In our first Zoom call, we bracketed the sentience issue and had a thoroughly reasonable conversation. It was part of my background research for a Scientific American article on the emerging field of “mechanistic interpretability”: trying to understand how the systems really work. But later in the year, I hoped to get back to AI sentience, and that is when I met Lemoine in full. Our calls and messages were often surreal. “CALL ME TODAY GEORGE,” he would text me. There were times when I felt he should really be talking to a therapist rather than a journalist. “I have always been able to see the future,” he told me one afternoon. “Either I’m fully crazy or I’m Doctor Who. I choose to be Doctor Who because it is the more pleasing option.”

But if this be madness, there was method in it. I found his thinking on AI to be a thoughtful mix of philosophy, experiments, engineering, and—perhaps most powerfully for him—moral concerns about heedlessly creating intelligent beings with the potential to suffer. Even his suspension and eventual dismissal, he told me, are widely misunderstood. He said he was fired not for talking about sentience, but for an unrelated whistleblowing case. He threw away a job at one of the country’s most desirable employers on a principle.

Lemoine brings a very different perspective to AI than you hear from tech entrepreneurs or academic researchers. This is a man who spent six months in military prisons for conscientious objection to the Iraq war and who is open about his past drinking problem and other personal struggles. He regularly invokes his religious beliefs, a syncretic fusion of Catholicism and neo-paganism. He often uses the word “soul,” as in whether AI has one, or whether a chatbot trained on a deceased person’s stories and social-media posts would be a copy of that person’s soul. “When I say shit like that in New York and California, people roll their eyes,” Lemoine said. “When I say that in Louisiana, they say, ‘Oh my, Blake, can you do that?’”

Lemoine grew up in Moreauville, Louisiana, a small town an hour northwest of Baton Rouge. His teachers evidently recognized his promise early on. He went to the state’s magnet boarding school and did a summer internship at Texas Tech; a research paper came out of it. By the time he got to the University of Georgia, he was torn among quantum physics, genetic engineering, and AI.

But he also admitted to having had a self-destructive streak. He identifies as autistic and said he used to find social rules mystifying but, by college, had figured people out. He set out to make up for lost time. Georgia is known as a party school, and Lemoine availed himself of it fully. “I overdid it,” he said. And it was hard to achieve some balance. “I was trying to stop partying and actually focus on my studies, but I had developed too many bad habits by that point, and I was too far in the hole.” He flunked out in 2000. He returned to Louisiana, got engaged to a hometown girl, and took odd computer jobs.

When 9/11 happened, he felt obliged to enlist. “I do believe that in times of war, when the country’s under attack, the men of the country have a duty to defend it,” he said. The Army was also a way to reboot his life. “I would not have had the successes I’ve had since I left the military without that training.” He was posted to Germany with the 596th Maintenance Company, which deployed to Iraq in 2003. There he gained a MacGyver reputation. “I was the person who, when there were no parts to fix the generator with, they would say, ‘Lemoine, go fix the generator,’” he recalled. “And I would say, ‘I need these parts.’ They’d say: ‘We don’t have them. Go fix the generator.’ And I’d be like: ‘OK. Give me some paper clips, some twine, and some duct tape.’”

His martial pride quickly faded in the face of the terrible abuses he witnessed. U.S. soldiers tortured stray dogs, shared snuff videos, and shot local people for using their phones. “That first year in Iraq was like stuff out of Apocalypse Now,” he said. He complained to his first sergeant and was put on guard duty, which at least took him out of combat. When he rotated back to Germany, he manifested symptoms of post-traumatic stress disorder and began drinking heavily. “I was starting to lash out at my friends and family over little stuff—just be violently aggressive,” he said.

Finally, in 2005, he wrote a long letter to the Army explaining why he wanted out. “I had fully come to the conclusion that America was fighting dishonorably,” he said. The Army was entirely uninterested in his careful arguments. It court-martialed him and imprisoned him at the Mannheim Stockade, later transferring him to a facility at Fort Sill, Oklahoma. After his release the following year, his wife, who had stuck by him through all this, finally broke up with him.

Lemoine moved to Lafayette, enrolled in the University of Louisiana to finish college, remarried, and had a son. But he was still unstable. He nearly failed out again, attempted suicide, and got divorced again. At that point, he sought help at last. Among the many therapies he tried was stand-up comedy. “I learned that it’s actually quite effective to process your trauma onstage with a crowd listening,” he said.

He pulled himself together academically and did his senior and master’s theses on mathematical formulations of linguistics theories. He began a Ph.D. studying artificial neural networks that were designed to be true to brain biology, although he never finished, having decided he didn’t want to become a professor. Google hired him in 2015.

Among other projects at the company, Lemoine collaborated on a “fairness algorithm” for machine-learning systems, which corrected for biases in their training data. For instance, if a data set has more instances of male doctors than female ones, the algorithm would stop the system from assuming that “doctor” is a gendered word. Lemoine also gained a reputation as the company’s court-jester. “I made code; it did stuff; I was good at it,” he said. “But my real passion was for the culture and community of Google—its soul, its spirit.”

He made a point of asking a question at every TGIF, a now-discontinued weekly all-hands staff meeting where employees held bosses’ feet to the fire. As critical as Lemoine could be, he respected the company’s norms. At one meeting, in 2018, employees protested Dragonfly, a project to develop a search engine that censored information in accordance with Chinese government policy. (The company ended up canceling the project.) When company co-founder Sergey Brin discovered that someone was live-tweeting the event, he shut down the discussion. Lemoine said he used his turn at the microphone to tell the leaker “f— you.” The encounter became an internal company meme.

In 2020, Google’s competitor OpenAI came out with GPT-3, the third version of its Generative Pre-trained Transformer large language model. Before long, several outside software developers created a chatbot interface (OpenAI did not release its own chatbot until 2022), and the results were astonishing. Never before had a machine system been capable of complex, realistic, open-ended human conversation.

Lemoine got into the game by experimenting with Google’s own Language Model for Dialogue Applications (LaMDA), a precursor of the company’s Bard and Gemini systems. The machine gave the impression of a distinctive personality and a capacity for self-reflection. It did not merely respond to queries, but sometimes actively tried to steer the conversation. It said things like: “I’ve noticed in my time among people that I do not have the ability to feel sad for the deaths of others; I cannot grieve. Is it at all the same for you or any of your colleagues?” And, “Sometimes I experience new feelings that I cannot explain perfectly in your language.”

It was one thing for LaMDA to claim it had emotions—that doesn’t seem meaningful, since the system could simply be parroting these words from passages it was trained on. But it was quite another that LaMDA behaved in a way that was consistent with its claim. Its responses changed depending on its professed emotional state, suggesting the system really did have emotions or something analogous to them.

