Herbert Ives – relativity, the aether, and theory choice


Table of contents

  1. Einstein’s relativity ¶ Einstein’s relativity and Maxwell’s equations (and Newton) ¶ Michelson–Morley (and Larmor and Lorentz and FitzGerald and all) ¶ Einstein’s relativity – where did that come from? ¶ Widespread acceptance
  2. Ives's aether theory ¶ John Bell's electromagnetic SR ¶ Ives and electromagnetic SR
  3. Choosing an explanation ¶ Ives, Einstein, and axioms ¶ Where do postulates fit? ¶ Lorentz and starting-points ¶ So, how do I choose which explanation to use?
  4. Some broader context ¶ Ives and fringes ¶ Distaste for SR, anti-semitic and otherwise ¶ Logic and social aspects of Science ¶ Bibliography

The 1938 Ives-Stilwell experiment is one of the experiments regularly listed as corroborations of Special Relativity, and occasionally repeated at higher precision. What is much less often mentioned is that Herbert Ives saw this measurement as a confirmation of a rival theory to Special Relativity, which he labelled the Larmor–Lorentz theory, and which he developed and advocated for, for the rest of his life, specifically as a competitive alternative to Einstein's relativity.

It's a pity this story isn't better known, because it's a fascinating and instructive one which covers

Ives seems like a peripheral figure in 20th-century physics: his experiment was interesting but his science was wrong. His optics work was influential and mainstream; his sideline on Relativity, however, became slightly cranky, though it remained on the professional side of whatever borderline there is, there. I'd like to rehabilitate him, not to claim that his physics was right, but to claim that the dead-end he explored was an interesting one, historically and pedagogically. Doing that requires giving you a bit of context, so we have to take quite a roundabout route to (what I think is) the payoff.


Pictures, first, though.

Portrait of Herbert Ives (Public domain, via Wikimedia Commons) stilwell 128
Herbert Ives and GR Stilwell

Ives did his work along with ‘G R Stilwell’, who seems to have been Ives's technical assistant at Bell Labs (I don't know his full name, and pictures of him are seriously hard to find – this is a scan from the Bell Laboratories Record of February 1930). I've included pictures of some of this story's other participants as well, since they don't get enough of a look-in in the standard accounts. I should warn you, though, that there is a great deal of facial hair coming up, sometimes in bird-nesting quantities.)

The Ives–Stilwell Experiment

The Ives-Stilwell experiment measures the transverse Doppler effect for light (Ives and Stilwell, 1938). The Doppler effect is the change in frequency of light or sound when the source is moving towards or away from you. This is the well-known sound effect of an ambulance speeding past you; but it also happens for light, which changes colour, becoming bluer or redder, when its source is moving towards or away from you. When the source is passing by you – that is, when the waves are getting to you transverse to the direction of motion – there's no frequency shift.

Unless you're moving very rapidly, that is. Special Relativity says that there is a transverse Doppler effect for light, when the source is moving at speeds near that of light, even though it's much smaller than the effect along the direction of motion – the longitudinal effect. The transverse effect is small enough that it's very hard to measure without being overwhelmed by the much larger longitudinal effect, but in 1938 Ives, and his laboratory assistant Stilwell, devised a variant of this observation, which picks up the same underlying effect in a precisely measurable way. It's now recited as one of the standard corroborations of Special Relativity (SR), and that's right and proper, because the observed behaviour of the apparatus is precisely what SR predicts. Well done, Ives and Stilwell!

What is much less often mentioned is that Ives designed this experiment as a refutation of SR, and for the rest of his life maintained that that's exactly what it did, and that it ruled out SR in favour of his version of what he insisted on referring to as the ‘Larmor-Lorentz’ theory.

So what's going on here: does the Ives–Stilwell experiment cause problems for SR or not? The answer is no, it doesn't, but if we dig into the reasons why not, that digging leads us to some interesting places.

Einstein’s relativity

Before we can really talk about Ives, though, we have to talk about Einstein’s relativity. My goal here is not to give you a serious introduction to relativity – I hope you have at least a passing awareness of the key ideas there – but to give you a compressed overview, which points to the places where Einstein's SR, revolutionary though it certainly was, was continuous with the concerns of the day, and the work of other scientists.

