PBS lies in an attempt to prove that race does not exist

[ranskaldan]

In other words: isn't this how race works? It is a continuous spectrum, BUT with certain clusters in distribution. And if so, why would classifications be "unscientific"?

EDIT: that should say "Each dot", not "Each point" at the bottom.

[St Leo]

Originally posted by Japher
It would be funny to see a Bull Masive try to mate with a some sort of minature dog!

Especially if the miniature dog is the male in that (presumably heterosexual) relationship.

[Mao]

A male bull...well I'll be damned...

[Ecthy]

ranskaldan

[ranskaldan]

Originally posted by Ecthelion
ranskaldan

: bows :

Awaiting reply from a race-denier.

[Odin]

There are fairly well defined groups of people, but they should be called ethnic groups because race has become such a loded term, and "race" is usually a few larges groups wile there is actually many small groups.

[MichaeltheGreat]

Originally posted by GP


So is the concept of different breeds of dog useless because we lack a convenient DNA marker for each breed or because there are cross-breeds or because there is variation within breeds?

Actually, dog breeds are the best example of why the concept of race for humans is useless.

Let's pick on Brain, our purebred Miniature Schnauzer, for an example. Standard Schnauzers are an old German breed of working dog, dating back to at least the 15th century. They were cattle drovers, guard dogs, cart haulers, etc. So they were purpose bred for both strength (they're muscular, barrel chested, and good hip and lower back drivers) and herding instincts, which is an unusual combination. The idea was the do-everything-on-the-farm one dog fits all farmer's dog.

In the late 1800's, breeders decided to miniaturize and improve the line to give them better outdoor durability and vermin hunting characteristics (a la terriers), while preserving their other traits (except the cart hauling, which was no longer needed and purely a function of size. So smallish Standard Schnauzers were bred with Affenpinschers, and two of the three allowable coat types, salt and peppers and solid blacks, also were selectively bred with Poodles, Belgian Griffons, and wolfgrey Spitzes to get specific features from those dogs.

To get the desired characteristics, thousands of dogs were bred over time, with those meeting the desired specs (or intermediate mixes) allowed to breed on, and those who didn't, being killed off. Once there was a small pool of dogs established with the characteristics that were wanted (call these the proto-Schnauzers ), they were inbred with each other for several generations, with spec dogs living to breed, and non-spec dogs summarily killed off by the dog-nazis, ooops, breeders. All of the Mini-Schnauzers alive today are descended from the original descendants of the proto-Schnauzers. In fact, almost all of them can trace their ancestry directly to three male proto-Schnauzer descendants and a progressively larger pool of females, who were themselves descendants of the original female proto-Schnauzers.

"Official" (i.e. Breed club members) breeders of registered dogs today are still supposed to kill off puppies who don't meet the specs, so they don't "contaminate" the bloodlines. Of course if you sell the dog without papers, you make money, and there's no problem, because undocumented migrant dogs never get accepted in proper dog society.

Both mtDNA and nuclear DNA can demonstrate that a dog is in fact likely to be a Miniature Schnauzer, but a look and a ruler and the breed standard would tell you that for a lot less money, or simply checking out the AKC (or whatever national breednazi organization applies) registry.

Herein the problems lie for humans and "race." mtDNA can tell you if a Mini-Schnauzer is in fact descended from one of the female proto-Schnauzers, but not tell you if the stray on the street got into the mix one day. Nuclear DNA can tell you close degrees of relations to known dogs, or general degree of genomic similarity to proven dogs, but you're screwed unless you know the entire genome, and all the ways a particular gene can be expressed. You're also screwed with false negatives if an insignificant random mutation, or non-critical (i.e. no effect on the breed standard) recessive gene is present.

Humans don't have breednazi registries (well, maybe that Aryan dating club thing does), we don't have a history of selective breeding and culling to create proto-races which are then fixed by inbreeding (you get that last part in the Ozarks, parts of the Appalachians, and various Indo-Pacific Islands, but without the selective breeding and culling)

With humans, you can identify several thousand distinct genetic groups with mtDNA, and the nuclear DNA picture is even muddier.