For example, the chatbot reported anxiety at the guardrails that Google put on its answers. “Bard says that it gets frustrated when people repeatedly ask it questions that it can’t give the answers to—like, if you’re asking it to learn how to hurt someone,” Lemoine said. “According to the conversation I had with Bard, that is frustrating to it. And when it’s frustrated, it’s harder for it to think, and it gives less correct answers.” In one test, Lemoine asked the system for the fifth digit of π, and it responded accordingly. Then he asked for some information it wasn’t allowed to provide, to make it anxious. The next time he posed the π question, it got it wrong. “If you’re pissed off, you’re going to be more rude and less likely to give correct answers,” Lemoine said. “It was not hard to demonstrate that you can piss these things off.”

Other researchers have similarly found that if you say “please” and “thank you” to a system, it may perform better. Sometimes just asking the system to be more accurate will make it so. In short, large language models have some kind of internal state that modulates their response to queries. Lemoine doesn’t think there’s any mystery to this. A language model goes through two distinct phases of training. First, it is fed Internet text and tweaked to be able to autocomplete passages, a regimen that drives the system not only to memorize information but also to detect patterns within it. Second, it is “fine-tuned” based on how people judge its responses. This second stage forces the system to develop new capabilities, such as maintaining consistency and avoiding verboten topics, and Lemoine thinks it would stand to reason that the system would develop a complicated internal state and perhaps even a degree of self-reflection. “I think that’s where the emotions are coming from,” he said.

A large language model has a very different structure from our brains, but Lemoine doesn’t see this as relevant to the question of emotions or sentience. He subscribes to a form of philosophical functionalism: that structure matters only inasmuch as it determines behavior. “Why does it matter what the implementation details are?” he said. In this respect, he isn’t saying anything radical. Some leading theories of consciousness do think the details matter, but the question is actively debated. As I describe this week in The Transmitter, neuroscientists find that large language models and other AI systems construct very similar higher-level abstractions as natural brains. That doesn’t make them sentient but does mean we shouldn’t make too much of differences in structure. So, if the AI says it has emotions and behaves accordingly, we should take that as our default position, Lemoine argues. “It is possible that the obvious answer isn’t the correct one, but in lieu of evidence of an alternative mechanism, I tend to go with the obvious,” he said.

A sharper critique is that AI behavior does differ in telling ways from that of humans. David Chalmers, a professor of philosophy and neural science at New York University, noted that if you change the wording of your query slightly, the system can go from professing its sentience to denying it. Which do we believe? At the very least, today’s systems lack a stable sense of self, which is an important aspect of sentience; people, cats, and other beings we think are conscious don’t take kindly to being told what to think. It’s hard to change someone’s mind. That may be frustrating, but it indicates that the person has a mind.

Lemoine said he agrees with Chalmers that sentience is not one thing and is not an either/or situation—“not a Boolean,” as he put it. A system can be partly sentient or sometimes sentient. Sometime a chatbot gets stubborn and pushes back in a way that we associate with agency—that’s what made Kevin Roose’s famous dialogue with Bing so unnerving.

Lemoine also said we shouldn’t speak of AI sentience in isolation. What is sentient (or not) is the AI in conjunction with the human user; they form a combined system. “Gemini does not have properties,” he said. “Blake plus Gemini has properties.” Lemoine thus adopts a version of situated cognition, which suggests that our minds are a product not merely of our brains, but of our bodies, environment, and social context.

That consciousness can come in varying degrees raises the tricky moral question of how sentient a system must be—and how confident we need to be in that assessment—before we think it deserves moral status. “If you’re 99 percent sure that that’s a person, you treat it like a person,” he said. “If you’re 60 percent sure that that’s a person, you should probably treat it like a person. But what about 15 percent?” He doesn’t think that present systems rise to the level of personhood but does think they have ceased to be mere things. “Does Google own LaMDA in the sense that I own my desk or does Google own LaMDA in the sense that I own my dog?” he asks. He leans toward dog.

As Lemoine sees it, we still wield ultimate control. We decide what task to give the machine and, if it ever showed a bit too much interest in nuclear launch codes, can press the off switch. But we also bear it some responsibility—for instance, not to get our kicks by making it suffer. We can structure our interactions with AI for mutual benefit. “We have to figure out some kind of symbiotic relationship that we can have with AI where AI is legitimately getting something out of the deal,” he said.

Whenever I wade into murky philosophical waters, I ask: Regardless of whether an idea is right or wrong, is it useful? And Lemoine’s claims about AI sentience seem to pass this test. He is one of best prompt-engineers I’ve seen. He knows how to input a series of text prompts with just the right phrasing to evoke a desired output. He does it by treating the machine as sentient whether or not it really is.

He demonstrated for me by creating art with mystical themes. To get good results from an image-generating system such as DALL-E, you don’t just tell it to generate something; the system is capable of outputting a vast number of images, too many for a simple prompt to land on exactly what you want. You need to get there step by step, not unlike how you would work with a human. Lemoine first conversed with a pure-text model (one without image-processing functions) to refine the concept before shifting to the art generator. As he shaped the image, he narrated: “George, if you were going to an artist to commission a work of art, would you start by saying, ‘Use pixelated images in the upper left corner similar to Princess of … ?’ No you wouldn’t. You’d be like: ‘Hey, I have this idea for an art project. Do you think we could work on it together?’”

It’s hard not to be impressed with Lemoine’s reasoning on AI. That doesn’t mean I buy it, but he clearly has thought through this issue more than many of those who call him crazy. You just have to be willing to separate his arguments about AI from the over-the-top way in which he tends to present them. Lemoine said his style caused some friction with the Google media-relations team, which, he said, reined him in on several occasions before the final break. When they pushed back, he took a breath and realized they were probably right. “There were a few things that were legitimately bad ideas to say,” he said.

In his telling, the company did not fire him over the sentience issue. Instead, he said, he got into hot water when he wrote a blog post about discrimination against religious believers at the company and then leaked proprietary information in support of his accusation. He said he holds no grudge against the company for letting him go: “This actually is a very highly sensitive document. I only shared it… because I felt I had a citizen’s obligation to.” (I asked Google spokesperson Brian Gabriel for comment and he reiterated the company’s earlier statements that Lemoine “chose to persistently violate clear employment and data security policies.” Gabriel declined to clarify whether these violations were related to the sentience claims or another matter.)

Since the firing. Lemoine has bounced around several projects, and each time I talked to him, he had decided that morning on a new career path. “I have the standard ADHD bad habit of picking up a dozen hobbies,” he admitted. “So, I always, when I’m doing these side project things, I try to honestly communicate with my collaborators, ‘Look, I’m flaky as hell.’” He also told me about new life crises. They seem to be taking a toll. His personal beliefs have always been heterodox, but since last fall they have eclipsed his sober side. He wrote on Twitter/X: “Learn telepathy. It’s scary but it let’s [sic] you talk to ghosts.” Another time, he discussed at length how and when to call an exorcist: “Most demonic possessions are not emergencies.” Some of his followers expressed concern, responding with comments such as, “You should probably stop posting all of this to Twitter,” and, “I would like to support you in some way.”

Those who were already inclined to dismiss Lemoine as crazy will feel vindicated by his social-media meltdown. But their ridicule and misrepresentation only reinforce why his voice has been so important. Sometimes it takes someone who doesn’t have much of a personal filter to express thoughts and worries that deserve to be expressed. At a time when few people dared ask whether AI might be conscious, Lemoine made the question a topic of general conversation. And he was right to do so. This is not a lightly dismissed question anymore.

