Re: Sociobiological Concepts

From: Stevan Harnad (harnad@coglit.soton.ac.uk)
Date: Thu Oct 02 1997 - 16:57:36 BST


> From: Jon Wright <jjw195@soton.ac.uk>
>
> What is evolution? Evolution is a process of progressive development
> such that organisms which prosper in a given environment, continue to
> flourish while the less well-suited organisms die out.

Evolution is not development in the sense that a child grows and
develops (nor is it in any sense "progress"). It is something that
happens across time, like development, but it is a blind causal process
like erosion. When you read Dawkins's "The Selfish Gene," you will see
that the genes are replicators. They are recipes or blueprints or
programmes for a trait or traits (blue eyes, sweet tooth). They differ
from real recipes, blueprints or programmes in that they are
"self-executing". What traits they code for (usually something to do
with making certain proteins under certain conditions), they also cause
to grow, during development, or in adulthood.

Organisms are not replicators: They are the VEHICLE for the replicators,
which are the genes.

The way evolution works is simple and perfectly mechanical and causal,
like the erosion of cliffs by the pounding sea: The replicators that
code for traits that make their vehicles survive long enough to
reproduce are the "successful" ones. Success just means that they are
able to replicate themselves, but this depends on the survival and
reproduction of their vehicle (us, for example).

Evolution is just the blind outcome of this process across time.
Organisms (vehicles) that reproduce by splitting into two (cloning) and
growing till they rich the splitting size again are successful when
this is enough. Suppose there is also some random variation in how big
they must be to be able to split. The big and small splitters would
both survive and reproduce equally well, so nothing much happens.

Then, let's say, the climate changes, it gets colder, which wipes out the
normal prey for a predator, say a sea newt. So the newt switches to
eating the creature that reproduces by splitting, but the newt is
short-sighted, so it can only see and catch the big ones. This means
that the big-splitters are at a disadvantage: as they are eaten, the
genes inside them that code for splitting size do not get to replicate,
so they disappear, and the small splitters, the successful ones,
prevail.

Variations like those that code for big- or small-splitting may have
occurred because of some random change in the genes long ago. That
"silent" change now becomes important for surviving and reproducing.
Creatures that reproduce sexually have many more chances like that, to
adapt better to a new environment, because their genetic material is
based on cutting the genetic "card deck" into two halves (by meiosis,
the creation of sperm and egg; all other cell divisions are mitotic,
like cloning, because they all splitters have the exact same genetic
material). In sexual reproduction, the two half decks, shuffled, join
to become the new organism. If any genes in the new organism make it
die soon, or make it survive and reproduce longer, this will be
translated into "failure" or "success".

And so it goes: Nothing succeeds like success. The replicators replicate
because their vehicles survive and reproduce.

> It need not be
> an organism which evolves but also an ontogenetic behaviour considered
> `successful' by others of the species. Survivors survive, while the
> weaker (=less well adapted) die out.

It's not that the organism (vehicle) evolves; nor is it exactly the
replicators (genes) that evolve: Organisms survive and reproduce, and
replicators replicate if what they code for helps their vehicle survive
and reproduce better. This PROCESS is called evolution.

> The process of Natural Selection occurs through the natural variation
> within members of a species, such that the better survivors have more
> chance of reproducing and passing on their characteristics to the next
> generation.

That's a standard definition of evolution, For this course, you should
think of it in terms of replicators (codes that pass themselves on) and
their vehicles.

> Success can be measured in terms of how long a lineage a
> particular gene/ trait stretches.

Yes, I suppose, but it's kind of arbitrary. Individual success is based
on survival and reproduction, so every creature on earth now is in this
sense equally successful, because they are all still here. All
replicators inside them are likewise successful. But I'm not sure
whether time is the right measure for success. Some creatures live long
lives, some short; some have many offspring, some few. Is a creature
that has successfully reared one offspring to reproductive age in 20
years more or less successful than the fruitfly that has bred millions
of generations in those 20 years?

Perhaps it's easier to define failure than success: Any gene that codes
for a trait that fails to get it successfully to the next generation is
unsuccessful.

> Cloning is a successful form of
> reproduction as long as the parent is well adapted to the prevailing
> environment but sexual reproduction enjoys an advantage over asexual
> reproduction in the case of subtly varying conditions, since genes of
> two individuals are combined to create a unique new individual which
> may turn out to be more successful in its environment and consequently
> pass on better qualities to the next generation.

Correct. Sex evolved to allow more variation (because, as usual, it was
more successful than cloning).

