Do A Person with Second Language or L2 Learners Attain the Same Level of Language as Native Speakers?

The question of the end-point of L2 acquisition was already implicit in the question about age but has been raised more explicitly in recent years: what is the final state that L2 users can reach in the knowledge of a second lan- guage? Despite the interlanguage assumption that L2 learners have independent grammars, the final state of the L2 learner has frequently been seen in terms of whether L2 learners can achieve the same competence as a native, often called “ultimate attainment.”

The starting point was a study by Coppieters (1987) who gave grammatical- ity judgments to near-native and native speakers of French on nine syntactic structures. Though the near-natives hardly deviated from the native speakers on some structures, on others they differed more, for example, tense contrasts. Even these advanced L2 users could therefore still be distinguished from native speakers. Their ultimate attainment differed from that of the native speaker.

Birdsong (1992) criticized the Coppieters research on a number of counts and essentially redid the experiment with near-native speakers of French with English as L1 and native speakers. He found that, while it was true that the near- natives differed from the natives as a group, when treated as individuals 15 of the 20 near-natives were within the native speaker range, while in the Coppieters study none were. That is to say, in effect five people should not have formed part of the near-native group. The L2 attainment of these speakers did not differ from that of native speakers.

White and Genesee (1996) continued this approach by comparing native speakers of English and L2 learners, divided into near-native and non-native groups, who were given a timed grammaticality judgments test of questions such as Which one are you reading a book about? and Who did you meet Tom after you saw? There were no differences between the natives and near-natives in accuracy and speed, with the exception of sentences such as Which movies do the children want to rent? The conclusion is that “Ultimate attainment in an L2 can indeed be native-like in the UG domain” (White and Genesee 1996: 258).

The balance of the research to date suggests that a small proportion of L2 learners can acquire the same knowledge of a language as native speakers, just as a small group seem able to acquire a native-like accent. But the question remains whether closeness to the native speaker is an appropriate yardstick to measure them by. Birdsong (1993: 717) construes “ultimate attainment in L2A [second language acquisition] in terms of whether non-natives can display evidence of possessing native linguistic norms.” But bilinguals use languages for different purposes than monolinguals and have a total language system of far greater complexity in their minds; why should L2 users be measured against the knowledge of a person with only one language? As Sridhar and Sridhar (1986) point out, “Paradoxical as it may seem, Second Language Acquisition researchers seem to have neglected the fact that the goal of SLA is bilingualism.” Indeed it is evident that L2 users can become more proficient than average L1 users, as we saw with Conrad’s writing. L2 users for instance make fewer spelling mistakes in English than 15-year-old native children (Cook 1997d). Relating L2 ultimate state to native speakers may be convenient but does an injustice to the overwhelming majority of L2 users, who are thereby seen as failures for not achieving something which is, by definition, not an achiev- able target. The unique status of the two languages of the L2 user has been abandoned in favor of seeing whether the L2 is a defective version of the L1.

How Important Is Transfer to L2 Learning?

Transfer from the first to the second language involves both use and acquisition, i.e. it may affect both the processes of speaking in the short term and the processes of learning over a period of time. The influence of the first language on the second is obvious from our everyday experience; most native speakers of English can tell whether an L2 user comes from Japan, Germany, France, or Spain.

Some early research, however, attempted to minimize the role of L1 influence. Grammatical morphemes research, for example, suggested that people with different L1s had similar acquisition sequences (Dulay et al. 1982). Dulay and Burt (1974) tried to quantify transfer mistakes vis-à-vis developmental mistakes, claiming that only 24 out of 513 mistakes by Spanish learners of English could be ascribed to their L1.

In the days when linguists considered all languages varied from each other in arbitrary ways, each pair of languages had to be compared from scratch through contrastive analysis. Now that most linguists are concerned with overall relationships between languages, transfer can be seen to utilize overall sys- tematic relationships between languages. Take the example of writing systems. These are mostly held to fall into two groups: meaning-based systems as in Chinese characters and sound-based systems as in the alphabetic system used for English (Paap et al. 1992). L1 transfer involves carrying the characteristic of the L1 writing system over to the L2. Chinese L1 speakers acquiring the Japanese syllabic writing system (kana) rely more on visual strategies, English users on phonological strategies (Chikamatsu 1996); Chinese L1 students have difficulty processing non-words in English, showing their phonological process- ing is under-developed (Holm and Dodd 1996). Speakers with meaning-based L1 writing systems are better at visual reading tasks in English than those with sound-based L1s (Brown and Haynes 1985). As in other areas, L1 transfer can be a help as well as a hindrance.

