AS well as his Nobel prizewinning work on the structure of proteins, Max
Perutz was a long-time contributor to both The New York Review of Books
and The London Review of Books. His essays spanned the spectrum of
science, scientists and humanity. This volume is, with one or two exceptions, a
collection of reprints of the best of Perutz’s contributions to these
journals.

Peter Medawar once said that, “Science at all levels of endeavour is a
passionate enterprise and the pursuit of natural knowledge a sortie into the
unknown.” Perutz demonstrates this admirably in these essays, which include:
detective stories, tales of conflict and battle, a Nobel laureate’s geriatric
illusion about a cure for cancer, a woman’s love affair with crystals, an attack
on social relativists, an account of phantom perils threatening to poison us
all, the effect of a gruesome fascination with poison gas and much, much more.
Along the way we encounter Linus Pauling, Peter Medawar, Leo Szilard,
François Jacob, Fritz Haber and Lise Meitner, to name but a few. Perutz
introduces the giants of 20th-century science gracefully: writing with the
lucidity and precision that he brought to his own work on proteins. There is
something for everyone here.

THEORETICAL computer science is overflowing with impossibility results,
ranging from the impossibility of an algorithm for telling when a computer
program will stop (the halting problem) to the impossibility of deciding the
truth or falsity of a general Boolean expression (the SAT problem). Such results
are the outer limits of complexity—in some sense problems of infinite
complexity.

None of this bothers practising numerical analysts. They want to know just
how much does it cost in computational resources to solve numerically a
particular problem to a prescribed degree of accuracy, given a specific amount
of information? Information-based complexity (IBC) is a theory devoted to
answering this and Joseph Traub is a founder . So this series of lectures
summarising the current state of the art is noteworthy.

Speaking more formally, IBC is the branch of computational complexity that
studies problems for which the information available is partial, contaminated or
comes at a price. In fewer than 140 pages, Traub and Arthur Werschulz cover an
astonishing amount of material: numerical integration, Fredholm integral
equations, nonlinear optimisation, linear programming, theorem verification and
path integration. For good measure, they throw in more philosophical issues,
such as the role impossibility theorems play in limiting what we can know about
the real world, along with a discussion of how the model of computation one
chooses affects what is and is not “computable”.

Clearly written, filled with interesting examples, important theorems and
tantalising conjectures, this is destined to be a classic.

KARL MARX once said: “Philosophers have sought to interpret the world. The
point, however, is to change it.” Marx got it the wrong way around. Politicians
have sought to change the world; the real task is to explain it. Paul Ormerod’s
Butterfly Economics is a big step forward in helping us interpret
(read, “understand”) the world of “real” economics.

Traditionally, economic theory is founded on the notion that supply matches
demand through the medium of prices. It is an equilibrium-centred view of
economic processes. The economic machine is continually seeking to reach
nirvana, in which the “right” price is set so that all buyers and sellers are
happy. Of course, this is a pure fiction: no real economic system ever gets near
such a state. The situation is much worse than this.

In real economies, the buyers and sellers are wildly heterogeneous in their
resources: what they have to sell, what they want to buy, the level of risk they
are willing to assume and so forth. Real economies consist of millions upon
millions of these agents, each with their own goals and resources, interacting
with each other through institutional and social mechanisms to enact their
economic decisions. These individual decisions and their consequent
interactions, in turn, generate higher-level, aggregate quantities such as price
changes, business cycles and many other economic phenomena of the sort we follow
in the newspapers every day. In the jargon of systems theorists, such phenomena
are “emergent” properties of the economic system.

Ormerod’s fascinating and informative book tells the tale of an entirely new
view of economic life, based on the paradigms coming out of the complex system
theory community. Economists such as Brian Arthur of the Santa Fe Institute and
Paul Krugman of Stanford University are creating new theoretical frameworks for
considering the behaviour of real economies in the brave, new economic worlds in
which information and knowledge are the prime commodities on offer, not guns and
steel.

The characterising features of the “new economy” are things such as positive
feedback, in which small initial advantages lead to firms like Microsoft taking
over an entire industry. Butterfly Economics tells many fascinating
stories to show how a nonequilibrium-centred view of economic processes leads to
theories that are far more in harmony with the facts than the classical
perspectives. Ormerod says, for example, that much of the control that
governments think they exercise over the economy and society is an illusion.
Rather, these systems are inherently difficult to predict and manage. Hence, the
title “butterfly economics”, reflecting the famous metaphor from chaos theory of
a butterfly beating its wings in Brazil today causing a tornado in Kansas next
week.

