A Corpus-based Study on the use of Phrasal Verbs in Malaysian Secondary School Textbooks

Phrasal verbs (PVs) are one of the most notoriously puzzling aspects of English language instruction. Despite their difficulty and idiosyncrasies, they are of high relevance for ESL/EFL learners because mastery of PVs is often equated with language proficiency. Different from prior researches, this content analysis study seeks to identify the PVs used in the Malaysian upper secondary school textbooks and the frequency count of each PV in each textbook using a corpus linguistic approach. The most frequently PVs in each textbook were compared to the list provided by Biber et al (1999)’s Longman Grammar of Spoken and Written English corpus. Findings from this study revealed that the selection and presentation of these combinations in the secondary school textbooks used in Malaysia depended more on authors’ intuition rather than on empirical findings and pedagogical principles. There were no clear explanations of Phrasal Verbs in both textbooks that were analysed. Despite the large number of PVs in the corpus, their presentation in the textbooks was far from satisfactory with some being over-repeated at the expense of some others. This paper makes a few suggestions to further improve the present treatment of PVs in the textbooks used at upper secondary schools in Malaysia. It is suggested that accurate definitions and appropriate selection and presentation of PVs should be considered. Rather than relying on intuition, Malaysian textbook writers must consider integrating the use of corpus into their selection of PVs to be presented to students.


INTRODUCTION
By the end of the past century, it has become evident that multi-word items (hereafter MWIs) have pedagogical value for ESL/EFL learners. Multi-word items, chunks or prefabs is a sequence of two or more words which semantically (constitute a meaningful, inseparable unit (Moon, 1997). Therefore, to continue the linguistic research, many ELT materials writers have included a lot of MWIs into their reference materials. Now, it is not unlikely to find a great number of collocations, idioms, phrasal verbs (PVs) in ELT textbooks.
As more and more MWIs are appearing in the ELT coursebooks, it is important to examine the basis on which these MWIs are being chosen. Many studies argue that ELT texts writers are adopting a pragmatic approach as they rely on their intuition, experience, and common sense. Thus, the accuracy and usefulness of such a process is questionable (Kamarudin, 2013;Zarifi & Mukundan, 2012;Moon, 1997;Sinclair, 1991;Sinclair & Renouf, 1988). Hence, the selection and presentation of many MWIs is not based on empirical studies. Also, most coursebook writers do not differentiate between coursebook and syllabus which may be problematic for ESL learners (Sinclair & Renouf, 1988).
There is no doubt that MWIs are so important for ESL learners if they wish to attain fluency in English. Learning MWIs from coursebooks, however, is still quite far from satisfactory (Zarifi, 2013).
There was a common belief that multi-word lexical items are peculiar to speech and informal writing. Now, however, they could be observed in all registers from street slang through formal speech to the most academic forms of language (Cornell, 1985). According to Wyss (2003), multi-word lexical items represent a native-like fluency. However, mastering MWIs, particularly, PVs, is not an easy task for ESL/EFL learners to achieve. Celce-Murcia and Larsen-Freeman (1999) point out that it is unlikely for ESL/EFL learners to master PVs especially for those whose L1 is a non-Germanic language as it is rare to find the verb-plus-particle combination outside the Germanic family. This difficulty can be aggravated by both complex syntax of PVs. Some PVs, for example, are separable while others are not, and most of them have more than one meaning. Unfortunately, such difficulty can push the L2 learner to avoid producing such linguistic form.
The pedagogical concern in the current study was whether the selection and presentation of combinations in ESL textbook materials for higher secondary school level in Malaysia are empirically justified. This is deemed imperative as various studies (Mukundan, 2004;Zarifi & Mukundan, 2012;Zarifi, 2013;Kamarudin, 2013;Kamarudin, Rahim, & Arifin, 2017) hold that Malaysian prescribed ESL textbooks were prepared through a process that involved assumption and as such the writers intuited what they considered to be the only relevant and acceptable language input for the target ESL learners. More specifically, it is often observed that coursebooks present PVs in a way that implies that there are no systematic patterns with these combinations or fail to create learnable patterns or even create patterns of the wrong kind (Side, 1990).
The main objective of this study is to examine the use of PV found in Malaysian upper secondary textbooks. In this present study, the generally agreed upon classification of PVs made by Celce-Murcia and Larsen-Freeman (1999) was adopted. Thus, using this classification, this study was aimed to investigate how the PVs are presented and being used in in the English language textbooks; KBSM English Form 4 (F4) by Tan Phaik Lee, Angelina Ng Kim Leng and KBSM English Form 5 (F5) by Rani Parasuraman, Carole Chung Mei Choo, Shadhana Popatlal as prescribed for higher secondary school level in Malaysia. The emphasis was specifically on the textbook for form 4 and form 5 as the students at this stage are preparing for a major government examination that will determine their qualifications to enter tertiary education.

