. 6
( 7)


But indeed, if we think of “structure” in terms of the analysis of mech-
anisms as explored in chapter 4, we will have all the advantages of her
structuralist perspective and none of the patent disadvantages. That is,
we reject “voluntarism” because agents are always both enabled and con-
strained, whether they are the typical actors of “structures” (“¬nanciers,”
members of the “dominant class,” etc.) or the acts of monarchs and pow-
erful others who, in C. Wright Mills™s terms, make decisions of major
social importance. We reject rei¬ed “structure” and do not give it causal
status. Instead, social mechanisms are sustained (and transformed) by
the actions of persons, represented typically. But while acts are both con-
strained and enabled, they are not “determined.” This holds as regards
the actions of typical actors and as regards the causally consequential acts
of critical individuals or groups. Contingency remains a critical feature
of the actual outcome.
Skocpol was not insensitive to this problem. It would be addressed
in her narratives of each of the cases, both with promissory notes of
mechanisms and by identifying critical decisions by critical actors.15 For
example, after noting that “most Bourbon kings had survived debt and
bankruptcy,” she wonders why “the troubles of Louis XVI developed
into a major crisis” (1979: 63). The mechanism (in the way of a promis-
sory note) is fairly straightforward: in the past the Chambers of Justice
had been able to cancel debts to ¬nanciers. But when “high accoun-
tants” became “nobility” and merged with the traditional ruling class,

15 Skocpol notes that Sewell was “correct to say that I should have devoted more space in
1979 to discussing the methodological connections between the comparative structural
and the conjunctural-narrative aspects of the investigations and presentations that went
into States and Social Revolutions” and notes also that “a very similar point was made
years ago by Peter Manicas, 1981, in one of the best reviews originally written about
my book” (1994: 333). Manicas appreciates her generosity but would now be clearer
that the problem was trying to combine structural / conjunctural analysis “ including
the contingent acts of key identi¬able agents “ with “explanatory generalizations” aimed
at providing “the suf¬cient causes of social revolution.” On the other hand, if Sewell
believes that “the narrative achievements of States and Social Revolutions exist “ or could
have been arrived at “ apart from the macrocausal comparative analysis” (1994: 313)
he is, as she insists, “dead wrong.” The macro-causal analysis did not, as Sewell rightly
sees, provide the necessary and suf¬cient conditions for social revolutions. But at least
as I read him, he does not deny the importance of comparison properly conceived.
Appendix B 169

this was no longer possible. When the costs of war mounted, there was,
accordingly, a ¬nancial crisis. But a ¬nancial crisis, too, might have been
managed. In turn, in 1787, another mechanism comes into play: “news of
the monarchy™s ¬nancial peril precipitated a general crisis of con¬dence
within the dominant class” (1979: 64). But the narrative is still radically
incomplete. Thus, “as everyone knows, the summoning of the Estates-
General [by the king] served not to solve the royal ¬nancial crisis but to
launch the Revolution” (1979: 65). Her narrative is replete with instances
of this sort. Indeed, no explanation of a social outcome can do entirely
without reference to the causal mechanisms sustained and transformed
by the actions of (typical) persons and the causal consequences of major
actors. And the story will always be complex “ and incomplete.

The role of comparison
As regards the uses of comparisons in historical sociology, we may con-
clude with a useful observation of Skocpol. She wrote that:
“comparative history” is commonly used rather loosely to refer to any and all
studies in which two or more historical trajectories of nation-states, institutional
complexes, or civilizations are juxtaposed. In this very broad sense, the term refers
to studies with very different kinds of purposes. Some . . . are meant to show that
a particular general sociological model holds across different national contexts.
Other studies . . . use comparisons primarily to bring out contrast among nations
and civilizations taken as synthetic wholes. But there is a third version . . . in which
the overriding intent is to develop, test, and re¬ne causal, explanatory hypothe-
ses about events or structures integral to macro-units such as nation-states.
(1979: 36)

We agree with Skocpol that the ¬rst purpose, to show that a general
model holds, cannot be sustained (see chapter 5). The second purpose is
too narrow. First, there is no reason to restrict comparisons to nations or
civilizations understood as synthetic wholes. Of considerable importance
is the effort to compare social mechanisms, paths and processes, both for
their “resemblances” and for their differences. Second, one obvious goal
is to explain these differences “ as Weber and Moore insisted. The third
version, then, is correct “ if one abandons the covering law model and
instead thinks of “causal explanatory hypotheses” as hypotheses about
social causal mechanisms.
The comparative method, accordingly, serves a variety of roles, but it
does not yield explanatory generalizations: it does suggest causal hypothe-
ses (hypotheses about causal mechanisms) and, as Skocpol says, “com-
parative historical analysis does provide a valuable check, or anchor, for
theoretical speculation” (1979: 39). She was correct in describing her
170 A Realist Philosophy of Social Science

effort as following the path set out by Barrington Moore, but as I have
tried to show (chapter 5), Moore was fundamentally Weberian in his
approach to historical explanation. His narratives gave causal accounts
of three paths to modernization; Skocpol offered narratives of three paths
to modern social revolution. As argued, in the different paths, there will
always be “resemblances” and there will always be important differences.
Skocpol™s conclusions were modest. She asked whether the “broad
resemblances” she identi¬ed could be applied beyond the three cases of
her focus, and answered, “unequivocally ˜no™.” She gives two reasons.
Firstly, as Mills insisted, “the mechanisms of change . . . vary with the
social structure we are examining” and second, “patterns of revolution-
ary causation and outcomes are necessarily affected by world-historical
changes in the fundamental structures and bases of state power as such”
(1979: 288). But why stop here? Why not assert that this is true of all
history, all cases, all structures? “Resemblances” are just that: they are
neither suf¬cient nor necessary conditions. They do not explain.
Sewell, in summarizing what he called, “eventful sociology,” summa-
rizes matters well enough:
Sociology™s epic quest for social laws is illusory, whether the search is for time-
less truth about all societies, ineluctable trends of more limited historical epochs,
or inductively derived laws of certain classes of social phenomena. Social pro-
cesses . . . are inherently contingent, discontinuous, and open-ended. Big and
ponderous social processes are never entirely immune from being transformed
by small alterations in volatile and local social processes. “Structures” are con-
structed by human action, and “societies” or “social formations” or “social sys-
tems” are continually shaped and reshaped by the creativity and stubbornness of
their human creators. (1996: 272)
Appendix C Rational choice theory and
historical sociology

Rational choice theory (RCT) has become an important part of the
debate not only in general sociology but in historical sociology in par-
ticular. This appendix picks up on several themes of previous chapters
and in terms of them examines a very recent and lively debate, occur-
ring both in the pages of The American Journal of Sociology (AJS) and in a
recently edited volume.1 This will, hopefully, both sharpen the issues and
provide further evidence of the pervasive legacy of empiricist philosophy
of science.

The AJS debate: realism and causality
Craig Calhoun (1998) notes correctly that a key point of difference
between this debate and the earlier debates is agreement by the contend-
ing parties on the centrality of causal explanations in historical sociology,
and, in particular, the acceptance of “the more recently fashionable label
of ˜realist™ philosophy of science” (1998: 847). There are two critical
questions here.
In this debate, in marked contrast to earlier ones, no one on the panel
defends what Skocpol called “interpretative historical sociology.” In their
1991 essay, Kiser and Hector had asserted that there is “wide agree-
ment . . . across social science that causality is the ¬rst requirement of
an adequate explanation” (1991: 4). This is a highly dubious empirical
claim (as noticed by both Somers and Calhoun).2 It rules out, by ¬at,

1 The symposium in The American Journal of Sociology (vol. 104, November 1998) had its
genesis in a panel on history and theory at the 1989 meetings of the American Sociological
Association. Drafts circulated among a number of important writers. The result was the
AJS Symposium which has contributions by Margaret Somers, Edgar Kiser, Michael
Hector, Craig Calhoun and Jack Goldstone. In what follows, unless otherwise noted,
references to these writers are from the published symposium. See also Roger V. Gould
(ed.), 2004.
2 Calhoun (1998) is probably correct that even among those who think of explanation in
terms of causes, there is little critical re¬‚ection of what this means “ a point addressed
in much of the present volume. Indeed, as he later writes, following Boudon, “most of
what passes for causal analysis in the social sciences is in fact identi¬cation of more or
less ˜weak implication™ between statistical variables” (1998: 866).