Readers looking for the nitty-gritty details of the theory of the “new
economy” will have to go elsewhere for salvation, however, because Ormerod’s
goal is to explain more by example and anecdote than by precept. Nevertheless,
as befits a former staffer on The Economist, he fills his book with
stories reinforcing the notion that the economy is more like a living organism
than a machine. He opens with ant foraging. A traditional economist, seeing an
ant nest equidistant from two identical heaps of sugar, would expect the ants to
divide their energies randomly between the heaps. Biologists know better. The
first ant coming home with sweeties informs others. The more ants visit any
particular site, the greater chance that yet more will visit it in future.

So if you’re looking for a fascinating, entertaining introduction to the
economics of the 21st century, here’s an excellent starting place. In fewer than
200 pages, you’ll learn more about the way real economies work than by reading a
bookshelf full of academic tomes. And you’ll have a lot of fun along the
way.

Straightaway one of the biologists at the table shot back, “That’s a
no-brainer. Darwin’s On the Origin of Species is the only choice
possible.” While I sagely nodded my head, I was thinking to myself how dull and
disappointing this answer was. Not because Darwin’s classic isn’t worthy of
desert-island consideration, but simply because it is such a predictable choice.
It’s like nonscientists saying that they would take the Bible or
Encyclopaedia Britannica. What I was hoping to hear about was some book
that I did not already know of, that might point me in directions I had never
considered exploring.

While the conversation shifted to other matters, the desert island idea
stayed with me, and I began formulating a plan to ask some of my friends the
same question—but with a twist. Instead of allowing them the entire scope
of printed history since Gutenberg first moved type around, I would only allow
recently published books. I thought that a period of 25 years would be just
about right: not so long that obvious candidates would have had time to become
true classics, and not so short that fashions of the moment might sway someone’s
choice.

It is clear how difficult it is to come up with a selection in this game.
Whatever book you choose, it has simultaneously to satisfy several conflicting
criteria. Foremost among these is that it should be a book that can be read and
reread, each time with both pleasure and intellectual benefit.

So I decided that any volumes that entertain the first time round, but hold
no appeal for an encore performance, were out. I also decided to exclude books
that focus on a single theme which is explored and explicated in painstaking
detail, because works such as these do not leave enough for the imagination to
work on. What is needed to alleviate the rigours of solitude is a book from
which you can draw an almost infinite supply of meaty questions to ponder while
awaiting rescue from the island—or planning your escape. Moreover, these
questions should be implicit in the material of the book rather than being
simply stated outright in, say, a long list of unsolved problems. And, of
course, the writing style, format of the book and structure of presentation
should be such that the reader never tires of going through it.

With these criteria in mind, I asked a dozen or so of my friends and
colleagues the edited version of my question—that is, “If you were to be
banished for life to an uninhabited desert island, what one science book
published in the past 25 years would you take with you and why?”

So without further ado, you will find the (almost) unedited, unexpurgated
replies that I received on the next page.

What a library! And my own selection? Let me confess that my own choice was
Doug Hofstadter’s already classic volume, Gödel, Escher, Bach. So
when the very first, from Ian Stewart, came up with that title, I was slightly
annoyed. How could he have picked the very book that I wanted to use? Maggie
Boden’s first choice was also Hofstadter’s tour de force.

On the principle that we should be offering as broad a selection of books as
possible, I asked Maggie to name an alternative book, forcing myself back to the
drawing board as well.

After racking my brain for several days and casting my eye up and down the
bookshelves at home, it finally dawned on me that Hofstadter had written a
desert island book that I like even better than Gödel, Escher,
Bach. So my choice for the long, one-way journey to nowhere is the book
Hofstadter first published in 1985, Metamagical Themas(Penguin/Basic
Books in the US, 1996, ÂŁ16/$23, ISBN 0140179968/0465045669). This
volume, which for the most part consists of the columns he wrote for
Scientific American magazine, contains thought-provoking facts and
speculations on virtually every page.

With entertaining and challenging treatments of topics ranging from the
genetic code, the music of Chopin, the aesthetics of different typefaces and the
peculiarities of the Rubik cube, this is exactly the type of book from which one
can draw intellectual sustenance for a desert-island lifetime. So with
Metamagical Themas in my backpack, I might even welcome my stay on that
desert island, as an opportunity to yet once again revisit these fascinating
topics.

