Hackers produce more than software, they produce hackers


(Previously The Identity Games of Hacker Culture)

By Lars Risan
Version 2.1



Abstract

This paper discusses the social cohesion of Open Source projects. In Open Source programming most of the programmers who contribute to a project are very loyal to it, even if the Open Source licenses basically states that anybody can do what they want with the software, as long as they do not deny others the same right. The paper explores how loyalties to projects are created in social processes that also produce the identities of the programmers. In dialogue with Eric Raymond’s seminal writing on Open Source, the paper partly draws on evolutionary psychology. Hence, two models are developed; symmetrical and complementary identity games.



Content:
1. Introducing Linux
2. Explaining Linux
3. Does a natural history of competing agents explain the emergence of open source software?
4. Complementary versus symmetrical identity games
5. Runaway sexual selection
6. The human brain
7. Human protean behavior
8. Hacking as ``protean courtship''?
9. Strange Loops: As meaningless as life itself
10. Getting the ideology right
11. Darwin revisited
12. Creating loyalty to projects

14. References
15. Notes



1. Introducing Linux

The history of Linux is the adventure of a young Finish student, Linus Torvalds, and a large network of programmers who made an operating system for PCs that is more flexible, more robust and needs less resources than the giant Microsoft Windows. And, best of all, it is free! These programmers often call themselves «hackers», they are often computer science students, and their «hacks» consist of constructive programming, not destructive crimes.

Together with other free programs -- known as freeware, or Open Source -- Linux is today vital to the Net. "Apache" is one of the most successful Open Source projects. A press release on their web site reads: "Software produced by the Apache HTTP Server Project serves over 61% of all public Internet web sites, according to the June 1999 Netcraft Web Server Survey (http://www.netcraft.com/survey/). Hundreds of users have contributed ideas, code, and documentation to the project. Every eight seconds another Apache based web site joins the existing 3.5 million on the Web." (See http://www.apache.org/) You very rarely send an E-mail without using some Open Source software along the route.

How could a collection of informal networks of mostly unpaid students write a set of programs that is vital to the most important (at least the most hyped!) innovation of the 90 ties, the Net, and that may be about to undermine the monopoly of Microsoft? An why do they do it for free?

These questions are raised and answered by Eric S. Raymond in his seminal papers The Cathedral and the Bazaar (CB) and Homesteading the Noosphere (HtN). The current paper is a comment and a further exploration of Raymond's papers.

2. Explaining Linux

Speaking in very broad terms, we might say that Eric Raymond's The Cathedral and the Bazaar (CB) is about how hackers make Open Source software, whereas his Homesteading the Noosphere (HtN) is concerned with how that software make hackers. Open Source software, according to Raymond, produce hackers by being a testimony the of the skills of it's maker. Those testimonies (both in the form of working binaries and as credits in the version history of the source) are part of a modest «reputation game» that produce hackers as (better or worse) hackers. An effect of that game is cooperation, and HtN explains how that effect comes about. In HtN however, the reputation game in it's fundamental nature is based on a competitive instinct of individual human beings. Basically, I will question the primordial role that Raymond place on competition (seeing cooperation as a mere effect), but I will build on and further develop the analysis of HtN, rather than pose a radical alternative to it.

My purpose is double. I want to bring the explanation of how hackers become hackers one step further, for the very sake of the explanation. This is the «purely academic interest» of this paper, as one say. If one is sympathetic to the Open Source movement, one may of course hope that these explanations will help future hackers to improve the social design of their Open Source projects -- although I do have my doubts to the feasibility of such design. What characterize Open Source projects is precisely the lack of social design. Will intensive analytic thinking on the organization of these projects hamper them more than promote them? I don't know the answer to this. It may be the case or it may not, depending on a plurality of factors that I cannot analyze here.

There is, however, another way that the following explanations may be working, rather than as recipes for how to cook good Open Source projects. The explanations of The Cathedral and the Bazaar and Homesteading the Noosphere have become working tropes in the marketing and the ideological promotion of Open Source software. Thus the short step from Raymond's evolutionary explanations to Apple's naming of their Open Source project; «Darwin». And thus the short step from Raymond's decentralized and bottom up explanations to his own classical-liberal political position. Following this, I want to make better explanations of how hackers become hackers because, as I will show, these explanations provide us with better ideologies (or better marketing tropes, depending on how this paper may be used). Again, my additions to the explanations of HtN does not pose a radical alternative to the ideology of Raymond, they merely twist it a bit here and there.

3. The main problem

The main point of the paper is an answer to the following question. Does a natural history of competing agents explain the emergence of Open Source software? What HtN explains, and explains well, is what we quite simply may call the wonder of Linux : A seemingly anarchistic, very little organized group of hackers have made an extremely robust operating system. Eric Raymond calls this organization a «bazaar», drawing on the bottom up qualities of such markets. An operating system is a complicated system; everybody thought one needed a strictly hierarchical and well organized group of people, a «cathedral-style» of an organization in Raymond’s terminology, an organization based on scientific management (Hannemyr 1997), in order to make them as robust as they ought to be. The emergence of Linux has proved that this is not the only possible organization of large scale programming.

Central to all Open Source programs are the various licenses. These licenses basically state that all source code (the program as it is readable by human beings) should be distributed together with the binary code (the program as it is executable by the computer), and that anybody who receives the program has the right to modify the source code, and to further distribute the program, changed or unchanged. In HtN, Raymond shows that in practice hackers are less «free floating» than this compulsory liberalism and freedom suggests. In addition to the of Open Source licenses, there is a set of quite strict rules of ownership in the hacker community. This ownership is not explicit and juridical, it is a part of the informal culture of hacking. Someone always has the main responsibility of every Open Source project. Hackers who make changes to a program («the Program»), and wish to publish these changes, always sends them to this moderator. The moderator decides when new versions of the Program is to be released. In the «Version history» of the Program people who have made contributions are credited.

This system of ownership and credit has two major effects. The first effect is really the «wonder of Linux» that HtN explains: These rules of hacker culture prevents Open Source projects to fork, to multiply into as many projects as there are users. «Linux» has remained one project, still to be controlled by Linus Torvalds. This entirely informal system of ownership and credit is faithfully followed -- very few hacker projects have forked -- even if the explicit rules of ownership basically states that anyone can do what he or she likes with the Program as long as others are not denied the same right. Thus, Raymond concludes that the hacker community has a «Promiscuous Theory» and a «Puritan Practice» (HtN). I may add that the facts that Linux works brilliantly, and that just a couple out of thousands of Open Source projects have forked is a wonderful testimony to the classical-liberal credo that «given freedom, people will take responsibility».