Einstein’s relativity and Maxwell’s equations (and Newton)

The odd thing about Special Relativity... the two odd things... amongst the odd things about Special Relativity (SR), are both its disconnection from what came before, and the swiftness of its acceptance into the mainstream.

Isaac Newton James Clerk Maxwell Albert Einstein
The three giants: Isaac Newton (at the British Library, Paolozzi, after Blake); Maxwell (pre-beard, really!, image from https://www.clerkmaxwellfoundation.org); Einstein (from the fine [Einstein for Beginners book](https://en.wikipedia.org/wiki/Einstein_for_Beginners))

These two statements are true in essence, I maintain, even though neither of them is really true historically.

SR solves a quite specific problem, which was well-known at the time. Maxwell's Equations very successfully describe electricity, magnetism, and light, tying these apparently various things together into a single theory which was one of the towering achievements of 19th century physics. But in doing so they appear to pick out a reference frame which is ‘at rest‘ in some absolute sense.

Newton supposed, and discussed at some length in a scholium near the start of his Principia, that there exists some frame which is in an absolute sense ‘at rest’, and with respect to which other frames were absolutely in motion. It was impossible to pinpoint this frame, and I think even Newton would acknowledge that it was at least very difficult to identify it, but it was necessary for Newton that it exist in some abstract sense.

I may be moving relative to you, but neither of us can go further than that and assert that I or you are actually the one moving, in some absolute sense. It seems vaguely plausible that there is some state of motion which really is ‘at rest with respect to the universe’, or something like that, but it appears to be impossible to determine by experiment when you're actually in that state of (non)motion. The idea seems intuitively necessary but physically redundant.

In 1865, James Clerk Maxwell proposed (Maxwell, 1865) a theory of how light worked, which magnificently unified a collection of separate but related theories. Maxwell's Equations are one of the towering achievements of 19th century theoretical physics.

Michelson–Morley (and Larmor and Lorentz and FitzGerald and all)

Maxwell's Equations seem to afford some support to the idea that there might be a state of absolute rest, since the most natural way of interpreting them, at the time, was that they were vibrations in a space-filling substance called the ‘aether’ (or ‘ether’), which was therefore an excellent candidate for this long-redundant notion of the absolute rest frame. That in turn implied that electricity, magnetism and light would behave differently – measurably differently – if you were moving with respect to the aether.

Albert A Michelson Edward W Morley Joseph Larmor Hendrik Antoon Lorentz George Francis FitzGerald
Albert A Michelson, Edward W Morley, Joseph Larmor, Hendrik Antoon Lorentz and George Francis FitzGerald

The aether had to have odd properties. It had to be very rigid, in order to support the high frequency vibrations of light, predicted by Maxwell's equations and experimentally verified by, for example, Hertz. But it also had to be very tenuous, in order to provide no resistance to the planets in their orbits. Newton's laws, working on the assumption that the planets are orbiting in a vacuum, had worked extremely well so far – logically, in the process of solving a problem for electrical theory, the idea of the aether created a problem for Newtonian gravitation. So does it exist? How do you measure the aether?

If you are on Earth, you are moving in a near-circular orbit around the Sun. That means that, during the course of a year, you are bound to be moving at a variety of different speeds with respect to this aether, and so you should be able to detect these changes in behaviour of light, at different times of year. The famous experiment to measure this was the Michelson–Morley experiment, but there were a number of similar measurements made at the same time. This is a technically difficult measurement to make, but the expected result was within the capabilities of the equipment. And the experiment found... nothing: it appeared that the earth was stationary with respect to the aether at all points in its orbit around the sun. So the earth dragged the aether with it? A nice idea, but firstly, that would be yet another peculiar and otherwise unmotivated property of the aether; and secondly, the resulting shear in the aether would have been detectable in stellar aberration observations, and wasn't.

Although the Michelson–Morley experiment is the best known, other experiments of the time, including the Kennedy–Thorndike and Fizeau experiments provided further puzzling observations, with no satisfactory physical explanations.