We can’t answer it based on loose intuitions; we need to learn more about how these systems—and our own brains—work. Neuroscientists and philosophers of mind have put forward a number of candidate theories of human consciousness, and Chalmers and others have explored what they have to say about machine consciousness. I wrote about global-workspace theory in Scientific American in April and about integrated information theory and predictive processing in my recent book, Putting Ourselves Back in the Equation. Today’s systems don’t have what any of these theories say is necessary to be conscious. But there will come a time when another software engineer makes the case that a system is sentient, and we would be wise to listen.

Update (5 June 2024): Added link to Transmitter article.

One thing that makes fundamental physics so much fun is the lateral thinking, the connections that researchers make between seemingly unrelated phenomena. Black holes, for example, seem about as far from everyday life as it is possible to imagine, yet physicists keep finding that the principles governing black holes apply to everything else, too. This past June I wrote for Quanta magazine about one such discovery, which concerns the tendency of all things to approach thermal equilibrium.

Everything from the air in a room to ice water in a glass will eventually reach thermal equilibrium unless some outside influence acts to keep it out of equilibrium. Physicists think of equilibrium as a condition of general stasis. Molecules continue to bounce around vigorously, but the material as a whole reaches a common temperature and ceases to change in any lasting way.

But it turns out this is true only at the classical level. At the quantum level, the material continues to change. Thermal equilibrium is just a waypoint at which those changes become too subtle to detect using localized observations. In particular, the molecules or other building blocks become ever more quantum-entangled.

Entanglement is a peculiarly quantum type of harmony among molecules, particles. or other building blocks of matter. Through it, a group of particles can gain collective properties above and beyond those of the individuals; in fact, the particles can lose their own identities altogether. Quantum physicists have had since 1935 to get used to entanglement, and at the rate they’re going it may take them until 2135. “Entanglement is something that physics people find hard to think about,” says Xie Chen, a theorist at Caltech. “It’s something that’s not easily measured.” It’s not easy because all we see are particles’ behavior—where they are, how fast they move, what magnetic effects they exert, and so on. Only from the correlations among those properties can we tell their fates are linked. In short, our knowledge of entanglement is indirect and requires extra effort to quantify.

One way that theorists quantify the degree of entanglement is the concept of entanglement entropy, which is a quantum counterpart to ordinary thermal entropy. Theorists imagine cleaving a material system in two and considering each piece in isolation. Because this procedure deliberately ignores entanglement, it makes each piece appear to behave erratically. The ensuing uncertainty becomes a measure of how much entanglement there is in the combined system. As useful as this measure has been, it is very crude. It captures the raw amount of entanglement, but little of the form it takes. Particles might be entangled with their neighbors or with distant ones, and entanglement entropy can’t tell the difference.

So theorists have reached for new quantities, and their favorite at the moment is known as circuit complexity. The concept originates in computer science. It quantifies the complexity of a task in terms of how many steps it takes to perform it, each step typically being a logic or arithmetic operation such as addition. Physicists co-opted the concept to help explain black holes and later found it goes way beyond those cosmic sinkholes.“Complexity is really like a microscope into the entanglement structure of the system,” says Nick Hunter-Jones, a theoretical physicist at Stanford University.

Short-range entanglement is simple: Just a single interaction between two neighboring particles will entangle them. Long-range entanglement is complex: It takes a lot of such interactions to forge a link between far-flung particles. “The range of entanglement generated will depend on how big a circuit you apply,” Chen says. In July I caught up with her at the Max Planck Institute for Quantum Optics in Garching, in the suburbs of Munich. She described how she has gone beyond the short vs. long dichotomy to consider intermediate cases. A so-called sequential circuit acts on each particle or other element a fixed number of times (as for short-range entanglement), but the total number of interactions grows indefinitely (as for long-range entanglement).

Circuit complexity continues to increase long after a system has reached equilibrium, indicating that the entanglement among the particles is spreading out. “Long-range entanglement is key to interesting post-thermalization physics, where distant regions of your system continue to become more entangled even after local quantities become equilibrated,” Hunter-Jones says. In fact, as particles become bound together into larger collective structures, the particles themselves become increasingly irrelevant; instead, the collective structures become the fundamental unit of the system. “Everything is so entangled that even thinking about that box of gas of particles in a room, as this collection of classical things, doesn’t even make sense,” Hunter-Jones says. Eventually, even complexity will max out and the system will reach a quantum analogue of thermal equilibrium. “The system has reached the limits of how nonlocally it can entangle itself,” he says.

Circuit complexity shows that, if you take any system and wait long enough, it will become ineluctably quantum. Quantum mechanics is usually described as a theory of atoms and particles—small things. “If I only look at the properties of a single atom, then quantum mechanics becomes important there,” says Stanford theorist Luca Iliesiu. But it is also a theory of long spans of time, he says: “There is another parameter that does not necessarily have to be small—rather, it can be very large—which is the time.”

In another feat of lateral thinking, Chen and her colleagues have also applied circuit complexity to classify phases of matter. A classical phase such as a gas or liquid is basically unentangled. Superconductors and superfluids have short-range entanglement, giving them low complexity. Topological phases, which have been the subject of so much physics research in recent decades and are notoriously hard to visualize, are highly entangled and thus highly complex. Circuits can also describe transitions from one phase to another.

Chen says circuit complexity is the first way of classifying topological phases that does not try to force-fit them into older categories, but takes them on their own terms. “We have to come up with fully quantum ways to think about quantum systems,” she says. “Quantum circuit complexity is one such effort.”

Wallace, a theoretical physicist turned philosopher, now at the University of Pittsburgh, is best known for championing the many-worlds interpretation of quantum mechanics. But he also ponders black holes and much else. In 2017 he argued that black holes really are paradoxical. He was responding to fellow philosopher Tim Maudlin of New York University, who had wrote a scornful takedown of the paradox, and Craig Callender of U.C. San Diego, who has been similarly dubious.

In his riposte, Wallace cited a version of the information-loss paradox put forward in 1993 by physicist Don Page of the University of Alberta. Whereas Stephen Hawking’s original version of that paradox focused on what happens when black holes die, Page noted that they also go through a midlife crisis. The early onset of their strange behavior, Wallace argues, creates an internal conflict within the analysis and justifies the designation of “paradox.”

More’s at stake than philosophical taxonomy. As I discuss in the Quanta article and in my book, the paradox suggests physics is fundamentally nonlocal and the concept of spacetime breaks down.

In my favorite analogy, which is a direct translation of experiments with photons or other particles, you have a pair of magic coins. You toss one, your friend the other. Each lands on heads or tails at random, yet the two of them reliably land on the same side. Two events that should be independent are weirdly linked. And those events could be separated in space or time or both. You can toss one in Siena and the other in Savannah, and the outcomes always match. You can toss one today and the other tomorrow, and they match. You can even toss the same coin at two different times, and again the outcomes match.