And if you are wondering what Hofstadter himself would choose to read if
exiled to the island, the answer lies on p 51.

A Logical Journey by Hao Wang, MIT Press, ÂŁ33.95/$40, ISBN 0 262
23189 1

AS Kurt Gödel lay on his deathbed in a Princeton hospital in January
1978, he was described to his physician as “the greatest logician since
Aristotle”. The words were spoken by Robert Oppenheimer, director of the
Institute for Advanced Study, where Gödel had spent the bulk of his
professional life. While no one would deny this claim, A Logical
Journey marshals ammunition for the even broader one that Gödel was,
perhaps, also one of the greatest philosophers—at least of the 20th
century, if not since Aristotle. This is quite an assertion, but when it comes
from Hao Wang, a man who was himself no slouch in the philosophy business, it
deserves some attention.

By now, almost everyone is familiar with Gödel’s work on incompleteness
in mathematics, as well as the possible implications of this work in areas such
as artificial intelligence. But Gödel was selective in what work he
published. And after his death a vast body of material outlining his views on
subjects such as Platonism, the nature of logic, the existence of God, logical
positivism and phenomenology was discovered in his unpublished papers.

Wang was fortunate enough to have had numerous personal conversations with
Gödel, especially during the last decade of Gödel’s life, and this
book draws heavily on these discussions to present a fascinating glimpse into
Gödel’s mind. Wang had completed the draft manuscript of this book when he
died suddenly in 1995. The draft was brought to publication by several of his
associates, and this has left the book with a less than polished
flavour—one that is somehow more spontaneous and informal than what I’m
sure would have been the case if Wang had lived to see the book into print.

I liked this somewhat less finished style, and its more colloquial tone than
in Wang’s earlier writings on Gödel’s life. This book is about as close as
one can get to a private conversation with Gödel himself on a whole host of
issues in both philosophy and mathematics. So if you’d like to know what it was
like to talk to the great man, this is the place to find out.

Thought Contagion by Aaron Lynch, Basic Books, New York,
$24, ISBN 0 465 08466 4

AT the end of his enormously popular book, The Selfish Gene, Richard
Dawkins introduced the idea of a kind of sociocultural “gene” that would carry
information about cultural artefacts from one brain to another, much as
biological genes carry information about the physical structure of an organism
from one body to another.

Dawkins termed this hypothetical cultural gene, a “meme”, citing catchy tunes
or political ideologies that move from one host—a human brain—to
another, much like the spread of a contagious disease. Thought
Contagion is an extended account of how such beliefs, coded as memes,
spread through society via the medium of the human mind.

The twist here is that while traditional sociology and psychology focuses on
how people acquire ideas, memetics turns the question on its head and asks: how
do ideas acquire people?

Of course, Dawkins’s idea of using concepts and arguments from Darwinian
evolution to explain social phenomena is not a new one. For instance, Herbert
Spencer and the Social Darwinists employed much the same strategy decades ago to
justify some rather dubious social programmes. So, for that matter, did the
Nazis. And even in the world of science, the notion of both “culturgenes” and
“sociogenes” have been introduced by Charles Lumsden and E.O. Wilson, and Carl
Swanson, respectively, to serve much the same function as the meme.

But what distinguishes the notion of the meme from these exercises in
intellectual transference is the idea that one might base an actual science of
human belief—its origins, transmission, and end—on the concept of a
meme. Thought Contagion is a kind of manifesto in support of this dream
(or should it be nightmare?).

As an indication of the scope of Thought Contagion, here is a (very)
partial list of the bewildering variety of social, economic, and cultural
concepts that the book accounts for in terms of memes: folklore, love for
children, inter-faith marriage, crucifixion, homosexuality, bottle-feeding,
astrology, breast fetishes, single parenthood, class rebellion, jihads, street
gangs, talk radio, jingoism, masturbation, male dominance, birth control,
polygyny, drug abuse, abortion and chastity. This Borges-like list gives you the
general idea: the world according to Aaron Lynch is totally driven by memes. But
what exactly is a meme? Lynch suggests that it is “Like a software virus in a
computer network or a physical virus in a city”, adding that thought contagions
proliferate “by effectively `reprogramming’ for their own transmission”.