One reasonably successful suggestion was made by George Francis FitzGerald (FitzGerald, 1889), and elaborated by Hendrik Lorentz (Lorentz, 1895; Lorentz, 1904) and Joseph Larmor (Larmor, 1900) (and see (Lorentz, Einstein, Minkowski et al., 1952) for reprints of some of these): if we suppose that the aether compresses material objects along their direction of motion, and slows down clocks in motion, by a factor dependent of the speed of the object, then it's possible to square Maxwell's Equations with actual observation. This length contraction, and time dilation, is still referred to as the Lorentz, or FitzGerald–Lorentz contraction, and if one assumes this effect, then one can match at least some of the aether observations. But the Lorentz contraction is largely unmotivated – it's just pulled out of a hat: it seems to work, but doesn't follow from anything very much.

Einstein’s relativity – where did that come from?

This was the perplexing situation in 1905, when Einstein produced his famous paper (Einstein, 1905). This paper summarises the situation very briskly in its first two paragraphs:

It is known that Maxwell's electrodynamics – as usually understood at the present time – when applied to moving bodies, leads to asymmetries which do not appear to be inherent in the phenomena [...]

Examples of this sort, together with the unsuccessful attempts to discover any motion of the earth relatively to the ‘light medium’, suggest that the phenomena of electrodynamics as well as of mechanics possess no properties corresponding to the idea of absolute rest. [...] The introduction of a ‘luminiferous ether’ will prove to be superfluous[...].

That very first sentence of the paper is acknowledging that Maxwell's equations do not fit in with our mechanical intuitions about how things should behave when they're moving. Einstein is here saying that what is wrong might be, not Maxwell's equations, but instead our intuitions. And the second paragraph is saying that, to the extent that these intuitions lead us to suppose the existence of a ‘luminiferous ether’ (and those are Einstein's scare-quotes), those intuitions may be leading us on a wild goose chase.

Einstein is addressing a then-current problem, but he's (very boldly) not making an incremental addition to existing work, but instead starting from scratch somewhere else, and coming at this problem from an unexpected direction. In just a few pages of this paper, he's able to reproduce the relevant results of FitzGerald, Lorentz and Larmor, and show that Maxwell's equations really are consistent with a slightly adjusted version of our Galilean intuitions.

Einstein's unexpected starting point is a pair of postulates which, he implicitly asserts, capture vital features of our universe. Rephrased, these two famous postulates are

  1. You can't tell you're moving (that is, there's no experiment, of any type, that you can do that would let you conclude you are absolutely moving – all you can do is look out of a window and see that you're moving relatively to someone else).
  2. Any experiment will always measure the same speed of light (that is, the measured speed of light doesn't increase or decrease if you're moving towards or away from the source of the light).

The first of these isn't surprising – it's generally regarded as Galileo's Relativity Principle - and if you add to this an axiom such as ‘there is no maximum speed in nature’ or ‘speeds add in the obvious way’, then you can deduce something that is effectively Newton's mechanics, which worked quite acceptably until folk started bringing light into the story.

It's the second axiom that is astonishing, and goes flatly against our intuitions. There are multiple alternatives to the second postulate, including ‘our universe has a maximum speed’. When you follow through the consequences of these ideas, you discover not only the length contraction and time dilation that Larmor and co explained with a rather special-pleading aether theory, but you're obliged to think about time itself in a different way. It's no longer the simple universal clock that served us so well in the centuries since Newton, but now an observer-dependent thing, so that different observers will give different answers to the question ‘what time did X happen at?’

No-one fails to be disoriented at this, and most folk carry on being disoriented, even after they've learned enough to be able to do calculations in this new world. But ‘surprising’ doesn't mean ‘wrong’: Einstein's Special Relativity is correct (we'll save the sophisticated quibbles for later), and the robust physicist’s answer to the exclamation ‘but this goes against my intuition!’ is ‘so... get some better intuition!’

Widespread acceptance

SR was broadly accepted pretty quickly, with luminaries such as Planck, Ehrenfest and Sommerfeld regarding it as ‘established’ by 1911; the rest of the academy naturally followed. General Relativity appeared in 1915, and by 1920 most of the key experimental tests of both Special and General Relativty had been completed.

Portrait of Dayton Miller Georges Sagnac
Dayton Miller and Georges Sagnac (https://www.traces-ecrites.com/auteurs/georges-sagnac/)

The history of SR is more intricate than this (of course), but it remains broadly true that expert acceptance of SR was both widespread and notably swift.

