by Rod Ellis
To counter Bongaert’s claim, it is necessary to show that there are at least some differences between very successful late-starting learners and native speakers. Abrahamsson and Hyltenstam (2009) found evidence of this. Using ten measures of pronunciation, vocabulary, and grammar, they reported that none of the late-starting native-like learners in their study scored in the same range as native speakers on all of the measures although they did on some. In contrast, some of the early-starters (younger than 12) did succeed in performing identically to the native speakers. Hyltenstam and Abrahamsson (2003) argued that the subtle differences between near-native and native proficiency, which they found evidence of, give support to the CPH, but they also noted that these differences ‘are probably highly insignificant in all aspects of the second language speaker’s life’ (p. 580).
Reaching a conclusion about the CPH
On balance, the research suggests that it is unlikely that post-puberty L2 learners are capable of achieving completely native levels of proficiency—at least when highly sensitive measures of L2 proficiency are examined. However, to claim that there is a critical period for learning a second language, it is necessary to demonstrate that there is a clear discontinuity between a period when full competence is possible and a period when it is not possible. Birdsong (2006) concluded that there is no such discontinuity. He undertook an analysis of data from a wide range of studies and commented ‘in all analyses of pooled data from the early and late arrivals, age effects persist indefinitely across the span of surveyed age of arrival’ (p. 14). In other words, age does have an effect, but it is continuous in nature (i.e. there is no clearly defined ‘window of opportunity’ for learning a second language). In contrast, other researchers (for example, Granena and Long 2012) claimed that there are clear discontinuities according to age in the acquisition of different L2 systems.
There is, however, another reason for circumspection about the CPH. Birdsong (2006), drawing on Cook’s (1991) idea of multicompetence, pointed out that ‘it is more reasonable to argue that minor quantitative departures from monolingual values are artefacts of the nature of bilingualism, wherein each language affects the other and neither is identical to that of the monolingual’ (p. 22). Such a view is clearly compatible with a connectionist account of language and learning (see Chapter 1). L2 acquisition does not just entail the development of a completely new and separate neural network, but an elaboration of an established network. That is, the neural networks of the two languages become interwoven. Thus, when someone learns another language, they do not become a native speaker of their first language and a non-native speaker of the second language but a multicompetent speaker of two languages. From this perspective, the CPH is perhaps conceptually misguided.
The ultimate attainment of child and adult L2 learners
We turn now to briefly examine the comparison that Singleton (2003) considered to be the only appropriate one—i.e. ‘between post-pubertal L2 beginners and those who begin to acquire an L2 in childhood’ (p. 10). In order to make a valid comparison, of course, it is necessary to ensure that both groups of learners had experienced a similar learning context and to take account of any difference in the number of years of exposure to the second language. There are a number of studies that satisfy these conditions.
Research shows that naturalistic learners who start as children achieve a more native-like accent than those who start as adolescents or adults. Oyama (1976) investigated 60 male immigrants who had entered the United States at ages ranging from six to 20 years and had been resident there for between five and 18 years. She asked two adult native speakers to judge the learners’ accents in two 45-second extracts taken from performance on a reading-aloud task and a free-speech task. Oyama reported a very strong effect for ‘age of arrival’ but almost no effect for ‘number of years’ in the United States.
Similar results have been obtained for the acquisition of grammar. Patkowski (1980) investigated 67 educated immigrants to the United States. He found that those learners who had entered the United States before the age of 15 were rated as more syntactically proficient than learners who had entered after 15. Furthermore, there was a marked difference in the distribution of the scores (based on native speakers’ ratings on a five-point scale) for the two groups. The adult group’s scores were evenly distributed, with the majority at midpoint on the rating scale. The child group’s scores clustered at the high end of the rating scale, with 29 out of 33 achieving a rating of four-plus or five. Patkowski also investigated the effects of the number of years spent in the United States, the amount of informal exposure to English, and the amount of formal instruction. Only the amount of informal exposure had any significant effect, and even this was negligible in comparison with the age factor.
Singleton (1989) concluded his review of studies that compared groups of learners starting as children and as adults in this way:
Concerning the hypothesis that those who begin learning a second language in childhood in the long run generally achieve higher levels of proficiency than those who begin in later life, one can say that there is some good supportive evidence and that there is no actual counter evidence. (Singleton 1989: 137)
In fact, Singleton felt this was one of the few definite conclusions he was able to reach in his comprehensive survey of age-related research.
Age and rate of acquisition
In general, older learners learn more rapidly than child learners at first except possibly in pronunciation. On the face of it, this contradicts the conclusion reached in the previous section but, in fact, it is compatible with it. Older learners only have an initial advantage but, over time, child learners catch up and surpass them.
