Language Contact: Sociolinguistic Context and Linguistic Outcomes

Abstract

In this chapter, we will first outline the sociolinguistic context of language contact between Hong Kong Cantonese, Standard Written Chinese (SWC) and English. Then we will exemplify the typical language contact phenomena in terms of salient patterns of linguistic outcomes, namely:

• Lexical borrowing or transference from SWC and English into Hong Kong Cantonese;

• Translanguaging in speech: more commonly at the intra-sentential level (traditionally termed ‘code-mixing’) than the inter-sentential level (code-switching); and

• Translanguaging in writing: written Chinese may range from formal Hong Kong Written Chinese (HKWC), where SWC has been infused with classical Chinese (wenyan) and Cantonese elements, to informal colloquial written Cantonese, which is modeled on the norms of speech, including the free insertion of English words.

The various linguistic outcomes of language contact exemplified in this chapter, first in speech then in writing, seek to demonstrate that all the languages and language varieties within a plurilingual’s linguistic repertoire are treated as a composite pool of semiotic resources to make meaning. Our focus is on the transference of English open-class words into Cantonese, including the use of terms of address mixed with the salutary expressions ‘Sir’ and ‘Madam’ among members of the Hong Kong disciplinary forces, and ‘Sir’ and ‘Miss’ when referring to teachers in the education domain. By virtue of languages being closely bound up with specific sociocultural attributes, language choice invariably has the semiotic potential to symbolize or index speaker/writer identity. This is the basis for social motivation of translanguaging. In the absence of evidence of language choice being consciously motivated by a wish to signal and/or negotiate speaker/writer identity, translanguaging to English typically results from one or more of the following linguistic motivations: (a) to fill a lexical gap, (b) to avoid semantic incongruence, and (c) to use field-specific terms acquired through English-medium instruction, hence the medium-of-learning effect (MOLE).

2.1 Impact of English on Hong Kong Cantonese

For over 100 years until the 1970s, Hong Kong was looked upon as a haven for successive generations of economic migrants from the war-torn and socially insecure parts of mainland China. Political instability and socioeconomic despondency drove many mainlanders to leave their homelands in search of job opportunities and brighter prospects in this British colony (So 1998; Tsou 1997; Zhang 2009). Many of them originated from the province of Guangdong, where Cantonese is the dominant regional lingua franca. After the Second World War, many of the new arrivals considered this “borrowed place [in] borrowed time” (Hughes 1976) a stepping stone in transition toward some other dream destinations. Some managed to leave, while many more had no choice but to call Hong Kong home. Life gradually stabilized after their families settled down and their children grew up to become linguistically Cantonese-dominant through socialization and education, including those born to non-Cantonese-speaking parents (e.g., Hakka [Kejia] and Chaozhou [Teochew]), especially after the universalization of vernacular primary education since 1971 (So 1998, pp. 157–159). This is why, roughly since the 1990s, younger generations of Chinese Hongkongers increasingly report Cantonese as their usual language, as shown in (by-)census figures since 1996, see Table 1.1, Chap. 1).¹

As of the mid-2010’s, practically all home-grown Chinese Hongkongers below 50 are bilingual in Cantonese and English to different extents. This language profile may be explained by recurrent patterns of language choice in society since 1970s, especially in the home and school, and in the domains of friendship and media. Under the nine-year compulsory education policy since 1978 (extended to 12-year in 2012), all parents are obliged to send their children to primary school. Although English is formally taught from Primary 1 (age 6), most P1 students have already had up to three years of basic English literacy training in kindergarten. Through formal education, school-age children of Chinese descent who are already more or less Cantonese-dominant gradually develop plurilinguality to include English and SWC, which, with few exceptions, is taught in Cantonese in Hong Kong (and Macao) Special Administrative Region (SAR). Since English is seldom used among Chinese Hongkongers for intra-ethnic communication (see Chap. 6), education is an important and arguably indispensable means for fostering students’ knowledge and grasp of English (spoken and written) and Chinese literacy.

In terms of language learning outcomes, the amount of home support is an important variable and predictor. In general, in those households where parents can afford setting aside resources to provide extra support, such as engaging a private tutor (native English-speaking tutor often preferred where possible), hiring an English-speaking domestic helper, and increasing exposure to English through games and other literacy-focused activities, their children tend to stand a better chance of attaining a higher level of English proficiency. Indeed, many thrifty parents are reportedly keen on setting aside precious money to buy their children such extra support, in the hope that they would not ‘lose at the starting line/point’.² According to an Oxford University Press survey conducted by Richard Wong concerning the attitudes of middle-class parents with one or two children up to age six toward home support for their children’s English language development, over half of the 950 respondents wished that English be their children’s first language (Ngai 2015; Lui 2015). While all of the respondents hired a Filipino domestic helper to alleviate their household chores, most expected their helper to assist their children with English, even though in some cases, the helper’s English accent and accuracy was a matter of concern (e.g., saying *no have for ‘do not have’, *eat rice for ‘have lunch/dinner’ and *open gun for ‘shoot’). Quite a few parent respondents indicated they did not mind spending up to HK$3000 (ca. US$400) per month, just to give their children extra exposure to English in one way or another (Ngai 2015; Lui 2015). Parents from more affluent families clearly have more options, including sending their children to elite boarding schools in UK or USA. The lead of a news story entitled ‘Price is worth paying for an elite schooling’ is very instructive in this regard: “Chinese parents are willing to pay hefty fees for British boarding schools, and some are sending children away as early as the age of seven” (Zhao 2015). John Ing, head of London-based Quintessentially Education, which had an office in New York and which opened an office in Hong Kong in May 2015 to “cash in on the demand”, indicated that “Hong Kong and mainland China contribute more students than any other single group” (Zhao 2015). Ing also noted that some Korean and mainland parents would not mind sending their children to UK or USA at an early age (e.g., age 3), which he advised against. There is also some evidence that young people’s life chances, as defined by their ability to secure a place in a local English-medium university, correlate strongly with their family income. Such a socioeconomically sensitive ‘English divide’ appears to be attested by the findings of one 2011 study conducted by K.-L Chou concerning the university enrolment of young people (Fung 2013). Chou’s findings showed that, in 2011, the university enrollment rate of 19- and 20-year-olds from the richest 10% of households was about 3.7 times higher than those from households with incomes of less than half of the median level, whereas two decades earlier, in 1991, the difference was only 1.2 times (Fung 2013). In terms of the actual numbers of admittees, 48.2% from the wealthiest families were at university compared with 13% from families living in poverty (SCMP editorial 2013).

The typical language profile of Chinese Hongkongers outlined above is crucial for understanding various language contact phenomena between Cantonese, English and SWC. In the rest of this chapter, we will examine the impact of English on Cantonese and Hong Kong Written Chinese (HKWC, Shi 2006; Shi et al. 2014), as shown in Cantonese-dominant Hongkongers’ informal social interaction with one another in speech, and an excerpt of a sample newspaper column featuring the informal use of written Chinese.

2.2 Plurilingual Interaction: Mobilizing All Linguistic Resources to Make Meaning

Where no linguistic norms prevail to restrain their language choice, plurilingual speakers will naturally mobilize all their linguistic resources to make meaning and, in the process, they are guided, often subconsciously, by an awareness of the social role and linguistic repertoire of the interlocutor(s) they are interacting with. For illustration, let us begin with one instructive example from my field notes, a ‘slice of life’ I observed over 10 years ago.

On my way home one day, entering the lift of the building where I lived, I overhead a short English conversation between a 7- or 8-year-old boy and (presumably) a Filipino domestic helper, who was carrying a school bag on her shoulder that in all likelihood belonged to the boy. I did not know them other than finding their faces familiar and that they lived on a higher floor. The boy was visibly excited about something that had happened to him during the day. From the segment of his mini narrative I heard during the 30-second lift ride, there was one sentence that I retained with interest and jotted down in my field notes after I got home:

(1)	…I tou saliva on the spider…
	‘…I spit on the spider…’

I did not have the larger context to tell the circumstances under which the young boy would utter that sentence in (1). What was clear to me was that the boy came from a middle-class family, a socioeconomic status that was typical of families in my neighborhood, of which one visible indicator was one or more English-speaking domestic helpers they hired. That young boy’s English was fluent and he could make meaning clearly with the domestic helper, whose English accent was reminiscent of people from the Philippines. Their conversation in the lift was entirely intelligible to me, although I missed the rest of the details in that young boy’s story.

Several points are of interest in (1) from the linguistic point of view. First, for bilingual interlocutors who understand Cantonese, the insertion of [tou], pronounced like in Cantonese high level tone tou ⁵⁵, is comprehensible in context as the lexical equivalent of Cantonese 吐 (tou ³³, ‘spit’). Second, whereas (1) is syntactically well-formed, it does not sound so idiomatic, in that Standard English would require the use of the verb spit instead of tou saliva (‘spit saliva’), the object of spit being semantically subsumed and therefore redundant (compare: color in *yellow-color car). Third, the use of the low-frequency word saliva suggests that the young boy was mapping the Cantonese V-O verb phrase 吐口水 (tou ³³ hau ³⁵ seoi ³⁵, literally ‘spit mouth water’) onto English. Compare:

(2)	我 吐 口 水 落 隻 蜘 蛛 度
	ngo 23	tou 33	hau 35 seoi 35	lok 22	zek 33	zi 55 zyu 55	dou 22
	1sg	spit	saliva	on	CL	spider	place
	‘I spit on the spider.’

In sum, (1) is a clear example showing how, despite an apparent gap in a bilingual speaker’s English lexicon (i.e., the English verb spit), that speaker would turn to the equivalent in some other language within his or her linguistic repertoire to make meaning (in this case Cantonese). Whether it was due to a momentary lapse of memory or ignorance of the verb spit, the young boy probably first acquired that meaning in Cantonese, which is expressed in a V-O phrase (吐口水, tou ³³ hau ³⁵ seoi ³⁵), which motivated him to look up the meaning of or asked his caretaker for that everyday expression 口水 (hau ³⁵ seoi ³⁵, ‘saliva’), whose equivalent in English is a low-frequency word that is hardly needed by his age-relevant English-L1 peers. In terms of communicative effectiveness, other than the flouting of a grammatical norm in Standard English, the intended meaning of (1) was not adversely affected or lost in what was virtually bilingual speech involving both English and Cantonese.

Traditionally, the embedding of Cantonese 吐口水 (tou ³³ hau ³⁵ seoi ³⁵) in an English sentence as in (1) may be variously analyzed as ‘code-switching’, ‘code-mixing’, ‘code-alternation’ or ‘lexical borrowing’, among others. In general, if the source language (SL) element follows the pronunciation norms of the SL (here, Cantonese) rather than that of the recipient language (RL), it is usually analyzed as an instance of switching, mixing or alternation. By contrast, if its pronunciation has been accommodated to the phonological system of the RL (here, English), it is more customary to analyze it as an instance of lexical borrowing. The problem is that determining whether the pronunciation of a given SL word (string) deviates from the SL or approximates the RL is often not as straightforward. To avoid terminological complications, we will follow Clyne (1997, 2003), and use the related terms ‘transfer’ and ‘transference’, as follows:

A ‘transfer’ is an instance of transference, where the form, feature of construction has been taken over by the speaker from another language, whatever the motives or explanation for this. ‘Transference’ is thus the process and a ‘transfer’ the product. (Clyne 2003, p. 76)

Accordingly, 吐口水 (tou ³³ hau ³⁵ seoi ³⁵, ‘spit saliva’) in (1) will be analyzed as an instance of transfer from Cantonese into English at the lexical level. As Clyne (2003, p. 76) has made clear, transference may take place at different levels – phonetic/phonological, prosodic, tonemic, lexical, morphological, syntactic, semantic, graphemic, in any combination. For instance, as a result of the transference of English words into Hong Kong Cantonese, its phonological system has gradually been expanded to accommodate those ‘loanword syllables’ (e.g., [k^hɔ⁵⁵] < call; [wɛn⁵⁵] < van; [dzel ⁵⁵] < gel), which are “non-occurring syllables or unused syllables which represent both accidental and systematic gaps in the syllabary” (Bauer and Wong 2010, p. 7). Notice that traditionally there were no Cantonese syllables ending with the lateral [l]; owing to the transference of English words like feel, gel and sell, the phonological system of Hong Kong Cantonese has been expanded to include the loanword syllables [ɪl] and [el]. From 1997 to 2006, such loanword syllables increased from 40 (Bauer and Benedict 1997) to 49 (Bauer 2006), and was further extended to 78 in 2010 (Bauer and Wong 2010; cf. Li et al. 2016). Lexically, while it cannot be denied that the extent of integration varies from one case to another, including ‘nonce borrowing’ or ‘nonce loans’ that would never occur again owing to a low level of acceptability in society (for a critical discussion, see Onysko 2007, pp. 37–38), Clyne’s (2003) terminological distinction has the advantage of freeing us from a concern, namely, to what extent a given context-bound SL transfer has been integrated into the RL. This in turn allows us to focus on the possible reasons behind specific instances of transference in context.

