The Theory of Honest Signalling

This is an old version of the site. Unless you have some particular reason to want the old veevrsion, you'll probably do best to head over to the
new one.


Introduction: Part 1

Introduction: Part 2

The Basic Problem

The Basic Solution

Honest signalling in biology

Zahavi's handicap principle

Grafen's model

Attracting mates

Begging for food

Deterring predation

Contesting resources

Autumn color

Honest signalling in economics

Conspicuous consumption


The mathematics of honest signalling

Signalling as a game


Other resources

Carl T. Bergstrom

Using Mathematica

Contact Information

Department of Zoology
University of Washington
Box 351800
Seattle, WA 98195-1800

Honest signalling among rivals:
Contests over resources

How can animals use signals when contesting resources?

Biologists have long marvelled at how seldom animal conflicts over food, territories, mates, or other resources escalate to all-out fighting. Prior to unrestrained aggression, animals often engage in extended bouts of signalling, making prominant display of their weapons (e.g. antlers) or running through a reportoire of highly-stereotyped aggression signals. These displays are presumably designed to convey information about the aggressive intent and/or the fighting ability of the participants. But why should these signals be reliable? Why don't opponents lie, and pretend to be the "tough guy" all the time? At least three closely related arguments have been proposed.

First, prominant signals of fighting ability may be impossible, or impossibly costly, to fake. For example, a full-grown bull elk's rack of antlers may weigh in excess of 40 pounds - far more than a young, weak, or sickly individual could hope to carry. Costly traits of this sort would presumably function in a manner similar to sexual selection ornaments.

Second, Zahavi (1975) proposed that gestures or displays which put an individual at a relative disadvantage if attacked, such as exposing a flank, may serve as reliable signals of fighting ability. If I want to show an opponent that I am not frightened by him, one approach is to turn my back and "dare him" to try a sucker punch. Taking a risk like this can be an honest signal, because a stronger fighter can afford to send it, whereas a weaker fighter could be knocked out right away if he tried this tactic.

Third, some signals involved in competitive or aggressive interactions may be produced without directly imposing a cost or increased risk on the signaller. These signals - such as the dark throat-patch on a dominant male sparrow - are called ``conventional signals'' or ``badges or status.'' Such signals are thought to remain honest because of their indirect consequences: signal receivers will severely punish those who are caught sending misleading signals. For example, Rohwer (1977) and Moeller (1987) found that when a sparrow's throat badge was artificially augmented with a dark dye, making him appear more dominant than he truly was, other sparrows soon saw through the deception and vicously attacked the dyed individual.


Geist (1966) and Zahavi (1975) were among the first to consider unfakable siganls of fighting ability. In the same 1975 paper, Zahavi also proposed that risks could serve as reliable signals. Adams and Mesterton-Gibbons (1995) provided a formal model of this scenario; interestingly, they find that some bluffing is likely at equilibrium. Rohwer (1975) proposed that variable plumage characters serve as "badges of status" in several bird species. Enquist (1985) and Maynard Smith and Harper (1988) published two of the first game-theoretic models of aggressive signalling. Hurd (1997) highlighted the distinction between costly signals, risks, and badges of status.

[ Previous Page ] [ Next Page ]

Last modified September 4, 2002
Copyright © 2002 Carl T. Bergstrom