Stability and hybridity in refugee legal advice meetings: Discursive structure as a resource for lawyer–client intercultural communication

This article offers new insights into the discursive structuring of legal advice communication. Drawing on interactional data from eight legal advice meetings between one immigration lawyer and several different clients concerning the reunification of refugee families, the article employs communicative activity type (CAT) as a meso-level analytical approach to reveal the dynamically structured discursive organisation of these meetings. I show that whilst the stable discursive structure of the legal advice meeting evident in these data broadly confirms existing pedagogic models of legal advice communication, three different kinds of hybridity are also evident, revealing flexible use of the discursive structure in everyday practice. I also show that this stable but flexible discursive structure functions as a resource to support intercultural communication in this immigration advice context. This finding contrasts with analyses of intercultural communication in institutional gatekeeping interactions, which have argued that discursive structure functions as a barrier. The present study demonstrates the importance in discourse analysis of considering the purpose of an intercultural interaction when interpreting the meanings and functions of hybridity in discursive structure. The CAT analysis enhances our understanding of existing legal advice communication research, and functions as a heuristic for viewing legal advice as a form of institutionally grounded intercultural communication.


Introduction
This article establishes a hitherto missing link between the existing research on communication between legal advisors and clients in face-to-face legal advice meetings and pedagogical literature on the structuring of legal advice communication. It employs an activity-type approach (Levinson 1979;Linell 2010) to analyse rarely accessed empirical data from intercultural legal advice meetings concerning refugee family reunions in the UK, opening up a broader, and yet more situated, understanding of the organisation of discourse in legal advice. The paper also explores the interculturality of legal advice communication in this context of migration, challenging existing research-based conceptions of discursive structure in institutionally grounded intercultural interactions.
I firstly show that in this interactional setting, the discursive structure of legal advice communication (LAC) exhibits both stability and a range of forms of hybridity. Secondly, I illustrate how this discursive structure functions as a resource that supports intercultural communication between lawyer and client, and I argue for a focus on the functions of discursive hybridity in this cooperative professional-lay interactional context. This contrasts with studies of institutional gatekeeping interactions such as job interviews and asylum interviews, which have focused on the meanings carried by hybridity and have characterised discursive structure or 'activity type' (Levinson 1979) as problematic for, or even a barrier to, intercultural communication in evaluative settings (Gumperz 1992;Maryns and Blommaert 2002;Roberts 2009).
I begin by situating the study within the practical context of refugee family reunion legal advice and the field of intercultural communication. Section 2 explores the nature of LAC (defined here as face-to-face interactions between a lawyer and a client, acting as an individual, for the purpose of seeking and giving advice on a legal issue 1 ) according to existing empirical research and leading pedagogical models. Section 3 outlines the methodology and explains the significance of communicative activity type analysis (Linell 2010). The empirical data are analysed in Section 4, revealing both the stable core discursive structure of these interactions and how they manifest three different kinds of hybridity, and showing how each of these features functions to support lawyer-client intercultural communication in the advice meeting. I end with a discussion of the issues arising for theory and practice.

Situating the study
The data examined come from advice meetings about the legal processes governing refugee family reunion in the UK. The spouse (or civil partner or 1.2 Interculturality in legal advice I draw on a conceptualisation of culture as a dynamic construct associated with individuals' multiple (ascribed or avowed) social group identities (Holliday 1999;Gee 2012), determined by how those identities are negotiated and made salient in interaction (Chen and Collier 2012). LAC is viewed here as inherently intercultural: a legal advisor is part of a social group of experts trained in the terminology, communicative practices and institutional workings of the law and these membership statuses are salient within, and impact on, interaction with a lay client outside of this group. This legal-lay divide is well documented in legal communication scholarship (Conley and O'Barr 1990;Sarat and Felstiner 1995) and is one dimension of interculturality in LAC.
In the refugee family reunion advice setting, other differences in social group identities (ethnic, socio-economic, language-group based) may also impact on communication, bringing in multiple potentially salient dimensions of interculturality. Given these multiple dimensions, how do the parties negotiate a sufficient degree of mutual understanding, and what impact does the structure of the interaction have? The key contention of this paper is that the stable but flexible discursive structure of LAC found in this context 94 stability and hybridity in refugee legal advice meetings functions as a resource to support intercultural communication and mutual understanding between lawyer and client across these dimensions.

Literature review: Structure in legal advice communication research and pedagogy
This section, in exploring the key themes emerging from empirical research on LAC, highlights the lack of research focusing on lay-legal discursive structure. It also suggests that empirical investigation of actual legal advice practice could inform and enrich the structural model underpinning the Anglo-Western pedagogical approach to teaching LAC skills.