Other writing system research has looked at L1 transfer in spelling. Adult Spanish learners of English show characteristic Spanish transfer mistakes involving the double letters <rr> and <ll> and transpositions involving <l> or <r> (Bebout 1985). Some 38.5 percent of English spelling mistakes made by 10-year-old Welsh / English bilingual children can be attributed to interference from Welsh, whether from phonological interference in the L2 pronunciation, orthographic interference from Welsh sound / letter rules, or transfer of cognate words (James et al. 1993). Different L1s produce characteristic spelling mistakes in English; Japanese learners of English frequently confuse <l> and <r> as in walmer, grobal and sarary (Cook in press), perhaps because of their well-known pronunciation difficulties with the sounds /l/ and /r/, perhaps partly because of the way that English loan words are spelled in the kanasyllabic system in Japanese.

Research into phonological transfer has also progressed from lists of phonemes to more general aspects. In the acquisition of English stress assignment by speakers of Polish and Hungarian, 95 percent of the mistakes consisted of transfer of L1 metrical settings (Archibald 1993). English syllables are made to conform to the structure of the L1 by adding epenthetic syllables – [filoor] (Egyptian floor), [piliz] (Hindi please), and [iskul] (Bengali school) (Broselow 1992). The role of transfer may change during L2 development. Major (1990, 1994) claims that phonological transfer decreases over time while developmental factors first increase, then decrease.

The transfer of pragmatic speech functions from L1 to L2 has mostly been seen negatively. German learners of English produce requests that are too direct (Kasper 1981); L2 learners of English thank people in ways that are more formal than native speakers, Thank you very much rather than Thanks (Cook 1985b). Again research has gone from unique features of languages to universal schemes. There is an overall pattern to apologizing in any language consisting of explicit apology, explanations, denial of responsibility, and so on varying in weight and emphasis from one language to another (Kasper 1996, Bergman and Kasper 1993). An inappropriate linguistic form may be transferred; a learner may use the conventional Japanese way of refusing by making a statement of principle I never yield to temptations (Beebe et al. 1990), with odd effects in English.

Transfer has been looked at within the Competition Model (MacWhinney 1987). This claims that all languages make use of four cues for the subject of the  sentence  –  word  order,  agreement,  case,  and  animacy  –  but  these  are weighted differently across languages. Thus German speakers should rely on the agreement between verb and subject the horse hitS the cow; Italian speakers on the subject coming first THE  LAMB  a dog pats; and Japanese on the subject being animate The eraser THE COW kisses. L2 users indeed tend to carry over the weightings from the first language and only gradually lose them over time, whether Japanese animacy affecting English (Harrington 1987), or Dutch agree- ment affecting English (Kilborn and Cooreman 1987); Issidorides and Hulstijn (1992), however, showed that animacy may be an overriding factor with both English and Turkish learners of Dutch. Transfer here is carrying over the L1 weightings for processing the sentence to the L2.

The interpretation of transfer within the Universal Grammar (UG) theory has taken the most general point of view. If both L1 and L2 represent different choices from the same possibilities laid down by UG, the question of transfer is whether the L1 choices carry over into the L2 knowledge. The UG model has the great advantage of providing an overall descriptive syntactic model within which the two languages can be compared, even if UG theory changes constantly. It will be discussed in the next question.

In general, transfer has become a less overt concern in SLA research and has been subsumed within other issues concerned with the relationship between the two languages. Weinreich’s original definition indeed allowed transfer to go in both directions: “Those instances of deviation from the norms of either language which occur in the speech of bilinguals as a result of their familiar- ity with more than one language” (Weinreich 1953). The L2 may also have an effect on the user ’s L1. In phonology this has been a popular subject of investigation. For example, Voice Onset Time (VOT) is the moment when the voicing of a stop consonant starts relative to the release of the consonant (Liberman et al. 1967). In the English /g/ in got the voicing starts before the tongue release, to be precise an average VOT of –50 milliseconds; in the /k/ of cot it starts after the release, a VOT of +80 milliseconds. Even though two languages seem to have the same phoneme, this may disguise differences in VOT. For example Spanish /g/ is −108 milliseconds, Spanish /k/ +29 milli- seconds, both different from the typical values in English. L2 research has shown that French learners of English have a longer VOT for the voiceless /t/ sound in their L1 French than monolingual speakers (Flege 1987). Similarly L1 meanings for words may be influenced by the L2; a monolingual speaker of Korean uses paran sekj (blue) to mean something greener and less purple than a Korean who also knows English (Caskey-Sirmons and Hickerson 1977). Even language functions transfer from L2 to L1 (Kasper 1996); Locastro (1987) for example found English speakers of Japanese using aizuchi (nodding for agree- ment) when talking English.

Transfer in the sense of the relationship between the two languages in the same mind is at the heart of second language acquisition. If people simply acquired an L2 in the same way as their L1 there would be no need for a sep- arate discipline of SLA research. A major factor in the different courses of L1 and L2 acquisition must be the developing links between the two languages In a sense any investigation of L2 learning or use that does not involve this relationship is not SLA research.