The Ineffectiveness of the Provision of Input on the Problematic Grammatical Feature of Articles

This study examined the value of giving specific input on the use of articles on an undergraduate English for Academic Purposes (EAP) course. This topic was chosen as previous cohorts had generated a noticeable amount of errors in their use of articles, and developing written grammatical accuracy was one of the course’s aims. Participants were provided with input on the definite, indefinite and zero articles. This involved providing images of authentic article errors, and catering for the interpersonal learner, with emphasis on student: student interaction. Also, a newspaper article was incorporated to illustrate how grammar works at a textual level, and because research has shown that accuracy in the use of the definite article increases during contextualized tasks exploiting authentic materials learners can relate to. However, the lack of effectiveness of highlighting this grammatical feature is illustrated by the students’ post-input writing, which generated more errors than the pre-input task, pro rata. In the post-input essay, there was one error for every 25 of the noun phrases, the equivalent of an error every 131 words. In comparison, the rate of errors in the preinput work was one in every 28 of the noun phrases, or an error every 153 words. Therefore, due to the 16.7% increase in post-input errors, in future, I will not focus on this grammatical feature when attempting to improve written accuracy.


Introduction
Research on the effectiveness of English for Academic Purposes (EAP) courses on students' language skills has resulted in some academics coming to the conclusion that such courses are effective in improving the accuracy of the participants.However, others are of the belief that improvements are minimal, rendering the provision of such input ineffective.Therefore, research was conducted to determine whether explicitly focusing on the English article system resulted in undergraduate students using this aspect of the language with greater accuracy in their post-input written work.This is relevant as accuracy is a requisite of academic writing and correct article usage not only indicates mastery of the language, but also exactness in thought and expression (Miller, 2005), with Master (1997, p.216) claiming that making errors can "suggest imperfect knowledge", resulting in a belief that the writer has an inadequate grasp of their subject.
The issue is exacerbated as the is the most frequently used English word, with a ranked fifth (Sinclair (1991), in Master, 2002, p.332), with the definite and indefinite articles accounting for one in every ten words in the average academic text (Berry, 1991).Therefore, proficiency in article usage provides non-native speakers with an increased level of accuracy, and limits highly noticeable errors.Hewings (2001) comments on the use of articles as being one of the most difficult grammatical features of English to master, due to their complexity, and, according to Garcia Mayo (2008), their frequency exacerbates the issue, making the application of rules problematic, if, indeed, not daunting.This is especially true for speakers of languages which do not have English article equivalents (Han, Chodorow, & Leacock, 2006).For example, definiteness in Mandarin Chinese is marked on noun phrases either by the use of demonstratives, or the word order (Quek, & Wong, 2007).The issue is exacerbated by the fact that the rules are notorious for non-native English speakers because of the exceptions (Han, et al. 2006).Cultural usage, which refers to article usage with a noun which is a unique and well known referent in a particular speech community (Liu, & Gleason, 2002) is especially problematic, irrespective of proficiency (Wong, & Quek, 2007).Furthermore, articles' unstressed nature makes it difficult for them to be noticed (Master, 2002), while their varying usage, dependent on pragmatic intent, such as the context, and pre-existing knowledge of the speaker (Chrabaszcz, & Jiang, 2014), further complicates matters, as is their use being dependent on "the interaction of many heterogeneous factors" (Chodorow, Gamon & Tetreault, 2010, p.420).

Flourishing Creativity & Literacy
As a result, learners can fail to achieve proficiency in their usage, with even advanced non-native speakers making errors after all the other elements of the language have been mastered (Master, 2002), with some educators believing that the system is so difficult to acquire that rules cannot be taught (Krech & Driver, 1996).However, it could be argued that because most errors do not affect comprehension (Swan, 2005), learners focused on being comprehensible rather than accurate can lack the motivation to master this language feature.This can be the case even though publications tend to focus on this ubiquitous language issue.For example, Swan (2005) dedicates 16 pages divided into 10 sections, one of which is further divided under 19 subheadings, while Hewings (2001) has 7 sections of rules and practice activities on 14 pages.Brown (1999) claims that the most common error is due to the omission of articles from singular countable nouns.This theory is supported by Thomas' findings on the zero article's overgeneralization, or equivalently, the failure to use any article, as being the most common source of errors for learners at all three levels of proficiency; low, mid, and high (1989).They can be so predominant that in the research conducted by Han, et al. (2006) on 668 TOEFL essays of native Chinese, Japanese, and Russian speakers, errors were made in one in eight noun phrases, while 12% of 4550 email sentences contained at least one article error (Chodorow, Gamon & Tetreault, 2010).Consequently, the aim of this research was to determine whether the provision of input on a salient grammar point, namely articles, as recommended by Ferris (2004), Master (1997) and Pica (2007), resulted in an improvement in written accuracy due to there being fewer errors made post-input, compared to the number of errors made pre-input, and comparing the data with the number of errors made by students on the same course who had not been provided with input on article usage.

