kc p torcal 51 ties ot occupational numbers. Tlrc constructionist hi nki ig in e rid all too often stops thete. cii love i or tint rs ii
las
o
,
ass inbling
n image
c
When
you
make
a constructionist moVe,
dwayc
gi. in
rlth consc uc a realist turn. Once you’ve opened things up real cotrs nenacs in tint tionism, it is time to ngnrc- out inc -
construction. it Ove the Still, one wonders whether tile heurstc tin Hilt papers that ci other way, against constructionism. Are there push uS o liherarely undercut social construction papers roar an ry ward realismb One is Daniel Chambliss s c1ssic “Mund scudirig Dmpeucise Excellence.” Clianibliss spent fi’e years local to thi national leve a SWimflllng. He coached from the c 5 with the b interviewed dozens of swimmins and rraieied to disco’ teams in th country. His core research problem x as ing. His cell the nature of “talent” and “excellence” in swimm tion, a tral conclusion was that talent was a complete tabrica romanticize meaningless construction designed to cover and what he called ‘the mundaniry of excellence
Superb performance is really a confluence in dozens of nThH P. skills or aCtivitieS, each ore learned or sconbied pari, is ci to have been carefully drilled into habit and Lileli aiC acrt’J dina.’ gether in a synthesized whole. There is nothing extraor or superhuman in any one of those actions’ aul all the>’ are done consisientl> and correctly. cad duced excellence (Chambliss 1959:111)
he fact that together.
pro
ann fliane Great champions are people who work on the detuis motivatiot s sure they do all of them right all of the time Their Olympics as were also “mundane. They didn’t aim to win the e their much as to polish up their backstroke next week, improv Ii sleep habits over the next month, arid eat more careful1>. , Chambliss short, their goals were nearby, not fat off Indeed
FxicTA1 Hiajiusrics
191
Mf lii0J) 01 1 )Iscos LIlY
he h g meets into a big meet, asi in l making every single day meant rvery s n Ic rat c a prier c Th big ones then
h a uc i u id ni
t ci ci
piut s
oInhii fliOfC 111210 [IIC’
)rIicrS
wimil ic
ii
rr b ss’s m rs’
s n
ry
to t
u
rgue rha
1
particular social
that it umirrucrion (talent) is simply a label hr the success all Differsupposcills explains. It does not relcr to anything at xplan u ccss in swimming. tt ct ces a s i s a I ii. of the in n set tins muse as a behaviorist (tie, because es. i3ur the opposite of a sisrcrice on looku g at si iab practic culturali t me, and talent (as e cv 0 is r w e ss old he better or ins d rc rh routine whas mis ot doing a turn thing as it sleeping well) i lilt 50 11 tCll a cultural or subjective have s iT pl it if c n a, at unreal ty (So tills move could Believing in L e 1 sri iic ldrcd P One of my commonplace lists.) SOtfle one thing that ci iit rne tis Pc hiving ti it th’rt must he although tiialccs for COIlslstt lit succcss in swimming and that, u ly what t s, w should give it a single w d i 1 cc nsrrnerion that gets na,flc: talent. It u a ralier in’l’IC oca the real sources of consistent ill nit way of our u lderstandinf \ C
1
c
i
S
11 11111
from more N on surprisingly, (Tamlib us s p1cc drew fire him for trnngIy cOnstructionist ss tints Ira DeNota chided had of— i nany othc r he ds (Chambliss ha it ii of Winning ter d ccn tat is c reneraliz.a ions) even ilic standards wmr S in I audiences (for exam sic r ticgc dated betwcei p ‘rI ated an Ii in hi Ss s arg ir icnt, she felt, “cultiv ii Ii arts imagery of excellence and rank r pirar ‘ly tic ‘in ncr i for treating dig (192.102). T ta I she a tackcd Chat ibliss Chambliss ss inning itscli not as soc tally c onsuucted hut as real. ..
,
d herue 1 UICC arbitrary çr 1 admitted, in response, that ii WaS L. times cially constructed) that we rated swimmers by their al stead of their beauty (the way we rate diversj or the ce’hnic precision of their strokes. But once that decision ii manu, supecu still can define excellence provisionally as consistent m Ot ority in meeting that standard” (1992:105). In tile Wccc are real W I. Thomas, “(l)f men define situations as reai, they in their consequences” (Thomas and Thomas 192S:572). the Chambliss’s paper and the debate it inspired show well ic fluid power of the heuristic that flows from the reallst!consrrc ied, tionist debate. Everyone involved in the debate was idenrtf ds and at least by the majority of the discipline, with metho committed subfields that are considered to be overwhelmingly fairly con to constructiomsm. But even within tills small and ed an sistent group of scholars, Chambliss’s realist turn produc extraordinarily heated debate. that The realist/constructionist debate and the heuristic debate flows from it are probably the most familiar of the ic use of heuristics I discuss. It is esscntial to realize that heurist two such a debate does not aim at debunking or demo1ishug— moves, common reasons for making constructionist and realist topics to respectively. The idea of heuristics is to open up new invoke con find new things. To do that, sometimes we need to prestige. structionism, as have the students of occupational given by Sometimes we need a little realism, such as we are heuristics Chambliss. in both cases—indeed, with all of the they arc discussed here——the idea is to open things up. Once trashinc open, the excitement lies in following new leads, i or our opponents.
‘1
i.
(orc LXTt
J[
)L
ic1s\1 AD
tn&AzT/ L 14CURLSTICS
)Vk CV
1NO\cQ [lxi
ALISM
r no i g co texts is the last c t s i oki i the debates irisi ug irons issues Ot soaial ontology. The issue I Ii ‘a on ought t rr c o imp ove and deepen is ci oc’ Ic cowled ;e it a pr thlem. The contextual strategy is to ‘Ii
i
‘a
n o r in di e me r o see ho v it is embedded ii the I rgtr ‘,o(, iIl norid, Th nonee litexi ualizing strategy is to ii by finding comparable 01 ei of en g scu 1 te r urn ts or problems elsewhere. r e soci sciences i oncontextualiz e I u y 1 r ing S rice program of most of the scientizers,” Contextualizing is cit t c Ce he n lie pr gram ) those resisting scientiza tiors But when one looks at research, it is clear that the moves ft og so ig he on ext o ofexpi citly choosing to ignore an important context occur throughout social science heuristics. As I etoti, sh 11 in has at ssoves syainst the grain, since they illustrate tin puss er of such a heuristic particularly well. I hee’in cv Is he’ move of en extuaIizin. SCA analyses hi
generally asoid tIns For example, throughout this book I have re’s ‘ut d xai p1 s ft m the laige literature on stratification, most of them looksng it individual achie ement in terms of the f ii dividuals This is a noncontextualjz 5 ii ticular atrrhut s ategy, ‘Ihe’ parrneters measuring thc effects of education r c .upati )n n tath r’s occupetioi on a respondent’s current income cit acliic’s e ment are estirnate’d on the assumption that nly tie Ic spo ide i it own terihu es have an effect, nothing ing
sri
else, riot his Ii iends’ types of emploc ment nor Ins extended so cial netnoiks is is \lark Grenoetter’s Cetrinl a Jof model)
nor such “market’ contexts as proportions or various eyj es o jobs avaitable in his locality. But as early as the 1970s, an important line of researc c cc gan to argue that workers were grouped ifltO sectors anu that that sectoral context had a major effect not only on absolute achievement br.st also on the ways in which education, oLcupa non, and so on determined that achievement That seetoral tra
dition is well ilitistrated by E. M. Beck, Patrick Roman, aiid Charles Tolbert s “Stratification in a Dual Economy.’ The pape predicts annual earnings based on the usual array of vanabies. sex, race, age, education, occupational prestige, union membership, work stability, and parents’ schooling and occupation What is unusual is that the data (national survey data) are split into “core” and “periphery” sectors by the individuals’ type or’ employmenr. The equations were estimated separately for the two sectors, and tests were then done to see whether the effects varied between the two sectors. The paper finds that nor only are rewards vastly different in the two sectors but so also is the size of the effects of various variables on those rewards. Context matters. The individuals should not be seen as an undifferenti ated mass but rather as grouped into these two quite separate labor markets. Note that this paper could also be seen as employing the splitting heuristic of the last chapter. The context is not partic ularly elaborate; it’s simply a matter of seeing the workers in tion Ecology of two groups. A clearer example is “The Popula Organizations” by Michael Hannan and John freeman, which launched an entire paradigm of organizational analysis. I have already mentioned this celebrated paper as an example of
FRACTAL IILLRIST1CS
i9
11(105 01’ I)iscovuty
borrowed was i mxici of context, Rathct than hinkina a orhanit’ations st iding alone Hannan and F it entan arrucd. sic should imagine them in a competitive if h is as adahtinL consciously tint n ( It V I it it I at to ens tioum mii opportunirles and threats, is e should think of ons trained md unable to adapt, thei future deter tOt in as
L
or us
i
\‘hai
is is
mi ILd in iact I y rite tt mpetitlvt ressures of their peers. pI yin to m ii n odds borrow ni directly from biology, I -Iannai. and l’rccman pr posed that wt recast our theory of Orgtuiztions LI1IJft [is ate p iLl ne suoned above, the n L c m a I nat Fannan and freeman paper included many formalities that the ater limeratt re discarded. Bu at rht heart tf the article was an iiisisttiiu at ontext as tst most important determinant of an rga izat )n S c e to V Nor surprisingly, researchers following
rraoition studied the founding arid dissolution of organ ts ZtiOiis e\tensiVCiL Oid sought to pur into practice concep ike r a iii a d g tieta ist at r in, a radical contextualism of he aeti a) I irniulation is as lost; routine methods is ial ii that had not been consj ic ,iously contextualized (durational o
this
methods) emieIgcd as the standard methodology for population 1 g s u lie \V hou d ieeaH F owexer, that the tradition ha Uis is ith a r idical t ontextualizing titus e, invoked in the con text ii tornsi ned ods. where such ioves were unusual. Ant this r elegant illwtiat ion of ontextualizing, from yet anI i ti od is Elizabeth Butts ‘Urban Pam i it e t a it ) ,
a shortened s et,it,d or what is :camc tiC, hook 1ami1) and lose ties to several literatures LO .\ ru ni Butts paper has a•r hay t seotini r h it partit ukir, he literarures on social net-
clt wa works and on the household division of labor. Her reear London done under the auspices of the Tavistock institute in csciga research center dedicated largely to psychological ins and tions. Her original aim was to “further the sociological h de psychological understanding of families, and her researc s to that sign—intensive study ot twenty families—bore witnes interest in depth of analysis. sion. The research itself resulted in an extraordinary conclu to the the family division of labor was very closely related not both but psychological qualities of the husband or the wlfe or ’s social rather to the degree of connectedness of the family pattern network. Bott did not find the strong ‘subcuiture” Ai mentioned in my discussion of George Farkass paper. ss and though professional families tended to have more jointne status, sharing of household tasks than families of lower social What there were plenty of exceptions and counterexamples, s with did turn up as an absolute regularity was that familie net strong role separation had tight, fuil connccted Social works around them. Again, context mattered. her Bott did not specify which way causality ran, although mined argument leans toward saying that the networks deter What is the household divisions of labor rather than vice versa. bent important for us is that rather than following Tavistock’s turned for psychological or psychodynamic explanation, Bott ncr attempt outward, to social connections—social context—in It is to discover the origins of the household division of labor. uandtarive paper mote 1 instructive that when Farkas did his c been than twenty years later, this move toward context had forgotten.
M
Ho
I)ist
:HIOL s ii
Faac IAL
sri v
h’nk that riO e a a always made toward cornxt. hü cc ometin es ii is essential to resist the pres sure tu stud) coiitext, a are most tamiliar with this, as I said, Htir the anricontextual i SC \ and ci ii I a iii ci ii is impulse 1, tair y common iii histoty as well. There it rakes the We sli
loin of g
t
lit
It
Id
t
p ek to ra documents and rediscovering history
p An cxrcl cut teceUt example of tins is Amanda Vickery’s ex traordinary CL /1/ maui I )axgIiltr V ckery wanted to know what ally a pent d to ge nicel women in thc eighteenth century. C ii ma! a c ii sts t w mcn in this period have emphasized the 1
t U
U Tic it
1
n from h cit r >1 iii product ion the creation of idle don as ia tv, arid the rise of separate spheres’ for men sri I o ii ii V k ry a tacked this radition on several grounds. F irst, sch ilars has e traced thes d eiopments to wildly vary i op period5 bet s eon the sixteen cli and nineteenth centuries. ia n
val
if
so ft
riginate in the simple (and erroneous) deduction that some Ci d of gradual transition must exist bat st CL II the starting point of the medieval productive
S a
d thi r
tip urnents
usually
-nineteenth-century system ut absolutely separate spheres scholars deduced this transition and then in pi se I it on dCta Indeed, Vickery ilso attacks the nit tlmdoi g c 1 i o s of these accounts, which are usually based p
y
i.
I
i
cud
porn
of the
Lit
(which are higl ily selective). She also notes ti LI tht urge to it i e iLmily crc nds with respect to related political and economic contexts has led to an aigtnng from con 1 out yu Ii cs, s y o production in the in a a C t’ ii e’ P 10
t jilt
SQtitQa’,
dustiiai resolutic n) a the family tas in the assumption that as 1)1 odin ion niovd out of the household into factories, women a l In a played a si ia ii r I lit i hi Note that these cirgrI
HtUR1S KS
o i to tILL ments use a wide variety of heuristic nsoes in dditi contextualizing one nil: Vickery sets aside contextual phenomena SUCh as rha mnmJO an cia . consumption, the transformation of th 1 story o making of social life. She also sets aside the “larget’ Is 1e tile Pita England iii the eighteenth century. Robert WaIpo nucrscc. 050 the Seven \‘ears’ War, the American War of indcpe Rather she industrial revolution—these, too, are nearly absert. sand f starts with the immediate situation itself, with thou one hundrcu pages of diaries and letters written by more than And Proc 1 omen (and a few men) in the north of England. pictuta of thc these myriad details, she builds up a complex gcnriiii. everyday world of these women, under the headings dent Cat>H “love and duty,” ‘fortitude and resignation, ‘‘pru ty.” C I omy,” “elegance,” ‘civility and vulgarity and proprie contextual de course, the women are seen in extraordinary lOcal it. ‘e see tail. But we see the larger context only as heF Saw events only what the documents discuss. That eliding of larger of charac is indeed part of the books empirical message part terizirig the world of experience these women knew. .
‘
‘
ge one set In some ways, then, what the book does is eachan sts) fos a of Contexts (the larger social processes seen by tHeori neighbornooa. other (the experiential contexts of eseryday life: this sense, his friends, correspondents, retailers, and so on). in What results is a book o torians never fully decontextualize. portrait oi extraordinary strength of detail, a book whose assembled into larger argu women s lives utterly resists being way betwec i ments. Again and again, Vickery finds a middle marriaga faa the poles of prominent theoretical debates. On example, she concludes:
NI
l)is
ion
fiaLiaL hacaisi ii
sins
Marriage carried the potential both tot harmonious license and for miserable servitude, as it long had done. The patriarchal
I
ti
d asserted,
argued
ol
t
no Ic
s
fan
inc
t cCCsSivc stages in
as Lawrene Stone has
her these were, as Keith ‘rightson has sensibly poles sIan enduring cntlnuum in marital relations tar
v
s am
narr age we
onat
ini pm
it
id
i
t’
a
of
i
it
d I
tl
iagc a a p
tin acy of male author ical and emotional part
netslrp. iViker 1995:56) i g vi 1 p tier c oi the grasp of general an rorr impor nit at s of political and social ,urn. .s, a. elI history, and taking it on its own terms as experience, Vickery b sth lecontcxrualize and reconrextualiaes The result is to B
ci i
im narize a d reduce to an abstract I t I inding that can be inserted irit theoti ical debates about fam il and gender in Europe. This example makes it especially lear F it ri issue I cOnt xt is always mplex Most removals ize another. When you ts t tmp t it ot c i Xl i
c
t c’
sc the totirextualizi rig noncontextualizing heuristic. you must indeed be deepi aware of this multiplicity of contexts. Even he z s st ap aren ly no icontextualizi g of SCA work is still tuatit g it
VII.
CRUiCL
Vi
ii
subje ts s mewhc r
an i
d
C0STRX1\T tt
n
,
we
mc
a heuristic pairing that
the hingt we tend to wh 1 ha e calle I proLlarnat ises take as problematic iii the world. Some people think the world sonic in terms of con s to b understood in tetms of choic i t odt cing this lebate in Chapter Two, .rai its A I i o c ‘‘.
choice arid constraint ieinh ii pit other social scientists, but I nsti cC
me opposition of
rcuiir
1.. a)
cescea COL
against many debate, too, as a fruitful sourc of heuristic if Ovus disc usn I shall illustrate the powet of this hee ristie by anoice cod some particularly extrauidinary work dnc n is the constraint in recent social science. vise setch example of add ic tempt of economists to develop an econom c model iw it seem non. Addiction is a coimundrurn for economics bca negatit c ic to involve the choice of a behavior Rnowfl to flas C setting higr wards. Economists have tried co account foe it by am ad discount rate for rewards, so that the near-term pleasures 11—overwhclm the far-off tand 4 y diction—even if they me 51 in value possibly large) costs because those costs are discounted pfesCing rh the further off they are. Such a model, although held as addiction concept of choice even in so unpromising a addicts often try i: does riot effectively explain the fact that limit their future behavior. Lleueg inslie In a brilliant argument, the psychologise addictive bChaViO 5 an I has shown that we can account for if we are willing to other kinds of “temporary preferences” 2 a mim inside the individual u99
create a “picoeconomics economics in two ways economics differing from standard choice rules and dis First, in is governed not by the standard
but by different rules and counting curvc’s of the economists rather than expolien different kind of discounting (hyperbolic this picoeconomics are not inch tial). Second, the “actors” in states, with then mnteres viduals but successie motivational they govern. The varying according to the future periods arena in which th iorig diet’s internal life is then an economic compete to “buy” the “reward’ run selves and the short-rein self
.\1F iHODi OF
FKaCTL HeuRisTiCs
DiscovERs
I
g uccot Yng to rht wishes ot one of them as op )O Ito c TI c riultipi n nal states sustain the possi— i of i ent d hype olic discount curves, which C[JsS OJIL ways cxpoicuta1 ones cannot, guarantee th creation ot the temporary preferences” toward \\ hi Is the addict is arnbis alent,
have ignored the consuaints posed to mobility by che asaitahit a ity of jobs; others have not. Blau and Duncan’s menci. 1 frequently mentioned in these pag Occupational Structure ultimately takes little notice of the const ‘aims on nof iii y choices. By contrast, the segmented_labor-market Itteracure de veloped a strong focus on constrained mobility between the
I hi mud—i titus ourchoiced’ the eConOmiStS by turning rat i divi i is into i It ecoflom of choices, al hsyin 1 nc yE different rults from those of S icii’c c i mu’s iiisIie cam o this algument over a long lwr1oI Of unit, Of course. (He ss as not addicted to the shortterm pleasure of w ruing attacks on economists but took nearly us enty years to assemble the entin argument into a book,) His
primary and secondary labor markets. Writers like Seymour Spilerman (1977) focused on the career as a sequence of th” re’ suits of many constrained choices over a lifetime. indeed, some writers try to envision the interlocking of two choices. choice of job by person and choice of person by job. This is the theme ot the job matching” literature in labor economics. What ch
r nis
[>dlmav
1
shov
i
se pow” ad e
Os’
mt
o
omn
ng even
choice into tad to choice as a model of
At the same time, otheis has e of
more
course tried
diversity of this literature makes clear is that a verm’ useful heuristic indeed is to question the role of choice and constrainL in one’s research problem. Thinking about these in new way’ can open up whole new terrains for investigation and research
to suggest
limited forms or constraints on choice’. As I have mentioned seseral times Herbert Simon’s work is associated with the no-
of ‘b
led rat on ty’ (19 2). It staits out from choice s I mo s r man all’ in iii inquires into the impact of train thcn n ong t eta the most important and flOSS srtidmc’d mnvol\ e information von can t always get enough information to make a “rational decision,” and the informa tion on do hCt costS you something. which must be taken out .