The second effect of the ownership and credit system of Open Source software -- an effect which is central to HtN's explanation of why OS-projects stick together -- is that these rules produce pride, personal identity and mutual acknowledgment of effort and skills. Hackers are credited for their work and become someone in the eyes of their significant others. Hackers do not hack for «free», they do it for some kind of social reward.

HtN describes the specific way in which hackers compete for social status as a gift culture.

In gift cultures, social status is determined not by what you control but by what you give away. Thus the Kwakiutl chieftain's potlatch party. Thus the multi-millionaire's elaborate and usually public acts of philanthropy. And thus the hacker's long hours of effort to produce high-quality open-source code. (HtN)

The hacker community may very well be compared to gift cultures, but there is a special case of gift culture from the annals of anthropology that even better than the potlatch highlights the, in fact, extraordinary organization of Open Source programming. The case in question is is the social organization around the Big Men of Melanesia, as described by anthropologist Marshall Sahlins (Sahlins 1963).

The power of the Big Men was very fragile. They where only able to keep their small «states» together as long as they where able to interest their followers with gifts. Their status as Big Men had to be won by their personal ability to win trust and respect of followers, and thereby to collect gifts from these followers. Later, these gifts were redistributed in order both to win the trust and respect of new followers, as well as reconfirming old relationship of trust and respect. So, being a Big Man was a costly affair, and the Big Man of a network could be one of the poorest persons of this network. When an established Big Man died, his network disintegrated, and new or neighboring Big Men had to reconquer the particular territory that the late Big Man had controlled.

Sahlins contrasts these Big Men of Melanesia to the Chiefs of Polynesia. The Chiefs had established «bureaucracies» to help them collect taxes. These bureaucracies survived the individual Chiefs, and the Chief position itself was inherited from father to son. Thus, the networks of the Chiefs were relatively stable across generations. The power of the Polynesian Chiefs was institutional, whereas the power of the Melanesian Big Men was strictly personal.

A company has its institutional Chief, the CEO. This position remains the same, even if the persons filling it will change. The leading hackers -- the moderators of Open Source projects -- however, are much closer to Big Men, given the way they maintain their position by collecting and redistributing software.

The size of the networks that the Big Men may control is a notable difference between the Big Men of hackerdom and those of old time Melanesia. Hackers can build much larger networks, and the reason is quite simple. The Big Men of Melanesia collected and redistributed material gifts. These resources were limited. If a Big Man wanted to win the respect and trust of, say, a neighboring village, he had to offer them gifts for some time, before the relation was established and he started getting anything in return. The gifts spent on this neighboring village had to be collected form his established followers in his own village. But if these followers felt that he gave too much to third parties of what they gave him, they could just turn away from him and save their gifts for another, competing Big Man. The Big Men thus had to balance the gifts spent on conquering new territory versus maintaining old ones. A finite amount of material resources thus limited the size of these networks.

The Big Men of Open Source software does not have this limitation. They can upload their program to the net, and give away an infinite amount of gifts. These computer code gifts work the opposite way of a scarce material recourse. When more people download and start using these gifts two things happen: First, the gifts are subjects to an opposite inflation; they generally become more valuable, not less, when they are easily and widely distributed, because many users makes many programs more useable. Second, the particular gifts to those who have contributed to the distribution -- the gift, that is, of having their names and credits listed in the version history -- also becomes more valuable when the program is distributed to many rather than to few. They then become remembered by many, as contributors to success histories, rather than being forgotten as participants of tiny technical inventions. Linus Torvalds is not the «CEO of Linux». There is not any such chief position, it cannot be, according to the GPL-license that copyrights the software. Torvalds is an informal Big Man, in a global network of information economy, of which size the old Big Men of Melanesia could only dream about. I am fascinated by the fact that Torvalds’ position is respected, not only by his fellow hackers, but also by firms (such as IBM) that now invest huge amounts of money on the Linux platform.

The analogy to the Big Men of Melanesia illustrates the informal quality of what Raymond describes and explains as a «reputation game» (HtN). In this explanation the reputation game is strictly competitive; hackers compete for social status, and Raymond goes on to explain this drive for competition by reference to Natural History (hackers compete because they, as human beings, have an innate drive to do so. I will discuss this sociobiology at length further down.) While not denying the central importance of this reputation game, I’d still like to complicate it. There is more to this «game» than competition. There can be many kinds of reputations, among many kinds of «peers». Not all of them are the result of competition. Here are a few:

First we need to distinguish the Big Men of Open Source projects from the «followers», the more or less famous and eager debuggers and patch writers of these projects. The very kind (not only the degree) of reputations among these two groups need not be similar. Moreover, there is the reputation that leaders may have among other leaders, as well as the reputation they may have among their followers. Also, these two reputations can be qualitatively different. Finally I could mention the difference between the reputations that followers may have with one particular leader, as well as the one they have among other followers. The contexts of these reputations may vary to a great extent. There may be the relatively global reputation of Linus Torvalds, and there may be the local reputation that a student of computer science has among his friends in the PC-lab.

The following is an example of how the reputation game can be at work between the «Big Men of hackerdom». It verifies Raymond's competitive model, but is unusual in that it is so immodestly and explicitly concerned with reputation per se.

At http://www.gnu.org/gnu/linux-and-gnu.html Richard Stallman writes about «Linux distributions» and the difference between «Linux», the kernel, and «GNU», the large database of free software (of which the linux kernel is a tiny, albeit central, part). Stallman is worried about the fact that in a so called «Linux distribution» the Linux kernel makes up only about 3% of the programs. About 28% of the distribution are other GNU programs. Stallman writes, «So if you were going to pick one name for the system based on who wrote the programs in the system, the most appropriate single choice would be «GNU».» Stallman also writes:

``Many users are not fully aware of the distinction between the kernel, which is Linux, and the whole system, which they also call «Linux». The ambiguous use of the name doesn't promote understanding.»

Well it is true, concerning the facts of the case, that «this doesn't promote understanding». But it doesn't promote «GNU» either, And promoting «GNU» is quite obviously the concern of Stallman in this document. Somebody is not getting their due credit, and if, as Stallman write, we were «going to pick a single name», and we picked «GNU», then the single person whose reputation would have increased the most, would have been Stallman, at the expense of Torvalds. Fighting for his own reputation was not necessarily Stallman's intention behind writing this document. But if the -- according to Stallman -- more fair naming became standard, the effect would nevertheless have been an increse of Stallman's fame.