172 A Realist Philosophy of Social Science

all interpretative sociology, a still important alternate conception of the
human sciences.
As argued in chapter 3, interpretativists are not wrong in insisting on the
importance of meaning and hence of ideas in their explanatory efforts:
the question is whether, as Weber insisted, these need to be part of a
larger causal argument. Interpretativists rightly rejected the dominating
empiricist framework for explanation, including Humean causality and
the covering law model, but seeing no alternative consistent with their
understanding of human action, they threw the baby out with the bath
This is the second critical feature of the AJS argument. Both parties,
Kiser and Hector, and Somers, say that they are “realists.” This suggests
that the once dominating empiricist philosophy of science is on the wane “
at least among some sociologists. As with Kiser and Hector™s claim regard-
ing explanation, this too is doubtful “ at least if realism includes some
key theses regarding explanation. Realism, understood as an ontological
position, holds that the real is not restricted to what is “in experience.”
Accordingly, as argued in chapter 1, causal mechanisms can be non-
observable. This is a healthy step in the right direction, but will not, of
itself, excite either most social scientists or most ordinary people. As a
consequence of its ontological position, realism is also an epistemological
position which can (and generally does) accept now-standard Kuhnian
criticisms of empiricist epistemology. Both parties to the AJS debate sug-
gest they have very different views on this “ a point to which I return. But
the most obvious problem with their realism is the idea of causality, and,
in consequence, their “realist” views of explanation.
It is not clear whether the writers in the symposium have fully ingested
a realist concept of causality, the burden of chapter 1 of this volume.3
As argued in chapter 1, on a realist conception, causes are productive
powers that bring about outcomes. They may be expressed in singular
causal statements, e.g., “Sam crushed the cracker,” or as promissory
notes of mechanisms: “The iron rusted because it oxidized.” The mech-
anism here, of course, is fully developed in molecular chemistry. In neither
case are causal statements to be understood as law-like contingent regu-
larities. It follows, accordingly, that explanation does not take the form of
the covering law model. This is clearly seen by Somers. Kiser and Hector

3 Calhoun notes that Kiser and Hector make no use of the work of Roy Bhaskar. Somers
does, but rejects exactly what is at the heart of Bhaskar™s account: “that causality is found
in an entity™s essential properties (1998: note 13, 743f.). See below. Neither party seems
aware of the extensive development of realist causality in the work of Rom Harr´ (1970,
1975), and with special reference to social science, Varela and Harr´ (1996).
Appendix C 173

are less clear. Thus, Calhoun reads them as follows: “When Kiser and
Hecter refer to explanation, they take it for granted that it means causal
explanation in a covering law model” (1998: 856). There is a good deal of
textual evidence for this in their 1991 paper, but little in their 1998 essay.
In the 1991 essay, the problem would seem to result from their under-
standing of causality “ which in the intervening years may have shifted.
For example, in 1991, they wrote: “Causal uniformity implies the exis-
tence of law-like relationships that hold between events . . . In essence,
causal explanation works by subsuming events under causal laws . . . and
causal laws, in turn derive from general theories” (1991: 6).
Since realist causality is not understood in terms of law-like relations
between events, and thus causal explanation is not subsumption of events
under law-like statements, these remarks leave little room for doubt “ at
least as regards their posture in 1991. Somers, with considerable justi-
¬cation, insists that Kiser and Hector incoherently accept both the idea
of scienti¬c explanation by the covering law model and of explanation by
means of “abstract models” of mechanisms “ a distinctly realist move.
Chapter 1 of the present volume offers arguments that indeed, these are
inconsistent. The fundamental problem can be summarized here: the
covering law model holds that the explanandum is a logical consequence
of the explicans; the realist view defended in this volume insists that this
is the wrong relationship: we explain when we know what produced the
The problem is compounded by the fact that, despite their frequent use
of the term, none of the writers in the symposium provides a clear idea of
what they mean by “causal mechanism.” Kiser and Hector note that, at
least in the natural sciences, “mechanisms typically explain outcomes by
invoking phenomena at a level of analysis lower than that of the outcome
4 Calhoun says that “rational choice theory has played an important role in encouraging
more emphasis on causal mechanisms alongside covering laws or causal relationships,”
suggesting that he accepts that the two ideas can cohere. But he also notes that “Kiser and
Hector are actually somewhat vague about what is involved here” (1998: note 5, 851).
As Calhoun suggests, much is at stake here. The remainder of his note is helpful, but
especially his observation that “their language shifts a bit between seeing the mechanisms
as necessarily deduced from lawlike statements, and as simply being lawlike statements
but of a different order from relationships of implication or correlation” (1998: note 5,
851). Seeing mechanisms as “necessarily deduced from lawlike statements” suggests fun-
damental misunderstanding. Law-like statements entail nothing about the nature of the
mechanisms. What they are is the problem of theory and, as always, theory is under-
determined by “facts,” including con¬rmed “law-like” statements. Nor is it helpful to
say that mechanisms are “law-like” statements of “different order,” even where aspects
of mechanisms can be represented mathematically. See chapter 1 and references there.
But it is correct to say that causal statements are of a different order from relationships
of implication (as in the D-N model) or “correlations.”
174 A Realist Philosophy of Social Science

of concern” (1998: 790). But in part at least, because of their lingering
commitments to the covering law model, there is no attention paid to how
a realist understanding of causality and theory gives us understanding.
Somers does better by providing an example. Thus, “since food has calo-
ries, and calories are energy, when we reduce our intake then the body
has less energy to draw from external sources so it has to turn to internal
sources, which are stores of fat, and the body uses up fat when it draws
that energy, and so forth” (1998: 770). This is, as she says, a “causal
narrative.” It takes for granted a host of undetailed causal mechanisms
represented by a host of theoretical ideas. Presumably, there will be an
analog in social science.

General theory?
Kiser and Hector (1998) see themselves as following a more or less
standard realist account of theory. Thus, they endorse the following
two de¬nitions: “[Theories] are causal explanations providing intelligi-
ble answers to why-questions about empirical facts”; and, “theory can
be taken to mean a set of assumptions or postulates with which one
approaches some part of the empirical world, and a set of propositions,
emerging from the assumptions and relating the concepts, about the way
this part of the world ˜works,™ which is checked against observations of
that world” (1998: 793, note 20). These formulations smack somewhat
of deductivism, but are suf¬ciently vague to be interpreted in realist
terms. As above, we need to know more about how theory represents
“mechanisms” and how they explain.
But there is a prior problem, regarding the question of “general the-
ory” and exactly what Kiser and Hector mean by this. It is important
to be clear about this since it may well be that while there are perfectly
adequate theories of speci¬c historically grounded mechanisms which
produce speci¬c historically bounded outcomes, there may be no use-
ful or interesting general theories which apply willy-nilly to such out-
comes. This is Somers™s view of the matter, shared, as noted, by Skocpol
(chapter 5, above) and the present writer. The position of Kiser and
Hector is far from clear.
Compare here, molecular chemistry and Marx™s Capital. Molecular
chemistry is surely a general theory in the sense that it provides an account
of mechanisms which apply to all matter qua chemical, everywhere and
anywhen: the mechanisms are omnitemporal. But the theory articulated
in Marx™s Capital applies only to capitalist societies. Nor indeed can the
mechanisms theorized there be “generalized” to offer an understanding
of non-capitalist political economies.
Appendix C 175

Kiser and Hector write: “Although philosophers of science consider
mechanisms to be abstract and omnitemporal (Bunge 1996), our critics
[viz. Somers, etc.] prefer them to be historically speci¬c” (1998: 796).
But, contrary to Kiser and Hector, philosophers of science must surely
consider whether the absence of omnitemporal mechanisms in social sci-
ence may be the consequence of a key point of difference in the ontologies
of the natural and the social world. A main goal of the present volume
has been to show that explanation (and understanding) in both domains
requires an account of mechanisms. But it has also been a major theme to
note that there are important disanalogies which stem from ontological
Kiser and Hector want us to be clear that “the principal claim of [their]
article was that general theory is useful in historical explanations since it
is an important source of causal explanations. General theories provide
both the omnitemporal laws [sic] that animate contextual models . . . and
the guidelines necessary to attack particular substantive problems” (1998:
793). This is puzzling, but especially the idea that we need “omnitemporal
law” and that these “animate contextual models.” Kiser and Hector do
not help their cause, partly at least because of confusion over whether or
not they remain committed to the covering law model.
But a footnote follows the text just quoted. It opens new possibilities.
They write, quoting Kuhn: “ ˜General theories™ supply the group with
preferred or permissible analogues and metaphors. By doing so they help
to determine what will be accepted as an explanation and as a puzzle-
solution; conversely, they assist in the determination of the roster of
unsolved puzzles and the evaluation of the importance of each” (Kiser
and Hector, 1998: note 21, quoting Kuhn, 1970: 184). The text quoted
from Kuhn does not speak of “general theories.” It speaks speci¬cally
of “models,” “heuristic” and “ontological,” and in the context Kuhn is
plainly referring to what he was calling, following his abandonment of
the fuzzy notion of a “paradigm,” “the disciplinary matrix.” This is an
important Kuhnian point, now very much a part of current sociology of
science and realist theory of science (Pickering, 1992).
If “general theories” are “disciplinary matrices,” then the puzzlement
can be overcome. Thus, within the Kuhnian frame, the social sciences
are marked by the existence of competing disciplinary matrices. Kiser and
Hector aim at offering what was termed in the present volume, a meta-
theory for a human science: an effort to set out what should be accepted
as an explanation and as a puzzle-solution in social science. Indeed, they
are very clear in asserting that, for them, “all good sociological explana-
tions must consist of separate arguments pertaining, ¬rst, to the motives
of individual actors and, second, to models of the contexts within which
176 A Realist Philosophy of Social Science

their action takes place” (1998: 799“80). On this reading, their “general
theory” tells us where to look for causes and what a social mechanism
must include. And on this reading, RCT is but one of a family of possible
“agent-as-mechanism” theories “ including strong versions which effec-
tively deny history. Of course, agents omnitemporally are motivated and
have productive powers “ indeed, this assumption is “essential” to our
understanding of persons. But it allows for the generation of a number
of alternative theories of mechanisms including, plainly, those which are
historically grounded and limited in their application. It provides, as Kiser
and Hector claim, a source which can “animate contextual models . . . and
the guidelines necessary to attack particular substantive problems”
(1998: 793). Similarly, ruled out are alternative “general theories” “
disciplinary matrices, meta-theories “ that look elsewhere for causes, for
example, in social structures.5
Also puzzling is their idea that “the mechanisms derived from general
theories are generalizable “ they can be used in different substantive areas
and historical periods” (1998: 706). Here they may be thinking of strong
ahistorical versions of RCT. In any case, one wonders whether generaliza-
tion is being confused with abstraction. That is, moving from the concrete
to the abstract (abstraction) should not be confused with moving from a
particular instance to a generalization about all such instances. Thus, it
would be an error to argue that, for example, the mechanism identi¬ed
which explains unemployment in Japan can be applied to understanding
unemployment in the USA “ or worse, the People™s Republic of China.
On the other hand, in the case of the USA and Japan “ but not China “
Marx™s highly abstract account of the mechanism of capitalism is true
of all and only capitalisms. It would be implicated in developing the
mechanisms of unemployment in Japan and the USA “ but again, not
of China. That is, since for good historical reasons, Japanese and US
capitalism are concretely different in critically relevant ways, the mecha-
nism which explains Japanese unemployment will be different to the one
which explains unemployment in the USA even if capitalist mechanisms
are working in both. Similarly, assume that a principle of an abstract
mechanism is the (Hobbesian) idea that people seek power. If this is not
vacuous (as is often the case), then to give it explanatory bite, one must
specify the particular conditions of persons, their speci¬c motives and
capacities. For example, even if at some high level of abstraction people
seek power, the social mechanism which explains monarchic succession

