‘We Want Fork but No Pork’: (Im)Politeness in Humour by Asian Users of English as a Lingua Franca and Australian English Speakers


This study explores the conversational humour of Asian multilinguals using English as a lingua franca (ELF) – specifically, their use of (im)politeness strategies to humorously maintain, neglect or affront their target’s face. The Asian ELF data come from the Asian Corpus of English (ACE), which comprises naturally occurring data of English being used as a lingua franca by Asian multilinguals. These data are compared with similar instances drawn from existing Australian humour studies. Our qualitative analysis reveals significant insights into how (im)politeness (both actual and pretended) are utilised as humour strategies by Asian ELF speakers compared with Australian English users.


Introduction
Although the study of conversational humour is not new, most extant studies have examined humour in first languages rather than in lingua franca contexts. And most lingua franca humour studies that do exist are concentrated in Europe or Scandinavia; with a few exceptions (e.g. Walkinshaw, 2016 on teasing), studies of conversational humour among Asian speakers of English as a lingua franca (ELF) are thin on the ground. Still more uncommon is research contrasting pragmatic aspects of lingua franca humour with those of firstlanguage humour, which is the focus of the current research project. We look specifically at producers' orientation to (im)politeness in conversational humour, specifically their attention to the positive and negative face of the targets of their humour and other involved participants. Our study explores a data set drawn from the Asian Corpus of English, a million-word spoken corpus of English being used as a lingua franca by Asian multilinguals spanning a variety of social, professional and geographical contexts. ACE is freely accessible at http://corpus.eduhk.hk/ace/ (ACE, 2014).
In this paper we report on the findings from the ACE data set, contrast them with findings from a range of existing studies of Australian English humour, and suggest some reasons for the differences. Although the scale of this research is limited, it offers preliminary insight into the nature of humour in English as a lingua franca in Asia which is, we believe, fertile ground for further exploration.

Humour in English as a Lingua Franca
In this segment we briefly define conversational humour and explore its interface with face and (im)politeness. We then define English as a lingua franca and outline some of humour's identified functions in that sphere, as well as discussing how humour is accomplished in ELF contexts. Finally, we set out the categories of humour that emerged from our analysis of the ACE data set. Due to space restrictions our treatment of these topics is necessarily brief, serving as a definition and overview rather than a comprehensive literature review.

2.1
Defining Conversational Humour Conversational humour is not normally considered a genre or activity as such. Rather it is realised though a frame or key (Dynel, 2009(Dynel, , 2011b which is selectively activated in spoken interaction for the sake of amusing the producer and at least some (though not necessarily all) other interactants. It is spontaneous, context-bound, and not normally recyclable. (An exception is conventionalised stock conversational witticisms (Norrick, 1984) such as as clear as mud.) Humour has entertainment value, but as Hay (2000) points out, it often serves other pragmatic functions too: increasing solidarity, maintaining or subverting power, or attenuating face threatening acts (Shardakova, 2012). Humorous intention may be identified through contextualisation cues (e.g. laughter particles, changes in intonational pitch or volume, elongation) which can signal the speaker's intent to project humour and support the hearer's reception of the utterance within a humorous frame (Holmes and Marra, 2002).
As this is a comparative study of the use of humour, it is important to consider whether humour is universal. Goddard (2018) has noted that words such as joke and tease are English -specific and have no direct translations even into closely related languages such as French and German. He therefore suggests describing humour as a research field "that studies culturally shaped situations in which people laugh or can laugh with good feeling in response to something someone says or does" (2018: 505). In their study of the differences between Chinese and Canadian students' perspectives on humour, Yue, Jiang, Lu and Hiranandani (2016) argue that humour is a universal but that people perceive and use it differently. Their findings suggest that Chinese value humour (e.g. in a comedy act) but not as an individual's personality trait. In contrast, the Canadians in the study rated humour as significantly more important than the Chinese subjects and perceived themselves as being more humorous than the Chinese. Our data below present examples of humour being used by native speakers of Australian English and Asian multilinguals using English as a lingua franca, and we contrast how each uses humour as a pragmatic strategy to advance, maintain or neglect the face-wants of its target/s and other present and absent recipients.

