Here is a report in the UK Times, which includes an obvious but nevertheless important comment from Armand Leroi.

http://www.timesonline.co.uk/ tol...icle3031104.ece

Strange -- PNAS gives me an "Article not found" message when I follow the link. Even when I sign in to my august institution's subscription.

Evolutionary theory requires that a rapidly evolving population be (1) isolated, so that the new mutations are NOT diluted out in the larger population and (2) that selective pressures be radically reduced so that the normal functioning of natural selection, which is to eliminate mutations, does not occur.

So, in what human populations are these conditions being met?

Evlutionary theory requires (1) that an evolving population be isolated so that new, beneficial mutations are not diluted out by outbreeding with the larger population and (2) that selective pressures be substantially reduced so that the normal functioning of natural selection, which is to eliminate mutations, does not occur.

How does the purported recent human evolution conform to these conditions? In what subpopulation and where?

Neither conditions (1) or (2) are generally true. If a mutation is beneficial, then it will tend to rise in frequency, not be diluted out. (This depends on population size, which is a major point of the present work -- the larger the population, the more likely a beneficial mutation is to rise up and take over.) While demography can have all sorts of interesting effects, isolation of an evolving population is certainly not required. Second, the normal functioning of natural selection is not to eliminate mutations, but to modulate the frequencies of mutations in accordance with the selective (dis)advantages they confer. That is, deleterious mutations tend to be eliminated, and beneficial mutations tend to fix in the population.

Evolutionary theory requires that a rapidly evolving population be (1) isolated, so that the new mutations are NOT diluted out in the larger population and (2) that selective pressures be radically reduced so that the normal functioning of natural selection, which is to eliminate mutations, does not occur.

that is certainly a bizarre reading of evolutionary theory.

p-ter's critical analysis of the acceleration paper's genomic and statistical methods and assumptions strikes this non-expert as absolutely devestating, if it in turn doesn't have major holes in it. I'm in no position to judge.

I hope those sufficiently knowledgeable to do so will fully engage this argument. Perhaps even one or more of the article in chief's authors.

Perhaps even one or more of the article in chief's authors.

Hawks responds to P-ter's post, among other things, here.

Does anyone have a better link to the paper. The one on this page is broken.

Never mind I found it

PDF link http://www.google.com/url?sa=t&c...-MteeFu-1aQG- NA

HTML link
http://72.14.205.104/search?q=ca...clnk&cd=3& gl=us

David B. writes: Here is a report in the UK Times, which includes an obvious but nevertheless important comment from Armand Leroi.

He's referring to this:

Armand Leroi, Reader in Evolutionary Biology at Imperial College, London, said: “In principle, this could have led to speciation if it had continued. In practice, it has got to be the case that that cannot happen now. The reason is that this study has looked at largely separated populations in the past, but everything about human history since the Industrial Revolution weighs overwhelmingly against separation and thus against speciation too. Huge increases in gene flow are going to wipe this trend out.

How is this "obvious"? Please show me evidence of "huge increases in gene flow" into China, Japan, India, or West Africa. "Huge increases in gene flow" between continents are hardly a necessary or unstoppable consequence of the Industrial Revolution, except in the minds of those for whom "huge increases in gene flow" (though only to some areas) are politically desirable.

except in the minds of those for whom "huge increases in gene flow" (though only to some areas) are politically desirable.

can we not start with the politics yet? some of us are interested in the science first and would like that discussion to proceed before the usual suspects start cranking. gene flow happens to be a very powerful parameter for the record, surprisingly low levels can prevent divergence.

Consider how they're detecting selection-- they look for alleles that are at high frequency (but not fixed) and have extensive LD around them. Since haplotype length decreases with time, by definition, these are young alleles. Any old allele with a strong selection coefficient has long gone to fixation and is not detected (or more carefully, there's much less power to detect it). P-ter,

There are biases pointing in both directions. Recent selective events would disproportionately involve alleles that are just starting to fall or rise in frequency. This variation would be less likely to emerge above the noise in the data. This bias would favor alleles that are older and clearly responding to selection.

ok, since i'm enjoining a focus on science, gene flow

1) in neutral situations you need 1 migrant people populations per generation to prevent divergence. i won't get into that since we're not talking neutral here.

