Misconceptions about quantum physics (that I learned from scientists)
Surely the fault is in us, and not in our wavefunctions.
Topic: quantum physics
Thesis: wave-function collapse is the most misleading and misled idea in popular quantum mechanics.
Disclosure: the Academic, an intellectual cherub, born in grace, cites primary academic sources.
I, an intellectual Cainite, born in sin, cite Wikipedia.
In this world all of us are divided by chasms of belief. We disagree on whether the unfolding human catastrophes we see around us derive from malice or mere incompetence; we disagree on our ethical frameworks, our values and tolerances, our visions of a just society. Yet despite our many points of disagreement with even our closest allies, there is perhaps one bridge that unites us all, one point of contact between all the denizens of the global mental world, and that is a failure to understand quantum mechanics.
For as long as I have tried to understand quantum mechanics, certain vexing questions have rattled about in my mind. In conversations with students of the field, more patient thinkers on the topic than I have succeeded, through their instruction, in knocking those questions to the darker corners, so that in moments of bright optimism I could actually convince myself that I understood what was going on.
I want to address some of the things I got wrong as a result of these conversations, and show, ultimately, how I got them right. The title of this article overstates things a bit: these misconceptions were not necessarily taught. Rather, they came about as the result of a collision between my thinking and the instruction of people who had spent time studying the topic. Nevertheless, these misconceptions appear to be common enough that it is worth the time to document them and show how they are wrong. Perhaps some future student of the topic will go less far astray than I did.
Groundwork
I am writing this for people who have spent significant time grappling with weird ideas of quantum mechanics, for example superposition, entanglement, and wavefunction collapse; and accordingly, the guiding examples of the double-slit experiment and Schrödinger’s Cat. An understanding of the underlying mathematics is not required; my own understanding of it is cursory at best.
For a refresher on the basic ideas, I recommend these explainer videos:
Superposition, entanglement, and collapse (via quantum computing): tinyurl.com/veritasium-quantum-computing
Entanglement: tinyurl.com/veritasium-entanglement
Schrödinger’s Cat: tinyurl.com/scishow-schrodingers-cat
Double-slit experiment: tinyurl.com/dr-quantum-double-slit (still the best explainer I can find on the issue; note that the use of an “eye” to represent the measuring device is misleading because the measuring device must interact with what it is looking at in a way that an eye would not.)
Misconception Number One: “consciousness never figured into the scientific consensus on quantum physics”
The misconception: the “New Age” crowd, that woo-woo cadre of shysters and hucksters, is responsible for the misunderstanding that wavefunction collapse is a result of conscious observation. They took advantage of the uncertainty and obscurity of quantum mechanics, combined with its surprising results, to assert that consciousness fills the gap in the explanation of the experiments, thereby leaving the door open for unscientific ideas like manifesting reality. End of misconception.
“Consciousness causes collapse” is the three-word summary of the von Neumann-Wigner interpretation [wiki]. It is no longer a mainstream view, but it is false to assert that no prominent physicists advocated it.
Initially, this is surprising, given the apparent absurdity of the view. To bring it down to just one aspect: quantum-probabilistic effects are necessary to nuclear fusion, the process that powers all stars. If “consciousness causes collapse” were true, stars would be in a superposition of shining and not-shining until observed by something. Accordingly, earth would be in a superposition of shined-upon and not-shined-upon, life in a superposition of formed and not-formed, humans in a superposition of evolved and not-evolved. These superpositioned humans, being only potential observers until actuated by wavefunction collapse, would not be able to will themselves into being. We have a first-mover problem for observation.
We could escape this by saying that:
Potential observers can retroactively collapse the wavefunction that brings them into being, creating a stable time-loop (some may find this easy to believe, but to me it borders on the mystical); or
Something other than human observation can cause wavefunction collapse, be it a rare spontaneous accident or the scrutiny of an intermittently interested God. At this point the interpretation loses all explanatory power: if wavefunctions can collapse arbitrarily, or for incorporeal reasons, why assume that human consciousness is necessary for collapse at all?
