Neither Modernist Nor Postmodemist A Third Way

Mara Beller
Abstract. In this paper I undertake an analysis of the heritage of Kuhn and Feyera-

bend as compared with the main tenets of the logical positivism, and identify the
components of logical positivism that directly lead to relativism. I argue that the notion of consensus creates major problems in historiography and philosophy of science, preventing a description of scientific change. I further argue that the concept of
creative disagreement should be introduced into studies of science not only as a historical actuality, but also as a basic epistemological and methodological presupposition. I trace the grip of the notion of consensus in social studies of science to Durkheim's heritage, focusing on the representatives of the Strong Program in sociology
of science. I also argue that Thomas Kuhn inherited the same Durkheimian view of
society through Ludwik Fleck. Finally, I briefly outline a dialogical alternative to the
current historiography-an alternative in which the notion of disagreement plays a
fundamental epistemological role.
Let us begin with a fairy tale. Once upon a time there was a Glorious Reign
of Reason on Earth. The Golden Age lasted until as recendy as thirty to forty
years ago, before the "enemies" of reason invaded academia and dethroned
the Goddess of Truth. "Reason today is not what it used to be"-this nostalgic outcry underlies the statements of the defenders of reason and truth.
The collection of essays, The Flight from Science and Reason (Gross et al. 1996)
offers many examples:
Susan Haack (1996) opened her contribution by describing the intellectual
integrity and love of truth, which characterized philosophical inquiry until
such thinkers, as the noted philosopher Richard Rorty, joined the bandwagon
of antireason. "Until the 70's"-lamented the sociologist Stephen Cole, "the
good Mertonian Sociology of science shielded the objective context of science from sociological analysis, till voodoo Sociologists or in other words,
social constructivists, seized control of the field (1996). According to Mario
Bunge in the same volume, over the past three decades enemies of learning
and rigor inftltrated the hallow walls of academia. Up until the mid 60's such
traitors of reasons were prompdy ostracized. And Alan Sokal disclosed that
the aim of his hoax (1996) was to protest the recent betrayal by academics
and humanists of the worthy enlightened heritage of rationality and truth.
M. Carrier et al. (eds.), Knowledge and the World: Challenges Beyond the Science Wars
Springer-Verlag Berlin Heidelberg 2004

266

Mara Beller

In this paper I offer a different perspective on the heritage of 20th century

physics and philosophy. Contrary to the claim that reason is not what it used
to be, I will argue that reason never was what it appears to have been. Thus,
the assumed divide between the defenders of reason who wish to resurrect
the ideas of modernity, and post-modernists who supposedly have dispensed
with these ideas, is much less substantial than is generally perceived.
This paper is divided into three parts.
In the first part I connect the heritage of Kuhn and Feyerabend with logical
positivism, and identify the component of logical positivism that directly
leads to relativism. In the second part I analyze the notion of scientific community, used in the social studies of science, and will uncover its Enlightenment roots. Both defenders of reason and its enemies are trapped, I will
claim, in the same conceptual predicament. In the third part I outline what I
perceive as an alternative that is neither modernist nor postmodernist-it is a
"third way."
Let us begin with the early logical positivism. Logical positivism in its early
incarnation was aimed to eliminate the metaphysical (identified with nonsensical) from philosophical discussion. It was used by Philipp Frank, a leading
logical positivist, as a weapon against the irrational "metaphysical" trend of
the 30s, (50 years before similar postmodernist trend of the 80s), which connected acausality of quantum physics with religion, mysticism and politics.
According to Frank, the pre-World War II logical positivism recognized only
two sets of legitimate terms-theoretical symbols and observational entities.
The observable facts are deduced from theoretical principles of the theory by
a long chain of mathematical deductions and operational definitions. It is the
urge to take an illegitimate shortcut, to connect the general principles with
common-sense terms directly, which introduces the "irrational" and "spontaneous" into quantum physics, and connects physics with religion and mysticism (Beller and Fine 1994, pp. 19-20).
After the war, in a different setting, we encounter a new face of logical
positivism. Frank, now a Harvard professor and a founder of the Unity of
Science movement, spelled out the necessary connection between science and
values, between epistemology and ideology. The previous dyadic scheme of
theoretical symbols and observed phenomena was supplemented, following
Peirce, by a third, pragmatic dimension which comprises the relations between the scientist and the symbols that he (or she) invents. This ultimately

Neither Modernist Nor Postmoderrust-A Third Way

267

led to the consideration of extra-scientific, socio-psychological criteria in

evaluation of theory choice (Frank 1957, chapter 15).
Frank argued as follows: "Logical consistency and observational agreement are insufficient criteria for unequivocal theory choice". One often has
to invoke additional epistemic criteria, such as simplicity and economy. Yet
the simplicity criterion is neither obvious nor stable over time. Theoretical
physicists raised in a mathematical atmosphere found Einstein's theory of
gravitation incredibly beautiful and simple, while experimental physicists
found these formulas extremely complicated. Thus, the notion of simplicity,
according to Frank, is historically and culturally conditioned (ibid.).
Yet even the criteria of simplicity does not narrow our choices to a unique
one. In the case of Copernican theory, as opposed to the Ptolemaic theory,
other reasons then epistemic simplicity-religious, moral, and politicalplayed a crucial role in theory choice. Or, in Frank's own words:
" ... fitness to support desirable conduct on the part of citizens
or, briefly, to support moral behavior, has served through the
ages as a reason for the acceptance of a theory... . We have
learned ... that the general principles of science are not unambiguously determined by observed facts. If we add reqwrements
of simplicity and common sense, determination becomes narrower, but it does not become unique. We can still require their fitness to support desirable moral and political doctrines. All these requirements together enter into the determination of a scientific
theory" (Frank 1957, pp. 355-355).
The words of Frank, the logical positivist, are peculiarly reminiscent ofperhaps sometimes indistinguishable from-the arguments of social constructivists, feminists and left intellectual critics of science today. When the
philosopher and feminist Helen Longino argues that we can have both empirically adequate and value desirable theories, she faithfully echoes Frank's
analysis. Sokal (1996), in his parody, ridiculed the words of Andrew Ross who
stated that scientific theories should be "publicly answerable and of some
service to progressive interests." Sokal could have, as well, preceded these
words by a quotation from the oracle of reason himself-the logical positivist
Frank.
To say that social constructivism can be seen as an offspring from a version of logical positivism is not to assert that Kuhn's and Feyerabend's impact
was dispensable. For it is in their writings that another philosophical thesis-

