OK, I get the point. Sean Carroll really doesn’t care about problems of the ontology of personhood in the Many World Interpretation. I figured that, as a physicist, these would not be at the forefront of his mind, which is fair enough. But philosophically they are valid questions – which is why David Lewis thought a fair bit about them in his Model Realism theory. It seems to me that a supposedly scientific theory that walks up and says “Sorry, but you are not you – I can’t say what it is you are, but it’s not what you think you are” is obliged to take questions afterwards. I wrote my article in Aeon to try to get those questions, so determinedly overlooked in many expositions of Many Worlds (though clearly acknowledged, if not really addressed, by one of its thoughtful proponents Lev Vaidman) on the table.
But no. We’re not having that, apparently. Sean Carroll’s response doesn’t even mention them. Perhaps he feels as Chad Orzel does: “Who cares? All that stuff is just a collection of foggily defined emergent phenomena that arising from vast numbers of simple quantum systems. Absent a concrete definition, and most importantly a solid idea of how you would measure any of these things, any argument about theories of mind and selfhood and all that stuff is inescapably incoherent.” I’m sort of hoping that isn’t the case. I’m hoping that when Carroll writes of an experiment on a spin superposition being measured by Alice, “There's a version of Alice who saw up and a version who saw down”, he doesn’t really think we can treat Alice – I mean real-world Alices, not the placeholder for a measuring device – like a CCD camera. It’s the business of physics to simplify, but we know what Einstein said about that.
All he picks up on is the objection that I explicitly call minor in comparison: the matter of testing the MWI. His response baffles me:
"The MWI does not postulate a huge number of unobservable worlds, misleading name notwithstanding. (One reason many of us like to call it “Everettian Quantum Mechanics” instead of “Many-Worlds.”) Now, MWI certainly does predict the existence of a huge number of unobservable worlds. But it doesn’t postulate them. It derives them, from what it does postulate."
(I don’t quite get the discomfort with the “Many Worlds” label. It seems to me that is a reasonable name for a theory that “predicts the existence of a huge number of unobservable worlds.” Still, call it what you will.)
I’m missing something here. By and large, scientific theories make predictions, and then we do experiments to see if those predictions are right. MWI predicts “a huge number of worlds”, but apparently it is unreasonable to ask if we might examine that prediction in the laboratory.
But, Carroll says, “You don’t hold it against a theory if it makes some predictions that can’t be tested. Every theory does that. You don’t object to general relativity because you can’t be absolutely sure that Einstein’s equation was holding true at some particular event a billion light years away.” The latter is a non-sequitur: accepting a prediction that can’t be tested is not the same as accepting the possibility of exceptions. And you might reasonably say that there is a difference between accepting a theory even if you can’t get experimentally at what it implies in some obscure corner of parameter space and accepting a theory that “predicts a huge number of unobservable worlds”, some populated by other versions of you doing unobservable things. But OK, might we then have just one prediction that we can test please?
I was dissatisfied with Carroll’s earlier suggestion that you can test MWI just by finding a system that violates the Schrödinger equation or the principle of superposition, because, as I pointed out, it is not a unique interpretation of quantum theory in that regard. His response? “So what?” Alternatives to MWI, he says, have to add to its postulates (or change them), and so they too should predict something we can test. And some do. I understand that Carroll thinks the MWI is uniquely exempt from having to defend its interpretation in particular in the experimental arena, because its axioms are the minimal ones. The point I wanted to raise in my article, though, was that the wider implications of the MWI make it less minimal than its advocates claim. If a “minimal” physical theory predicted something that seemed nonsensical about how cells work, but a more complex theory with an experimentally unsupported postulate took away that problem, would we be right to assert that the minimal theory must be right until there was some evidence for that other postulate? Of course, there may be a good argument for why trashing any coherent notion of self and identity and agency is not a problem. I’d love to hear it. I’d rather it wasn’t just ignored.
“Those worlds happen automatically” – sure, I see that. They are a prediction – sure, I see that. But this point-blank refusal to think any more about them? I don’t get that. Perhaps if Many Worlders were to stop, just stop, trying to tell us anything about how those many unobservable worlds are peopled, to stop invoking copies of Alice as placeholders for quantum measurements, to stop talking about quantum brothers, to say simply that they don’t really have a clue what their interpretation can mean for our notions of identity, then I would rest easier. And so would many, many other physicists. That, I think, would make them a lot happier than being told they don’t understand quantum theory or that they are being silly.
I’m concerned that this sounds like a shot at Sean Carroll. I really don’t want that. Not only is he a lot smarter than me, but he writes so damned well on such intensely interesting stuff. I’m not saying that to flatter him away. I just wanted to get these things discussed.
Friday, February 20, 2015
Many Worlds - a longer view
Here is the pre-edited version of my article for Aeon on the Many Worlds Interpretation of quantum theory. I’m putting it here not because it is any better than the published version (Aeon’s editing was as excellent and improving as ever), but because it gives me a bit more room to go into some of the issues.
In my article I stood up for philosophy. But that doesn’t mean philosophers necessarily get it right either. In the ensuing discussion I have been directed to a talk by philosopher of science David Wallace. Here he criticizes the Copenhagen view that theories are there to make predictions, not to tell us how the world works. He gets a laugh from his audience for suggesting that, if this were so, scientists would have been forced to ask for funding for the LHC not because of what we’d learn from it but so that we could test the predictions made for it.
This is wrong on so many levels. Contrasting “finding out about the world” against “testing predictions of theories” is a totally false opposition. We obviously test predictions of theories to find out if they do a good job of helping us to explain and understand the world. The hope is that the theories, which are obviously idealizations, will get better and better at predicting the fine details of what we see around us, and thereby enable us to tell ever more complete and satisfying stories about why things are this way (and, of course, to allow us to do some useful stuff for “the relief of man’s estate). So there is a sense in which the justification for the LHC derided by Wallace is in fact completely the right one, although that would have been a very poor way of putting it. Almost no one in science (give or take the [very] odd Nobel laureate who capitalizes Truth like some religious crank) talks about “truth” – they recognize that our theories are simply meant to be good working descriptions of what we see, with predictive value. That makes them “true” not in some eternal Platonic sense but as ways of explaining the world that have more validity than the alternatives. No one considers Newtonian mechanics to be “untrue” because of general relativity. So in this regard, Wallace’s attack on the Copenhagen view is trivial. (I don’t doubt that he could put the case better – it’s just that he didn’t do so here.)
What I really object to is the idea, which Wallace repeats, that Many Worlds is simply “what the theory tells you”. To my mind, a theory tells you something if it predicts the corresponding states – say, the electrical current flowing through a circuit, or the reaction rate of an enzymatic process. Wallace asserts that quantum theory “predicts” a you seeing a live Schrödinger’s cat and a you seeing a dead one. I say, show me the equation where those “yous” appear (along with the universes they are in). The best the MWers can do is to say, well, let’s just denote those things as Ψ(live cat) and Ψ(dead cat), with Ψ representing the corresponding universes. Oh please.
Some objectors to my article have been keen to insist that the MWI really isn’t that bizarre: that the other “yous” don’t do peculiar things but are pretty much just like the you-you. I can see how some, indeed many, of them would be. But there is nothing to exclude those that are not, unless you do so by hand: “Oh, the mind doesn’t work that way, they are still rational beings.” What extraordinary confidence this shows in our ability to understand the rules governing human behaviour and consciousness in more parallel worlds than we can possibly imagine: as if the very laws of physics will make sure we behave properly. Collapsing the wavefunction seems a fairly minor sleight of hand (and moreover one we can actually continue to investigate) compared to that. The truth is that we no nothing about the full range of possibilities that the MWI insists on, and nor can we ever do so.
One of the comments underneath my article – and others will doubtless repeat this – makes the remark that Many Worlds is not really about “many universes branching off” at all. Well, I guess you could choose to believe Anonymous Pete instead of Brian Greene and Max Tegmark, if you wish. Or you could follow his link to Sean Carroll’s article, which is one of the examples I cite in my piece of why MWers simple evade the “self” issue altogether.
But you know, my real motivation for writing my article is not to try to bury the MWI (the day I start imagining I am capable of such things, intellectually or otherwise, is the day to put me out to grass), but to provoke its supporters into actually addressing these issues rather than blithely ignoring them while bleating about the (undoubted) problems with the alternatives. Who knows if it will work.
_____________________________________________________________________
In 2011, participants at a conference on the placid shore of Lake Traunsee in Austria were polled on what the conference was about. You might imagine that this question would have been settled before the meeting was convened – but since the subject was quantum theory, it’s not surprising that there was still much uncertainty. The conference was called “Quantum Physics and the Nature of Reality”, and it grappled with what the theory actually means. The poll, completed by 33 of the participating physicists, mathematicians and philosophers, posed a range of unresolved questions, one of which was “What is your favourite interpretation of quantum mechanics?”
The mere question speaks volumes. Isn’t science supposed to be decided by experiment and observation, free from personal preferences? But experiments in quantum physics have been obstinately silent on what it means. All we can do is develop hunches, intuitions and, yes, favourite ideas.
Which interpretations did these experts favour? There were no fewer than 11 answers to choose from (as well as “other” and “none”). The most popular (42%) was the view put forward by Niels Bohr, Werner Heisenberg and their colleagues in the early days of quantum theory, now known as the Copenhagen Interpretation. In third place (18%) was the Many Worlds Interpretation (MWI).
You might not have heard of most of the alternatives, such as Quantum Bayesianism, Relational Quantum Mechanics, and Objective Collapse (which is not, as you might suppose, saying “what the hell”). Maybe you’ve not heard of the Copenhagen Interpretation either. But the MWI is the one with all the glamour and publicity. Why? Because it tells us that we have multiple selves, living other lives in other universes, quite possibly doing all the things that we dream of but will never achieve (or never dare). Who could resist that idea?
Yet you should. You should resist it not because it is unlikely to be true, or even because, since no one knows how to test it, the idea is not truly scientific at all. Those are valid criticisms, but the main reason you should resist it is that it is not a coherent idea, philosophically or logically. There could be no better contender for Wolfgang Pauli’s famous put-down: it is not even wrong.
Or to put it another way: the MWI is a triumph of canny marketing. That’s not some wicked ploy: no one stands to gain from its success. Rather, its adherents are like giddy lovers, blinded to the flaws beneath the superficial allure.
The measurement problem
To understand how this could happen, we need to see why, more than a hundred years after quantum theory was first conceived, experts are still gathering to debate what it means. Despite such apparently shaky foundations, it is extraordinarily successful. In fact you’d be hard pushed to find a more successful scientific theory. It can predict all kinds of phenomena with amazing precision, from the colours of grass and sky to the transparency of glass, the way enzymes work and how the sun shines.
This is because quantum mechanics, the mathematical formulation of the theory, is largely a technique: a set of procedures for calculating what properties substances have based on the positions and energies of their constituent subatomic particles. The calculations are hard, and for anything more complicated than a hydrogen atom it’s necessary to make simplifications and approximations. But we can do that very reliably. The vast majority of physicists, chemists and engineers who use quantum theory today don’t need to go to conferences on the “nature of reality” – they can do their job perfectly well if, in the famous words of physicist David Mermin, they “shut up and calculate”, and don’t think too hard about what the equations mean.
It’s true that the equations seem to insist on some strange things. They imply that very small entities like atoms and subatomic particles can be in several places at the same time. A single electron can seem to pass through two holes at once, interfering with its own motion as if it was a wave. What’s more, we can’t know everything about a particle at the same time: Heisenberg’s uncertainty principle forbids such perfect knowledge. And two particles can seem to affect one another instantly across immense tracts of space, in apparent (but not actual) violation of Einstein’s theory of special relativity.
But quantum scientists just accept such things. What really divides opinion is that quantum theory seems to do away with the notion, central to science from its beginnings, of an objective reality that we can study “from the outside”, as it were. Quantum mechanics insists that we can’t make a measurement without influencing what we measure. This isn’t a problem of acute sensitivity; it’s more fundamental than that. The most widespread form of quantum maths, devised by Erwin Schrodinger in the 1920s, describes a quantum entity using an abstract concept called a wavefunction. The wavefunction expresses all that can be known about the object. But a wavefunction doesn’t tell you what properties the object has; rather, it enumerates all the possible properties it could have, along with their relative probabilities.
Which of these possibilities is real? Is an electron here or there? Is Schrödinger’s cat alive or dead? We can find out by looking – but quantum mechanics seems to be telling us that the very act of looking forces the universe to make that decision, at random. Before we looked, there were only probabilities.
The Copenhagen Interpretation insists that that’s all there is to it. To ask what state a quantum entity is in before we looked is meaningless. That was what provoked Einstein to complain about God playing dice. He couldn’t abandon the belief that quantum objects, like larger ones we can see and touch, have well defined properties at all times, even if we don’t know what they are. We believe that a cricket ball is red even if we don’t look at it; surely electrons should be no different? This “measurement problem” is at the root of the arguments.
Avoiding the collapse
The way the problem is conventionally expressed is to say that measurement – which really means any interaction of a particle with another system that could be used to deduce its state – “collapses” the wavefunction, extracting a single outcome from the range of probabilities that the wavefunction encodes. But the quantum mechanics offers no prescription for how this collapse occurs; it has to be put in by hand. That’s highly unsatisfactory.
There are various ways of looking at this. A Copenhagenist view might be simply to accept that wavefunction collapse is an additional ingredient of the theory, which we don’t understand. Another view is to suppose that wavefunction collapse isn’t just a mathematical sleight-of-hand but an actual, physical process, a little like radioactive decay of an atom, which could in principle be observed if only we had an experimental technique fast and sensitive enough. That’s the Objective Collapse interpretation, and among its advocates is Roger Penrose, who suspects that the collapse process might involve gravity.
Proponents of the Many Worlds Interpretation are oddly reluctant to admit that their preferred view is simply another option. They often like to insist that There Is No Alternative – that the MWI is the only way of taking quantum theory seriously. It’s surprising, then, that in fact Many Worlders don’t even take their own view seriously enough.
That view was presented in the 1957 doctoral thesis of the American physicist Hugh Everett. He asked why, instead of fretting about the cumbersome nature of wavefunction collapse, we don’t just do away with it. What if this collapse is just an illusion, and all the possibilities announced in the wavefunction have a physical reality? Perhaps when we make a measurement we only see one of those realities, yet the others have a separate existence too.
An existence where? This is where the many worlds come in. Everett himself never used that term, but his proposal was championed in the 1970s by the physicist Bryce De Witt, who argued that the alternative outcomes of the experiment must exist in a parallel reality: another world. You measure the path of an electron, and in this world it seems to go this way, but in another world it went that way.
That requires a parallel, identical apparatus for the electron to traverse. More, it requires a parallel you to measure it. Once begun, this process of fabrication has no end: you have to build an entire parallel universe around that one electron, identical in all respects except where the electron went. You avoid the complication of wavefunction collapse, but at the expense of making another universe. The theory doesn’t exactly predict the other universe in the way that scientific theories usually make predictions. It’s just a deduction from the hypothesis that the other electron path is real too.
This picture really gets extravagant when you appreciate what a measurement is. In one view, any interaction between one quantum entity and another – a photon of light bouncing off an atom – can produce alternative outcomes, and so demands parallel universes. As DeWitt put it, “every quantum transition taking place on every star, in every galaxy, in every remote corner of the universe is splitting our local world on earth into myriads of copies”.
Recall that this profusion is deemed necessary only because we don’t yet understand wavefunction collapse. It’s a way of avoiding the mathematical ungainliness of that lacuna. “If you prefer a simple and purely mathematical theory, then you – like me – are stuck with the many-worlds interpretation,” claims MIT physicist Max Tegmark, one of the most prominent MWI popularizers. That would be easier to swallow if the “mathematical simplicity” were not so cheaply bought. The corollary of Everett’s proposal is that there is in fact just a single wavefunction for the entire universe. The “simple maths” comes from representing this universal wavefunction as a symbol Ψ: allegedly a complete description of everything that is or ever was, including the stuff we don’t yet understand. You might sense some issues being swept under the carpet here.
What about us?
