Is race a valid concept?

[Arete]

Great. Can you show how race according my definition fails this?

 

First, if you're talking about genetically related groups, then generally the correct terminology would be populations, rather than races. In population genetic context, the term "races" is generally used to describe different karyotypes within a species.

 

Once you've established this, you can then use a qualitative measure of genetic distance to define populations. Rosenberg et al. Already did this.

 

This would then represent a generally widely accepted pattern of human genetic variation, which I doubt many members here would disagree with.

Edited May 31, 2016 by Arete

[Mikemikev]

First, if you're talking about genetically related groups, then generally the correct terminology would be populations, rather than races. In population genetic context, the term "races" is generally used to describe different karyotypes within a species.

 

That's just semantics. I'll call them races (populations define by descent or similarity), you use the PC superset euphemism "populations" (any grouping), whatever. Can you answer my question?

[Saxon]

There are not 18 ethnic groups in Tang's samples.

 

They only have: White American (non-Hispanic white), African-American (non-Hispanic black), Hispanic, and Taiwanese.

 

However, Hispanic is not an ethnic group. Although they point out those that self-identified as Hispanic were Mexican.

 

How exactly does this prove race? There is no continental clustering of "major races".

[Mikemikev]

 

So why did you say that when I asked how you defined race? I didn't ask what the cause was.

 

That's a complete lie. You asked what the 5 races were. This is the result of applying a definition . I told you there is no true level of analysis and no fixed number.

 

Tang 2005" is not very precise at all. I tried googling but just came up with lots of seemingly irrelevant results. I did ask you to clarify but apparently you couldn't be bothered to be precise.

A precise reference would be, for example:

Tang, N. and Cousins, C. (2005), Working Time, Gender and Family: An East-West European Comparison. Gender, Work & Organization, 12: 527–550. doi: 10.1111/j.1468-0432.2005.00287.x

 

 

I guess I excpectd people to have the initiative to google "tang 2005 race". Maybe they did and are just wasting my time. Please don't be too surprised why I no longer respond to you.

There are not 18 ethnic groups in Tang's samples.

 

They only have: White American (non-Hispanic white), African-American (non-Hispanic black), Hispanic, and Taiwanese.

 

However, Hispanic is not an ethnic group. Although they point out those that self-identified as Hispanic were Mexican.

 

How exactly does this prove race? There is no continental clustering of "major races".

 

Nonsense

Edited May 31, 2016 by Mikemikev

[Arete]

That's just semantics. I'll call them races (populations define by descent or similarity), you use the PC superset euphemism "populations" (any grouping), whatever. Can you answer my question?

 

it has nothing to do with "political" correctness, rather than simple correctness. Population has a specific meaning in the context of biology, and is the scientifically correct term. Redefining terms as you have done above leads to confusion, as seen throughout the thread.

 

The other issue with your "definition" or race, is that its has no a priori quantitative thresholds. You can arm wave and say that you define them by "ancestry" but without some methodology that's somewhat meaningless. Must populations in your model be panmictic? Have an F_ST > 0.2? an N_m < 25? Without some form of objective criteria for population divisions, the boundaries become subjective, arbitrary, and thus devoid of use in a quantifiable, scientific sense.

Edited May 31, 2016 by Arete

[Strange]

That's just semantics.

 

Of course it is semantics, we are discussing the definition of terms. It is not "just" semantics, it is totally semantics and nothing but semantics. (The phrase "just semantics" makes me wonder is people know what the word means.)

[Arete]

Nonsense

 

" Subjects identified themselves as belonging to one of four major racial/ethnic groups (white, African American, East Asian, and Hispanic)."

http://www.sciencedirect.com/science/article/pii/S0002929707625786

 

That'd be four, not 18. They simply used the SIRE model to justify setting K=18 in the STRUCTURE analysis.

Edited May 31, 2016 by Arete

[ajb]

Can you explain why it is not 100% effective?

Well, for one, how far back do you go in this definition?

 

 

Can subspecies or genus be 100% effective?

As you know, looking at subspecies is not easy. So what? This just adds to the care that needs to be taken with the definitions.

 

Can anything?

I think so, but the question is too general.

 

What does it mean to be effective in this context?

This is a better question...

