Updated 29 Mar 2021
Retrocausality is the idea that the chain of cause-and-effect can sometimes be reversed in time. From relativity to quantum physics to parapsychology, retrocausal phenomena have long been theorised but have never yet stood the test of, er, time. But now, that is beginning to change. The weird and baffling consequences of quantum mechanics are such an awkward fact of life that in desperation to explain them, some physicists have been revisiting the idea of retrocausality. It helps to explain a lot of quantum puzzles which are hard to explain any other way. On the other hand, it can also open the way to major paradoxes of its own.
Thermodynamics, which describes the flow of heat and energy, was the first of the three modern pillars of physics to be established. As such it is a "classical" or macroscopic theory whose foundational ideas are drawn from the old world view of Galileo and Newton. One of its unspoken assumptions is that causality, the flow of cause-and-effect, by definition moves forwards in time: it is by observing which event happens at the earlier time that we define the cause and the later event is defined as the effect. This mindset is enshrined in the Second Law of Thermodynamics where disorder, or entropy, always increases as time passes.
Retrocausality flatly contradicts the unspoken assumption about cause-and-effect in time and as such must, at a classical level, violate the Second Law. But is the assumption actually valid? Is the Second Law a wholly universal one? As the next two pillars, relativity and quantum mechanics emerged, they began bending and sometimes even breaking the old classical ideas. A new science of quantum thermodynamics is only now beginning to emerge and, as one might expect, is showing signs of surprises in store.
The first modern scientific discussion of retrocausality came in the wake of Einstein's theory of relativity. Although the relativistic nature of time could lead different observers to see events happening in a different order, such events could not be causally connected. However it was soon realised that an object travelling faster than light would appear to be moving backwards in time, and Einstein's equations did not explicitly forbid this. Hence the classic limerick from 1923;
There was a young lady named Bright
Whose speed was far faster than light;
She went out one day
In a relative way
And returned on the previous night.
Since relativity sets the speed of light as an absolute limit, at first such speculation seemed a bit nonsensical. Later it was realised that relativity also allowed faster-than-light objects whose lowest speed limit was that of light. Such "tachyon" particles behave just like the good Miss Bright of the limerick. But there is not the slightest evidence that such particles actually exist.
Perhaps the most notorious retrocausal phenomenon in physics is the idea of time travel, with its paradoxes such as Miss Bright's, or going back to kill your own mother before you were born.
A wormhole in spacetime, creating a closed timelike loop along which the matricidal maniac might travel at a more leisurely speed, is consistent with relativity theory. However the creation of such a wormhole requires an assortment of highly speculative things such as dark energy and dark matter, and such gargantuan quantities of them, that Stephen Hawking suggested a kind of cosmic censorship in which the laws of physics do not allow such things to exist.Moreover, even if such a thing could be created, there are reasons to believe that getting onto the timelike loop and one end, and off at the other, might prove impossible.
Retrocausality first appeared in quantum physics when Paul Dirac proposed the existence of an anti-particle to the electron, the positron, and realised that it could be understood as an electron travelling backwards in time. An embellishment to this was the idea, discussed by John Wheeler and Richard Feynman, that there was in fact only a single electron-positron, which bounced to and fro in time so it looked like a huge number of particles. But they could not explain why the positron is much less common, so they dropped the idea. We nowadays accept it as a different kind of particle, moving forwards in time.
A later variant came with the sum-over-histories approach to the quantum wave equation, notably with the work of Feynman. Until then, deriving the wave equation for a particle meant ignoring some of the maths for the sake of convenience and just picking the bits that gave the right answer. Feynman followed up the maths that included every conceivable path in spacetime between the quantum's start and end points. Not just in space, note, but in spacetime. That means not just everywhere but everywhen. To make the theory work, the beginning and end of time both have to be blocked off so that the quantum cannot escape altogether but effectively bounces back. When all these paths are added together (summed), they interfere with each other and most of them cancel each other out. Amazingly, what gets left behind is the usual quantum wave function, the bit that physicists were already picking out. It's a neat way of avoiding that cherry-picking exercise, at the expense of allowing the entire future of the universe to affect the present moment. Nothing else useful came out of the idea, so it remained a minor curiosity of little practical relevance.
What really changed things for quantum theory was the discovery that quantum entanglement acted instantaneously across spacetime; it was the "spooky action at a distance" that Einstein so hated. If you set up two entangled particles and at some point force the properties of one, then you also force the properties of the other. Measuring the second a moment later confirms this. But it is possible for a relativistic observer, say in a passing spaceship or near a black hole, to see the measurement event occurring before the forcing event. They are faced with a choice between either the measurement forcing the decision on forcing, or retrocausality with the effect happening before its cause. The first option has grave consequences for any rational theory at all, for how can a particle the other side of the room force a choice made via some arbitrary experimental technique? The second option is generally held to be an example of retrocausality. Technically the principles of causality, realism and localism are to some extent interchangeable and some physicists prefer to sacrifice one of the others rather than temporal causality. Local realism - the combination of locality and realism - is already ruled out by experiments investigating the consequences of Bell's theorem, so why drag causality into it too?
Here, the world of science seems to be split. Some physicists take their cue from classical thermodynamics, rule out retrocausality as a matter of principle and interpret the quantum equations accordingly. Others are prepared to consider retrocausality if it turns up in the maths. Quantum weirdness includes things like wave-particle duality, uncertainty, superposition of states and nonlocality. Many of these aspects have made similarly contentious entrances over the years and mainstream opinion is hardening towards the view that, whatever form any such weirdness might be recast in or reinterpreted as, if it appears in the maths and is borne out by experiment then it must ultimately be accepted as an intrinsic fact of nature.