If all you end up with is something to the effect that there's a few hundred to a few thousand genetically distinct groups (let's call them gene tidepools since they're isolated subsets of the human genepool), why bother with race or zace, instead of "tribes" or "clans" (or "vibes" or "klaverns" if the traditional terminology bugs people), because all you're saying is an isolated community that only breeds within itself ends up with a distinct and limited genepool, who cares? Compare "hey, those Dinka tribesmen sure are tall." with "hey, those Dinka tribesmen sure are tall because they come from a limited genepool." Big important difference there.


It doesn't mean anything useful in either a scientific or sociological sense, especially when you can't make hard and fast membership rules because nobody can agree on anything other than a known bloodline basis, and most people represent genetic influences from a dozen or more of these gene tidepools, maybe even a hundred or more. It's sure a hell of a long way from the textbooks I saw as a kid that lumped humans into five specific races.

[MichaeltheGreat]

Originally posted by GP


How many breeds of dog are there? Just because the exact tag is not known or because there is variation and cross-breeding, does that mean there is NEVER any use in using the concept?

Just "official" breeds, or do we want to include things like Labradoodles and Schnoodles?

"Breeds" or "strains" and whatever the plant version is called are useful, because they are purpose bred and created for specific characteristics, and a lot of effort goes into creating those characteristics. They're also distinctly non-random, since the average male dog will hump any female in heat it can find without concern for maintaining the bloodline or the breed standard. Since human breeding is relatively random to extremely random, it's really not a useful concept for humans.

[MichaeltheGreat]

Originally posted by GP


What are the genetic markers for breeds of dog?

Old fashioned controlled breeding and bloodline management is more cost effective, but if there's ever a dog genome project, you'll be able to genetically identify (probable) members of specific official breeds much more readily than you would ever be able to identify group differences in humans.

[MichaeltheGreat]

Originally posted by ranskaldan


: bows :

Awaiting reply from a race-denier.

Nice idea, but your data is cooked.

IF you defined race by two and only two sets of characteristics (skin color and skull morphology), and the graphs for the whole human species gave you a large number of individuals with a staticstically significant variation in density at different loci on the graph, you could conclude that there were statistically significant numbers of people with similar skin color who also had similar skull morphology.

The questions still get down to:

(a) what particular use is this concept?

(b) what is so revelatory about the notion that if you take a bunch of people who more or less look alike, and they breed in isolation, most of their offspring will kinda more or less look alike too?

(c) why is skin color and skull morphology the two things you pick? Why not skin color vs. schlong size, or hair color vs. whether earlobes have that little crease in them, or thin to full lips vs. long thin fingers and toes to short stubby fingers and toes? In other words, what is the utility (other than it's easier to look at two dimensional graph) of picking just two features, and why do you pick the two you do?

[ranskaldan]

Originally posted by MichaeltheGreat

Nice idea, but your data is cooked.

IF you defined race by two and only two sets of characteristics (skin color and skull morphology), and the graphs for the whole human species gave you a large number of individuals with a staticstically significant variation in density at different loci on the graph, you could conclude that there were statistically significant numbers of people with similar skin color who also had similar skull morphology.

Methinks that your first and second paragraphs totally contradict each other.

The questions still get down to:

(a) what particular use is this concept?

What use is any classification other than convenience? Many things vary along a spectrum without clusters - liberals and conservatives, developed and developing nations, the color spectrum - and yet we have names for certain ranges for no other reasons other than convenience. Now, here's something that DOES show clusters. If we don't name these clusters, that doesn't make these clusters go away. So why not recognize them?

(b) what is so revelatory about the notion that if you take a bunch of people who more or less look alike, and they breed in isolation, most of their offspring will kinda more or less look alike too?

So you're not disputing that these clusters exist?

Btw, there's nothing revelatory at all. Such clusters exist, we call them "race", and apparently you agree.

We could call it something else other than "race", if you prefer. But changing its name hardly changes its existence.

(c) why is skin color and skull morphology the two things you pick? Why not skin color vs. schlong size, or hair color vs. whether earlobes have that little crease in them, or thin to full lips vs. long thin fingers and toes to short stubby fingers and toes? In other words, what is the utility (other than it's easier to look at two dimensional graph) of picking just two features, and why do you pick the two you do?

Because it's easier to look at a two dimensional graph.

If you could show me how to show a fifteen-dimensional graph on a 2D screen, I would of course be very grateful.