2.2
Face and (Im)politeness in Humour A brief definition of politeness, impoliteness and face is in order. According to Brown and Levinson (1987), some acts (e.g. criticism) are inherently threatening to another's face, defined by Goffman (1967: 5) as "an image of self delineated in terms of approved social attributes". Politeness is constructed by employing strategies to reduce the threat to the recipient's face, and so minimise discord. Conversely, impoliteness pursues the opposite aim: social disruption or damage to a person's self-identity by means of face-affront, faceneglect or face-attack (Culpeper, 1996(Culpeper, , 2011. Our analysis investigates producers' attention to or neglect of the positiveface and negative-face wants (Brown and Levinson, 1987) of targets or other recipients. Positive face refers to "the want of every member that [their] wants be desirable to at least some others …[including] the desire to be ratified, understood, approved of, liked or admired" (1987: 62). It is operationalised through positive politeness strategies, wherein a face-threatening act is performed in a manner that supports the recipient's positive face, e.g. by using in-group identity markers such as mate, seeking agreement, assuming common ground and so on. Conversely, negative face refers to "the want of every competent adult member that [their] actions be unimpeded by others" (Brown and Levinson, 1987: 62). It is attended to through negative politeness strategies which acknowledge or redress any impingement or inconvenience to the recipient, such as apologising, hedging, providing reasons for the imposition and so on. Brown and Levinson's (1987) model permits us to analyse humour as an (im)politeness strategy in interpersonal interactions, and to examine some of the pragmatic devices by which (im)politeness is accomplished.
Of course, such analysis is less straightforward than it first appears, because recipients may be unsure whether the humour is meant as entirely jocular, or whether there is a malicious undercurrent to it. Indeed they may evaluate mock-impoliteness as conveying implicit aggression (Culpeper, 2011), even if they realise that malice was not the producer's intention (Mitchell and Haugh, 2015). Because spoken corpus data do not always allow analysts to accurately infer recipients' actual feelings (Kotthoff, 1996), the current analysis should be interpreted judiciously.
Another crucial point -particularly when discussing (im)politeness among Asian English users -is the criticism of traditional universalistic notions of (im)politeness and face by Japanese and Chinese scholars, who argue that these primarily reflect Western individual-focused notions of behaviour. Ide (1989) argues that in Japan individual rationality is superseded by a desire to discern the appropriate social or collective convention. Similarly, Mao (1994) underscores the interdependent nature of face in China. Gu (1990) propounds the emic concept of mianzi, recasting face in China as a social and shared phenomenon rather than a universally individual-focused concept. There has so far been little research into this relational phenomenon in other Asian cultures, or to what extent it may permeate Asian ELF discourse. We will explore this as appropriate below.

2.3
Defining English as a Lingua Franca English as a lingua franca (ELF) is English used as a mode of communication between interactants with no shared first language. Socio-pragmatically, ELF is commonly characterised as endonormative, with interactional norms continually being (re-)constructed and (re-)negotiated by interactants rather than being pre-established (Seidlhofer, 2011). ELF users' shared identity as learners and users of English as an additional language (Hülmbauer, 2010) licenses them to suspend their culture-informed norms of socio-pragmatic appropriateness when engaged in ELF talk. That capacity originates in: a) their awareness that limitations in their own and/or their interlocutors' linguistic resources may disrupt production or interpretation of pragmatic items; and b) their consciousness that socio-pragmatic notions of appropriateness differ markedly among lingua-culturally divergent interactants.
ELF is a contact language which can be seen as the outcome of social and individual bi-and multilingualism so that the use of one language shows traces of the other. One must be careful to avoid a monolingual mindset "that assumes 'normal' languages to be pure" (Schaller-Schwaner and Kirkpatrick, 2020: 233). ELF can be linguistically influenced by any of the languages in any ELF speaker's linguistic repertoire. "ELF is thus subject to unbounded variation. It is therefore not a stable easily definable 'thing'" (Schaller-Schwaner and Kirkpatrick, 2020: 234). ELF, as a form, is "… inherently, chronically, irremediably variable" and "is also inherently hybrid in nature" (Firth, 2009: 163). This unboundedness and hybridity is illustrated in our examples from the Asian Corpus of English. Influenced as it is by other languages, ELF provides an excellent context to investigate to what extent humour is culturally shaped. How is humour realised in ELF? To what extent is the humour comparable to humour realised in native English-speaking contexts? Of course, humour in these contexts will also differ, and here we use data from native speakers of Australian English to compare with Asian ELF users.

2.4
Humour in ELF Interactions We here explore some identified functions of humour in ELF, and how humour is accomplished despite ELF interactants' differing linguistic capacities.
One function of humour in ELF is to maintain or enhance solidarity and rapport. A salient example is Matsumoto's (2018) study of the rapport-enhancing functions of laughter in ELF student-teacher miscommunications in American multilingual writing classrooms. An earlier study (Matsumoto, 2014) looked at how humour was collaboratively co-constructed by ELF speakers in an American academic setting, drawing for her analysis on contextual cues such as shared laughter, smiling and abruptness of talk. Matsumoto found that the interactants used humour primarily as a solidarity-boosting device and to create supportive and collaborative talk after disagreements. Habib's (2008) study of humour and disagreement among female ELF users in the US found that humour (particularly jocular teasing) and disagreement were used to maintain and display interactional rapport among interactants (cf. Walkinshaw, 2016;Walkinshaw and Kirkpatrick, 2014). Rogerson-Revell (2007) and Moody (2014) highlight a similar function in business contexts in South East Asia and Japan respectively. In Pullin Stark's (2010) study of business meetings in Switzerland, humour promoted solidarity between the meeting chair and the participants, temporarily overriding hierarchical distance. It also helped establish common ground and solidarity among the culturally and linguistically diverse participants. Humour may also disarm threats to interactants' face in high-stakes contexts. Mežek's (2018) study of PhD oral defences among ELF interactants in Sweden found that humour relieved tension and promoted a collegial, non-confrontational environment. Another function of humour is the (often implicit) maintenance of power relationships. In Pullin Stark's (2010) study, authority figures used humour as a cloaking device to maintain their authority and reduce the face-threat of their critical utterances. In an English language learning context in the US, Davies (2003) explores how language teachers exploit humour as an off-record means of conveying authority and asserting/maintaining control. Lastly, humour may be used to promote acquisition of L2 pragmatic knowledge: Habib (2008) studied how a group of female ELF users in the US used humour for presenting and processing learned information about other cultures and pragmatic aspects of other people's behaviour.
Failed attempts at humour and their interactional ramifications among ELF users have also been studied (e.g. Bell, 2002Bell, , 2007Bell and Attardo, 2010). There are several levels at which attempted humour might fail: language may not be successfully processed at the locutionary level; meaning, connotation, or pragmatic force may be misjudged; the humorous frame may not be recognised (i.e. the joke may be missed); the hearer may not grasp the incongruity of a joke, or may fail to appreciate it, or fail to support the joking frame being used by the other co-present interactants. Yet incomplete comprehension of humour in a second language is no obstacle to appreciating its rapportenhancing value (Bell and Attardo, 2010). Furthermore, lingua franca humour normally succeeds in generating amusement among its recipients. Bell (2002Bell ( , 2007 attributes this to accommodation strategies (cf. Cogo, 2010;Jenkins, 2011) (e.g. avoiding complex linguistic constructions) by which interactants maximise their shared interpretation of meaning. Also germane are Pullin's (2018) findings that non-standard language does not impede humour among ELF interactants, and that ELF users' humour seldom draws on culture-specific knowledge. Similarly, Bell (2002Bell ( , 2007 points to a paucity of humour about potentially sensitive topics. In sum, and as we illustrate below, humour is a common occurrence among ELF speakers, a dynamic interactional activity (Bell, 2007) which is constantly being (re-)constructed and (re-)negotiated as the interaction plays out (Cheng, 2003;Matsumoto, 2014). Even lower-level English language learners/ users co-construct and react to humour to the extent of their linguistic capacity (Davies, 2003). Participants thereby indirectly contribute to group rapport through their alignment with, and extension of, the humorous frame in play.

