Explaining the Diversity in Malay-English Code-Switching Patterns: The Contribution of Typological Similarity and Bilingual Optimization Strategies

Round 1

Reviewer 1 Report

See attached .pdf file

Comments for author File: Comments.pdf

Author Response

1. This reviewer feels obliged to raise a question about the length and discourse structure of the MS: it takes until p. 24 (Sec. 3.3, “The current study”) in an MS totalling 43 pages before the author’s new data is presented and analysed, with the five research questions listed on p. 25. The results, discussion and conclusion sections are relatively brief, by contrast.

Suggestion: since the main focus, as stated in the title, is on Muysken’s Congruent Lexicalization (CLX) category, perhaps the (albeit fascinating) sections on Muysken’s three other categories applied to Malay-English code-switching could be shortened.

Answer: Thank you very much for this. We agree that the study was a little imbalanced in that we mention congruent lexicalization prominently in the title but discuss many different types of code-switching in the body of the article. We think there is a need for a broad overview of the variability of code-switching patterns, as the evidence is currently scattered over a wide range of different publications, that are difficult to access for international audiences. We have therefore changed the title to a more generic one and cut out parts of the literature review which focus exclusively on congruent lexicalization. Section 2.3 (What does the evidence for congruent lexicalization consist of?) has been deleted. The abstract now reflects the fact that the paper will summarize the main literature on Malay-English code-switching. This is also relevant for future researchers wishing to develop online tools to study code-switching, as the paper can be an important source for authentic examples that can be used in experimental designs.

2. If there is time and inclination for quite a major redrafting, section 2.2 could be rewritten using examples from the data corpus of this study, and placed in the findings section. This would help redress the imbalance noted in item 1. above.

Answer: we appreciate the reviewer’s point, but we do not think it is common to add analyses of one’s own data to the literature review. The latter normally covers evidence available through publications. Once the overall picture has been sketched, we can investigate whether the types of code-switching found in the literature are also reflected in our own data. It would be particularly difficult and highly unusual to put the quantitative analyses in the literature review.

3. Notwithstanding the above points, Muysken’s 4 CS types are well explained and exemplified.

Answer: thank you very much for your kind comments.

4. At the end of section 3 (Methods), pp. 1-2 of the 2nd part, there should a summary or signalling sentence. It is odd to finish this major section with Tables 2 and 3.

Answer: Thank you. We have added a transition sentence after the tables. The tables have been renumbered because the Table from Muysken (2014) has been deleted.

5. In the Results section (4.), the research questions 1 and 2, 4 and 5 are addressed appropriately with analysis of data examples. But there is no specific reference to research question 3. The issue of similarities and differences between the types of code-switching found in Majid’s (2019) corpus is not discussed further.

Answer: thank you for flagging this up. We have now added a summary of the similarities and differences at the end of the section 4.3

Micro-level issues noted:

1st part

Line 125, p. 3: “lexica” => “lexis” - changed

Line 141 “lexical as well as functional”: reverse the order of adjectives => “functional as well as lexical” - changed

Spelling of cited author’s name (McLellan√ / MacLellan x); also inconsistent in inclusion of 1st name in reference list (J. vs James). Also elsewhere in the reference list…. see notes below  - changed

Lines 988-1006, p. 23, referring to Cantonese particles (wei, lor): this is tangential to the argument here and could be dropped

Answer: We think it is important to point out that there are switches of discourse markers in the data and that these are best analysed as backflagging (not insertion). We would therefore like to keep these examples.

Layout and lack of line spacing (this may not be the fault of the author; sometimes these features can be lost in transition): e.g. p. 15, sec. 2.4.12, also 2.4.13, 2.4.15. Also in the 2nd part, section 4.4 seems wrongly aligned.

Answer: thank you – This probably happened when the manuscript was transferred to the template. We have revised the alignment and transferred the text to a new template.

Line 1105, p. 25: missing word: “Musyken’s /\ framework”, or similar

Answer: thank you – now corrected

Reviewer 2 Report

General Comments

Overall, more data is required to support the claims made with regard to the findings.  Seeing that the number of participants is limited to two and the interaction is mainly one-way, assuming that such interaction took place in a lecture hall or room during classes (as this was unclear in section 3.22 on p. 25), it seems to me the data is inadequate to make a definitive claim that CLX is a frequent code-switching pattern in Malay-English code-switching types, especially since only the data of one of them showed this fact.  The second participant’s data has CLX as second most common among the various code-switching types. 