Phrasal Verbs
A phrasal verb (PV) is defined as a combination of a verb proper and a morphologically invariable particle that function as a single unit lexically and syntactically (Darwin & Gray, 1999;Quirk, Greenbaum, Leech, & Svartvik, 1985). Cowie and Mackin (1993) consider the PV as a combination of a verb + particle functions as a single unit of meaning. In the same vein, Celce-Murcia and Larsen-Freeman (1999) argue that PV consists of two parts functioning as a single verb. Therefore, it is sometimes called a two-word verb (Siyanova & Schmitt, 2007). Besides, the term 'phrasal verb' is the most common term used by the pedagogic circles including reference materials such as textbooks, course books, dictionaries (Kamarudin, 2013). Liao and Fukuya (2004) examined the use of PVs by Chinese EFL learners. They, however, divided PVs into only literal and figurative with two proficiency levels, that is, intermediate and advanced learners. Based on their results, Liao and Fukuya (2004) argue that not only the structural differences between L1 and L2 that make learners avoid PVs but also the semantic complexity of such language form makes it difficult for EFL learners to master it.
According to Biber et al. (1999), a PV consists of a verb followed by an adverbial particle. This terminology is followed in this study. Biber et al. (1999) also suggest three criteria with which PVs can be identified: (a) whether there is an idiomatic meaning, (b) whether particle movement is possible, and (c) how the wh-question is formed. Although Cornell (1985) observed that large numbers of PVs are nonidiomatic in the sense that their meaning can be easily deduced from the two constituting parts as in run away or steam off, PVs are usually idiomatic in meaning. Cornell (1985) argues that because many PVs are nonidiomatic, it should not be a problem for ESL/EFL learners to deduce the meaning from their constituents. Because of this idiomaticity, a PV can usually be replaced with a single-word verb. For instance, put out does not mean 'put in an outward direction'. It rather means 'to extinguish'. Therefore, the PV put out can be replaced with the single-word verb extinguish. According to Celce-Murcia and Larsen-Freeman (1999), although some PVs do not allow particle movement (e.g., run into, come across), PVs usually allow particle movement when they are followed by a noun phrase. However, when the object of a PV is a pronoun, the adverbial particle usually comes after the object pronoun. Finally, in sentences with a PV, wh-questions are usually formed with what or who, and the particle cannot be separated from its lexical verb.
Following the same classification with different definitions, Celce-Murcia and Larsen Freeman (1999) also identify three different types of PVs. First, literal, in which the meaning of the PV is a composition of both the verb proper and particle (e.g., stand up, sit down). Second, aspectual, in which the meaning is not transparent but not idiomatic either since the particle has an aspectual meaning (e.g., run on, hurry up). Particles on and up have a sense of continuity. Third, idiomatic, in which the meaning is completely different from the sum meaning of its constituent parts (e.g., run out = to exhaust one's supply).
According to Vethamani and Perumal (2008), today, the English language is viewed as an integral part of the Malaysian students' education. Malaysia's curriculum places great importance of English, almost equal status as the national language. Therefore, the majority of the higher institutions adopt English as the main medium of instruction. This is mainly becuase that teaching materials are in English, lectures are conducted in English, and student assignments are required for submission in English as well (Zamin & Raihana, 2017). As a result, the secondary level English must be able to prepare students for their tertiary level as they need to be more academically proficient in English. The English language curriculum from the primary up to the secondary level in Malaysia has always prioritized the use of the Standard British English model, an infusion of some localised words could still be observed in Malaysian English (Wong, 1991). Some of the examples of Malaysian English taken from Platt, Weber, and Ho (1984) are I don't like heaty food (I don't like hot food), open the light (switch on the light) and close the tap (Turn off the tap). These examples reveal that the use of PVs may be a problematic area among Malaysian ESL learners. Wong (1991) further states that interferences and over-generalisation are two factors that could affect the learning of the second language. Thus, an ESL/EFL learner not thoroughly familiar with PVs would say 'I want to extinguish my cigarette' instead of 'I want to put out my cigarette'. Both verbs are correct and carry the same intended meaning. However, in an informal context put out will be more appropriate. Armstrong (2004) points out that although PVs are largely found in current ELT textbooks, EFL learners find them difficult to understand. PV structures and their correct use are also one of the problematic grammatical items for Malaysian ESL learners (Akbari, 2009;Zarifi & Mukundan, 2014;Kamarudin et al., 2017). Because of this difficulty, EFL/ESL learners tend to avoid English PVs and prefer to overuse single lexical items where a PV would be much more appropriate. In their error analysis of Form 4 English compositions, Rosli and Edwin (1989) found that students in both urban and rural areas had problems using verb forms including the PVs. Therefore, they concluded, English language teachers needed to be careful when teaching this grammatical item to students.

