Perception and Production of Thai Learners on English Prepositions

In this paper, I adopt Best’s (2001) Perceptual Assimilation Model (PAM) to account for how Thai learners acquire English prepositions in prepositional phrases and propose the ranking order of English preposition acquisition into three different categories. The ranking is as follows: Category A is a one-to-one semantic mapping between English and Thai prepositions, therefore ranked first suggesting that they would be the easiest to be acquired. Category B is a one-to-many semantic mapping between English and Thai prepositions. Acquisition of Category B should be more difficult than Category A because one English preposition can have more than one correspondence in Thai. Category C is one-to-null mapping between English and Thai prepositions. A preposition that exists in one language can be null in another language. Category C would rank the lowest in terms of acquisition. The participants consisted of 20 graduate students in the MA program at a university in Bangkok. They were placed into medium (8-10) and low (6-7) proficiency levels of English by a placement test called the Language and Instructor System (ELLIS) administered via computer. The two tests used in this study were a grammatical judgment test for English prepositions and a writing test. The two tests were exactly parallel in each item. The correlation between their awareness in spotting incorrect prepositions and the ability to use correct ones were measured using Pearson’s correlation coefficient. The results were consistent with the ranking proposed here. It showed that L2 Thai acquirers of English prepositions were able to judge grammatical and ungrammatical sentences correctly with respect to the ranking A>>B>>C. However, no significant difference of the correlation between perception and production in all categories was found. The results further revealed that both medium and low proficient participants were able to perceive and produce dependent prepositions more accurately than independent prepositions.


Background and Rationale
How Thai students learn English as a foreign language has been long studied with respect to their phonological, morphological, syntactic and semantic errors during the course of their learning (Bennui, 2008;Intratat, 2001;Noojan, 1999;Thep-Akrapong, 2005;Tawilapakul, 2002).Explanations usually fall into three basic categories: frequency of errors, types of errors, and causes of errors.The findings are based on two analyses: contrastive analysis, which focuses on similarities and differences between native language and target language (Buren, 1984) and error analysis, which focuses on identifying patterns of errors in interlanguage grammar (Selinker, 1972).However, the theory of universal grammar and second language acquisition turn to principles and parameters of UG in that UG is accessible in both L1 and L2 acquisition (Felix, 1991) while Tsimpli and Roussou (1991) argues that in some circumstances UG is not accessible with some functional categories in L2 acquisition, be it tense, modals, agreement, plurals, determiners and prepositions (Finney, 2005).This is because functional categories are independent of UG.If this hypothesis holds, second language learners will not have direct access to them.In light of much research on second language acquisition, we are led to be convinced.A number of works; for instance, have investigated how English language learners from Asian speaking countries have acquired English prepositions by adopting CA and EA with participants whose first language are Asian languages (i.e., Kim & Ahn, 2003;Castro, 2013;Al Khotaba, 2013).Kim and Ahn's work looked at how Asian speakers acquire English prepositions following the Natural Approach proposed by Krashen and Terrell (1983).
Most participants, regardless of their proficiency level, chose the correct answers on the must-be-used prepositions.This suggested that the must-be-used prepositions are easier to acquire than optional and obligatory deletion ones.Castro and Al Khotaba; on the other hand, focused their studies on examining types of errors and causes of errors.Castro found that the majority of the prepositional errors made by Filipino undergraduate students were accounted for by intralingual errors.Khotaba found no differing results on Arab graduate students in Malaysia.The errors in their use of prepositions were diagnosed as interlingual errors.Best's Analysis (2001) In this paper, I assume Best's (2001) Perceptual Assimilation Model (PAM).Best proposes that the assimilation of L2 sounds/phones into the native system of phonemes as follows:

