1 / 34

Classroom input to accelerate feature reassembly of English generics

Classroom input to accelerate feature reassembly of English generics. An experimental study. Background…. Linguistic theory….

stefanie
Télécharger la présentation

Classroom input to accelerate feature reassembly of English generics

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Classroom input to accelerate feature reassembly of English generics An experimental study

  2. Background…

  3. Linguistic theory… Differences between languages have been defined in terms of phonological, syntactic and semantic functional features that languages choose to realize through different functional and lexical items. (Chomsky, 1995)

  4. English articles: A linguistic description English has two overt articles: the and a and one covert article (Ø), and all three forms have: • grammatical properties ‘e.g. number’, • semantic properties ‘e.g. generic vs partitive’ and • selectionalproperties ‘selects singular, plural or mass nouns’ (Radford 2004).

  5. Second language acquisition… • Features are ‘primitive, elemental units that make up the lexical items of every language’ (2008:173). • A second language learner will bring to the acquisition task an already fully assembled L1 grammar that is different from the L2 following a Full transfer/ Full access view (Schwartz and Sprouse 1996). • Learners have to select new features for the L2 lexical items if they do not exist in L1 and reassemble features that exist in the L1 but are assembled differently.

  6. A Contrastive Analysis of Features not Languages

  7. Articles Languages with articles Article-less languages Japanese Chinese Korean Russian Check, Serbian • English • Germanic • French • Spanish • Icelandic, Arabic, Hebrew

  8. Semantic features of English articlesThe, A & Ø • I saw a car in the driveway. The car was red and shiny. • The computer guy is coming at 10. • I want to speak to the winner of this race whoever that is! • I saw a kitty in the pet shop and I want to buy her for my daughter. • Lions usually hunt alone. • A lion usually hunts alone. [± definite] [± specific] [± generic] Ionin, Ko & Wexler (2004) Ionin & Montrul (2010)

  9. What about article-less languages? The features definite, specific, unique, generic are all available but established through context or other grammatical means (demonstratives) they just don’t have dedicated morphology (articles). The semantic features are not new to them when they learn a language with articles… just the morphological expression of those features. Slabakova (2008)

  10. Arabic articles • All spoken Arabic varieties have an overt definite article as a bound prefix form al and an indefinite covert article Ø (Bardeas 2009): Al-baitØ Bait The house A house • The distribution of the Arabic definite article al matches the English article the in anaphoric, encyclopaedic, associative definite contexts with no restriction on number, gender or argument position. • The difference is with the indefinite determiners as English has two morphemes encoding the feature [number]: a [–plural] and Ø [+plural] while Arabic encodes [―plural] onto one morpheme Ø

  11. Slabakova’sCline of Difficulty (2009) • Articles express many universal meanings available to all languages (e.g. familiarity, uniqueness, referent tracking…). The possible mappings between the category D and the morphemes in languages would be different as languages choose different distributions and interpretations. • The difficulty level starts with the easier morpheme to morpheme mapping with no assembly required and then difficulty increases when the process requires re-assembly then more difficulties are anticipated when the acquisition process requires the learner to map the grammatical feature expressed by context in L1 to a morpheme in L2.

  12. English generics (Krifka et al, 1995) Sentence level generics NP level generics

  13. Arabic Generics • Genericity in Arabic (sentence-level or NP-level1) is only possible with singular or plural NPs marked by the definite article al which makes Arabic similar to Romance languages (Fassi-Fehri2004).

  14. Acquisition of English articles by L1 Arabic learners Arabic is similar to Spanish and Romance languages in using definiteplurals with generic readings, and studies on the acquisition of Englisharticles by L1-Arabic learners specified difficulties in the acquisition ofbare plural and indefinite singular generics (e.g. Almahboob 2009; Azaz, 2014; Sarko 2009).