In plurilingual interaction, when all linguistic resources within a speaker’s repertoire are used to make meaning, their language outputs naturally contain elements which are traditionally associated with different languages, and more or less discrete styles, genres or registers within the same language. Where two or more natural languages are involved in plurilingual interaction, depending on the structural distribution of such elements at the utterance (spoken) or sentence (written) level, one could usually identify the dominant or matrix language, into which elements of the embedded language(s) are inserted. Broadly speaking, transference may take place inter-sententially at clause level, or intra-sententially within a clause, both of which are exemplified in the following excerpt adapted from a bilingual radio programme on Metro Radio.

(3)	An Excerpt of a Radio Programme (Li 2001a, p. 9; my approximate translation on the right)
	Date and time:	Channel:	Speaker:	Gender:
	24-7-2000, 3:30 pm	Metro Radio (精選104)	Disc Jockey	Male
	(i) 我希望呢 今日喺呢個長途電話訪問裏面呢 真係可以面對面, 咀對咀, 唔係, phone 對 phone 問佢 Chanel O’Connor 一條問題 (ii) once and for all, just tell us, are you … or are you not…? (iii) 呢首作品歌曲名字叫做 ‘No man's woman’. (iv) How’s your weekend? (v) 呢個weekend你過成點呢? (vi) Anything special? (vii) 我聽之前節目啲聽眾講都非常之唔錯喎, 有人話去南灣, 有人話去大嶼山, 鬼咁健康! (viii) 我就去咗見Sasha…		(i) I hope, in the long-distance call today, I can really, face to face, mouth to mouth, no, phone to phone, ask her, Chanel O’Connor, a question, (ii) once and for all, just tell us, are you … or are you not …? (iii) The name of this song is ‘ No man’s woman ’. (iv) How's your weekend? (v) How are you doing this weekend? (vi) Anything special? (vii) I heard from a few (radio) fans earlier that they’re doing fine, some said they went to South Bay, others to Lantau Island, so healthy! (viii) As for me, I went to see Sasha…

For convenient reference, the disc jockey’s utterances are reproduced below in linear order:

• (i) 我希望呢 今日喺呢個長途電話訪問裏面呢 真係可以面對面, 咀對咀, 唔係, phone 對 phone 問佢 Chanel O’Connor 一條問題³

• (ii) once and for all, just tell us, are you … or are you not…?

• (iii) 呢首作品歌曲名字叫做 ‘No man's woman’.⁴

• (iv) How's your weekend?

• (v) 呢個weekend你過成點呢?⁵

• (vi) Anything special?

• (vii) 我聽之前節目啲聽眾講都非常之唔錯喎, 有人話去南灣, 有人話去大嶼山, 鬼咁健康!⁶

• (viii) 我就去咗見Sasha…⁷

In this excerpt consisting of eight more or less discrete utterances of varying lengths, some are entirely in English (ii, iv, and vi); one only in Cantonese (vii); the rest of the four utterances have Cantonese as the matrix language, with English words inserted (i, iii, v, and viii). If we think of alternation between languages metaphorically as an operation of switching, we may say that an inter-sentential, clause-level switch from Cantonese to English occurs between (i) and (ii), between (iii) and (iv), and between (v) and (vi), while a switch from English back to Cantonese is found between (ii) and (iii), between (iv) and (v), and between (vi) and (vii). Those scholars who analyze intra-sentential code-switching as ‘code-mixing’ would regard utterances (i), (iii), (v), and (viii) as instances of ‘mixed code’, which is characterized by the insertion of elements from the embedded language (here, English) into the matrix language (here, Cantonese). Such a pattern of language use, blending Cantonese and English seamlessly in what may be called ‘infotainment discourse’, is rather typical of disc jockeys’ speech style as well as talk shows hosted by bilingual commentators of local radio or TV programs. Beyond such bilingual programs on broadcast media, however, the speech style of the disc jockey is less typical of the way Chinese Hongkongers speak and therefore less often encountered in society. This is largely because expressing ideas at clause length entirely in English, as shown in utterances (ii), (iv) and (vi) above, is less common – except in (ii), which is arguably triggered by an imagined question raised to Chanel O’Connor, a non-Cantonese speaker.

Unlike the young boy in (1) and the disc jockey in (3), who were ready to interact with others in English spontaneously, the majority of Chinese Hongkongers, children and adults alike, tend to use much more Cantonese than English in their bilingual interaction with one another. The syntactic structures are clearly Cantonese (‘matrix language’, Myers-Scotton 1993a), with short English expressions inserted (Muysken 2000; cf. ‘embedded language’, Myers-Scotton 1993a), typically in accordance with the grammatical requirement in Cantonese. That is, an English noun or noun phrase is inserted where a Cantonese noun or noun phrase is expected; with few exceptions, very much the same is true of English words from the other open word classes: verbs (or verb phrases) and adjectives (or adjective phrases). One good illustration is (4), which is an excerpt of a conversation between a male and a female speaker in their early 30s. That excerpt was carefully reconstructed by a group of three students who were present when the conversation took place; they had been trained to collect and record field work data using pen and paper without the support of an intruding tape-recording device (an approximate translation is provided on the right hand side).⁸

(4)	Place: On an MTR train compartment	Relationship: Couple or close friends	Age: About 30 years old	Gender: 1 Male & 1 Female
	(i)	M: 呢個樓盤唔錯喎! [pointing at the brochure]9	M: This [new] commercial property looks nice! [pointing at the brochure]
	(ii)	F: 係咩? 有無會所 ?10	F: Is it? Is there a club house?
	(iii)	M: 有呀! 個club house 仲好大添呀,又有泳池11	M: Yes, there is! The club house is real big, and there’s also a swimming pool.
	(iv)	F: 個club house有咩玩呀,有無健身室 ?12	F: What facilities are there in the club house, is there a gym?
	(v)	M: 有呀! Gym room 一定有啦,而且仲有好多健身器材添呀!13	M: Yes! Certainly, there is a gym [room], and a lot of fitness equipment as well!
	(vi)	F: 哦,咁都OK喎…我地可以一齊做gym如果我地住呢度! [pointing at the brochure]。近海喎,個view一定勁正14	F: Oh, sounds OK…we could do [exercise in the] gym together if we live here! [pointing at the brochure]. So close to the sea, the view must be super fantastic.

Structurally speaking, the language use pattern of this extract is representative of Hong Kong Cantonese-English ‘mixed code’ (中英夾雜, zung ⁵⁵ jing ⁵⁵ gaap ³³ zaap ²²/zhōngyīng jiázá) in many ways. First, except for ‘OK’, which is a frequently used ‘discourse marker’, the English elements are mainly nouns inserted within a clause (i.e., intra-sententially rather than inter-sententially), a tendency which is relatively more common compared with the insertion of English verbs, adjectives and adverbs into Cantonese. Second, the English insertions are sometimes used in free variation with their Cantonese equivalents. For instance, club house and gym [room] occur twice; both were initiated by the male speaker in (iii) and (v), after the female speaker first mentioned their Cantonese equivalents in her preceding questions, that is, 會所 (wui ²² so ³⁵) in (ii) and 健身室 (gin ²² san ⁵⁵ sat ⁵⁵) in (iv). In (iii) and (v), the male speaker’s switch to English may have been influenced by the printed information in the brochure he was browsing, which was most likely bilingual in Chinese and English. Third, it can be seen that many of the English insertions are monosyllabic, for example, club house, gym room, and view. Monosyllabic English words (MEWs) may also be combined with Cantonese morphemes to form verb phrases, as in the case of 做 gym (zou ²² gym, ‘do gym [exercise]’) in (vi). As Li et al. (2015, 2016) have shown, the frequent transference of MEWs into Cantonese is probably due to a typological characteristic in the recipient language, Cantonese, such that high-frequency MEWs are treated collectively by Cantonese-L1 Hongkongers like Cantonese morphemes. We will have more to say below about this salient pattern of transference into Cantonese.

Local university students are among those whose informal Cantonese is commonly embedded with English words. Very often, English expressions are preferred because no semantically satisfactory translation equivalents (are thought to) exist. This is clearly the case of words like project and presentation (see, e.g., the video production, ‘Multilingual Hong Kong: Present 一個 project’ by Chen and Carper 2005). Apart from lexical gaps in Cantonese, very often an English term is used largely because its putative equivalent in Cantonese is semantically incongruent and therefore not useable from the speaker’s or writer’s point of view. This is one of the findings in an experimental study conducted by Li and Tse (2002) who, following the ethnomethodological principle of ‘revelation through disruption’, instructed 12 English majors not to use English for one day, with a view to seeing whether and if so under what circumstances English was considered indispensable in their context-specific interaction with others (cf. Li 2011a, b). One instructive example was reported by a female participant (F3), who was tempted to invite a new male acquaintance to play wargames in the countryside. In Hong Kong, wargame is rendered in colloquial Cantonese as打野戰 (daa ³⁵ je ²³ zin ³³, literally ‘fight wild battle’), which, however, is also commonly used in soft-porn literature referring to illicit sexual activities. Being mindful of the artificial no-English-allowed rule of speaking on the day of the experiment, F3 used Cantonese (i.e., daa ³⁵ je ²³ zin ³³) to invite that new male acquaintance to ‘fight wild battle’ with her, which turned out to be extremely embarrassing for both. At the subsequent focus group interview where participants could give fuller details of ‘rich’ events that left them a deep impression, F3 pointed out somewhat emotionally that she would have no doubt followed the common parlance and used the code-mixed expression daa ³⁵ wargame if she had not been obliged to observe that funny no-English-allowed rule of speaking. Similarly, many examples of calquing in Cantonese, when first introduced, tend to have limited currency and a low level of social acceptance. This is clearly the case of various renditions of ‘mobile phone’ in the 1990s, when different Chinese translations on both sides of the Taiwan Strait and in the two SAR’s competed for currency, including 流動電話 (lau ²¹ dung ²² din ²² waa ³⁵, ‘flow phone’), 移動電話 (ji ²¹ dung ²² din ²² waa ³⁵, ‘move phone’), and 手提電話 (sau ³⁵ tai ²¹ din ²² waa ³⁵, ‘hand-held phone’). It took about five years, around the beginning of the new millennium, when communities across Greater China collectively settled for the bisyllabic 手機 (sau ³⁵ gei ⁵⁵/shŏujī, literarily ‘hand machine’ or ‘hand phone’; compare Handy in German and 핸드폰 [haendeupon] in Korean). Another high-frequency example is the calquing of the word deadline as 死線 (sei ³⁵ sin ³³), which was typically regarded as a joke when it first occurred some 15 years ago in the 1990s. Today, there is some indication that死線is in the process of being naturalized and increasingly felt to be acceptable, as shown in its use in more or less formal HKWC texts, with or without scare quotes (angle brackets in Chinese texts). For example:

(5)	傳 亞 視 續 牌 今「死 綫」 蘇 錦 樑 拒 評 論 [headline]
	cyun 21 aa 33 si 22 zuk 22 paai 21 gam 55 sei 35 sin 33	Sou 55 gam 35 loeng 21	keoi 23 ping 21 leon 22
	rumour ATV extend licence today deadline	Greg So	decline give comment
	‘Rumour has it that ATV’s “deadline” for extending its licence is today
	Greg So declined to give comments.’ (Sky Post, 晴報, 31/03/2015, p. A2)

(6)	…過 晒 交 建 議 書 的 死 線…
	gwo 33 saai 33 gaau 55 gin 33 ji 23 syu 55	dik 55	sei 35 sin 33
	pass completely submit proposal	NOM	deadline
	‘…the deadline for [ATV to] submit a proposal [to extend the licence] has passed…’
		(Headline Daily, 頭條日報 2/4/2015, p. 4)

Examples such as these (wargame, mobile phone, deadline) suggest that avoiding unwanted semantic loss or gain is one of the main reasons for preferring the English expressions while using Cantonese/HKWC, resulting in lexical transference or mixed code (cf. Li 2001b; for more examples and discussion of borrowing from English, see Shi et al. 2014, Ch. 6 and 7).

Other linguistic motivations of lexical transference in Hong Kong mixed code may be illustrated with the following examples featuring conversations between university students.

(7)	Place: University Computer Centre	Relationship: Schoolmates	Age: About 22 years old	Gender: Both female (F1, F2)
	(i).	F1: 琴日in 成點呀?15	F1: The interview yesterday, how was it?
	(ii).	F2: 我一去到佢就叫我簽約咯, 根本就唔算係second in!16	F2: [You know what?] As soon as I got there, he asked me to sign a contract, [it was] not at all a second interview!
	(iii).	F1: 咁咪好囉, 咁易就請, 你就好啦, 咁快搵到嘢做。17	F1: That’s great, you got the job so easily; good for you, found a job so quickly.
	(iv).	F2: 我根本就唔想做, 你睇下我呢份聘書, 成張學校通告咁, 根本就唔pro。唔講住喇, 做完呢份report先講啦!18	F2: I don’t really want to take it up. Look at my employment letter; it’s like a school announcement, not professional at all. Don’t chat about it [now]; [let’s] get this report done first!

