Could Mind Forge the Universe?

Markus Mueller could mistakenly be accused of being a supreme individualist. After all, the quantum physicist’s radical new take on reality seems to suggest that the world we see emerges from our observations. He admits that it’s a tough idea for people to wrap their heads around, or for him to articulate: "No one’s even made up a science fiction story that would roughly explain it," Mueller says. "Our theory says that only observations are fundamentally ’real’."

Mueller, who is based at the Institute for Quantum Optics and Quantum Information in Vienna, Austria, and his colleague philosopher Michael Cuffaro of the University of Western Ontario, in Canada, are questioning one of our most basic assumptions that we operate on as human beings: that reality is objective and pre-dates us. Mueller and Cuffaro aren’t saying that the world is an illusion or unreal in some way, but they are saying that something more fundamental is behind it. With the help of an FQXi grant of over $100,000, they have been exploring what this approach can reveal about the origin of reality and how fundamental physical laws arise.

Quantum theory already forces physicists to take the role of the observer seriously. For example, before we measure a quantum system, it can hold contradictory properties, such as being in two different energy states. When we observe it, we force the system to assume a particular state, inherently tying the act of observing into reality itself. Mueller and Cuffaro argue that they are just following these hints to their logical end. Their aim is to develop a framework to describe reality without assuming the existence of ordinary objects with properties, governed by physical laws. But how do you set about the ambitious task of constructing a universe, without any building blocks or instructions for how to assemble it?

"Mathematically when you ask, ’What if there are no laws of nature?’ that’s like asking ’What if all you have is mathematics itself? Can that give you something that allows you to predict what you’ll see—that is, some probabilities of observations, without assuming anything else?’," says Mueller. "That was my starting point."

There was already mathematics in place to help the pair of researchers to do this. The field of "algorithmic probability" is already employed by those working on artificial intelligence and machine learning. It describes the probability of an observation occurring over a whole range of possibilities, and programmers use it to create robots that can ’learn’ from observations of their environment and make decisions about what action to take based on past outcomes.

Brain Dump

Using this mathematical framework, Mueller has derived a coherent description of reality from nothing more than descriptions of sets of observations, made one after another. He starts with an observation that is encoded somehow. What you use to encode it doesn’t matter too much, so let’s say we use binary strings of zeroes and ones. The observation could be a raw dump of your brain at a moment in time describing your state now as a long string of bits. Mueller now borrows the mathematical tool from artificial intelligence, and asks: how likely is it that a random computer program would produce these bits out of mere chance? This is called the ’algorithmic probability’ of that observation.

Mueller postulates that what happens next depends on its algorithmic probability—conditioned on what has been observed in the past. This process, he argues, can forge reality itself. Future observations that don’t fit in with what we’ve seen in the past—perhaps worlds with five or six dimensions—are thrown out because their likelihood is so much smaller than us continuing to observe a world with three spatial dimensions. Observations consistent with what we’ve seen in the past, those that fit into a description of the world in the usual sense, with regular laws—these are the most likely. Hence we get an ’emergent world’ from our observations.

"It turns out that algorithmic probability is a very useful notion," says Mueller. Physical laws and regularities emerge in the same way that robots eventually learn to give the right answers. "Laws tend to stabilise and you see regularities around you and this world with these laws of nature," Mueller explains. "It’s the same kind of mathematical origin of both."

Renato Renner, a theoretical physicist based at ETH Zurich, in Switzerland, notes that there is a long history in physics of trying to formulate theories from a first-person perspective. But unlike most attempts, Mueller has succeeded in developing "a well-defined theory within which non-trivial results can be derived," says Renner. "This is very remarkable!" In a rough sense, Renner says, Mueller has provided a quantitative version of Occam’s razor: it assigns high probabilities to possible future events that have a simple explanation.

Boltzmann Brains

As well as explaining why there are laws of nature at all, Mueller’s approach can also resolve some notoriously difficult puzzles in cosmology and quantum theory. One of the most bizarre is known as the "Boltzmann Brain problem". This notion dates back to the end of the 19th Century when physicist Ludwig Boltzmann published a theory stating that the entropy, or disorder, of a closed system always increases. Think of a hot coffee in a box of air: the disorder in this system will increase over time as the heat energy from the coffee gets shared with the air around it. But there is always the possibility that part of the system will fluctuate from disorder to order.

Boltzmann reasoned that the extraordinary complexity of life on Earth, including our brains, were the result of these random quantum fluctuations, and that they could occur anywhere in the universe. Taking it further, he predicted that we should see Boltzmann brains, self-aware entities, spontaneously popping in and out of existence. Stranger still, cosmologists in recent years have calculated that many models of the universe suggest that Boltzmann brains should vastly outnumber normal human brains. So how can you be sure you are not a Boltzmann brain, and that your next experience won’t be deep space?

Mueller’s theory provides a reassuring answer: continuing to experience business as usual on Earth is far simpler than suddenly finding yourself in space. And if something is information-theoretically simpler, then algorithmic probability says it’s more likely.

The ideas are in their infancy—Mueller is yet to publish in a peer-reviewed journal—but the framework has stimulated a lot of interest, and conference invitations have come flooding in. Rob Spekkens, a theoretical physicist based at the Perimeter Institute in Canada, thinks that algorithmic information theory will find many applications in physics in the future and applauds Mueller for being one of only a few people taking up the challenge. "It is very original and thought-provoking stuff," he says.

But there are wrinkles in the theory, says Ruediger Schack, a mathematician at Royal Holloway, University of London, in the UK, who works on an alternative interpretation of quantum theory, which also gives the observer a central role, called "QBism." While Schack admires the mathematical rigour of the project, he worries about the lack of definition given to agents—the people making observations—in Mueller’s approach. "The agents of Mueller’s theory are strangely diminished entities, defined by a random process," says Schack. "Surely the defining feature of agents should be that they act?" In Mueller’s theory, by contrast, the concept of action, as well as the world the agents act on, are secondary, emerging features of the theory. "This just doesn’t feel right to me," says Schack.

Mueller concurs that the notion of ’action of an agent’ is not part of the fundamental description in his theory; however, he does not see this as a flaw. "Choosing one’s actions is a secondary notion that should not be a fundamental part of any physical theory," says Mueller, "in the same way that emotions or free will should not." This view is in line with conventional thinking in cosmology, for instance, he says.

Renner, who has seen a draft paper of Mueller’s on the topic, says that it is one of the most thought-provoking pieces he has read in the last couple of years (arXiv:1712.01826). "In view of the conceptual problems we have with our current physical theories, notably with quantum theory," Renner says, "such radically different approaches are, in my opinion, desperately needed."