A compter du 15 septembre 2023, la Fondation Maurice Allais change de statut et est dénommée Fondation Maurice Allais sous égide de la Fondation Mines Paris

As of September 15, 2023, the Maurice Allais Foundation will change its statute and will be known as the Maurice Allais Foundation under the aegis of the Mines Paris Foundation

The physicist





By Jean-Bernard Deloly

A golden rule: the primacy of experiments

In tandem with the works which won him the 1988 Nobel Prize in Economics, Maurice Allais devoted a great part of his life to Physics.

His aim: better knowledge of its most fundamental laws.

To reach this goal, empirical evidence is paramount: progress is always due to the detection of phenomena which cannot be explained in terms of the current theories, not only because they question them, but also because it gives new information for the construction of new ones .

In terms of broad principles, this has no longer been really disputed for nigh on two centuries. But the concrete reality remains less satisfactory.

  • Since all that occurs in the domain that is most immediately accessible to us (on the macroscopic level, on Earth) has been considered for decades now to have been explained by the classical theories, completed as necessary by relativistic corrections, the pursuit of phenomena potentially able to invalidate the reigning theories has been largely confined to the extreme domains: astrophysics and particle physics.
    Aside from the difficulties – and costs – entailed by conducting effective observations, a major problem is that, even when a new phenomenon is detected, it is non usually easy to reach any decisive deduction from it, a typical example being the problem of “dark matter” (see boxed text “Dark Matter”).
    The chances of bringing to light an experimental fact which unambiguously invalidates the prevailing theories are therefore extremely limited.

Dark matter

It has emerged that there was a difference between the mass of galaxies calculated on the basis of directly observed matter and their mass calculated on the basis of the observation of stellar movement by application of the laws of gravitation: for all the galaxies analysed the latter is some ten times greater than the former.

Is this due to the presence of “dark matter”, or to the fact that the laws of gravitation are inexact (in which case general relativity, which incorporates these laws, is also inexact, with all the conceptual upheavals that this would lead to)? Since the phenomena involved occur millions or even billions of light-years away and involve factors that are known only via the models that have been conceptualized of them, we shall probably have to wait a while yet for the answer to this question.

  • Moreover it is generally accepted that the two new theories of the 20th century – quantum mechanics and relativity theory – although they are conceptually incompatible, have never been experimentally falsified, each in its own field of application. Against this background, therefore, the construction of a unitary theory of physics comes down to a purely mathematical problem: how to construct a theory which incorporates both.
    Although it has mobilized hundreds of researchers, among the most brilliant on the planet, over several decades, this approach today seems to have led to an impasse: see, for instance, Lee Smolin’s remarkable 2006 work The Trouble with Physics: The Rise of String Theory, the Fall of a Science, and what Comes next.
    The fact is that for a new theory to be validated, it must be experimentally tested. It is not enough for it to be compatible with all known experimental facts: it must also provide at least one new prediction concerning an experiment yet to be conducted, or the explanation of an experimental fact which current theories are unable to explain, and to date this does not apply to any of the theories advanced.

But what if this impasse was due to the fact that at least one or other of the theories of relativity and of quantum mechanics was false?

Two notions must be sharply distinguished: the efficacy of a theory (i.e. its ability to account for reality more or less completely, in other words its empirical usefulness), and its exactitude. Relativity and quantum mechanics are indisputably effective theories, which is what justifies their current use.

But efficacy does not involve truth, and the history of the sciences provides examples of theories today recognized as false, or indeed which have entirely collapsed, and yet which were remarkably effective in their day.

For instance the theory of epicycles, despite being based on a fixed Earth, enabled the apparent movements of the planets and eclipses to be predicted with astonishing precision, which is how it came to be the accepted theory for fifteen centuries – until Kepler’s laws deprived it of all utility.

Yet it was a conceptual dead-end, and all research conducted in a context assuming its truth was evidently destined for this reason to prove fruitless.

So what about restoring pride of place to the experimental research?

When Maurice Allais insisted, with that utter independence of mind which is the hallmark of his entire work, on seeking phenomena which cast doubt, as directly as possible, on the current theories, this was exactly the approach he adopted.

To this end, he not only conducted a number of experiments himself, between 1953 and 1960, in the fields of mechanics and optics, but also rescued from oblivion some previous observations, identifying aspects of their findings which had not been recognized by the original researchers: the interferometric observations of Dayton C. Miller at Mount Wilson (1925-1926), and the optical observations of Ernest Esclangon at the Strasbourg Observatory (1926-1927). Both these observations were incompatible with the postulate of the invariance of the speed of light.

The results found were:

  • From observation of the motion of a pendulum: the existence of discrepancies in relation to the laws of classical mechanics, linked to the configuration and movements of the stars, and entirely inexplicable, in view of their order of magnitude, by relativistic corrections:
    – periodic components linked to the motion of the Earth and the Moon (the more significant being the lunar components);
    – a very pronounced anomaly on the occasion of the solar eclipse of 30th June 1954.
    It is noteworthy that this was a historic day, as the apparition of gravitational anomalies during eclipses had been suspected for several decades, thanks in particular to the research of Italian physicist Quirino Majorana. Programmes of observation using gravimeters or inclinometers had been organized, but without result.
    Maurice Allais was therefore the first to evince the existence of deviations from the known laws of mechanics under eclipse conditions, although others had been exploring along the same lines for a long time: this is one of the reasons why the discovery became celebrated and later came to be known as the “Allais Effect”.
  • From the observations of Dayton C. Miller and of Ernest Esclangon as well as optical experiments organized by Maurice Allais himself (sightings on sighting marks and collimators): the existence of similar phenomena (inexplicable periodic components) in the field of optics. 
  • The presence of consistencies, some of them remarkable, between the mechanical and the optical anomalies.