There are two features of interest in (7). First, there is a tendency for polysyllabic English words to be clipped to just one syllable, with their denotation and connotation remaining intact. Thus the verb interview in (i) is reduced to monosyllabic in [pronounced in high level tone in ⁵⁵], while the noun phrase second interview also gets simplified as second in in (ii). Likewise, three of the four syllables in the adjective professional are deleted, making the initial syllable pro the de facto exponent of that meaning in (iv) (compare gymnasium ➔ gym). There is no evidence of any communication problem, suggesting that the clipping of long English words to one syllable is widely recognized and used. In Li et al.’s (2015, 2016) studies of the ‘Monosyllabic Salience Hypothesis’ (MSH), it was found that in a Hong Kong Chinese newspaper corpus of about 600,000 characters (Li et al. 2014), such a tendency to reduce or truncate polysyllabic English words to one syllable is statistically more marked with verbs and adjectives compared with nouns (cf. Luke and Lau 2008). This is especially clear with regard to polysyllabic English lexemes which are identical in spelling and pronunciation except for their word class. For instance, whereas the noun copy is usually rendered bisyllabically as [k^hɔːp⁵⁵p^hiː²¹], as a verb copy is systematically truncated to one syllable as [k^hɔp⁵⁵]. Similar contrasts occur with fail, minor, major, reply, report, and tips (see Table 5, Luke and Lau 2008, p. 353).

Second, while there is a standard, semantically congruent translation of report (報告, bou ³³ gou ³³), this English word is still preferred in (iv), probably because it is felt to be more specific when making reference to a particular course assignment (compare: project report, lab report). This is consonant with Li’s (2011a, b) study involving data obtained from participating university students after going through a ‘one day with only Cantonese’ (Hong Kong) or ‘one day with only Mandarin’ (Taiwan) experiment. One of the key findings in these ‘one day’ studies is that when technical concepts or academic/school jargon are first introduced or learned in language X (here English), those terms or jargon tend to be cognitively mediated by language X, even though their translation equivalents in another language (language Y) have subsequently been encountered. Such a psycholinguistic motivation, termed ‘medium-of-learning effect’ (MOLE, Li 2011a, b), may be traced back to Gibbons’ (1987) observation of ‘learning effect’, which he conjectured was one of the main reasons why students at Hong Kong University were so prone to mixing English into their Cantonese, resulting in a language use pattern he called ‘MIX’. Interestingly, being field-specific, MOLE is consistent with Fishman’s (1972, p. 439) “topical regulation of language choice”, for “certain topics are somehow handled ‘better’ or more appropriately in one language than in another in particular multilingual contexts”. As for the various “mutually reinforcing factors” leading to such topical regulation, Fishman explains as follows:

Thus, some multilingual speakers may ‘acquire the habit’ of speaking about topic x in language X (a) partially because this is the language in which they are trained to deal with this topic (e.g., they received their university training in economics in French), (b) partially because they (and their interlocutors) may lack the specialized terms for a satisfying discussion of x in language Y, (c) partially because language Y itself may currently lack as exact or as many terms for handling topic x as those currently possessed by language X, and (d) partially because it is considered strange or inappropriate to discuss x in language Y. (Fishman 1972, pp. 439–440, emphasis in original)

In a footnote on the same page, Fishman explains his point (b) further as follows:

This effect [i.e. lacking the specialized terms for a satisfying discussion of x in language Y] has been noted even in normally monolingual settings, such as those obtaining among American intellectuals, many of whom feel obliged to use French or German words in conjunction with particular professional topics. English lexical influence on the language of immigrants in the United States has also been explained on topical grounds. (Fishman 1972, p. 439)

In light of the topical regulation of language choice, which in turn may be accounted for by MOLE, it is not difficult to understand why the conversation in (8) between two hall-mates regarding their ideal choice of a minor in their undergraduate degree studies is sporadically ‘sprinkled’ with academic/school jargon in English (highlighted), including the word ‘minor’ itself.

(8)	Place: University hostel	Relationship: Hall-mates	Age: About 20 years old	Gender: Both female (F1, F2)
	(i).	F1: 你有無諗過讀minor呀?19	F1: Have you thought about studying a minor [subject]?
	(ii).	F2: 我想讀Psychology做minor呀…你呢?20	F2: I want to study Psychology as minor…what about you?
	(iii).	F1: 無呀…我唔想讀多成十五個credits 呀.21	F1: No. I don’t want to study an extra 15 credits.
	(iv).	F2: 我都係……但係好似好有用咁……22	F2: Neither do I……But [it] looks very useful [to minor in Psychology].
	(v).	F1: 都係…咁你諗住讀咩科呀?23	F1: True…Which courses will you choose?
	(vi).	F2: 無呀… Basic principles of Psychology… Abnormal Psychology…Movie and Psychology…不過我都未諗定呀.24	F2: Not sure…Basic principles of Psychology… Abnormal Psychology……Movie and Psychology……but I haven't decided yet.
	(vii).	F1: 其實我都有諗過minor Global B !25	F1: Actually I also thought about studying a minor in Global B.
	(viii).	F2: Global B? 咩嚟 ?26	F2: Global B? What’s that?
	(ix).	F1: Global Business囉.27	F1: Global Business.
	(x).	F2: 都好吖! 好似好有用咁!28	F2: That’s good! Sounds very useful.
	(xi).	F1: 唔知呀…都係睇定吓先…29	F1: Don’t know…better wait and see.

Except for course titles in (vi), all the other English terms have SWC equivalents:

副修	fu 33 sau 55	‘minor’
學分	hok 22 fan 55	‘credit (point)’
心理學	sam 55 lei 23 hok 22	‘psychology’
環球企業管理	waan 21 kau 21 kei 23 jip 22 gun 35 lei 23	‘Global Business Management’

For students like the two hall-mates in (8), however, the corresponding Chinese terms have little currency in speech, probably because at English-medium universities, Chinese (Cantonese/HKWC) is seldom used in the public discourse of university administration, internal communication with students (by email or on the intranet), and course titles such as those listed in program handbooks. That is probably why it sounds strange to use Chinese to refer, for example, to miscellaneous school jargon (e.g., credits, GPA, major, minor, program), names of locations within the university (e.g., Computer Centre, Learning Commons, Red Zone), academic disciplines (e.g., phonology, robotics) and course titles (e.g., Abnormal Psychology, Global Business), even though their Chinese equivalents may be cognitively retrievable by the speaker/writer at the time of speaking or writing. Here again, the truncation of polysyllabic words is found with regard to course titles: Global B usiness Management ➔ Global B (cf. Ed ucation Psy chology ➔ et ⁵⁵ saai ²², often abbreviated in writing as ‘Ed Psy’).

MOLE is not at all restricted to educated users at tertiary level. The following conversation extracted from a dialogue between a private tutor and her 10-year-old Primary 4 (Grade 4) tutee is similarly ‘littered’ with technical terms and academic jargon in English (here: English grammar).

(9)	Place: Tutee’s home	Relationship: Private tutor & Tutee	Age: Tutor (F1) over 20; Tutee (M1) about 10	Gender: One female (F1) & one male (M1)
	(i).	F1: 你琴日個test做成點呀? 老師有無問tenses呀?30	F1: Your test yesterday, how was it? Did the teacher ask about tenses?
	(ii).	M1: 老師出咗present tense同埋past tense之嘛.31	M1: The teacher tested [us] present tense and past tense only.
	(iii).	F1: 咁preposition呢?32	F1: What about preposition[s]?
	(iv).	M1: 好似無出喎。33	M1: Didn’t seem to occur [in the test].

Regardless of whether the tutor and tutee were aware of the Chinese equivalents such as the following:

測驗	caak 55 jim 22	cēyàn	‘test’
時態	si 21 taai 33	shítài	‘tense’
現在式	jin 22 zoi 22 sik 55	xiànzàishì	‘present tense’
過去式	gwo 33 heoi 33 sik 55	guòqushì	‘past tense’
介詞	gaai 33 ci 21	jiècí	‘preposition’

the medium-of-learning effect (MOLE, Li 2011a, b) helps explain why these English terms come to mind more readily compared with their Chinese equivalents, while as a consequence, the topical regulation of language choice (Fishman 1972) accounts for the naturalness of referring to English grammar terms in English. The key to both phenomena, however, is the language of instruction. For instance, mainland Chinese students who learn English through the medium of Putonghua (i.e., Putonghua as the medium of instruction, or PMI) would find it perfectly natural to use the corresponding Chinese terms to refer to various categories and aspects of English grammar. There is one very instructive example in my ‘one day with no English’ data (Li and Tse 2002). One female participant (HEF9) was a non-Cantonese-speaking exchange student from mainland China, who had been in Hong Kong for only 4 months. From her reflective diary written in simplified Chinese characters (see (10) below) and the subsequent focus group sharing, she indicated that before coming to Hong Kong, she had rarely found it necessary to insert any English words into her Putonghua. That changed after studying in Hong Kong for about four months. Probably influenced by the intensity of bilingual interaction involving Cantonese and English in the SAR, she gradually became aware of an increasing practice of inserting English expressions of various lengths into her Putonghua, which motivated her to take part in the ‘one day’ experiment. One interesting example she gave was the abbreviated course title ‘CCIV’, referring to ‘Chinese civilization’, which she said she could not help saying (pronounced in four syllables) every time she referred to it. More interesting still, in her reflective diary she used that example to justify what she called ‘the first impression hypothesis’:

(10).

当一个人第一次接触一个新词汇是用英文时,则这个词留在他脑海中的印象就是英文,以后使用英文来表达这个词的机会比较大些。 例如:我第一次接触到中国文化中心的课程时,就是CCIV,则在以后的表达中我一直使用CCIV来表达,本次实验是我第一次用中文来表达,非常不习惯,不自然. (HEF9)34

‘When a person first encounters a new term in English, the impression of this term in that person’s mind will be in English, and so later the chance of using that English term will be higher. For example, the first time I came across the course offered by Zhōngguó wénhuà zhōngxīn [literally ‘Chinese Civilization Centre’] is CCIV. After that, I have always used CCIV to refer to that course. [In] this experiment I used the Chinese term [of this course] for the first time, [which is] unnatural and [I am] not used to it at all.’

This mainland Chinese participant’s (HEF9) reasoning or conjecture was shared by a few other participants (Li 2011a, pp. 231–232). In essence, it is not unlike an observation made by F4, a Hong Kong participant in Li and Tse’s (2002, p. 174) ‘one day with no English’ study, namely ‘the first one who entered is the master’ (先入為主, sin ⁵⁵ jap ²² wai ²¹ zyu ³⁵/xiān rù wéi zhǔ). Such a ‘First-Impression Hypothesis’ (FIH) may be formulated as follows:

When a concept C is first encountered in language X, and provided X is the widely preferred language for expressing C in the community, then C tends to be cognitively mediated through the language X (Cx), even if a direct translation of C is subsequently encountered in language Y (Cy). (cf. Li 2011a, p. 230)

The first-impression hypothesis (FIH) predicts that Cx (new concept C introduced in language X) – if proved to be a popular (rather than idiosyncratic) choice of its users – would be cognitively more salient than Cy (concept C subsequently available in language Y), as shown in the strong tendency of concept C being more readily retrieved in language X than in language Y. Additional empirical evidence for FIH and MOLE includes the naturalness of using a specific language when being trained in a particular sport, for example, Japanese for judo (e.g., te waza, ‘to throw’), Korean for taekwondo (e.g., chi-gi, ‘punch’), French for fencing (e.g., marche!), English for modern dance (e.g., freeze!). Examples such as these were reported by Taiwanese student participants who were inconvenienced by being prevented from using their usual languages on various speech events, including at training sessions during the ‘one day with only Mandarin’ experiment (Li et al. 2010). More research is needed to ascertain the psycholinguistic validity of the First-Impression Hypothesis (FIH) and the medium-of-learning effect (MOLE).

2.3 Influence of English on Hong Kong Written Chinese (HKWC)

Traditionally, being a ‘dialect’, Cantonese is officially considered as inappropriate for writing. This is why, strictly speaking, ‘written Cantonese’ (Snow 2004, 2008, 2010, 2013) is a linguistic anomaly that must be ironed out through education and, if surfaced in students’ class work or homework, be banned and corrected with SWC-based school literacy. This approach appears to have worked in Cantonese-speaking regions across the border in Guangdong province, China. Whereas Cantonese as a vernacular has continued to thrive (e.g., radio and TV programs and opera), Cantonese elements in print are seldom used in mainland Chinese public media, print or electronic (Snow 2004). This is not the case in Hong Kong, however. Owing to political insulation from the mainland from 1949 to the early 1980s, and the British colonial government’s “benign indifference” toward Chinese language education and use in society (Bauer 1995), Cantonese has flourished in a large number of domains:

In satisfying the social, cultural, and linguistic needs of Hong Kong’s predominantly Chinese community of six million inhabitants [over 7.1 million as of April, 2016], Cantonese has become widely used as the language of radio news programs and plays, TV news broadcasts and soap operas, live theatre, popular songs and novels, newspaper cartoons and serialized stories, and mass advertising. (Bauer 1995, p. 246)

To Bauer’s list of domains may be added debates in the Legislative Council (Legco), which have shifted from English to Cantonese after the return of Hong Kong’s sovereignty to China in July 1997, and court trials in Cantonese. The latter, though by no means commonplace today, are no longer seen as a novelty after the first court case was heard in colonial Hong Kong about two decades ago in December 1995 (Buddle 1995). The tremendous vitality of Cantonese in society, including its use as the medium of instruction in school from primary to secondary, helps explain why written Cantonese elements in Hong Kong have been given social space to grow, notably in informal sections or genres of Chinese newspapers, such as columns, infotainment, advertisements and cartoons, and electronic communication platforms such as MSN, SMS, and social media like Facebook, Twitter and Whatsapp.