So far so well, Raymond's explanations hold. But then, consider the following, slightly constructed example of how the relationship between a leader and a follower might work:

One hacker, let's call him Jim, downloads an early version of Linux. He finds some bugs, corrects them and send the corrections to Torvalds. Torvalds sends him a «thank you» message, and incorporates the corrections into the new release. This happens again, perhaps several times. Now, the more famous Linux (and thus Torvalds) gets, the more proud Jim becomes -- because he helped Torvalds. Jim can start to boost about being one of Torvalds helpers. At a conference where Torvalds is invited, Jim also attend. He claps vigorously after Torvalds' talk. He promotes Linux to all the people he knows. Jim has become a faithful follower of Torvalds.

In this case there does not seem to be a competition about a scarce social resource (high social status). The more famous Torvalds becomes, the more Jim's «follower-of-Torvalds» identity is strengthened. Moreover, by promoting Linux and, every now and then, fixing a bug on the latest release of the Linux kernel, Jim is, by strengthening the quality of Linux, also strengthening Torvalds' position. So, the more leader Torvalds becomes, the more devoted follower Jim becomes. The more devoted follower Jim becomes, the more leader Torvalds becomes (unless or until, for example, Jim becomes a total fanatic, loosing all critical sense). They do not compete about leadership or followership, they produce the conditions for each others leadership and followership.

Raymond, as I mentioned above, explains the, to him, competitive reputation games by reference to Natural History:

``Human beings have an innate drive to compete for social status; it's wired in by our evolutionary history. For the 90% of that history that ran before the invention of agriculture, our ancestors lived in small nomadic hunting-gathering bands. High-status individuals (those most effective at informing coalitions and persuading others to cooperate with them) got the healthiest mates and access to the best food. This drive for status expresses itself in different ways, depending largely on the degree of scarcity of survival goods.» (HtN)

If a human instinctive drive to «compete for social status» was to be at work in the interactions between a leader and a follower, this instinct would prevent the process, not drive it. If Jim really started to compete with Torvalds, what would be the likely result? Linux would fork instantly, as Jim would release his own unix clone, «Jimux», hoping it would be a worthy alternative to Torvalds' «Linux». (And GPL-license gives him the full right to do this. He could however not redistribute his own version under the name of «Linux TM », as Torvalds has protected this name as an ordinary trademark.)

This is not to say that human beings may not have an «instinct» to compete for social status, only that this possible instinct, as all instincts, is activated in certain contexts and not in others. Competition for social status surely take place in the hacker community, between leaders as well as between followers, but we need something else than competition to explain the dynamics of the many leader-follower relationships among hackers.

4. Complementary versus symmetrical identity games

One possible way to go is to describe leader-leader (and follower-follower) relationships versus leader-follower relationships in terms of Gregory Bateson's symmetrical and complementary relationships (Bateson 1972). Let's first look at symmetrical relationships.

In competitive relationships, such as between Stallman and Torvalds, the agents are, to use a sport metaphor, running for the same goal. They agree on this goal. (If one runner tries to run the distance as fast as possible, while the other one tries to use as long time as possible, there would be a rather silly «competition».) So there is an element of symmetry in competitions: The resources that the agents search for are the same, and there is a necessary win-loose situation. Somebody gets more gold, others get less. If all the agents got more than plenty of the desired recourse, competition would cease. (In sport and in some political ideologies there is an additional symmetry: The competition should be fair. Thus the starting conditions should also be equal, people should have «equal possibilities». Fair competitions may of course be relatively rear.) Competitions between humans (I don't know about animals) produce «social status», or more broadly, personal identities. You «become someone» by competing. Competitions are thus instances of what we might call symmetrical identity games.

The interaction between a follower and a leader is very different form symmetrical identity games. Resources are something we have in our environments. The follower has a leader in his environment, the leader has a follower in his environment. There is a possible positive feedback between these two: The more you get of the one, the more you may get of the other. In these cases there is a sort of «win-win» situation, because no one is losing. But since there is no loser, it is also a bit inaccurate to speak of winners. So rather than speaking of two winners in a win-win situation, something that suggest that the two agents are somehow equal in a competition, a symmetrical game, I prefer to speak about complementary identity games. The two agents are complementing each other, mutually producing the conditions for each other's existence.

I think that both symmetrical and complementary identity games are part of the social dynamics of the hacker community. I also think that complementary identity games are much more important, at least quantitatively speaking. There are many more followers than leaders. They are all producing their identities as complementary to some leader. A leader such as Linus Torvalds probably communicates much more with his followers than with other leaders, he is involved in much more complementary identity production than in symmetrical identity production.

In HtN all identity production is described as symmetrical (competition for social status). These symmetrical processes are explained in a naturalistic way, as part of Natural History. They are instances of a Darwinian evolution that is also presented as entirely based on symmetrical (that is competitive) processes. So, how do I incorporate the complementary identity games into an understanding of hacker culture?

There are several possibilities: First, I can deny that complementary identity games can be explained naturalistically. I can insist that the production of hackers is a purely «social» process, that «nature» has nothing to do with this case. (After all, I am a social anthropologist!)

Secondly, I can question the usefulness of evolutionary explanations on more methodological reasons. For example: Sociobiologists, or «evolutionary psychologists» as they call themselves nowadays, can tell almost any kind of story and make it come out «true». This is 1) because among the heterogeneous and varied social life of today there is always some life form available to support almost any kind of hypothesis, and 2) because the paleontological record is extremely scarce. The gaps are huge and possible interpretations are many. So, again, stick with a purely «social» explanation. There is a Science War going on, and I (again as a social anthropologist) should better stay on the right (that is the left?) side of the war.

I am not happy with these denials of evolutionary explanations, anymore than I am happy with the confrontations of Science War. Like it or not, I think we are destined to understand ourselves as subjects in some kind of tale of origin. I don't know if this destiny is a deeply founded cultural twist; the «destiny» of the West, given by a Judeo-Christian linear World View, or if it is something else. But wherever it comes from, not telling tales of origin is merely giving others the room to tell these stories. So I rather attempt to tell better histories of origin. Thus I opt for my third alternative:

I will insist on an evolutionary explanation of why hackers hack, but I will broaden up our understanding of Natural History, to include complementary processes as a part of it. There are other fundamentals to include in our Great Story of Becoming than competitive individuals. The one I will explore here is called sexual selection .