But let’s stick with it. What are these parallel worlds like? This hinges on what exactly the “experiments” that produce or differentiate them are. So you’d think that the Many Worlders would take care to get that straight. But they’re oddly evasive, or maybe just relaxed, about it. Even one of the theory’s most thoughtful supporters, Russian-Israeli physicist Lev Vaidman, seems to dodge the issue in his entry on the MWI in the Stanford Encyclopedia of Philosophy:
“Quantum experiments take place everywhere and very often, not just in physics laboratories: even the irregular blinking of an old fluorescent bulb is a quantum experiment.”
Vaidman stresses that every world has to be formally accessible from the others: it has to be derived from one of the alternatives encoded in the wavefunction of one of the particles. You could say that the universes are in this sense all connected, like stations on the London Underground. So what does this exclude? Nobody knows, and there is no obvious way of finding out.
I put the question directly to Lev: what exactly counts as an experiment? An event qualifies, he replied “if it leads to more than one ‘story’”. He added: “If you toss a coin from your pocket, does it split the world? Say you see tails – is there parallel world with heads?” Well, that was certainly my question. But I was kind of hoping for an answer.
Most popularizers of the MWI are less reticent. In the “multiverse” of the Many Worlds view, says Tegmark, “all possible states exist at every instant”. One can argue about whether that’s the quite same as DeWitt’s version, but either way the result seems to accord with the popular view that everything that is physically possible is realized in one of the parallel universes.
The real problem, however, is that Many Worlders don’t seem keen to think about what this means. No, that’s too kind. They love to think about what it means – but only insofar as it lets them tell us wonderful, lurid and beguiling stories. The MWI seduces us by multiplying our selves beyond measure, giving us fantasy lives in which there is no obvious limit to what we can do. “The act of making a decision”, says Tegmark – a decision here counting as an experiment – “causes a person to split into multiple copies.”
That must be a pretty big deal, right? Not for theoretical physicist Sean Carroll of the California Institute of Technology, whose article “Why the Many-Worlds formulation of quantum mechanics is probably correct” on his popular blog Preposterous Universe makes no mention of these alter egos. Oh, they are there in the background all right – the “copies” of the human observer of a quantum event are casually mentioned in the midst of the 40-page paper by Carroll that his blog cites. But they are nothing compared with the relief of having to fret about wavefunction collapse. It’s as though the burning question about the existence of ghosts is whether they observe the normal laws of mechanics, rather than whether they would radically change our view of our own existence.
But if some Many Worlders are remarkably determined to avert their eyes, others delight in this multiplicity of self. They will contemplate it, however, only insofar as it lets them tell us wonderful, lurid and beguiling stories about fantasy lives in which there is no obvious limit to what we can do, because indeed in some world we’ve already done it.
Most MWI popularizers think they are blowing our minds with this stuff, whereas in fact they are flattering them. They delve into the implications for personhood just far enough to lull us with the uncanniness of the centuries-old Doppelgänger trope, and then flit off again. The result sounds transgressively exciting while familiar enough to be persuasive.
Identity crisis
In what sense are those other copies actually “us”? Brian Greene, another prominent MW advocate, tells us gleefully that “each copy is you.” In other words, you just need to broaden your mind beyond your parochial idea of what “you” means. Each of these individuals has its own consciousness, and so each believes he or she is “you” – but the real “you” is their sum total. Vaidman puts the issue more carefully: all the copies of himself are “Lev Vaidman”, but there’s only one that he can call “me”.
““I” is defined at a particular time by a complete (classical) description of the state of my body and of my brain”, he explains. “At the present moment there are many different “Levs” in different worlds, but it is meaningless to say that now there is another “I”.” Yet it is also scientifically and, I think, logically meaningless to say that there is an “I” at all in his definition, given that we must assume that any “I” is generating copies faster than the speed of thought. A “complete description” of the state of his body and brain never exists.
What’s more, this half-baked stitching together of quantum wavefunctions and the notion of mind leads to a reductio ad absurdum. It makes Lev Vaidman a terrible liar. He is actually a very decent fellow and I don’t want to impugn him, but by his own admission it seems virtually inevitable that “Lev Vaidman” has in other worlds denounced the MWI as a ridiculous fantasy, and has won a Nobel prize for showing, in the face of prevailing opinion, that it is false. (If these scenarios strike you as silly or frivolous, you’re getting the point.) “Lev Vaidman” is probably also a felon, for there is no prescription in the MWI for ruling out a world in which he has killed every physicist who believes in the MWI, or alternatively, every physicist who doesn’t. “OK, those Levs exist – but you should believe me, not them!” he might reply – except that this very belief denies the riposte any meaning.
The difficulties don’t end there. It is extraordinary how attached the MWI advocates are to themselves, as if all the Many Worlds simply have “copies” leading other lives. Vaidman’s neat categorization of “I” and “Lev” works because it sticks to the tidy conceit that the grown-up "I" is being split into ever more "copies" that do different things thereafter. (Not all MWI descriptions will call this copying of selves "splitting" - they say that the copies existed all along - but that doesn't alter the point.)
That isn't, however, what the MWI is really about – it's just a sci-fi scenario derived from it. As Tegmark explains, the MWI is really about all possible states existing at every instant. Some of these, it’s true, must contain essentially indistinguishable Maxes doing and seeing different things. Tegmark waxes lyrical about these: “I feel a strong kinship with parallel Maxes, even though I never get to meet them. They share my values, my feelings, my memories – they’re closer to me than brothers.”
He doesn't trouble his mind about the many, many more almost-Maxes, near-copies with perhaps a gene or two mutated – not to mention the not-much-like Maxes, and so on into a continuum of utterly different beings. Why not? Because you can't make neat ontological statements about them, or embrace them as brothers. They spoil the story, the rotters. They turn it into a story that doesn't make sense, that can't even be told. So they become the mad relatives in the attic. The conceit of “multiple selves” isn’t at all what the MWI, taken at face value, is proposing. On the contrary, it is dismantling the whole notion of selfhood – it is denying any real meaning of “you” at all.
Is that really so different from what we keep hearing from neuroscientists and psychologists – that our comforting notions of selfhood are all just an illusion concocted by the brain to allow us to function? I think it is. There is a gulf between a useful but fragile cognitive construct based on measurable sensory phenomena, and a claim to dissolve all personhood and autonomy because it makes the maths neater. In the Borgesian library of Many Worlds, it seems there can be no fact of the matter about what is or isn’t you, and what you did or didn’t do.
State of mind
Compared with these problems, the difficulty of testing the MWI experimentally (which would seem a requirement of it being truly scientific) is a small matter. ‘It’s trivial to falsify [MWI]’, boasts Carroll: ‘just do an experiment that violates the Schrödinger equation or the principle of superposition, which are the only things the theory assumes.’ But most other interpretations of quantum theory assume them (at least) too – so an experiment like that would rule them all out, and say nothing about the special status of the MWI. No, we’d quite like to see some evidence for those other universes that this particular interpretation uniquely predicts. That’s just what the hypothesis forbids, you say? What a nuisance.
Might this all simply be a habit of a certain sort of mind? The MWI has a striking parallel in analytic philosophy that goes by the name of modal realism. Ever since Gottfried Leibniz argued that the problem of good and evil can be resolved by postulating that ours is the best of all possible worlds, the notion of “possible worlds” has supplied philosophers with a scheme for debating the issue of the necessity or contingency of truths. The American philosopher David Lewis pushed this line of thought to its limits by asserting, in the position called model realism, that all worlds that are possible have a genuine physical existence, albeit isolated causally and spatiotemporally from ours. On what grounds? Largely on the basis that there is no logical reason to deny their existence, but also because accepting this leads to an economy of axioms: you don’t have to explain away their non-existence. Many philosophers regard this as legerdemain, but the similarities with the MWI of quantum theory are clear: the proposition stems not from any empirical motive but simply because it allegedly simplifies matters (after all, it takes only four words to say “everything possible is real”, right?). Tegmark’s so-called Ultimate Ensemble theory – a many-worlds picture not explicitly predicated on quantum principles but still including them – has been interpreted as a mathematical expression of modal realism, since it proposes that all mathematical entities that can be calculated in principle (that is, which are possible in the sense of being “computable”) must be real. Lewis’s modal realism does, however, at least have the virtue that he thought in some detail about the issues of personal identity it raises.
If I call these ideas fantasies, it is not to deride or dismiss them but to keep in view the fact that beneath their apparel of scientific equations or symbolic logic they are acts of imagination, of “just supposing”. Who can object to imagination? Not me. But when taken to the extreme, parallel universes become a kind of nihilism: if you believe everything then you believe nothing. The MWI allows – perhaps insists – not just on our having cosily familial ‘quantum brothers’ but on worlds where gods, magic and miracles exist and where science is inevitably (if rarely) violated by chance breakdowns of the usual statistical regularities of physics.
Certainly, to say that the world(s) surely can’t be that weird is no objection at all; Many Worlders harp on about this complaint precisely because it is so easily dismissed. MWI doesn’t, though, imply that things really are weirder than we thought; it denies us any way of saying anything, because it entails saying (and doing) everything else too, while at the same time removing the “we” who says it. This does not demand broad-mindedness, but rather a blind acceptance of ontological incoherence.
That its supporters refuse to engage in any depth with the questions the MWI poses about the ontology and autonomy of self is lamentable. But this is (speaking as an ex-physicist) very much a physicist’s blind spot: a failure to recognize, or perhaps to care, that problems arising at a level beyond that of the fundamental, abstract theory can be anything more than a minor inconvenience.
If the MWI were supported by some sound science, we would have to deal with it – and to do so with more seriousness than the merry invention of Doppelgängers to measure both quantum states of a photon. But it is not. It is grounded in a half-baked philosophical argument about a preference to simplify the axioms. Until Many Worlders can take seriously the philosophical implications of their vision, it’s not clear why their colleagues, or the rest of us, should demur from the judgement of the philosopher of science Robert Crease that the MWI is ‘one of the most implausible and unrealistic ideas in the history of science’ [see The Quantum Moment, 2014]. To pretend that the only conceptual challenge for a theory that allows everything conceivable to happen (or at best fails to provide any prescription for precluding the possibilities) is to accommodate Sliding Doors scenarios shows a puzzling lacuna in the formidable minds of its advocates. Perhaps they should stop trying to tell us that philosophy is dead.
In my article I stood up for philosophy. But that doesn’t mean philosophers necessarily get it right either. In the ensuing discussion I have been directed to a talk by philosopher of science David Wallace. Here he criticizes the Copenhagen view that theories are there to make predictions, not to tell us how the world works. He gets a laugh from his audience for suggesting that, if this were so, scientists would have been forced to ask for funding for the LHC not because of what we’d learn from it but so that we could test the predictions made for it.
This is wrong on so many levels. Contrasting “finding out about the world” against “testing predictions of theories” is a totally false opposition. We obviously test predictions of theories to find out if they do a good job of helping us to explain and understand the world. The hope is that the theories, which are obviously idealizations, will get better and better at predicting the fine details of what we see around us, and thereby enable us to tell ever more complete and satisfying stories about why things are this way (and, of course, to allow us to do some useful stuff for “the relief of man’s estate). So there is a sense in which the justification for the LHC derided by Wallace is in fact completely the right one, although that would have been a very poor way of putting it. Almost no one in science (give or take the [very] odd Nobel laureate who capitalizes Truth like some religious crank) talks about “truth” – they recognize that our theories are simply meant to be good working descriptions of what we see, with predictive value. That makes them “true” not in some eternal Platonic sense but as ways of explaining the world that have more validity than the alternatives. No one considers Newtonian mechanics to be “untrue” because of general relativity. So in this regard, Wallace’s attack on the Copenhagen view is trivial. (I don’t doubt that he could put the case better – it’s just that he didn’t do so here.)
What I really object to is the idea, which Wallace repeats, that Many Worlds is simply “what the theory tells you”. To my mind, a theory tells you something if it predicts the corresponding states – say, the electrical current flowing through a circuit, or the reaction rate of an enzymatic process. Wallace asserts that quantum theory “predicts” a you seeing a live Schrödinger’s cat and a you seeing a dead one. I say, show me the equation where those “yous” appear (along with the universes they are in). The best the MWers can do is to say, well, let’s just denote those things as Ψ(live cat) and Ψ(dead cat), with Ψ representing the corresponding universes. Oh please.
Some objectors to my article have been keen to insist that the MWI really isn’t that bizarre: that the other “yous” don’t do peculiar things but are pretty much just like the you-you. I can see how some, indeed many, of them would be. But there is nothing to exclude those that are not, unless you do so by hand: “Oh, the mind doesn’t work that way, they are still rational beings.” What extraordinary confidence this shows in our ability to understand the rules governing human behaviour and consciousness in more parallel worlds than we can possibly imagine: as if the very laws of physics will make sure we behave properly. Collapsing the wavefunction seems a fairly minor sleight of hand (and moreover one we can actually continue to investigate) compared to that. The truth is that we no nothing about the full range of possibilities that the MWI insists on, and nor can we ever do so.
One of the comments underneath my article – and others will doubtless repeat this – makes the remark that Many Worlds is not really about “many universes branching off” at all. Well, I guess you could choose to believe Anonymous Pete instead of Brian Greene and Max Tegmark, if you wish. Or you could follow his link to Sean Carroll’s article, which is one of the examples I cite in my piece of why MWers simple evade the “self” issue altogether.
But you know, my real motivation for writing my article is not to try to bury the MWI (the day I start imagining I am capable of such things, intellectually or otherwise, is the day to put me out to grass), but to provoke its supporters into actually addressing these issues rather than blithely ignoring them while bleating about the (undoubted) problems with the alternatives. Who knows if it will work.
_____________________________________________________________________
In 2011, participants at a conference on the placid shore of Lake Traunsee in Austria were polled on what the conference was about. You might imagine that this question would have been settled before the meeting was convened – but since the subject was quantum theory, it’s not surprising that there was still much uncertainty. The conference was called “Quantum Physics and the Nature of Reality”, and it grappled with what the theory actually means. The poll, completed by 33 of the participating physicists, mathematicians and philosophers, posed a range of unresolved questions, one of which was “What is your favourite interpretation of quantum mechanics?”
The mere question speaks volumes. Isn’t science supposed to be decided by experiment and observation, free from personal preferences? But experiments in quantum physics have been obstinately silent on what it means. All we can do is develop hunches, intuitions and, yes, favourite ideas.
Which interpretations did these experts favour? There were no fewer than 11 answers to choose from (as well as “other” and “none”). The most popular (42%) was the view put forward by Niels Bohr, Werner Heisenberg and their colleagues in the early days of quantum theory, now known as the Copenhagen Interpretation. In third place (18%) was the Many Worlds Interpretation (MWI).
You might not have heard of most of the alternatives, such as Quantum Bayesianism, Relational Quantum Mechanics, and Objective Collapse (which is not, as you might suppose, saying “what the hell”). Maybe you’ve not heard of the Copenhagen Interpretation either. But the MWI is the one with all the glamour and publicity. Why? Because it tells us that we have multiple selves, living other lives in other universes, quite possibly doing all the things that we dream of but will never achieve (or never dare). Who could resist that idea?
Yet you should. You should resist it not because it is unlikely to be true, or even because, since no one knows how to test it, the idea is not truly scientific at all. Those are valid criticisms, but the main reason you should resist it is that it is not a coherent idea, philosophically or logically. There could be no better contender for Wolfgang Pauli’s famous put-down: it is not even wrong.
Or to put it another way: the MWI is a triumph of canny marketing. That’s not some wicked ploy: no one stands to gain from its success. Rather, its adherents are like giddy lovers, blinded to the flaws beneath the superficial allure.
The measurement problem
To understand how this could happen, we need to see why, more than a hundred years after quantum theory was first conceived, experts are still gathering to debate what it means. Despite such apparently shaky foundations, it is extraordinarily successful. In fact you’d be hard pushed to find a more successful scientific theory. It can predict all kinds of phenomena with amazing precision, from the colours of grass and sky to the transparency of glass, the way enzymes work and how the sun shines.
This is because quantum mechanics, the mathematical formulation of the theory, is largely a technique: a set of procedures for calculating what properties substances have based on the positions and energies of their constituent subatomic particles. The calculations are hard, and for anything more complicated than a hydrogen atom it’s necessary to make simplifications and approximations. But we can do that very reliably. The vast majority of physicists, chemists and engineers who use quantum theory today don’t need to go to conferences on the “nature of reality” – they can do their job perfectly well if, in the famous words of physicist David Mermin, they “shut up and calculate”, and don’t think too hard about what the equations mean.
