Dakota J. Thomas-Wilhelm, University of Iowa, Iowa City, Iowa, USA

In current second language acquisition (SLA) research, there is a natural gap between theoretical research (referred to as GenSLA research in this article) and pedagogical research. GenSLA research tends to focus on solving theoretical puzzles with few to no classroom implications for the results of the studies, while pedagogical research is often very instruction-focused and, at times, detached from some of the generative theoretical underpinnings that are essential to all SLA research. (Here, pedagogical research refers to research conducted by instructors themselves. It is often also referred to as teacher research, action research, practitioner research, or exploratory practice; Allwright, 2005.) An additional difficulty is that these two groups also perceive one another’s research to be inaccessible or irrelevant (Whong et al., 2013).

The study discussed in this article seeks to fill this gap by examining the learning task that first-language (L1)–Mandarin (n = 77), second-language (L2)–English learners face when acquiring L2 English articles and noun types. Following the work by Lopez (2019), I create a language pedagogy that is informed by findings and theories within GenSLA. I call this pedagogy linguistically informed instruction (LING), which uses semantic universals to teach L2 English articles and noun types: [±definite] for articles (whether both the speaker and the listener can identify the noun and answer the question “Which one?”; e.g., the is definite and a/an is indefinite); and [±count] (whether the noun can have a plural form) and [±atomic] (whether a noun can be divided into separate parts; e.g., furniture is an uncountable, but atomic, noun that is composed of chairs and tables) for nouns.

The foundation of this study is rooted into two primary questions in the field of GenSLA concerning language transfer and universal grammar, namely if the learner is transferring anything from the L1 and what this transfer consists of and if the learner has access to universal grammar and can actually acquire L2 feature specifications. The project discussed here approaches these questions via Lardiere’s feature reassembly hypothesis (2008), where the learners face difficulty in remapping relevant features from the L1 to the L2. The greatest source of difficulty is said to be in the transfer and reassembly of the same features on different lexical items from the L1 to the L2. This project seeks to explore the acquisition and recognition, as measured by an acceptability judgment task, of L2 English article and noun type features through the use of three different instructional environments for explicit grammar teaching and acquisition in an ESL grammar classroom. Two research questions guide the study:

1. Are there any significant differences between the three instructional contexts at the different testing times?

2. How does the explicit L2 knowledge develop over time within the different instructional contexts?

The Study

The project and data collection took place at a public university in the midwestern United States, and data were collected from four different L2 knowledge instruments at pretest, immediate posttest, and delayed posttest times. Only the acceptability judgment task is presented in this paper.

Instructional Contexts Under Investigation

The three instruction contexts are LING (n = 35), “traditional” instruction (TRAD, n = 23), and a control context which received no explicit instruction (NOEX, n = 19). Note that participant group sizes differed because of retention issues over the 5 weeks of the project. All participants were L1-Mandarin, L2-English language learners with ages ranging from 17 to 23, and each group was a mixed-gender class.

The LING group received a 1-hour workshop with explicit instruction of English articles and noun types using materials that were developed and informed by recent research in GenSLA. These materials taught the particularities of English articles using the feature [±definite] and noun types using the features [±atomic], [±countable], and [±plural] to encourage feature transfer, reassembly, and acquisition. Following are examples of features taught in the LING workshop.

a. [+definite]: Both the speaker and the listener can identify the noun and answer the question “Which one?”

b. [–definite]: Only the speaker, not the listener, can identify the noun.

c. [±count]: The feature of a noun that determines whether it can have a plural form

d. [±atomic]: The feature of a noun that determines if it is made up of individuals and can be divided into separate parts

These materials also included a side-by-side comparison/contrast of articles and noun types in the two languages under investigation. (E.g., while there is no equivalent of the English definite article the in Chinese, there are other determiners, such as zhèi [this] and nèi [that], which mark definiteness.) Materials also noted common mistakes/notable differences of which learners should be particularly aware. (E.g., an ungrammatical construction of many homeworks can be corrected in two possible ways: [+count]: many assignments, or [–count]: a lot of homework).

In contrast, the TRAD group received a 1-hour workshop with explicit instruction of English articles and noun types using the assigned textbook from their grammar course, Grammar & Beyond 4. Though these chapters are in the participants’ assigned textbook, they are not assigned to students in the curriculum for the grammar course during their regularly scheduled classes; only students in the TRAD group received explicit instruction using these materials during the 1-hour workshop. NOEX did not receive any instruction on English articles and noun types.

