The 'Speech Act Method': Studying Power and Influence in Conversation Interaction and a Critique of Conversation Analysis

This article presents a method for studying influence and power in conversation interaction. The 'speech act method' categorises conversation into actions that influences participants objectives, perceptions and actions. It is argued that influence is a process where an individual changes his or her intentions and actions on the basis of the likely reactions of others. Power is understood as the capacity to effect action. The method takes three stages. The first stage requires that the researcher create an account of interaction among participants. The second requires that the researcher consider the consequences of these interactions for individuals and groups. The third requires that the researcher theorise upon the objectives and perceptions that form the motivation for participants' actions. This article also examines a commonly used method for studying conversation interaction: conversation analysis. It is argued that conversation analysis is premised on a number of erroneous assumptions that hinder the study of power and influence in interaction. A note on gender language: ‘He’ and ‘She’ have been used separately and alternately to denote a hypothetical person. I felt other methods for talking about a person without discriminating on the basis of gender interrupted the article’s flow. A Methodology for Studying Influence and Power in Conversation Interaction In this section I conceptualise power and influence within a theory of interaction informed by the symbolic interactionist perspective. Interactionists posit that individuals act toward things and other humans on the basis of the meanings that those things have for them (Blumer, 1969, p.2). I would argue more than this; individuals act toward things and people on the basis of their personal objectives and the likely reactions of significant others. Although this implies that all actions are based on reasoning, this is not to say that one always reasons before each and every action. Within a given setting one may learn to take an action assuming that a particular objective will be achieved in doing so. That is one learns to take an action for a reason without actually examining that reason. In deciding how to act, an individual must first hold to a notion about what it is that he wishes to achieve. That is he must have an objective or objectives. These can be fleeting or long-standing but either way the individual uses the objective to choose an appropriate action. That is an individual assesses the suitability of an action for achieving the objective in question. In carrying out this assessment the individual speculates on the likely reactions of significant others to the action he might take. A situation may arise where the individual considers that taking an action to achieve one objective may invoke a reaction in others that would deny him a more important objective. Consequently he may seek an alternative setting where both objectives can be achieved, a new objective in itself. If he decides to stay where he is, he will have to decide which of the two objectives he will orient his actions to. This shows that one's consideration of others reactions can lead to one altering one's objectives as well as one's actions (see Blumer, 1969, p.8). I can now offer a definition of influence: the process where an individual changes his intentions and/or his actions on the basis of the likely reactions of others. In considering the likely reaction of others an individual will create an idea of what those others' objectives are. She will also estimate the capacity or power of individuals for certain types of action and reaction. This may involve her assessing their physical and financial capacity, their social support networks and their access to material things, e.g. weaponry. She will also take into account how others have acted in the past, whether that information be gained through first hand experience or through a secondary source. Influence and power are linked such that an individual may change his objectives and actions if he observes: • A change in the intentions of certain powerful individuals. • A change in the powerfulness of an individual. • The absence of particular powerful individuals or the presence of new ones. Individuals may attempt to influence others to carry out certain types of action. For example an individual whom we shall call B might convey to A that if she does not carry out a given task then he will deny her a certain important objective. For example a slave driver may threaten a slave with death if she does not work harder; a senior officer may threaten to overlook a junior officer in the next round of promotions if she does not alter the detail of her report. The individual may be successful in influencing others. A may act to achieve the objectives of B so that B does not deny her the important objective. However individuals can be unsuccessful too. This can happen when A does not seek the objectives that person B threatens to deny her. For example the slave may decide that it is better to die than to live, and choose not to work harder. A may not buy into the perception of power that person B is putting across about himself. The junior officer may perceive that the senior officer has little influence on who is promoted and may therefore refrain from altering the report. Social interaction is fundamentally about how we are influenced by one another in day-to-day life. In order for social interaction to occur, an individual must hold objectives and at the same time