It’s true that the equations seem to insist on some strange things. They imply that very small entities like atoms and subatomic particles can be in several places at the same time. A single electron can seem to pass through two holes at once, interfering with its own motion as if it was a wave. What’s more, we can’t know everything about a particle at the same time: Heisenberg’s uncertainty principle forbids such perfect knowledge. And two particles can seem to affect one another instantly across immense tracts of space, in apparent (but not actual) violation of Einstein’s theory of special relativity.
But quantum scientists just accept such things. What really divides opinion is that quantum theory seems to do away with the notion, central to science from its beginnings, of an objective reality that we can study “from the outside”, as it were. Quantum mechanics insists that we can’t make a measurement without influencing what we measure. This isn’t a problem of acute sensitivity; it’s more fundamental than that. The most widespread form of quantum maths, devised by Erwin Schrodinger in the 1920s, describes a quantum entity using an abstract concept called a wavefunction. The wavefunction expresses all that can be known about the object. But a wavefunction doesn’t tell you what properties the object has; rather, it enumerates all the possible properties it could have, along with their relative probabilities.
Which of these possibilities is real? Is an electron here or there? Is Schrödinger’s cat alive or dead? We can find out by looking – but quantum mechanics seems to be telling us that the very act of looking forces the universe to make that decision, at random. Before we looked, there were only probabilities.
The Copenhagen Interpretation insists that that’s all there is to it. To ask what state a quantum entity is in before we looked is meaningless. That was what provoked Einstein to complain about God playing dice. He couldn’t abandon the belief that quantum objects, like larger ones we can see and touch, have well defined properties at all times, even if we don’t know what they are. We believe that a cricket ball is red even if we don’t look at it; surely electrons should be no different? This “measurement problem” is at the root of the arguments.
Avoiding the collapse
The way the problem is conventionally expressed is to say that measurement – which really means any interaction of a particle with another system that could be used to deduce its state – “collapses” the wavefunction, extracting a single outcome from the range of probabilities that the wavefunction encodes. But the quantum mechanics offers no prescription for how this collapse occurs; it has to be put in by hand. That’s highly unsatisfactory.
There are various ways of looking at this. A Copenhagenist view might be simply to accept that wavefunction collapse is an additional ingredient of the theory, which we don’t understand. Another view is to suppose that wavefunction collapse isn’t just a mathematical sleight-of-hand but an actual, physical process, a little like radioactive decay of an atom, which could in principle be observed if only we had an experimental technique fast and sensitive enough. That’s the Objective Collapse interpretation, and among its advocates is Roger Penrose, who suspects that the collapse process might involve gravity.
Proponents of the Many Worlds Interpretation are oddly reluctant to admit that their preferred view is simply another option. They often like to insist that There Is No Alternative – that the MWI is the only way of taking quantum theory seriously. It’s surprising, then, that in fact Many Worlders don’t even take their own view seriously enough.
That view was presented in the 1957 doctoral thesis of the American physicist Hugh Everett. He asked why, instead of fretting about the cumbersome nature of wavefunction collapse, we don’t just do away with it. What if this collapse is just an illusion, and all the possibilities announced in the wavefunction have a physical reality? Perhaps when we make a measurement we only see one of those realities, yet the others have a separate existence too.
An existence where? This is where the many worlds come in. Everett himself never used that term, but his proposal was championed in the 1970s by the physicist Bryce De Witt, who argued that the alternative outcomes of the experiment must exist in a parallel reality: another world. You measure the path of an electron, and in this world it seems to go this way, but in another world it went that way.
That requires a parallel, identical apparatus for the electron to traverse. More, it requires a parallel you to measure it. Once begun, this process of fabrication has no end: you have to build an entire parallel universe around that one electron, identical in all respects except where the electron went. You avoid the complication of wavefunction collapse, but at the expense of making another universe. The theory doesn’t exactly predict the other universe in the way that scientific theories usually make predictions. It’s just a deduction from the hypothesis that the other electron path is real too.
This picture really gets extravagant when you appreciate what a measurement is. In one view, any interaction between one quantum entity and another – a photon of light bouncing off an atom – can produce alternative outcomes, and so demands parallel universes. As DeWitt put it, “every quantum transition taking place on every star, in every galaxy, in every remote corner of the universe is splitting our local world on earth into myriads of copies”.
Recall that this profusion is deemed necessary only because we don’t yet understand wavefunction collapse. It’s a way of avoiding the mathematical ungainliness of that lacuna. “If you prefer a simple and purely mathematical theory, then you – like me – are stuck with the many-worlds interpretation,” claims MIT physicist Max Tegmark, one of the most prominent MWI popularizers. That would be easier to swallow if the “mathematical simplicity” were not so cheaply bought. The corollary of Everett’s proposal is that there is in fact just a single wavefunction for the entire universe. The “simple maths” comes from representing this universal wavefunction as a symbol Ψ: allegedly a complete description of everything that is or ever was, including the stuff we don’t yet understand. You might sense some issues being swept under the carpet here.
What about us?
But let’s stick with it. What are these parallel worlds like? This hinges on what exactly the “experiments” that produce or differentiate them are. So you’d think that the Many Worlders would take care to get that straight. But they’re oddly evasive, or maybe just relaxed, about it. Even one of the theory’s most thoughtful supporters, Russian-Israeli physicist Lev Vaidman, seems to dodge the issue in his entry on the MWI in the Stanford Encyclopedia of Philosophy:
“Quantum experiments take place everywhere and very often, not just in physics laboratories: even the irregular blinking of an old fluorescent bulb is a quantum experiment.”
Vaidman stresses that every world has to be formally accessible from the others: it has to be derived from one of the alternatives encoded in the wavefunction of one of the particles. You could say that the universes are in this sense all connected, like stations on the London Underground. So what does this exclude? Nobody knows, and there is no obvious way of finding out.
I put the question directly to Lev: what exactly counts as an experiment? An event qualifies, he replied “if it leads to more than one ‘story’”. He added: “If you toss a coin from your pocket, does it split the world? Say you see tails – is there parallel world with heads?” Well, that was certainly my question. But I was kind of hoping for an answer.
Most popularizers of the MWI are less reticent. In the “multiverse” of the Many Worlds view, says Tegmark, “all possible states exist at every instant”. One can argue about whether that’s the quite same as DeWitt’s version, but either way the result seems to accord with the popular view that everything that is physically possible is realized in one of the parallel universes.
The real problem, however, is that Many Worlders don’t seem keen to think about what this means. No, that’s too kind. They love to think about what it means – but only insofar as it lets them tell us wonderful, lurid and beguiling stories. The MWI seduces us by multiplying our selves beyond measure, giving us fantasy lives in which there is no obvious limit to what we can do. “The act of making a decision”, says Tegmark – a decision here counting as an experiment – “causes a person to split into multiple copies.”
That must be a pretty big deal, right? Not for theoretical physicist Sean Carroll of the California Institute of Technology, whose article “Why the Many-Worlds formulation of quantum mechanics is probably correct” on his popular blog Preposterous Universe makes no mention of these alter egos. Oh, they are there in the background all right – the “copies” of the human observer of a quantum event are casually mentioned in the midst of the 40-page paper by Carroll that his blog cites. But they are nothing compared with the relief of having to fret about wavefunction collapse. It’s as though the burning question about the existence of ghosts is whether they observe the normal laws of mechanics, rather than whether they would radically change our view of our own existence.
But if some Many Worlders are remarkably determined to avert their eyes, others delight in this multiplicity of self. They will contemplate it, however, only insofar as it lets them tell us wonderful, lurid and beguiling stories about fantasy lives in which there is no obvious limit to what we can do, because indeed in some world we’ve already done it.
Most MWI popularizers think they are blowing our minds with this stuff, whereas in fact they are flattering them. They delve into the implications for personhood just far enough to lull us with the uncanniness of the centuries-old Doppelgänger trope, and then flit off again. The result sounds transgressively exciting while familiar enough to be persuasive.
Identity crisis
In what sense are those other copies actually “us”? Brian Greene, another prominent MW advocate, tells us gleefully that “each copy is you.” In other words, you just need to broaden your mind beyond your parochial idea of what “you” means. Each of these individuals has its own consciousness, and so each believes he or she is “you” – but the real “you” is their sum total. Vaidman puts the issue more carefully: all the copies of himself are “Lev Vaidman”, but there’s only one that he can call “me”.
““I” is defined at a particular time by a complete (classical) description of the state of my body and of my brain”, he explains. “At the present moment there are many different “Levs” in different worlds, but it is meaningless to say that now there is another “I”.” Yet it is also scientifically and, I think, logically meaningless to say that there is an “I” at all in his definition, given that we must assume that any “I” is generating copies faster than the speed of thought. A “complete description” of the state of his body and brain never exists.
What’s more, this half-baked stitching together of quantum wavefunctions and the notion of mind leads to a reductio ad absurdum. It makes Lev Vaidman a terrible liar. He is actually a very decent fellow and I don’t want to impugn him, but by his own admission it seems virtually inevitable that “Lev Vaidman” has in other worlds denounced the MWI as a ridiculous fantasy, and has won a Nobel prize for showing, in the face of prevailing opinion, that it is false. (If these scenarios strike you as silly or frivolous, you’re getting the point.) “Lev Vaidman” is probably also a felon, for there is no prescription in the MWI for ruling out a world in which he has killed every physicist who believes in the MWI, or alternatively, every physicist who doesn’t. “OK, those Levs exist – but you should believe me, not them!” he might reply – except that this very belief denies the riposte any meaning.
The difficulties don’t end there. It is extraordinary how attached the MWI advocates are to themselves, as if all the Many Worlds simply have “copies” leading other lives. Vaidman’s neat categorization of “I” and “Lev” works because it sticks to the tidy conceit that the grown-up "I" is being split into ever more "copies" that do different things thereafter. (Not all MWI descriptions will call this copying of selves "splitting" - they say that the copies existed all along - but that doesn't alter the point.)
That isn't, however, what the MWI is really about – it's just a sci-fi scenario derived from it. As Tegmark explains, the MWI is really about all possible states existing at every instant. Some of these, it’s true, must contain essentially indistinguishable Maxes doing and seeing different things. Tegmark waxes lyrical about these: “I feel a strong kinship with parallel Maxes, even though I never get to meet them. They share my values, my feelings, my memories – they’re closer to me than brothers.”
He doesn't trouble his mind about the many, many more almost-Maxes, near-copies with perhaps a gene or two mutated – not to mention the not-much-like Maxes, and so on into a continuum of utterly different beings. Why not? Because you can't make neat ontological statements about them, or embrace them as brothers. They spoil the story, the rotters. They turn it into a story that doesn't make sense, that can't even be told. So they become the mad relatives in the attic. The conceit of “multiple selves” isn’t at all what the MWI, taken at face value, is proposing. On the contrary, it is dismantling the whole notion of selfhood – it is denying any real meaning of “you” at all.
Is that really so different from what we keep hearing from neuroscientists and psychologists – that our comforting notions of selfhood are all just an illusion concocted by the brain to allow us to function? I think it is. There is a gulf between a useful but fragile cognitive construct based on measurable sensory phenomena, and a claim to dissolve all personhood and autonomy because it makes the maths neater. In the Borgesian library of Many Worlds, it seems there can be no fact of the matter about what is or isn’t you, and what you did or didn’t do.
State of mind
Compared with these problems, the difficulty of testing the MWI experimentally (which would seem a requirement of it being truly scientific) is a small matter. ‘It’s trivial to falsify [MWI]’, boasts Carroll: ‘just do an experiment that violates the Schrödinger equation or the principle of superposition, which are the only things the theory assumes.’ But most other interpretations of quantum theory assume them (at least) too – so an experiment like that would rule them all out, and say nothing about the special status of the MWI. No, we’d quite like to see some evidence for those other universes that this particular interpretation uniquely predicts. That’s just what the hypothesis forbids, you say? What a nuisance.
Might this all simply be a habit of a certain sort of mind? The MWI has a striking parallel in analytic philosophy that goes by the name of modal realism. Ever since Gottfried Leibniz argued that the problem of good and evil can be resolved by postulating that ours is the best of all possible worlds, the notion of “possible worlds” has supplied philosophers with a scheme for debating the issue of the necessity or contingency of truths. The American philosopher David Lewis pushed this line of thought to its limits by asserting, in the position called model realism, that all worlds that are possible have a genuine physical existence, albeit isolated causally and spatiotemporally from ours. On what grounds? Largely on the basis that there is no logical reason to deny their existence, but also because accepting this leads to an economy of axioms: you don’t have to explain away their non-existence. Many philosophers regard this as legerdemain, but the similarities with the MWI of quantum theory are clear: the proposition stems not from any empirical motive but simply because it allegedly simplifies matters (after all, it takes only four words to say “everything possible is real”, right?). Tegmark’s so-called Ultimate Ensemble theory – a many-worlds picture not explicitly predicated on quantum principles but still including them – has been interpreted as a mathematical expression of modal realism, since it proposes that all mathematical entities that can be calculated in principle (that is, which are possible in the sense of being “computable”) must be real. Lewis’s modal realism does, however, at least have the virtue that he thought in some detail about the issues of personal identity it raises.
If I call these ideas fantasies, it is not to deride or dismiss them but to keep in view the fact that beneath their apparel of scientific equations or symbolic logic they are acts of imagination, of “just supposing”. Who can object to imagination? Not me. But when taken to the extreme, parallel universes become a kind of nihilism: if you believe everything then you believe nothing. The MWI allows – perhaps insists – not just on our having cosily familial ‘quantum brothers’ but on worlds where gods, magic and miracles exist and where science is inevitably (if rarely) violated by chance breakdowns of the usual statistical regularities of physics.
Certainly, to say that the world(s) surely can’t be that weird is no objection at all; Many Worlders harp on about this complaint precisely because it is so easily dismissed. MWI doesn’t, though, imply that things really are weirder than we thought; it denies us any way of saying anything, because it entails saying (and doing) everything else too, while at the same time removing the “we” who says it. This does not demand broad-mindedness, but rather a blind acceptance of ontological incoherence.
That its supporters refuse to engage in any depth with the questions the MWI poses about the ontology and autonomy of self is lamentable. But this is (speaking as an ex-physicist) very much a physicist’s blind spot: a failure to recognize, or perhaps to care, that problems arising at a level beyond that of the fundamental, abstract theory can be anything more than a minor inconvenience.
If the MWI were supported by some sound science, we would have to deal with it – and to do so with more seriousness than the merry invention of Doppelgängers to measure both quantum states of a photon. But it is not. It is grounded in a half-baked philosophical argument about a preference to simplify the axioms. Until Many Worlders can take seriously the philosophical implications of their vision, it’s not clear why their colleagues, or the rest of us, should demur from the judgement of the philosopher of science Robert Crease that the MWI is ‘one of the most implausible and unrealistic ideas in the history of science’ [see The Quantum Moment, 2014]. To pretend that the only conceptual challenge for a theory that allows everything conceivable to happen (or at best fails to provide any prescription for precluding the possibilities) is to accommodate Sliding Doors scenarios shows a puzzling lacuna in the formidable minds of its advocates. Perhaps they should stop trying to tell us that philosophy is dead.
Monday, February 16, 2015
General relativity's big year?
For the record, my op-ed in the International New York Times.
______________________________________________________________
You might think that physicists would be satisfied by now. They have been testing Einstein’s theory of general relativity (GR), which explains what gravity is, ever since he first described it one hundred years ago this year. And not once has it been found wanting. But they are still investigating its predictions to the nth decimal place, and this centenary year should see some particularly stringent tests. Perhaps one will uncover the first tiny flaw in this awesome mathematical edifice.
Stranger still is that, although GR is celebrated and revered among physicists like no other theory in science, they would doubtless react with joy if it is proved to fail. That’s science: you produce a smart idea and then test it to breaking point. But this determination to expose flaws isn’t really about skepticism, far less wanton nihilism. Most physicists are already convinced that GR is not the final word on gravity. That’s because the theory, which is applied mostly at the scale of stars and galaxies, doesn’t mesh with quantum theory, the other cornerstone of modern physics, which describes the ultra-small world of atoms and subatomic particles. It’s suspected that underlying both theories is a theory of quantum gravity, from which GR and conventional quantum theory emerge as excellent approximations just as Isaac Newton’s theory of gravity, posed in the late seventeenth century, works fine except in some extreme situations.