Humour Types of ELF Speakers in the ACE Data Set
This section outlines the three main types of humour that have been identified as performing an (im)politeness function in the ACE data set: solidarity humour, disaffiliative humour, and self-denigrating humour. Examples of each type are examined in later sections.

2.5.1
Solidarity Humour Solidarity humour refers to humour employed specifically to foster positive politeness (Brown and Levinson, 1987), without intending any actual facethreat. (Though face-threat may be pretended, as we explain below.) Three sub-categories are outlined: mock-impoliteness, witticisms, and autotelic humour. Though broadly disparate, these three humour types share a common orientation to solidarity and enhancement of respondents' positive face. We outline each type.

2.5.1.1
Mock-Impoliteness A common sub-type of solidarity humour is mock-impoliteness, that is, ostensibly impolite utterances whose obvious untruth and impoliteness signal to the recipient that the utterance is non-serious and that the producer intends the opposite meaning to that conveyed in the linguistic realisation (Leech, 1983). Despite targeting some sensitive aspect of the recipient's character, it is (ideally) jointly understood that the humour performs a positive politeness function, reflecting and fostering solidarity between producer and recipient (Culpeper, 1996;Kienpointner, 1997), and that any face-threat is pretended.

2.5.1.2
Witticisms A witticism is defined as a "clever and humorous textual unit interwoven into a conversational exchange, not necessarily of humorous nature" (Dynel, 2009(Dynel, : 1287. Witticisms can include stylistic figures, puns, allusions, distortions or quotations, or register clashes (Dynel, 2009). They tend to be spontaneously uttered, and are typically apropos of the (often non-humorous) conversational context. Their primary social advantage is in allowing producers to display their skill at amusing repartee to their co-interactants. Additionally, the shared amusement generated fosters social intimacy and affiliation in relationships (Lampert and Ervin-Tripp, 2006;Haugh, 2012), and consequently group cohesiveness (Long and Graesser, 1988). Although witticisms can serve purposes additional to -and even subversive of -their jocular function, for instance implying disaffiliation, or moderating or disguising a given moral or affective stance (Haugh, 2017), we confine our focus to witticisms as "benevolent, inoffensive wit" that is "devoid of face-threat" (Dynel, 2016: 129).
Some studies have explored different types of witticism in Australian humour. One is Haugh and Weinglass' (2018) examination of jocular quipsdefined as "playful or light-hearted comments on, or responses to, another speaker's just prior serious talk" (2018: 534) -in Australian getting-acquainted talk. Also, Goddard (2017) mentions quips as part of Australian deadpan jocular irony, wherein humorous utterances are produced within an ostensibly serious frame, with no prosodic cues to signal intended irony.

2.5.1.3
Autotelic Humour Autotelic humour is conceptualised by Dynel (2017) as a cooperative nonliteral, non-bona-fide mode of communication (Raskin and Attardo, 1994) in which the interactional goal is not to communicate information but to provoke a humorous reaction (cf. Raskin and Attardo, 1994), e.g. through joking, irony, absurdity (including jointly-constructed fantasy scenarios), kidding or leg-pulling (Dynel, 2017;Vincent Marrelli, 2006). The overt incongruity of autotelic humour signals to recipients that the utterance intends humour and that truthfulness is suspended. Autotelic humour thereby promotes solidarity politeness through shared amusement, as well as shared insider-knowledge of the witticism's non-truthful nature. Because scope inevitably exists for interpretation as serious or even malicious, we confine our analysis to instances of humour which are (as far as discernable) entirely autotelic.
Some Australian studies have investigated humour's autotelic functions. Goddard (2017) explores what he terms jocular deception, wherein a speaker directs an untrue utterance at an unsuspecting interlocutor to provoke an amusing reaction, an act locally termed geeing up or stirring up the target (Rowen, 2012). Haugh (2016a) examines post-teasing claims of non-serious, non-literal intent ('just kidding') among Australians. Haugh (2017) also examines jocular language play in Australian talk, but his focus is not on its autotelic function but on its utility for negotiating, disguising or modulating a particular affective or moral stance.
Relatively few studies have explored unambiguously disaffiliative humour in Anglo-Australian interactions. Haugh's (2017) study of language play as (dis)affiliation in conversational interaction mentions several such instances, each targeting a non-present third party. Other instances of disaffiliative humour among Australians are discussed by Haugh (2010) and Haugh and Bousfield (2012), in each instance targeting co-present interactants.
A problem in analysing disaffiliative humour is ascertaining intentionality: what is received by a target as genuinely disaffiliative humour may have been intended by its producer as mock-impoliteness (and vice-versa). Again, the limitations of spoken corpus data analysis inevitably moderate the conclusions that can be made about instances of disaffiliative humour.