Close analysis of the publications on these issues brings out in full the solidity of Maurice Allais’s research. It also highlights the fact that today, more than fifty years after his experiments and more than three quarters of a century after those of Miller and of Esclangon, no convincing challenge has yet been made to his conclusions nor indeed to those of his predecessors.

What Maurice Allais probably discovered, that was a veritable mine of new phenomena which must be taken into account by any theory that aspires to be unitary.

In the first place, of course, there are all the phenomena which are still to be discovered by reprising and pursuing in greater depth his experiments in the fields of mechanics and optics – which were prematurely cut short.

But Maurice Allais, on a more general level, was also the first scientist to seek out methodically the existence of deviations from the known laws of mechanics and electromagnetism linked to the configurations and movements of the stars.

He was also the first to seek, and bring to light, links between optical anomalies and anomalies in the field of mechanics.

And he was the first to emphasize the crucial importance of long-term observations (see boxed text) – and the first to conduct such observations in a systematic way.

Hence, of all the observations carried out prior to 1930 – after which the ascendancy of the theory of relativity over the scientific community was so overwhelming as to make it practically impossible to call into question the principle of the constancy of the speed of light – the observations of Miller and Esclangon were the only ones spread over a period of about a year: all the others had been one-off observations or at most sequences of measurements spaced over a few days.

They were also the only ones from which it has been possible to conclude that variations in the speed of light do indeed occur – variations which, in the present case, display a significant sidereal diurnal periodic component (23 h 56 min). [1]

The crucial importance of long-term observations…

  • No safe conclusion can be drawn from short-term observations: if anything is observed, it cannot be characterized with enough precision to be interpreted. For instance, the speed variations of some 5 to 10 km/s detected during most of the numerous observations carried out by Michelson and Morley between 1887 and 1930, which were considerably lower than the expected speeds of several hundred km/s corresponding to the movement of the Earth, were systematically interpreted as noise.

  • It was because Miller’s observations were sufficiently numerous and spaced out over time that this “noise” was discovered to include a significant diurnal component and that this diurnal component turned out to be sidereal diurnal (23 h 56 min) rather than solar diurnal (24 h) – a fact which is evidently crucial for its interpretation.
  • And it was because Maurice Allais’s experiments lasted a month that he was able to distinguish, in the anomalies in the precession of a pendulum, the 24 h 50 min lunar diurnal component from a component of approximately 24 h.


In the light of recent observations, there are strong grounds for thinking that this pursuit of deviations from the known laws of mechanics and electromagnetism linked to the configuration and motion of the celestial bodies, which to date has only been the object of marginal research, ought to prove highly fruitful.

Thus not only can the existence of abnormal phenomena under solar eclipse conditions and the impossibility of explaining them in conventional terms be today regarded as established, but (a) they are not limited to anomalies of pendulum movement, nor even to anomalies in the field of mechanics (abnormal deviations of torsion balances have regularly been observed, modifications of the frequency of atomic clocks have been recorded…), and (b), which is especially disconcerting, they do not seem to be limited to solar eclipses as such phenomena are also found during lunar eclipses as well as in planetary alignments (an eclipse is simply one particular alignment of celestial bodies).

One of the most exasperating problems of the present day is our inability to establish with greater exactitude than 10-4 the value of the gravitational constant G, as the only outcome of the last century’s progress in reducing uncertainty ranges has been that to date these uncertainty ranges no longer overlap at all. It is worth noting that the only continuous long-term measurement of G (Gershteyn, 2001, using a torsion pendulum) brought to light, as far as can be judged, a variation in G linked to the movement of the Earth (presence of a significant diurnal sidereal component).

Let us emphasize that all the experiments involved belong to the macroscopic domain, on fixed terrestrial sites, and under environmental conditions displaying no extreme characteristics. While they certainly demand high levels of competence, organizational capacity and motivation, the means to be implemented remain relatively inexpensive in comparison with what is at stake, and in technological terms are highly accessible.

The regularities within the solar system

It is impossible to pass over Maurice Allais’s contribution to our knowledge of a phenomenon identified more than two centuries ago, but which has still not been explained: the existence of striking patterns of regularity within the solar system.

Since the end of the 18th century it had been noticed that the distances between the Sun and the 6 planets then known (Mercury, Venus, Earth, Mars, Jupiter and Saturn) displayed remarkable relations to one another, which led to what is known as the Titius-Bode Law.

It later came to light (L. Gaussin, 1880) that relations of the same kind are found within the satellite system of those planets which have a significant number of satellites (Jupiter, Saturn and Uranus).

Using updated data, Maurice Allais reprised and completed this research, taking account in particular of the density of the central planet or star, which had not previously been done.

This led to a unique law applicable at the same time to the Sun and these three planets.

[1] Reminder: at the end of a sidereal year the Earth is once more in the same position relative to the fixed stars, whereas at the end of a solar year it has returned to its original position relative to the Sun. The sidereal year is therefore one solar day shorter than the solar year and the sidereal day is about 4 minutes shorter than the solar day.