2) obviously there was some boundary between east asia and europe which allowed different light skin alleles to be driven to selection very rapidly. remember that the probability of fixation of a selected allele assuming large pop is 2s, where s is selection coefficient. that's for one mutant. obviously even if very modest gene flow from one population to another occurs then you have lots of copies of a rare variant which is rising in frequency in one population introduced into another. in other words, if the logic of acceleration is correct we may be accelerating even more due to moderate levels of between region migration because the effective breeding population just went up. and no, we don't need huge population transplants on the scale of the new world. small, but reasonable sized eurasian communities could do it.

surprisingly low levels can prevent divergence

Yes. But I would think that those conditions have always existed in the continental Old World. Genetic divergence must have been for environmental reasons.

However, I do think that Leroi is right because not only has the magnitude of gene flow increased everywhere, but environmental conditions are converging.

I would think that selective forces changed dramatically during the last 50-200 years (varying by population). In that case, what once was under powerful selection could now be nearly neutral.

But I would think that those conditions have always existed in the continental Old World. Genetic divergence must have been for environmental reasons.

i would too. but look at the way SLC24A5 just stops at the altai. we know that there was selection in east asia for light skin. but they tend to be different loci. so that means that selection was going 'faster' than gene flow in spreading around favorable mutants. and in this case the driver was likely the same. it isn't an either/or, it looks like TYR (another skin color gene) does span both areas, though i think it is somewhat older....

Recent selective events would disproportionately involve alleles that are just starting to fall or rise in frequency.

they've already conditioned on allele frequency. conditional on an allele being at, say, 70%, they have less power to detect it if it's old (because of the LD decay).

We had five peer reviews of the paper. They were all more or less useful but what they brought up were either minor matters of clarification, else utter nonsense, else of the "something is wrong here but I don't know what" genre.

Isn't it interesting that the only scientific serious criticism that we have gotten, the kind of criticism that peer review ought to provide, comes from a blog? Seems to me there is food for thought about what the internet is going to do to scientific publishing.

On a substantive matter, what is the error rate of the imputed phase data? Is it of a kind that would systematically bias inference? I have always thought it safest just to ignore it because I didn't trust imputation.

Henry Harpending

Hi Henry--

Thank you for your comments--I do think this paper contains some fascinating ideas, and that it will inspire some interesting research.

regarding phasing, the hapmap CEPH and yoruba data are in trios, which allows one to use the information from the child to help infer the phase of the parents. this is what the HapMap consortium did, and it's the gold standard--something like 0.1% switch errors. for data not in trios (like the hapmap asians, for example), the best algorithms do significantly worse (~5% error rate).

numbers are from here:
http://www.journals.uchicago.edu.../10.1086/ 500808

For the untutored, what does phased and unphased refer to? Where is the best primer?

what does phased and unphased refer to?

Someone who knows more can give a more precise answer, but here is the informal answer.

The LDD test looks at linkage disequilibrium, which is a fancy term for what we could call the Red Rover effect: if a gene is targeted by selection, a force is applied to it, and its neighbors are pulled along with it. Clearly that depends on how close they are to it -- the force decays the farther you are from the impact site.

So you might think you could just look for that pattern in the genome, and anytime you saw it, say, "Hey look, that gene just got hit!"

The trouble is, we have two copies of every gene, so there are really two games of Red Rover being played. If the data are unphased, imagine there being a two-floor gym with see-through floors, with a Red Rover game played on each floor, and you're looking straight down on it from the roof.

The players on floor 1 are exactly underneath the players on floor 2 -- so if you saw what you thought was the "that guy just got hit" pattern, how could you be sure? Maybe the players on both floors are arranged so that the overlapping of their patterns makes a Red Rover pattern, not necessarily that someone was just targeted.

You couldn't tell which people were from which floor, if you were looking straight down, so you can't be sure that people who appear to be dragged along are all on the same floor, near a teammate who was just hit.

One solution to this, which one of the authors of this paper thought up, was to just look at homozygotes. That's like the team on floor 1 essentially having a copy of itself on floor 2, and what happens to one happens to the other. In this case, you can be much more sure that if you see a Red Rover pattern, it's not the superposition of two non-Red Rover patterns, since there is effectively just one team for homozygotes.

