Thai EFL University Students’ Productions of the English Past Counterfactuals and Their Influences from Interlanguage Fossilization

The English past counterfactuality is difficult for non-native learners to learn and understand due to cross-language discrepancies. Oftentimes, individuals unavoidably develop the interlanguage and this language system becomes fossilized regardless of the amount of exposed time and input. The purpose of the study was twofold: (a) to examine 126 Thai EFL university students’ production of the English past counterfactuals through the grammar translation method (GT); (b) to qualitatively investigate how interlanguage fossilization exerts an influence on the students’ production. The results showed that their ability to use the English past counterfactuals was relatively low, whereas the errors of its structural elements could be found across the entire production. It was further observed from the interview that the low productive ability resulted from the L1 transfer and the training migration. The pedagogical implications for foreign language instructors are discussed.


Introduction
Every language has different theoretical formations and ways to express conditional sentences (Rosa & Leow, 2004;Traugott et al., 1986). This differentiation has attracted scholarly attention to investigate formational structures, meaning, objectives, and how they are acquired. English conditional sentences are regarded as cognitively and linguistically subtle as they can convey different meanings depending on forms and discursive functions (Folse, 2017;Norris, 2003). Some types of conditional sentences in English such as the past counterfactual conditionals are considered difficult to acquire due to its structural, syntactic, and semantic complexity (Celce-Murcia & Larsen-Freeman, 1999;Mindt, 1996). This complexity encompasses several linguistic aspects namely forms, meanings, and subject-verb and tense-relational agreement (Nekoueizadeh & Bahrani, 2013). Therefore, speakers of other languages whose verbal inflection regarding temporal aspects is uncommon in other linguistic systems such as Thai and other Asian languages are likely to encounter difficulties in precisely expressing their conditional statements in English (Conroy & Cupples, 2013;Katip & Gampper, 2016).
Particularly in the Thai language, according to Jaszczolt and Srioutai (2011), the expression of past-time reference can be made in several ways and does not necessarily rely upon an explicit lexical or grammatical marker of temporality like in English. Rather, it is the combination of information gained from "standard, presumed meaning or pragmatic inference" (p. 275) that facilitates a speaker to assess the temporal location of the situation at hand. This indicates that the semantics of past-time reference is represented differently both in English and Thai. Considering this issue along with the complex notions of conditionals, past counterfactual conditionals may be among the most inexplicable grammar teachings for Thai EFL learners. In the light of these factual arguments, we argue that the traditional grammar-translation (GT) teaching method would be useful for the teaching of English past counterfactuals as it gives the accurate meaning of the L1 counterparts (Khan, 2016), thereby ensuring the learners' accurate understandings of this complex notion of temporality in English and Thai. Besides, most Thai learners are familiar with the GT method (Noom-ura, 2013;Tieocharoen & Rimkeeratikul, 2019).
Hence in this study we employed a grammar-translation task to quantitatively measure the problematic patterns identified in the production of past counterfactual conditionals among Thai learners. The patterns emerged in the present study will further inform language teachers of detailed mistakes frequently made by the learners after we have learned that the capability to correctly use and produce English past counterfactuals by Thai learners was relatively low (Hinnon, 2015;Nonkukhetkhong, 2013;Sattayatham & Honsa, 2007).
In addition, language fossilization, a condition where target language acquisition remains static, is one key linguistic problem found during the process of second language learning. Research shows that second language (L2) learners are unable to acquire further linguistic features of L2 and several errors remain prevalent and repeated (Hartshorne et al., 2018;Selinker, 1972). Fossilization may result in demotivation and frustration for both learners and instructors. Given the vital importance of fossilization, studies on language learning have emphasized this phenomenon (Yang & Xu, 2019). Hence, fossilization can be another difficulty pertaining to the inability to acquire the past counterfactual condition. Despite a number of issues regarding interlanguage fossilization that may potentially exert an influence on the erroneous patterns of past counterfactuals produced by Thai EFL learners such as first language interference, the unavailability of verb inflection in the Thai language system, and deficient vocabulary knowledge (Hinnon, 2015;Nonkukhetkhong, 2013;Sattayatham & Honsa, 2007), these challenges have not been explicitly discussed and remain to be further investigated. As a result, this study set out to qualitatively explore how the aforementioned issues regarding interlanguage fossilization would influence the production of English past counterfactual statements among Thai learners.