The hope is, then, that if we can find some dark corner of the universe where GR fails, perhaps because the gravitational fields it describes are so enormously strong, we might glimpse what extra ingredient is needed – one that might point the way to a theory of quantum gravity.
General relativity was not just the last of Einstein’s truly magnificent ideas, but arguably the greatest of them. His annus mirabilis is usually cited as 1905, when, among other things, he kick-started quantum theory and came up with special relativity, describing the distortion of time and space caused by travelling close to the speed of light. General relativity offered a broader picture, embracing motion that changes speed, such as objects accelerating as they fall in a gravitational field. Einstein explained that gravity can be thought of as curvature induced in the very fabric of time and space by the presence of a mass. This too distorts time: clocks run slower in a strong gravitational field than they do in empty space. That’s one prediction that has now been thoroughly confirmed by the use of extremely accurate clocks on space satellites, and in fact GPS systems have to adjust their clocks to allow for it.
Einstein presented his theory of GR to the Prussian Academy of Sciences in 1915, although it wasn’t officially published until the following year. The theory also predicted that light rays will be bent by strong gravitational fields. In 1919 the British astronomer Arthur Eddington confirmed that idea by making careful observations of the positions of stars whose light passes close to the sun during a total solar eclipse. The discovery assured Einstein as an international celebrity. When he met Charlie Chaplin in 1931, Chaplin is said to have told Einstein that the crowds cheered them both because everyone understood him and no one understood Einstein.
General relativity predicts that some burnt-out stars will collapse under their own gravity. They might become incredibly dense objects called neutron stars only a few miles across, from which a teaspoon of matter would weigh around 10 billion tons. Or they might collapse without limit into a “singularity”: a black hole, from whose immense gravitational field not even light can escape, since the surrounding space is so bent that light just turns back on itself. Many neutron stars have now been seen by astronomers: some, called pulsars, rotate and send out beams of intense radio waves from their magnetic poles, lighthouse beams that flash on an off with precise regularity when seen from afar. Black holes can only be seen indirectly from the X-rays and other radiation emitted by the hot gas that surrounds and is sucked into them. But astrophysicists are certain that they exist.
While Newton’s theory of gravity is mostly good enough to describe the motions of the solar system, it is around very dense objects like pulsars and black holes that GR becomes indispensible. That’s also where it might be possible to test the limits of GR with astronomical investigations. Last year, astronomers at the National Radio Astronomy Observatory in Charlottesville, Virginia, discovered the first pulsar orbited by two other shrunken stars, called white dwarfs. This situation, with two bodies moving in the gravitational field of a third, should allow one of the central pillars of GR, called the strong equivalence principle, to be put to the test by making very detailed measurements of the effects of the white dwarfs on the pulsar’s metronome flashes as they circulate. The team hopes to carry out that study this year.
But the highest-profile test of GR is the search for gravitational waves. The theory predicts that some astrophysical processes involving very massive bodies, such as supernovae (exploding stars) or pulsars orbited by another star (binary pulsars), should excite ripples in space-time that radiate outwards as waves. The first binary pulsar was discovered in 1974, and we now know the two bodies are getting slowly closer at just the rate expected if they are losing energy by radiating gravitational waves.
The real goal, though, is to see such waves directly from the tiny distortions of space that they induce as they ripple past our planet. Gravitational-wave detectors use lasers bouncing off mirrors in two kilometres-long arms at right angles, like an L, to measure such minuscule contractions or stretches. Two of the several gravitational-wave detectors currently built – the American LIGO, with two observatories in Louisiana and Washington, and the European VIRGO in Italy – have just been upgraded to boost their sensitivity, and both will start searching in 2015. The European Space Agency is also launching a pilot mission for a space-based detector, called LISA Pathfinder, this September.
If we’re lucky, then, 2015 could be the year we confirm both the virtues and the limits of GR. But neither will do much to alter the esteem with which it is regarded. The Austrian-Swiss physicist Wolfgang Pauli called it “probably the most beautiful of all existing theories.” Many physicists (including Einstein himself) believed it not so much because of the experimental teats but because of what they perceived as its elegance and simplicity. Anyone working on quantum gravity knows that it is a very hard act to follow.
______________________________________________________________
You might think that physicists would be satisfied by now. They have been testing Einstein’s theory of general relativity (GR), which explains what gravity is, ever since he first described it one hundred years ago this year. And not once has it been found wanting. But they are still investigating its predictions to the nth decimal place, and this centenary year should see some particularly stringent tests. Perhaps one will uncover the first tiny flaw in this awesome mathematical edifice.
Stranger still is that, although GR is celebrated and revered among physicists like no other theory in science, they would doubtless react with joy if it is proved to fail. That’s science: you produce a smart idea and then test it to breaking point. But this determination to expose flaws isn’t really about skepticism, far less wanton nihilism. Most physicists are already convinced that GR is not the final word on gravity. That’s because the theory, which is applied mostly at the scale of stars and galaxies, doesn’t mesh with quantum theory, the other cornerstone of modern physics, which describes the ultra-small world of atoms and subatomic particles. It’s suspected that underlying both theories is a theory of quantum gravity, from which GR and conventional quantum theory emerge as excellent approximations just as Isaac Newton’s theory of gravity, posed in the late seventeenth century, works fine except in some extreme situations.
The hope is, then, that if we can find some dark corner of the universe where GR fails, perhaps because the gravitational fields it describes are so enormously strong, we might glimpse what extra ingredient is needed – one that might point the way to a theory of quantum gravity.
General relativity was not just the last of Einstein’s truly magnificent ideas, but arguably the greatest of them. His annus mirabilis is usually cited as 1905, when, among other things, he kick-started quantum theory and came up with special relativity, describing the distortion of time and space caused by travelling close to the speed of light. General relativity offered a broader picture, embracing motion that changes speed, such as objects accelerating as they fall in a gravitational field. Einstein explained that gravity can be thought of as curvature induced in the very fabric of time and space by the presence of a mass. This too distorts time: clocks run slower in a strong gravitational field than they do in empty space. That’s one prediction that has now been thoroughly confirmed by the use of extremely accurate clocks on space satellites, and in fact GPS systems have to adjust their clocks to allow for it.
Einstein presented his theory of GR to the Prussian Academy of Sciences in 1915, although it wasn’t officially published until the following year. The theory also predicted that light rays will be bent by strong gravitational fields. In 1919 the British astronomer Arthur Eddington confirmed that idea by making careful observations of the positions of stars whose light passes close to the sun during a total solar eclipse. The discovery assured Einstein as an international celebrity. When he met Charlie Chaplin in 1931, Chaplin is said to have told Einstein that the crowds cheered them both because everyone understood him and no one understood Einstein.
General relativity predicts that some burnt-out stars will collapse under their own gravity. They might become incredibly dense objects called neutron stars only a few miles across, from which a teaspoon of matter would weigh around 10 billion tons. Or they might collapse without limit into a “singularity”: a black hole, from whose immense gravitational field not even light can escape, since the surrounding space is so bent that light just turns back on itself. Many neutron stars have now been seen by astronomers: some, called pulsars, rotate and send out beams of intense radio waves from their magnetic poles, lighthouse beams that flash on an off with precise regularity when seen from afar. Black holes can only be seen indirectly from the X-rays and other radiation emitted by the hot gas that surrounds and is sucked into them. But astrophysicists are certain that they exist.
While Newton’s theory of gravity is mostly good enough to describe the motions of the solar system, it is around very dense objects like pulsars and black holes that GR becomes indispensible. That’s also where it might be possible to test the limits of GR with astronomical investigations. Last year, astronomers at the National Radio Astronomy Observatory in Charlottesville, Virginia, discovered the first pulsar orbited by two other shrunken stars, called white dwarfs. This situation, with two bodies moving in the gravitational field of a third, should allow one of the central pillars of GR, called the strong equivalence principle, to be put to the test by making very detailed measurements of the effects of the white dwarfs on the pulsar’s metronome flashes as they circulate. The team hopes to carry out that study this year.
But the highest-profile test of GR is the search for gravitational waves. The theory predicts that some astrophysical processes involving very massive bodies, such as supernovae (exploding stars) or pulsars orbited by another star (binary pulsars), should excite ripples in space-time that radiate outwards as waves. The first binary pulsar was discovered in 1974, and we now know the two bodies are getting slowly closer at just the rate expected if they are losing energy by radiating gravitational waves.
The real goal, though, is to see such waves directly from the tiny distortions of space that they induce as they ripple past our planet. Gravitational-wave detectors use lasers bouncing off mirrors in two kilometres-long arms at right angles, like an L, to measure such minuscule contractions or stretches. Two of the several gravitational-wave detectors currently built – the American LIGO, with two observatories in Louisiana and Washington, and the European VIRGO in Italy – have just been upgraded to boost their sensitivity, and both will start searching in 2015. The European Space Agency is also launching a pilot mission for a space-based detector, called LISA Pathfinder, this September.
If we’re lucky, then, 2015 could be the year we confirm both the virtues and the limits of GR. But neither will do much to alter the esteem with which it is regarded. The Austrian-Swiss physicist Wolfgang Pauli called it “probably the most beautiful of all existing theories.” Many physicists (including Einstein himself) believed it not so much because of the experimental teats but because of what they perceived as its elegance and simplicity. Anyone working on quantum gravity knows that it is a very hard act to follow.
Holding Rome together
Here’s my latest Material Witness column for Nature Materials.
____________________________________________________________________________
Calling it the world’s earliest shopping mall is perhaps a qualified accolade, but Trajan’s Market in Rome is certainly a remarkable structure. These vaulted arcades, built early in the second century AD and perhaps originally administrative offices, have withstood almost two millennia of moderate-scale earthquakes. They aren’t alone in that: the Pantheon, Hadrian’s Mausoleum and the Baths of Diocletian in Rome have all shown comparable longevity and resilience. What is their secret?
The structures use concrete made from the pyroclastic volcanic rock of the region: coarse rubble of tuff and brick bound with a mortar made from volcanic ash. It is this mortar that provides structural stability, but the properties that give it such durability have only now been examined. Jackson et al. [Proc. Natl Acad. Sci. USA 111, 18484 (2014) – here] have reproduced the mortar used by Roman builders and used microdiffraction and tomography to study how it acquires its remarkable cohesion.
The Roman mortar was the result of a century or more of experimentation. It used pozzolan, an aluminosilicate volcanic pumice found in the region of the town of Pozzuoli, near Naples, which, when mixed with slaked lime (calcium hydroxide) in the presence of moisture, recrystallizes into a hydrated cementitious material. Although named for its Roman use, pozzolan has a much longer history in building and remained in use until the introduction of Portland cements in the eighteenth century.
The production of the volcanic ash–lime cement was described by the Roman engineer Vitruvius in his book an architecture from the first century BC, and Jackson et al. followed his recipe to make modern analogues. They found that the tensile strength and fracture energy increased steadily over several months, and used electron microscopy and synchrotron X-ray diffraction to look at the fracture surfaces and the chemical nature of the cementitious phases. Among the poorly crystalline matrix are platey crystals of a calcium aluminosilicate phase called strätinglite, crystallized in situ, which seem to act rather like the steel or glass microfibres added to some cements today to toughen them by providing obstacles to crack propagation. Unlike them, however, strätlingite resists corrosion.
Since the cement industry is a major producer of carbon dioxide liberated during the production of Portland cement, there is considerable interest in finding environmentally friendly alternatives. Some of these have a binding matrix of similar composition to the Roman mortar, and so Jackson et al. suggest that an improved understanding of what makes it so durable could point to approaches worth adopting today – such as using chemical additives that promote the intergrowth of reinforcing platelets.
Of course, the Roman engineers knew of the superior properties of their mortar only by experience. A similar combination of astute empiricism and good fortune lies behind the medieval lime mortars that, because of their slow setting, have preserved some churches and other buildings in the face of structural shifting. They tempt us to celebrate the skills of ancient artisans, but we should also remember that what we see today is selective: time has already levelled the failures.
Monday, January 26, 2015
Secrets of exploding sodium revealed
Here’s the longer version of my latest news story for Nature. I love this stuff. I saw the experiments being done by Phil M when I visited Pavel a couple of years ago, and have been waiting for the work to come together ever since. Could you possibly need any more evidence that chemistry rocks?
____________________________________________________________________________
There’s more than exploding hydrogen in the violence of the reaction of alkali metals with water.
It’s the classic piece of chemical tomfoolery: take a lump of sodium or potassium metal, toss it into water, and watch the explosion. Yet a paper in Nature Chemistry reveals that this familiar piece of pyrotechnics has not previously been understood [1].
The explosion, say Pavel Jungwirth and his collaborators at the Czech Academy of Sciences in Prague, is not merely a consequence of the ignition of the hydrogen gas that the alkali metals release from water. That may happen eventually, but it begins as something far stranger: a rapid exodus of electrons followed by explosion of the metal driven by electrical repulsion.
Neurologist and chemical enthusiast Oliver Sacks offers a vivid account of how, as a boy, he and his friends carried out the reaction on Highgate Ponds in North London with a lump of sodium bought from the local chemicals supplier [2]: “It took fire instantly and sped around and around on the surface like a demented meteor, with a huge sheet of yellow flame above it. We all exulted – this was chemistry with a vengeance.”
Highly reactive sodium and potassium react with water to form sodium hydroxide and hydrogen, and the reaction liberates so much heat that the hydrogen may ignite spontaneously. The process seems so straightforward and understandable that no one previously seems to have felt there was anything else to explain.
But as Jungwirth says, there is a fundamental problem with the conventional explanation. “In order to have a runaway explosive behaviour of a chemical reaction, very good mixing of the reactants needs to be ensured,” he says. But the hydrogen gas and steam released at the surface of the metal should impede the further access of water and quench the reaction. Why doesn’t it?
This, Jungwirth admits, was only a part of the original motivation for looking more deeply into the reaction. The experiments were conducted by his colleague Philip Mason, and he says that “an equally important part is Phil's love for exciting experimentation and the easy availability of our balcony, where the first experiments were carried out.” There Mason set up a high-speed video camera to film the process, although the final movies were shot in the lab of coauthor Sigurd Bauerecker at the Technical University of Braunschweig in Germany.
Despite its notoriously explosive nature, the reaction of sodium with water is in fact extremely erratic: sometimes it explodes and sometimes it doesn’t, largely because of surface oxidation of the metal. “The basic trick Phil came up with is to use liquid metal – a sodium/potassium alloy that is liquid at room temperature”, says Jungwirth. But getting a reliable explosion has its hazards. “A face shield is a must”, he adds. “Phil took it off once to blow out a small fire and a tiny piece of metal exploded into his face: luckily lower part of it, so he only had a few scratches on his cheek.”
The movies revealed a vital clue to what was fueling the violent reaction in the early stages. The reaction starts less than a millisecond after the metal droplet, released from a syringe, enters the water. After just 0.4 ms, “spikes” of metal shoot out from the droplet, too fast to be expelled by heating.
What’s more, between 0.3 and 0.5 ms, this “spiking” droplet becomes surrounded by a dark blue/purple colour in the solution. The reason for these two observations became clear when Jungwirth’s postgraduate student Frank Uhlig carried out quantum-mechanical computer simulations of the process with clusters of just 19 sodium atoms. He found that each of the atoms at the surface of the cluster loses an electron within just several picoseconds (10^-12 s), and these electrons enter the surrounding water where they are solvated (surrounded by water molecules)[3].
Solvated electrons in water are known to have the deep blue colour observed transiently in the videos – although they are highly reactive, quickly decomposing water to hydrogen gas and hydroxide ions. What’s more, their departure leaves the metal cluster full of positively charged ions, which repel each other. The result is a “Coulomb explosion” in which the cluster bursts apart due to its own electrostatic repulsion, a process first explained by Lord Rayleigh in the late nineteenth century.
This explosion creates the spikes known as Taylor cones, the researchers say. They support that idea with less detailed simulations involving clusters of 4,000 sodium atoms, which also break up with spike-like instabilities at the surface.
“Four thousand sodium atoms is still a very tiny piece of matter, and I do not think we see proper Taylor cones in the simulations”, says Jungwirth. “At best, we see a microscopic version.”
Inorganic chemist James Dye of Michigan State University, a specialist on solvated electrons, is full of praise for the work. “I have done the demonstration dozens of times and wondered why sodium globules often danced on the surface, while potassium leads to explosive behaviour”, he says. “The paper gives a complete and interesting account of the early stages of the reaction.”
