Date of Degree

9-2015

Document Type

Dissertation

Degree Name

Ph.D.

Program

Biology

Advisor(s)

Ofer Tchernichovski

Subject Categories

Biology | Music

Keywords

Birdsong; Butcherbird; Groove; Music; Rhythm; Thrush nightingale

Abstract

Birdsong is a complex, learned behavior that, like music, has meaningful units at multiple timescales. Birds perform by constructing extended presentations of their phrase repertoire. Each bird's repertoire is built from small units, such as syllables, or groups of syllables with characteristic pitch, rhythm, and timbre. Like a musician each bird has its unique structure of performance that communicates its individual identity. Also contained within a bird's performance, is information about its group identity and species identity. Like a musician's performance, a bird's singing affects the behavioral state of listeners'birds perform to attract mates and defend territory.

Subjectively, many can appreciate birdsong as musical but what evidence is there that birds have music? What parameters can be chosen to test the presence of musicality in birdsong? Are there quantitative ways to demonstrate musicality in birdsong?

In this study I test quantitatively for the presence of musical structure in birdsong by homing in on two distinct features: structural balance and groove. Music is known for its characteristic balance between complexity and regularity. Groove, in the context of genres such as jazz offers a unique, visceral parameter that is known to vary in nuanced ways. I test for musical features based on understanding of how these two parameters manifest in music.

Like music, birdsong affects the behavioral state of conspecifics, but what is it in the acoustic signal that serves to affect the behavioral state of bird listeners in a desired manner? By investigating extensive song databases of birds' singing performances, I developed methods that facilitate a deeper understanding of what structures are present within song performances and why they may arise. A key feature of these methods is the capacity for multimodal data processing, as well as analysis at micro and macro levels simultaneously. This facilitates an understanding of the relationship between units and performance level structure. I studied two species to test for the presence of musicality within their vocalizations. In the Australian pied butcherbird I investigated temporal regularity in phrase types and demonstrated a characteristic balance analogous to that found in music. In the thrush nightingale I studied regularity in song rhythms and found that performance nuances used in groove rhythms follow similar principles in the context of music and birdsong alike.

Australian pied butcherbird song phrases are built from the rearrangement of shared motifs (syllables or stereotyped groupings of notes). If the function of these motifs is to increase the repertoire of different phrase types, then transition probabilities between phrase types should capture most of the structure of singing performances. Alternatively, phrase types can be seen as varied presentations of shared themes, as often is the case in music. If this is the case, temporal regularity in performing shared motifs should be observed beyond phrase types, as if the transitions between phrases are designed to 'organize' those motifs over longer time scales.

I tested which of those two views can explain more statistical regularity during entire singing performances of wild Australian pied butcherbirds, including thousands of song syllables recorded without interruption for each bird. I found that all birds produced several highly stereotyped phrase types. Most phrase types produced by each bird had shared motifs. Throughout the performance, the temporal gap between a motif's reappearance was much more regular than what was expected by chance. In contrast, regularity in the performance of phrase types was much weaker. I developed a statistical estimate of the extent to which transition probabilities between phrase types are 'optimized' to maximize regularity in the repetition of shared motifs. I found that the phrase-types syntax is selective in achieving a regular repetition of shared motifs over the entire singing performance of the bird. This effect was stronger in birds with a richer song repertoire, suggesting the intriguing possibility that birds may regulate the temporal diversity of dominant themes in their singing performance in a manner that takes their repertoire size into account.

The thrush nightingale is a distant relative of the pied butcherbird so it would be surprising to find similarities in the deep structure of the two species. I test whether or not thrush nightingales distribute motifs throughout a performance uniformly as butcherbirds do. I found that thrush nightingales exhibit more regularity in their distribution of phrase types than what is expected from chance. However, I failed to find a distribution of motif types that was balanced against repertoire size. The thrush nightingale ends many of its song phrases with buzzes (or rattles). Upon closer inspection these buzzes emerge from a diversity of repetitive rhythmic patterns of clicks. These clicks are repeated at a regular pace, or in rhythmic groups of two, three, and four or more and they sound like the complex grooves of a jazz drummer.

I tested whether or not these patterns contain timing relationships that coincide with small integer ratios and found a no significant bias for small integer ratios. I tested whether or not the range of rhythmic ratios used could be explained by any systematic trend. I tested whether or not thrush nightingales, like jazz drummers adjust their 'swing ratio' according to tempo. Swing ratio is a term that describes the non-isochronous manner in which jazz musicians interpret eighth note rhythms, using a 'long-short' pattern instead of equal timing between beats. Jazz drummers tend to use a longer long segment at slow tempos and more even segments at fast tempos. I found that thrush nightingales have a significant tendency to adjust the swing ratio in the same manner.

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.