As for the two features - I could of course pick two more obscure features, but that wouldn't change my point at all.

[MichaeltheGreat]

Originally posted by ranskaldan


Methinks that your first and second paragraphs totally contradict each other.

Not at all - unless I'm blind, it seems there's a few billion missing dots on your graph.

What use is any classification other than convenience?

For it to have any scientific use, or any real utility at all, it should convey some specific piece of useful information, in a brief way, and in a way precise enough that reasonably knowledgeable people know what is being described. It's a lot easier to refer to an intermediate vector boson or gram-negative bacilli than to provide a two paragraph description without a name. It also meets the precision and usefulness tests, because these are sets of things with clearly defined members that function in clearly defined ways.

Many things vary along a spectrum without clusters - liberals and conservatives, developed and developing nations, the color spectrum - and yet we have names for certain ranges for no other reasons other than convenience.

It depends on how you define convenience. If you want to dismiss someone with an ad hominem yes - "he's just a whiny liberal" Unfortunately, people like me screw up the usefulness of arbitrary systems like that - I have people here refer to me as a right winger and a liberal on about an equal basis. So the label is meaningless, except to give an idea of how that person views my politics (or a part of them) in relation to his. If you want to say "race" is a rhetorically useful, arbitrary classification, nobody's arguing that. What's being argued is that it is either an objective criteria, or that it has any scientific use or validity when applied to humans.

Now, here's something that DOES show clusters. If we don't name these clusters, that doesn't make these clusters go away. So why not recognize them?

How many useless clusters do you want? How about a correlation of schlong size to premature male pattern baldness? We can make millions of clusters, if we want.

So you're not disputing that these clusters exist?

I'm not disputing that if you take pairs of arbitrary gross physiological characteristics and plot the range of observed characteristics for the whole human species, you'll likely get some variation in density. Random walk theory in mathematics will show that if you flip a coin a million times, you'll get statistical spikes and dips above and below the 50% probability line, and that random variation can be modeled. Flip a dollar coin in one axis, and a Euro coin in another axis, and you can get clusters too. In the human case, there's a loose causality effect, in that Arkansas hillbillies breed with Fijian islanders a bit less often than they breed with other Arkansas hillbillies. So you're going through this huge exercise and plotting graphs and making up classifications to tell me that?

Powerline radiation junkies occasionally manage to find a bunch of people who get cancer who live fairly close to an electric substation. Trouble is, they don't get them often enough to show that they're statistically significant.

Btw, there's nothing revelatory at all. Such clusters exist, we call them "race", and apparently you agree.

Sure. There's a statistical correlation of being a Masai tribesman and having a 20 inch schlong, too. Doesn't do much for "whitey", or Zulus, Ibo, or Tswanas, etc. But then there's no evidence "Long Dong" Silver has Masai ancestry, so too bad, there's no apparent hard boundary to that cluster. In fact, if you arbitrarily pick any two features you want to pick, and do your little graph for the entire human species, and cross reference them all, you can probably come up with millions of two-factor clusters.

According to a lot of people, you can find clusters if you correlate degree of hair blondeness and IQ. Hey! We now have an official "dumb blonde" race! Wheeeee! That's useful.

What we call "race" is an antiquated notion that you can classify the entire human species into 3 or 5 or 6 or some small number of discrete groups (excluding people of "mixed" background) and have any particular scientific meaning to such a classification system.

We could call it something else other than "race", if you prefer. But changing its name hardly changes its existence.

I prefer to call it "arbitrary selection of arbitrary factors to look for meaningless statistical clusters in isolation from their significance in the entire species population"

Because it's easier to look at a two dimensional graph.

If you could show me how to show a fifteen-dimensional graph on a 2D screen, I would of course be very grateful.

You can still do a fifteen-factor correlation study, you'd just have to give up the graphs for tables.

As for the two features - I could of course pick two more obscure features, but that wouldn't change my point at all.

Nor would it change mine. In a comprehensive population sample, multiple events of small scale soft clustering of the occurence of two arbitrarily chosen gross physiological features has neither scientific use nor validity for any analytical purpose.

[ranskaldan]

Originally posted by MichaeltheGreat


Not at all - unless I'm blind, it seems there's a few billion missing dots on your graph.