Further the data set is rather small for a corpus-based study.  In addition, since the setting of the study is that of an English language class, this would lend itself to particular syntactic categories.  Thus, a larger data set covering more varied contexts is required to make these claims stick.  

The English proficiency of the participants (which was not measured) might also play a role in the instances of code-switching in the data set.

Comments for author File: Comments.pdf

Author Response

General Comments

Overall, more data is required to support the claims made with regard to the findings. Seeing that the number of participants is limited to two and the interaction is mainly one-way, assuming that such interaction took place in a lecture hall or room during classes (as this was unclear in section 3.22 on p. 25), it seems to me the data is inadequate to make a definitive claim that CLX is a frequent code-switching pattern in Malay-English code-switching types, especially since only the data of one of them showed this fact. The second participant’s data has CLX as second most common among the various code-switching types.

Further the data set is rather small for a corpus-based study. In addition, since the setting of the study is that of an English language class, this would lend itself to particular syntactic categories. Thus, a larger data set covering more varied contexts is required to make these claims stick.

The English proficiency of the participants (which was not measured) might also play a role in the instances of code-switching in the data set.

Answer: we agree that the data set cannot be representative for the types of code-switching that are found across Malaysia and have added a sentence in the conclusion section to explain this. However, we have looked into the differences between the types of switching found here and the switches which are attested in the literature. The few differences we found are discussed in section 4.3. It seems the types of code-switching we find are very similar to those found in the literature, except for those found in Facebook or other social media, where word-internal switching is found too. It is likely that there will be interpersonal variability in the frequency with which the different types of code-switching are found,  and this is reflected in the two participants chosen for this study. We have mentioned in the conclusion that a limitation of the current study is that participants’ proficiency in both languages  was not measured, but this was not the case in any of the other studies of Malay-English code-switching either. A very different set up would be needed to investigate the relationship between language proficiency and code-switching patterns.

Review: Specific Comments

Analysis of data in other studies

The translation in (22) could be more consistent and detailed on the whole.

Example (22): Tu form three seorang (line 612 on p. 14)

DET form three a person

A form three boy

Answer: thank you – corrected

Under closed class items / demonstratives, provide an example for the type mentioned in lines 855-857 (p. 20).

Answer: example (3) is an example of code-switching of lah. I have added a reference to this example.

Provide an example for the condition stated in lines 936-938 (p. 23).

Answer: references to two examples with usages of kalau have been added.

Actually, Malay grammar (books) recognize that ‘kena’ is used to form passives in Malay. As such, this is not due to Perciller (2016) alone (lines 886, p. 20). Thus, the sentence in line 886 should include the phrase ‘in code switching’ after ‘a passive’.

Answer: We have added a reference to a Malay grammar book Karim, Onn, Musa & Mahmood, 2008). The sentence has been rephrased.

Lines 1044-1046: Wouldn’t normal informal face-to-face interactions be informal too?

Answer: Of course face-to-face interactions can also be informal. We have therefore taken out the direct comparison with face-to-face conversations. It is possible that the remoteness of facebook exchanges is the real reason why code-switching is different in this medium. This is explained in the following sentences.

Methodology

State the location of the data collection exercise.

Answer: Data collection took place in Negeri Sembilan. This has been added.

If the participants’ proficiency levels were not measured, could this have affected the extent or level of the code-switching and their use the types of code-switching and also the bilingual optimization strategies on the part of the participants?

Answer: this is possible but as explained above this issue cannot currently be investigated. A limitation has been added to the conclusion section.

Analysis of data for present study

Line 54 (p. 30): The sentence could be better rephrased. ‘… appears to flag that an explanation is coming.’

Answer – corrected

Lines 77-79 (p.31): What could be a possible explanation for the contrast in the direction of code-switching?

Answer: this has now been rephrased. The reason is given in the next sentence: the technical words that need to be used in the classroom are mainly from English.

Could it be that the speakers’ L1 is more dominant than the L2 (English).

Answer: that is entirely possible but cannot be asserted on the basis of the current evidence.