References
1. Mason, P. E. et al., Nat. Chem. http://dx.doi.org/10.1038/nchem.2161 (2015).
2. Sacks, O. Uncle Tungsten, p.123. Picador, London, 2001.
3. Young, R. M. & Neumark, D. M., Chem. Rev. 112, 5553-5577 (2012).
Friday, January 23, 2015
Are you ready? Then I'll begin...
The beginning of a play or book is so hard. I was reminded of this last night while watching the RSC’s new production in Stratford upon Avon, Oppenheimer. It’s a pretty good play, as I’ll say in my review in Nature soon. But I had first to get over the hump of the opening lines, where Oppenheimer reads from Niels Bohr’s 1934 book Atomic Theory and the Description of Nature: “The task of science is both to extend the range of our experience and to reduce it to order.” It seems an unobjectionable claim, even a rather good one. But as spoken by an actor dressed in period style as Oppenheimer, it seemed a terribly stagey and self-conscious opening. It was as if he were saying “The play’s starting now, and it’s about science, and now you have to believe that I’m Oppenheimer, OK?”
I had the same feeling at the start of Michael Frayn’s Copenhagen when I first saw it years ago. As I recall, the actress playing Margrethe Bohr marched on stage, struck a pose and said “But why?” And I thought “Yeah, yeah, so we are supposed to allow that the play is starting in mid-conversation and to ask ourselves, Why what?” But Copenhagen is brilliant, and so is Frayn, so what’s my problem here?
It’s all about that transition to another reality, and how to make us believe in it. Once Oppenheimer was underway, there was no problem – there was still the odd stagey moment in that production, but on the whole we can get inside the narrative quite comfortably once we are acclimatized. But how do you avoid that awkward instant at the start, where the actors have to say “We’ve started pretending now”?
This matters to me even more with books. I won’t say that I judge them by their first line, but that first line is certainly a hurdle that they have to clear. If it feels as though it has been worked on, burnished, set in place like a jewel for us to admire, then I am off to a bad start. New writers seem to be told that first lines matter a lot, and in a sense they do – but this doesn’t mean that a first line has to strive to be brilliant and lapidary, to compete with the astonishingly over-rated opening lines of Pride and Prejudice or War and Peace. Getting it right with a memorable first line, like Camus in L’Étranger or Dickens in A Christmas Carol, is far more difficult than is generally acknowledged, and more often these attempts just come across as contrived and self-conscious. How much better it is to go for the effortlessly mundane: “Stately, plump Buck Mulligan came from the stairhead, bearing a bowl of lather on which a mirror and a razor lay crossed.” Surely what is far better is that the opening page or so is captivating. If you can create one as jaw-dropping as Dickens in Bleak House, it doesn’t matter what the heck your very first line is.
But theatre: that’s another challenge. Here you’ve got the added problem that there are real people standing in front of you pretending to be different real people, and you know that and they know you know that. So how to start weaving the illusion without a jolt?
One of the best answers I ever saw was in Theatre de Complicité’s Mnemonic, when Simon McBurney just began by talking to us, as the audience. It seemed like a preamble to the start of the play, but gradually we realized that this actually was the play. Arguably that was a trick or gimmick, but it contained a more general solution: don’t try too hard. A Brechtian approach won’t work for every play, but at the very least it seems a good idea to relax and not to feel you have to ensnare the audience from the very first utterance. At that point at least, there’s really no risk we will be bored.
I had the same feeling at the start of Michael Frayn’s Copenhagen when I first saw it years ago. As I recall, the actress playing Margrethe Bohr marched on stage, struck a pose and said “But why?” And I thought “Yeah, yeah, so we are supposed to allow that the play is starting in mid-conversation and to ask ourselves, Why what?” But Copenhagen is brilliant, and so is Frayn, so what’s my problem here?
It’s all about that transition to another reality, and how to make us believe in it. Once Oppenheimer was underway, there was no problem – there was still the odd stagey moment in that production, but on the whole we can get inside the narrative quite comfortably once we are acclimatized. But how do you avoid that awkward instant at the start, where the actors have to say “We’ve started pretending now”?
This matters to me even more with books. I won’t say that I judge them by their first line, but that first line is certainly a hurdle that they have to clear. If it feels as though it has been worked on, burnished, set in place like a jewel for us to admire, then I am off to a bad start. New writers seem to be told that first lines matter a lot, and in a sense they do – but this doesn’t mean that a first line has to strive to be brilliant and lapidary, to compete with the astonishingly over-rated opening lines of Pride and Prejudice or War and Peace. Getting it right with a memorable first line, like Camus in L’Étranger or Dickens in A Christmas Carol, is far more difficult than is generally acknowledged, and more often these attempts just come across as contrived and self-conscious. How much better it is to go for the effortlessly mundane: “Stately, plump Buck Mulligan came from the stairhead, bearing a bowl of lather on which a mirror and a razor lay crossed.” Surely what is far better is that the opening page or so is captivating. If you can create one as jaw-dropping as Dickens in Bleak House, it doesn’t matter what the heck your very first line is.
But theatre: that’s another challenge. Here you’ve got the added problem that there are real people standing in front of you pretending to be different real people, and you know that and they know you know that. So how to start weaving the illusion without a jolt?
One of the best answers I ever saw was in Theatre de Complicité’s Mnemonic, when Simon McBurney just began by talking to us, as the audience. It seemed like a preamble to the start of the play, but gradually we realized that this actually was the play. Arguably that was a trick or gimmick, but it contained a more general solution: don’t try too hard. A Brechtian approach won’t work for every play, but at the very least it seems a good idea to relax and not to feel you have to ensnare the audience from the very first utterance. At that point at least, there’s really no risk we will be bored.
Wednesday, January 21, 2015
Beyond relativity
Sorry folks: Prospect has asked that my latest piece in the February issue, a survey of the centenary year of general relativity, remains "premium content" - which means I can offer only a teaser here. (Cue debate about paywalls and blogs - but we've all got to survive...) I'll be putting up some more on this topic soon, though.
____________________________________________________________________
One hundred years ago, Albert Einstein presented a paper to the Prussian Academy of Sciences that explained gravity. It is one of the four fundamental forces in the universe, although in 1915 only one of the others – the electromagnetic force – was known. (The other two act inside the atomic nucleus.) But Einstein’s paper offered a radically different way of thinking about gravity. Rather than being an invisible force between two massive objects, he described it a distortion induced by the masses in the very fabric of time and space (spacetime). This warping dictates the paths that objects take under gravity’s influence: Newton’s apple fell to earth because it was, in effect, slipping down the slope of bent spacetime. In the curved space around the sun, the planets execute orbits rather like marbles running around the rim of a bowl.
This geometric interpretation of gravity is the central idea of Einstein’s theory of general relativity. It is widely considered to be not only his greatest intellectual achievement but also the epitome of a beautiful theory. Ernest Rutherford said that the theory “cannot but be regarded as a magnificent work of art”, and Einstein was not shy of advertising its virtues himself: “Scarcely anyone who fully understands this theory can escape from its magic”, he wrote.
But the centenary celebrations for general relativity will not simply be looking back. For 2015 will be a banner year for some big, ambitious experiments that aim to probe the theory. They are looking for one of the most spectacular of the theory’s predictions: ripples in spacetime called gravitational waves...
[The rest will be on Prospect's site very shortly.]
____________________________________________________________________
One hundred years ago, Albert Einstein presented a paper to the Prussian Academy of Sciences that explained gravity. It is one of the four fundamental forces in the universe, although in 1915 only one of the others – the electromagnetic force – was known. (The other two act inside the atomic nucleus.) But Einstein’s paper offered a radically different way of thinking about gravity. Rather than being an invisible force between two massive objects, he described it a distortion induced by the masses in the very fabric of time and space (spacetime). This warping dictates the paths that objects take under gravity’s influence: Newton’s apple fell to earth because it was, in effect, slipping down the slope of bent spacetime. In the curved space around the sun, the planets execute orbits rather like marbles running around the rim of a bowl.
This geometric interpretation of gravity is the central idea of Einstein’s theory of general relativity. It is widely considered to be not only his greatest intellectual achievement but also the epitome of a beautiful theory. Ernest Rutherford said that the theory “cannot but be regarded as a magnificent work of art”, and Einstein was not shy of advertising its virtues himself: “Scarcely anyone who fully understands this theory can escape from its magic”, he wrote.
But the centenary celebrations for general relativity will not simply be looking back. For 2015 will be a banner year for some big, ambitious experiments that aim to probe the theory. They are looking for one of the most spectacular of the theory’s predictions: ripples in spacetime called gravitational waves...
[The rest will be on Prospect's site very shortly.]
Thursday, January 08, 2015
Computer becomes "unbeatable" at poker
Here's a longer version of my story on Nature News on the new poker-playing computer program.
_________________________________________________________________
A computer algorithm has perfected the art of playing a popular version of the gambling card game
A new computer algorithm can play poker, in one of its most popular variants, essentially perfectly. Its creators say that it is virtually “incapable of losing against any opponent in a fair game.”
This is a step beyond a computer program that can beat top human players, as Deep Blue famously did against Garry Kasparov in chess in 1997. The poker program devised by Michael Bowling and colleagues at the University of Alberta in Edmonton, Canada, along with Finnish software developer Oskari Tammelin, plays perfectly, to all intents and purposes. This means that this particular variant of poker, called Heads-up Limit Hold’em (HULHE), can be considered “solved”. The algorithm is described in a paper in Science [1].
The strategy they have computed, says computer poker researcher Eric Jackson in Menlo Park, California, is so close to perfect “as to render pointless further work on this game.”
“I think that it will come as a surprise to experts that a game this big has been solved”, says Jackson. “I follow the work on computer poker closely and I did not expect that heads-up limit was going to be solved this soon.”
A few other popular games have been solved before, including checkers, which a team from the same computer science department at Alberta (including Neil Burch, coauthor of the new study) cracked in 2007 [2].
But poker is harder to solve than checkers. Chess and checkers are examples of perfect-information games, where players can have perfect knowledge of all past events in a game. In poker, there are some things a player doesn’t know: most crucially, which cards the other player has been dealt.
Devising a strategy to deal perfectly with that uncertainty is very hard. This hidden information is what gives a poker player the opportunity to bluff: to face the other player down with a relatively weak hand.
But although bluffing looks like a very human, psychological element of the game, it’s not. You can calculate how to bluff optimally. “Bluffing falls out of the mathematics of the game”, says Bowling. If you’re dealt a jack, say, it is possible to figure out how often you should ideally bluff with it. Some of the early pioneers of game theory, such as John von Neumann, aimed to develop mathematical strategies for bluffing.
The real challenge for a poker algorithm is dealing with the immense number of possible ways the game can be played. Bowling and colleagues have looked at one of the most popular forms, called Texas Hold’em, in which the dealer deals two cards to each player (face down) and also a set of face-up “community cards”. Players can bet, raise or fold after each deal. With just two players, the game becomes Heads-up, and it is a “limit” game when it has fixed bet sizes and a fixed number of raises. There are 3.16×10^17 states that HULHE can reach, and 3.19×10^14 possible points where a player must make a decision.
The new algorithm involves calculating all possible decisions in advance, so that they can just be looked up as a game proceeds. This is done in a learning process: the strategy begins by making decisions randomly, and is then updated by experience as the algorithm attaches a “regret” value to each decision depending on how poorly it fared. It takes a little more than 1500 training rounds to make the program essentially invincible.
This “counterfactual regret minimization” (CFR) procedure has been widely adopted in the Annual Computer Poker Competition, which has run since 2006. But Bowling and colleagues have now improved the procedure by allowing it to re-evaluate decisions considered to be poor in earlier training rounds.
The other crucial innovation was the handling of the vast amounts of information needing to be stored to develop and use the strategy – of the order of 262 terabytes. This volume of data demands disk storage, which is slow to access. The researchers figured out a data-compression method that reduces the volume to a more manageable 11 TB, and which adds only 5% to the computation time from the use of disk storage.
“I think the counterfactual regret algorithm is the major advance”, says computer scientist Jonathan Shapiro of the University of Manchester. “But they have done several other very clever things to make this problem computational feasible.”
Although the new algorithm plays “perfectly”, it is not necessarily unbeatable, since there is a large chance element in poker: it depends on the hand you’re dealt. The algorithm “may in fact lose after any finite number of hands, if it were unlucky”, says Bowling. But it always wins in the long run.
What’s more, there are two other limitations. It only “weakly” solves HULHE, meaning that it plays perfectly only if the strategy has been played throughout, and not if the computer player is suddenly dropped into the middle of a game (a situation that is in fact not really clearly defined for poker, as it is say for checkers).
And it is only what the researchers call “essentially solved”, meaning that it is not strictly unbeatable: there is an extremely small margin by which, in theory, it might be beaten by skill rather than chance. But this margin is negligible in practice. “Even if a human could identify the perfect counter-strategy to exploit our solution”, says Bowling, “and even if they could play this counter-strategy without error, and even if they spent 70 years only playing poker (over 60 million hands), they still couldn’t be statistically confident that they are winning [by superior play rather than chance].”
There has been heated debate about the use of computers (poker bots) for online poker games. “For quite a while now there has been a struggle between people who write bots and try to get them to play surreptitiously on online sites, and the sites who try to detect them and ban them”, says Jackson. But Bowling says that their algorithm will have little direct effect on this, because the popularity of online HULHE has declined as human players got better. “While it may dry up the last vestige of online HULHE play just due to the perception that humans can query a perfect strategy, this impact is going to minimal”, he says. But it may respark a philosophical discussion about bots in online poker.”
“I definitely think the ideas in this paper will be fruitfully applied to other forms of poker, such as no-limit”, says Jackson. “And more generally to other games, whether with dice or cards or whatever, that have imperfect information.”
But the stakes might be higher still for imperfect-information “games” beyond mere play and gambling. The stock market seems an obvious candidate, but Bowling explains that there the unknowns are too great. “The deck of cards in the stock market, which define the distribution at chance events, is not common knowledge and not even really knowable.”
He says that it might be useful, however, for portfolio management. “We’ve been investigating robust decision-making where the goal is to optimize a particular risk measure (such as value-at-risk)”, he says. “Such robust decision-making scenarios can often be cast as a game having an almost identical form to some of our poker games, with our solution techniques be immediately applicable.” The team’s current explorations beyond poker are, however, focused on supporting medical decision-making, in collaboration with diabetes specialists.
1. Bowling, M., Burch, N., Johanson, M. & Tammelin, O. Science 347, 145-149 (2015).
2. Schaeffer, J. et al., Science 317, 1518-1522 (2007).
_________________________________________________________________
A computer algorithm has perfected the art of playing a popular version of the gambling card game
A new computer algorithm can play poker, in one of its most popular variants, essentially perfectly. Its creators say that it is virtually “incapable of losing against any opponent in a fair game.”
This is a step beyond a computer program that can beat top human players, as Deep Blue famously did against Garry Kasparov in chess in 1997. The poker program devised by Michael Bowling and colleagues at the University of Alberta in Edmonton, Canada, along with Finnish software developer Oskari Tammelin, plays perfectly, to all intents and purposes. This means that this particular variant of poker, called Heads-up Limit Hold’em (HULHE), can be considered “solved”. The algorithm is described in a paper in Science [1].
The strategy they have computed, says computer poker researcher Eric Jackson in Menlo Park, California, is so close to perfect “as to render pointless further work on this game.”
“I think that it will come as a surprise to experts that a game this big has been solved”, says Jackson. “I follow the work on computer poker closely and I did not expect that heads-up limit was going to be solved this soon.”
A few other popular games have been solved before, including checkers, which a team from the same computer science department at Alberta (including Neil Burch, coauthor of the new study) cracked in 2007 [2].
But poker is harder to solve than checkers. Chess and checkers are examples of perfect-information games, where players can have perfect knowledge of all past events in a game. In poker, there are some things a player doesn’t know: most crucially, which cards the other player has been dealt.
Devising a strategy to deal perfectly with that uncertainty is very hard. This hidden information is what gives a poker player the opportunity to bluff: to face the other player down with a relatively weak hand.
