Reading and Writing: An Interdisciplinary Journal 16: 573–609, 2003.
© 2003 Kluwer Academic Publishers. Printed in the Netherlands.
Language speciﬁc speech perception and the onset of reading
MARCS Auditory Laboratories, University of Western Sydney, Sydney, Australia
Abstract. In two studies the relationship between the onset of reading and language speciﬁc
speech perception, the degree to which native speech perception is superior to non-native
speech perception, was investigated. In Experiment 1 with children of 4, 6, and 8 years,
language speciﬁc speech perception occurred maximally at 6 years and was positively related
to reading ability for age and language comprehension level. In Experiment 2, with an
expanded range of ages and various stimulus and task changes, the relationship between
reading and language speciﬁc speech perception still held, and maximal language speciﬁc
speech perception occurred around the onset of reading instruction for three different sets
of speech contrasts, but not for a control set of non-speech contrasts. The results show
that language speciﬁc speech perception is a linguistic rather than an acoustic phenomenon.
Results are discussed in terms of early speech perception abilities, experience with oral
communication, cognitive ability, reading ability, alphabetic versus logographic languages,
phonics versus whole word reading instruction, and the effect of age versus instruction.
Key words: Perceptual re-organisation, Phoneme awareness, Phonemic perception, Reading,
Reading instruction, Speech perception
Considerable research has been conducted on the relationship between
reading ability and phonological processing abilities in children. Wagner
and Torgesen (Wagner & Torgesen, 1987; Wagner, Torgesen, Laughton,
Simmons & Rashotte, 1993; Wagner, Torgesen & Rashotte, 1994) identi-
ﬁed three aspects of phonological processing. These are (a) phonological
awareness, consisting of phonological analysis (phoneme segmentation and
categorisation tasks) and phonological synthesis (phoneme blending); (b)
retrieval of phonological codes from long-term store, either in isolated or
serial speeded naming tasks; and (c) phonological recoding to maintain infor-
mation in working memory (using tasks such as digit span and sentence
repetition recall). All three are relatively stable over development (Wagner
et al., 1993, 1994; Hansen & Bowey, 1994), and related to reading ability.
There is evidence for (a) a causal relationship between phonological aware-
ness and reading ability (Calfee, Lindamood & Lindamood, 1973; Bradley
& Bryant, 1985; Goswami & Bryant, 1990; Wagner et al., 1994), and for a
reciprocal causal relationship between phonological awareness and reading