The structure of everyday narrative in a city market : An ethnopoetics approach

This paper considers the value of Hymesian ethnopoetics as a means of analysing everyday narrative in conditions of mobility and change. The paper offers an account of the development of ethnopoetics as a means to make visible and valorize narrative in the Native American oral tradition, and as a method of revealing culturally specific relations of form and meaning. Hymes’ ethnopoetic approach viewed narrative structure as a reflection of a cultural tradition of meaning-making. Hymes’ analysis proposed that traditional narrative was a culturally shaped way of speaking, and analysis of narrative structure could reveal and recreate culture. His orientation rested on an assumption that the culture of a group was more or less stable and fixed. This paper adopts an approach to analysis based on ethnopoetics, representing everyday narrative dramatically, organized not only as lines and verses, but also as scenes and acts. Representation in scenes and acts makes visible the dynamic nature of the narrative. The paper asks whether Hymes’ ground-breaking work on ethnopoetics still has currency and purchase in 21st-century conditions of mobility, change, and unpredictability. Analysis of everyday narrative in a city market concludes that, notwithstanding the complexity of notions of ‘culture’ and ‘language’ in such conditions, ethnopoetics can be productively applied to everyday contexts for the analysis of narrative.


INTRODUCTION
Ethnopoetics, developed in analysis of Native American oral narrative, may be extended to the analysis of narrative in everyday encounters. Analysis of the ethnopoetics of narrative has hitherto largely been conducted in the context of folk stories transcribed as dictated to researchers (Hymes 1981(Hymes , 2003. The narratives Hymes focused on were told in Northwest U.S. Indian languages to anthropologists Boas, Sapir, and others. Hymes accessed transcripts of narrators presenting the most valued stories (myths) of these societies, frequently dictated slowly, to a relative stranger (the anthropologist), with no native audience responding. The stories were presented as prose, with facingpage and/or interlinear translation. It was against this background that Hymes introduced analysis of measured verse patterns, based on Jakobson's (1960) notion of 'equivalence' in poetic verse. By these means, Hymes aimed to demonstrate the validity and value of oral cultures, arguing that it is possible to arrive at an arrangement of a transcript that reflects the rhetorical or poetic structure of an oral performance, and that, in doing so, a native voice may be restored or recovered.
However, the context of this paper is quite different, as we introduce ethnopoetic analysis of narrative as it occurs in everyday speech in a busy city market. In contexts of 21st-century mobility, 'language' and 'culture' are complex, as people appear to take any linguistic and communicative resources available to them and blend them into complex linguistic and semiotic forms. In a globalising world, narrative and culture are less stable and fixed than they once might have been, as resources are characterised by internal and external forces of perpetual change operating simultaneously and in unpredictable mutual relationships (Blommaert 2014). In this context, speakers juggle the limits of face-to-face intelligibility with new styles of expression made up of ever-changing linguistic resources (Parkin 2016). This leads us to question whether Hymes' ground-breaking scholarship on ethnopoetics is equipped to enhance our understanding of everyday narrative in situations of complex interrelationships. In the remainder of this paper, we analyse an interaction between a Chinese butcher and his customer, audio-recorded in the course of four months of detailed ethnographic observation in a busy city-centre market in the U.K. In doing so, we consider whether ethnopoetic analysis, rooted in Jakobson's analysis of poetry and Hymes' analysis of folk tales, has the potential to enhance our understanding of language in contemporary social life. Jakobson (1960) argued that there were six functions of language: referential, phatic, metalingual, conative, emotive, and poetic. Jakobson (1960: 353) insisted that the functions he identified were not separate, but co-existed in shifting hierarchies of dominance: 'The diversity lies not in a monopoly of some one of these several functions but in a different hierarchical order of functions'. Jakobson argued that, whilst the predominant orientation of many messages may be to the referential function, we should be alert to the accessory participation of other functions, and that 'the scrutiny of language requires a thorough scrutiny of its poetic function ' (1960: 356). Jakobson pointed out that the poetic function could not be studied out of touch with other functions, but should be considered in the context of referential, phatic, metalingual, conative, and emotive functions. He proposed that when we pay attention to repetition and 'sameness' in a text (including an oral text) we are able to make comprehensible its poetic structure (1960). For Jakobson (1985: 42), any noticeable reiteration of the same grammatical concept becomes an effective poetic device. He pointed out that repetitions such as parallelisms, whether based on sound, or on grammatical categories, or on lexical categories, are a natural result of the raising of equivalence to the constitutive device of the sequence. In this paper, we propose that such structures are identifiable in everyday narrative speech. Through attention to structures of repetition we are able to hear more clearly the voices of speakers in the city market, and to make more visible the social relations between them.

