Read doi:10.1016/j.langsci.2005.11.014 text version

Language Sciences 28 (2006) 604­615 www.elsevier.com/locate/langsci

Individual differences in language attainment: Comprehension of passive sentences by native and non-native English speakers

Ewa Dabrowska *, James Street

Department of English Language and Linguistics, University of Sheffield, Sheffield S10 2TN, UK Accepted 2 November 2005

Abstract This paper challenges a widely held assumption in linguistics ­ that all normal speakers master the basic constructions of their language ­ and argues that proficiency with a particular structure depends on individual speakers' linguistic experience. Our argument is based on an experimental study testing speakers' ability to interpret passive sentences. Since full passives are used predominantly in written texts, more educated speakers have more experience with the construction, and hence might be expected to perform better. In order to determine whether the type of linguistic experience matters as well as the sheer amount, we also tested non-native speakers. Highly educated adult second language learners have the benefits of schooling, but quantitatively less experience with passives than native English speakers, and hence should perform worse than native speakers if proficiency is merely a function of the amount of exposure. We tested sentence comprehension using a modified version of a task developed by Ferreira [Ferreira, F., 2003. The misinterpretation of non-canonical sentences. Cognitive Psychology 47, 164­203]. Participants were asked to identify the agent in four types of sentences: plausible active, implausible active, plausible passive, and implausible passive. We found that both of the highly educated groups and the less-educated non-native group performed at ceiling in all conditions. The less-educated native group performed at ceiling on the plausible sentences, but had difficulty with implausible actives (65% correct) and especially implausible passives (36% correct). These results suggest considerable (possibly education-related) differences in level of attainment among native speakers. However, the performance of the less-educated non-native group indicates that this effect is not solely attributable to the number of passives in the speakers' experience. We suggest that

*

Corresponding author. Fax: +44 112220240. E-mail address: [email protected] (E. Dabrowska).

0388-0001/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.langsci.2005.11.014

E. Dabrowska, J. Street / Language Sciences 28 (2006) 604­615

605

processing implausible non-canonical sentences depends to some extent on metalinguistic skills, which may be enhanced by explicit L2 instruction. Ó 2006 Elsevier Ltd. All rights reserved.

Keywords: English passive construction; Linguistic competence; Linguistic performance; Sentence comprehension; Sentence processing

1. Introduction Most linguists assume, either implicitly or explicitly, that all native speakers have more or less the same mental grammar. It is, of course, well established that there are vast individual differences in lexical knowledge and knowledge of archaic, very formal and highly literary grammatical constructions (e.g. Little did I know that. . .); all speakers, however, are thought to share the same `core' grammar. However, this is very much an assumption and not an empirically established fact. Considering the theoretical and practical significance of any differences in native speaker competence, there is surprisingly little research addressing this issue, and the few studies which have been conducted appear to support the claim that there are considerable individual differences, at least with regard to complex syntactic structures (see Chipere, 2003 for a review). One such study, conducted by Dabrowska (1997), for example, tested participants' abil ity to process four different sentence types: sentences containing complex NPs (1), the tough movement construction (2), and two types of sentences with parasitic gaps (3­4). (1) (2) (3) (4) Paul noticed that the fact that the room was tidy surprised Shona. John will be hard to get his wife to vouch for. It was King Louis who the general convinced that this slave might speak to. The nervous-looking student that Chris met after being told his girlfriend wanted to jilt took the 11 o'clock train.

Participants were asked simple questions about the sentences (e.g. What did Paul notice? What surprised Shona? for the complex NP sentence). Dabrowska found a very strong relationship between education and performance on all sentence types. One might argue that such differences reflect differences in linguistic competence; on the other hand, it is also possible that the less-educated participants' failure to respond correctly is attributable to processing factors such as working memory limitations. It should be pointed out that the testing procedure was designed to minimise the effects of performance factors: the sentences were presented in both spoken and written form, and participants given as much time as they needed. Therefore, there is an important sense in which some speakers' failure can be regarded as a failure in competence. On the other hand, it is undeniable that the sentences used in the study taxed the participants' processing capacities and thus it cannot be ruled out that working memory limitations were at least partly responsible for the poor performance of the less educated groups. This issue was taken by Chipere (2001), who tested two groups of 18-year-olds from the same school: a low academic ability (LAA) group comprising pupils who scored `D' or below in their GCSE English examination, and a high academic ability (HAA) who scored `A' in at least five GCSE subjects, including English. In the first phase of the experiment,

