-
James S. Coleman as a Guide to Social
Research
Author(s): Charles
Tilly
Source: The American
Sociologist, V
ol. 28, No. 2 (Summer,
1997), pp. 82-87
Published
by: Springer
Stable URL:
/stable/27698831
Accessed:
23-03-2015 15:00 UTC
Your
use
of
the
JSTOR
archive
indicates
your
acceptance
of
the
Terms
&
Conditions
of
Use,available at
/page/info/about/policies/
JSTOR is a not-for-profit service that
helps scholars, researchers, and students
discover, use, and
build upon a wide
range of contentin a trusted digital archive. We
use information technology and
tools to
increase productivity and facilitate new forms of
scholarship.
For more information about
JSTOR, please contact support@.
Review Essay
James S. Coleman as a Guide
to Social Research
Charles Tilly
Despite his outstanding empirical work
and his exceptional clarity as a theorist of
collective
choice,
James
S.
Coleman's
program
for
social
science
would
take
it
in
the
wrong
direction.
Whatever
else
it
does,
social
science
must
explain
social
processes.
Coleman:
1)
neglected
to
specify
causal
mechanisms,
2)
promoted
an
incomplete
and
therefore
misleading
psychological
reduction ism,
and 3) advocated a form of general theory?rational
choice analysis?that cannot, in
principle or in practice, provide
adequate explanations for the great bulk of social
processes.
Overview
Precisely by producing
superb work in a flawed vein, James S. Coleman set
a bad example
for us lesser social
scientists. He neglected to specify causal
mechanisms, promoted an incomplete
and
therefore misleading psychological reductionism,
and advocated a form of general
theory
——
rational choice
analysis
——
that has for some
time been enticing social scientists into blind
alleys.
All
this
he
performed
with
panache,
attracting
attention
to
social
science
as
a
basis
for
public
policy in a way that
few of his predecessors had managed. His legacy: a
generation of researchers
who
misconstrue the social sciences' explanatory
mission because they have adopted a vivid but
ultimately unsound view of how social
processes operate.
This
essay,
inspired
by
Jon
Clark's
James
S.
Coleman
(1996)
but
drawing
especially
on
Coleman's
own
Foundations
of
Social
Theory
(1990),
neither
deals
with
the
many
thoughtful
articles in Clark's compilation nor
reviews the whole of Coleman's corpus. In order to
make my
argument sparely, and thereby
clarify legitimate grounds of debate, I will
ignore many admirable
features of
Coleman's work, including its boldness, its wide
range, its strong empirical grounding,
its clarity of exposition, and its
incessant concern for human welfare. The essays in
Clark's volume
amply
document
these
merits.
Although
I
cheer Coleman's
insistence
that
social
science
should
concentrate on the
explanation of collective phenomena rather than
individual behavior, I will pass
over
the ingenious ways in which he derived
sophisticated collective arrangements from simple
combinations
of
individual
decision-making.
Because
it
characterizes
his
theoretical
and
programmatic writings less well than
his many empirical analyses, I will save for
another occasion
my complaints that
Coleman: 1) conducted analyses based on the
assumption of a world arraying
itself
as a set of independently specifiable continuous
variables, and 2) offered variance-reduction
as a mode of explanation, although I
think the two features seriously limit the value
of his major
empirical investigations
as models for social research.
This
critique
will
also
neglect
debatable
ideas
in
the
book
under
review,
such
as
Martin
Bulmer's
surprising
assertion
that
and
social
policy
as
academic
fields
are
relatively
uneasy
bedfellows
(Clark
1996:
103).
Surely
of
all
social
science
disciplines,
sociology
most
rapidly changes
direction in response to current public-policy
concerns, quickly taking up violence
in
times
of
violence,
education
in
times
of
educational
reform,
and
adolescent
childbearing
in
times
of
public
outcry
over
pregnant
teenagers.
While
each
of
the
book's
thoughtfulessays
provides
ample incitements to comment?affirmative or
otherwise?in this review, it will suffice to
challenge
the
endorsement
of
Coleman's
explanatory
program
that
pervades
Jon
Clark's
semifestschrift.
Taking for granted James Coleman's many virtues,
let me concentrate on dangers
in the
theoretical pro gram he advocated during his later
years.