ETHNOPOETICS
Hymes (1981) applied Jakobson's analysis of equivalence to Northwest U.S. Indian language folk tales collected and transcribed by researchers. The folk tales were perhaps somewhere between Jakobson's characterisations of 'poetry' and 'ordinary speech'. An oral genre, Hymes recognised that the folk tales were nonetheless organised in terms of patterned sequences of lines (2003). In his analysis, Hymes demonstrated that oral narratives are 'organized in terms of lines, verses, stanzas, scenes, and what one may call acts ' (1981: 309). Hymes claimed that analysis of the poetic structure of narrative 'will add to understanding of language itself and contribute to the many fields of inquiry for which the use of language in telling stories is a part ' (2003: viii). Discovering lines and relations in narrative: can lead to understanding and interpretation otherwise not possible. We can recognise artistry and subtleties of meaning otherwise invisible. For a true account of the human capacity for verbal art, this is crucial. (Hymes 2003: 96) Here, Hymes speaks of the potential of ethnopoetics to bring to light the verbal art of discourse. Whereas Jakobson had privileged the poetics of verse, and Hymes the poetics of traditional Native American folk tales, we will explore the potential of ethnopoetics to bring to light the poetics of everyday speech in a city market. Hymes (1981) insisted that we must work to make visible and audible something more than is evident on first hearing. Ethnopoetics, he said, 'helps us to see more of what is there' (Hymes 1996: 182). Whereas representing narrative as prose tends to hide its characteristic form, ethnopoetic analysis unearths the underlying poetic structure that is the essence of narrative (Hornberger 2009: 349). Bauman (2013) described Hymes' work on ethnopoetics as a game-changing, perspective-altering way of conceptualising texts, revealing their poetic form, and ultimately, through that analytical process, elucidating their cultural meaning: 'there is no question that Hymes' discoveries were truly pathbreaking, opening up new territories in oral poetics and attracting large numbers of further explorers and new settlers into the territory' (2013: 177). Attention to the poetic structure of narrative can challenge received assumptions about the nature of language and the ways that individuals engage in and use language. An ethnopoetic approach to narrative takes up Jakobson's notion of equivalence. Hymes viewed narrative as structured in 'equivalent' lines and groups of lines (verses, stanzas, scenes), and argued that the organisation of lines in narratives is an implicit patterning that creates narrative effect. The principle of equivalence implies a text that is a sequence of units (Hymes 1994). In addition to equivalent units (and repetition and parallelism), there is 'succession'. Succession is not a matter simply of linear sequence, of counting. Successive units give shape to action. In particular, patterns of succession can be ways of coming to an ending point. Hymes (2003) refers to 'intonation contours' as structures which organise lines. Verses may be signalled by a grammatical feature such as reported speech in a narrative, or turns at talk. Hymes (1994) suggested that a verse is easily recognised in speech, being marked by one of the main intonational contours of the language. Such verses form sequences, and do so in terms of a small set of alternatives. Repetition of words, phrases, or grammatical structures may also mark equivalence. Hymes' approach is one in which narrative is re-organised through attention to prosodic features, syntactic features, morpho-grammatical features, phonetic features, and lexico-syntactic features (Blommaert 2006a). Lines then combine into larger units, verses, stanzas, scenes, and acts, and Jakobsonian equivalence is the formal principle that identifies such units. A transition from one unit to another can be marked by a shift in intonation or prosody, a change in the dominant particles used for marking lines, a change in verb tense, or a lexical change. Hymes argued that artistic patterns in narrative such as parallel structures, rhythmic repetitions and lexical oppositions indicate a high level of formal skills and sophistication. Scollon and Scollon (1981) proposed a more explicit definition of narrative structure, distinguishing between lines (utterances separated by pauses), verses (units that correspond to sentences), stanzas (units that share the same perspective on action, participants), and scenes (units that indicate changes in participants, actions, location). Hymes (1981) developed ethnopoetic analysis as part of a more general project which aimed to show that oral cultural traditions were as valid as written genres. Ethnopoetics was a means of investigating cultures and their specific ways of understanding reality. It was also a means of achieving a better understanding of local knowledge (De Fina and Georgakopoulou 2012). Hymes (2003) also argued that close analysis of narrative structure can reveal culturally specific relations of form and meaning. His study of narrative traditions among Native Americans became part of a more general project that involved looking at the ways in which texts are organised as both reflecting and recreating cultural traditions of meaning-making. Scollon and Scollon (1981) similarly argued that the structure of narrative mirrors cultural interaction patterns. They studied Athabaskan narratives, and concluded that their narrative structure mirrors interaction patterns in Athabaskan life. Moore (2013) analysed Wasco-Wishram Chinookan oral narrative, and elaborated on Hymes by taking into account the shifting linguistic environment in which narration takes place. Moore proposed that we need to see how poetic structures in discourse not only emerge in contexts of verbal interaction, but also help to (re)shape those contexts in particular ways. He suggested a transcription format that enables us to represent on the page these situated aspects of poetic form that unfold in everyday interaction.