Widespread, but not as complete as many histories might suggest. In the US, Dayton Miller was still claiming in 1933 that his variant of the Michelson–Morley measurement was detecting the aether; in France in the late 1920s, Georges Sagnac insisted that the ‘Sagnac effect’ was doing the same; and a school of German physicists (about whom more later) insisted that there were theoretical inconsistencies with SR. Ives was amongst these hold-outs.

Ives's aether theory

Ives’ biography is not intricate (the account here is based on the brief Wikipedia page and on a detailed discussion of his work in (Lalli, 2013)). Ives spent most of his career as a researcher at AT&T's Bell Labs, and was a recognised authority on optical physics, including a spell as president of the Optical Society of America. He was by no means a fringe figure.

Two things follow from this. First, it illustrates that Ives was very much a lab physicist (and inventor), and although he was not shy of maths, he was aligned with what seems to have been a broad German–British–American school of physics whose style was anchored in practice and in visualisable models. This contrasts with a much more mathematical school of theoretical physics, of which Poincaré might be regarded as the exemplar. Second, it's unsurprising, with this background, that Ives saw the ‘Larmor–Lorentz theory’ as an aspect of electromagnetism; in contrast, Einstein approached the problem from mechanics, and framed his postulates as covering all of physics, with no restrictions or exceptions.

John Bell's electromagnetic SR

The link with electromagnetism is not unreasonable, of course – the problem arose first in that context, and Einstein cited those problems in his 1905 paper. Much later, in 1976, John Bell claimed that electromagnetism might be a more accessible path to teaching SR than the usual approach (this article is (Bell, 1976), but it's usefully reproduced in (Bell, 2004); this article is incidentally more famous as the source for ‘Bell’s spaceship paradox’). In that article, Bell described how the electric field of a moving charge appears to an observer with respect to whom the charge is moving. The field of that charge is compressed along the direction of motion, compared to a stationary charge. The amount of that compression is characterised by a rescaled space coordinate, which depends on the velocity of the particle, and when that is combined with a suitably rescaled time coordinate, the electric field of the moving particle, when expressed in those coordinates, is seen to be the same as that of a stationary one. Bell then argues that the rescaled position and time coordinates are the physically relevant ones for the stationary observer, describing the moving charge, and that if the moving charge were, for example, the electrons in the atoms of a moving observer's body, and in their wristwatch, then the stationary observer would measure the moving observer to be length-contracted, and time dilated, whilst the moving observer would notice no effect at all.

John Bell
John Bell

The ‘rescaling’ here turns out to be precisely the Lorentz transformation that Lorentz and Larmor had derived, and that Einstein had re-derived, and what this indicates is that, if you want to describe a physical object, composed of atoms with electrically charged components, in a way which is fully compatible with Maxwell’s equations, then you must do it using the Lorentz transformation to translate between what you see, and what the moving observer sees. You cannot use the simpler recipe for switching between frames, that you would expect from your study of particle mechanics (a recipe that in this context is referred to as the ‘Galilean transformation’). It must be said that you almost can, and you wouldn't be far wrong for anything moving at less than about 80% of the speed of light, but faster than that, and your calculations give the wrong answers.

Galilean mechanics is therefore wrong, but you'd never notice that unless you were looking at things moving nearly as fast as light. What's the first thing that physicists studied that moved near light speeds? Light! Which is why SR first appeared in the context of electromagnetism.

Another way of putting that is that, if you believe Maxwel's equations (and you most certainly should), then you (and Maxwell) are forced to use the Lorentz transformation when you are talking about high speed moving objects. And two consequences of that are that moving observers can't tell they're moving, and, when you do the relevant calculation, the speed of light that that moving observer measures, in their frame, is the same as that measured by the stationary observer. In other other words, it is possible to frame Einstein's postulates as an indirect consequence of Maxwell's equations.

Bell's claim is that this would be a useful way of teaching SR, in contrast to the more abstract route, which starts with Einstein's postulates and works forward from there. He asserts, reasonably, that this approach, starting with Maxwell and working out from there, demystifies the Lorentz transformation and length contraction. Myself, I'm only partially convinced by the teaching point: Bell's argument depends on some advanced EM theory at a crucial point (deducing ‘the electric field of a moving charge’ is a notoriously non-trivial exercise, and the fact that Bell thinks this is ‘the easy way’ probably tells you quite a lot about Bell's powers and tastes as a theoretician). That means that there would remain a great big ‘it can be shown that’ at the centre of the argument, which makes this approach generally unsuitable for a lecture at the stage when students are most usefully taught SR.