  15. Arabic & English generic interpretation features

  16. A cline of difficulty on Arabic & English generic interpretation

  17. The learning task of L1 Arabic learners of L2 English? • They will start out by perceiving similarities between the English article theand the Arabic article al as they both carry the feature set: [+definite], [+generic], [–plural]. • They may initially transfer their complete L1 feature set allowing: [+definite], [+generic], [+plural]. • They will have to reassemble the feature set [+definite], [+generic] to only include [–plural] (definite singular generics). • They will have to reassemble the feature sets: [+indefinite], [–generic], [―plural] onto two morphemes in L2-English ((a and Ø); the reassembly will involve allowing the feature [+generic] to be part of the two indefinite sets in L2-English: bare plurals and indefinite singulars. • They will find more difficulty with indefinite singular generics for two reasons: (a) learners have to acquire the new morpheme a and then map the [–plural], [+generic] features onto this morpheme, and (b) more evidence in the input is expected for bare plural generics in English as they are used more freely as both NP and sentence-level generics.

  18. Accelerating the learning task? ‘feature reassembly may be slow to occur or may not occur at all if the relevant evidence for the formal o semantic feature is rare or contradictory in the linguistic input’. (Cho & Slabakova, 2014, p160). Input is suggested to accelerate the reassembly process and recover from L1 transfer (Azaz, 2014).

  19. How is this different from previous studies on article instruction? Many studies were conducted to investigate the effect of instruction on the acquisition of English articles (e.g. Akakura 2011; Master 1994). However, those studies provided general instruction on the use of articles in all contexts to groups of learners from different L1 backgrounds. They are examples of what Whong, Marsden and Gil (2013) describe as classroom research with a property theory ‘gap’ lacking the ‘fine-grained understanding of linguistic generalizations’ (2013:208)

  20. Genre Analysis & Focus on form (Swales, 1990) (Long, 1990) • Based on Whong’s (2007) recommendation, two trends in English language teaching are used in this study: Swales’ (1990) ‘genre analysis’ and Long’s (1990) ‘focus on form’. • With genre analysis, learners will be exposed to texts that are reinforced with the target generic structures, giving them a form of ‘implicit input’. • Another group will be provided with the same reinforced texts with additional ‘explicit’ focus on form on sentence-level and NP-level grammatical rules given after the engagement with meaningful texts.

  21. Research Questions • Will the results of Najdi Arabic second language learners of English reflect the levels of feature reassembly difficulty on the two generic contexts as proposed in the difficulty cline? • Will the classroom input provided by this study result in improvement from the pre-tests and, if so, which type of input would be more effective: implicit or explicit? • Will the effect of the instruction hold over the eight weeks delayed post-test period?

  22. Participants • 54 L1 Najdi Arabic learners of English. • Never lived in an English speaking country and learned English primarily through formal instruction. • The Oxford Quick Placemen Test (OQPT) (Oxford 2001) placed the L1-Najdi Arabic learners in three proficiency categories: elementary, lower intermediate and upper intermediate.

  23. Methodology

  24. Instrument 1: Forced choice task (Ionin, Ko & Wexler, 2004 & Hawkins et al, 2006) • The task included sixteen items covering three contexts: • (a) [+definite] [―plural] (eight items), • (b) [–definite] [+plural] [+generic] (four items) • (c) [–definite] [–plural] [+generic] (four items).

  25. Instrument 2: Sentence repetition task(Pierce & Ionin, 2011 & Snape & Yusa, 2013) • It involved listening to sentences produced by a native speaker as many times as needed (to avoid load on working memory), then writing down the sentences on paper. • The goal of including this task is to investigate the ability to produce the English articles in different contexts, giving us another perspective on the knowledge of English articles as learners filter language that they heard through their own grammatical system.

  26. Instruction • The instructional material was designed to engage students in meaningful input through providing analysis of authentic texts as examples of the ‘research genre’ on introductions, methods and results (Swales 1990). • The texts were reinforced with the targeted noun types and provided in class through a sound file recorded by a native speaker following a stress pattern that stressed the nouns. • Following this engagement in meaning, explicit grammatical input on English generics was given to the explicit instruction group.