606

E. Dabrowska, J. Street / Language Sciences 28 (2006) 604­615

participants were tested on comprehension and recall of complex NP sentences. The LAA group performed significantly worse than the HAA group on both tasks. Chipere then divided the LAA participants into two subgroups which were given different types of training. Half of the participants took part in a memory training programme in which they were asked to repeat complex NP sentences. The other half were given comprehension training. This involved explicit instruction about the sentence type used in the experiment, followed by a practice session in which participants had to answer comprehension questions and were given feedback on their performance. Both groups were then tested again with new complex NP sentences. Chipere found that memory training resulted in improved performance on the recall task, but not the comprehension task, whilst comprehension training lead to an improvement in performance on both tasks. These results suggest that the LAA group's poor performance on the initial comprehension test was due to lack of experience with the particular grammatical structure used in the experiment rather than processing factors.1 In this study, we investigate individual differences in speakers' ability to interpret a much simpler construction: the full passive. The passive construction was chosen for two reasons. First, it does not place such a heavy burden on working memory as the constructions used in the earlier research, since it does not involve embedding. Secondly, while knowledge about passives is undeniably part of `core' grammar, individual speakers differ in the amount of experience they have had with this structure. In particular, since full passives are found predominantly in formal written texts, speakers who are more familiar with this kind of discourse (i.e., speakers with more formal education) will have had more exposure to this structure than other speakers. Thus, if proficiency with a particular structure depends on the amount of exposure to it, more educated speakers might be expected to perform better on a task tapping knowledge about the passive construction. Our second aim is to determine whether the type of linguistic experience matters as well as the sheer amount. In particular, we wanted to test the possibility that highly educated speakers would perform better regardless of the amount of experience with the English passive. To investigate this possibility, we compared the performance of native- and non-native speakers, the logic being that whilst highly educated adult second language learners of English will have had the benefits of schooling, they will have had quantitatively less experience with passives than native English speakers. To test comprehension, we asked participants to identify the `do-er' (i.e. agent) in a number of sentences describing various transitive events. This type of task has been used in several earlier studies, including MacWhinney et al. (1984), who tested comprehension of simple transitives, and Sasaki (1997), who used ditransitive and causative stimuli. Ferreira (2003) employed a similar method to test comprehension of passive and cleft sentences. She found that undergraduate students made a number of mistakes on passive sentences, and that performance on this structure was affected by plausibility: participants supplied the correct answer 88% of the time with plausible passives such as The man was bitten by the dog and only 74% with implausible passives such as The dog was bitten by the man. Performance on unbiased passive sentences (e.g. The woman was visited by the man) fell in

1 It is worth noting that participants presumably had command of the component parts (embedded sentences, NPs with subordinate clauses), but not the particular combination of constituents that makes up the sentences used in the experiment.