5 This reading may well be rejected by Kiser and Hector. If so, then perhaps there remains
a lingering covering law view.
Appendix C 177

will be very different to the one explaining succession of the chief execu-
tive in a republic.
There are, as noted, alternative “general theories” (“disciplinary matri-
ces”). Parsonian “general theory” is one obvious candidate.6 It gives an
omnitemporal conception of society as an ensemble of connected sys-
tems, each having distinct functions. It tells us the explanation proceeds
by identifying the concrete instantiations of these and then showing how
dysfunctions arise between, for example, the structure of the personality
system and the social system. Indeed, while she might deny it, Somers
inescapably also has a “general theory” “ understood here as a meta-
theory which de¬nes what a good sociological explanation must consist
of. As noted, Kiser and Hector insist that “all good sociological explana-
tions must consist of separate arguments pertaining, ¬rst, to the motives
of individual actors and, second, to models of the contexts within which
their action takes place” (1998: 799f.). This is precisely what is denied
by Somers. Unfortunately, the debate got mis-couched as a debate over
rational choice theory. We need to clarify this and then to set out her
alternative. Finally, there is the question of what role, if any, narrative has
to play in explanation in historical sociology.

Assessing the debate
We should notice, ¬rst, that the Kiser / Hector conception of what all good
sociological explanations must consist of is perfectly consistent with the
argument of chapter 4 (above); second, that it rejects alternative con-
ceptions, e.g., functionalist theory; and that, third, as Somers argues, it
inevitably includes a number of ontological commitments. One can reject
these, to be sure, but if so, then some alternative set of commitments
must be made.7 Somers understands perfectly well the ontological com-
mitments of the Kiser / Hector metatheory “ even if they are sometimes
expressed by her in a somewhat distorted fashion.
6 No doubt the idea of “general theory” is a sociological invention spurred by Parsonian
universalistic theorizing. Merton (1957) famously argued that sociology needed “theo-
ries of the middle range,” theories which were suggested by “general theory,” but were
designed to address the historically concrete. And of course, Merton remained within
structural functionalism.
7 Somers writes: “the protestations of some social scientists notwithstanding, all theories of
knowledge make a more or less explicit ontological choice between either the individual or
the social structure as the basic unit of social analysis” (1998: 750). The argument of the
present volume took the side of “the individual” as the sole causal power in society, but
insisted also that a Giddens-type explication of social structure gave us all that we needed
to avoid the pitfalls of “methodological individualism.” But it would seem to follow that
if Somers rejects agents as causes, then social structure does the explaining.
178 A Realist Philosophy of Social Science

Thus, she writes that there is “a commitment to a causal ontol-
ogy in which agential intentionality is posited the a priori causal
force / mechanism at work in the social world” (1998: 750). Indeed, agen-
tial intentionality is posited as the causal force at work in the social world,
but, we may ask why is this posit a priori? Surely, everything in experience
says that agents are causes and that action is “intentional,” even if it is not
always self-conscious and even if its consequences were not intended?8
Some of her criticisms seem sound, but not versus the fundamental
idea that agents are causes or that causal mechanisms must be models of
the contexts within which action takes place. Her criticism is apt versus
a version of rational choice theory “ at least as she understands it. Thus,
she attacks “essentialism” and a Hobbesian “pre-social” conception that
posits “not only ¬xed solipsistic identities but ontological entities that are
born preequipped to act through essential inherent causal mechanisms
(reasons as causes) that drive action on their own autonomous momen-
tum” (1998: 764). But there is nothing in an agent-centered ontology “
including interesting versions of RCT “ which requires that persons are
“pre-social” or that they have ¬xed solipsistic identities. And, of course,
people do act for reasons, and as argued in chapter 2, reasons are quite
properly conceived as causes.
To be sure, a good deal of RCT is Hobbesian and does err in precisely
the ways that Rousseau said that Hobbesian theory erred.9 On the other
hand, if suf¬ciently weakened, then while it sees that agents are causes,
it is not clear whether it is any longer the theory that it claims to be.10

“Theoretical realism” and “relational realism”
These terms are Somers™s and, as Kiser and Hector rightly see, “despite
her claim that we share a commitment to epistemological realism, our
philosophical differences with Somers are profound” (1998: 88). The
main concern in this section is to examine the ontology which Somers
offers as the alternative to the “agents as mechanism” ontology of Kiser

8 Nor is it clear what could be meant by saying that the causal powers of agents is “exoge-
nous” (1998: 750), except perhaps that she thinks of agent causality in mentalistic terms.
Thus, “the explanatory work of the theory is carried out by this invariant causal mech-
anism of a dispositional agential intentionality that necessarily (in the absence of con-
straint) causes intents to convert to actions be they rational or irrational” (1998: 751).
Again, actors have reasons for their actions and reasons are quite legitimately causes.
But this does not require a Cartesian metaphysics. Indeed, it is inconsistent with one.
9 In his too often unread “Discourse on the Origin of Inequality.”
10 Among examples are the writings on so-called “exchange theory,” for example, Blau,
1964 and Coleman 1988, 1990. See also Chai, 2001.
Appendix C 179

and Hector. But to do this we need to say more about the two realisms
as Somers sees them.
First, Somers seems to believe that a post-positivist, post-Kuhnian epis-
temology distinguishes “theoretical realism” from “relational realism.”
But this is hardly clear. On the one hand, despite her rejection of the idea
that theory aims to represent reality as it is in-itself, she also holds that
“relational realists believe in the importance of determining which theo-
ries more closely represent reality” (1998: 745).11 So there is no argument
between Somers and Kiser and Hector on this score. And she acknowl-
edges that realists who hold to the view that theory aims to represent
reality, for example, Bhaskar, are also “agnostic about the absolute truth
of any given theory about the world” (1998: 744). So, despite occasional
hyperbole in her attacks aimed at Kiser and Hector™s “theoretical real-
ism,” this too is a non-issue. Finally, both parties can agree on by now
standard criticisms of the empiricist epistemology, including the turn
taken by post-Kuhnian sociologists and philosophers of scienti¬c knowl-
edge. They can, accordingly, reject Baconian conceptions of induction,
Popperian notions of falsi¬cation and assent to the Kuhnian idea that
“only certain types of mechanisms are deemed plausible by the relevant
scienti¬c community at any given time” (Kiser and Hector, 1991: 6).
Difference in the two views is most pronounced when Somers says,
for example, that “there are no universally valid principles of logical
reasoning; there are only problem-driven ones” (1998: 766), and in her
denial of “essentialism,” “a philosophy which looks to the ˜essence™ of
things for information about their ˜true™ nature and behavior” (1998:
764). These are substantial differences, and indeed, Kiser and Hector
infer, not unreasonably, that if logic is “culture bound,” then Somers is
committed to a radically relativist epistemological position “ a position
which in other places she would seem to disavow. More importantly, per-
haps, given her views on these issues, it is not clear that her “relational
realism” can sustain a plausible account of scienti¬c practice as that is

11 She also writes that the two realisms differ in that “while theoretical realism attributes
an ontological truth to the theoretical phenomenon (e.g., the theory of electrons or the
theory of market equilibrium), relational realism focuses on the relational effect of the
phenomenon itself (e.g., the impact of the hypothesized electron on its environment or of
the hypothesized market forces on an observable datum)” (1998: 745). This is puzzling.
The remark smacks of an empiricist understanding of theoretical entities as “convenient
¬ctions.” But surely, if the electron is to have “impact,” then it must exist. Mechanisms
can explain empirical outcomes only if theorized entities and their powers are part of the
ontology of the theory. On the other hand, Somers™s criticism of explanations provided
by neo-classical theory takes exactly this position. She writes that one cannot explain
prices in terms of managerial decisions based on marginal costs, exactly because no
manager “has the slightest idea of what the marginal cost of producing something really
is.” She rightly insists: “For a cause to explain, the cause really has to exist” (1998: 770).
180 A Realist Philosophy of Social Science

currently understood by realist philosophy or sociology of scienti¬c
knowledge (Hacking, 1992, 2000). Otherwise, since they draw on now
widely accepted criticisms of positivist philosophy of science, some of the
attacks on Kiser and Hector™s efforts to identify criteria for theory accept-
ability are well-placed.12 At the same time, much of the initial plausibility
of Somers™s alternative to the views of Kiser and Hector is stimulated
by her effort to draw from the work of Kuhn, including, importantly,
her main claims regarding “path dependency and causal narrativity in
explanatory structure” (1998: 731).13
But before looking at this important line of thought, we need to exam-
ine Somers™s relational realist ontological alternative to the ontology of
“agents-as-mechanisms.” There is, unfortunately, very little to help us
here. She writes:
Beginning with the postulate that we are neither monads nor self-propelling enti-
ties but “contingent, transitory connections among social constructed identities”
(Tilly 1995: 1595), a relational pragmatist ontology takes the basic units of social
analysis to be neither individual agents (agent, actor, person, ¬rm) nor structural
wholes (society, order, social structure) but the relational processes of interaction
between and among identities. (1998: 767)
As already noted, there is nothing in the program of defenders of
an agent-centered approach which requires that persons are “monads”
(self-contained, complete and independent) and, consequently, there is
nothing which requires denying either that identities are socially con-
structed or that persons stand in relations that are themselves contingent
and changing. Of course, if self-propelling means capable of acting, then
persons surely are self-propelling. So it is not clear that her postulate
counts as an argument “ even the beginnings of an argument in favor of
her alternative ontology.
It is clear that Somers wants to reject views that “solve” agent / structure
dualism by eliminating either pole of the dichotomy; but for all her good
intentions, she seems very much to generate a structuralist ontology which
(following on her rejection of the ontology of Kiser and Hector) does
eliminate agency.14 Indeed, it is dif¬cult to see what sort of mechanism
she has in mind. She writes that “the basic mechanisms of causality are
12 See Goldstone™s helpful comments on these issues. Some of Somers™s criticisms of
“theoretical realism” depend on what are probably distortions of the views of Kiser
and Hector, as they insist.
13 Somers must be complimented for offering some novel insights here, but one can doubt
very much her claims regarding much of Kuhn™s legacy. For example, a good deal of
anti-realist, even post-modernist anti-science thinking owes to readings of Kuhn. Some
of the writings of Richard Rorty are good examples of this.
14 Some of her remarks on this score are dif¬cult to understand. Thus, her ¬rst “limiting
principle” is that “belief in the causal power of a theoretical social dynamic (e.g., gender,
Appendix C 181