Prediction's from
A non-native phone may be perceptually assimilated to the native system of phonemes in one of three ways: (1) as a categorized exemplar of some native phoneme, for which its goodness of fit may range from excellent to poor (2) as an uncategorized consonant or vowel that falls somewhere in between native phonemes (i.e., is roughly similar to two or more phonemes) (3) as a nonassimilable nonspeech sound that bears no detectable similarity to any native phonemes (p.777).
Assuming PAM, I propose the ranking order of English prepositions acquired by Thai learners into three different categories.The ranking is as follows: 1) Category A is a one-to-one semantic mapping between English and Thai prepositions.This suggests that the direct translation is identical in both languages.The Thai version comes from the first lexicon that appears in a Thai dictionary which is presumably used most frequently.2) Category B is a one-to-many semantic mapping between English and Thai prepositions.One English preposition can have more than one correspondence in Thai and 3) Category C is one-to-null mapping between English and Thai prepositions.
In Category A, each morpheme is assimilated equally well or poorly to a single native morpheme, in Best's term Single Category assimilation (SG).We predict that Category A will be the easiest one to be acquired due to its identical semantic relationship between English and Thai prepositions.Category B is a one-to-many semantic mapping between English and Thai prepositions.One English preposition can have more than one correspondence in Thai.Best explains that when there is one of the two fits better than the other, it will be termed a Category Goodness difference (CG).Category C is one-to-null mapping between English and Thai prepositions.If both non-native phones are uncategorized speech segments, they will be termed uncategorized Speech Segments (UU) (Best, McRoberts, & Goodell, 2001). 1) Degree of accuracy in perceiving prepositional errors in English will correlate with the ranking order A>> B>> C regardless of the participants' level of English proficiency and linguistic backgrounds.
2) Degree of accuracy in using English prepositions will correlate with the ranking order A>> B>> C regardless of the participants' level of English proficiency and linguistic background.
3) Thai learners of English would be able to perceive and produce dependent prepositions more accurately than independent prepositions.4) There is a correlation between syntactic perception and production.
The hypotheses predict the following: 1) Functional categories are independent of UG evidencing from low degree of accuracy in correctly perceiving and producing functional categories.
2) Parameters are associated with functional categories.Parametric variation between L1 and L2 results in the transfer of the L1 value to L2 as in the proposed ranking.

Participants
The participants for this study consisted of 20 graduate students in the MA program of Linguistics and Educational Linguistics at the faculty of Humanities, Srinakharinwirot University.Language and Instructor System (ELLIS) administered via computer was used to test the participants prior English proficiency which has 12 levels made up of 3 components; listening, vocabulary and grammar: Level 1-2 (ELLIS Basics) are for the beginners and Level 3-7 (ELLIS Intro) are an introductory course, while Level 8-10 (ELLIS Middle Mastery) are an intermediate course and Level 11-12 (ELLIS Senior Mastery) are an advanced course for prospective English teachers and those whose expertise lie in English.Some students (n = 13) were placed into Level 8-10 (medium competence level in English), while others (n = 7) were placed into Level 3-7 (low competence level in English).

Instruments
The two tests employed in this study were a grammatical judgment test for English prepositions (see Table 1) and a writing test.The carrier sentences were actually from advertisements seen either on billboards or print magazines in Bangkok.The participants are exposed to the incorrect items on a daily basis.On the error recognition test, there were 10 minimal pairs of sentences made up of 5 pairs of independent prepositions and another 5 pairs of dependent prepositions, which must occur only after some verbs, adjectives and nouns.The subjects were asked to look at them and identify whether they were aware of the wrong use of English prepositions from the carrier sentences and able to identify them.Their options were the followings: 1) one sentence was right.2) one sentence was wrong.3) both sentences were right and 4) both sentences were wrong.Those sentences were checked by two English native speakers and corrected.To avoid memory effect, one week later they were asked to translate Thai sentences into English.An English verb, adjective or noun was provided in each sentence to control for the accuracy of preposition use.Also, they were instructed to use prepositions in every single sentence even though no Thai prepositions exist in some sentences.The two tests are exactly parallel in each item.The correlation between their awareness in spotting the incorrect prepositions and the ability to use the correct ones were measured.
The other instrument measured the ranking order of English preposition acquisition proposed for Thai learners.One-to-one mapping ranked above the one-to-many mapping in a tableau.In turn, one-to-many mapping dominates one-to-null mapping.The order is as follows: one-to-one >> one-to-many >> one-to-null.The sentences in the two tests, the first measuring the perception and the second measuring the production, were classified and put into the ranking categories such as A, B, and C. English proficiency and linguistic backgrounds.The results revealed that there was a slight difference in their judgment of prepositional error detection at 4.48% from medium to low in category A. As for category B, the gap was further apart at 8.43%, while the medium proficient participants perceived category C much better than (12.54%) the low proficient ones.However, the fact that their means were not high indicated that their English proficiency could be a determining factor of their inaccessibility of UG in functional categories.
The analysis correctly predicts that the participants performed better in category A because there is no parametric variation between L1 and L2; in category B, it gets more difficult to perceive the errors as parametric variation increases.The most difficult environment in correctly identifying errors is in category C, in which certain English prepositions do not operate in L1 at all.