E. Dabrowska, J. Street / Language Sciences 28 (2006) 604­615

607

between these two extremes, averaging 79%; and participants virtually always supplied the target answer with active sentences, regardless of their plausibility. Ferreira's results thus indicate strong effects of plausibility for passives, but not actives. To see if this finding can be replicated, we also manipulated plausibility in our own study, testing participants on four types of stimuli: plausible actives, implausible actives, plausible passives, and implausible passives. 2. Method 2.1. Participants Forty adults (18 males and 22 females) ranging in age from 18 to 50 participated in the experiment. They were divided into four groups (graduate native, graduate non-native, non-graduate native and non-graduate non-native), with 10 participants in each group. Participants belonging to the two graduate groups had at least 15 years of formal education and were studying for a higher degree (MA or PhD) at the University of Sheffield. The non-graduate participants had no more than secondary education. The non-graduate nonnative group were native speakers of Arabic and had come to the UK as asylum seekers or refugees. The graduate non-native participants came from a variety of linguistic backgrounds. All the non-native speakers were first exposed to English in adolescence or adulthood and were proficient users of the language. The non-graduate non-native participants were all studying English at the Carl Duisberg Language Centre in Sheffield. The nongraduate native participants were employed as shelf stackers, checkout assistants or cleaners at a large supermarket in Sheffield. 2.2. Materials There were 80 experimental items, each describing a simple transitive event. Twenty of the sentences were active and plausible (e.g. The dog bit the man), 20 were active and implausible (e.g. The man bit the dog), 20 were passive and plausible (e.g. The man was bitten by the dog) and 20 were passive and implausible (e.g. The dog was bitten by the man). The 80 items were divided between four versions of the test so that each version contained five sentences of each kind, with no repeats of the same action involving the same participants: in other words, each of the four examples given above appeared in a different version. The sentences in each version were presented in a semi-random order. A complete list of sentences used in one version of the test is given in the appendix. One quarter of the participants in each group were tested with a different version. Thus, each participant heard five sentences of each kind, and every variant of each sentence was presented to one fourth of the participants. 2.3. Procedure Participants were asked to listen carefully to a series of sentences and, after each sentence, identify the `do-er' (i.e. agent). Some participants required clarification of this upon which it was explained that in a sentence like The dog chased the cat, the dog would be the do-er, because it is the thing that did the action. The interviews were conducted at the place were the participants worked or studied and were as informal as possible.

608

E. Dabrowska, J. Street / Language Sciences 28 (2006) 604­615

Participants had as much time as was necessary to answer the questions, but most completed the task in about 6 min. 3. Results The results of the experiment are summarised in Table 1. As can be seen from the table, all four groups performed at ceiling in some conditions, making parametric analyses problematic. Therefore the data were analysed using non-parametric tests: Mann­Whitney U for between-participant comparisons and Wilcoxon Signed Ranks for within-participant comparisons. 3.1. Between-participant comparisons These were computed by comparing the overall scores for the entire test. The non-graduate native group performed significantly worse than the graduate native (Mann­Whitney U = 1.000, z = À3.900, p < 0.001), graduate non-native (Mann­Whitney U = 0.000, z = À3.918, p < 0.001) and non-graduate non-native groups (Mann­Whitney U = 5.000, z = À3.445, p < 0.001). The non-graduate non-native group performed slightly worse than the graduate natives (U = 21.5, z = À2.448, p = 0.029) and the graduate non-natives (U = 22.0, z = À2.355, p = 0.035). Thus, as expected, graduate speakers performed better than non-graduates. This was true for both native and non-native speakers. The effects of native speaker status, on the other hand, were different from those anticipated: both of the graduate groups performed equally well, but, surprisingly, the non-graduate non-native group performed better than the non-graduate natives. 3.2. Within-participant comparisons Since both graduate groups and the non-graduate non-native group had performed close to ceiling in all conditions, the only significant differences in performance on different sentence types were observed in the non-graduate native group. (The non-graduate nonnative were slightly better on plausible passives than implausible actives, but the difference is not statistically significant.) Non-graduate natives performed significantly better on plausible sentences than on implausible sentences (Wilcoxon Signed Ranks: z = À2.821, p = 0.005). They also performed significantly better on active implausible sentences than on passive implausible

Table 1 Proportion of correct responses on the `do-er' task Group Plausible actives M Graduate native Graduate non-native Non-graduate native Non-graduate non-native 100 98 98 94 (SD) (0) (6) (6) (13) Implausible actives M 100 100 64 90 (SD) (0) (0) (30) (11) Plausible passives M 100 100 98 98 (SD) (0) (0) (6) (6) Implausible passives M 96 98 36 94 (SD) (13) (6) (26) (10) Overall M 99 99 74 94 (SD) (3) (2) (12) (6)

