Haifa, born 1953

David Deutsch

The brain is a computer, and any universal computer can run any program, so an artificial mind must be possible. What we lack is not the hardware but the program. A person is a universal explainer, able to create any explanation there is, and nobody yet knows how creativity works.

A woodcut of a traveller on hands and knees pushing his head through the starry dome of the sky to see the wheels and clouds of the cosmos beyond
The Flammarion engraving, from Camille Flammarion's L'atmosphère, 1888. A traveller reaches the edge of the sky and looks through. For Deutsch, people are the creatures who can always look further: there is no limit in principle to what explanations can reach.Artist unknown. Public domain, via Wikimedia Commons.

The view

Deutsch is a physicist at Oxford and one of the founders of quantum computation. His 1985 paper on the universal quantum computer extended the Church–Turing idea into physics: a universal quantum computer could simulate any physical process at all. His books The Fabric of Reality (1997) and The Beginning of Infinity (2011) build a world view on four strands: quantum physics in its many-worlds form, the theory of computation, evolution, and the philosophy of Karl Popper. His argument about time is the centre of the sister site, Philosophy of Time.

His view of the mind follows from computational universality. The brain is a physical object, and every physical process can be simulated by a universal computer. So whatever a brain does, a suitably programmed computer could do too. Artificial general intelligence, or AGI, is therefore possible in principle, as a matter of physics. What makes humans special is not the brain's material but its software. We can create new explanations, and our range of possible explanations is unlimited. Deutsch calls people universal explainers.

The argument

AGI's missing ingredient is not computing power but a theory of creativity. After Deutsch, "Creative Blocks", Aeon (2012)

A key idea in The Beginning of Infinity is the jump to universality. Many systems improve step by step until, at some point, a small change makes them universal. Early writing used a separate sign for each word; the alphabet could spell any word, including ones not yet invented. Roman numerals struggle with large numbers; place-value notation can write any number. Babbage's machines were designed for particular tables until his Analytical Engine could, in principle, compute anything computable. Try the figure below.

Special-purpose machines: each does one job, built into its hardware.

Deutsch argues that human minds made the same kind of jump. Our ancestors' capacities grew gradually, but at some point people became able to explain anything that can be explained. There is no deeper kind of mind that could understand things we could not; there is only more knowledge to create. This also explains why AI research, in his view, has struggled. Programs that perform tasks, however well, are like special-purpose machines. What makes a person is the ability to conjecture new explanations and criticise them, which no one yet knows how to program. The Turing test, he argues, judges by outputs and so misses the point: it is the creativity behind behaviour that matters, not the behaviour itself.

A finely detailed black-and-white portrait of a man seated at a desk in a workshop, woven in silk, with the inscription À la mémoire de J. M. Jacquard
Portrait of Joseph Marie Jacquard, woven in silk, 1839. The picture is not drawn but woven, by a loom following thousands of punched cards. The same loom could weave any pattern its cards described. Charles Babbage, who owned a copy, borrowed the punched cards for his Analytical Engine.Public domain, via Wikimedia Commons.

Where he meets Dennett

On the basics they agree. Minds are what brains do; there is nothing supernatural in the skull; and since a brain's work is computation, a machine could in principle do it too. Both are confident that artificial minds are possible, both reject dualism, and both owe a debt to Darwin and to Popper. Dennett's Popperian creatures, testing hypotheses before acting, are close to Deutsch's picture of knowledge growing by conjecture and refutation.

Where he parts

Dennett was a gradualist. Minds come in kinds and degrees, built up from competence without comprehension, and there is no sharp line between the simplest controller and a human thinker. Deutsch stresses a sharp divide. Either a system is a universal explainer or it is not, and there is a qualitative gap between people, of any physical make, and every other thing we know of. He has suggested that consciousness may come with creativity, which would make it far rarer among animals than Dennett supposed.

They also disagree about qualia. Deutsch is not an illusionist. He treats qualia as real and still unexplained, and he suspects that explaining them will go hand in hand with explaining creativity. Dennett thought the explanation was already in outline: once the brain's representations and their uses are understood, there is no further thing to explain. Deutsch would answer that a theory that cannot tell us how to build a creative, feeling mind is not yet the explanation, however good its outline.

Deutsch and Dennett side by side
DeutschDennett
What is the mind?A program running on the brain: knowledge-creating software.What the brain does: a pattern of functions.
Is phenomenal consciousness real?Yes. Qualia are real and still unexplained.No, not as philosophers define it. It is an illusion produced by the brain.
Could a machine have a mind?Yes: the universality of computation means AGI must be possible.Yes, in principle, if it does what we do.
Is there a hard problem?Unsolved; it awaits a theory of creativity.No. There is the Hard Question: and then what happens?

See every view compared →

Further reading