David Deutsch:

"The Beginning of Infinity" (Viking, 2011)

(Copyright © 2012 Piero Scaruffi | Legal restrictions )
The Israeli physicist David Deutsch continues the discussion that he started with his previous book, The Fabric of Reality" (1997), in which he tried to unify theory of evolution, theory of computation, theory of knowledge (epistemology) and theory of matter (quantum theory); but this is a much more profound meditation on the meaning of science and on the human condition. Its breadth is... breathtaking.

Deutsch begins by attacking empiricism: experience "selects" a theory, it does not form it. It was empiricism that started the scientific revolution, by providing a framework (observation followed by generalization) from which knowledge of what has not been experienced can be derived from what has not been experienced; but, unfortunately, empiricism was based on induction (the abovesaid "generalization"), which turns out to be an unreliable way to predict the future based on the past. Deutsch's viewpoint is that theories "evolve" through a process of guessing and testing: a scientist guesses that something might be true and tests his guess; if it works, we have a new theory; if it doesn't work, back to the drawing board. The goal is to come up with an explanation.

It is the "explanation" that allows us to create a long chain of interpretation that allows us to understand phenomena that are very far from our everyday experience. Deutsch paints a vivid picture of how astronomical observation consists of, in practice, just looking into devices built out of humble materials found here on Earth but, ultimately, allows the mind to visit distant worlds. The mind can "leave" the body for this spectacular voyage because of the "explanations" that guarantee it is not just being tricked by drugs. There is something special about this process that makes humans truly unique.

Deutsch disagrees with those who think that there is nothing special with this typical planet of a typical star of a typical galaxy (the principle of mediocrity): humans have created conditions (for example, low-temperature refrigerators) that are very rare in the universe. Deutsch emphasizes that most of the universe is very different from the environment that humans created in their cities, homes and especially laboratories. What made humans unique is that they can use technology to expand the capabilities of their brains. Whatever caused their brains to evolve the way they did, those brains are now capable of going beyond their original survival functions. And the reason that technology does that is the explanatory knowledge that only humans have learned to master. What is unique about explanatory knowledge is obvious if one considers that humans managed to survive in a wildly hostile biosphere and even moved to environments that are not the ones in which they were shaped by natural selection. Other species have knowledge too, but the kind of straightforward knowledge that limits their reach to their habitat. The explanatory knowledge of humans allows them to do things that their brains were not programmed for and their bodies could not do without help from technology.

Thus Deutsch calls humans "universal constructors": humans can transform anything into anything as long as they understand the natural laws that govern the universe and then build technology according to those laws. If the universe is governed by deterministic rules, then the human brain can eventually (explanation after explanation) get to understand everything. We are capable of everything, and in particular of understanding everything. There is nothing that obeys the laws of nature and that we cannot understand. If such a thing existed, it would have to be "supernatural" by definition: not obeying the laws of nature. In a universe governed by mathematical laws, the reach of explanatory knowledge is infinite, and therefore so is the reach of a brain that is capable of acquiring explanatory knowledge recursively (one explanation leading to another one). There is no limit to human knowledge. There is no limit to human creation.

We used to live in a world of supernatural entities. Then we discovered that we are capable of explaining how the world works because the world obeys mathematical laws that our brain can master. Therefore we became universal beings, capable of understanding anything that happens in the universe ruled by those laws of nature. A higher intelligence (incomprehensible to us) can only exist if the universe is not explicable, i.e. if there exist supernatural beings. In other words, the cognitive closure of human brains would imply that the ancient superstitions were correct. This is a powerful argument against the theory that human brains are limited by a "cognitive closure" that cannot be overcome.

Historically, it wasn't always this way. At some point humans became what their brain made possible: universal constructors. He links this conceptual revolution with the Enlightenment, that spawned rebellion against authority in many fields, including science: those rebels looked for explanations, not just dogmas. Furthemore, the Enlightenment introduced the notion that progress is good and should be a universal ideal to achieve by society. So the chain of explanations that leads to more and more knowledge became not only possible but an explicit goal of humankind.

The effect of brains like ours (brains that can deal with explanatory knowledge and therefore become universal constructors who can alter their environment at will) is significant. Deutsch points out that astrophysics is incomplete without a theory of people because people can alter the course of events that the laws of nature alone would cause. If we don't consider the actions of people, we can't understand why the Earth is the way it is.

Furthermore, the effect of knowledge-processing brain is profound also because an intelligent deliberate transformation is actually more likely to happen than a spontaneous one: the chances that an object is changed by a spontaneous transformation (one caused only by the laws of nature) are relatively low in the grand scheme of things, whereas the chances that an object is changed by an intelligent being for whatever purpose are very high.

In other words: the universality of the laws of nature coupled with a universal constructor like the human mind yields unlimited knowledge growth.

Deutsch argues that the well-known limitations on mathematics and computations (Godel's Theorem and the likes) do not affect his claims that humans can extend knowledge indefinitely: one can understand a mathematical statement without proving it. Basically, the proof is a technicality. If you can't prove it, it doesn't mean you don't "know" something. The only limit to human knowledge that Deutsch sees is the impossibility to predict future knowledge: the fact that we can find solutions to any problem does not mean that we already know those solutions.

The physics of knowledge representation is that the physical system of the brainsomehow contains a model of the physical system of the entire universe. As more and more knowledge is acquired that model becomes a better and better approximation of the universe (and viceversa). The world is knowledge-friendly because it obeys to laws of nature. The human brain is knowledge-capable. Starting with the Enlightenment, the human brain decided to exploit this capability. The spirit of the Enlightenment provided the motivation that was missing. Knowledge allows the brain to build technology to help acquire more knowledge even about facts that the brain cannot experience directly. This chain reaction leads to an endless creation of knowledge.