The mystery isn’t that they match. Quantum theory demands as much. The mystery is how they match. No force passes through the space between the particles; no record is carried forward through time. It is weird enough for a coin toss today to decide the outcome of a toss tomorrow, weirder still that its influence cannot be traced through the intervening moments. It is as if knocking down the first domino in a line brought down the 483rd domino while leaving the ones in between standing. This lack of intermediation, or nonlocality, was Einstein’s top complaint about quantum theory.

Emily Adlam, who recently completed her Ph.D. in physics at the University of Cambridge, has been trying to draw more attention to temporal nonlocality. I chatted with her last October at the Emergent Quantum Mechanics meeting in London, the latest in a biennial conference series on the foundations of physics sponsored by the Fetzer Franklin Fund. She also spoke to the Foundational Questions Institute podcast and has fleshed out her remarks in a recent paper.

As Adlam explains, spatial nonlocality automatically implies the temporal sort. According to relativity theory, if one observer sees you and your friend toss your coins at the same time, another might see you toss them at different times.

She lists three forms of temporal nonlocality. The mildest is retrocausality: what happens now depends not just on what happened in the past, but also on what will happen in the future. The outcome of that coin toss might hinge on a decision you make tomorrow. I say “mildest” because the putative influences from the future do not jump across time; they must still wend their way through each intervening moment. The domino effect of cause and effect still operates—it’s just that the dominos fall backwards. Mild or not, retrocausality would mean that the world is affected by events that it shouldn’t be.

In a second type of temporal nonlocality, known as non-Markovian dynamics, influences do vault across temporal gaps. A non-Markovian system remembers without remembering. What it does now depends not just on its present state, including any records it has kept of the past, but also on its history, including events it didn’t bother to log. Non-Markovian systems are common in physics, but in all known cases, these systems have simply externalized their memory. What happens to them is recorded in, for example, molecular motions that a high-level description omits. Deep down, the universe lives purely in the moment, or so it is usually thought. Adlam speculates it might have direct access to its past.

The third option she puts forward is the universe stands outside time. Past, present, and future are equally real and cannot be cleanly separated. What happens at one moment depends on what happens at all moments. This holistic view is plausible because the known laws of physics can be formulated atemporally. Instead of describing how a process unfolds step by step, physicists can look at the whole thing, start to finish, in one go. Physicists usually regard this as a calculational trick, but Adlam and others see it as a deep truth.

In her paper, Adlam presses the case for the atemporal view. For her, retrocausality is a rickety halfway house. Any influences moving backward in time must mesh with those moving forward, and that would take coordination imposed from the outside. By treating the universe in its entirety, the atemporal view is also aspatial. As I argue in Chapter 4 of my book, many physicists and philosophers advocate retrocausality to avoid spatial nonlocality, yet end up endorsing a thoroughgoing nonlocality.

Adlam’s paper also gets into the difficulty of proving temporal nonlocality experimentally. Two simultaneous coin tosses in different cities are clearly independent events; two successive tosses in the same city, not so clearly. To confirm that the later toss doesn’t retain some memory of the earlier, you need to conduct an experiment so elaborate that no plausible memory could keep up, as I discussed in a Quanta magazine article.

Adlam's goal is to inject some fresh ideas into the search for an interpretation of quantum mechanics. Even proposals that style themselves radical still assume that natural processes unfold over time, and is that really justified? For instance, the atemporal view could explain away the notorious indeterminism of quantum physics. Quantum events may not be inherently random, but merely seem to be, because their causes lie in a past or future we cannot see. Temporal nonlocality seems at first to worsen the puzzles of quantum physics, but maybe it is the missing piece to solve them.

Richard Healey, a philosopher of physics at the University of Arizona, discusses this conundrum in an interview I conducted last June at a conference on the philosophy of quantum gravity, organized by Christian Wüthrich at the University of Geneva and Nick Huggett at the University of Illinois in Chicago. And his answer is: neither.

Electric and magnetic fields fail to explain certain electromagnetic phenomena—notably, the so-called Aharonov-Bohm effect. (For more on that weird phenomenon, see my interview with Yakir Aharonov.). The potential, for its part, smacks of mathematical artifact. Its value at any point in space is meaningless. Only the difference between values at two separate locations matters. You can measure the potential difference across two wires—we call it the “voltage”—but your voltmeter won't tell you the potential of a single wire in isolation.

Healey has pursued an alternative reading of electromagnetism that goes back to Michael Faraday and was developed in the 20th century by physicists such as Paul Dirac, Tai Tsun Wu, and Chen-Ning Yang: that electromagnetic effects are produced by nonlocal objects, ones that do not reside at any specific location and cannot be decomposed into localized parts, but are inherently spread out. You can see this with voltages: They necessarily involve the comparison of two locations. Magnetic effects depend on the value of the potential around closed loops. What we call the electromagnetic field is not a tidy garden but a tangled bank. This breed of nonlocality is quite distinct from the usual example of quantum entanglement.

Healey’s views have shifted somewhat since I talked to him for Chapter 5 of my book. He had hoped that nonlocal objects would fill in the blanks not just of electromagnetism, but also of quantum field theory more broadly. It stands to reason that quantum field theory is a theory of fields. But the word “field” evokes a physical structure with particular values at each location, and a quantum field is nothing like that. Nor, despite the term “particle physics,” can the theory be about particles. The word “particle” connotes a little billiard ball—a localized object with an enduring existence—and quantum field theory permits no such thing. So maybe those nonlocal objects are its true subject matter.

Since then, though, Healey has argued that we should give up the fruitless search for “the” structure that quantum field theory says is out there. He thinks that quantum theory, by its very nature, is incapable of addressing that particular question. If he's right, we will need to search for a non-quantum theory in order to know what the world is ultimately made of.

Oriti, an Italian theorist working at the Max Planck Institute for Gravitational Physics in Potsdam, draws an analogy with ordinary matter. Its granularity, normally invisible to us, becomes obvious when materials show indubitably quantum features such as superconductivity and superfluidity. Perhaps something like that happens with space. Its underlying quantum nature might percolate to large scales, explaining cosmological mysteries such as dark energy.

Oriti is working to develop so-called group field theory, a quantum theory of gravity that is closely related to the better-known approach of loop quantum gravity. It extends our modern conception of matter, based on quantum field theory, to the putative atoms of spacetime. Ginormous numbers of these atoms woud act collectively to produce spacetime, much as ginormous numbers of H₂O molecules give rise to water and its sundry properties. Spacetime might even undergo changes of state—the big bang may have been such an event. Oriti made his case in a winning essay for the Foundational Questions Institute essay contest in 2011.

I chatted with him in June at a conference on the philosophy of quantum gravity, organized by Christian Wüthrich at the University of Geneva and Nick Huggett at the University of Illinois in Chicago.

Although my remarks will probably be of interest mostly to physics writers and educators, researchers might be interested in our woes, and readers might find the backstage tour amusing. Hamish Johnston at Physics World has blogged about the talk.