268

Mara Beller

still more corrosive to the rationality of science than "under-determination of

theory by data"-is developed to its utmost. It is the thesis of theoryladenness of observation (there are no bare facts, but only theory-dependent
interpretations of phenomena) which leads to an almost irrefutable relativism.
The version of "irrefutable" relativism runs as follows:
Scientists adhering to different paradigms will literally live in different incommensurable worlds, containing different facts. No rational argument, including appeal to observational evidence, can support the choice of one paradigm over another: the pre-revolutionary paradigm will be supported by prerevolutionary facts, while the post-revolutionary paradigm by the revised,
post-revolutionary facts. Knowledge thus clearly is relative to the conceptual
framework one is operating within-one can arbitrarily choose the framework according to one's own whim. Science becomes indistinguishable from
non-SClence.
For those who fInd this blanket relativism unpalatable, the problem is at
least well defIned: something is wrong either with the notion of holism of
frameworks or with the notion of theory-Iadenness of observation, or both.
Do we have to accept the notion of theory-Iadenness of observation and of
underdetermination of theory by data? I think we do-these notions are supported by ample historical evidence, from the age of Copernicus to the times
of Heisenberg and Schrodinger. But do we have to accept the notions of holism and subsequently of incommensurability? Both historical evidence and
philosophical argument clearly indicate-no, we do not.
The notion of holism underlies Kuhn's notion of incommensurability. All
throughout his career, especially in his later writings, Kuhn emphasized the
importance of holism for understanding the nature of revolutionary change.
Paradigms are holistic because they require a number of interrelated changes
of theory to be made at once-only at the price of incoherence can such
changes be made piecemeal, one step at the time (Kuhn 1987).
According to Kuhn himself, intrinsic holism implied by the Duhem-Quine
thesis played a formative role in his understanding of science (1993). Social
constructivists such as David Bloor and Batry Barnes often invoke this thesis
in their defense of relativism. According to this thesis, an experiment does
not contradict a single theoretical statement (which consequently can be saved
from refutation), but undermines the whole theoretical framework at once.
My argument is that, contrary to Kuhn and others, this holism does not
imply a necessary logical connection between the different aspects of the theory. The fact that a theory is holistic in the sense that one rather than the

Neither Modernist Nor Postmodernist-A Third Way

269

other aspect can be disproved does not mean that they all stand or fall together. Quite the contrary is the case-the fact that we can choose at will
which one of the elements is suspect without implying others, means that the
theoretical frameworks do not have a logical cohesion implied in the notion
of holistic paradigms (Beller 1997).
It also means that paradigms are not necessarily incoherent, if modified
piecemeal. The following example will suffice: According to Bohr, the
acausality and contextuality (experimental set-up dependence) of measurement results are intrinsically bound together. Bohr's numerous thought experiments reveal a supposedly strict logical connection between the two. In
Bohm's version, contextuality of measurement is retained, yet the "inevitability of acausality" is challenged-Bohm's version can be a deterministic one.
Not surprisingly, a closer study of the history of science undermines the
notion of the holism of paradigms. Rather than dogmatic commitment to a
rigid set of ideas, creative scientific work-whether in the Copernican, the
Chemical or the Quantum Revolution-is characterized by the ingenious
mingling and selective appropriation of ideas from distinct paradigms. Not an
autistic incommensurability, but fundamental communicability underlies scientific effort (Beller 1997, 1999).
In a local interactive setting, theory-Iadenness of observation assumes, I
want to suggest, a constructive, rather than subversive role. Rather than serving to initiate the irrational disruptive break with the past, theory-Iadenness of
observation allows to extend and to consolidate the acquired knowledge. I
have argued elsewhere that such was Heisenberg's reinterpretation of intraatomic orbits-not as a continuous path, but as a sequence of discrete points
subject to uncertainty relations (Beller 1999).
In such ongoing practice of interactive scientific theorizing, some of our
deeply entrenched philosophical notions become strikingly inadequate, and it
is they who lead to the pseudo-problem of relativism. Thus, the notion of
"theory choice" is the key-term of the philosophy of science. Yet theories are
not chosen by some epistemic rules of comparison; theory-choice is an after
effect of the ongoing communicative practice of scientific theorizing. In the
setting of such local, open-ended practice of science, relativism becomes not
so much refuted, as simply beside the point.
It would, of course, be misleading to maintain that Kuhn and Feyerabend
introduced the concept of holism. As we know today there are close links between the ideas of logical positivists and the neo-Kantian emphasis on holism, and Kuhn's work relies extensively on these ideas, notably on Carnap's

270

Mara Beller

work. In particular, we can find in Carnap's work a verSlOn of theoryladenness of observation, of semantic holism, and of incommensurability as
untranslatability.(Friedman 1987, Earman 1993).
It was Kuhn who combined the epistemological insights of logical positivism with the sociological notion of scientific collective. The holism of the
positivistic notion of conceptual structures was canonized in the Kuhnian notion of paradigms. The unholy alliance between these rigid structures and
homogeneous scientific community resulted in a redefinition of scientific
truth: "As in political revolutions, so in paradigm choice-there is no standard higher than the assent of the relevant community." Kuhn's followers,
notably those of the Strong Program of Sociology of Knowledge, were naturally led to the inadequate notion of scientific truth as no more than consensus.
I will now turn to an analysis of the notion of a scientific collective and of
consensus in the Kuhnian and post-Kuhnian studies of science.
The notion of agreement, or consensus, is a basic notion of the studies of
science. While for the positivistic philosophy the notion of consensus was
implicitly presupposed in the very notion of a conceptual scheme, model or
framework, Kuhn's achievement is taken to be in the explicit introduction of
the notion of scientific community and of consensus-a paradigmatic agreement on the shared methods, goals, metaphysical presuppositions, epistemic
virtues, exemplars, etc. With a few rare exceptions, the post-Kuhnian historiography and philosophy of science is characterized more by a reduction in
scale Oocalization) of the Kuhnian description of science than in its basic reevaluation. If there seems to be a consensus in studies of science, it is of the
indispensability of the notion of consensus.
Philosophers of science as different as Philip Kitcher (1993) and Ernan
McMullin (1991), sociologists of science as divided on basic questions as
David Bloor (1976) and Stephen Cole (1992), all seem to agree that the notion of consensus, or collective agreement, is the basic notion for evaluation
and explanation of scientific results. According to Kitcher, in order to understand the progress of science, "we need to be able to articulate the relations
among successive consensus practices" (1993, p. 87). To describe the range
of scientific activities is to pay "particular attention to the role of tradition, convention, agreement, and the social processes wherein such things are sustained or
undermined," claim Barry Barnes, David Bloor, and John Henry (1996, p. 10)
in their recent textbook on the sociology of knowledge. The militant oppo-

Neither Modernist Nor Postmodernist-A Third Way

271

nent of the Strong Program, sociologist Cole, agrees nevertheless with these
authors, approvingly citing Ziman's opinion that the goal of science "is a consensus [italics in original] of rational opinion over the widest possible field"
(Cole 1992, p. 102).
All these authors agree that knowledge is "communally held," that knowledge is a "collective good," and that eventually all the disputes are resolved
and disagreements eliminated. The difference between Kitcher and Stephen
Shapin (1994), or between Cole and Bloor, concerns the ways by which disputes are resolved and consensus is achieved-by rational processes for
Kitcher and Cole, by social mechanisms hardly related to the validity of
knowledge under discussion, for Shapin and Bloor. To understand science
sociologically (or philosophically) in the post-Kuhnian era is to understand
the collective aspect of science. Shapin makes this point forcefully:
"If sociology is the study of the collective aspects of human
conduct, then a basic role for the sociological study of scientific
knowledge is showing in what ways that knowledge has to be
understood as collective good and its implication as a collective
process. If there is a fundamental and irreducibly sociological
point to be made about scientific knowledge, it is this one"
(Shapin 1995, p. 302).

Closely connected with the notion of consensus is the notion of "solidarity."

Richard Rorty sees the scientific solidarity achieved as a result of "unforced"
agreement. According to Rorty, solidarity is a substitute for "objectivity" and
is a model for the democratic society in general.
The notion of a collective, in which there is a paradigmatic agreement, or
where such an agreement is posited as a goal is, of course, a thoroughly
monological notion. The positivistic legacy of European rationalism, with its
presupposition of self-sufficiency of a single consciousness and universality
of Reason, allowed a consolidation of monological ideology in all spheres of
cultural activity-philosophy, science, psychology, literature and even sociology (Bakhtin 1984, pp. 82-83; Sampson 1993, pp. 31-65). The notion of consensus, of knowledge as collectively shared goods, as communally held beliefs, entered into historiography and sociology of scientific knowledge from
the French tradition in the sociology of knowledge with a heavy reliance on
the writings of Emile Durkheim. This Durkheimian concept of knowledge
and of society is shared by Merton, by Kuhn and by Bloor in spite of other
basic differences between them.

