can emerge. Thus, the occurrence of call sharing at any of these scales implicates vocal learning. However, alternative mechanisms can cause similar
patterns (see below), necessitating experimental tests to demonstrate
unequivocally that vocal learning does occur. Without this experimental
proof, vocal learning can only be conjectured.
A number of experimental protocols have been used to test for both
forms of vocal learning and to identify the mechanisms involved. Perhaps
the most common and most convincing for learned acquisition are acoustic
isolation experiments and deafening (e.g., Winter et al. 1973; Heaton and
Brauth 1999). Social isolation experiments provide evidence of when social
interaction is essential to normal call acquisition (e.g., Hughes et al. 1998).
Many experiments on song learning provide either recorded or live call
tutors and then quantify the extent of acoustic similarity between tutor
and pupil, which can test either learned acquisition or social modification,
depending on the call repertoire at the start of the experimental manipulation (see Kroodsma and Miller 1996); unfortunately, this protocol has
seldom been used to study call learning. Data from experiments where
pupils copied call tutors who produced aberrant calls (including humans)
clearly demonstrate the power of social interactions (e.g., Brittan-Powell
et al. 1997). An important experimental protocol to test social modification
is to alter the social environment (i.e., create new social groups) and quantify acoustic changes. When convergence or divergence occur in novel social
groups, social modification is strongly supported (e.g., Farabaugh et al. 1994;
Boughman 1998).
In all of these observational and experimental studies, careful characterization of vocalizations is required to ensure that the acoustic parameters
measured capture the relevant variation. Determining the appropriate
measurements to make can be complex. Ideally, prior experiments have
identified the acoustic features that the animals use to discriminate among
individuals, groups, or dialects.
1.4. What Alternative Mechanisms Can Produce Patterns
Similar to those Found in Learned Vocalizations?
Alternative mechanisms such as maturation of the vocal tract, body size,
genetics, and local ecology can affect the structure and development of
vocalizations and can produce patterns similar to those found for learned
vocalizations, although these mechanisms do not involve vocal learning.
Thus, these mechanisms need to be ruled out when tests of vocal learning
are being done.
As animals mature, vocal tract morphology changes and motor control
improves. These maturational changes can result in subtle to striking
changes in vocalizations (e.g., Scherrer and Wilkinson 1993). However,
these changes have a purely biophysical basis and are not influenced by
social interactions and thus are not due to vocal learning. Age-related
changes in vocalizations can provide substantial information to listeners
142
J.W. Boughman and C.F. Moss
patterns (see below), necessitating experimental tests to demonstrate
unequivocally that vocal learning does occur. Without this experimental
proof, vocal learning can only be conjectured.
A number of experimental protocols have been used to test for both
forms of vocal learning and to identify the mechanisms involved. Perhaps
the most common and most convincing for learned acquisition are acoustic
isolation experiments and deafening (e.g., Winter et al. 1973; Heaton and
Brauth 1999). Social isolation experiments provide evidence of when social
interaction is essential to normal call acquisition (e.g., Hughes et al. 1998).
Many experiments on song learning provide either recorded or live call
tutors and then quantify the extent of acoustic similarity between tutor
and pupil, which can test either learned acquisition or social modification,
depending on the call repertoire at the start of the experimental manipulation (see Kroodsma and Miller 1996); unfortunately, this protocol has
seldom been used to study call learning. Data from experiments where
pupils copied call tutors who produced aberrant calls (including humans)
clearly demonstrate the power of social interactions (e.g., Brittan-Powell
et al. 1997). An important experimental protocol to test social modification
is to alter the social environment (i.e., create new social groups) and quantify acoustic changes. When convergence or divergence occur in novel social
groups, social modification is strongly supported (e.g., Farabaugh et al. 1994;
Boughman 1998).
In all of these observational and experimental studies, careful characterization of vocalizations is required to ensure that the acoustic parameters
measured capture the relevant variation. Determining the appropriate
measurements to make can be complex. Ideally, prior experiments have
identified the acoustic features that the animals use to discriminate among
individuals, groups, or dialects.
1.4. What Alternative Mechanisms Can Produce Patterns
Similar to those Found in Learned Vocalizations?
Alternative mechanisms such as maturation of the vocal tract, body size,
genetics, and local ecology can affect the structure and development of
vocalizations and can produce patterns similar to those found for learned
vocalizations, although these mechanisms do not involve vocal learning.
Thus, these mechanisms need to be ruled out when tests of vocal learning
are being done.
As animals mature, vocal tract morphology changes and motor control
improves. These maturational changes can result in subtle to striking
changes in vocalizations (e.g., Scherrer and Wilkinson 1993). However,
these changes have a purely biophysical basis and are not influenced by
social interactions and thus are not due to vocal learning. Age-related
changes in vocalizations can provide substantial information to listeners
142
J.W. Boughman and C.F. Moss