FOSSILIZATION, OR WHEN L2 DEVELOPMENT COMES TO A STOP (BUT DOES IT?)

There is no guarantee that all learners will converge with the target grammar system along the pathways we have described in the preceding six sections. Much to the contrary, many L2 users may continue developing without aligning with the

Table 6.9

The emergence of questions in L2 English according to Pienemann et al. (1988)

Stage Description Illustration

1 Words and fragments with rising intonation A ball or a shoe?

2 Canonical word order with rising intonation He have two house in the front? The boy threw a shoes? 3 Fronting of a questioning element (wh-word, Where the little children are?

do, something else) What the boy with the black short throw? Do the boy is beside the bus?

Is the boy is beside the bus? 4 Inversion in two restricted contexts: (1) Where is the sun?

(1) in wh-questions with copula (2) The ball is it in the grass or in the sky? (2) in yes/no questions with auxiliaries other Is there a dog on the house? thando

5 Inversion expands to the full range of How many astronauts do you have? target-like contexts What is the boy throwing? 6 Negative questions Doesn’t your wife speak English?

Question tags You live here, don’t you?

Questions in embedded clauses Can you tell me where the station is?

Note: All illustrations are from Spada and Lightbown (1993, 1999); questions for stages 1 through 5 were

produced by Francophone 10- to 12-year-olds in intensive English programmes in Canadian schools during task-based oral interactions with a researcher. Questions for stage 6 are unattested examples invented by Spada and Lightbown.

target representations, while many may stop along the way, perhaps permanently. The term fossilization was coined by Selinker (1972) and is used to characterize cases of ‘permanent lack of mastery of a target language (TL) despite continuous exposure to the TL input, adequate motivation to improve, and sufficient opportunity for practice’ (Han, 2004, p. 4). There are several well-studied cases of adult learners whose development appears to fit the fossilization diagnosis, at least at first blush.

One is Alberto, the 33-year-old immigrant worker from Costa Rica whose naturalistic acquisition of English was studied by Schumann (1976; see Chapter 4, section 4.2). He appeared to be unable to move beyond basic English. After ten months in Boston, Schumann noted that, in terms of morpheme accuracy (cf. Table 6.4), his suppliance of –ing and copula was relatively accurate but below conventional mastery levels of 80 per cent while his suppliance of –ed was virtually zero. In terms of questions (cf. Table 6.9), he only produced uninverted ones and thus remained at stage 3. With regard to negation (cf. Table 6.2), he remained at the pre-verbal first stage for the entire ten months of the study. Hoping he could destabilize what seemed to be an unusual lack of growth, Schumann delivered an intensive regime of one-on-one instruction about English negation to Alberto over seven additional months. This was to no avail. After 17 months in the L2 environment, of which seven months also included instruction, Schumann reports that Alberto remained in the pre-verbal stage of negation, at 20 per cent overall accuracy in this area, the same level he had exhibited before instruction.

Not all learners who allegedly fossilize are naturalistic, and not all fossilization occurs at only incipient levels of development. The study of Patty by Lardiere (2007) is a well-known case of an instructed learner who achieved a very high level of com- petence in the L2 but nevertheless seems to have ceased developing in one specific area of the L2: bound verbal morphology, particularly –ed and third person singu- lar –s (cf. Table 6.4). Patty, an L1 speaker of Hokkien and Mandarin, moved to the United States at age 22 and was 31 years old when Lardiere first interviewed her, nine years after that move. About nine years later, at age 40, she was interviewed again, and once again two months later. Over two decades of being surrounded by English in graduate school and later in her workplace, Patty developed advanced English abilities, including rather high levels of accurate article usage (84 per cent accuracy for the and 75.5 per cent for a), despite this being an area of great difficulty for many L2 users from no-article language backgrounds (as Mandarin and Hokkien are). But in interview data after a decade, and later, two decades, of residence in the L2 envi- ronment, Patty continued to supply two of the morphemes that are typically mas- tered last (cf. Table 6.4) at extremely low rates: about 35 per cent for regular past –ed and about five per cent for third person singular –s.

