The Confusion over Evolution
Review of The Ant and the Peacock by Helena Cronin,
The Miner's Canary by Niles Eldredge, and On Methuselah's
Trail, by Peter Douglas Ward.
by Stephen Jay Gould
liver
Cromwell delivered history's most famous rebuke to the hero-worshiping that
irons all subtlety into flawless cardboard:
Mr. Lely, I desire you would use
all your skill to paint my picture truly like me, and not flatter me at all;
but remark all these roughnesses, pimples, warts, and everything as you see
me, otherwise I will never pay a farthing for it.
Helena
Cronin, in The Ant and the Peacock, displays a raw talent clearly
equal to that of our finest portraitists, but has placed herself into a
position even worse than Mr. Lely's. Cromwell's painter at least faced the
subject himself; Cronin has produced an uncritical gloss upon a false and
simplistic view that never was more than a caricature of Darwinian theory.
As its most deliciously radical component, Darwin's
original theory proposed a causal mechanism for evolution by natural
selection among organisms struggling for personal reproductive
success—and nothing else. Consider the impact of this cleansing upon
the older tradition of natural theology—the creationist principle that
sought to prove not only God's existence, but also his attributes of power
and goodness, from the excellent design of organisms and the intrinsic
harmony of ecosystems. Darwin acknowledged these aspects of nature, but
labeled them as sequelae, or side consequences, of the only causal force
operating in evolutionary change: organisms struggling for themselves alone.
Quite a contrast: up from below in the "selfish" interest of organisms vs.
down from above as directly imposed by a wise creator.
Inevitably, I suppose, Darwin's success in pulling down
the level of causality from an overarching God to a struggle among organisms
led some evolutionists to explore a kind of ultimate reductionism in viewing
genes themselves as the struggling units, and organisms as mere vehicles
constructed for their machinations. Under such a view, called "gene
selectionism," nature's truly causal competition takes place among different
forms of a gene, each "struggling" to leave more copies of its own version
in future generations of a population. In classical Darwinism,
organisms struggle for reproductive success. (If, for example, short
plants are favored by natural selection, then the runts in Mendel's pea
patch produce more surviving offspring, per individual and on average, than
the tall plants.) In gene selectionism, the plants are passive vehicles and
the struggle occurs among genes. (If you can dredge up your high school
biology, you will remember that T and t ["big t" and "little t" even in
biological jargon] are different forms [called alleles] of a gene at a
chromosomal position [called a locus] influencing the height of the
resulting pea plants. Under gene selectionism, little t is struggling with
big t to leave more copies of itself in the next generation. Selection is
then viewed as working for little t alleles, not for short plants.)
Several of my colleagues toyed with this formulation
during the 1970s, while Richard Dawkins provided a popular version in his
book The Selfish Gene (1976). This hyper-Darwinian reductionism (and
pervasive adaptationism, though from the gene's point of view) contained
some interesting ideas and made a stir within the field. But gene
selectionism—as a hard causal claim rather than a colorful
metaphor—also received sharp and devastating criticism both from
biologists and philosophers.[1] Even Richard Dawkins
backed away in his next book (The Extended Phenotype, 1982), which
opened with the fatal concession that natural selection among genes and
ordinary Darwinian selection among organisms may be viewed as equally valid
modes of description for the same causal phenomenon. (This stunning
admission of relativism flatly contradicted Dawkins's previous claim for
true and exclusive causality at the genic level.)
Helena Cronin, a philosopher by original profession, has
gathered much interesting material in The Ant and the Peacock, but
her book suffers grievously from the curious and vociferously advocated
scheme that she has chosen as her vehicle of presentation. In short, she has
somehow received the impression that genic selectionism has accomplished the
greatest revolution since Darwin and has swept away all opposition within
the field (she proclaims "revolution" as often as Marx or Thomas Kuhn, and
labels the new orthodoxy "modern Darwinism").
Now I freely confess my own strong preference for the
other side of this debate—for a model that views selection as operating
simultaneously at several levels of a genealogical hierarchy including
genes, organisms, local populations, and species. In other words, I argue
that no natural entity can properly be described as the exclusive "unit of
selection"—as Cronin and Dawkins would claim for genes, and classical
Darwinians for organisms. Nature is organized as a hierarchy—genes in
organisms, organisms in populations, and populations in species. Entities at
each level of the hierarchy can act as biological "individuals," and
Darwin's process of selection can therefore occur at all levels, with none
dominant in all situations. Genes may "struggle" for increased
representation in the gene pool of a population, but species may also
"struggle" for increased membership of their branches in the "species pool"
of an evolutionary lineage.
But whatever one's personal inclinations, no one can deny
the sociological fact that relatively few experts accept the theory of near
exclusivity for gene selection, and no amount of blithe verbal assurance
about the validity of the theory can convert it into a successful revolution.
Most of my colleagues continue to defend Darwin's view that selection works
nearly exclusively on organisms.
