Monday, March 4, 2024

Poincare was Five Years ahead of Einstein

It is funny to see historians try to credit Einstein. Here is a 2013 essay I had not seen before:
On some points, such as the principle of relativity or the physical interpretation of the Lorentz transformations, Poincaré’s contributions preceded by at least 5 years those of Einstein’s published in 1905. On the other hand, many of their contributions were practically simultaneous. In 1905 Poincaré published an abridged version of his “Sur la dynamique de l’électron” [3] (which preceded the work of Einstein); the expanded version of the article appeared in 1906 [4].

What are the conceptual differences? According to Darrigol,

Einstein completely eliminated the ether, required that the expression of the laws of physics should be the same in any inertial frame, and introduced a “new kinematics” in which the space and time measured in different inertial systems were all on exactly the same footing. In contrast, Poincaré maintained the ether as a privileged frame of reference in which “true” space and time were defined, while he regarded the space and time measured in other frames as only “apparent.” …Einstein derived the expression of the Lorentz transformation from his two postulates (the relativity principle and the constancy of the velocity of light in a given inertial system), whereas Poincaré obtained these transformations as those that leave the Maxwell–Lorentz equations invariant [1].
These are conceptual differences that have no actual experimental consequences as far as electromagnetism and optics are concerned. As Lorentz commented, the difference is purely epistemological: it concerns the number of conventional and arbitrary elements that one wishes to introduce in the definitions of the basic physical concepts.

Are we then dealing with a case of simultaneous discovery?

No, it was not simultaneous. Poincare was five years ahead of Einstein. Poincare was years ahead of Einstein with the relativity principle, rejection of the aether, local time, synchronizing clocks, interpreting Michelson-Morley, Lorentz group, mass-energy equivalence, four-dimensional spacetime, and relativistic theories of gravity. Einstein's only claim to originality is to certain epistemological differences of no physical significance.

Credit Einstein with those obscure conceptual differences if you want, but there is no mathematical or physical value to any of them. Mostly they consist of mathematical misunderstandings by Einstein and other physicists. For example, there is nothing erroneous about choosing a privileged frame on a symmetric space. It does not break the symmetry. Those who criticize Poincare for occasionally choosing a privileged frame are just mathematically ignorant.

The essay concludes:

As Darrigol suggests, it seems wiser to concede that Lorentz, Poincaré and Einstein all contributed to the emergence of the theory of relativity, that Poincaré and Einstein offered two different versions of the theory, and that Einstein gave form to what today is considered the best one.
No, Einstein's is not considered best today. Nearly everyone prefers the spacetime formulation that Poincare advanced in 1905 and Minkowski perfected in 1907. Einstein was still rejecting it in 1911, and did not even speak positively about it until after that, and never really accepted the geometrical significance.

A lot of historians concede that Lorentz and Poincare had all the mathematics of special relativity, and all the physical consequences, but insist that Einstein had a more modern viewpoint or superior understanding, leading to how we understand the theory today. But that is false. The Poincare-Minkowski geometrical interpretation has been preferred by nearly everyone but Einstein, since 1908.


  1. Dear Roger,

    0. I have not properly studied the history and philosophy of relativity. I just read a few bits here and there, as part of my general reading. Then, recently, I taught myself the special theory of relativity (STR for short) only to the extent that it's a prerequisite for the relativistic quantum mechanics and QED.

    With that said, let me still share my thoughts.

    1. The Lorentz transformations (LT for short) arose from the study of electrodynamics (ED for short). There was a move towards synthesis of apparently disparate areas of physics, and classical ED was the latest achievement in this regard: not just electricity and magnetism but also optics came together in a single theory. (Thermodynamics had become a branch of mechanics.)

    2. As to relativity, I guess (but don't know) that the Galilean transformation (GT) was already characterized *as* a kind of a *relativity* theory. With GT, the PDE (Newton's) law kept the same *form*, but the time coordinate still remained independent of the spatial ones.

