A WORLD WITHOUT CAUSE AND EFFECT Logic-defying experiments into quantum causality scramble the notion of time itself.

BY PHILIP BALL

lbert Einstein is heading out for his This finding1 in 2015 made the quantum the constraints of a predefined causal structure daily stroll and has to pass through world seem even stranger than scientists had might solve some problems faster than con- Atwo doorways. First he walks through thought. Walther’s experiments mash up cau- ventional quantum computers,” says quantum the green door, and then through the red one. sality: the idea that one thing leads to another. theorist Giulio Chiribella of the University of Or wait — did he go through the red first and It is as if the physicists have scrambled the con- Hong Kong. then the green? It must have been one or the cept of time itself, so that it seems to run in two What’s more, thinking about the ‘causal struc-

other. The events had have to happened in a directions at once. ture’ of quantum mechanics — which events EDGAR BĄK BY ILLUSTRATION sequence, right? In everyday language, that sounds nonsen- precede or succeed others — might prove to be Not if Einstein were riding on one of the sical. But within the mathematical formalism more productive, and ultimately more intuitive, photons ricocheting through Philip Walther’s of quantum theory, ambiguity about causation than couching it in the typical mind-bending lab at the University of Vienna. Walther’s group emerges in a perfectly logical and consistent language that describes photons as being both has shown that it is impossible to say in which way. And by creating systems that lack a clear waves and particles, or events as blurred by a order these photons pass through a pair of flow of cause and effect2, researchers now haze of uncertainty. gates as they zip around the lab. It’s not that this think they can tap into a rich realm of pos- And because causation is really about how information gets lost or jumbled — it simply sibilities. Some suggest that they could boost objects influence one another across time doesn’t exist. In Walther’s experiments, there the already phenomenal potential of quantum and space, this new approach could pro- is no well-defined order of events. computing. “A quantum computer free from vide the first steps towards uniting the two

590 | NATURE | VOL 546 | 29 JUNE© 20172017 Mac millan Publishers Li mited, part of Spri nger Nature. All ri ghts reserved. ©2017 Mac millan Publishers Li mited, part of Spri nger Nature. All ri ghts reserved.

