D O C . 3 2 1 R E V I E W E L S B A C H 3 2 5
do not find this standpoint natural. Let a physical theory consist of the parts (ele-
ments) A, B, C, D, which together form a logical whole that correctly connects the
pertinent experiments (sensory experiences). Then the tendency is that less than all
four elements, e.g., A, B, D, still say nothing about the experiences, without C; no
more so A, B, C, without D. One is then free to regard three of these elements, e.g.,
A, B, C, as a priori and only D as empirically determined. What always remains
unsatisfactory in this is the arbitrariness of the choice of elements to be designated
as a priori, even disregarding that the theory could be replaced at some point by
another theory that substitutes some of these elements (or all four of them) with
others. One could be of the view, though, that through direct analysis of human rea-
son, or thought, we would be in a position to recognize elements that would have
to be present in any theory. But most researchers would probably agree that we lack
a method for recognizing such elements, even if one were inclined to believe in
their existence. Or should one imagine that the search for a priori elements was a
kind of asymptotic process that advances along with the development of science?
Elsbach is fully aware of these problems and limits epistemology to the analytic
examination of physical science as it presently exists. He rightfully admonishes
this reviewer as well as H. A. Lorentz for having asserted “epistemological as-
pects” in the decision between physical theories where one ought just to have spo-
ken of “logical
arguments.”[7]
In chapter 6, “The Logical Standards of Judgment,” it is set forth with great acu-
men and clarity that one should not set two theories in parallel by comparing their
claims but rather by juxtaposing the arguments that are supposed to support these
claims. This analysis very much deserves to be noticed and taken to heart. Much
has been perpetrated on this point, particularly in assessments of relativity theory
from philosophical points of view.
Chapters 7 and 8 treat Cohen’s and Natorp’s views about time and space. I have
not been able to make much out of Cohen’s elaborations; Natorp’s way of arguing
is transparent, but in my view it bypasses the essentials. Elsbach draws from his ac-
count the following arguments:
1. Wanting to determine the structure of space by measurement is based on cir-
cular reasoning. For, in order to perform the measurement, Euclidean geometry is
already presupposed.
2. If a deviation from Euclidean geometry were established by experiment, this
does not necessarily need to lead to an alteration of the geometrical laws because
the agreement between theory and experience can also be produced by an alteration
of the physical laws.
3. Even if non-Euclidean geometry were introduced, that would be based only
on technical considerations about the simplicity of the calculation, whereby the real
[col. 1690]