not within discrete agents, but in the pathways of agential interaction . . .”
(1998: 768). The key word here would seem to be “pathways” since
plainly the “agents-as-mechanisms” view focuses on “agential interac-
tion.”15 What, then, is the force of “pathways”? Worse, she seems to step
into a very Hegelian sort of trap. She writes:
In place of a language of essences and inherent causal properties, a relational
realism substitutes a language of networks and relationships that are not pre-
determined but made the indeterminate objects of investigation. Relational
subjects are not related to one another in the weak sense of being only empir-
ically contiguous; they are ontologically related such that an identity can only
be deciphered by virtue of its ˜place™ in relation to other identities in its web.
(1998: 767)

On its face, this seems inconsistent with the text quoted from Tilly
regarding “contingent, transitory connections among social constructed
identities.” It would seem, at least, that if relations are “contingent and
transitory,” they cannot be ontologically related. Indeed, her view seems
at least to takes us straight to the age-old ontological debate pressed by his-
torical idealism: the pertinence of “internal” versus “external relations.”16
Hegel (and Hegelian Marxists) hold that all relations are internal: the
truth is the whole. This is, accordingly, a profoundly relational concep-
tion of reality. Empiricists, by contrast, hold that all relations are external
and nothing is essential in the world. Realists hold that there are internal
relations, e.g., sister / brother, capitalist / wage worker, and these are a
critical part of understanding their “dynamics.”17 How all this bears on
questions of “identity,” however, is quite another matter.

utility maximization, class struggle) is independent from belief in any one particular
theory” (1998: 766). This would seem to mean that her favored ontology is theory-
neutral “ which, of course, it is not. If sexism and class struggle cannot be understood
in terms of the actions of agents working with materials at hand, then how are they to be
understood? Similarly, “a relational realist would use pragmatic reasoning to argue that
despite fate or fashion of any particular theoretical concept such as ˜sex roles,™ ˜sexual
division of labor,™ or ˜gender™ “ each of which represents a different causal conception
of an unobservable postulated reality “ we have reason to believe in the causal force of
that which terms variously attempt to signify largely for one reason: When we dress a
baby in blue, we can observe that people treat that baby differently than when we dress
that same baby in pink” (1998: 744). Fine, but which theory of mechanism gives us an
understanding of this?
15 In a footnote, she observes that her relational realism “¬nds an analytic home in network
theory” (1998: note 29, 768), but one can still ask for the mechanisms, and one can still
offer an agent-as-mechanism account.
16 A relation aRb is internal if both a and b are what they are by virtue of standing in
the relation aRb. By contrast, a relation aRb is external if a and b are what they are
independently of R.
17 Similarly as regards “natural kinds.” Something would not be salt unless it dissolved in
water. See Kripke, 1982.
182 A Realist Philosophy of Social Science

The metaphysics of history
It seems clear enough that Somers wants very much to preserve contin-
gency in history, but it is not clear how this is to be understood. The
problem is the idea of path dependence. Goldstone points out, ¬rst, that
the term is appropriated from contexts in which path dependence is a
property of a system understood in a strict sense.18 But unless Somers
is assuming that society is a system in this sense, then we must assume
that this is rather more a metaphor. But the main idea of path depen-
dence as metaphor is not obvious: it cannot be, presumably, that once
one has started on a path one is fated to continue on it. This makes
“paths” freeways for which there are no exits!19 Something more like
the following would seem appropriate: since “structure” is incarnate in
action, and this is always the product of history, there will be both change
and continuity in history. Path dependency then would seem to reduce
to the near truism that even with change, there is continuity and that
the past cannot be undone. Dramatic changes do occur, of course, and
these do sometimes mark a new path in the sense that some important
former possibilities are foreclosed and others become possible. Beginning
a war is an obvious example. As regards the systems Goldstone refers to,
we can speak of “another run.” But this makes no sense in history. Nor
is there any set of “initial conditions” from which history proceeds, not
only because there is no Day One, but because every act and event con-
tributes to the “making” of history. On the present view, since it draws
on assumptions alien to historical analysis, “path dependence” is at best
a misleading metaphor. “Paths” are historical legacies which enable and
constrain current action; but since we are not “locked into paths,” they
are not determining of the future.
Somers also rightly sees that, in history, there are neither laws nor sets
of conditions from which one makes deterministic calculations. Thus,
Goldstone suggests that

18 Goldstone summarizes: “Path dependence is a property of a system such that the out-
come over a period of time is not determined by any particular set of initial conditions.
Rather, a system that exhibits path dependency is one in which outcomes are related
stochastically to initial conditions, and the particular outcome that obtains in any given
˜run™ of the system depends on the choices or outcomes of intermediate events between
the initial conditions and the outcome” (1998: 834). But nothing in this involves the
effort to explain the choices or outcomes of intermediate events.
19 See also Lawson (1997: 251) who writes: “My worry is that, without due care to the
way it is presented, the approach defended allows, even encourages, the inference that
once an account is provided of how a form of social organization or a technology became
established, this is more or less the end of serious inquiry . . . It facilitates the view that
once a technology or social structure is in place then it can be treated as locked-in for
good; that the past is not only ever present but also all determining.”
Appendix C 183

what Somers and other critics of RCT are reacting to is the tendency of some
practitioners of RCT to grossly simplify the actual complexity of initial conditions
in order to make deterministic calculations of social outcomes . . . Many sets of
initial conditions and interactions are indeterminate “ like the path dependence
of a Polya urn. For most RCT theorists, such problems are uninteresting . . . This
difference in interests is precisely what causes the con¬‚ict or miscommunication
between theoretical economists and economic historians. (1998: 840)
But the difference is here much more than a matter of “interests.” The
deeper reasons are differences in explanation and in the metaphysics of
history. That is, RCT theorists, like mathematical economists, accept the
D-N model, are ahistorical and fail to see the consequences of the absence
of closure. One has a “determined” outcome by calculating an outcome
from a premise set, but, as the historian insists, this is not how historical
outcomes get produced or explained.

“General laws” and mechanisms
To explain some actual outcome, one needs to go back in time and identify
sequentially the pertinent causes as they combined to produce the out-
come. This will require a narrative which links critical actions and events
with ongoing social processes grasped in terms of social mechanisms.
Goldstone argues that we need “laws” to do this. He offers illustratively
the RCT “law” that “people seek to maximize their well-being, usually
de¬ned in terms of wealth, power, or status, through their interactions
with other people” (1998: 833). First, there is a persistent temptation in
this case (and others like it) to reduce laws to tautologies: anything counts
as an instance of power-seeking.20 Second, this “law-like” assertion is a
generalization, not a law, since, if it is not a tautology, there is noth-
ing “nomic” about it.21 To be sure, it may still be a useful generalization:
some acts do proceed from power-seeking motives. Goldstone insists that
20 There is a section in Weber™s Roscher und Knies (1975) where he has great fun with the
idea of “laws” in historical explanation. He refers to a verse by the humorist Busch which
goes as follows: “Whoever is pleased whenever he is distressed makes himself, on the
whole, unpopular.” Weber notes that Busch correctly sees that this is not “a necessary
truth”: it is a non-nomic generalization. And he notes also that anybody who has been
properly socialized will have a battery of such generalizations “ otherwise they could
hardly carry on in society. “Would it,” then, “make scienti¬c sense for [the interpretative
disciplines] to formulate special generalizations and so-called ˜laws™ that are intended
to achieve abstraction? . . . Can this project be expected to produce new useful insights
germane to their concrete problems? In general it is not in the least self-evident that this
must be the case” (Weber, 1975: 107).
21 No end of damage has been done with the term “law-like.” Most are mere generalizations,
which, even if true, can lay no claim to necessity “ the critical attribute of nomicity. There
are plenty of generalizations in history, but no historical “laws” exactly because agents
always could have acted otherwise.
184 A Realist Philosophy of Social Science

