Darwin's Dangerous Idea (22 page)

"biological laws" and what is possible, impossible, and necessary in the version of those varieties of possibility and necessity, and then develop our world. Recall that we needed to get clear about these issues because, if we restricted notion of biological possibility in terms of it. The law of large are to explain the way things
are,
it must be against a background of how numbers and the law of gravity, for instance, are both deemed to hold things
might
have been, or
must
be, or
couldn't
be. We can now define a unreservedly and timelessly over the space. Clamping physical law lets us say restricted concept of biological possibility:

flat out, for instance, that all the different genomes are physically possible—

because chemistry says they are all stable, if encountered.

x
is biologically possible if and only if
x
is an instantiation of an accessible Keeping logic and physics and chemistry clamped, we could choose a genome or a feature of its phenotypic products.

different starting point. We could choose some moment on Earth five billion years ago, and consider what was biologically possible then. Not much, Accessible from where? By what processes? Ah, there's the rub. We have to because before tigers could become possible (on Earth), eukaryotes, and then specify a starting point in the Library of Mendel, and a means of "travel."

plants producing atmospheric oxygen in large quantities, and many other Suppose we were to start where we are today. Then we will be talking, first, things had to become actual. With hindsight, we can say that tigers were in about what is possible
now
—that is to say, in the near future, using whatever fact possible all along, if distant and extremely improbable. One of the virtues means of travel are currently available to us. We count as possible all
actual
of this way of thinking of possibility is that it joins forces with probability, contemporary species and all their features—including the features they have thus permitting us to trade in flat all-or-nothing claims about possibility for in virtue of their relations with other species and their features— plus claims about relative distance, which is what matters for most purposes. (The anything that can be obtained by traveling from that broad front either just "in all-or-nothing claims of biological possibility were all but impossible [hmm, the course of nature"—without human manipulation—or with the help of that word again] to adjudicate, so this is no loss.) As we saw in our such artificial cranes as the techniques of traditional animal-breeding (and, exploration of the Library of Babel, it doesn't make much difference what our for that matter, surgery), or via the fancy new vehicles of genetic engineering.

verdict is about whether it is "possible in principle" to find some particular After all, we human beings and all our tricks are just another product of the volume in that Vast space. What matters is what is practically possible, in one contemporary biosphere. Thus it is biologically possible for you to have a or another sense of "practical"—take your pick.

fresh turkey dinner on Christmas Day, 2001, if and only if at least one This is certainly not a standard definition of possibility, or even a standard instantiated turkey genome has produced the requisite phenotypic effects in
sort
of definition of possibility. The idea that some things might be "more time for dinner. It is biologically possible for you to ride a pter-anodon before possible" than others (or more possible from over here than from over there) you die if and only
if Jurassic Park-ish
technology permits
that
sort of is at odds with one standard understanding of the term, and some genome to get expressed in time.

philosophical critics might say that this is simply not a definition of
possi-No matter how we set these "travel" parameters, the resulting notion of biological possibility will have an important property: some things will be 120 THE POSSIBLE AND THE ACTUAL

Possibility Naturalized
121

bility,
whatever it is. Some other philosophers have defended views of whether we could count nonactual possible objects. One of the virtues of the comparative possibility (see especially Lewis 1986, pp. 10ff.), but I don't proposed treatment of biological possibility is that, thanks to its "arbitrary"

want to fight over it. If this is not an account of possibility, so be it. It is, then, formal system—the system arbitrarily imposed on us by nature, at least in a proposed
replacement
for a definition of possibility. Perhaps after all we our neck of the woods—we can count the different nonactual possible don't need the concept of biological possibility (with its required all-or-genomes; they are Vast but finite in number, and no two are exactly alike.

nothing application) for any serious investigative purpose. Perhaps degree of (By definition, genomes are distinct if they fail to share a nucleotide at any accessibility in the space of the Library of Mendel is all we need, and is in one of several billion loci.) In what sense are the nonactual genomes
really
fact a better concept than any all-or-nothing version could be. It would be possible? Only in this sense: if they were formed, they'd be stable. But nice, for instance, to have some way of
ranking
the following in terms of whether or not any conspiracy of events could lead to their being formed is biological possibility: ten-pound tomatoes, aquatic dogs, flying horses, flying another matter, to be addressed in terms of accessibility from one location or trees.

another. Most of the genomes in this set of stable possibilities will
never
be That will not be enough to satisfy many philosophers, and their objections formed, we can be sure, since the heat death of the universe will overtake the are serious. Briefly considering them will at least make it clearer what I am building process before it has made a sizable dent in the space.

claiming and what I am not claiming. First of all, isn't there something Two other objections to this proposal about biological possibility cry out to viciously circular about defining possibility in terms of
accessibility?

be heard. First, isn't it outrageously "gene-centered," in anchoring
all
(Doesn't the latter term just reintroduce the former in its suffix, and still considerations of biological possibility to the accessibility of one genome or undefined?) Well, not quite. It does leave some definitely unfinished busi-another in the Library of Mendel? Our proposed treatment of biological ness, which I will simply acknowledge before moving on. We have supposed possibility flatly ignores (and hence implicitly rules impossible) "creatures"

that we are holding some concept or other of
physical
possibility clamped for that are not end points of some branch of the Tree of Life that has already the time being; our idea of accessibility presupposes that this physical taken us as far as we are today. But that just
is
the grand unification of possibility, whatever it is, leaves us
some
elbow room—some openness of biology that Darwin discovered! Unless you harbor fantasies about sponta-pathways (not just a single pathway) in the space. In other words, we are neous creation of new life forms by "Special Creation" or (the philosophers'

taking on the assumption that
nothing stops us
from going down any of the secular version) "Cosmic Coincidence," you accept that every feature of the pathways that are open so far as physics is concerned.11