2.5.3
Self-Denigrating Humour The third type is self-denigrating humour, wherein a producer ostensibly targets their own face for shared humour value among the interactants (cf. Lampert and Ervin-Tripp, 2006). The face-threat is disarmed by its ready interpretation as conveying humorous intent. Zajdman (1995: 337) explains: "it is assumed that nobody in his/her right mind is hostile towards him/herself. Therefore, when a self-denigrating FTA is performed, this is interpreted as humorous." Crucially, self-denigrating humour may burnish one's self-image by demonstrating one's ability not to take oneself too seriously.
The ability to self-denigrate is claimed to be highly valued in Anglo-Australian discourse (Béal and Mullan, 2017;Goddard, 2006Goddard, , 2009Goddard, , 2017Haugh, 2012;Peters, 2007;Sinkeviciute, 2014Sinkeviciute, , 2019: it reflects a cultural principle emphasising ordinariness, familiarity and friendliness, along with the admonishment to play down one's admirable characteristics, abilities or achievements (Goddard, 2006), and deflect or reject praise of these. It may also appease interlocutors' negative face, since it suggests (albeit humorously) that any face threat to a recipient is the fault of the producer.

Methodology
We here describe the Asian English data set. We then explain how instances of conversational humour were identified.

3.1
The Data Set The Asian English data set is drawn from the Asian Corpus of English (ACE), a million-word archive of naturally-occurring spoken interactions among Asian ELF users from Brunei Darussalam, Cambodia, China, Indonesia, Japan, Korea, Laos, Malaysia, Myanmar, the Philippines, Singapore, Thailand, and Vietnam. The three primary domains of talk are: Education, Leisure and Professional (sub-categorised as Professional Organisational, Professional Business, and Professional Research and Science) (ACE, 2014). The ACE data set used for the current study comprised five hours of recorded conversation drawn from the entire ACE corpus. There were nine interactional sequences with 31 speakers in total, in which 59 instances of humour were identified. All but 30 minutes of the 5 hours of talk in the ACE data set took place over a meal or social gathering.

3.2
How Was Humour Identified? Examples of apparent humour were selected intuitively. To support the likelihood that the various interactants intended and received these as humorous we identified contextualisation cues in each sequence (Holmes and Marra, 2002). These were: paralinguistic, e.g. laughter (Glenn, 2003); prosodic, e.g. pauses, sudden shifts in pitch, intonation or tempo, elongation of syllables or vowel sounds (Norrick, 1994(Norrick, , 2004; and discoursal, e.g. further humorous comments on the same topic by the producer or others, or responses from cointeractants indexing the utterance's humorous nature or content. The data set was cross-checked by a second rater for increased reliability. However, we acknowledge that humour's inherent ambiguity (Boxer and Cortés-Conde, 1997) makes prosodic analysis an inexact science, and our reading should be treated provisionally in light of other possible interpretations.

4
The Data Here we explore examples of the three types of humour typically employed in the ACE data set to orient to (im)politeness: solidarity humour (comprising mock-impoliteness, witticisms, and autotelic humour); disaffiliative humour; and self-denigrating humour. In each example we identify the nationalities of the ELF participants from the ACE data set to demonstrate the diversity of their linguistic backgrounds. We then explore similar instances occurring among Australian English speakers. The transcription conventions are listed in the Appendix.

Solidarity Humour
We here explore three sub-types of solidarity humour which we outlined previously: mock-impoliteness, witticisms, and autotelic humour.