VIII.
CONFLICT AND
CONSENSUS
The pairing of conflict and consensus also has its uses as a heuristic. Most of us have pretty clear preferences on this OHE. and it is both useful and important to try to rearrange then Adherents of the two poles differ in their view of human cia ture, in where they think conflict comes from, and in what they
v en im v
oward choice and moves to
haiateristic
I stratification literature as
think are the effects of conflict Consensus thinkers hold that people are disorderly and greedy and that social conflict comes from these qualities of hunsan nature. They are nor interested in where conflict comes from (that is obvios to them) bat onlm in how ii is to he restrained or contained. By contrast, conflict thinkers hold that people are by nature orderly and that social
ms cii. Some mm ho has e studied the status—attainment process
conflict is foisted on people by wrongful social institutions.
our ti ( t
di s
reward
2 large 1
rature has examined these
maE’ y n b
atd tunstiaint is
Fuacrac Han u:vics
I ci in
c
ri
se ki
in thi ha dlv fi iak
institutions When we contrast seems extreme, hut the pieces of s s woe cx unpies will u I 5 Ii a
it
nit tt
I
ig is of ti
it
strong.
I eat.
Ii
of wo k playing with the cookie t and LOnscnsuS debate orne from the man) writers who iv ye studied in ‘erld SurniePs sv rds Thr Swial Oidcr of the ia. lh if sso tbc ri of louis, the med iorn turnof-the 1 Pe ha
ii
n sr
iii
otis ‘xai
s
that they are places without a social na ned itman nature run i a let Iii lii n hi’ aw for s dee. \Vith no S icial control to restrain them, these forces t e b orbit retorrncrs: poverty i i .1 e ic is ii emit my proenessis ‘sm,
is
lit pekssness. crime, divorce, and v
on
This
was an absolute
its It ii I But ill lam I’. hyte s bri [ham Srmvet C ornr Sudetj showed nw B )st ins Nc n Ii id s fear (I by city leadeis, was in fact a 3 orderly corn munmnv s ith Its own elaborate rules and in Iuyhl
C
151
nun mis ‘h tc Sr Ldie I b wling leagues among the ‘corner b vs ant fo md that h vs howling scores correlated exactly
st
with their soLlal status, maintained by implicit rules and con nr us. lie ‘at n ned be murnbcrs c tine (an illegal lottery on rse races) and thund it to be a stady employer and stabiliL in c ii ii i y II i net He stud d local politics and found cliat
s cc lruptioim played an imporiant role in facilitating and r gulatir ig u i ity i e. in ii err he found a highly or di rl y SOt iii sysiem, One with functk fling institutions and rules ni
Ic J f rein I vi lie ules of 1 h nonn nugrann Loflimunini. Sunless nai/ of tar )iiim did the sank thing in a n ore cc mplex ueighh whood, with three ci inn grouj 5, i C hcago s Neai \\‘est Side in the I 9 0s. 6 i
at
Again, underneath the disorder so cmphasiced in ccaisee c views of city structure was an elaborate and complex set of su cial rules—different, to be sure, but elaborate and extieniely
in hv
i
By contrast, Mark Suchrnan and Mia Cahill, writing about lawyers in California’s Silicon Valley in the 1990s (1996) facd a different situation. Here, the standard view of th situac on was a conflict-theory one. Lawyers were widely regarded in the literature as disturbing influences who introduced dvcrsarmat ism and got in the way of simple market relationships. Entre preneurs and venture capitalists, in this view, would get along fine if their lawyers weren’t introducing so much complexity, formality, and contentiousness into relationships that had been smooth, informal, and relatively unconflicted. Suchman ntcr viewed dozens of entrepreneurs, lawyers, and others and found that far from introducing disorder, lawyers were in tact among the most important Jinilitators of entrepreneurial lift in Simon Valley. Their contingent fee structures reduced important un certainties for entrepreneurs. Their opinion letters helped man age uncertainties for investors. They served as gatekeepers ano builders for the informal funding and entrepreneurial networks that built the community. In short, where the conflict theorists had seen lawyers as a disturbing force, Suchman and Cahill— taking a more consensual point of view—saw them as an essen tial ordering institution of the community. So in both cases, we have authors who moved toward seeing order wheie others saw disorder, but n the first case the others were a body of consen.cas thinkers and in the second a body of nt conflict thinkers. The same move—but in two quite differe intellectual contexts.
Mi
a’ toys
i
I’RACTAL
Dur o tin
ov is riiauc by James Kuklinski and his col leagu S in an in lot or ial paper or p >1 rical tolerance (1991). furs ap r like most of the literature on political tolerance, is Pu it oii ci ‘OflSctiSU iramework’, r sscmr-s that people have inst nitrous as tolerance, ike at l Its ike ml ha ,s at to
A liflkr or
ii
(IL ib ratio r, rid civil ihcrties are nn en dry tO keep those likes and dislikr fron exteasrvc a RI do, tructive cxpression. Kuklin
sk and Iris vile spires start with i cuirure-to-hehavior move, insist ny that ss e lust s urve trot just respondents likes and di uk .s f PC r it gr u S but ils their views of groups’ po r e t L 1 a ‘us So ri e quest ions i ml u led nor ust do you approve (or the vu Klux Klan or whomever) but also “as vu d yot approat’ of tire Communists’ holding a rally?” (or the Klan, r whoever, holding a rally or reaching a class, or wi at ye y 13t K ul Ii iski and hi colleagues also introduced di rr c is tO ti Os to subje ts S mc subjects received no in-
or
tv
mmui is s
answer from the gut without reflection, and sonar were told to think carefully about tire (onsequences of the actions irisolsed. It turned out th r eflcct jot sri orig ly err/vied t )le ance TI e vs at hert, thi a. w is to investigate empirically sri or
rot
s
ov to
o a
answer,
some
were told
t(
one C he •r icr tl estr ning institutions of a con Sutisus svsu In actually supported that system. What turned out consensu s” sal LieS, tolerance and do as as that iv o important Iib t’ tj 1, ci ii o Iw ri onflic t. The heuristic move was not s( m ci tt as a d c inllic thaeor is it was a simple empirical wi
Ct let
Si Ot
n
i
u
P isis o a ‘oust isus iheor
An ecitrivalent most on the (ouflict side is made in Ronald Nat rue of he l’it in, a paper that many decades after inc s its pr hi anon was cogi izcd as o e of the classic papers of the ‘
Hut
RISSICS
century in econon’ucs. It starts with am of tray e.rhiet 1irs tics, asking, “Why are (here firm? Put a intl icote cou pletely, it asks, “if the price mechaitisrrr aOd me uaarlcet aiR so perfect, who should we organize an,’ activities in tixothei way—by command and nonmarket coordination as We do so firms?” This is problematizing the obvious. Classical cLonomin. ‘
usually took firms for granted. Coase did not But the way it which he proceeded fit the conflict,consensus heuristic qulic mcelI. Classical microeconomics in some ways squares the confirm, consensus circle On ifle one I’iand, it assumes people at greedy, as in consensus theory. On the other, it aigues with mal 1 conflict theory, that this (greedy) nature produces an opt
state of affairs in its “natural” state unless we muck it up win disorderly and wrongful institutions. It’s not important here that we reconcile this apparent difficulty in the distinction h_ tween conflict and consensus. (My argumnt would be tha “economic man” is only partly greedy, greedy in a controlled way, microeconomics tacitly assumes an enormous amount of control in economic relations.) Here, Coase plays with only the second, conflict side of economics. Are firms in fact dsordeul, and wrongful institutions, mucking up the optimal system o markets? Coase says no: firms exist because there arc “costs to using the price mechanism to make decisions. For example there are the costs of writing contracts for purchase and sale There are the costs of marketing goods and sets ices. There are the costs or’ writing specific contracts for long-term needs tria may turn out differentI than initially imagined. Iii shorn firms emerge, Coase says, because people choose to organize activities in the cheapest way possible, and sometimes th
kio,cme HLLRLSTkCS
\le 111005 (II DiS( 05 SRS
s
a.
wa
is a.
i bin fir
fl
(He ilso
aretully includes an
of drurneEIt tot wiis a.e dont organze evcryihing this way,
ourscd p c t iech Iso as the absolute pun iple tor resou u e a] location y savi ig that people sometimes boost on pr ee p oi rids nor t USC the rict mechanism. What appeais to be a violatit is of the univ sality of price logics—the rhc cC ntrarv the clearest evidence of is o f lirni it x a. 11
t
I’ a.
i
ie
dcej er lesel, rue price principle is recursive; what appeared it an even usrilv its own removal In short, flow of i rr r v al inst rutk s c[ it ering up the free o b mark’ t base I mt ract on is imit an irrational, disorder-creating free flow of price tea liStitUtioli. t s itsi if an expres ‘ion 1 the
itS operatloil at a
SoIling.
u. playu p with the parts of the conflictj ommsenstms debate Time s arious examples here show how useful can b o question one’s beliefs about the orderliness or dis rdetlii ca. h hdi v r, bi ut dv nature of institutions as -
a
,
it i
is
ti
ai s f ‘o ti oi d sorde o This whole series of problems of nk one’s research tfrs a wide sarietv of different was s to rethi t/consensus questions. Like all the other fractal debates, conflic I S S tis t I eu i ti 0 i
IX. Ic
la.S( CNI)F,\’I A\’I) SEi’t AICD KNOWLLDGI
rI crso mom debate ages in social science over the question of tra iscend ‘or ot situated. Much of this itt I n wie Ip Si lebare is di is. cmi by polita. al COflCLfij, But we are concerned hr not with a. hat drives ibis debase hut with how the debate cit ai I e, ima. d a a bent sue Ii is best to start with an un s
a derstandinh of how the political vcisio a operates in tcii s heuristics The logic of much of contemporary social scieii research begins with the recogmrioll that “X is true’ has onen meant in soci1 science “X is tine for white men of the mi Idle class (or, worse yet “X is Ctm.IC for a few collegc students i ikd Thousards ot rescar tiers have inSisted 00 iO /eStigtola whether this or that truth holds up in other groups, be they women, blacks Vietnamese immigrants old people, working class Latina mothers, or whomever. The heuristic moves of such work are fairly stiaightforward. The first step is to sa char
to”).
what appearen to be general knowledge is in tact situated or lo cal knowledge. The second step is to seek other forms of local knowledge’. Once such research makes the original move against transcendence, it usually moves direLtly into the realm here’ of the additive heuristic: “X was true thete; is X true Most often, it finds a negativc- answer, which then presents am opportunity for a wide variety of heuristic possibilities. But if the move against transcendence is by fr the most comnmn of heuristic moves employing the sii,uated/transcerm dent debate, there are some others as well, and ii is useful to il lustrate them. I shall use two simple illustrations, both famous and classic papers about issue-s of transcendence. The first is one of the most influential papers in the literature or stress, the endlessly of cited “Social Readjustment Rating Scale (SRRS) article be Thomas Holmes and Richard Rahe, This paper is a classic es of clin cause it made a very big bct on transcendence, Decad events that ical research had uncovered a long list of’ crucial
Di
r
F RAC IAL HR U RISTICS
LY
shaped people’s uses, It was obvious trom this work that for any one of these life events, “the psychological signiticance and aried s leiy wi Is the a iem” (l96:2 16). But the siotit task (“how important is at i i a I v
term danger; knowledge becomes ungrounded. But rrorn Out point of view, past bets on transcendence are always ripe targets for the situated-knowledge heuristic. What u’ire those system atic variations Holmes and Rahe noticed? Have the got1 i
and fairly di erse sample ut people Tli- results is Cic’ averaged as a rating scale. The crucial rnovc was as foliows
bigger with time? Are there subpopuiations that ‘ealiy ditt sharply? Every move on this scale creates opportunities to mo
is es
I I (in a
tl
h c ph
1 (JO
t
e of
S
‘rctc
mdc o he
ad sa
(>
i
iten
,
a large
pruu c o assign a different it is he degrce of similar
me popuhutons in the sample that is impressive. The high degree of consensus also suggests a unixersal agree— men’ beiwl en groups and amore, individuals about the signif in
betis
n
I
I
Ic cv nts unde
t
ly that tr inscends
e, sex sari Li t 105, t ducation, social cS I cneration American relirion and race. (1 96:21)
class,
Holmes and Robe bet on transcendence ças did Hodge, S egcl a d Rossi in the o upatlo il-prestige paper mentioned ha ci) I he subs uent literature employing t i u it these cale saluc
is enor
variation that was strength of a bet on
n though the issue of the remains for investigation. The
MOLIS,
ignored
c
can produce such a em In us ai 01 lit 0 1 sterestmp work The danger is that u w xi rly what that work is worth because we ci an iave h siglii o the ariation i lecided to ignore. Note, by the was. that the authors were perfectly aware of what they \\
transcendence is
that
it
crc doing. Indeed, the SRRS was backed up by a long clinical sic) radition o op back to the great turn hno iat 1 muss Ad P Meyer. Most of those who I I , i ii s, of to irse. This is a longsit! thu SRRS la ked that awaret
in the opposite direction. My second example is another famous bet on cranscendenc but one that argues paradoxically that because a tact is tran scendent, it is actually unimportant from a research point of view. In “Age and the Explanation of Crime,” one of the clam sics of modern criminology, Travis Hirschi and Michael Gnu’ fredson argued that the relationship between age and crime is so systematic and so invariant that there is no point in doing research that looks at variables that explain the distribution of age and crime. Since the distribution doesn’t vary anywhere,
st
can’t be explained by things that do. This paper is a quite un usual type: the definitively negative paper. For our puipose what is striking about
it
is that rather than making a transcer
dence argument in order to facilitate further work—the com mon move, illustrated by the Holmes and Rahe paper just discussed—it essentially makes a transcendence argument to strike clown further research. Age, Hirschi and Gottfredson say, is more or less uninteresting with respect to crime because it. relationship to crime is completely invariant, of the transcendent,’situated debate and its associated heuristic moves, I come to the end of my discs, s mon of fractal heuristics. My aim here has been to show how these profound debates, which generate so much noise an I
WITH MV DIaCUSSION
t 1.1 II
‘a
S .1
1
151 it I itS
in w iunt about paraht,ms and presuppositions, ri i u to let s punt )f mtw o Ps with which to pen c’ u o cc s C i t w idea, and arguilscnts. An astonis hing tw nb r 01 iciic tt, itt picces of social science have madc their to, A put c1y by pIIy op with th ‘se t arious deatcs in cxcit in wy a Thcie s no rccisu the Student should riot use the iu oc
Sat ic
dl a
tot Is Y ci
51000 01
p t a sc isc of ii sc dr bates and, above r hr ii mt as some hitig to get right or take a po otltetwise etch in stone bat as something to play
Ctiapieu S’eii UEAS ANI) PUZZL
itt ic
with. ‘rhest debarts are the most sophisticated tools for pro chic uì t ‘W so ial St ito :e And toy good student can get in on
—________
I.
__
TESTS OF IDEAS
lL OTHER PEOPLE Ill. LITERATURE
IV.
TASTE
V. PERSONAliTY VI. PUZZLES
WE HAVE NOW BEEN THROUGH four chapters of heuristics that generate new ideas. But not all of these new ideas will be good ideas. How do we know which are good and which are bad?
Part of the answer depends on what we mean by a good idea. Sometimes ‘good idea” means an idea worth retaining for the moment. (And it’s worth remembering “the moment’ could mean a lot of different things—five minutes, an after noon, until I think of something better, and so on.) But some times a “good idea” means good on some absolute scale. A good idea is good because it’s right or because we really believe it. Obviously, an idea has to see some testing before we decide it’s good in this second sense. There are several different ways to recognize and develop good ideas when we see them. First come tests we set ourselves. Critique starts at home, as everyone knows. So we need to dis cuss some personal ways to test ideas, to get a personal sense of 211
\iTO)D OF
I
i cci
he
tlic wen.