    3. Then, given the spirit of synthesis, the following steps would seem feasible:

    3.1. Known facts:

    ED governs charged bodies.

    To make the PDEs of ED keep the same *form* in all frames, LT is required. LT involves space and time transforming into each other. LT reduces to GT in the limit of low relative speeds for the frames (not necessarily the objects).

    3.2. Alternative development 1:

    Take a leap in the spirit of synthesis, and assume that all material bodies, including those which are not charged on the gross scale, are composed of some elementary parts that are themselves charged. Possible because electric charge has two polarities.

    Hence, on the gross scale, if any PDE for any physical phenomenon is to keep its form in all frames at high speeds too, then it has to be compatible with LT. In short, other laws of physics are regarded as emergent from those of ED because ED governs the elementary building blocks of all objects.

    3.3. Alternative development 2:

    All the known dynamical laws are expressed with space and time at the most basic level (whether as domain or as independent variables).

    LT involves space-time transforming into each other.

    Take a leap in the spirit of synthesis, and declare that the space-time transformations involved in LT apply to *form*s of *PDEs* of *all* physical phenomena --- known *and* unknown.


    Modern thinkers present the second alternative as an inevitable development, even an indisputable one. I don't think so, and I don't buy their conclusion.

    IMO, the first alternative is a more guarded and a better approach. Note the emphasis on the ``gross scale,'' on the ``emergent'' phenomena, and the hypothetical / postulatory character of the generalization. (If later on STR was found to work with QM phenomena too, then that was to be regarded as a merely fortuitous finding, but not one that was already included in the generalization. Inductive generalization is subject to change when context changes. RQM should've been treated only as a hypothesis until validated. Later on, it was.)

    For establishing priority, I would say that whoever got the first alternative earlier, is to be given the proper credit for STR.

    If people agree to accept Alternative 2 but merely tentatively, then that's a welcome development. Yet, IMO, priority would still remain with the developer of the first alternative. Reference to maths doesn't settle any basic issue in physics.

    IMO, aether and all are just distractions. IMO, arguments for priority on who geometrized it first, are inapplicable. This is physics.

    If necessary, I will clarify my position later. (I might even improve some words/minor elements of my argument/position!)


  2. Lorentz was closer to Alt dev 1, while Poincare was Alt dev 2. For Einstein it is hard to say, and it is not clear that he understood the difference.

    1. Thanks.

      I really don't have the time to *carefully* look into this entire history... Too much back-log on too many other on-going topics...

      But since you replied, I rapidly looked at the Wiki for ``Lorentz ether theory''; went to the section on ``Lorentz transformation'' therein; took the in-page link for ``Poincare 1905 b'' in the references; and pursued the out-side link which lands at:

      In this paper, on p. 1507, Poincare clearly says:

      ... But that's not all: Lorentz, in the work quoted, found it necessary to complete his hypothesis by assuming that all forces, whatever their origin, are affected by translation in the same way as electromagnetic forces...

      Even if I assume that Lorentz's particular hypothesis wasn't correct in all respects, and further, even if I assume that the subsequent passages in this paper by Poincare also may not be correct (I really don't have the time to look into it all), the quoted passage, by itself, shows that their thinking (Lorentz's and Poincare's) was sufficiently advanced along the lines as indicated by the second para in my above Alt. dev 1.

      If so, it would seem that the priority goes to Lorentz and Poincare. (Poincare's paper is dated 05 June 1905 (perhaps the received date); Einstein's 30 June 1905 (the received date).)

      Of course, making this tentative conclusion firm (and deciding on the priority between Lorentz and Poincare) would take a lot more study. ...Sigh. Just don't have time now, and won't have time for another year or two! (Plus, it's not a topic of very great interest to me, anyway!)

      But within the above-mentioned limitations, the conclusion seems inescapable.

      May be a many months later (*after* my studies of QED are over), I will write a post or a paper as to why I say that considerations as in Alt dev 1 must be regarded as being decisive for settling priority.