James
Coleman
codified
and
exemplified
understandings
of
social
processes
and
their
explanation that have become
sociological orthodoxies. To repeat, he:
1. neglected to
specify causal mechanisms;
2. promoted an incomplete and therefore
misleading psychological reductionism; and
3. advocated a form of
general theory?rational choice analysis?that has
for some time been
enticing social
scientists into blind alleys, where they have
wandered aimlessly, falling victim to
local thugs and confidence men selling
various brands of individual reductionism.
To be sure,
Coleman himself forewarned readers not to trust my
judgment in these matters.
Placing me
in the excellent company of Arthur Stinchcombe,
Neil Smelser, Harrison White, and
Michael Hechter, Coleman identified me
as a re viewer who
Social Theory (Clark
1996: 6-7). Coleman's indictment shields me from
the suspicion of having
held
my
tongue
until
he
could
not
reply,
but
it
also
indicates
that
he
would
have
dissented
enrgetically from my three claims about
his work. Caveat ergo lector. Let me draw my
illustrations
of these points from
Foundations, which Coleman himself identified as
his summary and synthesis
of previous
work.
1.
Causal
Mechanisms
Social
science
in
general
suffers
from
the
poverty
of
its
well-
specified,
plausible,
verified
causal
mechanisms:
recurrent
cause-effect
sequences
of
general
scope.
As
compared
to
the
constant invocation of
mechanisms in, say, molecular biology or
seismology, social science rarely
points
confidently
to
causal
mechanisms
that
recur
across
a
wide
range
of
circumstances.
Coleman's
work provides little relief from that poverty. He
takes a pragmatic view of explanation:
intended
pragmatic
criterion promises no means of discovering
generally valid causal mechanisms.
Coleman
does
not,
however,
retreat
into
agnosticism
or
particularism.
On
the
whole,
he
actually explains a
social phenomenon by its service to the interests
of its participants or to their
perceptions of those interests. The
trouble starts there: Coleman does not specify the
mechanisms
that translate a set of
interests?or perceptions of interests?into their
realization in operating social
arrangements.
While
allowing
that
states
ordinarily
come
into
being
and
acquire
their
basic
organizational forms
(their
example, Coleman provides no
account of the process by which state agents
continue to receive
compliance and
resources from underdogs while underdogs somehow
coexist with states that hurt
their
interests (Coleman 1990: 346-49).
Again,
Coleman
rightly
points
out
three
changes
since
World
War
II
in
the
conditions
affecting
American high schools' authorities over their
students: declining parental authority over a
wide
range
of
children's
behavior;
decreased
consensus
within
the
populations
of
districts
from
which high schools draw their students;
increased use of litigation by those who challenge
school
authority (Coleman 1990:
350-51). But he offers no description of the
causal chain by which these
conditions
actually weakened the positions of teachers and
admin istrators, much less how they
translated into changed relations among
day-to-day participants in high school life. For a
specialist
in
the
organization
and
effects
of
schooling,
the
omission
is
remarkable.
As
Aage
Sorensen
comments more
generally:
Coleman did not follow his own
principles in his educational research. Instead,
he went from
some theory
to
no theory about school effects from 1962 to 1966,
and then went back to some
theory in
1982. The integration of theory and evidence was
never completed. There is no explicit
statement about how schools may
influence learning in any of the studies; only
suggestions about
how private schools
may be more effective than public schools. (Clark
1996: 209)
Throughout
Coleman's work, in fact, we discover a strong
preference for arguments about why a
given array of actors
might
choose a given social arrangement over
specifications of the causal
processes
resulting in such arrangements. Consequence: a
decisive shift away from explanation.
2. Psychological
Reductionism
Coleman began
his magnum opus with dissent from prevalent
social-scientific concentration
on
individual
behavior.
He
clearly
distinguished,
furthermore,
between
two
ways
of
explaining
individual
behavior:
1)
examination
of
processes
internal
to
individuals,
and
2)
statistical
association of
individual behaviors with
of
that
behavior
arguing
correctly
that
neither
gives
analysts
much
purchase
on
social
phenomena (Coleman
1990: 1-2). Even more decisively, he rejected the
holisms (
he
called
them
in
1987:
Swedberg
1990:
49-50)
in
which
social
systems
either
acted
autonomously
or
shaped
individual
behavior.
He
failed
to
mention
the
other
alternative:
a
relational approach in which
interactions, trans actions, or social ties
constitute the starting points
of
social
analysis.
Those
rejections,
explicit
and
implicit,
drove
him
to
a
peculiar
variety
of
psychological
reductionism.
How
so?
Social
phenomena,
in
Coleman's
most
elementary
account,
emerge
from
the