Hypothesis 2
Evidence from the difference in performance between perception and production in category A, B, and C further indicate that acquisition of production is more difficult to be than perception.Both groups of participants produced English sentences using correct prepositions in the order predicted A >> B >> C.There was no difference between the medium proficient participants and the low proficient ones in category A. The former performed moderately better than the latter at 5%.Interestingly, both groups' performance substantially decreased in category B from 35.9% to 23.81%.However, the medium ones were still able to use more correct English prepositions than the low ones at 12.09%.Notice that the medium group performed category C (51.28%) much better than category B (35.9%).The low group's performance considerably dropped at 23.81% and 21.43% in category B and C respectively.The Thai L2 acquirers perform in ways that conform to the model Figure 5.The means that showed percentages of medium and low proficient participants' productions on English prepositions

Hypothesis 3
The third hypothesis was confirmed as shown in Figures 7 and 8.Both medium and low proficient participants were able to perceive and produce dependent prepositions more accurately than independent prepositions except that the low proficient ones produced independent prepositions more accurately than those with dependent ones on the test sentences in category A.
Figure 7.The means that showed percentages of medium proficient participants' perceptions and production of English independent and dependent prepositions A1, B1, C1 = independent prepositions A2, B2, C2 = dependent prepositions Figure 8.The means that showed percentages of low proficient participants' perceptions and production on English independent and dependent prepositions A1, B1, C1 = independent prepositions A2, B2, C2 = dependent prepositions Table 4 gives us insightful data on how Thai participants were more accurate in their grammatical judgment of dependent English Preposition Phrases than the independent ones in category A e.g."You will be satisfied with our service."and "If you are interested in our product, contact us."As for category B, they were able to perceive (95%) and produce (85%) significantly better in the phrase "Thanks for using our service."than the other sentences of all categories.This high accuracy can be accounted for by frequency of exposure which impacts the choice of preposition.
However, the participants performed worst on judging the use of the preposition in the sentence "Reward your dearest mom with fabulous gifts.Ladies dept discounts at 10-30%."And none of them was able to use the correct preposition in the sentence "We discount at 30-50%."They all chose the preposition "from10 to 30%".In light of this result, we can conclude that the participants employed direct translation of the preposition used in their L1 when such translation was there to grasp.In Schieber's psycholinguistic experiment (2013), she found that when there was a mismatch between L1 and L2 translation of the preposition, the reaction time was longer and accuracy was less correct.The explanation basically comes down to conceptualization how second language learners can conceptualize the language.

Hypothesis 4
To test whether the relationship between perception and production is bidirectional , Pearson's correlation coefficient was performed.There was no support for hypothesis 4 in all categories, which predicted a correlation between ease or difficulty in perceiving ungrammatical or grammatical use of English preposition associated with producing them.No significant difference was found from the correlation between perception and production of category A (r = -.062,p = .796)and C (r = .135,p = .569).There was, however, a weakly positive correlation between perception and production of category B (sig = r = .261,p = .266).Flege, (1995) stated if one cannot perceive a speech sound in a language, s/he will not be able to produce that speech sound either.This suggests that the perception and the production of speech sounds are correlated.Conversely, some prior studies conducted, such as by Chanintaratheip (2013), indicated that the perception and the production of English final consonants by Thai learners of English did not bear any relationship to each other.Likewise, Altenberg (2005) revealed that the perception and the production of English consonant clusters by native speakers of Spanish acquiring English were not correlated.The research findings, therefore, are tacitly consistent with those of the two research studies mentioned above.The results support a view that functional categories in L2 acquisition are independent of UG and L2 learners, regardless of their proficiency level, are not yet able to master prepositions.

Conclusions
In this paper, I adopt Best's Perceptual Assimilation Model to account for the acquisition of English prepositions by Thai learners.Best's model is extended to three ranking categories of L2preposition acquisition according to their semantic correspondence between L1 and L2 prepositions.The results show that the more L1 and L2 are assimilated, the easier it is for native speakers to perceive and produce English prepositions.However, when there is a greater degree of parametric variation between L1 and L2, functional categories in L2 are evenmore difficult to acquire.One methodological flaw was the researcher's assumption that English native speakers as a control group would score 100% and this might pose to be problematic in validating the research results.

Figure 4 .
Figure 4.The means that showed percentages of medium and low proficient participants' perceptions of English prepositions

Figure 6 .
Figure 6.A comparison between medium and low proficient participants

Table 1 .
Grammatical judgment test on independent English preposition phrases (PP)

Table 2 .
Grammatical judgment test on dependent English preposition phrases (PP)

Table 4 .
Score ranking of the two tests