Darwin's Dangerous Idea (40 page)

Read Darwin's Dangerous Idea Online

Authors: Daniel C. Dennett

BOOK: Darwin's Dangerous Idea
10.97Mb size Format: txt, pdf, ePub

common with the strategy known to engineers as
reverse engineering
( Den-There is a book on how to detect fake antiques (which is also, inevitably, a book on nett 1990b). When Raytheon wants to make an electronic widget to compete how to
make
fake antiques ) which offers this sly advice to those who want to fool the with General Electrics widget, they buy several of GE's widgets and proceed

"expert" buyer: once you have completed your table or whatever (having utilized all the usual means of simulating age and wear) take a modern electric drill and drill a to analyze them: that's reverse engineering. They run them, benchmark them, hole right through the piece in some conspicuous but puzzling place. The would-be X-ray them, take them apart, and subject every part of them to interpretive buyer will argue: no one would drill such a disfiguring hole without a reason (it can't analysis: Why did GE make these wires so heavy? What are these extra ROM

be supposed to look "authentic" in any way ) so it must have served a purpose, which registers for? Is this a double layer of insulation, and, if so, why did they means this table must have been in use in someone's home; since it was in use in bother with it? Notice that the reigning assumption is that all these "why"

someone's home, it was not made expressly for sale in this antique shop ___Therefore it is authentic. Even if this "conclusion" left room for lingering doubts, the buyer will questions have answers. Everything has a
raison d'etre;
GE did nothing in be so preoccupied dreaming up uses for that hole it will be months before the doubts vain.

can surface.

Of course, if the wisdom of the reverse engineers includes a healthy It has been claimed, with what plausibility I do not know, that Bobby Fischer has used the helping of self-knowledge, they will recognize that this default assumption of same strategy to defeat opponents in chess, especially when time is running out: make a optimality is too strong: sometimes engineers put stupid, pointless things in deliberately "off-the-wall" move and watch your opponent waste precious time trying to their designs, sometimes they forget to remove things that no longer have make sense of it.

214 BIOLOGY IS ENGINEERING

Artifact Hermeneutics, or Reverse Engineering

215

pretty-good version of the machine on hand, either an earlier model, or a 1969, p. 126). This is the very engine of Darwinian evolution. If, however,

"mockup" or scale model that we have built. We examine it carefully, and try we look back with tunnel vision at your family tree, we will find many out various alterations: "If we just bend this jaw up a little bit like so, and different organisms, with a wide variety of strengths and weaknesses, but, move this zipper-bit over a tad like so, it would work even better." But that is curiously enough,
their
weaknesses never led a single one of them to a not the way evolution works. This comes out especially clearly at the premature demise! So it looks from this angle as if evolution can't explain molecular level. A particular molecule is the shape it is, and won't tolerate even a
single feature
that you inherited from your ancestors! Suppose we much bending or reshaping. What evolution has to do when it improves look back at the fan-out of your ancestors. Notice first that eventually it stops molecular design is to make
another
molecule—one that is almost like the fanning out and begins to double up; you share
multiple
ancestors with one that doesn't work very well—and simply discard the old one.

everybody else alive today, and are multiply related to many of your own People are advised
never
to switch horses in midstream, but evolution ancestors. When we look at the whole tree over time, we see that the later,
always
switches horses. It can't
fix
anything, except by selecting and dis-more recent ancestors have improvements that the earlier ones lacked, but all carding. So in every evolutionary process—and hence in every true evolu-the crucial events—all the selection events—happen offstage: not a single tionary explanation—there is always a faint but disconcerting odor of one of your ancestors, all the way back to the bacteria, succumbed to something dicey. I will call this phenomenon
bait-and-switch,
after the shady predation before reproducing, or lost out to the competition for a mate.

practice of attracting customers by advertising something at a bargain price Of course, evolution
does
explain all the features that you inherited from and then, when you've lured them to the store, trying to sell them a substitute.