The last couple of decades sexual selection has had a revival, after it was first introduced by Charles Darwin, later to be mostly neglected by biologists for a century. In the following I will briefly retell Natural History with sexual selection as a fundamental part of it. I will then situate hacking within that history.

5. Runaway sexual selection

In 1871 Charles Darwin introduced the concept of «selection in relation to sex», or sexual selection (Darwin 1871). In the more famous natural selection (Darwin 1859), the selection pressure is external to the species. The selection pressure is the environment, the surrounding «nature''; hence natural selection. In sexual selection the selective pressure is internal to the species. It is the opposite sex. In a Review on Sexual Selection Geoffrey Miller writes:

Darwin (1859, 1871) realized that his theory of natural selection through differential survival could not explain extravagant male traits such as the peacock's tail, because such traits actually decrease survival ability. Rather, he reasoned that in a sexually-reproducing species, any heritable traits that help in competing for sexual mates will tend to spread through the species, even if they compromise survival somewhat. (Miller, in press 1: 2)

In 1930, biologist R. A. Fisher proposed the existence of a process he called runaway sexual selection. Here is how this process might work; I am using the emergence of the peacock and peahen on earth as an example.

The peafowl is member of the pheasant family of birds. Experiments has shownthat peahens have a preference for males with relatively longer tail feathers. Let us assume that among the «proto-pheasant», long time ago, the females came in two types: 1) Choosy females that preferred males with relatively longer tale feathers (than that of the average of males). 2) Non-choosy females with respect to longer tail feathers. Then, at some point in time, some males with longer tail feathers emerged on the scene. The choosy females would choose longer tailed males, and thus produce offspring inheriting both the «longer male tail feather genes» and the «choosy female choice genes». The non-choosy (with respect to male tail length) females would produce offspring with both longer and shorter tail feathers. On the average, the tail would not grow in this latter population. In the sub population inheriting both the longer male tail and the choosy female choice, however, the male tails would have a tendency to grow as generations passed. This is because the females taste was relative: They would always go for the males with the longest tail feathers. Soon a new race -- a species to be -- of pheasants would have budded off from the three of evolution. Fisher wrote about this (using plumage as example):

The two characteristics affected by such a process, namely plumage development in the male and sexual preference in the female, must thus advance together, and so long as the process is unchecked by severe counter selection, will advance with ever-increasing speed. (quoted in Miller, in press 1: 9)

The termination of a runaway sexual selection will at some point have to take place, one way or another. But if external resources are in relative abundance, sexual selection may take the development of elaborate traits quite far. The males of the now extinct Irish Deer, for example, had antlers that measured up to 13 feet, 3, 5 meters, tip to tip. This was very likely a result of a runaway sexual selection. The huge male antlers probably became the destiny of the species as well. They became a severe obstacle, not in themselves, but in relation to an always changing environment (and I have seen two possible environmental changes proposed; a change from low vegetation to taller bushes and trees, and/or the advent of a new predator; the homo sapiens).

Runaway sexual selection may also loose its momentum for reasons that do not involve natural selection. If, for example, some female birds have a taste for red-chested-males, then the males may start to invest some of their energy on producing redder chests. At some point, however, the average male will have a saturated red spot on their chest. Redness cannot possibly be more red than bright red. Producing more red pigment will not produce a redder chest and thus will not be of any help in attracting females. Those who invest in more than average red pigment will not get a pay off in terms of more than average reproduction. Thus the runaway process will terminate smoothly, not abruptly. To my knowledge, most color ornamentation of birds and fishes are in bright, saturated colors, suggesting precisely that their emergence was through runaway sexual selection.

Biologists are also exploring the possible beginnings of (runaway) sexual selection. Miller reviews this research, and just to give one example of the many evolutionary mechanisms that are at play:

Basolo (1990) showed that female platyfish prefer males with colourful plastic `swords' glued on the ends of their tails, suggesting that this preference also pre-dated the evolution of such ornaments in their close relatives the swordtails. (Miller, in press 1: 8)

So the choosiness may have predated the elaborated ornament. This choosiness, again, may have had a previous function. Miller suggests: «Birds that eat blue berries may evolve blue sensitive eyes, which would tend to favor blue ornaments». (Miller in press 1: 8).

Natural selection works together with the «creative» principles of mutation and crossover to produce change with direction. But the selective mechanism itself is «conservative». It selects against a given trait and works in situations of scarcity. There is no «survival of the fittest» in natural selection, it is just the death of the unfit.

Sexual selection, on the other side, selects actively for a given trait, it is a more «creative» selection. It may take a species into new spaces of shapes and behaviors, and it works in situations of relative material abundance; it has to be relative material abundance for organisms to invest so much energy on spectacular ornamentation.

So, let's look at the process of sexual selection in terms of symmetry and complementarity. Surely, there is an important symmetry in these processes. In the case of colorful birds, the symmetry is in the relationships between the males, as their relationships to the females are similar. They compete to obtain favor with the females. But there is an equally important complementary relationship, namely the relationship between the male and the female. This relationship, on an evolutionary scale, has something important in common to the relationship between the «Big Men» of hacking and their followers. By the very act of following Big Men, followers produce Big Men as we know them. The presence of followers in the environment is a precondition for Big Men to emerge, just as the presence of choosy female birds is a precondition for elaborated male birds to emerge.

But then, there is a major difference between these two scenarios. In the case of hacking, the relationship is mutual; just as Big Men need followers to emerge, so does followers need Big Men in their environment to emerge. In the situation described above, with choosy female birds and increasingly elaborate males, we have one half of such a loop: The females produce ever more elaborate males, but males need not produce ever more choosy females. (If their choosiness is relative, it need not change as the male trait grow. Females just keep on choosing as they always have done; the most elaborate male.)

But if males were also choosy and females also produce ever more elaborate traits, then the situation would be a kind of complementary identity game: the two agents, on an evolutionary scale, would produce each other mutually by being each other's environment. Such cases exists; one of them may be the evolution of the human brain.

