Civilizations and E-Mail

March 11, 2013 11:03 AM

A team of computer researchers analyzed ten million Yahoo! e-mails and noticed a phenomenon: "E-mails tend to flow much more frequently between countries with certain economic and cultural similarities".

The paper, titled "The Mesh of Civilizations and International Email Flows" was written by researchers at Stanford, Cornell, Yahoo! and Qatar’s Computational Research Institute. It places the research in the context of Samuel Huntington’s much-maligned “Clash of Civilizations" theory:

"In this respect we cautiously assign a level of validity to Huntington’s contentions, with a few caveats. The first issue was already mentioned – overlap between civilizations and other factors contributing to countries’ level of association. Huntington’s thesis is clearly reflected in the graph presented in Figure 3, but some of these civilizational clusters are found to be explained by other factors in Table 5. The second limitation concerns the fact that we investigated a communication network. There is no necessary “clash” between countries that do not communicate, and Huntington’s thesis was concerned primarily with ethnic conflict."

We see in the study that nations placed into the same "civilization" per Huntington’s definitions do e-mail each other more frequently than nations e-mail nations e-mail outside their civilization. The authors themselves, however, add “the advancement of an explanation is premature.”

posted by spaltavian (26 comments total)

you don't say?
posted by clarenceism at 11:18 AM on March 11 [1 favorite]

Ig Nobel nominee?
posted by Melismata at 11:21 AM on March 11 [2 favorites]

I need a team of researchers to tell me that "people are more likely to communicate with people who are like themselves"?

Really?

Because I could have told you that for a lot less money.
posted by DWRoelands at 11:22 AM on March 11 [2 favorites]

This does seem a little bit duh. What I'd like to know is where they got 10 million Yahoo emails from.

Also if there is any information on what water feels like or the temperature of fire, plz let me know asap. I'll take my answer off the air.
posted by DU at 11:22 AM on March 11 [2 favorites]

"certain economic and cultural similarities"

Like speaking the same language perhaps?
posted by Ardiril at 11:22 AM on March 11 [2 favorites]

What I'd like to know is where they got 10 million Yahoo emails from.


My "Spam" box, most likely. Next they'll tell us how all civilizations have hot girls that live locally.
posted by DWRoelands at 11:24 AM on March 11 [10 favorites]

Of all the research ideas you could come up with for a great data set like this, trying to cobble together a pseudo-empirical justification for Samuel Huntington's xenophobia must rank near the worst.
posted by RogerB at 11:24 AM on March 11 [8 favorites]

In other science news all the cheerleaders only talk to the jocks, and the nerds sit over there by themselves.
posted by chavenet at 11:27 AM on March 11 [4 favorites]

Most of my email comes from Nigeria.

I guess I'm an outlier.
posted by mazola at 11:28 AM on March 11 [3 favorites]

I need a team of researchers to tell me that "people are more likely to communicate with people who are like themselves"?

It's one thing to suspect that something is true, another to demonstrate it with data.

Theories must be validated by experiment (or, when that's not possible, at least research using real-world data) if they're to be worth anything at all.
posted by Kadin2048 at 11:29 AM on March 11 [11 favorites]

Ardiril: Like speaking the same language perhaps?

Not really. While some of the "civilizations" speak mostly the same language, such as the Latin American and Sinic groups, others do not, such as Western, African and Orthodox civilizations.

Seems like a better criticism would be simple geography, though America and Canada are substantially separated from the rest of the West.
posted by spaltavian at 11:29 AM on March 11

"certain economic and cultural similarities"

Like speaking the same language perhaps?

According to the article, the US is in a node with Israel, Switzerland, and Italy, and the former UK colonies (which would share at least a second language lingua Franca) don't have a very close connection.

So, no. Not according to the article which is the subject of the post. You might try reading it.

Having read it, the article still sounds iffy --- why drag in Samuel Huntington? What metric are they possibly using for "dislike of uncertainty"? Weird.
posted by Diablevert at 11:29 AM on March 11 [2 favorites]

I wonder how representative Yahoo mail is of all e-mail interactions.

I would expect e-mail patterns to weakly follow social network graphs. I'll be looking for some mention of that when I read the paper.
posted by jeffamaphone at 11:39 AM on March 11 [1 favorite]

RogerB: Of all the research ideas you could come up with for a great data set like this, trying to cobble together a pseudo-empirical justification for Samuel Huntington's xenophobia must rank near the worst.

That's not an accurate summation of the study. They used Huntington's civilization definitions in part to test out his theory and see if that would translate in some froms of personal behavior. I think tying it so specifically to Huntington's paper was possibly a mistake, but mostly due to knee-jerk reactions.

Just the quoted portions in the post make it clear the author's aren't claiming to have justified "The Clash of Civilizations", let alone xenophobia.

But for everyone is so eager to dismiss Huntington's essay (as I largely do), don't you think a critical take, with empirical backing, is exactly the way to do it? And that we shouldn't pretend a study didn't happen even if it, in a very tiny, partial, somewhat supports the thing we're tyring to refute?
posted by spaltavian at 11:40 AM on March 11 [3 favorites]

Notice that I did not write "the same predominant language". That was intentional.
posted by Ardiril at 11:43 AM on March 11

That's not an accurate summation of the study.