E. Dabrowska, J. Street / Language Sciences 28 (2006) 604­615

609

sentences (Wilcoxon Signed Ranks: z = À2.066, p = 0.039). There was no significant difference in performance on active plausible and passive plausible sentences, where, like the other participants, non-graduate native speakers performed at ceiling. 4. Discussion 4.1. Methodological issues The experiment generated a surprising finding: non-graduate native speakers performed very poorly on implausible sentences. This was especially true for the implausible passives, where they were actually below chance. We will discuss the possible reasons for this in the final section of this paper. First, however, it is important to establish that this was a genuine difficulty with this sentence type and not a mere artefact of the experimental procedure. The less educated respondents almost certainly had relatively little experience of formal testing, so it is reasonable to ask whether this could have affected their performance. However, the excellent performance of the less-educated non-native speakers argues against such an explanation. This group were also unused to the testing situation and in addition had even more reasons to be nervous, since they were tested in a foreign language in a foreign country. It is also worth noting that all participants did very well on the plausible sentences. Finally, there is some indirect evidence from Ferreira's study which also argues against such an explanation. Ferreira's participants were psychology undergraduates, who routinely participate in experiments for course credit, and hence clearly were no novices to the testing situation. Yet they scored 74% correct or implausible passives ­ i.e., much better than our low-education group (36%), but considerably worse than our graduate students (96%). Thus, there appears to be a continuum of proficiency with passive sentences, with participants with more schooling performing better than less educated participants. A second possibility is that the non-graduate native participants had misunderstood the task and chose the most plausible interpretation. This explanation, however, is also unsatisfactory. It does not explain why participants should perform better on implausible actives than passives (the effects of plausibility should apply in both types of sentences), or why only the less educated natives should adopt this strategy. Furthermore, if it was the implausibility of the test sentences that was to blame for some speakers' difficulties with the test sentences, they should perform well on unbiased passives such as The driver was seen by the runner, where both participants could plausibly be construed as `do-ers'. We did not include such sentences in our experiment, but Ferreira (2003) did, and found that performance was only slightly better than on implausible passives (see above). Thus, we can be reasonably confident that the non-graduate native group's poor performance on implausible passives does reveal a genuine linguistic problem, although, of course, this will need to be validated by further research using different stimulus sentences and different methodologies. 4.2. Native speakers sometimes process sentences non-syntactically Our experiment has generated two main findings. First, we found that native speakers sometimes process sentences non-syntactically. This result replicates Ferreira's (2003)

610

E. Dabrowska, J. Street / Language Sciences 28 (2006) 604­615

findings and, like Ferreira, we would argue that it indicates that speakers often rely on simple processing heuristics rather than syntactic algorithms. In other words, syntactic processing is often quite shallow and yields what Ferreira calls `good enough' representations rather than a complete syntactic parse. In addition, our results suggest that it is less educated speakers who are more likely to disregard syntactic cues and that they do so even in active sentences (though considerably less often that in passives). A possible reason for this is differences in linguistic experience: since full passives occur mostly in formal written texts, graduate participants will have had more experience with them, and hence will have developed stronger mental representations of this construction. 4.3. Some second language learners process use syntactic cues more reliably than less-educated native speakers Secondly, we found that both of the non-native groups performed better than the nongraduate native group. This suggests that the differences in the sheer number of passives in a speaker's linguistic experience cannot be the only relevant factor, since it is extremely unlikely that the non-graduate non-native group had encountered more exemplars of this construction than the non-graduate native group.2 Thus, it seems that the type of linguistic experience also matters, at least when processing implausible sentences: experience with decontextualised language (written texts, classroom) helps to increase metalinguistic awareness, which in turn leads language users to pay more attention to formal cues. Metalinguistic skills appear to be more developed in the non-native group, perhaps as a result of explicit instruction in the second language. There are other cases in the literature of non-native speakers outperforming natives on tasks tapping morphosyntactic abilities. Chipere (1998), for example, found that nonnative graduates achieved higher scores than native graduate speakers on comprehension of complex syntactic structures of the kind studied by Dabrowska (1997), i.e., sentences with complex NPs, tough movement and parasitic gaps. Sasaki (1997) tested comprehension of canonical and non-canonical actives and causatives by native and non-native Japanese speakers. His results revealed large individual differences in both groups, with the non-native speakers achieving slightly higher overall scores than the natives. It is worth noting that the non-native participants in this study were only intermediate-level students, but, significantly, they had received explicit instruction in the relevant structures. 4.4. Reasons for non-graduate native speakers' poor performance on implausible passives The non-graduate native speakers' poor performance on implausible passives could be explained by appealing to processing factors. Passive sentences place higher demands on the processing system because of their non-canonical word order. Implausible sentences are also difficult to process because the syntactic and semantic cues are in conflict. It follows that implausible passives are particularly demanding, and consequently they may be too difficult for speakers with limited processing resources.

