External validity and experimental design: the sensitive phase for song learning

Abstract

In the past decade there has been widespread acceptance of the idea that the use of live, socially interactive tutors in song learning experiments produces results that differ dramatically from, and are more valid than, results obtained with tape tutors. The influential series of experiments by Baptista & Petrinovich (1984,Animal Behaviour,32, 172–181, 1986,Animal Behaviour,34, 1359–1371; Petrinovich & Baptista 1987,Animal Behaviour,35, 961–974) have promoted this view, and their results have been widely accepted. Baptista & Petrinovich's results have led them and others to question whether there are age-limited sensitive phases for song learning, as Marler's (1970,Journal of Comparative Physiology and Psychology Monographs,711–25) early studies on the white-crowned sparrow,Zonotrichia leucophrys, suggested, and led them to argue that interactive tutoring produces essentially open-ended learning. I argue here that this view is mistaken, and that the sensitive phase for song learning in the white-crowned sparrow is identical for live and tape tutors and is restricted to the first few months of life. Baptista & Petrinovich presented live and tape tutors for different periods to their subjects. Type of tutor (live versus tape) is confounded with duration of exposure to tutors. As a consequence, their experiments confound two different forms of song learning that occur at different ages, and give the mistaken appearance of an extended sensitive phase when live tutors are used.