272

Mara Beller

In the Durkheim's tradition the terms of "collective representation," "collective consciousness" and of "collective mind" are basic. The problem with
this conceptual heritage is not merely that such a notion of homogeneous
monological society, where "all hearts beat in unison," in which all minds are
indistinguishable is valorized. The problem is that such Durkheimian solidarity is achieved at a price. Such a society is that of order, and order is achieved
by discipline, by dominance. "Discipline," writes Durkheim, "has its justification in itself. It is good that man should be disciplined" (quoted by Catlin
1968). The notion of authority is central in sociological studies of science.
Collectivization of an individual mind is achieved by control, by execution of
authority. Sociologists of science, such as Shapin and Bloor, talk in terms of
"cognitive order."
It is not clear what the term of "cognitive order" means, when we consider societies other than the tribal and primitive religious communities, or
"closed societies" analyzed by Durkheim, especially scientific communities,
characterized by fruitful uncertainty, freedom to doubt, open-endedness, lack
of stability, constant change. Cognitive order is achieved, in contrast, by coercion and indoctrination. In Mary Douglas' characterization, our "colonization
of each other's mind is the price we pay for thought" (Douglas 1975, p. XX).
The notion of the "cognitive order" in Durkheim's static terms was
adopted, in a self-conscious way, by the founders of the Strong Program in
the Sociology of Knowledge. Durkheimian notions developed for closed,
primitive tribal societies, were adopted by Bloor and others wholesale, without critical reexamination, programmatically.
As Bloor clearly stated, Durkheim's study of religion provides "models
and sources of inspiration" for the sociologist of science (Bloor 1976, p. 4). It
is this uncritical acceptance of the analogy of science and religion that provided the inspiration for the most far-reaching symmetry thesis of the Strong
Program: true and false beliefs should be treated (explained) in the same way.
It is the notion of "belief' that made the transition from the Durkheim's approach to the study of science possible.
This widespread use of the notion of belief in the studies science is highly
inadequate-it implicitly blurs the difference between science and nonscience. The notion of a best informed (temporary) presupposition is much
more suited to describe the changing, open-ended dynamic nature of science
than the static notion of belief. Of course, the symmetry thesis of the Strong
Program, when we replace the notion of belief with that of best informed

Neither Modernist Nor Postmodernist-A Third Way

273

presupposition, falls apart-why should the well informed and badly informed statements be accounted for in the same way?
While I cannot develop this issue in the confines of this paper, it is worthwhile to mention that another principle of the Strong Program-that of
causality-also was directly appropriated from Durkheim (compare with
Durkheim 1968, pp. 89-140). So was the program of reduction of all philosophical questions to sociological, as well as the Durkheim's positivistic definition of social fact with subsequent almost pathetic tension between the naive
realism in the description of the social and unrestrained relativism in the description of the natural. The central notion of truth as shared belief, or truth
as consensus, also was adopted from Durkheim. Bloor criticized the philosopher's notion of truth as teleological. Yet the notion of "consensus" as a future normative state of affairs (despite initial disagreements) is no less teleological then the notion of a predetermined truth of a theory, towards which
the scientific efforts inevitably lead. The sociological and the positivistic notions of truth are merely simplistic mirror images of one another.
Durkheim's solidarity is totalitarian in its essence, as is the Kuhnian notion
of scientific collectivity, characterized by a monological discourse during the
paradigmatic regime. In this tradition, there is no place for human individuality, scientific or otherwise. The individual is disruptive to society. The only interesting question in such accounts is that of the techniques of "colonization," or, in more politically correct terminology, strategies of persuasion. As
Shapin put it, "the gap between individual, or public, must be filled by persuasion" (Shapin 1995, pp. 305). This approach gave rise to an enormous
amount of studies of rhetoric in science. In the confines of Durkheim's
framework, most such studies focus more on strategies of persuasion rather
than on the communicative underpinnings of good reasoning. The Durkheimian terms, developed for tribal societies, do not supply tools that can distinguish between evidence-based knowledge and superstition, between science and fiction-the major problem of the current sociology of science.
While Latour's dynamic account of science in action considerably transcended the static timeless structures of the Strong Program, it still retained
this notion of creation and transmission of knowledge as the process of colonization and dominance. In Latour's account, the cognitive order is achieved
by relentless politics of intellectual conquest, by imperialism of networks of
alliances.
Pledging for a society of liberal democracy, Rorty envisaged a culture
which is "secular through and through. It would be one in which no trace of

274

Mara Beller

divinity remained ... no room for the notion that there are non-human forces
to which human beings should be responsible". Rorty identifies the notion of
"devotion to truth" as a notion that retains religious remnants (Rorty 1991,
pp. 38-39). I would rather argue that it is the notions of "consensus", of a
"collective spirit", "collective mind" and the like that retain the mythical,
non-human forces of the primitive tribal societies. Durkheim himself called
these terms "hyperspiritual" (Durkheim 1968, p. 34). The notion of a "groupmind" (or of a "paradigm" for that matter) has clear non-secular roots and
God-like attributes. "Consensus" is nothing else but the opinion of the "hyperspiritual", omnipotent group-mind. The notion of consensus as normative
notion has no place in a free inquiry.
Let us pause and realize the similarity between the position of the logical
positivists and the post-Kuhnian sociologists of science. Both employ holistic
and static terms for understanding of scientific theorizing (what I mean is that
both can pay lip-service to the dynamic nature of science, but lack the conceptual tools to describe it). Both share the monological vision of Enlightenment: a single consciousness is sufficient for scientific knowledge-be it an
individual mind in the case of positivism, or the super-mind of a collective
for social constructivists. There is a fundamental unanimity between mindsbetween rational minds because of the adherence to a valid scientific method
(positivism),or unanimity between members of a scientific collective because
of authority, tradition, persuasion or even coercion ( social constructivism).
Despite Kuhn's substantial reservations from the Strong Program (Kuhn
2000, chapter five), the roots of his conception of scientific collectivity are
similar, if not identical, to that of Bloor.
In order to see the sources of Kuhn's notion of scientific community, it is
instructive to compare his approach with Ludwik Fleck's pioneering Genesis
and Development if a Scientific Fact (Fleck 1979). Kuhn's debt to Fleck seems to
be more extensive than is acknowledged by Kuhn. Many of Kuhn's notions
are remarkably similar to and easily traceable to Fleck. More than a superficial
similarity can be found between Fleck's proto-ideas (Fleck 1979, pp.23-27)
and Kuhn's paradigmatic ideas, between Fleck's thought styles ("autonomous,
style permeated structures'') and Kuhn's paradigms, between Fleck's "tenacity
of systems of opinion" and Kuhn's notion of resistance of paradigm to
change. Even the notion of incommensurability is clearly stated by Fleck: " ...
direct communication between the adherents of different thought styles is
impossible" (Fleck 1979, p. 36).