perceive others as holding objectives and as having the power to act and react. A methodology for studying power and influence in interaction should pay attention to this. Denzin (1970) has argued that research in the symbolic interactionist tradition should: • Focus on the meanings of participants that inform and are emergent from interaction. • Investigate how participants negotiate conceptions of self and rules of participation with reference to their setting; • Attend to both stable and emerging forms of interaction. However the preceding discussion also suggests that researchers should pay close attention to the objectives and intentions that participants bring to the setting being studied. It is also useful to explore the relationship between participant's perceptions of one another and their generation of objectives and actions. This article combines these elements in a three-stage research method for studying power and influence in interaction. Before method is presented I show how a commonly used method for studying conversation, conversation analysis [CA] is premised on a number of erroneous assumptions that hinder the study of power and influence in interaction. Critique of Conversation Analysis CA was developed in the early 1960s from within the sociological position of ethnomethodology. Ethnomethodology seeks to understand how every-day social routines and practices are accomplished and produced in particular settings. CA is premised on the notion that talk has a natural organisation that shows different features in different settings. Analysts seek to discern this organisation and find the procedures people use to produce it (Psathas, 1995, p.2). They have argued that when people talk, that talk contains features that influence how subsequent speakers talk. Therefore talk is considered as context shaped, both made and understood in relation to what has been said previously (Drew & Heritage, 1992, p.18). It is also considered as context renewing, that is contributing to the context against which further utterances are constructed and made sense of (Drew & Heritage, 1992, p.18). Analysts have also sought to understand how people interpret the meaning of others talk in terms that are relevant to the task they are trying to achieve (Hutchby & Woofitt, 1998, p.39). Proponents of CA present their approach to studying everyday language use against a critique of 'all efforts to classify and code units of interaction' (Psathas, 1995, p.8; Zimmerman & Boden, 1993, pp.5-6; Wilson, 1993, pp.26-27). Analysts assume that using pre-constructed concepts, often of a quantitative nature, to study the field of interaction leads researchers to ignore those socially meaningful features of the 'local context' that cannot be predicted a priori (Psathas, 1995, p.8). Most go further in claiming that any use of preconceived sociological categories necessarily militates against the production of grounded research and pre-determines the resulting analysis (see Wilson, 1993, p.27). For example Heritage (1997, p.163), commenting on the study of talk in institutions, rejects the notion that 'pre-existing institutional circumstances... enclose interaction' preferring the view that, ''context' is both a project and a product of the participants' actions'. Conversation analysts not only refrain from imposing social categories and theories but also claim to clear their mind of any pre-formulated notions when analysing the data; i.e. they operate an analytic mentality of 'unmotivated looking'. They argue this allows them to adopt 'an openmindedness and a willingness to be led by the phenomena of study’ (Psathas, 1995, p.21). Analysts focus on what participants 'orient to in their discussions' (Stokoe & Smithson, 2001, p.219). Conversation analysts have studied people orienting to a number of different features: 1. Particular sequential orders of conversation interaction. 2. The everyday tasks in which they are involved. 3. The identities and characteristics of others and themselves. 4. Social categories. This occurs when a participant in a conversation seeks to impose an interpretative context through the introduction of a key term (along with its implicit inferences and connotations). Take for example a group discussing what should be done to a young boy who has committed a crime. A sympathetic person may describe the boy as a disadvantaged child as a means of getting participants to focus on the social reasons behind the actions. He is then said to orient to the category of disadvantaged children. CA stresses that if one wishes to argue that participants are orienting to a particular feature, that feature 'must be shown to be relevant to the participants' (Stokoe & Smithson, 2001, p.219). Participants' orientations must 'be demonstrable in their actions, in what they actually say and do, and not assumed by the analyst' (Psathas, p.95). However the discursive practices of 'attending to', 'making relevant' or 'orienting to' remain unspecified in CA (Stokoe & Smithson, 2001). A debate has arisen about whether participants have to explicitly mention a term for it to be demonstrated that they are orienting to it (see Stokoe & Smithson, 2001). This debate arguably reflects analysts struggle to recognise that CA is concerned with two distinct social processes. The first is about how participants explicitly introduce terms into a discussion to socially construct events, problems and solutions. The second is about how people organise themselves and attempt to organise others' participation during conversation. Often people do not explicitly mention how they are organising themselves or others. Hence