Corpus Analysis
The PV combinations have been recently studied in different general corpora. For instance, Gardner and Davies (2007) studied the PVs in the British National Corpus (BNC). Findings indicate that the combination of a small number of 20 lexical verbs with 8 adverbial particles (160 cases) accounts for more than one-half of the 518,923 PV occurrences in the mega-corpus. Further analysis indicates that only 25 PVs account for nearly one-third of all PV occurrences in the BNC, and 100 PVs account for more than one half of all such items. Besides, findings show that some words like 'out, up, down, and back' is more likely to function as particles rather than as prepositions. Conversely, some other words such as 'under, by, and across' are unlikely to occur as particles but most frequently do as prepositions. Certain forms of verbs, such as 'pick, point, and carry' seem to occur more frequently in PV combinations than to act as lexical items.
On the other hand, there was an abundance of studies that examined PVs in general and learner corpora (Schnider, 2004;Waibel, 2007;Gardner & Davies, 2007;Trebits, 2009;Akbari, 2009;Zarifi, 2013;Kamarudin, 2013). As far as the Malaysian context is concerned, a survey on relevant research on PV constructions shows that some studies focused on learner performance while others chose to examine language materials. Instructional materials did not receive the same importance from researchers in dealing with the use of these puzzling combinations in the instructional materials (Koprowski, 2005;Gardner & Davies, 2007;Zarifi & Mukundan, 2012;Kamarudin, 2013). Therefore, it could be concluded that the pedagogical values of these ELT materials are very limited to ESL learners.
Some studies have been done on ESL learner corpus to compare the frequency of PVs used by ESL learners with those in the real use of language (Akbari, 2009;Kamarudin, 2013;Zarifi & Mukundan, 2014;Zarifi & Mukundan, 2015). For example, Akbari (2009) utilised the corpus of the English language of Malaysian School Students (EMAS) to analyse the writing of students in Forms 1 and 4. Akbari (2009) adopted Celce-Murcia and Larsen Freeman's (1999) classification of PVs (i.e., literal, aspectual, and idiomatic) and used qualitative approach as well as descriptive statistics by using MonoConc Pro 2.2 (Barlow, 2003) to find distribution and types of PVs as well as errors and avoidance of these structures. The study revealed that the use of PVs was more frequent by the students at a higher level. Both groups used less idiomatic PVs. The students at the lower level, Form 1, used more avoidance behaviour than Form 4 students which indicates that the proficiency level influences the avoiding of PVs.
Kamarudin (2013) conducted a survey among teachers and students while using corpus analyses of PVs in the EMAS corpus and used the BoE as a reference corpus. She focused on (1) learners' understanding of PVs, (2) teachers' perceptions of vocabulary teaching, and (3) the treatment of PVs in Malaysian reference materials (i.e., textbooks and two dictionaries). Besides, she examined the patterns of PVs and possible elements to produce them. Using CLAWS tagger, she carried out POS tagging to identify PVs which then were transferred to WordSmith software for further analysis.
The study administered a test of PV for learners and a questionnaire for the teachers. ANOVA and post-hoc test were used to examine the differences of understanding PVs at the three different proficiency levels, low, medium, and high. The study found that students' understanding of PVs is at an average level.
In another corpus study, Zarifi and Mukundan (2014) carried out a corpus-based content analysis of the EMAS corpus. To identify the creativity and unnaturalness of using PVs among school students, the study used WordSmith Tools version 4.0 and then tagged and lemmatised the PVs found to accumulate all their inflectional forms. The study used a dictionary to determine the acceptability of PVs. Also, PVs which did not appear in the dictionary entry were examined against the BNC. Zarifi and Mukundan (2014) found that although learners generally prefer to use PVs, they unfortunately often combine unusual forms of non-literal PVs. The study suggests that students should be provided with appropriate materials and activities that enable them to produce PVs, particularly idiomatic ones, more effectively. Koprowski (2005) examined the selection and presentation of PVs and other MWIs in three ELT coursebooks. He considered the usefulness of any lexical item should be based on corpus frequency and range, so he compared the items against the Bank of English corpus (BoE) to establish frequency data. He found out that (1) MWIs were chosen haphazardly as there were no specific criteria for selection; (2) most items appeared in the coursebooks had low frequency and range values. Consequently, the study claimed that (3) items were subjectively selected and lacked any empirical study since ELT coursebooks writers had chosen the items based on their own experiences and intuition. Zarifi and Mukundan (2012) examined the presentation of PV constructions with research findings in textbooks for Malaysian ESL learners at the secondary level (Forms 1-5). They developed the Zar-Test of Initialisation which can be applied in three stages to identify the various types of PVs. Moreover, they offered Focus Framework and Cognitive Load for the evaluation of the use of PV constructions. The study used WordSmith version 4.0 and the Oxford Dictionary of Phrasal Verbs as the instruments. They found that there was an inconsistency between the PVs selected and presented in the textbooks and the actual use of these PVs in BNC. Moreover, the study found that PVs used in the textbooks had extremely low frequency counts in general English and vice versa. Thus, Zarifi and Mukundan (2012) argued that the process of selection and presentation of PV constructions in Malaysia secondary school textbooks are based on authors' intuition rather than on research findings and pedagogical principles. In a similar study, they explored the grammatical treatment of PV items in Malaysian ESL secondary level textbooks Zarifi and Mukundan (2014). The findings of the study revealed that there appear to be "no guiding principle underlying the selection, presentation and sequencing of different patterns associated with them, bringing further home the observation that the development of ELT textbooks is more intuitively than empirically motivated" (p. 649). Zarifi and Mukundan (2015) also examined the semantic treatment of PV combinations in a pedagogical corpus of Malaysian ESL textbooks (Form 1-5). They found that although the corpus has a large number of PV combinations, most of these constructions were barely presented, that is, they were repeatedly used with the same meaning. The study also points out that some items were presented with their rare and infrequent word meanings. Therefore, the selection and presentation of the word senses of different PV combinations proved to be more intuitively than empirically motivated.