(J
fRY
sums 4ND 1U.IJIS
w rt h c’lano a ng and dcvelopi up. Second 1 t ‘ts ways yii p the idea out on others. p d a hers are pretty m an I t c it, wc ofte s behave as if
t’ ai
ictiur
oUl
1)m
us and ot
ideas weic targets. We dis mm tn rn ss uk the ohiiato “chat may be, but I think hue inttllctu life s neither a shoot-out nor a sequence of caudum opinions. It a mutual challenge, with equal empha siS 1 //aS and a cn Oh thoughts can help you see wi
wc
c Inc
v
i
t
C s h a
nt rn m
bad abc ‘st 0 i
YOL
I OWi
.lea’ witi respect to existing
topic. I
au mall, I said at the outset aced in the complaint I heard from many scitdeis char I have nothing new to say’ Now that you’ve read a hook s se orth of \vax s to bud new ideas, the literature shuuid no began seem so fniphtcning1 complete and compre hem core I to usc i literature in order to evalu hook
tam
at
1
s
i
ci iii
lop
ut
rat
c v
It
Yw
t’
if
y
sea,
ave t u iderstand how social a t to have deas that make n.
cial world as puzzling.
uttcatir g puzafemer
t
s
/IJI
CI
ing concern.
I TFTS Of IDLtsS Obviously, rhe brst test of an idea is to rd it OUL. CO lull pm o some data. In practice, most ideas come from ioolmn at data 1 the first place. Only when one is using formal methods do diem
come from dataless thinking, and eVeO 5s.Cf furmal nwthous the ideas usually come more from reflecting on comrnousen knowledge than from pure deduction. Most people get ft stimulus from thinking about data they ye already got oc e i pinical things they already know. Once you’ve go an idea, you need to try it OLs on Some case data. So if you’re an ethnographer studying welfare-so-v”ocd training programs and youve begun to notice that the trainer’ rhetoric emphasizes getting rid of race-stereotyped manual
-
isms you start looking for other indications of ovmt oc u vcri race retraining in othet parts of your data. If you ate Barringto Moorc studying the histories of the revolutions thai. led t
peupu v lie write ti 1 ISIs leads Us naturally to two broader topics: how we de \ elop good taste in ideas and hess we come to know OLif intel leireca] personahties. The question or taste is crucial, in the
modernity and you notice that in America and France the old rural aristocracy was undercut completely but ILl German. it survived and even dominated politics, you start lookir for other cases and see if you can predict wheteser a govcrnmcn
inrcilecrual taste i the best passport to good C p ss no gee ithom a r scans of travel, So
turns fascist based on now its rural aristocracy fared Juriug
ii IS
tLifl
idea
It
S
th
plus!
x
i
l
mor e
C s
Ic
r
in,
it
u
Ily i nyc way
c I ii or mci easy, These
an
L’ach if us has habits of iinking more dangerous or
of taste and person thc Issue of puzzlement. Having good idcas also incails being able to see certain things in the sm
.111CC
ormg ma, finally, to
c toph s,
rnodcrnizati on. It’s not just a matter of looking for other cases of a phenoir mien or a relationship you’ve identified. Its also a quesciol looking for other implications that your idea kas fec data. Sup OSC you’re a survey analyst studying married women’s jabor force participation and you suddenly get the idea that
ir
s
MLTiioDs OF 1)151 O CR1
1UES \u
I
need to guarar tee a skill set and an expe ftcord o that she can support herself in case of divorce. mix Iruri that idea that thi. lon erm overall trend in i h f pi I at o old or late closely with tOe long trim oserall trend in the divorce rate. That correla non tbllows logically tron your new idea because jf women iv i
y
nina
s
nellie
d
id
suffer divorce’s ceo nornic lossi, then (on oui arrzui sent) there 5 not the same r cc cssitv br them to have work skills as a precaution. You s I’ w it or c a ii spli at the ndividual level) at on vitf Ite nyc en uces unaffcc ed by divorce çwomen With inherited SI edith, SaV SI oflt has e to get the re t
1’
f
g
C
t
i
)tC
iliro ph v rk, and so you Fheor implies that they v I c .s c V 1 iL 1 cI icy will also hi. for other reasons, ot course) Both ul these empirical predictions can be tested formalls or nfurmahlv. W
1
d
y I im he ons nost ci any in formal ci In Is. Ii i his usually pi sducc clear predictions. The for rind arpunleilts in Siliehlitig’s famous iIIic,’omcth ri book have it i pin ons er tialhc ja ss, t social movements and ri a so In I J, ) se old sa the greatest virtue of for mal methods n their copious production 0 f implications. Flu
it
leas has e implication for data, whatever the c I s Yo I o i d get i t ha nt of continually gen at op these imi I icai ions and of continually moving your ideas on new Cases or data. It should become a matter of see I i e x p tI at p es 0 almost automatically when ii t
all
ik
Gould n ,iS 1
e “young
an
a
lea. ily friend ai d colleague the late Roger master at this. YOU IS ould utter an idle truism, op1
rei lways e uh
)
her’s harshest critic’s,” and
lie immediately would respond, “\‘ll, if Jan s Cft., dw, i oughr to be the ease that dissertation defenses ill D caner u graduate stlidents than having lunch with their Irience ) you really mean that peoples harshest critics ci dna thci so that older peoples harshest critics die nLLer 4idei I ple” and so on. Note coat ‘ust because an idea fails a law c these tests—makes a few had predictions, doesu t wock in i couple of casesoesn t mean that we musr chrcSI it our. Mi often, ss’e get new wrinkles in our ideas than way: We leatn hu to move them around a bit, expand one pa t at the expense o another. crhar’s what Roger would have bren suppesLa u making the generalization that peers are always tile liarshesn critics. It’s rather like decorating a room’ you try it, step bce move a few things, step back again, try a serious reoiganiz tion, and so on. This continuous monitoring and resting of your ideas rests more than anything else on a firm command of logic. Thc basic logical forms—implication, inverse, c0nvene and so un—need to be hardwired into your mind so that the process of monitor ing goes on in the background, like the antivirus software on your computer. It is a matter of practice as much as anything else. If your logic software hasn’t been updated reccocly a ru view might be worthwhile. Being able to quickly think up three or four implications (positive and negative) of a social theory is a crucial skill, in order to be tested, all of theSc ideas and implications must be framed in such a way ihat they can be wrong. In d great if your idea works most of the rime, but if it works al the time. you should start to suspect it. Its liked cc be a and therefore not terribly interesting. (Although sometimes it -
2H
SlEllioDs ci
tDiiAs AND
tun to tutu a etunru on ts head. as ss e se seen.) It is quite sur prising how mans researchers—even graduate students in their ru d s oj uet sic ts th an be wrmg For ex an Iccc c p c sals I ci e fbi I n go ig to take a iieounst itutiunal
mental hospital foundings” or
01
viess
Sr
s p ci cii i-zs exuc insults by omb ing a Goffmanian a n I t I iii CI H t n roar to lang sage” are ‘ion interest n heause thc do nut pnopose an idea that can be .
runi l hey h di doss is
r lassitving a phenomenon or, seen t is a r i ly 11 is r sy r rhro Similark a user ci prednates are iii general uninteresting, cccii if they are consequential. Thus, the idea that this or that a it a! t der ol -s ay, o ‘co intancy—is socially n
—
d ion t titt 1 irk ster ting Everything is so iall\ constructed in some sense, and probably even in a rela nv is st aug case Thi inn-i stiec questions invoke how ii uc
n nit
it
$
c
i
constructed nature
c
ut
u ed
of accounting
u Jai he conseqnences of experts are. Watch out for
unvc-rsal gre-U ‘cares. no ii
r to
hat good ideas have real ci rernat i c-s. not simple negations. It is better to be thinking “A is true or B is true’ than ‘A is true or A is icot true.” If you h a enui pi a ss c t to solve t, not simply to t ii
sa
s t
.
r
v
r
.ciia
pa
i
so
it
on
loesti’
work. Thinking
wnrhout alremniatises is a particular danger in cthnography and
hi
analysis. where the natural human desire to develop i Hi ‘U (i tu H pesen a cohesive inrurpne tar ii at the end of the research) pronpts us to notice univ those aspects ol reality that accord with our cLirrent ideas, I
t miii
Is
org
i
y
;
i t
ot
t
sr
ard iua itirar ye work,
I-’u ZeES
which often tests ideas against things char are caked, all coo 1it erally, null hypotheses. The majority of published quanriratie C articles do not have two real alternatives that are boih dcci a the writer. Most of the time, the writer s sympatisies are cleat well ahead of time, and the suspense is purely rhetorical. The writer’s ideas are tesred against random chance, even though nobody really thinks pure randomness occurs much in social life. All of this is wrong. An idea always does its best lf it has a real alternative. Always maintain tuo basic ideas about your project, and try to be equally attached to both.
Truisms are not a lost cause, however, it is a useful challenge to try to make a truism into an idea that can be wrong. Sup pose we wanted to make something out of the old joke that the leading cause
of divorce is marriage.
To
make
this
meaningful,
one has only to reconceptualize marriage as formalization of a relationship and divorce as breakup or damage,
and we have the very interesting hypothesis that formalizing a love relation ship decreases some aspect of its quality and hence makes it more likely to dissolve. This, too, is a platitude (not only in the nontechnical literature on romance but also in Weber’s formal version of it as ‘routinization of charisma”), but it is nor defi nitionally true and could be empirically right or wrong. its a much better idea than the bald statement that “marriage is the leading cause of divorce,” if a little less amusing Not being able to be rong is rhtis a sign of a bad idea. It goes without saying that having no empirical referent am all is also a sign of a bad idea. An idea of the form ‘The populationecology theory of organizations is really just a version of con flict theory” is not very interesting. One could for various
reasons want to write a polemical paper about it, but it’s not a
2iS
Maiitoo,ot Discossai
.1 i unlc’ssvc u.rr it ntotheempirical asstr a tlit the , puhit on-eco gy theoty of oiganizuzions .JflR3e “S Iii Se.,i13 tnc. conflict theory. Although somewhat S aiLie. tilia Versiuti has the bezinnings cit a good idea in it. The nt iou us lasihator e ‘rcsL I’hc second is an rzq Ia on un clist yofscc i scienc A .ood idea, then, ouht to haie some referent in the real world This is nun to deny the utility of pure social theory. but hcv ma yof ialtle r>e. isistsoftclabeling.Allreal C. Sc ii w cal wot ii dii. attc.mj t. tc make sense of siciat isc rid. no niattti how abstractly construed. A stu dt.nr is stlI acli ised to stay clear ot writing pure theory. It’s an )lleti iitat’ to v uq. s c I ,unv t )lL esiculdnotethatitisalso a bat ,ti if idea works t o will or too quicldy Usually this means that the idea is just relabeling something that is already known or accc.pted When vo z have an idea—say, that a certain tic el t is a. idal y ott -mo t of the time you are chic gt act La thc b havnrisregularandcon tfll sistent. The notion of norms doesn’t add anything to the act of reguLiriti unless it involves the positive assertion that the reg Ian i u ‘ed y )bligatory, c nergent rules But then you as r ci’ cia iorstratinl that these rules actually a ist indtpendc. it at the behai ior thei enjoin. It’s this existence qucstion that is trudal, and if you don’t fight it out, your work jus• ros ad ig tar labels for so nething simple. I L It a&ta ala vit insocialsc.ienccbecauseit’s a isa ap aning ovels thouc havins to do much. Often when yuu’ic ust read a new theorist, that theorist’s language is iLi 5cc-rn supremely compelling because of its novelty, but ov
a cx
ii
IDEaS aND
Puzz.a
219
then it will turn out to be the same old stuff with ncw nanaco Much of sociology fell in love with Pierre Bourdieu’s wo d rut 1kv, for example, but most of the time when the term is usea b) others in sociology, it simply means “regular benas iot’ It’s )U5L a new word flit something we have talked about fbi a lonb time. To the extent that it is new, it involves the assertion that. the behavior involved is in some way self-perpetuating, that doing it regularly creates the possibility and me likelihood that we will do it even more That is a stronger assertion—on that must be considered empirically—but of course it too, i s mechanism 1 quite old and kmilist (Stincncombe called th “historicist explanation,” for example (L968]i Ideas that reclassify something are also usually pretty unia teresting. “Social work is ieally a profession ia an intercst ing topic polemically, but as a research idea it is going to be interesting only if by seeing social work as a profession, we can understand something profoundly puzzling about it. Pot example, we might think that demonstrating that social work was really a profession might explain why its practitionets work for so little money. But then the strong form of the idea would be some more general statement, such as “People are al ways willing to exchange prestige for salary, and being thought professional confers high prestige.” This is quite different from “Social work is really a profession.” By themselves, then, classi ficatory ideas aren’t interesting, but they often conceal an interesting question. So the proper cballenge to present to ‘t classificatory idea is Why do I think this classification matters What is really at issue? Note, too, that in the largest scale, re classifications are often analogies, which are among the most powerful of heuristic gambits. Saying that the family was really -
ii
I
I
s(
1
Y
liwas iso Pu ,
in bt Let
i
t
iit
kc
og
ii
ke iriy other heljed win Gary economics,
d i Ie: di ens ed so f r are shortoterm na [1 csc ire iot t lie only o mes. One ot the roost impor te s u a ‘nod idea, itadless to say, 15 that it still seems I
it ri
(a
mr got
ss en got
tic
been Union sonieihing h
L
n
Fur
to
tia ubei I nze
The cr
ii
\i
w
trr m ii w
gt n
else tar
en ugi
bvi nis Lu t
a
hc next d ty oi ss hen you’ve mess dais and come back to it.
t r
,
piactice va often
to] lows the
forget
it.
that no good v t i .i sinek i ig he uractice of genera ma s or colt ye students dncluding me) notwithstand ing. If it In ii a ft r m in a really ai away from it, so that von e t1rgotteil iniportant sarrs of it—you can’t come al is
\ en
that enables you to see xl or tint. A good dea is one that sta s faithful ‘at lit with orhcr i leas There’s no other way to d it
it I
hethc
t
whsmm
xxi
est ha ti
ii
In the Ii
ni
t
corollary
t
it
Iurside
haul. the best
s
‘e
pe rsotial criterion for a
good idea I e piniosopliem Tm e Lakatos thirty yeats ago I 9U). A good idea s one that is nondegenerating.” i n dir is I dyes i sc to in t deas, to more puzzles, to ti e
nt
n.
n
I I
iore p ssihihits, Its cur e is upward. At the same time, it a ‘t her s it w th e rid lenly eserything is solved” eel irly that conies rrom ii tnsms and rt labelings. A good idea is
I
Ic
e’
55 ant it to
La and
to u
I s(
to a
works
aes
docsn
t work when
we
we least expect it to. I ii iatriv, nc knows good ide is by the solid feeling they cii user unit A good idea will make on feel secure while os to t he grunt work that tc k 5 tip the majority of research Lam i1 1)1 tnt ai SC da spending lonel time in sometimes it
is
hen
La
ethnographic settings slogging thiouph archivd d )LUI me When you do these things with d good de if, our i.ad,
I Ca
know why you are doing them. That gives you the uflfl.i I and endurance you otherwise teen. ‘v hen aL uoU guiding idea, you feel despeiaLe you hoya roar sonecriuw ill emerge magically from the next pae of coetfciaritt the next incomprehensible docuirient or conversant to Indee’ students often throw themselves into the detail work to hica from their feeling that there isn’t a big idea. Don t. \ork c idea
the idea, and the grunt work wd become much rrmolc beatabi 11. O’rHER Propi,L Once an idea has passed our
OWi prelsminary
screening,
it
needs to be tried out on others, Sometimes this exercise will be formal, sometimes informal. From the start, trying out ideas on otliems is ditfere it ate trying them out on y ourseif. Others do nor hear your idees the -
way you hear them yourself. Its not jLLSt that they disagree something like that. Rather, inside our own heads, our ideas are sustained by a lot of assumptions and things taken mo granted that we are unaware of. Ii s like singing. An intra ment but the voice is heard by performer arid listener ni cli same way: through the ear. But your voice ‘caches your ea much through the inner passages of the head s through
tOe
outer car, so it never sounds the same to you as to someone else That’s why singers are always listening to recordings of them
sels es, tryitig to hear what others hear So, too, with ideas. They neVer sound he unite mu others. And it is crucial to remember that for all save a handful of Lu it is their sound to others that matters: to teachets to readers
1DCA ANu 1’ij0LrS
pi oksio
or popifkir audienec c ma wish to persuade. Ih inure iirroiant ciffloflg us find this a hard lesson to learn. o IP ty fa r id perfect 1 isightful, t I t i It-! opP in t car her hen when you nrsl-nt them. iou must tind a better way to communicate, ) he ist 0 Wi 1 be no S 0 p it y (It lot sou d the sante to others is i,il
.
a
sO\ inc that coo ssill ads ays find yourself leaving out rue l aspt ts of our k s hen y ni talk to other people In ( ly t or p t wha he people say in re ponsc to coot idea ts hat the add, what they want clarified, s hat rid niistuiderstand-- that coo will he able to figure out (S Ott id ssc tiol r s ci he ea So listu carefilly Ot C S C lieu iot arifi tioti At th stunt t tine, however, it is tru that an idea that re tire i oa an nit f ox lanat n is orobably nt t a good t at of
a V st ely jt ho 1k, n t e teed for expla nation u telling too that. ote that these two arguments push in ditieront direc tiolis. The first says you should figure out from o Ii u i d CXf La 1 t add or to nove in order to naPe votti dmi rL )n tl at argument, the more problems others Ilevi with c our idea, the more you can figure out about Tb ott so hai I v liv o cl oo inch explaining, iii
a
I
1000
wttli
to
tht r
tn
hal it, p
o
iso
ooe
lit
)hletns other people the weaker your idea is. The skill of learning -and is skill just Ike any other—lies in oi ‘P u i in se vi tradicroty processes i tire
urree tic kite i or
first t
is
the more important
ot
a
if nl
s a s
nit
ia
the
No matter how people eat r under-
two,
stand you. its nut due to thea stupidity, diOecst. il’• so on, but to your inability to articulate yooe ida pruperkc The reason for making this assumption is not that it is necesseol correct; they may well be stupid, disinterested, and so .n. hi t ,
the assumption enables you to get the maximum oat of rherr Every social scientist learns this from dealing with hind ref r ces (people who view articles for publication in journals usually they are unidentified colleagues at other universities. One’s first reaction anger.
their criticisms is to scream and yell n But even if thry are fools, the way they misunderstood
you tells
to
how to write better for others. Some of us don’t get angry at negative comments. \Ve find them overwhelming and collapse before them But even if you you
believe someone who says your idea is junk, you should assume that the reason this smart person tnought your idea- vas wrong was that you didn’t say it right, not that the idea itself is bad That enables you to use others comments to improve youi idea, to raise it to its highest possible level. It may turn out in he much better than you thinle. The things you learn from this process of clearing up others presumed misunderstandings are fairly specific. You learn first about intermediate steps that you left out of your arganient: these are hidden stages you may not have noticed and may in volve real difficulties. You also learn about the background as sumptions that you make—often as part of your general way of thinking about the world——-that others do not necessarily share If you are careful, you will also learn a great deal about th Spe cific (and often contradictory) meanings that people give to words. For example, I called my book about professions Tin 5steni of Profeisioiis,
more or
less because I liked the sound of
I
‘4
o s
Discui sin
used for au early papet on the subject. i v b ik had a title allowcd m to feel it was more luring the five yeats it took t wr te it But I 1St v red that mana p ipic unffr horn the word cys t t te hook arguc s that tha re is some kind of grand inm bali uid the wa piofessious stork, as it all of society’s pores a us weta pai-t of a huge plan. in rct, the book says pre Sc the ret LfsL of that. hut I had torgotcen what the word sjs , means to most readers. Thus, one should remember that social science is a place w here most of the basic concepts— idncit, strue cure, culture, nation, and so on—do not have anything like generally accepted definitions. Indeed, this is al a u.’ the tint place to look for misunderstanding: the defini tions of the words you are using to state your idea. Note that .1 haven’t said much yet about whether ocher peo ple think your idea is good. I have talked only about the fact that they are likely to misunderstand it. It is important not to take other people’s first react ions to your ideas at face value. TI t ‘nu s berhr they think it’s a great idea or a bad one. If c y thu I t s great, it could easily be that they don’t under tan i note than ou do and that it’s really a bad idea i o 1 c misunderstood. Or it could be that they i t ii itch and are agreeing in order to be polite. ii b tl a ou hav€ an overbearing personality and e n I e ause it’s too n nab work for them to dis e I it it ‘I t icy think it’ i lousy idea’ they could have cia ) a wg i icr; cii y Id nae ui drstood it hut is a I t a s, h Id b lisnussive people who never v ti i bu thernselv i so n, don’t take the first vraci s Oily 1
a
‘a
.