in this area it is common practice in CA to argue that participants understand events in terms they do not speak about. According to Schegloff (1999, p.570) 'CA’s insistence on ‘relevance to the parties’ does not need to be met by showing the parties talking about the thing whose relevance is in question, but by showing that the parties are oriented to it in doing whatever they do.' The concept of adjacency pairs illustrates this. The concept describes that when a first person starts a conversation with a second, then the first person expects that the second will say something consistent with their opening utterance. The second person is aware of this expectation and usually fulfils it by producing a second utterance that relates to the first. If the second person does not do this then the first will say something to highlight this aberration (see Psathas, 1995, p.18; Hutchby & Wooffitt, 1998, p.42). The examples used by analysts in demonstrating participants' orientation to the order of adjacency pairing do not show participants talking about the term. No first speaker ends their opening utterance by saying something like, 'Now I have just completed the first part of an adjacency pair I expect you to produce a second part that is relevant'. It is of course feasible that the first speaker would hold such an expectation, albeit an ineffable one but to state that this is so the analyst would have to create a term that the participant does not use. For some this type of practice contradicts CA principles for it implies an interpretative role for the analyst. Stokoe & Smithson (2001, p.223) note 'if something is implicit in conversation it is up to the analyst to reveal something that, logically, is not available directly.' Billig (1999, p.546) similarly observes that in CA 'Although participants are ostensibly studied 'in their own terms', they are not to be written about in such terms'. Some conversation analysts do however admit the use of interpretative assumptions in inferring what people are orienting to from what they say. ten Have (1999, p.39) has explained how analysts commonly operate a policy of 'analytic induction' to reach their conclusions. Analysts begin by formulating a provisional rule of interaction and when a deviant case is identified the rule is reworked to cater for it (Psathas, 1995, p.50). If there is a distinction to be made between what conversation analysts do and the methods they commonly refute, then it is that analysts explicitly formulate hypotheses only after they have started to consider the data. What remains the same however is that analysts use interpretative assumptions held prior to the analysis to formulate hypotheses about the data during the analysis. This is one small step from admitting that analysts bring pre-conceptions and categorisations to the data. Hence some conversation analysts argue they are privy to and use only the interpretative resources used by the participants they are studying to create propositions about what those participants are orienting to (see Psathas, 1995, p.49). 'Common-sense knowledge itself is treated as ahistorical and homogenous, as if all members have access to the same resources' (Stokoe & Smithson, 2001, p.226). Drew & Heritage (1992, p.18) for example claim that the analyst has an ‘intuitive sense of the particular social identities or attributes which the parties treat as significant in the course of their interaction’. Adhering to this position allows CA to maintain its claim as a naturalistic science where the analyst simply observes things as they are. It also allows it to avoid the uncomfortable fact that analysts may misunderstand what participants are orienting to and that CA is an exercise in using empirical data to test theoretical presuppositions. More importantly this position makes illegitimate any attempt to consider the objectives and perceptions of participants that are not readily suggested by the interaction data; a crucial tool in studying power and influence in interaction. A second methodological weakness of CA is that it tends to take agency away from participants. This derives from its fundamental assumption that participants interact in orderly ways and they do this by orienting to the order of interaction that is appropriate to the setting. Hutchby & Woofitt (1998, p.1) argue ‘The way in which utterances are designed is informed by organized procedures, methods and resources which are tied to the contexts in which they are produced’. Psathas writes about the ‘natural organization’ of talk (p.2) that does not depend on particular persons or particular settings (p.45). He argues that analysts can ‘find... the machinery, the rules, the structures that produce and constitute orderliness’ (p.2). Heritage (1997, p.163) talks of 'a common set of socially shared and structured procedures' that produce action. Despite these claims, CA does recognise that participants will sometimes not follow the pattern of interaction posited by the analyst (see Zimmerman & Boden, 1993, p.10). However it uses such occasions to emphasise the normative nature of the form of interaction to which the deviating participant should be orienting. For example Heritage asserts a 'fundamental theory about how participants orient to interaction' (1997, p.162) consisting of three fundamental features of conversation interaction. Although he admits that participants can depart form these fundamentals he notes only that 'participants can be held morally accountable for such departures'. Similarly Zimmerman & Boden (1993, p.10) admit that participants may deviate and evade the expectations of others but only to