For Cronin, ants and peacocks are synecdoches for two
great issues mentioned in her subtitle—altruism, epitomized by the
behavior of ants in their communities, and sexual selection, the alleged
raison d'être of the peacock's flamboyant and burdensome tail. In
Cronin's view both altruism and sexual selection are explained by the theory
of selfish genes; and I agree that the gene's perspective has been useful in
dealing with these two substantial problems. But I shall argue that the
peacock only needed classical Darwinism to account for its tail, while the
key question evoked by the ant's altruism has not been resolved. If the
approach of gene selectionism is false, then Cronin's title makes little
sense and all her fascinating flowers of insight (primarily on the
differences between Darwin and Wallace) languish in barren soil.
This attempt to validate gene selectionism by applying it
to ants and peacocks fails for three main reasons (which I consider in turn
in the rest of this review): 1) the general theory is bankrupt; 2) the two
chosen examples form a false and disparate pairing that either does not need
or does not fully illustrate the theory; 3) the theory, even if valid in its
own limited realm (which it is not), cannot serve as a paradigm for all, or
even much, that evolutionary theory must explain. The fine books of Ward and
Eldredge illustrate why this is so.
Fallacy of the General Theory
Since Darwinism is a theory about differential
reproductive success ("survival of the fittest" in the old cliché of
limited utility), and since organisms are plainly doing the struggling and
reproducing "out there" in nature, why would anyone want to relocate the
action at the level of genes encased within these organisms? This question
engages a central issue in Darwinian theory: On what kind of object does
natural selection work? What, in short, is a "unit of selection"? Cronin
gives her answer in no uncertain terms:
So when is an "adaptive unit" really an
adaptive unit? When is a category that's seen by us, seen by nature, too?
The answer must be: When it's a unit that selection can work on. For
classical Darwinism this would have been difficult to specify precisely. But
for modern Darwinism, a unit is obviously a gene and the ramifying tree of
all its phenotypic effects [by "phenotypic effects," she means the manifest
characteristics of the organism including its anatomy,
etc.].
But if gene selectionism is so self-evidently true, why
hasn't it swept the field, smiting all opposition before it? In a remarkable
passage, Cronin admits the stubborn persistence of alternative
interpretations, but brands them as obtuse because she so clearly grasps the
true logic of genuine Darwinism—and it says what she says it says,
period.
Let me forestall the mutterings of
disagreement that can already be heard in the background. By no means all
modern Darwinians would accept my characterisation. But I am dealing with
the theory, not with how individuals have chosen to interpret it. One must
distinguish between the fundamental tenets of a theory (what the theory
actually says) and how it is viewed by some practitioners (what is said
about it). I am dealing with the former.
Cronin's confidence arises from her misuse of an
important distinction between "replicators" and "interactors" made by the
philosopher of science David Hull and others. Natural selection is a theory
of "differential reproductive success"; therefore, according to one school
of thought followed by Cronin, only natural objects that replicate
themselves faithfully can be units of selection. For if an object doesn't
replicate itself faithfully, then it cannot be a reliable transmitter of the
characteristics that make for superior reproductive success. Now an
organism, the traditional "unit of selection" in Darwinism, fails by this
criterion because, in sexual reproduction, offspring contain only half the
genes of each parent. What good is a replicator that dilutes itself by half
in each generation?
The genes themselves, on the other hand, replicate
faithfully (except for rare mutation) into future generations. A fecund
organism is passing copies of its genes, not its body, into the next
generation. Under this view, genes are replicators, while bodies are, in a
sense, their servants. Bodies interact with the environment and engage in
the "struggle for existence" via differential reproductive success. Bodies
are interactors (Richard Dawkins prefers the more loaded and almost
pejorative term "vehicles"); genes are replicators. Therefore, only genes
are units of selection. Cronin remarks.
Genes, then, can be replicators whereas
organisms, groups and other levels of the hierarchy cannot. Natural
selection is about the differential survival of replicators. So genes are
the only serious candidates for units of selection.
This superficially attractive argument collapses from two
major fallacies. First, it is simply not true that only genes replicate with
adequate faithfulness. I accept the point that sexual reproduction dilutes
the integrity of organisms in replication. But adequate replication returns
at higher levels of the hierarchy—populations and species—because
splitting at these levels is analogous to asexual reproduction. For example,
species split into daughter species that resemble their parental populations
far more than any other (descendant dogs are more like ancestral wolves than
like any other species). Species are therefore good replicators, and some
evolutionary lineages can be more successful than others because their
species give rise to more successful branches.
Second, the replicator criterion is at best insufficient,
and at worst entirely mistaken. A simple appeal to vernacular usage tells us
that a lower unit (a gene, for example) can't be an exclusive agent if all
the action occurs at higher levels (organisms, for example)—and the
properties that generate this action are "emergent" characters of the higher
level—that is, not a simple summation of features built by the lower
units (genes).[2] Now, manifestly (and gene
selectionists do not deny this), organisms are primary objects struggling
for reproductive success in nature. How, then, can "hidden" genes be the
true agents if organisms are doing the fighting, cooperating, generating,
and dying? Gene selectionists respond that all the relevant properties of
organisms can be described as results of the various genes involved in their
construction. Such properties, the argument continues, are therefore only
the complex manifestation of genetic action.
But many, undoubtedly most, properties of organisms are
not simple summations of contributions from several genes. They are products
of interactions among genes and therefore they cannot, in principle, be
adequately predicted or known at the level of genes. Since selection acts on
such emergent properties of organisms, genes cannot be exclusive units of
selection.
