by Rod Ellis
Factors Definition and commentary
Cognitive factors Intelligence Intelligence is ‘a general sort of aptitude that is not limited to a specific performance area but is transferable to many sorts of performance’ (Dörnyei 2005: 32). Intelligence has often been treated as a general ability but H. Gardner (1993) proposed that there are multiple intelligences (for example, mathematical intelligence, spatial intelligence, and linguistic intelligence).
Language aptitude Language aptitude is the special ability for learning a second language. It is considered to be at least partly separate from general intelligence. Language aptitude has been theorized as involving a number of distinct abilities—phonemic coding ability, grammatical sensitivity, inductive language-learning ability, and rote-learning ability (Carroll 1965).
Learner beliefs Learners form ‘mini theories’ consisting of the beliefs that they hold about language learning. Beliefs can be classified in terms of whether they reflect an experiential or analytic approach to learning. Learners also hold beliefs about their own self-efficacy as language learners.
Conative factors Motivation Motivation is a complex construct that involves the reasons or goals learners have for learning a second language, the effort they put into learning, and the attributes they form as a result of their attempts to learn. Various theories of the role played by motivation in L2 learning have been proposed. Early theories distinguished ‘instrumental’ and ‘integrative’ motivation and ‘extrinsic’ and ‘intrinsic’ motivation. Theories of motivation have continued to develop and currently emphasize its dynamic nature and the importance of context.
Willingness to communicate Willingness to communicate is defined as ‘the intention to initiate communication, given a choice’ by MacIntyre, Baker, Clement, and Conrad (2001: 369). It is viewed as influenced by a number of other factors and as the immediate antecedent of communication behaviour.
Affective factors Language anxiety Different types of anxiety have been identified: (1) trait anxiety (a characteristic of a learner’s personality), (2) state anxiety (apprehension that is experienced at a particular moment in response to a definite situation), and (3) situation-specific anxiety (the anxiety aroused by a particular type of situation). Language anxiety is seen as a specific type of situation-specific anxiety. It can be facilitating (i.e. have a positive effect on L2 acquisition) but is generally seen as debilitating (i.e. have a negative effect).
Mixed factors Personality Personality is generally conceived of as being composed of a series of traits such as extraversion/introversion and neuroticism/stability. An array of different personality characteristics such as self-esteem, openness to experience, and risk-taking have been claimed to be significant in language learning.
Learning style Learning style refers to the preferred way in which a person sets about learning in general. It reflects ‘the totality of psychological functioning’ (Willing 1987: 6) involving affective as well as cognitive activity. A variety of learning styles have been considered relevant to language learning (for example, sensory preferences, inductive vs deductive, synthetic vs analytic).
Table 3.1 Key individual difference factors in language learning
Language aptitude
Language aptitude is traditionally viewed as a ‘special talent’ for language learning. However, it is perhaps better defined as a conglomerate of abilities that interact dynamically with the situation in which learning takes place (Kormos 2013). Learners who possess these abilities, who are able to use them appropriately in the learning situations they find themselves, and who are motivated to do so are likely to achieve a high level of proficiency in a second language.
Language aptitude is not the same as intelligence. Sasaki (1996) conducted a study with Japanese learners of English and concluded that although the two constructs were related they were also in part distinct. Intelligence was found to be related to one aspect of language aptitude in particular—language analytical ability—but not to other aspects. Language aptitude is best seen as a complex construct involving a number of distinct abilities. We will now examine what these abilities are, starting with the early work by Carroll (1965) and then taking a closer look at more recent models of language aptitude.
Carroll’s model of language aptitude
Carroll saw language aptitude as determining ‘the amount of time a student needs to learn a given task, unit of instruction, or curriculum to an acceptable criterion of mastery under optimal conditions of instruction and student motivation’ (Carroll 1990: 26). His research in the 1950s was directed at designing tests that would indicate which learners were likely to be successful in terms of how rapidly they could learn a second language. His method of enquiry was to develop a series of tests and then select those tests that correlated most strongly with measures of L2 proficiency. He ended up with the five tests comprising the Modern Language Aptitude Test (MLAT) (Carroll and Sapon 1959). These tests measured the four abilities shown in Table 3.2.
Ability Definition
Phonemic coding ability The ability to code unfamiliar sounds in a way that they can be remembered later.
Grammatical sensitivity The ability to recognize the grammatical functions of words in sentences.
Inductive language-learning ability The ability to identify patterns of correspondence and relationships between form and meaning.
Rote-learning ability The ability to form and remember associations between L1 and L2 vocabulary items.
Table 3.2 Carroll’s model of language aptitudeNOTE 2
The MLAT was designed to predict which learners would be successful when taught by the pattern-drilling method of language teaching popular at that time. Later, researchers asked whether a new model of language aptitude and a different battery of tests were needed to account for the abilities required by more communicative approaches to teaching and for learning in naturalistic contexts. Krashen (1981), for example, argued that language aptitude only predicted ‘learning’, not ‘acquisition’ (see Chapter 1). Carroll (1990), however, remained ‘somewhat sceptical about the possibilities for greatly improving foreign language aptitude predictions beyond their present levels’ (p. 27). By and large, his scepticism has been shown to be well founded, as tests developed later (for example, Grigorenko, Sternberg, and Ehrman 2000; Kiss and Nikolov 2005) have not been found to be notably more effective.