Neither Modernist Nor Postmodernist-A Third Way

275

But while Kuhn's philosophical debt can be traced to other sources (Beller
1997, 1999), it is Kuhn's incorporation of the notion of a scientific community into philosophical discussion which made Kuhn's contribution unique
and significant. In this respect, as Kuhn himself acknowledges, his debt to
Fleck is direct: " ... Fleck's work made me realize that those ideas [those of
concern to Kuhn-M. B.] might require to be set in the sociology of scientific community" (Kuhn 1962, pp. vi-vii).
What is the source of Fleck's notion of collectivity and of scientific community? Fleck himself stated that he directly relies on "Durkheim's school in
France," which stressed " ... the importance of sociological methods in the
investigation of intellectual activities" (Fleck 1979, p. 46). Fleck approvingly
cites Durkheim's ideas about "force exerted on the individual by social structure," "superindividual and objective character of ideas belonging to the collective," the notion of "collective intellect" that determines the religious and
cosmological ideas of a tribe. He also cites Levi-Bruhl, "a student of Durkheim," who, like Durkheim, held that the sociological study of primitive religions is useful for understanding "our categories and logical principles" (Fleck
1979, p. 97). In is not an individual who thinks, but rather "his social community" (Fleck 1979, p. 47, citing Gumplowicz).
In their recent book Bloor, Barnes and Henry confidently declared that
their image of scientific knowledge was firmly established in the studies of
science. They called their model the "normal model.": "Knowledge in this
normal version was the possession of the members of a culture or subculture,
transmitted from generation to generation as a part of their tradition, and dependent for its credibility on their collective authority" (Barnes et al. 1996,
p. 111). They admitted, in fact proudly announced, that this mode originated
from a "small literature on the sociology of knowledge which was in fact
mainly concerned with religious doctrines and political ideology ... and, perhaps most valuable of all, anthropological studies of knowledge in tribal societies". And while initially this model seemed inapplicable to scientific
knowledge, "it did nonetheless prove possible to assimilate scientific knowledge ... into normal view" (Barnes et al. 1996, p. 111). The biggest inspiration
and support came, of course, from Kuhn, who "identified scientific knowledge as the possession of a collective, incarnate in their tradition of research."
Kuhn's work "served to vindicate the sociological conjecture that the normal
model would apply everywhere, to any body of knowledge ... " (Barnes et al.
1996, p. 112).

276

Mara Beller

It is not by chance that this model is called "normaL" The Kuhnian and
the Strong Program notion of collective (as well as that of Merton) inherited
the distinction, so central for Durkheim, between the "normal" and the
"pathological" (Durkheim 1968, p.47-76). There are "norms" of Merton,
and, of course, the "normal" science of Kuhn. Genuine expression of individuality is pathological. Durkheim identified the "normal" with the "average" (remember the Kuhnian army of obedient puzzle solving). The demonstration of individuality during the paradigmatic regime of Kuhn is not merely
deviance, it is a waste of effort, and, as any abnormality, must be illogical and
irrationaL Any genuine novelty (revolutionary advance) can only occur by inexplicable irrational jumps during the time of disorientation and chaos. The
revolutionaries, endowed with the divine irrationality (a pale version of a Romantic mad genius), are exempted from the paradigmatic discipline.
The current historiography of science must labor hard to undermine these
false Durkheimian dichotomies between individual and collective, between
order and chaos, between normal and pathological. Lack of stability, genuine
change is not to be identified with chaos, and creative novelty is not to be
perceived as pathological. In order to do so, we have to break the opposition
between autistic individuality and homogenous collectivity. Such an alternative, I will briefly argue, is provided by a dialogical approach (developed in
Beller 1999). In the dialogical approach, in which it is impossible to conceive
of any individual as being outside the relations that link her to others, such
false dichotomies do not arise.
In the dialogical approach disagreement, rather than consensus or agreement, provides the basis for scientific interaction and for ensuing change.
There is a widespread mythology, shared by Kuhn, Rorty and many postmodernist, that a high degree of consensus characterizes natural sciences as
opposed to social sciences and humanities.
In Rorty's words: "If we say that sociology or literary criticism 'is not a
science,' we shall mean merely that the amount of agreement among sociologists or literary critics on what counts as significant work, work which needs
following up, is less than among, say, microbiologists" (1991, p. 40).
Such sensibilities seem to be natural, almost obvious, and they fit equally
well the positivistic and the Kuhnian historiography. And it is exactly this
overarching complacent acceptance of the notion of consensus, despite the
basic differences between positivists and social constructivists, that makes
this notion suspect.

Neither Modernist Nor Postmodernist-A Third Way

277

When Cole (1992) undertook an empirical comparative study of degrees

of consensus in natural and social sciences, he merely expected to confirm
the prevalent opinion of a high degree of consensus in the natural, as opposed to the social sciences. Yet he repeatedly failed to confirm this hypothesis. He found, for example, that there was a very similar degree of consensus
(around 60+%) in a group of biochemists and in the group of sociologists
about what counts as a significant contribution and which work needs following up. Still more revealing is Cole's study of the degree of disagreement
among qualified reviewers about grant proposals in the natural sciences submitted to National Science Foundation. The level of disagreement in evaluation of these proposals was so high that Cole could find no better criterion
than "luck" for accounting whether the proposal was funded or not (Cole
1992, pp. 82-101).
In a sense, this conclusion can hardly be considered sensational. Previous
studies of "science in action," such as Latour (1987), Pickering (1984) or
Collins (1992), reveal the prevalence of the scientific controversy at the frontier. Yet the state of disagreement, they concede, is a temporary state of affairs. The valid, or rather the accepted, part of science is the one that remains
after all disputes are resolved. The conclusion is the same whether one assumes, with the positivists, that results are accepted because they are valid, or
with constructivists that results are considered valid because they became accepted.
I want to argue, in contrast, that disagreement is neither a necessary evil,
nor a temporary aberration. My central claim is that disagreement is the creative principle of scientific practice. In this respect, a study by Henry Small
(1974, referred to in Cole 1992) is revealing. He found that for articles submitted for publication to chemistry journals, the articles that received the
most criticism from referees, later became most cited. The most stimulating
and fruitful statements are those with which we disagree.
I do not claim that the question about how a reasonable agreement can be
or is achieved, is devoid of value. Yet the study of reasonable agreement cannot be based on the notion of consensus. The sociological studies of the degree of disagreement at the frontier of science suggest the naturalness, the
reasonableness of disagreement. Yet consensus excludes such reasonable disagreement. Consensus means just that-unreasonableness of disagreement,
irrationality of disagreement. By excluding legitimate disagreement, the state
of consensus cannot result in a rational scientific change. No wonder
Kuhnian accounts have no epistemological tools to account for scientific de-

