menu3.gif
Computational Science Research Center
Home Subscribe Contact Donate
menu4.gif
About the CSRC Academic Programs Faculty CSRC Students Acsess Program CSRC Colloquium
   
 
eventsmenu.gif
CSRC Colloquium
Upcoming Colloquium
Colloquium Archive
Acsess Program
Other Events
  Colloquia Archive

Untitled Document

DATE: Friday, November 9th, 2007
TITLE: Acoustical Challenges and Rewards of Studying Dialect Development in the Killer Whale, Orcinus orca
TIME: 3:30 PM
LOCATION: GMCS 214
SPEAKER: Ann Bowles
Senior Research Scientist
Hubbs-SeaWorld Research Institute
ABSTRACT: Dialects are unusual among mammals, documented only from cetaceans and a few primates. In killer whales, matrilineal pods have unique dialects that have been the subject of considerable research. However, development of the dialect and change over time are difficult to study due to the challenges of isolating callers in the wild. My lab is measuring dialect development in the killer whale under controlled conditions to determine how it is acquired (learned? innate?) and how it functions in the social system. Based on what we’ve learned, killer whales acquire their dialect in stages surprisingly similar to human language development, suggesting that the species could be a useful system for studying the evolution of language. However, there are considerable acoustical and signal processing challenges associated with collecting the needed data. The whales usually don’t exhibit obvious behaviors when they call. Their repertoire is complex (7-15 elements) and they use multiple, simultaneously-operating sources to produce the sounds. We are collecting data in a relatively reverberant environment, which makes localization challenging. Historically, with one hydrophone, we could only attribute calls to whales when they exhibited rare behaviors such as bubbling (<10% of calls, possibly biased). Currently, using simple localization cues (level) and an 8-element array in SeaWorld’s Shamu Backstage pool, we have raised the attribution rate to over 50%. The additional data are showing that individuals share calls with common features, as described from the wild, but that stereotyped variants are probably individual ‘signatures’ that are not shared, raising the possibility of counting callers as well as identifying pod of origin. Also, killer whales appear to use different elements of the repertoire in different social contexts when young, but go through a ‘learning window’ or crystallization of the repertoire as adults. To obtain strong support for these working hypotheses, we must be able to localize calls when whales are swimming together. This will require improvements in call sampling, attribution, and localization.

HOST: Tom Cole
   

Computational Science Research Center :: 5500 Campanile Drive :: San Diego, CA 92182-1245 :: (619) 594-3430
©2007 Computational Science Research Center, SDSU - All rights reserved.

Last updated: February 21, 2008 8:38 AM