In an extensive study, Kamarudin et al. (2017) examined the descriptions of common PVs in Malaysian high school English Textbooks (Form 1-5). The study examined the definitions, selection, and exercises on PVs found in textbooks. The study reported some interesting findings. Firstly, the school textbooks do not treat this important language form (i.e. PVs) appropriately or adequately. Secondly, only a small section in the textbooks discussing PVs (while PVs can be found in only one page in both Form 1 and Form 5 textbooks, 2 pages in Form 4, and 3 pages in Form 3, Form 2, however, lacks any discussion on this important language form). Thirdly, many of the PVs presented to learners are also not carefully defined and clearly explained with good examples of PVs. Fourthly, important information comcerning PVs that learners need to understand is either lacking or missing (i.e., transitivity and separability of PVs) which supports the claim made by Zarifi and Mukundan (2014). The selection of PV items to be included in the textbooks seems to lack any pedagogic principle and mainly based on writers' intuition (Zarifi & Mukundan, 2012;Kamarudin, 2013;Zarifi 2013). Finally, the five textbooks lack enough practice to reinforce understanding. Thus, the study calls on textbooks writers in Malaysia not to overlook this important language form and be cautious regarding the definitions, selection, and presentation of PVs presented to learners as most of PVs in the textbooks seem to be selected and presented subjectively.
In another recent study, Kamarudin and Zamin (2018) studied the description of PVs in reference materials; the descriptions of PVs in school textbooks Form 1 -Form 5 and the descriptions of PVs in learner dictionaries (i.e., Kamus Dwibahasa Longman and Kamus Dwibahasa Oxford Fajar). The study found that both the school textbooks and learner dictionaries under investigation do not treat PV appropriately and adequately as an important language form.
There is only a small section in the textbooks discussing PVs. Many of the PVs presented to learners are also not carefully defined and clearly explained with good examples of PVs. The selection of PV items to be included in the reference materials seems to be highly subjective, and mainly based on writers' intuition, rather than authentic language data (i.e., corpus-based frequency counts). In other words, PVs are presented to learners without sufficient consideration as to their frequency of occurrence in real life situations since they were selected based on writers' intuition and common sense.
To sum up, the review of literature has shown that PVs have been studied extensively to find out how ESL learners use these complicated items and more recent studies focused on the use of PVs in learner corpora. According to the literature, ESL/EFL learners have difficulties in using PVs; however, advanced learners perform better. The researches showed that the level of English proficiency influences the utilization of PV combinations. Most studies are concerned with showing the weaknesses of learners in using these structures; however, a few studies have focused on how these weaknesses might have resulted from the presentation of PVs in ELT materials. Consequently, it is a pedagogic concern to examine whether EFL textbook writers are considering the research findings and pedagogical principles in presenting the most useful PV combinations.

METHODOLOGY
The current study is a corpus-based content analysis of two secondary level textbooks prescribed for Malaysian students; for form 4 (F4) by Lee and Leng (2003) and form 5 (F5) by Parasuraman et al. (2003). It tries to, through a descriptive quantitative approach, look for the possible patterns with the PVs and their constituents in the target corpus. The quantitative content analysis suites the purpose of this study as Conrad (2005) argues that identification of patterns of language use necessarily involves making a quantitative assessment since it is presupposed that patterns used typically by native speakers are most likely to be more frequent than unusual or uncommon uses of language.
The present study takes into consideration that corpus-based research can provide quantitative and empirical evidence, hence, it can perfectly fulfill the purpose of identifying solid criteria for the selection of PVs. Biber and Conrad (2001) and Granger (2003) agree that, however, there is no single corpus that can extensively and exclusively provide a comprehensive combination of language exemplars, a corpus is deemed a much better starting point than an invented list of exemplars. Moreover, a list of most frequently language forms in their authentic contexts can be obtained from a quantitative corpus-based approach.