IDiLAS
I
Ni) Pt. LAcES
The first hint that you are past the stage of first reacrio comes when you yourself feel confident that you cai state youi idea clearly effectively, and bricfi, The trucial moment conies when other people are able to repeat your idea to you in such a
way that you recognize it and agree with their presentation of it. For an undergraduate trying out ideas for a course paper this is going to happen after talking to four or fit e people and hammering out the details. For a graduate student writing a dissertation proposal, this is going to happen after many weeks and many drafts. Whenever it comes, the ability of others to restate your idea clearly is the watershed. Then you can start to put some faith in their judgment. Of course, you still have to factor in their per sonalities. Arrogant people like only their own ideas. Negative people don’t like anything. Pollyannas like everything. You have to reset your meter based on the person you’re talking to. If the negativist thinks it is nor the worst idea he or she has ever heard, maybe that’s good news. This relativism is true, b the way, for faculty just as much as for anyone else; there are faculty of all types, from thoughtlessly arrogant to hopelessly negative to mindlessly supportive. Although only their own graduate students really know how to read particular faculty members, it’s wise to be aware that each has a unique style. You can probably guess most of it, and you need to second-guess the rest. You will find that it is useful to build up a small group of people who air sympathetic but thoughtfully critical (‘Ihe way to do this, of course, is to play the same role for them.) It s also important to keep pedalling your ideas in many differe it places. Your friends get used to you (they start to know, and
1D5A5 A “10 1 uazau
Ideu assumptiot ) a id i diiig a group of pmople H your luslopii p iJeas
ik
1 ultimately get
i
can do It
Is
who S
will listen to,
di
iIost
inipor—
alit the hardest
who become sLrious sd hats, the nit mate test of a dea is tIn taxi-dri r itsi if von art or your way someit S
a Lere
present
tc
L’r is
hat son
Jr let
n
idea ud von cannot
our
talking about
aiC
the person
iii
the
is
adjacent
hi
‘ii
iuklao
cx
1 enough so that your taxi
aircraft
SCat
can understand
ht t ‘s interesting. ou don’t really understand your idea tet. You arent ready to prescnr it This holds no matter lion coiiiplea your idea iS. if Vuu cali’t state it in everyday iw for an as crept person with no special interest in it, you and see is
or
undersi
end
it
sti use ihrniahsnis.
I
u
e talked so tai ahot suhmittnig sour ideas to your own nt and sour lriens judgment. But what about the
a
1
isi
if
a Ii
r
vet Even for those is tirking in the most ab this is the aDsolLite test of understanding.
t
mis idea
i
tins
i
u i i i
t
it
i
y
10 fltCi
is the hardest hit. It always seems that every ossihii hL said has hero aid There is no lac to start. Moteove when you do think
t
and new, tht I ter itu e s reaction (via 1 or ismissive omprchciidin
i
i
t
i
alize is 1 r tel
1 w
ions pun1 ished worki [‘or tin—
b
i
i
I i
dcl ha r
irid
ti
“
i
.
piobibly true that t least it the level
I mu is lilu y to be think pr vent lieu themselves f om saying id again it i w ways with new
evidence, in new contexts. indeed, that s what a huge propu tion of excellent social science scholarship is: saying the o[u things in new ways (if we didn’t say them again and aged’ we’d forger them, which would be a had thing.) \*‘har taculty know that students do not know—arid what enabler them LU accomplish this turning of old things into new ones—is die conventional nature of Lhe literature. They know which aid things can be resaid and, indeed, which old things neiai to ne resaid. They know how the literature defines the border be tween restating something and stating something new. This system of conventions is mostly invisible to under graduates and even to most graduate students. Suppose you take a stratification course. You read the stratification litera ture. There are a lot of questions that occur to you about that literature that most people writing in it don’t seem to worry about. For example, why should we judge somebody’s success by how well he or she was doing in a particular year? Why should we assume that everybody judges success by the same scale? Why do we think about a family’s social status by ask ing the job of the husband? Indeed, why is measuring soc,ai status more important than measuring, say, personal judg well-being or satisfaction? And so on. Occasionally, these things do get written about, of course. But in the main the stratification literature goes on happily envisioning new
ments of
puzzles and issues without thinking about these questions for a second. They are ignored by common agreement. Yet the seem of burning importance to an undergraduate. and iightl SO.
As I have said throughout, literatures work by making mm plifying assumptions about some things so that researchers can
Mi. I Hi 105 oi Dot. 1)5
-
LilS
1i)EAs -,J5i kLLtS
Oilcx anal5 ses ot other dot gs. That’s the nature of the its osol I to do ‘o si. renct—by any method ho ph ying assumptions They tare h N 5 rt tiring pie tro bogging down in
oL
w
.
rimiurlu, c mrvv anals sts make
about how to behavior, and etluoraphers make assump ons about ‘ion informants do and do not twist the truth. And I issui is iOu I go wel ) mid the methodological r i t y t P c try cc 1 f t c substance, assumptions
actituuu rcLire
tht
(f stratifi anon test arch. F Os Lit V hr on these cons en nuns so well that they are usually UllC LtliO\% Ore ()t them as coflvuntiofls, As a result, many ideas that Oct or imrneoiarels to underyri duates seem ridiculous to I t ltv. ‘W i ss cd y a igo t h hat did hr matter,’ “That’s st lI(
I
‘
F
it
it
.
s
at
set. in
I
t.onvct
h
d an
mc ut 1 in substancc, and tr art typi oh reactions to ‘
I
t ntr
lii e ohs ous questions to a bright undergraduate. 1 01 these may turan that the fa ul ty member has forgotten Al that sour idea is a legitimate question because it has been set a
i
t
r i ai 1
I
ialls by the lirera at course ht
(These statements don’t hey m y
t
an
hapteis I’ cc through Six, a good idea ,dies one r another k these conventions. But a good jda doesn’t try to push several cons entioris at once. So. to continue the stratification exarnplc, it would be interesting to ci
o c that
a.
s hx haj I ns to the standard r datiunshIp between educa xi i
h’s
nc conrrihute
to
a
if
‘
is tin n I
i
ec rh
Wii
s job prestige
tot of tan fly social status ith, Such a study would
soon av agt of ) the literature precist ls by opening up one of its
conventional assumpticus to rurtiei analysis. but. soc n° 5 sa C changed indicators on both sides oF the relationship ii t s., moving to the wife’s job prestige as the status nidica’oi U. also changing education from degrees or cats in sa iOO standard indicators) to a true outcome variable, lific
ti 3_-t. i
scores, for example fon the assumptioil tnac the CA’i acccoj.s measures prior achievement and schooling irare than u sures sciioohng-independent talent), This would res
in s
attention to the college bound, as wcrl as changing s ceptual idea of the meaning of education, A d now U e ma begins to lose its relation to the traditional stratincation inoic hcarion 1 cure, where it is conventional to think abOot scrai terms of breadwinner employment and where
it
is
in
cslstsjmar\
terms of credentials twith t1ier nb e ii. rect link to occupation and income) rather char achicvcm i± scores (which measuri a less actualized but pciIiap inot.c consider education in
eral resource) So you would have done a doubly brtiar but one hanging in midair as far as iltetatutes ate aonccriceJ. Conventions play an important role in all marhods and in eratures. A historically inclined student might b inteiesteb changes in the patterns of lawyers careers ovet the twentiecl century and decide to approach it by reading twenty oi. biographies of lawyers in order to develop a schematic modci lawyers’ lives. But a faculty adviser would probably
inOhe sl
largely conventional judgment that the student should either toward a quantitalise analysis, digging up simple
inO th1aL
matiun fur a much larger but random sample
throughout the period, or toward a detailed study of tw c three lasvycns suitably spaced through the century. Thc conven non is either to be fuily scientiFic, with a aefensinle
stiatefli
-lii
\tei
non., oi 1)is
u
SWS
1btAS .ND tCLLis
and agr±d-upon career measures, or to be deeply interpretive. Ytt against the first plan. one could easily argue that changes in 1
n
w
of
mea
I at c )dlng aregories, like orL g lu Li’ irm, mea..t so .ethig completely differcot in 900 than they did in 200( in that sense, there is no bit’ rego at i of bs that v 11 enable meaningful cod
o
rh
nt
t
Ai aguast ic se ond plan, one could argue that its samplIng is so arbitrary that any conclusions are spuriom onethe ess, tfe conventions are that you probably i d e t V 01 Ot dir inrcrpretiv version, but you will hasc troubh writing about twenti to thirty lawyers’ uses iii the middle i
i
ng h sc i on a —d dami ed if you ont hing y w iatevr r else ii does, the best work s tie c set )fl5 r rhi same rim be
mv
her o these damned-if Everybody agrees that neatly always overturns the general preference is
io
cspc ally whe me is starung out So you n obey [he convennons and have people think you unadven rurous or disobey them and have people reject or rnisunder— d r a un ) stu is, the best way to learn th research Conventions is tf coLirse to look at current work, and the easiest ss av to generate feasible ideas is to clone an cx i ig jet Iy angir g one detci getting a new variable, a ig the rime ernd exarr med adding some more cases. (1 his is the additive heuristic of Chapter Three.) But this in vii s ii ha of inidi y..
I s t vay it I t ii’ a iei a, wnich is, after all, the dilemma of creativity in social science writ small. It is impor tant, nonetheless, to know aoout th problem of consentions, It atrs
i
I
P
o u
-tsr ndi
t
ho v the professionals
in
giVei uc, your world—meaiiing people whu than you do. be they aider students or rPcair, —v,H va
your ideas. Often, faculty push students toward blh wi i ventions for the very good reasati that un onvenhional v much harder. Students research plans are eaten
ci
uncealisLic
a
the extreme, and faculty are trying to escou ‘agi studcr u terests while helping make the research more feasible Urg n students to learn conventional research models arid to cvnii conventional papers is a way of doing that. A student iced: be aware of this complex tension between convention, origin ity. and feasibility—and to be vvilling to make some rump mises if necessary. IV. Tks1 C Cons enrions and the problem of knowing them bring us to the matter of taste 3 Lidging ones ideas becomes much easier wh
one begins to acquire scholarly taste. By taste, I mean a gen eral, intuitive sense of whether an idea is likely to be a goud one or nor. It is of coarse important not to become a slave ones caste, to try new things as one rites new toods. But dcv 1 oping a sense of taste makes things a lot easier. The foundation of good taste—like the foundation of goud heuristic—is broad reading. It is not necessary that all th reading be of good material, only that it be broad and that it always involve judgment and reflection. A musical metaphor is again useful. A good pianist always practices not only zeci nique and repertoire but also sight-reiiding. Broad e..drig L social scientists is the equivalent of sight-reading rOt pianists. A pianist practicing sight-reading grabs a random pieCe music and reads it through, playing steadily ott in spite
ci
f
1
id
1
on
St
CRY
So,
tc o
iDeAs AND PuzzLes
should you jus pick up pieces of social wiener ui sociology or whatever and just read through Ii in, wherhi r yo i knois the iet i Is of the methods, see the Cl ip xit s I tlt dl L in ut )r c n like the style of analysis. I t ot vious uav t i do this is to r up recent issues of jour in sniP s
i
in
r
1
1
,li he 1 ‘ion kern many things from sudi broad reading. You learn hi ,o ‘s f r wa Ii ii disci 1 or You learn the conven t
tu us
t
ir
if ca h zone. and too hgule
out n hich you like and ott don t 1 Pc. You learn what interests you and what d s t f UI 1 I n let your interests dictate iou y u von r react Olb usc as you should disregard, when you are ‘Ii cc p c v i ho i v th vhi h you disagree. When no tind you don t like a paper s methodology and you think its to ce s d ii’r ial SC 1st force vot tself to go on and ask what there s that u on get out of it- perhaps some facts, a hy pothesis, even in the worst casei some references. In the best di p iaty jt usc s, cry err Ic ‘11 have something to reach on. even hose ar it ics that li completely outside your own P In Wi
icli
,
—
fins is also
i usetul rule tor seminars and lectures, which are an the us tu1 lac tc dcv iop ‘on taste. There is no point in sit nig thu ugh a lrcture ot talk whose methods von hate, self righteously telling y’ot rself about the “positAist morons” or ii sm odet i ill ‘I it r what ver, All that does is rein— to] cc v )ur
rej idices
and troth
Judge a talk or a i us I tri ig to do. This hard, it. you will gain a much surer sense of both tO stt ng us nd rut we knesses of ‘our own preferences. You xi ill bcct ie ahit to yatF er useful ideas, theories, facts, P
c” but by is orkint’ at r
‘
you nothing.
and methodological tricks from material that used to tcli you nothing You will, of course, run into plenty of bad stuff: bad books bad papers, bad talks. The symptoms are usually pretty clean pontification, confusion, aimlessness, overreliance on authoro ties. Other signs are excessive attention to methods rather tha substance and long discussions of the speaker’s or writers poSi tions on various important debates. But even bad material can teach you things. Most important, it can teach you how to set standards for an article or talk on its own terms. What was the writer trying to accomplish? For the truly terrible, what should the writer have been tr11ig to accomplish? This last i 5 the ques tion that enables you to judge material on its own grounds by imagining the task it should have set itself. Of course, it is also important self-consciously to read good work. Oddly enough, good work will not teach you as much as will bad. Great social science tends to look self-evident after the fact, and when it’s well written, you may not be able to sec what the insight was that instituted a new paradigm. What you take away from good work is more its sense of excitement and clarity, its feeling of ease and fluidity. Nor that these are very imitable. But they set an ideal. How does one find such good work? At the start, you ask people you know—faculty members, friends, tellow students. You also look at influential material, although—again oddly— there is plenty of influential material that is badly argued and opaque. Soon your taste will establish itself, and you can rely more on your own judgment. There is no substitute for prac tice and, in particular, for “sight-reading.” You just need to learn to read and make judgments, always working around
vIi:i no
no p
c us
hwxs AND PuazLEs
paL u
C
d work f om w wk you sim
self where you are on each scale. In is true, thougir, as Mr Va in every dus u says in Pride and Prejudice, that “[t)here is non a tendency to some particular evil, a natural defect w ci h .