> Significant variation in the make-up of a species may take 100's of
> generations so the current traits and tendencies should only be
> interpreted in terms of the Environment of Evolutionary Adaptiveness
> (EEA) or the Original Environment. These concern the `distal' causes of
> evolution.

Yes, distal (or ultimate) causes are the ones that made a trait
advantageous in the EEA: sugar gave energy to escape predators (say), so
those ancestral children who ate all the scarce sugar whenever they had
a chance fared better, and their sweet-toothed genes are here to tell us
about it. The proximal cause of sugar-preference in children is the fact
that it tastes good.

Think of the distal cause as the (evolutionary) end, and the proximal
cause as the (psychological) means.

> An ESS is a trait which can exist and prosper and not be invaded by
> cheats. If a person helped any stranger they met, then there would be a
> `help debt'. If the rule was that you would try to help strangers in
> order that they they would help you (to pay off the debt) then it would
> result in a net population of helpful people. Cheats could not simply
> expect help as they have to earn it. Individuals would try to stay `out
> of debt' and help as many others as they could.

Not quite: First of all, this is mixing up distal and proximal causes;
you also do not give an evolutionary scenario for the advantages and the
success of the strategies. I'll invent one: Let's say that the
environment gets very wet, so that people often end up in water over
their heads.

Now we don't need any special explanation for the help that mothers and
fathers would give to their children in this situation; that's ordinary
selfish gene work: Helping your children helps your genes into the next
generation. But what about the help you may give to nonrelatives or
strangers? If you were rushing home in this waterworld and saw a
stranger calling for help, then if you invested time and energy into
helping that person, your own children could be drowning in the
meantime without your help.

Obviously this scenario must be repeated over and over, and hence what
we are talking about is a TENDENCY to help (or not-help) nonrelatives.
On the face of it, those who have the tendency to do so are
disadvantaging their own genes, and hence that strategy would not meet
with success. Moreover, in a world where there were helpers and
nonhelpers, the nonhelpers would have a double advantage: they don't
deprive their own genes of help, AND they get free help from others.

So if it were just a help and nonhelp gene, the nonhelp gene would soon
be the only one left, so there's an end to altruism.

But you suggested that if there were some way to ensure that there was
reciprocity, so everyone gave as good as they got, that might be an
ESS. But there's a problem with the proximate mechanism for this: What
would be the guarantee or even the likelihood that you were dealing
with a reciprocating person? Rarely do life events come in balanced,
controlled tit-for-tat pairs, so you can see clearly whether you are
helping a reciprocater or not. Besides, if there WAS a signal -- say, a
red-cross emblem on the forehead of reciprocaters, then of course this
strategy could be easily invaded by nonhelp genes that produce a false
red cross!

Other have thought that mutual helpfulness may be useful to the group, or
the species: But what is a species? It is a population of interbreeding
organisms that compete for the same resources. So there is no "good of
the species," only the competing vehicles for selfish genes.

One group effect might be something like cooperative hunting, where, for
example, the wolf pack have more success if they hunt in groups; but
there is no question of giving up the meat to other wolves (other than
mates and offspring). It's only the hunt itself that is collaborative.

This is a very controversial topic. Selfish Gene theory has only one
possible mechanism for helping nonrelatives: That they are not really
nonrelatives. Most of the members of the tribes of people in the ESS
would be related to a greater degree than arbitrary pairs of people, To
avoid inbreeding, they would (like today's hunter-gatherers) marry from
a nearby tribe rather than their own, but those nearby tribes wouldn't
be complete strangers either; with spouse swapping, the degree of
relatedness between the two tribes would be at the 2nd-3rd cousin
level, which is far enough to avoid the dangerous effects of
inbreeding, but not far enough to be a total stranger either.

So this is the source of the theory of "inclusive fitness": The
intuition pump here is the saying: "I'll lay down my life for:
2 siblings, 4 cousins, 8 2nd-cousins, 16 3rd cousins... etc."

In the EEA, we shared a lot of our genes with the people we saw most in
our lives, and it was for the cumulative success of our genes that we
cooperated as much as we did (and do now). This means that in the same
way that many genes in the same vehicle may have to "collaborate" in
order to make it into the next generation, so do the genes distributed
across our kin: In the case of parents and children this is obvious;
for siblings, almost as obvious. So for inclusive fitness, just
extrapolate this to more distant kin.

