A `Tree' Is Not a `Train': Mistaken Analogies in Pacific Archaeology
Terrell, John Edward, Antiquity
In 1988, I wrote in ANTIQUITY on two ways of thinking about prehistory in the Pacific Islands (Terrell 1988). One is the view that the past `looks like' an enormous family tree, and each branch on this tree shows us when the forebears of a given human `race', `ethnic group', `people' or `culture' went off on their own separate journey through time and space. The other view was harder to label. I used Darwin's image of biotic communities as like `an entangled bank'. According to this view, the Pacific Islands are a geographic array of local and island-wide populations that have been more or less in touch with one another, and who have followed more or less linked historical pathways of local adaptation and change.
These two ways of thinking about the past are basically those that Robert Dewar (1995) has labelled as `trees' and `nets'; John Moore (1994) has called `cladistic' and `rhizotic' models; and Alan Templeton (1998) has characterized as the `candelabra' and `trellis' approaches to human evolution. These are contrasting ways of thinking about the past; they are not directly testable hypotheses. They are best seen as the extremes of a range of likely historical situations. Their `truth' lies in how clearly these antithetical ideas help us think about the past (Clark & Terrell 1978).
In January 1999, Nature published `Lizards took express train to Polynesia' by the zoologist Christopher Austin (1999). I had been a reviewer for Austin's manuscript. I had suggested to Nature -- obviously unsuccessfully -- that he had based his argument on a misreading of what I had written in 1988. He had deduced an hypothesis about the settlement of the central and eastern Pacific from what I had said that made little sense given what is known about the human biology, linguistics, and archaeology of the Pacific Islanders. He had attributed this supposed `entangled bank' hypothesis to me. And he reported that he had tested his idea using data on the mitochondrial DNA diversity of Lipinia noctua, a small, human-transported (`hitchhiking') lizard found on many Pacific Islands. Finding a poor fit between lizard genetics and what he thought was my hypothesis, he concluded that the genetic diversity of L. noctua `matches the predictions' of another proposition about the colonization of Oceania -- the biologist Jared Diamond's idea that the ancestors of the Polynesians had migrated from Asia to Polynesia on an `express train' or `speedboat' (Diamond 1988; 1997).
Now the antithesis of an entangled bank, net or trellis model of evolution would seem to be a family tree, clade or candelabra -- not a train or a boat. Furthermore, L. noctua is a species native to the New Guinea region. It is not found in Asia. Austin's total lizard sample (n = 29 from 15 locations) merely shows that the genetic diversity of L. noctua on New Guinea and nearby islands (where L. noctua may date back to the Tertiary) is greater than the variation he had sampled from places farther out in the Pacific -- specifically Vanuatu (n = 2), the Fiji Islands (n = 3) and Polynesia (n = 8). How on earth could his biological data tell us anything about whether -- or how quickly -- the ancestors of the Polynesians had or had not come out of Asia?
Missing the boat
I was dismayed when I learned from Nature that they had accepted Austin's manuscript. I am the only person he cites to support his claim that `the entangled-bank hypothesis argues that movement into the central and eastern Pacific was a gradual event, occurring over an extended period from different Melanesian populations' (1999: 113). I have never said that `the entangled-bank hypothesis predicts that L. noctua populations in the central and eastern Pacific should be paraphyletic [i.e., should have multiple origins] as a result of several dispersal events that reflect the different human migrations from Melanesia' (1999: 113). Therefore, I had assumed that Nature would not accept his paper. …