lop rd judgi
t
ib o
e
Ia s is a crucial step iv n all n c n nrs scattered i
iv s Fvc ()ugliOut this I p ter u Igir V i w s ideas s the hardest sk of all, The v i be im skilled at it is to acquire genera] taste and i he ueful y at d painfully t irn that taste on your os ii thinkiny, Th k’ll of rn ng to find good and bad things in the work of otlu r i h I cc h Ip in finding the ood and bad things in s oui owi von i
V Pcxsux. c; Ti 5
Part of developing a taste for goon ideas is getting a sense of our on n urengrlis and weaknesses as a thinker. You must eventuall\ learn no second-guess sour scholarly judgments. This sec (md-guessing comes from understanding your wider chara rer as a researcher and thinker: your intellectual person ahir, Your intellectual personality is based on your everyday character, of course, but builds on it in surprising ways. The strengths and weaknesses of your intellectual character deci sively influence the way you evaluate ideas and. indeed, every thi hg about he was you think It is important to realize from the start that every aspect of Our mallet roal character lik esery aspect of your everyday hai acrer, is both a strength dnd wcakiw ss in the everyday ss odd, s hat is preci ins loyaht ii ( Se context is mindless ob stinac in another, Tht sai, Lcd less is true in the re search w uld what is darn g nak ny at one point is dangerous s i ucness t anoth S a us c inc de somc character traits as ella t tial virtu d via s Yr 0 1 d to figure out for your-
riot even the best education can overcome.’ F ii of as least one great weakness; understand it, and you cons a lonr way toward controlling it. Let us consider some important qualities of irrellectual character. Take orderliness, for example. It is painfully obvi u that orderliness is absolutely necessary for any major resea c project. A keen sense of research design, a mania about careful records and filing, a deliberate discipline of analysis—these atc the avatars of orderliness necessary to undertake any major Ccsearch enterprise, from an undergraduate paper to a multiinvestigator project. But orderliness can also be important within thinking itself. It is very helpful to have an orderly nsind. When you write our a big, long list of ideas, it’s very useful to have the habit of rearranging the ideas every now and then into categories, changing the category system from time to rinse, to make it better and better. So in writing this chaptet. I first wrote down dozens of free associations about judging ideas. Then I put them into a set of categories; there seemed Cu be some about talking to yourself, some about talking to oth ers, and so on. Later (after adding sonse niore ideas). I put thOsc categories in an order for writing, figuring to move from the individual to the group and the literature and from the specific qualities to more general ones. Once I saw this emerging out line I saw that I needed to split up one category and relabel r few others. I then sat down to write the chapter, cleating care gories within my headings (for example, the different types o
personality qualities) and setting those in order as I came to
I
vi et
Ui
Li
ii CC
SC
L
IDIOk NL)
Fl
sot icd wh 1)
t
s a use ul s tatcgy for m because I I vu got i long list of somewhat related ideas
it
St
i
C
Ohs 1005k orderliness or thought is a good c]Liality in mild I ut a n mi in ch r u ris ic t F as problems It is it the riot of tite rechissiheation papers mentio ned earlier, pa pets wIt thy urn is to pull sorn idea 0 ph nomenon out of lie pieco iltolc and put it nto anotlict. Ptgeoi ibolers also have a hud t mc ii nch ng phenomen 1 ooia It puzzling Their main on
ri
$
Ctt hg
hi igs
th
r
boxes, EveH worse. stunt n re’s th pigconlioiei has personal idiosyncratic set of ox t ii ii C0f t nt Ii n iu 1 pigeonliolers often r ike tilt t$ and detort them considerably to get them into as lab e aj te it [ci tin ,s ambiguous and open. on
spet
‘
‘Pci
this ahn icy
to
levc things
unresolved
is absolutely neces
tu Ii t he ‘1 bus iirdcrli ness is a i uah lv char can cut both ways. So. 1 t 10 s I al p rice Ear, loyalt to ideas. On he one hand, ceitain 101 alce to clcas is a gleat strength. Often a good idea ciusflt show its colors Or a svl ile t Ic c “sists r evades. Loyalty i \ U ide s ii the face of various kinds of criticisms is a streiich ‘\t the same note it can hecc me a liahihnv. You have k wv vi gi e t p on idras, wE en to set them aside and riot c on, Most of us have a little museum of cherished notion s at to I e ed fo his or that reason, much agait si cur viii In’s OK no keep these ideas in a personal mu try
,
u v d 1 b 4 l tiire A itnhc r Ibit i ty that cuts built w igs is habit. There are in h ii Ha a v t t sd’ i is useful to have the habit 1 out unicicoily vs rib ing the log Cai structure of ones ideas u ‘
Paees
before considering them further, It is useful to have It hab I listening to others as Well as oiieselr. ft uSera to Lsow ct conventions and usual disciplines of ones lesea ch arca, ii same time, habit caii become parali LOg. it can lead one cc cept dead conventions, It can hide Ui gaihs of Imaginauc completely. Also two faced is DreOdlh oi interest Teie is comet ±u wonderful about a giect breadth of interest, an ability to scc the many things relevant to any gisen issue Breadth at ifltte can open the doors to powerful analogies. It can bring d tant methods to new uses. At the same rinse, excessive breadth ann depth)
of
the need
inteiest can, like habit be utterly paralyzing. u the to
say everything one knows in every single papem
-
ts
the most common single disease among young researchers. And excessive breadth of interest can jead to a variety ol artier pathologies, to pigeonholing,
such diverse 3 interest ;
because only
thaL
can
deal with
to arbitrary argument, because it
Will
bring things together somehow; no slicer paralysis, becausc the range of topics is too great. Related to breadtn of interest is another uaiiry with vary ing impact: imagination Ii may seem odd at the end of a bool aimed at increasing imagination to mention chat it’s possible to
he too imaginative. hut
it
is worth reflecting o
i imaginatioi
There is more titan a grain of truth in Edison’s “genius is
99 percent perspiration and F percent inspiiation” Ideas need to be worked out. The svorking
Out iS flOl
easy.
it
to
is all too
one’s ideas becaus” “others don the imaginatis e links I have made,” and so on. wiost of the
comfbrtable to avoid recasting sc-c
rime
when your ideas don’t survive the tests presentcd ca [let
in this chapter, they’re had ideas. If they don n
sustai
nindeed
Ineas
1.1501
/ Cs
tendency to see things as alike (by
cudlcies. Li s best tà alLernate haLO eel. 1 with so nienI c styles if you Cafl. uaiicies of 1 ablicly cv iGent c Come ruicto LILe moic intellectual personality Ut Iesc, by fat ihe iriost rciportant selt-conndencc in genelal everyone in acaocniia Lliii&KS C 01 site can judge the self-conhdene of orheis by noting how macic
making analogies): others see things as dirferent bv making sonriel irc rots of Bell the in rio i t Many ‘ears Li ‘e Ic ic b s trot g hat they nit similar 1 ) engineers worked for to c so
they talk. In ldct, there’s couch else involved in talking too much. People can talk a lot because they know a lot or because they come from talky cultures ot because the ae try ng t peisuade themselves that they have something to ay a in
D bosses. This quai or seeing (]ittereflct-S is captured in the oh seeing siiiiilariti that m poiog is a niatlitmatician who oath tus jt (0 U ci H y, y A louglinut a d a cof iice a plane intersecting poP up liv t I alen
some cases, s1mpl’ because they are airuganc. There is probably nothing’ more importara than cuOiifl to a good sense of your own degree of seir-confliene. It pLCity easy to tell it you have too much self-confidence, if you can’t quickly th nk of two or three people who have recently taugi you something important about a topic y ou tnought you ktievv
cdl uut tor—aretu1 elaboration, rhcv’re probably just
flimsy
anJogics ss ith nothing in them. So watch out fur congratulat irna ii ‘moo can I e a cover—up for youru. on y ii -
it
e
r
p
Sonic people has a
S Cusses it\
idm
so
.
an I D
(ditre rences)
alits
engineers
lilicrence
at issue
tbr
,
th in can intersect
ris a
disconne ftJ parts, something that
an’t happen with a pencil or a tennis bail, which are topologi n to
ch orl r hut
to doughnuts or coffee
) Hi
ati emauc us. Things
c u
us iou c aljk to them.
C C
ooi
;1\
tilt
As the Bell Labs
,
makes clear, this quality of seeing
feterence
silnhlarities or differences takes on much of its color relation al
tram tl
habit
I
d
p isioi
ry a 5c
y
t
otlic Ds
t
(
arounc II
tool
imar
on. To be an S person in iat you’a treated
as a vi
in a group of Ss can de
be
tei on
plodding pgconhokr cc as Someone with his or tin ground. It is worth Irving to figure our your
‘cal Ii
t
ou
as
a
l)o
i
is
t
b n
up
hxh I ni
ci
tis d
ir si
in
s t m ti
keci
antics’
build down from
Qi do you see differ
11 he de uls straight? As
-
ost of the 1 self-confident. If you do ir talking in most of your classes or in groups of itiends, you arc probably too self-confident, if ou don t have to rewrite mon P your papers three or four times, you are pobb[y too sd; well, you are probably
tOo
confident. if you cant rake cniLiciSisl, you are probably iGO selt confident. Generaill, overconfident students ate unaware of their overconfidence, If they do recognize tneir tendency to domineer, they may put it down to other things: educational advantage, prior study, desire to help others, and so on. By con trast, students who tack selt-cunndence are usually quite awace of their timidity, but the’ often do riot see ic a their problem so much as that of other students, who (they chtnie domineer in an odd way, people who have too much self confidence have much the same problem as people who havc’ too ittle
IDEAS AND Puzzars
ti e feedback necessary to learn from others. e much self-confidence don t pay attention to s e to say, even it they give then time to say it.
kets or networks or nested dichotomies or whatever oni tasc it don is for the moment. It’s like falling in love. Everyrhitig read seems to ht the analogy perfectly just as everyrhing about
ore lose most ut what other people have to tell I his rriakes their os ii intc Ilecrual de elopinent harder. y air onI} As good as their own ability to judge and im pros their ideas Tl]e\ dun find out about facts that others happen n base noticed They don t hear that others have tried
the person you fall in love with seems to fit perfectly with vou interests and desires. Feelings can be just as sriong for otne
Neitl
n
I eoi vi c c i
certaif
ntellectual paths and tound them Liseless. It’s as if Mark Granoverter’s job seekers cm Chapter Four) were trying ut
hnd jobs .m their own, without all the weak ties—you can do it, but it takes a long time. The short-run reward for such to
ople is aiwas s being right. But the long-run costs are great. 1 hey deny thcmselves the help others can give. Only truly out standing talent can make mLich headway with such a handicap, and even then only at the price of incredible labor. People who lack self-confidence also iuse what others have to tel them, but not because they don’t listen. Rather, they lis ten too rnuLh, never risking their own ideas independently. As cli v often end up following the lead of something outci
I rs
s es——a book, a friend, a teacher— and never really i for themselves. They can do well under certain
I a
ii
e
I be
c
i i )
(
ci i
c
otis—particularly if they are students of an er, but they cannot learn to think on their i y ior risk then wn ideas, r I about ii in nions of ideas, Having ri i i tional b i it s You need to recognize ( sak, verF hosc of us who analogize t i re are r nents whet we get into an nd cvervtiliilf the w rid looks like mart
c
styles of intellectual personality. The pigeonnoler can ponue ss ith sweet indecision, which might be th best of four or five ways of viewing patrimonial bureaucracies, all th while spree lating on the many details one might use to place them hetteL as a type of administration or, perhaps better still, LO break them down into patrimonial bureaucracies set up as such arid patrimonial bureaucracies deriving from the gradual drealc down of rules in meritocratic administrative systems. Every in tellectual personality has its moods of excitement, when hard work becomes pleasure and Edison’s 99 percent perspiraiflois suddenly disappears into the 1 percent genius. As in love, so here, too, it is worth surrendering yourself LU the excitement for a while, maybe for a good, long while. In dulge yourself. Wallow in your ideas. But remember that ulcu mately ideas are for communicating to others, so you have tu stand back and judge them, just as you have to stand back arid decide whether to move in with or marry someone you love. An idea you become serious about is just like somebody you live daily. You see it wea ing a bathrobe and slippers, without its makeup or aftershave But you should feel you can never come to the end of it, thai u
with. You get familiar with it. You use
it
retains the sudden enticement and novelty that grabbed you cc begin with, that it continues to challenge and provoke. You shouldn’t move in with an idea that doesn t have that kir d endless power and excitement.
Dicovzit
.HL I i1( Os Ot
iht
iOiAS ANC
hoc sc ggesrs so ethin else important. Re at tni y i de wd u h end time alone, tilot I5tl 1. l I t neat mm ito TV, no talking 00mm ‘t. Do what on most to create a private world in \V1iiLiI 10(1 ti gut to koow your idea in depth. For me, it n am OliHI walkii aIUUl iiid t ilk nit aloud to an in \c m
a
ii
b
ip ni oh
CsIi n
b
whatever
or
I st
t
say
hich
id
i iings
(My in isibic companion e
or
is very usurul.)
then or get bor Someho, talking my resay
ideas dirouglt to somuone imaglnurv makes mu more conscious bo tiers v hmr 11 m. (0 ourse, its also great fun; an -
1
va
skDws
h(wfart
push you and
lUt I
You w iii do mmething difibrent. no doabt: perhaps sit in a mitain p10cc and look at certain scenery, perhaps clear your 110(1 LerL musu Iwfore si my do vii to think, perhaps swk c p nit is t t ideas—like the social real ty I d Hsc d he op iii ng p g of this hook —have to be is ooud to he won. They don t just shoxi up thll dressed and to ueji out dir a loi ely
ant y
toll
ntion not
ci clung
part
on the town. And they
fit.
S
l1
o[ tl
tSriiiLs Us to
rn final topic: the qticstion or puzzles.
In the ver\ beginning, I suggested that one of the odd qualities suci n-nec that w oftL n s 1 t a project with only a reD
1 in s it ii area I mdi ig thc teal puzzle and 1 icliiij s sol on oem togctl as wc go foratd. I now oeed to clarify that idea. e
‘c
r
PLzzs
What does it mean to say that we start out with general interest and aren t clear at first what our puzzle isi Consider the rare reverse case once in a while, a research project starts witl a striking, clear, puzzling fact. I once noticed that status rank ings of professionals within professions were different from sta
tus rankings of professionals by those outside. Professionals thcmselves give highest respect to colleagues who have little to do with clients: consulting physicians, lawyer team leader elite researchers. The public, by contrast, gives highest respecL to front-line, hardworking professionals in the thick of client problems: pri mary-case physicians, courtroom attorneys, clas room teachers. Why should this be? I was working on the psychiatric profession at the time, and this empirical puzzle simply occurred to me one morning while I was thinking about the fact that high-status psychiatrists talked to upper-middleclass clients with minimal difficulties while low-status psy chiatrists worked in mental hospitals with mostly lower-class clients with huge difficulties, as I and most people then imag med most psychiatrists did. It was one of those rare occasions when there is an obvious empirical puzzle and a straight march of the research from puzzle to solution. Most of the time, however, clear puzzles don’t appear n data. We are more likely to start out by playing at normal sci ence with our data, trying out all the old additive tricks: What is the effect of another variable? Does such and such a finding hold up in another setting? At the same time, we are generally being urged on by the general (and insoluble) problems that probably got us into social science in the first place: Why does society have the statuses that it has? How does real social
M1:’riioos ol Dis&.os em
Ia
\V
ua
dii Cs di d ision of labor? How are pilci td alucs 5 tahli e h ed Inn resting as these problems are, tism re nearv devoti ot real content. \Ve can’t directly reason Ix he ‘en the cry w i s i i 1 en have infinitely con L
est price s
.
e r
ii
t
Sn
s, s c al eF ange division of labor, uc—-none of these has a iixed, context-free meaning. roust ;rni’n a c find ourselves with a general concern of t
S
ss and
1 at v c n s as ‘elivar t to that gen en 1 i lien hut c i that hi mg ng the concern and the data together wtll lead us to a more specific puzzle and a solu tion, ‘1kw i ti issu is how wc recognize a pi zzie in this amor 1
1
it
L he en sting
est and d it
,\ n i
‘I
p with ideas, finding things puzzling is very much a matter ot taste and knowledge. The knowledge part is mvi s. Y I eat tell whethi something is puzzling unless a t t to It di F re t o i whit it i That expectation i
‘
rcsts on what
alread kno’a So the basis for finding things surprising, is knowing about things that arent surprising. This s wh Ut di gra ‘1 ian’ slajorS ri ode survey ‘ourses and why you
to have trouble seeing puzzles. Out minds a a v ci 0 nalizers, and seeing puz.les means, in peat snuro”s, o powerful pattern-making machine
ut fOOte JnecISei. 1CLEii1
di ift a bit, Note that this is another
lace vim ,ie e
essivc
confidence gets in the way. Self-confident people, particu ted of the arrogant varletl. aren t happy running the engine an rome fur a bit. But that idling often helps in seeing pueieu i.e a
-
ug the instant answem is what leads to success Some of us rely on external puzzle geneiaruis. bus iUr many social scientists, recOgnitoan Originates n OOliti,,aL or moral commitments. The I Pbus was a time Os niomig fomtu k rods—- ,..n cal and moral commitments—of many ±rfe those who entered social science in that pet od usually han sense that inequalit, war, social change. ad so uO, were bum ing concerns. No matter what tile paiticular duection of ti en commitments, these people caine to social sewn c tl e’tdy thinking that these phenomena were heepl) minceesting. 115cc might have thought inequality was wrong. or they mht
been angry with people who thought inequality wes em but they all thought inequality was extremely important. aiic
gra I in it gm s (ought to) I ye general examinations. You have to know the background before von can see that something doesnt it itlto it. Note that this explains why people itt c e al heo v ne ret rue up w th much. If you don snov mvii sg ab ur dii woild, it s hard to see what
one always sees the same puzzle The result is what one ol’ misy female colleagues dismisses as “reseaich of the loins add
puts
women and
of
the worh I call out for explanation. You end up writing
in many ways puzzling. The danger of the moral-political
cause t 1
r zzlcs
s r seat s
it tog
a vii
it i
solved a, well Seeing things
hug to live with ambiguity. If your
first instinct ss i cli ans unusual fact is to am it into a category a lion ize n nr us of y ii favorite idea, you are going
it
Stit.