4.1.1
Mock-Impoliteness The following (Example 1.1) is an instance of mock-impoliteness in the ACE data set. S1 and S2 are Malay females and S3 is a Chinese female (not talking in this excerpt). The participants have been discussing plants in their dorm rooms.
(Ex. 1.1) S1: i want to buy a plant but i thought i was going to i was going to go off and i thought she will kill it {laughs} S2: i will kill it ah S1: {laughs} i thought i thought The mock-impoliteness (in the form of jocular mockery) occurs in the sequence's first turn: S1 mentions she is hesitant about buying a plant for her room because she has been planning to go away and fears S2 would kill the plant, presumably through neglect or misadventure. The utterance constitutes jocular mockery: a negative assessment of S2's capacity to maintain plants is delivered bald on record -she will kill it -with utterance-final laughter particles to signal S1's non-serious intent. Yet the utterance's unmediated delivery in fact operates as a cue to its target that the mockery is superficial or pretended (Culpeper, 1996;Kienpointner, 1997): S1's actual goal is likely to be relationship affirmation and positive politeness. Although we cannot make definitive judgments about a recipient's genuine reaction to a producer's utterance (Zajdman, 1995), S2's instant self-denigrating agreement -I will kill it ah -does constitute participation in the humour sequence targeting her and appears to signal her acceptance of it as intending solidarity. The mock-emphasis of her utterancefinal discourse marker ah, a feature of Malaysian English roughly synonymous with you know (Low and Deterding, 2003), also illustrates the playful nature of her response, as well as its linguistic hybridity (discussed further in Section 5). S1 then laughs, projecting appreciation of S2's humorous retort and confirming the jocular nature of her original mock-impolite utterance.
Example 1.2 is an illustrative instance of mock-impoliteness (again in the form of jocular mockery) in Australian English (example from Béal and Mullan, 2017).
(Ex. 1.2) S2: what do you have first thing in the morning S1: oh coffee yeah S2: when you wake up S1: yeah S2: oh no. wake up. tea::::::::::::::::::::. : S2: I have it in bed S1: oh no I would never get up I would never get up I would just [on Saturdays we do that] S2: [it's beautiful ] it's beautiful S1: but we stay in bed (.) so S2: it is the HIGHLIGHT of my day S1: {laughter S1} [that's very sad Heather S2: [no {laughter S1 and S2} S1: no you need a life I think S2: {laughs} no I'm serious This sequence is notable for its use of hyperbole, a magnification of reality which describes the world "in terms of disproportionate dimensions" (Haverkate, 1990: 103, cited in McCarthy andCarter, 2004). In interpersonal interaction it may convey affectionate humour between intimate, friendly interlocutors (Slugoski and Turnbull, 1988), as occurs in Example 1.2. S1 and S2 are discussing what they drink when they first wake up. S2 expresses a craving for tea (emphasised by elongating the vowel sound: tea::::::::::::::::::::) which she has in bed, an activity that S1 rejects for herself because she would never get up. Having characterised drinking tea in bed as beautiful (underscoring her point with repetition), S2 now infuses her narrative with humorous overstatement (McCarthy and Carter, 2004) and sudden increased volume: it is the HIGHLIGHT of my day. Laughing at S2's hyperbolised characterisation, S1 delivers a mock-impolite indictment of S2's claim: that's very sad Heather. Her utterance is clearly hyperbolic, since sadness is seldom occasioned by slugabed tea-drinkers, and the hyperbole marks her impolite characterisation as non-literal. S2's response contains a token protest ('no'), but its kernel is supportive laughter which signals her acceptance of the hyperbolic utterance as humorous and non-literal (McCarthy and Carter, 2004). S1 then resumes her humorous turn with no you need a life I think. Her utterance, a variation on the cognate idiom get a life, exhorts S2 to stop focusing on unimportant things and start living a fuller existence. Again, S1's negative assessment constitutes relationship-affirming hyperbole rather than literal truth, and that is how S2 receives it, responding with further laughter.
In terms of potential face-threat, the contrast between these two representative examples of solidarity-oriented mock-impoliteness is quite marked: the ACE example is innocuous, but the Australian one is nearer the knuckle, casting the target's pleasant habit as a deficiency to be rectified. The potential for misinterpretation as connoting malicious intent is greater.

4.1.2
Witticisms Example 2.1 presents a witticism in the ACE data set. The participants are eating in a restaurant. S2 and S5 are Vietnamese females, S3 and S4 are Indonesian females and S7 is an Indonesian male.
(Ex. 2.1) S4: why you cannot oh you cannot eh ok erm do you have fork S3: i would like too S5: and i have fork fork too S2: yeah me too me too S4: ah we want fork but no pork SS: {laughter} The humour sequence is initiated by S4, who notices a group member requires a fork. When she asks a waiter if he has one S2, S3 and S5 all state that they need forks as well. S4 plays on the rhyming sounds of words when asking the waiter for a fork: we want fork but no pork. Additionally, her wordplay references that some of the participants are Muslim and do not eat pork. The witticism and the recipients' shared mirth constitute social play (Long and Graesser, 1988), addressing their positive face as in-group members partaking in a shared activity.
In Example 2.2 from Australian English (from Béal and Mullan, 2017), participant S1 has just arrived at researcher S2's house for a recorded interview. She spots the recording device to be used for the interview and (after confirming that the tape is rolling) utters a witticism: (Ex. 2.2) S1: is the tape rolling? S2: yeah S1: dobr'iy vecher {laughs} S2: hello Dobr'iy vecher means good evening in Russian. S1's abrupt code-switch to Russian (which she has been learning) appears to reference that the conversation she is about to commence is being recorded -a humorous allusion to the Russian security service's covert recording practices. The humorous utterance occurs at the very beginning of the speech event (i.e. the interview), thereby setting up a play frame (Norrick, 1993) which moves the interlocutors from quotidian information transfer into a more jocular, solidarity-focused exchange.
These examples demonstrate that witticisms perform solidarity-enhancing functions in the ACE data set just as in Australian humour research, and are not the preserve of native(-like) speakers of English.

