/sci/ - Science & Math

[Anonymous No.6323626]

So what is the scientific consensus behind human races? All of my research either points to some unscientific "BUT WE ARE ALL ONE RACE/RACES DONT EXIST" bleeding heart bullshit or to some just as unscientific "THIS IS WHY BLACK PEOPLE ARE DUMB AND I AM JUSTIFIED IN WANTING TO ERADICATE THEM" racist garbage and I'm really interested in developing a scientifically rigorous model of race.

[kitty !kitty/EnNg No.6323629]

something close to dog breeds

[Anonymous No.6323635]

>>6323629
Actually no, dogs have a much higher rate of speciation and the genetic distance between any two dogs of different breeds is orders larger than the genetic distance between any two humans.

OP it's pretty much considered impossible to define rigorously and uniquely. There was one main proponent who gave finally gave up around one or two decades ago.

[Anonymous No.6323637]

>>6323626

different human races are definitely subspecies.

Darwin said this:

>There is, however, no doubt that the various races, when carefully compared and measured, differ much from each other,—as in the texture of the hair, the relative proportions of all parts of the body, the capacity of the lungs, the form and capacity of the skull, and even in the convolutions of the brain. But it would be an endless task to specify the numerous points of difference. The races differ also in constitution, in acclimatisation and in liability to certain diseases. Their mental characteristics are likewise very distinct; chiefly as it would appear in their emotional, but partly in their intellectual faculties. Every one who has had the opportunity of comparison, must have been struck with the contrast between the taciturn, even morose, aborigines of S. America and the light-hearted, talkative negroes. There is a nearly similar contrast between the Malays and the Papuans, who live under the same physical conditions, and are separated from each other only by a narrow space of sea.

Whites. for example, have a lot (some say around 20%) of neanderthal genome and blacks don't. So we're definitely a different subspecies.

[Anonymous No.6323647]

>>6323637
The last century called, they want their pseudoscience back.

[Anonymous No.6323650]

>>6323637
define species

[Anonymous No.6323653]

>>6323647
>creationist trolling

So you're saying that Darwin and his concept of evolution are nonsense? Go to bed dude. God demands it.

[Anonymous No.6323655]

>>6323650
I said subspecies. and look it up in a dictionary.

[Anonymous No.6323659]

>>6323653
Saying that these antiquated hypothesis don't make valid arguments in an era of genomes.

[Anonymous No.6323660]

>>6323655
Pop sci retard detected. Gb2rebbit

[Anonymous No.6323661]

>>6323626
Basically what happened was anthropology was used to justify some of the Nazi atrocities and so the consciously and intentionally took a step away from things which were scientifically verifiable and went with absolute political correctness

[Anonymous No.6323662]

>>6323660
> Gb2rebbit

uh oh, that sure told me.

[Anonymous No.6323665]

[Anonymous No.6323667]

>>6323661
Not realizing the social sciences are considered pseudoscience on /sci/.

gb2/lit/

[Anonymous No.6323668]

>>6323667
I'm just explaining the current toxic culture in that field

[Anonymous No.6323674]

>>6323626
/pol/ what are you doing? GO BACK TO YOUR CAVE PSEUDOSCIENCE

[Anonymous No.6323676]

>>6323674

You don't understand, I'm interested in this BECAUSE /pol/ is full of pesudoscience racist dreck and reddit is filled with bleeding heart liberal nonsense.

[Anonymous No.6323681]

>>6323668
Social science threads on /sci/ are considered trolling.

[Anonymous No.6323687]

>>6323676

You want to be able to argue about retarded shit with other regards? I hope you haven't convinced yourself you can get either group to change their ways through logic and reason.

[Anonymous No.6323690]

>>6323687

It's that I know what they spout is retarded shit and I want to educate myself on the realities of the situation, not on heavily politicized narratives. I want to learn this shit for my knowledge.

[Anonymous No.6323691]

>>6323681
biology and evolution is not social science you retard.

[Anonymous No.6323699]

Aren't races simply humans that evolved over time to their environment?

For example, darker skinned people lived in areas with increased levels of sunlight, or the Incans lived high enough that their lungs developed increased capacity.

