June 2011

                                         THEORIES ARE FOR TURKEYS.

Before philosophers were  relegated to welfare beneficiary status, like ballet dancers, opera singers and other high culture performing artists,  they did identify a problem we have to deal with on a daily basis.

It's the problem of induction, often called Hulmes problem after the eponymous 18th century Scottish philosopher and Adam Smith compatriot, but was identified much earlier than this-  certainly by Epicurius , 2nd century AD Greek empiricist, and probably around a camp fire a long long time prior to then.

It can most usefully be explained by example:

A turkey is cared for and fed regularly, and for every day this happens it's expectation that this will also happen the next day strengthens. Then Christmas comes along (or Thanksgiving if you're an American turkey).

From this reasoning , philosophers argued that no matter how often certain things occur in certain ways, it's never possible to be sure that they always will.  Following this line to it's logical end, they eventually concluded that all prediction is impossible.

They then became hung up on finding something, anything, that could be declared absolutely true (like Descartes: "I think therefore I am", which they soon decided isn't ), toyed with "the only absolute is that there are no absolutes", began to have doubts even about this, went all sulky for a while, and then suffered a meltdown into post modernism, which curiously holds that no opinion or fact can have more value than any other while at the same time insisting that various left wing views do have.   

But I digress.

On the other hand, science, by avoiding such hubristic absolutism, has found that some things can be predicted to a useful extent, and by the process of reductionism (dividing problems up into smaller and smaller elements until the fog of overlapping influences clears) has inexorably expanded our understanding of the universe,  even to insights into puzzling aspects of human behaviour, like philosophy.

Science uses a Popperian approach:

Theories are developed from observations then tested  by experiment or further observation.  Karl Popper was (contradictorily) an Austro-British  philosopher who, while sitting out WW2 in the safe haven of  Christchurch's Canterbury University (in the next town up from Ashburton)  refined this useful approach to problem solving.  His insights; that to be useful, a theory has to be falsifiable (that is, testable) and that no number of positive results can prove a theory, but that just one negative can disprove it, underpin scientific endeavour.

This theory based approach to acquiring knowledge is sometimes called platonic, which is only marginally appropriate.  Mr Plato was very fond of theories but didn't believe in testing them, so was wrong about just about everything.

But even with the testable part added, engineers often don't have the opportunity to indulge in a theory based approach.  Their priority is finding solution to particular problems.  Understanding the universe is a bi-product- welcome and useful, but not the first order of business.  Many, if not most of the world's useful inventions take shape before adequate theories to explain them exist- scientists came along later to mop up this sort of detail. 

.

Single line kite design should be amenable to theory- they've been around for more than long enough to be thoroughly understood by now-  but, regretfully, they're not, or at least not yet to as useful an extent as I had come to believe.

I've been forced to this view after almost a lifetime of working on a theory based approach by the four bridle pilot kites I'm currently developing.

Except for pilots, I've generally worked with much more complex kites and it's been easy enough to convince myself that the apparent failure of current theories to accurately predict the effect that various changes will have is because of the confusion of overlapping influences.

The four bridle pilots are a reductionists dream kite and are now at the development stage where I'm trying one small change at a time- an ideal laboratory for testing whether the theories I have are any use or not.

Before test flying each change, I note what I expect the effect will be; if it's an asymmetric change, will it pull the kite to right or left for example. 

After a few months, my success rate in this predictive game has been barely indistinguishable from random guesses.

The answer therefore is very clear: the theories are mainly bunk.

Bugger, wasted life, Aaaagh!

But two questions are begged:

The first; why single line kite flying hasn't yet yielded to theory (traction kites have) is easily answered: Turbulent flow is complex because it cannot be predicted just by defining every parameter (velocity, density, pressure differences, shapes of adjacent surfaces etc) but is dependent also on what happened just prior-which depends on just prior to that, and so on back to forever.  Add to this the relationships around self stability and that the shape of a soft kite (which determines it's lift and drag) is a function of various pressure differences which in turn depend on the shape in a never ending disequilibrium, and there are all the excuses that will ever be needed to explain failure.

The second; why myself and others congenitally believe that we have useful predictive theories when apparently we (or at least I) don't have, is less comfortable to answer.

For evolutionary reasons (self  belief confers a survival advantage even when misplaced) , we are all too ready to believe we know why things are happening.  That so many mutually exclusive religions all have their staunch followers is evidence enough of this- and even the tending-secularist Plato was a prisoner of this hard wiring.

Mea culpa.

So, where now for single line kite development?

It seems that while we all fondly believe we know what we are doing, we don't actually know at all:- but because our various wrong theories are so disparate,  the net effect is basically one of random changes. 

Most random changes fail, but those that succeed become memes (cultural genes) that quickly go forth and multiply (get copied) in the cultural equivalent of biological evolution.  This produces useful results- and there's no question that kites are more reliable now and better fliers that they were even 10 years ago.

But the development of a predictive theory of single line kite stability still seems to me to be a worthwhile and possibly achievable  goal.

So I'll just go on being a turkey;  believing that I can understand what's happening- but perhaps more warily now, with an eye on the calendar and an ear for the axe. 

But then again, philosophers would say that it's not possible for me to know whether I really am or am not actually a turkey.

Peter Lynn,  Ashburton, June 1 '11