But although bluffing looks like a very human, psychological element of the game, it’s not. You can calculate how to bluff optimally. “Bluffing falls out of the mathematics of the game”, says Bowling. If you’re dealt a jack, say, it is possible to figure out how often you should ideally bluff with it. Some of the early pioneers of game theory, such as John von Neumann, aimed to develop mathematical strategies for bluffing.
The real challenge for a poker algorithm is dealing with the immense number of possible ways the game can be played. Bowling and colleagues have looked at one of the most popular forms, called Texas Hold’em, in which the dealer deals two cards to each player (face down) and also a set of face-up “community cards”. Players can bet, raise or fold after each deal. With just two players, the game becomes Heads-up, and it is a “limit” game when it has fixed bet sizes and a fixed number of raises. There are 3.16×10^17 states that HULHE can reach, and 3.19×10^14 possible points where a player must make a decision.
The new algorithm involves calculating all possible decisions in advance, so that they can just be looked up as a game proceeds. This is done in a learning process: the strategy begins by making decisions randomly, and is then updated by experience as the algorithm attaches a “regret” value to each decision depending on how poorly it fared. It takes a little more than 1500 training rounds to make the program essentially invincible.
This “counterfactual regret minimization” (CFR) procedure has been widely adopted in the Annual Computer Poker Competition, which has run since 2006. But Bowling and colleagues have now improved the procedure by allowing it to re-evaluate decisions considered to be poor in earlier training rounds.
The other crucial innovation was the handling of the vast amounts of information needing to be stored to develop and use the strategy – of the order of 262 terabytes. This volume of data demands disk storage, which is slow to access. The researchers figured out a data-compression method that reduces the volume to a more manageable 11 TB, and which adds only 5% to the computation time from the use of disk storage.
“I think the counterfactual regret algorithm is the major advance”, says computer scientist Jonathan Shapiro of the University of Manchester. “But they have done several other very clever things to make this problem computational feasible.”
Although the new algorithm plays “perfectly”, it is not necessarily unbeatable, since there is a large chance element in poker: it depends on the hand you’re dealt. The algorithm “may in fact lose after any finite number of hands, if it were unlucky”, says Bowling. But it always wins in the long run.
What’s more, there are two other limitations. It only “weakly” solves HULHE, meaning that it plays perfectly only if the strategy has been played throughout, and not if the computer player is suddenly dropped into the middle of a game (a situation that is in fact not really clearly defined for poker, as it is say for checkers).
And it is only what the researchers call “essentially solved”, meaning that it is not strictly unbeatable: there is an extremely small margin by which, in theory, it might be beaten by skill rather than chance. But this margin is negligible in practice. “Even if a human could identify the perfect counter-strategy to exploit our solution”, says Bowling, “and even if they could play this counter-strategy without error, and even if they spent 70 years only playing poker (over 60 million hands), they still couldn’t be statistically confident that they are winning [by superior play rather than chance].”
There has been heated debate about the use of computers (poker bots) for online poker games. “For quite a while now there has been a struggle between people who write bots and try to get them to play surreptitiously on online sites, and the sites who try to detect them and ban them”, says Jackson. But Bowling says that their algorithm will have little direct effect on this, because the popularity of online HULHE has declined as human players got better. “While it may dry up the last vestige of online HULHE play just due to the perception that humans can query a perfect strategy, this impact is going to minimal”, he says. But it may respark a philosophical discussion about bots in online poker.”
“I definitely think the ideas in this paper will be fruitfully applied to other forms of poker, such as no-limit”, says Jackson. “And more generally to other games, whether with dice or cards or whatever, that have imperfect information.”
But the stakes might be higher still for imperfect-information “games” beyond mere play and gambling. The stock market seems an obvious candidate, but Bowling explains that there the unknowns are too great. “The deck of cards in the stock market, which define the distribution at chance events, is not common knowledge and not even really knowable.”
He says that it might be useful, however, for portfolio management. “We’ve been investigating robust decision-making where the goal is to optimize a particular risk measure (such as value-at-risk)”, he says. “Such robust decision-making scenarios can often be cast as a game having an almost identical form to some of our poker games, with our solution techniques be immediately applicable.” The team’s current explorations beyond poker are, however, focused on supporting medical decision-making, in collaboration with diabetes specialists.
1. Bowling, M., Burch, N., Johanson, M. & Tammelin, O. Science 347, 145-149 (2015).
2. Schaeffer, J. et al., Science 317, 1518-1522 (2007).
Tuesday, January 06, 2015
The birth of the scientific journal
This is an extended version of a piece written for Research Fortnight. To celebrate its 350th anniversary, Phil. Trans. is also soon to publish a special issue containing some of its “greatest hits”, along with accompanying commentaries explaining their significance and impact. I have written a piece for it on Alan Turing’s classic 1952 paper on morphogenesis, which I’ll put up here when the time comes. The exhibition described below is small but fun, and if you’re in the neighbourhood of the Royal Society, well worth a look.
_______________________________________________________________________
The scientific journal is 350 years old this year. As the first real scientific journal, the Philosophical Transactions of the Royal Society, which published its first issue in January 1665, can claim to have set the scene for the entire scientific literature of today, which now counts its titles in the tens of thousands.
This history is explored in a new exhibition at the Royal Society in London to mark the anniversary. It has been put together by a team at the University of the University of St Andrews in Scotland, led by historian Aileen Fyfe, that has studied the development of the journal. The exhibition includes a copy of the first issue, the referee’s report on Charles Darwin’s sole publication in the journal (a minor work of 1839 about Scottish roads) and the handwritten manuscript submitted by James Clerk Maxwell in 1865 in which he proposes that light is an electromagnetic wave.
“Phil. Trans. was central to the whole idea of a scientific journal”, Fyfe says. Yet you need only glance at the first page of the first issue to see that the resemblances with the modern scientific journal were at that stage remote. Among this ‘accompt of the present Undertakings, Studies, and Labours of the Ingenious in many considerable parts of the World’, one could find the following:
"An account of the Improvement of Optick Glasses at Rome. Of the Observation made in England, of a Spot in one of the Belts of the Planet Jupiter. Of the motion of the late Comet predicted… An Experimental History of Cold… A Relation of a very odd Monstrous Calf. Of a peculiar lead-ore in Germany… Of the New American Whale-fishing about the Bermudas. A Narrative concerning the success of the Pendulum-watches at Sea for the Longitudes… A Catalogue of the Philosophical Books publisht by Monsieur de Fermat, Counsellour at Tholouse, lately dead."
In its early days Phil. Trans. published all kinds of strange, curious and often fanciful accounts of phenomena related to the Royal Society by its network of “virtuosi”: men (almost without exception) interested in the natural world, inventors, travellers, dilettantes and armchair philosophers. The selection of what to include and exclude was made solely by the Royal Society’s energetic secretary, the German natural philosopher Henry Oldenburg.
Oldenburg was the original networker, a multi-linguist who cultivated connections with all the “experimental philosophers” of seventeenth-century Europe. His approach is exemplified in a letter he sent in 1667 to the Italian naturalist Marcello Malpighi on Sicily:
"We earnestly beg you to be so good as to let us know of all that is noteworthy – of which there is so much in your island – concerning plants, or minerals, or animals and insects, especially the silkworm and its productions, and finally concerning meteorology and earthquakes, known to you or to other ingenious men."
The avowed intention of the Royal Society was to collect facts without rushing to formulate theories about them – witness Isaac Newton’s famous (and somewhat disingenuous) “hypotheses non fingo”. Yet Oldenburg’s choices reflected the spirit of his times, in which wealthy collectors and antiquaries stocked their cabinets with “curiosities” – strange and bizarre objects from around the world. Like them, the collectors of ‘facts’ at the Royal Society were often drawn to reports that were entertaining, amazing or strange rather than necessarily informative.
The editorial power wielded by Oldenburg – his contemporary Robert Hooke, demonstrator for the Royal Society, called him a “dog” for perceived biases in his record-taking – was inherited by his successors as secretary. So, however, was the considerable financial burden of producing the Transactions. But when in 1752 the Royal Society first took official responsibility for the journal (it had previously been something more akin to a news-sheet), the organization felt that it needed to think about its reputation. In the face of complaints about the poor quality of some of the content, which placed sensation before reliability, the council members figured that they needed a mechanism for making editorial decisions that were seen to be fair and well grounded.
At this time (and for many years subsequently), papers in Phil. Trans. were first read at the Society’s meetings before being published. So it was decided that Fellows would hold a secret ballot on whether, after hearing a contribution, it should be included for publication, with or without modifications. (Fyfe doubts whether this procedure was always followed to the letter.) It was a kind of peer review, after a fashion – albeit one conducted by what amounted to show of hands among a tiny clique.
In 1832 that procedure for collective editorial decision-making was extended when the Society began to solicit written reports from two reviewers – the first real instance of what we would now recognize as real peer review. Fyfe notes that some other societies were starting to introduce this system for their journals around the same time, but Phil. Trans. was certainly the most prestigious title to do so, and that the use of two referees was standard practice by the mid-nineteenth century. “Phil. Trans. became a modern scientific journal in the nineteenth century”, she says – indeed, it more or less created the template for what that meant.
The secretary for most of that century’s second half, the physicist George Stokes, was instrumental in this increasing professionalization of the publication process. His role, and that of his successors, was now becoming something like that of a journal editor as we know it today. It was during this period that commercial scientific journals began to flourish, such as the idiosyncratic Chemical News edited and published by the equally idiosyncratic William Crookes, and most famously Nature, started in 1869 by the astronomer Norman Lockyer. While these commercial ventures could publish what they liked – the peer review system at Nature was still very informal in the 1960s – learned journals such as Phil. Trans. were concerned to show their objectivity and impartiality: attributes that any modern scholarly journal now likes to claim.
This, however, was all relative. For one thing, until the 1970s, if you wanted to submit to Phil. Trans. but were not a Fellow of the Royal Society then you needed the blessing of someone who was. This meant that you needed to be plugged into the right networks, and it encouraged systems of patronage, even nepotism: Lord Kelvin was particularly active as a sponsor of submissions, often those of his former students. Schemes of this kind still persisted in recent times. Notably, the Proceedings of the National Academy of Sciences USA only admitted regular submissions from non-members of the Academy without an NAS sponsor in 1995; and not until 2010, after much criticism, did the journal do away with the principle of “communicating” submissions via Academy members, which almost guaranteed publication.
What’s more, the Phil. Trans. referees came from a limited pool. In the mid-nineteenth century, about half of them were members of the Royal Society council, and all the others had to be Fellows.
The interesting question is how much these developments changed the nature of what was published. Pre-selection procedures by the Fellows doubtless excluded a lot of bad material, so Fyfe thinks that one of the main consequences of formal peer review was not that it raised the quality of published research so much as that it encouraged authors to develop a particular literary style to improve their chances: to reduce speculation and observe the brevity, sobriety and even blandness that some would say afflicts the scientific literature today. (Darwin’s paper was criticized by geologist Adam Sedgwick for its loquaciousness.)
With alternatives to the “standard model” of peer review now proliferating, from the “techniques-only” assessments of PLoS ONE and Scientific Reports to the increasing acceptance of preprint servers as venues of de facto publication, it seems particularly timely to consider how science publication evolved and acquired its customs and habits. Perhaps peer review has become something of a shibboleth. Certainly it seems sometimes to have mutated from a routine check and trash filter to a dictatorial, almost paranoid gatekeeper: biologists complain that no referee seems to consider they have done their job unless they have suggested half a dozen additional experiments. There is surely something in the famous suggestion that Watson and Crick’s 1953 paper would not have found favour with Nature’s reviewers today. And the broadening of reviewing networks, while surely beneficial in many ways, hasn’t eliminated accusations (some well founded) of favoritism, discrimination and bias towards big-name labs. There is a fine balance to be fund between rigour and permissiveness, one that can fall foul of conservatism and petty box-checking as much as caprice. The story of Phil. Trans. opens a lively window on that discussion.
_______________________________________________________________________
The scientific journal is 350 years old this year. As the first real scientific journal, the Philosophical Transactions of the Royal Society, which published its first issue in January 1665, can claim to have set the scene for the entire scientific literature of today, which now counts its titles in the tens of thousands.
This history is explored in a new exhibition at the Royal Society in London to mark the anniversary. It has been put together by a team at the University of the University of St Andrews in Scotland, led by historian Aileen Fyfe, that has studied the development of the journal. The exhibition includes a copy of the first issue, the referee’s report on Charles Darwin’s sole publication in the journal (a minor work of 1839 about Scottish roads) and the handwritten manuscript submitted by James Clerk Maxwell in 1865 in which he proposes that light is an electromagnetic wave.
“Phil. Trans. was central to the whole idea of a scientific journal”, Fyfe says. Yet you need only glance at the first page of the first issue to see that the resemblances with the modern scientific journal were at that stage remote. Among this ‘accompt of the present Undertakings, Studies, and Labours of the Ingenious in many considerable parts of the World’, one could find the following:
"An account of the Improvement of Optick Glasses at Rome. Of the Observation made in England, of a Spot in one of the Belts of the Planet Jupiter. Of the motion of the late Comet predicted… An Experimental History of Cold… A Relation of a very odd Monstrous Calf. Of a peculiar lead-ore in Germany… Of the New American Whale-fishing about the Bermudas. A Narrative concerning the success of the Pendulum-watches at Sea for the Longitudes… A Catalogue of the Philosophical Books publisht by Monsieur de Fermat, Counsellour at Tholouse, lately dead."
In its early days Phil. Trans. published all kinds of strange, curious and often fanciful accounts of phenomena related to the Royal Society by its network of “virtuosi”: men (almost without exception) interested in the natural world, inventors, travellers, dilettantes and armchair philosophers. The selection of what to include and exclude was made solely by the Royal Society’s energetic secretary, the German natural philosopher Henry Oldenburg.
Oldenburg was the original networker, a multi-linguist who cultivated connections with all the “experimental philosophers” of seventeenth-century Europe. His approach is exemplified in a letter he sent in 1667 to the Italian naturalist Marcello Malpighi on Sicily:
"We earnestly beg you to be so good as to let us know of all that is noteworthy – of which there is so much in your island – concerning plants, or minerals, or animals and insects, especially the silkworm and its productions, and finally concerning meteorology and earthquakes, known to you or to other ingenious men."
The avowed intention of the Royal Society was to collect facts without rushing to formulate theories about them – witness Isaac Newton’s famous (and somewhat disingenuous) “hypotheses non fingo”. Yet Oldenburg’s choices reflected the spirit of his times, in which wealthy collectors and antiquaries stocked their cabinets with “curiosities” – strange and bizarre objects from around the world. Like them, the collectors of ‘facts’ at the Royal Society were often drawn to reports that were entertaining, amazing or strange rather than necessarily informative.
The editorial power wielded by Oldenburg – his contemporary Robert Hooke, demonstrator for the Royal Society, called him a “dog” for perceived biases in his record-taking – was inherited by his successors as secretary. So, however, was the considerable financial burden of producing the Transactions. But when in 1752 the Royal Society first took official responsibility for the journal (it had previously been something more akin to a news-sheet), the organization felt that it needed to think about its reputation. In the face of complaints about the poor quality of some of the content, which placed sensation before reliability, the council members figured that they needed a mechanism for making editorial decisions that were seen to be fair and well grounded.
At this time (and for many years subsequently), papers in Phil. Trans. were first read at the Society’s meetings before being published. So it was decided that Fellows would hold a secret ballot on whether, after hearing a contribution, it should be included for publication, with or without modifications. (Fyfe doubts whether this procedure was always followed to the letter.) It was a kind of peer review, after a fashion – albeit one conducted by what amounted to show of hands among a tiny clique.
In 1832 that procedure for collective editorial decision-making was extended when the Society began to solicit written reports from two reviewers – the first real instance of what we would now recognize as real peer review. Fyfe notes that some other societies were starting to introduce this system for their journals around the same time, but Phil. Trans. was certainly the most prestigious title to do so, and that the use of two referees was standard practice by the mid-nineteenth century. “Phil. Trans. became a modern scientific journal in the nineteenth century”, she says – indeed, it more or less created the template for what that meant.
The secretary for most of that century’s second half, the physicist George Stokes, was instrumental in this increasing professionalization of the publication process. His role, and that of his successors, was now becoming something like that of a journal editor as we know it today. It was during this period that commercial scientific journals began to flourish, such as the idiosyncratic Chemical News edited and published by the equally idiosyncratic William Crookes, and most famously Nature, started in 1869 by the astronomer Norman Lockyer. While these commercial ventures could publish what they liked – the peer review system at Nature was still very informal in the 1960s – learned journals such as Phil. Trans. were concerned to show their objectivity and impartiality: attributes that any modern scholarly journal now likes to claim.
