"... for you must accept your inhabitants of America as a young people: younger a thousand years, at least, than the rest of the world."

-- Francis Bacon

This essay is intended to serve as background research for a future alternate history. It's on the web, rather than just on my hard drive, just in case I don't turn out to be the author. If you can run with any of this, go for it with my blessing.

Question One: Why, at the moment of contact, was America behind Eurasia?

Was it just a matter of Eurasia being colonised first? Or of some climatic change which permitted the rise of civilisation occurring in Eurasia before it occurred in America?

Or was it, as Jared Diamond's Guns, Germs and Steel would have us believe, a consequence of large scale geography and domesticable species? This is going to be a very dodgy analysis, and I'm not pretending it should stand beside serious work on the subject.

Question Two: How Large Was the Gap?

Suppose we wanted to have an alternate history in which Eurasia arrived to find an America which was every bit as advanced as they were? One way to solve this would be to move back in time every date in the Americas: the date of colonisation, of every important climatic event, of every technological and cultural advance. But how far would we have to move America back? The same analysis applies if we want an America as much advanced over Eurasia as historically Eurasia was over America.

Assumption: Progress Can Be Measured.

In order for this analysis to make any sense at all, progress has to be a univariate quantity. That is, we have to be able to peg any culture at any time on a line, and feel that adequately describes their state of advancement.

Anthropological orthodoxy (as understood by a physicist, anyway) is that this is a Very Bad Thing. Many anthropologists would view a statement such as "The Australian aborigines were primitive compared with the European colonists around 1800" as unscientific. I personally regard this attitude as casting aside a useful (if rough) approximation for the sake of (at best) a not always necessary precision and (at worst) political correctness (whatever that means).

The issue, as I see it, is whether the univariate approximation leaves enough information to be of any use. I'll use statistical arguments below to argue that it does.

The Raw Data: Dates of Key Advances.

I extracted these advances from a timeline in the front of The Times Atlas of Archaeology. A few are controversial, some are probably out of date, others I probably misunderstood and wrote down wrong. Believe them at your own risk.

Chances are there are errors here, so please tell me if you find any.

The dates are generally the first development of an idea anywhere in the area concerned. Obviously not everywhere civilised got these ideas as soon as they were invented.

I made up the last entry myself, arbitrarily assuming that the ``Inca'' state of Tawantinsuyu and Achmaenid Persia represented the same level of development. It turns out to be an outlying datum, so I'll ignore it in all the analysis.

Analysis

For the analysis I used a simple linear regression. There's two obvious ways of doing this: graphing the date of discovery in America against the date of discovery in Eurasia and vice-versa. The two methods give slightly different results for reasons that make one dubious about trusting the results. The results do, however, give you some faith in the idea of approximately univariate progress. The correlation (technically this isn't exactly a correlation, but it's the way to think of it, I think) between the dates is 0.9568: a value of 1 would indicate the theory matched perfectly, while a value of 0 would indicate there was nothing to it whatsoever (and ``advances'' came in random order).

I'll quote the lines of best fit in terms of how many years before contact an advance was discovered. The arbitrary contact date is 1500 AD, so a date of 500 BC will be represented as 2000 years before contact.

I also tried plotting the difference between Eurasian and American dates against the average of those dates. I haven't used these figures in analysis, which is perhaps a mistake. The idea here is to fiddle the x and y axes in order to get rid of the correlation, but there's still a residual correlation of 0.4345.

A more rigorous analysis suggests the quantity

has a mean of zero and no particular trend in time.

Qualitative Conclusions

First, all the statistical approaches are giving the same answers, more or less. So there isn't anything particularly subtle in my choice of variables. The main comment here is that there's no sign America was stalled, progress-wise. It wasn't progressing as fast as Eurasia, and Diamond (for instance) has a number of ideas about why. But there's no particular reason to believe that American societies wouldn't have developed computers and so forth eventually, even if all of Eurasia had been scoured from the Earth by some exotic mechanism. Of course, it's an unreasonable extrapolation from our data to suggest they would have done so: the question of why one Eurasian society had an industrial revolution and others didn't is a whole subject in itself.

The American progress rate, as a percentage of that in Eurasia, was in the low eighties. When all is said and done that's not so different from saying that the progress rates were roughly the same. I think any theory that says ``there is something distinctive about Eurasia, civilisations of the level of Eurasia would have had a much harder time arising anywhere else'' should be viewed with some doubt. The moral of these calculations is that humans are an adaptable species and will make do with less suitable crops if necessary: second rate domesticable species (or whatever) slow us down, but they don't stop us, at least not on the scale of continents and millennia. It's actually a rather hopeful message.

Quantitative Conclusions

We can extract any number of numbers from this, some of which may be useful. In fact we can extract two versions of each number, using the two different trendlines above. I'll quote both values in brackets and the average after the brackets, I suppose the average is the useful quantity but the numbers in brackets give a very vague idea of a minimum uncertainty.

