Second Language Learning and Language Teaching

Evidence for the effects of age on L2 learning
Evidence in favour of the superiority of young children, however, has proved sur-
prisingly hard to find. Much research, on the contrary, shows that age is a positive
advantage. English-speaking adults and children who had gone to live in Holland
were compared using a variety of tests (Snow and Hoefnagel-Höhle, 1977). At the

---

end of three months, the older learners were better at all aspects of Dutch except
pronunciation. After a year this advantage had faded and the older learners were
better only at vocabulary. Studies in Scandinavia showed that Swedish children
improved at learning English throughout the school years, and that Finnish-speak-
ing children under 11 learning Swedish in Sweden were worse than those over 11
(Eckstrand, 1978). Although the total physical response method of teaching, with
its emphasis on physical action, appears more suitable to children, when it was
used for teaching Russian to adults and children the older students were consis-
tently better (Asher and Price, 1967).
Even with the immersion techniques used in Canada in which English-speaking
children are taught the curriculum substantially through French, late immersion
pupils were better than early immersion students at marking number agreement
on verbs, and at using ‘clitic’ pronouns (‘le’, ‘me’, etc.) in object verb construc-
tions (Harley, 1986). To sum up, if children and adults are compared who are
learning a second language in exactly the same way, whether as immigrants to
Holland, or by the same method in the classroom, adults are better. The apparent
superiority of adults in such controlled research may mean that the typical situa-
tions in which children find themselves are better suited to L2 learning than those
adults encounter. Age itself is not so important as the different interactions that
learners of different ages have with the situation and with other people.
However, there are many who would disagree and find age a burden for L2 learn-
ing. These chiefly base themselves on work by Johnson and Newport (1989), who
tested Chinese and Korean learners living in the USA and found that the earlier
they had arrived there, the better they were at detecting ungrammatical use of
grammatical morphemes such as ‘the’ and plural ‘-s’, and other properties of
English such as wh-questions and word order; indeed, those who arrived under the
age of 7 were no different from natives. DeKeyser and Larson-Hall (2005) found a
negative correlation with age in ten research studies into age of acquisition and
grammaticality judgements, that is, older learners tend to do worse.
Usually children are thought to be better at pronunciation in particular. The
claim is that an authentic accent cannot be acquired if the second language is
learnt after a particular age, say the early teens. For instance, the best age for Cuban
immigrants to come to the USA so far as pronunciation is concerned is under 6, the
worst over 13 (Asher and Garcia, 1969). Ramsey and Wright (1974) found younger
immigrants to Canada had less foreign accent than older ones. But the evidence
mostly is not clear-cut. Indeed, Ramsey and Wright’s evidence has been challenged
by Cummins (1981). Other research shows that when the teaching situation is the
same, older children are better than younger children even at pronunciation. An
experiment with the learning of Dutch by English children and adults found imi-
tation was more successful with older learners (Snow and Hoefnagel-Höhle, 1977).
Neufeld (1978) trained adults with a pronunciation technique that moved them
gradually from listening to speaking. After 18 hours of teaching, 9 out of 20 stu-
dents convinced listeners they were native speakers of Japanese, 8 out of 20 that
they were native Chinese speakers.
It has become common to distinguish short-term benefits of youth from long-
term disadvantages of age. David Singleton (1989) sums up his authoritative
review of age with the statement:
The one interpretation of the evidence which does not appear to run into contradic-
tory data is that in naturalistic situations those whose exposure to a second 
language begins in childhood in general eventually surpass those whose exposure
Individual differences in L2 users and L2 learners
148

---

begins in adulthood, even though the latter usually show some initial advantage
over the former.
Adults start more quickly and then slow down. Though children start more slowly,
they finish up at a higher level. My own view is that much of the research is still
open to other interpretations (Cook, 1986). The studies that show long-term disad-
vantages mostly use different methodologies and different types of learners from
those conducted into short-term learning. In particular, the long-term research has,
by coincidence, mostly used immigrants, particularly to the USA, but the short-term
research has used learners in educational systems elsewhere. Hence factors such as
immigration cannot at present be disentangled from age. Age in itself is no explana-
tion if we cannot explain which aspect of maturation causes the difference, whether
physical, social, cognitive or linguistic.