Key themes in legal advice communication research
The focus of existing research on LAC is largely micro-analytic, although there are some ethnographic studies offering a longitudinal perspective on lawyerclient relationships in particular contexts (Sarat and Felstiner 1995;Masson 2012). The research coalesces around three closely interconnected key themes: asymmetry in interactional power dynamics, processes of legal-lay translation and the role and importance of relational work.
Lawyers' privileged access to a specialised legal sphere of knowledge and associated discourses generates asymmetry between legal professionals and laypersons in legal interactions (Rosenthal 1974;Gibbons 2003). Interactional dominance by lawyers in the lawyer-controlled space of the advice meeting (Baynham et al. 2018) is a primary theme, although one study in the context of US divorce cases contests the view that lawyers are always dominant over clients (Sarat and Felstiner 1995). Micro-analytic studies examining interactional or pragmatic features indicative of dominance (e.g. topic control, interruptions, questioning techniques) indicate a connection between the institutional context of advice and the interactional style adopted. Lawyers in more strictly regimented institutional environments, such as legal aid or other governmentally funded services, adopted a more authoritarian style, controlling the scope and duration of talk, distancing themselves from clients and positioning themselves as gatekeepers (Bogoch and Danet 1984;Bogoch 1994Bogoch , 1997Trinch 2001). Other lawyers, situated in volunteer-staffed advice clinics, or in the non-governmental not-for-profit sector, in contrast demonstrated a more 'participatory' approach involving a 'discourse of facilitation' (Dieckmann and Rojas-Lizana 2016: 168) and adoption of an advocate positioning to address inherent imbalances in knowledge and expertise (Trinch 2001;Dieckmann and Rojas-Lizana 2016;Baynham et al. 2018).
This participatory approach is characterised by processes of legal-lay translation at two levels. At the level of terminology, lawyers explain legal terms to clients in everyday language through 'linguistic flattening' (Dieckmann and Rojas-Lizana 2016;Baynham et al. 2018: 43). At the level of discursive strategies, Maley et al. (1995: 46) report bi-directional translation between 'the discourse of legal principles, rules and categories and the discourse of acts and personal relationships' . In lay-to-legal translation, lawyers orally reformulated clients' socially-oriented narratives into legally identifiable categories (characterised by Baynham et al. 2018: 33 as 'matching and mapping'). In legal-to-lay translation, lawyers used lay discourse strategies to bring the legal world closer to the client's world. These included using value-laden and colloquial language to affiliate with the client, vocally dramatising the scenario or possible outcomes for the client and employing softer suggestion and non-directive guidance in the manner of the 'helping professions' (Maley et al. 1995: 53) of welfare or counselling. Similar strategies are reported by others (see Trinch 2001;Masson 2012;Dieckmann and Rojas-Lizana 2016;Baynham et al. 2018).
In addition to the translation function, a lawyer's use of lay discourse strategies also performs relational work (Maley et al. 1995). Other lawyer microinteractional strategies for relational work include using pronouns to express affiliation, and explaining and justifying directives to clients in order to lessen the face threat involved (Trinch 2001). In immigration legal advice contexts, legal advisors are reported to use a 'lexis of appreciation and encouragement' (Baynham et al. 2018: 35) to put clients at ease, and to share personal anecdotes in efforts to dismantle positions of dominance (also observed in Masson 2012).
There are strong hints in two studies that a meso-level approach to analysis, linking the activities undertaken in legal advice to the range of microinteractional strategies used, could lead to a more holistic understanding of these findings. Maley et al. (1995) characterise lay-to-legal translation as part of a first, fact-gathering phase, and legal-to-lay translation as a feature of a later, advice-giving phase. Trinch (2001) also reports that interactional styles shift within meetings according to the task and the (advocate or gatekeeper) role adopted by lawyers within each task. These observations indicate that discursive structure is a significant dimension of legal advice meetings; however, neither study described here, nor others I have found, exploit the opportunities of attention to this meso-level of communication. In short, the lack of attention to structure and structural mapping (Sarangi 2010) in LAC research represents an explanatory lacuna. Yet, in a related context -namely, the Anglo-Western pedagogical literature on LAC -structure is foregrounded. For this reason, it is beneficial to consider the contributions made in that domain.

