ANIMAL SIGNALS - MODELS AND TERMINOLOGY
Citation
Jm. Smith et Dgc. Harper, ANIMAL SIGNALS - MODELS AND TERMINOLOGY, Journal of theoretical biology, 177(3), 1995, pp. 305-311
Categorie Soggetti
Biology Miscellaneous
SICI code
0022-5193(1995)177:3<305:AS-MAT>2.0.ZU;2-Q
Abstract
Three classifications of signals are discussed. The first is based on
the type of information that they convey. A 'self-reporting signal' pr
ovides information about some property of the signaller, while an 'oth
er-reporting signal' conveys information about an object or organism o
ther than the signaller. The second classification is based on the pro
cesses that maintain the reliability of the signal. A 'minimal signal'
is one whose cost is no greater than that needed to transmit the info
rmation effectively; such signals can only be reliable if signaller an
d receiver place the possible outcomes of the interaction in the same
rank order. A 'cost-added signal' is a signal whose cost is greater th
an that required to transmit the information. The intensity of the sig
nal is correlated with some quality of interest to the receiver. This
correlation exists because, in the past, receivers have responded diff
erently to signals of different intensity. An 'index' is a signal in w
hich there is, again, a correlation between intensity of signal and qu
ality, but one which does not depend on past coevolution, but on physi
cal necessity. For example, if size is correlated with fighting abilit
y, a signal conveying information about size is an index; in species i
n which the male feeds the young, courtship feeding is an index of his
ability to do so. Since in both cost-added signals and indices, there
is a correlation between signal intensity and some relevant quality o
f the signaller, both are examples of evaluation signals; this term ca
n be used when the mechanism responsible for the correlation is uncert
ain. A third classification of signals is based upon the relationship
between their form and that of the object being signalled about. A 'sy
mbol' is arbitrarily linked with its object, an 'icon' resembles its o
bject and an 'index' is physically related to its object. It seems unl
ikely that many, if any, biological signals are symbols. (C) 1995 Acad
emic Press Limited