FEATURE NEWS

cornerstone theories of physics and resolving this logic has unravelled over the past decade, and 1, the second qubit experiences a causal one of the most profound scientific challenges as researchers have realized that it is possible superposition of both sequences — meaning today. “Causality lies at the interface between to imagine quantum scenarios in which one there is no defined order to the particle’s tra- quantum mechanics and general relativity,” simply can’t say which of two related events versal of the gates (see ‘Trippy journeys’). says Walther’s collaborator Časlav Brukner, a happens first. Three years later, Chiribella proposed an theorist at the Institute for Quantum Optics Classically, this situation sounds impossible. explicit experimental procedure for enact- and Quantum Information in Vienna, “and so True, we might not actually know whether A or ing this idea5; Walther, Brukner and their it could help us to think about how one could B happened first — but one of them surely did. colleagues subsequently worked out how to merge the two conceptually.” Quantum indeterminacy, however, isn’t a lack implement it in the lab1. The Vienna team uses of knowledge; it’s a fundamental prohibition a series of ‘waveplates’ (crystals that change a TANGLES IN TIME on pronouncing on any ‘true state of affairs’ photon’s polarization) and partial mirrors that Causation has been a key issue in quantum before a measurement is made. reflect light and also let some pass through. mechanics since the mid-1930s, when Einstein These devices act as the logic gates A and B challenged the apparent randomness that Niels AMBIGUOUS ACTION to manipulate the polarization of a test pho- Bohr and Werner Heisenberg had installed at Brukner’s group in Vienna, Chiribella’s team ton. A control qubit determines whether the heart of the theory. Bohr and Heisenberg’s and others have been pioneering efforts to the photon experiences AB or BA — or Copenhagen interpretation insisted that the explore this ambiguous causality in quantum a causal superposition of both. But any outcome of a quantum measurement — such mechanics3,4. They have devised ways to create attempt to find out whether the pho- as checking the orientation of a photon’s plane related events A and B such that no one can ton goes through gate A or gate B first of polarization — is determined at random, say whether A preceded and led to (in a sense will destroy the superposition of gate and only in the instant that the measurement ‘caused’) B, or vice versa. This arrangement ordering. is made. No reason can be adduced to explain enables information to be shared between A and Having demonstrated causal indetermi- that particular outcome. But in 1935, Einstein B in ways that are ruled out if there is a definite nacy experimentally, the Vienna team wanted and his young colleagues Boris Podolsky and causal order. In other words, an indeterminate to go further. It’s one thing to create a quantum Nathan Rosen (now collectively denoted causal order lets researchers do things with superposition of causal states, in which it is sim- EPR) described a thought experiment that quantum systems that are otherwise impossible. ply not determined what caused what (that is, pushed Bohr’s interpretation to a seemingly The trick they use involves creating a special whether the gate order is AB or BA). But the impossible conclusion. type of quantum ‘superposition’. Super­positions researchers wondered whether it is possible to The EPR experiment involves two particles, of quantum states are well known: a spin, for preserve causal ambiguity even if they spy on A and B, that have been prepared with inter­ example, can be placed in a superposition of up the photon as it travels through various gates. dependent, or ‘entangled’, properties. For and down states. And the two spins in the EPR At face value, this would seem to violate the example, if A has an upward-pointing ‘spin’ experiment are in a super­position — in that idea that sustaining a superposition depends (crudely, a quantum property that can be pic- case involving two particles. It’s often said that on not trying to measure it. But researchers tured a little bit like the orientation of a bar a quantum object in a superposition exists in are now realizing that in quantum mechanics, magnet), then B must be down, and vice versa. two states at once, but more properly it simply it’s not exactly what you do that matters, but Both pairs of orientations are possible. But cannot be said in advance what the outcome of what you know. researchers can discover the actual orientation only when they make a measurement on one of the particles. According to the Copenhagen interpretation, that measurement doesn’t just AN INDETERMINATE CAUSAL ORDER LETS reveal the particle’s state; it actually fixes it in that instant. That means it also instantly fixes RESEARCHERS DO THINGS WITH QUANTUM the state of the particle’s entangled partner — however far away that partner is. But Einstein SYSTEMS THAT ARE OTHERWISE IMPOSSIBLE. considered this apparent instant action at a distance impossible, because it would require faster-than-light interaction across space, a measurement would be. The two observable Last year, Walther and his colleagues which is forbidden by his special theory of rela- states can be used as the binary states (1 and 0) devised a way to measure the photon as it tivity. Einstein was convinced that this invali- of quantum bits, or qubits, which are the basic passes through the two gates without immedi- dated the Copenhagen interpretation, and elements of quantum computers. ately changing what they know about it6. They that particles A and B must already have well- The researchers extend this concept by cre- encode the result of the measurement in the defined spins before anybody looks at them. ating a causal superposition. In this case, the photon itself, but do not read it out at the time. Measurements of entangled particles show, two states represent sequences of events: a par- Because the photon goes through the whole however, that the observed correlation between ticle goes first through gate A and then through circuit before it is detected and the measure- the spins can’t be explained on the basis of pre- gate B (so that A’s output state determines B’s ment is revealed, that information can’t be existing properties. But these correlations don’t input), or vice versa. used to reconstruct the gate order. It’s as if you actually violate relativity because they can’t be In 2009, Chiribella and his co-workers came asked someone to keep a record of how they used to communicate faster than light. Quite up with a theoretical way to do an experiment feel during a trip and then relay the informa- how the relationship arises is hard to explain like this using a single qubit as a switch that tion to you later — so that you can’t deduce in any intuitive cause‑and-effect way. controls the causal order of events experienced exactly when and where they were when they But what the Copenhagen interpretation by a particle that acts as second qubit3. When wrote it down. does at least seem to retain is a time-ordering the control-switch qubit is in state 0, the parti- As the Vienna researchers showed, this logic: a measurement can’t induce an effect cle goes through gate A first, and then through ignorance preserves the causal superposition. until after it has been made. For event A to gate B. When the control qubit is in state 1, the “We don’t extract any information about the have any effect on event B, A has to order of the second qubit is BA. But if measurement result until the very end of the happen first. The trouble is that that qubit is in a superposition of 0 entire process, when the final readout takes

©2017 Mac millan Publishers Li mited, part of Spri nger Nature. All ri ghts reserved. ©2017 Mac millan Publishers Li mited, part of Spri nger Nature. All ri ghts r29ese rJUNEved. 2017 | VOL 546 | NATURE | 591