[Anonymous No.6323701]

>>6323691
You're right, but what you're posting is neither. Anthropology and sociology have no rigor or controlled experiments. It's witchdoctor tier.

[Anonymous No.6323723]

>>6323699
Not quite. I mean intuitively it is, but the problem occurs when trying to define different races. Suppose you paid attention to traits like intelligence, disease, fitness, etc.. and fed the data to a machine learning clustering algorithm. You would end up with an entirely different set of racial groups. A much more useful partitioning at that. However you wouldn't be able to tell what race people are just by looking at them. Your neighbors could be from an entirely different race and you wouldn't know. Worse still clustering algorithms, and clustering in general aren't convergent. You could get lots of different partitionings with the same exact data. In other words current notions of race are not only largely useless but there's nothing special about them.

[Anonymous No.6323746]

>>6323723

Sounds like you're selecting for the wrong traits. If you pick the right traits you will find a unique and distinct pattern.

[Anonymous No.6323810]

>So what is the scientific consensus behind human races?

Not much, take individualistic genetic differences, apply them to groups, raise or lower the extremes of those differences in moderation, spread out those values statistically, select several traits that work as visual identifiers for each group, keep them all genetically similar to interbreed but separate enough to retain their genetic differences to some degree.

Basically it all boils down to a bunch borderline autists arguing over the varying degrees of systematics and the eventual semantics that comes from it without quite knowing all the specifics to it.

[Anonymous No.6323825]

>>6323746
Clustering algorithms don't work on specific traits only. They work on ALL traits you give them. The more traits you give them the better. Regardless, they're not convergent. It literally does not make sense for them to be convergent.

[Anonymous No.6323828]

>>6323626
When mankind gathered at the Tower of Babel, and meant to make a tower to the gods, YHWH disbursed them by confusing their languages and they scattered throughout the earth, as intended.

Oh, I'm sorry, you said "scientific". Well, then, we don't know, but we're looking into it, we have a few good theories, nothing just yet, but we'll get right on it!

[Anonymous No.6323833]

>>6323828
>we have a few good theories, nothing just yet, but we'll get right on it!

Actually almost all of the hypothesis have been discarded in the hard sciences. Only the social sciences still take them serious but it's not surprising considering no one takes the social sciences seriously.

Also, OP
http://www.nyu.edu/gsas/dept/philo/faculty/block/papers/Heritability.html

[Anonymous No.6325542]

>>6323637
Whites don't have 20% Neanderthal. It's around 2.7%. I have 3.2% and I'm in the 99th centile of europeans according to 23andme.

[Anonymous No.6325550]

>>6325542

quick search shows this:

>Surprise! 20 Percent of Neanderthal Genome Lives On in Modern Humans, Scientists Find

http://news.nationalgeographic.com/news/2014/01/140129-neanderthal-genes-genetics-migration-africa-eurasian-science/

[Anonymous No.6325554]

>>6323699
Yes.

>Not quite. I mean intuitively it is, but the problem occurs when trying to define different races. Suppose you paid attention to traits like intelligence, disease, fitness, etc.. and fed the data to a machine learning clustering algorithm. You would end up with an entirely different set of racial groups. A much more useful partitioning at that. However you wouldn't be able to tell what race people are just by looking at them. Your neighbors could be from an entirely different race and you wouldn't know. Worse still clustering algorithms, and clustering in general aren't convergent. You could get lots of different partitionings with the same exact data. In other words current notions of race are not only largely useless but there's nothing special about them.

This is all wrong. Groups are not decided on traits. They are decided on genetics. If you factor the analysis on genetics, you get this result or something very much like it.

[Anonymous No.6325562]

>>6325550
Which agrees with me... "people living outside Africa today are composed of some 1 to 4 percent Neanderthal DNA.".

[Anonymous No.6325567]

>>6323833
>no one takes the social sciences seriously.
>links to a paper from a newspaper written by a philosopher on the topic of not-philosophy.

The writer starts out with a strawman too about The Bell Curve.