“
Such research is
SOUCCC ‘of
lioL
puzzles h that
terrible :rmncresing be’
soon becomes elenticSS normal scsi 0Cc, Te toot
political source for puzzles work
1111 if’ o e
JIoc
etc
puzzles to grow perpetually within one’s biozoer concert no you can start with the puzzle of explainuig \mli’ wome and men seem so often to behave differently but then go OIl 0
Sir IHUDs (ii Discoi FR\
liit
about whs t is that xs ithiii women s groups s e often see tepeared mane of the patterns 01 difference that we see between hr es he ohj ‘k in r 0 hc in nsion with the ri I in uzzi hoe vet, id o tend to force a choice uori v
to either tjck v ith the origilial puzzlt or allow the subpuz ‘les rak on logi uf H eir sn Among the best of the 1 11 n t va cc is p cent ly the tension be c i rhi se two log ics that Uris es their creatis itv. For somi pcs)pie—this is more characteristic of generations ,
ial v ilc puplcxing because they are p< Icxed V i lie it oss posit on n it fhe most common form of thIs attirudt mdcx mcinteststself iii what xe usually call It n t r arc I This it scareh soti ated by and focused on n pu 1 lai cnn 01 tril utc o the researcher: gender, ethnieitv, race. or whatever. Often identity research takes the thin’ of Is there ‘ri sorrow like my sorrow H’ n which case we tr
)s
to s
cot w akn sse I thc political-moral puzzles I lust mentioned The strength is strength of commitment and depth of intetest, I lie weakness is the danger of bias and relent ssl ni i ins c i r ivil d en ( c i als bc I is ci to st dy divorce or disability or sclloolina or wealth hecatisc’ of immediate personal experiences that mix nc t be clenti y rd ted 1 you alk with faculty mem o
xx 11 in i stnp isn utimber whose this Icind. it is sobering chat usually these motivated tacaltv members rite reacting to tin i
,
i
liOnS dtiiths air’ ot
c\priumll
ia y ‘c t r ‘i ol ov vas ,ht when lit said that “all fan dies art alike, hut an tnihappm family is unhappy at’— haN its own Isllioii. ‘To tidgr h 1 social science practice. there is iiii
I iN
ut
join crest
s
al
t pc sitive experiences
Lit-
tie is written about them, although
a
school of “well ‘haitig
search has finahi taken root on the frontiers of psvchoiog
‘
re
and
econol nics The most important weakness of these personal mot1vatiol is not one from which students suffer. It is, rather, a piuDleni for middle-aged faculty. if we hgure our our basic pm ale. w don’t have a new source for problems. Perhaps r is this that e plains the surprising number of social scientists who undertake passionate research as oung professionals and then go to sleep intellectually in middle age, as their personal problems loot smaller in a life filled with marriage, children, stunents, ho hies, prolessiunal and Institutional eminence, and so on There are, then, personal sources for puzzles as well as social ones. All of these various sources can be dangerous because the give us particular desires for particular kinds of resulis, becausc they can get mindlessly routine, and because they are good only as long as the personal and social concerns last. But they also can provide an energy and passion that drixe our need to understand a puzzling world. These are the driving forces b hind most great social scientists. There are those, finally, who simply find the social world in trinsically interesting and puzzling, just as some of us wanted Lack it) know all about snakes or tadpoles as little kids people. And to he blunt, very rare people. Foi eveiy petson xshose passion for social science comes from irub disinterested curiosity, there are dozens whose passion arose onigunaliy from personal and social concerns. Faculty who are deeply puzzled about the social world xx ithout hc ing a personal or social agenda are often the hardest to come to know. Theic passion ately disinterested curiosity seems strange to the majority ot’
03
In I
ci.
concerns
IiL atCS,
sot
sit Discotgi
inn
1
sen ce front pcisoual and social But men are ctlWdVS among the most creative. s meth g tf at all good iu tHai t is 1 h u
ye
socia
undergradu 0ficlcLatC stticnt, or sent If protessors. hatever its h s UUZ/ enienc hcc mes compulsion to ligure our the I W et y i I I fac I y ho have it, learn St
(1
vs whet
arc wgn ni
th
ig
a
front tHem. The much
tOoT 031cc
1i SOC I
ie rc
1 HI
Ott C
is
iii has c thea faults often great ones, hut
teas Ii and are thumsels es willing to learn. w t I nd y an own gifts of 1o i I h pIt
to
oat it
ii
minc lioss ci er, that there arc active and ci en tal lists v ho I i t I ye th s err ative puzzlement soc ii nrc s ciii scic ice not for love mc iibers who d tI a ci 11 go1n throuah cons entional motions often Hut tot a lii in usidt tab c success, a success they yalue more highly with 0 iH
si
Y
it
fl
iv 11 cc
in
hen by their behavior:
ho condesce ncltn and uninterested: another is but tuliVciltionii and stale. Vlien you go to ofhce IC Lit I 1 Ic or ic ir cm usin the bland, busy
coiC 15 sli Jft cflhiiiclit
ho
it
all i an. wers I r n i leas exti icate yourself i prod son titdc uul SucH people have nothing to teach you. I am 11 iv it tIny lad is 11 aginatiun. I said at the otiti1 5 itt ia a en vcta on between rigor and imag— ci i t Iflat 100. 1 LISt aS rigor cJfl Dc practiced and mastered, so can I Ilatilil Dc Os ciopcd and ilicrished I hope in this book to Iii cs o doing that But I have he i cm mxci i c e a tn[i SclO,,cStL d It a now for cc u to find the excitement that 5 cOflIOS svtth invent icy your ins n heuristics and reimagining the
n
-
additive heuriseir Tra heurisc :iovc
cT
c)ifiy
iccs dimension ot analysis making
mu in a
00cc icsl:. usc
as
ol
V men
logical
on
argument heuristic i he heuristic mc vs ci ai rums a familiar eryClmci tar dc new t,tle. The ma,n .irgucnei Ic urisi Os are prohieniati a
t
mo a
i.
i Op or Jenti rig mdi eal assuiri p:ians .nd rca oneefo ual ran behaviorism. The position that air cani-ot measure (or s.iIy) tiic nicaiiiii
that
One can study only behavior extcinal actions t sat ace sign manner. Op3sJ as cuituralism, reputable and reliable uta: 50. i,ll actor. objeeO or L. i. ogle .i 0 iii 1 stud A studs. case 0, .rs: causalits. The raasons things oar Causaiitx was thougit tormal, proximaic,
tan know only isgular
patrons,
s
orab un r 5
and final) and by Home to h
mi iheir causesl It a
,fitiS
neasciranic
in action
brands (macanal,
e
fl comp
shibboleth f standa
ci
sun
analysis. cluster
analysis.A tfoanriVumve technqse
macion about rcseinhlarmci-
i
th,s sorts noleca
i’iW
nssips
a/cs
distance between tOe ubjc cc
base I uaca—reoct u
techniques. mcsultc in a dimidc conictJconsensus. The debate user whether ui,ordcr in social I lent cedniatu0 ii 00 suIh, rri,in or theory icunboc institutions and oppressive
contextuabsm.
inc helif mar soHl Octs make no
‘eras
cii in
1 csl-en ahstrnreo ir
soctal Lime and spase.
sot al taca cisc surround correlational analysis. A form ii quantitative analysis ha
it
cu
ne
us
winy 01
rir
15
.
-
-
-
-
.
.
-
enrl> disurdeny indisiduis d,lnseiSsiis Jcaur). ocmmtcr I constructionism. The position chat the chings and the qc alitmn of things em realism /s Op/wacO inreracnoo. in anew reproduced continuously social reality lire
euhuralism, Thc i 1 Us
a,,,
sosition
thai rIle symbolic
5lLO5S
of cimicorm
uSO
n I rio
id
ft
c, behas ionstn,
—
ulteirir, I he
-
si
their lives
nod i
cc s:enr
I
Ii
sue
al
,ce 0.-.-
iUcr’
.0 e’
-
ane.
i a’
a
I
Oataredation
‘P
irf
a
Pt’’
c ctpt
qctea.
it 0
.,
[l,
ta
P a
.
tea
hetitt..r.c
c
inLar
.
‘at e the r
,‘
a
0.sei Ott
““‘
tap..
l.a.p..
c
i
a
.x
hits
ate
\tS
1inpi
or
ces
attic
t1atii ag the
con
We
aol split a
1 i
s
1
i i
t
i
Ic
t
ii
i
is d
ti
The cxia
of
interaction effect. In linear jatodets ri,r data, an effect that ,nv lres m,onmbitaaetonr of 0 abuts. such tIr sariahie X’ 5 er0.ct on sariable Z depeatis 0 the level if “ecal:oe vie vets,
These can he of var’ tOg types: mutttpbcattee efftats, suppressor
s A a., plaitat tim
c
splanatort a
e’q lanattsin
pregi at it
,5”i.ti’
tacit,,.
,it
tat antic.
ac
lie’,..
I.
‘
tacttit snalmata. A
cci,.
.rc
1
tm tta,iJA
1
Ba
s
te.tttmqeti.
o a
a
aS,. .
the 1960s and 190s, advoc
ing rlac formal analysis or literary texts.
atrateps or
depimtes met t icmls
‘.;
(it ttt.
miqm.mtng ntethod to the practices of tIme critiqued naerhod. method. A set of standard procedures and assumpriotas for carrying out some forna or
tip
oroos soc ml aiaalvsis ,m
t
.
mt
,t.
.
je,mt
‘](m,it
‘ct.m.mSe acat’ a
it .t ‘iii
nartlaocl
it t
l,ut,t
ta s c
o
cIt
wft
tc Ul.mt get total
mt ‘m ic m.
.,..omt.t
a.
r
cuïe,.
their meaning, usually far a particular actor, riinc, atad place. OppUieul to posisivis
nactacritique. The critique of tine method by another based on the application
mat
i
tt
l
(it.mtm\c dat.,. tot cite lie
tta ic-
att”,ta.est
di
ilii ‘.iantuaotae’tra
I’m. hniqctea
o
tall 1
t a
tio a
imgctact
It)
ia,,
i ttie ..lttat.Ltat
‘dot
cx
’lat,aric,t, 1 max
‘,
P
a.
ft..’ 1cm
pragmarn
ij
sjalanat on
t isac [to e
eattaplr.
t,,a
a
a’
_
as
p ft
it
rtaetit cia a..eo a
..r”
.
mm.
,tt. 5,’ t
In otate
if a
‘i
irlrerprocivism. The position tlaat socia facts cannot be measured a ichouc caking accoui of
tar to
—r
o urvilinear mtfet ts. arid so on. They create serious problems for nearlu’ all rypes of c inatiott and inierpretatior of coefficients. Opposed to mulct cf/or’ she effects o otd
literary sri’ucturalisin. A nauvement, dating largels from i
storeS Daseet cit. em’s
primary documents.
potident variables when taken stngly.
clii I sm
et
I
historical rsarration. A iaaerhod to a.aaivztng soci.d inc b 1 rensive readiiie
di
Oi to simpler
.t,.l_i,ta,citoOal
isa
at oi at
ciaipkx data into
11 Sn
ii ..te r in.
cc
ii,’ta
It vi
,.itp
tore
tot
p.
rf..ssc
1
.
at
so.
i a
p rot
I
I pl It toriaaal cod analysts of
at
cd faets
t
atctlaod c
tt isa i’
‘ot
iaertv
tttg
t
otaccpt on in soctal act
methodologis al india tnt and last orm them, face mulscm
iduahsrn.
relits
emergent;
-
Tlae
that all sorial phetitinmetia .rc namarel
mtapa’—
tact on,.l that of rltr incus idc.atS sIio afe held to eenie..,te
reduction.
methodology The dtscipline of invi.stigariog naethmads. l’he word is also often used as equivalent of ,ntihsd, as in the phrase “\\‘hat is your merhodoltigy a oaeamng ‘WI ac naethod did you use?”
lttacr ,oid liter lcvels, always
microeconomics, The branch of economics concerned with the behas jot of multitudes o identical actors in simple markets under simple constraints founded ott concepts o
roe
at
model. In quantitative social science, the mathematiral form relating variables to onc ao
ta
a..
the relationships antong supply, demand, price, and budget constraint. 0.
e instil.
It,.
ci it
an mc theta ,a,mtte
a atiah
...
It fr ot
tat
air
P.
patti’
adam
On
0.
tit,
ts o’
a
it
1
11 ‘a sr
t
tat I
tOe
s,
a
t
L
1.1
acti
a
tOe elasale debates of
I cii
at
t,t tiara itt ss
adci
i
a
etit
n at
a art sul
n
I
heat [tat. d . m
—
sattattles pius sonar error
iii (it
Ii
I
ters). as hich must hr estimated by some naathematical functioma of the dam.,.. multidimensional scaling.A qctantitariae techniqste that turns informarton ,.bom rm.sciaa
hiance or distance between a set of social oblects into a map of tht oh ccts usua,my art 1
ram,cici
c S
a
.tartae .mitim.’ imp ..menc
s.ott.
cOO ,e,.
.
‘..ttcrn... ‘here are laondrecla of
t:tt
l.tep0irttt
tim
Ic sot
rcj
at
10”
(ti
setettu
itatist 5
a.
the patcmro a
C
A t
tp
caner. ‘fypicalls, the relatiunsh,ps tnvolse scalar coefficients usho known as psranae
‘I ,eo,.r,.i ttmatliet., tat,. t,tmh tO
oat
t
tlmar taaojm.
)lcmna.m, ‘I’ll
a.,.. a’
n
ott
.o a .rtiica rolec ate
t
teitet al linear tiimcdcl t(.L\l a a..,
pertaca a
c tint
t
so
cc di
a estttaaatcd The
(LM c thor dirsetli miii
t titia
a
ci c
criods, usit
ota. t
sect
if
t
w fat a and
two or three dimensionst, reramning in rite map as much cit rIte original distance
nsation
as
ittaoi—
possible See also data-reduction cechmques.
multiple regression. An alternative name for the standard linear model for indepcoclc a, atad dependent variables, nor to be cortfused with sirdinary least squares (OLS peter aliecd least sqciares (GLS), naaximuna-likelihoud cstimatioia (MLEI, arid so on whicf are nantes for different sets of assulnptioos (aiad the algorithms associated wici tla as asseimprions) that arc used ro estiniare the parameters of these (aiacl other) mcdels. nan ative heuristic. ‘I’lw heuristic move that works by changing ths’ wa we usc
sro’
and events to describe the a cual lanooms. The important iaarrative heciristics atc p
cs
till
ds: 0
1
ny
5.
Ii Owl
5
ir
i
it
L
al esn
sat
Si
i
U
.
Is
ns
( la
•V
iiOi flJ t
st
I
051
t
i
cli
a.
t
S
III 1 doan
s
SS
strue i
S_i
cquai ohs
i
thc ha
,
>Vs.h
Ills.
stylr ad
tVdll
suro
.
ts.\IS)UtlHL.
S
inis ow. 5
1’
C
iCiSS
;VVS:
S.
Ii
s
S
15 hi 1
115
1
a
V
sO I I
i
c
5
Is
n
ci
.s
sJw
p
to
win
rrserlt
r
a
1
in
a.
t
h measure
i sti
ij
t
Lit
S
tact
S
It
interpre—
nit
uid
4110,0
hence
,
s.j
1 Tls
SIVIOC
tnnatc
It a.
ti
went
to
s. t 1 5
VlI.h
loll
1
HISS
a. ntactn. explanation. An or
srI
ill
TI
ISW
ivtsIae
e15
arm
I dat
Ut as.
o.
(.105411
.55
55
s.s•
. 5 .. isascpens.ler
V
ip
550
S
dl
I5 Ls
Un
a. U Us
tI
thico 5 sl a.! I
I 1)
t I 1)
.1 Inca
C. tiC aoOS
S
VS
sidLCsV_Ie IL
.
5iCS
5.
S
S
a.
V
U
is ,.ss.is
eLnO
iSa.
Se,s.CLCJ sample hr reopt .555CC 55.
s.IU4OSSSCVs..\
HiS
..t
iL
da.
S_jtlitShi Ii,
ariables
45
.5)
htt\5 eel) depns.ient and
ii
V
noethou
.0
oUCS
model
A impli hed lsiyhly aba.
ival method. A td
.nn
.
will ii alive
tincti
dit cc
VIV
iseai,iy, generalit
seciucncc
fl 5 5 u
at
r s it
ii
150001) eel
tL Crc
iarnii 55 .ns.
her is
iii
a.
ms..L
.15..
at
C
ia.
ii
CVI
VVII1a..S0
i
VVI
es
S. C
S
s_f
VS
is nil)
VS.)
A
met’ts 55 nlfs,rnllr ours
sails
a
eros in a i ariahie
llIC
‘V
ura 1 wju atio
.0.1
ariable, tIependeOl:
aw.i\ with thc di.
Ia tIll S
adigiti
is.tV)UIVVS
.
11)1)5
C
a id
o
i n inc
j ott iso
1(1
d ii i
o
I th
1111w
ieisd—nncar— TIOS Ic!
i.e
linear nsnidel:
aiim
II
I
is) tn
teo.
.
si
5551555
41011 is) litjn.
tIC
enS
Vtflt.d
.Vp!_HLs.ta.ss
OUSt
ifl 5115511
s11ec AcsilIsSci in
tile
sis.yis.
lie. 1
ossio.la
i.e
tICtI
othei qo. ha. oi ib, xplssnai r
time-series method, A method hacd
_Us,_
is..
VVa.
550/I
models of so
H
:s0s.
VS
1)
sin
solve
VS
S
S
:a.ph s siz A 5 I
S
ables. In economics, time-series tee hods are most cc n moIsly dc
Ss.
um shic. In sociolnys they are SHort s.fre 5 ptriuds but only one v
ii
ten js hot sviilo mans vasiabics, a cit elms. data Ii ow ci e Ui ffercs I
a ai aba. U
C
SC (15
COO
I ml
S
III
VSSSH
V5
,S
,tstiltd. s.a.h 1 I
ch
.ilisIlLii
Cs.-
I
sS
.SSVISVVVVIV_•_.
S•
S
V
V
.05501
VV
F
SL1’VilSi
VII
l
ii:m
t:Vs.
VS_S
VII
_
tb. t 1
V
5/
5
tote
us. i wi t
I
tOlUtits
ol
s.irs.d
‘h tLii.
5.-us.
I
i
i
rs.1
u
its ti
d I y
al
Ii
lOsil
tealitc
oa.1li
Sn
scislItLi It)
p1 c h
i
.
I in 1
i
f5 5 so
.U
s.ioinrs
I
.
c
IV
u
Sit!
SlUt
janatiisn.
•SV
Ao
U H
a.
tiiLssft..s
sit
5055551
S.
Ussni
5.5,
5555
St
1
S
.5
P11tH
,.VH
V
V
dl
S
S
IVV
t
ss
V5d51C
Si
V
intttas.tisjfl.
Iowcr
(1 lieisi I
by rig irded as a
t
ii
IV
Is
Oils I
ll
s
u
a
V
uual olluctiS
LV
teal IS
V
Vii
I
Sd
5
pi:iL 1
iieullstf_s
s.tnrs
.1 I
1
a. sa.n,
t
S
1 s/s
V
It
)iseUilflitli.l
‘si
a
ill
i iAs.r
iy S
S
y ti
LOIS
ut
Ida.s
lormalization. f ones.
5 pis.siiiy
001015
U
nttlIIidS.
SIC
0/))
s 5.1
VVV
s.t.midard s.acsos.!l .nmaliow ,(AI. 5 i!)
1051001
atlas
h 5 oar
si.isncJ 1 inro
S ‘nit
1 i
1 W 1
he
IViOC
fi,l di
55
th
I 1,01
I
t
tas..
nVSI ,,SUiny
,f
)s
il
VU
b.,t ss srk 5
he!iavn t
S
I
is.
V.Ss.1IFSCsifVU
V
sib
‘nisi
t s. t
I
IS
VS
15051551
.111
.1 S:CLs.rai I
VS
-
Vs.
51)11 1551
.1’ rs.prs-ss.-SStIII 1
i
is.w
551
V.
If
OslO
UCOF 15551)51 0
SVs
variable, depenaent. in waoU,srd IitlleiO.
I H
LI
St
001,100
SVIS5F 5 C SSS
si
H C
i
a
51555
s.flsil1O
s.tISVVs.ie
so
St
tic
5
extcnt
_.SsJ. LH.sts.)