So the von Neumann-Wigner interpretation has conceptual problems. Its initial development, beginning in 1932 and persisting at least through the rest of the ‘30s, is understandable: the importance of quantum mechanics to the structure of the cosmos was not well-understood at this time. Yet its persistence to the present day is evidence of the explanatory gap that still exists between the mathematics of wavefunctions and human intuition about quantum mechanics.
Misconception Number Two: “there is such a thing as wavefunction collapse”
The misconception: in the majority view among physicists, wavefunction collapse is a real event: it can be situated in spacetime, like the precise instant of the ticking of a clock. The event of collapse is known to be caused by the physical process of taking a measurement. End of misconception.
The first reason the above statement is a misconception is because of the first few words: “the majority view among physicists”. In fact the ontology (manner-of-being) of wavefunction collapse is a contentious issue among physicists, contentious enough that one may fairly deny the existence of a consensus on the matter.
Let’s consider some of the popular answers. The dominant interpretations continue to be the many-worlds interpretation and the Copenhagen interpretation. Let’s leave aside the many-worlds interpretation for now; we’ll come back to it later.
Since the Copenhagen interpretation is more a school of opinion than a single unified position, it is better to describe it by laying out some opinions that fall under the umbrella, leaving to one side how they accord with one another:
Wavefunction collapse is not real, only apparent to an observer making a measurement of the system.
Measurements always influence a system, so it’s wrong to use information from a measurement to ask what would have happened if the measurement had not been taken.
Collapse is a step in the formal process of translating a quantum system into classical terms.
When enough particles become involved in a system, quantum effects disappear.
The mathematical model of a quantum system is more reliable than any description of the system using common-sense terms. Avoid trusting language when analyzing quantum systems. Trust the math.
Taken all at a glance, the picture is fuzzy. Wavefunction collapse is triggered by observation, but associations of many particles can also destroy quantum effects? Wavefunction collapse bridges the formalisms of quantum and classical systems, but it also changes the predicted outcome?
An important clarifying distinction, proposed [wiki] early on by Neils Bohr, is between the “collapse” of physical systems that happens on its own, without measurement, and the “collapse” that determines which of several possibilities will be observed. The former, observer-independent collapse we’ll call “decoherence”, named as such by David Bohm in 1952 [wiki].
In “decoherence”, interacting particles force each other into a superposition of non-interfering classical states, regardless of whether an observer is present or not. In “collapse”, one of these classical states is chosen. Subsequently, or simultaneously, a classical observer comes along and notices it.
By making this distinction, we avoid some of the weirdness about the physical role of an observer. But collapse remains mysterious. By definition, it cannot be observed; an observation can, at best, only capture the immediate aftermath of it. Furthermore, it appears to be a process with a random element to it: via the work of H. Dieter Zeh beginning in the 1970s, we can mathematically model how a quantum system decoheres into multiple independent states (“dead cat” and “live cat”, in the infamous thought experiment) [wiki], but no model exists to explain how one of these states is “chosen” for the observer’s benefit.
Importantly, the moment of collapse, where one state becomes real and the others are forever ruled out, cannot be bound to a single instant: it occurs somewhere between decoherence, where the states become mathematically separated, and the moment of observation. In fact, it’s a little worse than that, because we can still ask: if an experiment has multiple observers, which observer forces collapse to have occurred? Is it the first? The last? Or is it relative to the observer?
This is actually the essence of the Schrödinger’s Cat paradox. By choosing a cat for his example, Schrödinger left it ambiguous whether the cat was an “observer” from the quantum standpoint. Can the cat determine “for itself” whether it is alive or dead in the box? Or does it remain in a superposition of states until a human checks on it? What if it were a child in the box? The real horror of it, the real existential queasiness of it, comes from the fact that, to the observer outside of the box, there is no way to distinguish between the two cases—that is to say, between a superposition of decoherent states and a single realized outcome.
In sum, we can say that the Copenhagen interpretation does not paint a picture of wavefunction collapse as a real event—or, if it does, it does not give that event a strict physical definition. Over and above the conceptual difficulties, the Copenhagen interpretation places an emphasis on proceeding with the mathematical work of quantum mechanics, and not worrying too much about the failures of common-sense intuition to follow along. In this it has actually been phenomenally successful, owing to what Eugene Wigner called “the unreasonable effectiveness of mathematics”. But its dominance in the field has also left the door wide open for intuitionist attempts at understanding quantum physics, including the von Neumann-Wigner interpretation: the already-addressed view, still persistent in the popular understanding, that consciousness causes collapse.