Ethnopoetic analysis attempts 'to unearth culturally embedded ways of speaking, materials and forms of using them, that belong to the sociolinguistic system of a group, and that have a particular place in a repertoire due to their specific form-function relationships' (Blommaert 2009: 269). A focus on individual narrators pays not only attention to what is said explicitly, it also allows the analyst to discover relations between lines and verses in the narratives. These relations are often not noticed or consciously produced by narrators and may therefore be considered to be the 'cultural dimension' of narration (van der Aa 2013). For Hymes, 'narrative inequality' derives from the fact that certain ways of speaking, certain ways of telling a narrative, may be dismissed and marginalised (Webster and Kroskrity 2013). Reconstructing the functions of narratives is a politics of recognition which starts from a restoration of disempowered people as bearers and producers of valuable culture, over which they themselves have control: for Hymes recognising one's language means recognising one's specific ways of speakingone's voice. Blommaert (2006a: 181) argues that ethnopoetics could be used for the analysis not only of traditional folk narratives, but also of narratives in institutional contexts such as police interviews, courtroom hearings, and asylum interviews. He proposes that there is room for exploring 'applied' topics for ethnopoetic analysisfor taking it beyond the study of folkloric oral tradition and into other spaces where narrative matters: 'It would be a great pity if a powerful analytic tool such as ethnopoetics would remain under-used because of it stereotypically being pinned on a small set of particular analytic objects' (Blommaert 2006b: 268). Tannen (2007: 101) argues that Jakobson's observations of pervasive parallelism in poetry 'apply as well to conversation'. She argues that repetition is a limitless resource for individual creativity and interpersonal involvement, and claims that it is 'the central linguistic meaningmaking strategy ' (2007: 101). Collins (2009: 335) goes further, saying a failure to treat narrative in this way is to treat everyday speech as 'the country cousin to the citified sophistication of deliberately composed prose'.
By attending to implicit, indexical patterns in narratives, applied ethnopoetics creates different criteria for assessing the validity of stories, because it reconstructs a different voice (Blommaert 2006a). Ultimately, what ethnopoetics does is to visualise the particular waysoften different from normsin which subjects produce meanings (Blommaert 2006b). Collins (2009: 334) points out that Hymes' analysis offers evidence of a richness of everyday storytelling that calls into question received dichotomies between speaking and writing, as well as the common contrast between narrative as 'mere anecdote' and analytic thought. In Hymes' analyses of both Amerindian and English narratives, he argues that to represent narrative as prose is to lose sight of its characteristic form, and to ignore this form is to misconstrue the nature of narrative and to lose sight of a common human potential. Ultimately, what ethnopoetics does is to 'enhance respect for an appreciation of the voices of others' (Hymes 1996: 219). Such an approach has the potential to rearticulate the voices of those whose speech in the language of the narrative may appear inarticulate, hesitant, or lacking confidence.
In summary, an ethnopoetic analysis of narrative was developed by Hymes as a means of investigating culturally specific relations of form and meaning in the structure of (oral) narrative. In order to make explicit the structure of narrative he developed a means of segmentation on the basis of parallelism, expressive phonology, syntax and lexicon. More recently, a relatively small number of scholars has appropriated Hymes' approach as a means of 'restoring' otherwise silenced or disenfranchised voices. Blommaert proposes that Hymes's ethnopoetic work is one way of addressing the main issue in ethnography: to describe (and reconstruct) languages not in the sense of stable, closed, and internally homogeneous units characterising parts of mankind, but as ordered complexes of genres, styles, registers, and forms of uselanguages as repertoires or sociolinguistic systems (Blommaert 2009: 269). Here lies a potential tension, however. Ethnopoetic analysis was developed largely in the assumption that 'cultures' were more-or-less homogenous, that members of a group were likely to tell stories in similar ways, and that narrative structures may be straightforwardly linked to the cultural practices of specific communities. However, in the short time since Hymes developed ethnopoetics as means of understanding culturally specific norms and practices, many global societies have become more mobile, and consequently more complex in terms of 'cultural norms' and 'cultural practices'. In contexts of 21st-century mobility, 'language' and 'culture' are more complex, as people appear to take any linguistic and communicative resources available to them and blend them into complex linguistic and semiotic forms. In a globalising world we may consider narrative and culture as less stable and fixed than they once might have been, and as resources characterised by internal and external forces of perpetual change, operating simultaneously and in unpredictable mutual relationships (Blommaert 2014). This leads us to question whether Hymes' ground-breaking scholarship is equipped to enhance our understanding of everyday narrative in situations of complex interrelationships. In the remainder of this paper we analyse an interaction between a Chinese butcher and his customer, audio-recorded in the course of four months of detailed ethnographic observation in Birmingham Bull Ring market, in the U.K. In doing so we consider whether ethnopoetic analysis, rooted in Jakobson's analysis of poetry and Hymes' analysis of folk tales, has the potential to enhance our understanding of language in contemporary social life.