Unlike ‘hard’ news stories, editorials and feature articles, the ‘soft’ sections of Hong Kong Chinese newspapers and magazines are generally exempt from strict monolingual norms and tend to favor a vernacular-driven writing style, whereby linguistic resources from conventionally discrete language varieties – Cantonese, Standard Chinese, Classical Chinese, English, as well as genres and registers within any of these – are mobilized to attain rhetorical effects that are otherwise impossible in SWC alone. Before English came into the picture, until the 1960s a writing style known as 三及第 (saam ⁵⁵ kap ²² dai ³⁵) involving the mixing of SWC and classical Chinese elements into Cantonese, was made popular by the political satirist 三蘇 (Saam Sou) and a few columnists (Wong 2002). As Snow (1991) remarks, saam ⁵⁵ kap ²² dai ³⁵ is:

a written language which combines classical Chinese, Cantonese and SC [Standard Chinese] (...). The beauty of this style is that it allows a writer a very broad range of registers. Classical Chinese creates a distinct impression of formality, and Cantonese creates the impression of slang, thus allowing the writer to make radical shifts of tone and create linguistic incongruity that is both amusing and arresting. (Snow 1991, p. 147)

Since the 1960s, the saam ⁵⁵ kap ²² dai ³⁵ writing style, which has progressively become even more hybrid with elements of English infused into the linguistic admixture, is arguably the unmarked writing style not only in those newspapers and magazines characterized by features typical of the popular press, but also the back pages of the ‘quality’ press as well. As Lin and Li (2015) observe:

this [saam ⁵⁵ kap ²² dai ³⁵] style has won the hearts of many readers (or ‘Like’ in the facebook era) who appreciate the subtle nuances and humour conveyed successfully by such a fluid performance through the mobilization of multiple linguistic resources (Wong 2002) to juxtapose multiple social views and voices. This trend has continued since the 1970s; to make meaning creatively, skillful writers who are trilingual in Cantonese, SWC and English would draw on the semiotic potential of elements from their whole linguistic repertoire, which is treated as a composite pool of resources rather than as compartmentalized languages or registers. (Lin and Li 2015, pp. 86–87)

As a translingual, heteroglossic writing style (Bakhtin 1935/1981; Bailey 2012), English-infused saam ⁵⁵ kap ²² dai ³⁵ may be seen as the Hong Kong press industry’s collective response to survive cut-throat competition. Given the preference of their Cantonese-dominant yet plurilingual readers, and in the face of the onslaught of many e-rivals, locally and internationally, editors of print and electronic dailies alike have no choice but to shape their writing style in ‘soft’ sections to appeal to their readers’ collective preference for vernacular-driven writing (Bell 1991).

Since literacy in colloquial Cantonese has never been officially standardized and is banned in formal writing through SWC-based literacy training in school, expressing colloquial Cantonese elements in print is sometimes a challenge. What happens if a particular Cantonese morpho-syllable has no known written representation? Research has shown that Cantonese-dominant Hongkongers would resort to all kinds of linguistic means in order to lend expression to their target morpho-syllables. The key is an age-old, highly productive strategy known as 假借 (gaa ³⁵ ze ³³/jiǎjiè, ‘phonetic borrowing’ or ‘phonetic loan’; literally ‘false borrowing’), which happens when an element from any language is borrowed not for its semantic content but only for its phonetic value (sometimes just an approximation, Li 2000). Table 2.1 shows some examples of SWC morphemes being borrowed for their (approximate) sound value to represent Cantonese morphemes in print.

Table 2.1 Examples of phonetic borrowing from SWC into Cantonese, and literacy problems engendered for non-Cantonese readers

As shown in the examples in Table 2.1, phonetic borrowing from SWC (or classical Chinese to a lesser extent) sometimes entails a semantic shift, as shown in (i) – (iv), while the usual SWC meaning of the phonetic loan is totally irrelevant, as in all the examples (i) – (viii). In other cases, there may also be an additional shift in word class or function, for example: adjective ➔ classifier in (i); noun ➔ verb in (iv); a shift in tone level, as in (iii), (v), (vi) and (vii); or a shift in the segmental from [f] to [b] as in (ii). All this creates literacy problems for non-Cantonese readers; such problems are further aggravated when phonetic borrowing from English is used to fill the orthographic gap, when no SWC morpho-syllable is deemed to be suitable. Based on written Cantonese data collected from the pocket-book literature in the late 1980s, Luke (1995) identified three common solutions, in descending order of significance: (a) phonetic loan from some existing Chinese character similar or identical in pronunciation to the target morpho-syllable, (b) coinage of a new character, or (c) phonetic loan based on some existing English word or letter (combination). The choice of these possible solutions, Luke (1995, pp. 107–108) argues, is guided by two underlying principles: ‘phonetic proximity’ and ‘Chinese-character-based written representation’, of which the former appears to override the latter in case they are in conflict. In other words, if the selection of a similar-sounding phonetic loan from Chinese is viewed as causing semantic interference or literacy problem, a similar-sounding English syllable – including individual English letters or non-words in Roman script – would be preferred (cf. Bauer 1982, 1988; Cheung and Bauer 2002). This helps explain the proliferation of script mixing, such as fing 開 (fing ²² hoi ⁵⁵, ‘shake off’), lur 飯應 (loe ⁵⁵ faan ²² jing ³³, ‘readily accept/agree’), and jit 我 (zit ⁵⁵ ngo ²³, ‘tickle me’) so commonly found in popular Cantonese novels (cf. Luke 1995, pp. 105–107). These Roman-script-based coinages are clearly modeled on English pronunciation rules (compare: wing, fur and sit), and are therefore intelligible to Cantonese-speaking readers with basic literacy skills in English. These examples show that often a phonetically satisfactory solution cannot be found in the stock of Chinese characters to represent the Cantonese morpho-syllable, in which case a writer may turn to the Roman script for a written representation. In sum, the unavailability of a standardized orthography does little to stop Cantonese-dominant Hongkongers from expressing vernacular-based ideas in writing (Cheung and Bauer 2002; Li 2000; Luke 1995). Table 2.2 shows a few examples how monosyllabic English words are borrowed for their sound only, while their meaning is supposed to be backgrounded or ignored.

Table 2.2 Examples of phonetic borrowing from English into written Cantonese

As shown in the examples in Table 2.2, the meanings of the recognizable English loanwords are totally irrelevant. And, like those examples cited by Luke (1995) in the late 1980s, when there is no suitable Chinese character to represent the target Cantonese morpho-syllable in print, Hongkongers biliterate in Chinese and English have no difficulty coining a romanized Cantonese word such as pok (iv), including homophones like the English letter D (Bauer 1982, cf. Bauer 1988). More recent examples of romanized Cantonese words include hea (he ³³, ‘laid-back’ or ‘tardy’), chok (cok ³³, ‘suffocating’) and chur (coe ³⁵, ‘hard pressed for time’), whose written forms are similarly modeled on English (compare: hea vy, chok ed and chur ch). Such pseudo English words, like phonetic loans from English, are of little help when readers of English are searching for clues how they contribute to the textual meaning, for they are Cantonese morphemes expressed in Roman script (Li et al. 2016).

All this is reminiscent of the pidginization of English words and expressions, as evidenced in Chinese Pidgin English (CPE) attested during the eighteenth and nineteenth centuries. As Shi (1993) has pointed out, during that period, many Cantonese-speaking merchants in Canton (today’s Guangzhou) were eager to learn some English in order to do business with English-speaking ‘supercargoes’ (i.e., trade representatives) and sailors at a time when trading activities with ‘red-haired barbarians’ were tightly regulated and highly restricted. Short of formal instruction, some authors with knowledge of English compiled phrasebooks to help Cantonese merchants articulate basic English words and practical business-related expressions needed to communicate with English speakers. Such English words and expressions were written in Chinese characters (e.g., 紅毛通用番語, Hung ²¹ mou ²¹ tung ⁵⁵ jung ²² faan ⁵⁵ jyu ²³, ‘Red-haired people’s common foreign language’). This was done by substituting (approximate) Cantonese syllables for those required by English words. Thus, for example, ‘sailorman’ and ‘wife’ were transliterated as些利文 (se ⁵⁵ lei ²² man ²¹) and 威父 (wai ⁵⁵ fu ²²), respectively (for more examples, see Ansaldo et al. 2010). As Shi (1993) explains:

The CPE item is represented by one or several Chinese characters. The semantic content of the characters is irrelevant in so far as they were chosen for their phonetic and phonological value. When being read aloud in Cantonese, these characters gave rise to a string of sounds which purportedly represented the phonological form of a CPE word. (Shi 1993, p. 460)

Through repeated practice, that is, reading such CPE expressions out loud, eager Cantonese learners would give the impression of speaking (pidgin) English. What is interesting is that the earlier practice of using character-based Cantonese syllables to transliterate English words is like a mirror image of the current practice of using SWC characters to transliterate Cantonese syllables and, when the linguistic outcome is deemed unsatisfactory, they would have no hesitation turning to the Roman script for inspiration, the purpose being to lend written forms to the Cantonese morpho-syllables.

Written Cantonese is not at all limited to informal use; formal written Chinese in Hong Kong has also been influenced by Cantonese and English to a large extent. As Shi et al. (2014) have made clear using corpus data collected from formal sections of Chinese newspapers, by virtue of distinct Hong Kong characteristics in local (Han) written Chinese,³⁵ it is more appropriately called ‘Hong Kong Written Chinese’ (HKWC). This is because HKWC exhibits massive influence from Cantonese and English and, as such, deviates lexico-syntactically from SWC considerably. Shi et al. (2014, p. 6) further define HKWC as follows:

‘Based on SWC, adorned with some classical Chinese elements, being deeply influenced by Cantonese and English, and used mainly in Hong Kong, HKWC is different from SWC with regard to its vocabulary subsystems, word meanings and interpretations, structural combinations, sentence structures and usages.’ (my translation)³⁶

The nature and extent of various sources of lexico-syntactic and lingua-cultural influence (影響) on HKWC are characterized as follows (Shi et al. 2014, p. 25):

Standard Chinese:	fundamental influence (根本的影響)
Cantonese:	comprehensive influence (全面的影響)
English:	deep influence (深刻的影響)
Classical Chinese:	heritage influence (傳承的影響)
Lexical innovations and other dialects:	limited influence (有限的影響)

Gradually taking shape since the 1970s and ‘becoming mature’ (“趨於成熟”) and stabilized in the 1990s (Shi et al. 2014, p. 13), HKWC exhibits features that diverge from those in SWC. Often the same character (combination) may mean different things or vary in terms of collocational constraints. For instance, under the influence of English frontline and grassroots, HKWC has evolved calques 前線 (cin ²¹ sin ³³) and 草根 (cou ³⁵ gan ⁵⁵), which are used to modify other words (e.g., 前線工作人員 (cin ²¹ sin ³³ gung ⁵⁵ zok ³³ jan ²¹ jyun ²¹, ‘frontline worker’); 草根階層(cou ³⁵ gan ⁵⁵ gaai ⁵⁵ cang ²¹, ‘grassroots level’), whereas the same meanings are expressed in SWC by 一線 (yīxiàn, jat ⁵⁵ sin ²²) and 基層 (jīcéng, gei ⁵⁵ cang ²¹) respectively (Shi et al. 2014, pp. 152–153).

Syntactically, there is also evidence of syntactic transference from English. For example, the Anglicized clause structure in HKWC: 是時候… (si ²² si ²¹ hau ²²…, ‘it is time to…’) is clearly a result of the high-frequency English structure ‘it is time (for someone) to do something’, especially in local Chinese media. This structure is sometimes fronted with a locative expression in the subject position, a syntactic feature which is not admissible in SWC. For instance:

(11)	香 港 是 時 候 重 新 輸 入 活 雞 了.
	Hoeng 55 gong 35	si 22	si 21 hau 22	cung 21 san 55	syu 55 jap 22	wut 22 gai 55	liu 23
	Hong Kong	is	time	again	import	live chicken	F.P.
	‘It is time for Hong Kong to import live chickens again.’
	(Slightly modified, adapted from Shi 2006, p. 310)

A few other differences may be traced back to cultural differences. For instance, the meaning of一樓 (jat ⁵⁵ lau ³⁵, ‘first floor’) follows British practice and refers to the floor above the ground floor (dei ²² haa ³⁵, 地下), whereas the same floors are referred to in SWC as 一樓 (yīlóu, jat ⁵⁵ lau ³⁵) and 二樓 (èrlóu, ji ²² lau ³⁵), respectively (Shi et al. 2014, pp. 30–32).