biosphere is one fruit or another of the Tree of Life (or, if not
our
Tree of Quine's questions (at the head of this chapter) invited us to worry about Life, some other Tree of Life, with its own accessibility relations). No man is an island, John Donne proclaims, and Charles Darwin adds that neither is any clam or tulip—every
possible
living thing is connected by isthmuses of descent to all other living things. Notice that this doctrine rules
in
whatever 11. This idea of elbow room is something we need to presuppose in any case, for it is the marvels technology can produce in the future, provided—as we have already minimal denial of actualism, the doctrine that only the actual is possible. David Hume, in noted—that technologists themselves, and their tools and methods, are firmly
A Treatise of Human Nature
(1739), spoke of "a certain looseness" we want
to
exist in our world. This is the looseness that prevents the possible from shrinking tightly around located on the Tree of Life. It is a small further step to rule in life forms from the actual. This looseness is presupposed by
any
use of the word "can"—a word we can outer space, provided they, too, are the products of a Tree of Life rooted, as hardly do without! Some people have thought that, if determinism were true, actualism ours is, in some nonmiraculous physical ground. (This topic will be explored would be true—or, to turn it around, if actualism is
false, in
determinism must be true—

in chapter 7.)

but this is highly dubious. The implied argument against determinism would be discon-Second, why should we treat biological possibility so differently from certingly simple: this oxygen atom has valence 2; therefore, it can unite with two hydrogen atoms to form a molecule of water (it
can
right now, whether or not it does); physical possibility? If we assume that "laws of physics" fix the limits of therefore, something is possible that isn't actual, so determinism is false. There are physical possibility, why shouldn't we attempt to define biological possibility impressive arguments from physics that lead to the conclusion that determinism is false—

in terms of "laws of biology"? ( We will turn to an examination of physical but this isn't one of them. I am prepared to assume that actualism is false (and that this laws and physical necessity in chapter 7, but in the meantime, the difference assumption is independent of the determinism/indeterminism question), even if I can't appears large.) Many biologists and philosophers of science have maintained claim to prove it, if only because the alternative would be to give up and go play golf or something. But for a somewhat fuller discussion of actualism, see my book
Elbow Room
that there are biological laws. Doesn't the proposed definition rule them out?

( 1984 ), especially ch. 6, "Could Have Done Otherwise," from which material in this note Or does it declare them superfluous? It doesn't rule them out. It permits is drawn. See also David Lewis' (1986, ch. 17 ) concurring opinion, about the related issue someone to argue for the dominion of some law of biology over the of the irrelevance of the issue of indeterminism to our sense that the future is "open."

Possibility Naturalized
123

122 THE POSSIBLE AND THE ACTUAL

adopted, it resulted in many millions of typewriters and ... the social cost space of the Library of Mendel, but it does put a difficult burden of proof on of change ... mounted with the vested interest created by the fact that so anyone who thinks that there are laws of biology
over and above
the laws of many fingers now knew how to follow the QWERTY keyboard. QWERTY

mathematics and physics. Consider the fate of "Dollo's Law," for instance.

has stayed on despite the existence of other, more "rational" systems.

[Papert 1980, p. 33.]12

'Dollo's Law' states that evolution is irreversible....[But] There is no reason why general trends in evolution shouldn't be reversed. If there is a trend The imperious restrictions we encounter inside the Library of Mendel may towards large antlers for a while in evolution, there can easily be a subse-look like universal laws of nature from our myopic perspective, but from a quent trend towards smaller antlers again. Dollo's Law is really just a state-different perspective they may appear to count as merely local conditions, ment about the statistical improbability of following exactly the same with historical explanations.13 If so, then a restricted concept of biological evolutionary trajectory twice (or indeed any
particular
trajectory), in possibility is the sort we want; the ideal of a universal concept of biological either direction. A single mutational step can easily be reversed. But for larger numbers of mutational steps... the mathematical space of all possible possibility will be misguided. But as I have already allowed, this does not trajectories is so vast that the chance of two trajectories ever arriving at the rule out biological laws; it merely sets the burden of proof for those who want same point becomes vanishingly small __ There is nothing mysterious or to propose any. And in the meantime, it gives us a frame-work for describing mystical about Dollo's Law, nor is it something that we go out and 'test' in large and important classes of regularity we discover in the patterns in
our
nature. It follows simply from the elementary laws of probability. [Dawkins biosphere.

1986a, p. 94.]

There is no shortage of candidates for the role of "irreducible biological CHAPTER 5:
Biological possibility is best seen in terms of accessibility (from
law." For instance, many have argued that there are "developmental laws" or
some stipulated location) in the Library of Mendel,
the
logical space of all

"laws of form" that constrain the relation between genotype and pheno-type.

genomes. This concept of possibility treats the connectedness of the Tree of
In due course we will consider their status, but already we can locate at least
Life as a fundamental feature of biology, while leaving it open that there may
some of the most salient constraints on biological possibility as not "laws of
also be biological laws that will also constrain accessibility.

biology" but just inescapable features of the geometry of design space, like Dollo's Law (or the Hardy-Weinberg Law of gene frequency, which is CHAPTER 6:
The R and D done by natural selection in the course of creating
another application of probability theory, pure and simple).

actual trajectories in the Vast space of possibilities can be measured to some
Take the case of the horned birds. As Maynard Smith notes, there aren't
extent. Among the important features of this search space are the solutions to
any, and we don't know why. Might it be because they are ruled out by a
problems that are perennially attractive and hence predictable, like forced
biological
law?
Are horned birds flat impossible? Would they have to be
moves in chess. This explains some of our intuitions about originality,
inviable in any and all possible environments, or is there simply no path at all
discovery, and invention, and also clarifies the logic of Darwinian inference