And just to round out the picture, LDD shows recent selection since, given enough time, the teammates will get back into a straight line, leaving just the one guy who was targeted to stand out by himself.

For the untutored, what does phased and unphased refer to? Where is the best primer?

Consider diploid SNP genotypes measured at locations #1 through #5 along a single chromosome for a single individual:

Locus 1 - AA
Locus 2 - AG
Locus 3 - CC
Locus 4 - GT
Locus 5 - TT

Based on that (unphased) data alone, you can't tell what the genotype along each chromosome is. There are four (phased) possibilities:

AACGT
AGCTT

AGCGT
AACTT

AACTT
AGCGT

AGCTT
AACGT

You can determine this experimentally by sequencing across the region. You can deduce with some accuracy which phase is most likely given enough samples of unrelated individuals from the population. Or given the genotypes from related individuals, you can make a even more accurate estimate of which phase is correct.

I agree that the concerns about the LDD test introducing a methodological bias because of increased power for detecting recent selection seem to be critical to this finding. I'm not familiar with the LDD test but simulation studies should be able to address how strong of a bias would be introduced using the metrics applied in the current PNAS paper. Was that done here (perhaps in the supplement which I couldn't find online yet)? Or in a previous paper describing the LDD test?

our estimates of the long-term rate are based on things like the number of coding differences that have accumulated between humans and chimps since the split, not from linkage disequilibrium methods.

I'm not familiar with the LDD test but simulation studies should be able to address how strong of a bias would be introduced using the metrics applied in the current PNAS paper. Was that done here (perhaps in the supplement which I couldn't find online yet)? Or in a previous paper describing the LDD test?

I agree that simulations would be an excellent way of addressing this bias (and understanding how much of the effect it accounts for). as far as I can tell, none were done. in the previous paper, there was some attempt to address the characteristics of the test, but I find them lacking (footnote 3), and they definitely don't address the bias against detecting older sweeps.

our estimates of the long-term rate are based on things like the number of coding differences that have accumulated between humans and chimps since the split, not from linkage disequilibrium methods.

yes, but if I'm reading correctly, you estimate the current rate from the LDD test. there must be some cutoff of the test that reduces the number of current selective sweeps to a level consistent with the long-term rate, no? I'm willing to believe that this cutoff would be absurd, but I'd still be interested in hearing what it is.

if you don't believe the numbers from the LDD test (I don't), it's still possible to argue for acceleration. but the paper is largely based on the numbers from the LDD test.

I'm willing to believe that this cutoff would be absurd, but I'd still be interested in hearing what it is.

As a back-of-the-envelope estimate, we'd be looking at a cutoff of 99.9975 instead of 99.5 to be consistent with the long-term rate. The result would be only around 30 recently selected genes per population instead of around 3000.

That's imperfect, because the biggest outliers tend to fall in a window more recent than 10,000 years, so even as we threw out clusters, we would be concentrating events within the most recent part of history. So they would still tend to look like acceleration, just over a shorter time frame.


The Voight-Pritchard data show an acceleration of at least 30 times, but to be more precise I'd have to model the ascertainment with the iHS statistic better.

john,

thanks, that makes sense.

Apologies if this is a bit of a sophmoric question, but I'm still attempting to get to grips with evolutionary biology beyond the pop sci level. Anyway, I was reading this paper recently:

Nei, M. (2007) The new mutation theory of phenotypic evolution. PNAS, 104, 12235-12242.

Nei, towards the end of the paper, says the following:

Most molecular evolutionists are well aware of the importance of mutation in protein evolution. Yet, many investigators are trying to identify even the slightest trace of natural selection using various statistical methods (99–102). Using these methods, a number of authors have reported that a substantial proportion of amino acid substitutions are caused by positive Darwinian selection (103–106). However, the statistical methods used are based on many assumptions, which are not necessarily satisfied with actual data (18, 107, 108). Furthermore, their estimates of selection coefficients are often of the order of 10–6 (100, 106) and are unlikely to affect gene function (18). Note also that although these authors emphasized natural selection, they are actually estimating the proportion of mutations that are adaptive.