This, however, was all relative. For one thing, until the 1970s, if you wanted to submit to Phil. Trans. but were not a Fellow of the Royal Society then you needed the blessing of someone who was. This meant that you needed to be plugged into the right networks, and it encouraged systems of patronage, even nepotism: Lord Kelvin was particularly active as a sponsor of submissions, often those of his former students. Schemes of this kind still persisted in recent times. Notably, the Proceedings of the National Academy of Sciences USA only admitted regular submissions from non-members of the Academy without an NAS sponsor in 1995; and not until 2010, after much criticism, did the journal do away with the principle of “communicating” submissions via Academy members, which almost guaranteed publication.
What’s more, the Phil. Trans. referees came from a limited pool. In the mid-nineteenth century, about half of them were members of the Royal Society council, and all the others had to be Fellows.
The interesting question is how much these developments changed the nature of what was published. Pre-selection procedures by the Fellows doubtless excluded a lot of bad material, so Fyfe thinks that one of the main consequences of formal peer review was not that it raised the quality of published research so much as that it encouraged authors to develop a particular literary style to improve their chances: to reduce speculation and observe the brevity, sobriety and even blandness that some would say afflicts the scientific literature today. (Darwin’s paper was criticized by geologist Adam Sedgwick for its loquaciousness.)
With alternatives to the “standard model” of peer review now proliferating, from the “techniques-only” assessments of PLoS ONE and Scientific Reports to the increasing acceptance of preprint servers as venues of de facto publication, it seems particularly timely to consider how science publication evolved and acquired its customs and habits. Perhaps peer review has become something of a shibboleth. Certainly it seems sometimes to have mutated from a routine check and trash filter to a dictatorial, almost paranoid gatekeeper: biologists complain that no referee seems to consider they have done their job unless they have suggested half a dozen additional experiments. There is surely something in the famous suggestion that Watson and Crick’s 1953 paper would not have found favour with Nature’s reviewers today. And the broadening of reviewing networks, while surely beneficial in many ways, hasn’t eliminated accusations (some well founded) of favoritism, discrimination and bias towards big-name labs. There is a fine balance to be fund between rigour and permissiveness, one that can fall foul of conservatism and petty box-checking as much as caprice. The story of Phil. Trans. opens a lively window on that discussion.
Monday, January 05, 2015
The First Emperor's rivers of mercury
This is a slightly extended version of my feature article in the latest issue of Chemistry World.
_________________________________________________________________________
The Chinese emperor had done all he could to become immortal, but in vain. His physicians had prepared herbal and alchemical elixirs, but none could stave off his decline. He had sent a minister on a voyage far over the eastern seas in search of a mythical potion of eternal life. But that expedition never returned, and now the quest seemed hopeless. So Qin Shi Huangdi, the first emperor of a unified China in the third century BC, had begun preparations for the next best thing to an endless life on earth. He would continue his cosmic rule from the spirit world, and his underground tomb would be a kind of palace for the afterlife, complete with its own army of life-size clay soldiers.
Those terracotta warriors lay hidden for two millennia beneath several metres of wind-deposited sandy soil a mile from the First Emperor’s burial mound at Mount Li (Lishan), to the northeast of the city of Xi’an in Shaanxi province of north-central China. They were rediscovered in 1974 by farmers digging a well, and Chinese archaeologists were astonished to find over the next decade that there were at least 8,000 of them, once brightly painted and equipped with clay horses and wooden chariots. As further excavation revealed the extent of the emperor’s mausoleum, with offices, stables and halls along with clay figures of officials, acrobats and labourers and life-size bronze animals, it became clear that the Han-dynasty historian Sima Qian, writing in the second century BC, hadn’t been exaggerating after all. Sima Qian claimed that 700,000 men had worked on the emperor’s tomb, constructing entire palaces, towers and scenic landscapes through which which the emperor’s spirit might roam.
The Terracotta Army was created to serve and protect China’s first emperor.
No one knows what other wonders the mausoleum might house, for the main burial chamber – a football-pitch-sized hall beneath a great mound of earth – remains sealed. Most enticing of all is a detail related by Sima Qian: “Mercury was used to fashion the hundred rivers, the Yellow River and the Long River [Yangtze], and the seas in such a way that they flowed.” This idea that the main chamber contains a kind of microcosm of all of China (as it was then recognized) with rivers, lakes and seas of shimmering mercury had long seemed to fantastic for modern historians to grant it credence. But if Sima Qian had not been inventing stories about other elaborate features of the mausoleum site, might this account of the tomb chamber be reliable too?
In the 1980s Chinese researchers found that the soil in the burial mound above the tomb indeed contains concentration of mercury way above what the soil elsewhere in this region carries. Now some archaeologists working on the site are quite ready to believe that the body of Qin Shi Huangdi may indeed lie amidst vast puddles of the liquid metal.
Yet it seems unlikely that anyone will gaze on such a sight in the foreseeable future. “We have no current plan to open the chambers”, says archaeologist Qingbo Duan of Northwest University in Xi’an, who led the mausoleum excavations from 1998 to 2008. “We have no mature technologies and effective measures to protect the relics.” So can we ever know the truth about the First Emperor’s rivers of mercury?
A harsh legacy
The construction of this immense mausoleum started fully 36 years before Qin Shi Huangdi’s death in 210 BC, when he was merely King Zheng of the kingdom of Qin – a realm occupying the valley of the Wei, a major tributary of the Yellow River, now in Shaanxi. Qin was one of seven states within China at that time, all of which had been vying for supremacy since the fifth century BC in what is known as the Warring States period. By finally defeating the last of the rival states, Qi in modern Shandong, in 221 BC, Zheng became Qin Shi Huangdi (“the First Qin Emperor”), ruler of all China.
Qin Shi Huangdi was China’s first emperor, and he hoped to use alchemical elixirs and medicines to sustain his life indefinitely.
Some etymologies trace the name “China” itself to the Qin dynasty (pronounced “Chin”), and so you might imagine that it would have a very special status in Chinese history. But the unified state barely outlasted the death of Qin Shi Huangdi himself – four years later it succumbed to a rebellion that became the much more durable Han dynasty (206 BC – 220 AD) – and it is regarded with little fondness in China today, for the First Emperor was a tyrant who ruled with brutal force. He compelled his subjects to achieve marvelous feats of engineering – he constructed the Great Wall from existing fragmentary defenses on the northern frontier, as well as the Lingqu Canal connecting the Yangtze to the Pearl River delta in the south, not to mention his own mausoleum. The First Emperor also introduced standardization of weights and measures and of the Chinese writing system. But in an attempt to expunge all previous histories and ideas, he ordered the burning of many precious documents and works of philosophy and poetry: a treasury of learning that was lost forever. The Qin rulers followed a philosophical tradition called Legalism, which advocated the ruthless suppression of all criticism and opposition.
Since much of what we know about the Qin era comes from Sima Qian, who was writing to justify the Han ascendancy over the previous rulers, it’s possible that Qin Shi Huangdi gets something of a raw deal. But there’s reluctant admiration in the way Sima Qian describes the magnificence of Qin Shi Huangdi’s tomb, which was unlike anything that had been attempted before. The Emperor didn’t just see himself as a worldly ruler – he considered his empire to be blessed by heaven, and he placed himself in the line of “sage-kings” going back to China’s mythical origins. Like all Chinese at that time, he believed that after death people’s spirits didn’t travel to some heavenly place removed from the physical world, but that the spiritual and mundane worlds coexisted, so that in some sense his rule would continue on earth after death. There was, then, nothing symbolic about all the trappings of power that would surround him in his tomb – they would be useful in the times to come.
“In ancient China, people believed the souls of the dead would live forever underground, so they would prepare almost everything from real life to bury for use in the afterlife”, says Yinglan Zhang, an archaeologist at the Shaanxi History Museum in Xi’an and deputy director of the mausoleum excavations from 1998 to 2007. Given what has already been unearthed, he says “there should be many other cultural artifacts or relics still buried in the tomb chamber or other burial pits around the tomb – maybe things beyond our imagination.”
The pits housing the Terracotta Army lie outside the 2 by 0.8 km boundary wall of the burial mound. Inside this wall are ritual buildings once containing food and other items that the emperor would need to sustain him. There are chambers full of stone armour that could protect against evil spirits, and it is possible that the emperor himself might not have been interred alone in the main chamber: Sima Qian says that officials were buried there with him, and it’s not clear if they were alive or dead at the time.
The mound itself was originally about 0.5 by 0.5 km (erosion has shrunk it a little), and the burial chamber lies about 30-40 m below the original ground surface. Its shape has been mapped out by measuring gravity anomalies in the ground – an indication of hollow or less dense structures – and by looking for changes in the electrical resistivity of the soil, which result from buried structures or cavities. In this way, Chinese archaeologists have figured out the basic layout of the tomb over the past several decades. The chamber is about 80 m east-west by 50 m north-south, surrounded by a wall of closely packed earth and – to judge from other ancient Chinese tombs – perhaps water-proofed with with stone covered with red lacquer. In 2000 researchers discovered that towards the edge of the mound a drainage dam helps to keep water away from the chamber. So there’s some reason to believe that the tomb itself might be relatively intact: neither wholly collapsed nor water-filled.
The burial mound of China’s first emperor, near Xi’an in Shaanxi province.
Measurements of the soil resistivity in the region of the chamber have also revealed another intriguing feature. They show a so-called phase anomaly, which is produced when an electrical current is reflected from a conducting surface, such as a metal. Could this be a sign of pools and streams of mercury?
The first detailed study of mercury levels in the mound were conducted in the early 1980s, when researchers from the Institute of Geophysical and Geochemical Exploration of the China Institute of Geo-Environment Monitoring sunk small boreholes into the soil over an area of 12,000 m2 in the centre of the mound and extracted soil samples for analysis. Whereas soils outside this central region contained an average of 30 parts per billion of mercury, the average above the chamber was 250 ppb, and in some places rose to 1500 ppb. A second survey in 2003 found much the same: unusually high concentrations of mercury both in the soil itself and the interstitial vapours between grains.
The grid of borehole samples allowed the Chinese researchers to make a rough map of how the high levels of mercury are distributed. “There is no unusual amount of mercury in the northwest corner of the tomb”, says Duan, “while the mercury level is highest in the northeast and second highest in the south.” If you squint at this distribution, you can persuade yourself that it matches the locations of the two great rivers of China – the Yellow and Yangtze – as seen from the ancient Qin capital of Xianyang, close to modern Xi’an. “The distribution of mercury level corresponds to the location of waterways in the Qin empire”, Duan asserts. In other words, the tomb might indeed contain a facsimile of the empire, watered by mercury.
The mercury levels in soils above the tomb chamber (top), and a map of China from the eleventh century AD (bottom) showing the rivers, especially the Yellow (north) and Yangtze (south). In Qin times the knowledge of China’s topography would have been much more rudimentary, but the locations of the main rivers would have been known roughly.
Zhang isn’t so sure that one can conclude much from the present-day mercury distribution, however. He thinks that the tomb chamber must have collapsed thousands years ago, just like the pits containing the Terracotta Army. “The mercury will have volatilized into nearby soils during this long time, so it would be impossible to show up detailed information that we can connect with particular rivers or lakes”, he says.
Silver water
In any case, just because the mausoleum apparently contains a lot of mercury doesn’t in itself verify Sima Qian’s account. It had other uses too, particularly in alchemy, which has some of its oldest roots in China. In the West this art was commonly associated with attempts to make gold from other metals, and some Chinese alchemists tried that too – in 144 BC the Han Emperor Jingdi decreed that anyone caught trying to make counterfeit gold should be executed. But Chinese alchemy was more oriented towards medicinal uses, in particular elixirs of immortality. Some believed that alchemical gold could have this effect: the Han emperor Xuandi in 60 BC appointed the scholar Liu Xiang to make alchemical gold to prolong his life.
Others thought that the elixir of life lay elsewhere – and perhaps mercury (in Chinese shui yin, literally “water silver”) was the key. Chinese legend tells of one Huang An, who prolonged his life for at least 10,000 years by eating mercury sulphide (the mineral cinnabar). Qin Shi Huangdi was said to have consumed wine and honey laden with cinnabar thinking it would prolong his life, and some have speculated that he might have hastened his death with these “medicines”. During the Warring States period, mercury was a common ingredient of medicines, being used to treat infected sores, scabies, ringworm and (even more alarmingly) as a sedative for mania and insomnia.
It had other uses too. Cinnabar itself is red, and it was long used in China for art and decoration – its artificial form, produced in the West since the Roman era, became known as the pigment vermilion. The mineral has been found on the “oracle bones” used for divination during the Shang Dynasty of Bronze Age China (second millennium BC).
Cinnabar (HgS) was widely used in ancient China for decoration, medicine and alchemy.
One of the most important uses of mercury at this time has a particularly alchemical tinge. Gold and silver dissolve in mercury to form amalgams, and such mixtures were used for gilt plating. The amalgam was rubbed on and heated to evaporate the mercury and leave behind a gleaming coat of precious metal. Such mixtures also featured in alchemical elixirs: the Daoist concept of yin and yang, the two fundamental and complementary principles of life, encouraged an idea that cold, watery (yang) mercury and bright, fiery (yin) gold might be blended in ideal proportions to sustain vitality. Such ideas, says Duan, “led astray the ancient scientific aspects of mercury use until a re-awakening in the Song dynasty” (10th-13th centuries AD).
Throughout antiquity cinnabar was the source of all mercury metal, which can be extracted simply by heating. There was a lot of cinnabar in China, particularly in the western regions such as Sichuan. Shaanxi alone contains almost a fifth of all the cinnabar reserves in the country, and there are very ancient mines in Xunyang county in the south of the province that are a good candidate source of the mercury apparently in the First Emperor’s tomb.
To extract mercury from cinnabar one need only roast it in air, converting the sulphur to SO2 while the mercury is released as vapour that can then be condensed. Since mercury boils at 357 oC, this process needs temperatures of little more than 350 oC, well within the capabilities of Qin-era kilns. Of course, anyone trying this method in an unsealed container – closed chambers weren’t used until the Han period – risked serious harm.
But despite there being a mature mercury-refining technology by the time of the Qin, and although Zhang attests that “the people of the Qin Dynasty had some basic chemical knowledge”, Duan argues that Chinese alchemy was still in its infancy in that period. In particular, he says, there is no good reason to think that the practice of soaking dead bodies in mercury to prevent their decay, common during the Song dynasty in the 10th-13th centuries AD, was used as early as the Qin dynasty. So even though mercury, either as cinnabar or as the elemental metal, has been found in tombs dating back as far as second millennium BC, it’s not clear why it was put there. Might its toxicity have acted as a deterrent to grave-looters? Probably not – the dangers of mercury fumes were not recognized until Han times. So if, as it seems, there’s a lot of mercury in Qin Shi Huangdi’s burial chamber, it’s unlikely to be either a preservative or an anti-theft device. (Sima Qian says that the First Emperor’s tomb was, however, booby-trapped with crossbows “rigged so that they would immediately shoot down anyone attempting to break in”, suggesting that if archaeologists were ever to try opening it up, they might face Indiana Jones-style hazards.)
Yet even if this mercury was indeed used for fantastical landscaping, Duan doubts that there can have been much of it. Based on estimates of mercury production from the Song era and allowing for the imperfections of the earlier refinement process, he thinks the chamber might have contained at most 100 tons of the liquid metal: around 7 m3.
We might never be able to check that. “Right now, our archaeological work is focused on deducing the basic layout” of the tomb, says Duan. Because even a small breach in the seal could admit water or air that might damage whatever lies within, even robot-based exploration of the interior is ruled out. “If the chamber was opened even using a robot or drilling, the balance of the situation would be broken and the buried objects would deteriorate quickly”, says Zhang.
So if we’re ever going to peek inside, it will have to be with better scientific techniques than are currently available. “I dream of a day when technology will shed light on all that is buried there, without disturbing the sleeping emperor and his two-thousand-year-old underground empire”, says Yongqi Wu, curator of the Qin Shi Huang Mausoleum Museum at the Lishan site. Maybe these concerns to preserve the unknown heritage will guarantee Qin Shi Huangdi a kind of immortality after all.