ETHNOPOETICS IN A CITY MARKET
Markets are places where we encounter difference. More than any other city spaces, they define human engagement with difference, with different people, different clothes, different goods, and different ways of speaking (Pennycook and Otsuji 2015). Markets offer 'an ideal setting to explore the relationship between economy and society, especially when we consider the ways that these markets reflect, but also shape, the nature and meaning of social and cultural diversity' (Hiebert, Rath and Vertovec 2015: 16). They entail encounters between people, frequently across lines of social and cultural difference. Watson (2009aWatson ( , 2009b argues that markets represent a neglected site of social connections and interaction in cities, which have been subject to limited analysis to date. Watson (2006aWatson ( , 2006b proposes that the sociocultural context of markets warrants textured investigation to make sense of how encounters across difference occur productively or antagonistically. The example we present in this paper is an interaction between a customer and market traders on a butcher's stall, recorded in November 2014. The research was conducted as part of a four-year project funded by the Arts and Humanities Research Council, 'Translation and Translanguaging: Investigating Linguistic and Cultural Transformations in Superdiverse Wards in Four UK Cities'. 1 The multi-site ethnographic project is directed by the second author, Angela Creese. Field researchers were the first and third authors, Adrian Blackledge and Rachel Hu. The overall aim of the project is to investigate how people communicate in superdiverse cities when they bring different histories, biographies, and trajectories to interaction. Over four years, linguistic ethnographic research was conducted in sixteen sites in four cities in the U.K.: Birmingham, Cardiff, Leeds, and London. The present paper is concerned only with material collected in the Bull Ring Indoor Market, in Birmingham. The specific focus of this paper is to consider whether Hymesian ethnopoetics is fit for purpose in analysing everyday narrative in the culturally diverse setting of the city market-place.
In the example presented here, Adrian Blackledge and bilingual researcher Rachel Hu were engaged in regular observation of communicative interactions at a butcher stall owned by a Chinese couple, Kang Chen and Meiyen Chew. Kang Chen was originally from Changle in Fujian, in the South of China. He had relatives in the U.K., and arrived in 2001. He opened the butcher stall in 2011. He told us he was able to get the business off the ground by selling to a niche market, stocking products such as fish balls, blood curd, and pig's intestine.
It was evident to the research team that a significant feature of the repertoires of both market traders and their customers was the deployment of gesture. Rymes (2014) adopts the term 'communicative repertoire' to refer to the collection of ways individuals use language and other means of communication to function effectively in the multiple communities in which they participate. Repertoire can include not only multiple languages, dialects, and registers in the institutionally defined sense, but also gesture, dress, and posture. Gesture, mime, and physical performance were part of the spatial repertoire (Pennycook and Otsuji 2015) of the market hall. However, this deployment of the corporeal voice, of gesture as repertoire, was not always equally available to all.

STRUCTURES OF EVERYDAY SPEECH
The interaction we present here is one of the many hundreds we observed during the field work period. Kang Chen and Meiyen Chew collaborated with the researchers for four months. For much of this time they and their assistant butcher, Bradley, wore small microphones attached to digital voice recorders. They gave written consent to have their names identified, and for their speech and other means of communication to be represented in academic contexts. Large signs in Chinese and English informed customers that their voices may be recorded for research purposes, and that they may choose to have these recordings deleted.
A Chinese woman in her sixties arrived at Kang Chen's stall, asking him where she could buy lamb (he did not sell lamb himself). Kang Chen pointed her to a stall along the aisle. At the time of his encounter with the customer, Kang Chen (KC) was wearing a digital voice recorder, so we were later able to listen to his interaction with the customer (FC). The interaction, Excerpt 1, is first presented in Chinese script and Roman script. English translation is provided below, and indicated by < pointed brackets >. The Chinese woman approached Kang Chen because she viewed him as a potential source of help and advice. Kang Chen was the only Mandarinspeaking butcher in the market, and the woman came to him because she was not familiar with the layout of the stalls. We saw on a number of occasions that Chinese customers would approach the butcher for help and advice about the market, and, on occasion, about other matters such as housing and welfare services. In this brief interaction Kang Chen was typically helpful, pointing to the meat stall along the aisle where the woman would be able to buy lamb. He recommended one stall, in particular, where the stall-holders were of Pakistani heritage. In pointing out the stall, Kang Chen refers to its owner with 'a mideast Asian face', and nominates him as a 'Cha'. This is a pejorative term used by some overseas Chinese to refer to people of Indian or Pakistani nationality/ ethnicity. Wessendorf (2010Wessendorf ( , 2014 notes that differences of origin, language, religion, and so on may be acknowledged as a point of connection. Here, the reference to the stall-holder with 'a mid-east Asian face', and the nomination 'Cha' may do more than represent the mutton butcher negatively, perhaps also aligning Kang Chen with the Chinese woman, as Kang Chen positions them as sharing the same (albeit discriminatory) values. Seven minutes later the woman returned to Kang Chen's stall, apparently in some consternation following her encounter with the mutton butcher (Excerpt 2). When the Chinese woman returned to the stall she was indicating something with her hand, tapping one of her thighs with the open palm of her hand. This seemed to be a source of amusement for Kang Chen.