Fossilization, particularly among very advanced learners, does not always need to involve basic syntax or morphology, but can also affect subtle areas where syntax and semantics interface. This was shown in a seven-year study by Han (2000, 2006) of Geng and Fong, two male adults from Chinese L1 backgrounds who were otherwise extremely advanced users and had enjoyed optimal learning circumstances. Geng and Fong had had formal English instruction for six years in

Fossilization, or when L2 development comes to a stop (but does it?) 135

their country, China, and had scored over 600 on the TOEFL before they moved to the United Kingdom to obtain their doctoral degrees. Upon receipt of their degrees, each continued living in English-speaking environments, and both actively published in English in international journals in their fields. In her analyses of grammaticality judgements as well as free writing over the seven-year span, Han found that both Geng and Fong consistently failed to supply passive in some cases where English requires it, as illustrated in (11), and they oversupplied it in other contexts where the active voice would be pragmatically and discoursally preferred in English, as shown in (12):

(11) I do not know whether these problems have solved in the newest release

(written by Fong in 1996; Han, 2000, p. 89)

(12) What I can do for you is to give you a list of professors … The list will be sent to you later

(written by Fong in 1996; Han, 2000, p. 94)

In addition, both L2 users showed indeterminacy in their knowledge of English unaccusativity, or the use of certain verbs in the active voice with a quasi-passive meaning (e.g. These doors will close at midnight). Specifically, on occasion Geng and Fong ungrammatically overgeneralized the passive voice by applying it to verbs with unaccusative meanings, as illustrated in (13), while other times they used unaccusatives grammatically, as shown in (14):

(13) Thanks to John’s blocking the event were stopped after 3/7/03

(written by Geng in 2003; Han, 2006, p. 69)

(14) The action already stopped on 1/6 probably after receiving our mail

(written by Geng in 2003; Han, 2006, p. 69)

The persistence of both the non-native-like and native-like solutions over the seven years is indeed suggestive of likely permanent cessation of learning in this one area of the L2.

In the end, the notion of fossilization, while strongly intuitive, has proved to be extremely problematic to pin down. Three are the most serious caveats raised by the experts (see reviews in Long, 2003; Han, 2004). One, complete and permanent cessation of learning cannot be conclusively demonstrated unless learners are followed over their lifetime, or at least over a very long period of time. Two, the studies rarely document in depth whether the so-called fossilized learners enjoyed truly optimal learning conditions, including: (a) sufficiently rich opportunities for exposure and practice; (b) positive attitudes towards the target language and society; and (c) the aid of (high-quality) instruction. Three, even if it could be demonstrated that fossilization exists, the reasons behind it have resisted any consensus thus far. For example, Schumann offered negative attitudes towards the target language and culture as the culprit (see Chapter 4, section 4.2); Han, Lardiere, and also Sorace (1993), proposed that fossilization is caused by a subtle ceiling that the L1 imposes on L2 development for even the most advanced learners;

Long (2003), reviewing a variety of studies, suggested that sensitivity to the input (or lack thereof) may be the best explanation for fossilization in general (cf. the discussion of noticing in Chapter 4, section 4.6); and Selinker and Lakshmanan (1992) stipulated that it is all of these causes in various combination that can lead to fossilization.

Perhaps the biggest challenge when trying to understand fossilization is that there are two different meanings confounded in the concept. Some researchers discuss fossilization as an inevitable universal characteristic of all L2 learning. Under this perspective, fossilization ultimately means that L2 grammars cannot reach an isomorphic state with the grammars of native speakers. All learners are expected to fossilize, and the fact that some do sooner and others do later is of peripheral interest (e.g. Coppieters, 1987; Sorace, 1993; see the discussion about ultimate attainment and age in Chapter 2, section 2.4). In the name of fossilization, other researchers investigate the premature cessation of learning, as manifested, for instance, in the radically different outcomes reported after just three years of naturalistic learning in the cases of Julie (Chapter 2, section 2.2) and Wes (Chapter 4, section 4.1). For SLA researchers who have begun to examine L2 learning through a bilingual prism (Chapter 2, section 2.8), inevitable fossilization is problematic because it rests on a tautological and intractable assumption, namely that bilingual grammars will never become monolingual grammars. For language teachers, premature fossilization is of utmost interest. Yet, contemporary research on fossilization has favoured the first notion and neglected the second one. This is not to say that individual variation in L2 learning has not received research attention. As we will see in the next three chapters, the study of cognitive, conative and affective sources of individual differences has a long history in SLA. However, differential learning rate, rather than the possibility of permanent cessation of learning, has been a central concern in the study of individual differences. Thus, until our understanding of rate and cessation of learning phenomena is better, a cautious attitude towards fossilization is advisable.

6.14 WHAT IS THE VALUE OF GRAMMAR INSTRUCTION? THE QUESTION