Unfortunately, at least for theoreticians, Lynch’s account of memes is one
based on example rather than on precept, and the only characterisation the book
gives of what constitutes a meme is that it is a self-replicating entity that
somehow encodes the cultural artefact in question and reproduces itself by
transfer from a host brain to another brain.

If you put this vague, almost useless, description of a meme beside the
book’s subtitle, “The New Science of Memes”, an eyebrow or two might be raised
over claims of the emergence of a new science of memetics. Unlike biological
genes, where both experimentalists and theoreticians have constructed good,
workable symbolic and mathematical characterisations of genes, memeticians like
Lynch have yet to put forth any theoretical constructs that would enable you
actually to study the creation and propagation of such entities in anything even
approaching a scientific manner.

To illustrate the difficulty, consider one of the book’s almost endless
supply of examples, what Lynch calls the “promiscuity meme”. Here is an
abbreviated version of his account of how this meme propagates. Those who
believe in having a crowded sex life often hold the proselytic advantage over
monogamists. And while both pluralists and monogamists would like to promote
their memes, monogamists tend to stop this promotion after they find one sexual
relationship. Pluralists, on the other hand, continue proselytising to potential
partners even after entering one or more sexual relationships.

So their messages can also reach bystanders, helping the pluralist meme jump
from man to man or woman instead of just man to woman to man. Ergo, Lynch
concludes, promiscuity memes enjoy episodes of great replication, with their
propagation advantages helping to fuel the explosive growth of free sex that
characterised the so-called Sexual Revolution.

Just to hammer home the point, let me cite another example from the book that
employs this same line of reasoning. Rival street gangs in large cities need to
acquire new members to stay in the game. All the gang members know this and
pressure nonmembers to join, thereby expanding the gang’s control over turf and
revenue. Gangs with the most hostile and criminal attitudes have the most
intense recruitment needs, which causes them to out-propagate rival, milder
gangs. Fuelled by such life-and-death motives in both the recruiters and their
targets, street gang memes replicate despite the best efforts of communities to
stamp them out. Or so goes the memetician’s story, anyway. Lynch also narrates
the tale of how ideas about a country’s military strength spread. Again, as with
the promiscuity meme, one has, he says, a “proselytic advantage”. The idea that
“my country is dangerously low on weapons” makes those who believe
this—the “hosts”—fearful. The “side effect of fear triggers
proselytizing”. Those who do not share this belief do not act to promote their
view.

I’ll leave it to the reader’s judgment as to whether these stories look as if
they contain the nucleus of a scientific attack on either promiscuity or street
gangs. Let me only say here that almost every example Lynch cites to support of
the meme hypothesis rests upon exactly this sort of storytelling, and give no
real insight into how one might construct an actual theory of memetics that
would in some way encompass—and explain—this snowstorm of social and
behavioural norms.

What I would like to see is something akin to the theoretical constructs and
modes of transmission that people use to describe mathematically the spread of a
contagious disease or the diffusion of various types of genes throughout a
population of hosts. On the surface, it seems that such a theoretical framework
is quite feasible for describing memes as well, but you will not find a real
indication of how it might work in this volume.

In the tradition of the experimentalist rather than the theoretician, tucked
into my copy of the book was a mimeographed sheet describing the use of the
World Wide Web as a medium of thought contagion. The idea is that if you belong
to an e-mail forum, in which each member’s messages are sent to all the other
members, then by putting the URL http://www.mcs.net/~aaron/thoughtcontagion.html
somewhere in your message, that Web address will pop up in colour as a
push-button link in the recipient’s e-mail on most computers. The
recipients can then just click their mouse on this button, and start reading
about Thought Contagion and sampling its opening chapter.

So here is an example of thought contagion in action, in which the meme of
the book itself is transmitted from mind to mind, via the medium of the global
electronic brain called the Internet.

But perhaps it’s unfair to hold this book to a scholarly standard that it
never claims to aspire to. Despite the fact that there is not even a sketch of
what an actual theory of memetics might be like, the work is not without merit.
It is true that the idea of an information analogue of a biological gene is
irresistibly appealing.

It is true that Thought Contagion contains such a wealth of possible
examples of beliefs transmitted by “something” from one brain to another that at
least a few of them might be taken seriously. And it is also true that Lynch has
dressed-up these examples in an interesting, easy-to-read-and-understand
fashion. So for all of these reasons I would recommend this book as an
excellent introduction to what memes are all about. But for an actual science of
memetics, readers will have to look elsewhere.