6. The human brain

Let's start with laughter. Laughter can be many things. It can be cynical, or it can be pretended. It can be false or it can be hysterical. But it can also be genuine. And we all know what that is from our own experience of it. A genuine laughter is most often, if not always, a reaction to some kind of expressive novelty or twist. Something is reorganized, put together in unordered ways, something is surprising. Human beings seems to be constantly starving for novelty in some form. We enjoy certain kinds of, and we are very able to make them. 1 Laughter is one, of many, expressions of the human capacity to enjoy the protean behavior of our friends. 2 Evolutionary psychologist Geoffrey Miller has caught considerably interest lately for his hypothesis that our protean behavior has played a central role in the evolution of the human brain (see Miller 1993, 2000, and Miller in press 1).

Arguing for his hypothesis, Miller turns to the paleontological records as well as to observations of what present time human beings tell us. Some brief points of our paleontological history is a useful starting point for the argument:


So, first our ancestors raised up on their posterior legs and became bipedal. They had skulls of the same size as chimpanzees, and they did all right as bipedal creatures on the African savannah. Then for about 3,5 million years they lived a life with an almost unchanged technology, a technology, or «extended phenotype» not very much more complex than that of the beaver or the spider (Miller 1993:219). The first million years or so, they lived happily, and spread out on the eastern part of Africa, with no major increase in skull size. Then, without any corresponding ecological change, without any change in technology, their skulls started to grow, from 500 cc and up to 1300 cc. As this increase of brain size took place, other primates inhabited the same ecological settings without developing a highly energy consuming brain. Then another 200.00 years passed on, and then they started to invent technology.

Thus, it seems that we did not need big brains in order to become bipedal, and we did not need to evolve big brains in order to invent better technology for our survival on the African savannah. Contrary to most theories of how we became human, there does not seem to be any external, ecological reason for the growth of our brains. It does not seem to be any «Nature» out there to have forced us to go for bigger brains. If we stick to the version of our paleontological history as listed above, there is no available explanation of our big brains in terms of Natural Selection. But, as the reader may have guessed, there is in terms of Sexual Selection, or at least in Miller's slightly twisted version of Sexual Selection.

7. Human protean behavior

The major mental capabilities of modern humans include language, music, art, complex sexual play, conceptual play and ideology (Miller 1993:219). These are uniquely human. 3 These capabilities are all connected to creativity, to laughter, and to novelty. They are extremely flexible, and their flexibility are utilized cross culturally. There is a story about the «Modern Man», I may add, where «Modernity» is seen as a unique culture of change, and celebrated as such. This «culture of change» is juxtaposed to the supposedly «traditional» life of non-Westerners. By modernist Westerners this stable life is seen as «primitive», whereas the more romantic Westerner often sees these «old cultures» as «natural», «harmonious» and eternal -- in ecological tune with Mother Nature since time immemorial. Countless anthropological studies have falsified these pictures. Even the supposedly most «Natural» of us, the people of New Guinea, are creatively using, inventing and rearranging their expressive repertoire, not faithfully reproducing a harmonious or unchanging culture (see Barth 1975). So change and creativity is a cultural universal. It is not something particular «modern». The very cultural diversity itself is a testimony of the universality of human creativity.

So, might it not be that some of our ancestors started to like novel and creative expressive behavior, whereas others started to utilize this preference in order to attract possible sexual mates? We would, or at least we could, get something like a runaway sexual selection. To like novelty is like the preference for longer tail feathers. It is a relative taste. The new bodily dance movement, the new song, the new joke is only new until someone else comes along and makes another newer, more creative dance movement, song or joke. Thus, those with better than average creativity had a better chance to reproduce, and the human ability to attract mates by protean behavior started to grow down the generations.

There is something peculiar about this particular kind of sexual selection. The bodily and cognitive capacity of the peahen to like longer (and more colorful) tails are quite different form the peacocks ability to grow longer and more colorful tail feathers. The cognitive and bodily human ability to act protean (say, to tell a good joke) is probably not very different from the ability to enjoy a good joke. This means that to choose someone who is good at telling stories, is also to chose someone who good at enjoying stories. To enjoy a good story is, evolutionary speaking, similar to enjoying long tail feathers; it is «choosiness». Thus, the dynamics of this runaway sexual selection is quite extraordinary. To select for the sexual ornament (protean behavior) of the partner, is also to select for the choosiness of the partner. Both males and females were choosing choosiness! Thus, the choosiness evolves along with the ornament in a positive feedback loop. Moreover, the sexual ornament -- i.e. the protean behavior -- is not just something to observe at a distance. It is interactive: To be good at dancing, playing music, conversing, etc. is to be good at doing this together with others, including the choosy partner.

This is an extremely powerful coupling of (sexually) selection pressures and selected traits. In the case of the peafowl we saw that the choosiness of the female did not need to evolve to the same degree as the elaborate trait evolved. It was sufficient that the choosiness was relative, always choosing the most elaborate male at the given generation. In the case of human protean behavior, the very choosiness evolves positively with the elaborate trait; and mutually, in both the female and the male. The better we became at telling good stories, the better able we became at enjoying good stories. The more man and woman both became creative language users, the more they both wanted creative language users.

It may of course be that men and women both choose and perform their protean behaviors a bit differently. The symmetry in choice and performance need not be perfect. Miller (1993:222-226) argues along classical sociobiological lines that this is the case: Males tend to broadcast their protean behavior more publicly (in order to, putting it crudely, «spread their genes''), females tend to «narrow cast» their protean courtship displays (to have a devoted father, «close to home», who will help bringing up their kids). As a social anthropologist I am brought up to think that this is, well ..., traditional «sociobiological crap» (fueling Science War). In this paper there is neither time nor space to defend or reject such a universal human sex difference, even if it does seem that the age and sex distribution of hackers (males in their late teenage and their 20 ties), and the way they broadcast their computational artistry, support Miller's thesis.

8. Hacking as «protean courtship''?

As argued above, complex technology entered the scene 200.000 years after the human brain had reached its present size. It did not drive this evolution. But it may be an effect of it. It may be that modern technology, in its paleontological beginning, started of as artistic bits and pieces in our protean courtship. Perhaps its first use was aesthetic rather than practical. Perhaps -- well many places, most certainly -- technology as art still is a protean courtship device, an erotic expression. Does this mean that technology in general is a courtship device? Can free software be reduced to an esthetical expression that is made in order to impress possible sexual mates?

In that case we would have a naturalist and causal explanation of a kind that we might call a «really is» reduction. Hacking is really a kind of sexual courtship. The identity games of hacking is a matter of producing sexual identities.