Somers is false to her assumptions since while she rejects “general laws,”
she nevertheless appeals to them. According to him, she must since “with-
out the assertion of a necessary or probable connection, there is no causal
account “ it just happened that way” (1998: 833). But Goldstone here
seems to be under the thrall of Humean causality in which true law-like
sentences provide the causal connection. He says as regards the law of
re¬‚ection that “positivists may state the general law, which accounts for
all observations, and be content at that. Realists, however may wish to
explain further what events or principles govern this result” (1998: 833).
But even given a perfect observed correlation between the angle of inci-
dence and the angle of exit, this does not “account for” the observations
(even if it allows for prediction). Of course, once we have the mechanism,
we can “account for” the observations. Indeed, given an understanding
of the mechanism, the relation is shown to be nomic: it had to be what
it is.
But even if we put this problem aside, Goldstone acknowledges that
our Hobbesian “law” can “tell us very little about why particular national
histories, e.g., turn out the way they do” (1998: 833). Thus, to explain
why England surpassed Holland in the eighteenth century, “we must
trace the action of this particular principle through the action of particular
historical actors in particular historical settings” (1998: 834). Indeed. But
we need no “laws” to say that their actions did not “just happen that
way.” We have a causal explanation of the particular decision of some
particular actor, when we have knowledge of the particular motivations
and expectations and particular conditions in terms of which the actor
made choices. Roughly, if persons have reasons for their actions, since
reasons are causes, actions need no covering laws. And while it may well
be the case that in some particular case, the actor was motivated to seek
power, there is no requirement that this “principle” is essential to the
explanation, nor that the “principle” holds in every instance.
Similarly as regards Goldstone™s observation that Somers assumes that
“earlier institutions inevitably leave their impress on subsequent ones.”
This generalization, called by Goldstone “the law of historical sedimen-
tation,” like the Hobbesian “law,” does not take us very far: again, we
need speci¬c social mechanisms. Thus, to explain both reproduction and
change, one needs (as Goldstone would seem to agree) to understand
the motives, expectations and conditions of action of typical groups and
key actors in fourteenth-century England, France or Holland. While our
narrative will no doubt contain references to the decisions of actors that
were, as C. Wright Mills put it, “of major social importance,” it was not
only them who “made history.” We cannot explain the French Revolu-
tion without reference to the acts of Louis XVI, but nor can we explain it
Appendix C 185

without understanding the actions of the sans-culottes of Paris, the French
nobility and the peasants of rural France. But this assumes (if unselfcon-
sciously) the idea of typical actors “ each acting with materials at hand.
The AJS debate suggests that there remains considerable unre¬‚ective
empiricist philosophy still afoot even in highly self-conscious and sophis-
ticated writers. If the foregoing analysis and criticism are near to being
correct, all sides to the debate, while nominally “realist,” fail to see the
deep difference between a realist notion of causality and the conven-
tional Humean constant conjunction view. Indeed, the consequences as
regards theory and explanation are enormous. Similarly, while Kiser and
Hector rightly insist on the causal priority of agency in explaining social
outcomes, rational choice theory as a “general” theory cannot do the
job that they intend. Finally, then, Somers is quite right to reject gen-
eral theory and to insist on the importance of history in explanation, but
her “relationalism,” which certainly seems agentless, would also seem to
create more problems than it solves.
Appendix D The neo-classical model

For the interested reader, this appendix adds some detail and argument
to the discussion of neo-classical economics. Following Hausman (1984),
it may be convenient to distinguish “equilibrium theory” from “general
equilibrium theory.” Static equilibrium theory may be de¬ned in terms
of the following fundamental assumptions. These comprise the core of
mainstream micro-economics:
1. For any individual A and any two options x and y, one and only one of
the following is true: A prefers x to y, A prefers y to x, A is indifferent
between x and y.
2. A™s preferences among options are transitive. (If A prefers x to y and
y to z, A prefers x to z.)
3. A seeks to maximize his or her utility where the utility of an option x is
greater than the utility of an option y for A if and only if A prefers x to
y. The utilities of options are equal just in case the agent is indifferent
between them.
4. If option x is acquiring commodity bundle x and option y is acquir-
ing commodity bundle y and y contains at least as much of each
commodity as x and more of at least one commodity, then all agents
prefer y to x.
5. The marginal utility of a commodity c to an agent A is a decreasing
function of the quantity of c that A has.
The foregoing postulates de¬ne the rationality of actors. The following
de¬ne the assumptions of production.
6. When we increase inputs into production, other things being equal,
output increases, but, after a certain point, at a decreasing rate.
7. Increasing all inputs into production in the same proportion increases
output by that proportion. The production set is weakly convex and
8. Firms attempt to maximize their pro¬ts by minimizing costs relative
to revenues.

Appendix D 187

As Hausman points out, this brief summary is both rough, misleading
and incomplete: rough because the theory can be stated more precisely,
misleading because neo-classical economists do not always make use of
all these postulates, and incomplete for two reasons. First, as part of
the background, commonly used mathematical techniques are assumed.
Indeed, the full power of the theory is the fact that it can be repre-
sented mathematically. This is a feature of what is sometimes attributed
to “physics envy,” where the understanding of physics is in terms of posi-
tivist theory of science. Second, and more obviously, a host of strong, but
narrower assumptions must be made. On the competitive market model,
one must assume that there are many buyers and sellers in every market
and that each may enter and leave easily. These are relaxed in conditions
of “imperfect competition.” This includes monopoly, the limiting case of
“imperfect competition,” and oligopoly, where several large ¬rms dom-
inate the market, there are barriers to entry and the suppliers produce
relatively similar products. General equilibrium theory assumes further
that everyone has all the relevant information, that there is an interde-
pendence among the many markets, and that commodities (including
labor) are in¬nitely divisible. This last is critical insofar as calculus is the
indispensable tool for writing and solving simultaneous equations which
de¬ne equilibrium. But, for example, if labor does not have a marginal
product, but only an average or step-wise product, the marginal produc-
tivity of labor is a mathematical ¬ction, useful for calculation but lacking
empirical reality.
In exemplary “deductivist” fashion, many propositions may be strictly
deduced from the foregoing, for example, that the price will rise when
demand exceeds supply. Let us sketch how this works.
As noted, propositions 1“5 de¬ne rationality. We then construct “indif-
ference curves” which represent preferences of actors between bundles of
two items, for example, apples and bananas. In a two-person exchange,
equilibrium is reached (the invisible hand) when both persons are ratio-
nal and pursue a minimax strategy. If, for example, we measure apples on
the vertical line and bananas on the horizontal line (¬gure 1) there is an
in¬nite series of bundles of apples and bananas to which O is indifferent.
Every bundle further from point O is preferred (curve O2 is preferred to
O1 ). The curves are convex because we assume that apples and bananas
have diminishing marginal utility. That is, after some amount “ one can-
not eat that many bananas “ there is a decreasing utility for each additional
consumption of the good. The same as regards X, drawing his indiffer-
ence map for this bundle convex to point x. If they are rational they will
not stop trading at either P or S since one or the other will see that the
188 A Realist Philosophy of Social Science



Figure 1.

exchange is not to his advantage. They reach equilibrium at T: there is
no way for either to get more without the other getting less, the de¬nition
of a “Pareto optimum.”1
If we introduce money as “numeraire,” we represent the price of (say)
apples on the vertical axis (¬gure 2). This is O™s demand curve for apples,
what she will buy at what price. It slopes down since buyers are rational.
Price represents opportunity cost, what other goods must be sacri¬ced.
Given some ¬xed amount of income (a budget), O™s problem is to allocate
her money so as to maximize utility (in accordance with the postulates
of rationality). We can represent a supply curve for X, what she will
supply at what price. Since she is also rational (and in a condition of pure
competition), the supply curve slopes up. The supply curve slopes up as
a condition of diminishing marginal productivity.
The concept of elasticity is pertinent here. A curve is elastic when the
quantity demanded (or supplied) is very responsive to price. Oxenfeld
offers that the contribution of the economist™s idea of elasticity of demand
“may be far below zero.” First, although theory acknowledges that this

1 Indifference curve analysis is owed to the economist-cum-sociologist Wilfredo Pareto. As
Dyke shows, in order to construct an indifference curve, one must assume that choices
are made pairwise and that this assumes the independence of irrelevant alternatives, that
preferences can be arranged ordinally. Thus, choices depend only on comparing one
bundle with another. The point is that we are rarely in such conditions. For example, the
choice between two breakfast cereals may be a “pure preference” but the choice between
a needed medication and a breakfast cereal generally is not. As Dyke points out, there are
many reasons for choosing one thing over another, including availability, future needs,
budgetary constraints, etc. Dyke™s account (1981: 114“116) is excellent.
Appendix D 189





0 1 2 3 4 5 Quantity
Figure 2.

elasticity is not uniform over the relevant stretches of the demand curve
(where it is not either vertical or horizontal) simplifying assumptions are
always made. This is a source of considerable mischief. Second, busi-
nessmen use an alternative and far simpler method: “the effect of price
reductions on unit sales can be stated as the number of added units that
would be sold as a result of given dollar ˜expenditures™ in the form of
price reductions”(Oxenfeld, 1963: 73).