Moreover, emergence occurs frequently at all levels of
biological organization. Selection acts simultaneously on all levels of
nature's hierarchy—on genes (Cronin's "exclusive" level), on cell
lineages, on organisms (Darwin's "nearly exclusive" level), on local
populations, and on species. Our intellectual task as evolutionary
biologists is to determine the relative importance and complex interactions
of these levels. Claims for theoretical exclusivity of one chosen level are
false and blinkering.
Why have some scientists been attracted to exclusive gene
selectionism in the light of these criticisms? I believe that the appeal of
the idea rests upon a central fallacy, one embraced by Cronin: the confusion
of bookkeeping with causality. Gene selectionists have correctly noted, but
fundamentally misinterpreted, an important property of evolving systems: all
evolution by selection, whatever its level of causation, is
recorded by change in the frequencies of genes (the lowest level of
the causal hierarchy). Since genes record all changes, some evolutionists
have been fooled into assuming that genes therefore cause all
changes. But scribes are not agents, and bookkeeping is not causality.
This recording of all change by genes reflects a general
property of hierarchies, not a statement about the locus of causality.
Disturb a hierarchy at any level, and all units at that level and below must
be shuffled—while higher-level units may be unaffected. For example,
suppose that the lineage of cats once contained two groups of five species
each. They compete entirely at the level of species selection and, 10
million years later, all species in group A are eliminated, while group B
has flourished and now contains fifteen species. The genetic makeup of the
entire lineage has altered as a result (for species in group A had different
genes in different proportions from species in group B)—but no causal
process has operated at the gene level.
But the obverse does not hold: lower-level impacts on
genes need not affect higher-level units at all. For example, some genes may
increase their number of copies within organisms by genic selection
alone—but organisms need not "notice" and need not be affected in any
way.[3] Now this asymmetry in
hierarchies—upper-level disturbances felt at all lower levels, but
lower-level impacts potentially invisible to higher-level units—makes
the lowest level an attractive place for bookkeeping, because all changes,
whatever their causal locus, are recorded there.
Bookkeepers are slighted in our metaphors as mere scribes
and recorders of action occurring elsewhere. What magnificent revenge the
error of gene selectionism must promise them! For their ledgers—the
tables of changes in genetic composition—have been misread by gene
selectionists as the fundamental cause of life's history. But revenge based
on error is both short and ultimately frustrating, for the deprecatory
metaphor turns out to be true in this case.
Fallacies of Cronin's Particular Arguments
Cronin's entire book promotes what I like to call the
Senator Aiken strategy for untenable positions. This fine legislator once
proposed a wonderfully simple solution for the morass of our military
involvement in Vietnam: Why don't we simply declare victory and get out!
Cronin does much the same. She proclaims victory, dogmatically and
vociferously, over and over again, for the gene-selectionist version of
strict adaptationism ("modern Darwinism" in her neologism). In one
remarkable passage she even tells us that Darwinism has triumphed on other
unknown planets because evolution can work in no other way!
We have seen why Darwinism was in 1859, and
still is, the best explanation of why living things are as they
are—not only, it turns out, on this planet, but in any world that
resembles ours in several fundamental respects.
In essence, Aiken's strategy must mask failure with a
claim for triumph. How then does Cronin manage to defend such a flawed and
discredited approach as pure adaptationism from the gene's point of view?
She snatches rhetorical victory from the jaws of defeat by a series of false
arguments and uncritical assertions. I give examples in just a few
categories:
1) False or misleadingly incomplete citation of empirical
data. To support her panadaptationist world view, the belief that virtually
all heritable changes in organisms are the result of adaptation, Cronin
tells several classical tales of natural history as triumphs of selection
(when the actual story is far more complex and ambiguous). For example, she
cites color banding in the land snail Cepaea as "just one example of
natural selection rescuing phenomena from the explanatory clutches of
genetic drift."[4] (Genetic drift refers to a random
process of increase or decrease in the frequencies of genes in populations.
Suppose that only ten individuals exist in a species of beetles, and that
three of them carry a favorable gene subject to positive Darwinian
selection. An elephant rambles by and squashes half the beetles, including
[by chance] all three with the favorable gene [which is now randomly
eliminated from the population, despite its Darwinian value].)
Now A.J. Cain and others demonstrated the power of
selection (based on visual discrimination among differently banded snails by
predatory birds in different habitats) where drift had previously been
assumed—and this work was a genuine triumph for Darwinism. But Cronin
never mentions the second chapter of this story—Cain's unsuccessful
attempts to explain so-called area effects (abrupt changes in banding types
from one region to another, but correlated with no evident change in
habitat) by selection based upon "cryptic environmental differences." Most
snail workers (I am one) now regard area effects as nonadaptive vestiges of
former historical movements of populations and habitats.[5]
2) Relegation of powerful and important exceptions to
peripheries. Cronin admits the vital principle that "some side effects of
adaptations, which become positively useful when conditions change, are
until then just lying around dormant." But she cautions against overuse of
this principle: "Such arguments, unless they are applied with
discrimination, could end up peppering the world with a multitude of
characteristics that have no Darwinian purpose (even though they eventually
get put to good use)." But the world is so peppered—and this seasoning
is a fundamental (and nonadaptive) feature of evolution. In fact, Cronin
actually admits the primary example in the very next sentence—features
in the human brain, arising as nonadaptive sequelae of its computing power,
as in our ability to read and write—and then buries the subject without
further commentary.
