| Human Nature Review ISSN 1476-1084 | Table of Contents | What's New | Search | Feedback | Daily News | Submit A Manuscript |
PDF of this article
Download Adobe Acrobat Reader
Email the reviewer
Reviewer's web site
Search for papers by Barabasi, A.
Author's web site
Publisher's web site
Send a response to this article
Search the web for related items
Contact the Editors

The Human Nature Review Human Nature Review  2002 Volume 2: 312-316 ( 19 August )
URL of this document http://human-nature.com/nibbs/02/linked.html

Essay Review

Linked: Barabasi ratifies Kauffman, demotes the social gene

By James Brody

A review of Linked: The New Science of Networks
By Albert-Laszlo Barabasi
NY: Perseus, 2002, 229 pp + 50 pp of notes & index.*

Sidney Redner reviewed Linked in the July 11th issue of Nature, p. 127-128: "...Barabasi presents an entertaining introduction to this vital field at a level that is generally accessible to the layperson interested in modern science." True enough. I, however, would go a large step further: Linked offers many heuristic possibilities if your interests are in genetic, neural, electronic, or social organizations. Much of the existing literature marvels at the beauty of complex organizations but doesn't take us past awe and eye candy. Yes, emergence exists but by what rules? Barabasi describes simple rules that may have tremendous power.

Nodes, hubs, and links

We already know that large things emerge from small things; larger things connect to each other in networks rather than in amorphous clouds. Species, molecules, professional societies, words, and home pages on the Internet all share important network properties.

First, there are nodes and connections between them, called links. Nodes can represent individual people, a business, a web page, or an amino acid. Links may consist of telephone lines, letters, Internet connections, or exchanges of substrates, products, or neurotransmitters...it doesn't matter because similar functional properties occur regardless of the participants or how many there are.

Second, the frequency of links for the nodes in random networks can be described by bell curves that have a mean and a spread of values, often symmetrical, on either side. Power laws, however, describe emergent networks like those found in living cells and in friendships. That is, a few nodes (about 20%) have many links but most nodes (about 80% of them) have only one or two links. The nodes with many connections are known as hubs.

Hubs make it possible to add massive numbers of nodes to the network with little effect on the number of connections needed to get from any one node to any other. These properties are scale invariant. For example, Google.com has billions of links and is a hub. The molecule ATP, followed by ADP and water, are hubs in cells. Wilt Chamberlain, a basketball player, claims to have averaged 1.2 different sex partners per day or about 20,000. He is either a liar or a hub.

Forty-three species were mapped for the average number of links between chemicals in their cells. Regardless of species size, but as Kauffman (1995) might have predicted, the number was approximately 3: a few molecules participate in the majority of reactions but most participate in only one or two. (According to Nature Science Update, July, 11, 2002, any two English words are connected on the average by only three degrees of separation!) Species are generally connected by 2 links on food chains. Milgram found that about 6 links separate any two Americans, scientists between 4 and 6, and 19 links will take you from any web page to any other (Barabasi, 2002). (If anything, the web is overly complex in comparison with a cell or an ecosystem and may reflect the absence of selection pressure.)

Fitness, number of links, and rate of acquisition

"Fitness" in networks refers to the number of links connected to a node: it is the number of links that determine your opportunities. There is also a "fitness connectivity product" for nodes: the number of links at a particular time multiplied by the rate at which the node recruits new links. This idea is similar to ones in biology when fitness can be either the contribution of a genotype to the next generation or the rate of increase of a genotype across generations (Keller, 1992, p. 120-121).

For example, Radio host Jordan Rich described his guest, Forry Buckingham, as handsome and marveled at the film and commercial success of this Boston actor. Buckingham commented that he enjoys working with lots of different people and, when doing a commercial, can manufacture a convincing relationship with a total stranger in 20 seconds (5/18/02, Jordan Rich Show, approx. 2 am, radio station WBZ, AM1030 Boston). Buckingham appears to be a good hub, one with a high fitness connectivity product.

"Fitness" in biology can apply to a lineage of organisms, to an individual, or to a specific gene scattered through relatives (Keller & Lloyd, 1992). In any case, the computations are difficult and often a matter for faith as much as for manipulation and measurement. In electronic networks, you simply increase or remove links to change the fitness of a hub: the relationships are apt to parallel those you might find between species on an island.