Now, looking through the Hawks paper, they don't seem to address this kind of critique of evidence for positive selection, as this AFAICT. Of course, there is no necessity for them to do so; if their statistical techniques and so on are strong, then the data stands on its own merits (particularly as I imagine they use different stats to papers 103-106, although I haven't read these yet). However, I was wondering what more knowlegable people than me think of Nei's argument here. In particular the point about the selection coefficients not being significant enough to impact gene function.

Thanks

For completeness sake, the references are:

18: Nei, M. (2005) Mol Biol Evol 22, 2318–2342.

99: Smith, NG & Eyre-Walker, A. (2002) Nature 415, 1022–1024.

100: Sawyer, SA, Kulathinal, RJ, Bustamante, CD & Hartl, DL. (2003) J Mol Evol 57, Suppl 1, S154–S164.

101: Zhang, J, Nielsen, R & Yang, Z. (2005) Mol Biol Evol 22, 2472–2479

102: Tang, H & Wu, CI. (2006) Mol Biol Evol 23, 372–379.

103: Clark, NL & Swanson, WJ. (2005) PLoS Genet 1, e35

104: Eyre-Walker, A. (2006) Trends Ecol Evol 21, 569–575.

105: Gojobori, J, Tang, H, Akey, JM & Wu, CI. (2007) Proc Natl Acad Sci USA 104, 3907–3912.

106: Sawyer, SA, Parsch, J, Zhang, J & Hartl, DL. (2007) Proc Natl Acad Sci USA 104, 6504–6510.

107: Hughes, AL, Friedman, R & Glenn, NL. (2006) Curr Genomics 7, 227–234

108: Subramanian, S & Kumar, S. (2006) Mol Biol Evol 23, 2283–2287.

SteveF's post is very perceptive and very nicely done with the inclusion of the references to which he refers. This is as un-sophomoric as I have seen.

We dealt with this at length in our response to reviewers. The short answer is that at megabase scales this doesn't work because of the numbers of mutations we are seeing. At kilobase scales this is a valid critique.

I will have to ask the rest of the gang for permission to post our response to reviewers about this: it may be of interest because Hawks wrote it and, like everything he writes, it is clear as a bell.

Henry

The Economist has an excellent and fairly long open access article on the Hawks, Cochran, Harpending et.al. acceleration paper.

they don't seem to address this kind of critique of evidence for positive selection,

I don't think they need to. the tests for selection described in those papers are over much longer periods of time (ie if you compare two genomes from different species, they want to know how many of the substitutions are adaptive, as opposed to hawks et al, who test within a species), and use completely different sources of information (substitution rate etc. versus linkage disequilibrium).

sabeti et al. have a nice introduction to all these methods and their distinctions here:
http://www.sciencemag.org/cgi/co...t/312/5780/ 1614

The UK Times today has comments on the acceleration paper by their science editor here

http://women.timesonline.co.uk/ t...icle3051334.ece

Not totally hostile, but critical of some of Henry Harpending's interpretations.

"The UK Times today has comments on the acceleration paper"

His point is to insist that conversation be kept politically correct, hence boring.

I wonder if it occurs to him (or anyone else) that our off-the-cuff about Swedes is much the same as Greg Clark's argument in "End of Alms".

Henry

"99.9 per cent of human DNA is shared by all" has become as much of a standard signal as references to Nazis have that people are about to be lectured on the need to be politically correct.

As long as you preface your comments with that stock phrase, you can probably go on to say pretty much whatever you want, as long as it is put in sufficiently abstruse languange for the drum beating activist crowd not to get the point, which means not very. The charming Irishman in Greg Clark pulls it off without having located the "industrial revolution gene", and points to a greater propensity for violence in hunter/gathers which is easily verified from observational data. I have no wish to see him beset by the Reign of Terror, it's just ironic when they fail to see essentially the same argument applied in a very slightly different context and feel entitled to lecture the speaker on the language he is permitted to use - it is code for "but....." or I have just lost this argument but have no intention of admitting it.

The reference to a "peace gene" is so chronically stupid as to disqualify the writer from ponitificating on the subject anyway - talk about dumbing down a subject in order try to prejudice people against a very interesting and potentially very useful thesis.