HKWC is used in formal sections of printed media such as news stories, editorials and feature articles. What about informal sections of the same newspapers and magazines like columns, infotainment, adverts, and cartoons? Is it possible to write in Cantonese exactly like the way one speaks, keeping all the vernacular-style features such as lexical transference from English intact? The answer is a resounding ‘yes’. Apart from Cantonese pocket-book literature exemplified and discussed by Luke (1995), such ‘soft’ content is often written entirely in colloquial Cantonese (cf. Snow 2004, 2008). Below we will illustrate ‘colloquial written Cantonese’ with promotional discourse data on half a printed page in Headline Daily (頭條日報, Tau ²¹ tiu ²¹ jat ²² bou ³³/Tóutiáo rìbào), a Hong Kong newspaper distributed free of charge except Sundays and public holidays.

Given Hong Kong Chinese readers’ collective preference for the kinds of heteroglossic written Chinese which “are clearly more characteristic of those of heteroglossic orality, rather than those of ‘proper’ compartmentalized monolingual school literacy” (Lin and Li 2015, p. 86), practically all Chinese newspapers and most magazines contain sections of more or less ‘soft’ content, covering a wide range of topics from popular culture and infotainment to tips on good food and latest fashion, from chatty ‘talk of the town’ gossiping and photo stories to illustrated travelogues and adverts disguised as recommendations or personal preferences, among many others. For this reason, colloquial written Cantonese data, typically blended with some English, is rich and easily collectable. To illustrate, I will conduct a focused analysis of five texts covered within the space of half a printed page of a tabloid-like daily that prides itself on being the free newspaper with the highest print circulation in Hong Kong: Headline Daily (Fig. 2.1). Altogether there are six short texts (labeled schematically from Texts 1–5 (Fig. 2.2), each of about 100–250 characters in length, appearing in the same column with the following title:

(12)	商 界 講 呢 啲
	soeng 55	gaai 33	gong 35	ni 55 di 55
	business	sector	talk about	these
	‘This is what the business world talks about’

Fig. 2.1

A half-page column from a local tabloid-like daily newspaper distributed free of charge(Headline Daily, 11-04-2015, p. 24)

Fig. 2.2

Schematic representation of the text layout (商界講呢啲, Headline Daily 11 Apr 2014, p. 24)

The bylined columnist is Cally, a pen name in English with no Chinese name mentioned. In terms of graphic adornment of the column, instead of a picture or portrait of the writer, right above the column title on the top left-hand corner is a cartoon figure featuring a smiling woman with a cup (presumably of coffee or tea) in hand, projecting an image of a smart, enlightened office worker at managerial rank. The presentation of this column is illustrated with appropriate pictures or images accompanying each of the short texts (Fig. 2.1).

To appreciate the variety and extent of transference from English into colloquial written Cantonese more fully, Table 2.3 provides a synopsis of all the five texts and lists all the clauses embedded with some English, with comments on various aspects of transference from English given under ‘Remarks’ on the right. This is followed by a detailed analysis of each of the 15 English-embedded clauses listed, from (13) to (27) below.

(13)	踴 躍 捐 書	閱 讀 樂 趣 開 心 share  (Text 1a & 1b, heading)
	jung 35 joek 33 gyun 55 syu 55	jyut 22 duk 22	lok 22 ceoi 33	hoi 55 sam 55	share
	enthusiastically donate ‘book’	read	joy	happy	share
	‘Donate “books” enthusiastically [for] the joy of reading [let’s] share [books] happily’

(14)	呢 個 活 動 梗 係 要 開 心 share 同 大 力 支 持 啦! (Text 1b)
	nei 55 go 33	wut 22 dung 22	gang 35 hai 22	jiu 33 hoi 55 sam 55	share tung 21	daai 22 lik 22	zi 55 ci 21	laa 55
	this CL	activity	certainly	must happily	share and	strongly	support	F.P.
	‘This activity [we] certainly must share [books] happily and support [it] strongly!’

(15)	詳 情 可 瀏 覽 新閱會 Facebook 專頁 www.facebook.com/shkpreadingclub。(Text 1b)
	coeng 21 cing 21	ho 35	lau 21 laam 23	San 55 jyut 22 wui 35 Facebook zyun 55 jip 22 […]
	details	can	surf	Sung [Hung Kei Properties] Reading Club Facebook page […]
	‘For details [you] can refer to the Facebook page www.facebook.com/shkpreadingclub.’

(16)	手 機 x  PS4™	隨 時 隨 地 喪 打 勁 Game  (Text 2, heading)
	sau 35 gei 55 x PS4™	ceoi 21 si 21 ceoi 21 dei 22 song 33 daa 35 ging 22 Game
	hand phone x PS4™	anytime anywhere mad play super game
	‘hand phone x PS4™	play super [computer] games like mad anytime, anywhere’

(17)	仲 可 以 睇 埋 friend 嘅 打 機 實 況 (Text 2)
	zung 22	ho 35 ji 23	tai 35 maai 21  friend ge 33  daa 35 gei 55  sat 22 fong 33
	also	can	watch in addition friend NOM play computer live
	‘[You] can also watch [how your] friends play computer [games] live.’

(18)	手 機 變 身 做 遙 控 mon  (Text 2)
	sau 35 gei 55	bin 33 san 55	zou 22	jiu 21 hung 33	mon
	hand phone	transform	as	remote	monitor
	連 接 同 操 控 PS4™ 上 嘅 game
	lin 21 zip 33	tung 21	cou 55 hung 33	PS4™ soeng 22 ge 33 game
	connect	and	control	PS4™ above NOM game
	‘The hand phone gets transformed and becomes a remote control monitor connected to the games on PS4™.’

(19)	買 埋 PS4™ 嘅 game 同 download 落 去 PS4™ 度 添 (Text 2)
	maai 23	maai 21	PS4™ ge 33 game tung 21	download	lok 22 heoi 33	PS4™ dou 22 tim 55
	buy	also	PS4™ NOM game and	download	onto	PS4™ as well
	‘[And you may] also buy PS4™ games as well and download [them] onto PS4™.’

(20)	牛 仔 褲 都 一 定 係 我 衣 櫃 入 面 嘅 must have item (Text 3)
	ngau 21 zai 35 fu 33	dou 55	jat 55 ding 22	hai 22	ngo 23	ji 55 gwai 22	jap 22 min 22	ge 33 must have item
	jeans	also	must	be	1sg	wardrobe	inside	NOM must have item
	‘Jeans are also must-have items inside my wardrobe.’

(21)	輕 易 着 出 個 人 style (Text 3)
	hing 55 ji 22 zoek 33	ceot 55	go 33 jan 21	style
	easily	wear show	personal	style
	‘[I] can easily wear [jeans and show my] personal style.’

(22)	咁 多 個 brand 入 面,我 就 最 鐘 意 texwood (Text 3)
	gam 33 do 55	go 33	brand jap 22 min 22	ngo 23	zau 22 zeoi 33	zung 55 ji 33	texwood
	so many	CL	brand among	1sg	as for most	like	texwood
	‘As for me, among all the brands, I like texwood the most.’

(23)	着 起 嚟 有 種 高 廋 嘅 效 果, (Text 3)
	zoek 33 hei 35 lai 21	jau 23 zung 35	gou 55	sau 33 ge 33	haau 22 gwo 35
	put on	have kind	tall	slim NOM	effect
	成 個 長 腿 oppa 咁!
	sing 21 go 33	coeng 21 teoi 35	oppa gam 35
	whole person	Long Leg	oppa seem like
	‘[When I] put [the jeans] on [I look] tall and slim, which [makes me] look virtually like Long Leg Oppa!’ (N.B.: ‘長腿Oppa’ refers to the Korean celebrity Lee Minho [李敏鎬].)

(24)	型 格 牛 仔 褲 新 登 場 每 日 都 Feel good (Text 3, heading)
	jing 21 gaak 33	ngau 21 zai 35 fu 33	san 55 dang 55 coeng 21	mui 23 jat 22 dou 55	Feel good
	trendy	Jean	new arrival	every day also	feel good
	‘new arrival trendy jeans, [I] feel so good every day.’

(25)	排 清 毒 素 Keep 住 輕 盈 又 Healthy (Text 4, heading)
	paai 21 cing 55	duk 22 sou 33	Keep zyu 22 hing 55 jing 21	jau 22	Healthy
	excrete	toxic element	keep ASP poised/light-weight	and	healthy
	‘Excrete toxic elements completely   Keep poised / light-weight and healthy’

(26)	幾 時 都 keep 住 健 康 啦。 (Text 4)
	gei 35 si 21 dou 55	keep zyu 22	gin 22 hong 55	laa 55
	anytime also	keep ASP	healthy	F.P.
	‘[that] keeps [me] healthy anytime.’

(27)	按 摩 纖 體 油 讓 Body fit 起 來! (Text 5, heading)
	on 33 mo 55 cim 55 tai 35 jau 21	joeng 22	Body fit	hei 35 loi 21
	massage slim body oil	let	body fit	ASP
	‘Massage slim body oil   let the body get fit!’

Table 2.3 Synopsis of Texts 1–5, headings, English-embedded clauses, and comments on transference from English

There are plenty of language contact features in these five texts (Fig. 2.1), of which Cantonese-English contact features are quintessentially illustrated by examples (12) to (27). First, there is no question that these five texts are written entirely in colloquial Cantonese, which is partly characterized by extensive lexical transference from English. The writer followed the principle of ‘write as one speaks’ very closely (e.g., the extensive use of the genitive marker or nominalizer 嘅, ge ³³, instead of its HKWC equivalent 的, dik ⁵⁵).³⁷ Consequently, while the written medium determines that the texts in Fig. 2.1 are intended for silent reading (i.e., a literacy activity), a Cantonese-literate reader who reads them out loud (i.e., rendered through orality) would give the unmistakable impression that he or she is talking mainly in colloquial Cantonese. As such, it is not difficult to explain the large amount of lexical transference from English and, to a limited extent, from Japanese kanji as well (e.g., 新登場, pronounced in Cantonese as san ⁵⁵ dang ⁵⁵ coeng ²¹, ‘new arrival’, see example 24).

Second, the amount of lexical transference from English varies considerably by topics, with Text 2 (computer games) and Text 3 (jeans) inserted with considerably more English compared with Text 1a and Text 1b (book donation), Text 4 (detox product) and Text 5 (massage oil). This is consonant with Li’s (1996) observation that code-mixing tends to be domain- or topic-specific, with field-specific English jargon being more difficult to avoid in such domains as business, show business (‘show biz’), fashion, non-local films and TV productions, non-local food items, and miscellaneous products reflecting or indexing a modern lifestyle. It can be seen that being a metropolis where ‘East Meets West’, Hong Kong has always been receptive to technological innovations, business practices, international entertainers and artists, cultural novelties from popular culture to fine arts, as well as novel ideas for a modern lifestyle. All this is clearly manifested in the consumption of trendy fashion, good food, fine wine, cutting-edge IT gadgets, tantalizing cosmetic and health care products, and sundry lingua-cultural, multi-media consumables of western origin. These are arguably intimately related to the everyday lives of those Hongkongers who are plurilingual and pluricultural (Coste et al. 2009) in their socio-psychological orientation. To illustrate, in late April 2015, Apple Watches were launched in Hong Kong, making front-page news due to speculation. Where this news story is covered in several local Chinese dailies, Apple Watch is mixed into the HKWC text, resulting in ‘mixed code’. For example:

(28)	Apple Watch 忽 然 炒 起
	Apple Watch fat 55 jin 21 caau 35 hei 35
	Apple Watch suddenly speculate surge
	‘[There is a] sudden surge in speculation of Apple Watches.’ (headline, Headline Daily 25-4-2015, p. A1)

According to a photo featuring a handwritten notice board posted in a shopping mall, the most sought-after models are Sport 38mm and Sport 42mm, which are referred to in Headline Daily as:

Sport 版 Apple Watch
baan 35
‘Sport version Apple Watch’

It would be difficult to imagine how people could talk about such new products initiated in the English-speaking world (English as a native or an additional language) that hit the market every once in a while, if they were not allowed to use English. Compared with its flashy, translocal trademark in English (i.e., Apple Watch), a calque of that trademark and product like 蘋果手錶 (ping ²¹ gwo ³⁵ sau ³⁵ biu ⁵⁵) may be intelligible, but it would belong to a lower scale and ‘order of indexicality’, invoking images and associations that are blandly local (Blommaert 2010). As such, 蘋果手錶would be communicatively far less effective and, if used, might risk being heard as a joke largely because the translocal indexicality to that prestigious new product would be lost in the translation.³⁸ Conversely, to the extent that no Cantonese/HKWC equivalent is useable, to be able to index the referent directly by using the original brand name in English is arguably the most efficient and effective way to enact one’s plurilingual and pluricultural identity. This is preferred so long as no higher-order context-specific norms or regulations governing language use prevail (e.g., Chinese-medium class, Cantonese news broadcast), or when the speaker is (suddenly) aware that the English term in question may not be intelligible to the interlocutor(s). That much has been clearly attested in a number of ‘one day with no English’ or ‘one day with only Cantonese/Mandarin’ studies (Li 2011a, b; Li et al. 2010; cf. Li and Tse 2002): intended speaker meanings may be lost if Cantonese-dominant Hongkongers are prevented from using English in their language output, in speech or writing. The same may be said of Taiwanese student participants trying, in vain, to keep to ‘pure’ Mandarin by suppressing Minnan Hua (閩南話) or English in various contexts.

