I think too much

So this week I invented computational history.

Perhaps because I read too much Asimov as a child and was unduly fascinated by psychohistory.

Perhaps because I was always going to be a historian until I dropped out to play with computers for a living.

Perhaps because my best friend is studying Computational Physics.

Perhaps because Crush is an historian and I am computerian and this way we have a project that combines our expertise and can be a testament to our love.

Perhaps because I have a first-edition copy of Wolfram's immense opus A New Kind of Science.

Perhaps because I have consumed too much Chlorpheniramine Maleate.

Regardless, I have been toying with some ideas.

Last week I attended the launch of a book dealing with Counter Factual or Virtual History:

"a way of examining the events of the past in order to form judgements about particular decisions made by individuals, or to evaluate circumstances that have occurred"

"a form of history which attempts to answer "what if" questions. It seeks to explore history and historical processes from the point of view of extrapolating a position in which certain key historical events did not happen or had an outcome which was different to that which did in fact occur"

These ideas instantly got me thinking: "What if you could simulate history"?

If you could create a model of the geo/socio/eco/politico/culturo system, then you could simulate historical events ... and more importantly, you could play with the variables in the system in order to gain insight and understanding.

I am not saying this would be easy, of course, but I think it could be done.

The model would only ever be a gross approximation, but the best part about histroy is that you can create a model and see if it results in a replication of actual historical events. This means that a model can be tested against actual experimental data, so to speak. So if (for example), you take a set of starting conditions, input these into the model and simulate the rise of Nazi Germany or the development of Colonial Powers, or the Middle East, then the model can be verified against the "real world" and then theses initial conditions (simular to a potential in molecular simulation*) can be used to play with variations in the historical conditions to gain insight into "what if".

Playing with some ideas this week and I realised the hardest part of the project is really in the assumptions one makes about the way things work ... the more I have delved into various theories about 'stuff,' the more I have realised that this is a slippery slope into a dark dark hole. I think that staying on a a very 'applied' level is the only answer here. Stick on the systems level, deal with the empirical information, verified against 'fact' (which is of course another slippery slope).

So that's my latest hare-brained scheme ... ambitious?

And more:
Because of the problem of assumption, I was actually thinking that what is needed is a simulation modelling language with an ability to create macro-style scripts. People could then buiuld and extend the model to handle particular sets of assumptions. So a Marxist would add a class dynamic, for example. To this end I looked at Common Lisp. And freaked out.

Has to be way of handling this in a language that I know.



* I think. Charlie gibbers at me and I pretend to grok what the hell he is talking about.