The students
A Qualifying English Test is taken by undergraduates entering The National University of Singapore (NUS) who do not possess the necessary English language qualifications upon enrolment.The test involves the production of an argumentative essay based on a journal article and/or a book extract.Having taken the examination, the students whose writing fails to achieve the necessary degree of proficiency are obliged to attend a 12-week course.Among this cohort were the 52 students participating in the study.They were from China, Malaysia, and Indonesia, as well as Singapore, and were enrolled in the Faculties of Arts and Social Sciences, Business, Computing, Engineering, and Science.

Input
During the 24 lessons conducted, the main aim of the English for Academic Purposes (EAP) course is to improve students' academic writing.This involves the provision of input on the researching, organizing and structuring of essays, as well as developing written grammatical accuracy.With regard to the grammar, the lecturers have the option of using grammatical rules presented on PowerPoint slides, together with text editing texts containing inserted errors, both of which focus on the following: run on sentences and comma splices, sentence fragments, connectors, nouns, parallel structures, pronouns, word forms, subject/verb agreement, modal verbs, verb forms, prepositions, and articles.Furthermore, there are related online grammar exercises, which form 10% of the course's continuous assessment mark.
However, to cater for a wider variety of learning styles, I provided alternative input.This included images of authentic article errors, an example of which is provided in Appendix A. Furthermore, the interpersonal learner was catered for with numerous activities being conducted in pairs or small groups.Peer interaction was the preferred means of input, as opposed to focusing on theory, as becoming proficient in English results from the provision of opportunities for meaningful interaction (Cives-Enriques, 2003), with learners having a greater sense of language awareness post-input, which Storch and Tapper (2009) claim is an essential aspect of language acquisition.Also, Sinha (2014) is of the opinion that communicative tasks help in the development of learners' communicative competence.Consequently, at every stage of the input, the students were encouraged to collaborate before providing plenary feedback.This included the introductory activity, because, as Brozo (2005) states, helping students discover and develop interest in a topic creates affective engagement, which enhances learning.
The use of an authentic text was incorporated in the supplementary material in order to show how language operates in "the real world" rather than in the mind of a textbook writer (Nunan, 1998, p.105), as over-reliance on the latter can result in language learning being made to be "…more, not less, difficult for learners" (p.105).This approach is supported by Richards and Reppen (2014), who are of the belief that using authentic texts raises awareness of how grammar works at a textual level, and Sinha (2014), whose research came to the conclusion that accuracy in the use of the definite article increases during contextualized tasks exploiting authentic materials, as the provision of material that learners can relate to is provided in this approach.The input is shown in Appendix B.
As previously mentioned, the cultural usage of articles tends to cause the most errors because of the rules being overly complex and numerous (Wong & Quek, 2007).This is why an extra worksheet was included, entitled "General Knowledge Quiz" (Appendix C), aiming to draw students' attention to this particular usage.An optional activity of reading a newspaper article focusing on the complexity of the issue (Abrahams, 2006) was also provided.

Data: pre-input
At the start of the course, data was generated via a task which involved the students writing an introductory essay on themselves and/or their English language journey.This activity was completed by the 48 students present in the first lesson, which, along with their subsequent output, was posted on their blogs, due to the benefits of using this medium for the submission of work (Blackstone & Wilkinson, 2011).
The objective was for the students to introduce themselves to their peers and the lecturer.A word range of between 250 to 350 words was allocated to the task, which, not being formally assessed, did not contribute to the students' continuous assessment mark.Though each essay, on average, consisted of 286 words, which totalled 13,744 words, the range was between 188 and 497.This included an estimated 2,483 noun phrases, with the figure being calculated from the analysis of 1,374 words.These generated 90 article errors, at an average of 1.9 per student, resulting in an error being made for every 28 of the noun phrases used.This is the equivalent of an error every 153 words.These are outlined in Table 1:   Table 1.It can be seen that article omission is by far the most common type of error, especially of the definite article.This is partly explained by the use of the compound noun, English language, which accounted for 25 of the total amount, or just over 50%.
It should also be noted that with regard to the data generated in both the pre-and post-input tasks, an error was categorized as being article-based even if the correction could involve forming the plural form of a noun (as is the case with the authentic material in Appendix A), or the inclusion of a determiner other than an article.