The organisation of the transcript in Excerpt 2 pays attention to patterns of equivalence. Following Moore (2009), we arranged the interaction to highlight ethnopoetic principles of verse analysis. This required some initial decisionmaking. First, we introduced line breaks based on prosody. We organised the narrative according to these 'intonation contours' (Hymes 2003: 304) in Mandarin speech. We also represent the narrative by pursuing the principle of 'equivalence', developed by Hymes (2003), following Jakobson (1960). Hymes proposed that in addition to intonation contours, verses may be signalled by a grammatical feature, such as the quotative. That is, reported speech in a narrative may signal a new verse: 'Turns at talk seem always to count as verses' (Hymes 2003: 304). The principle of equivalence implies a text that is a sequence of units which give shape to action. We can see verses in Kang's narrative by attending not only to intonation contours, but also to reported speech. In Excerpt 2, we have transcribed this conversation paying attention to intonation contours, equivalence, and parallelism. Following Hymes, we included reference to 'acts' and 'scenes'; stanzas are marked with capital letters (A); verses are indicated by spaces between groups of lines. Gestures are indicated in capital letters. The protagonists are the customer (FC), the butcher Kang Chen (KC), and the English assistant butcher Bradley (BJ). The interaction is represented in Chinese characters (to represent Mandarin speech) and Roman script (to represent English speech), and this is followed by an English translation. As before, translated speech is represented in pointed brackets.  (1941), cites a motto he said summed up and informed his work: 'Use all there is to use'. At times, Hymes had little to go on other than the transcripts of anthropologists from a former age. However, we have a rich account of communication in this market, constituted as a wealth of field notes, audio-recordings, video-recordings, photographs, interviews, and WeChat messages. Most of all, we have the invaluable experience of being there repeatedly and regularly for four months. In addition, each of the authors had experience of the markets as a commercial space before and after the field work period. Hymes (2003: 98) pointed out that 'as patterning emerges, it contributes to interpretation'. An ethnopoetic analysis enables us to view the unfolding action as a three-act play, in which the second actthe Chinese woman's encounter with the South Asian butcheris played out 'off stage'. In Act I, the Chinese woman approaches Kang Chen at his stall. In Act II, off stage, the Chinese woman attempts to purchase a leg of lamb at another butcher's stall. In Act III, the woman returns to Kang Chen's stall. In the analysis that follows we focus on Act III. Through segmentation of the transcript, we can view Act III in three scenes: • In Scene 1, the Chinese woman tells her story. • In Scene 2, Kang Chen re-tells the woman's story to her, offers her a brief lesson in the orders of discourse of the market, and makes fun of her communicative strategy.
• In Scene 3, as the Chinese woman exits, Kang Chen relates the narrative to his colleague, Bradley.

Act III, Scene 1
The Chinese woman tells a story to Kang Chen in which she has had to expend considerable effort in attempting to buy lamb. In the woman's account she went to the mutton stall and saw a sheep's head, and, finding herself unable to ask for lamb from the leg of the animal, had indicated her own leg.
In lines 1-4 the repetition of the first verse indexes the woman's anxiety about the interaction she has just experienced. In the first verse of the next stanza (B) Kang Chen seeks to reassure her (lines 5-6), but in the next verse implicitly reprimands her, asking two questions linked by an evaluative comment ('you were just there shopping'). He appears to have no truck with her complaints. However, his questions elicit the Chinese woman's narrative, and in the next stanza she elaborates, deploying metasemantic commentary (Silverstein 1993) to replay the off-stage action. In two verses she makes three points about the recent miscommunication. First, 'I don't know how to say it' indexes her feelings about her proficiency in English. Second, 'I don't understand them' indexes her feelings about her comprehension of English. Third, 'I wanted to say / I just want to buy leg of lamb' further indexes her feelings about her proficiency in English, and, more precisely, her feeling that she is unable to complete her purchase. A crucial part of the interaction here is the Chinese woman's re-creation, or recontextualisation, of her gesture to the mutton butcher as she tried to make her meaning clear in the off-stage scene. As part of her narrative she reproduces the same gesture she had deployed in Act II: she taps her own thigh as a means of relating that she had done the same in buying lamb. The woman tapped her thigh on several occasions in her narrative account, in correspondence with her narrative at lines 13, 18, and 23. As such, the gesture itself becomes a form of 'equivalence'. The gesture is recontextualised from the off-stage scene, and reproduced here on several occasions for dramatic effect.