3) Mis-citation of critics. Cronin quotes a line from me,
for example, and delights in her discovery: "If we are programmed to be what
we are, then these traits are ineluctable. We may, at best, channel them,
but we cannot change them, either by will, education, or culture." She then
makes the following sarcastic comment, assuming that these words, from an
article I wrote in 1978, represent my own view. "That's die-hard
intransigence for you! But, actually, those quotes come not from some ardent
proponent of an all-in-our-genes view but from Stephen Gould, a voluble
critic of selfish gene-ery in general and of its application to humans in
particular." I will admit to writing with less than optimal clarity on
occasion, but there can't be much doubt that in this passage I was
characterizing my opponents' views, not my own. Earlier, in the very
paragraph Cronin quotes, I labeled this view as "a crude biological
determinism." Of course the words convey "die-hard intransigence"—for
they are my description, my caricature I will even admit, of the
opposition.
4) Ignoring opponents. As I have noted, the pure gene
selectionism championed by Cronin is a marginal position among evolutionists
(this, of course, doesn't make it wrong). Yet Cronin falsely depicts this
view as a consensus. She performs this astonishing turnabout by simply not
discussing, usually not even mentioning at all, the numerous and
devastatingly effective critiques that invalidated gene selectionism after
its brief run of incipient popularity in the 1970s (see my
first footnote).
If Cronin's general account of gene selectionism is so
awry, then the chosen examples (ant and peacock), and attendant problems
(altruism and sexual selection), don't fit together either, for even
well-crafted pictures may jar when juxtaposed in a single false frame. She
claims that both are classical problems of old-style Darwinism (selection on
organisms), now triumphantly solved by the modern gene-selectionist
version.
In epitome, the peacock's tale (also tail) is a story of
delayed vindication for selectionism. Darwin developed an ancillary
mechanism, which he called "sexual selection," to explain competition among
members of the same species for access to reproduction, in contrast with the
usual form of "natural selection," or competition for limited resources to
sustain life. (Darwin developed the term and concept in The Origin of Species
[1859], but covered the subject most thoroughly in his 1871 treatise on
The Descent of Man and Selection in Relation to Sex—a work often
read only for its short and speculative thoughts on human evolution, but, in
the main, a long and copiously documented treatise on sexual selection
throughout the animal kingdom.)
Darwin delineated two modes of sexual selection, called
"male competition" and "female choice." In male competition—e.g., among
antlered deer—males fight like hell and the winners get the females. In
female choice males strut and preen, display and bellow, and females choose
to mate with the individuals that impress them most. Peacocks, in other
words, do not evolve their showy tails for direct victory in battle over
other males, but to win a beauty contest run by females.
Male competition never sparked any controversy, for it
looks so much like good old natural selection. What does it matter if two
male deer fight for access to food or to females? They still need good
weapons and nature remains red in tooth and claw. (I usually avoid this
clichéd line from In Memoriam; but 1992 is the one hundredth
anniversary of Tennyson's death, so I make an exception.) Female choice,
however, elicited a firestorm of criticism. Most of Darwin's contemporaries
rejected the concept, often vehemently. It was similarly ignored, and
curiously so, throughout the early excitement in twentieth-century studies
of animal behavior. Neither Julian Huxley nor the German ethologists had any
use for the idea. But female choice has roared back to acceptability and
prominence during the past twenty years—and I certainly agree with
Cronin that this reversion to Darwin's original concept represents one of
the most important contemporary themes in evolutionary theory.
But I strongly disagree that gene selectionism lies
behind this renaissance; hence, this half of Cronin's book, while full of
insight and interesting documentation, does not support or illustrate her
argument. I think that she has fallen into the classic error of equating
correlation with causality. It is true that gene selectionism had its fling
in the 1970s and that sexual selection began its renaissance at the same
time. But, during the same years, Watergate unfolded and The
Godfather won an Academy Award—and I really don't think that these
coincident events are causally related either.
You don't need gene selectionism to validate female
choice; plain old Darwinian selection on organisms works perfectly well (as
Darwin himself recognized in establishing the concept): females, in choosing
the most healthy and vigorous males, are bolstering their personal
reproductive success—the essence of the Darwinian game.
If female choice did not need gene selectionism, then why
did its vindication occur so recently? When this interesting history is
sorted out, I believe that the record will show a renaissance based
primarily on the removal of two longstanding impediments (both mentioned by
Cronin, but treated as incidental), rather than the emergence of any new
(and ultimately fallacious) theory such as gene selectionism.
Disrespect for the cognitive capacity of nonhuman animals
formed the first impediment: Male competition is just "doin' what comes
naturally," but female choice requires an additional mental step that seemed
too close to cognitive abilities supposedly unique to humans: that is,
females must survey a field and make a judgment based on some aesthetic
criterion of beauty in sight or sound. Most biologists weren't willing to
grant such capacity to animals. Now we know better, but the insight was a
long time coming.
The second impediment arose from an unwillingness to
place females in the driver's seat of evolutionary change. Again, male
competition just represents the big boys in ordinary control, but female
choice argues that males go to enormous lengths of adornment and
posturing—all to await the judgment of females. Like nearly every
science, evolutionary biology was almost entirely a male preserve until this
generation. You don't have to be an active sexist to ignore female choice in
such a male-dominated world. You may be the kindest male on earth and still
fall prey to a social and intellectual atmosphere that doesn't include the
concept of females in such control. (Sometimes, however, the source of bias
is more overt. In a line cited by Cronin, the prominent British biologist
and Darwinian opponent St. George Mivart proclaimed: "Such is the
instability of a vicious feminine caprice, that no constancy of coloration
could be produced by its selective action.")
I don't mean to sound like a mouthpiece of contemporary
political correctness, but I do feel that the delay in acceptability for
Darwin's well-formulated concept of female choice lies in the social
impediments of sexism and speciesism. And I suspect that this concept's
recent rise to prominence largely reflects the social and political
questioning of such biases, as well as the most important and salutary
sociological shift in recent science: the entrance of so many women into the
profession.
If the peacock doesn't fit into Cronin's chosen setting
of gene selectionism because the problem has indeed been solved, although
not by her particular approach, then the issue of the ant (altruism) is
quite different—for the ant does fit, but the problem has not been
solved. Ants and other social insects produce mostly sterile, though
genetically female, offspring—the "soldiers" and "workers" of hives and
hills. But how can such sterility evolve if the Darwinian game is
fundamentally about personal reproductive success? (Evolutionary
biologists define altruism in this limited and operational sense as behavior
that decreases one's own reproductive potential in the service of others.
The sterility of worker ants may represent an extreme example, but animal
behavior is rife with other cases that merely endanger an altruist for the
apparent benefit of others—warning calls issued by birds who sight a
predator, for example. Thus, the problem of altruism, so defined, has been
central in Darwinian theory.)
On this issue, I agree with Cronin up to a point, and I
greatly appreciate her incisive treatment. A solution to the problem of
altruism has provided the greatest success for viewing evolution from the
gene's point of view, while directly departing from Darwin's own focus on
organisms. The key insight, usually called "kin selection theory," was
provided by W.D. Hamilton in
the mid-1960s, though many hints and half-formulations can be found in
earlier literature.
What is the "individual reproductive success" of which
Darwin speaks? It cannot be the passage of one's body into the next
generation—for, truly, you can't take it with you in this sense above
all! "Reproductive success" can only mean the passage of more copies of
one's own genes into future generations. Such success is usually best
achieved by bearing a maximal number of surviving offspring
oneself—hence the usual Darwinian struggle for personal reproduction.
Usually, but not always. We share, on average, a certain percentage of genes
with our relatives, depending on closeness—one half with each of our
parents and full sibs, one fourth with our grandchildren and half-sibs, one
eighth with our first cousins.
Now suppose that I am in a position where I can either
die to save three full sibs or survive at the cost of their death. What
should I do in the Darwinian calculus? Die for the three sibs, of course,
for they represent, in sum, 150 percent of my genes, while I hold only 100
percent. Better for my genes if I go in order to let three of them live to
reproduce. My act may look altruistic from the organism's point of view, but
it is properly selfish and Darwinian from the gene's perspective. In short,
the theory of kin selection explains apparent acts of sacrifice as evolved
Darwinian adaptations in the cardinal interest of passing more copies of
one's genes to future generations.
The model is powerful because it suggests an eminently
testable research program: study the context of altruistic acts and see if
they are performed for the benefit of enough close relatives to overbalance,
through fecundity of kin, any individual loss of reproductive success. The
model has been tested and confirmed in a wide variety of cases, including
sterility in social insects—as among ants, where the sterile females
known as workers forgo their own reproduction to help their mother, the
queen, raise fertile sisters.
But we now come to the crucial, if admittedly parochial,
limit. We are primarily interested in the problem of altruism because human
beings seem so singularly capable of behavior in the apparent interest of
others. Manifestly, this behavior is frequently not directed toward close
relatives (especially since we don't choose to describe sacrificial acts
between parents and children as altruistic)—and therefore can't be
encompassed by the gene-centered argument of kin selection. Biologists may
choose an operational redefinition of the vernacular word "altruism" as
apparent sacrifice for actual genetic good, but this concept does not apply
to most human acts deserving the word (and one might even argue that the
vernacular notion demands no hidden selfishness at any level).
One might reply that the introduction of vernacular human
altruism into the argument is unfair. After all, if gene selectionism has
been successful for ants and other animals, then grant the victory and leave
us out of it (for the professional enlightenment would still be great, even
if the subject then failed to touch popular interest in our own condition).
But Protagoras was apparently right (even if he only spoke for half of us)
when he proclaimed that "man is the measure of all things," and few writers
for general audiences can resist the temptation of trying to extend their
perspective to this greatest prize of all. The longest chapter in Cronin's
book is titled: "Human altruism: A natural kind?"
How then might the acknowledged success of gene
selectionism be useful to us in understanding human altruism? Several
approaches have been tried, and none have succeeded. One might argue, as
Robert Trivers
and others have, that the calculus of kin selection won't work, but that a
different kind of selfish and adaptive advantage arises through the old
principle of "favor banking." If I am perceived as a good altruist through
acts that help nonrelatives, then other people are more likely to help me
when I am in need—so called "reciprocal altruism." Fine, but we
scarcely need Darwinism, or genetic arguments at all, to convince us that
humans are smart enough to figure out the advantages of "you scratch my back
and I'll scratch yours."
Or we might argue, as Cronin often suggests, that our
general altruistic urges evolved long ago by kin selection among small
groups of relatives, where neighbors were invariably kin, and the
evolutionary rule of "be nice to those close by" would suffice to guarantee
the Darwinian calculus. True altruism to nonrelatives would then be a
consequence of formerly advantageous behavior, now altered by a changing
social setting that makes neighbors of genetic strangers. I find this
argument unattractive on two grounds. It is, first of all, an untestable
speculation about unrecoverable behavior patterns of distant ancestors.
Second, historical origin and current status represent entirely different
problems in evolutionary biology. So what if the historical origin of
altruism were adaptive via kin selection? If we still do it, after centuries
of contexts unfavorable to the Darwinian calculus, then altruism is a
currently non-adaptive behavior (in the narrow Darwinian sense) demanding
some other explanation, presumably social rather than directly
biological.
In sum, the ant and the peacock are apples and oranges.
Gene selectionism neither unites the problems nor resolves the issue. We
don't require gene selection for the peacock, for Darwin resolved this
problem within his system of selection on organisms, though we needed
another century to dissolve social barriers impeding the acceptance of his
answer. Gene selectionism has been useful in explaining many examples of
what we call "altruism" in nonhuman animals (including ants), but it cannot
resolve the vernacular human style that remains our ethical glory and our
intellectual burden. The ant and the peacock don't belong together, and
Cronin's book is incoherent (in the literal, not the pejorative, sense). The
problems of sexual selection and altruism are as disparate as the outward
appearances of Cronin's chosen synecdoches—ants and peacocks.
The Strictly Limited Domain of Adaptationism, Whether
Gene or Organism Based
In praising the power of selection, Cronin writes:
The general point has been to illustrate how
resourceful and subtle a tactician natural selection can be…. Once this
is appreciated, non-adaptive explanations cannot be treated as other than a
last resort. And resolute adaptationists can be confident that "The use of
each trifling detail of structure is far from a barren search to those who
believe in natural selection." [The last line is a quotation from Darwin's
1862 book on orchids.]
Such confident effusions exemplify two crucial errors:
the false claim that selection dominates the domain of organic form, and the
flawed inference that this supposed domination provides, by extension, an
adequate account of evolution at all scales.
I have argued that gene selectionism is an ultimately
incorrect view of evolutionary mechanics. But suppose it were right. Would
gene selectionism then be the fully comprehensive theory of biological
change that its advocates tout so vociferously? As a paleontologist, working
with changes in units of millions of years rather than generations, I find
this strange assertion to be the most blinkered and untenable in the entire
catalog of strict Darwinian parochialisms.
Darwin himself relied crucially on such an extrapolative
vision: smoothly extend the adaptive struggles of generations across
millions of years in geological time, and you will obtain the entire,
wondrously ramified tree of life. Consider two famous passages from the
Origin of
Species:
It may be said that natural selection is
daily and hourly scrutinizing, throughout the world, every variation, even
the slightest; rejecting that which is bad, preserving and adding up all
that is good; silently and insensibly working, whatever and wherever
opportunity offers, at the improvement of each organic being in relation to
its organic and inorganic conditions of life. We see nothing of these slow
changes in progress until the hand of time has marked the long lapse of
ages.
The inhabitants of each successive period in
the world's history have beaten their predecessors in the race for life, and
are, in so far, higher in the scale of nature; and this may account for that
vague yet ill-defined sentiment, felt by many paleontologists, that
organization on the whole has progressed.
If this uniformitarian vision of extrapolation fails,
then we must conclude that while adaptationism may control immediate changes
in the overt forms of organisms, it cannot render evolution at other scales.
The main excitement in evolutionary theory during the past twenty years has
not been—as Cronin would have us believe—the shoring up of
Darwinism in its limited realm (by gene selectionism or any other patching
device), but rather the documentation of the reasons why Darwin's crucial
requirement for extrapolation has failed. Selectionism is not a general
model for evolutionary change at most scales.
In the world below organisms, at the scale of changes in
nucleotides of the DNA code, Motoo Kimura's theory of neutralism (based on
the prevalence of genetic drift, as defined earlier), combined with a better
understanding of genetic mechanisms, has demonstrated the neutrality of
much, if not most, alteration at minimal magnitude. Selectionists often
respond, as Cronin does, that their Darwinian preferences are not thereby
compromised because such neutral genetic changes do not alter the external
forms of organisms, and therefore couldn't be "seen" by natural selection
anyway. Cronin writes: "[Kimura's] theory also assumes that chance is an
evolutionary force but it is to do with changes at the molecular level that
have no phenotypic effects, not evolution in the sense that we are concerned
with—adaptive change." But how can you dismiss a process that probably
accounts for more than 50 percent of all genetic change by noting that such
alterations don't manifest themselves at the level that happens to interest
you most? This special interest, after all, is just a parochialism based on
human sizes and lifetimes, and on the history of our thinking. Nature,
working at so many other scales, takes scant notice and plays no favorites.
If we lived in the world of nucleotides, we would see the random ebb and
flux as fundamental and view occasional islands of adaptive coagulation at
larger scales as peculiar exceptions in an alien domain.
But the ultimate failure of Cronin's adaptationism, as
a general evolutionary model, appears most clearly when we consider the
paleontological record. Darwin's vision may prevail in the here and now of
immediate adaptive struggles. But if we cannot extend the small changes
thereby produced into the grandeur of geological time to yield the full
tree of life, then Darwin's domain is a limited corner of evolutionary
explanation. New documentation on the rapidity and intensity of mass
extinction (including the event that wiped out dinosaurs) has provided the
strongest argument for rejecting Darwinian extrapolation. Darwin clearly
understood the threat, and he struggled against the implications of mass
extinction in the Origin
of Species by trying to deny both their extent and rapidity. He
endeavored to spread them out in time and diminish their effects. He
attempted to render them as an intensification of ordinary competition
(inspired, perhaps, by an increase in rates of change for conventional
processes like mountain-building and change in sea level). But if mass
extinctions are true breaks in continuity, if the slow building of
adaptation in normal times does not extend into predicted success across
mass extinction boundaries, then extrapolationism fails and adaptationism
succumbs.
The Permian extinction (about 225 million years ago) may
have wiped out 95 percent of marine invertebrate species. The Cretaceous
extinction (about 65 million years ago) was probably set off by the impact
of a large extraterrestrial body. The adaptive struggles of millions of
previous years, whatever their intensity and the beauty of their results,
could not prepare organisms for a random catastrophe. A fish honed to
hydrodynamic perfection will still die if the pond dries up. Survival
through mass extinction requires the good luck of evolving features for one
reason in normal times, and then finding them fortuitously well-suited for
survival through unanticipated catastrophe.
Two of my colleagues, Peter Ward and Niles Eldredge, have
recently written short and incisive books on mass extinction. Taken
together, this pair provides a fine documentation for why Darwinian
selection cannot, by extrapolation, encompass the history of life. The books
differ greatly in both content and intent. Ward's On Methuselah's
Trail is a personal account of the fieldwork that convinced him about
the catastrophic character of mass extinctions, particularly the event that
occurred at the end of the Cretaceous period. Eldredge's The Miner's
Canary strongly doubts scenarios of extraterrestrial impact and focuses
on similarities between mass dyings of the past and the current human
assault upon biodiversity (hence the metaphorical title, invoking the
organic side of our chief industrial symbol for harbingers of death by
environmental poisoning). Both books recognize the special and dominant
character of mass extinctions as agents that changed the pattern of the
history of life.
Consider just one example, supreme in its parochial
importance—for I wouldn't be writing and you wouldn't be reading
otherwise. Why did mammals survive, but dinosaurs die, in the great
Cretaceous extinction, an event almost surely triggered by extraterrestrial
impact? The adaptationist and extrapolationist model strives to render such
a turnover as intensification of a process already underway in previous
normal times—the growing domination of mammals as a result of their
success in ordinary Darwinian competition against inferior dinosaurs. But
such a comfortable argument cannot hold. Mammals emerged at about the same
time as dinosaurs. Mammals lived for more than 100 million years in the
interstices of a world dominated by much larger dinosaurs; they made no
"progress" against these massive incumbents; no Mesozoic mammal was much
larger than a rat. (By contrast, the so-called "age of mammals" since the
death of dinosaurs has so far spanned only 65 million years.) The Cretaceous
catastrophe removed dinosaurs, but mammals survived and inherited an emptied
world—and they surely made the most of it.
If the comet or asteroid had not struck, I suppose that
dinosaurs would probably still be in command (why not; they had prevailed
for far longer against mammals, and mammals had been making no inroads).
Mammals, if they survived at all, would probably still be small creatures no
larger than rats, and small size precludes self-conscious intelligence.
Dinosaurs were not moving toward higher cognition in our form, and probably
could not do so. Thus you must thank the extraterrestrial impact for this
copy of the New York Review.
But why did mammals prevail and dinosaurs die? Doesn't
this fact point to some intrinsic mammalian superiority? Not necessarily. We
do not know the answer, but here is one plausible scenario for a partial
explanation: the rules change in mass extinction, and adaptive advantages of
the past may become dangerous deficits. Large populations provide a good
hedge against extinction, all other things being equal. Dinosaurs, with
their massive bodies, must have maintained species of small population size.
The world must contain far fewer elephants than ants, far fewer
brontosauruses than mouse-sized mammals. So perhaps mammals gained a crucial
edge by large populations maintained as a consequence of small body
sizes.
Now why were mammals small? Surely not because they knew
that a comet would hit 10 million years down the road, and that large
populations would then be useful. Presumably they were small for a negative
reason in Darwin's immediate world of competition: because dinosaurs had
usurped the ecological space of large terrestrial vertebrates, and relegated
mammals to a periphery. Yet the reasons for relative failure in normal times
may translate fortuitously to the crucial ingredient of success in
prevailing through a mass extinction. The Darwinian struggle does not
extrapolate to the tree of life.