The structure of legal advice communication in the pedagogical
literature Student and novice lawyers need guidance on how to engage in giving advice. The primary pedagogic models of LAC in the UK (Sherr 1998) and in the US 96 stability and hybridity in refugee legal advice meetings (Binder et al. 1991) are both underpinned by an ideology of client-centredness, within which encouraging the client to share information about their legal issue, active listening, rapport building and involving the client in taking decisions are central. Sherr's (1998) UK-based model of an initial lawyer-client advice meeting comprises 13 tasks to be performed in three key stages: (1) listening to the client's explanation of the issue, (2) questioning the client to explore the issue and (3) advising the client on the issue. Sherr (1998: 8-9) observes that this structure resembles that of lay-professional interactions with other 'helping professionals' such as doctors, commenting that such work 'naturally tends to fall into these sequential stages' . He also advocates a set of 18 skills required in legal advice work, reflecting key objectives such as developing the lawyerclient interpersonal relationship by showing ease, empathy and reassurance, managing interactional dynamics by 'facilitating the client to talk' whilst 'controlling the client in a helpful manner' (Sherr 1998: 120) and reaching across the legal-lay communicative divide by 'not overusing legal terminology' (Sherr 1998: 343).
There are evident parallels between Sherr's (1998) model of an initial meeting and the key themes in the empirical research literature discussed above. However, Gibbons (2003) highlights that these training models for structured communication are not always followed in reality. He outlines the 'fuzzy' phase structure of a 'lawyer-client consultation genre' (Gibbons 2003: 139) which evidences similarities to Sherr's (1998) pedagogic model but also departs from it, particularly in follow-on meetings. This brief account hints at the flexible use in legal advice practice of an idealised model of discursive structure, but no detail of discourse practices is given and the study from which it draws is unpublished. 2 There is thus a need for further research to explore how the discursive structure central to the pedagogical models relates to the reality of legal advice practice.

Methodology and analytical framework
This paper emerges from a linguistic ethnographic case study of LAC in the UK asylum and immigration context, fieldwork for which was conducted in 2015-2016 following ethical approval. Linguistic ethnography is characterised by qualitative case study research on language use in a specific social site. Ethnographic and linguistic/semiotic analysis methods are combined in order to investigate empirically both communicative practices in the site of interest, and the context(s) that shape, and are shaped by, such practices (Rampton 2007;Copland and Creese 2015). reynolds 97

Data collection and transcription
The fieldwork involved seven months of participant observation within an English city-based not-for-profit advice service offering free or low-cost advice on various legal matters, including refugee family reunion. Whilst working as an administrative volunteer, I gathered data through ethnographic observations and field notes about the advice service's daily working practices. Where each party involved had given their prior informed consent, I also gathered interactional data (observational notes and audio recordings) in selected legal advice meetings between clients and Julia, the service's asylum and immigration solicitor (all names are pseudonyms). Meetings sometimes also involved interpreters and/or accompanying support workers.
Audio recordings were transcribed, and where necessary translated into English. I undertook the English-language transcription myself, as a means of becoming familiar with the data (Bucholtz 2000; see Appendix for transcription conventions). Transcription and translation of Arabic-language data was completed by a qualified bilingual research assistant 3 and then subjected to my detailed review in conjunction with the audio files. In the transcription and translation process the Arabic-language data have gone through a further layer of transformation, in that Modern Standard Arabic is used in the transcripts rather than the spoken Arabic varieties that individuals were speaking (Sudanese by clients, and Libyan by the interpreter). Analysis is primarily based on the English translation, 4 but the Arabic transcription is provided where applicable, to respect the form of the original data to the fullest extent possible and to increase its accessibility to readers of Arabic. The detailed analysis that follows is based on eight meetings concerning refugee family reunion, averaging 42 minutes in length each, drawn from a larger corpus.

Communicative activity type
The discursive structure of the meetings was analysed by operationalising Linell's (2010: 42) construct of 'communicative activity type' , or 'CAT' . This construct is grounded in a broader theoretical literature concerned with the operation of communicative norms within recurrent social situations (Levinson 1979;Briggs and Bauman 1992;Sarangi 2000Sarangi , 2010Bauman 2006). Such communicative norms give rise to a stability in discursive structure that is particularly evident within purpose-driven institutional and professional interactions (Sarangi and Roberts 1999) such as job interviews or medical consultations.
My selection of the CAT construct in preference to other available constructs describing routinised communication 5 is driven by its focus on commu-nication taking place through interaction, including but not limited to the oral mode (Levinson 1979;Linell 2010), and Linell's specific theoretical approach to CATs. Linell emphasises the dialogic nature of CATs (Bakhtin 1981): i.e., meaning is dialogically created not only in-the-moment through the parties' engagement with each other, but also by individuals drawing on, and reaffirming or contesting, historically established patterns of communication (Briggs and Bauman 1992). For a full understanding of a particular CAT, according to Linell (2010: 43), analysts must consider the three aspects of (1) the pre-given or brought-along 'framing dimensions' of the interaction (e.g. purpose and participant roles), (2) the brought-about 'internal interactional organisations and accomplishments' of the interaction (e.g., sequential phase structure, topic shifts and dominance patterns) and (3) the sociocultural ecology of the CAT, in terms of its history and intertextual links to other CATs or activity systems.
Importantly, Linell (2010) emphasises that CAT structures will exhibit hybridity, and that manifestations of hybridity and their meanings should be a central focus of a CAT analysis. Hybridity in discursive structure can occur, for example, through intertextuality within CATs (such as one CAT being embedded within another CAT and reproduced through performance, reporting or reprisal -Clark 1996) or through the same discourse type 6 occurring across different activity types and vice versa, in what Sarangi (2000: 2) terms 'interactional hybridity' . Hybridity as well as stability was therefore attended to in the analysis.