In fact, the MLAT has proven to be a robust and useful instrument and survives as the most popular measure of language aptitude today. Carroll (1981) pointed out that the studies he carried out with the MLAT produced correlations between .40 and .60 with measures of L2 achievement—correlations that are stronger than those reported for any other individual factor other than possibly motivation. Gardner and MacIntyre (1992) commented: ‘in the long run language aptitude is probably the single best predictor of achievement in a second language’ (p. 215). It has also become clear that it is not just a predictor of performance on traditional language tests and formal classroom learning. It also predicts success in more communicative tests and in naturalistic, ‘acquisition-rich’ contexts. However, as we will see later, there are some learning conditions where it does not seem to play a significant role.
Carroll’s model of language aptitude has led to a number of questions being raised. It was used to characterize learners as having ‘high’ or ‘low’ language aptitude, but researchers have increasingly recognized the need to differentiate learners according to the specific cognitive abilities they possessed. Carroll’s claim that language aptitude was largely stable and not amenable to training was also challenged. It was questioned whether language aptitude played any role in how children learned a second language. Finally—and most importantly perhaps—researchers began to question whether the four abilities that the MLAT measured constituted a valid conceptualization of language aptitude. I will turn now to a consideration of these various issues.
Differentiating types of learners
The MLAT (and other tests of language ap
titude) provide measures of different abilities. One possibility, then, is that learners may be strong in some of these abilities, but weak in others. In other words, there may be more than one route to success. Wesche (1981) used language-aptitude test scores to distinguish those learners who were strong in language analytical abilities and those strong in memory. She conducted an experimental study to investigate whether learners benefitted most if they were matched to a type of instruction that she hypothesized would suit their aptitude type. The results provided support for this hypothesis. Skehan (1986) conducted a study of learners studying Arabic in the Army School of Languages in Britain. His analysis of the results of aptitude tests indicated that some of the learners demonstrated strength in grammatical sensitivity whereas others were strong in memory and ‘chunk-learning’ (i.e. learning formulaic sequences). He proposed that learners could be distinguished in terms of whether they were analytic or memory-oriented and showed that both types were successful.
Stable or trainable?
Carroll (1990) considered that language aptitude was a fixed ability that was not influenced by training. However, other researchers have taken a different view. Grigorenko et al. (2000), for example, claimed that ‘language aptitude is a form of developing expertise rather than a fixed entity at birth’ (p. 401). The test of these rival views of aptitude is whether it changes as a result of experience of learning languages or as a result of attempts to train the abilities involved. Nation and McLaughlin (1986) provided evidence that expert learners (i.e. learners with experience of learning several L2s) were superior to novice learners (i.e. learners with experience of learning just one language) in implicit learning. They suggested that this was because expert learners made more efficient use of learning strategies, but it is also possible that their superiority was due to their enhanced language aptitude. However, Harley and Hart (1997) did not find that learners in an early immersion programme with 12 years of experience demonstrated higher levels of aptitude than late immersion learners with only four years’ experience. They suggested language aptitude may function differently for young and older learners (see next section).
Skehan’s (2002) evaluation of the research relevant to this issue was that ‘for now, following Carroll, we will assume that aptitude does not change with the seasons’ (p. 79). One possibility, however, is that some of the abilities involved in language aptitude are amenable to experience and training while others are not. Kormos and Sáfár (2006) reported that intensive experience of language learning by a group of immersion Hungarian learners of English resulted in a significant increase in their phonological sensitivity. This is clearly an area in need of further research.
Language aptitude and age
In the last chapter, we saw that learners’ age has an impact on the rate, ultimate attainment, and (possibly) the process of L2 acquisition and that this may reflect the different approaches to learning employed by young as opposed to adolescent/adult learners; older learners, for example, may make use of their greater analytical skills. This suggests that the role of language aptitude varies according to the age of the learner.
DeKeyser (2000) investigated the hypothesis that language analytical ability contributes to the success of learners who start learning as adults, but not for those who start as children. He investigated 57 adult Hungarian-speaking immigrants to the United States, some of whom had arrived as children and others as adults. He found that language analytical ability was not a significant predictor of the childhood arrivals’ performance on a grammaticality judgement test, but that it was for the adult arrivals. Only those adult arrivals with high analytical ability scored in the same range as the young arrivals. In other words, analytical ability made a difference, but only for those learners who started to learn as adults. If it is assumed that child learners rely primarily on implicit learning while adults are engaged in explicit learning, this study suggests that the relevance of language analytical ability depends on the type of learning involved.
