New JoM-EMIT paper: Syntactic Structure in Birdsong by William Majoros

From: Bruce Edmonds (
Date: Thu Nov 15 2001 - 12:06:11 GMT

  • Next message: Wade Smith: "Re: Debunking pseudoscience: Why horoscopes really work"

    Received: by id MAA21039 (8.6.9/5.3[ref] for from; Thu, 15 Nov 2001 12:16:36 GMT
    Message-ID: <>
    Date: Thu, 15 Nov 2001 12:06:11 +0000
    From: Bruce Edmonds <>
    Organization: Centre for Policy Modelling
    X-Mailer: Mozilla 4.73 [en] (Win98; U)
    X-Accept-Language: en
    To: JOM announcements list <>
    Subject: New JoM-EMIT paper: Syntactic Structure in Birdsong by William Majoros
    Content-Type: text/plain; charset=us-ascii
    Content-Transfer-Encoding: 7bit
    Precedence: bulk

                       Syntactic Structure in Birdsong:
                  Memetic Evolution of Songs or Grammars?

                         by William Majoros
                           Celera Genomics
         1 - Anecdotal Evidence for Syntactic Structure in House Finch Song
         2 - Evolutionary Advantages of Grammar-Based Song Generation
         3 - Overview of the Experimental Approach
         4 - Methods
              4.1 - Sound Recording
              4.2 - Selection and Transcription of Recordings
              4.3 - Grammar Induction
              4.4 - Contingency Table Analysis
              4.5 - Simulation of Memetic Evolution of Songs
              4.6 - Regression Analysis
         5 - Results
              5.1 - Properties of House Finch Data
              5.2 - Properties of Simulated Data
         6 - Discussion
         7 - Conclusion


         In order to ascertain whether the units of memetic
         transmission and recombination in the birdsong of a particular
         species of finch exist at the level of individual songs or instead
         at the level of grammar models, a large number of songs were
         subjected to grammar induction techniques and the resulting
         grammars analyzed for key structural properties considered
         indicative of the existence of an underlying grammar. The
         weight of evidence provided by this approach was judged by
         applying the same analyses to the results of a genetic algorithm
         applied to a synthetic repertoire of songs and comparing the
         results. Memetic evolution applied directly to song elements,
         as simulated by the genetic algorithm, was found to generate
         models very similar to those inferred from the House Finch
         song, thereby demonstrating that the problem of discerning
         between these two competing hypotheses for explaining the
         syntactic structure observed in House Finch song is not easily
         solved using the limited data obtainable in the field. This
         underscores the importance of experimental rigor when
         studying memetic systems, and leaves open the question of
         how to confidently determine the actual level at which memes
         are operating in a particular system.

    Available at:

    This archive was generated by hypermail 2b29 : Thu Nov 15 2001 - 12:22:20 GMT