Empirically, this is a dubious reduction. The very large majority of hackers are males. Hackers are each others significant others, each others peer group. But the hacker community is not a gay community. One might argue the Freudian way: Their hacking is «sublimated sexuality», even if the hackers themselves do not recognize their hacking as «sexual behavior». Now as many critics of Freudian psychology has said many times (e.g. Popper, 1972), this is a kind of argument that cannot be proved anymore than it can be falsified empirically. It is a non-falsifiable tautology.

I find this criticism sensible, I do not want to say that hacking «really is» sexual behavior. Nevertheless; hacking is mostly done by young men in their twenties (some late teenage, some in their thirties). It is done by the same age and sex segment of people who, according to Miller, tend to broadcast their expressive novelties (Miller, 1993:222 - 226). So it may still be that the desires and skills which is needed to hack has an origin in runaway sexual selection, a sexual origin. But it need not be sexual by (genetical) nature. When the kind of behavior of which hacking is an example takes place in a sexual, courting context, it becomes sexual behavior, and it can be part of a loop of runaway sexual selection. To speak in a post-structuralist language; sexuality is performed or co-produced as we observe it in various relationships.

When the desires, pleasures and skills of art, politics, science and technology takes place in a sexual context they emerge or are co-produced as sexual, and can drive evolution, and has driven evolution. When these protean, expressive behaviors do not take place in sexual contexts, they simply are not sexual behaviors (and may, or may not, be part of biological evolution through natural selection). But they may still have their origin in a sexual context.

The degree to which hacking is courtship, is of course an empirical question, and may vary from one local setting to another, depending on how sexuality, of any kind, is involved in the performance of hacking. What is quite certain, if we follow Geoffrey Miller, is that the cognitive abilities to make the verbal, artistic and engineering expressions has a sexual origin.

*

I now turn to another connection between the runaway sexual selection that made us, and the complementary identity games that continues to make hackers. This connection is not historical, it is formal. The dynamical properties of the two processes are similar.

9. Strange Loops: As meaningless as life itself

Hackers cooperate, as Raymond describes (CB), to make very stable software, and they are able to do so because they are not in equal positions. Thus they use the software differently, and find different bugs. «Given enough eyeballs, all bugs are shallow» (CB). In this cooperation the hackers are complementary to each others. There are no losers. When cooperation is successful, all win. But this cooperation, even if it is complementary, is not the same as the complementary identity games.

Cooperation is about something -- a task, and this task is external to the cooperative relationship. This external task is the primary meaning of cooperation. To cooperate about the cooperation itself is meaningless.

identity_games.png
Competitive relationships are quite similar to cooperation. They have a meaning, and the meaning is outside the competitive relationship itself. The agents A and B do not compete about the competition itself, they compete about a scarce recourse, say consumers in a marked or gold medals in sport contests:

identity_games1.png
Runaway sexual selection and complementary identity games are different from these patterns. These processes are not «about» anything. Runaway sexual selection does not solve any problems outside the runaway process itself. The agents A and B, say «Man» and «Woman», did not necessarily select each other's elaborate ornaments because these ornaments were useful to some external task.
identity_games2.png
A complementary and self referential definition of agents. The relationship is complementary, even if the two arrows may symbolize identical processes, as when A and B are choosy about the same trait (e.g. protean behavior), and are both elaborating this trait (as well as the very choosiness). Douglas R. Hofstadter, in his famous book Gödel, Escher, Bach: An eternal golden braid, called such arrangements strange loops ( Hofstadter 1979).

Runaway sexual selections are processes that are sensitive to initial conditions. These initial conditions may have been totally meaningless, they may have been whims of evolution. Nevertheless, they may set of enormous changes in the new species-to-be. We might say that runaway sexual selection is as meaningless as life itself . 4 Life itself has no purpose, no task to be solved through cooperation, no scarce recourse to win in competition. It may be that life itself started out by a complementary process akin to runaway sexual selection and complementary identity games: Ilya Prigogine's theory of the origin of life, as I briefly understand it, is a story where large molecules organized themselves in «hypercycles». In these hypercycles, the macromolecules where able to reproduce. That is, in the primordial soup of the young Earth, before the first biological cell appeared on it, these molecules self organized by providing each other with each other's environment for (chemical) reproduction. Within the scientific fields known as «science of complexity» and «artificial life», these process has been studied extensively by the means of computer simulation. I cannot dive into that literature her; a Net search on words like "Science of complexity", "Stuart Kauffman", "Artificial Life", "self organization", "autopoeisis" and "Ilya Prigogine" will provide a number of possible starting points. For those who want to enjoy the riddles of such strange loops there is of course Hofstadter's book that I mentioned above (Hofstadter 1979).

There is one essay on «science of complexity» which should be mentioned here, as it deals with the Open Source movement. This is the paper «Linux: A Bazaar at the Edge of Chaos», by sociologist Ko Kuwabara. Kuwabara has interviewed some of the Linux kernel programmers and relates these interviews to some of notions developed within the science of complexity. The interviews are nice, but on the theoretical side Kuwabara restricts himself to the part of the science of complexity that stays firmly within a conventional discourse of competitive natural selection. Kuwabara follows Richard Dawkins' reduction of almost everything to the assumed competition between «selfish genes». He thus ignores all the rich literature both on the self-organization of systems and on the dynamics of sexual selection. 5 These are all works where the central focus is on the recursive dynamics of part that are complementary to each other, even if the term «complementary» is not central in this literature. Again, complementarity is lost and competition is celebrated.

10. Getting the ideology right

In the introduction of this paper I said that I had an ideological agenda. The language used to understand the Open Source movement works as tropes in the commercial and political promotion of Open Source. Consider first a brief repetition of the language used to understand and sell Open Source: In an article named «The wild bunch», New Scientist identify the hacker's way of programming as a «Darwinian dynamics». Raymond's speaks of «evolutionary programming» (CB), and Apple has termed their Open Source project «Darwin». In the New Scientist article, as we have also seen it in Ko Kuwabara paper, Open Source programming is «natural selection in action.» In (HtN) competition and symmetrical identity games is held to be the driving force behind the success of Open Source software.