Yeah, it's pretty retarded. Don't link people to crap like that in the future. :)

[Anonymous No.6325588]

>>6325542
>>6325550
>>6325562
If you guys actually kept up with research then you'd see that there's a lot of back and forth on this currently. Some peer reviewed papers attribute the overlap in DNA to a much older common ancestor to both homo sapiens and neanderthals. Others claim that there's evidence of isolated interbreeding in Italy. It goes on and on. We're on the bleeding edge of the cutting edge with this stuff right now so there's really no clear consensus. Pop-sci articles are retarded to be publishing articles the way they are, but then that's also what sells.

[Anonymous No.6325604]

The scientific consensus pretty much has to be "all races are the same" because anything more complex than this and people will either say science is racist or use what they don't understand to justify discrimination.

This position is completely understandable and I espouse it publicly to avoid stupid discrepancies.

I don't have a great biology education but I am reasonably literate in evolutionary biology and psychology. Its almost certain there are at least some few small biological differences. As for things like intelligence, behaviour and violence, this is empirically speaking based on environment. If you're rewarded for your behaviour it will continue, this applies for beating someone to a pulp or training for the international math Olympiad. The first scenario could be the case if you're born in a bad neighbourhood and have to defend yourself, the latter could apply for those going to good schools and who are encouraged towards intellectual pursuits. What you're rewarded for will determine how you end up as an adult. This isn't a very complex concept to grasp, but many on this board seem to struggle with it.

I hope people will read this without bias and at least some of you will able to grasp it this time.

[Anonymous No.6325601]

>>6325554
You fail at reading comprehension and genetics. Groups are decided on traits before hand. They chose all of those different races first and then found genetic markers for identifying each of those races. Then they modeled the markers in a tree. The problem is that their flow of logic follows like this.

>Assume this set of races exists.
>Find genetic markers for these races.
>Show that these genetic markers partition the set of humans.
>Therefore this set of races exists.

If you don't see the flaw in logic then I suggest you take a proof based math course. You cannot assume something is true to prove it is true. They could have just as easily chosen a different set of races.

[Anonymous No.6325608]

>>6325567
>Pointing out that arguing on X-level isn't rigorous and full of contradiction.
>Gives an example by posting a contradictory article that also argues on X-level.

Are you illiterate?

[Anonymous No.6325611]

>>6325604
>The scientific consensus pretty much has to be "all races are the same" because anything more complex than this and people will either say science is racist or use what they don't understand to justify discrimination.

Only idiots who've never worked in science believe this.

[Anonymous No.6325616]

>>6325611
From what I've seen it is, elaborate.

[Anonymous No.6325630]

>>6325611
From what I've seen scientists generally don't try to proliferate racist ideas for numerous reasons.

Maybe the scientists you know do, but that of course can't be generalized to everyone.

[Anonymous No.6325649]

>>6325616
>>6325630

The only controversies about scientific censorship since the turn of the millennium are stuff like this.
http://www.endsciencecensorship.org/

Outside the social sciences the term race isn't even taken seriously anymore because it's not rigorous. When one writes a study that affects people with some sets of genes they don't say it affects a specific race, they say it affects people with some set of genes. Pop-sci articles like to try and correlate that to real world terms so they say "it's more likely to affect asians" or something non-rigorous (are russians considered asian?).

[Anonymous No.6325771]

>>6325601
>Assume this set of races exists.

This is not how it works. There is no specific "set of races". The number of groups is always arbitrarily chosen. One can take a the same dataset and divide it into groups, going from N=2 to e.g. N=10 to see how it works.

The thing is that the groups that emerge from the data are NOT arbitrary. That's why these methods (explanatory factor analyses and the like) are said to allow the data to speak for itself.

One can see it here, where a guy goes thru N=3 to 15.

dienekes blog spot com/2010/12/human-genetic-variation-first.html

[Anonymous No.6325775]

>>6325604
g is mostly genetic, with heritability in the area of 80-90%. Shared environment is not important.

Racial differences in g are most likely partly genetic due to selection for different environments over the last 100k years or so.