No, and it wasn't intended to be — it's a summation only of the stupid part of the study, which includes a lot of the authors' framing. There's still plenty of interesting material in the paper for readers who can bear to skip around the Huntington bits. But the authors do themselves no favors by failing even to test the Huntington "civilizations" against any other groupings of countries (say: random groups; language groups; economic/trade groups; groups defined by per-capita shared citizenships or tourism data). Instead they just assume that Huntington's "civilizations" are the only testable grouping of countries — seemingly because they went in with it as a preconceived idea.

It's one thing to suspect that something is true, another to demonstrate it with data.

And another thing still to take a bunch of interesting data and recalcitrantly try to shoehorn it into a preconceived explanation that conveniently just happens to justify the explainer's political ideology!
posted by RogerB at 11:53 AM on March 11 [2 favorites]

In this respect we cautiously assign a level of validity to Huntington’s contentions, with a few caveats

One of those caveats is that uneven interconnectedness does not significantly increase the likelihood of ethnic conflict. Which is a pretty big caveat considering the thrust of Huntington's argument was that these civilizations will clash.
posted by MisantropicPainforest at 12:05 PM on March 11

I would guess that Yahoo e-mails tend to flow much more frequently between my mom's hacked account and thousands of people who will soon learn how to get CHEAP VIAGRA!
posted by etc. at 12:06 PM on March 11

I still use my Yahoo email address mainly because I've had it since 2000, making it almost certainly the only constant in my internet use since that time.
posted by The Card Cheat at 12:06 PM on March 11

Next they'll tell us how all civilizations have hot girls that live locally.

I believe this was the central thesis of Sex At Dawn.
posted by Horace Rumpole at 12:14 PM on March 11 [1 favorite]

"That's not an accurate summation of the study. They used Huntington's civilization definitions in part to test out his theory and see if that would translate in some froms of personal behavior. I think tying it so specifically to Huntington's paper was possibly a mistake, but mostly due to knee-jerk reactions. "

"A Post writer once called it “the most dangerous idea of our time“; elsewhere, scholars like Edward Said and Noam Chomsky have gone to lengths to shoot it down."

Social science especially does a hell of a lot more than just observe the phenomena it hopes to explain but when done well or done poorly also creates it.

Not only do we build the world that we imagine but there is a lot of truth in George McGovern's old saying that wars are just things that old men dream up for young men to die in. A world built according to this model really is the worst of possible outcomes for the end of the Cold War and it is nothing like inevitable unless we collectively give into it, plan it, and execute it. It is one thing for our old men to dream up terrorists and brown people sneaking onto planes with bombs, but this is something else entirely. This shallow outgrowth of casual racism, should we allow it to take root, would simultaneously represent the final end of both any real hope for peace and anything that could conceivably be worth fighting for.

This 'Clash of Civilizations' is a model that the authors not only don't verify in the way that the press isn't at all picking up on but really should never be used anyway lest it actually become useful.
posted by Blasdelb at 2:09 PM on March 11 [2 favorites]

I could have told you that for a lot less money.

Amen. This idea of "Let's prove the adage!" is part of the reason so few people take social science seriously. It's possible that there really is value to having data on something like this, but gee-golly-whiz do these folks ever do a piss-poor job of communicating it to dumb folks like me, who are left wondering (1) how much tuition do Stanford and Cornell charge to pay for nonsense like this, and (2) how many actually useful things could have been accomplished with this time and money.
posted by cribcage at 2:25 PM on March 11 [2 favorites]

Dunno... I think this study is ok. It's not groundbreaking but it's a small step to give empirical substance to a controversial theory. They tested it, it was a match, but it didn't have to be. It also doesn't claim to prove Huntington was right forever and ever. I thought it was an original, neat and straightforward approach, in terms of method.

You may think it's obvious, but part of what science does is systematically question the "obvious" and see if it holds water, and in this case it did. Not glamorous or revolutionary, just science.

And like spaltavian said, if you want to start arguing for/against a theory you (dis)agree with, falsifiable empirical studies are a good way to go. And you should be prepared to prove yourself wrong.
posted by ipsative at 3:21 PM on March 11

Amen. This idea of "Let's prove the adage!" is part of the reason so few people take social science seriously.

If you don't do research about conventional wisdom, how would you find out when conventional wisdom isn't true (Prisoner's dilemma, Stanford/Milgram experiments etc.)?
posted by ersatz at 5:19 PM on March 11 [1 favorite]

I'll venture to advance the theory that, when proving the obvious results in pecuniary rewards, the decline of scholarship is an inevitable corollary.

Whether that decline is a cause or a consequence will be the topic of my next monograph: subject to the proviso that it may not appear if someone burns the whole thing down.

P.S. Which, come to think of it, may explain Alexandria.
posted by Twang at 5:45 PM on March 11

Huh. I never thought the problem with Huntington was his broad general data so much as the point he was trying to use it to make. Of course parts of the world are more in communication within themselves than without, and of course those parts are going to have cultural similarities (though it's nice to have data.) What doesn't follow is the warmongering bullshit that came afterwards.
posted by Navelgazer at 6:17 PM on March 11