Now, the following example is another example of Darwinism. It is taken from the introduction to the anti trust hearing at the US senate the 3. Mars 1998, where the possible marked dominance of «some software firms» were discussed. The senator who headed the hearing said:

«Most of our witnesses today have been part of successful business ventures and have obtained tremendous wealth. Two of the witnesses we will hear from today, Mr. Jim Barksdale, CEO Netscape, and Mr. Scott McNeally, CEO Sun Microsystems, are direct competitors of Microsoft. I think all of our panelists will agree that our anti trust laws are designed to protect consumers, not competitors, but I hope they will also agree that at least over the long run, consumers can be harmed when one firm is able to use its current power to prevent the successful establishment of competing new technologies that drive innovation forward.» (Downloaded from http://www.fednet.net)

This argument is based on a Darwinian and competitive model of society: The consumers are the «selection pressure» in a natural selection between commercial firms. This will supposedly «drive innovation forward».

As consumer in a competitive marked I can choose between, say, Macintosh, Windows and, if I am devoted, Linux. That's necessary and that's fine. But it is potentially something very different to be a consumer of computers and to be a hacker of computers. Raymond writes: «Users are wonderful things to have, and not just because they demonstrate that you're serving a need, that you've done something right. Properly cultivated, they can become co-developer.» (CB) Co-developed software is owned as much by the «users» of the software as by the chief developer of the program. The users are co-developers of their own software , and thus of their own (daily) life in front of the screen. That can be something quite different from being a consumer, choosing between products that tend to become more and more similar, as the symmetrical competition between the market agents tends to drive these agents to rapidly adopt each other's innovations.

But if being a consumer in a market and a co-developer in a Open Source community both can be described by the same «social Darwinism», then something seems to be lost in our descriptions.

An important difference between a consumer and a co-developer is that only the latter takes part in, and is partly defined by, two complementary processes; the cooperation of making common software, described in (CB), and the complementary identity games described in this paper. We need to make this difference clear. Darwinian selection and evolution is more than competition. Hacking is more than competition. This is true, as a matter of fact, and it is better propaganda for Open Source software. I would much rather, if I could, become a co-developer than a consumer of software.

We need to revisit Darwinism so that we can speak about evolutionary programming without conflating hacking to the marked Darwinism of the senator above. I have explored another Darwinism in this paper, by exploring sexual selection. I'll make one more detour to biology, this time to the history of the discipline, precisely to revisit Darwinism.

11. Darwin revisited

Why is natural selection so easily equated with Darwinism, while sexual selection and the even more spectacular runaway sexual selection tend to be overlooked?

In 1859 Darwin published his book on Natural Selection. It stirred up a large debate, a quarrel between Darwinists and Christian creationists that are still running high in some places of the world. It made an absolutely enormous impact, partly because Darwin got many supporters, partly because his theory fitted to an existing social ideology; the economic liberalism of Herbert Spencer and others, and partly because atheism and materialism where well established, albeit controversial, philosophical positions of the time. Some of Darwin's supporters were men of powers, men of the right kind of decency, modesty and civilized manners (Haraway 1989). In 19th century Great Britain, Europe and North America, these men had what it took be trustworthy, to talk with authority.

In 1871 Darwin published his book on sexual selection. It did not stir up the same kind of debate. Some of its conclusions were simply overlooked, especially those concerning the mechanism of sexual selection, even if his discussion of sexual selection filled 578 pages of dry but solid empirical observations. «Sexual selection» did not become «Darwinism».

Why not? Miller suggests a reason why: Because the idea that choosy females could drive the evolution, and even define maleness, profoundly disturbed those who had supported Darwin 12 years earlier; the modest men of science, power and authority. (Miller 1993: 32-36) In 19th. century Europe and North America only men could vote. Only men could choose their government. Only men were in some sense (and to the men themselves) real subjects of society. Women were closer to the objects of nature. Darwin potentially turned this upside down, even if he was speaking of birds. He made nature look like a place where females were the choosy subjects, whereas men became the object of their choice. He potentially naturalized the opposite order to that of Victorian decent society. We might say that the idea of sexual selection was too controversial, too obscure, to become controversial. So it was silenced.

It probably took someone like Darwin to propose such a theory in 1871. Charles Darwin was of good families, but he was not very concerned with being a proper man of his time. (Hylland Eriksen 1997) He preferred to stay in his garden rather than to frequent the «right» places to see and to be seen. He was not fashionable, not culturally up-to-date, not going a lot to the theaters or the opera, not reading the books that would equip him with the right kind of references -- to be able to take part in witty and intelligent bourgeois conversation. He was really quite a bit of a «nerd», there in his gardens, where he meticulously studied birds and plants. If he had been very concerned with cultivating the right kind of Victorian, British malehood, it is doubtful that he ever would have dared to make the females of nature into choosy subjects, and the males into the objects of their choice. (Or, perhaps, if he was more culturally up-to-date, he would have known how to write about sexual selection in a way that would be more catchy and acceptable to his potential allied.)

Since 1871, (male) biologists have found alternative explanations to the male ornamentation Darwin wanted to explain. Generally these ornamentations has been considered to be the sole result of symmetrical competition between males, not involving any female choice, and not involving any complementary processes. This alternative is still defended by some biologists. For example, concerning the Irish elk (see above), Richard Dawkins writes:

If a trend is seen towards increasing antler size, this is because natural selection favors larger antlers. Individual stags with large antlers have more offspring than stags with average-sized antlers, either because they survive better (unlikely) or attract females (probably irrelevant) or because they are better at intimidating rivals (likely). (Dawkins, 1993)

It seems to me that the one and only thing that makes male «intimidating rivals» more «likely», and attracting choosy females «probably irrelevant», is the premise that the only kind of relationship in nature worth mentioning is symmetrical competition. It is an old, stubborn premise from the days of 19th. century sexism. 6

So, yes, talking of evolutionary programming is fine, but let us be aware of the many kinds of relationships that make up evolution.

12. Creating loyalty to projects

In describing how Linus Torvalds made Linux, Raymond has made a list of lessons. Here is one of them:

7. Release early. Release often. And listen to your customers. (CB)

Linus Torvalds released early, buggy versions of Linux very often, sometimes more than once a day. Raymond writes:

Linus was keeping his hacker/users constantly stimulated and rewarded -- stimulated by the prospect of having an ego-satisfying piece of the action, rewarded by the sight of constant (even daily) improvement in their work. (CB)

I have described this «ego-satisfying piece of the action» as a complementary identity game. But my point is that this process does more than to satisfy an ego, as if the «ego» was given before the satisfaction of it. Identity games produce «egos», or identities. By releasing early and often, Torvalds set off a positive feedback loop that did more than producing brilliant programs. He set off an effective way to produce programmers , or, more specifically, he started a powerful process to co-produce a leader (himself) and a community of loyal followers. It is important to note that most hackers are young people, in their early twenties, often in their late adolescence, when they start their hacking career. Psychologically, this is the (later) period of identity formation. The identities and loyalties of the young hackers-to-be are initially fluid. Hooked up in highly interactive and stimulating identity games, their loyalties become fixed. They become loyal and devoted followers of Linux, Perl, Apache or KDE.