The alert reader will notice the two mentions of Wigner, close together, and wonder if they are in fact the same Wigner. The answer is: yes, they are. If I could shake this Wigner awake from his eternal slumber, I would ask him what he considered to be the unreasonably effective mathematics of consciousness.
What about wavefunction collapse in other interpretations? There is also the so-called “pilot-wave” or de Broglie-Bohm theory [wiki]. This attempts to describe quantum mechanics in terms of classical measurements of a quantum wavefunction. It avoids the idea of collapse by positing a real configuration of the system relative to a classical observer, with the quantum wavefunction governing how that real configuration evolves. Hence it replaces the problem of collapse with the problem of explaining where classical observers come from and why it cannot be that there are alternative observers for each of several quantum outcomes, in a way that’s analogous to the “consciousness causes collapse” model.
Some models do attempt to extend quantum mechanics to describe wavefunction collapse as an objective physical event. These have their problems, though. For one example, take the Ghirardi–Rimini–Weber theory or GRW [wiki]. According to GRW, collapses occur with random outcomes, at random times; the weirdness of quantum experiments happens because the interaction of such a large number of particles with previously isolated quantum states means that at least one collapse is almost guaranteed to happen, a collapse which must then force the collapse of the entire system. All well and good—except that the proposed mechanism violates the law of conservation of energy.
What about the many-worlds interpretation? What does it say about wavefunction collapse?
Misconception Number Three: “many-worlds means mitosis”
The misconception: according to the many-worlds interpretation, wavefunction collapse causes the entire universe to split, producing a parallel universe for each possible classical outcome. End of misconception.
I confess that in this case I resist my own framing: it is difficult to plainly argue that the above is a misconception. The reason for this is that any school of thought held in common between people will contain some differences in view, so that any attempt at a summary must somewhat conflate what is and what should be the mainstream opinion of the school.
The history of the many-worlds interpretation [wiki] begins with Hugh Everett’s formulation in 1957 and its subsequent popularization by Bryce DeWitt. What’s key to know is that it was DeWitt who coined the term “many-worlds”, not Everett. Everett’s foundation was to suppose that the Schrödinger equation applies to the universe as a whole: that is, that the universe is one wavefunction, unbroken for all time. There is no such thing as collapse.
Does Everett’s view contradict DeWitt’s popularization? This question brings us to the heart of the problem with the popular conception of many-worlds. The answer depends on the event that causes splitting. Originally, Everett was forced by his model to assume that measurement played some role as the direct cause of splitting, because without it there seemed to be no way to break a superposition down into choices. Put simply, why does the superposition of Schrödinger’s cat have live cat–dead cat as the two “worlds”? Why not, instead, ¼ living–¾ dead? This is the so-called “preferred basis” problem [wiki].
So Everett and DeWitt did both commit themselves, however unwillingly, to a measurement event that plays a role in the division of worlds. Such an event is collapse by another name; at least, it fits the definition of collapse I have used so far. It also carries the same conceptual baggage as “collapse” in the Copenhagen interpretation, only magnified infinitely by the burden of generating universes wherever it should occur. As late as 1970, DeWitt is quoted as saying: “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 of itself.” [wiki] Perhaps by “quantum transition” he meant something different than “wavefunction collapse”; if he did, I am unable to determine what it might be.
So what I have catalogued as a “misconception” is actually an almost perfect description of the many-worlds interpretation as it stood fifty years ago. It was also in 1970 that the work of H. Dieter Zeh set many-worlds on its path to its modern foundation. Zeh is known for formalizing the idea of “decoherence” mentioned earlier, showing how a system that exists in a superposition of states, with each state influencing the evolution of the others, can evolve into isolated states that cannot interact.
Let’s sketch out this notion of decoherence. The first thing to know is that, when we model quantum wavefunctions, we’re modeling them in a type of mathematical space. This is alternately called the “configuration space”, “phase space”, or “possibility space”. For every possible state of the system, there is a corresponding point in phase space. By the nature of quantum mechanics, the system is never only at a single point. It is always spread out, if only slightly, over a region of phase space. It may be spread out in such a way that there are multiple distinct “peaks”. This would be what we’d expect when modeling a photon that has traveled through a barrier with two slits: a peak for Slit A, a peak for Slit B.
