However, female preferences for calls within the conspecific frequency
range are also temperature-dependent (Gerhardt and Doherty 1988).
Because these temperature dependencies scale differently under similar
temperature ranges, they are likely to be controlled by different neural
mechanisms (see van Dijk et al. 1997). Of course, genetic coupling and
coevolution are not mutually exclusive.
2.3.2. Motor-Sensory Linkages in Learned Signals
Linkages between production and perception can also result from developmental changes that selectively modify phenotypic patterns. A compelling example is the development of human speech. It has long been
known that human infants can distinguish many of the phonetic contrasts
in speech, even those in languages to which they have not been exposed
(e.g., Lasky et al. 1975). Adult speakers of different languages, on the other
hand, not only differ from one another in the location of perceptual boundaries between phonetic contrasts but also lose altogether the ability to
perceive some nonnative contrasts (e.g., Best et al. 1988; Logan et al. 1989).
The loss of sensitivity to selected nonnative contrasts occurs somewhere
near the end of the first year (Werker and Tees 1984).
The developmental changes in speech perception are roughly matched
to developmental changes in vocal production, and such findings have been
taken to support the theory (Liberman et al. 1967) that speech perception
is guided by a process that matches speech sounds to the vocal gestures
required to produce those sounds. This is the so-called “motor” theory of
speech perception. As with the genetic-coupling hypothesis, in motor
theory, the same pattern generators participate in perception and production. The modern version of the motor theory of perception (Liberman and
Mattingly 1985) stresses representational modularity as conceived by Fodor
(1983).
Psychologists have staged a series of strong assaults on the motor theory
(e.g., Lindblom 1991; Fowler 1996; Nearey 1997). Because the motor theory
of speech perception appears to make clear predictions regarding the
recruitment of specific motor pathways in speech perception, it would
appear possible to test predictions of motor theory with neurobiological
techniques. Unfortunately, instantiations of the motor theory of speech
perception do not specify the neural mechanisms implied by requisite
processes such as vocal gesture, information encapsulation, and so forth.
Absent these precise definitions, motor theory is not falsifiable at a neurobiological level. In addition, experimental neuronal data from humans with
the spatial and temporal precision required to test theories of perception
such as motor theory cannot yet be obtained.
The comparative approach can help resolve questions that might otherwise be difficult to resolve with the available human data. Humans are phylogenetically isolated with regard to the central feature of interest, vocal
344
T.Q. Gentner and D. Margoliash
range are also temperature-dependent (Gerhardt and Doherty 1988).
Because these temperature dependencies scale differently under similar
temperature ranges, they are likely to be controlled by different neural
mechanisms (see van Dijk et al. 1997). Of course, genetic coupling and
coevolution are not mutually exclusive.
2.3.2. Motor-Sensory Linkages in Learned Signals
Linkages between production and perception can also result from developmental changes that selectively modify phenotypic patterns. A compelling example is the development of human speech. It has long been
known that human infants can distinguish many of the phonetic contrasts
in speech, even those in languages to which they have not been exposed
(e.g., Lasky et al. 1975). Adult speakers of different languages, on the other
hand, not only differ from one another in the location of perceptual boundaries between phonetic contrasts but also lose altogether the ability to
perceive some nonnative contrasts (e.g., Best et al. 1988; Logan et al. 1989).
The loss of sensitivity to selected nonnative contrasts occurs somewhere
near the end of the first year (Werker and Tees 1984).
The developmental changes in speech perception are roughly matched
to developmental changes in vocal production, and such findings have been
taken to support the theory (Liberman et al. 1967) that speech perception
is guided by a process that matches speech sounds to the vocal gestures
required to produce those sounds. This is the so-called “motor” theory of
speech perception. As with the genetic-coupling hypothesis, in motor
theory, the same pattern generators participate in perception and production. The modern version of the motor theory of perception (Liberman and
Mattingly 1985) stresses representational modularity as conceived by Fodor
(1983).
Psychologists have staged a series of strong assaults on the motor theory
(e.g., Lindblom 1991; Fowler 1996; Nearey 1997). Because the motor theory
of speech perception appears to make clear predictions regarding the
recruitment of specific motor pathways in speech perception, it would
appear possible to test predictions of motor theory with neurobiological
techniques. Unfortunately, instantiations of the motor theory of speech
perception do not specify the neural mechanisms implied by requisite
processes such as vocal gesture, information encapsulation, and so forth.
Absent these precise definitions, motor theory is not falsifiable at a neurobiological level. In addition, experimental neuronal data from humans with
the spatial and temporal precision required to test theories of perception
such as motor theory cannot yet be obtained.
The comparative approach can help resolve questions that might otherwise be difficult to resolve with the available human data. Humans are phylogenetically isolated with regard to the central feature of interest, vocal
344
T.Q. Gentner and D. Margoliash