If people can't take the truth, they need to learn. If it doesn't fit their self-seving political agendas, tough.

Does the LDD test miss copy number selection events?

re: "99.9 per cent of human DNA is shared by all"

Based on the Venter diploid genome, Europeans differ by about 1%.

"Within the human genome there are several different kinds of DNA variants. The most studied type is single nucleotide polymorphisms or SNPs, which are thought to be the essential variants implicated in human traits and disease susceptibility. A total of 4.1 million variants covering 12.3 million base pairs of DNA were uncovered in this analysis of Dr. Venter’s genome. Of the 4.1 million variations between chromosome sets, 3.2 million were SNPs. This is a typical number expected to be found in any other human genome, but there were at least 1.2 million variants that had not been described before. Surprisingly, nearly one million were different kinds of variants including: insertion/deletions (“indels”), copy number variants, block substitutions and segmental duplications.

While the SNP events outnumbered the non-SNP variants, the latter class involved a larger portion (74%) of the variable component of Dr. Venter's genome. This data suggests that human-to-human variation is much greater than the 0.1% difference found in earlier genome sequencing projects. The new estimate based on this data is that genomes between individuals have at least 0.5% total genetic variation (or are 99.5% similar) The researchers suggest that much more research needs to be done on these non-SNP variants to better understand their role in individual genomics."
http://www.jcvi.org/press/news/ n..._2007_09_03.php

Related GNXP post on copy number variation.
http://www.gnxp.com/blog/2006/11...r- variation.php

"The reference to a "peace gene" is so chronically stupid as to disqualify the writer from ponitificating on the subject anyway"

I don't see anything wrong in referring to a 'peace gene', if there is some mutation (new allele) which reduces tendencies to violence. Evolutionary biologists use this kind of shorthand all the time. Presumably if Henry and co. believe that the change of the Swedes from Vikings to peacemakers is a result of genetic change, they must postulate the existence of 'peace genes'.

But the Swedes were pretty warlike as late as the 17th century, when under Gustavus Adolphus they invaded most of Germany and the Baltic states, so any genetic change must be very recent!

"But the Swedes were pretty warlike as late as the 17th century, when under Gustavus Adolphus they invaded most of Germany and the Baltic states, so any genetic change must be very recent!"

But they didn't have the wild and crazy berserk stuff very much by the time of Adolphus. I don't think that state warfare is the same thing as gang warfare, and the Vikings seem more like gangs.

Henry

But the Swedes were pretty warlike as late as the 17th century, when under Gustavus Adolphus they invaded most of Germany and the Baltic states, so any genetic change must be very recent!

add 100 years.
http://en.wikipedia.org/wiki/ Gre...at_Northern_War

I don't think that state warfare is the same thing as gang warfare

i think you can make the argument that the 30 years war was gang warfare ;-)

and if swedes have changed probably has something to do with cultural selection. between 900 and 1200 scandinavia went from being a collection of tribes and petty principalities to becoming integrated into the western european state system in the form we see them today.

But the Swedes were pretty warlike as late as the 17th century, when under Gustavus Adolphus they invaded most of Germany and the Baltic states...

A large part of those armies weren't ethnic Swedes, and in some battles mercenaries and/or Finns outnumbered them! This doesn't let the Swedes off the hook for warlikeness, but perhaps it demonstrates some dangers in deducing "national character" from history.

But the Swedes were pretty warlike as late as the 17th century, when under Gustavus Adolphus they invaded most of Germany and the Baltic states, so any genetic change must be very recent!

At the risk of piling on, I present some more pertinent examples:

But the French were pretty warlike as late as the 19th century, and look at them now, a bunch of wusses.

But the Germans were pretty warlike as late as the 20th century, and ...

But the Japanese were pretty warlike as late as the 20th century, and ...

The earliest reference I can find to the Swede illustration is Haldane. But we saw the same sorts of "changes" with Japan and Germany later in the 20th century, so why do folks still use the pre-war analogy as if it was a unique shift?

Aggressive nations or even gangs aren't necessarily made up of aggressive people. In fact people who are better at cooperating are better at organized violence.