Imperfect competition
But worse, in conditions of imperfect competition, not only is elastic-
ity not uniform across the relevant stretches of the supply and demand
curves, but in “the short run” the curves will be very inelastic: for some
goods, electricity, for example, there is a public monopoly and con-
sumers have no choice but to purchase the service. Similarly, produc-
ers infrequently increase output in response to higher demand in the
short run: they take windfall pro¬ts instead.2 But the point here is not
whether monopolies or oligopolies are inef¬cient as de¬ned by the theory
(below); only that in imperfect competition, the competitive equilibrium
model does not hold and that, accordingly, price is not the intersection
of the sloping demand and supply curves as these are constructed by the

2 The “long run” is a wonderful escape for theory, but as Keynes observed: “in the long
run, we are all dead.”
190 A Realist Philosophy of Social Science

As is generally acknowledged, there is no adequate economic theory of
oligopolistic competition which explains how prices and output are set,
except to say that corporations do not engage in what would be mani-
festly mutually destructive price-competition and that they remain in a
competitive environment in which they must constantly employ a host
of forms of non-price competition (Galbraith, 1968; Baran and Sweezy,
1968). On the other hand, since for the mainstream, imperfect compe-
tition continues to be considered “the special case,” the absence of a
rigorous theory is not seen to be a disaster.3 But of course, the behavior
of corporations which are not engaged in price competition is of con-
siderable importance to understanding contemporary market capitalism,
including “consumerism,” global inequalities and continuing problems
with the stability of the system.
In a purely competitive market, mainstream theory tells us that there
would be no payoff for advertising which provided more than information
since ¬rms responding to market demand would produce the same prod-
ucts and sell all their output at the equilibrium price. Similarly, on the
mainstream view, all costs “ including accordingly, the enormous costs of
the sales effort “ are “necessary.” But this provokes the question, raised
by Thorstein Veblen, of the rationality of the system: He wrote:
The producers have been giving continually more attention to the saleability of
the product, so that much of what appears on the books as production-cost should
properly be charged to the production of saleable appearances. The distinctions
between the workmanship and salesmanship have been blurred in this way, until it
will doubtless hold true now that the shop-cost of many articles produced for the
market is mainly chargeable to the production of saleable appearances, ordinarily

3 Schumpeter remarks that “from all the in¬nite variety of market patterns pure or perfect
monopoly and pure or perfect competition stand out by virtue of certain properties “ of
which the most important is that both cases lend themselves to treatment by means of rel-
atively simple and (in general) uniquely determined rational schemata “ and on the other
hand, that the large majority of cases that occur in practice are nothing but mixtures and
hybrids of these two, then it seems natural to accept pure monopoly and pure competition
as the two genuine or fundamental patterns and to proceed by investigation how these
work out” (1954: 975). Unfortunately, reality was sacri¬ced for rigor. See Schumpeter,
1954: 962“985, 1150“1152. See also Brakman and Heijdra (eds.), 2001. The editors
note that there were two “revolutions” in the theory of monopolistic competition, the
¬rst initiated by Robinson (1969) and Chamberlin (1962) in the 1930s and the second
by Dixit and Stiglitz in 1977. The ¬rst failed, according to them, because it lacked an
adequate model. The second attempt “introduced a formalization that had all the char-
acteristics of monopolistic competition but was easier to handle” (Brakman and Heijdra,
2001: 1“2). But they note also that there are serious problems with this formalization,
even if, for them, it shows promise.
4 Veblen, Absentee Ownership and Business Enterprise in Recent Times, p. 300, quoted by
Baran and Sweezy (1968: 133). The account of Baran and Sweezy, which also draws on
Chamberlin (1962) and Kalecki (1965), remains very useful.
Appendix D 191

Veblen, writing in 1923, saw only the beginning of this “blurring” which,
with the astronomical growth of advertising in the recent past, makes the
production unit the handmaiden of the marketing division. Baran and
Sweezy quote a sympathetic defender of this dramatic change:
In fact, broadly de¬ned as it properly can be, to include the whole range of
marketing operations from product design through pricing and advertising right
on to doorbell pushing and the ¬nal sale, selling or marketing is not only a symbol
of a free society, but is in ever-increasing measure a working necessity in our
particularly free society.5

To be sure, consumerism is a measure of “our particularly free society,”
and is “a working necessity,” but an explanation was already in Marx,
hinted at by Keynes and is the central theme of Baran and Sweezy. Com-
mon sense tells us that commodities produced must be sold if the system
is to reproduce itself, that unless consumption is somehow guaranteed,
the system falls into crises. The problem, as noted by Henry Ford, was
that it is not only necessary that wage earners want to buy what is being
produced, but that they have to have suf¬cient income to do this. For
Marxists, of course, this was a problem for which capitalism could pro-
vide no permanent solution (Harvey, 1987).

Micro- and macro-theory
This is a good place to include Keynes™s contribution. Despite what may
be an excusable misunderstanding of Keynes, he was no radical in his
theorizing, conserving most of the elements of the legacy derived from
Marshall and Pigou. Indeed, despite its pertinence to his main concerns,
his work shows no evidence of in¬‚uence from attacks in the early 1930s
on traditional price theory by Joan Robinson and E. H. Chamberlin.
“As Robert Lukas repeatedly pointed out even in the early 1970s, grad-
uate students were taught one thing during their Monday / Wednesday
microeconomic courses and another thing on Tuesdays and Thursdays in
the macro-economic courses” (Boettke, 1997: 36).6 Samuelson (1947)
was here the key ¬gure in making the system seem to work. But his syn-
thesis of the neo-classical tradition and Keynes was, contestably, coher-
ent. As Boettke writes: “Samuelson™s synthesis created a rather strange
mix of general equilibrium economics with Keynesian macroeconomics”

5 Dexter M. Keezer and Associates, New Forces in American Business, p. 90, quoted by Baran
and Sweezy (1968: 124). By contrast, one thinks of Vance Packard™s important The Waste
Makers (1960).
6 Rational choice theorists tend to take a more optimistic view and hold that in economics,
the micro“macro gap has been closed.
192 A Realist Philosophy of Social Science

(1997: 36). Indeed, for Baran and Sweezy: “the effects of a thoroughgoing
reintegration of the two levels of analysis “ the substitution of a monopo-
listic price system for the traditional competitive system, and the analysis
of its implications for the whole economy “ are nothing short of devastat-
ing to capitalism™s claims to be considered a rational social order which
serves to promote the welfare and happiness of its members” (1968: 56).
The link between the classical theory and Keynes™s view was in the labor
market. Presumably, on the standard view, where commodity and labor
markets are perfectly competitive, a wage reduction expands employ-
ment which expands consumption. There would always be a wage rate,
no matter how low, which produced full employment. The most criti-
cal departure from the neo-classical tradition was Keynes™s demolition of
Say™s law, the idea that in competitive markets general glut or unemploy-
ment could not occur because supply creates its own demand.7 But, as
Boettke points out: “if the labor market was in competitive equilibrium,
this implied that the full-employment output level had been achieved,
i.e., there was no macroeconomic problem” (1997: 36). This is Keynes™s
summary of his rejection of this idea:
When employment increases, aggregate real income is increased. The psychology
of the community is such that when aggregate real income is increased aggregate
consumption is increased, but not so much as income. Hence employers would
make a loss if the whole of the increased employment were to be devoted to sat-
isfying the increased demand for immediate consumption. Thus, to satisfy any
amount of employment there must be an amount of current investment suf¬cient
to absorb the excess of total output over what the community chooses to consume
when employment is at the given level. For unless there is this amount of invest-
ment, the receipts of entrepreneurs will be less than is required to induce them
to offer the given amount of employment. It follows, therefore, that, given what
we shall call the communities™ propensity to consume, the equilibrium level of
employment . . . will depend on the amount of current investment. The amount
of current investment, in turn, will depend on what we shall call the inducement
to invest; and the inducement to invest will be found to depend on the relation
between the schedule of the marginal ef¬ciency of capital and the complex rates
of interest on loans of various maturities and risks . . . There is no reason in
general for expecting [the equilibrium level of employment] to be equal to full
employment. The effective demand associated with full employment is a special
case, only realized when the propensity to consume and the inducement to invest

7 The “law,” named for Jean Baptiste Say (1803) was a lynchpin of both classical and neo-
classical theory. In Ricardo™s crisp formulation: “No man [sic] produces but with view to
consume or sell, and he never sells but with an intention to purchase some other com-
modity . . . By purchasing them, he necessarily becomes either the consumer of his own
goods, or the purchaser and the consumer of the goods of some other person . . . Produc-
tions are always bought by productions, or by services” (quoted from Robert Lekachman,
1964). This is a most useful introductory volume.
Appendix D 193

stand in a particular relationship to one another . . . But it can only exist when,
by accident or design, current investment provides an amount of demand just
equal to the excess of the aggregate supply price of the output resulting from
full employment over what the community will choose to spend on consumption
when it is fully employed. (Keynes, 1960: 27)

The “propensity to consume” is, of course, profoundly in¬‚uenced by
what Keynes terms “subjective factors,” or “those psychological char-
acteristics of human nature and those social practices and institutions
which, though not alterable, are unlikely to undergo a material change
over a short period of time except in abnormal or revolutionary circum-
stances” (Keynes, 1960: 91). Of course, the existence of “subjective
factors” leaves enormous room for some concrete empirical work. But
Keynes paid no attention, for example, to how advertising and the manip-
ulation of wants affect these “subjective” factors. In his book, Keynes,
not untypically, takes the subjective factors as “given,” and assumes that
“the propensity to consume depends only on changes in the ˜objective fac-
tors™,” on, that is, the “variables” de¬ned by neo-classical theory. But to
his credit Keynes was not, in contrast to Samuelson, a formalist who was
committed to mathematical economics. Keynes wanted models, but for
him, building them required “a vigilant observation of the actual working
of our system.” Indeed, “to convert a model into a quantitative formula
is to destroy its usefulness as an instrument of thought” (Keynes, 1984).
That conclusion can be strongly endorsed!
Keynesian state policies follow: governments can affect total spending
and total employment either by monetary policies, which lower inter-
est rates (for example, increasing the supply of money) or ¬scal policies
which expand total spending by increasing public spending, without rais-
ing taxes, or decreasing taxes without reducing public spending. This lim-
ited intrusion of government did not in any way threaten capitalist markets
as these were theorized by mainstream theory. Similarly as regards direct
“public spending” “ as long as it was in the interest of national defense “
so-called “warfare capitalism” (in contrast to “welfare capitalism”). But
the consensus which formed on this view of the matter did not last.
While the point cannot be pursued here, Samuelson™s neo-Keynesian
model managed to “square the circle” “by means of ˜wage stickness™ and
˜the money illusion™ ” (Boettke, 1997: 37). But here again the assump-
tions were not only ad hoc, but implausible, leaving the model vulnerable
to the Chicago School™s “hyperformalist attempt to purify the synthesis
by purging it of its Keynesian contaminants” (Boettke, 1997: 38). The
Neo-liberalism of the recent past was the consequence. Boettke summa-
rizes matters well:
194 A Realist Philosophy of Social Science

Samuelson™s reconciliation of the micro-economic ideal type with involuntary
unemployment was repudiated, along with Keynesian prescriptions, in favor of a
view that there could be no involuntary unemployment, hence that government
action was unnecessary. The result was a doctrinaire derivation of the laissez-faire
conclusions that had been overturned by the formalist revolution; economics was
now cleansed of Keynesian impurities that had been introduced in the interest of
realism. (1997: 38)8

Market ef¬ciency according to the model
But to return, supply and demand curves of commodities are simply
aggregations of the curves of an in¬nite number of individuals (as either
producers or consumers). The equilibrium price is the price where the
two curves intersect (see ¬gure 3). At this price, the market will “clear”:
everything brought to the market will be purchased.9



Figure 3.

