Vocal ontogenetic processes in bat pups. From babbling behavior to the interplay of social and vocal complexity

Abstract

The human language faculty is a complex system involving key components such as speech, syntax and semantics. Speech is the motor vocal output of language and our main communication modality in daily life. Speech acquisition requires the ability of vocal production learning and is characterized by a pronounced vocal practice behavior called canonical babbling. During canonical babbling, the infant acquires the basic speech subunits by matching the own oral output to the auditory input. Hence, canonical babbling is a prerequisite to acquire the phonetic repertoire of language. So far, babbling behavior appears to be a relatively rare trait in the animal kingdom and even rarer among mammals. In the thesis presented here, I investigated a conspicuous vocal practice behavior, called babbling, in a vocal production learner, the bat species Saccopteryx bilineata. During their ontogeny, pups engaged in a conspicuous eight-week long phase of vocal practice behavior during which pups acquire a part of their adult vocal repertoire through vocal imitation. The main finding of my dissertation shows that S. bilineata pups exhibit a babbling behavior which is characterized by the same features that define human infant canonical babbling. These findings suggest that similar features characterize a distinctive vocal practice phase across vocal production learners, probably arising from similar underlying mechanisms geared toward acquiring adult communication. In human infants, canonical babbling is shaped by vocal and/or behavioral social feedback. In S. bilineata, the pups´ social environment (i.e. the colony) is composed of their mothers, other females with their pups and adult males. The pups´ babbling behavior was accompanied by behaviors and interactions restricted to mother-pup dyads. Adult males produce two song types and provide the main auditory input for pups. My study showed that the social environment influences two main factors of the babbling behavior; the amount of vocal practice and the babbling (i.e. syllabic) composition. The maternal behavior influences the amount of vocal practice, both on a daily level and over the entire ontogeny. Both the mothers´ behavior and the number of singing males present in a colony influenced the babbling bout composition, especially the production of syllable sequences that were mainly composed of song syllables acquired through vocal production learning. This study demonstrates that similar to infant speech development, social feedback influences the babbling behavior in bat pups. A special form of social feedback in humans is the so-called infant-directed speech or "motherese". Infant-directed speech is characterized by universal features (e.g. increased pitch) and, moreover, recent research has shown that the timbre (i.e. the spectral quality of sound) of motherese is distinctly different than the timbre of adult-directed speech in humans. Motherese supports linguistic learning and has social benefits. In non-human animals, pup-directed vocalizations are very common, but so far there is hardly any research on whether there is a similar phenomenon as motherese in animal vocalizations. Females of S. bilineata produce one pup-directed vocalization, the maternal directive call. Other vocalizations produced by females are directed at adults. By investigating timbre in female pup-directed versus adult-directed vocalizations I could demonstrate that the timbre of the pup-directed female vocalization was significantly different from the timbre of adult-directed female vocalizations. This thesis thus includes the first description of a phenomenon reminiscent of motherese in bats. Pup-directed vocalizations are often female-biased, especially in bats. I discovered that males in S. bilineata produce a pup-directed vocalization. They produced isolation calls, a call which is otherwise only emitted by pups to solicit maternal care. It encodes information about individual identity and social group affiliation (i.e. vocal signatures). Until now, it was assumed that the group signature develops through call convergence of pups from the same colony towards a social group (i.e. colony) mean. My acoustic analysis demonstrated that the isolation calls of pups are acoustically more similar to those of males from their own social group than to those of males from other social groups. These findings allow the speculation that the group signature in isolation calls could be transmitted horizontally across generations. The aforementioned pup isolation call is composed of different syllable types and encodes two vocal signatures. The possibility to encode information in a vocalization is limited by the available acoustic space (i.e. the information in bit which can be encoded in an acoustic signal). Therefore, vocalizations that are composed of multiple elements have the potential to distribute information among those elements and hence increase the amount of encoded information. The multiple messages hypothesis states that different elements of a vocalization can convey different information. I demonstrated the novel finding that the multisyllabic isolation call encodes information about pup age, and as expected by the multiple messages hypothesis, the information about individual identity and age were encoded in different syllable types. Hence, the multisyllabic isolation call encodes multiple messages which are temporally segregated. The "social complexity hypothesis for communication" states that increasing social complexity should lead to increased vocal complexity, ultimately resulting in a positive feedback loop. An example of a measurable feature of vocal complexity is the amount of information encoded in specific vocalizations (e.g. vocal signatures). Bats are highly gregarious animals, exhibiting a variety of social systems and display a similar diversity concerning their vocal repertoires. Bats' social vocalizations can encode a variety of information, amongst which information about identity and social group affiliation is crucial in social interactions. Individual signatures in social vocalizations are the best-studied vocal signature type in bats. Considering the aforementioned arguments, bats provide an ideal study system to investigate the relation between social and vocal complexity. I analyzed how the information content of social vocalizations encoding an individual signature was related to the complexity of the social group. A phylogenetic comparative analysis across species revealed that increased social complexity (i.e. relevant social group size) was correlated with an increase in information content encoded in social vocalizations thereby confirming the positive feedback loop between vocal and social complexity.