Stanza D plays out across four verses (lines 14-24). In the first verse, Kang Chen repeats his verification that what the woman has purchased is lamb. He then, perhaps surprised at her consternation, checks that she went to the stall he had recommended. His question elicits a re-telling of her narrative in the next verses. Now (line 16) the narrator brings the scene to life through a combination of reported speech and metasemantic commentary. The first three lines (16-18) ventriloquate the character of the Chinese woman in interaction with the mutton butcher. The fourth line comments on the reported action: 'but he didn't understand it'. This metasemantic comment is structurally in parallel with the narrator's previous metasemantic comments. Adding to the woman's reported feelings about her proficiency in, and comprehension of, English, now her comment indexes her feelings about the mutton butcher's comprehension of her verbal and non-verbal communication. Reported speech continues to drive forward the narrative, but through repetition, as the woman-as-narrator voices her character in lines which are equivalent to what has preceded them. Throughout this verse, the woman taps her thigh in a physical reconstruction of the miscommunicative event. In the final line of the stanza the woman makes a more general expression of her frustration and consternation, evaluating the interaction which she has just rehearsed: 'oh my! What an effort, you know'.
Stanza E is less dramatic, and is governed by a question-answer pair, as the woman asks Kang Chen whether he sells lamb. The question is probably rhetorical, and an expression of her exasperation (akin to: 'if only you sold lamb . . .'). But Kang Chen takes the question literally, and apologises for the fact that he advertises lamb but does not stock it. Act III, Scene 1 concludes.
The story has been told, and the action replayed. But this is not the last of the narrative, as we see in the following scene.

Act III, Scene 2
Scene 2 opens in a different vein, as Kang Chen comments on the Chinese woman's story in an ironic tone, mocking her incompetent attempts to navigate the spatial repertoire of the market. Whether or not Kang Chen implies that the woman's 'body language' indexes prostitution, his joke in the second verse here suggests that the Chinese woman's bodily representation of 'leg of lamb' might position her as flirtatious. Despite relying solely on the Chinese woman's account, Kang Chen amplifies the story, ventriloquating the hitherto-silent character of the mutton butcher (lines 35-36). The ventriloquated voice of the mutton butcher performs the kind of masculinity which characterises the market: 'he thought you were asking him if you were pretty'. The narrated character of Kang Chen aligns with the narrated character of the mutton butcher, and both mock the narrated character of the Chinese woman. Key here, are the relations of the narrator to the narrated characters, and the relations of the narrator to the audience of the story (Wortham 2001;Wortham and Reyes 2015). Kang Chen positions himself in a particular way through his re-contextualisation of the action (Georgakopoulou 2007), deploying the narrative as a resource to ridicule the woman and align himself with the mutton butcher. In the next stanza the Chinese woman, somewhat indignantly, repeats the statement she had made at lines 18, 23, and 24 ('it was the meat from its body that I wanted'), as she had repeated the thigh-tapping gesture. For her this appears to serve as sufficient explanation.
In Stanza H, Kang Chen repeats three times the Chinese woman's statement that she wanted the meat from the body [of the lamb]. However, whereas for the woman the statement had served as explanation, for Kang Chen repetition serves as mocking accusation. The woman's words are reworked, re-accented, and held up to mockery. Kang Chen makes two points here: that the mutton butcher cannot understand the woman's Mandarin; and that there are other, non-linguistic means of indicating to a stall-holder which cut of meat is desired: 'I would have just pointed at it for you / and you would know'. Kang Chen offers a brief lesson -'free goods' in Goffman's (1967) termsinforming the Chinese woman about the orders of discourse in the market. Elaborate gesture may be acceptable, but where it is inappropriate simple pointing will achieve the desired outcome. Kang Chen seems to indicate that he would have been willing to assist the woman by teaching her this lesson. Gesture, the common order of discourse in this space, was deployed by the woman, but its deployment was off key. This may relate to the customer's gender, or her age, or both. Either way, Kang Chen's view is that, instead of 'using body language', the woman should have pointed to the meat she wanted to buy.
In the final part of Scene 2, Kang Chen returns to his proposition that the mutton butcher would have responded to the woman's gesture by thinking she was engaging in flirtatious performance. In the first verse of Stanza I, 'am I pretty' is multivoiced, as it represents (at minimum) the voice of (1) the Chinese woman as character, (2) the ironic / mocking voice of the mutton butcher, and (3) the parodic voice of Kang Chen. In addition, it is the voice of Kang Chen in the narrating event as he retells the story to the Chinese woman in the present interaction. Furthermore, 'you are dead gorgeous!' is multivoiced, representing the narrated ironic voice of the mutton butcher, and aligning with the narrating ironic voice of Kang Chen. Hymes (1994) pointed out that 'patterns of succession' can be ways of coming to an ending point. Here, the Chinese woman's straightforward statement, and its repetition three times in succession by Kang Chen, drive towards Kang Chen's multivoiced joke, and his uproarious laughter. On this note of mocking humour the scene is brought to a close, and the woman exits.