In closing this paper, it is also important to note that good Open Source programs are not made by programmers who try to be good «hackers». «Becoming someone» -- through complementary and symmetrical identity games -- is an effect of having made good programs, whereas good programs is not an effect of trying to become someone (Raymond discuss this at length in HtN). Thus, the explicit motivation that people may have for doing Open Source programming must generally be sought elsewhere than in a high degree of self-consciousness. The range of motivations is wide, spanning from making money, through the sheer fun of programing, to devoted idealism. The last motivation is not the least. The KDE desktop, for example, comes in a wide variety of languages, one of them the tiny second Norwegian language «Nynorsk» (New Norwegian). Microsoft and Apple has never bothered to translate their desktops into «Nynorsk», and the license of their desktops (Windows and Macintosh) prevent the supporters of «Nynorsk» from doing the translation themselves. The license of KDE however, together with the policy of releasing early and often as well as the entire policy of the KDE project, encourage the translation of the desktop into any kind of language, as long as some enthusiasts somewhere want to spend their long evenings doing the job. «Nynorsk» has its young, devoted practitioners and supporters, some of them within computer science. There is plenty of enthusiasm around, for a variety of causes -- causes that all have to be found locally, even if they have some shared traits -- and KDE 2.2 is the first desktop to be released in «Nynorsk».



14. References

Barth, Frederik
1975 Ritual and knowledge among the Baktaman of New Guinea, Oslo: Universitetsforlaget.

Bateson, Gregory
1972 Steps to an Ecology of Mind, New York: Ballantine Books.

Darwin, Charles
1859 On the Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life London : J. Murray.
1871 The Decent of Man, and Selection in Relation to Sex New York: D. Appleton and Company

Dawkins, R.
1993 «The evolutionary future of man» in Vol. 328, Economist , 09-11-1993, pp 87. (at www.clark.net/pub/ogas/evolution/EVPSYCH_dawkins_on_ev_future.htm )

Hannemyr, Gisle
1997 Hacking Considered Constructive, http://home.sol.no/~gisle/oks97.html

Haraway, Donna
1989 Primate Visions: Gender, Race and Nature in the World of Modern Science , New York: Routledge.

Hofstadter, Douglas R.
1979 Gödel, Escher, Bach: An eternal golden braid Hassocks: Harvester Press

Hylland Eriksen, Thomas
1997 Charles Darwin Oslo: Gyldendal

Kuwabara, Ko
2000 Linux: A Bazaar at the Edge of Chaos athttp://firstmonday.org/issues/issue5_3/kuwabara/

Latour, Bruno
1996 Aramis, or The love of technology Cambridge, Mass.: Harvard University Press

Miller, Geoffrey F.
1993 The evolution of the human brain through runaway sexual selection: The mind as a protean courtship device, Volume 1 Dep. of Psychology, Stanford University 1993
in press «A review of Sexual Selection and Human Evolution: How Mate Choice shaped Human Nature» in Handbook of evolutionary psychology: Ideas, issues and applications C. Crawford & D. Krebs (Eds.) Lawrence Erlbaum, in press.
2000 The Mating Mind: How sexual choice shaped human nature . New York: Doubleday.

New Scientist magazine
1998 «The Wild Bunch», vol 160, issue 2164, 12/12/1998, p. 42

Popper, Karl R.
1972 Objective knowledge : an evolutionary approach Oxford : Clarendon Press.

Raymond, Eric S.
(CB) The Cathedral and the Bazaar at http://www.tuxedo.org/~esr/writings/cathedral-bazaar/
(HtN) Homesteading the Noosphere at http://www.tuxedo.org/~esr/writings/homesteading/
Both papers are available in printed form:
1999 The Cathedral and The Bazaar Musings on Linux and Open Source by an Accidental Revolutionary , Beijing: O'Reilly

Sahlins, Marshall
1963 «Poor Man, Rich Man, Big Man, Chief: Political Types in Melanesia and Polynesia» in Comparative Studies in Society and History 5: 285-300


Notes


1 This is not to say that humans also seek certain kinds of stabilities, orders and predictabilities in their lives. A stable home, a stable identity, a stable history. This need not be in conflict to our enjoyment of novelties. A group of people who know each other well may keep on laughing at the same stories. A once-upon-the-time novel and funny event may be turned into a token of long lasting and stable friendships.

2 «Protean: Extremely variable, often changing, a protean policy || able to take on different shapes, after PROTEUS» (Websters Dictionary) Human beings are not the only animal to have evolved protean behaviors. Animals of prey, for example, are often good at it: The rabbit jumps in unexpected ways, to make it difficult for the fox to predict where it will run. One could ask: How did the protean behavior of human beings start, in the first place? I could always answer the question by telling a story in terms of some prior usefulness of the behavior. I will not do that here, but see (Miller 1993) for more details.

3 «Language» here «means «Chinese» and «Indian» etc., and not «body language», whereas «ideology» is more than political ideologies. It includes complex textual arrangements, such as religion and science, but more on this further down.

4 One of the first persons to propose that the human brain was a result of runaway sexual selection was the Norwegian philosopher Peter Wessel Zappfe in the 1930ies. In his pessimistic mode, he used this biological theory to explain the tragedy of being human: Over equipped with brain, human beings have lost their innocence for ever. We know we are dying. Zappfe also liked to climb mountains. Mountain climbing was to him as «meaningless as life itself».

5 Kuwabara does refer to people like Stuart Kauffman, but he does not take the implications of their work seriously, he does not really apply their theories, he merely quote them superficially.

6 Dawkins seems to have changed his opinion on this topic. On the back page of Millers book (2000) he is quoted: «one of the hottest topics in modern Darwinism. The idea that the human mind evolved as a sort of software peacock's tail has been mooted before, usually to be dismissed in favor of some alternative theory. Geoffrey miller has really been the one to run with that ball, and he now brings his ideas together in this thoughtful, witty and vividly written book. It is a work of advocacy, and Miller is beguilingly skilled advocate. But there's nothing wrong with advocacy if you are right. And I think he just may be.»