1551 1 seilontUn.).
SC A
5)55
treat
nil] udes
.Ss.s.
mL
standard
variables, independent.
in
UsiLsJ
causal
55511’
CII
1115
-
..S_SVltsm_
st n 4 mnaiysms. sli dar Causal 5
1w-s sI aricthet, depends. ot
S5SiS
)555
SIVI
I.
s5
analysis.
iVarVahie.
S
/ss,
se,
U
SsiCSds.s.t,
stairdaru e.s.ural analysis.
p
-
Noms
1
C1IPTER 1
As
5 tg
anrs: anrlir pi.lo st Evaiss—Prits haid nessi
0 oorkmn
of tn Oiitinfl ti
In the
mow
remarked,
is not a complete d o can do titIis orE, atsd L me 5
morIg
a
Oflec
nt-vs
peopw
se
lcse out been studied hetur he cannot help making an kfmoSs ledge... Ar ‘one an pn d oc a new tact: the di no is
idea.
(170:2-Is)
icai phraseology 5 thcore
of
hnre Lakatos ,i 90: -2r1,the rnos :mww 5
quality or research programs is thcir heuristic
pow
1 to keep producing er,’ their ahilir
OCo
ideas and ix inc the wa to new findings. 2. Among nlanv writers who have made the case for “beatv’ in scmentinc
Sw
Chandrasei-iar ii 99. s,
Svnr,a.si, semantics, and pragmatits arc the three fundamental sspectS 01 vs
-t
2 li
tw:s explanation
‘vorsis
is
an
for oem tin 0 methods
rs lutinc- prucc-o Ores lot
rsgorm civ
stems,
of,sii
s-sample. rs- Morris (1938). art-
chamsgmiig. Properly
gui ry. M etiiodology
S
spcdPmmg.
(Ii re—calls
i
j
a mk-noe
sS
c.
-
-
I ms-sos -us
Sc OSSIOiS
5 or standard causal analysis (SCA). then, is a method, vs hilt to write aL’s m Ethnograph ethnography or SCA is to write merhodolog. in practicr people ,mrs tact!
mds/sy m mean method,’
iv in
liat people osny, ihsse terms oloy t” ls,s,s, 5 to, c itt/s
loi m
o
-
ie/, w h 1 met/sari; they use ,x, !.su,Is/sm
1 and k-ti/S lslog.. a
I have tried to maintain
and inetliodoIog thrt oghoco
SIte!
1’ the lhmiliar seminar question, “What’s your method nOsy
not ustamarily use ses,hsJlsa!
‘
ow
the ad e
a
cli
.s tlios the adjective form mis,ed
e 8 t
traditional distinction betvvecn meeho I
(or
-
5. Scimetimes qoant tacise analysts do undertake dctailrd study of several case’, For an c ample see Paige (1975).
-
Li,
-
55
‘
,ici_at,,i
l’s
1 ‘
-
t
‘--
_ti
S_i
ii
ci
ti
iSIS
c
itt c.
si
-1
5
t
hi
s
5 it dii k and 0
Soft
Sc ‘is. it n, moos-is
tn
i i is
mujt
rln,r.ir d
resorne ii
is
torn,
ta to beyond
at
.
‘i’hctc, I ,tttitjst stiles .n chc iii h : their tusages 5 bo
i1 1 clue Hiss I ot
iii
hi i’
cit
is
‘I
s C
,
‘I
H
as i
t
an ods of shr has- bccu
r
t
bcii,
se
c
-its ise Ott cas,sjlci.
ii i
tiznc
S
i,
15
cii’
iii c-i
r
if
(5
itoil’
-
001
it
iC..C
is ii hLis 5
a.
-
die
ai’tCk ft
its
at
r r
‘Si
i,
Si’S
i
it_k
.,‘
rS
c Lsi,tntn
5
itt
isO
-i
Os-
s,-c,,• st
•‘‘
c
c,
, 555 i5s
mc and of
Of i
a
Hut
tics itisc ci.
I
i
ii
attic
ass
ii
to
a
1 ‘l’heie. Helmet Registration, and
C.”
So;itne
Mi
“
I 9’” “lnse;eae;onaliec’d ()raanizarions
t s ‘lii, 1”o,’inei
I
I 989
‘
(an
Viku
Suchman, C. C. and C. L. Canill 1996. “flie Hired Gan as Cacti, Laws. La
Press.
mg/nD
r dgc. ‘anibi dgc University Press
he Duality A C
lure and Pea e cc.
7 and Soil— Them
ac; talc
21:679—712.
Sorrles, G 1968. ‘The Ssera/ Order if the S/a,;i. Chicago’ Uis;vers;t o ‘Chicago Press I 98-s. “rhe (,umc;Litive ‘learcire of Lual Urban Culture. Air erican Jo sinai ;y or ‘
1°gj 90:283 SOt, C.
i,
ii
)
\l,,rr1s, (
Ii
s5,
1”,,’saiarr
y,
,
-
n 13 LDeostvi ;hi/ and Draw rat Boston: Beacon Press c_s I/Seer? 1 log:’. Vol 1, on 2. or bra natl’asal Enu,lp Jia 1
I. ‘or; a Soar;, c.d 0. “Ivurath. Chicago. C niuc’rmry of Chicago Press.
in
Szrerer, S 1984. “‘I he Origins of the Registra;’ sden,,ral s Social Classila’seioo of Ccen tions.” Stirs.!,,
Jv;”ii,l/ 0
SYriSi’;.,)
5:522—4C..
Slorl, ,j F liwi “Raeion,,i Expectations and the Theor if Price Mosemenes.” E;’snsmriri;a
Oa i ns o 5/a; SnooP W “r/P Am’. Loi;de’a: Hanaiiwo. 3. P 1961. 7L 3 Therhorn, G. I9”. “The Role of Capital and eh Rue of Dcmacr,ns’.” \n, C,.’! 1
l9Y 79-89 a S or ida I ausal na is, 9 7 i
about individual sir,
and heuristics, 83,84,90 91, 11-4— 16
)3, 56, 5
ily
anout ioowleoge, So—Il 5) It
and small—N comparison, 22 Ia euro ii
8,
s
uI
1 y, S ap al zatro fib. al
ii
I
1
ii
d I’i
IIL000(
i
so
lt’X
I
95
ieua1iiai’Lt\
oo
See aiu iperijk methsdr
about narratorS 58, 59, 1 1
43—l4, 51
59 57, 58,
1
about behavior,
euv ‘—a
)
-1-1—45,
48, 51, 53, 90,
about positisism. -il, v4, 49, 91
5 58, 59, 179—83
10)’ l.
ivervicsv about, 4:2 91—5 59. 108—71
(Ii
1,
(14, od.
about choice, -10—40. 5t) 51. 55, 56,
about teali,m, -In— i , 4$. Si, 5 a.
cumplexits of, .11,0 Le’nei’a, .0. cod
clt • 110
01,11
ideas.
il..:is
tliu.,,
ana
—
0,
6—’S. lii
lst,or,
narratise,
56, 58, 59, 60. 201—6
1
r, ow
I I’)
60-60
ai’,,ut
0515101,04
nd sti,all-i\
0 s,
4.
i
soIhil
99
.1.,ailsb, 09,
st4,t0rc,
.
Id
our:
t
(
tv,
4,
,)
515
t’,, 01
—
about construcrinnism, 46—-1. 48. 50—
DeNorii. Tia, 1911
tIll
)i
‘I
I
I
dependence. Sc, dependent s ariab1es:retlOO
1. 53, 55, 56, 5, 58, oD, 192—
dependent variables, 20, _i.
s
57
5
°2. 90. is:
109,
22 6.. ( c
i
4 1
i
S
cur sueu I(y I
and structtire, 56, 5°, 56 59
1st
s
1 0
1.’,
,)
Iii
,
‘.
—.
--
about culture. 4-s -iS, 48 51. 53, 55,
1
5 . 5
descripror heuristics, 1 ;S—46. it, l)esrijsifres, Main, l8°—i,
and cyclcs of critique, 60—5
DiMaggio, Paul, 129
about emergentism, 45—40,-iS, Si, 50,
disciplines.
analysis
‘-1’,,
ii
1
-
and standard causal analysis, 55, 5—5n demography. 44, ill —92
56. 5, 5, 59, l”9—8 a 4515 ‘1 i
4 4—4h
and social structure. 53, 55, 59
about context contextualism, 4— -48,
.ini stat,da,o .,a,,,s,d ..n,ovsis. (a_i ba
“
about sueial untolugs’.
about constraint, 48—49, 50, 51, 55, 56, 57, 59, oO, 198—201 51. So. 58, 18—91
nd
and small—N comparison. 58—59 56, 57, 59, 1 9-’Ss
about consensus, -18—49, 50, 51. 55,
05. b, 6’) ‘1,
0’, 0-1
about conflict, -i9—50, 51, 55. 50, 58,
5, 58, 59, 18—l
59. 90. 201—6
65
,oo,1
aOd
-:55! to,,,
8—9
0,
auto, 4 ne
cc;,
(19,
5. 59. 60. 198—201
•eueu,,
73 55
192—98
basic, 49—53
7171
7
ao,l t0tfl1aa,..
•
4 51,
9
about noiisontcsooalism, ‘7—48,
59, 171—79
2 (5—6’s, 65,
ly 6(
(91
cv:
and methods, 5,—60
debate
liii’’,
,l1d
and tests of ideas 213, 214
about analysis,
i
58—59, 60 2(1—6
—
no
t
8 i9
i
b c
d
1
/
55, 78, [68-
and puzzles, ‘o3 2ov
,6
59 183—8
about ioterprctivism, 4 a, -14, 45 11
and rnerh ids, 13 14 c
—46 48
-I
i
1°
tation 4 1
I ‘0)
si
$9,
38
8,6)
1,
and heuristics, 108, 119, 11
57, 183—87
about ethnography,
5S—9’
about foimalization, 59 60 fractal nature of 42 78 79, 111, 162— 210
141 166
zones of research in 292 .S,
a/a iftcrf/ir dnrrplrne
distini
tioris,
making, 96,
38
Douglas, Mary, 117
and generalization, 55
drarnarism, Borkcs five keys of 94 13
and heuristics, 78—79, 111
drunk driving, Gosfield’s analysis of
importance of 166
104—5, 134
21
IrsOliX
and innu I-’),
)ly
and
nl
ss 2
so ifli
69
5
characteristics oI 13—9,
i
and cycles of critique,
o
logical structure of, ii
oisc
overview of, 8
1,111
1
s6 18
and oltu Its
tique. 00 02 63-64,
oS. 6o n9 0—1. “2, “3,
S
a,
1
n, S
I
,
di
cr,ni’l,
I))
3, I
5, it
S
rormalization,
and saplanatuil programs. 27, 29—31
and causalir,, 27--29, 32
and abstraction, 2”
and conorereness, 27—28. 29, 3
and analysis, 59
and cycles of critique, 61 and data 34 36, 37, ,8
and behavior, Sy and causality 74
definition ol 5
arid choicc, 60
and ethnogiaphy 27, 29—31 32, Sn,
and conflict, 60
)4
sos cs, 168. 172, 1”’,
ann Rascal
nd I’S
00
cOos 55 7
let
I
tir,000.I
,i
and jots
,,,rds Ri, ha,,.,
‘0—
as
3 5
method.
0s5,
69,
‘.15
as. N
-0
and positis
4
1
rgil
,
56
aid ,mall
salysis 22, 68, 74
I
so
ctur,
ai I stat dat
I
and test o
S
.rinaii
evalUation,
Esans-Piitsha.
C,
and cycles of critique, 60—ui, 69, 71,
debate about, 59—60
pragmatIc. 8—14. 2
and ethnography, 61)—Cal, 68—09.
28, 32, 3”—38,
I
39, 40 syntactic 9 11—13, 27, 28, 29—30, 31, 32, 33 34, 35—37, 38, 39, 40
—
.0J
aOl
,.ii
a’s. ‘S
arOirn. silas hoi
Os-iirl\tics.
4’
1 as 2 3 dl 6 38 F F., 10--I”, 31— 42.
and heuristics, 76—7”, 90, 184—86, 196 and individualism, 59 and inferences, 68 and knowledge, 60
and universalizing, 28
measuiemeni, 59
and variables. 3”, 38. 39
metacritiqus. ol
7
arid narrariss, 59, “1,71. 6—”7’ overview about. 25—20
cxpcrinientai’sic-sgns. t19
Farkas, George. 100—01, 183. 195
and positivism, 59
explanat on,
Felson, Marcus, 105—6
and realism. 59
final causality. 95 9”
and small-N analysis,
Fischer, Claude, 124—25
and social onto1ogy 59
as
So modality
at mci 5 oa,
and ,ausalit
—
and explanatory programs, 2’-. n 34
factor analysts, 181—83
UXiStlOct.
68—
and explanation, ‘4
and temporality, 31), 54, 36, 37
15). C-ill,,
04.
23_74
and parterns, 28, 34, 26—27
arid standard causal analysis, 37—38
.3 55
osal analysis 68, 69
S HI) ..i
and culture, 59
and narrative. 2’, 29-’- 30, 32—3). 2-4,
31, 32, 33 3-4, 35. 36- 37, 38—
and iealisrr 5,
I
and contextualism, 60, 190
hhrid, 38
semantic. 8—9, 10--lI, 12, 27,28.29—
16s
iT
and ctinstraiot, 60
and hcuristics, 27
39, 40. 106—v
overview about, 15—17, 26 1511
liu,,1)
and narraris e heuristi,s, 152— 53
11000
and formalization, 27, 33—34, 3
39
4. ,-i, 26
nd narratne, 53, o5, (aS, .1,55
and consensus, 60
90 91
sin 168
i
and mcta,.r-c
S
8, 8
andi di 6 a sri aild Loon Ic
5
foreground. Ss 000text, levels formal caosalict. 95—9’
and “octs
lOs
explanatory programs, 26—40
follsiwahility, 3n
and assumptions, 2”, 32, 35
l’S( I
1 aLto explanatory programs: mcthods
Fogel, Robert 12°—_4, 1)1—60
description or, 14
5 —5
ad lornialicahon 60-61 6)3—69,
‘8-
Ira,
74
—8
and absrrascion, 2. 20. 34, 35, 36. 3
.lso,sru abo t.
tiSnOirs ,,i,ai,,Oics.
focus groups, 68
and switching questions,
yi
Fitzgerald, 1’. Scott, 166 Fleck, Lodwdc, 111
and s,
5
Fisher, Irving, 33 Fisher, Ronald, 3 —38
13
13
senses of 27
and
1,0.,. ‘1 ..arlatr,— 100’ OLin,
—
1—15
,f ‘.0
011 sileOct, 5
28 —29
12—1
‘
INDEX
and ci idard causal analyst
loS, 170,
179 83, ISo- 87 192, 196, a I I’
it ii
In,
u turaiss is 1,56
,icc
r
rrat,ds. .scnnttion
-
f4LiaI 5 RUiI.i
,nd
1
,,s
5—2
B
dame
)51, (ci, 0-,, 05,
ul
geise 58
i
,
caccalss
i
sarect
isti
1 2
206
as a
,
161
argument iseciristics. descriptive
Jccrtstici; riarratise heuristics;
a at
,
rib ilcUriSt 4.
14(
-
gaticti
n 1
cOd
18
tsx,ca TI4c[ILJt,l
and Icidis
-i
2
55,64,72
‘‘,
1)2 24
gtogra
t,citcalizac,,,n. I 4.—44, I ii, P2
S’
Idncl4n, ih
Goftnran. hi
sing, 50, ll
Girttfrsras is, \licnaas, ..io5
sri
14.17’
Ocoule
,0 7 r’r,
r,sr
21
-
15
histocy
and cycles of critique, 76 79, 111 and debate. 8—29, Iii
and cycles of critique 6-s
definition of, 79
and explanatory programs, .s7
‘6
a’ a:lvc.
16s 1
‘9
rai
P .irJ prn’cc
?8 129 192—
I 7-‘
14,51
I’S
-
8-4
J
‘cpn,
and methods, 14, 26 organization of 5 philosophy of, 32, a3
and formalization, 76—77, 90
and small-N comparison, 22
general, 6, 110—36, l3—61
Se a/is narrative
idea of, 80- 88
Hobbes, Thomas, ‘9
and inrcrprecivists, 78
Hochschild, Arlic, 140—41
and Ka it’s list of categories, 94 97—104
Hodge, Robert, 174—75, 208
and knowledge, 107, 120
Holmes, Thomas, 207 --9
Itmitatruns on practice of, 111
Horan, Patrick, 193
and lists, 92—109, 162, 190
Human Relations Area Files. 62
and mathematics, 8 1—83
hypotheses, 26, 78, 119, 217
and modeling 91
ideas, of
language,
and analogies, 23, 238—39
and normal science, 77 11 s 3, I
and pos ivisi
,
acid rules, i.ci
urn,
It
I lcggct
re.strrpretau
rcs crsch,
..