Though these methods of analysis address a linguistic phenomenon differently, Neuendorf (2010) holds that their ultimate findings can nicely fit together to lend a good instance of research method triangulation. With the above arguments in mind, the current corpus-based study, thus, used a mixed-method approach to investigate the frequency count and semantic presentation of PVs in a pedagogic corpus of EFL textbooks.

IJALEL 8(6):76-85
The textbooks which are approved by the Malaysian Ministry of Education, are the English textbooks used for the higher level of secondary school. The written and transcribed sections of the textbooks prescribed for secondary level students of F4 and F5 comprise the population of this corpus study. The corpus, consisting of about 150,885 running words, was studied to identify the frequent occurrences of the PV combinations and their presentation. This study adopts a comprehensive data sampling (Ary, Jacob, Razavieh, & Sorensen, 2006) as every PV occurring in the corpus is included in the analysis. To identify the usefulness and frequently used PVs in textbooks under investigation for learners, the current study made use of the most common PVs list provided by Biber et al. (1999) in the Longman Spoken and Written English corpus (LSWE).
Of major methodological issue in this corpus-based study was the extraction of PV forms from among the potential combinations. As English PVs are deemed as highly productive regarding the variety of the verb probers that can combine with the potential particles (Celce-Murcia & Larsen-Freeman, 1999), it is almost unlikely to list all the verb components of these combinations in the textbooks. On the other hand, since there are not so many particles, it is not uncommon to find some lists of English particles. The Collins COBUILD Dictionary of Phrasal Verbs (2012), for example, provides a comprehensive list of 48 possible particles. Also, it is noteworthy that since it is not impossible for a particle to occur as a preposition as well, at this stage no distinction was made between these grammatical forms as such combinations are needed to undergo a certain set of tests to qualify as a PV. In this preliminary stage, PV forms were identified from their orthographically similar structures of prepositional verbs. To this end, the research adopted the following set of tests provided by Darwin and Gary (1999): (1) Whether for the particle to be repeated without its lexical verb: (particles cannot be repeated, but prepositions can).
(2) Whether it is possible for adverbs ending in -ly to be inserted between the combination: elements (PVs do not allow such insertion, but prepositional verbs do). (3) Whether the particle can precede the verb prober: (PVs do not allow forefronting of the particle, but prepositions can be forefronted). (4) Whether the particle is stressed or not:(Particles in PVs are stressed, but prepositions in prepositional verbs are not). (5) Whether the verb prober with its particle form a single intonation unit:(PVs form an intonation unit, but prepositional verbs do not). (6) Whether the particle allows for answering the 'where' question: (The particle in PVs cannot answer 'where' questions, but prepositions can). After this stage, the findings were grouped 'lemmatised' to get the frequency of all the inflectional forms of each lexical verb counted together. Thus, lexical verbs were grouped in the same way as they appear in dictionary entries (McEnery, Xiao, & Tono, 2006). For instance, the forms 'find, finds, finding, and found' were grouped under the lemma "FIND".
The textbooks were thoroughly examined to find the PVs, the frequency of each PV, and the contexts in which they occur. To analyse the identified combinations in terms of their frequency of occurrence, the collected data were presented in a spreadsheet form. This form enabled the researcher to calculate the frequency of each PV and frequency rankings. Then, the obtained frequency distribution was compared to that provided by Biber et al. (1999). Finally, the semantic and structural characteristics of PVs were examined for further analysis.

Definition of VP in Textbooks
The F4 Textbook surprisingly presents PVs as "verb + preposition combinations" and it further pointed out that "the meaning of a phrasal verb changes when different prepositions are used" (F4, p. 143). The F4 Textbook did not provide examples in complete sentences to illustrate the definition provided. However, it provided three "verb + preposition" combinations taken out of text (e.g., look at, look for, look up) (F4, p. 143). In the F5 Textbook, PVs are similarly defined as "verb + preposition combinations" with further explanation that "some verbs combine with different prepositions to give different meanings" (F5, p.162). In addition to that, F5 Textbook provided three examples (cut down, cut into, cut off) in complete sentences (e.g., The authorities were told to cut down on the expenditure for development projects) (F5, p.162). It is apparent that these definitions in both textbooks were not clear and may lead to confusion as no clear distinction is made between the forms of LV+PREP and LV+AVP. There is no doubt that frequent prepositions do not occur as particles, thus their combination (V+PREP) acts as a prepositional verb (e.g., change into, keep to) rather than a PV. Moreover, this definition of "verb + preposition combinations" in both textbooks may result in learners assuming that combinations such as look at, go to, run across, are also examples of PVs, which is not true. Consequently, there is a need for clearer definitions to be presented to inform students that PVs are not just "verb + preposition combinations" but in fact, they are a combination of an LV and particle that carry a specific meaning.