Analysis process
The CAT analysis of each meeting involved the following multi-stage iterative process: (1) making notes on the purpose, participants, main interactional roles and main topics of the meeting, to identify the 'framing dimensions' (Linell 2010: 43); (2) annotating transcripts to add any significant gestures or other relevant points recorded in observational or field notes; (3) segmenting transcripts into a series of component 'sequences' (ten Have 2007: 122) of linked turns at talk; (4) classifying each such sequence according to what general activity that sequence was contributing to (e.g. legal advice, small talk), deriving a picture of the different activities engaged in during the meeting; (5) identifying smaller 'communicative projects' (Linell 2010: 36) or phases (groups of sequences during which one main topic is being dealt with) within each activity to uncover the sequential structure; (6) examining micro-interactional features 7 within each phase and across phases, and as part of this, identifying the discourse type(s) used within each phase; (7) noting instances of hybridity emerging in the discursive structure of the meeting, and examining the nature and functions of such forms of hybridity.
In the next section I evaluate the stability of the core structure in the legal advice activity, and discuss the three dimensions of hybridity evident within it, using specific examples from the data as illustrations.
4 Findings: Stability and hybridity in refugee legal advice meetings 4.1 Stability in framing dimensions and core structure The set of meetings analysed involved Julia, the lawyer, meeting with a range of refugee clients in a private meeting room at the advice service and advising clients on the legal process of family reunion visa applications. The meetings thus exhibited clear 'framing dimensions' (Linell 2010: 43) identifying them as legal advice meetings. Language, in spoken and written forms, was a central resource in each encounter, and interactions took place either in English (with the client using this as a second language), or in English and Arabic (mediated by trained community interpreters).
The main activity of legal advice exhibited a stable CAT structure, outlined below. Some meetings also involved one or sometimes two additional activities, which each supported the main legal advice activity in some way. Overall, the interactions were managed by Julia: as the host of the meeting and the legally expert party, she was positioned as the one in control of the interactional space, able to initiate (or, less frequently, permit initiation by others of) shifts between phases or into different activities (Baynham et al. 2018). 100 stability and hybridity in refugee legal advice meetings As manifest in the meetings analysed, and as shown in Table 1, the stable core of legal advice activity comprises the three central phases or communicative projects of information gathering (Phase 2 in Table 1), advice on the situation (Phase 5) and advice on the next steps (Phase 8). The meetings also featured opening (Phase 1) and closing (Phase 11) phases. This structure broadly reflects the legal pedagogical frameworks (Binder et al. 1991;Sherr 1998), although there, advice on the situation and advice on the next steps are amalgamated. The central tasks of information gathering and giving advice can also be identified in some of the empirical research on LAC discussed in section 2.1. In the meetings in the data set, however, a range of additional phases were sometimes evident (Phases 3, 4, 6, 7, 9 and 10) -an aspect of flexibility in the use of the CAT structure discussed further below.
Extracts 1 and 2 illustrate how the phases within the stable core structure are distinguishable through the lens of the discourse types used, participant roles and topics in each phase. Extract 1 is from Phase 2 information gathering early on in Julia's (J) meeting with a new client, Mebratu (M) (all text is in the original language unless otherwise indicated). The discourse type in this extract -closed questions and answers in the form of initiation-response-feedback (IRF) sequences (Sinclair and Coulthard 1975) -is typical, within the data set, of interaction in the informationgathering phase with first-time clients wanting to make a first visa application. For this type of query, Julia needs information about the location of family members, the status of family relationships and the availability of evidential documents, most efficiently gathered through closed questioning. Julia retains close interactional control, but the client still has an active role as the information provider.
The core phase of advice on the situation (Phase 5) is illustrated in Extract 2. Here, Julia is advising the Somali client, Aamina (A), on the reasons why her husband has been refused a visa to join her and her children, and what her options are.