278

Mara Beller

velopment except by an ad hoc postulation of irrational jumps from one

framework to another, incommensurable framework. "Incommensurability"
is a direct consequence of illegitimacy, "irrationality" of local disagreement.
If both positivists and social constructivists failed in their account of scientific change, they failed for the same reason-their resignation of exploring
the constructive (rather than merely temporal and inconsequential) role of disagreement in the advance of knowledge.
As an illustration, I will briefly summarize some aspects of a dialogical
analysis of Heisenberg's uncertainty principle (Beller 1999). In 1925 the young
Werner Heisenberg laid the foundations of the new quantum theory by
eliminating the electronic orbits of Bohr's model and replacing the classical
parameters of motion by algebraic constructs, which turned out to be matrices. Through this step, Heisenberg effectively abandoned Anschaulichkeit
(visualizability) of the quantum domain and dispensed with regular space-time
concepts.
A few months later, after the appearance of Schrodinger's theory a, and
especially his competing interpretation, Heisenberg often contemplated interpretive issues of the new quantum theory. How is one to interpret the new
formalism, when the space-time imagery no longer seems to fit? What can the
notion of "interpretation" in such circumstances mean? Are the discontinuous and acausal features of the new theory derivative, or fundamental? Does
the apparent incompatibility between the new formalism and experiment imply that quantum theory is in need of further elaboration and modification?
Or should one accept the formalism as complete and final, and to undertake a
suitable reinterpretation of some of experiments and experimental terms?
Heisenberg's elaboration of these issues took place in a context of intense
dialogue. I describe in detail how Heisenberg's opinion on these central issues
emerged gradually in his dialogues with Schrodinger, Bohr, Pauli, Dirac, Jordan, Born, Campbell and Senftleben.
The dialogical analysis demonstrates the indispensability of lesser scientists
in the process of the emergence of novelty. Two crucial components of Heisenberg's uncertainty paper were provided by "marginal" physicistsSenftleben and Campbell. Senftleben provided the formulation for Heisenberg's "definitive" resolution of causality issue: causality does not hold not
because quantum laws are inherently statistical, but because the present cannot be known precisely, and therefore the future cannot be predicted exactly.
Campbell, before Heisenberg, suggested that an intra-atomic electronic path
should be considered as a sequence of discrete irregular points.

Neither Modernist Nor Postmodernist-A Third Way

279

All these physicists supplied a vast resource of ideas and techniques that
Heisenberg borrowed. But no less, and perhaps more significant, were
Heisenberg's disagreements with his actual and potential interlocutorsdisagreements that triggered, enriched and constantly rearranged the complex
pattern of Heisenberg's quest.
Yet all these disagreements, to which I have alluded, could not have performed their creative, formative function, unless Heisenberg have not been
prepared to qualify, to modify and sometimes even to reverse his earlier
stand. Disagreement can lead to novelty, rather then to merely futile controversy, only if there is a basic open-mindedness, lack of dogmatism, element
of ignorance, uncertainty, and genuine doubt at the frontier of science. The
existence of doubt (disagreement with oneself), as well as disagreement with
others, is a necessary condition for the occurrence of scientific change. Again,
in Feynman's words: "I believe that to solve any problem that has never been
solved before, you have to leave the door to the unknown ajar. You have to
permit the possibility that you do not have it exactly right"(1998, pp. 26-27).
If the living scientific research is characterized by doubt and uncertainty,
we should be able to uncover the traces of doubt in a published scientific paper. A careful analysis might reveal vacillations, conflicting inner voices and
lack of confidence, despite the image of published paper as monological and
unequivocal structure. I have argued (Beller 1999) that Heisenberg's uncertainty paper is permeated with remnants of past struggles and conflicting inner VOlees.
I have also analyzed along these lines Bohr's Como lecture, where he first
announced his principle of complementarity, and which is considered as one
of the most obscure scientific texts. This lecture becomes however lucidly
clear when we realize that the text is not a monologicallinear argument, but a
juxtaposition of several arguments, addressed simultaneously to different adressees (Beller 1999, pp. 117-145).
Based on an extensive analysis of the process of discovery and the structure of scientific papers in quantum physics, I came to the conclusion of the
need to articulate a dialogical approach to the study of science. My proposal is
to treat an addressive response as the primary epistemological and social unit
for the analysis of science. Thus, scientific thought presupposes the existence
of an interlocutor to whom the thought is addressed and/or by whose statements the thought is triggered. Such addressivity is complex and rnultichanneled; scientists simultaneously respond to many interlocutors and at the
same time address potential or actual addressees. This open-ended multidirec-

280

Mara Beller

tional and constantly changing web introduces conceptual tensions and ambiguity that underlies the creative scientific effort. The notion of addressivity is
not a static concept, it is a temporal notion. It introduces locality, contextuality and historicity of scientific practice in a fundamental way.
Dialogical analysis incorporates conversation and communicability both as
a social reality and as an epistemological presupposition. From the dialogical
perspective, a creative scientist is linked fundamentally to the efforts and concerns of others. Science is simultaneously social and rational-the rationality
of science is dialogical and communicative.
A successful scientific result is one that is strongly embedded in the practice of the day. This embedding is achieved by simultaneously responding to
as many other scientific statements (from other papers, conversations, etc.) as
possible-drawing on many possible support links (confirmations), taking
into account many possible objections (refutations). Both the Carnapian idea
of confirmation and the Popperian idea of refutation are, again, onedimensional idealizations, and both can be concurrently present in a dialogical
response.
Not all aspects of scientific theorizing are equally well grounded-some
are more arbitrary than others. While Heisenberg's uncertainty formula, or
Schrodingers's equation (or modification thereof in GRW theory [Ghirardi et
al. 1986]) are indispensable for the working quantum physicist, he or she can
create without the Copenhagen Interpretation. Thus, the Copenhagen Interpretation is dispensable-many physicists, who are engaged in successful scientific theorizing, are opposed to it, while many others, subscribing to it, nevertheless do not employ it in their research. It is reasonable to consider such
dispensable parts of science as being more "subjective" and arbitrary than the
indispensable ones. It is illuminating to analyze these aspects of science in
terms of rhetorical strategies of persuasion, as is done in social studies of science.
Dialogism, let me emphasize, is not an explanatory theory of the growth
of knowledge. It does not suffer from major weakness of explanatory theories-how do we get from cause to effect (from Nature to theory-in positivism, or from social variables to scientific statements-social constructivism). Dialogism is a phenomenological approach for studying scientific
change (and also a methodology to for studying scientific texts): the flux of
dialogues is not an explanation of scientific change, it is the change itself]

Neither Modernist Nor Postmodernist-A Third Way

281

The dialogical approach that I developed for the description of the history
of quantum physics bears some similarity (as it turned out) to the ideas of
Mikhail Bakhtin, the Russian literary theorist and philosopher. The Bakhtinian approach is based on the concept of dialogue and indispensability of the
notion of the "other" in human existence and thinking. We are alive to the
extent that we are responsive (and responsible), as well as addressive to the
other.
Bakhtin singled out Dostoyevsky as the discoverer of the new literary
genre-the polyphonic novel. Dostoyevsky's heroes are not clearly defmed
characters-their truth, never a completed one, emerges in confrontation
with, and anticipation of, the other's point of view. Every idea, that of the
characters and that of the author, is a rejoinder in unfinished and unfinalizable dialogues carried out simultaneously.
Bakhtin contrasts such polyphonic thinking with monologism of natural
science. According to Bakhtin, in science there is no genuine place for the
other: the other in science is the one who errs. Contra Bakhtin, I argue that
the "other" is fundamental in science not less, and arguably even more, than
in other human endeavors. Dialogism is a social epistemology of science-it
presupposes the existence of the others mind as a necessary condition of scientific productivity.
Dialogism as social epistemology differs fundamentally from social constructivism. Dialogism does not hold that social determinants (in the sense of
the usual sociological variables) are the primary determinants of scientific
ideas. As opposed to social constructivism, dialogism does not conflate beliefs and facts-dialogism shies away from the notion of belief altogether.
Dialogism rejects the social constructivists' notion of the consensus theory of
truth: dialogism rejects the notion of consensus as mandatory and the notion
of truth as explanatory.
Dialogism is in agreement with the defenders of reason that a thorough
understanding of science is a precondition for its sociological and philosophical analysis. In fact, in the dialogical approach, one cannot even begin to analyze the mutual responses without going into minute details of scientific reasonmg.
Yet dialogism shares an important insight of social constructivism-the
blurring of a clear distinction between the context of discovery and context
of justification. Discovery and justification are in fact the same kind of dialogical process-permeated with responsiveness, adressivity and communicative rationality.