But the general point is a very good one. SR is intimately associated with EM and Maxwell's equations – to the extent that I claim that electromagnetism (via Maxwell's equations) is the first experimental test of SR. If Einstein hadn't existed, one can imagine a counterfactual history in which Larmor, Lorentz and co headed down the path that Bell described, situated the Lorentz transformation in that context, discovered (just as Newton did with his mechanics) that there was no need for a frame of absolute rest (so the idea of the aether became redundant, once again), finally deduced Einstein's postulates, and discovered that you could use those postulates, once turned around, from deductions into foundational statements about our universe, to work upwards from the foundations, hugely productively. That might have taken a couple of decades, but one can imagine the new theory of spacetime easing its way into the professional subconscious, without anyone experiencing the intellectual shock that 1905 presented.

Although he wouldn't put it this way, that alternative route – relativity without the shock – was what Ives was trying to develop.

Ives and electromagnetic SR

If anyone could have managed to bring about this alternative route, it would have been Ives. He had long experience with experimental EM and its numerous counterintuitive features, and he had the mathematical experience to be able to work with Maxwell's equations. In a series of papers (of which (Ives, 1945) is probably the key one), he examined the relationship between EM radiation and a perfectly reflecting particle. In this scenario, the demand that energy and momentum be conserved then obliges him to conclude that mass varies with velocity, and that length and time intervals vary with velocity, also. Quoting the first two paragraphs of this 1945 paper:

The Lorentz transformations were obtained by Lorentz as a succession of ad hoc inventions, to reconcile Maxwell's theory with the results of experiments on moving bodies. By Einstein they were derived after a discussion of the nature of simultaneity, and the adoption of a definition of simultaneity which violates the intuitive and common-sense meaning of that term. [...]

[...]. From [this argument] the Lorentz transformations can be derived. The space and time concepts of Newton and Maxwell are retained without alteration. ((Ives, 1945), emphases in original)

The paper does what Ives promises. He derives the Lorentz transformations as a consequence of the axiomatic inputs he describes. The mathematical conclusions are the same as Einstein's and thus – and this is a key point – Ives's theory is experimentally indistinguishable from Einstein's, as far as non-accelerated motion goes.

This extended Larmor theory – hell, let's just call it the ‘Larmor–Lorentz–Ives’ (LLI) theory – does seem to hang together.

But Ives doesn't stop there. In two subsequent papers ((Ives, 1939; Ives, 1948; see also Ives, 1947)), Ives develops a theory of gravity based on Newtonian gravitation acting on the relativistic mass of the planet corresponding to the speed it would have if it were falling inward, having fallen to its current orbital radius from infinity: ‘We shall assume first that an interferometer stationary in a gravitational field assumes dimensions, and "clock" rate as though in motion ’ (Ives, 1939) (in the 1947 paper he consistently calls this the ‘Lorentzian mass’, of course, but it is the familiar γm). With this assumption, he is able to make a connection to other work (specifically (Phillips, 1922) and (Barajas, Birkhoff, Graef et al., 1944), early in the exploration of GR, but apparently mainstream), identify a factor of $(1-2GM/rc^2)$ in the right place (which will be familar to those with knowledge of the Schwarzschild solution in GR), and from there proceed to the expression for the advance of the perihelion of Mercury, one of the classic tests of GR. It looks like Ives could at this point have claimed to reproduce the results of linearised GR, at least.

The network of papers is intricate, and I'm not confident I've got to the bottom of exactly which idea depends on which other one, and from whom, but the point remains that Ives is able to get first to the Lorentz transformation, and then onwards to the advance in the perihelion of Mercury, without going beyond what he asserts is a newtonian model, and without using arguments he considers to be intuitively unacceptable.

Choosing an explanation

Ives, Einstein, and axioms

What is going on here? Has Ives shown Einstein to be wrong? Or unnecessary? Or – probably the least negative interpretation Ives might accept – right for the wrong reasons?

No. The arguments Ives advances are thought-provoking, and the argument in the Lorentz transformation paper is interesting for demonstrating the tight interrelationships of electromagnetism, energy and momentum conservation, and relativity. The arguments are surprising and impressive: I would not have expected that approach to work. But I don't believe they add anything to our understanding of physics, apart from indirectly emphasising the tight relationships between SR and EM, in something like the pedagogic sense that Bell described.

