The idea that the universe constantly splits into alternate realities has had a surprisingly warm reception in science, philosophy and popular literature. But this baroque solution to the unique problems of quantum mechanics is simply not good science. Not only should we resist its strange appeal, it’s time to find another way, writes Philip Ball.
Alternative realities hold an irresistible allure. Whether it’s Dickens’ A Christmas Carol, Frank Capra’s It’s a Wonderful Life, or the quantum-computed parallel universes of Alex Garland’s recent TV series Devs, the possible lives that we can imagine having led but did not lead offer a stage for acting out our fears and fantasies. Like Robert Frost’s traveller confronted with diverging paths in a wood, we often wonder where the road not taken might have borne us.
It is no surprise, then, that the Many Worlds interpretation (MWI) of quantum mechanics seems to hold such attraction. Even though most physicists dismiss or even deride it, it is often eagerly embraced by physics popularizers and their audiences. Yet it can be hard to figure out how seriously some of its advocates really take it. I believe some physicists genuinely see it as an elegant solution to deep conundrums of the notoriously mind-bending quantum theory, and I sympathize with some of their reasoning. But when they start talking about “quantum brothers” (and presumably sisters, though Many Worlds has curiously few female advocates), or using “quantum apps” to make difficult decisions by triggering some quantum measurement in the conviction that they are thereby splitting off a world where they made the other choice, I have to wonder whether, indifferent to the philosophical complications, they are just enjoying the fantasy.
I have to wonder whether, indifferent to the philosophical complications, physicists are just enjoying the fantasy of the Many Worlds interpretation.
Why, though, would anyone decide that the theory of quantum mechanics reveals an unimaginably vast, possibly infinite, series of other worlds constantly disentangling from our own? In his recent piece for IAI News, Daniel Nolan explains that the notion arises from the probabilistic nature of the theory. As German physicist Max Born (who is too rarely given due credit) argued in the 1920s, quantum mechanics seems to have a curious property unlike any other theory in science. It seems incapable, in general, of predicting what the observed outcome of a quantum event will be, but only the probabilities of all the possible outcomes. For a classical object like a tennis ball or a space rocket, Newton’s equations of motions can tell us exactly what path it will follow under a given set of circumstances. But fire a photon of light at two slit-like openings in a screen, and quantum mechanics can only offer probabilities of the road taken.
Only when we measure the outcome of a quantum event can we get a definitive answer. A 50:50 probability of outcomes A and B before the event, say, somehow turns into one or the other (a 100% certainty) when we look at the event once it has happened. It’s like flipping a coin: there’s a 50:50 chance of heads or tails before we flip, but once we’ve flipped and looked, it’s either 100% heads or 100% tails. Yet we can confirm that the initial 50:50 probability was right by repeating the experiment – making many tosses, whereupon we'll find that we get the same number of heads, on average, as tails.
[NHY1]In most of science, probabilistic predictions reflect a lack of knowledge about the details. If we knew, at the outset, the positions and movements of every atom in the coin, the flipping hand, and the air, we should be able to predict the outcome of the coin toss with certainty. (That’s impossible in practice, but not in principle.) Yet in quantum mechanics the probabilities are fundamental: the theory seems to insist that there is nothing we can possibly measure that will enable us to do better than a probabilistic prediction of outcomes.
Quantum mechanics seems to insist that there is nothing we can possibly measure that will enable us to do better than a probabilistic prediction of outcomes.
The question is then what changes a probability before the event to a certainty after it. There seems to be some abrupt change that quantum theory itself can’t account for. The equation devised by Erwin Schrödinger in 1924 to describe how a quantum particle behaves contains no such transformation. It describes particles and their behaviour in terms of a mathematical entity called a wave function (because it resembles the equation for a classical wave). As Born showed, the wave function can be used to calculate the probability that we will measure a particular value of the particle’s properties – its position in space, say. But Schrödinger’s equation seems to insist that these properties always behave smoothly, whereas a measurement seems to induce an abrupt jump to a particular value: what Nolan refers to as a “collapse process” of the wave function.
Quantum mechanics doesn’t itself have any prescription for the “collapse of the wave function” – it has to be added by hand in an ad hoc fashion, as first done by the Hungarian mathematical physicist John von Neumann in the 1930s. That’s awkward and unsatisfactory – sure, the maths works okay, but we’re none the wiser about what actually happens in the real world to convert smooth quantum probability waves into unique observations. The Many Worlds interpretation was proposed by Hugh Everett in his doctoral dissertation at Princeton in 1957 as a way to avoid this problem. What if, Everett said, there simply is no wave function collapse that selects one possible outcome from all the possibilities? What if, instead, they all occur – but in alternative realities? In this view, as the photon reaches the double slits, it passes through one slit in one universe and the other slit in the other.