The most cited study addressing age and the rate of acquisition is Snow and Hoefnagel-Höhle (1978). This study investigated the naturalistic acquisition of Dutch by eight- to ten-year-old English-speaking children, 12- to 15-year-old adolescents, and adults over a ten-month period. The learners’ proficiency was measured on three separate occasions (after three months, after six months, and at the end of the study). With regard to morphology and syntax the adolescents did best, followed by the adults, with the children last. However, there were only small differences in pronunciation, and the grammar differences diminished over time as the children began to catch up.
Experimental research also indicates that in formal learning situations adults seem to do better than children—even in pronunciation—the area of learning that most favours children. For example, Cochrane (1980) investigated the ability of 54 Japanese children and 24 adults to discriminate English /r/ and /l/. The average length of naturalistic exposure was calculated as 245 hours for the adults and 193 for the children (i.e. relatively little). Before the instruction, the children outperformed the adults. However, when the two groups were taught the phonemic distinction, the adults benefited while the children did not.
Older learners also outperform younger learners in vocabulary development (Singleton 1999). No matter whether the context is a naturalistic one, a short-term instructional one, or a long-term instructional one, older children outperform younger children and adult/adolescent learners progress more rapidly than child learners.
Overall, then, older learners have an initial advantage over younger learners, especially in grammar and vocabulary. This can be explained by the more advanced abilities that come with the formal operations stage of cognitive development; older learners are better equipped to make use of conscious learning strategies. Children, however, have an advantage in implicit learning and over time this enables them to catch up and overtake older learners.
Age and the route of L2 acquisition
In Chapter 1, we saw that L2 learners manifest relatively fixed orders and sequences of acquisition. Thus, the final question we can ask is whether child and adult learners follow the same route in acquiring the grammar of a second language. In the first edition of this book, I concluded that the available evidence indicated that age did not affect the route o
f L2 acquisition. I cited studies that have reported the same order of acquisition of grammatical morphemes for adults and children (Dulay and Burt 1973; Bailey, Madden, and Krashen 1974). Similarly, Cancino, Rosansky, and Schumann’s (1978) study, which included child, adolescent, and adult learners found the same sequence of acquisition for English negatives and interrogatives. I concluded that learners process linguistic data in the same way, irrespective of how old they are.
This conclusion has received support from later research. Harley (1986), for example, compared early and later French immersion students. She found that the older starters demonstrated greater attainment of the French verb system than the younger ones, but that there was no difference in the order in which they acquired different verb forms. Alvarez (2006) compared the developmental sequence evident in child and adult learners’ performance of an oral narrative. As in previous studies, the adults demonstrated more rapid progress through the transitional stages of the sequences and had already reached a relatively advanced stage at the beginning of the study. However, Alvarez reported that from that stage onwards, the developmental pattern for the two groups of learners coincided closely.
However, it may be premature to dismiss the possibility that age affects the process of acquisition. Child learners are likely to rely on implicit learning while older learners are more likely to make use of explicit learning. It is reasonable to suppose that the more analytical approach of older learners will have some impact on the route of acquisition. A recent study provides evidence of this. Dimroth (2008) examined two untutored Russian beginners (aged eight and 14) of L2 German under similar conditions. She focused on the acquisition of negation and verb tense markings. The 14-year-old displayed a similar developmental pattern to that reported for adults. However, the eight-year-old child manifested a different order of acquisition (for example, tense was marked on lexical verbs before auxiliaries). Dimroth concluded that the older learner was dependent on a step-wise path to the target system while the younger learner was able to adopt target patterns immediately without going through analytical steps. She concluded that the process of L2 acquisition and not just the end product is influenced by ageNOTE 3.
Educational policy
Finally, we will briefly consider educational policy regarding the best age to introduce instruction in a foreign language in school systems. Recent years have witnessed a change in policy. In many countries, language instruction traditionally began in secondary school (i.e. around the onset of puberty), but increasingly countries are opting to start at the elementary-school level. This has been motivated by the belief that younger learners are better equipped to learn a second language than older learners. However, as we have seen, this is not quite correct as older learners develop more rapidly than younger ones and the advantage that child learners hold in terms of ultimate attainment only emerges after massive amounts of exposure to the second language. Provision for language instruction in most school systems is insufficient to enable the potential advantage that young starters hold to manifest itself. In other words, from the perspective of the research we reviewed above, it is unlikely that starting foreign language instruction early will confer any real benefit.