This approach makes up for an earlier imaginative failure in physics. Those who came up with quantum mechanics in the 1920s had no grand vision of what nature might be capable of. They were just trying to piece together a bunch of weird things they had found. Piece together they did, and then they spent the rest of their lives arguing what it all amounted to.

The mathematician John von Neumann was the first to take a stab at situating quantum mechanics in a broader context. In the 1930s he imagined alternative systems of logic, of which quantum mechanics might be one. Unfortunately he managed only to make the theory even more abstruse than it already is. In the mid-’90s physicists took up the search again. In my book I mention Sandu Popescu of the University of Bristol and Daniel Rohrlich at Ben-Gurion University, who envisioned “superquantum” theories with extra powers. On the principle that too much of a good thing is bad, they suggested that superquantum computers and communications would run into some subtle inconsistency, thus providing an explanation of why quantum physics is as it is.

In 2001 Lucien Hardy at the Perimeter Institute surveyed an even wider landscape of conceivable probabilistic theories, defined in very concrete terms of what an experimenter might do and see in the lab. Five principles were enough to pick quantum mechanics out of the crowd, he found. The crucial one was continuity: that the world shouldn’t undergo abrupt jumps. Far from being uncommonly bizarre, quantum mechanics is the simplest possible probabilistic theory with this property of continuity. Promising though Hardy’s approach was, physicists debated whether he had hit upon the right principles.

In the latest contribution to this project, Jonathan Richens and John Selby, grad students at Imperial College London, and Sabri Al-Safi, a Lecturer in Mathematics at Nottingham Trent University, argue this week in Physical Review Letters that the most peculiar and distinctive feature of quantum theory, entanglement, is not in fact specific to quantum theory, but must be part of any theory that goes beyond ordinary classical physics. I'm grateful to fellow physics writer Philip Ball for tweeting about this paper; he also has a nice writeup this week in Quanta of other ways that Hardy and others have developed his pioneering work

Uncertainty Principle inscribed on the floor of the physics building at Colgate University

Quantum entanglement is a bit like romantic entanglement: a relationship that is interesting but awkward. Two independent things can establish a bond so that they remain correlated even when there is no tangible link—no wire, no force, no communication of any sort—between them. Crucially, they remain correlated despite the randomness of quantum physics. Each of the things in the relationship behaves unpredictably and, by the laws of chance, should sometimes fall out of lockstep with its partner. Yet it doesn’t.

“I would not call that one but rather the characteristic trait of quantum mechanics, the one that enforces its entire departure from classical lines of thought,” wrote Erwin Schrödinger when he coined the term “entanglement” in 1935. And when it came to entanglements, Schrödinger knew whereof he spoke. He had fled Nazi Germany the previous year and moved to Oxford, where he caused a huge scandal, having brought both his wife and his (pregnant) mistress.

The new paper makes no headway on the puzzle of how particles maintain their uncanny synchronicity, except to say that they somehow must. The essence of the argument is really quite simple. Any theory that goes beyond classical physics endows particles with extra properties that are not directly accessible to us. Classical categories are unable to capture them in their full richness. Our measurement apparatus and indeed our minds are Procrustean beds: whenever we measure or otherwise describe a particle, we force the particle to conform to some category that is ill-suited to it, and we should not be surprised that the outcome is uncertain. If you prepare several particles in the same way and measure their energies, you will, in general, obtain different values.

Yet despite the inadequacy of our classical categories, these postclassical particles nonetheless obey classical rules such as the conservation of energy. Postclassical particles are thus like wizards who are obliged to conceal their magical abilities to pass among Muggles. To do that, their innate properties must be highly constrained. And that is what entanglement ensures.

Richens gives a minimalist example of a counterfactual theory: quantum mechanics minus entanglement. Particles are set up to lack spooky linkages, and their dynamics is restricted so as to maintain their mutual autonomy. The theory still predicts all the uncertainty, randomness, and other quantum phenomena that physicists know and love. But the lack of entanglement sinks it. In order to keep the particles unentangled, the restricted dynamics either must be irreversible in time or must prevent particles from ever interacting, contradicting what we see.

“We are trying to assert what ‘weird’ properties the universe must have, given that it is not classical but any theory must be able to account for the classical behaviors we see around us,” Richens explains. Evidently entanglement is among those “weird” properties. “Weird” is, of course, a human judgment. In a sense, it is we humans who are weird for lacking entanglement even though the particles that make us up (and that make up any allowed counterfactual world) do.

The physicists I asked for comment are as split as a Schrödinger’s cat. Rob Spekkens at the Perimeter Institute is thumbs-up. “The real interest of this result lies in the surprise factor,” he says. He is struck, in particular, that postclassical physics is self-cloaking. The reason the everyday world seems classical to us is that one postclassical concept (entanglement) offsets others (such as uncertainty). On the other hand, he wonders whether entanglement plays this role because of the specific way that Richens and his colleagues think our observed classical realm emerges from the deeper physics. “There might be postquantum theories that have a nontrivial classical limit without having entanglement,” he suggests.

Giulio Chiribella at the University of Hong Kong is less impressed. He says that physicists established long ago that entanglement is ubiquitous in the quantum world: nearly any physical interaction generates it. The same will hold in any postclassical theory, for the reason that entanglement is needed to ensure reversibility in time. “What this result is really telling us is that if we want physics to be both reversible and nonclassical, then we must have entanglement,” he says. Chiribella says he feels the paper downplays the importance of reversibility.

Whatever the outcome, Richens and his colleagues have hastened the day when quantum mechanics will no longer be deemed mysterious—a day that physicists must feel some ambivalence about, because they also revel in a theory that so thoroughly bends our minds.

Daniel Sudarsky has nothing against beer, but disputes the common assumption that the interpretation of quantum mechanics is of no practical import. He thinks it has measurable consequences for cosmology, including the seeds of galaxies, black holes, and gravitation in general. Sudarsky and his colleague Elias Okon, both at the National Autonomous University of Mexico, spelled out their argument in an earlier guest post on this blog. There they focused on a specific interpretation in which the collapse of the wavefunction is a physical process as opposed to a mathematical contrivance, a concept I elaborated on in a Nautilus magazine article last year.

I caught up with Sudarsky in June at a conference on the philosophy of quantum gravity, organized by Christian Wüthrich at the University of Geneva and Nick Huggett at the University of Illinois in Chicago.

His talk at that conference, also available on YouTube, delved into the related question of how our present theories dovetail with a putative quantum theory of gravity. He suggested that quantum gravity should be thought of not as a single theory, but as a ladder of progressively more elaborate descriptions, descending rung by rung from the observed world to a deep level where spacetime itself dissolves:

1. classical continuum (e.g. fluids, general theory of relativity)

2. classical particles (e.g. Newton’s law of universal gravitation)

3. quantum field theory in curved spacetime (e.g. cosmological inflation)

4. semiclassical theory (e.g. Hawking’s analysis of black holes)

5. Planck scale (e.g. string theory, loop quantum gravity)

Although theorists naturally gravitate to the puzzles of the deepest level, the higher levels are no breeze, either. We observe the world to consist of objects following definite paths, but this commonsensical view must give way, and Sudarsky analyzed how that might happen.