Eurasia achieved America's 1500 level about (1857, 1717) 1787 BC. This is the era of Babylon, the Hittites, Shang Dynasty China, Middle Kingdom Egypt. It roughly corresponds to some time during the first epoch of the board game History of the World, if that helps. It shows just how badly outclassed the Americans were when Cortes et al arrived, and why they would have folded up swiftly even without the assistance of Eurasian diseases.

An unmolested America would have reached Eurasia's 1500 level in (5612, 5271) 5442 AD. So this is the date we would expect the American Columbus-equivalent to land at Galway or Tokyo or whatever. This is a problem from an alternate history point of view, the gap in time is an absurdly big job to fill and can't be done unambiguously.

The moment I sees two lines I wonder where they intersect, and whether the intersection is physically meaningful. Are the Europeans ahead just because they progressed faster, or is there an element of a head start as well? The trend lines suggest that both cultures would have had the same advancement about (16750, 20375) 18562 BC. I think this is so far outside the domain of our evidence we can safely regard it as spurious.

That said, however, the time scale is vaguely comparable with climate changes associated with glaciation periods. It's tempting to adopt a paradigm that all societies were created equal in sophistication at some climatic year zero about twenty-thousand years ago, and then progressed at similar but somewhat differing rates. Nonsense, almost certainly, but plausible enough as a basis for an alternate history, and as likely to be right as any other wild guess.

If we want the Americans to fight the Europeans (and other Eurasians, for that matter) on equal terms, we need to give the Americans a head start of (4112, 3771) 3942 years. Or, equivalently, we can shove the Europeans back that far. Of course if we plan to use AD/BC then we really should leave the Eurasians where they are, but perhaps we'll want to use the Mayan calendar. :-)

These figures would only be slightly different if you want to shove the dates back a little, and have contact induced by Vikings landing in Vinland. Bear in mind that Vinland is this world will be an Amerindian country with, say, vaguely mediaeval technology: the Vikings will raid and/or trade on roughly equal terms much the way they did in Europe.

If instead we want the Americans to trample all over the Eurasians the way they were trampled historically, we need to give them twice that much start: (8224, 7542) 7883 years. This pits Americans as advanced as 1500 Europeans up against the Babylonians, Shang, Egyptians and so forth.

Where Do We Go From Here?

I don't really know. It could be fun to map out the progress of American civilisation over the four-thousand years after contact, but ...

There's a saying that there are no right alternate histories, only plausible ones. That still leaves the question of how to decide on a particular evolution when several course each seem possible: what a physicist would call the tunable parameters of the model. In this case the tunable parameters would be legion, the possible futures profuse. Any attempt to talk about particular empires is doomed.

Turning the problem around, suppose the Americans had colonised Eurasia. Suppose further that a bored alternate history enthusiast from that culture decided to write up an alternate history of Eurasia, assuming no colonisation from America. And suppose that [s]he's a first class genius, whose insights are mind-bogglingly accurate. What level of resolution could we hope for in this document?

The way I see it, the very best you could hope for are comments like these:

• ``When the mouldboard plough becomes available the Mediterranean will decline relative to northern Europe. This is because Mediterranean agriculture doesn't need that kind of plough but the soils are getting exhausted whereas northern Europe is still a treasure house.''
• ``By the eighteenth century western Europe will be settling down into nation states, whereas south-eastern Europe will still be an empire dominated by Central Asians of some kind.''
• ``China has no comparable civilisiations anywhere near it, and is fundamentally a land- and river-based culture. Western Europe, on the other hand, is seriously indented and will have a maritime tendency. This is where we get our New World discoverers.''

• ``The Sahara desert will insulate Sub-Saharan Africa from Eurasia, turning it into a backwater. When Europeans develop oceanic capabilities they will go to Africa and India first, then much later to the Old World (which they, of course, will call the "New World").

Even at this level, it would be staggeringly hard. But maybe there are easier tasks as well, that someone else can conceive and execute.

Alternately there are the light and fluffy approaches. You could, for instance, set out to achieve a particular objective in the most plausible fashion you can. Or you could carry in a philosophy like:

"I think of North America as Europe, Mexico as the middle East, the Yucatan as India, and South America as China. When a dynasty fell in one of those, the corresponding dynasty will fall in my world. There'll be a Mexican equivalent of Omar the conqueror, an Amerind from Michigan just like Julius Caesar, a terrifying nomad from the Argentine Pampas who styles himself Genghis Khan."

It would probably give a purist an apoplectic attack: it doesn't do much for my appetite. But as a cue for the free association needed to write fiction, perhaps it would have some benefit. And it would probably work, which is more than you could say for most methods of creating four-thousand years plus of fictional history.

Whatever the basis of your alternate history, you'll need a plausible-sounding point of divergence. I suggest modifying the climate. As soon as humans have made it into North America proper, bring back the ice age in Afro-Eurasia, and delay civilisation by however many years you need.

Anyway, if anyone can use these results, they're free for normal academic crediting rules (i.e. say whence they came). And please, give me feedback at David.Bofinger@dsto.defenceSpamProofing.gov.au (delete the spamproofing).