But then there must be an observer in each universe that sees the two outcomes. In other words, everything splits: the apparatus, the experimenter, the lab, the universe. Everett was discouraged from saying much about this multiplication of actual people (as opposed just to particles) in his thesis, but the notion was there. It was expanded on in the 1970s, and the proliferation of quantum selves is now the typical “Wow!” element of popular presentations of the MWI.
Lev Vaidman, John Ellis and David Wallace explore whether there are many worlds.
In the meantime, some of the questions left hanging by Everett – in particular, how the splitting” actually happens – have been clarified. The process by which a quantum event (happening to a particle, say) becomes registered as an observable outcome in some human-scale measuring device is known now to involve a phenomenon called decoherence, in which the quantum state of the particle gets mixed up (“entangled”) with those of the particles in its environment. Many Worlders argue that Everett’s splitting is a kind of unravelling, brought about by decoherence, in which the two (or more) possible outcomes inherent in the pre-event universe separate and can no longer influence one another. They are to all intents then manifested in separate universes, coexisting “in the same space” but oblivious to one another.[N
In this way, Many Worlders argue, the problems with wave function collapse evaporate. We just have to accept in return that the universe is constantly unravelling into parallel realities.
That’s too baroque and extravagant a price for many people. For one thing, these parallel worlds can never be observed – by definition, for it is only by becoming causally independent that they truly become other worlds. If, in contrast, the two outcomes don’t decohere but remain able to interact, they can “interfere”: a classic manifestation of quantum behaviour. But how, some ask, can we accept as truly scientific a proposition that can never be demonstrated to be true? Many Worlders reply that the truth of their interpretation is there for all to see in the Schrödinger equation itself; the onus is instead on other interpretations of the theory, which are forced to add something to the equation to explain how it “breaks” to give a unique outcome.
How, some ask, can we accept as truly scientific a proposition that can never be demonstrated to be true?
In other words, while some object that the Many Worlds Interpretation is poor science because it is too profligate with universes, Many Worlders say that it is in fact the most parsimonious interpretation in terms of assumptions. The truth is that neither of these positions is a strong argument.
A more serious objection to Many Worlds Interpretation is that it doesn’t explain how probabilities enter into quantum mechanics via Born’s rule for extracting an expected outcome from a wavefunction. How can we say that there is a 50% probability of outcome A (which we can verify experimentally) if the truth is that all outcomes always happen (in some world or other) with 100% probability? Some Many Worlders say that these probabilities should be understood instead as the “weighting” of the branches in a quantum split – what Nolan calls the “intensity”. But that doesn’t in itself mean much. If outcomes A and B have probabilities of 75 and 25%, say, this doesn’t make the universe in which B happens any less real, absolute or robust than that for A.
Some Many Worlders argue that we should understand these apparent probabilities as the chances that we should assign to ending up in each possible world. In other words, they are subjective probabilities describing individual experience, even though a “version of me” will end up in every universe with 100% probability. There are several variations on this idea: one asserts that Born’s rule[EM1][ itself can be derived from the subjective view of how a rational observer would bet on quantum outcomes. That’s to say, even though an observer would know (if the Many Worlds view is right) that versions of her will definitely end up in all possible worlds, each version will only be aware of one of them – and she can place bets before a quantum split about which outcome “she” will later observe.[EM3[P
Rather poignantly, the MWI banishes exactly what it seems to need in order to make sense.
But this is where the MWI comes tumbling down. For in attempting to banish the awkwardness of wave function collapse, the idea has become forced to install individuals – people, making conscious decisions! – as active ingredients that can “explain” how particles are seen to behave. That no longer seems so parsimonious, does it? In fact it is much worse than that – because, although few Many Worlders will even deign to consider the matter, the MWI obliterates our ability even to talk about individuals – to speak meaningfully about what “you” or “I” will observe. There is, in short, no meaningful and unambiguous way to connect the “you” before the quantum “split” to the “you” after it. We think we can imagine what that means, and some physicists even try to defend the notion with talk of malfunctioning Star Trek transporters or clones or the like. But looked at with philosophical clarity and honesty, the image of splitting selves placing bets on the future dissolves into incoherence. Rather poignantly, the MWI banishes exactly what it seems to need in order to make sense.
That the MWI advertises itself as an antidote to the sleight-of-hand of wave function collapse is a red herring too. For the whole concept is now redundant: a careful consideration of what happens when we make a measurement on a quantum system reveals that “collapse” is no longer a useful concept.There is now, in contrast, a fairly good picture of how the possibilities inherent in a quantum system gradually tip one way or another through the interactions with its environment that ultimately cause decoherence and give rise to definite, classical outcomes. Many Worlders seem keen to cling to this obsolete straw man of collapse, but quantum physics is moving on. That’s not to say we can explain all of its puzzles – but it now seems unlikely that the intriguing, inventive and bold suggestion of proliferating universes will help very much. It’s time to take another road.