In Ferreira's (2003) second experiment, participants interpreted subject clefts (It was the man that bit the dog) correctly, but they had problems with object clefts (It was the dog that the man bit), despite the fact that both constructions are very rare. This also suggests that frequency is only one of several factors determining speakers' proficiency with a particular construction.

2

E. Dabrowska, J. Street / Language Sciences 28 (2006) 604­615

611

As explained in Section 2, we assigned speakers to groups according to education. We did not control for IQ or working memory capacity, but it is reasonable to assume that the graduate participants had higher IQs than the non-graduates. It is also possible ­ indeed likely ­ that the non-graduate non-natives had higher IQs than the natives. The non-graduate native group did not go to university either because they chose not to or because they did not qualify. The non-graduate non-native group, on the other hand, were denied access to higher education due to reasons beyond their control (e.g. war, famine, religion, or gender). Thus, it is likely that some members of this group would have obtained a degree had conditions been more favourable. This indicates that there may be a relationship between IQ and command of the passive construction, and there is some evidence for such a relationship. Urban's (1982) elicited production study with German-speaking nine-year-olds revealed moderate to strong correlations between use of passives and intelligence (0.70 for boys and 0.60 for girls) and between use passives and German marks (0.85 for boys and 0.76 for girls). However, there are some serious problems with this proposal. First, Ferreira's undergraduates also had difficulty with implausible passives. If we were to apply the same logic to this finding, we would have to conclude that only people who are intelligent enough to qualify for a postgraduate degree have the cognitive resources to process implausible passives. Secondly, Chipere's work shows that low academic achievement speakers' poor performance on a more demanding structure is not attributable to limited processing capacity, which makes the proposal even less attractive. We therefore propose a modified version of the processing account. Implausible passives are difficult to process for the reasons outlined above, and less-educated speakers have less experience with the passive structure and therefore lack a well-entrenched verb-general passive schema. As accessing a poorly entrenched structure requires considerably more effort, less educated speakers are more likely to resort to non-syntactic strategies. This account predicts that less educated speakers' difficulties with the passive should disappear after training with this structure. Further research will be necessary to determine whether this is indeed the case. 4.5. Implications for child language research Our results suggest that there are considerable individual differences in native speakers' ability to process relatively simple constructions such as the full passive. This is surprising, given the widely held assumption that the full passive construction is mastered by age 4 or 5. In fact, full passives have been observed in the spontaneous speech of three- and even two-year-olds (albeit rarely), and a number of studies have demonstrated that four- to fiveyear-olds sometimes do very well on experimental tasks testing comprehension of full passives, at least on passives with prototypical action verbs. (Truncated passives, i.e. passive sentences without the by phrase, are produced and understood even earlier, but these could be variants of the adjective complement construction rather than the passive construction: cf. The window was broken and The window was open.) However, the fact that children perform above chance on passive comprehension tasks, and occasionally produce passive sentences themselves does not necessarily mean they know all there is to know about the passive construction. Our results, and a careful analysis of earlier research with children, suggest that we need to adopt a more sophisticated view of what it means to `master a construction'.