Evolvability: hubs and links between niches and occupants

First, evolvability describes the ease with which core structures add programs and options (Gerhart & Kirschner, 1997). Metazoans exploited small differences in niches by changing their own form more quickly than they could modify the niche itself. It can be said that metazoans kept their biochemical and genetic hubs but acquired bewildering variations in their appearances and conduct. Second, rates of evolutionary change may increase over generations. Thus, Gerhart and Kirschner (1997) submit that evolvability itself acquires greater speed and flexibility. Instinct, learning, and cultures are simply the latest finesses on evolvability, ones that allow us to modify our niche very quickly and no longer change our form. Third, there may be selective advantages to organisms when they build, maintain, and improve a niche cooperatively. Each improvement in the niche becomes a new platform for further refinements of both its occupants and itself (Laland, et al., 2000).

If evolution consists of responding to and managing a wider range of conditions, then evolvability is defined by the number of niches occupied and the rate at which new ones can be acquired. Thus, Gerhart & Kirschner's concept becomes identical to that of a fitness connectivity product. Not only did metazoans link to a lot of environments but they also recruited them at tremendous speed, a mere 600 million years in comparison with the 3 billion years of comparative stasis seen with prokaryotes. Humans accelerate that trend.

The concept of evolvability sets an unusual precedent: we compare metazoans not in terms of their direct competition with each other for a specific niche but in the number of niches they penetrate and their speed of doing so. Metazoans have not only adjusted to a wide range of environments but now take charge of them on a scale not seen since single-celled organisms catalyzed an oxygen shell for earth (Lovelock, 1979/1995; Turner, 2000; Wilson, 2002). Thus, links are established not only between organisms but also between organisms and settings. If this step can be taken, we can take the reciprocity models that have been applied to conspecifics, parasites, and mutualists and extend them to the relationships between organisms and their settings. Ridley (1996) gives examples of sustained relationships between a small human group and their water and fishing resources. His observations could also apply to farming, herding, and even mining. The trick in all of these situations may lie in permanency: cut back migration and we treat some of our cousins far better.

Spreading ideas or shutting them down

New medicines are marketed by delivering literature and supplies to physicians who are hubs, who are socially connected not only to lots of patients but also to lots of other physicians. A similar tactic applies to book sales, toy sales, hit records, professional recognition, and movie careers: use the hubs and you reach massive numbers of people with less investment.

Hubs are both useful to a network invader but the hub itself may be vulnerable through displacement. That is, a second node may replace the current owner of those links. Barabasi (2002) mentions the displacement of Newton by Palm Pilot and De Havilland by Boeing. In these situations, the fitness connectivity product takes on a valuable if sinister meaning. Your total number of links is less impressive if an outsider recruits links at a higher rate than you can.

Hubs are also important factors in attacking or protecting the network itself. Both emergent and random networks resist random attacks by viruses or by humans: many nodes can be destroyed with little effect on the remaining network. Emergent networks, however, are vulnerable to selective attacks on hubs.

Disease and terrorism spread by infecting or destroying hubs. For example, AIDS is carried by many people but the epidemic is sustained by hubs, people who have many sexual partners. (One AIDS carrier estimated that he had 250 different partners each year.) Plagues and social movements are hot coals that often go out when separated. Wall off or cure the hub and the epidemic cannot maintain itself.

Isolates vs. hubs

While a number of authors (Wilson, 1975; Margulis and Sagan, 1997, 2002; Sober & Wilson, 1998; Wright, 2000; Bloom, 2000) consider the formation of larger organizations from smaller ones, a drift identified with cooperation, symbiosis, or "group selection," singletons and isolates still abound at every organizational level. We can attribute them to "variation" and a form of developmental error or we may consider, understand, and eventually appreciate the conditions that favor independent players. Networks may offer hints of the conditions under which independent players might show adaptive advantages. (They may also simplify computerized analyses of varied reciprocity tactics, analyses now conducted with spurious assumptions of randomized contacts between participants.)

Hubs sometimes "eat" other hubs as in the case of corporate take overs. Black holes are perhaps the ultimate example of a hub that devours everything else. The concept fits governments as well as health maintenance organizations, entrepreneurs, bullies, and rapists. Once, however, a hub is completely dominant, evolvability diminishes and evolution stumbles. Development stalls in the presence of too much order (Kauffman, 1993; Maynard Smith & Szathmary, 1999). Variation, whether in social isolates or in suppressed clusters of DNA (Quiest et al, 2002) , might be insurance against sudden environmental changes that overwhelm the group's typical strategies.

Grounding the social gene

We commonly tip our hats to the idea that genes are socialists, each of them synchronized with every other gene (Mayr, 1964; Dawkins, 1976/1989; Gerhart & Kirschner, 1997). On the one hand, it is true that genes switch on and off in response to each other's activity and to outside experiences, activating a different set of genes from moment to moment. Genes travel in multiple sets so that if one set cannot manage a toxin, another set becomes active. All of this is old news, however, stories that we inferred some 30 years ago.