[Anonymous No.6326007]

>>6325771
Yes, but clustering is not convergent. It would help if you knew some actual math.

http://en.wikipedia.org/wiki/Cluster_analysis#Clustering_algorithms

The groups that emerge from the data are not arbitrary but they are not convergent (this means you can end up with tons of different partitions or tons of different numbers of groups). Worse still when you work it from the backend you have no guarantee that any of the groups you get are the same as the "races" you think they represent.

[Anonymous No.6326009]

>>6325775
>most likely
what is rigor?

[Anonymous No.6326015]

>>6326007
>>6325771

By the way, the data in that guys BLOG (lol) post is constructed using the mclust clustering algorithm. If he ran it several times he would get different clusters. He would also get different clusters if he used a different clustering algorithm.

[Anonymous No.6326035]

>>6326015
Perhaps slightly different results. But the overall patterns are the same every time. First split is always Africans and non-Africans, for instance. Northern europeans never end up with e.g. Thai. And so on.

[Anonymous No.6326067]

Can people who've never taken a graduate genetics course admit they have no idea what they're talking about and stop with the race bullshit?

[Anonymous No.6326075]

>>6326035
No, that's exactly the problem I'm talking about. The cluster that you call "Africans" on one run may not be the same as the cluster you call "Africans" on another run. It's possible the cluster might not even exist in some runs or that it absorbs other clusters.

Clustering Algorithms are part of a larger class of algorithms called Machine Learning Algorithms. Machine learning algorithms in general are typically used to predict stuff that you don't know but could figure out. Think of them like linear regression on steroids. They have several problems that affect all of them including stuff like over-fitting, under-fitting, and just bad results. When you implement an algorithm you normally have four or more separate data sets that you use both to "train" the algorithm and to test it. The data comes in two forms, one where you know the "answer" (the thing you're trying to predict, in this case what race an individual is) to it and one where you don't. You split the data that you have answers to into two or more sets (some for training and some for testing). Over-fitting occurs when you train your algorithm on the training data and it fits really well and makes really good predictions but then makes really bad predictions on the test data. In under-fitting it just sucks on both. Look at the picture for a very simplified example (normally your data has thousands of dimensions, not just two). There's a lot of very hard work that goes into making sure that an algorithm is implemented and tuned correctly to the data. Companies spend lots of time working on these and it's not a simple thing. Statistical clustering algorithms like mclust are prone to overfitting and since there is no proper test data (a list of samples who's races we do know) then there is no way to check for overfitting or underfitting.

cont in next post

[Anonymous No.6326102]

You know what else isn't rigorously defined? Life. We don't have a strict definition.

>It is a challenge for scientists and philosophers to define life in unequivocal terms.[26][27][28] This is difficult partly because life is a process, not a pure substance.[29][30] Any definition must be sufficiently broad to encompass all life with which we are familiar, and must be sufficiently general to include life that may be fundamentally different from life on Earth.[31][32][33]

>Since there is no unequivocal definition of life, the current understanding is descriptive. Life is considered a characteristic of organisms that exhibit all or most of the following characteristics or traits:[32][34][35]

And yet we don't quibble about it like we do with race. Race is just a useful way of grouping individuals by likeness.
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1196372/
Of 3,636 subjects of varying race/ethnicity, only 5 (0.14%) showed genetic cluster membership different from their self-identified race/ethnicity.

>>6326075
So certain algorithms are better at displaying the patterns we see than others. That's all that boils down to.

[Anonymous No.6326109]

The Bible says human races don't exist. We are all made in God's image, one race - human.

>yfw you realize the modern race denial movement is just rehashed Christardianity.

[Anonymous No.6326112]

>>6326075

Clustering Algorithms specifically are worse because unlike other types of machine learning algorithms, they are not convergent. Ultimately there is no way of telling if the clustering you've achieved is the best or if there exists a better one. You more or less just train it a ton of times and choose the one that works best for you. This guy performed neither of those on his blog. Worse he went ahead and named the clusters ahead of time as if they were the same ones. This is not just misleading but it's wrong.

It would be possible to configure a clustering algorithm so that the set of samples that you already decided on being called "Northern European" (before running any measurements or experiments) never end up being clustered with the set you call "Thai" (again for no legitimate reason). Since there's no possible test data either then there's no way to tell if there's anything wrong about it either.