_________________________________________________________________________
The Chinese emperor had done all he could to become immortal, but in vain. His physicians had prepared herbal and alchemical elixirs, but none could stave off his decline. He had sent a minister on a voyage far over the eastern seas in search of a mythical potion of eternal life. But that expedition never returned, and now the quest seemed hopeless. So Qin Shi Huangdi, the first emperor of a unified China in the third century BC, had begun preparations for the next best thing to an endless life on earth. He would continue his cosmic rule from the spirit world, and his underground tomb would be a kind of palace for the afterlife, complete with its own army of life-size clay soldiers.
Those terracotta warriors lay hidden for two millennia beneath several metres of wind-deposited sandy soil a mile from the First Emperor’s burial mound at Mount Li (Lishan), to the northeast of the city of Xi’an in Shaanxi province of north-central China. They were rediscovered in 1974 by farmers digging a well, and Chinese archaeologists were astonished to find over the next decade that there were at least 8,000 of them, once brightly painted and equipped with clay horses and wooden chariots. As further excavation revealed the extent of the emperor’s mausoleum, with offices, stables and halls along with clay figures of officials, acrobats and labourers and life-size bronze animals, it became clear that the Han-dynasty historian Sima Qian, writing in the second century BC, hadn’t been exaggerating after all. Sima Qian claimed that 700,000 men had worked on the emperor’s tomb, constructing entire palaces, towers and scenic landscapes through which which the emperor’s spirit might roam.
The Terracotta Army was created to serve and protect China’s first emperor.
No one knows what other wonders the mausoleum might house, for the main burial chamber – a football-pitch-sized hall beneath a great mound of earth – remains sealed. Most enticing of all is a detail related by Sima Qian: “Mercury was used to fashion the hundred rivers, the Yellow River and the Long River [Yangtze], and the seas in such a way that they flowed.” This idea that the main chamber contains a kind of microcosm of all of China (as it was then recognized) with rivers, lakes and seas of shimmering mercury had long seemed to fantastic for modern historians to grant it credence. But if Sima Qian had not been inventing stories about other elaborate features of the mausoleum site, might this account of the tomb chamber be reliable too?
In the 1980s Chinese researchers found that the soil in the burial mound above the tomb indeed contains concentration of mercury way above what the soil elsewhere in this region carries. Now some archaeologists working on the site are quite ready to believe that the body of Qin Shi Huangdi may indeed lie amidst vast puddles of the liquid metal.
Yet it seems unlikely that anyone will gaze on such a sight in the foreseeable future. “We have no current plan to open the chambers”, says archaeologist Qingbo Duan of Northwest University in Xi’an, who led the mausoleum excavations from 1998 to 2008. “We have no mature technologies and effective measures to protect the relics.” So can we ever know the truth about the First Emperor’s rivers of mercury?
A harsh legacy
The construction of this immense mausoleum started fully 36 years before Qin Shi Huangdi’s death in 210 BC, when he was merely King Zheng of the kingdom of Qin – a realm occupying the valley of the Wei, a major tributary of the Yellow River, now in Shaanxi. Qin was one of seven states within China at that time, all of which had been vying for supremacy since the fifth century BC in what is known as the Warring States period. By finally defeating the last of the rival states, Qi in modern Shandong, in 221 BC, Zheng became Qin Shi Huangdi (“the First Qin Emperor”), ruler of all China.
Qin Shi Huangdi was China’s first emperor, and he hoped to use alchemical elixirs and medicines to sustain his life indefinitely.
Some etymologies trace the name “China” itself to the Qin dynasty (pronounced “Chin”), and so you might imagine that it would have a very special status in Chinese history. But the unified state barely outlasted the death of Qin Shi Huangdi himself – four years later it succumbed to a rebellion that became the much more durable Han dynasty (206 BC – 220 AD) – and it is regarded with little fondness in China today, for the First Emperor was a tyrant who ruled with brutal force. He compelled his subjects to achieve marvelous feats of engineering – he constructed the Great Wall from existing fragmentary defenses on the northern frontier, as well as the Lingqu Canal connecting the Yangtze to the Pearl River delta in the south, not to mention his own mausoleum. The First Emperor also introduced standardization of weights and measures and of the Chinese writing system. But in an attempt to expunge all previous histories and ideas, he ordered the burning of many precious documents and works of philosophy and poetry: a treasury of learning that was lost forever. The Qin rulers followed a philosophical tradition called Legalism, which advocated the ruthless suppression of all criticism and opposition.
Since much of what we know about the Qin era comes from Sima Qian, who was writing to justify the Han ascendancy over the previous rulers, it’s possible that Qin Shi Huangdi gets something of a raw deal. But there’s reluctant admiration in the way Sima Qian describes the magnificence of Qin Shi Huangdi’s tomb, which was unlike anything that had been attempted before. The Emperor didn’t just see himself as a worldly ruler – he considered his empire to be blessed by heaven, and he placed himself in the line of “sage-kings” going back to China’s mythical origins. Like all Chinese at that time, he believed that after death people’s spirits didn’t travel to some heavenly place removed from the physical world, but that the spiritual and mundane worlds coexisted, so that in some sense his rule would continue on earth after death. There was, then, nothing symbolic about all the trappings of power that would surround him in his tomb – they would be useful in the times to come.
“In ancient China, people believed the souls of the dead would live forever underground, so they would prepare almost everything from real life to bury for use in the afterlife”, says Yinglan Zhang, an archaeologist at the Shaanxi History Museum in Xi’an and deputy director of the mausoleum excavations from 1998 to 2007. Given what has already been unearthed, he says “there should be many other cultural artifacts or relics still buried in the tomb chamber or other burial pits around the tomb – maybe things beyond our imagination.”
The pits housing the Terracotta Army lie outside the 2 by 0.8 km boundary wall of the burial mound. Inside this wall are ritual buildings once containing food and other items that the emperor would need to sustain him. There are chambers full of stone armour that could protect against evil spirits, and it is possible that the emperor himself might not have been interred alone in the main chamber: Sima Qian says that officials were buried there with him, and it’s not clear if they were alive or dead at the time.
The mound itself was originally about 0.5 by 0.5 km (erosion has shrunk it a little), and the burial chamber lies about 30-40 m below the original ground surface. Its shape has been mapped out by measuring gravity anomalies in the ground – an indication of hollow or less dense structures – and by looking for changes in the electrical resistivity of the soil, which result from buried structures or cavities. In this way, Chinese archaeologists have figured out the basic layout of the tomb over the past several decades. The chamber is about 80 m east-west by 50 m north-south, surrounded by a wall of closely packed earth and – to judge from other ancient Chinese tombs – perhaps water-proofed with with stone covered with red lacquer. In 2000 researchers discovered that towards the edge of the mound a drainage dam helps to keep water away from the chamber. So there’s some reason to believe that the tomb itself might be relatively intact: neither wholly collapsed nor water-filled.
The burial mound of China’s first emperor, near Xi’an in Shaanxi province.
Measurements of the soil resistivity in the region of the chamber have also revealed another intriguing feature. They show a so-called phase anomaly, which is produced when an electrical current is reflected from a conducting surface, such as a metal. Could this be a sign of pools and streams of mercury?
The first detailed study of mercury levels in the mound were conducted in the early 1980s, when researchers from the Institute of Geophysical and Geochemical Exploration of the China Institute of Geo-Environment Monitoring sunk small boreholes into the soil over an area of 12,000 m2 in the centre of the mound and extracted soil samples for analysis. Whereas soils outside this central region contained an average of 30 parts per billion of mercury, the average above the chamber was 250 ppb, and in some places rose to 1500 ppb. A second survey in 2003 found much the same: unusually high concentrations of mercury both in the soil itself and the interstitial vapours between grains.
The grid of borehole samples allowed the Chinese researchers to make a rough map of how the high levels of mercury are distributed. “There is no unusual amount of mercury in the northwest corner of the tomb”, says Duan, “while the mercury level is highest in the northeast and second highest in the south.” If you squint at this distribution, you can persuade yourself that it matches the locations of the two great rivers of China – the Yellow and Yangtze – as seen from the ancient Qin capital of Xianyang, close to modern Xi’an. “The distribution of mercury level corresponds to the location of waterways in the Qin empire”, Duan asserts. In other words, the tomb might indeed contain a facsimile of the empire, watered by mercury.
The mercury levels in soils above the tomb chamber (top), and a map of China from the eleventh century AD (bottom) showing the rivers, especially the Yellow (north) and Yangtze (south). In Qin times the knowledge of China’s topography would have been much more rudimentary, but the locations of the main rivers would have been known roughly.
Zhang isn’t so sure that one can conclude much from the present-day mercury distribution, however. He thinks that the tomb chamber must have collapsed thousands years ago, just like the pits containing the Terracotta Army. “The mercury will have volatilized into nearby soils during this long time, so it would be impossible to show up detailed information that we can connect with particular rivers or lakes”, he says.
Silver water
In any case, just because the mausoleum apparently contains a lot of mercury doesn’t in itself verify Sima Qian’s account. It had other uses too, particularly in alchemy, which has some of its oldest roots in China. In the West this art was commonly associated with attempts to make gold from other metals, and some Chinese alchemists tried that too – in 144 BC the Han Emperor Jingdi decreed that anyone caught trying to make counterfeit gold should be executed. But Chinese alchemy was more oriented towards medicinal uses, in particular elixirs of immortality. Some believed that alchemical gold could have this effect: the Han emperor Xuandi in 60 BC appointed the scholar Liu Xiang to make alchemical gold to prolong his life.
Others thought that the elixir of life lay elsewhere – and perhaps mercury (in Chinese shui yin, literally “water silver”) was the key. Chinese legend tells of one Huang An, who prolonged his life for at least 10,000 years by eating mercury sulphide (the mineral cinnabar). Qin Shi Huangdi was said to have consumed wine and honey laden with cinnabar thinking it would prolong his life, and some have speculated that he might have hastened his death with these “medicines”. During the Warring States period, mercury was a common ingredient of medicines, being used to treat infected sores, scabies, ringworm and (even more alarmingly) as a sedative for mania and insomnia.
It had other uses too. Cinnabar itself is red, and it was long used in China for art and decoration – its artificial form, produced in the West since the Roman era, became known as the pigment vermilion. The mineral has been found on the “oracle bones” used for divination during the Shang Dynasty of Bronze Age China (second millennium BC).
Cinnabar (HgS) was widely used in ancient China for decoration, medicine and alchemy.
One of the most important uses of mercury at this time has a particularly alchemical tinge. Gold and silver dissolve in mercury to form amalgams, and such mixtures were used for gilt plating. The amalgam was rubbed on and heated to evaporate the mercury and leave behind a gleaming coat of precious metal. Such mixtures also featured in alchemical elixirs: the Daoist concept of yin and yang, the two fundamental and complementary principles of life, encouraged an idea that cold, watery (yang) mercury and bright, fiery (yin) gold might be blended in ideal proportions to sustain vitality. Such ideas, says Duan, “led astray the ancient scientific aspects of mercury use until a re-awakening in the Song dynasty” (10th-13th centuries AD).
Throughout antiquity cinnabar was the source of all mercury metal, which can be extracted simply by heating. There was a lot of cinnabar in China, particularly in the western regions such as Sichuan. Shaanxi alone contains almost a fifth of all the cinnabar reserves in the country, and there are very ancient mines in Xunyang county in the south of the province that are a good candidate source of the mercury apparently in the First Emperor’s tomb.
To extract mercury from cinnabar one need only roast it in air, converting the sulphur to SO2 while the mercury is released as vapour that can then be condensed. Since mercury boils at 357 oC, this process needs temperatures of little more than 350 oC, well within the capabilities of Qin-era kilns. Of course, anyone trying this method in an unsealed container – closed chambers weren’t used until the Han period – risked serious harm.
But despite there being a mature mercury-refining technology by the time of the Qin, and although Zhang attests that “the people of the Qin Dynasty had some basic chemical knowledge”, Duan argues that Chinese alchemy was still in its infancy in that period. In particular, he says, there is no good reason to think that the practice of soaking dead bodies in mercury to prevent their decay, common during the Song dynasty in the 10th-13th centuries AD, was used as early as the Qin dynasty. So even though mercury, either as cinnabar or as the elemental metal, has been found in tombs dating back as far as second millennium BC, it’s not clear why it was put there. Might its toxicity have acted as a deterrent to grave-looters? Probably not – the dangers of mercury fumes were not recognized until Han times. So if, as it seems, there’s a lot of mercury in Qin Shi Huangdi’s burial chamber, it’s unlikely to be either a preservative or an anti-theft device. (Sima Qian says that the First Emperor’s tomb was, however, booby-trapped with crossbows “rigged so that they would immediately shoot down anyone attempting to break in”, suggesting that if archaeologists were ever to try opening it up, they might face Indiana Jones-style hazards.)
Yet even if this mercury was indeed used for fantastical landscaping, Duan doubts that there can have been much of it. Based on estimates of mercury production from the Song era and allowing for the imperfections of the earlier refinement process, he thinks the chamber might have contained at most 100 tons of the liquid metal: around 7 m3.
We might never be able to check that. “Right now, our archaeological work is focused on deducing the basic layout” of the tomb, says Duan. Because even a small breach in the seal could admit water or air that might damage whatever lies within, even robot-based exploration of the interior is ruled out. “If the chamber was opened even using a robot or drilling, the balance of the situation would be broken and the buried objects would deteriorate quickly”, says Zhang.
So if we’re ever going to peek inside, it will have to be with better scientific techniques than are currently available. “I dream of a day when technology will shed light on all that is buried there, without disturbing the sleeping emperor and his two-thousand-year-old underground empire”, says Yongqi Wu, curator of the Qin Shi Huang Mausoleum Museum at the Lishan site. Maybe these concerns to preserve the unknown heritage will guarantee Qin Shi Huangdi a kind of immortality after all.
Friday, January 02, 2015
There goes the neighbourhood
How would you like to live on Heinrich Himmlerstraat? Maybe not? Don’t worry – to my knowledge no street named after the SS Reichsführer exists. But if you do fancy living on a street dedicated to an ardent Nazi, the one pictured above will do you: it is Lenardstraat in Nijmegen. A Google search for this location will take you to the Nijmegen city website, where we’re told that “Philipp Lenard was a German physicist, and received the 1905 Nobel Prize in physics for his research on cathode rays and their properties.” Sounds worthy enough, huh? They forgot to mention that Lenard was also a fervent supporter of the Nazis who hosted Hitler in his home after the Führer-to-be was released from a Bavarian prison following the failed Beer Hall Putsch of 1923. Lenard was one of the main progenitors of “Aryan physics”, which he proclaimed as the only true physics, in distinction from the pernicious “Jewish physics” promulgated by Einstein and his acolytes. In 1924 Lenard and his associate Johannes Stark published an article called “The Hitler spirit and science”, in which they said that Hitler and his comrades “appear to us as God’s gifts from times of old when races were purer, people were greater, and minds were less deluded.” According to Lenard, all of the great scientists of former times (including Galileo!) were of Nordic-Aryan stock.
Should we hold all this against Lenard the physicist? Well, certainly it does not invalidate his important work on cathode rays and the photoelectric effect, any more than Stark’s adulation of Hitler should prevent us from recognizing his discovery of the Stark effect (which won him a Nobel too). But acknowledging scientific achievement and precedence is one thing; celebrating committed Nazis and anti-Semites by naming streets after them is another. (The University of Heidelberg quietly ditched the Philipp Lenard Institute – which once boasted of being “Jew-free” – after his death.)
You might think that the matter would be especially sensitive in the Netherlands, which suffered so greatly under the Nazis. Yet the Dutch academic who pointed out this road to me tells me that his efforts to raise the issue have got nowhere. “After seeking support from Dutch historians of science and sending a copy of my letter to the local press”, he says, “I received a letter from the mayor that a name change will not be considered since 80% of the street inhabitants resisted a change and nowadays nobody knows who he was.”
Well, this is who Lenard was: one of the most unpleasant of the distinguished scientists of the twentieth century. It’s of course a valid and difficult question at what stage (if ever) we draw a veil over the dubious character of a historical figure and simply recognize their achievements. But given the furious debate that still rages today in the Netherlands over how to think about wartime conduct, memories in this case seem disturbingly short.
Subscribe to:
Posts (Atom)