Differentiating Low Frequency PVs from High Frequency PVs
Further analysis indicated that learners were not only presented with an inaccurate definition of PVs, but also the use of low frequency PVs. The example cut into (F5, p. 162) was used to illustrate the combination of LV+AVP but was not very useful to students as this PV is not common in an everyday setting. What was supposed to be provided to the students are highly frequent PVs. Examples of LV+AVP form will not only help facilitate better understanding but also students may find them very useful in communication outside classrooms. Besides, the use of PVs like take off (remove clothing; leave the ground and fly), pick up (take somebody in a vehicle) should alternatively be considered as they are highly produced by native speakers, and, undoubtedly, more useful to learners. Kamarudin (2013) in her analysis, found that however the Malaysian learners produce the PV pick up extensively, they use it inappropriately as they cannot "distinguish the meanings and use of the PV pick up (to lift something up from somewhere) and the LV pick (to gather by plucking)" (p. 185). Examples taken from the EMAS corpus illustrate this: a. Sara and Siti are picking up flowers in the garden. b. Ah Meng decided to pick up the rambutans to eat.
The above examples clearly illustrated that students cannot differentiate the meanings of pick up and pick in which they are more likely to associate pick up with plant objects like flowers and rambutans. Having further examined the sentences, it indicated that the students meant to "illustrate the act of picking flowers/rambutans from the trees/plants and not to pick up flowers/rambutans from the ground/floor" (p. 186).

Frequencies of PVs
Results of the analysis of the F4 Textbook show that 213 PVs are used in the textbook, and all together they occur 469 times. To get the total number of words occurring in the F4 Textbook, the words that appeared in chapter one including words that appeared in its transcription's section were manually counted (5,325 words), and the resulting number was multiplied by the total number of the chapters of the textbook, which is 14. Thus, the total number of words occurring in the F4 Textbook is approximately 74,550 words (5,325 words X 14 chapters). It can be said that PVs occur approximately 6,479 times per million words. This overall frequency of PVs in the textbook is much higher than the frequency of PVs provided by Biber et al. (1999).
One major reason for such high frequency of PVs in the textbook is that some of the PVs are repeatedly used as imperatives in most cases of the textbook. For instance, find out occurs in variations of an imperative sentence, "Find out the origin of the following items using an encyclopaedia or a search engine" (p. 123). A closer examination shows that PV find out can be found in all chapters except chapter 12 and occurs 5 times in chapter 8 of the textbook and a total of 34 times throughout the textbook. Also, the textbook has two more PVs that are exclusively used as imperatives: write out (20 times) "Write out the possible reasons" (p. 203) and fill in (10 times) "Fill in the blanks with the suitable verb" (p. 143). Thus, the combined number of occurrences of these three PVs (64 times) makes up around 14% of the total number of occurrences of all PVs (469 times) in the textbook. Table 1 shows that there is a huge gap in the distribution of PVs throughout the F4 Textbook. The mean score is 34 PVs, but they are distributed disproportionately throughout the textbook. While the occurrence of PVs in some chapters goes over 50 PVs (chapters 6 and 9), it goes down in some others to less than 20 PVs (chapter 14). This finding is further confirmed with another fact that only five chapters (i.e., 1, 3, 6, 9, and 10) account for half of the total number of PVs. This difference in the distribution of PVs may be resulted from the nature of each theme as the chapters of the textbook are framed around different themes (e.g., people, social issues, science and technology, and environment).
However, these two findings question the principles, if any, on what basis the PVs have been selected and presented in the textbook. This further strengthens the claims made by Kamarudin (2013), Zarifi (2013), Zarifi and Mukundan (2012) that the selection and presentation of PVs used in prescribed textbooks for secondary schools in Malaysia lack any pedagogic or guiding principle.
The analysis of the F5 Textbook shows there is a rather consistency between both levels as the F5 Textbook includes 216 PVs (only 3 PVs more than F4 Textbook), and as illustrated by Table 2, they occur 489 times in total. Again, to know the total number of words occurring in the F5 Textbook, the researcher counted the word that appeared in chapter one including words that appeared in the transcription's section (5,098 words), and the resulting number was multiplied by the total number of the chapters of the textbook, which is 15. Thus, the total number of words occurring in the F5 Textbook is approximately 76,335 words (5,098 words X 15 chapters). Consequently, it can be said that PVs occur approximately 6,405 times per million words which is quite similar to F4 Textbook's count. As a result, the overall frequency of PVs in the F5 Textbook is much higher than the frequency rate of PVs (2,000 times per million words) provided by Biber et al. (1999).