8 Rational expectations theory also played a role. Human capital theory (invented by
Gary Becker) was the effort to place labor back within the price-auction framework.
See Thurow, 1983: chapter 7.
9 Thurow (1983) argues that the question is not whether or not markets clear or even
whether they are competitive. He insists that the real question is whether markets clear
based on ¬‚uctuations in prices (1983: 9). Similarly, as Hicks argues, we can assume
competition, but once we take seriously the idea that inventory plays a critical role in
pricing, we can see that:
The traditional view that market price is, at least in some way, determined by an equation
of supply and demand [has] now to be given up. If demand and supply are [no longer
to be] interpreted, as had formerly seemed to be suf¬cient, as ¬‚ow demand and supplies
coming from outsiders, it is no longer true that there is any tendency, over any particular
Appendix D 195

Equilibrium is, by de¬nition, a condition of ef¬ciency (“getting the
price right”). Ef¬ciency is Pareto optimality. “A distribution of goods or a
scheme of production is inef¬cient when there are ways of doing still better
for some individuals without doing any worse for others” (Rawls, 1971:
67). We can separate the two components here. Production is ef¬cient
when there is no way to produce more of some commodity without pro-
ducing less of another. Since, however, there is “consumer sovereignty,”
where markets are competitive, consumers have decided on the bundles
to be produced. But since we are at equilibrium, the distribution will also
be ef¬cient: there will be no distribution which improves the circumstance
of at least one with someone™s situation being worsened.
We need to see ¬rst there are many ef¬cient con¬gurations. Assume a
¬xed stock to be distributed between x and y.





Figure 4.

By de¬nition for all the points in the convex set within AOB, excepting
those on line AB, it is possible to improve either x or y (or both) without
worsening their opposite with the change (¬gure 4). At C, for example
y™s condition has been improved without change to x. At D, both x and
y are better off than they were at F. D is ef¬cient but so is G. Indeed,
all the points on line AB are ef¬cient since for any point on that line

period, for them to be equalized; a difference between them, if it were not too large, could
be matched by a change in stocks. It is of course true that if no distinction is made between
demand from stockholders and demand from outside the market; demand and supply in
that inclusive sense must always be equal. But that equation is vacuous. It cannot be used
to determine price, in Walras™s or Marshall™s manner. (Hicks, 1989: 11)
196 A Realist Philosophy of Social Science

it is impossible to alter the distribution so as to make some persons (at
least one) better off without at the same time causing the other™s situ-
ation to worsen. If (say) x has it all (point A), then he must lose if y
were to get any. Indeed, “the willingness to trade shows that there is a
re-arrangement which improves the situation of some without hurting
anyone else” (Rawls, 1971: 70). A is ef¬cient, since where x gets it all, y
has nothing to trade!10
A powerful motivation for policy choice in modern political economy
is the idea that markets are ef¬cient. However, it is not dif¬cult to show
that even if the theoretical conditions are met, it is highly dubious that
we would have ef¬ciency. Many different lines of argument are available.
One line of argument regards externalities. Roughly, externalities are
side-effects, spillover costs or bene¬ts for third parties. Polluting smoke
from a steel mill is a negative externality. Plainly, with externalities there
will be a misallocation of resources. Thus the (real) costs of producing
steel are not included in the supply schedule and thus, social utility is not
optimized. We can introduce a point made earlier about markets. We need
to know the type of property relation to determine if some “externality,”
for example, air pollution, is Pareto-relevant. This is nowhere a given: it
is always decided and raises the question of who decides?11
Second, there are various forms of micro-rationality which lead to
macro-irrationality. The Prisoner™s Dilemma, for example, shows that
two rational economic actors (actors who satisfy 1“5) will not end up
with the best result. One needs to show, accordingly, that such situations
never or rarely arise, for example, that in¬‚ationary pressure is not best

10 Thus, Samuelson: “The Invisible Hand will only maximize total social utility provided the
state intervenes so as to make the initial distribution of dollar votes ethically proper” (quoted
by Lachman, 1984: 310). Followers of Hayek are on the right track when they write:
“There is, of course, no such thing as an ˜initial distribution™ before the market process
starts [as is assumed by mainstream theorizing]. The distribution of wealth in terms
of asset values at any point of time is the cumulative result of the market process of
the past” (Lachman, 1984: 310). I say “on the right track” here because Lachman (like
Hayek) assumes, with Samuelson, that the market process is always a competitive market
process. It is for this reason, accordingly, that the Invisible Hand maximizes total social
utility only if the State does not seek to make “the initial distribution of dollar votes
ethically proper” (Lachman, 1984: 310).
11 Similarly as regards transaction costs. Boettke argues that “Coase™s project . . . has
been largely misunderstood by formalist neoclassical economics. Instead of highlighting
the functional signi¬cance of real-world institutions in a world of positive transactions
costs, Coase™s work has been interpreted as describing the welfare implication of a zero-
transactions-cost world” (1997: 21). On the contrary, it should lead to a consideration of
how the constitution of markets through the use of law may address the many problems
created by existing markets. As Boettke notes, Coase™s work must be contrasted with the
work of Posner (e.g., 2004) who is engaged in the opposite task of applying neo-classical
theory to issues in law.
Appendix D 197

explained in terms of rational strategies pursued by workers and con-
Similarly, if one assumes neo-classical theory, then in a perfectly com-
petitive market, “although all the ¬rms have a common interest in a higher
price for the industry™s product, it is in the interest of each ¬rm that the
other ¬rms pay the cost “ in terms of the necessary reduction in output “
needed to obtain a higher price” (Olsen, 1971: 9). Indeed, “the only thing
that keeps prices from falling in accordance with the process . . . is out-
side intervention. Government price supports, tariffs, cartel agreement,
and the like may keep the ¬rms in a competitive market from acting con-
trary to their interest” (Olson 1971: 10).12 But if, for example, lobbying
efforts are thought to be necessary to get the government™s help, an iden-
tical problem arises: just as it is not rational for a producer to restrict his
output, it is not rational for him to assume any of the costs of hiring the
lobbyist “ the so-called “free rider problem.”
Finally, there is what Hirsch (1976) has called the “adding up” prob-
lem. For “positional goods,” for example, driving your private car to
work, or high-rise development to attract buyers who want an ocean view,
“opportunities for economic advance, as they present themselves serially
to one person after another, do not constitute equivalent opportunities
for economic advance by all. What each one can achieve, all cannot”
(Hirsch, 1976: 4). There is an immediate advantage to standing on tip-
toe to get a better view, but if everyone does this, everyone is worse off. At
some point, as new drivers get on the freeway, the initial gains are lost in
traf¬c jams. More generally, for a host of “positional goods,” since “the
standard concept of economic output is appropriate only for truly pri-
vate goods, having no element of interdependence between consumption
by different individuals” (Hirsh, 1976: 7), there is no way to translate
individual improvement to overall improvement. Hence, individual max-
imizing behavior in these contexts is counter-productive. For each of us
the scramble is rational since individually we never confront “the distinc-
tion between what is available as a result of getting ahead of others and
what is available from a general advance shared by all” (Hirsch, 1976:
10). More than a distributional issue is here involved. Wider participa-
tion affects not only what one gets from winning the game, but the nature
of the game itself. If the goal of a bachelor™s degree is a better job and

12 Of course, this all assumes that corporations are engaged in price competition. But
we should not assume this. Indeed, not only are cartels and non-price competition
rational, but in many industries, to prevent either extra-large pro¬ts (by a monopolist)
or extremely low pro¬ts, for example, in agriculture, government will gladly step in
with regulatory commissions, price supports, acreage controls, etc. to assure that these
“deviant” industries maintain healthy pro¬t ratios. See Baran and Sweezy, 1968: 64“66.
198 A Realist Philosophy of Social Science

better income, then while getting a bachelor™s degree remains rational,
the consequence is the diminishing value of the degree. At some point,
in terms of costs and gains in income, everyone is worse off. We would
like to believe, of course, that the effort was not wholly instrumental and
that the educational experience bought more than a credential (Thomas,
More generally, Adam Smith and the tradition which followed him
was wrong. As Hirsch concluded: “Competition among isolated indi-
viduals in the free market entails hidden costs for others and ultimately
for themselves. These costs are deadweight cost for all and all involve
social waste” (Hirsch, 1976: 5). And indeed, as with all of the foregoing
problems, collective action provides the only solution.
A third line of argument is Arrow™s paradox. On the neo-classical view,
we have individual preferences which via the mechanism of the market
result in a social preference. Arrow established a series of conditions,
“social choice functions,” which restrict ways that the social preference
could be derived from individual preferences. All are fairly obvious: ratio-
nality, the idea that as more people prefer some alternative, then this
alternative ought not to lose ground as a social preference; that “irrel-
evant alternatives” must be independent (above); and there is “non-
dictatorship,” “ no individual™s preference automatically de¬nes social
preference. The independence condition says that “the social choice
made from any environment depends only on the orderings of individ-
uals with respect to alternatives in that environment.” (This precludes
substitutes and duplicates and is assumed in postulate 1, above.) Arrow
demonstrated that “no social choice function ful¬lling these conditions
could guarantee satisfactory results when there were more than two indi-
viduals in the society and more than two alternatives to choose from”
(Dyke, 1981: 113). This has implications for democratic theory; but it
shows, I think, now uncontestably, that a fundamental error of equilib-
rium theory is the assumption that market behavior consists of simply
pairwise choices between bundles: the assumption of the independence
of irrelevant alternatives.13
Finally, neo-classical theory assumes that ef¬ciency can be de¬ned in
terms of exchange values. But surely this is not a reasonable notion. An
economy could produce “ef¬ciently” (as de¬ned above) and wastefully
and destructively. Destructive but ef¬cient production violates the envi-
ronment, perhaps making it un¬t for human life. Wasteful but ef¬cient
production generates commodities which fail to serve human needs and

