One of the set reading pieces for the MSc in the History of Science, Technology and Medicine that I am studying is a paper by Peter Galison entitled “Einstein’s Clocks: the Place of Time”. In it, Galison challenges the commonly-held view that Einstein’s “day job” in the Swiss patent office was irrelevant to his role as the originator of the revolutionary Special Theory of Relativity (SR). He points out that, at around the time Einstein was working on his new ideas of space and time, large numbers of public clocks were being electromagnetically synchronised – in his home town of Bern, Switzerland, as well as on some railway lines – and that this might have acted as a stimulus to his thoughts about simultaneity. Furthermore, Einstein had some specialist knowledge of electro-mechanical equipment through his father’s firm, Einstein & Cie, which made electricity meters; and in his job at the patent office he would have seen patent applications for all manner of clock synchronisation devices.
This fascinating idea might appear to be an ideal vehicle for introducing, to students who will not necessarily have degree level physics, some of the basics of a conceptually very difficult theory – by wrapping them up in a nice story about turn-of-the-century Europe, complete with old Prussian generals concerned about the military implications of non-standardised time, and Swiss clockmakers worrying about deteriorating public confidence in public clocks that contradict one another. The point is not whether Galison’s somewhat off-the-wall thesis about public clock synchronisation stimulating Einstein’s development of SR is actually true; it is enough that the paper introduces the concepts. The trouble is that it introduces the wrong concepts.
The first rung of the ladder is there all right, albeit twenty pages in: Einstein, we are told, “aimed for a theory that would start with simple physical principles”. But Galison doesn’t tell us what those principles were, which is a pity, since one of them is the Principle of the Constancy of the Velocity of Light, and it is pretty well impossible to understand SR without it. Physics students learning SR are usually told about the debate which led to this principle – a debate about the medium in which light was thought to travel – the ether – and whether the motion of the earth with respect to the ether could be detected, which it certainly ought to be unless we adopt a somewhat geocentric theory of the cosmos. The pivotal experiment in this respect was the Michelson-Morley Experiment, and although it is commonly cited as a precursor to SR, there is usually an acknowedgement in physics textbooks that the extent to which it influenced Einstein is by no means clear. Nevertheless, whether or not it influenced Einstein, the experiment certainly influenced attitudes to the ether theory among physicists in general. Yet Galison does not mention this experiment; only once, nearly at the end of the paper, does he admit that “the ether couldn’t be measured to first order in the ratio of velocity to the speed of light … [and] it wasn’t there to second order …”
The trouble with not mentioning the Principle of the Constancy of the Velocity of Light is that it is that principle that is responsible for all the weird results about length and time; but the weirdness only comes in when you consider reference frames that are moving relative to one another. Simultaneity at points in the same reference frame is not a problem; synchronising clocks in Bern, Berlin and Paris is not a problem in relativity because these three cities are at rest relative to one another; it is simply an engineering problem.
What worries me is that readers without sufficient grounding in SR will come away with erroneous ideas. Does that matter? After all, the course is intended to teach people, not physics, but the history of science. But what is the history of science for? Granted it’s an interdisciplinary subject – but surely, just because we are working in an academic field that neeeds to be broad enough to embrace both history and science, this does not mean that we should accept a lowering of academic standards on either side of the inter-disciplinary divide? Surely the course shouldn’t be taught in a way that muddles up superficially similar, but actually very distinct, concepts like these? Here I am not necessarily saying that the faults in the science outweigh the interesting points made by Galison about the external influences on Einstein’s thinking – but the paper’s inclusion in the syllabus for this particular course is problematic.
I have a few other gripes about Galison’s paper, and now is a good time to get them off my chest. We can perhaps forgive him for saying that “work, mechanics teaches us, is a product of time and force”; after all, he is paraphrasing clockmaker Albert Favarger here, although he might perhaps have pointed out Favarger’s error, if only to save the reader from wasting time puzzling over it. No, my other criticisms stem from my experience as a railway telecommunications engineer, one of whose many tasks was … errr … the synchronisation of clocks. In that context, I was surprised to read about the electrical synchronisation of Swiss railway clocks as early as the 1890s. Many of the clocks in the British railway network had to wait until the 1980s to achieve this, and while the technology was undoubtedly available a century before that for synchronising clocks in geographically localised areas, the problem about doing this on the railways was the shortage of cable pairs. Until the arrival of reliable, cheap, waterproof, polythene-insulated cables, suitable for lineside locations, around the middle of the 20th century, getting a telecommunications circuit from A to B meant putting it on overhead wires that ran alongside the rails on multi-armed telegraph poles (those of us “of a certain age” remember that you could calculate the speed of a train by counting the number of poles passed in a certain time interval). Cable pairs on pole routes were very scarce and expensive to maintain, and tended to be used for essential services such as emergency telephones and signalling systems; clocks were not that important, at least not on this side of the Channel.
OK, that is not really a criticism, and I see that Galison does give a few references to back up what he says, although none of them look as though they deal specifically with railway clocks. However, I somehow doubt that railway clock synchronisation was achieved as easily or as quickly as he suggests. But the prize gaffe in the whole paper is not far away. Having listed all the sensible reasons for synchronising clocks, Galison offers us this: “Trains often barrelled through the countryside in opposite directions on single tracks, and a shunting error of timekeeping could and did lead to calamity. Remote time regulation … was all that stood between a smooth ride and smoking debris”. Errr, no. What stood between a smooth ride and smoking debris was the signalling system; and signalling systems, like clocks, are basically very simple mechanical or electromechanical networks that had been in existence for a good 50 years before Einstein started thinking about time. A signalman would never, ever, dispatch a train just because it was time for it to depart, and everyone who has travelled on a train surely knows this. Not only the signalman, but the signalling system itself, needs to be satisfied that the line is clear before the signal can be given; the technical term used for this is interlocking. So Galison, as well as possibly leading people up the garden path on the subject of special relativity, is painting a very inaccurate picture of railway safety, which might even worry potential passengers. Shame on him. And all he needed to do was ask an expert!