Path: utzoo!utgpu!water!watmath!clyde!cbosgd!ihnp4!upba!eecae!nancy!msudoc!straney
From: straney@msudoc.ee.mich-state.edu (Ronald W. DeBry)
Newsgroups: sci.bio
Subject: Re: Rape a reproductive advantage?
Message-ID: <256@nancy.UUCP>
Date: 22 Jan 88 01:13:36 GMT
References: <517@gtx.com> <248@nancy.UUCP> <2043@navajo.UUCP> <4424@ecsvax.UUCP>
Sender: usenet@nancy.UUCP
Reply-To: straney@msudoc.UUCP (Ronald W. DeBry)
Distribution: na
Organization: Michigan State Univ., Engineering, E. Lansing
Lines: 129

In article <2043@navajo.UUCP>, baxter@navajo.UUCP (Ray Baxter) writes:
>... 
> I would like to see one shred of data which shows that human rape is
> not heritable.  It is fine for you to say that it is a social problem,
> but you certainly do not know that is not a genetic problem.


You've caught me indulging one of my favorite biases.  I admit that I
know of no experimental evidence that directly addresses the
heritability of human (or even duck) rape.

Having admitted that, I will now try to take some advantage of the
situation by climbing on my soap-box and explaining why I like that
bias so much.

The bias is: When it comes to natural selection, I prefer to assume
that something is not, until it is shown that it is.  When people
start a discussion of a situation that is observed in nature by
listing reasons why the trait is advantageous, or disadvantageous,
they are making the opposite implicit assumption - that natural
selection IS responsible for the trait's frequency in nature.  I have
problems with this assumption at two levels.

The first is that it often (but not always) means that the person
making the assumption has further assumed that fitness values can (at
least at a gross level) be inferred from gene frequencies.  In other
words, they assume that since a "good" trait should become common,
then common traits must be good, and must have become common through
the action of natural selection.  This is what I still think prompted
the original question, which was why doesn't rape become more common,
since it appears to be a way for a male to pass genes on to the next
generation.  To counter by listing reasons why rape could be a
disadvantage is to continue to make the assumption that the frequency
of rape is governed primarily by natural selection, and to continue to
make the second implicit assumption, which is that it has a genetic
basis and is a genetic problem.

The general subject of whether a common trait must be adaptive is
complex, but a reference to Lewontin and Gould's classic "Spandrels of
San Marcos" paper is a reasonable start.  I should also point out that
many prominent evolutionary biologists believe that it is acceptable
to assume that a trait is adaptive *as a starting point*, and that
Lewontin and Gould's characterization of "the adaptationist programme"
is overly harsh.

Nonetheless, the second implicit assumption remains - that the trait
is heritable.

In article <4424@ecsvax.UUCP> hes@ecsvax.UUCP (Henry Schaffer) writes:

>  It isn't all that easy to demonstrate the degree of heritability of a
>quantitative behavioral trait.  Cf. the problems in trying to study the
>heritability of IQ in humans.  The combination of long generation time,
>small family size, inability to specify a breeding scheme, and the
>difficulty of measuring the trait being studied, all make it very difficult
>(at best).  To demonstrate heritability all you have to do is to get an
>estimate that differes significantly from 0.  However, to show that the
>behavior is "not heritable" (which I interpret as meaning that the
>heritability is 0) you have to measure the heritability accurately enough
>to show that the value is 0, and not merely some small value, say 5%.  Even
>in an experimental animal or plant population it is difficult (takes much
>recsources, effort, ...) it estimate a heritability with a std. error of
>much less than 5%.
>--henry schaffer  n c state univ


These comments bring up another point.  Which way are we going to make
the test?  What is the null hypothesis?  If we start with the
assumption that natural selection IS responsible, then we must show
that the heritability is zero in order to reject our starting
assumption.  You have correctly pointed out that such a demonstration
is damn near impossible, which suggests to me that the starting
assumption is a bad working hypothesis (disclaimer: yes, I do have
certain Popperian leanings, but only to the extent that I think a
proper scientific hypothesis should be falsifyable :-})

On the other hand, if we start by assuming that the trait is NOT
heritable, then that hypothesis is much easier to falsify.  If such
data were forthcoming, then I would be convinced that rape is, at
least in part, a genetic problem.


Having (In my own mind, anyway :-}) justified my intial assumption and
bias, I think that a couple of other glimpses into the complexity of
measuring the heritability of rape are worthwhile.


[If you aren't interested in the technicalities, the following will be
even more boring than the first part.  If, like Henry, you already
know more quantitative genetics than I ever will, please bear with me
for a paragraph, and correct me if neeeded]

First: It is easy to confuse "genetic" with "heritable", because
heritabilty comes in two forms - "broad sense" and "narrow sense".  "A
character can be 'hereditary' in the sense of being determined by the
genotype or in the sense of being transmitted from parents to
offspring, and the extent to which it is hereditary in the two senses
may not be the same" (Falconer, 1981 - Introduction to Quantitative
Genetics).  The narrow sense heritability is the important component
in determining the potential effects of selection.  In its simplist
form, this is the degree to which the phenotypic variance is
determined by the genetic variance, as opposed to the environmental
variance ("environmental" here subsumes every thing else, it doesn't
just mean eating Twinkies).

It is perfectly possible for a trait to be inherited from one's
parents, and yet have the trait not be heritable in the narrow sense.
So, merely showing a gentic basis does not mean that selection has
anything to work on.  For example, it certainly is true that the
ability to perform a sexual act is a component of the trait
"propensity to rape".  This ability is also certainly genetic.  The
difference is that essentially all males (to keep to this form of
rape) in the population possess the component "ability to rape".  If
this is the only genetic component, then the genotypic variance is
essentially zero, and so is the heritability.  In other words, the
*variation* in "propensity to rape" is not determined by genetic
variation, but only by environmental variation.


Another complexity is that the trait is *propensity* to rape, so it is
something which cannot be measure for an individual.  It can only be
measured in a group of genetically identical individuals.  So, even
though I feel that it is "more scientific" to assume that rape is not
genetic until it is shown otherwise, that proof is even more difficult
to obtain than even Henry suggested.  Even in a standard experimental
organism, such as _Drosophila_, such an experiment would be extremely
difficult to properly design.

Ron DeBry  Dept. of Zoology