612

E. Dabrowska, J. Street / Language Sciences 28 (2006) 604­615

First, children (and adults) may be relying on verb-specific schemas such as NP1 BE broken by NP2, NP1 BE eaten by NP2, and so on, rather than a general passive construction (NP1 BE V-en by NP2). This would allow them to produce and understand passive sentences with these verbs. Some evidence suggesting that this might be the case is offered by a study conducted by Gordon and Chafetz (1990), who tested comprehension of passive sentences by children aged 3;6 and 4;6. The children were tested twice a week apart and tended to get the same sentences right or wrong on the test and retest. In another study, Brooks and Tomasello (1999) taught children nonce verbs in active or passive structures and then elicited them in both constructions. After short but intensive training (two 30-min sessions containing about 100 models), 85% of younger children (age 2;11) and 95% of the older children (age 3;5) produced a passive with a novel verb if the verb had been modeled in the passive (with different NPs expressing the agent and patient roles). However, if they had only heard the verb in active sentences, none of the children produced a full passive, and only 12% were able to produce a truncated passive. On the other hand, there is also evidence of more general knowledge about the passive. Pinker et al. (1987) found that four-year-olds are able to act out the meanings of full passive sentences with nonce verbs which had been taught in the active. A more recent study by Savage et al. (2003) also suggests the existence of verb-general knowledge, but adds an interesting twist. Savage et al. presented children with pairs of pictures. The experimenter described the first picture using the passive construction (The car got smashed by the tree), and then asked the child to describe the second picture (which showed different characters performing different actions). Six-year-old children showed structural priming effects (i.e., they were more likely to produce a passive sentence when describing the second picture if the experimenter had used a passive to describe the first picture), but younger children did not. However, there was priming even in three- and four-year-olds in another condition, when there was lexical overlap between the two sentences ­ that is to say, when the experimenter's sentence contained lexical material that the child could use in her own sentence (It got VERBed by it). Thus, the results suggest that young children's constructions are lexically specific, while older children have more abstract representations. It seems that such knowledge develops very gradually. Sudhalter and Braine's (1985) study of the comprehension of passive sentences showed a steady increase in performance of about 5% per year between the ages of 4 and 10, with quite large differences in the number of correct responses to passives with different types of verbs (for example, three- to four-year-olds were correct on only 26% of the trials involving experiential passives such as was heard and was seen, and on 54% of the trials involving actional passives such as was cut and was dressed). The authors conclude that . . . the development of the comprehension of the actional and experiential passive is not an all-or-none process. In neither experiment did we see two distinct groups, one suggestive of competency and a second group demonstrating non-competency. Moreover, a fluctuating attention mechanism that degrades performance in passives cannot, by itself, account for all the response distributions. (Sudhalter and Braine, 1985 p. 469) Horgan (1978) makes similar observations about the development of the ability to produce passives. These findings, then, appear to support the view that knowledge of the passive construction is a matter of degree: with increasing exposure to passive sentences, learners' representations of the construction become more entrenched, which allows them to access

E. Dabrowska, J. Street / Language Sciences 28 (2006) 604­615

613

the construction more easily and hence results in more reliable performance on both comprehension and production tasks. Our results also indicate the existence of significant individual differences in native language attainment: while most of our respondents performed at ceiling on the comprehension task, some were not able to use syntactic cues to process full passives, which suggests that they had not fully mastered the passive construction. This may explain contradictory results in child language research. Some studies suggest that the passive construction is mastered relatively early. For example, Pinker et al. (1987, Experiment 1) found that four-year-olds performed at ceiling on a comprehension task involving unfamiliar verbs in the passive voice, and many were also able to produce passives with verbs which they have encountered only in the active. Maratsos et al. (1985, Experiment 1) report that fourto five-year-olds in their study responded correctly on 67% of the trials involving familiar action verbs, 40% of the trials with familiar mental verbs, and 47% of the trials involving unfamiliar (nonce) verbs (where chance performance was 50%). Finally, in an experiment conducted by Gordon and Chafetz (1990, Study 2), children aged from 4;2 to 5;6 were 58% correct on actional passives and 29% on non-actional passives with familiar verbs, while a younger group, aged from 3;0 to 4;2, did slightly better, scoring 69% and 41% respectively. Interestingly, the children who participated in the Pinker et al. study were recruited from day care centres affiliated with Harvard University; those in the Maratsos et al. study were from `generally middle class and professional backgrounds' (p. 170); while those who participated in Gordon and Chafetz's study came from more mixed backgrounds. It is difficult to draw any firm conclusions from such differences, since the studies employed different procedures for testing comprehension; but they may be at least partly attributable to differences in proficiency with the passive resulting from differing amounts of exposure to the construction. It is also possible, of course, that the Harvard children were more test-wise than the other groups; however, the differences in performance on active sentences were much smaller, which suggests that familiarity with the testing situation was not the only relevant factor. More generally, one might note that most language acquisition research is conducted on children from fairly privileged backgrounds. There are good practical reasons for this: it is usually easier to obtain parental consent from middle class parents; and it enables researchers to avoid complications resulting from children speaking a different dialect. An upshot of this is that the shelf-stackers and checkout assistants who could not interpret the implausible passives in our study are much less likely to have been recruited as subjects of a language acquisition study when they were children than the graduate participants. As a consequence, we do not know how much, or how little, children from less privileged backgrounds know about passives. A further methodological problem arises because most studies only report group data, which may mask substantial individual differences: even when the group results are relatively good, some individuals may be performing at chance. It is clear that language acquisition researchers should devote more attention to individual differences in the course of language development as well as the ultimate outcome. 5. Conclusion As we have seen, some native speakers have considerable difficulty interpreting implausible passives (and to some extent also implausible active sentences), which suggests that