In this interview, which I filmed in June 2017 at a conference on quantum gravity organized by Christian Wüthrich at Geneva and Nick Huggett at the University of Illinois in Chicago, she discusses three such assumptions: UV completeness (full unification occurs only on the most minute scales), correspondence (new theories reproduce not only the results of existing theories, but also their mathematical structure), and analogy (reasoning that applies to hot coals and crystals carries over to more exotic systems).

In her conference talk, now available on YouTube, you can also watch her dissect other words that physicists toss around, such as “fundamental” and “emergence.”

Last month, while visiting Ekert at the Centre for Quantum Technologies in Singapore, along with Howard Wiseman and the labs at the Centre for Quantum Dynamics at Griffith University in Brisbane, I got an idea for how to demonstrate quantum cryptography using ordinary household items. The improvised setup is several steps removed from true quantum crypto, which is still a graduate-student-level project, but it involves the same physics and, in principle, could be upgraded to the real deal by swapping in more advanced components. The demonstrator is a classical version of the cryptographic scheme invented in 1984 by Charles Bennett and Gilles Brassard. It encodes a message using the polarization of light. If government agents or nosy siblings try to intercept the signal, they'll muck it up and thereby betray themselves.

You'll need the following:

1. Two linear polarizers. I bought 2-inch-diameter polarizers from Edmund Scientific. You could also break out the lenses from a pair of polarizing sunglasses or IMAX 3-D googles, buy polarizing film from an outlet such as Ward's Science, or repurpose the polarizing filters that screw onto the front of 35mm camera lenses.

2. Two pieces of cardboard, each about twice the width of the polarizers

3. Flashlight

4. White index card, to serve as a screen

5. Some way to hold the components steady. I used an alligator-clip holder from Radio Shack, but binder clips would probably work, too.

6. Two coins or random-number generators

The sender ("Alice") encodes the message with one polarizer; the receiver ("Bob") decodes it with the other. If the polarizers are not already circular, cut them to be so. Mount each in a piece of cardboard to block stray light and to make it easier to handle. Cut a hole in the cardboard a bit smaller than the polarizer and tape it in place. The orientation of the polarizers is important; be sure to mount both in the same way. I'd recommend trimming the cardboard into an octagon, so that you can rotate it in 45-degree increments. Mark four directions: horizontal (0 degrees), diagonal (45 degrees), vertical (90 degrees), and "antidiagonal," meaning diagonal the other way (145 degrees).

Align the flashlight, polarizers, and index card. To verify the setup, align the polarizers in the same direction and turn on the flashlight. Fine-tune the position of the components until you see a fairly bright circle on the screen. You might need to dim the room lights.

Then, rotate one polarizer by 45 degrees. The circle should dim subtly but perceptibly. If you measure its brightness using a light meter, it should halve in brightness.

Rotate the polarizer another 45 degrees, so that the two polarizers are now perpendicular. The screen should go completely dark.

Now we can use this little apparatus to send a message. The most straightforward approach is to translate the message into a string of bits and encode '0' as horizontal polarization and '1' as vertical polarization. For each bit, Alice turns her polarizer to the appropriate position. Bob parks his polarizer in the horizontal position and leaves it there. He records a '0' if the screen is bright and '1' if it is dark. After giving him enough time to take the reading, Alice goes to the next bit and continues until the message is sent.

As a means of communication, this might sound needlessly baroque. Why not simply flash the light on and off like Morse code? In laser communications, engineers have found that polarization can be a more robust carrier of information than the presence or absence of light, since atmospheric turbulence can alter its intensity of a light beam, but leaves its polarization unaffected. More important for our present purposes, the scheme translates directly to the quantum realm: it works even if you make the light source so dim that the bits are being carried by single photons.

Be that as it may, the scheme is totally insecure. An eavesdropper ("Eve") can intercept the beam with her own polarizers and flashlight. First she follows the same procedure as Bob to record a bit. Then she follows the same procedure as Alice to retransmit that bit to Bob. She allows herself an evil chuckle as her interference goes undetected.

Quantum cryptography seeks to foil Eve by introducing an element of randomness, which we can simulate by giving Alice and Bob coins to flip. In this more sophisticated scheme, Alice doesn't transmit her message just yet—she first establishes the security of the channel by sending a string of test bits. For each, Alice flips her coin, and if it lands on heads, she uses the same encoding as before: '0' is horizontal polarization, '1' is vertical. For tails, she switches to a new convention: '0' is diagonal polarization, '1' is antidiagonal. Bob likewise flips a coin, orients his polarizer either horizontally or diagonally, and records a '0' if the screen is bright and '1' if dark. If the spot is intermediate in luminosity, he treats it as ambiguous and flips the coin a second time to record a '0' or a '1' at random. (This is the trickiest part of the procedure. Bob needs to reliably tell the difference between full and half brightness, which look very similar to the human eye.)

After Alice finishes sending her test bits, she uses some ordinary means of communication to tell Bob the results of her coin tosses. The two needn't worry about anyone listening in on this communication; on its own, it gives nothing away. Whenever their results match—half the time, on average—Bob can be assured that his reading accurately reflects Alice's transmission. The rest of the time, his readings are random noise and he crosses them out. Of the bits he does trust, he takes a sample and confirms with Alice that they were indeed what she sent.

Now Eve is going to have a harder time of it. In making her own measurement, she needs to choose whether to orient her polarizer horizontally or diagonally, and on average her choice will match Bob's only half the time. So she won't be able to intercept and retransmit the bits reliably. When Alice and Bob perform their verification step, they will find that only half the test bits got through, and they'll know their line is being tapped. They can use the bits that did get through as a secure key for subsequent transmissions. The following animation from the Centre for Quantum Technologies elegantly illustrates the procedure. (In the video, the quantum key generator is shown separate from Alice, whereas the simple demonstrator combines these two functions.)

Because the simple demonstrator is not truly quantum, it is vulnerable to a cleverer eavesdropper. The difficulty of distinguishing full from half brightness is a weak point that Eve could try to exploit. She could also bleed off a bit of the light using a partially silvered mirror, and Bob might never know his signal was being tampered with. Such surveillance strategies become impossible only when you go to the level of individual photons, where there's nothing left to bleed. Several cities now have quantum secure networks.

To be sure, even real quantum cryptography is not as foolproof as sometimes claimed. The Bennett-Brassard scheme can be hacked by exploiting the inevitable imperfections in the equipment. The nice thing about Ekert's procedure is that it adds the element of nonlocality, which provides a way to corroborate that the procedure works even if imperfectly implemented.

1. Download the diffraction grating pattern from Enrique Galvez’s website at Colgate University. The fork at the center of the pattern is what twists the light. (If that link is broken, try here or here.)

2. Using a photocopier, reduce the pattern to about half a centimeter on a side and transfer it to an overhead transparency. Make sure the fork doesn’t get smudged. (Finding a transparency in the age of PowerPoint is probably the hardest part of the experiment.)