reynolds 101
Extract 2 1 J: because they (.) expect to see marriage certificates 2 A: mmm hmm 3 J: er birth certificates etc. and that's that's why it's really difficult (.) to to make an application without those documents 4 A: mmm hmm ((sniffs)) 5 J: um (4) 6 J: so you have (2)  The discourse type illustrated here is a lawyer monologue consisting of the delivery of information, explanation and advice. Julia again has interactional control, but in contrast to Phase 2, client involvement is limited to minimal acknowledgement tokens. Compared with Phase 2, information is now passing the other way, from lawyer to client. In this advice phase, Julia applies her legal knowledge and expertise (such as the cost of filing an appeal, turn 14) to the facts of Aamina's situation, and shares this with Aamina. Moreover, Julia tries to explain the gap between the (legally enshrined) cultural expectations of the UK authorities, who expect families to have official certificates to confirm their relationships, and her client's social reality (turns 1 and 3): Aamina does not have any family registration documents, and has no possibility of obtaining these from Somalia, as she subsequently explains to Julia. Legal-lay cultural identities, and national cultural identities (together with accompanying expectations and realities about documentation of family relationships) are the salient cultural differences negotiated here. Accompanying this cultural translation (Carver 2014), a clear 'discourse of facilitation' (Dieckmann and Rojas-Lizana 2016: 168) is evident, as Julia not only phrases the problem in layperson's terms, but also linguistically accommodates to the second-language English of her client in her advice, using linguistic flattening and translation strategies (e.g. frequent pauses; comprehension checks at turns 8 and 16; simple phrases like 'it costs money' in turn 12; and reformulation of references to legal processes in different terms to facilitate understanding at turns 6, 8 and 10). Thus, lawyer strategies to communicate interculturally are evident in a number of dimensions of difference.
Phase 8, in which advice on the next steps is offered, typically also features a lawyer monologue and a passive role for the client, but with a topic focus on future tasks and actions instead of the current situation. The purposes and lawyer-dominated discourse types of these two advice phases in the data closely resemble what is expected in the pedagogic literature. Some use of specialist legal terminology is inevitable here, but lawyers are advised to keep it to a minimum, to explain legal terminology and processes in an accessible way and to check for client understanding (Sherr 1998: 56-60). In this, the discourse type resembles professional-lay advice giving in other settings (Heritage and Sefi 1992), and importantly already encompasses several known linguistic strategies for negotiating understanding in intercultural encounters (Bremer et al. 1996). These strategies are aimed at addressing the legal-lay divide but are also helpful in negotiating other dimensions of interculturality.
The two-way exchange of information, enabled through the discourse types used in these core Phases 2, 5 and 8, is a further part of how the discursive structure of LAC supports successful intercultural communication. The goal of LAC is to achieve mutual understanding between the lawyer and the client about the client's legal issue. This requires the exchange of information about, and the development of adequate shared understanding of, both the client's context and situation (achieved through information gathering) and how UK law applies to this situation and future actions to address it (achieved through advice).
In the account so far, two things are evident: the core discursive structure of the LAC is stable, and this core structure and its interactional characteristics support intercultural communication. In addition to the stability described above, however, the legal advice meetings exhibited a 'fuzzy' (Levinson 1979: 368) and flexible use of the CAT structure. Three different manifestations of this hybridity are explored below, with a particular focus on the functions of each of these hybridities in supporting successful lawyer-client communication.