Ives built his arguments by erecting his postulates in a different place from Einstein. He thought they were in a better place; he was in a minority.

In the 1945 paper mentioned above, Ives starts by acknowledging that Lorentz developed the eponymous transformations by an ad hoc assumption of length contraction (and we might remember that Planck made the breakthrough to the idea of the quantisation of energy, in 1900, by a similarly desperate ad hoc assumption). But he then goes on to make an equivalent assumption, where the (very reasonable) requirement that energy and momentum are conserved forces him to assume that a particle's mass changes depending on its speed. There's no reason for this assumption, other than that the sums don't work out otherwise, and in this sense it has the same logic as Lorentz's and Planck's innovations. This is the crucial postulational input into Ives's theory, just as the second axiom was Einstein's crucial input into SR.

Where do postulates fit?

Let's just stress that for a moment: Lorentz's, Ives's and Einstein's theories are all intelligible in a ‘postulational’ framework. They start with an axiomatic statement – for example, that the speed of light is always measured to have the same value – which is a statement about the universe which could be false (ie, it's not a mathematical statement, true a priori). Theoretical Physics consists of drawing out the consequences of that starting point, until you get to a statement which is experimentally accessible. If what the theory predicts doesn't come to pass, then the axioms are one of the things in the frame, when it comes to looking for explanations (yes, this is a broadly Popperian way of framing this whole enterprise – one can find substantially more sophisticated ways of describing what's happening here).

This change of mass with speed is a consequence of Einstein's SR, and, going in the other direction, the constancy of the speed of light is a consequence of Ives's theory, as would be any of the various alternatives to Einstein's second axiom.

That is why the ‘LLI theory’ and Einstein's SR are arguably the same theory, and why it would be impossible to find any observational or experimental difference between them. This is how Ives’s friend Howard Robertson (of FLRW fame) described Ives’ work:

I have at various times examined with care a number of Ives' theoretical papers attacking Einstein’s theory, and leading to apparent alternatives to the Lorentz equations of transformation. Originally I looked for errors in Ives' deductions, for I considered his postulates to be consistent with special relativity, and I did not see how he could otherwise arrive at conclusions so apparently at variance with the relativity theory. To my surprise I found that in each case his deductions were in fact valid, but that his conclusions were only superficially in contradiction with the relativity theory – their intricacy and formidable appearance were due entirely to Ives' insistence on maintaining an aether framework and mode of expression. Ives had, in fact, set up a theory which was completely equivalent in substance to the special theory of relativity. I sincerely admired his ability to carry through these intricate deductions, in spite of the complications caused by his adherence to the notion of a preferred frame tied to the aether – but was never able to convince him that since what he had was in fact indistinguishable in its predictions from the relativity theory within the domain of physics, it was in fact the same theory. My only present concern is that some who have not penetrated to the essence of Ives' theoretical work have seized upon it as overthrowing the special theory of relativity, and have used it as an argument for a return to outmoded and invalid ways of thought. [Howard Robertson, quoted in (Buckley and Darrow, 1956)]

The same is true of Ives's assumption of newtonian gravitation acting on a ‘Lorentzian mass’. This assumption gives the right answer, and though it is, again, thought-provoking, it's unmotivated – it appears as an assumption out of nowhere. Given that it's true, and would not be true in a purely newtonian gravitational model, it's a logically permissible place to insert your foundational axiom. But why would you choose to start there?

Lorentz and starting-points

Lorentz, I think, provides a key to a productive way of thinking about this. Lorentz and Einstein had a notably warm relationship, though they cordially and respectfully disagreed on the interpretation of relativity (far from being the rivals some might represent them as, Einstein seems to have been rather in awe of Lorentz (Kox, 1993), and Lorentz atypically warm with Einstein). Lorentz is quoted (Kox, 1993) as saying:

Einstein simply postulates what we have deduced, with some difficulty and not altogether satisfactorily, from the fundamental equations of the electromagnetic field. By doing so, he may certainly take credit for making us see in the negative result of experiments like those of Michelson, Rayleigh and Brace, not a fortuitous compensation of opposing effects, but the manifestation of a general and fundamental principle. Yet, I think, something may also be claimed in favour of the form in which I have presented the theory. I cannot but regard the ether, which can be the seat of an electromagnetic field with its energy and its vibrations, as endowed with a certain degree of substantiality, however different it may be from all ordinary matter.