Studies comparing the proficiency of learners who started at the elementary- and secondary-school levels support this claim. The Barcelona Age Factor Project (Muñoz 2006), for example, examined the acquisition of English by classroom learners of English in Catalonia (Spain), comparing students who began their study at the age of eight, 11, and 14 and controlling for exposure to English outside the classroom. Data from a battery of tests were collected on three occasions—after 200 hundred hours of instruction, 416 hours, and 726 hours. The main finding was that the older learners progressed faster than the younger learners, who failed to catch up over time. The advantage for the older learners was strong and durable on measures of grammar, but was less evident on measures of speech perception, listening comprehension, and oral fluency. Larson-Hall (2008) in a study of Japanese college students—some of whom had started between the ages of three and 12 and others at 12 or 13—also found no advantage for the early starters in grammar, but did find they were better able to distinguish the sounds /l/ and /w/.
The research on age and L2 acquisition and especially these studies of school learners suggests that policy makers would do well to consider carefully before introducing L2 instruction at the elementary-school level to avoid false expectations. At best, it might assist the development of a more nativelike pronunciation, but there is no clear evidence that it holds any advantage for other aspects of L2 acquisition.
Summing up
While learners who start learning as adults can achieve high levels of L2 proficiency, there is growing evidence that they fall short of total native-like competence. However, this may simply reflect the fact a bilingual’s ‘multicompetence’ is qualitatively different from a monolingual’s competence.
Controversy exists as to whether there is a critical period for language acquisition. However, even if there are no well-defined age limits for achieving native ability in a second language, starting age has been shown to correlate with ultimate achievement. To avoid the problem of whether or not there are clear discontinuities before and after a critical age, some researchers have opted to talk about a ‘sensitive period’ rather than a ‘critical period’.
Critical (or sensitive) periods have been found for different aspects of language—the period ends first for phonology, then for lexis and collocation, and finally for grammar.
The advantage of starting young for ultimate attainment only arises if learners have ample exposure to the target language. For this reason, doubts exist about the value of starting to learn a foreign language in a classroom at an early age.
Older learners acquire a second language more rapidly than younger learners in the initial stages, except in the case of pronunciation. This may reflect the fact that older learners make fuller use of conscious learning strategies while children rely more on implicit learning.
Whether age has an effect on the process of L2 acquisition is uncertain. Some research shows that starting age has no effect on the order and sequence of acquisition, but other research suggests that the analytical skills of older learners have an impact on how they acquire specific grammatical features.
Concluding comment
From a theoretical perspective, the key question is why young learners are capable of higher levels of attainment than older learners (after controlling for the number of years of exposure to the second language). One possibility is that younger learners are better equipped to engage in implicit learning and older learners rely more on explicit learning. Implicit learning is a slow process that requires massive exposure to the second language so no immediate advantage is apparent for younger learners. In fact, explicit learning may lead to more immediate success. However, over time, implicit learning wins out because it is more likely to enable learners to develop high levels of L2 proficiency.
Notes
1 This critique of Johnson and Newport’s (1989) study was itself critiqued by DeKeyser (2000). The debate testifies to the difficulty of drawing firm conclusions about the CPH.
2 Julie was clearly more native-like than Patty. A likely explanation for this is the language distance between these learners’ first and second languages. Arguably, the distance between Chinese and English is greater than that between English and Arabic. The role of language distance in L2 acquisition is considered in Chapter 6.
3 Further evidence that age affects the process of L2 acquisition can be found in Dijk, Verspoor, and Lowie (2011). They reanalysed the data in Cancino et al.’s (1978) study and concluded that the learning trajectory of the children in this study differed from that of the adolescents and adults. See Chapter 5.
3
Psychological factors and second language acquisition
Introduction
Psychological factors are traditionally divided into three principal types: cogn
itive, conative, and affective. Cognitive factors are those that influence the processing, storing, and retrieval of information. The cognitive factor that has attracted the most attention in SLA is language aptitude. Conative factors influence the learner’s ability to establish a goal and maintain effort to achieve it. In SLA, the key conative factor is motivation. Affective factors determine whether people respond positively or negatively to specific situations. For example, learners may vary in the extent to which they experience language anxiety. These psychological factors have been of interest to researchers because they help to explain differences in individual learners’ rate and success in learning a second language.
There is both an advantage and a danger of isolating specific factors for study. The advantage lies in the possibility of the in-depth investigation of a single factor and how it affects language learning. The main danger is that investigating individual factors will result in losing sight of the ‘whole’ learner. This is a point I will return to in the concluding section of this chapter.
Key psychological factors
Table 3.1 presents a brief description of the main factors that have been investigated in SLA in terms of these three dimensions. However, I will make no attempt to survey the research on all these factors but instead focus on three key factors generally considered to be representative of the three dimensions—language aptitude, motivation, and language anxiety. This is justified as the research has shown that these are the factors that impact most strongly on L2 learningNOTE 1. There is a scarcity of research that has investigated the direct influence of the other factors on the processes involved in language learning.