Data: post-input
The post-input analysis focused on a problem-solution essay of 800 words on the topic of globalisation, which formed 25% of the students" continuous assessment grade.The process writing approach was implemented for this task, because, as Brown (2001) states, it values learners' intrinsic motives, which, in turn, improves language skills.Article use was assessed in the first draft of 46 essays, which totalled 23,517 words, containing 4,426 noun phrases, with this figure calculated from the analysis of 2,351 words.It can be seen from Table 2 that errors were generated at the rate of approximately 4 per student, which calculates as one article error for every 25 of the noun phrases written, the equivalent of an error every 131 words.37 students completed the post-course questionnaire (Appendix D), with the results showing that 73% were of the belief that that there was a correct amount of grammatical input provided, while 22% thought there was not enough.Nobody was of the opinion there was too much of a grammatical focus, as the remaining 5% did not respond to the question.Regarding perceived grammatical improvements made, the results were positive, with 62% of the respondents believing they had improved "quite a lot", with the remaining 38% of the opinion that they had benefitted "to a certain extent".
Improving "a lot", "not particularly" and "not at all" all had a response rate of 0%.
Many of the students were realistic with regard to their language development, with numerous comments being made which were similar to the following: …it takes time for improvement as most of the mistakes made were habitual.…grammar can only be improved through practices [sic] and 3 months is too short to make a great improvement.…it takes a lot of efforts [sic]… Despite articles being the only grammatical area covered apart from relative clauses and prepositions, no one specifically mentioned their article usage having improved in the qualitative data.Most comments were with regard to essay structuring, while the grammatical areas specified were verb forms, tenses, conjunctions, sentence structuring, prepositions, subject-verb agreement and parallel structuring.A full breakdown of this data is provided in Appendix E.
Regarding the students' opinion on the effectiveness of the various forms of input, it can be seen that classroom activities were the most popular with 67% of the students finding them to be either very or quite effective, with only 8% regarding them as being not very effective.Again, it should be noted that despite the fact that article use featured in this classification, the data generated a zero response rate.A breakdown of the data is given in Table 3, below.

Discussion
While agreeing with Storch and Tapper's opinion that focusing on all the potential difficulties students face is not feasible, I believe that there should be a focus on problematic aspects of language because, as commented on by Chodorow, Gamon, and Tetreault, "The best way to learn is through continued exposure to many examples" (2010).Furthermore, Lindstromberg (1986), Berry (1991) and Master (1997) share this belief, making specific reference to the formal instruction of articles.However, the data generated in this instance fails to support the implementation of this approach, as shown in Table 4. Therefore, due to the supplementary activities on articles not having achieved their aim of improving the students' production of this problematic grammatical language feature, I will not continue focussing on this grammatical feature, in contrast to Miller (2005), who is of the belief that instructors should cover articles when teaching college level ESL students.Even though Tomlinson (1998) states that language acquisition is a gradual process and it is possible that the students did not have enough time between the input and production for learning to have taken place, I believe that this conclusion is justified by the dearth of comments in the qualitative data.
Furthermore, I will not focus on determiners in general, either, despite an increase in the number of errors with regard to this language point in the second writing task.While the shorter introductory task on a familiar topic saw a mere 6 determiner errors being made, which consisted of a solitary demonstrative error and 5 possessive pronoun errors, the essay saw an increase to 31 errors, which excludes those corrected with the definite article and the indefinite article.Of these, 18 were possessive errors, and 10 were demonstrative errors, with quantifying errors accounting for the other 3.However, despite this increase, which may have arisen through a lack of focus in the input, the amount of errors was relatively small in number, totalling much less than one per student.
Also, the question remains as to whether the provision of input on determiners would be effective, as the findings of Allen and Rochecouste (1997) revealed an increase in the number of pronoun errors, post-intervention, while Berry (1991) identified an unwarranted emphasis on certain types of use.
I will begin to provide a focus on countability though, due to it being a source of numerous errors with Brown (1999) commenting on the fact that it is common to regard all noun forms as being countable, including in Singapore.Unsurprisingly, therefore, the introductory task saw a total of 58 errors being made, with 14 instances of plural noun forms being used incorrectly, and 44 cases of uncountable or singular noun forms being used, as opposed to the plural form.In the essay, there was an increase to a total of 138 errors.Of these, 44 involved the incorrect use of the plural noun form, with 94 examples of uncountable or singular nouns being used erroneously.For example, "With sufficient researches [sic] …".This involves an increase in the number of countability errors from every 237 words to every 170, an increase of 39.3%.Again, it should be noted that these figures exclude countability errors which were corrected with an article, as opposed to making the noun form plural.
As a consequence of the volume of errors, I shall begin to provide input on countable and uncountable nouns, especially as formal academic writing, in general, uses more nouns than other written forms (Biber, Conrad, & Leech, 2002), and because encountering and producing structures during grammar instruction can, indeed, have a positive effect on written output (Master, 1997), (Nassaji & Fotos, 2004), (Morgan, 2014).

Table 2 .
Number of Errors Post-Input: The Essay

Table 4 .
Article usage, pre-and post-input