family groups nor groups within colonies shared call characteristics prior to
the introduction (Snowdon and Cleveland 1980). It is not clear what kind
of vocal changes would demonstrate social modification in this species
because the degree of similarity in calls does not appear to have functional
significance. A more recent experiment investigated vocal changes in
four newly formed pairs (Snowdon and Elowson 1999). No clear pattern
emerged for the four variables they measured. For each variable, some pairs
converged, others diverged, and others showed no change. The variable
most suggestive of social modification was peak frequency because three
pairs showed convergence, but in only two was it significant. Neither experiment was designed to explore call acquisition.
In 1997, Snowdon concluded that the weight of evidence does not support
a role for learning in the acquisition or social modification of vocalizations
in this or other species of primates. Snowdon and Elowson’s latest experiment (Snowdon and Elowson 1999) is consistent with Snowdon’s (1997)
position. In keeping with findings for vervet monkeys and some others
(Seyfarth and Cheney 1997), marmosets appear to learn the proper use of
alarm calls in response to predators and may learn to recognize individually
distinctive calls given by group mates (Hayes and Snowdon 1990). Learned
usage is a process distinctly different from either learned acquisition or
social modification, and we do not consider it vocal learning.
4. Dialects
Vocalizations often vary in structure geographically, sometimes with
sharp boundaries between call types. The scale of variation can be microgeographic (variation among contiguous individuals or groups) or macrogeographic (variation among groups separated geographically) (Krebs and
Kroodsma 1980). The types of geographic variation are not discrete but
form a continuum (Baker and Cunningham 1985). Geographic variation
has been termed dialects, although some authors prefer to reserve this
term for microgeographic variation resulting from vocal learning (Conner
1985). Here, we adopt a broader definition to include geographic variation
at all levels without regard to the mechanism producing such variation. We
take the latter stance primarily because the causes of geographic variation
are unknown for most mammalian vocalizations and bird calls, and also
because dialects can be biologically important even when they arise from
processes other than vocal learning. Studies of call dialects in birds and
mammals have focused primarily on describing dialectical patterns. Particularly in mammals, these studies often aim at finding evidence consistent
with vocal learning rather than determining the function of geographic
patterns.
Dialects can emerge as a natural consequence of learning vocalizations from neighbors (Catchpole and Slater 1995). Under this scenario,
4. Comparative Vocal Learning
183
the introduction (Snowdon and Cleveland 1980). It is not clear what kind
of vocal changes would demonstrate social modification in this species
because the degree of similarity in calls does not appear to have functional
significance. A more recent experiment investigated vocal changes in
four newly formed pairs (Snowdon and Elowson 1999). No clear pattern
emerged for the four variables they measured. For each variable, some pairs
converged, others diverged, and others showed no change. The variable
most suggestive of social modification was peak frequency because three
pairs showed convergence, but in only two was it significant. Neither experiment was designed to explore call acquisition.
In 1997, Snowdon concluded that the weight of evidence does not support
a role for learning in the acquisition or social modification of vocalizations
in this or other species of primates. Snowdon and Elowson’s latest experiment (Snowdon and Elowson 1999) is consistent with Snowdon’s (1997)
position. In keeping with findings for vervet monkeys and some others
(Seyfarth and Cheney 1997), marmosets appear to learn the proper use of
alarm calls in response to predators and may learn to recognize individually
distinctive calls given by group mates (Hayes and Snowdon 1990). Learned
usage is a process distinctly different from either learned acquisition or
social modification, and we do not consider it vocal learning.
4. Dialects
Vocalizations often vary in structure geographically, sometimes with
sharp boundaries between call types. The scale of variation can be microgeographic (variation among contiguous individuals or groups) or macrogeographic (variation among groups separated geographically) (Krebs and
Kroodsma 1980). The types of geographic variation are not discrete but
form a continuum (Baker and Cunningham 1985). Geographic variation
has been termed dialects, although some authors prefer to reserve this
term for microgeographic variation resulting from vocal learning (Conner
1985). Here, we adopt a broader definition to include geographic variation
at all levels without regard to the mechanism producing such variation. We
take the latter stance primarily because the causes of geographic variation
are unknown for most mammalian vocalizations and bird calls, and also
because dialects can be biologically important even when they arise from
processes other than vocal learning. Studies of call dialects in birds and
mammals have focused primarily on describing dialectical patterns. Particularly in mammals, these studies often aim at finding evidence consistent
with vocal learning rather than determining the function of geographic
patterns.
Dialects can emerge as a natural consequence of learning vocalizations from neighbors (Catchpole and Slater 1995). Under this scenario,
4. Comparative Vocal Learning
183