This study is worth investigating as it can provide educational practitioners with an in-depth understanding of how language fossilization plays a role in the production of English conditional sentences, particularly the past counterfactuals, of Thai EFL university students. The insights will also inform educators to design suitable strategies and materials to help L2 learners overcome difficulties in acquiring and utilizing English conditional sentences. The current study addresses the research questions as follows: 1. How do Thai EFL university students produce English past counterfactual sentences? 2. How do Thai EFL university students account for the production of English past counterfactual conditional sentences?

Conditional Sentences
English conditional structures. The conditional construction consists of if-clause (the antecedent or the protasis) and a main clause (the consequence or the apodosis) (Ferguson et al., 1986, p. 5;Huddleston & Pullum, 2010). A situation or action in the main clause happens if a particular event in the if-clause takes place. The fundamental structural meaning of conditionals is constructed with if p, (then) q (Ferguson et al., 1986, p. 5). The conditional sentences are inherently presupposed, conveying different meanings, such as cognitive reasoning, imaginative, and hypothetical (Celce-Murcia & Larsen-Freeman, 1999). The sequence of these clauses can be alternated without affecting the meaning (Sandford, 2003). Three different tenses can be applied to the if-clause, depending on the conditional types: present simple for type 1, past simple for type 2, and past perfect for type 3. Meanwhile future simple, present conditional, and perfect conditional are employed in those three if-clauses, respectively (Thomson & Martinet, 1986).
English conditional sentences can be categorized into four types and named differently according to purposive emphasis as follows: Type zero (or factual conditional), Type 1 (or future predictive), Type 2 (or present counterfactual), and Type 3 (or past counterfactual) (Katip & Gampper, 2016;Folse, 2017). Table 1 shows the four kinds of conditionals in English, with example sentences.
Each type of English conditional can refer to the different degree of the situational probability. Factual conditionals express 100% probability, indicating the true occurrence of the subsequence if the preceding event takes place. Predictive conditionals, hypothetical counterfactual and counterfactual conditionals hold 50%, 10%, and 0% of probability, respectively (Wu, 2012).
The inherent subtleties of conditionals, particularly imaginative, have received research attention (Berent, 1985;Chou, 2000;Jung et al., 2005;Ko, 2013;Norris, 2003). Jung et al. (2005) pointed out that learners confronted usage difficulties of conditionals, compared to other sentential structures due to uncertainty of appropriate verb forms. Norris (2003) also indicated that the acquisition of conditional sentences was one of the problematic issues among L2 learners. Several studies also attempted to find systematic ways to acquire conditionals from task performance of EFL learners with different L1 backgrounds (Berent, 1985;Chou, 2000;Ko, 2013).
Regarding semantic aspects, Celce-Murcia and Larsen-Freeman (1999) proposed that conditional type 1 reflects the future of predictive events while type 2 and type 3 manifest imaginative situations. Specifically, type 2 corresponds to present/future hypothetical and present counterfactual conditions whereas type 3 concerns past counterfactual conditional events (Celce-Murcia & Larsen-Freeman, 1999). Additionally, type 2 yields present or future time reference in the if-clause. Type 3, on the other hand, has past time reference. As for the counterfactual meaning, English conditions with Types 2 and 3 create a possible world that is contrary to the actual world (Declerck & Reed, 2006). The conditional Type 2 is counterfactual to present events and Type 3 is counterfactual to past events. Significantly, Declerck and Reed (2006) stated that not only does the third conditional indicate the past and present counterfactuals, but it also explains the future counterfactual. These clauses are in a subjunctive mood. They can be interpreted both indicatively and hypothetically (Folse, 2017), making conditional sentences, especially types 2 and 3, comprehensively problematic for learners of English.
Additionally, Ogihara (2013) posited the antithesis to the concept of counterfactual conditions in English based on the semantic theory. Not only do the past counterfactual clauses (type 3) indicate impossible beliefs in the past becoming true in the present, but they also bring up to fictitious events, such as imaginative actions or events expected to happen in the future; for example, "If his son had been born TOMORROW (instead of yesterday), John would have been ecstatic." This study highlighted the comprehension of the past counterfactual condition in terms of semantic understanding and structural patterns. Briefly, the structural complexity and tense uses turn conditional sentences into a problematic issue in comprehension and utilization among language learners compared to other sentential aspects (Lord, 2002;Mindt, 1996).
Thai conditional structures. L2 learners are more likely to rely on semantic and lexical information to comprehend and produce language, which consequently help them perform better in discourse-based tasks (Clahsen & Felser 2006;Conroy & Cupples, 2013;VanPatten, 1996VanPatten, , 2006. Contradictory to this argument, it was found that non-native learners from Asian L1, including Thai and Chinese, tend to relatively less use of and confront difficulties in employing modal perfect (MP) and counterfactual sentences, perhaps due to cultural (Conroy & Cupples, 2013) and grammatical differences (Katip & Gampper, 2016).
Particularly, Thai conditional sentences can semantically express four major intentional purposes: conditions, assumptions, alternatives, and persuasion (Niemboobpha, 2001). Thai language has no definite categorization in distinguishing different types of conditional sentences. In one conditional sentence, the meaning can be mainly expressed in three possible time references, either present, past, or future, owing to the non-existence of morphological inflection in the language system. Hence, contexts are required in order to determine the meaning interpretation of the sentences (Suteerapongsit & Pongpairoj, 2020). Usually, "if" is applied to the beginning of subordinate clauses. They state a subsequently true condition of the main clause. However, the omission of the "if" can oftentimes happen in abrupt speech (Smyth, 2002). For example, "รู ้ แบบนี ้ ไม่ ให้ ก็ ดี " /ɹú baep nɪ́ mâi hâi kɔ̂ di/ can be translated into "(If) I had known about it, I would not have given any more." Apart from omitting the "if," Thai does not have a morphological inflection, which is grammatically deviant from English.
Furthermore, one grammatical feature in Thai conditionals that differ from those of English language is the emphasis of frequent appearance of adverbial markers. These adverbs are employed to indicate the temporal conditions in a sentence, similar to the use of modal verbs in English. For instance, ถ้ าเขาไม่ กลั บบ้ านดึ กในคื นนั ้ น ก็ คงไม่ ต้ องเสี ยเงิ น มากขนาดนี ้ /táa kǎu mâi klàp dɯ̀k nai kɯn nán kɔɔ mâi tɔŋ sǐa ŋəən mâk kanàt ní/ If he had not returned home late that night, he would not have spent a lot of money on the taxi fare. In this sentence, /kɯn nán/ (that night) and the conjunction /kɔɔ/ (so) are used to indicate the third condition as the use of the modal perfect (MP) in English.
To measure Thai EFL learners' understanding of the structures of the three conditional types, Katip and Gampper (2016) examined a production of English conditional sentences using the gap-filling task. The past simple form was prevalently produced by the Thai EFL learners in both counterfactual types. Specifically, the highest number of errors was congregated in the past counterfactual. Similarly, Sattayatham and Honsa (2007) investigated grammatical errors Thai learners of English made in their English writing. The past counterfactual condition was the part wherein the errors were highly and frequently found. This could reflect the incomplete acquisition and internalization of the past counterfactual due to its grammatical complexity, when compared to other conditional types (Chou, 2000;Ko, 2013;Sattayatham & Honsa, 2007). Moreover, another difficulty the Thai learners confront is the inflection of verb forms. Thai is viewed as an isolating language in nature. In other words, the morphological inflection is non-existent in the Thai language system. Hence, verbs in Thai are not inflected to indicate time reference. This results in a possibility of meaning expression in different timeframes. The meaning is, therefore, contextually reliant (Suteerapongsit & Pongpairoj, 2020). Given the difficulty in acquiring L2 forms, conditional sentences appear challenging to Thai EFL learners. The linguistically, including grammatical, disparate issues can provoke the misinterpretation of the meaning and the erroneous production of English counterfactual sentences, particularly the past counterfactual. It is, hence, necessary to understand how Thai EFL learners produce English past counterfactual sentences.
Interlanguage fossilization in grammatical learning. Second language learners are challenged with reaching the proficient level equivalent to native speakers of the target language. Approximately 5% of them are able to attain this level (Selinker & Rutherford, 2014). Such a situation has been defined in the literature as fossilization. Fossilization, according to Selinker, is the phenomenon where linguistic knowledge of learners becomes fixed during the process of language learning. However, it is noteworthy that some L2 competencies could possibly be reached at the native-like levels during the process of learning whereas some may remain unattainable and become fossilized in the interlanguage (Vavilova & Broadbent, 2019). This results in a deviant production of linguistic features and grammatical forms produced by learners in the target language. A number of linguistic forms produced by the learners with fossilization tend to be wrong and persistent, though some of the fossilized features might be rectified (Yang & Xu, 2019).
The fossilization usually occurs in the interlanguage where linguistic items are being acquired and internalized. Interlanguage (IL) is a unique linguistic system resulting from a combined characteristic between native and target languages (Dai & Cai, 2001;Dai & Zhou, 2005). Initially, learners employed their native language, along with fundamental and familiar knowledge of the target language, in their L2 production. This system is constructed based on the learners' own assumptions to assist their production and comprehension of the target language. However, grammatical features employed by the learners' production are imperfect and remain correctible when exposed to the new L2 inputs, perceived dysfunctional, or appeared incomprehensible (Krashen, 1982;Selinker 1972).
The causation of fossilization has been classified into five processes: native language migration, over-generalization of linguistic rules, training migration, learning strategies, and communicative strategies (See Selinker, 1972). Fossilization is deemed common in second language learning. Yang and Xu (2019) indicated that the fossilization phenomena might be accounted for several constraints, such as the absence of a language learning environment, testing-oriented pedagogical emphasis, imbalance between teachers and students, and low proficiency level of learners. The fossilization influences a TL production, where certain L1 features appear. That is, certain linguistic features of TL are unable to be developed.
Fossilization can take many forms and emerge at any learning level (Han, 2013;Yang & Xu, 2019). With regard to grammatical perspectives, fossilization may be evoked in relation to L1 transfer (Han, 2013). Khurshid et al. (2016) studied the use of present tense among Urdu college learners. The empirical analysis manifested the presence of errors in usage. Learners were more likely to overgeneralize the rules and produce the target language with the assistance of their L1, particularly through direct translation from L1 to L2. The results highlighted the grammatical discrepancies between L1 and L2, which might lead to confusion and interference in learning development.  examined pragmatic fossilization in English usage among Persian university students. The errors of English pragmatic usage are attributed to L1 reliance, L1 transfer, limited target language exposure, overgeneralization of the target language, contextual unfamiliarity, and grammatical errors.
Furthermore, research by Tajeddin and Tabatabaeian (2017) investigated grammatical, lexical, and cohesive fossilization among EFL Iranian advanced learners. The participants exhibited incorrect usage of forms and failure of noticing, leading to the persistence of the utilization of nontarget-like structures. The findings also suggested the influence of L1 background on usage and language learning. Recently, Tang (2020) examined the acquisition and the use of English finite and non-finite verbs among Chinese EFL learners. The participants were divided into three proficiency levels: low, intermediate, and high. The problematic usage of the verbs in different syntactic types was found across the levels. The misuse of forms, inflection, and functions of verbs was particularly noted among the low-proficiency group. The error commission was attributed to the influence of L1. Learners appeared to transfer morphological and syntactic knowledge of L1 to assist their L2 production.
The review of the literature has acknowledged the use of conditionals as well as their structures, contexts, grammatical complexity, and L2 fossilization, which is a major barrier of non-native users and language learners. Many strategies have been consequently proposed to help learners improve their learning performance of the target language. However, the focus of the prior research has been emphasized on the use of verbs and tenses. There remains a paucity of studies investigating conditional aspects, particularly the past counterfactual sentences. To address the research gap, the current study aims to examine the production of the English past counterfactual conditionals by Thai EFL learners.

Methodology
Research design. The current study employed mixed method research design with explanatory sequential approach (Edmonds & Kennedy, 2017). This approach comprises two phases: quantitative and qualitative, allowing researchers to further elaborate the quantifying data from the first phase with the subsequently collected qualitative data in the second phase. In this study, a translation task of Thai sentences into English past counterfactuals was used to retrieve quantitative information representing linguistic performance regarding a production and the interpretation of English past counterfactuals from Thai sentences. The qualitative data was derived from a semi-structured interview regarding the influences, the interlanguage fossilization in particular, on the English past counterfactual production.
Participants. One hundred and forty-one undergraduate students (116 women and 25 men), enrolled in a compulsory fundamental English course at a public university in middle Thailand, were purposely recruited in this research. The average age of the entire participants was 19.96 years (SD = 3.21). Seventy-nine (56%) were pursuing their bachelor's degrees in nursing whereas 62 (44%) were pursuing bachelor's degrees in medical-related programs (e.g., paramedics, radiology, and medicine). According to their experiences in English learning, 70 students (49.6%) had 11 to 15 years, 53 (37.6%) had 6 to 10 years, 14 (9.9%) had 16 to 20 years, 3 (2.1%) had 0 to 5 years, and only 1 (0.7%) had 21 years or above. A majority of the students (72.7%) graduated from a high school program taught in Thai whereas the rest (27.3%) did from a bilingual or international program. The Common European Framework of Reference for Languages (CEFR) scale has been adopted to reform all aspects of English langue curriculum in Thailand since 2014 (Franz & Teo, 2017). With the adoption of CEFR, students who graduate from the Secondary Grade 6 are expected to be independent users (B1) (English Language Institute, 2015). Since all of the students in this study received a high-school diploma, their English proficiency level were at the independent-user level.

English Language and Its Status in Thai Educational System
English is designated as a compulsory foreign language subject delivered throughout 14 years of free fundamental education of Thai state schools, from pre-school to secondary levels (The Ministry of Education, 2014). Various pedagogical approaches, including grammar-translation (GT) and communicative language teaching (CLT), have been implemented to help reform English language curriculum for Thai educational institutions (Darasawang, 2007;Franz & Teo, 2017;Tieocharoen & Rimkeeratikul, 2019). In addition, different educational programs, such as English Program (EP) and English Bilingual Education (EBE), have been established. However, the improvement of English language proficiency among Thai students remains at the unsatisfactory level or even deemed deteriorating (Todd & Shih, 2013). In order to improve standards of English among Thai students, the Ministry of Education in Thailand has implemented the European Council's framework of reference for language proficiency (CEFR) into the recent reform English language curriculum for educational institutions across the country (The Ministry of Education, 2014). Students graduate from Thai educational institutions are expected to reach certain CEFR levels. Table 2 manifests the target CEFR levels for such students.
Different types of conditional sentences are introduced and reintroduced to students as they progress through their educational years, depending on topics focused on and textbooks utilized in particular Grade levels. In general, a single type of conditional sentences is provided in a lesson unit. However, the past counterfactual conditional structure is oftentimes featured in a lesson in conjunction to other conditional types in English textbooks indicating Immediate level, at minimal (Stephenson et al., 2017). In terms of teaching English grammar in Thailand, including conditional sentences, the grammar translation approach (GT) is predominantly adopted due to its benefits on grammatical comprehension (Noom-ura, 2013;Tieocharoen & Rimkeeratikul, 2019). For instance, GT explicitly explains grammatical rules along with sentential examples and exercises. The direct explanation bolsters learners' understanding whereas the exercises allow learners to practice, thereby becoming familiar with the usage (Harmer, 2017;Krashen, 1982;Larsen-Freeman, 2000;Thornbury, 1999). Concerning syntactic and semantic subtleties, the past counterfactual aspect is presented with structural constructions, usage, examples, and exercises to facilitate understanding and acquisition (Declerck & Reed, 2006;Garner, 2016;Stephenson et al., 2017). Upon the graduation of students' Grade levels, they at least become familiar with conditional structures and usage, including the past counterfactual. Data collection and instrument. Data regarding the counterfactual conditions were collected through a translation task on Google Forms. The research instrument consisted of two major sections: a demographic survey and a translation task. More specifically, students' demographic information included their age, gender, and educational backgrounds. The translation task was composed of ten Thai sentences, used to examine students' written production of the English past counterfactuals. All the sentences were designed and verified by a group of experienced English and Thai language instructors in terms of comprehensive clarity. In the task, students were asked to translate ten Thai conditional sentences into English. Forty minutes were allotted for a completion of the task. Prior to the data collection, the participants were informed about research purposes, task explanation, and identification anonymity of their responses.
Of the 126 students, the online interview was further conducted, based on convenience and accessibility, with 24 students to obtain insights into their understanding and production of the English past counterfactuals. The interview lasted for approximately 25 minutes. Prior to the interview, the participants were informed about the research purposes, the anonymity of their responses, and the record permission. The interview was on a voluntary basis, allowing them to withdraw at any time. The interview questions focused on the thinking process during performing the translation task, such as the selection of word choice and tenses and the interpretation of conditional aspects.
Data analysis. We first employed descriptive statistics to overview the frequency of patterns regarding the produced English past counterfactuals. Furthermore, we analyzed the qualitative data by means of open coding and thematic analysis (Braun & Clarke, 2020;Clarke & Braun, 2017;Koomen et al., 2016). The obtained data was transcribed, defined, and assigned codes for emerging ideas to generate initial themes to indicate significant broader patterns of meaning, or socalled potential themes. These themes were then refined, selected, and verified, respectively by the authors, to represent the possible influences of the interlanguage fossilization on the students' English past counterfactual production.

Results
Conditional sentences are a combination of two causal, consequential clauses: if-and main clauses -protasis and apodosis, respectively. The if-clause presents an outcome resulted from an eventual occurrence in the main clause Huddleston & Pullum, 2010). Possible English fundamental conditionals which are presented in grammatical references and EFL textbooks appear in four patterns, namely conditional types 0, 1, 2, and 3. Type 0 regards a real condition whereas type 1 refers to future or predictive conditional sentences. Types 2 and 3 correspond to imaginative conditions. Specifically, type 2 is associated with the present or future hypothetical and present counterfactual conditions whereas type 3 concerns past counterfactual conditions (Celce-Murcia, & Larsen-Freeman, 1999). Conditional sentences have been a radically perpetual problem in language teaching and learning owing to their syntactic and semantic complexity and variety, particularly the type 3 (Celce- Murcia, & Larsen-Freeman, 1999;Covitt, 1976). This study then sought to explore how Thai EFL university students produced the past counterfactual sentences in English.
One hundred and forty-one students responded to the translation task. However, the responses of the 15 students were incomplete and excluded, leaving a total number of 126 participants. Therefore, the remaining number of responses included for the analysis was 1,260 (each participant produced 10 responses).

How do Thai EFL University Students Produce English Past Counterfactual Sentences?
As seen in Table 3, two common patterns regarding students' production of the English past counterfactuals were identified, all of which were considered major errors as they were the principal components of the past counterfactuals such as the misuse of modal perfect (MP) (1,260 sentences -100%) and frequently applied tense patterns (1,151 sentences -91.3%). We additionally found two minor erroneous patterns, namely, 47.3 % of subject-verb agreement (596 sentences) and 43 % of missing copula verbs (541 sentences). However, we only reported detailed findings of the major errors as this research mainly focuses on the structural elements of the English past counterfactuals (Svalberg, 2018).
Misuse of modal perfect (MP). Despite a clear instruction of the task informing students to produce English past counterfactual sentences from the given Thai conditional sentences, 38.1 % of the answers were produced using type 1 conditional structure whereas 9.1 % were found to be in type 2. However, the type 3 conditional structure was not found in the entire answers (1,260 items). Samples of the answers composed using type 1 and 2 conditional structures are demonstrated in Table 4. It was observed that the full structure of the type 3 conditional sentence was not found across the 1,260 answers whereas the use of MP in their past counterfactual main clauses was substantially low (2.86 %). Some of the students' answers are exemplified in Table 5.
Students' frequently applied tense patterns in the English past counterfactual conditionals. The frequently identified pattern across students' production of the English past counterfactuals was (1) the present simple with the future simple (38.1%), followed by (2) the present simple with the present simple (32.1%), (3) the past simple with the future past (9.1%), (4) the present simple and the past simple (8.9%), (5) the present simple and the future past (6.2%), and (6) the past simple and the future simple (5.6%). Examples of each pattern are demonstrated in Table 6.
Interestingly, the past temporal markers were prevalently found in clauses regardless of tenses as in Examples 17,19,20,and 23. Despite the mismatch of tenses in producing conditional sentences, the use of past perfect tense was slightly found across students' answers. Examples of clauses in past perfect tense are illustrated in Table 7.

How do Thai EFL University Students Account for the Production of English Past Counterfactual Conditional Sentences?
In this study, we identified two major themes from the interviews. These themes, namely native language migration and training migration, to a certain degree, are relevant to Selinker's (1972) causes of fossilization. Given the second purpose of the study, how these three fossilization processes affect students' production of the English past counterfactuals are illustrated in the forms of assertion as follows: Assertion 1: Native language migration (L1 Transfer / interference). The use and the reliance of L1 are the most commonly adopted strategies among Thai EFL learners. This is due to the limited exposure of the target language and the dominant usage of the native language on a daily basis. The students were more likely to process their thoughts in L1 upon encountering or prior to expressing ideas in the target language. Interestingly, several of them mentioned the recognition of the past counterfactual and perfect tenses. However, they were uncertain about the aspects and the structures, thereby relying on their L1 to interpret and produce their English past counterfactuals. The following excerpts present that most of them began translating Thai sentences into English by using L1 knowledge whereas literal translation was the major strategy they used to perform the task. I looked at each sentence and compared them to what was available in Thai language. This involved a selection of tenses. For example, when I saw the past tense markers, I thought it would be the past simple. I wasn't sure about the counterfactual usage. I then decided not to use them and opted for simple ones. (Student 8,Interview, Additionally, some students specifically explained that they not only used their L1 to interpret the given statements, but also simplified and rearranged those statements into their own comprehensible prose. They then directly replaced each word in those simplified statements with simple or familiar English ones. They also employed back-to-back translation to check the appropriateness of their production. I first thought in Thai and arranged each word based on Thai structures. I literally replaced each Thai word with those of English I was familiar with or heard of.... I then used only what I could think of or existed in Thai language while performing the task. I also translated the produced sentences in English back

Examples of students' produced sentences
In type-1 structure (38.1%)* In type-2 structure (9.1%)* If Bangkok had not had a lot of cars, the air in the city would have been better.
(1) If Bangkok has few cars in the past, the weather in the city is better.
(2) If Bangkok had very few cars in the past, the weather in the city would probably improve.
If the rescue team had arrived earlier, many people would not have died.
( Though past-counterfactuals are complex in nature and one may question the validity of the findings as translation may be viewed as a more complex cognitive task (Krüger, 2016), we argue that the first assertion "native language migration" remains valid. Though we used the GT method to elicit learners' interlanguage development (Lu, 2010), we consider translation as a regular process of language transfer that occurs automatically and frequently with learners without a high level of foreign language proficiency. Non-native learners often display a high level of language transfer, particularly through an automatic habit of translating from their native language (Hartsuiker & Bernolet, 2018;Lu, 2010). Especially, during L2 acquisition, learners usually hold perceptions about what is transferable and construct their interim guidelines that govern their use of knowledge translated from L1 (Ellis, 1997;Khan, 2016;Schachter, 1988). Hence, in our views, the fact that the participants are thoroughly familiar with the translation, to some extent, compensates the cognitive complexity of the translation process as learners tend to learn more effectively once they feel comfortable (9) If BKK have a few cars as in the past, the air in the city could have been better.
(10) If Bangkok have a few cars as in the past, the air in urban could have been better. Ken would not have been sick if he had eaten good food.
(11) Ken wouldn't have been sick in the past if he ate quality food.
(12) Ken would not have been sick in the past if he ate quality food / nutrition. If the rescue team had arrived earlier, many people would not have died.
(13) Last week, if the rescue team comes first, a lot of people would not have died.  and familiar with a learning task, one of which is learning through translation (Nation & Macalister, 2010). Moreover, familiarization with a task can reduce the effects of complexity on working memory (Zhang et al., 2020).
Assertion 2: Training migration. Training migration was another factor affecting how the participants produced the English past counterfactuals. Training migration is defined as an application of acquired linguistic regulation. It involves recognizable items from training processes, such as exercises and procedural practices (Selinker, 1972;Selinker & Rutherford, 2014). English education in Thailand has generally been grammar-or exam-oriented (Darasawang, 2007;Noom-ura, 2013;Tieocharoen & Rimkeeratikul, 2019). The principally adopted training method is Grammar Translation (GT). It is believed that this teaching method could be reevaluated to provide an easier way to teach specific grammatical aspects (Thep-Ackrapong, 2005). Thai learners are trained to memorize and translate grammatical rules and scarcely exposed to the actual usage of the past counterfactuality in various contexts. Especially, the drawbacks of GT could be that students receive few opportunities to actively engage and interact with lessons and classes. Several interviewees in this study reported that, by means of the GT method, they were able to recognize the structural elements and aspects of the English past counterfactual while performing the task. Nonetheless, influenced by the parallelism of the translation between English and Thai, they expressed their uncertainty of the actual usage of the English past counterfactuality owing to its complex grammatical structures and the vague understandings of its semantic nuances. This pertains to the non-existence of this particular conditional mode in Thai language system, thereby leading to the learners' inability to produce the English past counterfactuals. These arguments are illustrated in the excerpts below: . . . I directly translated each Thai word into English. I arranged them according to basic [sentential] structures in Thai, including subject, verb, and object. I also thought over and over about conditional structures in both English and Thai. However, I was unsure about the usage [of the complex ones], so I just chose the simple ones (Student 5, Interview, 27-01-21).
. . .I started performing the task by simplifying the given sentences into my own language that could be easy to understand and directly translated Thai words into English words I knew. To check the correctness of the structures and tenses, I was slightly aware of them, but uncertain about them. So, I decided to literally translate them into English based on Thai [sentential structures]. (Student 6,Interview, We additionally observed that the training migration could be attributed to the teaching materials. The content and explanation of the counterfactuality aspects in most of the adopted materials are mainly aligned with native speakers' communicating views (Todd, 2015) and their use of language for expressing conditional purposes (Katip & Gampper, 2016;Stephenson et al., 2017;Suteerapongsit & Pongpairoj, 2020). The samples provided in several EFL teaching materials, oftentimes, involve fundamentally explicit scenarios that help students distinguish aspects and usage of each conditional type with ease. However, in fact, the appearance of the present and the past counterfactuals may not be explicitly distinguished in many contextual situations. The majority of the adopted EFL materials rarely address the semantically obscured aspect of this issue, though it is prevalently used in the English language communication and understood with ease among the natives (Katip & Gampper, 2016;Suteerapongsit & Pongpairoj, 2020;Todd, 2015). Such ambiguity may then provoke comprehensive and productive difficulties among learners whose pre-and past-or the concept of counterfactuality are not generally seen and used in their linguistic system.

Discussion and Conclusion
The current study investigated Thai EFL university students' production of the English past counterfactuals and the influences of interlanguage fossilization on their production. The results revealed that they appeared to be less familiar with the past counterfactual aspect. In general, they seemed to ignore or misuse the modal perfects and, instead, apply simple tenses to their produced statements such as the present simple and the future simple (type one) and the present simple and the present simple (type zero). This is congruent with results of Suteerapongsit and Pongpairoj (2020). They found that present and future wish clauses were relatively easy for Thai EFL learners to produce. However, an extensive number of responses were found that their conditional patterns were structurally mismatched, such as the use of the past simple tense in the if-clause and the future simple tense in the main clause.
Additionally, the interviews indicated that the primary factor contributed to the past counterfactual inability was the L1 interference. Negative transfer of L1 is defined as an inappropriate usage of L1 linguistic knowledge to supply insufficient abilities during L2 production (Yang & Peng, 2017). We observed that the interviewees were likely to interpret and translate the given statements in Thai (L1) throughout the process. Simple or familiar English lexicons were primarily selected and arranged into the sentences. More complex sentential structures and perfect tenses were reported to be beyond their consideration.
Minor common errors found across the participants' production of the English past counterfactuals could be attributed to the L1 transfer and the over-generalization of the grammatical rules. The participants failed to inflect verbs relevant to the singular subject in the present tense and omitted the copula verb in sentences where an adjective was present. They might be familiar with their L1 linguistic system, where the morphological inflection for verbs and auxiliaries to indicate the past counterfactuals is non-existent in nature. The findings showed that when they encountered unfamiliar linguistic structures and had interpretive difficulties, they tended to supply their insufficient knowledge of L2 by their L1.
The English past counterfactuals produced by the participants in this study appeared to be affected by L1 transfer and training migration (Selinker 1972, Selinker & Rutherford, 2014. The non-existence of morphological inflection and the grammar-oriented lessons could impede how the learners use MP and acquire the past counterfactual aspect effectively. The scarcity of counterfactual usage and the high dominance of L1 speaking environments may probably limit learners' opportunities to use the acquired L2 knowledge in various situations. Their interlanguage development might then be fossilized. The results are found incongruent with those of Ebrahimi et al. (2015) showing the improvement of English conditional performance among learners who received grammar-oriented instructions (Focus-on-Forms).
The mismatch of tenses found in the production is worth mentioning and discussing though in small percentage. Adverbs or adverbial phrases indicating past time were inserted in the learners' English past counterfactuals regardless of tenses. In the interview, several of them reported that they were not concerned about English conditional tenses, structures, and aspects. They subsequently produced the English past counterfactuals by directly replacing the given statements with simple or familiar English words. These erroneous production could reflect the interlanguage fossilization and the language transfer. The incomplete conditional acquisition and the use of adverbs for temporal indication might hinder how the learners interpreted and produced the English past counterfactuals.
Overall, the current study highlights how the Thai EFL learners acquire and employ the English conditionals, the past counterfactual in particular. The findings revealed that the ability of using MP and the past counterfactual was relatively low. Moreover, the study further implied that the learners' comprehension and production of the English past counterfactuals may be potentially affected by interlanguage fossilization, particularly the L1 negative transfer and the training migration. The findings can be also taken into account to maximize the effectiveness of the instructional pedagogies and to overcome learners' difficulties that may occur during the process of acquisition and utilization of the English past counterfactuals. Furthermore, we contend that the structural knowledge should be explicitly instructed concomitant with the explanation of its usage and that the extra time should be allocated to learners' acquisition and the understanding of the English past counterfactuals.
Lastly, the study has some limitations. First, data were collected from learners of a specific academic domain whereas the qualitative data were accumulated from a restricted number of participants, approximately 19% of those who engaged in the translation task. The findings may not thus be representative to students in other academic domains. Further studies may consider increasing the population size and collecting data from other academic disciplines to make findings more inferential and conclusive. Second, the translation task was employed in this study. The task was relatively passive in nature and could promote, to a certain extent, structural and grammatical competency domains. However, this task may be difficult for learners, particularly non-native speakers of English, to connect, retrieve, and apply the acquired knowledge to the actual communicative usage of the past counterfactuality. One aspect in using counterfactuals is for criticism (Conroy & Cupples, 2013). Speakers of English in collectivism cultures tend to have less propensity to criticize others than their individualist counterparts (Conroy & Cupples, 2013;Hofstede et al., 2017). The use of focused communication tasks, such as impromptu speaking or storytelling, may be alternatively considered for future research as they may be more effective in capturing learners' productive process of the English past counterfactuals.

Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.

Ethics Statement
The ethical permission is unnecessary applied to this study since animals were not involved. The collected data was also anonymized and subject to mere research purposes.