I
,
2-I
clarification of, 242—48
78, 100
classificatory, 219—20
Hannar Nlrcnaci, -
c hccir sti
c-i. 115,
(art, I )iiisU
comparison of previously published work against own, 212, 226—31
‘‘
and social ontulog’, 112
hIs. 1
us fl’s
51
,29. 10s
and alternatives, 216—17 and assumptions, 227—28
0—77
and reality, 103 S
and heuristic, 89, 184, 196 and metacritique, 73
and explanatory programs, 27
and narrative,
nruuj,’ tIeurv. ‘5 (Pfic1
-
and echnography, 76, 77, 78, 89, 90, 91
9-i, 106—’
hisror. Il’c-i
rrct m,rs
‘‘ii
,.
description of, 1
and methods, 83, 84, 91, 112
81 s8 P
c
histotical sociology, 22, 23, 72, i63 16
and Morris’s three modes a
Hirschi, Travis. 2u9
and commonplaces 92—109
and disciplines, 108
6.
rns about, 110—1, 1 s-i
r
and causality, 101, 103, 106
and social sciences, 82—83 of
-
of,
192— Sb 2
ciii
and burke’s five keys of dramatism, 9-4,
and difference between mathematics
2
-
iristics
at
to ct S
1,
l’s
s s ‘ant i_ic
)6 55
i
s-s
c-ru ‘-4.
tO
Irifils r. 20; —-0
,
118
i
Get-rca, ( isrtbrd. ass
sf hsurirtic
9-i, 95—9” historical narratson. Se Narracis e
naly
(call j-
See ohs eaumplec cause’,
tnt, A tcnc
a
of topics, 77 92—109
88, 92
and Aristotle’s four
05
Frccmac.iuhn, 114. 115, 129, 193—91
S
519.
of,
and analysis 90—91
1 (s.
aS,
and time, 30, 104
and action, 107 aim
u-dc
and theory, 84, 511
heuristics,
and context, 227
and space, 30, 10-i
and cunvencions, 227—31, 232. 2sS
and stages in intellectual life, b+—88
emotions of, 2-10—41
and standard causal analysis,
empirical referents for, 21— 18
89—90, 91, 119
6, 77,
explanation of, 222
INDEY
11 rlanut
Kant
stIr
ic
,t,,a
Ii
I
L’,
0’ C
p a ‘s,olc, )
1 2
--
I
-
id 1_I
21
I
fIL ,‘iz
2
ai
1.,3(
I
I,
i1i
i 1,
itcio
2
ast
4
is
0--
islosm. 16
trot
ti
and
iiterlal srrancls within, 6 OVerview about.
reaht of, 108
soc1 successfol methodological ra
o,J,,>’117,
d I
SS
nO
,,-,
2
1 3 .nO6—°
or ,,Iic, 711
.1 iL’
art
‘‘“‘14
so
oi.
2
—
0 cc ,,.1,1
l&
I
a
I,
111,0,1 ori
61 s
0,
It
,(
d
0,
I’-
(Lt
1
ti s
Koh s.
i1
pL’1[
ad Cc, Lacat d so>,,
ni-a i(5
ii
11
1CC-OS,
7
01’i, II, S
‘,
Ii,
a,
[Oc
‘ilV
Ii
tol
S
2’, ,i,,,,ISSi.
(Sf0,.
I a’
7
1,ii
1 C
a,.,c
4
1,,,
S (51
1,115
>2 11
15
,,1i15
di oci 1
Lain no
so
ii
IS
s
7’ (I s
iii
logic, 215
Mills. C- Wright, 99 Mirowski, P., 131
Lucas Robert, i”6
mobility literature, 7-4, 1 7(1— 31, 200 21)1
171,
53
modality, and Kant s list of categorics 100
modeling, 14, 33—34, 37,71,73,9
So also specifii mac/ti
Markovian tradition, o-, 131
Mohr, John, 170
Marx, Karl, 32, 50, 57 107 material causality, 95_97
Moore, Barrington, 23, 175, 213
mathematics, 81—83
Morris, Charles. three modes of languahe
moral commitments, 243—46
Mayhew, P. 134 35 measurement, social, 43, K’, 59.
t I,rcc tnodrs of, 94,
Oil,) lpLezJIc rcicarcbt,’
9”
Mandecille, Bernard 15
of 94, 106—7 2
Meeker. M., 13 metacrltique, 65, 69—70, 73-75, 75
,ltcilt 701, is
Old
iJIc
55
1 arc, 2so1
la,, a 5 ,,ot - 51, ri is
‘ii
I r
ca
iaitcllc iCy
mcthod
motivation,
experience, 2-,6—-7
II
nO 11-,riso! ,,1
,s.
I
IL. 1,1
s8eciju
Meyer, Aul% 208
Lorrain. Francois, 185 —86, 194
market research, 36
,‘oas Sh—h9, ill Janiso,
list,
anaI us, 58
2c—0
microeconomics, 34 35, 61, 71
212, 226—31
So—- 57)
dcas, 2 I S
ol
I is>
Kwa
Sos a1o
literatore, ideis compared to published,
MalinowsL Bronislaw, 16, 30— 31, ,,osal
of.
types of, 5
See
IlnSparlsons.
lard
si
a/u, upeiiJn list: topic
lumping, 144—40, 169, l81—b
I
IS
if--IN
a d
17 LII, a
it,
I —26
Lions, 15 suir b 5 iiiry
2
its
arid ‘i-
0(0
social function, 10” —8
literary strut tuialisni, 59 57’—
I Ii ,rna1izatisn, 0>
an
—
hybrids of. >5, ‘77
npical, 92—109, l3
deh,,ic ahoot, 30 51, 55. So K S sO 201 6
..
11
overuse of, 108
of
1a,s It”
1
12
s.
1’
and heuristics, 92—109, 162, 190
,See
,Oic
and heuristic, 83 44, 91
iOoti0ns about, (OH
3)
os 515
Lip5ct, Se moor Martin, 1 75 lists,
17
rO
N9, 1
hoist
,.
.
(l,,,l7
tOO
V (1
€7,
1”g
‘
I
dcl,
1
5
s
C., ItO
1’
icL’Ss,ti
I st of categoties of 94,
10, lii’, 1
i’
27
Sb
7
7
[isol
L
H
1 -47,
49
6
LasSaid,
121
151
27’
--
IlnIl 150 2
1,11
,
methods,
nO
1c- 5014 c Ri
il
I
lb
,
(Ia id
basic questions of, 13
multicase analysis, 22
borrowing, [13, 116, 118—20
multiple regression, 112, 119
categorization Lit, [5
Moth, John, 176—77
HI
‘4
s2. 59 01 -01,
and commonalitii s in sotial science, 5
Nadet, Ralph, I i-i
conceptual methods of, 5— 6
Namier, Lewis. 3(4
critique ot’, 6
narrativc
cycles 1ir rali,c s,i ado
5
c
([OSI ty
46, 4
classification of, 28
I “3--li 142—44, LI, iSs ia
personal and social, 245
Mozart, Wolfgang Amadeus, 87—88
of
critique of. 42, “0. 111
and analysis 43—44, 51, 171—79
and debates, 5 3—61>
arid behavior. Sn
and explanatory programS, 26 40
and choice, 56
I
a
ri c
1 rttiv
I4
t
101
I
b
i
philosophy of history, 32, 13
Isurpose, sod Botkc15 us cism 104—6
pugeonholcrs, 136, 241
isuzzlcs 212—1,,
philosophy, 100, 103
1 15
StiFle lOt
s6
t
4’-)
qiality and Kant
P1st,: 107
st,iti
,,-t
‘,,rr,,o -uI.. K F,
1 .01-
1-,’,
01,
.
‘‘4, 1,)
‘
‘0
ICII ,:titui:, it,
as
,
i
Sf,
s
“
n,udaiur
5:,
lb
,)
135—87 192-
IS
not not I
unsiics,
1
1
S
t,.,ri is
s bitt
political to1ernce irerlrLitr. 204
quasi-expei’irnentch
dirs:gns, 11”)
Polva. George, 81—82. 88. 92
tiestiOOS. 1 c and .‘vrisrncle s foot causes,
positivism,
case—study analysis 1)1 1 and heuristics, 37, 9s,, 123—26
So 5
,98 92,26’
and fornialisanon, 5’)
posing, 13, I—i—IS
gambits b—
uzi rig the. I 20—21, rchiem t 5
-
iJIson Mar u oin,tI otner
i
a d
I S-u--OS Teas 1,
i
COl
12
aol des
21—0
anlp:r
I ‘
rs,
15
as’s sI ,,ii,,b,ci
)I.,
“‘‘
-
PatTeR 1. P.:rs,:it, Pc,
Ic
‘1.)
-
rd cuus,d
random-error surveyors,
and standard causal analysts. 5”,
2
randomness, 21”
possibilit SIC modality
range of appiicctinn. I ‘0’
poverty studies of, 10. 173—74
Rashevsky Nicutas. 71
Powell, Walter, 129
rational-choice models,
posser, Parsons’ analysis of 116
rational decisions, 201)
17
4 I
I
wrstuila ui 1
2
24, 225, 25
1
91, 114, 156
pragmatic explanatory program, 8—13, 27 28, 32, 37 38, 39, tO, 106 7
rational-expectations hypothesis 1”(
pragmatic rolaitons,
realism/reality,
prediction. 20, 214—15
reading, and ideas, 2 11—34 and constroctionism. 50
17
163—66. 187—01 clehare about, 46—4, 4)’,
St.
5., 58. 19, i8—91
prnblematics, 198—99
and e.nnugraphy,
prnolemarizing rlie,,bs- ion,, 1 2i1 21,
anti fornsalizarion, 39
1 35—3b, 148—49. 15g—60
ddliS SO
Perras, I as c -5
t
Rahe. Richard, 2(6’—’) o
Price. R. H.. lob 5
)‘i”,,’
tan ic
narratise, “2,
cnd Morris’s modes of langucagc, 106
S. 51) hr—S 1
(.
.0,11 Sc due
importance of, 4-I
definition of, 106
S)
-15
—i .,,,,
1 1, 11,1(0Cc,
tt,
,ctru::ruJI\ os,
S
small-N analysis of. Quintikan, 92
and interpretivism, 13. 45. -). 51,
dl
12.-. e. 1s5 -0. iio—-*9. 159—
S
large-N analys’s of, 14 and methods, 13 14—IS
It.8—” 1 ,t’v,,,u’,. 1
“C
,,,
5, 59
and uartacivc 1”'
iv.ti!’ri 4ooss.
44, -4), 51
tahuhi
rot:. Si’,.
f ccor’,.
pusituse suJology. 10))
1)’ 98
—
flSUt,t,
S
anaJ’sus. 1
political histort, debate about, 56
and heoristics, “8, 100, 168—7 1 art
riuli
i’,v,i
St e
quantity, and Kant
and eGo ographv, 168
i
u
ci
1’C-t ..‘,etr
1,
i)
aica Ste
of c,aeg
t
quan:ltct’ve
debate about 43,
ii
Ii
poitical eoiSmitrnCnr’,, 245— 40
and tcles of critique, 72, 7,
itn ci
s
political science. 5, 22, 23, 39
rissa ii ti 9
SC. CSarV e 01- I’SSltV
9—- ii)
tC)fl.
—
104
p1ra0ry. .Sse quarui ty II)
3( 24
Piore Mich tel 154—Ss
1
oh’,i”o, 1 01
iii.
I
Sf58
keys f don
,o’id
5-i
heuristics, lOS, o3 bi.,
problems, as stoiree uf new idea,, 6
and brioss ledge, 1 bs
psychology, 28, 36, 33
and nasrativa 56
48 5
—
INDI’
I
ii
1
id
I I
1
r,
5 a
r fi
I
5’)
hniquss i6
Schclli
fh imas 2
ecplaoatioo as puq ost of
and contcxtualism 58 6 60 2 s
-
is
heuristic cc
and culture 58 aid tycles of criciquc 68,
7, ‘4
n ccl )1 88 8) $ 3 88 9,)5
0 0
saner s
eaadg ichsngi
cial Sc Ic cc
Ci 5/
dies 1
C
ay
ii
SOC
cc
ill shn s
1 35- 6 1,6)
0 2s5
2
planatory progrin 8 9, 10— 29 ,i ,i, 3, 5 )6
a
i
(
10
c
I at
11
nodcls )5
sIt cent c
Ot
t
c,as
C •
,8)
atic
7
8—3), 10, 61
15
qici t
gc
cii
S a
fI
Pau
ii i
i.
1
tat i
and dchatcs, 58—59
social science h story
and tthnography, 68, 74
social theory pure, 2
and formalization 74
sociology
C
r
as hybrid method, )8
comparativc historic, 1
and intcrprctmvism 58
of come 16
and knowlcdgt, 58-59
ai d echnof,sap
meracritique of 75
and cxi lai acoi pingiarl
and isarracive, 58, 0 71 7 1
and forinalizatio i 25
overview about, 2 1—23, 26
and heuristics, ‘8 cOO,
h
i
and rcahim 58
interpietivist,
ad st,,ndard aus1 na1ysn, 58, ‘2
acid K,,n,ia i,rla,iooal
73, 74
organization of, 5 positise, 78 100
1, 1 6 200
ye hcur s ics, 90
al
>chiv
tichi 11
1
a
1 a
if
4
rai dat
1
i a
ratcgy /5
S
uid cii s idc
usc sits 5
c
iS
I cit r uctionisrs 58
ID
,Sri a/iC spcrific iiabjir/d
social ontology
si—48, 51, )5, 56, 57,
22
6
6
3’
Spencer, Herbert, 12 5 Spderman, Seymour, 201
59, 179—83 and formalization, 59
sphicing, 144 46
and heuristics, 112, 179—83, 192 98
standard causal analysis (SCA)
and methodology, 4
and abstraction 27
and narrative, 56
at d behavior 5
and standard caosal analysis 57
and chcocc, 5 aisd cunfl,cc 58
social iealmty. Sir Social ontology aims goals of 8 12—13 111
5) t I
of
,i
5$
,
16
and standard causal analysis
8
siid consc isus 58
social scienci,
r
,
aid realists and cocistrucc onists
space, as heuristic, 7, 104
122 1
t-g
ard small- ‘I analysis 22 _3
—Ii, 125
social function, lists of, 10
I
64
social capital, 186 87 social change
S
8
and strucruic, 58
dcbace about
sis
a cI ad)
I
and culture 179—83
8(
)
I,
i
‘i},
socioiogy of science, 111
csert
n 1 ‘9
244
and behavior, 1 9—8,
uared k cwlcdge, 50 5) 55,56 57, )fs 206 2
vi
0, 7
wet)
ingle ca c a lalys
e
4
I a•5 2)8 if,,
nuato
C
t
Burkes fivc k”ys of drama 104 (
i ilirit
Ii
1
5
4 if’,
,
nachineiy of 3—4
small—N comparison. Sir small—N anal sis
‘) nods ol Iani,uage 106 7 nn beds, 20
t
rid uricginstloi
12
and consciai its, 7
coinmonaliues in, 5
and contextualism, 5
conccpts and definitions in, 224
md coltu,c, 57
as conversation 3 248
and ycles of critique, 62—63 64 61
diflt rcnce between mathematics and, 82 8, distinctions among disciplines in 5
(9 70,
-)
73, 7i
debate about,)) 7-58 and cmcrgeotism 57
IN bE)
,1:,aIs,”
ISa
.1.,
.
C
hy ),nac,,ii
ar
1
Canon
‘
(
:0
S’ -
‘
at
,rn
an
ii at
and
o9 ‘4
S
9)i
aS
,ir.d .iareatnc. 051
5’, 5)
it,
Idistoc, Lou, 2 n
and nariatise,
,,rid small-N anai,sis, 22
I’LL
icy
surv
it S
G”i
Ii
OVe’ieW
2i)2
14
S
1 lanara”y program, ex
P
5 C
4
5. 11—13.
Vieket, Amanda, 196—95
1 u6
transcendent knowledge. 50—SI. 59. 60.
West, Candace, 1 2
See a1 universal knowledgr
White, Harrison, 24, i30—31 148, 1>
truisms, 123—26, 2 15—16, 217
derini000 a. 106
185—86, 18’?, 194
truth, Th. 108. 2U len u 5 age,
a flindri at
Ca
Whyte, Wiliiam F., 168, 202
universalizing. 2$
Woolgar, Steve, 125—26 ssunld history, 5’?
nisersaI knnwled$e. 55, 5’. Sb—Si)
ssorld-systeaas tho:y. i-s-i
universal predicarts,
Woicnaan,C B. 166
unity Sec quantity 1$ 1 5” 55
Satcrrr-r. Sianui,
•iids
6
244--I lavisrock lnsr1tute, :a’sirus en ‘1 ‘to I
10,
Ott, in ni
on l,eur,sr.c 1’
1
‘ii
)
5 nan
5
-i’)
nato
inn.
stli
iS,
1
1’
xtu 6
ans
it.
lb ut
511
,-
c I
(as,
1), 09
0,
‘?
C
1W s
I.
1
anal
3,12
values, social, 49 50
119
a mithods,
,,.cP lltiusros,
N-i,
pure
ci
1
15
,
2
s,84 91 ,2
ale 107,
IS, 2-i.;
,nd card
leon
woilcl-syat
i
Phtrhoril. “C
r
and iron ,
IlcdnstiSs,
emptr am, 21$ ron
h 5 ain mdci 24, 1
a))—” e.
146,
wrong ideas, 21 5—i6. 2i
156
tcrnporality, 31, 34 56, 5” rests of ideas. 211—12, 21—21
and
or
.6
IJ
and argument
ii S
j07,,L0
sacali
7
,
,..,Lisa.)’
Ic
Wrightsun, Keith, 198
93
220
ren
\\/a,ner, sX( Lluyo. 1 Weber, Max. 32, t)9. h’
206—9
9,4(161
t nns
ssnchesis,
variables; standaid
causal analysIs
92—95, 10—9 01,
5
totality. See quantity
38
tel
Ot,
01
Sec u/au dependent variables’ independent
rcporter6 list
:ira,onc
C 155
162
2. 28. 29—30, 31. 32, 33—3d.
diP 071
Li.
ts., 14. 163, 164—
anaLy
i—l4,22$
nioL
5”. “U. ‘2,
“9
6
it
heuristics of 7, 92-S 109 135, 137,
65
n6 More ill
lie;
2i—24
on
n
Synlu St Lr’. 07 (‘1
2u
tupiCs.
Sr
1
and maccal heurisric 181 —92
Ic. 1
Suu.itrnan. 6107k, 20
7
N’)
)U,
u1isca,nn 3,, Reieron
burnt
:1
—‘
‘intervening,” 2)
idci.in’,m
‘l,nta_r,C,c,L.e
timc,
urlm)cis. o”. OS.
,laoatnr’ progcamr, a 1 and c . h, :5 7
‘Ihucvdidcs, 30 S
5 c L-s or
interuependrncc of.
urcy
a: 0
iSa St 190
Tolbont, (,hards, 193
10;
055 iSOpt.
,0j,z
anek sis.
iOU
Thomas W I
‘10cqciesLle, Ale’ns dc, o2
i,,
lzcd act
,‘,j ,iijs
paris Ins, CS
t
taosal
i[id, f
and
‘150mb. D. S
its,
1
11
07
a
(Si ‘a,i, 1 “‘i
—“6
fhcrnsrrom, Scephon. 1”
“—
-i
‘.
ariahles,
Zimmerman, Don, 1”2 Znaniecki, I’lorian, 183—84