Information gathering: hybridity through strategic use of multiple discourse types
In Phase 2, the lawyer aims to obtain all necessary information from the client to enable advice to be offered. The pedagogic literature recommends that lawyers invite clients to tell their story in their own words, using open questions to facilitate a client narrative that the lawyer should listen to actively, before clarifying missing facts, inconsistencies or details with the client through the iterative use of follow-up closed or open questions (Binder et al. 1991: 41-77;Sherr 1998: 12-43). The expected pattern is the mixing of two discourse types: reynolds 103 open questions and responses, and closed question-and-answer sequences. Extract 1 showed, however, that this approach was not always adopted.
In fact, Julia only used the expected pattern of open question inviting a client narrative in one out of the eight meetings; this was the one she entered with no prior information about the client's needs. Generally, Julia had some idea of the nature of the issue (this was recorded on a form by reception staff on booking the appointment and passed to Julia; for returning clients, Julia also had records of previous meetings and correspondence), and the discourse type that Julia chose to use in information gathering varied according to this. Extract 3 below follows directly on from Extract 1, in which Mebratu stated that his was a traditional marriage. It evidences a shift in discourse type when Julia's prior information is revealed to be incorrect. Turns 1-7 show Julia continuing the closed question IRF sequences observed in Extract 1, working on an assumption (evidenced by Julia's raised voice in surprise at turn 7 when a certificate is produced) that Mebratu's 'traditional' marriage is undocumented. This, however, shifts when Mebratu, recognising that he is not being given an opportunity to explain his situation, interrupts the interactional pattern with an interjection (turn 8). Julia quickly cedes control of the talk, and in a short narrative Mebratu reveals his actual situa-tion of needing advice on a refusal (turns 9-16). This client exercise of agency prompts Julia to ask (turn 17) for the visa refusal decision letter -the key piece of information needed to assess the legal situation -and on its production a discourse type of silence, accompanied by reading of a document, begins (turn 20). This discourse type is typical of the information-gathering phase where clients have had applications refused; the refusal decision letter becomes the primary source of information, sometimes supported by ancillary questions (as in turns 21 and 22). The data from the information-gathering Phase 2 illustrate that a multiplicity of strategically used discourse types can occur within this phase or communicative project, contrasting with analyses which equate one communicative project with one discourse type. The function of this hybridity is communicative efficiency: Julia strategically chooses whatever discourse type, or combination of discourse types, will be most efficient to achieve the sub-purpose of information transfer from client to lawyer according to the information available to her. Importantly, her strategic choices are intertextually informed: they depend on information gathered through prior interactions between the client and the advice service (such as the client's linguistic profile and interpreting needs, noted when an appointment is booked), and also on the procedural stage of the client's legal matter and availability of key documents.
Thus, we see that communicative forms used in interaction are likely to be connected not only to the immediate micro-interactional context within the meeting, but also to the wider series of activities and events that the legal advice meeting is a part of. Connected events were brought into the meeting through written entextualisations thereof (Bauman and Briggs 1990), such as the visa refusal decision letter or appointment booking form. Where key documents were available, the reading of these became a discourse type in itself, functioning to orient Julia quickly to the temporally and spatially distant, but topically proximate, events and texts comprising the trajectory of the client's legal issue.
In relation to Linell's (2010: 43) foregrounding of the sociocultural ecology of CATs, the data thus highlight that a legal advice meeting should be analysed as one event in a longer intertextual chain of linked events (Rock et al. 2013). This focus is reflected in existing studies examining the longer-term lawyerclient relationship (Sarat and Felstiner 1995;Masson 2012) but is missing from many micro-interactional studies of LAC (cf. Baynham et al. 2018).

Hybridity through iteration and recurrence of phases
Hybridity is also manifest in the selectively evident additional phases (3, 4, 6, 7, 9 and 10) within the legal advice CAT structure, marked 'Possibly' in Table  1. These illustrate the 'structure potential' (Hasan 1989: 64) of CAT structures: any model discursive structure will have some elements which are obligatory, reynolds 105 some optional, and some which are iterative (Gibbons 2003: 11), leading to variable manifestations in practice. The six additional phases were not evident in all meetings in the data set, but were important elements within some. They were characterised by fuzziness of their boundaries, with talk often shifting back and forth across adjoining phases (represented by semi-circular arrows in Table 1). They represent a second type of hybridity in this CAT structure.
These phases and their iteration served a function of enhancing communication. For example, in initial advice within information gathering (Phases 3 and 4), the immediacy and impact (the pragmatic force) of advice was increased by Julia interjecting it immediately in response to a relevant piece of information from the client, before returning to information gathering. Further, the question phases (Phase 6 and 9), which often overlapped with the preceding core advice Phases 5 and 8, enabled the parties to clarify and negotiate understanding of the advice given through the lawyer monologue. This is illustrated in Extract 4, in which Julia gives Ahmed (A) advice about evidence needed to support his wife's visa application. This extract also features an interpreter (Int.)

Extract 4
Original language Translation to English 1 J: in order to-to bring your wife, you'll need to have as I said her passport 2 A: mmm 3 J: um your marriage certificate translated In this extract, the shift from advice Phase 5 to the question Phase 6 occurs when Ahmed questions what sort of evidence is needed (turn 5). Achieving mutual understanding across the linguistic divide about what 'evidence of contact' Ahmed has with his wife, and what is acceptable to the UK authorities, is complex and extends well beyond this short extract. When family members keep in touch by phone, the evidentiary ideal is to provide copy itemised statements (contextually cued by Julia with the words 'phone bills' , turn 7) of a phone associated with the caller, on which specific details of calls made to the family member(s) can be highlighted. Statements of this kind are not obtainable from some card-based telephone payment services, however, and 'calling card' services (where the customer purchases a calling card with a PIN enabling access to a certain amount of talk time and accesses the service by dialling a central service number) are considered poor evidence, because the caller cannot evidence the number that he or she has been connected to. Julia and Ahmed, via the interpreter, try to negotiate shared understanding of what sort of phone service Ahmed uses, and what sort of evidence he can provide, from different cultural positions (established member of UK society vs. recently arrived refugee) associated with different phone usage and payment practices. Also salient here in the legal-lay dimension is the lawyer's awareness (and client's lack thereof) of the legal requirement for documentary evidence of a continuing relationship. Both parties use a series of clarifying questions, involving key words (e.g. Ahmed's use of kuruut alrasiid -'mobile cards' , turn 13) as contextual cues, to relate the other's description to their own sphere(s) of experience. They negotiate a shared understanding, across both dimensions of interculturality, of what is needed for the application. In this process the interpreter's mediation between these different spheres of experience, by careful interpreting of key phrases, is central.
Advice is still being delivered in the question phases, but the client becomes an active party in the talk, bringing a change in the interactional dynamic and discourse type. As well as the clarification function, the hybridity evident in these question phases can also enable a significant client-initiated topic shift, since the client has the chance to raise matters of concern to him or her. Thus, it can also serve to enhance client agency and participation. Importantly, client topic changes are oriented to, rather than dismissed: clients are viewed as legitimate contributors to the talk in LAC. This is emphasised in the pedagogic literature for two reasons. First, a client topic change will often be relevant to the advice because 'clients rarely if ever reynolds 107 stray onto irrelevant material' (Sherr 1998: 28). Second, lawyers should attend to client concerns even if they are not legally relevant, because doing so will help clients to feel supported and to focus better on the legal aspects of the issue (Binder et al. 1991). The flexibility that these optional iterative phases offer supports intercultural communication by opening up spaces in the interaction for the negotiation of understanding across legal-lay and other salient cultural differences, and for client voice to be heard, facilitating genuine exchange and mutual understanding.

The client's decision: interactional hybridity through multiple role
performance A third dimension of hybridity in the legal advice CAT structure was observed in Phase 7, the client decision phase: the performance of multiple roles by the lawyer. In this phase, Julia's adoption of a passive, client-supporting position accompanies a more counselling-focused discourse type, involving emotionally oriented talk and relational work. Extract 5 below is from a meeting in which the client, Khalid (K), must decide between two difficult options following visa refusals for his wife and five children. He can either lodge appeals against all six refusals (a very costly and lengthy process) or make a quicker, new application for his wife and younger children, whilst appealing the refusals for his two eldest children who have reached 18 and too old to apply again (cheaper, but risking the family becoming separated if -as is likely -the new applications are approved before the appeals are determined). Julia explicitly positions Khalid as the decision maker with the words 'SO: (.) it's up to you' , and then provides interactional space for Khalid to voice his emotions.

Extract 5
Original language Translation to English 108 stability and hybridity in refugee legal advice meetings The counselling-oriented talk continues for some minutes, interspersed with further client questions in an overlap between Phases 6 and 7, as Khalid reflects on, and eventually reaches, a decision with emotional and informational support from Julia. The interaction in these two overlapping phases bears resemblances to the 'interactional hybridity' (Sarangi 2000: 13) of genetic counselling talk. Sarangi argues that in genetic counselling, activity types of troubles telling and service encounter, and discourse types of advising and informing, are conflated.
Here, also, the lawyer identity involves the performance of several interactional roles -information provider, empathic listener and counsellor -in a delicate balancing act aimed at supporting the client whilst ceding the decision to him. There is a layering of 'frames within frames' (Linell 2010: 53) here, with other activities embedded within the legal advice activity, generating the kind of discursive hybridity reported by Maley et al. (1995). A key function of this role hybridity is that it aims to counter asymmetries of knowledge and power in the lawyer-client relationship by building a relationship of trust and mutual engagement (Binder et al. 1991;Sherr 1998). This function is also important in intercultural interaction more generally (Spencer-Oatey 2008), further underlining its particular value for the present data, which features more than one type of intercultural divide.