Instrument and Procedure

The research project data collection and instructional intervention took place over a 5-week timeline outside of regularly scheduled classes where, in Week 1, the pretest (T0) data were collected; in Week 2, the researcher administered the instructional intervention to the LING and the TRAD groups and collected immediate posttest (T1) data; and, in Week 5, the delayed posttest (T2) data were collected. Data collection instruments consisted of a preparticipation questionnaire that collected biographical and language use data, an English placement test, and the acceptability judgment task.

The acceptability judgment task consisted of 60 items testing four noun types in five different article/number contexts. The participants were asked to rate items on a 7-point Likert scale ranging from 1 (not natural at all) to 7 (very natural). Of the 60 items, 45 were grammatical and 15 were ungrammatical.

Data Analysis and Results

To explore the first research question, whether there are any significant differences between the instructional contexts at different testing times, a series of one-way ANOVAs were run, followed by Tukey post-hoc analyses.

As for the acceptability ratings of grammatical sentences at T0, differences were found between LING and NOEX, but not between LING and TRAD. At T1, significant differences emerged between LING and TRAD, but not between LING and NOEX. At T2, a significant difference between LING and both TRAD and NOEX was found.

With regard to ungrammatical sentences at T0, significant differences were found between LING and NOEX, but not between LING and TRAD. At T1, statistical differences emerged between LING and both NOEX and TRAD. At T2, a statistical difference between LING and TRAD, but not NOEX, was found.

In order to explore the second research question, the development of explicit L2 knowledge over time, a series of paired-sample t-tests were run within each group to compare acceptability scores of grammatical and ungrammatical sentences from T0–T1, T0–T2, and T1–T2.

As for the acceptability ratings of grammatical sentences by LING, there was a significant increase in ratings from T0 to T1 and T2. For TRAD, there was a nonsignificant decrease in ratings from T0 to T1, but no differences from T0 to T2. As for the NOEX group, there was no statistical difference between T0 to T1, but there was a statistical decrease in ratings from T0 to T2.

With regard to ungrammatical sentences, LING showed a statistical decrease in the ratings from T0 to T1 (here, a decrease in acceptability ratings of ungrammatical sentences would be trending in the direction of rejection), but these data virtually reversed and there was no statistical difference between T0 and T2. For TRAD learners, there was a statistical increase in the acceptance of ungrammatical sentences from T0 to T1, but there was no statistical difference from T0 to T2. NOEX showed no statistical differences in acceptability scores from T0 to T1 or T2.

These data show that between-group differences in grammatical and ungrammatical sentences emerged between the LING and the TRAD groups and also the NOEX group. Within-group improvement in linguistic knowledge was found most notably in the LING group, especially at immediate posttest. While the TRAD learners demonstrated some improvement, the NOEX group showed no improvement, and at times declined improvement, across testing times.

Conclusion and Further Research

Though results found significant improvement in linguistic knowledge in the LING group from pretest to immediate posttest, these results are often not maintained to delayed posttest. It is important to also bear in mind that articles are challenging to acquire and, often, do not interfere with comprehension. The kind of intervention applied here is limited to 1 hour. These data do suggest that if linguistically informed instruction was implemented in a systematic and prolonged way, using a more syntax- and semantics-focused approach to teaching challenging grammatical concepts, it may have a greater effect on linguistic knowledge.

If classrooms and instructional contexts became learning environments that encouraged and facilitated feature (re-)assembly, then acquisition may happen quicker and more efficiently for not only challenging elements of language, but all. For example, instructors could use the features [±count, ±atomic, ±plural] to teach noun types and plural –s in English by having students conduct linguistic analyses of the vocabulary they are using and/or learning. Instructors might also choose to include a side-by-side comparison/contrast of articles and noun types in the two languages under investigation. These are both examples of linguistically informed materials that can also be applied in an ESL grammar classroom.

With regard to this project, further research and analysis will investigate breaking the analysis down by article context and noun type, working to fit a linear model with the data, and comparing these data to other measures of implicit linguistic knowledge that were also collected in the same project.

References

Allwright, D. (2005). From teaching points to learning opportunities and beyond. TESOL Quarterly, 39, 9–31.

Lardiere, D. (2008). Feature-assembly in second language acquisition. In J. Liceras, H. Zobl, & H. Goodluck (Eds.), The role of formal features in second language acquisition (pp. 106–140). Lawrence Erlbaum Associates.

Lopez, E. (2019). Teaching the English article system: Definiteness and specificity in linguistically-informed instruction. Language Teaching Research, 23(2), 200–217.

Whong, M., Gil, K.-H., & Marsden, H. (2013). Introduction: Generative second language acquisition and language pedagogy. In M. Whong, K.-H. Gil, & H. Marsden (Eds.), Universal grammar and the second language classroom (pp. 1–13). Springer Netherlands.