Animal Signals

John Maynard Smith's Conclusion

His final letter of 7 September 2000, summarized in my own words — followed by my reply

By September 2000 our exchange had settled into agreement about the biology, with one difference remaining about terminology. In his final letter John summarized where we had arrived. What follows is my summary of his conclusions, in my own words, with a few of his own phrases kept in quotation.

John accepted that we agree about the evolutionary mechanisms – but may disagree about which words to attach to them. He restated my use of "amplifier" for what it is: a signal that helps a receiver perceive the quality of the signaler, with two features – it makes differences in quality easier to perceive, and it is hard to fake. He saw that the term reflects how such a signal widens the fitness differences (in my terms, the S-component, i.e., potential benefits by signal recipients) between high- and low-quality signalers, and he credited my 1992 model, with Cohen and Shmida, for explaining why even low-quality individuals produce the signal despite its cost.

He distinguished this from signals that merely make quality appear visually larger – arching the back, lowering the larynx (and I can add here – standing fully erect) – which are unreliable when they first appear but become reliable once fixed in the population: the category I had labeled "index."

His central point was about meaning and priority. When he and Harper introduced "index" in 1995, he wrote, they had used it in almost precisely the sense in which I had earlier used "amplifier." Their emphasis had been on the "cannot be faked" property, in order to separate such signals from handicaps. Had he recognized at the time that my "amplifier" already covered this, he would have noted the equivalence – though he explained that he would still have introduced "index," because it carries an established sense in semiotics, "signifying in virtue of a causal relation between sign and object." For him, "amplifier" did not by itself convey "hard to fake."

He closed by noting that no terminology could be made "official," that he hoped at least not to add to the confusion, and that "our exchange of letters has been very helpful to us."

My reply — 17 September 2000

Dear John,

Thanks for the thorough correspondence. It looks like your book will be an important contribution to this field, and to eliminate confusion that has been around for quite a while.

You've got it right about the thing that is amplified (i.e., the signaller's S-component).

My 1989 paper, JTB, was the first genetic model (a paper that came out of my dissertation, 1987, at the Univ. of Arizona with R. Michod) of amplifiers. And yes, the main puzzle, of course, was why should low quality males signal (I have mentioned later in my papers the question of what evolves first, the signal or the quality – this is important for the understanding of why should poor quality signals have fully developed signals). However, the 1989 papers was a sexual selection model. In the 1992 paper (with Cohen and Shmida) I aimed to generalize this concept to other contexts of communication, and to look at the potential of attenuators to evolve.

I have used the term "index" the first time in my 1997 paper (general model).

I wish you all the luck, John.

Cheers,
Oren Hasson