It is important to note, however, that DeKeyser’s study only investigated one aspect of language aptitude—language analytical ability. It is possible that other abilities (for example, phonemic coding ability and memory) are just as important for children as for adults. Harley and Hart’s (1997) study of French immersion learners in Canada supports this hypothesis. They reported a positive correlation between memory ability and L2 proficiency in the early starters, but between language analytical ability and proficiency in the learners who began in grade seven. However, Harley and Hart pointed out that this result could also be explained by the kind of instructional approach that the two groups had experienced. That is, the instruction for the early-starters emphasized memory-related activities whereas the instruction received by the late-starters was more demanding of analytical skills.
Aptitude and type of learning
Both DeKeyser’s (2000) and Harley and Hart’s (1997) studies suggest that the importance of language aptitude depends on the type of learning. But we have also seen that this is a controversial issue. Now we will consider two studies that investigated the role of language aptitude in implicit and explicit learning.
De Graaff (1997) investigated the relationship between two measures of language aptitude (grammatical sensitivity and memory) and the learning of simple and complex grammatical structures by adult learners, some of whom received explicit instruction—where the rules were explained—and others implicit instruction—where there was no rule explanation. Language aptitude correlated with the test scores of both groups of learners for both grammatical structures and there was no difference between the two groups. In other words, language aptitude proved to be an explanatory factor irrespective of the type of instruction.
Robinson (1996) also reported a study that showed language analytical ability correlated with measures of learning resulting from (1) an explicit condition where learners were given an explanation of the grammatical rules; (2) a rule-search condition where they were asked to search for the rules in sets of sentences that exemplified them; and (3) an implicit condition where they were asked to simply read and memorize sets of sentences. Language aptitude was implicated in the learning resulting from all three conditions. However, there was no relationship between analytical ability and learning in (4) an incidental learning condition where learners completed a meaning-focused task that simply involved exposure to the target structures. Arguably, it is this condition that corresponds most closely to the kind of implicit learning that takes place in a real-world context. Thus, this study can be interpreted as showing that a key component of language aptitude is a factor in explicit and memory-based learning but not in implicit learning.
Some studies (for example, Erlam 2005), however, suggest that aptitude may not necessarily play a role even in explicit learning. There are two possible hypotheses regarding the relationship between aptitude and explicit instruction. One is that learners with greater aptitude (especially language analytical ability) will be better equipped to handle instruction involving rule explanation. The other hypothesis is that careful rule explanation can compensate for differences in aptitude. It is not yet possible to determine which of these hypotheses is correct. Overall, no clear conclusion can be reached about the role of language aptitude in different types of learning. It would seem likely, however, that some abilities (for example, phonological ability and memory) are important in both implicit and explicit learning.
Reconceptualizing language aptitude
The developments that we will now consider were of two main kinds. The first involved an attempt to relate language aptitude to concepts as noticing, noticing-the-gap, and pushed output (see Chapter 1). In other words, language aptitude was now examined in relation to the process of acquisition and not just to its product. The second approach entailed attempts to develop new ways of measuring aptitude, in particular, by incorporating working memory into the model.
Skehan (2002) proposed a model of language aptitude that links different c
omponents to four macro stages in the process of language acquisition: (1) noticing, (2) patterning, (3) controlling, and (4) lexicalizing (see Table 3.3). In the case of (1) the relevant abilities are those involved in processing input; (2) involves analytic ability; (3) involves those abilities associated with controlling existing L2 knowledge whilst; (4) involves the memory abilities associated with converting rule-based knowledge into ready-made chunks that facilitate easy communication (a process Skehan called ‘lexicalizing’).
Stage Processes involved Aptitude components
1 Noticing Learner directs attention at some specific feature in the input. auditory segmentation; attention management; working memory; phonemic coding
2 Patterning Learner constructs a hypothesis (implicitly or explicitly) about the feature, subsequently extends the domain of the hypothesis before recognizing its limitations and restructuring it and integrating the new representation into the interlanguage system. working memory; grammatical sensitivity; inductive language learning ability; restructuring capacity
3 Controlling Learner is able to use the integrated feature with increasing ease and accuracy. automatization; proceduralization; retrieval processes
4 Lexicalizing Learner is now able to produce the feature as a ‘lexicalized element’ (i.e. it is accessed as a whole rather than by applying a rule). integrative memory; chunking; retrieval processes
Table 3.3 Skehan’s (2002) model of language aptitude and L2 acquisition
Skehan’s model expands on the traditional model by identifying a whole range of micro-abilities associated with specific acquisition processes. Robinson (2002) further developed this approach by attempting to show how what he termed ‘basic cognitive abilities’ (for example, ‘pattern recognition’ and ‘rote memory’) contributed to what he called ‘aptitude complexes’—such as ‘noticing-the-gap’—which are associated with specific ‘task aptitudes’—such as whether a communicative task calls for one-way or two-way communication. In this way, Robinson sought to identify the different abilities required to benefit from Task-based Language Teaching (see Chapter 11). These new models are clearly programmatic. The constructs they propose are in need of tighter definition, but they enrich our understanding of language aptitude and also afford specific hypotheses about how language aptitude relates to L2 acquisition.