your ancestors, but not by explaining why
you
are lucky enough to have Unlike that practice, evolutionary bait-and-switch is not really nefarious; it them. It explains why today's winners have the features they do, but not why just seems to be, because it doesn't explain what at first you thought you
these individuals
have the features they do.17 Consider: You order a new car, wanted explained. It subtly changes the topic.

and specify that it be green. On the appointed day, you go to the dealership We saw the ominous shadow of bait-and-switch in its purest form in and there it sits, green and new. Which is the right question to ask: "Why is chapter 2, in the weird wager that I can produce somebody who wins ten this car green?" or "Why is this (green) car here?" (In later chapters we will consecutive coin-tosses without a loss. I don't know in advance who that look further at the implications of bait-and-switch.) somebody is going to be; I just know that the mantle will pass—has to pass, The second important difference between the processes—and hence the as a matter of algorithmic necessity—to
somebody or other
so long as I products—of natural selection and human engineering concerns the feature of execute the algorithm. If you overlook this possibility and take my sucker bet, natural selection that strikes many people as most paradoxical: its utter lack it is because you are too used to the human practice of tracking individuals of foresight. When human engineers design something (forward engineering), and building projects around identified individuals and their future prospects.

they must guard against a notorious problem: unforeseen side effects. When And if the winner of the tournament thinks there has to be an explanation of two or more systems, well designed in isolation, are put into a supersystem, why
he
won, he is mistaken: there is no reason at all why
he
won; there is this often produces interactions that were not only not part of the intended only a very good reason why
somebody
won. But, being human, the winner design, but positively harmful; the activity of one system inadvertently will no doubt think there
ought
to be a reason why he won: "If your clobbers the activity of the other. The only practical way to guard against

'evolutionary account' can't explain it, then you are leaving out something unforeseen side effects, since by their very nature they are unforeseeable by important!" To which the evolutionist must calmly reply: "Sir, I know that is those whose gaze is perforce restricted to just one of the subsystems being what you came in here wanting, but let me try to interest you in something a designed, is to design all subsystems to have relatively impenetrable little more affordable, a little less presumptuous, a little more defensible."

boundaries that coincide with the epistemic boundaries of their creators.

Has it ever occurred to you how lucky you are to be alive? More than 99

Human engineers typically attempt to insulate subsystems from each other, percent of all the creatures that have ever lived have died without progeny, and insist on an overall design in which each subsystem has a single, well-but not a single one of your ancestors falls into that group! What a royal defined function within the whole.

lineage of winners you come from! (Of course, the same thing is true of every The set of supersystems having this fundamental abstract architecture is barnacle, every blade of grass, every housefly.) But it's even eerier than that.

We have learned, have we not, that evolution works by weeding out the unfit?

Thanks to their design defects, these losers have a "pathetic but praiseworthy 17. "But this is not to explain why, e.g., contractile vacuoles occur in certain protozoans; tendency to die before reproducing their kind" (Quine it is to explain why the sort of protozoan incorporating contractile vacuoles occurs"

(Cummins 1975, in Sober 1984b, pp. 394-95).

216 BIOLOGY IS ENGINEERING

Artifact Hermeneutics, or Reverse Engineering
217

vast and interesting, of course, but it does not include very many of the process by which a genotype gets "expressed" in a fully formed phenotype.

systems designed by natural selection! The process of evolution is notori-And the
design
processes that shaped the developmental processes that shape ously lacking in foresight. Since it has no foresight at all, unforeseen or the "finished products" were largely inaccessible to the sorts of intrusive unforeseeable side effects are nothing to it; it proceeds, unlike human en-observation and manipulation that good science (or good reverse gineers, via the profligate process of creating vast numbers of relatively engineering) thrives on. You could look at the sketchy historical record, and uninsulated designs, most of which are hopelessly flawed because of self-run it in fast-forward (like "elapsed-time" photography of plants growing, defeating side effects, but a few of which, by dumb luck, are spared that weather developing, etc.—always a nifty way to make the patterns visible ), ignominious fate. Moreover, this apparently inefficient design philosophy but you couldn't "rewind the tape" and run variations on the initial conditions.