 

In biological classifications, or any classification really, you are looking for 'equivalence classes'. That is you want some relation between all the objects you want to classify. You want to say that if a~b then they are in the same class. You want that if a~b and b~c then a~c. And you want the converse. You then want to use this to define your equivalence classes, ie. all the sets of objects that are identified. You also need that these classes are mutually exclusive. That is any object cannot belong to more than one to the equivalence classes. Moreover, you need every object to belong to an equivalence class.

 

All attempts to do this with race seem to fail.

 

One can do this loosely, as we do; look at 'equal opportunity monitoring forms' where you identify your ethnicity -- but this is not carefully biologically defined.

 

So all concepts in science must always give predictions with absolute certainty with perfect mapping between definition and measurement? Are you just setting up absurd standards that would have you laughed out of the room for any other subject?

 

Where did I say that?

 

I have discussed theoretical physics and how this 'matches' experiments in other threads. I am, I assume, more aware than most of the issues here.

 

Classifying by descent isn't biology now?

Did anyone actually say that?

 

The point is it seems not enough to define race.

[Strange]

 

That's a complete lie. You asked what the 5 races were. This is the result of applying a definition .

 

I asked because I was trying to understand your definition.

 

 

I told you there is no true level of analysis and no fixed number.

 

And yet you insist it is not vague and subjective.

I guess I excpectd people to have the initiative to google "tang 2005 race". Maybe they did and are just wasting my time. Please don't be too surprised why I no longer respond to you.

 

As I say, I did try that Google. And found a large number of results none of which appeared to be relevant. You seem unwilling or unable to provide a specific reference or a link. Why is that?

[Mikemikev]

 

it has nothing to do with "political" correctness, rather than simple correctness. Population has a specific meaning in the context of biology, and is the scientifically correct term. Redefining terms as you have done above leads to confusion, as seen throughout the thread.

 

The other issue with your "definition" or race, is that its has no a priori quantitative thresholds. You can arm wave and say that you define them by "ancestry" but without some methodology that's somewhat meaningless. Must populations in your model be panmictic? Have an F_ST > 0.2? an N_m < 25? Without some form of objective criteria for population divisions, the boundaries become subjective, arbitrageurs, and thus devoid of use in a quantifiable, scientific sense.

 

Why would it it have arbitrary quantitative thresholds? Would that make it more informative or predictive? The sine qua non of a scientific concept. Let's take your Fst 0.2 for example. Why would you arbitrarily make that a requirement? You know several species would fail that?

 

Taxa defined by descent are meaningless? You really want to claim this?

This is a better question...

 

In biological classifications, or any classification really, you are looking for 'equivalence classes'. That is you want some relation between all the objects you want to classify. You want to say that if a~b then they are in the same class. You want that if a~b and b~c then a~c. And you want the converse. You then want to use this to define your equivalence classes, ie. all the sets of objects that are identified. You also need that these classes are mutually exclusive. That is any object cannot belong to more than one to the equivalence classes. Moreover, you need every object to belong to an equivalence class.

 

All attempts to do this with race seem to fail.

 

 

Not at all. All individuals in one race must share more ancestry versus all individuals in another race. Its really quite simple.

 

And yet you insist it is not vague and subjective.

 

What's the true number of taxa for all living things? Clearly it varies by level of analysis. The subjectivity is in your question.

Edited May 31, 2016 by Mikemikev

[ajb]

All individuals in one race must share more ancestry versus all individuals in another race. Its really quite simple.

You still need to find some clear boundaries in order to put people into one bin or another. I doubt it is at all as clear as you suggest.

[Arete]

Why would it it have arbitrary quantitative thresholds?

 

Because genetic differentiation is a continuum. Any classification of populations is arbitrary, in a sense. You use a priori delimitation for the same reason you set significance thresholds a priori. What's your proposed alternative, qualitative delimitation? Random selection? Whatever you feel like on a particular day?

 

Would that make it more informative or predictive?

 

Yes. An example of such a threshold in practice would be the standard use of 97% similarity at the 16S locus to define a bacterial species.

 

 

Let's take your Fst 0.2 for example. Why would you arbitrarily make that a requirement? You know several species would fail that?

 

That was the point. What's your cut off of "descent" going to be?

 

Taxa defined by descent are meaningless? You really want to claim this?

 

Nice strawman.

 

You can arm wave and say that you define them by "ancestry" but without some methodology that's somewhat meaningless.

Edited May 31, 2016 by Arete

[Mikemikev]

You still need to find some clear boundaries in order to put people into one bin or another. I doubt it is at all as clear as you suggest.

