Diigo Links

Thursday, February 11, 2010

Update: Economists have it 20 both ways in England


As debate rages at Marginal Revolution over whether economics can be called a science, I read me some Stuart Kauffman over at his NPR blog. I now have doubts of my own. Can we rely on the classical definition of "Science" to get us "much further". The classical definition of "Science" per Popper requires repeatable experiments which produce consistent results (i.e. predictive power). This is "classical" in the sense that it's a pre-Einstein definition.

I think Kauffman, without naming it, is considering a way of doing business which is science-like, but lacks some of the constraints of a classical definition of science. I would hope it may lead us back to science in the same way that quantum mechanics leads to classical mechanics (also keep in mind that the quantum => classical model is not complete.)

So Kauffman steps boldly out into the wild, a place where you find tarot cards, black magic, phrenology and their like. A dangerous place to go for anyone looking for progress. The question is loudly begged (by any reputable scientist), can anything good come of this?!

Whether good can come of it or not, there's a market for it. Climate science is in this space. And people all over the world are gagging for fresh climate science. But for all intents and purposes, climate science, isn't. There is no way of testing a climate. There is no experiment which proves anything, which provides true predictive power. (Listen to this TED talk by david_deutsch - towards the end)

Much like economics, climate science is far better at proposing possible explanations for past events than predicting future events. In fact, climate science cannot and will never predict what's going to happen to the climate. It merely deals in probabilities. The main problem climate science will identify is that the horizon of a sort of laminar flow of climate is coming closer. What lies beyond that horizon is surely turbulence, but for how long, to what end, we won't know until after the fact. At which point climate scientists will say "we told you so" without ever actually saying what would happen - only predicting that "things would likely happen". But that's little more than coin-flipping compared to true predictive science. With enough opinions in the room, certainly some important climate scientists will be "right". But we could say the same for conspiracy theorists. There has to be a difference between the two. Or does there?

Kauffman is hinting towards a field of discovery which admits "ideas" and "perspectives". The difference between Kauffman's idea and anarchy (the kind of useless anarchy feared by scientists), is that ideas are to be lightly held. Hold all and any ideas of the adjacent possible. Then use attributes of the group of adjacent possible to define directions of understanding which might fall outside the lockstep of science. This is currently a well received way-of-working in the art world, but it's anathema to science.

Kauffman invokes a bit of Gödel to make his point. I'll extend a bit on that. In brief, Gödel "....proved that from the axioms of arithmetic, there could exist mathematical statements which, if true, were not derivable by deduction from the axioms of arithmetic. These are called formally undecidable statements. And Gödel showed that if you added such undecidable statements as new axioms of an enriched axiom set, new undecidable statements would arise - ad infinitum. In short, mathematics has the property that, given an axiom set, not all the true statements given that set are deducible from that axiom set." Kauffman calls into question whether rational decisions can be made based on this revelation of incompleteness. I would like to propose the question: Is "completeness" a universal axiom? Or rephrased; does the concept of "complete" have anything to do with the relationship mathematics has with the universe? Science relies heavily on mathematics as a rational base. And "the rational" saves us from a series of pitfalls that come with "the non-rational". But by what basis can we be certain that "the rational" is the sole explainer of the world around us? Are we using training wheels to interrogate the universe? Can we take them off? What would that look like?

I don't propose that the universe can be better explained by "the irrational". I propose that we may have created a false duality when we separated the world into either "rational" or "irrational". It's quite possible that the universe is more analogue and less digital than that.

The most common retort to this type of idea is; "look at this computer you're typing into/ reading from. It's here due to modern science. Every step of your developed-country-middle-class existence has been propped up by modern science. This is the basis by which we can confidently, if not certainly, rely on "the rational" to explain the world around us." While I do agree the methods of modern science have been productive and illuminating, what I'm wondering is whether we're not becoming victims of our own success, being lured towards a, albeit comfortable, local minimum.

Is it a mistake to combine the tenants of Gödel and Popper to form a statement: "Carry on doing what's worked in the past. We know it's incomplete, but it's the best we have. Only madness and ruin can be found outside." Should we not consider the combination of Gödel and Popper might form the statement: "We've been working with an incomplete model, maybe that's a hint that it's the wrong model. Should we keep pushing the same buttons harder, or should we change the rules to get a view above the field of play?

The sort of step I'm suggesting is a natural progression from what I call "Einstein’s breaking of the ruler". Prior to the common acceptance of the theories of relativity, it was commonly accepted that there was one universal clock, and one universal ruler. "Frame of reference" didn't come into the picture. Einstein changed that by making frame of reference primary to measurement. This changed everything. Kauffman is telegraphing a similar shift. I don't think he's quite nailed it yet, and he isn't the first to come up with the idea (clearly), but I think we're on the cusp of a paradigm shift just as challenging to rational thinkers as relativity was in the early 1900's.

String theory, I'm not a fan. I'm not a fan for all the reasons found in Peter Woit's book and blog "Not even wrong"- summarized as: "string theory has not provided predictive experimental results commensurate with its profile". It's just not yet what we might call "good science". A lot of work and money have been sunk into string theory, not a lot of results have come from it. Unlike relativity, a series of amazing inventions have not condensed from the revelations of string theory. And, until now, I've taken all of this as an indication that string theorists have gone down the rabbit hole (i.e. pure math). But my resolve is starting to waiver.

Kauffman is talking about a post-classical science. I'm agreeing with him. String theory is probably an example of the transition towards this post-classical science, along with climate science and economics. These disciplines are taking into consideration phenomenon beyond classical description. While climate science and economics continue to hammer away using classical methods, string theory is forging ahead into the wilderness. They're finding solutions to problems we don't have. They're discovering a fabric on their way to describing the garment. This isn't to say that string theorists have it quite cooked and ready to come out of the oven. But I'm starting to see a space for the rule-breaking that's going on.

So what governance model can we apply to a post-classical science to prevent results which read like post-modernist-science-fiction-poetry? I think we have to name it. Then I think it's important to declare (using the defined name) when, as a scientist, you're going beyond commonly accepted classical bounds of science. In the spirit of the "Gedankenexperiment" I've taken the opportunity to name post-classical science in German "Topf Rissbildung" or pot cracking (wink, wink). (see "backreaction" to Woit here) Named and claimed, the author then leaves the ideas to the market to decide their worth. And it would be up to scientists who partake in Topf Rissbildung to build credibility in the practice. It would also be up to the scientific community to allow/build a cultural space for the practice of Topf Rissbildung.

At the moment there is very very little patience in the scientific world for the publishing of any creative results or conjectures. This sort of writing is left to science fiction authors and bloggers (?!). And I agree that that's for good reason. But I also maintain that given proper governance, a productive arm of science could be built to encompass free-thinking and association (Cosmology ?). And I don't think "it's worth a shot", I think it's happening - unattended, unwanted, undefined, unconstrained. It's a future of science which will make today's scientists very uncomfortable, but which will provide the future of science a class of results unobtainable to the current discipline. May we watch and see.

No comments: