Minds & Machines

DOI 10.1007/s11023-016-9390-5

BOOK REVIEW

Kim Sterelny: The Evolved Apprentice: How Evolution

David Cole1

Received: 13 January 2016 / Accepted: 13 January 2016

One species rules the world—homo sapiens. Our nearest kin species are not even
close runners-up; they are losers. The most optimistic estimate of chimpanzee
numbers world-wide is under 300,000, and for gorillas it is around 1/3 of that. How
did this happen? The main differences between sapiens and our nearest surviving
kin species are not size, basic body configuration, or fecundity—though there are
differences, and these may contribute to our relative success. However our main
distinctive features appear to be in cognition and social behavior. What evolutionary
breakthrough allowed sapiens to dominate the planet?
This rise of sapiens has occurred in a remarkably short time. Seventy-five
thousand years ago sapiens was nearly extinct; now sapiens is the great extinctifier.
There are many competing theories as to what catapulted sapiens to the top in such
relatively short time. One theory is that there were specific breakthrough
technologies—e.g. cooked food. Others hold that it was social behavior, such as
grandmother contributions to child-rearing. On these accounts, the difference is
largely cultural, a discovery that can be spread by culture. Others hold that what
gave victory to humans in the evolutionary game of thrones was a very specific
genetic mutation, especially the mutation needed for language (as Steven Pinker
argues in The Language Instinct), or a mutation affording us social cognition. These
theories, whether attributing the successful innovation to culture or mutation, see the
change as dramatically sudden, a Big Bang, as Pinker calls it.
Kim Sterelny argues there was no such Big Bang, no sudden technological
breakthrough, no mutant first orator whose descendents displaced all linguistically
dumb con-specifics. Rather there were small co-evolved changes in the way sapiens

& David Cole

123
 D. Cole

fostered learning that eventually led to Harvard and planet domination. Sapiens rule
begins with whimpers, not a bang.
Sterelny, at the Australian National University, is a leading philosopher of
biology. He shifted from early books on language and philosophy of mind to focus
generally on biology and evolution in the 1990s, with the co-authored Sex and
Death, published in 1999. Other major publications include a 2001 collection, The
Evolution of Agency, Dawkins versus Gould (2001), Thought in a Hostile World:
The Evolution of Human Cognition (2003), and more, including a collection
Cooperation and Its Evolution (2013), published the year after The Evolved
Apprentice.
The Evolved Apprentice (2012), based on his 2008 Paris lectures in the Jean
Nicod series, is an extended argument in favor of a relatively gradualist account of
human cognition focused on cooperation. Much is via negativa—arguing against
rival accounts. The evolution of cooperation has a core prisoners’ dilemma problem:
for each member of any group there is always a prima facie advantage to obtaining
the fruits of others’ labor while not paying the cost of contributing. For cooperation
to evolve, there must be some way of getting around this Free Rider problem. With
information sharing, there is the additional serious problem of deception. A free
rider merely doesn’t contribute; a deceiver uses information to harm others (e.g. a
chimp might issue an alarm call to distract others from a tasty find). Hence much of
Sterelny’s discussion addresses these issues.
In the course of his book Sterelny discusses the work of many, including Dan
Sperber, Marc Hauser, Sarah Hrdy, Pete Richerson, James O’Connell, Stan Bowles,
Herbert Gintis, Robert Frank, Kristin Hawkes, Robert Boyd—and a host of others.
In general, he treats this work as providing partial insights into a complex story,
though sometimes he argues against accounts, e.g. against extrapolating from
chimpanzee violence and the archeological record as evidence for a somber
portrayal of our early ancestors as ultra-violent.
It seems obvious that cooperative activity is an eminent, perhaps pre-eminent,
human trait and basis for our biological success. There are no supermen; the
transformation of the planet has required layers upon layers of coordinated
cooperative activity. Other social species also cooperate and have division of labor,
but most are mono-maniacs limited to a very short list of cooperative activities.
Humans, by contrast, swarm this way in one millennium to build pyramids, that way
to build roads and aqueducts the next, we build ships for awhile, and then ever
larger swarms build trains, planes and automobiles, and finally on-line virtual
communities. And we destroy collectively as well, attacking each other in war, and
harvesting other species for lunch. As Ambrose Bierce famously defined us, man is
an animal whose ‘‘chief occupation is extermination of other animals and his own
species, which, however, multiplies with such insistent rapidity as to infest the
whole habitable earth and Canada.’’
Sterelny holds that the key to this infesting fitness is social learning and niche
construction (p. 174). These developed by gradual coevolution ‘‘driven by positive
feedback loops rather than by a key innovation’’ (p. 29). An important development
is extended childhood and apprentice learning (p. 34ff). One of the niches humans
build is an information-rich learning niche. This is relatively low cost but affords

123
Kim Sterelny: The Evolved Apprentice: How Evolution Made...

valuable knowledge flow. But this process is slow: chapter 3 treats of the long
150,000 year period of apparently modern humans before the displacement of
Neanderthals and the advent of agriculture. Sterelny speculates that Neanderthals
were caught in a fitness trap, reliant on ambush hunting of game that disappeared
(pp. 62–71). We were not so limited.
Sterelny has an extensive discussion in chapter 4 of two prominent forms of
human cooperative activity, child rearing and hunting. While conceding that
cooperative reproduction is important, Sterelny is a critic of Hrdy and Wrangham
and their elevation of helpful grandmothers to breakthrough status. Hunting too is
typically cooperative, and a male specialty. Sterelny argues against accounts that
denigrate the importance of hunting—hunting is time-consuming but economically
important. Anthropological evidence that hunting by contemporary hunter-gatherers
is low yield does not, he suggests, bear on the evolutionary situation in which game
was likely much more plentiful. Sterelny also argues that hunting is not (just) a form
of costly signaling in which males take on gratuitous challenges—the Peacock’s tail
is the paradigm—to signal fitness to would-be mates. Hunting is a group activity,
and the rewards are shared. Thus the competitive costly-signaling model does not fit
well. However recent analysis of frog croaks—the less fit still advertise their
relative puniness with higher-pitched croaks, apparently to signal that they are not
the very least fit—may raise problems for Sterelny’s account of the function of male
hunting. Hunters, like non-alpha motorcycle gang members, may signal that they
are man enough.
Solutions to the Free Rider and trust problems are central topics of chapters 5 and
6. The solution lies in part in specific genetic traits, including those that endow us
with emotions. Emotions dispose to behavior, but are often not rational—
vengefulness, e.g., impels us to disastrous retribution. But emotions also commit
us to cooperation—we can be steadfastly loyal, or in love. We feel guilt and remorse
and humiliation when we fail to meet social expectations. Further, we advertise our
emotions. On the face of it, this is paradoxical, a considerable disadvantage in
dealing with others. But clearly it is very useful in signaling to wary potential
cooperators. Shifty eyes cost us poker games but this disposition is elsewhere an
advantage: if others can know whenever we are dishonest because we must
advertise our dishonesty, in the absence of these involuntary signals we thereby
make ourselves trust-worthy and desirable as cooperators.
Information sharing, the central form of human cooperation in Sterelnys account,
gets its own chapter. Information is valuable, but consumers face the two serious
problems of free-riders and deception. Free-riders consume but do not provide
information, but deception is the more serious problem. Sterelny discusses Dan
Sperber’s work, but argues Sperber’s analysis is too Machiavellian. There are not
always incentives to deceive, and there are other important uses of meta-
representation (representing representations, such as our own or other’s claims
and beliefs) besides deception detection. We often need to assess claims and beliefs
(including our own) for unintended error, not just deception. There are many
channels of information sharing where deception is not a problem, including
information leakage, as when a hunter displays through his successful practice how
to track deer, or when Sony displayed to all how to make compact portable cassette

123
 D. Cole

recorders by placing them on store shelves. Language is just one of many sources of
information. We are inveterate gossips and monitors of slackers and the unreliable.
Sterelny concludes that information consumption sometimes but not always turns on
trust, and that the cognitive resources we have, including folk epistemology, for
assessing trust-worthiness are general purpose.
A penultimate chapter turns to social norms and rules, and the important role they
play in shaping cooperation. Sterelny argues, against moral nativists, that we can
articulate—and change—moral principles. Contra nativism, there is considerable
variation in moral codes and substantial change over time. Further, we attempt to
convince others to hold moral principles, whereas the rules of grammar, innate or
acquired, are opaque to most users of language. By contrast to nativism, Sterelny
offers a model of humans as moral apprentices. Adults provide copious moralizing.
Children learn by a form of abstract pattern recognition, and are given explicit moral
training, including stories, adult models, and morality play.
The last chapter discusses group selection in the context of war and the previous
long discussion of forms of cooperation. Sterelny argues that group selection is
plausible, and enhanced by aspects of culture that he has emphasized throughout the
book: ‘‘Group selection is made more powerful by any process that makes groups
internally homogeneous, and different from other groups. The combination of
information pooling and cross-generational learning has just these effects’’ (p. 177).
However group selection is not essential to explain the evolution of cooperation: the
potential individual gains from cooperation and reciprocation are sufficient to
enhance the fitness of cooperators. Sterelny ends the book by arguing contra Bowles
and others that there is not strong evidence for human violence in the late
Pleistocene period. However violence did likely increase as agriculture displaced
hunter-gatherer life in the Holocene. The sharply increased inequality raises an
unsolved problem of how such societies were (and are—it is a timely topic) stable.
It also raises questions that aren’t addressed here: did the radically changed
conditions, unequal distribution of resources, and required skill sets of the Holocene
importantly affect human evolution?
In this book Sterelny makes a compelling case for the importance of cooperation,
learning, and information transfer and culture in the evolution of humans. Teaching
and learning are what make humans unique—a thesis that academics may find
congenial. A remaining problem for Stelelnys Fabian approach is how minimal the
underlying biological changes might have been to provide the infrastructure for
these changes in social behavior. For example, it is fairly clear that complex
grammar processing, essential to human linguistic communication and the resulting
information flow that looms large in Sterelnys account, does require that genes
produce specific neural machinery, as well as dispositions to produce and consume
the verbal information thereby made possible. Furthermore, if cooperation requires
meta-representation, and meta-representation is all or nothing and also requires
extensive neural machinery, then specific key mutations may have been necessary to
support the social behaviors that are central to Sterelny’s account. Thus the co-
evolution picture and niche-construction story may be correct but still require
breakthrough mutations. At this point we certainly don’t know enough about what
exactly all those Brodmann areas do. But perhaps a breakthrough is eminent.

123