A phase space can have any number of dimensions. The more particles are involved in a system, the more dimensions are required to model it.
In a phase space with only a few dimensions, the evolution of one “peak” can influence the evolution of another. This explains the appearance of interference patterns in the original two-slit experiment, even when just one photon participates.
In a phase space with many dimensions, distinct peaks quickly dissipate into the surrounding chasm of possibility. They get lost in it, and can’t interact anymore. The different peaks in the system are like travelers in a world of many rooms. If each room has only two doors, then two travelers who begin in the same room are highly likely to encounter each other again; if each room has trillions of doors, the chance is negligible.
When translated into classical terms, the diverging peaks may be describing two very similar possible systems, occupying the same spatial coordinates; but in the phase space, disagreement on the position of just a few particles flings them so far apart that they are fated to evolve independently forever.
When we take a classical measurement of a simple quantum system, we give it reams of new dimensions as the particles of the measuring device interact (entangle) with the simple system. The two peaks, previously overlapping, now have so many degrees of freedom that they diverge permanently. The new compound system of particle-with-measuring-device still has two peaks: in the two-slit example, one for Slit A and one for Slit B. Each of these systems may then interact with other systems, such as an observer, creating still more divergent peaks of world-state.
We can see how naturally this account fits into many-worlds. I, the observer, see the photon pass through Slit A, because I, the observer, occupy the peak in phase space representing Slit A; but the other peak, and the other me, still exists as a “parallel world”. Far from seeming conceptually overloaded, this now becomes the more natural interpretation. The demon of “collapse” is finally and thoroughly exorcised.
But since decoherence seems to so neatly replace “collapse” in the many-worlds model, it’s important to note how it fails to fill the anticipated role. Decoherence shares the drawback with collapse that it is not a definite event that can be situated in spacetime. As the system evolves, states overlap less and less until they become effectively decoherent: the degree of their interference may hurtle exponentially toward zero, but (seemingly) is never fully gone. [I corresponded about this on Quora with retired physicist Peter Voke, who was kind enough to answer my questions.]
If we fit this into DeWitt’s 1970 explanation of many-worlds, it becomes strange. When I observe the result of the two-slit experiment, I split the whole universe—but for some unknown interval, the whole universe is only “somewhat” split into parallel universes, and always capable in some remote future of colliding together again?
Zeh himself, perhaps concerned with the same difficulty, proposed to extend the many-worlds interpretation with a “many-minds interpretation” [wiki]. In this model there is no talk of splitting into parallel universes, at least not for physical things. But the minds of observers, persisting in time, “follow” one peak in phase space instead of the other.
To me—and though I do not cite anyone on this view, neither do I contend that it is original—the simplest way out of the difficulties with many-worlds and decoherence is to suppose that, when a system becomes decoherent, nothing comes into being. The existing system becomes metaphorically “folded”, with each fold evolving independently, according to the same rules as the original. The decoherent system evolves through time, interacting with other systems and producing larger entangled systems with larger multiplicities of entangled states. Each state within the superposition may, according to the mathematics of uncertainty, evolve into its own distinct superposition, which may in turn produce another decoherence and another multiplication of possibilities.
I like to call this the “many-islands” or “archipelago” interpretation. It’s not the universe that splits, but entangled systems within it, which in turn interact with other systems to create expanding islands of superposition. At the same time, as each island evolves, parts of it become disentangled from each other: that is, the state of one part of the system ceases to depend on the state of some other part. In this way the islands break up, so that the universe never settles into one single entanglement, and the multiplication of possibilities goes on.
Again, that’s all well and good; but an experimenter watching her photons may well ask: “What about me?”
“Why”, she asks, “do I experience only the island-state where the photon was observed passing through Slit A, when the other one was equally possible?”
To discuss this question properly, we must break free of the constraints of the format I have chosen for this article. I will follow up with another article about what many-worlds, or many-islands, means for our understanding of consciousness. Subscribe to this newsletter to be notified the moment it arrives.
Update: the follow-up post has arrived! Check it out: