An Empirical Study on the Omission of Third-Person Singular -s in Writing

Enlightened by studies in inflectional morpheme in L2 speech, this article reports a survey on the use of third-person singular -s in writing by 6 L1 Chinese L2 English learners enrolled in Lancaster University in 2014. The results show that these undergraduate students omitted almost half -s marking in obligatory contexts but made little or no mistake when they actually used third-person singular -s. The students were more successful in using be form than -s form. The grammar of these students seemed to be “fossilized” in -s marking since they made no improvement after attending a pre-sessional program at Lancaster University. Whiling echoing the findings of previous studies on either Chinese or other English learners (adults or Children), this article concludes that the acquisition third-person singular -s is a challenge for adult Chinese English learners in writing. The implications for English language teaching in China are also discussed.


Introduction
Omission of verbal inflection in L2 (Note 1) acquisition has become a focus of study in the field of second or foreign language research in recent decades (Zobl & Liceras, 1994;Lardiere, 1998aLardiere, , 1998bLardiere, , 2000Prevost & White, 2000;Ionin & Wexler, 2002;Franceschina, 2001;Yuan, 2001Yuan, , 2003Chang, 2005;Hopp, 2010;Jiang, Novokshanova, Masuda, & Wang, 2011;Jensen, Slabakova, Westergaard, & Lundquist, 2020;Nguyen & Newton, 2021;Wakabayashi, Kimura, Matthews, Akimoto, Hokari, Yamazaki, & Otaki, 2021;etc.). The explanations for the reasons behind this omission vary. According to Ionin and Wexler, there are two approaches to the issue: some researchers "attributes L2 learners' use of nonfinite verbs to an impairment of functional categories and/or features in L2 grammar" (Ionin & Wexler, 2002, p. 99), while other researchers argue that "the grammar of L2 learners contains abstract categories and features and that the problem lies in mapping from the abstract feathers to the corresponding surface morphology" (ibid, p. 100). In Chang's (2005) term, the first approach to the issue is actually based on Failed Functional Features Hypothesis, while the second one is based on Missing Surface Hypothesis. Jensen et al. (2020, p. 3) argues that acquiring properties of the functional morphology is the most challenging part of learning a second language. Wakabayashi et al. (2021, p. 735) suggests that some verbal morphemes are extremely difficult to acquire in learning English, a typical example is third-person singular -s.
While the argument continues, we find most of the studies up till now have been focused on L2 speech. As teachers of English as a foreign language for several years, we often find similar situation in our students' writing. However, we are not very sure how serious the issue is in their written output. Due to this very reason, we are interested in finding out whether omission of verbal inflection really exists in L2 English writing by Chinese students. If yes, how serious is the situation? We will certainly discuss the possible reasons for what we will find, but again we will focus on a discussion of the implication for English language teaching in China.
Owing to the limit of time and data available, we are especially interested in finding out whether the acquisition of third-person singular -s is difficult for Chinese L2 English learners. Our subjects are some L1 Chinese students enrolled in Lancaster University in 2014. Our data comes from LANCAWE, Lancaster Corpus of Academic Written English. our focus of the part is a review of the studies by Ionin and Wexler (2002) and Lardiere (1998aLardiere ( , 1998b. In the third part, we elaborate on the current study in terms of the purpose (followed by hypothesis), the subjects, the methods and the results of our survey. We discuss the results of the survey and possible explanation for the findings in the fourth part of the article. In the last part, we draw a tentative conclusion and point out the implication of our survey for English Language teaching in China.

Theoretical Background of the Survey
In order to find out whether syntactic and morphological development go side by side in L2 acquisition, Lardiere (1998aLardiere ( , 1998b carried out a 9-year longitudinal study on Patty, an adult Chinese L2 speaker of English who had been living in the USA for nearly 19 years and had been fully exposed to L2 English-speaking environment for at least eight-and-a-half years since the first recording (Lardiere, 1998a, p. 12). The data Lardiere (1998aLardiere ( , 1998b examined was based on three audio recordings (conversation with Patty) taped when Patty had been living in the United States for about 10 years, 18 years, and 18-and-a-half years respectively. Besides other discoveries such as evidence in verb-raising, two of the main focuses are the evidence of Patty's rate of suppliance of past tense marking -ed and agreement marking on third-person singular -s in obligatory context. Table 1 summarizes the result of Lardiere's (1998a;1998b) study. As can be seen, the results showed that Patty's rate of using of -ed marking in obligatory context was as low as approximately 34%. Still worse was the fact that she was using -s marking between 0% and 4.76% at the same time. Meanwhile, Lardiere also surveyed the suppliance of pronominal case marking (Table 2), and indicated that "we do not find a single incorrect choice of subject pronoun case" (Lardiere, 1998a, p. 17 Based on the above facts, Lardiere argued that "we again find confirmation of a robust dissociation between the development of inflectional affixation and syntactic knowledge of formal features" (Lardiere, 1998b, p. 369).
To find out the reason behind omission of -s marking in L2 acquisition, Ionin and Wexler (2002) studied speech samples from 20 L1 Russian children who were living in the USA or Canada and acquiring/learning English there. The children ranged in age from 3 to 13 and varied in length (most of them less than 3 years) of staying in the USA or Canada. The children were able to speak and understand English at the time of study. The data they collected were conversations with each child which lasted 30−60 minutes. The topics of the conversations included talking about their friends or schoolwork, and describing pictures in storybooks.
The results of their study were illustrated in Tables 3 and 4. As can be seen from Table 3, learners produced nonfinite forms in place of finite forms with omission of inflection high across categories. A note to be added is that the categories of irregular past-tense verbs or irregular third-person forms (e.g., has) were not counted in Table 3. Table 4 gave evidence to the fact that there were very few tense/agreement errors in the data. Table 3 also gave support to their hypothesis that the learners were more successful in the acquisition of suppletive agreement (be auxiliary and be copula) than that of affixal agreement (third-person -s and past tense -ed).  Before introducing our study, we must acknowledge that the forementioned literature is only part of the two studies by Lardiere (1998aLardiere ( , 1998b and Ionin and Wexler (2002), in which they tried to prove the correctness of Missing Surface Hypothesis and to suggest a morphological approach to the discrepancy between L1 and L2 acquisition. We are not quite sure of the correctness of their studies as there are also other scholars (for example, Franceschina, 2001) who challenged the morphological approach by providing some new data collected from a near-native L2 speaker of Spanish Martin who seemed to have no problem with past forms, but have difficulty in case-marked forms. However, we found that both studies are based on spoken data, and neither study provided information about adult learners at school. As teachers of English for years, we are more interested in finding out whether omission of verbal inflection (especially the third-person singular -s) in L2 acquisition also appears to be an issue for adult Chinese L2 English learners. The following part introduces our survey based on data from LANCAWE.

Current Study on the Use of -s Marking in the Writing of Chinese English Learners
As has been pointed out in introduction of this article, the purpose of our study is to find out whether the third-person singular -s is extremely difficult for Chinese English learners to acquire.

Hypothesis
Based on the previous sudies by Lardiere (1998aLardiere ( , 1998b, Ionin and Wexler (2002), Nguyen and Newton (2021) and Wakabayashi et al. (2021), indicating that morphological inflection is a big problem for many L2 English or EFL learners (adults or children) in speech and writing, we made four similar hypotheses for adult Chinese English learners.
Hypothesis 1: The L2 learners will produce nonfinite forms in place of finite forms in their writing.
Hypothesis 2: There will be little or no incorrect finiteness inflection in the writing of the L2 learners Hypothesis 3 The L2 learners will be more successful in the acquisition of suppletive agreement than of affixal agreement.
Hypothesis 4 The issue of omission of affixal agreement -s cannot be solved even after an intensive training course like the Presessional Program at Lancaster University.
For the first hypothesis, we will survey the omission of -s marking, be auxiliary and be copula in obligatory context and calculate the percentage of omission over all obligatory contexts.
For the second hypothesis, we will count the number of inappropriate uses of -s marking, be auxiliary and be copula and calculate the percentage of misuse over all instances.
For the third hypothesis, we will compare learners' grasp of suppletive agreement with that of affixal agreement in -s.
For the fourth hypothesis, we will compare the L2 English/EFL learners' omission percentage of -s marking before and after the Presessional Program at Lancaster University.

Subjects
The subjects in this study were six L1 Chinese undergraduate students enrolled in Lancaster University in 2014. Before coming to the U.K., they had been learning English as a foreign language in China for at least 8 years.
They are enrolled in the presessional program either because their total IELTS (International English Language Testing System) score was below 6.5 or because the total score reached 6.5 but at least one of the sub-skills scores below 6. Of the six subjects we studied, five of them scored 6.0 in IELTS, one scored 6.5, but her speaking was only 5.0. All of six students (two males and four females) were randomly selected from a group of 43 Chinese undergraduate students who attended a one-month presessional program and took the tests (T2 and T3).

Data Collection
English writing samples were obtained from each of the six Chinese students who attended a one-month ijel.ccsenet.org International Journal of English Linguistics Vol. 12, No. 2;2022 presessional program at Lancaster University. Their writings during that period were recorded in LANCWE. The corpus contains five pieces of writing produced over four weeks by each of these students: two tests (T2 and T3) and three assignments (A4−A6). The tests data were selected since we thought they were more natural than the assignments as they were written within a limited time. We learned that the writing tasks completed under test conditions in 2014 were three retired IELTS academic writing tests, each containing a descriptive task (Task 1) and an argumentative task (Task 2). As task 1 usually requires test-takers to describe or explain one or some tables/graphs within about 150 words, we think it is not representative in indicating the examinees' writing capability. Therefore, we selected two writings (Task 2 in T2 and T3) for each of the 6 students as we consider these data to be long (at least 250 words per task) and representative (without too many constraints on test-takers) enough.
The data (12 argumentative tasks by 6 Chinese students) were downloaded from LANCWE website and jointed together in a plain-text format with all the errors kept intact. The results used to certify hypotheses 1 to 3 came from T2 (6 texts in all), and the both T2 and T3 (12 text in total) were used to testify hypothesis 4.
Although Ionin and Wexler distinguished regular thematic verbs from irregular ones like "does" and "has", as they indicated in their study (Ionin & Wexler, 2002, p. 109) that the difference is not very significant, we decide to include "does" and "has" in the counting of -s. Another note is that we do not include -ed marking in our study because of lack of data. Formulaic expressions (e.g., you know) were not counted or computed.

-s Omission
We firstly examined the use of three types of verbal inflection morphemes in obligatory context: the third-person singular -s, be auxiliary and be copula in 6 texts selected from Test 2 (T2). As is mentioned above, slightly different from what Ionin and Wexler (2002), we did not distinguish irregular verbs in third-person (e.g. has, does) from regular verbs. Table 5 gives the number of omissions in obligatory contexts, the total number of obligatory contexts and percentage of omission over all obligatory contexts. As Table 5 shows, the rate of omission of third-person -s in obligatory context was 48.28%. At first glance, the figure is not high as compared with that of Table 1. Yet as the learners in our study had been learning English at school for at least 8 years and had a total IELTS score of at least 6, this number is really high. Therefore, our first hypothesis that the L2 learners will produce nonfinite forms in place of finite forms in their writing is supported. An example such error is given below: 1). So if every body own at lease one car, it would be much crowder and we just have a limit place, how can we make such a lot land for the cars?

-s errors
We also counted the number of errors in third-person -s, be auxiliary and be copula in contrast to the total number of their actual use. Table 6 gives the results of tense/agreement errors in morpheme use.
As can be seen from Table 6, percentage of inappropriate use over all instance of use in both suppletive agreements is 0.00%, and in affixal agreement -s 9.09%. Similar to the results of Table 2, the figures are very low. Nevertheless, our second hypothesis that there will be little or no incorrect infiniteness inflection in the writing of the L2 learners is supported. As a matter of fact, there is only one instance of inappropriate use of -s agreement, which is given below: 2). We should open our mind and keep it up to date, then we can become the more modern people and help the technology develops more quicker.  As Table 5 indicates, the percentages of omission of be auxiliary, be copula and third-person -s in obligatory contexts are 25.00%, 8.7% and 48.28% respectively. Considering the fact that the total number of obligatory contexts for be auxiliary is only 4, the figure 25% does not make much sense. If we combine the number of omission in obligatory contexts for be auxiliary and be copula, the percentage of omission over all obligatory contexts for suppletive -be form will be 10% {[(1+4)/(4+46)]×100%=50%}. The discrepancy between 10% error in be form and 48.28% error in -s form is significant. Thus, our third hypothesis that L2 learners will be more successful in the acquisition of suppletive agreement than of affixal agreement is supported. Examples of omission of be auxiliary and be copula in our data are given below: 3) a. Omission of be auxiliary (the only case available) This kind of traditional culture will ø living as long as the world be on the opposite side, it will lost.

b. Omission of be copula
Secondly, more and more cars on the road nowadays make the people who ø always on foot in danger when they walk the road, especial the children and olders.

Changes of Omission of -s Marking Before and After an Intensive Training Course
As Lardiere (1998a;1998b) suggested, Patty's suppliance of past tense morpheme in obligatory contexts "fossilized" at a very low rate of approximate 34% over nearly 9 years of partially or fully exposure to L2 English environment. That is to say, she made about 66% error of -ed marking in obligatory contexts. Enlightened by this finding, we presupposed that Chinese English learner would not achieve any significant improvement even after an intensive Pressional Program at Lancaster University. Table 7 illustrates the result of this longitudinal study. Table 7. Omission of -s in obligatory contexts before and after presessional program As Table 7 shows, the percentage of omission over all obligatory contexts by the subjects were 48.28% before they attended the Presessional Program, yet the figure did not change much (50.00%) by the end of the program. Thereupon, our fourth hypothesis that the issue of omission in affixal agreement cannot be solved even after an intensive training course like the Presessional Program at Lancaster University is supported. We will discuss all the results mentioned above in next part.

Discussion
The results given in the previous part show that adult Chinese English learners produced high rate (nearly 50% , as indicated in Tables 5 and 7) of errors in employing -s marking in obligatory context in their writing. This should not have happened considering the fact that the learners had been learning English in China for at least 8 or 9 years and writing is usually thought to be formal and involves self-monitoring and self-correction in the process. Although Patty (Lardiere, 1998a(Lardiere, , 1998b made about 66% errors (approximately 34% suppliance) of -ed marking and over 95% (0.00% to 4.76%) errors of -s marking in obligatory contexts (Table 1), the Russian Children (Ionion & Wexler, 2002) made 78% errors of -s marking and 58% errors of -ed marking in obligatory contexts (Table 3), they are errors in speech after all. As teachers of English for several years, we know Chinese students make much more mistakes in -s agreement in speaking than in writing; we never expected the situation to be so serious. Meanwhile, similar to what Lardiere (1998aLardiere ( , 1998b and Ionion and Wexler (2002) had indicated, we also found the students were "competent" enough to use -s marking when they actually used it. According to Table 6, only one inappropriate use of -s out of eleven (9.09%) instances of use was found. A further study of the example (2) in 3.3.2 indicates that the use of "develops" might be a result of "attraction of verb agreement" by its adjacent noun "technology" as described by Bock et al. (2001). In this sense, the writer still remembered to observe the rule for third-person singular -s, though he failed in achieving correct expression by misunderstanding the role of "technology" in the context.
The students in our data obviously had some difficulty in observing the rule for agreement in third-person singular -s in obligatory contexts, meanwhile they also somewhat "knew" the rule since they were able to use -s marking correctly once they actually decided to use it. According to Ladiere, this suggests that "syntactic and morphological developments are autonomous in L2 acquisition" (Ladiere, 2000, p. 120). However, she also acknowledged that "much work is still needed in order to understand the complex algorithms which map syntax to morphology" (Ladiere, 1998a, p. 23). Perhaps Hopp is right when he suggests "non-native and native grammars and processing systems are fundamentally identical, with L2 systems being computationally less efficient due to L1 influence" (Hopp, 2010, p. 901).
We have also noticed from the results of both Table 5 and Table 6 that the subjects in our study made more errors in third-person singular -s in obligatory contexts (48.28%) than be form (auxiliary and copula) as a whole (25.00%, 8.7% respectively, and 10% in total) in their writing. Although the figures comparatively decreased as compared with that of Table 3, the results echo what Ionion and Wexler (2002) , Nguyen & Newton (2021), and Wakabayashi et al.(2021) suggested in that L2 learners are more successful in the acquisition of be form than -s form. Zobl and Liceras compared the order of acquisition in L1 and L2, and found out that different from L1 learners, L2 learners acquire auxiliary be prior to third-person singular -s. Both L1 and L2 learners acquire copula be prior to third-person singular -s (Zobl & Liceras, 1994). In other words, L2 learners acquire be form prior to third-person -s form. For Lardiere, "the maping from feature to form is arguably much 'easier' for case on pronouns (in English) than for tense and agreement on verbs" (Lardiere, 1998a, p. 22). Ionin and Wexler suggested that "L2 learners initially associates morphological agreement with verb-raising and, thus, acquire forms of be before inflectional morphology on in situ thematic verbs" (Ionin & Wexler, 2002). Table 7 gives evidence that Chinese English learners made almost no improvement in verbal agreement in English after a one-month presessional program at Lancaster University. We could not infer from the study that the presessional program was a failure because we did find evidence of improvement in students' writing in terms of more complex structures and richer vocabulary applied in T3 than in T2. However, we might be able to conclude from our study that the grammar of these Chinese learners seems to have "fossilized" with regard to -s marking. According Wu (2001), there are two approaches to the explanation of fossilization in L2 acquisition: one attributes this to the miss of "critical period", the other believes that there exists "a latent language structure" in every learner, L2 grammar is fossilized only because the structure has not yet "reactivated". As teachers of English, we are more inclined to believe that education is capable of reactivating the latent language structure in our students. We will further discuss this point in next part.

Conclusion and implication
Before drawing a conclusion, we would like to point out the limitations of our study.
Firstly, owing to the limit of time and data, the result of our study will by no means be comprehensive and representative as only 6 subjects and 12 texts are examined.
Secondly, most literature in this aspect study more natural language such as speech while we studied written language which might or might not be suitable for second language study as far as the applicability of UG in L2 is concerned.
However, this study does find something interesting. Omission of -s marking in obligatory contexts did exists in L1 Chinese L2 English learners at a high rate (nearly 50%) though they made little or no mistake when they actually used third-person -s. The students were more successful in using be form than -s form. The grammar of these Chinese learners seems to have "fossilized" with regard to -s marking.
In short, what some other researchers have found concerning -s morpheme in speech also exists in written English. Third person singular -s is proved to be a big challenge for Chinese English learners.
Implication can be drawn for language teaching in China: In China, Children differ in age when they are taught English. As far as we have learned, in some developed provinces, children are taught English in kindergarten while in some underdeveloped provinces, children begin to learn English at least 13 years old because of financial reasons. If "critical period" does exist, then policy-makers need to reconsider the situation.
In order to make the textbook easy to learn, as we find out, most texts and tasks used in our class are usually ijel.ccsenet.org International Journal of English Linguistics Vol. 12, No. 2;2022 adapted in a way that they are not authentic. As a result, Chinese students are either misled or bored with the uninteresting texts and impractical tasks. Therefore, with unauthentic texts and tasks, how can the "latent language structure" be "reactivated"?
As teachers of English for years, we also notice that with the advent of communicative approach (which was introduced to China in 1990s), language teaching and learning in China are more meaning-oriented rather than structure-oriented (which might be a result of misunderstanding of communicative approach). We used to think that "He goes to school at 7 every morning" is as easy as "He is a student" in either meaning or structure. However, the results of our study show this is not true. Therefore, equal attention should be paid in English class in meaning and form. Meanwhile, much needs to be done in finding out reasons behind the omission of inflectional morpheme in spoken and written output, and ways to prevent students from "fossilizing" their grammar in affixal morpheme like -s.