AIRS 4th Annual Meeting: 2012

Title: Can Singing Aid Language Acquisition in Post-Puberty Learners?
Authors: Henrietta Lempert (University of Toronto), Assunta Ferrant (University of Toronto),
Ju Hee Lim (University of Toronto), Rachel Williams (University of Toronto), Natalie Kwok
(University of Toronto)

Abstract

Introduction: Children seem to absorb new languages with ease, whereas learning a foreign
language is a major struggle for most adults. However, although adults perform at chance level
when required to extract novel words from a continuous stream of spoken syllables, they
evidence learning when exposed to sung syllables. But can singing facilitate language-learning
in adults beyond aiding their perception of statistical regularities in the speech stream? If so,
which aspects of language does it aid and how?
Method: We examine these issues with a novel language comprised of 14 words divided into
three classes. The language comes in two versions, a suffix dialect (hifto wadim vabie) and a
prefix dialect (ohift wadim ievab). In one learning condition, the study sentences are sung and
in the other condition, spoken. Participants (Introductory Psychology students) repeat 24 study
sentences according to condition and are tested for rule learning with 32 spoken sentences (16
legal sentences that did not occur during study trials and 16 illegal instances that violate one of
the rules).
Results: Currently available results for first language English females (n = 33) indicate a
comparable number of false negative responses in the speaking and singing conditions (false
negative = incorrectly rejecting a legal sentence). In the prefix dialect, correct rejection of illicit
sentences did not differ on Trial 1 in the singing and speaking conditions (Ms = 8.9 and 8.4) but
on Trial 3, performance was reliably superior in the singing than speaking condition (Ms = 11.0
and 8.9). In the suffix dialect, speaking was more likely than speaking to elicit correct rejections
on Trial 1 (Ms = 9.1 and 7.4) but by Trial 3, performance in the two conditions did not differ (Ms
= 9.2 and 9.1). The results suggest that perception of linguistic features differs in sung and
spoken sentences and that the complementary use of both approaches is important for second
language acquisition in adult learners.