5. Acoustics and Social Behavior
249
speed, and size with their echolocation, but insect echoes contain insectspecific information of the wing beat frequency, length, types, and structure,
providing prey-species information (Kober 1988; von der Embe 1988). In
this case, information is encoded within time intervals and changes in intensity. Could such information about prey or conspecifics be encoded in the
dolphin's environment or from ensonified objects? Do dolphins hear fish
buried in the bottom by their movement, and can they determine whether
the fish is a preferred species or size by such signatures? Roitblat et al.
(1995) present data that suggest dolphins exploit frequency and time
domain information and make fine discriminations of buried objects based
on high-amplitude components in signals.
The decoding of conspecific social signals is also a puzzle. Signature
whistles have long been thought to contain individually specific information. Repetitive broadcasts of "signature" whistles may serve one such function. However, other possible levels of information contained in a whistle
include species identity (Wang et al. 1995), family or pod dialect (Ford
1991), and emotive states (Caldwell and Caldwell 1967). Conspecifics may
also use their own signature whistle to contact another member of their
group. If dolphins can discriminate between the owner of a "signature"
whistle and a conspecific using the same whistle, other information such as
"timbre" or voice qualities, or at least contextual information about the
sender, must be available. Identification features need not be limited by
contour shapes of whistles, as evidenced by click codas (Watkins and
Schevill 1977) and signatures contained in echolocation trains (BarrettLennard et al. 1996). Nor has the signature or "voice" information been
seriously looked at within the diverse category of delphinid burst-pulsed
vocalizations. Burst-pulsed sounds contain and modulate the same acoustic
parameters utilized by individually unique rhesus monkey copulation calls
(Hauser 1993). Such information may be more available to conspecifics
than we realize, making the listening and hearing of a social exchange of
vocalizations rich in its potential complexity of extracted information.
Embedding of information may be encodable through the modulation
of acoustic parameters and decodable by conspecifics, and possibly by
predators.
The structural approach in determining relevant units of analysis in communication systems involves choosing between multiple acoustic parameters, including duration, amplitude, pitch, contour, and shape, and applying
the appropriate analysis methods. The number of possibly relevant parameters is often infinite and undefined, but methods in hardware and software for analogous acoustic and visual signal analysis exist and have been
used for a multitude of species (Clark et al. 1987; Hauser 1996).
It is critical that we begin to think about the perceiver's perspective
rather than simply the production features of a signal. Subtle modulations
of gross signals allow additional information to be communicated or at least
detected. Modulated prosodic features such as spacing between signals can
249
speed, and size with their echolocation, but insect echoes contain insectspecific information of the wing beat frequency, length, types, and structure,
providing prey-species information (Kober 1988; von der Embe 1988). In
this case, information is encoded within time intervals and changes in intensity. Could such information about prey or conspecifics be encoded in the
dolphin's environment or from ensonified objects? Do dolphins hear fish
buried in the bottom by their movement, and can they determine whether
the fish is a preferred species or size by such signatures? Roitblat et al.
(1995) present data that suggest dolphins exploit frequency and time
domain information and make fine discriminations of buried objects based
on high-amplitude components in signals.
The decoding of conspecific social signals is also a puzzle. Signature
whistles have long been thought to contain individually specific information. Repetitive broadcasts of "signature" whistles may serve one such function. However, other possible levels of information contained in a whistle
include species identity (Wang et al. 1995), family or pod dialect (Ford
1991), and emotive states (Caldwell and Caldwell 1967). Conspecifics may
also use their own signature whistle to contact another member of their
group. If dolphins can discriminate between the owner of a "signature"
whistle and a conspecific using the same whistle, other information such as
"timbre" or voice qualities, or at least contextual information about the
sender, must be available. Identification features need not be limited by
contour shapes of whistles, as evidenced by click codas (Watkins and
Schevill 1977) and signatures contained in echolocation trains (BarrettLennard et al. 1996). Nor has the signature or "voice" information been
seriously looked at within the diverse category of delphinid burst-pulsed
vocalizations. Burst-pulsed sounds contain and modulate the same acoustic
parameters utilized by individually unique rhesus monkey copulation calls
(Hauser 1993). Such information may be more available to conspecifics
than we realize, making the listening and hearing of a social exchange of
vocalizations rich in its potential complexity of extracted information.
Embedding of information may be encodable through the modulation
of acoustic parameters and decodable by conspecifics, and possibly by
predators.
The structural approach in determining relevant units of analysis in communication systems involves choosing between multiple acoustic parameters, including duration, amplitude, pitch, contour, and shape, and applying
the appropriate analysis methods. The number of possibly relevant parameters is often infinite and undefined, but methods in hardware and software for analogous acoustic and visual signal analysis exist and have been
used for a multitude of species (Clark et al. 1987; Hauser 1996).
It is critical that we begin to think about the perceiver's perspective
rather than simply the production features of a signal. Subtle modulations
of gross signals allow additional information to be communicated or at least
detected. Modulated prosodic features such as spacing between signals can