13 Conversely, there is the question of whether there ever is “market failure.” See Pitelis,
Appendix D 199

wants, or fails to do so as well as it might. Star Wars technology is a good
example of the former; poor quality housing an example of the latter.
But, of course, since, contrary to neo-classical theory, consumers are not
sovereign, this is to be expected.
But as argued in this appendix and in chapter 6, mathematical eco-
nomics is hardly the best choice as model of a successful social science.
Paradoxically, the reason for this is exactly what has made it so attractive:
it seems to be more like physics than any other social science. This volume
has argued, however, that the actual practices of the successful natural
sciences are poorly described by empiricist philosophy of science and that
the individuals who constitute the social mechanisms which would give
us an understanding are badly mis-theorized by mainstream theory.

Abrams, Philip (1983) Historical Sociology. Ithaca: Cornell University Press.
Achen, Christopher (1982) Interpreting and Using Regression. Beverly Hills: Sage.
Achinstein, Peter (1981) “Can There be a Model of Explanation?” Theory and
Decision 13. Reprinted in David-Hillel Ruben, Explanation, Oxford: Oxford
University Press.
Alexander, Jeffrey (1987) “The Importance of the Classics,” in Anthony Giddens
and Jonathan Turner (eds.), Social Theory Today. Oxford: Polity Press.
(1998) Neofunctionalism and After. Oxford: Basil Blackwell.
Anderson, Perry (1974) Lineages of the Absolutist State. London: New Left Books.
Anderson, P. W., Arrow, Kenneth J. and Pines, David (eds.) (1988) The Economy
as an Evolving Complex System. Redwood City: Addison-Wesley.
Appadurai, Arjun (1996) Modernity at Large: Cultural Dimensions of Globalization.
Minneapolis: University of Minnesota Press.
Archer, Margaret (1995) Realist Social Theory: The Morphogenetic Approach.
Cambridge: Cambridge University Press.
Aronson, Gerald (1984) A Realist Philosophy of Science. New York: St. Martin™s
Ashley, Richard K. (1989) “Living on Border Lines: Man, Poststructuralism,
and War,” in James Der Derian and Michael Shapiro (eds.), International /
Intertextual Relations. Lexington, MA: Heath.
Atkins, Peter (2003) Galileo™s Finger. New York: Oxford University Press.
Augier, Mie (1999) “Some Notes on Alfred Schutz and the Austrian School of
Economics: Review of Alfred Schutz™s Collected Papers, Vol. IV. Edited by
H. Wagner, G. Psatha and F. Kersten,” Review of Austrian Economics 11.
Bakan, Joel (2004) The Corporation: The Pathological Pursuit of Power. New York:
Free Press.
Balough, Thomas (1982) The Irrelevance of Conventional Economics. New York:
Baran, Paul A. and Sweezy, Paul M. (1968) Monopoly Capital: An Essay on the
American Economic and Social Order. New York: Monthly Review Press.
Barnes, S. B. (1977) Interests and the Growth of Knowledge. London: Routledge &
Kegan Paul.
Barnet, Richard J. and Cavanaugh, John (1994) Global Dreams: Imperial Corpo-
rations and the New World Order. New York: Simon and Schuster.
Barnet, Richard J. and Muller, Ronald E. (1974). Global Reach: The Power of the
Multinational Corporations. New York: Simon and Schuster.

References 201

Barone, Enrico (1908) “The Ministry of Production in a Collectivist State,” in
F. A. Hayek (ed.), Collectivist Economic Planning, London: Routledge &
Kegan Paul.
Bauman, Zigmunt (1998) Globalization: The Human Consequences. New York:
Columbia University Press.
Bendix, Reinhard (1976) “The Mandate to Rule: An Introduction,” Social Forces
(1978) Kings or People. Berkeley: University of California Press.
(1990) Force, Fate, and Freedom: On Historical Sociology. Berkeley: University
of California Press.
Berger, Peter L. and Luckmann, Thomas (1967) The Social Construction of Reality.
New York: Anchor Books.
Berle, Adolf A. and Means, Gardiner C. (1968) The Modern Corporation and
Private Property, revised edition. New York: Harcourt, Brace and World.
Bhaskar, Roy (1975) A Realist Philosophy of Science. 2nd edition, 1978. Atlantic
Highlands: Humanities Press.
(1978) “On the Possibility of Social Scienti¬c Knowledge and the Limits of
Naturalism,” Journal for the Theory of Social Behavior 8, 1.
(1979) The Possibility of Naturalism. Atlantic Highlands: Humanities Press.
(1986) Scienti¬c Realism and Human Emancipation. London: Verso.
Bickerton, Derek (1990) Language and Species. Chicago: University of Chicago
Biggart, Nicole Wesley (2002) Readings in Economic Sociology. Oxford: Blackwell.
Blau, Peter (1964) Exchange and Power in Social Life. New York: John Wiley.
Bloor, David (1976) Knowledge and Social Imagery. London: Routledge & Kegan
Boettke, Peter J. (1997) “Where Did Economics Go Wrong? Modern Economics
as Flight from Reality,” Critical Review 12.
Bohm, David (1984) Causality and Chance in Modern Physics. New edition.
London: Routledge & Kegan Paul.
Bonham, Vence L., Warshauer-Baker, Esther and Collins, Frank S. (2005) “Race
and Ethnicity in the Genome Era: The Complexity of the Constructs,” Amer-
ican Psychologist 60, 1.
Boudon, Raymond (1998a) “Social Mechanisms Without Black Boxes,” in Peter
Hedstrom and Richard Swedberg (eds.), Social Mechanisms: An Analytical
Approach to Social Theory. Cambridge: Cambridge University Press.
(1998b) “Limitations of Rational Choice Theory,” American Journal of
Sociology 104, 3.
Bourdieu, Pierre and Wacquant, Lo¨c, J. D. (1992) An Invitation to Re¬‚exive
Sociology. Chicago: University of Chicago Press.
Brakman, Steven and Heijdra, Ben J. (eds.) (2001) The Monopolistic Competition
Revolution in Retrospect. Cambridge: Cambridge University Press.
Brenner, Robert (1977) “The Origins of Capitalist Development: A Critique of
Neo-Smithian Marxism,” New Left Review 104.
(2003) The Boom and the Bubble: The US in the World Economy. London: Verso.
Brodbeck, May (ed.) (1968) Readings in the Philosophy of the Social Sciences. New
York: Macmillan.
202 References

Brown, Harold I. (1977) Perception, Theory and Commitment: The New Philosophy
of Science. Chicago: University of Chicago Press.
Bourgois, Phillipe (1997) “In Search of Horatio Alger,” in Craig Reinerman and
Harry G. Levine (eds.), Crack in America: Demon Drugs and Social Justice.
Berkeley: University of California Press.
Bunge, Mario (1979) Causality and Modern Science. New York: Dover.
(1996) Finding Philosophy in Social Science. New Haven: Yale University
(2003) “How Does it Work? The Search for Explanatory Mechanisms,” Phi-
losophy of the Social Sciences 34, 2.
Burawoy, Michael (1989) “Two Methods of Search in Science: Skocpol versus
Trotsky,” Theory and Society 18.
(1998) “The Extended Case Method,” Sociological Theory 16, 1.
Burian, Richard (1989) “The In¬‚uence of the Evolutionary Paradigm,” in Max
Hecht (ed.), Evolutionary Biology at the Crossroads. Flushing, New York:
Queens College Press.
Calhoun, Craig (1996) “The Rise and Domestication of Historical Sociology,”
in Terrence McDonald (ed.), The Historic Turn in the Human Sciences. Ann
Arbor: University of Michigan Press.
(1998) “Explanation in Historical Sociology: Narrative, General Theory, and
Historically Speci¬c Theory,” American Journal of Sociology 104, 3.
Calhoun, Craig, Gerteis, Joseph, Moody, James, Pfaff, Steven and Virk, Inder-
mohan (eds.) (2002) Contemporary Social Theory. Oxford: Basil Blackwell.
Camic, Charles (1987) “The Making of a Method: A Historical Reinterpretation
of the Early Parsons,” American Sociological Review 52.
Campbell, D. A. and Misanin, J. R. (1969) “Basic Drives,” in P. H. Mussen and
M. R. Rosenzweig (eds.), Annual Review of Psychology, vol. 20.
Cartwright, Nancy (1989) Nature™s Capacities and Their Measurement. Oxford:
Oxford University Press.
Cavalli-Sforza, Luigi Luca (2000) Genes, Peoples and Languages. New York: North
Point Press.
Cavalli-Sforza, Luigi Luca and Cavalli-Sforza, Francesco (1995) The Great
Human Diasporas. Cambridge, MA: Perseus Books.
Chai, Sun-Ki (2001) Choosing an Identity: A General Model of Preference and Belief
Formation. Ann Arbor: University of Michigan Press.
Chakrabarty, Dipesh (1989) Rethinking Working Class History: Bengal 1890“1940.
Princeton: Princeton University Press.
Chamberlin, E. H. (1962) The Theory of Monopolistic Competition: A Re-


. 6
( 7)