Maybe Swedes then as now are just talented at cooperating.

Of course the Clark evolutionary forces were happening all over Europe, including Sweden, so we can suspect they really were getting less interpersonally violent over time as well.


(See chapter 8 of Bobbi Low's Why Sex Matters)

"Aggressive nations or even gangs aren't necessarily made up of aggressive people"

Gangs usually are, not nations. See for example the Al Valdez book. Vikings were aggressive nasty people by all accounts.

Finns were some of the best fighters in WWII but they don't seem to be aggressive people. But I will leave the matter of the Finns to Razib..........

Henry

Well, if you listen to those Swedes, Finns are above all else notorious for violence and pointless machismo. Hey, reminds me of my new favourite story! A while ago, while out drinking, I ended up talking to a couple of Swedish gay guys. After I had spent a good while making fun of them for being SWEDISH AND GAY (HAHAHA), I finally asked them WTF were they doing there anyway, since we do have gay places even in Finland. Well... "We looove talking to Finnish straight guys, you're just so... violent... and homophobic. That's so hot!"

As for Swedish history, they didn't stop being aggressive as a nation until losing Finland (that's 1809). They were finished as a Great Power after the Great Northern War, but like many former powers, they refused to change their self-conception and kept fighting until they got their ass handed back to them in such a major way that they had to wake up and the utter shock turned them into the pansies they've been ever since. In the 18th century, Sweden tried a few disastrous invasions of Russia to regain territory and even went for absolute monarchy in the end - the usual last straw of past powers, "we're weak because of these effeminate talkers, we need a real man to lead us!" Even after that, they could've kept Finland, if they wouldn't have been living in the past: they overestimated their strength and didn't go for alliances, they believed Finns to be beneath them and refused to appease the budding nationalist movement (pushing a lot of Finns to actually favour Russia) and the King believed in the rule of Kings, despising Napoleon and turning the only offer that could've protected Sweden into the deal that sealed its fate.

Anyway, nations are not aggressive because of aggressiveness of the nationals. This is a basic mistake of nerdy academic types who do not have sweet memories of dominance over puny weaklings to draw from. When you're talking about people who are aggressive in their personal lives, it's people who *enjoy the feeling*. Fist-fights are fun for the participants (unless they lose really badly), but war is mainly numbers on paper. The aggressive types do not go for that - they're aggressive for personal emotional or financial gain, not because they like the abstract idea of "being aggressive". The only time the psychology comes into play is whenever one leader gets enough power to start wars that will be equated with his name in history, then you can play for personal glory. (Also, it works when you're a bunch of bandits pretending to be fighting "wars", like the Vikings or some gang today.) At other times, wars are ideas of emotionally detached nerdy types who can derive joy out of some change on paper.

Gangs usually are, not nations.

I didn't say they were, I said aggressive nations, and even gangs, etc, aren't necessarily made up of aggressive people. (e.g. pre-war Japan and Germany)

I would disagree that Vikings were more like street gangs than military forces. They were successful conquering armies which established settlements over much of Europe and beyond.

Interpersonally, I don't know, but Wikipedia actually makes them sound more similar to Swedes than not:

Despite images of Viking marauders who live for plunder and warfare, the heart of Viking society was reciprocity, on both a personal, social level and on a broader political level. The Vikings lived in a time when numerous societies were engaged in many violent acts, and the doings of the Vikings put into context are not as savage as they seem.

But like I said, I do believe Swedes became less violent over time in the same way Clark argues for England ('middle-class' evolution references above). I just never agreed with Haldane's premise. One, for the reason above, but, two, because Vikings weren't synonymous with Swedes, just as Yakuza aren't representative Japanese. Haldane had no comparable psychometric data for ordinary Swedes of 1000 years ago. We don't know how they compare to Swedes of today or other medieval ethnic groups.

This discussion is drifting around (as usual).

Henry's original argument was (roughly):

a) the Swedes used to be violent (Vikings, etc)

b) now they aren't

c) this may be due to a genetic change.

I pointed out that the Swedes were still violent in the 17th century, so the change must be very recent (and maybe even more recent, according to Jakkelli).

