Would it be possible to Computer Model Migration?

[Flavivs Aetivs]

Modern Migration Theory and Ethnogenesis are two very complex topics that have been extensively studied since Edward Gibbon first studied the Fall of Rome several centuries ago.

In the modern age of computers, and with use of statististics, population densities, factors such as military or technological superiority, and of course the approximate locations and political cohesion of individual "barbarian" groups to actually make a computer model of modern migration theory and ethnogenesis?

Obviously it would be an extraordinarily complex program, but if we can model the weather back then, why not statistical models of population movements, which we undoubtedly know more about?

[Lyceum]

Nope. We do anyway but nope for several reasons.

Basically, migration is complex as hell and we use several different models (so for the spread of proto-Greek into Greece we tend to use a wave model). The problem is these aren't usually single time events (stative) but diachronic (durative) so...well basically it's not at all neat and computers tend to want neatness. You can find lots of pretty diagrams in several books, however, just don't put to much faith into them.

Like I said, there is no real migration "theory" but several different models and people get behind the one they want. I suspect that the kind of date you're talking about, being so late, may be somewhat more amenable but in general think of the mid 20th century as a watershed which makes us much less sure about many things as more data comes up.

[Flavivs Aetivs]

Nope. We do anyway but nope for several reasons.

Basically, migration is complex as hell and we use several different models (so for the spread of proto-Greek into Greece we tend to use a wave model). The problem is these aren't usually single time events (stative) but diachronic (durative) so...well basically it's not at all neat and computers tend to want neatness. You can find lots of pretty diagrams in several books, however, just don't put to much faith into them.

Like I said, there is no real migration "theory" but several different models and people get behind the one they want. I suspect that the kind of date you're talking about, being so late, may be somewhat more amenable but in general think of the mid 20th century as a watershed which makes us much less sure about many things as more data comes up.

I understand that people have tried to model just "movements" but do these models account for everything? Differences in regional climates? Changes in weather patterns? Advancement of technology? Military superiority of groups they interact with? Known political organization and complexity, as well as policy between different groups?

You'd be looking moreso at the AI of a videogame than a comparatively simple wave model, applied to each individual grouping. (And obviously without the human input). It could be done using set events and only one or a few nations as a variable.

The mathematical model would be immensely complex.

[Lyceum]

More promising but still problematic, so much is either unknown or unquantifiable or unreliably quantifiable. Something like advancement of technology is not always linear nor a singular value. Something like political complexity is a mind field and something we'd stay away from, to give you a few examples.

You might be able to create several diagrams and then take something from the way they intersect, but it would be speculative and shaky at best. Also a wave model really isn't a simple thing when applied to peoples and languages lol. These things even for one value get complex pretty fast when you beign to take into account the human element. You can look at how archaeologists try to account for the increasingly well known Eastern element in Greek votive deposits for an example. Very easy to create a simple map of find spots and then grade that map diachronically and or shade it by frequency. That tells you very little. How did these objects get there, why? what purpose? how do these change over time etc? These are the more interesting questions.

I think, with the major caveat that all the problems of examining humans are kept in mind, you can possibly start using these models if corroborated with other approaches. Essentially I get what you're saying but there are reasons this isn't more widespread.