Digging The Foundation

As I’ve mentioned a few times with the production of Foundation for Apple TV I’m planning to re-read the first four books (which I started in my teens) along with the last three books (plus I hope the three non-Asimov volumes if I can get them at a reasonable price) over the next few years. At some point I hope to watch the series too – when it comes out on DVD.

Of course getting hyped about the series (and the books) I’ve been doing some digging on YouTube for Foundation related videos and, naturally, I’ve been thinking about the possibility of PsychoHistory. Is such a thing even possible? Can you (accurately) predict the Future and if not why not? When I read the first book back in the 1970’s the answer would be, I think, an easy ‘No’. Today, with some caveats and reservations, I’d have to say that the answer is a qualified ‘Yes’. Naturally there are those, probably a lot of people, who would say that accurately predicting future events is impossible. After all MANY have tried and they’ve almost always failed. Science-Fiction is (as we know) littered with reasonably accurate predictions but I seriously doubt that if you compare them to the number of SF predictions that did NOT come to pass that the apparently accurate ones represent anything other than statistically lucky guesses. So, what’s stopping History being more predictable?

I think that PsychoHistory would rest on three pillars. Firstly there’s the computing power to cope with the undoubted complexity of historical events. When the books were written this was beyond imagining. When I read the books such a thing was clearly science-fiction. Today….? We’re nowhere near that level of computing power needed BUT such power in doubling around every two years and still looks like it’s not going to slow down any time soon. If we need 100x or even 200x present day computing power we don’t really have that long to wait. 1-2-4-8-16-32-64-128-256. So, in EIGHT generations of computing – that’s sixteen years – we can expect our computers to be 128x as powerful. Two years later they’ll be 256x as powerful as today. So I really don’t think computing power is going to be a limiting factor here.

Second is data. You’re going to probably need to crunch a LOT of data to make accurate predictions days, months or even years (to say nothing of centuries) into the future. But with the amount of data we already have and the amount flooding into ‘the system’ every minute of every day I really don’t think that data, or the lack thereof, will be a limiting factor either.

The stumbling block is the third pillar. As far as I can tell we simply don’t understand how History ‘works’. The history we all know and the history we enjoy reading about are crafted narratives carved from world events into stories that make sense and that we can relate to. They follow ‘one damned thing after another’ and link things together in causal chains that satisfy the way we think about cause and effect but I think this kind of narrative has a very low (if not actual zero) predictive potential. The reason it all looks coherent, I think, is because the historian is looking backwards and can ‘explain’ things because they have access to the causal chain they want to talk about from beginning to end. They can tell the story because they know how it ends. If they’d been given the data which ran out before the event chain finished how many of them could predict the actual ending? Not many I think and not consistently. What we need to do that is to understand the mathematics of history – which we don’t…. at all. Without such a branch of science the idea of PsychHistory will remain nothing but science-fiction. But…..

There ARE glimmerings however. Firstly there is a long history (semi-pun intended) of those who try to discover historical trends and where those trends are leading us. To be honestly most of these are little more than guesswork – even from world class historians – but the effort is being made. Part of the problem is, of course, that we’re dealing with very complex systems here with billions of interactions, chaotic forces and Black Swan events. However, several areas of endeavour are looking at, modelling, and attempting to predict exactly that. They are: weather forecasting, behavioural economics and stock market analysis. All three are dealing with large complex systems that we struggle to understand and great efforts are being made to predict all three. These I think are the seeds of PsychoHistory – especially in the area of Economics which is attempting to model human behaviour on a very large scale. But can people really be ‘modelled’ that well so that human history can be accurately predicted. Surely humans are too complex and too unpredictable for that to happen. What does predictability say about Free Will?

I think I’ve already addressed the complexity issue. Not matter how complex human behaviour is – at the individual or mega scale – computing power will quickly catch up to address the issue. Also, to be honest, I don’t think that human behaviour on any scale is THAT complex. Free Will is an interesting issue. If you haven’t seen it already I’d recommend Series 3 of the TV series Westworld which looks at this idea. One scene in particular really jumped out at me. One of the main human protagonists asked the main robot protagonist if Free Will existed. She said that it did but that it was ‘fucking hard’. I think that’s right. If we do indeed have Free Will (something my personal ‘jury’ is still out on) we tend not to use it all the time or probably very much at all. We operate a lot by habit or preference, each running our ‘loops’ only seeming to make decisions. I think that’s probably the case most of the time. Interesting one element of the Westworld series was the existence of ‘outliers’ which the ‘system’ couldn’t always predict presumably because they exercised their Free Will more often or more aggressively than the rest of humanity. I think I remember them making up around 1% of the population.

Although I am not, honestly I’m not, making this into a fourth Knowledge Stream here (no, I’m not) I will from time to time pick up a book on historical analysis, weather prediction or human based algorithms, stock market analysis and behavioural economics. But it’s not a project, definitely not. Three long running projects are enough I think. There’s no way I’m adding a fourth, I’m just not…. So don’t even ask, don’t even go there….. So the label over on the right called PH (for PsychoHistory) is just a way of keeping everything in a searchable location. Not a project… Just a label.