ccording to historian of science Paul Harman, "the mechanical theory of the optical ether established a paradigm for the programme of mechanical explanation." However, until this paradigm was firmly in place, debates raged over the nature of light and the possible mechanisms of its transmission.
Before the wave theory was established as the canonical explanation of optical phenomena, scientists involved in debates over the production and the interpretation of these phenomena could be divided into two groups: emissionists and wave theorists. Emissionists believed light to be a sequence of rapidly moving particles subject to forces exerted by material bodies. Wave theorists, however, thought of light as a spreading disturbance in the omnipresent ether. By the 1830s, most optics-oriented members of the scientific community recognized the power of the wave theory for explaining contemporary experiments; emissionists could boast no such success.
However, as historian Jed Buchwald has pointed out, the rise of the wave theory of light was more complicated than that. Although it is certainly true that waves replaced light particles in this conceptual shift, another, deeper process also occurred. If waves in the ether became new tools of explanation, wave fronts also replaced rays as tools of analysis. In other words, to be considered a competent wave theorist at this time required an understanding not only of light as ethereal disturbance, but also of the nature of rays and their relation to light beams. Specifically, before 1830 many physicists found it difficult to understand how beams, as collections of discrete, countable rays, could be reconciled with waves, and especially wave fronts — an understanding that was crucial to appropriate deployment of the mathematical apparatus that helped make wave theory successful (that is, satisfactorily quantitative).
In emission theory, single light rays could not be polarized; polarized light resulted from sufficient numbers of rays in a given beam being lined up in the same way. However, in wave theory, it is possible to say meaningfully that a ray is polarized. In that case, polarization refers only to the state of the wave front (and to a particular asymmetry in it) and each asymmetry can correspond to only one ray. But because, with wave theory, a beam of light is not considered a collection of rays in the first place, the rays (as we're using them here) have only an analytic significance. For emissionists, polarization refers to collections of items (rays), whereas for wave theorists, the beam and the ray are identical and singular — and the wave front is more important than both.
In France, the Laplacian J. B. Biot constructed a quantitative theory of light based on these emissionist assumptions. His colleague and eventual nemesis, D. F. J. Arago, produced a theory that, Biot claimed, merely reproduced his own results. But, as Buchwald has shown, Biot's accusations were misplaced: Biot simply could not understand that the foundations of Arago's theory were not only different from his emissionist principles, but fundamentally incompatible with them. Arago's protoge, D. F. J. Fresnel, later performed a polarization experiment that, at least according to Arago's polemical point of view, conclusively demonstrated the invalidity of Biot's original position. Nevertheless, Biot retained his tenacious selectionism (emissionism) blinded him to Fresnel's implicit (though major) point — Fresnel's principle of interference (that what appears to be polarized light may be made of several different waves, not all of which are necessarily entirely polarized or unpolarized or even partly polarized) meant that rays cannot be counted; it is not possible to divide light.
From 1830, the central problem in optics was to determine the mechanical quantities comprising the light medium. For example, in order to explain light polarization, Fresnel supposed that the vibrations producing light consisted of both longitudinal and transverse vibrations; in polarized light, longitudinal ones are missing. But this meant that transverse waves had to travel through a fluid medium, which was impossible, as S. D. Poisson, another Laplacian, pointed out. Fresnel responded to this objection with the hypothesis that the ether was rigid. A rigid ether was susceptible to mechanical modeling; one could construct it in pieces.
By contrast, A. L. Cauchy pursued a thesis that assumed an ether with the properties of an elastic solid, but the mechanical structure of this elastic solid ether was open to question. Despite its immense difficulties, Cauchy's theory allowed for the use of differential equations in the description of wave phenomena. This approach appealed to British physicists such as John Herschel, G. B. Airy, W. Whewell, and especially George Green, because it dispensed with the need to explicitly model and define the material microstructure of the light medium.
After 1860, the rapid development of spectroscopy (the study of radiation by dispersing its component frequencies and attaching a relative intensity to each) and refinements in diffraction gratings increased the validity of the wave theory of light. In keeping with the rise of an electromagnetic, rather than strictly mechanical, ontology characteristic of the later nineteenth century, Maxwell's work explained light waves as oscillations in the equilibrium configuration of electric and magnetic fields, and he derived a wave equation for the spread of electromagnetic effects, the speed of whose propagation came remarkably close, he observed, to that of light. Decisive confirmation of the wave theory of light would wait until 1887/88, however, when Heinrich Hertz demonstrated that Maxwell's waves could be reflected, refracted, and polarized (just like light) and that they travel at light's velocity.
Last modified 15 April 2002