A matter of principles. Ernst Cassirer's reading of the theory of relativity as a principle theory

Considerable attention—partially raised by the so-called ‘neo-Lorentzian interpretations’ of SR (Brown, 2005)—has been drawn recently to Einstein’s distinction between principle and constructive theories. Janssen (Janssen, 2000) regards Einstein’s principle-strategy as a ‘physics of desperation’: failing to find constructive models, one might resort to principles that put constraints on possible models (Janssen and Renn, 2007). Howard has shown that Einstein’s possibly most valuable contributions to 20th-Century philosophy of science (Howard, 2005) were deeply rooted in the 19th century tradition of the ‘physics of principles’ (Howard, 2007), as found in Helmholtz (Bevilacqua, 1993) or Poincaré (Giedymin, 1982), possibly via the mediation of Violle or Kleiner (Stachel, 2000). Trace of the constructive/principle opposition can even be found in Lorentz’s work (Frisch, 2005), or in Hilbert’s and Minkowski’s belief in the validity of universal principles that all physical theories must satisfy (Corry, 2003). No attention, however, has been drawn to philosophers’ reception of Einstein’s distinction. In particular in spite of the renewed interest for Cassirer’s interpretation of relativity (Ryckman, 2005; Neuber, 2012), it has not been noticed that his reading of relativity theory is entirely based on his appreciation of Einstein’s principle-strategy theory. I will proceed as follows: (1) Marburg neo-Kantianism appeared to have been inclined early on (Laßwitz, 1890; Cohen, 1896; Cassirer, 1902; Buek, 1904) to exploit the Kantian overtones of 19th century ‘physics of principles’: principles put requirements on experience rather than being its byproduct, a conception that well fitted with the Marburg conception of the a priori (Stadler, 1876; Cohen, 1877; Cohen, 1883; Cohen, 1885; Cohen, 1907), paving the way (Cassirer, 1907) to the idea of the primacy of the validity of logical requirements over the existence of things (Cassirer, 1910; Natorp, 1910; Buek, 1912). (2) Einstein resorted to the same 19th century tradition of the ‘physics of principles’, by comparing the relativity principle to the second law of thermodynamics (Einstein, 1907), before Boltzmann’s statistic interpretation (to Sommerfeld, 1908) insisting thereby in its explanatory deficiency, as well as in its constraining role (Einstein, 1910; Einstein, 1911b; Einstein, 1914b; Einstein, 1914a): having to choose among too many constructive models (embarasse de richesse) Einstein, 1918a, it is better to search for requirements that such models have to satisfy (Einstein, 1917; Einstein, 1918b). (3) When Cassirer turned his attention to Einstein’s theories around 1920 (Cassirer, 1920; Cassirer, 1920/1921; Cassirer, 1921), he precisely emphasized Einstein’s principle-strategy—the ‘physics of principles’ being already part of the ‘Marburg’ conceptual framework. Relativity principles put constraints on physical laws—lawlike statements will not qualify as physical laws unless they satisfy such constraints—, mimicking thereby the neo-Kantian conception of the relation between transcendental and empirical laws. (4) Critical toward the neo-Kantian reading (Einstein, 1924a; Einstein, 1924b; Howard, 2010, cf.), Einstein regarded the principle-strategy as a provisional methodological device, resorting to the constructive strategy in his quest for a unified field theory. (Dongen, 2010). Cassirer on the contrary regarded relativity as the end of an historical evolution from ‘physics of models’ to the ‘physics in principles’ (Cassirer, 1929), the latter playing a relevant role in Cassirer’s later epistemology (Cassirer, 1936; Cassirer, 1937/1938). I will conclude arguing that Cassirer’s strategy towards a neo-Kantian reading of relativity theory does not consist in weakening the geometrical structure that might be considered a priori, as it is usually claimed, but in emphasizing the fact that relativity principles are not mere byproducts of physical laws, but requirements that we put on them (Wigner, 1949).