Abstract
In studies of cumulative cultural evolution in non-human animals, the focus is most often on incremental changes that increase the efficacy of an existing form of socially learned behaviour, such as the refinement of migratory pathways. In this paper, we compare the songs of different species to describe patterns of evolution in the acoustic structure of bird songs, and explore the question of what building blocks might underlie cumulative cultural evolution of bird song using a comparative approach. We suggest that three steps occurred: first, imitation of independent sounds, or notes, via social learning; second, the formation of categories of note types; and third, assembling note types into sequences with defined structures. Simple sequences can then be repeated to form simple songs or concatenated with other sequences to form segmented songs, increasing complexity. Variant forms of both the notes and the sequencing rules may then arise due to copy errors and innovation. Some variants may become established in the population because of learning biases or selection, increasing signal efficiency, or because of cultural drift. Cumulative cultural evolution of bird songs thus arises from cognitive processes such as vocal imitation, categorization during memorization and learning biases applied to basic acoustic building blocks.This article is part of a discussion meeting issue ‘The emergence of collective knowledge and cumulative culture in animals, humans and machines’.
Highlights
Bird songs have many of the characteristics of speech and language: they are learned from conspecific models, consist of sounds that can be divided into distinct categories, and are assembled into sequences that may vary according to syntactic rules on the level of syllables, phrases, songs or bouts [1,2,3,4,5,6]
Successive rounds of vocal learning can build upon, change, elaborate, and increase the efficiency of song features, without calling on an evolutionary drive to produce an increased number of distinguishable signals. Reduced to their most basic building blocks, bird songs consist of a set of notes
General cognitive functions such as species-specific sensory predispositions, memorization and simple learning biases serve as building blocks, working together to group the notes into categories during social learning
Summary
Bird songs have many of the characteristics of speech and language: they are learned from conspecific models, consist of sounds that can be divided into distinct categories, and are assembled into sequences that may vary according to syntactic rules on the level of syllables, phrases, songs or bouts [1,2,3,4,5,6]. Depauperate songs of their own species [60,61]—or even hearing only recordings of their own developing vocalizations [62]—sing syllables and songs that are closer to the species-typical versions sung by normally reared birds These results raise the possibility that population-wide categories are the consequence of ‘lumpiness’ in the acoustic space that a species uses, with birds being genetically predisposed to learn certain subsets or ranges of acoustic features. With its attendant variation in the form of noise, copying errors and innovation, makes the first scenario possible: different lineages could produce variations on a single sound (perhaps as as one lineage producing a version that emphasizes the output of the left side of the syrinx and another emphasizing a higher frequency produced with the right side) These two variants, diverging over time, could come to distinguish the two lineages (figure 2a).
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