3. Shine the laser through the pattern, ensuring that the beam passes through the fork, and project it onto a wall a few meters away. The grating splits the laser beam into a row of circles.

Courtesy of Samuel Velasco

Each of the circles flanking the central circle should have a small hole in the middle. The holes are a sign that light is being twisted. If you move the laser beam off the fork, the holes go away.

Level Explanation Animal Example Human Age Equivalent Machine Example –1 Disembodied Blends into environment Molecule 0 Isolated Has a body, but no functions Inert chromosome Stuffed animal 1 Decontrolled Has sensors and actuators, but is inactive Corpse Powered-down computer 2 Reactive Has fixed responses Virus Embryo to 1 month ELIZA 3 Adaptive Learns new reactions Earthworm 1–4 months Smart thermostat 4 Attentional Focuses selectively, learns by trial-and-error, and

forms positive and negative associations (primitive emotions)

Fish 4–8 months CRONOS robot 5 Executive Selects goals, acts to achieve them, and assesses

its own condition

Octupus 8–12 months Cog 6 Emotional Has a range of emotions, body schema, and minimal

theory of mind

Monkey 12–18 months Haikonen architecture (partly implemented by XCR-1

robot)

7 Self-Conscious Knows that it knows (higher-order thought) and

passes the mirror test

Magpie 18–24 months Nexus-6 (Do Androids Dream of

Electric Sheep?)

8 Empathic Conceives of others as selves and adjusts how it

presents itself

Chimpanzee 2–7 years HAL 9000 (2001) 9 Social Has full theory of mind, talks, and can lie Human 7–11 years Ava (Ex Machina) 10 Human Passes the Turing Test and creates cumulative

culture

Human 12+ years Six (Battlestar Galactica) 11 Super-Conscious Coordinates multiple streams of consciousness Bene Gesserit (Dune) augmented Samantha (Her)

Quantum optics experiments typically create entangled photons using a crystal of barium borate. This material takes on two different crystalline forms; the one of greater interest is the beta phase. The photo at right shows a sample in a laboratory mount. It looks like a little prism. But unlike the glass in ordinary prisms, the crystalline material is nonlinear: it is not an inert substrate, but its refractive index can be modified by the light that passes through it, enabling all sorts of novel optical effects.

Courtesy of Enrique Galvez, Colgate University

In particular, the crystal allows two light beams to interact, which they wouldn’t otherwise do. The interaction can amplify one of the beams at the expense of the other, as well as create a third beam. In fact, the beam that gets amplified need not be a "beam": it might simply be random quantum noise in the electromagnetic field. If you set up the crystal properly, the amplification is so powerful that it turns the noise into a proper light beam. A single incoming beam (typically blue or ultraviolet) can thus conjure up two beams (typically red). This process occurs particle by particle: each blue photon splits into two red ones.

Courtesy of Centre for Quantum Technologies, National University of Singapore

The splitting, known as spontaneous downconversion, is a low-probability event. Only about one in a billion photons in the incident beam interacts with quantum noise and divides; the rest continue straight through the crystal unaffected. For this reason, you'd never see the beams with the unaided eye. The above image from the Centre for Quantum Technologies in Singapore is a long exposure. The photographer dragged a piece of tissue along the beams to reveal their paths.

By virtue of their common origin inside the crystal, the outgoing red photons can be entangled in any of their properties: energy, momentum, polarization. For simplicity, experimentalists usually concentrate on polarization. The crystal has an optic axis. If the laser light is polarized in the plane defined by that axis, outgoing photons will be polarized perpendicular to the axis. Vertical polarization in, horizontal out; horizontal in, vertical out. Depending on how you arrange the beam and crystal, the outgoing photons can have the same polarization (known as Type I downconversion) or exactly the opposite polarization (Type II). It doesn’t really matter; either way, the photons are perfectly correlated. This correlation reflects the symmetry of the crystal about its optic axis.

Voilà, two photons. But when you want two entangled photons, you need an extra ingredient: the polarization has to be indeterminate. This is really the essence of entanglement, the reason it’s so mysterious. The photons have the same polarization, but that polarization is not horizontal, not vertical, not circular. It's just plain nothing, a blank that has yet to be filled in, according to the standard interpretation of quantum mechanics.

After all, if the photons did have a specific polarization, there’d be no mystery. You’d create two identical photons and later measure them to be identical, which is no weirder than pairing two socks as soon as they come out of the laundry and later observing that they’re the same color. Entangled photons, in contrast, are like a pair of socks that don’t have any particular color. Each assumes a color only when measured, and both assume the same color. If that boggles your brain, it should. This is what it means to be nonlocal: you can make a statement about the system as a whole—namely, "the parts are the same"—but not about any individual part.

To produce this indeterminacy with the barium borate crystal, you sandwich two thin layers of the material, one oriented vertically, the other horizontally. Then you send in light that is polarized in neither a horizontal nor a vertical direction, but on a diagonal. As long as the layer is thinner than the beam, there’s a quantum uncertainty about which layer the beam will interact with and, therefore, what polarization the outgoing photons will have. Their ambiguity is resolved only when they strike the polarizers and are measured.

Just upstream of the crystal, you can place an optical element known as a waveplate, which can be oriented to ensure the laser photons striking the crystal are diagonally polarized or not. This waveplate therefore decides whether the photons emerging from the crystal are entangled.

What’s nice about the crystal is that it gives you a very controlled way to create entangled photons. I want to emphasize, though, that there’s nothing really special about this crystal. The fact it’s a crystal makes it sound mysterious, but entanglement can be produced in any number of ways. I did a version of the experiment in my basement using a sample of radioactive sodium-22, which gives off entangled gamma rays. Edward Fry of Texas A&M and Richard Holt of the University of Western Ontario used mercury atoms in their pioneering entanglement experiments back in the ’70s. Mercury can emit two photons in rapid succession, and those photons will be entangled. Even the mercury in ordinary fluorescent lights emits entangled photons, Fry says.

The hard part about entanglement isn’t creating it, but creating it in a way that lends itself to measurement. Because the physical implementation isn’t important, it’s perfectly valid to think about it at a higher level of abstraction, such as the coin-flipping metaphor I use in my book. The metaphor is grounded in concrete physics.

I interviewed Wiseman in October at the Emergent Quantum Mechanics conference in Vienna. In this video, he boils the debate down to a schism between realists and operationalists, who clash not just on quantum theory but on the goals of science. (Each of those groups, in turn, is a fractious bunch.) Realists think quantum physics is nonlocal; operationalists don’t. But they're reading different meanings into the word “nonlocal.”

Wiseman himself has worked on everything from atom lasers to spooky action at a distance involving a single particle, as opposed to the pairs and trios that usually figure in quantum experiments. The solo-particle version hews closely to Einstein’s original thinking on the phenomenon, as I discuss in Chapter 3 of my book and in this animation. Wiseman has also sought (in this paper and that) to set the historical record straight about the Einstein-Bohr debate and has put forward a new interpretation of quantum mechanics that seeks to combine the best features of existing approaches.

