Montag, 13. Juli 2009

Stephen Wolfram - ANKS - Not a New Review -

A New Kind of Science A New Kind of Science by Stephen Wolfram

My rating: 2 of 5 stars
Wonderfully printed, easy to read, marvelous to look at, pretentious piece of quack.

It was hard to pass the Introduction and the first Chapter, when you hear the author praising himself, his own importance and why literally 'everybody' in science and why science itself is 'wrong' and did not 'see' what very self-aware author sees, but nobody else.

The book is a massive piece of rarely to find print. I was self-published by the author to assure the quality of the printing. It has many black-and-white Illustrations and pictures to accompany the text. The text, when it is not about self-praising, is rather easy to read and easy to understand for non-academics. So what is it about?

It talks about the 'big' picture in nature and science. Stephen Wolfram is a well known 'Wunderkind' in theoretical physics. He made his PhD in 1979 at the age of 20. In the mid-1980s he founded Wolfram Research Inc. and invented 'Mathematica' the leading software for technical computing and symbolic programming.

Wolframs main thesis of the book and the solution he offers are easy to understand:

"... very simple programs produce great complexity. For all it takes is that systems in nature operate like typical programs and then it follows that their behavior will often be complex. And the reason that such complexity is not usually seen in human artifacts is just that in building these we tend in effect to use programs that are specially chosen to give only behavior simple enough for us to be able to see that it will achieve the purposes we want" page 3

This is the crux of the book, there, luckily on page 3.

If you read the 12 Chapters (846 pages plus around 250 pages of notes and additional massive Index, than it is most likely, that you do it because you like an intellectual exercise or because you love books and don't mind the many times annoying voice.

I am not going to discuss the results of such a vast amount of research and work that went into proving and verifying his idea. There are plenty of scientists from his fellow theoretical physics to mathematicians and others that reviewed and challenged Wolfram's book over the past years. There is even a great discussion a the slashdot-forum that mocks the authors writing and is funny and hilarious to read for nerds like me. May reviews and articles are online and it is easy to find them, if you want to.

Beyond the Masters Voice, this is a lovely book to look at. An artifact in these times of trash-quality publishing. It is not a masterpiece of printing. You cannot compare it with bibliophile editions. But for science books, if you have seen a physical copy of it, it is more on the upside, which tells you where we are nowadays, when it comes to print.

There is a complete online-version of this book now available for free on the Wolfram Research Homepage. The whole thing. For the curious minds.

Unfortunately, somebody on Google buzz triggered the worst in me and made me write some sort of review/critique in 2010. I attach it here, since it found some readers.

disclosure: I am neither an academic, nor a scholar. Take my disappointing comments for what they are .

The main result (...) - that programs based on simple rules can produce behavior of great complexity - seems so fundamental that one might assume it must have been discovered long ago. But it was not, and it is useful to understand some of the reasons why it was not .” Wolfram ANKS page 42.

The premise of Wolfram’s book was to introduce to all of us “A New Kind of Science” - did he achieve that? Have we found out anything new or fundamentally different (as in ‘wrong’) with non-Wolfram “science”? Is his the underlying ‘principle’ - the one thing τι ὃ οὐ κινούμενον κινεῖ ("something which moves [other things] without [itself] being moved [by anything]" (Aristotle Metaphysics).

You mention Scott Aaronson's review of the book and that was the first review I have read. I felt grateful that somebody from Wolfram’s ‘crowd’ (on his level, way above mine) took the time and effort to debunk the pretentious ‘findings’. I lack that level of inside baseball. Keep reading, if you must.

“... it’s validity depends on how it is formulated.” Scott Aaronson, p. 99

What poisoned my experience with the whole book is his tone, not the more or less interesting, already ‘well-known’ phenomena that may or may not apply to the degree he thinks are true. Wolfram goes after everything and everybody. There is a well known tradition in Western Philosophy (which used to include all of ‘Science’) to make “scientifick” (rhetorical) statements like: “ (1) This is what my predecessors found out. (2) They were all wrong. (3) This is how the world works according to my body of work.”

Umberto Eco wrote in 1995 a book about the “The Search for the Perfect Language (The Making of Europe)” Eco, a linguistic professor and true “semiologist” (not the Dan Brown “DaVinci-Code, Langdon sort), explores the old European idea and cultural origins of a ‘universal language’, a concept that touches many of the aspects Wolfram is trying to get at, in his own way.