Henry countered that 'state violence' is different from personal or gang violence.

Well, maybe it is, but if so, the absence of state violence from the Swedes in the last century or two doesn't prove that the Swedes are less personally violent than in Viking times. Of course, the Swedes no longer sail around in longboats raiding monasteries, but we hardly need to postulate a genetic reason for that.

Measuring the propensity for personal violence is no easy matter, even when modern statistics are available. E.g. homicide rates are not much of an indicator, since they depend on availability of weapons, standards of medical care for the injured, and so on. Britain has one of the lowest homicide rates in the world, but I wouldn't claim that the British are especially peaceful, especially during football matches or around pub closing time.

One, for the reason above, but, two, because Vikings weren't synonymous with Swedes

If you're talking about those Vikings that raided and settled all over Western Europe, they weren't Swedes at all!

The Swedes were bottled up behind the Danes and the Norwegians, they couldn't just go raid England whenever they wanted to. In activity, Swedes were the least close to the movie stereotypes of berserk raiders. They had only eastern Europe to go to and they went around as traders, mercenaries and sort of urban settlers, actually opposite to the idea of barbarians raiding civilized lands - the Swedes were politically more advanced than northeastern Europeans and went around founding states in the middle of perpetually warring undercivilized Finnic and Slavic tribes. (Some of us still "remember" the Swedes as the "Rus" - in Finnish, Sweden is "Ruotsi". They even founded Russia, the bastards! It's like all they do is part of some cosmic plot against Finns.)

Thanks to Henry for his response and also p-ter for his link to the Science article. I shall have a peruse.

David B, did the Swedes actually ever raid a monastery? As far as I know, they had their churches raided, not the other way around. That's one reason why they expanded east - to stop us barbarians raiding them.

As for Sweden's post-Great-Northern-War wars of aggression, Sweden first "attacked" Russia during the 1740s, during another one of those now-forgotten mini-world-wars. This war was such a disaster that it's hard to call it Swedish aggression, though. It started off with the Swedes declaring war on Russia and their intention to take back their lost territory and waiting around for months without actually doing anything until the Russians finally attacked and ended up occupying all of Finland. (Yes, seriously. We don't have to make up stuff to make them look stupid.)

They had a second attempt near the end of the century when Russia was busy fighting someone (Turkey, maybe). This time they actually worked up a pretext, with the classic "a few Swedes dress up in Russian uniforms and kill some Finns" maneuver. They got nowhere and only escaped the massively humiliating peace terms Catherine the Great had planned for them when the French Revolution intervened and she suddenly had to buy allies.

Both wars are pretty well forgotten in Sweden, but remembered somewhat bitterly in Finland. The first one for the utter failure of Swedish defense which brought Russian occupation to Finland and the second one for destroying much of the remaining feelings of loyalty to Sweden. By then Finns were so fed up fighting Sweden's wars that Finnish officers actually rebelled and talked to Catherine on their own - the plot was foiled, but at least the King had to fear for his life when commanding his troops...

Finns were some of the best fighters in WWII but they don't seem to be aggressive people.

Riight... less than 100 years from one of the bloodiest civil wars in modern times, and counting!

Well, if you listen to those Swedes, Finns are above all else notorious for violence and pointless machismo.

Jaakkeli, if your Swedish is up to it, you might enjoy this Swedish TV comedy, part 4, about 6 to 11 minutes in.

"David B, did the Swedes actually ever raid a monastery? As far as I know, they had their churches raided, not the other way around. "

I stand corrected. It was those nasty Danes and Norwegians.

On the general question of violence in Sweden, this http://en.wikipedia.org/wiki/ Cri...Crime_in_Sweden
might be useful.

It occurs to me that there is a correlation between personal violence and consumption of alcohol. The colder European countries (including Russia) have a reputation for drunkenness and associated casual violence. In Norway (I think) the sale of alcohol is tightly restricted for this reason. The Mediterranean countries, in comparison, drink plenty of wine but don't get steaming drunk.

Oh yes. Evidently it is not the proportion of people who drink that counts, it is the subset of those who get dangerously drunk that correlates with personal violence.

It's the Danes who did the monastery-raiding, the ones who recently apologized for invading Ireland.