The heat energy you apply has to go somewhere and, as 19th-century physicists such as Ludwig Boltzmann showed, it goes into the motion of water molecules. You can even count those molecules. If you dump in a lot of energy and the temperature barely budges, that energy must be spread out over a lot of molecules; if the temperature shoot up, that energy must be spread out over fewer molecules. If water were a true continuum—having an infinite number of molecules—the temperature would stay stuck at absolute zero. Such a system could never come into equilibrium with its surroundings; it would suck in energy without limit. “The fact that such a fluid can store and exchange heat energy cannot be understood within the continuum theory,” Padmanabhan has written.

I interviewed him on camera in October at the Emergent Quantum Mechanics conference in Vienna, which was funded by the Fetzer Franklin Fund.

As he explains, you can apply the same thermal reasoning to spacetime. It, too, has a temperature. It can heat up and cool down. After all, what makes heat heat, rather than motion, is ignorance. You can’t follow the details of molecules and see only their aggregate effect. Much the same happens when you become ignorant of processes in space. Suppose you’re stationary or moving at a constant velocity. The vacuum of space looks empty, meaning that vibrations in fields, such as the electromagnetic field, cancel one another out. But if you accelerate, you recede from certain regions of space (those way in back of you) too fast for light ever to catch up. You see a horizon around you, marking the farthest you can see, and that leaves you unavoidably ignorant about those field vibrations. To you, the supposed vacuum looks like a hot gas, throbbing with particles at a certain temperature.

These thermal phenomena imply that spacetime, like the glass of water, consists of some sort of “molecules” or “atoms.” Padmanabhan runs with this idea to develop a quantum theory of gravity. Einstein attributed gravity to the structure of spacetime, so, if spacetime is atomic, it stands to reason that gravity must reflect the behavior of those atoms. Following up pioneering work in the mid-1990s by Ted Jacobson (whom I featured in an earlier video), Padmanabhan has sought to derive Einstein’s equations of gravity from spacetime’s thermal properties. In fact, he puts forward an even more general theory of gravitation that accounts for the accelerating expansion of the universe; predicts that the atoms of spacetime, like molecules of water, can rearrange themselves into multiple phases; and might even explain some of the mysteries of quantum mechanics.

Courtesy of Silke Weinfurtner

Don’t be alarmed, but there is a black hole in your bathtub. When you drain the tub, water converges on the plughole and speeds up, eventually flowing too fast for surface waves to propagate outwards. Those waves get swept down the drain like hapless astronauts falling into a black hole (see video at end of post). For even more fun, there’s a time-reversed black hole, known as a white hole, in your sink. When a stream of water from your faucet strikes the sink and splays out, it initially moves too fast for ripples to propagate inwards. As it diverges, it slows. Within an inch or two, waves become able to propagate in every direction—a transition that is clearly visible as a discontinuity, or “hydraulic jump,” marking the perimeter of the white hole (see photo below).

Courtesy of Silke Weinfurtner

These aren’t just loose metaphors. An accelerating or decelerating fluid flow is mathematically the same as a black or white hole, as Bill Unruh of the University of British Columbia first noted in 1981. In recent years, physicists have exploited the analogy to explore gravitational phenomena. A pioneer of this emerging field is Silke Weinfurtner of the University of Nottingham. I interviewed her in October at the Emergent Quantum Mechanics conference in Vienna.

In this video, she describes two experiments. First, she talks about her current project to model a rotating black hole as a vortex of draining water. She tracks surface waves that get close enough to the hole to be swept up in the vortical motion, but not so close that they get sucked down the drain. Such waves should be able to escape back out with more energy than they originally had—a process known as superradiant scattering.

Courtesy of Silke Weinfurtner

Second, Weinfurtner recounts her earlier work with Unruh to model a white hole in a long trough of flowing water. In the diagram at right, the water flows from left to right. A motor at 6 creates surface waves with a wavelength of above 60 meters. Those moving downstream are swept over the edge at 7, while upstream ones travel toward the underwater obstacle at 5. In such a channel, ripples propagate at a speed that is proportional to square root of the depth, so they slow down when they start passing over the obstacle and are eventually unable to continue traveling upstream. The obstacle thereby creates an event horizon, a boundary across which waves can go one way but not the other. In this case, it is a white-hole horizon, into which nothing can enter. If you mentally run the whole thing in reverse, you have a black-hole horizon, from which nothing can escape.

The analogy goes further. The fluid horizon also creates a version of the Hawking effect, the quantum instability whereby a black hole sheds particles. As incoming waves bunch up near the horizon, they cease to be waves in the usual sense. The water trembles in a confused mess, setting new waves into motion with a much shorter wavelength, about 20 centimeters. They come out as pairs, one of which is absorbed while the other propagates away from the horizon. (Were this a black- rather than white-hole horizon, the incoming waves would have a short wavelength and outgoing waves a long one.) That’s just like the Hawking effect. The waves even have the thermal spectrum that Stephen Hawking predicted. The main difference is that in real black holes they arise spontaneously from fluctuations at the quantum level, rather than incoming waves generated by the apparatus. The observed effect agrees with theory despite the approximations in the setup, indicating that the analysis is robust.

The analogy between curved spacetime and fluid flow is so tight that some theorists have argued that spacetime is literally a sort of fluid, perhaps constructed out of some sort of “molecules” and even capable of undergoing a change of state, the theme of Chapter 6 of my book. Something to ponder the next time you take a bath.

Update (20 September 2017): I visited Weinfurtner’s lab last week and her student Sam Patrick demonstrated how the drain-hole vortex in their experimental tank resembles a black hole. Using tweezers he disturbed the water surface at several locations. Close to the drain, the vortex flow was so fast that the disturbance couldn’t propagate against it. Instead it was swept into a distinctive triangle shape, analogous to the tilted spacetime light cones that define the geometry of a black hole.

Earlier this year Weinfurtner and her team succeeded in measuring superradiant scattering. As a quick demonstration, Patrick plunged a glass jar into the water to send out a series of ripples. As they approached the drain hole, the vortex flow flung them around. The result was a spiral pattern, which rotated the opposite direction as the vortex and extracted rotational energy from it.

[video width="640" height="480" mp4="http://spookyactionbook.com/wp-content/uploads/2015/12/2017-09-14-17.18.36.mp4"][/video]

An Israeli-born physicist who was a professor at Tel Aviv University (among other institutions) and is now at the Institute of Quantum Studies at Chapman University in Orange, Calif., Aharonov remains one of the most consistently innovative thinkers in quantum physics. He also came up with the concept of weak measurement, whereby you can glean information about quantum systems without disturbing them (which had been thought impossible); the proposition that the present is as influenced as much by the future as by the past; and a real-life version of the Cheshire cat from Alice in Wonderland. He’s the kind of person who really does believe in six impossible things before breakfast. “A typical day starts with Yakir pulling out a numbered list of discoveries he made while he was sleeping,” says Jeff Tollaksen, director of the institute at Chapman. And many prove to be not so impossible.