Abstract: East African vervet monkeys give short (125 ms), harsh-sounding grunts to each other in a variety of social situations: when approaching a dominant or subordinate member of their group, when moving into a new area of their range, or upon seeing another group. Although all these vocalizations sound similar to humans, field playback experiments have shown that the monkeys distinguish at least four different calls. Acoustic analysis reveals that grunts have an aperiodic F0, at roughly 240 Hz. Most grunts exhibit a spectral peak close to this irregular F0. Grunts may also contain a second, rising or falling frequency peak, between 550 and 900 Hz. The location and changes in these two frequency peaks are the cues most likely to be used by vervets when distinguishing different grunt types.

Abstract: Two rhesus and two Japanese macaque infants were cross-fostered between species in order to study the effects of auditory experience on vocal development. Both the cross-fostered and normally raised control subjects were observed over the first 2 years of life and their vocalizations were tape-recorded. We classified 8053 calls by ear, placed each call in one of six acoustic categories, and calculated the rates at which different call-types were used in different social contexts. Species differences were found in the use of “coo” and “gruff” vocalizations among control subjects. Japanese macaques invariably produced coos almost exclusively. In contrast, rhesus macaques produced a mixture of coos and gruffs and showed considerable interindividual variation in the relative use of one call type or the other. Cross-fostered Japanese macaques adhered to their species-typical behavior, rarely using gruffs. Cross-fostered rhesus subjects also exhibited species-typical behavior in many contexts, but in some situations produced coos and gruffs at rates that were intermediate between those shown by normally raised animals of the two species. This outcome suggests that environmentally mediated modification of vocal behavior may have occurred, but that the resulting changes were quite limited.

Abstract: Historically, a dichotomy has been drawn between the semantic communication of human language and the apparently emotional calls of animals. Current research paints a more complicated picture. Just as scientists have identified elements of human speech that reflect a speaker's emotions, field experiments have shown that the calls of many animals provide listeners with information about objects and events in the environment. Like human speech, therefore, animal vocalizations simultaneously provide others with information that is both semantic and emotional. In support of this conclusion, we review the results of field experiments on the natural vocalizations of African vervet monkeys, diana monkeys, baboons, and suricates (a South African mongoose). Vervet and diana monkeys give acoustically distinct alarm calls in response to the presence of leopards, eagles, and snakes. Each alarm call type elicits a different, adaptive response from others nearby. Field experiments demonstrate that listeners compare these vocalizations not just according to their acoustic properties but also according to the information they convey. Like monkeys, suricates give acoustically distinct alarm calls in response to different predators. Within each predator class, the calls also differ acoustically according to the signaler's perception of urgency. Like speech, therefore, suricate alarm calls convey both semantic and emotional information. The vocalizations of baboons, like those of many birds and mammals, are individually distinctive. As a result, when one baboon hears a sequence of calls exchanged between two or more individuals, the listener acquires information about social events in its group. Baboons, moreover, are skilled “eavesdroppers:” their response to different call sequences provides evidence of the sophisticated information they acquire from other individuals' vocalizations. Baboon males give loud “wahoo” calls during competitive displays. Like other vocalizations, these highly emotional calls provide listeners with information about the caller's dominance rank, age, and competitive ability. Although animal vocalizations, like human speech, simultaneously encode both semantic and emotional information, they differ from language in at least one fundamental respect. Although listeners acquire rich information from a caller's vocalization, callers do not, in the human sense, intend to provide it. Listeners acquire information as an inadvertent consequence of signaler behavior.

Abstract: Male humpback whales (Megaptera novaeangliae) produce long, structured sequences of sound underwater, commonly called “songs.” Humpbacks progressively modify their songs over time in ways that suggest that individuals are copying song elements that they hear being used by other singers. Little is known about the factors that determine how whales learn from their auditory experiences. Song learning in birds is better understood and appears to be constrained by stable core attributes such as species-specific sound repertoires and song syntax. To clarify whether similar constraints exist for song learning by humpbacks, we analyzed changes over 14 years in the sounds used by humpback whales singing in Hawaiian waters. We found that although the properties of individual sounds within songs are quite variable over time, the overall distribution of certain acoustic features within the repertoire appears to be stable. In particular, our findings suggest that species-specific constraints on temporal features of song sounds determine song form, whereas spectral variability allows whales to flexibly adapt song elements.