NEWS FEATURE TRIPPY JOURNEYS place,” says Walther. “So the outcome At the quantum level, the normal rules of cause and eect need not apply. two decades, some physicists and of the measurement process, and the To demonstrate this, researchers have sent photons through a series of mathematicians, including Hardy10 gates with an ambiguous causal order. The set-up is analogous to a train’s 11 NATURE time when it was made, are hidden path through a pair of gates governed by a two-way switch. and Brukner , have sought to clarify but still affect the final result.” things by building ‘quantum recon- Other teams have also been cre- structions’: attempts to derive at The switch in this If the qubit switch is NIK SPENCER/ ating experimental cases of causal case is a quantum set to state 0, the least some characteristic properties ambiguity by using quantum optics. bit, or qubit, which train passes through of quantum-mechanical systems — For example, a group at the University selects the route station A and then such as entanglement and superposi- that the train 0 station B. This route of Waterloo in Canada and the nearby takes. A has a de nite tions — from simple axioms about, Perimeter Institute for Theoretical causal order. say, what can and can’t be done with Physics has created quantum circuits 1 the information encoded in the states that manipulate photon states to pro- (see Nature 501, 154–156; 2013). duce a different causal mash-up. In “The framework of causal models effect, a photon passes through gates provides a new perspective on these A and B in that order, but its state questions,” says Katja Ried, a physi- is determined by a mixture of two B cist at the University of Innsbruck in causal procedures: either the effect of Austria who previously worked with B is determined by the effect of A, or the University of Waterloo team the effects of A and B are individu- on developing systems with causal ally determined by some other event The train now ambiguity. “If quantum theory is a acting on them both, in much the passes through theory about how nature processes Alternatively, the station B and then same way that a hot day can increase qubit can be set 0 station A, again and distributes information, then sunburn cases and ice-cream sales to state 1. A with a de nite asking in which ways events can without the two phenomena being causal order. influence each other may reveal the directly causally related. As with the 1 rules of this processing.” Vienna experiments, the Waterloo And quantum causality might go group found that it’s not possible to even further by showing how one assign a single causal ‘story’ to the can start to fit quantum theory into state the photons acquire7. the framework of general relativity, Some of these experiments are B which accounts for gravitation. “The opening up new opportunities for fact that causal structure plays such a transmitting information. A causal central role in general relativity moti- superposition in the order of signals vates us to investigate in which ways If the qubit is set travelling through two gates means to a superposition it can ‘behave quantumly’,” says Ried. that each can be considered to send of 1 and 0, it is The system has an “Most of the attempts to under- information to the other simultane- impossible to say ambiguous stand quantum mechanics involve whether the train 0 sequence in time: ously. “Crudely speaking, you get goes through A an inde nite trying to save some aspects of the two operations for the price of one,” A or B rst. causal order. old classical picture, such as particle says Walther. This offers a potentially 1 trajectories,” says Brukner. But his- powerful shortcut for information tory shows us that what is generally processing. needed in such cases is something Although it has long been known Any attempt to more, he says — something that goes that using quantum superposition measure the path of beyond the old ideas, such as a new and entanglement could exponen- the train destroys the way of thinking about causality itself. B superposition and tially increase the speed of computa- the train follows only “When you have a radical theory, to tion, such tricks have previously been one de ned path. understand it you usually need some- played only with classical causal struc- thing even more radical.” ■ tures. But the simultaneous nature of pathways in a quantum-causal superposition It’s not terribly complicated to build the Philip Ball is a freelance writer in London. offers a further boost in speed. That potential necessary quantum-circuit architectures, 1. Procopio, L. M. et al. Nature Commun. 6, 7913 was apparent when such superpositions were either — you just need quantum switches (2015). first proposed: quantum theorist Lucien Hardy similar to those Walther has used. “I think this 2. Brukner, Č. Nature Phys. 10, 259–263 (2014). at the Perimeter Institute8 and Chiribella and could find applications soon,” Brukner says. 3. Chiribella, G., D’Ariano, G. M., Perinotti, P. & 3 Valiron, B. Phys. Rev. A 88, 022318 (2013). his co-workers independently suggested that 4. Oreshkov, O., Costa, F. & Brukner, Č. Nature quantum computers operating with an indefi- UNITY IN THE UNIVERSE Commun. 3, 1092 (2012). nite causal structure might be more powerful The bigger goal, however, is theoretical. 5. Chiribella, G. Phys. Rev. A 86, 040301(R) (2012). than ones in which causality is fixed. Quantum causality might supply a point of 6. Rubino, G. et al. Sci. Adv. 3, e1602589 (2017). 7. MacLean, J.-P. W., Ried, K., Spekkens, R. W. & Last year, Brukner and his co-workers entry to some of the hardest questions in physics Resch, K. J. Nature Commun. 8, 15149 (2017). showed9 that building such a shortcut into an — such as where quantum mechanics comes 8. Hardy, L. Preprint at http://arxiv.org/abs/ information-processing protocol with many from. quant-ph/0701019 (2007). 9. Allard Guérin, P., Feix, A., Araújo, M. & Brukner, Č. gates should give an exponential increase in Quantum theory has always looked a little Phys. Rev. Lett. 117, 100502 (2016). the efficiency of communication between ad hoc. The Schrödinger equation works mar- 10. Hardy, L. Preprint at http://arxiv.org/abs/ gates, which could be beneficial for computa- vellously to predict the outcomes of quantum quant-ph/0101012 (2001). 11. Dakić, B. & Brukner, Č. in Deep Beauty: tion. “We haven’t reached the end yet of the experiments, but researchers are still arguing Understanding the Quantum World through possible speed-ups,” says Brukner. “Quantum about what it means, because it’s not clear Mathematical Innovation (ed. Halvorson, H.) mechanics allows way more.” what the physics behind it is. Over the past 365–392 (Cambridge Univ. Press, 2011).

592 | NATURE | VOL 546 | 29 JUNE© 20172017 Mac millan Publishers Li mited, part of Spri nger Nature. All ri ghts reserved.