4.1.3
Autotelic Humour Example 2.3 is a representative example of autotelic humour in the ACE data set. S1 and S2 are both Malay and S3 is a first language speaker of Cantonese. All are females. The three participants have been talking about how to tempt a chap when he comes to their apartment. A non-bona-fide (Raskin and Attardo, 1994) fantasy humour frame is constructed by S1, who jokingly suggests that S3 should wear a swimsuit and offers to teach her how to wear one. S3 laughs at the incongruity of S1's suggestion: swimsuits are an odd choice of costume for welcoming a man into one's apartment. S1 extends the fantasy frame by partially repeating her prior humorous utterance (wear a swimsuit), explicating that S3 should be wearing the item when her suitor opens her door: wear a swimsuit h then he open-. S3 interjects with an expression of amused incredulity at S1's suggestion: h really [S1] {laughs}. In sum, S1 is employing a positive politeness strategy by introducing an incongruous fantasy scenario for humour value. No face-threat is evident. Notice also that both S1 and S3 transfer the discourse particle lah into their lingua franca talk from their first language, exemplifying the linguistic hybridity ELF can display.
Example 2.4 from Australian humour literature comes from Haugh (2017). It is a getting-acquainted interaction between two colleagues in their thirties. S2 has been complaining to S1 that she is unable to work in a pub or aged care facility in Australia because she lacks a basic qualification. As a narrative emerges about the need for a course or qualification for even basic jobs, S1 introduces a playful frame, making an absurd (Kotthoff, 2007) claim that there is a course for getting out of bed. Absurdity is a common form of autotelic humour, dealing is it does in "notions that fly in the face of reality, common sense, or logic, and/or violate the universal rules of the world" (Dynel, 2017: 93). Absurdity is often jointly constructed in interactional exchanges, as here: S2 takes up and extends S1's line of absurd humour with her rejoinder that there is a course for tying one's shoelaces. S1 responds with laughter and agreement. On a roll, S2 then further extends the absurd humour sequence, claiming that OHS (occupational health and safety) checks are in place for making breakfast.
These two examples illustrate that autotelic humour occurs in both the ACE data set and in Australian humour studies. Note though that unlike the Australian example the ACE example is limited to a few turns, and is constructed by one individual rather than jointly by several interlocutors. This typifies autotelic humour in ACE, which tends to be unilaterally contrived and brief in duration.

4.2
Disafilliative Humour In Example 3.1 from ACE, three Thai females (S1, S2 and S3) and a Burmese female (S4) are talking about how Arabic sounds to their ears. S1 and S3 work in customer service at the same company.
(Ex. 3.1) S1: if you listen er the way (.) arabic arabic arabic person talk to customer is very HARSH {loud gibberish} SS: {laughter} : S4: talking arabic S1: ah yeah S4: that's funny SS: {laughter} S1: like need to fight a war {laughs} S1 initiates the disaffiliative humour sequence, evoking her co-worker's shared experience of people speaking Arabic: if you listen er the way (.) arabic arabic arabic person talk to customer is very HARSH {loud gibberish}. The move is intended to foster a sense of complicity and in-group membership: the outgroup are the (non-present) Arabic-speaking customers who are the butt of S1's disaffiliative humour. S1 describes the sound of Arabic as harsh, articulated at increased volume for emphasis, and followed with loud gibberish intended to represent how Arabic sounds to her. Her exaggerated representation constitutes disparagement humour, which "denigrates, belittles or maligns an individual or social group" (Ford and Ferguson, 2004: 79). General laughter ensues, marking the interactants' acceptance of the disparaging humorous frame; S4 directly indexes it: (that's funny). S1 then employs a simile to illustrate her interpretation of the paralinguistic properties of spoken Arabic: like need to fight a war. This too constitutes disparagement humour, referencing a stereotype of Arab people as warlike for humour value.
We now turn to an example from Australian English. In Example 3.2 (from Béal and Mullan, 2017), the Australian participants are engaged in a retelling about a conference dinner.
(Ex. 3.2) S2: there was this bloke next to me though who was another ex-student in the eighties S1: mm S2: (???) old woman (.) and um and he um he's turned into this trumpet fanatic {laughter} S2: and then on the other side (.) so trumpet Malcolm on one side S1/S3: {laughter} S2 describes having sat next to a fellow attendee known for continually practicing his trumpet. She constructs a disaffiliative characterisation of the target as a narrative device to add humour value to her anecdote: her use of the phrasal verb turned into rather than become suggests an abrupt, unexpected metamorphosis, which is clearly non-literal. She then characterises the target as this trumpet fanatic, representing his pastime as excessive or manic. S2's disaffiliative characterisation is hearable as constituting genuine mockery, i.e. demeaning or denigrating an aspect of someone's character for the gratification of other co-present interlocutors (Everts, 2003). The ensuing general laughter signals the co-interactants' acceptance of S2's disaffiliative frame as jocular. S2's continued retelling of the event encodes further disaffiliative mockery (Everts, 2003): she formulates a reductive nickname for her target (trumpet Malcolm) based on his purported obsession. Further laughter from her co-interactants attests to their continued acceptance of S2's disaffiliative characterisation as jocular and unmarked.
Both the above instances are multi-party interactions, as is common in disaffiliative humour (Dynel, 2016). Both also target non-present others. This is typical in disaffiliative humour scenarios in the ACE data set, but possibly less so in Australian humour: Haugh's (2017) study targeted absent others, but Haugh (2010) and Haugh and Bousfield (2012) both examined instances where the disaffiliated target was co-present.