However, a closer examination for PVs used in the F5 Textbook shows that such high frequency of PVs in the textbook is, again, resulted from the fact that some PVs are repeatedly used as imperatives. The PV find out, for example, regularly appears in variations of an imperative sentence, "Find out the meanings of the phrases in the boxes" (p. 199). This PV can be found in most chapters and occurs 6 times in chapter 12 of the textbook and 25 times throughout the  textbook. Besides, the textbook has two more PVs that are exclusively used as imperatives: fill in (43 times) "Fill in the blanks with the above phrases" (p. 145) and log onto (11 times) "Log onto these sites for more information" (p. 31). Thus, the combined number of occurrences of these three PVs (79 times) makes over 16% of the total number of occurrences of all PVs (489 times) in the textbook. There is also an inconsistency in the distribution of PVs throughout the F5 Textbook. However, this difference is less evident than that of the F4 Textbook. Also, the mean score of PVs in both F4 and F5 textbooks is almost the same (34 and 33, respectively). However, a close examination of the distribution of PVs in the F5 Textbook shows that the frequency count of PVs in the second part of the book is higher than that of the first part with nearly three-fifths of the total occurrence of the PVs throughout the textbook occur in the second half. Moreover, the number of PVs in chapter 15 (55 PVs) is higher than that of both chapters one and two (24 PVs each).

Pattern of Distribution of PVs
Another finding shows that the 5 chapters with the highest count of PVs (15, 9, 11, 10, and 3) account for only one-fifth of the total number of occurrence of PVs which indicates that the writers of F5 Textbook, unlike the F4 Textbook's writers, to some extent succeeded in distributing the number of PVs proportionately throughout the F5 Textbook. However, the distribution of PVs in some chapters needs to be reconsidered as it has been mentioned earlier that the number of PVs in some chapters (chapter 15) is more than twice that of some others (chapters 1 and 2). The PV fill in was the most mentioned with 45 occurrences while carry out and find out were mentioned 18 and 25 times respectively. A total of 135 PVs were only mentioned once. Some of the examples were act out, block out, blow away, blow up, break off, break out, bring into, build up, burst into, chop down, clean up, clear up, close down, close up, collide into, come forward, come in, come into, come out. As it is shown, only 33% (70 items) of all PVs in the F4 Textbook and 37% (81 items) of all PVs in the F5 Textbook have been repeated. In other words, most of the PVs (67% = 143 items and 63% = 135 items) in both F4 and F5 textbooks respectively are treated as hapax legomenon, occurring once in the corpus. Thornbury (2004) argues that words need to be repeated for a minimum of seven times over an array of intervals to have a good chance of being remembered.
Examining both textbooks, we can find two fractions of 3% (14 items) and 2% (11 items) in F4 and F5 textbooks respectively that have met this threshold. Consequently, it could be said that students might not benefit much from the big number of PVs in both textbooks as most PVs have not been adequately presented which consolidates the findings made by Zarifi andMukundan, (2012), Kamarudin, (2013), Zarifi (2013).
Similarly, Celce-Murcia and Larsen-Freeman (1999) argue that repeating the different forms of language over a certain period is pedagogical although some researchers question the pedagogical usefulness of corpora based on frequency lists (Widdowson, 1990;Howarth, 1998). Römer (2005) considers that the process of selecting significant language forms or syntactic structures in order to be presented to students should be guided by the frequency counts. He contends that selecting any language form to be provided for EFL students without any empirical basis is both difficult and doubtful.
Similarly, Kennedy (2002) argues that it is useful to provide early learners with language items of high frequency. Therefore, to help learners to master PVs, they need to be provided with the highly frequent combinations and revisit them repeatedly over an interval period. In the same vein, Biber et al. (1999) provide a list of PVs that occur over 40 times per million words in either the conversation register or the fiction register. The list is provided in Table 3 below which includes 31 PVs divided into six different semantic domains namely; activity, mental, communication, occurrence, copular, and aspectual.

Use of Common PVs
A close examination of the textbooks under this study shows that both textbooks failed to include all 31 common PVs. Moreover, around 20% (6 PVs) of the most common PVs are not included in either textbook under investigation. Table 4 above shows that the F5 Textbook has a slightly larger number of high frequency PVs compared to the F4 Textbook. From the list of 31 high-frequent PVs, the F5 Textbook listed 24 of them, and 18 were found in F4 Textbook. This indicated that some of the most frequent PVs that are widely used by native speakers in everyday settings and thus very useful for learners are not listed in both textbooks under investigation. It also illustrated that the PVs come off and turn out are of the most common PVs in the list and can be found in three sub-corpora in academic and non-academic discourse and written and spoken registers. In addition, the Collins COBUILD Phrasal Verbs Dictionary considers the two PVs among the key PVs in the dictionary and provides 8 different senses of come off and 9 different senses of turn out which suggests that these two PVs have a wide range of meanings and usage in everyday settings, and, therefore, are very useful for ESL/EFL learners. However, come off and turn out are not included in either textbook under study. Similarly, the PVs come along, come off, shut up, sit up, and turn out are also not included in the textbooks, which supports the claim made by Kamarudin (2013) that textbook writers for Malaysian secondary schools failed to provide many of the most frequent and important PVs in learners' textbooks. Both the selection and presentation of the PV combinations in both the F4 and F5 textbooks turned out to be inconsistent with the native speaker's use of these forms.