Lines 80-81 (p. 31): The distribution of 39 Malay adverbs in English versus 21 English adverbs in Malay does not seem balanced as claimed by the author(s).

Answer: we have replaced “more balanced” with “less asymmetrical”. As explained in Muysken (2000), insertional code-switching is often asymmetrical in that one language provides more lexical items than the other. The same is true in Deuchar (2020) who shows that for 96% of bilingual clauses, Welsh is the matrix language, and English items are inserted into this clause. By comparison with the Welsh-English data set, Malay-English code-switching is much less asymmetrical.

(82) is unclear (line 100, p. 31). The author could clarify this with some context.

Answer: the translation has been changed to “As short as this is ok.”.

The translation for (87) (lines 130-132, p. 32) is unclear.

The bare Malay verb ‘gaduh’ is ‘fight’.

The translation seems to be ‘He’s like if you have a fight with your girlfriend’.

Answer: thank – we have adjusted the translation.

The general translation for ‘pun’ in (90) (line 156, p. 33) is ‘even’ rather than ‘full’. As such, the translation is more like:

‘Even though/ Although I had already explained it, maybe the reader does not understand, so why don’t I give (an) example’ or 2

Answer: thank you – now changed.

In the analysis of (90) on p. 33, the focus is on the aspectual marker ‘dah’.

It is noted that another interesting code-switched word is ‘pun’. If we consider this to be equivalent to ‘even’, then it is a subordinating conjunction which is repeated in Malay after the subordinate clause ‘even I dah explain’ and this seems to an instance of BFL.

Perhaps this instance of code-switching could be discussed after (100) (line 226, p. 35) under section

4.4.7 Conjunctions and markers.

Answer: thank you – we have added an example containing pun in the section on conjunctions and discourse markers.

The translation for (92) (lines 178-179, p. 33) is unclear.

It should be ‘So that when you look at the answer key, you can guess how to actually get that particular answer’.

Answer: thank you – now corrected

The analysis for (105) (see lines 260-261, p. 36) is erroneous. ‘Yang’ marks the subject, not the direct object, of the relative clause.

Answer:  thank you. We have now referred to another example where yang is used to mark the object (in the new numbering that is 109) and changed the description of the other examples involving yang.

Do switches of Malay function words in English utterance qualify as CLX?

What is a possible reason for switches of Malay demonstratives, modals and relative pronoun ‘yang’ occurring in only one direction? (see lines 288-290, p. 37).

What is a possible reason for switches of English pronouns in Malay being more frequent than switches in the other direction? (see lines 290-292, p. 37).

Answer: this issue is being addressed in section 4.5. This section has been renamed “explaining the diversity in patterns”, to clarify the explanations will be given in this section.

Bilingual optimization strategies

The explanation in lines 308 (p. 37) should be for ‘tak sejuk’ and not ‘takut sejuk’.

Answer: thank you – now corrected.

Other comments

Line 201, p. 5: Table 1 The (‘t’ should be in caps).

Answer: this table has been deleted, as copyright clearance could not be obtained.

(6) (lines 230-231) is the other way around from the description in the text: The Sranan word is in regular font and the Dutch expression is in italics.

Answer: thank you – now corrected.

Is ‘criterium’ (lines 157 & 289, pp. 4 & 7) the intended word here? Or is the intended word ‘criterion’?

Answer: criterion is the intended expression. Now corrected.

Lines 718-719 (p. 16) – remove the word ‘switched’ on line 719 (redundant).

Answer: thank you – now removed.

Line 998 (p. 23) …in these bloggers’ data set … (replace ‘this’ with ‘these’).

Answer: thank you – now changed

Lines 1029-1030 (p. 24): The sentence could be better rephrased.

Line 1032 (p. 24): The sentence could be better rephrased.

Line 1041 (p. 24): The sentence could be better rephased after ‘(66) and (67) …’.

Answer: these have all been rephrased

References

The reference in line 347 (p. 38) is (Wong & Chan, 2005), not (Wong & Chan, 2011). See the reference on p. 43. And following the other Chinese names listed (Wong, 2012 & Wong, 1981), I believe the reference should be:

Wong, B. E. & Chan, S. H. (2005). English relative clauses: what Malay learners know and Use. Pertanika Journal of Social Sciences and Humanities, 13(1), 107-115.