Lorentz makes a strong case, and an entirely legitimate case, that his starting point is more straightforwardly intelligible; Bell would probably agree with him. But is that enough? The phrase ‘fundamental principle’ seems key, here. Lorentz seems to be taking a conventional and still current view in the physical sciences, that an explanation should start – should find its foundations – at the most fundamental level available, and then work ‘up’ towards more complicated, and more experimentally and intuitively accessible, ways of thinking about a question. Lorentz seems to be agreeing that Einstein's starting point is more fundamental, and in that sense preferable in principle, and that the phenomenon of length-contraction, and the ideas which necessitate discussion of an ‘aether’, are details which emerge half-way up the ladder.

Lorentz's assumption of the formula for length-contraction does select the same correct physical theory as Einstein's second postulate, but Lorentz himself here uses the word ‘fortuitous’ in connection with that, in contrast to Einstein's identification of a ‘fundamental principle’, from which Lorentz's formula is one of a number of conclusions which emerge. It's this choice of starting point, and this ‘fortuitous’ property, that makes Lorentz's theory, bluntly, less good than Einstein's.

I would go as far as to speculate that Lorentz would agree that Einstein's starting point is the ‘real’ explanation, even though Lorentz himself would find it more productive, in terms of his intuition and understanding, to think of the physics in terms of an aether.

So, how do I choose which explanation to use?

Are we saying here that ‘Einstein is right and (Larmor–Lorentz–)Ives is wrong’? Mostly yes, but I wouldn't go to the stake for it. All physical theories are attempts to explain the world around us. Some are ruled out because they predict things that are not the case: no, the earth isn't flat, and if you meaure how long it takes a fast-moving muon to decay it will take longer than Galileo would expect. Other theories are abandoned because they are clearly heading in a wrong direction: Ptolemy's epicycles, or caloric, didn't have a decisive experimental refutation, but they ran into the ground when a more productive theory came along.

This LLI theory didn't have time to get much traction before it was replaced in the communal understanding by Einstein's starting-point. But the reason it was abandoned, and forgotten other than as a footnote preface to Einstein's work, is not because it was wrong (as Robertson stressed, it's experimentally indistinguishable from much of SR), but because it was unproductive. Nothing follows from Ives, it sheds little new light on what's happening, and it doesn't create new foundations upon which to build new theories.

Ives showed just how far the aether idea can go – and it can go a lot further than I, for one, thought it could. It also, I suppose, shows that if you're willing to go along with the idea that things are squashed and slowed down by their passage through the aether (because... reasons) then you could potentially develop an intuition for what happens at relativistic speeds which would not end up inconsistent with reality. But that would be like going to the seashore and looking only at the sand, whilst the great ocean of truth lies all undiscovered before you.

When we ask how Science-with-a-capital-‘S’ chooses one theory over another, or changes its mind, we are actually asking three related but distinct questions:

These questions are therefore in the spectrum of questions between ones of scientific logic, and ones of, broadly, the sociology and history of science. Those questions touch on different parts of the scientific enterprise, broadly considered, but they overlap -- they are not disjoint from each other.

There's more we could say about this, but this blogpost is long enough already. So...

Some broader context

One of the reasons Ives is interesting is because of his location in what one might call the intellectual geography of early 20th-century physics.

Ives and fringes

Ives was not a crank. A crank is a crank not because of what they believe, nor how fervently, nor how doughtily they stand up for it it in proud opposition to an oppressive consensus. No, a crank becomes a crank when it appears the belief becomes more important than the reality. Ives took a (very) firm line on what did and didn't count as an acceptable foundation for a physical theory. This isn't to suggest that Ives was merely a dispassionate seeker after truth – he seems to have been motivated by at least some personal animus against Einstein, and in correspondence with Butterfield (about whom, more in a moment) he referred to the ‘destructive part’ of his papers (Lalli, 2013, p.55) – But however convinced be was that SR was a wrong approach, in the ‘constructive part’ of his criticisms of it he mostly stayed within the rules of engagement: for a theory to be wrong, it has to be shown to be inconsistent either with itself or with nature (the ‘mostly’ is important, there: the logic of Ives's objections stayed conventional, but it was clear to his contemporaries that the motivation for his work was a personal animosity that went beyond the professional (Lalli, 2013, p.88–89))..