614

E. Dabrowska, J. Street / Language Sciences 28 (2006) 604­615

they may be relying more on semantic and pragmatic knowledge than on syntactic cues in language processing. Furthermore, our experiment revealed considerable differences in the level of ultimate attainment among native speakers which appear to be dependent on education. The unexpectedly good performance of the less educated non-native group suggests that this effect is not solely attributable to differences in the sheer number of passive sentences in speakers' experience, since it is very unlikely that this group encountered more passives than the corresponding native group. Thus, it seems that the ability to process implausible non-canonical sentences also depends to some extent on metalinguistic skills, which may be enhanced by explicit instruction in the second language. Appendix. Sentences used in one version of the experimental task The The The The The The The The The The The The The The The The The The The The solicitor was sued by the doctor. golfer hit the ball. boy was scared by the ghost. mouse chased the cat. policeman was caught by the hooligan. food was ruined by the cook. detective investigated the suspect. model was painted by the artist. teacher was tested by the student. fox killed the hunter. voter was deceived by the politician. patient treated the doctor. mosquito was killed by the camper. waitress served the customer. thief pursued the policeman. horse kicked the man. man bit the dog. soldier was protected by the boy. dog was chased by the cat. servant served the queen.

References

Brooks, P.J., Tomasello, M., 1999. Young children learn to produce passives with nonce verbs. Developmental Psychology 35, 29­44. Chipere, N. 1998. Real language users. Retrieved 19 October 2003. Available from: http://cogprints.ecs.soton.ac.uk/archive/00000712/. Chipere, N., 2001. Native speaker variations in syntactic competence: Implications for first language teaching. Language Awareness 10, 107­124. Chipere, N., 2003. Understanding Complex Sentences: Native Speaker Variations in Syntactic Competence. Palgrave, Basingstoke. Dabrowska, E., 1997. The LAD goes to school: A cautionary tale for nativists. Linguistics 35, 735­766. Ferreira, F., 2003. The misinterpretation of noncanonical sentences. Cognitive Psychology 47, 164­203. Gordon, P., Chafetz, J., 1990. Verb-based versus class-based accounts of actionality effects in children's comprehension of passives. Cognition 36, 227­254. Horgan, D., 1978. The development of the full passive. Journal of Child Language 5, 65­80.

E. Dabrowska, J. Street / Language Sciences 28 (2006) 604­615

615

MacWhinney, B., Bates, E., Kliegl, R., 1984. Cue validity and sentence interpretation in English, German, and Italian. Journal of Memory and Language 23, 127­150. Maratsos, M., Fox, D.E.C., Becker, J.A., Chalkley, M.A., 1985. Some restrictions on children's passives. Cognition 19, 167­191. Pinker, S., Lebeaux, D.S., Frost, L.A., 1987. Productivity and constraints in the acquisition of the passive. Cognition 26, 195­267. Sasaki, Y., 1997. Individual variation in a Japanese sentence comprehension task: Form, functions and strategies. Applied Linguistics 18, 508­537. Savage, C., Lieven, E., Theakston, A., Tomasello, M., 2003. Testing the abstractness of children's linguistic representations: Lexical and structural priming of syntactic constructions in young children. Developmental Science 6, 557­567. Sudhalter, V., Braine, M., 1985. How does comprehension of passives develop? A comparison of actional and experiential verbs. Journal of Child Language 12, 455­470. Urban, K.K., 1982. Was hat die Verwendung des Passiv mit Intelligenz und den Deutschnoten zu tun? Papiere zur Linguistik 26, 85­91.

Information

doi:10.1016/j.langsci.2005.11.014

12 pages

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate

1240315


You might also be interested in

BETA
"La pasta se come con salsa": Acquisition of middle constructions in Spanish
doi:10.1016/j.langsci.2005.11.014
I
doi:10.1016/j.wocn.2005.05.001
The effect of lexical priming on sentence comprehension: An fMRI study