Monday, April 9, 2012

Not a free speech issue

If a private publication sacks a writer because they do not wish to be associated with his declared opinions, this is not a free speech issue. It is a branding issue, it is an employment issue, but it is not a free speech issue.

This is particularly true of a publication that is explicitly ideological in its role. (Using ‘ideological’ in its most general sense of having an explicit normative view of the world.)

The publication in question is the National Review, the premier magazine of American conservatism founded by the doyen of postwar conservative movement writers and activists William F. Buckley Jnr.

The author who was sacked was expatriate British journalist and writer John Derbyshire. He was sacked for this article in Taki’s Magazine.

If Derbyshire were to be prosecuted for publishing the article, that would indeed be a free speech issue. But National Review deciding that it does not wish to be associated with particular opinions, that is a matter of branding, of ideological identity, not free speech (as is, for example, being claimed here).

Colour awkward
The politics of race are endlessly fraught in the US and deciding that American conservatives are ipso facto racist is a common conceit among American liberals (using 'liberal' in its peculiar American usage) and folk further left. (That sometimes reaches the level of deciding that conservative support for black figures, such as Condi Rice, Justice Thomas or Herman Cain, is itself a manifestation of racism.) Without getting into the history of American conservatism and race, there are certainly historical reasons why the National Review might have some sensitivity on such matters.

There are also generational shifts. Younger conservative activists are post-civil rights folk. They accept civil rights as a positive feature of American history and a usually very against revisiting opposition to civil rights: young conservtive bloggers actively campaigned against Senate Majority leader Trent Lott after his implicit endorsement of Strom Thurmond’s opposition to civil rights, which led to Sen. Lott’s resignation as Senate Majority Leader.

Either way, the National Review’s decision makes perfect sense. Particularly when leading the ideological assault on a serving black President.

Probability wrong
So, what’s wrong with Derbyshire’s article?

A pretty standard thing: a statistical tendency is not a defining characteristic. It is perfectly true that the homicide rate among black Americans is much higher than it is among other Americans. This is primarily a problem for other black Americans.

It is true that more American whites are murdered by blacks than blacks are murdered by whites. But if a minority group has a much higher homicide rate than a much larger group, it is to be expected that more members of the majority will be murdered by members of the minority than vice versa.

So, Derbyshire has a point then? No, Derbyshire is a statistical illiterate. The chances of any particular black person being a perpetrator of homicide is extremely low. A propensity to violence is not a defining, or even a likely, characteristic of any given black person; particularly a black person who is not young and male. But even a young male black is far more likely to be no risk than high risk. (Their high incarceration rate has everything to do with the insane war on drugs—the systematic, and grievously failed, attempt to deny people dominion over their own bodies—very little to do with any tendency to violence.) Indeed, young black males are far more likely to be the victim of homicide than a perpetrator any white person is.

Even if one accepts a genetic explanation for higher rates of hyper-aggressiveness in African and African-migrant populations, it is still not a likely characteristic of any given African or person of African descent. (Yes, I know we are all ultimately of African descent; I mean descended from folk who were not part of the prehistoric out-of-Africa diasporas.) It is Derbyshire citing averages, when the issue is likelihoods, that demonstrates his statistical malfeasance.

Focusing on someone’s blackness is not focusing on a risky characteristic. It is actually very bad advice to give a young person.

Which is another reason for National Review to sack Derbyshire.

[Cross-posted at Skepticlawyer and at Critical Thinking Applied.]


  1. You write:

    "Indeed, young black males are far more likely to be the victim of homicide than a perpetrator."

    That may, for all I know, be true. But the link you give for it only provides figures on victims, not perpetrators, so cannot by itself support the claim. Do you have something else that does?

    1. David, thank you well spotted. I was in error and have corrected the post accordingly.

  2. From the NCVS, I found that in 2008, the rate at which blacks committed interracial violent crime against whites was six times the rate at which whites committed interracial violent crime against blacks. This is certainly higher than that expected from higher rates of criminality among blacks alone.

    1. Your mathematics is off. Since blacks are seven times more likely to commit homicide than a white person, while a black person is six times more likely to be murdered than a white person, the comparative interracial rate is in fact almost precisely the rate one would expect.

    2. That may or may not be true--but I looked at interracial violent crime, which does not include homicide, in 2008; the results are as I reported them.

    3. I have no quarrel with your numbers, it is the inference you draw from it. The rate of interracial crime is a reflection of comparative "internal" crime rates, there is no evidence of more intense "targeting" of whites by blacks.