Ironically, Cronin does seem to grasp this issue in her
final paragraph, if only through a glass most darkly. After four hundred
pages of panadaptationism, she finally recognizes that evolutionary theory
must solve other problems as well—particularly the issue of shifting
diversity through time—and that adaptation may not provide the basis
for all answers. Darwin did wrestle brilliantly and triumphantly with the
problem of adaptation, but he had limited success with the issue of
diversity—even though he titled his book with reference to his relative
failure: the origin of species. Cronin records and admits this irony in the
last line of her book: "But, in the midst of such success, there was one
problem that remained just outside his [Darwin's] grasp. It
was—poignantly—the problem of the origin of species." When strict
Darwinians drop their reliance on adaptation and extrapolation, and when
they break bread with paleontologists in the different realm of time in
millions, they will then engage this unresolved problem face to
face.
Notes
See especially E. Sober's The
Nature of Selection (MIT Press, 1984), E. Sober and R. C. Lewontin,
"Artifact, Cause and Genic Selection," Philosophy of Science, Vol.
49, pp. 157–180; E. Lloyd, The Structure and Confirmation of
Evolutionary Theory (Greenwood Press, 1988); and P. Godfrey-Smith and
R. C. Lewontin's forthcoming "The Dimensions of Selection," Philosophy
of Science.
"Emergence" is a complex and
contentious subject, with a long pedigree, in both the philosophical and
biological literature. I use the term here in the narrow technical (virtually
statistical) sense. A feature is emergent at any level if its construction
requires nonadditive interaction among the factors and components that build
it. In other words, if I can make a larger-scale entity D by just adding
components A, B, and C together, then nothing about D is emergent—and
D can be explained by reduction to its components. But if the building of D
requires interactions among A, B, and C that are not inherent in the
components considered separately, and cannot be predicted from knowing A, B,
and C alone, then D has emergent features and cannot be explained by
reduction to its component parts. Organisms clearly have emergent properties,
since their features of anatomy, physiology, and behavior are products of
complex and nonadditive genetic and environmental interactions—and not
the summation of genes considered separately. Therefore, selection operating
on organisms cannot be reduced to selection upon genes, and the "gene
selectionism" of Cronin's self-proclaimed "modern Darwinism" fails.
Incidentally, the concept of emergence helps
us to understand why the nature-nurture issue is such a false dichotomy.
Genes influence many aspects of human behavior, but we cannot say that such
behavior is caused by genes in any direct way. We cannot even claim that a
given behavior is, say, 40 percent genetic and 60 percent environmental, and
thereby defend at least a partial old-fashioned genetic determinism. Genes
and environment interact in a nonadditive way, yielding emergent features in
the resulting anatomies, physiologies, and behaviors.
The best example of legitimate
gene selection may be provided by a phenomenon known as "selfish DNA." Some
genes can make copies of themselves, and these copies may then move to other
locations among the chromosomes (so-called transposons, or "jumping genes").
This process constitutes positive selection for these genes at their own
level, since the process augments the number of copies of these genes among
the chromosomes of an individual—just as ordinary Darwinian selection
on organisms increases the number of offspring of favored individuals within
a population. But the organism need not "notice" as the copies of selfish
DNA increase, for these additional copies are often without function. In
fact, gene selection can be most effective when organisms do not "notice"
the increase—for if the increase of genetic copies impedes the organism
in any way, negative selection from the ordinary Darwinian level of
organisms may bring the process to a halt.
Cronin gives away her biases (and
any pretense of balanced argument) in such rhetorical flourishes. Why is
genetic drift an "explanatory clutch"? The phenomenon is perfectly
respectable and powerful, if not entirely applicable to Cepaea. In
other giveaway passages, we learn that "nonadaptive explanations cannot be
treated as other than a last resort." (Why? They are also permissible in
theory and robustly present in nature.) When Darwin strays from the
selectionist straight and narrow, his words "need to be interpreted more
generously" so that our chief icon may continue to spearhead the strict and
uncompromising version of his theory. (Again, why? Darwin was not a pure
selectionist.) When Wallace departs from selectionist principles, "he has a
lot to answer for." (But what is his sin?) We learn that Darwin "lets us
down" when he proposes group selection for human moral conduct (a proper
exception to his general orthodoxy of selection on organisms). When, among
the founders of population geneticists, Fisher and Haldane make some invalid
statements about group selection, they are excused because they were true
Darwinians at heart. But when Sewall Wright, third member of the trinity,
speaks in the same vein, he is not exonerated, for he truly believed in a
form of group selection. Cronin writes: "I did not include the other major
founding father of modern Darwinism, Sewall Wright, among the honorable
exceptions." (Now Cronin, Dawkins, and me, bundled all together, couldn't
fill Sewall Wright's left pocket insofar as intellectual power is
concerned—and it is simply unseemly for any of us to speak of his views
in terms of honor or dishonor.)
See S.J. Gould and D.S. Woodruff
in "History as a Cause of Area Effects: An Illustration from Cerion
on Great Inagua, Bahamas," Biological Journal of the Linnean Society,
1990, Vol. 40, pp. 67–98.
[ Stephen Jay Gould, "The Confusion over Evolution,"
The New York Review of Books,
November 19, 1992, pp. 47-54. ]
Home Page |
Further Reading |
Site Map |
Send Feedback