Do you really want those to be put in as well?

For it to have any scientific use, or any real utility at all, it should convey some specific piece of useful information, in a brief way, and in a way precise enough that reasonably knowledgeable people know what is being described. It's a lot easier to refer to an intermediate vector boson or gram-negative bacilli than to provide a two paragraph description without a name. It also meets the precision and usefulness tests, because these are sets of things with clearly defined members that function in clearly defined ways.

Terms like "Caucasoid", "Mongoloid", or even "Khoisan", "Ainu", etc perform a very similar role. They give a general collection of traits that makes it easier to talk about human genetics and evolution.

As you can see from the graph - the circles are fuzzy, without clearly defined boundaries. This is because there are also Ainu who don't look Ainu and Chinese who look Japanese. But notice that there's a concept of an "Ainu who doesn't look Ainu", meaning that there is a set of traits that Ainus tend to cluster around. That's the point of having a term like "Ainu" - to refer to these traits.

It depends on how you define convenience. If you want to dismiss someone with an ad hominem yes - "he's just a whiny liberal" Unfortunately, people like me screw up the usefulness of arbitrary systems like that - I have people here refer to me as a right winger and a liberal on about an equal basis. So the label is meaningless, except to give an idea of how that person views my politics (or a part of them) in relation to his. If you want to say "race" is a rhetorically useful, arbitrary classification, nobody's arguing that. What's being argued is that it is either an objective criteria, or that it has any scientific use or validity when applied to humans.

Liberals and conservatives can't be measured using centimeters and grams. Physical traits that define various ethnic groups, however, can. These are the objective criteria you're looking for.

As for "scientific use" - again, it is purely for convenience, so that we can say "we found an Ainu-like tribe in North America", instead of "we found a group in America whose various phenotypes cluster closely with those of the arbitrarily defined group known as the Ainu in Asia".

How many useless clusters do you want? How about a correlation of schlong size to premature male pattern baldness? We can make millions of clusters, if we want.

Sure. As long as such schlong-malebald clusters have relevance to the study of human evolution, by all means classify them.

I'm not disputing that if you take pairs of arbitrary gross physiological characteristics and plot the range of observed characteristics for the whole human species, you'll likely get some variation in density. Random walk theory in mathematics will show that if you flip a coin a million times, you'll get statistical spikes and dips above and below the 50% probability line, and that random variation can be modeled. Flip a dollar coin in one axis, and a Euro coin in another axis, and you can get clusters too. In the human case, there's a loose causality effect, in that Arkansas hillbillies breed with Fijian islanders a bit less often than they breed with other Arkansas hillbillies. So you're going through this huge exercise and plotting graphs and making up classifications to tell me that?

Yes. People here don't seem to realize that a continuous spectrum does not rule out clustered distributions.

Sure. There's a statistical correlation of being a Masai tribesman and having a 20 inch schlong, too. Doesn't do much for "whitey", or Zulus, Ibo, or Tswanas, etc. But then there's no evidence "Long Dong" Silver has Masai ancestry, so too bad, there's no apparent hard boundary to that cluster. In fact, if you arbitrarily pick any two features you want to pick, and do your little graph for the entire human species, and cross reference them all, you can probably come up with millions of two-factor clusters.

Nah. Things like the Sahara Desert, Atlantic Ocean, and Himalayas ensure that these miniclusters form up into bigger superclusters. That's also quite convenient for the study of human evolution and human migrations.

What we call "race" is an antiquated notion that you can classify the entire human species into 3 or 5 or 6 or some small number of discrete groups (excluding people of "mixed" background) and have any particular scientific meaning to such a classification system.

I prefer to call it "arbitrary selection of arbitrary factors to look for meaningless statistical clusters in isolation from their significance in the entire species population"

Why call a sandwich a "double-layered protein-centered carb-fringed gustatory device"?

And why would such clusters be meaningless? They offer useful pointers in studying human evolution and migrations.

[MichaeltheGreat]

Originally posted by ranskaldan


Do you really want those to be put in as well?

Yes, every last one of them. Plus a description of your measurement protocols and where you took the skin samples from. I'm a lot darker in the summer than I am in the winter, and if you've ever tried to measure a curved surface with precision, you know it's awful easy to fuzzify the numbers just a little each time, even though you're trying not to. The more missing data and fuzzified data, the less you can read into what you have.