Another detour shows how universality (of knowledge representation) emerged in history. A writing system based on an alphabet is universal: it admits not only every word so far known, but also every future word. Apparently, the alphabet was invented only in the region around Syria and Lebanon. No other place evolved an alphabet independently and even in that place it took a long time to achieve that feat. Ditto for the positional system to represent numbers, also a universal system (capable of representing all possible numbers) invented (in a rudimentary form) by the Babylonians but hardly used by anyone (not even by the Romans). The ancient world was not interested in universality, even if it was capable of it. It took the spirit of the Enlightenment to create the motivation to adopt universality wherever possible. Starting with the Enlightenment humans became addicted to abstract knowledge. The culmination of this program has been the invention of the computer, which is a "universal" machine, capable of simulating anything in nature. That is also the reason why Deutsch disagrees with critics of Artificial Intelligence, who think that the computer is simply a metaphor for the mind, but not equal to the mind: the digital computer (the Universal Turing Machine) is as universal as the human mind and therefore truly different in nature and power from any previous metaphor of the human mind. One key step was the transition from analog machines (that are not universal because of measurement errors that pile up as knowledge "propagates") to digital machines (where that problem is avoided by representing knowledge according to the equivalent of an alphabet).

The other process known to create knowledge is the process of biological evolution. Deutsch shows the similarities between human brains and DNA molecules, both capable of storing and replicating knowledge. However, biological knowledge is non-explanatory, and therefore its reach is not infinite. The way knowledge is achieved is also different, although they both boil down to "trial and error": biological knowledge is constructed via random mutations (that are then selected by the environment), whereas human knowledge is constructed via guesses/conjectures (that are then selected by observation). And one could even argue that "random" mutations are not so random and that human thoughts are more random than one would imagine. The replicator (the entity that propagates knowledge) in one case is Mendel's gene, in the other case is Dawkins' meme.

The way this other knowledge process emerged is still shrouded in mystery, but basically, once replicators were accidentally born, they joined forces in groups. Such groups are organized in a way that each member contributes to the chemical reactions that allow the whole to replicate itself (with all its members). That was the birth of the first living organism. "Genomes are group of genes that are dependent on each other for replication". The genetic code itself (the way to encode all of this) had to evolve until it reached a point beyond which it did not need to evolve anymore (it hasn't evolved for billions of years) while still allowing for organisms to evolve. The code, that had originally encoded just a single-celled organism, stopped evolving, but it was now powerful enough ("universal") that it could encode a virtually infinite range of (multi-cellular) organisms. That is the power of representation systems when they are universal. They can describe a lot more than what they were originally built for.

Deutsch attacks both reductionism (according to which one has to go to the lowest possible level in order to find the real explanation of a phenomenon) and holism (according to which one can only consider the phenomenon in its entirety). Deutsch instead admits a plethora of valid levels of explanation, multiple explanations of the same phenomena at different levels, all of them "fundamental" in their own way.

There's more. This is many books in one.

The part of the book that deals with Quantum Mechanics is by far the most difficult to grasp (and probably the most controversial). Deutsch subscribes to Everett's multiverse model against the Copenhagen interpretation of Quantum Mechanics: if something can be in multiple states according to the laws of Quantum Mechanics, then it means that it is in those multiple states, each in a different universe. Each universe is roughly ruled by the very deterministic laws of classical Newtonian Physics. Quantum weirdness appears because we are seeing the effects of multiple universes working at the same time. A key concept is "fungibility": it means that a set of objects can be considered as a set of identical objects. For example, if i lend you one dollar and a few days later you give me back one dollar, we assume that the state of the world is the same as it was before the borrowing even though the dollar bill that you return to me is not the one that i gave you. However, if i lend you a book and you return me a different book, i would be pretty upset: not all books are alike the way all dollar bills are alike. Photons are fungible: you can't tell one from the other. The atoms of lasers are fungible: they are all the same thing. It turns out that fungible objects can deviate from each other and become different entities... in different universes of the multiverse. And that is the origin of the apparent randomness that an individual in one universe observes. If one views a particle as a multiversal object, randomness and uncertainty disappear: a particle has multiple positions and multiple speeds in multiple universes. The wave associated to a particle is not due to the duality of particles and waves: a particle is distributed across many universes, and therefore it "is" a wave in the multiverse.

Then there is a book that deals with Philosophy. Several chapters belong to this thread. One is even an imaginary dialogue with Socrates.

There are even a chapter about politics (in which he argues that coalition governments are not a good idea) and a chapter about aesthetics (in which he argues that there are absolute criteria for beauty). After some more comments on the state of Western civilization, and a wildly creative explanation of why creativity evolved (he thinks it helped spread memes), Deutsch ends the book with powerful critiques of Jared Diamond's theories (See "Collapse: How Societies Choose to Fail or Succeed" (2004) and "Guns, Germs and Steel" (1997)) as well as of Kurzweil's theory of the "singularity". Deutsch sees no need for a singularity, for a moment when the machine becomes more "intelligent" than the human mind: the human mind can grow forever, even exponentially, and machines will simply remain the tool to grow knowledge exponentially. At best the machines will be as intelligent as the human mind that built them and uses them. Humans will always be able to understand their machines.

TM, ®, Copyright © 2012 Piero Scaruffi