carries a tremendous bonus that is relatively unavailable to the more efficient, Now, thanks to computer simulations, it is possible to
study
the hypotheses top-down process of human engineers: thanks to its having no bias against about the design process that have always lain at the heart of the Darwinian unexamined side effects, it can take advantage of the rare cases where vision. Not surprisingly, they turn out to be more complicated, and beneficial
serendipitous
side effects emerge. Sometimes, that is, designs themselves more intricately designed, than we had thought.

emerge in which systems interact to produce more than was aimed at. In Once the processes of R and D and construction begin to come into focus, particular (but not exclusively), one gets elements in such systems that have we can see that an affliction of shortsightedness that has often misled multiple functions.

interpreters of human artifacts has multiple parallels in biology. When we Elements with multiple functions are not unknown to human engineering, engage in artifact hermeneutics, trying to decipher the design of items of course, but their relative rarity is signaled by the delight we are apt to feel uncovered by archeologists, or trying to recover a proper interpretation of the when we encounter a new one. (A favorite of mine is found in the Diconix ancient monuments that we have grown up with, there is a tendency to portable printer. This optimally tiny printer runs on largish rechargeable overlook the possibility that some of the features that puzzle us have no batteries, which have to be stored somewhere; they fit snugly inside the function at all in the finished product, but played a crucial functional role in platen, or roller.) On reflection, we can see that such instances of multiple the process that created the product.

function are epistemically accessible to engineers under various salubrious Cathedrals, for instance, have many curious architectural features that have circumstances, but we can also see that by and large such solutions to design provoked functional fantasies and fierce debates among art historians. The problems must be exceptions against a background of strict isolation of functions of some of these features are fairly obvious. The many "vises" or functional elements. In biology, we encounter quite crisp anatomical isolation circular stairways that twist their way up inside the piers and walls are useful of functions (the kidney is entirely distinct from the heart; nerves and blood ways for custodians to gain access to remote parts of the building: to the roof, vessels are separate conduits strung through the body, etc.), and without this say, and to the space between the vault and the roof where the machinery is readily discernible isolation, reverse engineering in biology would no doubt hidden that lowers the chandeliers to the floor so that candles may be be humanly impossible. But we also see superim-position of functions that replaced. But many of the vises would be there even if no such later access apparently goes "all the way down." It is very, very hard to think about had been anticipated by the builders; it was simply the best, or maybe the entities in which the elements have multiple overlapping roles in only, way for the builders to get the building crew and materials where they superimposed subsystems, and, moreover, in which some of the most salient needed to be during construction. Other passageways leading nowhere inside effects observable in the interaction of these elements may not be functions at the walls are probably there in order to get fresh air into the interior of the all. but merely by-products of the multiple functions being served.18

walls (Fitchen 1961). Medieval mortar took a long time— years, in some Until recently, biologists who wanted to be reverse engineers had to cases—to cure, and as it cured it shrank, so care was taken to keep wall concentrate on figuring out the designed features of "finished products"—

thickness minimized so that distortion was minimized as the building cured.

organisms. These they could collect by the hundreds or thousands, study the (Thus those passageways have a similar function to the heat-dissipation "fins"

variations thereof, take apart, and manipulate
ad lib.
It was much more on automobile-engine housings, except that their functions lapsed once the difficult to get any epistemic purchase on the
developmental
or
building
buildings reached maturity.)

Other books

Blackbirds by Chuck Wendig
I Heard That Song Before by Mary Higgins Clark
Dangerous to Know by Dawn Ryder
For Camelot's Honor by Sarah Zettel
Great Sex, Naturally by Steelsmith, Laurie
Paul McCartney by Philip Norman
Bound to Moonlight by Nina Croft