Thank you - now corrected.

Reviewer 3 Report

Interesting, detailed study of CS/CLX, with many examples; good research background, description of the situation, and analysis of data; conclusions reasonable. –– Few minor remarks: (a) L997ff: don't know any study for that, but 'hor' might rather be Hokkien in origin, cf. Taiwanese -hoh; not an issue. (b) L1063: "Soon", not "Tek Wooi Sun". (c) L1072; L350, L361 etc.: There are (many) wrong hyphenations (use manual hyphens "CTRL + –" to enforce a position). (d) L0159: given > give. L333: askena > as kena. (e) References: sorting order: Smith after Kellermann; Vollmann, Ralf & SOON Tek Wooi (with f, and Soon = surname).

Author Response

Interesting, detailed study of CS/CLX, with many examples; good research background, description of the situation, and analysis of data; conclusions reasonable. ––

Thank you very much for your kind comments.

Few minor remarks:

L997ff: don't know any study for that, but 'hor' might rather be Hokkien in origin, cf. Taiwanese -hoh; not an issue.

Answer: Tay et al. (2016) refer to Kwan-Terry (1991) who argues that hor is very likely an element transferred from Cantonese. We do not know of other sources claiming different origins for hor.

 L1063: "Soon", not "Tek Wooi Sun".

Answer: thank you – now corrected.

(c) L1072; L350, L361 etc.: There are (many) wrong hyphenations (use manual hyphens "CTRL + –" to enforce a position).

Answer: thank you – we have now set the document to no hyphenation.

(d) L0159: given > give. L333: askena > as kena.

Answer: now corrected

(e) References: sorting order: Smith after Kellermann; Vollmann, Ralf & SOON Tek Wooi (with f, and Soon = surname).

Answer: thank you -now corrected.

Reviewer 4 Report

The article is interesting and provides additional insights how code switching occurs in English and Bahasa Malaysia. Many CS studies have been conducted  but only few papers in Malaysian context focus on congruent lexicalization. However, this article needs to shorten its introduction. Keeping it concise will help readers understand better the overall direction of the paper and the arguments presented. Also, there is a need to shorten the methodology section and remove the unnecessary details in the "instrument and data analysis section." Tables 2 and 3 in the methodology can be placed in the appendices.

Author/s should provide justification why only 2 participants selected and why male and female? Why there were no inter-raters? How will you ensure the validity and reliability of the data collection and analysis? How did the author/s create a corpus? In the analysis it has been mentioned that the data were taken from a corpus.

In the results section, it shows a table on syntactic categories of switches but there was no syntactic analysis at all. Most of data presented were just described by the author/s. Presenting a syntactic analysis will help clarify the occurrence of CS.

 

Author Response

The article is interesting and provides additional insights how code switching occurs in English and Bahasa Malaysia. Many CS studies have been conducted  but only few papers in Malaysian context focus on congruent lexicalization. However, this article needs to shorten its introduction. Keeping it concise will help readers understand better the overall direction of the paper and the arguments presented.

Answer: thank you – we have now revised the introduction to make it clearer what the focus is. The focus is now less on congruent lexicalization.

 

Also, there is a need to shorten the methodology section and remove the unnecessary details in the "instrument and data analysis section." Tables 2 and 3 in the methodology can be placed in the appendices.

Answer: we have revised the tables, and deleted one. Table 1 now contains details of the ways in which we used the criteria for distinguishing between types of code-switching. As one reviewer asked for clarification of the ways in which we used the criteria, we would like to keep this table in the text.

Author/s should provide justification why only 2 participants selected and why male and female? Why there were no inter-raters? How will you ensure the validity and reliability of the data collection and analysis? How did the author/s create a corpus? In the analysis it has been mentioned that the data were taken from a corpus.

Answer: See also our answer to reviewer 1. As explained on p. 46, a second author coded sixteen examples from the data. Any discrepancies were resolved in discussions among the team members.

In the results section, it shows a table on syntactic categories of switches but there was no syntactic analysis at all. Most of data presented were just described by the author/s. Presenting a syntactic analysis will help clarify the occurrence of CS.

Answer: syntactic analyses are presented in the following section (now 4.3).