The "signal" for helping kin would not be a kinship symbol on the
forehead, because that could be invaded: Instead, the proximate means
by which we accomplish the distal ends of inclusive fitness is a social
and perceptual one: Not only do parents "imprint" on their children, and
vice versa, so that it's because they love the baby that they have
known since birth that makes them treat it well rather than badly,
but children imprint on their extended families and, to an extent, with
the children they are raised with (even if they are not relatives).
This imprinting makes them want to help and discourages them from
wanting to mate with their nuclear or extended family. That seems to be
an ESS.

Instead of a genetic symbol, which could be invaded, it is the normally
expected environment that singles out the people you want to help: Our
ancestors would not have survived long if they were surrounded from
birth by strangers instead of relatives. Being in a nondangerous group
throughout youth is a precondition on making it to adulthood at all (or
at least it was in the EEA): It is familiarity that mediates our
helpfulness, but only because in the EEA the familiar people would have
been mostly kin to varying degrees. And that is an ESS because there
would be no way for a stranger to infiltrate a tribe undetected and
masquerade as a member. As strong as are our tendencies to help our
familiars (who were mostly kin in the EEA), we have equally strong,
homicidal tendencies toward strangers, or even familiar people of
another "kind."

This was behind the "ethnic cleansing" in Bosnia, for although Serbs and
Muslims grew up side by side, so powerful is our xenophobic tendency
that neighbours were prepared to murder lifelong neighbours when they
got the chance. Here too is an uncertain point in the
sociobiological/cognitive common ground: With familiar neighbours, the
EEA familiarity-detectors should have made us behave toward them as
distant kin; but the cognitive side has been nursing, during all those
years of familiarity, a story about Serbia under the invasion of Turks
centuries ago, which made many ready to betray and kill their
neighbours (as "turks"): Does cognition "over-ride" the dictates of the
selfish gene in cases like this? Or is cognition just a more complex
"signal" than familiarity?

One possibility about what might be happening here is that a distal ESS
of helping those in our extended family, as signalled by their
familiarity, is being over-ridden, not by cognition. but by another
ESS, namely xenophobia: Attack and/or escape from what is UNfamiliar,
or familiar only as adversaries.

Hunter/gatherer tribes have alliances with others nearby tribes trade
cousins for marriage and are mostly on good terms; and then there are
other tribes, usually living further away, with whom there is on and
off warfare. Xenophobia is triggered by the unfamiliar -- especially in
threatening form, like a warrior in full war regalia. But a different
language or skin colour can elicit the same xenophobia. (The ethnic
mixes in which we live today were not there in the EEA; there a
different skin colour, language or dress signalled danger.)

> As for the Modus Ponens (sp?) I was a bit confused. Swans are white - Y.
> This is a swan - Y. Ok. This is white - N. ???

I'll switch to something that is really true (since it's not true that
all swans are white; The example is actually a syllogism rather than
just modus ponens, but never mind):

ACHILLES: Do you agree that it's true that:

              (1) All human beings are mortal
          (i.e., they will all die some day)?

TORTOISE: Yup. I sure do.

ACHILLES: And do you also agree that it's true that:

          (2) Colonel Qadafi is a human being.

TORTOISE: Yup, I sure do!

ACHILLES: And do you also agree that therefore it is true that:

         (3) Colonel Quadafi is mortal.

TORTOISE: Nope. Why should I?

ACHILLES: Hang on. Let's go back and try this again: Didn't you
          agree that (1) was true?

TORTOISE: Yup.

ACHILLES: And didn't you agree that (2) was true?

TORTOISE: Yup. Sure did.

ACHILLES: And don't you agree that if (1) is true and (2) is true,
          then (3) must be true too? (4)

TORTOISE: Yup, sure do!

ACHILLES: Well then, (3) is true, right? (5)

TORTOISE: Nope. Why should it be?

ACHILLES: (Perspiring and getting a little red.). Let's go through this
          another time: You agree that (4) was true? That if (4) was
          true, (5) has to be true.

TORTOISE: Yup, sure do!

ACHILLES: Well then (5) is true, right?

TORTOISE: Nope, why should it be...

This dialogue can go on forever, as Xeno's Paradox (about Achilles never
being able to catch up with the Tortoise) can. What it shows is that it
is not possible to get anyone to be logical by logical means alone. A
proof is a proof, but for it to be binding on someone's beliefs, they
already have to be committed -- emotionally -- to the truth of simple
2-step deductions like these. If someone is not emotionally committed to
them, he can't be forced into accepting them by logic alone.



This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:23:07 GMT