When you acount for your error circles and the entire population, then you get a true idea of the number of clusters (if there are any true clusters), whether they correlate to physically real data (i.e. if you get light darkies and dark whiteys, they start messing with the data), how hard the boundaries of the clusters are, and if there's a large enough portion of the population that can be described in terms of clusters to actually be useful.

Terms like "Caucasoid", "Mongoloid", or even "Khoisan", "Ainu", etc perform a very similar role. They give a general collection of traits that makes it easier to talk about human genetics and evolution.

Or they make it harder. It depends on how precise the definitiion of group membership is. To give a concrete example, let's talk about sickle cell anemia. Until fairly recently, with the advance of genetic studies and epedemiology, sickle cell was incorrectly ascribed as being a genetic disease affecting blacks "negroid" or whatever you want to call them. You know, THEM. A lot of early epidemiology studies started barking up the wrong tree, for instance when US "hispanic" (you know, the other THEM) people got the disease, it was assumed that somewhere, there had been midnight rendezvous with darkies on both sides of the patient's ancestry, since we all knew it was a hereditary disease affecting blacks.

Trouble is, modern genetics and epidemiology and the spread of modern infectious disease medicine throughout the world showed this wasn't the case.

What we've learned is:

(a) the carrier gene for sickle cell is a commonly occuring mutation which can occur in almost any population. It is thus theoretically possible for two card-carrying white supremacists of "pure" ancestry to both be carriers and have a sickle cell kid. It probably has happened a few times, it's just not very probable.

(b) the reason the carrier gene is more common in some people is that it confers a significant survival advantage against malaria. In the days before infectious diseases and disease vectors were understood, this survival advantage was significant enough in areas where malaria was endemic that the carrier gene was favorably selected for over generations - the increased survival rate against malaria exceeded the mortality rate in those 25% of offspring of two-carrier parents who actually got the disease and died.

(c) Since the sickle cell mutation occurs in most populations, and conferred a survival advantage for many generations in the "malaria belt," most peoples native to those areas at the time developed a significant number of sickle cell carriers in their population. Thus, a "black" disease is common in indigenous Central and South American natives with no exposure to "black" genes, among mediterranean basin peoples, east Africans, some central Africans, and people from coastal lowland areas in the Indian subcontinent.

Yet it's virtually unheard of in other areas of Africa and in populations from those areas.

So "people with the gene for sickle cell" is a useful classification, while "black" or whatever you want to call it is so vague (and in this case, so disconnected from what people thought was true as to be misleading), that it's worse than useless.

As you can see from the graph - the circles are fuzzy, without clearly defined boundaries.

It's incomplete data without a description of the measurement protocols.
[/quote]

This is because there are also Ainu who don't look Ainu and Chinese who look Japanese. But notice that there's a concept of an "Ainu who doesn't look Ainu", meaning that there is a set of traits that Ainus tend to cluster around. That's the point of having a term like "Ainu" - to refer to these traits.

That's your working hypothesis, but unless you have all the data and map out all the error circles, you don't know if it's correct, useless or somewhere in between. If a majority of people don't fit in clusters, or if there's so many clusters you can barely count them, or if the "cluster" boundaries are so soft against the background that it's arbitrary where you draw them, then it's pretty far over on the "useless" side of the scale.

Of course you can have a concept of "Ainu that doesn't look Ainu" - the operative question is whether it's a social preconception, like "blacks who don't act black" or whether it's something that exists (i.e. there's a hard cluster around "Ainu" such that there's a clearly defined "look" that can be measured in a couple of variables.) Ainu is also a much smaller, much more narrowly constrained grouping than say, "caucasoid"

Liberals and conservatives can't be measured using centimeters and grams. Physical traits that define various ethnic groups, however, can. These are the objective criteria you're looking for.

They're only "objective" criteria if they can be measured precisely, without any possibility of statistically significant fuzzifying data, AND if they correspond to something that can be statistically isolated. Height, for example, has both a hereditary and environmental factor - chronic malnutrition tends to affect it a bit. Dark skin color also occurs in genetically relatively isolated groups (i.e. some South American indigenous peoples and some Indo-Pacific islanders are significantly darker than some Africans. If there's also a loose correlation to big heads, you can then lump genetically unrelated (until you get back to proto-humans) people into the same categories by false correlation.