We have suggested that the woman's experience of the sales encounter is at least partly constituted through relations of gender. Throughout our time in the markets, we observed that an important dimension of the spatial repertoire of the environment was the performance of masculinity. This was most clearly evident in the part of the market devoted to meat sales, where a large majority of the butchers were men. Kapchan (1996) describes women's emergence into the male-dominated marketplace in Morocco, and analyses how gender and commodity relations are experienced and interpreted in women's aesthetic practices. The Birmingham meat market was similarly male-dominated. In fact, Kang Chen's wife was a rare example of a woman butcher. The stall to which Kang Chen directed the Chinese woman was staffed exclusively by men. The public performance of masculinity was often evident in the 'shout-outs' of butchers as they advertised their wares to potential customers. The performance of masculinity was also characteristic of 'back-stage' comments and jokes. A further category of significance in the Chinese woman's narrative is that of age. Although the woman makes no reference to her age, Rachel Hu's field notes categorise her as 'A Chinese woman in her sixties'. This metacomment offers an important context for the narrative. The interaction may have had a different emotional outcomefor better or worseif the woman had been forty years younger.

Act III, Scene 3
Kang Chen's third re-telling of the story is delivered to his assistant, Bradley, as soon as the Chinese woman leaves the scene. A conventional transcription of this scene in ordinary running script would be as follows (Excerpt 3): KC she see the lamb head over there being sold by the Asian men you got, you got any, any, any say Chinese yea language and she go lamb I want here, I want here! [TAPS LEG WITH HAND] I say is it somebody say, like, I'm I'm any pretty? like that hahaha! hahaha. yeah? if you know how to say it, you just say any goat, lamb meat? yeah? that, that, that easy! As we have seen, Hymes considered that to represent narrative as prose is to lose sight of its characteristic form, and to ignore this form is to misconstrue the nature of narrative and to lose sight of a common human potential. At first sight this version of the story, told in 'English', is 'a story told with minimal linguistic resources' (Blommaert 2006a: 182). Kang Chen's English is heavily accented, and hesitant. He told us more than once that his English was 'poor', and 'not very good'. However, Kang Chen performs a narrative which includes the voices of the Chinese woman as narrator, the Chinese woman as character, himself as character, and the mutton butcher as character. The narrative is not straightforwardly comprehensible without access to the other versions, but it creates a short drama, and includes a cautionary tale. The butcher transforms the interaction of the narrating event through the deployment of humour, structured through ironic reported speech. These structures are not entirely invisible in the 'running script' version of the transcript. However, they become starkly visible when we pay attention to poetic structures of repetition and equivalence. Hymes (2003: 112) notes that in some traditional oral storytelling traditions narrative is 'built on quoted speech ironically deployed'. Kang Chen's story builds on the quoted speech of the woman, ironically reproduced, as he verbally represents and ridicules her words. In the first verse of Scene 3, 'she's say you get lamb meat' refers to the Chinese woman's initial approach to Kang Chen, asking him whether he sells lamb (this is clear if we interpret 'get' as 'got'). The next verse takes us to a different scene, the interaction between the Chinese woman and the mutton butcher, as she approaches his stall and sees a lamb head. Kang Chen provides context, categorising the traders as 'Asian men'. Kang Chen's narrative gives the Chinese woman voice, deploying ventriloquation as a resource ('you got any any any'). Kang Chen-as-narrator makes an evaluative metasemantic comment on the Chinese woman-as-character, pointing out that she was speaking Chinese to a stall-holder who had no comprehension of that language ('say Chinese language'). Kang Chen continues to voice the woman's character, but now in a way that sets up the joke, as he says 'I want here', to relate that the woman had attempted to communicate her wish to buy a leg of lamb by indicating her own leg with her hand. Kang Chen recontextualises the woman's action, tapping his own leg to illustrate the story.
In Stanza K, Kang Chen-as-narrator represents the voice of Kang Chen-ascharacter. He recontextualises his own joke, repeating it for the benefit of Bradley. There is more than one layer at work here, as Kang Chen-as-narratorin-the-present voices Kang Chen-as-character interacting with the Chinese woman as both narrator and character. As in the Mandarin version of the story mockingly told to the Chinese woman, 'I'm any pretty' is multivoiced, representing the voice of the Chinese woman as character, the ironic voice of the mutton butcher ('somebody'), and the parodic voice of Kang Chen. It is also the voice of Kang Chen in the narrating event as he retells the story to Bradley. Analysing Kang Chen's narrative in lines, verses and stanzas based on Hymes' approach to ethnopoetics allows us to see what at first sight looks like a disjointed and barely comprehensible account as a highly structured piece of impromptu narrative.
Bradley responds with laughter, which may either indicate that he has understood the funny story, or that he understands that laughter is an appropriate response. He offers no further comment. Kang Chen, on the other hand, gives a final evaluative coda in Stanza L: 'if you know how to say it, you just say any goat, lamb meat? yeah? that, that, that easy!'. In relating his second re-telling of the story, Kang Chen had said to the Chinese woman that she should have pointed at the meat she wanted to buy. In doing so, he was (in between mocking comments) educating her in the ways of the market's orders of discourse. Here, in her absence, he proposes that the most basic knowledge of English is enough to get by in the market. At the same time, the narrating event of the interaction with Bradley typically unites the two butchers in a joke at their customers' expense (Wortham and Reyes 2015).

