4 0 4 D O C . 5 2 G E O M E T R Y A N D E X P E R I E N C E
spaces is the physical space. He argued that this question is decided by convention: Euclidean geom-
etry will always be preferred on the grounds of its simplicity.
An extensive analysis of Poincaré’s conventionalism can be found in Ben-Menahem 2001. For
Poincaré’s early influence on Einstein, see Holton 1988, pp. 202–207. See Howard 1984 for a discus-
sion of Schlick’s conventionalism and its influence on Einstein’s views on scientific methodology.
Friedman 2001 analyzes the context for Poincaré’s thesis and Einstein’s position toward it. He points
out that unlike in Poincaré’s situation, where a choice between equivalent geometries has to be made,
there is no such choice in the case of Einstein’s use of Riemannian geometry. Hence, Einstein did not
face the same problem of underdetermination of the geometrical model that Poincaré had faced.
[14]In the manuscript, “Deutung” and “Identifizierung” are deleted and replaced by “Äquivalenz.”
[15]Einstein’s reading of Poincaré stresses that only complete theories, not individual theorems, can
be compared with empirical evidence (holism). See Schlick 1915, p. 151, for a similar interpretation.
Howard 1990 argues that this reading is influenced by Duhem 1906. A holistic interpretation of the-
ories can also be found in Planck 1913: “[A theory] normally consists of a whole series of separate
theorems in combination. . . . Therefore, each conclusion of the theory results from the combination
of several of its theorems. Hence, for each failure of the theory there are generally several theorems
that could be held responsible, and almost always different possibilities present themselves that offer
a way out of the problem” (“[Eine Theorie] besteht vielmehr in der Regel aus einer ganzen Reihe von
einzelnen miteinander kombinierten Sätzen. . . . Da mithin eine jede Schlußfolgerung der Theorie aus
dem Zusammenwirken von mehreren Sätzen derselben hervorgeht, so können auch für jeden Mißer-
folg, zu dem die Theorie geführt hat, in der Regel mehrere Sätze verantwortlich gemacht werden, und
es bieten sich fast immer verschiedene Möglichkeiten dar, um den rettenden Ausweg zu gewinnen.”)
See also Fine 1986 regarding Einstein’s holism.
[16]The manuscript has “fundamentale” deleted before “selbständige.”
[17]Einstein might have had in mind Weyl’s unified field theory in this passage, which gave up the
direct connection between the spacetime metric and clocks and rigid bodies, and which he had
criticized on physical grounds (Einstein 1918g [Doc. 8]). Einstein’s ambiguous position toward
Poincaré’s conventionalism was mirrored by his similar stance toward the underdetermination thesis
in general (see Einstein 1919g [Doc. 28], note 8). See also Einstein’s response to Eduard Study’s crit-
icism of Poincaré in his letter to Eduard Study, 25 September 1918 (Vol. 8, Doc. 624).
[18]The manuscript has “wesentlich” deleted after “ruht.”
[19]In the manuscript, this paragraph was originally marked with “a)” and followed by the deleted
paragraph: “b) Wenn zwei Strecken einer und derselben dritten gleich sind, so sind sie unter sich
gleich.”
[20]In the manuscript, “Vorraussetzungen” is followed by the deleted passages “für deren Zutreffen
in der Natur sich gewichtige Gründe aus der physikalischen Erfahrung anführen lassen. Der schärfste
von diesen wollen wir Wir wollen nur ein einziges anführen.”
[21]The argument for the existence of a Riemannian metric from the existence of sharp spectral
lines is also used in Einstein 1918g (Doc. 8) against Weyl 1918a.
[22]After “daß wir,” the manuscript has the deleted passage “die Ergebnisse der Riemann’schen
Geometrie sinnvoll auf die physikalische vierdimensionale Welt übertragen können.”
[23]For Einstein’s attempt to apply general relativity to the exploration of the constitution of ele-
mentary particles, see Einstein 1919a (Doc. 17).
[24]Einstein’s recurring doubts about the applicability of continuum theories to microphysics are
discussed in Stachel 1993b.
[25]Einstein first explored the consequences of general relativity for cosmology in Einstein 1917b
(Vol. 6, Doc. 43).
[26]See Doc. 19, note 72, for Einstein’s argument that a spatially infinite universe requires a van-
ishing mean mass density.
[27]See the discussion of Mach’s principle in Einstein 1918e (Doc. 4), note 5, for the reasoning
behind the claim that inertia is reducible to interactions between masses only in a finite universe.
[28]A well-known example for this is Willem de Sitter; see Vol. 8, the editorial note, “The Einstein-
De Sitter-Weyl-Klein Debate,” pp. 351–357. See also Einstein 1918e (Doc. 4), note 11, for a similar
remark by Einstein.
Previous Page Next Page