282

Mara Beller

Dialogism rejects the basic modernist idea of a self sufficiency of individual rationality, as defenders of reason still have it. Yet it centers on scientific
individuality-individuality that is completely eliminated from postmodern
studies of science which are based on the notions of homogeneous community, consensus and solidarity.
Thus, dialogism is neither modernist nor postmodernist. It is a third way.

References
BAKHTIN, MIKHAIL (1984), Problems of Dostqyevsky's Poetics. Edited and translated by Caryl Emerson. Minneapolis: University of Minnesota Press.
BARNES, S. BARRY (1974), T. S. Knhn and Sodal Sdence. London: Macmillan.
BARNES, S. BARRY, DAVID BLOOR, and JOHN HENRY (1996), Sdentiftc Knowledge---A SodologicalAnafysis. Chicago: The University of Chicago Press.
BELLER, MARA (1997), "Criticism and Revolutions," S dence in Context, vol. 10,
no.l, pp. 13-37 and vol. 10, no.6, pp. 241-255.
BELLER, MARA (1999), Quantum Dialogue---The Making of a Revolution. Chicago:
The University of Chicago Press.
BELLER, MARA and ARTHUR FINE (1994), "Bohr's Response to EPR." In:
FAYE/FoLSE (eds.; 1994), pp. 1-31.
BLOOR, DAVID (1976), Knowledge and Sodal Imagery. Chicago: The University
of Chicago Press.
BUNGE, MARIO (1996), "In Praise of Intolerance to Charlatanism in Academia." In: GROSS et. al. (eds.; 1996), pp. 96-115.
CATLIN, GEORGE E. G. (1968), "Introduction to the Translation". In: DURKHElM 1968.
COLE, STEPHEN (1992), Making Sdence: Between Nature and Sodery. Cambridge,
MA: Harvard University Press.
COLE, STEPHEN (1996), '<y"oodoo Sociology: Recent Developments in the
Sociology of Science". In: GROSS et. al. (eds.; 1996), pp. 274-287.
COLLINS, HARRY M. (1992), Changing Order: Replication and Induction in S dentific
Practice. Chicago: The University of Chicago Press.
DOUGLAS, MARY (1975), Implidt Meanings: Essqys in Anthropology. London:
Routledge & Kegan Paul.
DURKHEIM, EMILE (1968), The Elementary Forms of Religious Lift. London:
George and Unwin Ltd. [First published in 1915].
EARMAN, JOHN (1993), "Carnap, Kuhn and the Philosophy of Scientific
Methodology." In: HORWICH (ed.; 1993), pp. 9-36.
FEYNMAN; RICHARD P. (1998), The Meaning of It A/~Thoughts of a Citizen Sdentist. Reading: Addison-Wesley.

Neither Modernist Nor Postmodernist-A Third Way

283

FAYE, JAN and HENRY J. FOLSE (eds.; 1994), Niels Bohr and Contemporary Philosopf!y. Dordrecht: Kluwer.
FLECK, LUDWIK (1979), Genesis and Development of a Scientijic Fact. Chicago:
The University of Chicago Press [First published in 1935].
FRANK, PHILIPP (1957), Philosopf!y of Science. Englewood Cliffs: Prentice Hall.
FRIEDMAN, MICHAEL (1987), "Carnap's Aufbau Reconsidered," Nous,
vol. 21, pp. 521-545.
FRIEDMAN, MICHAEL (1999), Reconsidering Logical Positivism. Cambridge: Cambridge University Press.
GALISON, PETER L. and DAVID J. STUMP (eds.; 1996), The Disunity of Science:
Boundaries, Contexts, and Power. Stanford: Stanford University Press.
GHIRARDI, GIANCARLO, ANTONIO RIMINI, and TULLIO WEBER (1986),
"Unified Dynamics for Microscopic and Macroscopic Systems," Pf!ysical
Review D34, 470-491.
GROSS, PAUL R., NORMAN LEVITT, and MARTIN W. LEWIS, (eds.; 1996), The
Flight from Science and Reason. New York: New York Academy of Sciences.
HAACK, SUSAN (1996), "Concern for Truth: What it Means, Why it Matters."
In: GROSS et. al. (eds.; 1996), pp. 57-63.
HORWICH, PAUL (ed.; 1993), World Changes: Thomas Kuhn and the Nature of Science. Cambridge, MA: MIT Press.
KITCHER, PHILIP (1993), The Advancement of Science: Science without Legend, Oijectivity without Illusions. Oxford: Oxford University Press.
KRUGER, LORENZ, GERD GINGERENZER, and MARY S. MORGAN (eds.;
1987), The Probabilistic Revolution, vol. 1 and 2. Cambridge, MA: MIT Press.
KUHN, THOMAS S. (1962), The Structure of Scientijic Revolutions. Chicago: The
University of Chicago Press.
KUHN, THOMAS S. (1987), ''What Are Scientific Revolutions?" In: KRUGER
et al. (eds.; 1987), pp. 7-22.
KUHN, THOMAS S. (1993), "Afterwords." In: HORWICH (ed.; 1993),
pp. 311-341.
KUHN, THOMAS S. (2000), The &ad Since Structure. Chicago: The University of
Chicago Press.
LATOUR, BRUNO (1987), Science in Action: How to Follow Scientists and Engineers
Through Society. Cambridge, MA: Harvard University Press.
MCMULLIN, ERNAN (1991), "Rhetoric and Theory of Choice in Science." In:
PERA/SHEA (eds.; 1991), pp. 55-76.
PICKERING, ANDREW (1984), Constructing Quarks: A Sociological History ofPartide Pf!ysics. Chicago: The University of Chicago Press.
PERA, MARCELLO and WILLIAM SHEA (eds.; 1991) Persuading Science: The An
of Scientific Rhetoric. Canton, MA: Science History Publications.
RITZER, GEORGE (21996), Classical Sociological Theory. New York: McGraw-

Hill.

284

Mara Beller

RORTY, RICHARD (1991), Oijectivity, Relativism and Truth, Philosophical Papers,

Volume 1. Cambridge: Cambridge University Press.
SAMPSON, EDWARD E. (1993), Celebrating the Other, A Dialogic Account of Human Nature. Boulder: Westview Press.
SHAPIN, SlEVEN (1994), A Social History of Truth. Chicago: The Chicago University Press.
SHAPIN, STEVEN (1995), "Here and Everywhere: Sociology of Scientific
Knowledge," Annual Review of Sociology, vol. 21, pp. 289-321.
SMAIL, HENRY (1974), Characteristics ofFrequentfy Cited Papers in Chemistry. Final
Report on Contract no. NSF-C795. Philadelphia: Institute for Scientific
Information.
SaKAL, ALAN D. (1996), "Transgressing the Boundaries-Toward a Transformative Hermeneutics of Quantum Gravity," Social Text, vol. 46/47,
pp.217-252.