Clustering algorithms are used all over nowadays, they are what google runs it's news clustering service off of (clusters news articles by story). Even then it is not perfect (as it sometimes throws articles into the wrong story). Now take into account that in that Google made that implementation AND there is an overabundance of test data out there to check how well the algorithm is working. This blog post isn't at all rigorous.

>>6326102
>And yet we don't quibble about it like we do with race.
Yes we do. Even moreso than race.

>So certain algorithms are better at displaying the patterns we see than others. That's all that boils down to.
No, you fundamentally misunderstand what I said. The same algorithm on one training session can perform extremely well and on another training session can perform extremely poorly.

[Anonymous No.6326119]

Are There Really Different Races?
http://www.answersingenesis.org/articles/nab/are-there-different-races

[Anonymous No.6326127]

>>6326112
>>6326075

All that said, clustering algorithms are definitely not useless in regards to genetics when you've got test data. One might for example be interested in finding clusters that are prone to certain diseases. This would in a sense be analogous to races but no one wants to be told that they're in a race defined solely as "being prone to developing diabetes".

[Anonymous No.6326137]

>>6326112
>No, you fundamentally misunderstand what I said. The same algorithm on one training session can perform extremely well and on another training session can perform extremely poorly.
Fair enough, man. Then repeat the simulation a bunch of times and see what is the most likely outcome. Problem solved. So a few times in 1000 runs you don't get the clustering that that makes sense.

[Anonymous No.6326141]

>>6326112
>No, you fundamentally misunderstand what I said. The same algorithm on one training session can perform extremely well and on another training session can perform extremely poorly.
Fair enough, man. Then repeat the simulation a bunch of times and see what is the most likely outcome. Problem solved. So a few times in 1000 runs you don't get the clustering that makes sense, that's fine.

[Anonymous No.6326147]

>>6326141

Exactly. Except that in this case we have to decide beforehand what races exist and what samples fit into those races so that we can tell which clusterings make sense and which don't.

In other words, clusterings A and B both have our 50,000 samples split into 30 different clusters. However, the clusters in A and B are different sizes and contain different samples. How do we know which one is right if we don't already know which race a sample is "supposed" to be?

This is where the circular logic in >>6325601 fits in.

[Anonymous No.6326161]

>>6326147
So instead compare full genomes to see how they cluster by similarity and then see if there are markers for it. Problem solved.

[Anonymous No.6326170]

>>6326161
Right, in which case we would get a different set of races than what we have now. Although they would be a lot more useful and interesting than the garbage we have now.

[Anonymous No.6326171]

>>6326119
What are you doing man?

[Anonymous No.6326176]

>>6326161
And they did just that.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1934381/

[Anonymous No.6326185]

>>6326170
I wouldn't be so sure about that. These populations were partially isolated through history. I'd be really surprised if taking entire genomes we see something else.

[Anonymous No.6326199]

>>6326185
Why do you think that? Partitioning on Y-DNA and MT-DNA aren't really that similar after all.

[Anonymous No.6326204]

>>6326199
But that's not the entire genome, just haplogroups. Even these biomarkers you referred to would mean there is some more dissimilarity between populations already just because those sequences wouldn't match up.

[Anonymous No.6326210]

>>6326176
This isn't exactly the same thing as it is not trying to determine races but populations (which do have test data), but thank you for the link.

[Anonymous No.6326215]

>>6326204
You're right, I was just pointing out that in the same way those two are biased so could our conventional notion of races be biased. It's not like we're seeing an entire genome when we look at someone's skin color.

[Anonymous No.6326220]

>>6326210
Aren't those synonyms more or less in this context? Is it just the history of the word which is so bad? It's just a word, don't let it frighten you. You're stronger than that.

[Anonymous No.6326225]

>>6326215
I don't just look at skin color to determine someone's race, I think that's a bit of a strawman. Look at albinos. I can still tell when looking at an albino if they're African, Asian of European, just to use the basic scheme. I met an albino Ethiopian once, cool guy, though I could tell he was not a European guy.