When comparing these forms in the textbooks against the list provided by Biber et al. (1999), it suggests that more purposeful PV items of immediate use (e.g., put on, go off) need to be included. It also shows that the used forms need to be more effectively recycled and presented for the learners to master them with much less effort and more efficiency. Thus, the findings revealed by the analysis of PVs used in both textbooks again consolidate the conclusions made by Zarifi and Mukundan (2012), Kamarudin and Moon (2018), and Zarifi (2013) that PVs used in the prescribed textbooks for secondary school in Malaysia lack any empirical study.
In general, the analysis of PVs in the F4 and F5 textbooks indicated that many PVs were presented to learners without sufficient consideration as to their frequency of occurrence in real life situations. Nation and Waring (1997) argue that frequency information is highly significant as "learners get the best return for their vocabulary learning effort" (p. 17). They contend that it is most likely that learners will encounter items later outside the classroom setting. Thus, it seems pedagogically important to appropriately select and present core phrasal verbs suggested by Cornell (1985) to ensure learners are provided with the most useful PVs they may encounter in everyday settings.
On the other hand, as a result of the unsystematic selection of PVs in the F4 Textbook, the students were provided with low frequency PVs like gobble up and scoop out. However, most common PVs, which in the same time form a difficulty for students (come over, get back, get in, put on) are not adequately addressed in the F4 Textbook. From a pedagogical point of view, it seems necessary for textbook writers to consider the frequency information when selecting and presenting PVs to students. Moreover, corpus studies usually point out that EFL learners should be presented with the most frequent PVs instead of the less common ones (Celce- Murcia & Larsen-Freeman, 1999;Gardner & Davies, 2007). In addition, emphasis should be given to core meanings of PVs as they are more useful to language learners which is missing in the present school textbooks under investigation (Zarifi & Mukundan, 2012;Kamarudin, 2013;Zarifi, 2013).
It has been mentioned that the difficulty L2 learners face with PVs is divided into productivity -separability -polysemy. Firstly, regarding productivity, the analysis of both textbooks shows that the textbooks are missing any information regarding the productivity and flexibility of PVs, and how new PVs could be formed just by combining a lexical verb and a particle. Secondly, as far as the separability issue is concerned, the two textbooks do not explicitly highlight this feature of PVs. All PVs found in both textbooks are found only in the form of two-word combination (LV+AVP) (Please turn off) and not (LV+X+AVP) (Please turn it off) or (LV+X+X+AVP) (Please turn the computer off). Finally, concerning polysemy, most PVs have many different senses (Celce-Murcia & Larsen-Freeman, 1999;Gardner & Davies, 2007); however, the F4 and F5 textbooks, in most cases, use the PVs with only one sense as it is the case with the PV pick up (to collect) as in "Jamil has to pick up the bunga telur and banquet from the Mak Andam's house" (F4, p. 13).
The above analysis has revealed that, in general, the selection of PV items to be included in the Malaysian English Total 18 24 language textbooks for higher secondary school level (F4 and F5) is done subjectively and on a basis of intuition rather than empirical evidence, such as frequency counts based on authentic language data (i.e. corpus-based). Although many PVs are appearing in learners' textbooks, they, however, lack significant information that could inhibit the learners' ability to understand them. Consequently, Malaysian textbook writers should be more careful with the selection and presentation of PVs, and, at the same time, ensure that this language structure is presented to learners "in a manner that avoids unnecessary confusion and loss of time for both student and instructor" (Darwin & Gray, 1999, p. 66).

CONCLUSION
To address the first objective of the study, the above analysis has revealed that both the school textbooks for form 4 and form 5 under investigation do not treat PV appropriately and adequately as an important language form. There is only a small section in each textbook discussing PVs. Many of the PVs presented to learners are also not carefully defined and clearly explained with good examples of PVs. The selection of PV items to be included in the reference materials seems to be highly subjective and mainly based on writers' intuition and common sense, rather than authentic language data (i.e. corpus-based frequency counts). In terms of the presentation of PVs, it was found that PVs are illustrated to learners without sufficient consideration as to their frequency of occurrence in real life situations. Learners need to be presented with high frequency PVs as there is a high possibility that they will encounter such PVs more frequently in the future. This would help them to understand PVs better, and eventually, be able to use them appropriately in their written or spoken discourse. This suggests that a systematic selection of a 'core of phrasal verbs' is necessary to ensure learners are presented with PVs that are most useful for them in the world outside the classroom. It is hoped that the findings of the current study will be useful to all stakeholders involved in the teaching and learning of English in Malaysia. The findings will help to improve awareness among textbook writers, curriculum designers, teachers, and students in general on the correct use of PV in the process of teaching and learning. Accordingly, appropriate measures can be taken to improve the present scenario of PVs treatment in textbooks and teaching and learning in Malaysian schools.