4.3
Self-denigrating Humour In an illustrative example of this type of humour from ACE, S1 is a Korean male and the other two participants are Malay females.
(Ex. 4.1) S1: hey how are you? S3: fine S1: {laughs} S2: s-sleepy S1: sleepy S2: sleep i: {laughs} S1: sleep i S2 incongruously responds to S1's formulaic salutation how are you? with sleepy. Sleepy is a dispreferred response (Pomerantz, 1984) because it deviates from the conventionalised second pair part fine. S2's unexpected self-assessment is also hearable as implying low energy or tiredness, and therefore connotes potential social ramifications: other in-group members may feel compelled to adjust their plans to accommodate S2's fatigue. Yet pragmalinguistic and paralinguistic cues in her response mark it as non-serious and intending humour. A pragmalinguistic cue is the very incongruity of S2's sudden deviation from the conventionalised ritual exchange. Incongruity frequently signals the humorous nature of an utterance, as it appears to do here. The paralinguistic cue is encoded in S2's repeated utterance of the key word sleepy: she artificially elongates the final syllable the second time (Norrick, 1994(Norrick, , 2004 and interpolates laughter particles to underscore that her self-assessment is a humorous exaggeration. Her interlocutors' laughter marks their reception of her utterance as humorous. In a corresponding Australian example (from Béal and Mullan, 2013), interactants make self-denigrating comments about the interest value of the photographs they are about to show. S1 and S4 are guests arriving at the home of the hosts, S2 and S3.
(Ex. 4.2) S1: so we've brought some happy snaps to bore you with/ S3: great : S4: well we've got all of ours here S3: (???) to bore you to death with ours yeah Guest S1 humorously denigrates the interest value of the photos he has brought along by describing them diminutively as boring happy snaps. His utterance constitutes self-denigrating deadpan irony, both of which devicesself-denigration and deadpan irony -have been framed as features of Anglo-Australian humour (Béal and Mullan, 2017;Goddard, 2006Goddard, , 2009Goddard, , 2017Haugh, 2012;Peters, 2007;Sinkeviciute, 2014Sinkeviciute, , 2019. The self-denigration lies in S1's characterisation of his own photos as boring, while the irony is that S1 mentions them because of their presumed interest value to others. The deadpan delivery is also salient: there are no prosodic clues such as shifts in volume or pitch to indicate ironic or humorous intent (cf. Goddard, 2006). Deadpan delivery is effective here because it "cannot be read as conveying any undertone of criticism. The apparent implication may seem negative but the conveyed social message, if properly understood, is entirely positive." (Goddard, 2017: 64). Host S4 mentions that they too have photos to show, whereupon S3, the other host, extends and participates in S1's humorous frame by interjecting: to bore you to death with ours yeah. S3's humorous rejoinder serves two parallel functions. First, it demonstrates affiliation (Steensig, 2013) to S1's original humorous utterance (via partial repetition -to bore you). Second, it appropriates the utterance's form and intended function to her own identical purpose, i.e. to humorously denigrate the interest value of her own photos and minimise face-damage if they are judged uninteresting.
As these representative instances illustrate, the functions of self-denigrating humour in the ACE data set appear similar to those identified in Australian humour studies: to generate amusement by targeting one's own face, and to enhance one's self-image by demonstrably not taking oneself too seriously.