 

The boundaries are inherent in the definition. It seems like you've just caught the "received wisdom" that race is a wooly concept, without actually thinking it through. By definition individuals will share more or less ancestry and can be classified into discrete groups.

[CharonY]

You even say "more or less" which indicates a continuum. Where in this continuum do you put the boundary? How and when does a continuum become discrete and how is it inherent?

[Strange]

The boundaries are inherent in the definition. It seems like you've just caught the "received wisdom" that race is a wooly concept, without actually thinking it through. By definition individuals will share more or less ancestry and can be classified into discrete groups.

But, as you say, those groups are of variable number. So the choice of where to draw the boundaries is subjective and, ultimately, arbitrary.

[Mikemikev]

 

Because genetic differentiation is a continuum. Any classification of populations is arbitrary, in a sense. What's your proposed alternative, qualitative delimitation? Random selection? Whatever you feel like on a particular day?

 

Yes. An example of such a threshold in practice would be the standard use of 97% similarity at the 16S locus to define a bacterial species.

 

That was the point. What's your cut off of "descent" going to be?

 

Nice strawman.

 

 

Human variation is not a continuum. If it was I would agree division would be arbitrary (but still valid as a operationalization of a spectrum).

 

Is this a continuum?

 

http://www.scs.illinois.edu/~mcdonald/PCA84pops.html

 

Your 97% rule for species is irrelevant. That's based on creating a manageable number of names. Race is a scalable infra subspecific concept to describe variation. Do you have a better one?

Edited May 31, 2016 by Mikemikev

[ajb]

The boundaries are inherent in the definition.

As Arete states, you need to develop some methodology.

 

And as an example, where would you place Barack Hussein Obama II ?

 

Wikipedia tells us he is 'African American'; but what does this mean? His mother is 'mostly English ancestry' (mostly? how far back?) and his farther was from Kenya. The 'African American' bin fits if we use skin color and curly hair... but what about his English side?

[Strange]

Is this a continuum?

 

http://www.scs.illinois.edu/~mcdonald/PCA84pops.html

Interesting graphic. Can you provide some context?

Edited May 31, 2016 by Strange

[Mikemikev]

Interesting graphic. Can you provide some context?

 

Its from Doug MacDonald.

[Saxon]

Human variation is a genetic continuum. Note the plot Mikemikev uploads misses many global populations - so what it is doing is inserting false divides into the continuum. Mikemikev then turns around and says these gaps are the racial divides. It's fallacious.

 

I have a study from 2009 which samples over 100 populations from Africa alone. The plot Mikemikev posts has only about 10 African populations. As soon as you increase the population samples, all the gaps disappear.

Edited May 31, 2016 by Saxon

[Mikemikev]

Human variation is a genetic continuum. Note the plot Mikemikev uploads misses many global populations - so what it is doing is inserting false divides into the continuum. Mikemikev then turns around and says these gaps are the racial divides. It's fallacious.

 

I have a study from 2009 which samples over 100 populations from Africa alone. The plot Mikemikev posts has only about 10 African populations. As soon as you increase the population samples, all the gaps disappear.

Nonsense.

[Strange]

 

Its from Doug MacDonald.

 

 

Fascinating. Can you provide some context? Like a link to the web page or article it comes from?

[Saxon]

Yes, its real nonsense your plot only samples 10 African populations where there are hundreds. See Tishkoff et al 2009 who sampled 121 African populations.

 

Why are you quoting a study with only 10 African populations? Because when you use few population samples you create false genetic divides (that are filled when more populations are sampled).

 

It's obvious for everyone to see you have confirmation bias.

[Mikemikev]

Yes, its real nonsense your plot only samples 10 African populations where there are hundreds. See Tishkoff et al 2009 who sampled 121 African populations.

 

Why are you quoting a study with only 10 African populations? Because when you use few population samples you create false genetic divides (that are filled when more populations are sampled).

 

It's obvious for everyone to see you have confirmation bias.

Oh really. The South African Bantu and the Kenyan Bantu cluster very closely. What are these other populations that "fill in the gaps" (according to you)?

[Saxon]

There is no genetic divide between Africa and Eurasia. See Tishkoff et al 2009 who used 121 African population samples and 60 Eurasian population samples. What you see is a smooth genetic gradient between these two continents. Therefore racial classification is arbitrary.