  27. Results Forced choice task Sentence Repetition task The results of the ANOVA showed a significant difference (p = < 0.05) between all groups of second language learners and native English speakers on: indefinite singular generics [F = 24.244, p = 0.0001] and bare plural generics [F = 7.603, p = 0.0001]. However, a significant difference was also found with the two non-generic contexts: definite singular [F = 12.602, p = 0.0001] and definite plural [F = 7.784, p = 0.0001]. Unlike the forced choice task, this does not support the predictions of the cline. The one-way between groups ANOVA of the forced choice task showed a significant difference (p = < 0.01) between all groups of second language learners and native English speakers for the two generic contexts: indefinite singular [F = 12.831, p = 0.0001] and bare plural [F = 9.979, p = 0.0001]. The definite non-generic context (in which the target article is the) showed that all groups were target- like when compared to native speakers. The explicit group (M = 6.2, SD = 1.74, p = 0.1), implicit group (M = 6.54, SD = 1.62, p = 0.25) and uninstructed control (M = 6.4, SD = 1.56, p = 0.27). This supports the prediction that the two generic contexts are more difficult for L1-Najdi Arabic learners.

  28. T-test between pre& post test results After the intervention, the paired sample t-tests of the two experimental groups exhibited significant improvement on the total accuracy scores of both instruments when compared to their pre-tests and this was not found with the control group results. The forced choice t-test results of the implicit group (p = 0.032) and explicit group (p = 0.0001) were significant at the < 0.05 level with no significant increase for the uninstructed control group (p = 0.264). The sentence repetition task t-test also confirmed a significant increase in the scores of the implicit group (p = 0.001) and explicit group (p = 0.0001) with no significant increase for the uninstructed control group (p = 0.402).

  29. Statistical Analysis (Linear Regression Model)

  30. Results of the Linear Regression Model • The first model shows a positive relationship between language proficiency and the results of the instrument across groups and times with significant p < 0.001 levels. As language proficiency increased, the accuracy scores increased on all categories of the test. • The second model was conducted to confirm the significant effect of instruction on the test categories by testing the relationship between the dependent variables and the interaction effects between the two independent variables: group and time with the reference categories; control group and pre-experiment measures. The analysis confirms the explicit instruction group’s results in the paired sample t-tests, showing significant improvement when compared to the control and pre-test in the post-test of the two generic categories: indefinite singular (p < 0.05) and bare plural (p < 0.01). • However, this effect was only maintained in the delayed post-test results of the generic bare plural (p < 0.05) in the forced choice task and only the definite non-generic singular (p < 0.05) in the sentence repetition task.

  31. Forced choice task regression plots Generic Indefinite Singular Generic Bare Plural

  32. Sentence Repetition task regression plots Generic Indefinite Singular Generic Bare Plural

  33. Results of the explicit group on the [+definite] context • The mean scores for the implicit and control groups on this context show a significant increase from the pre-test for both groups (p = 0.02) and this was sustained with the delayed post-test. On the other hand, the results of the explicit group significantly dropped from the pre-test (M = 6.2, SD = 1.7) in the immediate post-test (M = 5.4, SD = 1.5, p = 0.04), then significantly increased in the delayed post-test (M = 6.18, SD = 1.5, p = 0.006).

  34. Discussion & Limitations • The instruction that aimed at accelerating the reassembly process may have given us a picture of the developmental stages that learners will go through which involve a stage of non-target definite results. • The effectivenessof one type of input could be described as leading to the development of effortless knowledge. • More investigations into the acquisition of generic contexts by using different offline and online instruments to get a better understanding of the underlying knowledge of learners. • Care must also be taken in matching groups for language proficiency to support the effect of one type of instruction over the other.

More Related