The first type of such potential inconsistencies is a mathematical one, such as requiring that $0 = 1$, or a violation of a truly foundational idea such as the Second Law of Thermodynamics. The second is where a theory makes a concrete physical prediction which can be straightforwardly tested and agreed to not match the experimental fact. This is where Ives failed. There were no mathematical inconsistencies to be found in SR, and as we've seen above, the LLI theory appears to be experimentally indistinguishable from SR. The only remaining space for an inconsistency is what we regard as the foundational ideas, and what we regard as derived ones, and that is to a large extent a matter of taste. If you want to, you can decide that LLI's notion of the aether gives you more insight, and stop thinking about SR: in doing so, you'll be taking a fringe perspective, and you'll cut yourself off from large chunks of the last 100 years of the development of physics, but you won't be wrong in any absolute (that word again!) sense.

Being in a conceptual minority might sometimes be unproductive, but it's a respectable place to be. One of Einstein's late contributions to quantum mechanics was the thought experiment known as the EPR paradox, very adroitly surfacing one of the puzzles with the interpretation of quantum mechanics. Einstein remained convinced, to the end of his life, that QM remained in some sense deeply mistaken, but his scientific interventions in that line remained fully well-informed, and firmly professional in style and content.

Distaste for SR, anti-semitic and otherwise

Not all fringe perspectives are harmless, though, and we can't ignore Elmore Everest Butterfield. Butterfield was an industrial chemist who provided continued encouragement to Ives, and networking support. He was also, however, a persistent and unconcealed anti-semitic conspiracy theorist, and supporter of the ‘aryan physics’ movement created in Germany from the 1910s, by Philipp Lenard, Johannes Stark and others. Butterfield did not attempt to conceal his motivations from Ives, but although Ives can be convicted of failing to drop almost his only supporter, he seems to have avoided echoing any of Butterfield's language back at him. Though Ives's objections to Einstein clearly arose from a mix of technical and emotional sources, it's not convictably clear that anti-semitism was part of it.

Ives wasn't alone in his distaste for SR. There was an efflorescence of outraged rejections of SR in Germany in the first decades of the 20th century. Nazi anti-semitism clearly provided fertile ground for this, but it also fits in with an outburst of not necessarily political anguish at what seemed to be the undermining of a sturdily practical anglo-german tradition, or style, of physics, by a new ‘theoretical’ style developing elsewhere in Europe. There is a fuller account of the relationship between Ives and Butterfield in (Lalli, 2013, pp.80–90), and discussion of the non-technical motivations for opposition to SR in (Goenner, 1993), and at book length in (Wazeck, 2014).

Logic and social aspects of Science

When it comes down to it, Ives is interesting and important because he is the gateway to a list of counterexamples to the impression one would get when learning SR (and which I think I have given when teaching it), that when Einstein wrote the 1905 paper, everyone went ‘oh! that's clearly the right answer’, and promptly rewrote the textbooks. I think it's still fair to say that, as revolutions go, SR ‘won’ unusually comprehensively and quickly; but there remained technical objections to the idea for decades to come, and particularly strong emotional objections which still persist here and there.

Ives therefore helps remind us – and this is a pretty obvious point, but an important and repeatedly forgotten one – that the history and process of science isn't exclusively logical: one of the mysteries of the scientific endeavour (I'm not going to call it a ‘method’) is that it manages to distil logically supportable conclusions from a process which includes many social components. This historical process is the object of the ‘Sociology of Scientific Knowledge’, which for example (Collins and Pinch, 2012) is a gateway to.

The logic of science does not have a social dimension. In roughly Popperian terms, Science (with a capital S), deems ideas to be true when repeated attempts to disprove them have failed (and all of the words ‘repeatedly’, ‘attempts’, ‘disprove’, and ‘failed’ could be given more texture; you could have a go with ‘true’ and ‘when’, as well). But there’s more to science than just its logic – the questions of when this ‘deeming’ happens, and how that conclusion is arrived at, or rejected, the questions of just how that scientific logic is realised in the actual scientific community, are inevitably social ones, since answering those questions inevitably involves actual people, and their preconceptions.


Norman, 2024 May 12