Third, as shown in (29), it can be seen that, from the structural point of view, practically all of the English elements are inserted at syntactic positions where corresponding Cantonese elements are expected (cf. Muysken 2000).

(29)		Cantonese with lexical transference from English (example cited above)		‘Pure’ Cantonese
	(i).	開心share	(13), (14)	開心分享 39
	(ii).	Facebook 專頁	(15)	臉書專頁40
	(iii).	喪打勁Game	(16)	喪打勁電子遊戲 41
	(iv).	睇埋friend 嘅打機實況	(17)	睇埋朋友嘅打機實況42
	(v).	做遙控mon	(18)	做遙控顯示器 43
	(vi).	買埋PS4™ 嘅game同download落去PS4™ 度添	(19)	買埋PS4™ 嘅電子遊戲同下載落去PS4™ 度添44
	(vii).	我衣櫃入面嘅must have item	(20)	我衣櫃入面嘅必有物品 45 (more colloquially: 我衣櫃入面一定要有嘅嘢)
	(viii).	個人style	(21)	個人風格 46
	(ix).	咁多個brand入面,我就最鐘意texwood	(22)	咁多個牌子入面,我就最鐘意蘋果牌 47
	(x).	成個長腿oppa咁!	(23)	成個長腿哥哥咁!48
	(xi).	每日都Feel good	(24)	每日都感覺良好 49
	(xii).	Keep住輕盈又Healthy	(25)	保持住輕盈又健康 50
	(xiii).	幾時都keep住健康啦。	(26)	幾時都保持住健康啦。51
	(xiv).	讓Body fit起來!	(27)	讓身體 健康起來!52

That is, in place of a Cantonese verb or adjective (e.g., 分享, 健康), an English verb or adjective (e.g., share, fit) is used; where a Cantonese object noun or noun phrase (e.g., 電子遊戲, 身體) is expected, the object position is filled by an English noun or noun phrase instead (e.g., Game, Body). Previous analyses of similar ‘code-mixing’ in social interaction, especially among young bilinguals, pointed to the speaker/writer intentionally trying to impress others by projecting a Hong Kong bilingual identity (e.g., Pennington 1998), one who is no ignorant bumpkin from the countryside but someone who is informed and up-to-date about the latest trends, fashion, cultural novelties and social practices among peers. Indeed, such an analysis is consistent with a Hong Kong-wide perception bordering on a stereotype, that a person (especially a new acquaintance) who appears to invoke English words in the middle of Cantonese indiscriminately may be perceived as a show-off, deliberately drawing attention to one’s western identity and, therefore, wants to be seen as modern, trendy, and fashionable – an identity enacted instrumentally through English so to speak.⁵³

As shown in (23), some of the non-Chinese expressions have no Chinese equivalents (e.g., the model of the computer game: PS4™, which is pronounced in English, and oppa in ‘長腿 oppa’ (coeng ²¹ teoi ³⁵ ou ²² paa ⁵⁵), the latter being the romanized form of the Korean word오빠, an intimate term of address for a girl’s male (usually older) lover. In Text 3, where ‘長腿 oppa’ is mentioned, reference is made to the Korean star Lee Minho (李敏鎬), who is popularly known to his Chinese fans by that nick-name (‘long-leg oppa’). Most of the other English expressions, if rendered into written Cantonese, would be either longer by up to three syllables (e.g., the Chinese equivalents of share, Game, friend, mon, style, brand, feel good, and fit), or sound too formal or semantically incongruent because they are Putonghua-based and thus belong to a different register (e.g., the Chinese equivalents of [computer] game, mon, and download). The preference for ‘Keep 住’ (keep zyu ²², ‘keep up’, ‘maintain’), which has the same number of syllables as the more formal-sounding保持 (bou ³⁵ ci ²¹), is arguably similarly motivated by a concern for the alignment of register (i.e., colloquialism).

By contrast, where ‘impression management’ matters, in that trendiness (i.e., being ‘in’ and savvy) are primary concerns in plurilingual interaction, being able to refer to the brand names of western products in English subtly projects an impression of the speaker/writer as someone who is ‘in the know’ and has sophisticated tastes. This type of sociolinguistic positioning appears to be enacted by the use of texwood in Text 3, which sounds translocal and is much better known and preferred in common parlance among Cantonese speakers than蘋果牌 (ping ²¹ gwo ³⁵ paai ²¹, ‘Apple brand’) and 牛仔褲 (ngau ²¹ zai ³⁵ fu ³³, ‘jeans’). For a similar reason, must have item (also Text 3) conveys a sense of principle and level of personal taste that would be too banal if calqued in stilted, Putonghua-based 必有物品 (bai ⁵⁵ jau ²³ mat ²² ban ³⁵) or, worse, in unsophisticated, colloquial Cantonese (一定要有嘅嘢, jat ⁵⁵ ding ²² jiu ³³ jau ²³ ge ³³ je ²³). Likewise, for a speaker/writer to subtly reinforce an identity of being an expert, using keywords in English that are intelligible to the reader is one convenient method. This appears to be the motivation of using Body and Healthy by the writer of Text 4, where a health product is promoted, as there is hardly any semantic loss or gain compared with their equivalents, 身體 (san ⁵⁵ tai ³⁵) and 健康 (gin ²² hong ⁵⁵), respectively.

The same may be said of the preference for Facebook rather than its Chinese equivalents among Hong Kong users. There are signs, however, that Chinese translations (e.g., 臉書 lim ²² syu ⁵⁵, 臉譜 lim²²pou³⁵, 面書 min ²² syu ⁵⁵) are becoming accepted in printed media. As of mid-2016, these equivalents of Mandarin or Putonghua origin still sound too formal when pronounced in Cantonese, but there is a good chance for one of these Chinese translations to gain community-wide currency eventually – much like it took years since the mid-1990s for din ²² jau ²¹ (電郵, ‘email’), the abbreviated form of din ²² zi ³⁵ jau ²¹ gin ³⁵ (電子郵件), to become naturalized in colloquial Cantonese.⁵⁴

2.4 Terms of Address: Lexical Transference in Colloquial Cantonese

Additional evidence of linguistic motivation behind lexical transference from English may be found in the use of code-mixed terms of address. This may be illustrated with one widely publicized news story. In March 2015, at a community award ceremony, the Chief Secretary Mrs. Carrie Lam was quoted as saying that she was a fen ⁵⁵ si ³⁵ (fan 屎, ‘fans’) of the Hong Kong Police Force. Of the many terms of address she was used to hearing, including those associated with her previous civil service positions, none pleased her more than ‘Madam’, the salutary address to a female officer in the Hong Kong disciplinary forces (e.g., police, immigration, and correctional services). The Chief Secretary was quoted as saying:

(30)	「最開心喺街上見到前線警務人員稱呼我為『Madam』,因為我會感覺到自己係呢一支引以為傲嘅警隊嘅一分子。多謝你!各位阿Sir、Madam!」 55	‘What pleases me most is to see front-line police personnel addressing me as ‘Madam’, because that makes me feel that I’m a member of the police force [that we are] so proud of! Thanks to you all, Sirs and Madams!’
	(明報新聞網, 28-3-2015)	(Ming Pao News Net, 28-3-2015)

There are two points of linguistic interest in this news story, both related to the plurality of countable English nouns transferred into Cantonese. First, despite the subject being singular (the Chief Secretary), fans appears in plural form (i.e., a fen ⁵⁵ si ³⁵, usually rendered as ‘fan 屎’ in writing). On the other hand, the last phrase in this quotation is clearly a collective term of address to all male and female police officers, present or in absentia: 各位阿 Sir、Madam!(‘Fellow Sirs and Madams!’), and yet the singular forms were preferred. These patterns, all bisyllabic (fen ⁵⁵ si ³⁵, 阿 Sir, Madam), may be accounted for by the Cantonese-specific ‘bisyllabic constraint’, especially for nouns (see below).

Given that the training of new officers in various disciplinary forces is conducted in English, a legacy of British colonial practice and, under the ‘one country, two systems’ postcolonial arrangement, the widespread use of ‘Sir’ and ‘Madam’ as salutary terms of address in the SAR is understandable. To my knowledge, no corresponding Chinese address terms, spoken or written, have ever been used for that function in Hong Kong; they are therefore part and parcel of the socialization process of becoming members of the ‘discourse system’ (Scollon and Scollon 1995) of the SAR disciplinary forces in question. What is interesting is that a similar pair of terms of address is commonly used for teachers in the education domain, especially from secondary level onwards, a practice which is more typical of staff and students in English-medium schools, but also in Chinese-medium schools to some extent (Table 2.4).

Table 2.4 Mixed terms of address for teachers in Hong Kong schools

Notice that in speech, these terms of address appear to be subjected to a ‘bisyllabic constraint’, which holds that monosyllabic units (morphemes or names) are preferably adorned with an appellation prefix or suffix (Li et al. 2015; cf. Luke and Lau 2008), while units longer than two syllables tend to be reduced to two. Accordingly, bisyllabic terms of address are commonly heard and found in the informal sections of the Chinese press (e.g., 阿 John, 阿 Mark, 阿 Bill, 阿 Jack, 阿 Jane, 阿 Kate, 阿 May, but not *阿 Peter or *阿 Janice [Peter, Janice preferred]; Benjamin is either pronounced in three syllables, or similarly clipped to two as 阿 Ben). Probably due to the bisyllabic constraint, ‘Sir’ is seldom used in isolation; rather, it is prefixed by aa ³³ (亞 or阿), hence aa ³³ soe ²¹. The same constraint explains why the general term of address for female teachers is mi ⁵⁵ si ²¹ or mit ⁵⁵ si ²¹ (more commonly heard than monosyllabic ‘Miss’), while the specific one for a teacher surnamed Lam is Ms. Lam (rather than *mi ⁵⁵ si ²¹ Lam). It is further noteworthy that the formulas of the specific code-mixed terms of address have different word orders depending on the gender: the formula for female teachers follows the word order ‘[Miss] + [Surname]’, while the word order of the formula for male teachers is the reverse: ‘[Surname] + [Sir]’. This is probably because ‘[Sir] + [Surname]’ is not an option, for in British English, Sir, like Lord, is conventionally used to signal knighthood when prefixed to a name (compare, e.g., Sir Run Run Shaw, also known as Sir Run Run; Sir Ti-Liang Yang, or Sir Ti-Liang in English).

The terms of address for teachers in Table 2.4 were already commonly used in the 1970s. While I am not aware of any study or discussion of their origin, the following hypothesis seems plausible: when first used, these mixed terms of address were probably intended for disambiguation purposes, for the corresponding specific Chinese terms of address are gender-neutral. For instance, Lam ²¹ lou ²³ si ⁵⁵ (林老師, ‘Teacher Lam’) may be used to address (2nd person) or refer to (3rd person) a male or female teacher. When there are two or more teachers of opposite genders surnamed Lam in the same school, the mixed terms of address as shown in Table 2.4 may conceivably serve a quick and effective identification purpose. Certainly, the speaker/writer may choose to use the Chinese teacher’s full name (typically consisting of three syllables in Hong Kong), but that would be considerably longer (e.g. Lam ²¹ gin ³³ man ²¹ lou ²³ si ⁵⁵, 林建民老師, ‘Teacher Lam Kin Man’), and so less preferred probably for that reason. The hypothesis outlined above suggests that the ‘code-mixed’ address formula probably began with specific terms of address, referring to Chinese teachers with specific surnames. This formula was subsequently extended to include an address formula for general purposes (i.e., without a surname) following the bisyllabic constraint pertaining to Hong Kong Cantonese. In other words, far from being arbitrary, these mixed address formulas were linguistically motivated for disambiguation when used by the first plurilingual teachers, before catching on Hong Kong wide, including in informal genres of public discourse (like the case of secretive triad language, Canto films featuring the school context must have helped popularize the mixed terms of address for teachers in this regard). Support for the above hypothesis is partly evidenced by the fact that other ‘address-sensitive’ English nouns such as teacher and principal are seldom, if ever, transferred into Cantonese (e.g., *Lam Teacher, *林 Teacher, *Teacher 林, *Teacher Lam; *Principal 王, *Principal Wong, *王 Principal, *Wong Principal). On the other hand, while Panel (short for ‘Panel Chair’, the teacher-in-charge of a school subject, e.g., English, Chinese, Liberal Studies) is transferred into Cantonese (pen ⁵⁵ nou ³⁵ or pen ⁵⁵ lou ³⁵, ‘panel’), it is seldom accompanied by an appellation affix.