[Anonymous No.6326231]

>>6323626
>unscientific "BUT WE ARE ALL ONE RACE/RACES DONT EXIST" bleeding heart bullshit
We are one race you dumb faggot, who ever told you some humans are inferior to others because of genetics is pseudo scientific and should go to >>>/x/ or >>>/pol/

[Anonymous No.6326232]

>>6326220
No, populations are more rigorously defined.
http://en.wikipedia.org/wiki/Population_genetics

Depending on who you ask you'll get different lists of races. "black", "white", "mexican", "european", "caucasoid", "negroid", "mongoiloid", "hispanic", "amerindian", "african", "asian", "japanese", "sub-saharan", "indian", "siberian", etc.. as examples. It's even worse when you ask these different people to classify people as different races. I remember passing by /pol/ out of curiosity during the protests in Greece and there were threads arguing about whether or not Greeks were considered "white".

Populations are a partitioning based on some somewhat quantifiable data. It isn't necessarily the only partitioning or the best one, but it is a useful one for its purposes.

[Anonymous No.6326237]

>>6326225
Even then you're still only seeing a small fraction of their genome. You might even be mixing in some information that isn't genetic like their accent/dialect, clothing, eating habits, beliefs, name, etc..

[Anonymous No.6326238]

>>6326232
/pol/ is shit. It's better not to pay attention to them. What about that picture posted here >>6326102, it seems like s synonym taking that definition.

[Anonymous No.6326245]

>>6326237
That's true but that small fraction does have an impact. Though the guy was wearing Western clothing and had a British accent. I could tell based on his facial structure he wasn't Anglo Saxon.

Sometimes I can even tell what specific ethnicity someone is fairly accurately.

[Anonymous No.6326248]

>>6326238
That picture says that all populations can be thought of as races but not all races can be thought of as populations.

[Anonymous No.6326253]

>>6326245
Ethnicity and race are different things anon. Ethnicity refers to social/cultural stuff, not necessarily genetic stuff. Anglo-Saxon might not even be a distinct race in another partitioning. Either way, the issue isn't whether or not you can use that small fraction to determine if antiquated notions of race can be discerned about an individual. The issue is that if you did partition individuals based on the entire genome via an algorithm then said algorithm may see things that we don't see and may partition them differently.

[Anonymous No.6326254]

>>6326248
I think you got that in reverse. All races are populations but not all populations are races. But specifically races are populations which are more genetically similar to each other than compared to other populations which is what that study did. So yeah in this context it's synonymous using that definition.

Anyway, man I enjoyed this conversation but I have to run. It was a pleasure.

[Anonymous No.6326269]

>>6326254
I understood it the opposite way. In particular the sequence of sentences right after the underlined section.

>Such fine distinctions are rarely useful, however.
It is not useful to say that each village is a race and that all the villages together are a race.
>It is usually more convenient to group populations into larger units that still qualify as races in the definition given.
Populations are "large" races, it is more useful to talk about populations specifically than just any race (as defined above).

[Anonymous No.6326270]

>>6326254
>>6326269

and yea, I was hesitant to participate in the conversation but it went better than I expected.

[Anonymous No.6326668]

>>6326102
>>6326075
This issue reminds me of some people who quibble endlessly about the factor indeterminacy problem in factor analysis. It's a waste of time, since the math works out good enough for solving real life problems, and the factors identified correlate ~1 with each other.

>>6326185
We don't. Turns out that traditional racial classifications, analyses based on a few markers, and those based on SNPs, and full genomes now are in good agreement.

The people who don't like this finding are the ethnic nationalists (e.g. Hitler Nazis) who wanted a pure breed, and the feel good-anthropologists (who deny races). Evidence isn't good to any of these groups. Since there are races/populations/ethnic group/subspecies/whatever (exact term used is arbitrary), this doesn't square with the ideologically based view of (2) and the negative effects of inbreeding effects doesn't square with (1).

[Anonymous No.6326676]

>>6326668
What about the rigorous scientists who don't agree that it works out "good enough for solving real life problems"?

>Turns out that traditional racial classifications
Like "white" and "negroid"?