The “theory of science”, the underlying ‘grammar’ we all want to rely upon, is always - implicitly tainted with our suggestions, presumptions, inaccuracies - and personal idiosyncrasies ... even (and especially) in science! Mathematical models don’t suddenly ‘appear’ out of the blue. They all went through a process that has a ‘historic’ element or string to it. Think of non-coding functional RNA, or junk-DNA - that really isn’t junk after all, as we now know . We used to call that junk-DNA, since we did not know what it was good for. Wolfram goes to the Cellular Automata model to explain the world. His ‘simple programs’ are not really simple, they are not well understood. ‘Translating’ the idea of a cellular automaton in ‘machine instructions of a computer’ is problematic in a very fundamental, philosophical sense. "Traduttore, traditore." (The word ‘translate’ means ‘to betray’).

True, since Leibniz we constructed many interesting mathematical problems and found very accurate descriptions for natural phenomenas - but also always reflecting the ‘Zeitgeist’, the technology ‘du jour’, on which we create the models of the world every time new. Why is Wolfram’s computation as a framework for explaining the phenomenon of universality so different from La Mettrie?

Maybe it is my personal lack of imagination, but seeing him inventing ‘A New Kind of..’ LEGO bricks and trying to explain all complexity, classical physics, quantum mechanics, and mathematical phenomena through ‘his’ idea of LEGOs (you can also read λόγος logos) reminds me very much of the Stoa and less of a more serious - less broad - underlying principle of all nature.

But things are not that simple. And he is mixing (deliberately?) theory and experimental outcome, again and again in his book.

In Chapter 11 Wolfram returns to discussing cellular automata, Turing machines, register machines, etc...

The main point (...) is that essentially all of these various kinds of systems - despite their great differences in underlying structure - can ultimately be made to emulate each other. This is a very remarkable result (...) In a sense its most important consequence is that it implies that from a computational point of view a very wide variety of systems, with very different underlying structures, are at some level fundamentally equivalent. For one might have thought that every different kind of system that we discussed (...) would be able to perform completely different kinds of computations. But what we have discovered here is that this is not the case. And instead it has turned out that essentially every single one of these systems is ultimately capable of exactly the same kinds of computations. And among other things, this means that it really does make sense to discuss the notion of computation in purely abstract terms, without referring to any specific type of system. For we now know that it ultimately does not matter what kind of system we use: in the end essentially any kind of system can be programmed to perform the same computations. And so if we study computation at an abstract level, we can expect that the results we get will apply to a very wide range of actual systems .” Wolfram, ANKS, page 674.

I would love to leave you alone with this passage. I would love to not mention Kant’s “Kritik der Reinen Vernunft” 1781 (Immanuel Kant - Critique of Pure Reason ) where Kant distinguishes between ‘Synthetic’ and ‘Analytic Judgement’. Wolfram is doing nothing less but justifying his ‘experiments’ by proposing to not make a difference anymore between theoretical (abstract) automatons and computational results of the physical kind he did, because all are in the end ‘nothing else but’ his kind of universal cellular automatons.

I would insist, a Turing machine is a theoretical device. Turing machines do not ‘behave’. They do not ‘translate/interact or compute’ into similar systems concepts - no matter how similar they are. On the other hand, ‘systematic computer experiments’ do ‘behave’. Most of them according to their authors. Furthermore - another basic/poor thought of mine: ‘Cellular automata’ as a theoretical concept unequals ‘cellular automaton programs’ as in ‘executed by compilers’. The results of actual physical computation - as accurate and as many digits long they may be - are NOT comparable to results as in ‘ideas’. Ask the hardware engineers, who spend their lives designing CPUs or the poor souls who try to ‘predict race conditions’ in their massively parallel programs. You mix theory and the physical world, you end up with weird math. Ask for that your theoretical physics and quantum physics department. Combining those two parts and introducing them into your ‘little’ world, of innocent cellular automatons is turning the concept upside down - it results in unpredictable complexity.

Especially the last chapter, proposing a “Principle of Computational Equivalence” is nuts, and has no basis - again, according to my humble brain-cells. The Church-Turing thesis is an abstract model in the classic Popper sense. But Wolfram goes into ‘metaphysical’ territory (not in Aristotle’s or Kant’s sense, but in today’s lingo - beyond the grounds of classical and/or quantum physics. Do I have to paint graphs and cite peoples work of the past 100 years to ‘debunk’ his ‘thesis’? I cannot.

If you assemble LEGO bricks long enough with a set of simple rules, you are able to ‘see’ some things you might not have anticipated. You may ‘see’ a bigger pattern. You may experience ‘gravity’ etc... - yes, I am polemic and simplistic. There is nothing ‘new’ about my method. I do not see anything new neither in Wolfram’s findings (described at length with nice pictures) in ANKS. Oh, and did I mention his pretentious tone?

Sorry for the length, grammar/spelling and lack of more elaborate, better versed English.July 19, 2010

View all my reviews >>