about the caller’s age and thus may have important functions, even though
learning is not involved.
Body size is an important determinant of vocalization frequency in a
number of taxa (e.g., Hauser 1993; Fitch and Hauser, Chapter 3). Individuals of similar body size often produce vocalizations of similar frequency;
individuals of different size produce vocalizations that differ in frequency.
Larger individuals typically produce lower-frequency vocalizations. The
effects of body size on call frequency can profoundly influence consequent
function of vocalizations and can contribute to individual distinctiveness
and to signature function. Size dimorphism can be part of the basis for
sex differences in vocalization frequency (e.g., Gouzoules and Gouzoules
1989b, 1990). Combined with maturational effects, body size can contribute
to age-specific patterns in vocalizations (Gouzoules and Gouzoules 1995).
Dialects can occur as a consequence of populations differing in body size
because of adaptation to local environments or genetic differences (Barclay
et al. 1999). These differences might be very important for identifying individuals, for classifying social partners and opponents, and even for deciding
who to mate with, but vocal learning through social interaction need not be
involved. Body size can have large effects on individual fitness; thus, finding
that calls indicate body size is an important discovery, even when calls are
not learned.
Social groups are often organized along kinship lines. Genetic variation
among social groups can generate parallel variation in vocalization characteristics, facilitating recognition of kin (e.g., Rendall et al. 1996). Kin recognition based on vocalizations may function importantly in many aspects of
social behavior. However, vocal learning is not necessary to produce these
patterns and may even be costly. Learning can introduce copy errors that
can decrease similarity among kin. Additional error can result if the individuals who serve as the tutors are unknowingly unrelated. Vocalizations
open to learning might be more easily copied by nonkin who usurp a
group’s resources through their cheating. Selection against cheaters can
favor vocalizations that directly reflect underlying genetics—a type of
relatedness marker that is uncheatable. Therefore, when groups consist of
relatives, genetic explanations need to be carefully considered and tested
before vocal learning can be inferred. This is especially true when there is
no previous experimental evidence demonstrating that the species under
consideration learns vocalizations. In these situations, vocal learning might
be a viable hypothesis, but we cannot be sure it occurs without direct
experiments that control for the influence of relatedness. Even when vocalizations are not learned, finding that individuals use heritable vocalizations
to recognize kin would be extremely important (e.g., Grafen 1990; Medvin
et al. 1992).
Dialects can also result from processes that do not involve vocal learning. For instance, species distributions can be disjunct because of geographic
barriers that reduce matings between populations, creating population substructure. Sustained isolation can give rise to sufficient genetic differentia4. Comparative Vocal Learning
143
learning is not involved.
Body size is an important determinant of vocalization frequency in a
number of taxa (e.g., Hauser 1993; Fitch and Hauser, Chapter 3). Individuals of similar body size often produce vocalizations of similar frequency;
individuals of different size produce vocalizations that differ in frequency.
Larger individuals typically produce lower-frequency vocalizations. The
effects of body size on call frequency can profoundly influence consequent
function of vocalizations and can contribute to individual distinctiveness
and to signature function. Size dimorphism can be part of the basis for
sex differences in vocalization frequency (e.g., Gouzoules and Gouzoules
1989b, 1990). Combined with maturational effects, body size can contribute
to age-specific patterns in vocalizations (Gouzoules and Gouzoules 1995).
Dialects can occur as a consequence of populations differing in body size
because of adaptation to local environments or genetic differences (Barclay
et al. 1999). These differences might be very important for identifying individuals, for classifying social partners and opponents, and even for deciding
who to mate with, but vocal learning through social interaction need not be
involved. Body size can have large effects on individual fitness; thus, finding
that calls indicate body size is an important discovery, even when calls are
not learned.
Social groups are often organized along kinship lines. Genetic variation
among social groups can generate parallel variation in vocalization characteristics, facilitating recognition of kin (e.g., Rendall et al. 1996). Kin recognition based on vocalizations may function importantly in many aspects of
social behavior. However, vocal learning is not necessary to produce these
patterns and may even be costly. Learning can introduce copy errors that
can decrease similarity among kin. Additional error can result if the individuals who serve as the tutors are unknowingly unrelated. Vocalizations
open to learning might be more easily copied by nonkin who usurp a
group’s resources through their cheating. Selection against cheaters can
favor vocalizations that directly reflect underlying genetics—a type of
relatedness marker that is uncheatable. Therefore, when groups consist of
relatives, genetic explanations need to be carefully considered and tested
before vocal learning can be inferred. This is especially true when there is
no previous experimental evidence demonstrating that the species under
consideration learns vocalizations. In these situations, vocal learning might
be a viable hypothesis, but we cannot be sure it occurs without direct
experiments that control for the influence of relatedness. Even when vocalizations are not learned, finding that individuals use heritable vocalizations
to recognize kin would be extremely important (e.g., Grafen 1990; Medvin
et al. 1992).
Dialects can also result from processes that do not involve vocal learning. For instance, species distributions can be disjunct because of geographic
barriers that reduce matings between populations, creating population substructure. Sustained isolation can give rise to sufficient genetic differentia4. Comparative Vocal Learning
143