Songbirds can learn context-free grammars
Timothy Gentner at UCSD claims in a talk abstract that songbirds can learn CFGs.
”
Neuronal and behavioral mechanisms for processing complex acoustic communicatio\
n signals
Timothy Gentner, Assistant Professor
Department of Psychology
University of California, San Diego
http://psy.ucsd.edu/~tgentner/index.html
Because of their acoustically rich, learned vocal repertoires,
songbirds are excellent organisms in which to study the
mechanisms that underlie the perception, representation,
songbirds are excellent organisms in which to study the
mechanisms that underlie the perception, representation,
and cognitive processing of complex acoustic communication
signals. This talk describes how a species of songbird,
European starlings, learn to recognize and classify the
songs of conspecific males — from the stimulus features
that control song recognition behavior, to the representation
of behaviorally salient auditory objects at the cellular
level. I will show that the acoustic units (“motifs”) in,
and only in, songs that birds have learned to recognize are
represented in the spiking patterns of single neurons and
populations of cells. In addition, I will show that starlings
can learn arbitrary rules that describe the sequential
patterning of song motifs, even when the patterning rule
is a context-free grammar. Thus, starlings can learn to
process recursive syntactic structure, a behavior previously
thought limited to humans. Together this work suggests how
learning may exert powerful effects on the construction of
neural representations for spectro-temporally complex auditory
objects and on the perception of how those objects are
patterned in time.
”
I guess I’ll believe it when I see the journal article…
April 26th, 2006 at 8:31 pm
Will someone ask this Tim why doesn’t he work with parrots? Why reinvent the wheel? Parrots learn language, grammer, even syntax, and understand. These birds sometimes live to 70 years, so, if they are kept around a couple a decent people, they learn! Starlings? Foolish.
May 31st, 2006 at 10:16 am
What a “Pretty Polly” says may not tell us much about how she normally strings things together, but Rodger and Rosebrugh (Animal Behaviour, 1979, 27, 737-749) showed that one could describe the morning song of the wood pewee (Contopus virens) by a right-linear grammar; and singing is the birds natural, normal behaviour.