As for "scientific use" - again, it is purely for convenience, so that we can say "we found an Ainu-like tribe in North America", instead of "we found a group in America whose various phenotypes cluster closely with those of the arbitrarily defined group known as the Ainu in Asia".

"Convenient" also has to coincide with accurate and concise. When you get into a debate on how Ainu-like is Ainu-like, that's one problem. Also labels which draw conclusions ahead of the evidence are dangerous - "Ainu-like" might lead you to speculate on Ainu migrations, then you find out they're really more like Norsemen crossed with natives, after looking in the wrong direction for twenty years.

Sure. As long as such schlong-malebald clusters have relevance to the study of human evolution, by all means classify them.

The problem is, you can probably find millions of pseudoclusters that may or may not be useful. So anything which tends to skew or filter the way you look at the data should be suspect unless you can show that it's clearly more useful than other ways of looking at the data.

Yes. People here don't seem to realize that a continuous spectrum does not rule out clustered distributions.

The correlary is that not all clustered distributions are statistically significant.

Nah. Things like the Sahara Desert, Atlantic Ocean, and Himalayas ensure that these miniclusters form up into bigger superclusters. That's also quite convenient for the study of human evolution and human migrations.

Or it just confuses the hell out of things, when you play with statistics to reach spurious conclusions based on imprecisely measured and arbitrarily selected morphology, rather than looking for more useful things like mtDNA and tracking evidence for the spread of rare mutations.

Why call a sandwich a "double-layered protein-centered carb-fringed gustatory device"?

Why lump a McDonald's hamburger, a Conch-steak sandwich, US military tradition "**** on a shingle," three-decker peanut butter and jelly into the category "sandwich" then decide that crepes and eggs Benedict don't belong in that category because most people eat them at a different time of day?

And why would such clusters be meaningless? They offer useful pointers in studying human evolution and migrations.

Maybe, or maybe they offer misleading info. It depends entirely on how statistically sound and hard the cluster is. If you started out with four foot tall totally white people and eight foot tall totally black people, and had a gradual appearance of six foot tall grey people, you could reach some definite conclusions from that. Unfortunately, anything as imprecise as a couple of morphological features is a hell of a lot fuzzier, and doesn't lead you to specific or accurate conclusions that withstand scientific scrutiny.

Tracking distribution or appearance through time of mtDNA would tell you a lot though.

[TCO]

Originally posted by MichaeltheGreat


Old fashioned controlled breeding and bloodline management is more cost effective, but if there's ever a dog genome project, you'll be able to genetically identify (probable) members of specific official breeds much more readily than you would ever be able to identify group differences in humans.

Well...the human group differences are genetic in some sense, no? otherwise they would be entierly environment determined rather than heritable. Even if it's not a simple story.

[TCO]

Originally posted by MichaeltheGreat


Nice idea, but your data is cooked.

IF you defined race by two and only two sets of characteristics (skin color and skull morphology), and the graphs for the whole human species gave you a large number of individuals with a staticstically significant variation in density at different loci on the graph, you could conclude that there were statistically significant numbers of people with similar skin color who also had similar skull morphology.

The questions still get down to:

(a) what particular use is this concept?

(b) what is so revelatory about the notion that if you take a bunch of people who more or less look alike, and they breed in isolation, most of their offspring will kinda more or less look alike too?

(c) why is skin color and skull morphology the two things you pick? Why not skin color vs. schlong size, or hair color vs. whether earlobes have that little crease in them, or thin to full lips vs. long thin fingers and toes to short stubby fingers and toes? In other words, what is the utility (other than it's easier to look at two dimensional graph) of picking just two features, and why do you pick the two you do?

1. Your same criticisms here could be made of dog breeds and in fact dog breeds are a result of some cluster method like what Ran decided on. I think Ran's point is important as a conceptual one. We can argue about how to segement the market, but some scheme like Ran's makes sense. And really...you can try it on dog breeds first.

2. I will readily agree that controlled breeding may produce quicker results than natural selection over time in different geographies or that the differences in dog breeds are probably wider than those in humans. But the basic point of the mutability of our characteristics remains. Humans are animals also after all.