Today's headlines are different. First, one "gene" does not talk directly to every other gene: 80% of the activity will be accomplished by less than 20% of the genes. Second, the average number of interconnections between genes will probably be 3 or fewer. Third, decisions are decentralized and modified according to local conditions. Fourth, the popularity of a node may give some clue to its metabolic or developmental importance. Thus, the paths between genes will probably be linear, of small size, and easily defined.

The same argument holds also for neurons and their organizations.

For once, the words, "new science," on a book cover may be deserved and consilience may emerge not from biology but from physics. Thus, you should read this book and loan it to no one. It is NOT merely an arm decoration for befuddled grad students who cannot get a date.

Even sociologists like it.


* Most of this material was adapted from J. F. Brody (2002) If Darwin had been a woman: alternatives to the received view of evolution. Poster session with Howard Bloom and J. Scott Turner, presented at the Human Behavior and Evolution Society annual meeting, Rutgers, NJ, June 19-22. Full manuscripts available on request.


Bloom, H. (2000) Global Brain: The Evolution of Mass Mind from the Big Bang to the 21st Century. NY: Wiley.

Dawkins, R. (1976/1989) The Selfish Gene. New York: Oxford.

Gerhart, John & Kirschner, Marc (1997) Cells, Embryos, and Evolution. Malden, MA: Blackwell.

Kauffman, S. (1993) Origins of Order: Self-Organization and Selection in Evolution. NY: Oxford.

Kauffman, S. (1995) At Home in the Universe: The Search for the Laws of Self Organization and Complexity. NY: Oxford.

Keller, E. (1992) Fitness: reproductive ambiguities. In E. Keller & E. Lloyd (Eds.) Keywords in Evolutionary Biology. Cambridge, MA: Harvard University Press, pp. 120-121.

Keller, E. & Lloyd, E. (Eds.) Keywords in Evolutionary Biology. Cambridge, MA: Harvard University Press.

Laland, K. N., Odling-Smee, F. J., & Feldman M. W. (2000) Niche construction, biological evolution and cultural change. Behavioral and Brain Sciences. 23(1): 131-146.

Lovelock, J. (1979/1995) Gaia: A New Look at Life on Earth. NY: Oxford.

Margulis L. & Sagan D. (1997) Slanted Truths: Essays in Gaia, Symbiosis, and Evolution. NY: Springer-Verlag.

Margulis L. & Sagan D. (2002) Acquiring Genomes: A Theory of the Origins of Species. NY: Basic.

Maynard Smith, J. & Szathmary, Eors (1999) Origins of Life: From the Birth of Life to the Origin of Language. NY: Oxford.

Mayr, E. (1964) From molecules to organic diversity. In E. Mayr (Ed.) , 1976, Evolution and the Diversity of Life: Selected Essays, Cambridge, MA: Belknap, pp. 64-72.

Quitsch, C, Sanster, T A, & Lindquist, S. (2002) Hsp90 as a capacitor of phenotypic variation. Nature, 417, 618-624.

Ridley, M (1996) The Origins of Virtue: Human Instincts and the Evolution of Cooperation. NY: Penguin.

Sober, E. & Wilson, D.S. (1998) Unto Others: The Evolution and Psychology of Unselfish Behavior. Cambridge, MA: Harvard University.

Turner, J. Scott (2000) The Extended Organism: The Physiology of Animal-Built Structures. Cambridge, MA: Harvard University Press.

Wilson, E. O. (1975/2000) Sociobiology: The New Synthesis. Cambridge, MA: Belknap, Harvard University Press.

Wilson, Edward O. (2002) The Future of Life. NY: Knopf.

Wright, R. (2000) Nonzero: The Logic of Human Destiny. NY: Pantheon.

Buy Linked: The New Science of Networks from Amazon United States of America Amazon.com  Amazon United Kingdom Amazon.co.uk  Amazon France Amazon.fr  Amazon Deutschland Amazon.de  Amazon Japan Amazon.co.jp Amazon Canada Amazon.ca

Computer-generated translation of this page French français German deutsch Spanish español Portuguese português Italian italiano Russian Russian JapaneseJapanese Chinese (Traditional) Chinese (Traditional)Arabic Arabic― also try this alternative fast translation service.

© James Brody.


Brody, J. (2002). Linked: Barabasi ratifies Kauffman, demotes the social gene. Human Nature Review. 2: 312-316.

US -

Amazon.com logo

UK -

Amazon.co.uk logo

The Human Nature Review