The superposition of quantum states has recently been shown to include causal states along the flow of time, with different causal orderings superimposed such that no definite causal flow exists. A "Bell's Theorem" for quantum retrocausality has also been proposed, in principle allowing us to test it experimentally. For now, quantum retrocausality remains a work in progress.
It seems reasonable to expect that, as with the relativistic undermining of simultanerity, quantum retrocausality will prove strictly limited in scope. Relativity does not affect the underlying temporal ordering of causally related events. However in a quantum context the above recent findings suggest that we might need to distinguish the causal sequence of events from their temporal flow. This would allow for example that, while the sequences of quantum events in time may be locally inconsistent between relativistic obervers, the underlying causal sequence remains unaffected.
But modern information theory has not really been aware of this distinction. For example the laws of thermodynamics, relativity and quantum mechanics must obviously apply to physical information. But it is less obvious how they should apply to symbolic information; none of the these three pillars of physics has anything to say about the meaning of a symbol, or can express the idea that it might be false. (Note that in classical information theory, garbled or corrupted information occurs when physical information is mixed with unwanted physical information from another source. It says nothing qualitative about what happens to the symbolic meaning carried by the signal).
On this basis, there is no immediate reason why symbolic information should be bound by thermodynamic causality, i.e. why it should be time-bound as opposed to merely logic-bound. A symbol may propagate backwards in time without breaking any physical laws, but to achieve that its physical carrier must also be able to do so. And that, of course, must depend on how quantum retrocausality pans out.
The behaviour of information in quantum systems is an even more recent crossover between these various theories. Crucially, it is beginning to bring insights into how they all hang together and how some of the unresolved issues within them relate to each other. In classical thermodynamics, information always increases (or, in some versions, it is the potential capacity to store information increases). So, to work backwards from an outcome to its cause, typically requires more information to work with than simply following the cause through to its effect does. But in quantum mechanics the amount of information remains constant throughout. Clearly, the concepts of "information" in these two disciplines must be different. One hopes that the still-emerging field of quantum thermodynamics may one day clarify such issues.
We already know that the amount of quantum information needed to describe any given system is less than the amount of classical information required for a thermodynamic description. This is attractive to computational modellers, who are always seeking elegant shortcuts to speed their calculations. But it can lose any sense of direction in time, challenging our ideas on our measurement or perception of that direction. And that brings us back to one of the central conundrums of quantum mechanics, the role of the conscious observer in determining the outcome of any experiment, along with its accompanying paradoxes such as Schrödinger's cat and Wigner's friend.
Where there is a quantum experiment, there is a conscious observer, and where there is such an observer, there is a mind built of quantum interactions. Might retrocausality occur in such a context?
Precognition, as postulated in theories of parapsychology, is an example of retrocausality at the macroscopic level. No psychic phenomenon of any kind has yet been reliably demonstrated to mainstream satisfaction. But then, as I have come to discover, most have not been reliably falsified either. Sceptical investigators have been every bit as guilty of bad science as the proponents. The scientific case is not as closed, either way, as many would have us think.
If quantum effects can indeed travel backwards in time and carry symbolic information with them (which is still debatable), and if brain physiology relies on quantum effects for its function (which it does) then technically this might leave open the door to retrocausal mental phenomena such as precognition. In particular, JW Dunne theorised that precognition references one's own future experiences rather than any external event. Such a model at least confines the phenomenon within a single human brain and thus avoids the need for arbitrary actions-at-a-distance, rendering any conceivable mechanism that much simpler and less implausible. Other potentially retrocausal psi phenomena which inherently extend beyond the brain, such as clairvoyance and prophecy, would require far wider manifestations of quantum weirdness and are to that extent far less plausible.
We are still a long way from a useful theory of quantum retrocausality, or of reconciling it with information theory. We are even further from any hope of discovering it at work in a conscious brain. A further caveat, if one were needed, is that so far related weirdness effects have been identified only in a limited number of well known, if incompletely understood, physiological phenomena such as ion channels across nerve synapses. Attempts to provide quantum-weird paradigms for unexplained aspects of the mind, such as memory, have so far proved untenable. In the main they seem to have been ill-conceived, with their proponents apparently lacking sufficient appreciation of quantum theory or neurology or both. The implication of microtubules in particular has given the field a bad name.
Expanding the influence of quantum weirdness to demonstrate precognition would be an enormous step. How likely it might be must depend on one's personal point of view. Nevertheless, its presence in the brain at an inter-molecular level perhaps lowers the credibility barrier a little to the idea that, as the field of quantum causality develops, its recognition of temporal retrocausality might allow modern physics to remain consistent with precognition.
1. Viatko Vedral; "Law and Disorder", New Scientist, 7 April 2018, pp.32-35.
2. Adam Becker; "Blast from the Future", New Scientist, 17 February 2018, pp.28-31.
3. Kelly Oakes; "When causality Breaks", New Scientist, 18 January 2020, pp.34-7.
4. Zych, M., Costa, F., Pikovski, I. and Brukner, C.; "Bell’s Theorem for Temporal Order", Nature Communications, 10, 3772, 2019. DOI 10.1038/s41467-019-11579-x
5. Leah Crane; "Quantum Thinking Can Ignore the Flow of Time", New Scientist, 28 July 2018, p.7.
6. JW Dunne; An Experiment with Time; AC Black, 1927 (many later editions).