Jean Bricmont on Mara Beller

I will limit myself to two general remarks, one socio-historical, about Mara
Beller's fairy tale, the other philosophical, about the nature of my disagreement with a good part of the post-Kuhnian work in history and sociology of
science and the possibility of a "Third Way".
First of all, there was never (unfortunately) a Glorious Reign of Reason on
Earth. I will not attempt to discuss how much progress or regression "reason" has made in the society at large since, say, the 1950's. That would involve far too many factors. Let me focus only on the academic world, and
particularly on the social and human sciences, the humanities, etc. To use a
science fiction cliche, I think that someone who had been "frozen" in the 50's
and "woke up" today, would find that part of the academic world quite
changed. Not necessarily negatively, by any means: the present academic
world is far more open-minded, on all kinds of issues, than during the Cold
War. But if we narrow down our interest further and focus on the perception
that faculty and students have of science and technology, then I think the
changes are quite radical and not entirely positive. Again, there used to be a
fair amount of science worship and that had to be corrected. But the present
climate, what one might call the postmodernist Zeitgeist, seems to me to be a
strange combination of extreme ignorance about what science actually saysleading to exaggerated fears about what technology, especially biotechnology,
will do-and bizarre confusions about the epistemology of science. The publication of Sokal's hoax, as well as part of the ensuing debate, illustrated those
problems. Now, why does it matter? Well, in my university, and probably in
others too, more students hear about science in "philosophical" or "sociological" courses where they will learn about Popper, Kuhn, Feyerabend or Latour than in "scientific" courses where they will learn about the work of
Newton, Einstein or Darwin!, And that is particularly true of those students
who become high school teachers, politicians or journalists. In this way, the
postmodernist Zeitgeist spreads beyond the walls of academia to society as a
whole. Although it is difficult to measure the effects of all this, let me remark
that recent polls show that at the same time that the influence of organized

That is one reason why it might be useful to create more courses of the "science for poets"
type that exist in American universities.

286

Comments

religion is collapsing in Western Europe 2, beliefs in various superstitions such

as astrology, telepathy or alternative therapies are thriving. Concerning these
"therapies", it seems that a large fraction of the population, including many
doctors, do not understand what makes them "alternative" (the lack of systematic testing) and why one should object to them. It is of course difficult to
assess how much that situation owes to the various confusions about the
natural sciences encouraged by postmodernism. But it is to be expected that,
at the very least, those confusions do not help. Furthermore, in the discourses
about education, the emphasis is constantly put on the relativity of our
knowledge, on the importance of uncertainty, and on how emphatically all
this has to be communicated to students (who are probably already skeptical
enough about notions such as truth and objectivity).
This brings me to my second point, the notion of consensus in science. It
is true that, as Mara Beller points out, both positivist and post-positivist philosophers of science have emphasized the importance of consensus. I will not
discuss Mara Beller's criticism of that notion, which would take me too far
afield. Rather, I want to emphasize what seems to me a crucial difference between the positivists and the post-positivists, namely their respective attitudes
toward the basic question: where does consensus come from? Roughly speaking, before postmodernism achieved prominence, the dominant view was that
consensus arises because we are (sometimes) able to find out (some aspects
of) how things really are, and a community of rational scientists comes to
agree among themselves because they share, not common beliefs, but a common knowledge. Since this idea is nowadays often dismissed as naive, let me illustrate it with one example, which was actually mentioned during the conference: how do humans make babies? What are the respective roles of the
mother and the father? This is something that no mildly educated person in
the modern world could fail to know. But this is not the case in "other cultures" (or in the not-too-distant past in Europe), where the role of the sperm,
for example, may be misunderstood. Now, is our "theory" about how one
makes babies "underdetermined by data"? As a point in philosophy of science, the answer is obviously "yes"3. It is, in principle, possible that all our
2

For example, in my country (Belgium), which is not atypical in that respect, about 10% of
the population still goes to church weekly, and the figures are much lower among the
youth.
I want to emphasize that this is a purely philosophical remark, basically saying that radical
skepticism cannot be refuted, and not a historical one. Mara Beller seems to understand it
in a different sense, since she says that underdetermination of theories by data is supported

Neither Modernist Nor Postmodernist-A Third Way 287

theories about the world be false. But there are many theories-like the one
about how to make babies-where this theoretical possibility is just that: a
purely theoretical possibility. There are theories, of course, that are more
speculative, particularly those involving "unobservable quantities" where underdetermination is genuine-we discuss that briefly in our paper. Mara Beller gives the example of the "Copenhagen interpretation of quantum mechanics" versus alternative theories such as Bohm's. Assuming that the
Copenhagen vulgate can be considered as a theory, she is certainly right that
there is a genuine underdetermination here. But this has nothing to do with
the general philosophical thesis about theories being underdetermined by
data. It is a specific problem, in a specific context. So, whenever people want
to show that our scientific theories and the consensus about them are "socially constructed" in a nontrivial sense (e.g. in the sense that the beliefs about
making babies in other cultures are socially constructed), they have to deal directly with the issue at hand, confront the evidence offered by scientists, explain why they think it is insufficient, why alternative theories are preferable
or at least equally plausible, etc. But they can obtain no support from general
skeptical theses, such as the underdetermination thesis.
To conclude, I would say that one of the main cultural problems that we
face today, and the main one posed by postmodernism, is the destruction of
empiricism. I do not mean empiricism as a theory about how the mind works
(the idea that it is a tabula rasa at birth, etc.), which is almost certainly false,
but as an attitude toward what constitutes genuine knowledge. That only the
patient confrontation between clearly stated theories and carefully gathered
empirical evidence can qualify for that status; and the observation that, by following empirical methods, one has obtained a surprisingly deep understanding of the world, unmatched by what is reached by any other "method", such
as introspection, intuition, revelations, etc.
Bertrand Russell once observed that the Humean "self-destruction of rationality" was followed by a "great outburst of irrational faith"4. The destruction of positivism in the 20 th century may well have a similar effect. And between the friends and the enemies of evidence, reason and logics, I am afraid
that there is no "Third Way".

4
5

by "ample historical evidence". But I am not sure what that can mean-why is not the
theory that theories are underdetermined by data also underdetermined by data (i.e., by historical evidence), no matter how ample the latter are?
BERTRAND RUSSELL, p. 646.
Needless to say, I do not count Mara Beller among those enemies.

288

Comments

Jay Rosenberg on Mara Beller

Prof. Beller's paper is very rich, and there is consequently a great deal to say
about it. One thing that I want to say at the beginning is that I find Beller's
"dialogical" approach to the study of scientific change quite congenial. Disagreements can indeed be constructive and creative, and it is hard to quarrel
with the claim that "science is simultaneously social and rational". But one
point that I do want to stress is that all those dialogues and disagreements are
about something-the proper scientific understanding of the world, to put an
admittedly-contentious label on it-and one thing that troubles me about
Beller's approach is how infrequently that even gets mentioned. Sometimes,
to put it metaphorically, it is nature who disagrees with one of our theoriesthat is, sometimes our theoretically-shaped observational expectations are
strikingly disappointed-and we would not be engaged in scientific inquiry if
such "disagreements" weren't then centrally relevant to evaluating the continued acceptability of the implicated theories.
Permit me an extended analogy. The point of soccer is scoring goals. Insofar as you are playing soccer, your aims have to do with scoring goals: You
are trying to help your team to score goals, and you are trying to help prevent
the opposing team from scoring goals. Scoring goals is what I shall call the internal aim of soccer.
There are, of course, lots of reasons for joining a soccer team and for playing soccer. You might be interested in improving your physical fitness and
stamina; you might be seeking companionship or a pleasant recreational diversion or trying to measure up to your parents' expectations or hoping to
impress your friends with your sports prowess; if you're sufficiently skilled,
you might even be aiming at a career as a professional soccer player and then
perhaps also at wealth and fame. Analogously, there are lots of reasons for
forming or sponsoring soccer teams or leagues and for supporting or promoting soccer playing activities. Businesses might be seeking higher profits through effective advertising and improved public goodwill. Cities or states or nations
might be interested in increasing tourism and stimulating the economy, or
they might be aiming at enhanced civic, regional, or national pride or patriotism, or trying to distract a potentially rebellious public from political scandals
and corruption. All of these, and indefinitely many more, can be external aims
of soccer. But if what we're interested in is how well or poorlY someone plays
soccer, then it is his performance vis-a-vis the internal aim of soccer that

