What (Good) is Historical
Epistemology? Thomas Sturm reflects on a conference on historical epistemology, held at the
MPIWG in July 2008, which brought together historians and philosophers of science.
• AUG 31, 2008 • Thomas Sturm •
• DEPT. I
Philosophical epistemology aims to clarify what knowledge is, whether we possess
any of it, and how we can justify our knowledge claims, including scientific ones.
While epistemology is a strong branch of current philosophy, its universalistic
pretensions have often been criticized. In particular, it has been suggested that
knowledge is situated in contexts (biological, social, historical, material) and that
epistemology cannot afford to ignore these contexts. One such challenge, which has
recently attracted many historians of science, has been named “historical
epistemology”. Yet there are several different versions of this approach. The
conference aimed to clarify and evaluate these in talks and discussions with
internationally leading historians of epistemology and philosophers and historians of
science. The conference attracted over 120 guests from Europe, America, and Asia,
who work in disciplines as diverse as philosophy, history of science, physics, geology,
economics, sociology, psychology, art history, and philology.
The guiding task was to clarify what versions of historical epistemology exist and the
pros and cons each of them presents. What kind of historical enterprise is historical
epistemology? What are its basic assumptions, and what are their rationales?
Moreover, in what sense is such a focus on epistemic categories and practices itself a
form of epistemology (or philosophy of science)? As papers and discussions were
based on studies about specific topics that exemplify or test one or another version of
historical epistemology, the conference covered a wide variety of issues. These
included the historicity of epistemological categories and standards (such as the
replication of experiments in the seventeenth and eighteenth centuries, the relation
between perception and judgment, or different models of explanation and causal
inference); the historicity of epistemic objects, that is, the “birth, life, and death” of
real or apparent objects of research (like phlogiston, the electron, memory, or the
economy); and models of scientific development, which were either guided by a neo-
Kantian framework or tried to deal with alleged cases of incommensurability by
means of theories of concepts from recent cognitive science.

The way the program was organized reflected three versions of historical
epistemology, as they are practiced by researchers at the MPIWG. Each has its own
points of contact to philosophical epistemology and the philosophy of science: (1)
According to Lorraine Daston, historical epistemology raises “the Kantian question
about the preconditions that make thinking this or that idea possible” (1994, 284), but
views these preconditions as thoroughly historical. Thus, not only our knowledge and
evidence changes or grows throughout history, but our understanding of what can
count as knowledge can be historicized as well. (2) Hans-Jörg Rheinberger’s version,
again, focuses strongly upon the material – especially experimental and technological
– conditions under which scientific knowledge develops, and claims that this goes
along with a shift away from studying the cognitive subject’s conceptual grasp of
objects towards a “reflection on the relation between object and concept, which starts
from the object to be known” (2007, 11f. transl. TS). This touches strongly on the
realism/anti-realism debate in the philosophy of science. (3) Jürgen Renn views
historical epistemology as an historically founded theory of long-term developments
of scientific knowledge. This addresses the philosophical issue of scientific progress,
but pursues it in the form of a naturalistic epistemology centered on an empirical
explanation – based on models of cognitive science – of how scientists come to know
certain things.
These versions of historical epistemology are not strictly competitive, but
complementary. They can also overlap. For instance, as became apparent in the
debates over epistemic things, their “lives” are often connected with issues of long-
term scientific developments. That certain objects become interesting for researchers
at some points, and forgotten or completely ignored at others, for reasons that may not
look entirely rational, raises the question as to whether scientific developments
actually do entail – to use Kuhn’s term – “revolutionary” shifts. Likewise, the
question as to whether certain steps in long-term scientific developments are
legitimate cannot be answered independently of what one believes, or what the
relevant actors believed, to be rational procedures.
This has an important consequence. Historical epistemology is destined to involve
second-order considerations: One cannot simply reconstruct the development of
scientific knowledge as such. There must also be a parallel program of reconstructing
what the agents thought were permissible or recommendable steps, or how they
understood such concepts as knowledge, evidence, observation, probability,
objectivity, and proof. This understanding of historical epistemology converges also
with Michael Friedman’s neo-Kantian approach: He argued that in order to solve
Kuhnian problems of revolutionary gaps in scientific developments, during, say, the
Einsteinian revolution, one should study not merely the history of the relevant

research, but also the philosophical frameworks that guided certain important steps in
that revolution.
Several philosophers at the conference, however, tried to establish a different
connection to the history of science. They were all inclined towards naturalism, the
view that epistemology should use the empirical sciences to understand how
knowledge grows and can be improved (and should give up attempting to look for a
priori presuppositions of knowledge). Peter Barker, Michael Heidelberger, Philip
Kitcher, and Sandra Mitchell defended this approach. Kitcher even introduced a new
version of historical epistemology, claiming that philosophers abandon a static view
of knowledge and its justification, adopting instead a dynamic picture of science,
looking to history for reliable methods of revising beliefs. That does not necessarily
imply a subordination of historical epistemology to naturalized epistemology. As
Mitchell noted, the kind of naturalistic epistemologist who accepts that science
changes historically must accept that his or her own naturalistic conceptions of
science also can, or even should, continue to change as well: “If epistemology has a
history, it also has a future.” That, however, invites a further question: Can this
approach explain scientific change without using neokantian assumptions or without
losing all substantive naturalistic ingredients?