2.5 Code-Switching, Code-Mixing, Translanguaging, Translingual Practice

Plurilingual interaction is among the most actively researched language contact phenomena to date (see, e.g., Chan 2008, 2009; W. Li 1994, 2002, 2005; Myers-Scotton 1993a, b, 2002; Muysken 2000), typically based on analysis of naturalistic speech data involving language pairs that belong to typologically unrelated language families (e.g., Chan 2009). Code-switching (CS), by far the most widely used term, may be defined as “the alternating use of two languages in the same stretch of discourse by a bilingual speaker” (Bullock and Toribio 2009, p. xii). Some scholars prefer to adopt CS as an umbrella term and make a distinction between switches at clause boundaries: inter-sentential CS, and switches within a clause: intra-sentential CS (e.g., Kamwangamalu 1992; Myers-Scotton 2002). Others prefer to use CS as a generic term to cover both intra-sentential and inter-sentential switches (e.g., Chan 2008; Clyne 2003). Still others prefer to speak of ‘code-alternation’ (e.g., Auer 1995), while a few insist on using ‘code-mixing’ throughout (e.g., Muysken’s 2000 monograph Bilingual speech: A typology of code-mixing).

As García and Lin (in press) have observed, echoing Grosjean (1989),⁵⁶ the term CS reflects earlier scholarly attempts to characterize and understand CS from a largely monolingual, monoglossic perspective (e.g., Auer 2005; Gumperz 1982; Myers-Scotton 2002; Weinreich 1953/2011). Following Bakhtin (1935/1981; cf. Bailey 2012), García and Lin (in press) argue that CS should give way to translanguaging, which is much better suited as a theoretical construct for capturing the dynamic nature of plurilingual interaction involving any language varieties, including bilingual interaction in the classroom context:

Code-switching, even to those scholars who see it as linguistic mastery (…), is based on the monoglossic view that bilinguals have two separate linguistic systems. Translanguaging, however, posits the linguistic behavior of bilinguals as being always heteroglossic (…), always dynamic, responding not to two monolingualisms in one, but to one integrated linguistic system. It is precisely because translanguaging takes up this heteroglossic and dynamic perspective centered on the linguistic use of bilingual speakers themselves (…) that it is a much more useful theory for bilingual education than code-switching. (García and Lin, in press, p. 3)

The monolingual, monoglossic perspective mirrors a popular belief in multilingual societies that CS is linguistically anomalous, or even pathological, reflecting the plurilingual speaker/writer’s inability to maintain watertight language boundaries, in speech or in print. Such a perception often gives rise to feelings of shame on the part of ‘code-switchers’, out of a concern for ‘failing’ to resist or suppress CS. As a correlate of strong social disapproval in many communities where switching between two or more languages is commonplace, CS tends to attract a bad name or pejorative label, for example, Spanglish, Tex-Mex (Spanish-English), Franglais (French-English), Bahasa Rojak (Malay-English), Taglish (Filipino-English), Japlish (Japanese-English), Konglish (Korean-English) and Hongish (Hong Kong English). For a long time, the Singaporean government has encouraged people to speak ‘good English’ and refrain from using Singlish, a ‘low’ sociolect (or ‘basilect’ in the lectal continuum) used by Singaporeans – educated and uneducated alike – for signaling shared ethnolinguistic or national identity. CS in Hong Kong is no different in this regard; it is commonly and apologetically referred to as mixed code or Chinglish (i.e., ‘half Chinese, half English’), reflecting community-wide perceptions as well as disapproval of the seemingly ‘random’ and ‘disorderly’ mixing of languages. Until recently, most of the researchers engaged in analyzing plurilingual speech data in Hong Kong have preferred using the term CS (e.g., Li 1996; Li and Tse 2002; Lin 1996, 2006; Lin and Li 2012; Lin and Man 2009), largely to avoid aggravating the marked, society-wide perception of CS being a product of unprincipled ‘mixing’,⁵⁷ which in turn is strongly suggestive of the speaker’s or writer’s failure to keep to a ‘pure’ language.

In the past three decades since the mid-1980s, there has been a lack of consensus regarding the terminology used to describe or categorize specific plurilingual communication phenomena. Divergence of definitions and the absence of clear delineation of such terms as CS, CM, and borrowing often gave rise to different interpretations of the same plurilingual interaction data, making it difficult to reconcile theory-driven and context-sensitive analyses across diverse datasets involving different language pairs (or, increasingly, trios), which in turn makes overarching generalization difficult to reach. In addition, various approaches to analyzing plurilingual speech data and explanatory models have been advanced, with different theories premised on specific ‘researcher categories’ (as opposed to ‘code-switcher categories’) competing for ascendency. In general, researcher categories are those theory-driven constructs that are held to be relevant and valid in support of a preferred analytical framework or explanatory model (i.e., to offer a coherent account of “what’s going on?”). There are two main theoretical approaches to date: conversation analysis (e.g., Auer 1995; W. Li 1994, 2002, 2005; W. Li and Milroy 1995) and the Markedness Model (e.g., Myers-Scotton 1993b; Myers-Scotton and Bolonyai 2001).

With the help of ample CS data in East Africa, notably multilingual Kenya, Myers-Scotton (1993b) demonstrates how, in transactional communication involving identity checking or negotiation, typically featuring a speaker vested with institutional power vis-à-vis a stranger (e.g., a university gate-keeper checking visitors’ identities; a bus conductor verifying passengers’ bus fares), switching from a local vernacular to Swahili or English is an effective way to index one’s ethnolinguistic group membership or social attributes. Myers-Scotton (1993b) provides plenty of instructive illustrations showing how CS, being socially motivated in contexts where speaker identities are negotiated or contested, is a useful communicative resource that may be deployed – a ‘rational choice’ so to speak – to optimize a plurilingual speaker’s communicative intent. Conversation analysts, on the other hand, insist that any attribution of specific speaker meaning to CS can only be established through meticulous sequential analysis of various conversational cues, including suprasegmental features (volume, pitch, pace of delivery, etc.) and the duration of pauses, if any. This is why, as a prerequisite, any speaker meaning attributed to a code-switcher must be based on carefully transcribed conversational segments following a rigorous transcription protocol (ten Have 2007). W. Li and Milroy (1995), for instance, show that in plurilingual interaction between a Cantonese-dominant mother in a Chinese community, UK and her British-born daughter over the dinner table, the latter (Ah Ying) signals dispreference (i.e., expressing reluctance) by responding to her mother’s question in a different language:

(31)	(Dinner table talk between mother A and daughter B.)
	A:	Oy-m-oy faan a? Ah Ying a?
		(Want some rice?)
	B:	(no response)
	A:	Chaaufaan a. Oy-m-oy?
		(Fried rice. Want or not?)
	B:	(2.0) I'll have some shrimps.
	A:	Mut-ye? (.) Chaaufaan a.
		(What?) (Fried rice.)
	B:	Hai a.
		(OK.)	(W. Li and Milroy 1995, pp. 287–288)

From this excerpt, it can be seen that Ah Ying’s (B’s) dispreference or indirect refusal in her second turn is doubly marked:

B’s indirect refusal is marked in two steps – first a two-second delay before delivery, a commonly occurring signal of an imminent dispreferred response; then the choice of English which contrasts the code choice in the immediately preceding turn by the mother. The child’s final acceptance of the mother’s offer of rice is in Chinese [Cantonese], which corresponds to the language choice of the mother, but differs from the one she has used to mark her indirect refusal. (W. Li and Milroy 1995, p. 288)

Linguistic and paralinguistic resources in bilingual conversation being seen as potentially indexical of speaker meanings, adherents of CA believe that all interpretive analysis must be grounded in the dynamic turn-by-turn sequence. This is why rigorous transcription protocols are followed and high standards are set for the coding of speech data.⁵⁸ In general, CA data processing:

requires repeated examination of detailed transcripts of audio and, when available, video, recordings. The data collection and transcription process can be very time-consuming, depending on many factors, including the level of detail of the transcript (e.g. Are changes in gaze and body position noted? Are overlapping talk, latching, breathing, pauses, laughter, etc. noted?), the number of participants involved (…), and the linguistic repertoire of the researcher (...). (Cashman 2008, p. 290)

Both the Markedness Model and the CA analytical frameworks have enhanced our understanding of some of the typical motivations behind CS. Owing to divergent theoretical orientations and a lack of standardized terminologies, however, the role of CS in plurilingual interaction – how it impacts on the lexico-syntactic structures of the languages in question, the dynamic, socio-pragmatically sensitive meaning-making potential of moment-by-moment speaker concerns or motivations – is still being debated. There are other researchers who refuse to see these two approaches as being mutually exclusive; rather, it may be demonstrated that pre-existing social structures such as gender, race, religion, ethnicity, professional or institutional identity are brought to bear in bilingual interaction whereby ‘identity-in-interaction’ is dynamically co-constructed (Gafaranga 2005; cf. Cashman 2008, p. 292). In general, the stronger the evidence of negotiation of identity being an interactional focus (social motivation, e.g., between shopper and salesperson at a department store selling luxury items; between doctor and patient during a consultation session; between police officer checking a person’s identity cards), the more likely it is for language choice to be bound up with context-specific speaker meanings or functions.

2.5.1 Social Motivation: Negotiation of Identity

Blommaert (2010, cf. 2005, pp. 203–204) has argued convincingly that, to the extent that language communication in any context invariably indexes the speaker’s or writer’s social attributes vis-à-vis those of his or her interlocutor(s), social interaction necessarily amounts to an ‘act of identity’ (Le Page and Tabouret-Keller 1985). The enactment of speaker identity through language choice – social motivation in short – is most relevant when there is evidence in bilingual interaction where the speaker’s language choice, deliberately or involuntarily, indexes association with specific social attributes of particular target groups, and that such symbolic associations are contested verbally, suggesting that negotiation of identity is foregrounded. Negotiation of identity is especially commonplace in transactional communication between speakers whose social roles are marked by a clear power differential, for example, when a person’s identity is challenged by a police officer or gate-keeper of an institution, or when a customer with a deep pocket feels the salesperson’s service is just not good enough (cf. Myers-Scotton 1993b).

In extreme cases, language choice in multilingual environments can be a matter of life and death. For instance, Blommaert (2010) provides a detailed analysis of the life story of a plurilingual young boy called Joseph, who grew up in crisis-ridden Rwanda in Africa. Born to middle-class parents, a Tutsi mother and a Hutu father in Rwanda in 1986, Joseph was brought to Kenya by his mother and studied in an English-medium school there. As a child he also picked up some Swahili. When he was five years old, his mother took him back to Rwanda, where he learned some Kinyarwanda from a Hutu servant. Throughout his childhood, his English-speaking parents insisted that he speak only English and discouraged him from mixing with other children who spoke the local languages. This is why and how Joseph grew up to become English-dominant. Shortly after his return to Rwanda, his mother was murdered and, six months later, his father was also killed and the house was burned down. Joseph managed to flee and find his uncle who lived in the Democratic Republic of Congo (DRC), where he picked up some Runyankole (‘Kinyankole’), which he would speak with his uncle, who, in addition to English, was also conversant in Kinyarnwanda and French, the “normative, standardized, and literate” languages in Rwanda (p. 167). Joseph’s childhood ordeal predated the brutal ethnic warfare between the Tutsi and Hutu culminating in the widely reported Rwanda genocide in 1994.