USING ALL THERE IS TO USE
Returning to Hymes' motto, in analysing this very short stretch of everyday speech we have used whatever resources are available to understand the politics of interaction in the indoor market. In 'using all there is to use' we rely not merely on an audio-recording of a single interaction, but the accumulated experience of repeated service interactions, shout-outs, banter, compliments, complaints, haggles, rows, jokes, and so on observed over several months. This is, at one level, the context in which the interaction is situated. In Hymes' (2003: 99) terms, in the generation of knowledge out of this material we 'work back and forth between details and contexts'. Following the Chinese woman's adventure in her attempt to purchase lamb, Rachel Hu asked her to reflect on the experience, and quoted her words in her field note: I wanted to buy some lamb leg but I could only see those sheep heads. So I patted my leg to show them that it was the leg meat I wanted, but they didn't know what I meant. Ah, what a big joke! The customer's metasemantic commentary provided both context and detail, emically corroborating what we heard as we listened to the audio-recording. In our analysis we are able to draw on the evidence of our eyes (recorded as field notes) and our ears (recorded as a digital audio file), and the evidence of the customer's perspective. The Chinese woman's commentary on the incident, out of the hearing of Kang Chen, mirrored Kang Chen's comments to Bradley after she had departed the scene.
One of the means by which we work back and forth between details and context is through the organisation of transcripts of spoken discourse in lines, verses, stanzas, scenes, and acts. Hymes (2003) proposed that presentation of transcripts in terms of lines and verses slows down our reading of the action and makes visible the shaping artistry of narrators. Through the segmentation of the transcript here into levels of discourse (lines, verses, stanzas, scenes, and acts) in terms of repetition, parallelism, and succession we pay close attention to narrative structures, and 'such analysis contributes to a general theory of the competence and practices involved in oral narrative itself' (Hymes 2003: 311).

CONCLUSION
In this paper, we have reviewed Hymesian ethnopoetics, developed in specific cultural contexts as a means to understand the links between narratives and socioculturally mediated ways of apprehending reality. We have further considered whether Hymes' approach has purchase in conditions of contemporary cultural diversity and complexity. We argue that adopting at least some aspects of Hymesian ethnopoetics in analysing speech in a busy city market has enabled us to identify the dramatic structure of everyday narrative, to make visible the significance of reported speech in shaping narrative, and to notice the role of repetition to alert us to the ways in which we organise narrative. Ethnopoetics offers insight in analysis of the structure of narrative. It enables us to see and hear more of what is there by segmenting narrative and organising it dramatically.
Ethnopoetic organisation allows us to see more clearly verbal artistry in the sociolinguistic practices of a butcher and his customer. Through such means, we were able to make sense of the social relations at work in the market. Within structures of repetition, parallelism and succession, we were able to point to indexicalities which underpinned the interaction, including the gendered spatial repertoire of the market, and the normative non-verbal orders of discourse. An ethnographic approach to observation of social life enabled us to link non-verbal to verbal communication. Corporeal gesture, like speech, was patterned and repetitive. Ethnopoetic analysis was alive to the succession of scenes, and to the movement of discursive action. Repetition and recontextualisation were fundamental to the structure of the interaction, and were made visible through lines of equivalence. And the overall structure of the interaction, as a three-act drama in which scenes played out both present and historical action, were also noticed through attention to segmentation. As ethnopoetics represents everyday narrative dramatically, it is organised not only as lines and verses, but also as scenes and acts. Representation in scenes and acts makes visible the poetic progression (Wortham and Reyes 2015), the dynamic nature, of the narrative. At first sight, and certainly on a first and subsequent hearing, the interaction was fastpaced, barely audible, and indistinct. Presentation in lines of equivalence, parallelism, and succession has allowed us to engage with the competence and practices involved in everyday oral narrative, and to better understand the social relations between people in the market-place. Further studies are required before we can say that ethnopoetics offers a rich seam for interrogating ethnographic data, and we can fully respond to Jakobson's (1960) imperative to 'embrace the field of poetics'. But initial evidence appears to demonstrate that ethnopoetics offers rewards in the analysis of narrative in everyday communication.

NOTES
1. Available at http://www.birmingham.ac.uk/generic/tlang/index.aspx 2. Transcription conventions: (xxxx) unclear speech ! animated tone or exclamation (.) a brief interval within an utterance (2) a brief interval within an utterance, in seconds [word] paralinguistic features and situational descriptions < > English translation of speech in 'Mandarin' A indicates beginning of stanza Act III, Scene 1 indicates beginning of scene GESTURE describes a gesture made by an interactant