Activity types in intercultural interactions
This study has uncovered an important theoretical issue for discourse analysts of institutional intercultural communication. Activity-type structures have been characterised as culturally specific, and use of them as a communicative behaviour signalling cultural belonging (Gumperz 1992), because the communicative norms they encode are part of a particular social context, familiar to those who frequent that social context. Take, for example, a game of football (soccer): all the players will normally know, and try to abide by, the rules. Moreover, in any given performance of an activity, departures from the generic form into hybridity, creating what Briggs and Bauman (1992: 149) term 'intertextual gaps' , carry social meaning for members of the cultural community -submission to or contestation of authority, for example. Imagine that one football player deliberately handles the ball: all players would recognise this as an offence, because they all have the interpretive and performative resources to interpret the intertextual gap in the same way. Within a social group, therefore, hybridities carry meaning.
In intercultural encounters, however, interactional purpose becomes key to the significance of hybridities. Imagine that the football game is a trial for a place on a team, but the offending player does not know the rules, and is reynolds 109 instead playing according to the rules of American football: (s)he is unlikely to win the place. In such an intercultural gatekeeping interaction, where evaluation against a (social, cultural or legal) norm is the primary purpose, there is less room for flexibility and the negotiation of mutual understanding across cultural gaps. In fact, the meanings attached to intertextual gaps between the insiders' expectations and the outsider's performance are often of key relevance to the evaluation, with exclusionary effects (Gumperz 1992;Maryns and Blommaert 2002;Roberts 2009).
In contrast, in co-operative interactions with an expert-lay advisory purpose such as refugee legal advice meetings, differences may be expected and the meaning of any hybridity may be oriented to differently. If the purpose of the football game were to teach the newcomer the rules, the function of the handball offence as a learning opportunity would be as important as its meaning to the players. The findings of this study demonstrate firstly that these intercultural legal advice interactions are relatively structurally open, in that they feature hybridity and flexibility in discursive structure (an openness that as noted above has also been identified in the expert-lay advisory context of genetic counselling -Sarangi 2000); and secondly, that participants orient to such hybridity non-evaluatively, unlike for gatekeeping interactions. In sum, the functions that different forms of hybridity perform in these intercultural communicative environments of professional mediation (Sarangi and Slembrouck 1996), such as improving communicative efficiency or enabling interlocutors to develop a more trusting relationship, are as important as the meanings that they carry -a point which analysts of institutional intercultural communication focused on gatekeeping situations have so far overlooked.

Discursive structure as a resource for intercultural legal advice communication
The analysis has demonstrated both the core stability, and the inherent hybridity, of the legal advice communicative activity type (CAT) structure underpinning refugee family reunion legal advice meetings. We have seen that this CAT structure, empirically evidenced for the first time, comprises three stable core phases of information gathering, advice on the situation and advice on the next steps, complemented by six additional phases which are selectively and flexibly used. The core phases function to ensure two-way information exchange between lawyer and client, enabling the negotiation of mutual understanding (supported by lawyer use of various strategies of linguistic and discursive accommodation) about the legal issue and options for action, whereas the additional phases function to enhance communication within an associated core phase, or to foreground the client voice.
What has not been foregrounded before is the way that hybridities in the CAT structure, of three different kinds in these data, also support lawyerclient communication in interactions where a range of different cultural identity positions (in addition to legal and lay) are made relevant. The analysis demonstrates how the lawyer in this case draws flexibly on the guiding discursive structure, and orients to the functions of features of hybridity therein, to support the negotiation of understanding across these different positions.
The CAT analysis undertaken thus reveals the stable but flexible discursive structure of LAC to be a central resource for successful intercultural communication across legal-lay, and other, dimensions of cultural difference. In the context of the refugee legal advice meetings analysed, the structures and strategies which support legal-lay communication were vital in bridging other salient forms of interculturality -for example in resolving differing understandings and expectations about necessary evidence grounded in different sociocultural realities, or in relation to a language gap with a second-language speaker.

Conclusion
It was noted at the start of this paper that there was a lack of attention to structure and structural mapping in LAC research, representing an explanatory lacuna. The schematic map provided by this meso-level analysis of LAC functions as a tool through which to better situate and interpret discrete, focused findings from micro-interactional discourse studies of LAC and relate them to the wider literature and pedagogical models. Moreover, through this analysis new insights have arisen regarding the utility of conceptualising and analysing intercultural interaction at this intermediate level of the organisation of discourse. These insights were facilitated by the choice of CAT analysis, used for the first time on legal advice discourse, and the focus on an interactionally complex situation, in which more than one dimension of interculturality contributed to the range of needs of client and advisor.
The findings will be of use to legal practitioners, legal education scholars and those interested in the discourses of professional mediation more broadly. Two points are of particular importance for the legal profession. First, in structuring LAC, lawyers need to allow for both a stable core structure and communicative flexibility in order to facilitate genuine exchange and the negotiation of mutual understanding. This is something that is not sufficiently emphasised within the pedagogical models. Second, it is the responsibility of the lawyer, as the power-holder in this asymmetrical communicative context, to manage the interaction successfully (Bremer et al. 1996;Linell 2010). Both points underline the importance of good communication skills and awareness training for legal professionals.  (3) longer pause or silence (number represents duration in seconds) : sound stretching (prolongation of sound) -cut off word (part of word only spoken) UTTERANCE utterance spoken more loudly than surrounding utterances°u tterance° utterance spoken more softly than surrounding utterances >utterance< utterance spoken more quickly than surrounding utterances Notes 1 Legal advice given in writing, and legal advice sought by clients acting as representatives of organisations or companies, are outside the scope of this paper. 7 These micro-interactional features include 'topics, turn organisation and feedback patterns, topical progression methods (e.g. question designs), dominance patterns, participant positionings, degree of (in)formality, and the role of artefacts' (Linell 2010: 43).