Discussion
We have examined how humour is used by ACE interactants to attend to (im) politeness targeting their recipients' positive or negative face. How does their speech behaviour balance against comparable data from Australian humour studies? We examine first the similarities and then the differences. In terms of similarities: the ACE participants used self-denigrating humour that superficially threatens the speaker's own face but actually indicates an ability not to take themselves too seriously (cf. Walkinshaw, Mitchell, and Subhan, 2019), a trait also valued in Australian humour (Goddard, 2009;Haugh, 2010Haugh, , 2012Haugh and Bousfield, 2012;Sinkeviciute, 2014). The ACE data set also contained examples of language play such as witticisms or autotelic fantasy/ absurd humour. This is interesting as it shows how ELF users are able to play with the language, whether or not they are highly proficient in it (Pullin, 2018). Language play has also been identified as occurring in Australian discourse (e.g. by Haugh, 2017;Haugh and Weinglass, 2018).
As to differences: in the ACE data set, most humour sequences are readily interpretable as 'doing' solidarity, i.e. promoting amicable interpersonal relations. Instances which potentially connote malicious intent are rare. Even mock-impoliteness sequences avoid potentially sensitive topics (Bell, 2002(Bell, , 2007 and typically occur among close friends during a laughter-infused sequence -thus optimising the likelihood of a jocular reception by their target. Riskier realisations such as jocular abuse, "a specific form of insulting where the speaker casts the target into an undesirable category or as having undesirable attributes using a conventionally offensive expression within a non-serious or jocular frame" Bousfield, 2012: 1108), are entirely absent from ACE. A possible underlying factor is ELF users' consciousness that they and their interlocutors operate in an additional language, with potentially differing capacity to accurately produce and interpret such humour types within the intended jocular frame. We theorise that ELF speakers in the ACE data set, sensitive to linguistic and cultural differences as an obstacle to understanding (Cogo, 2010), make cautious and conservative use of such pragmatically fraught formulations to avoid inadvertently causing face damage and disturbing group rapport. And do these findings suggest that the ACE interactants are guided by a communal, interdependent face/(im)politeness convention, as Gu (1990), Ide (1989), and Mao (1994) argue of Japan and China? Their predominant orientation to solidarity and rapport management in conversational humour does tentatively point to a group-focused approach to face-management rather than one prioritising the individual. Pending further research though, we can only speculate about this possibility.
In contrast to the ACE data set, Anglo-Australian humour is noted for its (actual or pretended) face-affronting character (Béal and Mullan, 2017;Goddard, 2006;Sinkeviciute, 2014Sinkeviciute, , 2019Wierzbicka, 1997). Goddard (2006) outlines relational-affirming strategies such as rubbishing your mates or taking the piss by means of jocular abuse or deadpan jocular irony (cf. Goddard 2017; Sinkeviciute, 2019). Likewise, Haugh's (2014) exploration of Australian humour describes a characteristic tendency to jocular mockery and teasing. Nor are sensitive topics necessarily off the table. An example from Haugh (2016b) is illustrative: a female interactant mentions serving as best man (traditionally a close male friend) at a friend's wedding. Another interactant remarks: you were the best man he could find, humorously implying that she was only asked because the groom could not find a man to ask. The jocular frame notwithstanding, the utterance is interpretable as diminishing the value of the target's role in the wedding. Wierzbicka (1997: 205-6) links the phenomenon to an Anglo-Australian cultural characteristic: the "preference for saying 'bad things' rather than 'good things' about people in general and about the addressee in particular, not because one thinks 'bad things' or feels 'bad things' towards them, but because of the cultural ideals of roughness, toughness, anti-sentimentality, anti-emotionality, and so on." Related is the claimed Anglo-Australian cultural ethos not to take oneself too seriously (cf. Goddard, 2009;Haugh, 2010Haugh, , 2012Haugh and Bousfield, 2012;Sinkeviciute, 2014Sinkeviciute, , 2017: targets of potentially face-threatening humour are expected to take the 'nipping' (Boxer and Cortés-Conde, 1997) jibes or teases in good part rather than to display offence.
Finally, let us explore linguistic hybridity in the ACE data set. We mentioned two instances (Examples 1.1 and 2.3) where discourse markers from the speakers' first language are transferred into their use of ELF. Here we provide a further instance of linguistic hybridity. In Example 5.1 S2, a Malay female, has just finished telling an anecdote about being too embarrassed to ask the handsome male invigilator for permission to go to the toilet during an exam. Her friend, also a female Malay, chides her for feeling embarrassed, suggesting she should be happy for the opportunity to talk to the handsome invigilator. The third participant, S3, is a female first language speaker of Cantonese, a language the other two are familiar with. These are the same participants as in Example 2.3 above.
(Ex. 5.1) S2: because i want to go toilet you know for the first when when the i don't start i just started cos i was late mah so i went in in the bathroom S1: you were late S2: ah S1: what time S2: about five minutes late ah S3: oh S2: so all of them started already so i don't want to go toilet lah so i just sit there and then once i sit down i feel like i feel the S1: toilet S3: o k you going to there S2: then i plan lah maybe thirty minutes later lah i will go then the invigilator is too too leng zai {cantonese: handsome} ah i don't dare to ask you know S1: {laughs} S2: i don't know how to ask S3: too long zai S1: too leng zai S2: too S1: handsome S3: lam lam {trying to understand equivalent in cantonese} o k k S1: leng zai S3: lam zai S1: [l-e-n-g {spells the word}] S3: [ah] lang lang zai o k k S1: a:h S3: {softly} i didn't know about that S1: then how S2: how to ask so pai seh {hokkien: embarrassing} ah h S1: NO ah NO no you should be happy ah h There is a great deal of code-mixing in this example. Not only are there frequent instances of discourse markers from the speakers' languages, S2 uses the Cantonese word for handsome, leng zai. This causes some negotiation, as the first language speaker of Cantonese, S3, makes sure she has heard correctly. S2's final comment here also involves code-switching, but from a different Chinese language, Hokkien, when she describes the situation as pai seh or embarrassing. This example demonstrates the intrinsically multilingual nature of ELF.

Conclusion
In this study, we have investigated (im)politeness and face in conversational humour among Asian multilinguals using English as a lingua franca, contrasting our findings with equivalent studies of Australian English speakers. Some key findings are: (i) Self-denigrating humour is a common feature of the ACE data set, as it is in the studies of Australian humour discussed above. Producers of such utterances often degrade their own face through humour, cognisant of increasing their positive self-image as being able to laugh at themselves. (ii) Although some humour scenarios in the ACE data set pretend mild face-threat, the majority present patently non-face-threatening humour.
Conversely, studies suggest that Australian humour is more likely to encode some form of pretended face-threat, often quite on-record and susceptible to interpretation by targets as malicious. (iii) Those instances in the ACE data set which do encode a pretended face threat largely avoid sensitive topics and are clearly situated within an established jointly-constructed jocular frame. In Australian conversational humour sensitive topics may be raised for humour value, though usually also within a jocular frame. (iv) Although ACE contains instances of disaffiliative humour which encode actual face-threat, their targets are invariably absent and the potential for face-damage correspondingly low. By contrast, some Australian humour studies present instances of disaffiliative humour targeting co-present interlocutors, with higher risk of face-damage.
What this study also shows is that users of ELF are perfectly able to construct and respond to humour in their interactions. Thus, despite operating in a language that is not their first and interacting with people from a variety of linguistic and cultural backgrounds, these ELF users can use humour in ways that are frequently comparable with the use of humour by native speakers. This exemplifies how a desire to communicate in a collegial manner (House, 2009) can overcome variability and hybridity. Indeed, these ELF users exploit variability and hybridity to engender humour. We hope that this initial investigation into the use of conversational humour by Asian users of ELF and its comparison with native speakers' use of humour may stimulate further comparative studies in the use of conversational humour.