Fate had it that Joseph landed in UK, where he sought asylum at the age of 14. His application was rejected on linguistic grounds, however. As Blommaert (2010) explains, Joseph’s language choice in response to British immigration personnel’s critical interrogation was interpreted based on their static view of the political geography of central and eastern Africa, a view which is conservative and completely out of place in an increasingly globalized world. When assessing Joseph’s application for asylum, the British immigration officers showed no sensitivity to his actual life circumstances, in particular the virtual absence of opportunities for proper schooling and therefore his inability to display any knowledge of standardized, literate forms of Kinyarwanda or French. Blommaert comments on the relationship between Joseph’s “thoroughly distorted conditions of life” (p. 156) and his ‘truncated’ linguistic repertoire as follows:

Joseph also appears to be quite aware of the indexical values of some of these languages: English sets them [speakers of English] apart and suggests a superior level of ‘civilisation’ (…). Runyankole suggests an identity as a foreign Hutu rebel (…), and he himself has very negative attitudes towards that language (…). Runyankole, in the crisis-ridden Rwandan context in which his story is set, naturally signalled enemy identities to those [especially Kinyarwanda speakers] whom he encountered on his way. (Blommaert 2010, p. 169)

This critical linguistic awareness made Joseph very cautious of language choice when caught in chance encounters with strangers from different ethnolinguistic backgrounds, especially the national language Kinyarwanda because it “may in itself be an expression of political allegiance” and that “in circumstances of violent conflict [it may] require dissimulation or denial for one’s own safety” (Blommaert 2010, p. 167). Joseph’s ordeal, as described and discussed in Blommaert’s (2010) critique, epitomizes the intimate link and symbolic relationship between language and identity in crisis-ridden multilingual settings awash with human miseries due to “war refugeeism” (pp. xi–xii), where the local and possibly transnational languages are perceived as indexing one’s friends or enemies, and where language choice or verbal performance is strongly indexical of the kind of person one is interacting with. It also brings home an important insight brought about by eminent scholars like Bourdieu (1991), Bernstein (1971) and Dell Hymes (1980, 1996), as Blommaert points out:

[T]he world of language is not just one of difference but one of inequality; that some of that inequality is temporal and contingent on situations while another part of it is structural and enduring; and that such patterns of inequality affect, and articulate around, actual, concrete, language forms such as accents, dialects, registers and particular stylistic (e.g. narrative) skills. (Blommaert 2010, p. 28)

In more mundane, urban multilingual environments, identity negotiation is also clearly evidenced when a plurilingual person is trying to make a complaint on the phone, where all the information pertaining to the complainant can only be deduced from the speaker’s voice and other language-related cues. In a context like multilingual Hong Kong, it is well-known that when a complaint is made by phone, those who speak fluent English with a native-like accent are more likely to be taken seriously (e.g., public utilities companies like telephone or electricity services or financial institutions like banks and credit card companies). In the absence of any evidence of speaker identities being negotiated (typically between peers), and when communication is content-focused, the reason for invoking English may lie elsewhere. If we compare the ‘pure’ Cantonese version of the text segments mixed with English in (29), some of the linguistic motivations for using English will become clear (cf. Li 1996).

2.5.2 Linguistic Motivation: From Code-Switching to Translanguaging and Translingual Practice

When a plurilingual speaker/writer is absorbed in meaning-making, and provided negotiation of identity is not foregrounded in plurilingual interaction, all the language varieties, accents, and registers within that speaker/writer’s repertoire are treated as a composite pool of semiotic resources to make meaning. This is the background against which the term CS is increasingly felt to be inappropriate as it unduly underscores and reinforces a monoglossic ideology or bias, as if bilinguals were “two monolinguals in one person” (Grosjean 1989; cf. ‘monoglot ideology’, Blommaert 2005; García and Lin in press). Further, it has been observed that there is hardly any limit to the speaker’s or writer’s creativity and criticality in the moment-by-moment decisions of language choice and plurilingual performance (W. Li and Zhu 2013), the only constraint being an awareness of whether the linguistic repertoire is matched by or shared with that of the interlocutor(s), which in turn informs the appropriacy of fleeting language choice decisions, especially in dynamic contexts where change in the configuration of participants is unexpected and difficult to predict (e.g., at a cocktail party involving plurilinguals on the move, hopping between loosely formed groups of plurilingual party-goers).

Another problem is related to the use of the word ‘code’ to refer to those highly salient linguistic practices. While CS and CM makes us think of ‘switching’ or ‘mixing’ as unusual or marked, and thus in need of explanation, the choice of ‘code’ in both terms is increasingly felt to be out of place by virtue of its meaning and use in other collocations. W. Li (2011) provides an instructive example in this regard. In his qualitative study of a small network of transnational Chinese university students in London, one of the informants (Chris) who characterized himself as a “heavy code-mixer” questioned why that everyday linguistic practice he engaged in came to be called “code-mixing”:

I mix Chinese and English openly; have to, really. No secret about it. . . . Why is it called code-mixing? Is it some secret message? (W. Li 2011, p. 1229)

Chris’s objection to the term ‘code-mixing’ was probably guided by the common collocation ‘secret code’, as in Morse Code and Da Vinci Code, whence the query about the apparent link between ‘code-mixing’ and secrecy.

In addition, recent breakthroughs in neuroscience research have also called into question whether ‘switching’ and ‘mixing’ are the right metaphors. There is empirical evidence showing that the languages in a bilingual brain remain activated and cognitively accessible even though only one language is used (Hoshino and Thierry 2011; Thierry and Wu 2007; Wu and Thierry 2010; cf. Lewis et al. 2012, p. 643; Paradis 2004). This suggests that bilingual speech production is too dynamic to be characterized as ‘switching’ or ‘mixing’, and too simplistic to do justice to the “spur-of-the-moment” creativity of bilingual interactants (W. Li 2011). All this helps explain a gradual convergence of views about the need for more adequate and appropriate terminology.

Canagarajah (2013a, b) proposes the term ‘translingual practice’ to dispel the monolingual bias which is historically entrenched in Anglo-European modernity and colonialism of the past four centuries (cf. García and Lin in press; Lin and Li 2015), and draws attention to the translingual nature of writing performance as the unmarked state of what is elsewhere referred to as bilingual or multilingual literacies. Thus when the primary objective of the plurilingual writer is to perform pluri-literacies (e.g., Arabic, English and French when writing one’s literacy autobiography, Canagarajah 2013a, pp. 1–2), such an objective would clearly be defeated if carried out monolingually and solely in English (narrative). To Canagarajah, however, the term ‘translingual’ is not limited to the flouting of boundaries between more or less discrete language varieties, for translingual practice is often embedded in what is traditionally conceived of as monolingual output in writing or speech. This happens, for example, when individual writers engaged in various literacy activities draw on a range of styles, genres and registers within the same language to achieve various communicative purposes, typically guided by and in response to an inward call for linguistic creativity that knows no boundaries (cf. ‘code-meshing’).⁵⁹ As for translingual speech performance, Canagarajah (2013a) observes that:

In a specific speech event, one might see the mixing of diverse languages, literacies, and discourses. It might be difficult to categorize the interaction as belonging to a single language. Languages are treated as resources and used freely in combination with others for people’s communication purposes. (Canagarajah 2013a, p. 40)

In short, with the term ‘translingual practice’, Canagarajah challenges the assumption embedded in traditional terms like bilingual or multilingual literacies that speakers or writers should adhere to, as if there were solid or stable boundaries between discrete languages within their linguistic repertoire. Rather, he demonstrates that since antiquity, translingual practice is intrinsic to all human communicative activities, beyond those contexts which are traditionally labeled as bilingual or multilingual. The conceptual thrust of translingual practice echoes Bakhtin’s (1935/1981) critique of monolingual ideology using the term ‘heteroglossia’ decades earlier, as Lin and Li (2015) observe:

Like translingual practice, the notion of heteroglossia focuses on breaking away from the ideology of discrete, unitary languages and breaking through the centralizing forces driven by ideologies of monolingualism and linguistic purism that are dominant in the literature of language education and government language education policies. (Lin and Li 2015, p. 82; cf. García and Lin in press; Lin 1996, 2006)

Key to Bakhtin’s notion of heteroglossia (literally ‘different voices’) is that words invariably index signs of social worlds, past and present, while the use of words in any contexts unavoidably echoes multiple voices embedded in the myriad genres and contexts where they have been used (cf. Bailey 2012, p. 506). Similarly, Blommaert (2010) argues cogently that ‘code-switching’ is largely an artefact of ‘the Saussurean synchrony of language’, an influential construct since the dawn of modern linguistics about a century earlier albeit with no social reality, which should therefore be abandoned and replaced with the ethnographic concept of ‘voice’, referring to how people actually deploy their linguistic resources when making meaning in context (p. 180). Seen in this light, what is generally referred to as ‘code-switching’ is more fruitfully re-conceptualized as “moments of voice in which people draw resources from a repertoire that contains materials conventionally associated with ‘languages’,” reflecting thereby “heteroglossic practices in which different voices are being blended” (Blommaert 2010, p. 181). As a default mode of communication, heteroglossia is at work in the speech of monolinguals or plurilinguals alike. Put differently, to speak is to be engaged in ‘heteroglossic speech’ (p. 181). To get at the meanings of individual speakers/writers, therefore, rather than belaboring which languages individual words or signs belong to, it is more fruitful to understand the “social tensions and conflicts between these different signs and voices” (Lin and Li 2015, p. 82), and what additional meanings are carried and instantiated in these voices.

A similar conceptual reorientation has been the concern of other scholars, who prefer using the term ‘translanguaging’, albeit with different emphases. Originally used in the classroom context as a bilingual pedagogical practice in Wales where students are guided to attend to language input (i.e., reading and listening) in language A, and use that input to generate output (i.e., speaking and writing) in language B (Williams 1996; cf. Cummins 2008 with regard to the Canadian context), translanguaging has been used somewhat differently depending on the scholar. In her monograph Bilingual Education in the 21 ^st Century, García (2009, p. 45) defines translanguaging as “multiple discursive practices in which bilinguals engage in order to make sense of their bilingual worlds” (emphasis in original), while Baker (2011, p. 288) refers to “the process of making meaning, shaping experiences, gaining understanding and knowledge through the use of two languages”. For W. Li and Zhu (2013, pp. 519–520), the scope of meaning-making potential of translanguaging is organized around the prefix trans-, whose semantic spectrum is extended to cover “three dimensions of flexible and dynamic multilingual practices” (p. 519; cf. ‘multi-competence’, Cook 1991, 2012):

• I. ‘trans-system/structure/space’, including across modalities such as speaking, writing and singing;

• II. ‘transformative’, encompassing the dialectic relationships between attitudes, beliefs and identity formation; and

• III. ‘transdisciplinary’, reflecting the holistic nature of plurilingual performance which is at the same time informed and produced by social practices. (W. Li and Zhu 2013, p. 519)

W. Li and Zhu (2013) have demonstrated that, to capture and fully account for the creativity and criticality so typical of plurilingual interaction, all three dimensions need to be addressed. Inspired by the scholars whose contributions were briefly reviewed above, I will use the term translanguaging to refer to speakers’ or writers’ use of linguistic resources that are traditionally categorized as belonging to different languages, varieties, more or less distinct genres, registers, and styles. Following W. Li and Zhu (2013), the emphasis on trans- is meant to capture the holistic multi-system, multi-modality, multi-identity and multi-disciplinary nature of the linguistic performance when plurilinguals are engaged in communicative meaning-making, in writing or in speech. As such, translanguaging is viewed as a natural extension of ‘languaging’, whether the speaker/writer in question is monolingual or plurilingual, regardless of the actual number of natural languages and competencies within his or her repertoires. When reference is made to individuals’ translanguaging collectively as social practice, the term translingual practice will be used.

For our purpose in this book, therefore, the conventional meaning of bilingual is subsumed under plurilingual, in that it refers to speakers/writers who have two or more languages within their repertoires (cf. ‘multi-competence’, Cook 1991, 2012), which are typically ‘truncated’ rather than ‘complete’, including our ‘mother tongue(s)’:

No one knows all of a language. That counts for our so-called mother tongues and, of course, also for the other ‘languages’ we acquire in our life time. Native speakers are not perfect speakers. (…) And there are always [linguistic, semiotic] resources that we do not possess. (Blommaert 2010, pp. 103, 105, emphasis in original)

Also included in our truncated repertoires is language-specific awareness of more or less distinctive styles (e.g., formal vs. informal), genres and registers pertaining to each of the conventionally defined languages, which evolved as a function of “our biographies and the wider histories of the communities” in which we have lived (Blommaert 2010, p. 105). Except when there is a need to actively monitor one’s language output and to observe context-specific norms of appropriacy, typically where the ‘crossing’ (Rampton 1995) or ‘mixing’ of languages is socially disapproved (e.g., news broadcast, speech delivered at a formal ceremony), the following premises are taken to be axiomatic when a plurilingual is engaged in meaning-focused interaction with other plurilinguals with a similar language profile:

• (a) Speakers/writers draw on all their linguistic resources in more or less discrete language varieties, styles, genres, and registers within their repertoire;

• (b) Speakers/writers have a low awareness of boundaries between language varieties, styles, genres, and registers, and feel minimally constrained by them;

• (c) Speakers/writers expect their translanguaging to be understood by their interlocutor(s), and that it will be reciprocated.

These premises are consistent with the Council of Europe’s characterization of plurilingual and pluricultural competence, which refers to:

the ability to use languages for the purposes of communication and to take part in intercultural interaction, where a person, viewed as a social actor has proficiency, of varying degrees, in several languages and experience of several cultures. This is not seen as the superposition or juxtaposition of distinct competences, but rather as the existence of a complex or even composite competence on which the social actor may draw. (Coste et al. 2009, p. v)

As such, plurilingualism “is focused on the fact that languages interrelate and interconnect particularly, but not exclusively, at the level of the individual. It stresses the dynamic process of language acquisition and use, in contrast with coexistence and balanced mastery of languages” (Piccardo 2013, p. 601; cf. Council of Europe 2014). When interacting with others, plurilinguals typically draw from whatever linguistic resources within their truncated repertoires, often resulting in translanguaging, with transference as a natural outcome at different linguistic levels, contributing thereby to their context-specific communicative purpose and goal (Clyne 2003).