« August 27, 2011 | Main | September 19, 2011 »

Saturday, September 10, 2011

Reading List: The Beginning of Infinity

Deutsch, David. The Beginning of Infinity. New York: Viking, 2011. ISBN 978-0-670-02275-5.
Were it possible to communicate with the shades of departed geniuses, I suspect Richard Feynman would be dismayed at the prospect of a distinguished theoretical physicist committing phil-oss-o-phy in public, while Karl Popper would be pumping his fist in exultation and shouting “Yes!”. This is a challenging book and, at almost 500 pages in the print edition, a rather long one, but it is a masterpiece well worthy of the investment in reading it, and then, after an interval to let its implications sink in, reading it again because there is so much here that you're unlikely to appreciate it all in a single reading.

The author attempts nothing less ambitious than a general theory of the creation of knowledge and its implications for the future of the universe. (In what follows, I shall take a different approach than the author in explaining the argument, but I think we arrive at the same place.) In all human endeavours: science, art, morals, politics and governance, technology, economics, etc., what we ultimately seek are good explanations—models which allow us to explain a complex objective reality and make predictions about its behaviour. The author rejects the arguments of the relativists and social constructionists that no such objective reality exists, as well as those of empiricists and advocates of inductive reasoning that our models come purely from observation of events. Instead, he contends that explanations come from conjectures which originate in the human mind (often sparked by experience), which are then tested against objective reality and alternative conjectures, in a process which (in the absence of constraints which obstruct the process, such as reliance on received wisdom instead of free inquiry) inevitably converges upon explanations which are minimal and robust in the sense that almost any small change destroys their predictive power.

For example, if I were so inclined, I could invent a myth involving gods and goddesses and their conflicting wills and goals which would exactly replicate the results of Newton's laws of mechanics. But this would be a bad explanation because the next person could come up with their own myth involving an entirely different pantheon which produced the same results. All of the excess baggage contributes nothing to the explanation, while there's no way you can simplify “F=ma” without breaking the entire structure.

And yet all of our explanations, however elegant and well-tested, are simply the best explanations we've found so far, and likely to be incomplete when we try to apply them to circumstances outside the experiences which motivated us to develop them. Newton's laws fail to describe the motion of objects at a substantial fraction of the speed of light, and it's evident from fundamental conflicts in their theoretical structure that our present theories of the very small (quantum mechanics) and the very large (general relativity) are inadequate to describe circumstances which obtained in the early universe and in gravitational collapse of massive objects.

What is going on here, contends Deutsch, is nothing other than evolution, with the creation of conjectures within the human mind serving as variation and criticism of them based on confrontation with reality performing selection. Just as biological evolution managed over four billion years or so to transform the ancestral cell into human brains capable of comprehending structures from subatomic particles to cosmology, the spark which was ignited in the brains of our ancestors is able, in principle, to explain everything, either by persistence in the process of conjecture and criticism (variation and selection), or by building the tools (scientific instruments, computers, and eventually perhaps our own intellectually transcendent descendents) necessary to do so. The emergence of the human brain was a phase transition in the history of the Earth and, perhaps, the universe. Humans are universal explainers.

Let's consider the concept of universality. While precisely defined in computing, it occurs in many guises. For example, a phonetic alphabet (as opposed to a pictographic writing system) is capable of encoding all possible words made up of the repertoire of sounds it expresses, including those uninvented and never yet spoken. A positional number system can encode all possible numbers without the need to introduce new symbols for numbers larger or smaller than those encountered so far. The genetic code, discovered as best we can determine through a process of chemical evolution on the early Earth, is universal: the same code, with a different string of nucleotides, can encode both brewer's yeast and Beethoven. Less than five million years ago the human lineage diverged from the common ancestor of present-day humans and chimpanzees, and between that time and today the human mind made the “leap to universality”, with the capacity to generate explanations, test them against reality, transmit them to other humans as memes, and store them extrasomatically as oral legends and, eventually, written records.

Universality changes all the rules and potential outcomes. It is a singularity in the mathematical sense that one cannot predict the future subsequent to its emergence from events preceding it. For example, an extraterrestrial chemist monitoring Earth prior to the emergence of the first replicator could have made excellent predictions about the chemical composition of the oceans and its interaction with the energy and material flows in the environment, but at the moment that first replicating cell appeared, the potential for things the meticulous chemist wouldn't remotely imagine came into existence: stromatolites, an oxygen-rich atmosphere, metazoans, flowers, beetles, dinosaurs, boot prints on the Moon, and the designated hitter rule. So it is with the phase transition to universality of the human mind. It is now impossible to predict based on any model not taking that singularity into account the fate of the Earth, the Sun, the solar system or the galaxy. Barring societal collapse, it appears probable that within this century individual wealthy humans (and a few years thereafter, everybody) will have the ability to launch self-replicating von Neumann probes into the galaxy with the potential of remaking it in their own image in an eyeblink compared to the age of the universe (unless they encounter probes launched by another planet full of ambitious universal explainers, which makes for another whole set of plot lines).

But universality and evolutionary epistemology have implications much closer to home and the present. Ever since the Enlightenment, Western culture has developed and refined the scientific method, the best embodiment of the paradigm of conjecture and criticism in the human experience. And yet, at the same time, the institutions of governance of our societies have been largely variations on the theme of “who shall rule?”, and the moral underpinnings of our societies have either been based upon received wisdom from sacred texts, tradition, or the abdication of judgement inherent in multicultural relativism. The author argues that in all of these “non-scientific” domains objective truth exists just as it does in mechanics and chemistry, and that we can discover it and ever improve our explanations of it by precisely the same process we use in science: conjecture and criticism. Perversely, many of the institutions we've created impede this process. Consider how various political systems value compromise. But if there is a right answer and a wrong answer, you don't get a better explanation by splitting the difference. It's as if, faced with a controversy between geocentric and heliocentric models of the solar system, you came up with a “compromise” that embodied the “best of both”. In fact, Tycho did precisely that, and it worked even worse than the alternatives. The value of democracy is not that it generates good policies—manifestly it doesn't—but rather that it provides the mechanism for getting rid of bad policies and those who advocate them and eventually selecting the least bad policies based upon present knowledge, always subject to revision based on what we'll discover tomorrow.

The Enlightenment may also be thought of as a singularity. While there have been brief episodes in human history where our powers as universal explainers have been unleashed (Athens and Florence come to mind, although there have doubtless been a multitude of others throughout history which have left us no record—it is tragic to think of how many Galileos were born and died in static tribal societies), our post-Enlightenment world is the only instance which has lasted for centuries and encompassed a large part of the globe. The normal state of human civilisation seems to be a static or closed society dominated by tradition and taboos which extinguish the inborn spark of universal explanation which triggers the runaway exponential growth of knowledge and power. The dynamic (or open) society (1, 2) is a precious thing which has brought unprecedented prosperity to the globe and stands on the threshold of remaking the universe as we wish it to be.

If this spark be not snuffed by ignorance, nihilism, adherence to tradition and authority, and longing for the closure of some final utopia, however confining, but instead lights the way to a boundless frontier of uncertainty and new problems to comprehend and solve, then David Deutsch will be celebrated as one of the visionaries who pointed the way to this optimistic destiny of our species and its inheritors.

Posted at 23:14 Permalink