Neither Modernist Nor Postmodernist-A Third Way 289

counts. And none of these external aims could be external aims of soccer if
there wasn't independentlY such a thing as the game of soccer per se.
I am reactionary enough to think that science is like soccer, that is, that
scientific inquiry also has an internal aim. Of course a long story needs to be
told at this point, but for present purposes a convenient short label for this
internal aim will suffice, so I shall just call it "understanding nature". The
point of scientific inquiry is understanding nature. Insofar as we are engaged
in scientific inquiry when we formulate, evaluate, and adopt or reject particular theories, our aims have to do with understanding nature.
Beller asserts that religious, moral, and political considerations have historically played crucial roles in theory choice, and she quotes Philip Frank as
claiming correlatively that "fitness to support desirable conduct on the part of
citizens or, briefly, to support moral behavior, has served through the ages as
a reason for the acceptance of a theory". But such claims are at least ambiguous and, I think, as they most naturally understood, in fact mistaken. It is certainly undeniable that religious, moral, and political considerations have been
adduced in debates over, for instance, the promulgation of scientific theoriese.g., Copernican cosmology or Darwinian evolutionary theory-and even
more frequently in debates over the applications of such theories-witness the
current disputes in Germany regarding pre-implantation genetic assessment
of in vitro human embryos. Religious, moral, and political considerations have
certainly shaped decisions regarding which projects of scientific inquiry to advocate, to promote, to sponsor, or to fund, e.g., the human genome project, but not
the superconducting supercollider. But all this shows only that religious,
moral, political, and other (e.g., social or economic) interests have served as
external aims of science.
What is not at all clear, however, is that such considerations have ever
been regarded by the practitioners of scientific inquiry, i.e., by scientific theorizers themselves, as "reasons for the acceptance of a theory" or grounds for
theory choice. If what we're interested in is the scientific merits of a particular
theory, then it is the theory's performance vis-a-vis the internal aim of science
that counts, that is, its contribution to our understanding of nature. That's
what we need to keep in mind when we assess such claims as Feyerabend's
notorious conclusion, which Frank presumably would have endorsed, that "in
the conflict between Galileo and the Church it was the Cardinal Bellarmine's
judgement, and not Galileo's, which was the better one". Perhaps it was indeed arguably good social policy to promote geocentrism and suppress heliocentric theories. Perhaps permitting the open advocacy and unrestrained dis-

290

Comments

semination of Copernican cosmology would indeed have "disrupt[ed] the

whole fabric of moral and religious life". But none of this implies that heliocentrism wasn't the scientijical!J better theory, i.e., the one that Galileo had the
better scientijic reasons to accept and adopt-and that, of course, was Galileo's
own unremitting conclusion. "And yet it moves."
Beller takes Frank as her prototypical logical positivist, but it is worth remembering that the positivists were a diverse group. Not every thesis endorsed by Frank would be equally congenial to, say, Schlick or Neurath or
Carnap, not to mention such late-come epigones as Ayer. In particular, I
would hope that they would all roundly criticize the normative non sequitur
that Frank apparently drew from the failure of observational data, even when
supplemented by considerations of such "epistemic virtues" as simplicity and
economy, to uniquely determine a choice of theories, namely, that "We can
still require their fitness to support desirable moral and political doctrines". (parenthetically, I'm quite skeptical about the claim that simplicity and economy are epistemic virtues, but that's another story for another time.) Of course, we can require that candidate theories fulfill af!Y requirement that we might think upeven such silly ones as that they be formulated in rhyming couplets-but that
has nothing to do with their merits as scientijic theories. Analogously, since impressionable children often treat professional sports figures as role models,
we can require that members of major league soccer teams be well-groomed,
polite, soft-spoken, unfailingly friendly, morally upright, and religiously devout-but none of that has anything to do with how well they play soccer. If
that is what we are interested in, then what counts is only their offensive and
defensive performance vis-a.-vis scoring goals. Cardinal Bellarmine and perhaps Frank were concerned with scientists and their intellectual achievements
as moral, social, or cultural role models. That's a perfectly legitimate thing to
be concerned about-but it shouldn't be confused with the question of how
those scientists and their theories measure up as such, that is, in terms of the
internal aim of scientific inquiry.
Beller, of course, is neither Frank nor Rorty, and she does not align herself
with the kind of strong social constructivism that most naturally resonates
with normative claims of the sort that I have just been critically exploring. As
Beller herself insists, "Dialogism as social epistemology differs fundamentally
from social constructivism. Dialogism does not hold that social determinants
(in the sense of the usual sociological variables) are the primary determinants
of scientific ideas." Rather her "central claim", she tells us, "is that disagreement is the creative principle of scientific practice", and although I have

Neither Modernist Nor Postmodernist-A Third Way 291

qualms about the definite article, I have quite happily already conceded that
"legitimate", "reasonable" disagreement is a creative principle of scientific
practice, by no means precluded by the norms of scientific inquiry per se.
That said, however, I find that I remain quite unclear about what, if anything, Beller takes to be the norms of scientific inquiry per se, and, indeed,
about whether she is even prepared to acknowledge anything falling under
that rubric. "Dialogism," she writes, "is not an explanatory theory of the
growth of knowledge", but rather "a phenomenological approach for studying scientific change (and also a methodology for studying scientific texts)".
But descriptive phenomenology is one thing; methodological heuristics, another; and normative epistemology-even normative social epistemologystill quite another.
The crucial question is how we are to understand, e.g., Beller's assertion
that "a successful scientific result is one that is strongly embedded in the
practice of the day", where such "embedding" is achieved dialogicallY, "by simultaneously responding to as many other scientific statements (from other
papers, conversations, etc.) as possible--drawing on many possible support
links (confirmations), taking into account many possible objections (refutations)". Is strong dialogical embedding in the scientific community supposed
to be constitutive of or criterial for scientific success as such, or is it rather that
such embedding has in fact proved methodologicallY fruitful (perhaps even
methodologically indispensable) in helping to cull acceptable from unacceptable theories? Do the relevant "support links" and "objections" have the
form of scientific statements, e.g., expressions of agreement and disagreement
regarding the result in question, or are they, as the parenthetical remarks
would have it, "confirmations" and "refutations", that is, the obseroations and
experimental outcomes, standing in normative evidential relationships to that ostensible result, that may be reporled in publications or at congresses and colloquia?
The ambiguities are striking. How Beller would resolve them remains unclear. What I suspect, however, is she would be inclined to regard her thesis
that "the rationality of science is dialogical and communicative" as both normative and constitutive rather than as merely phenomenologically descriptive-or even as trivially correct, for, after all, in light of Wittgenstein's compelling conclusions regarding "private languages", how could any form of
rationality fail to be "dialogical and communicative"? And if this is so, then I
would argue that the affinities of her "dialogism" with social constructivism,
whether in the positivist guise endorsed by Frank or the postmodernist ver-

292

Comments

sion advocated by Rorty, are much closer than Beller would evidently like to
admit.

References
RUSSELL, BERTRAND (1991), History of Western Philosopf?y. London: Routledge
& Kegan Paul [First published in 1946].