point out that they will be held accountable for doing so. In assuming that participation is determined by the type of interaction that is appropriate to the setting, the analyst neglects the study of how particular patterns of interaction are the result of people attempting to order each other's participation, sometimes successfully and sometimes not. Further it neglects to look at how participants actions are tied to their objectives and their perceptions of the likely reactions of others. That is CA fails to recognise that social interaction is as much about what is going on inside people’s heads as it is about what comes out of them. There is therefore a good case for arguing that conversation analysis is not an appropriate discipline for studying power and influence in interaction. As argued in the preceding section, a useful methodology would allow the exploration of the relationship between people's interaction and their respective objectives and perceptions. Next I detail how such a methodology could be encompassed in a three-stage method called 'the speech act' method. The 'Speech Act' Method The notion of the speech act was developed following Austin's observation that not all utterances are simply statements of fact, some are ‘performatives’ (1962, p.5). Austin defined a performative to be where ‘the uttering of a sentence is, or is part of, the doing of an action’. For example, a bride uttering ‘I do’ at a wedding performs the act of accepting a man to be her husband. Austin went on to claim that all utterances were concomitantly: • Locutionary: an ‘act of saying something’ (p.94, p.99). • Illocutionary: ‘an act in saying something’ (p.99) • Perlocutionary: producing ‘certain consequential effects upon the feelings, thoughts, or actions of the audience, or of the speaker, or of other persons’ (p.101) The 'illocutionary' act is of particular significance because it provides a basis for categorising conversation into acts and thus accounts of interaction. Searle (1969), a student of Austin's, (1969) argued that the illocutionary aspect of an utterance, what he called a ‘speech act’ was the basic linguistic unit of communication and meaning. He produced (1985, p. vii) a taxonomy of speech acts: • Assertives – telling people how things are. • Directives – trying to get people to do things. • Commissives – committing ourselves to doing things. • Expressives – expressing our feelings and attitudes. • Declarations – bringing about changes in the world through our utterances. Following Austin's distinction between stating and doing, Searle argued that a speech act could function as an assertive and an indirect directive or commissive concomitantly, e.g. ‘Sir, you are standing on my foot’. This led to him positing a distinction between the ‘literal sentence meaning’ of what someone says and the ‘utterance meaning’. The notion of speech acts can be usefully built into the interactionist perspective developed earlier. However this requires a departure from Searle’s position. Searle's distinction between 'literal meaning' and 'utterance meaning' suggests that when one listens, one interprets the meaning of some set of words constituting an utterance through some other set of 'words'. Or from the speaker's perspective, as Searle (1985, p.30) puts it, 'the speaker utters a sentence, means what he says, but also means something more' or 'means something else' (p. 31). Doubtless, for one to communicate to another how he, she or someone else understands a given utterance, one uses words and often words different to those uttered (see for example a dictionary). However how one actually understands is not an identical process to communicating how one understands. One understands by turning what Austin called 'locution' into non-linguistic objectiverelevant information (perceptions of the intentions and power of relevant others and objectives to be sought). To better explain I introduce the notion of 'lingual memories' (Becker, 1995; Johnstone 2002, p.133). Johnstone suggests that people make generalisations about meaning based on past linguistic experiences. Following on from this I would argue that 'lingual memory' is a set of non-linguistic, experience-based meanings linked to particular 'conversationally situated' locutions. In a conversation the listener tries to make sense of a locution by transposing the meaning it was seen to take in previous situations on to the current situation. This transposition constitutes dialectic between the meaning of the locution and the participant's contemporary objective-relevant information. The product of the dialectic is renewed objective-relevant information. So speech acts are ways of communicating how people perceive others actions. The conversation analyst constructs a speech act by considering whether the meaning invoked by the conversationally and historically situated locution indicates that the speaker has performed a relevant act. Before I proceed to detail the three stages of the 'speech act method' I should like to make clear that the construction of speech acts is an essential part of, but not the same thing as 'the speech act method'. Stage 1: Collecting Empirical Data on Conversation Interaction The first step in conversation analysis is to create an account of interaction. The best way of doing this is to tape record and then transcribe the interaction (see Edwards & Lampert, 1993). The next task is to reduce the data given by the transcripts into a concise account of the patterns and variations in people’s interactions. The data should be broken down into 'speech acts' of relevance to the research questions. The analyst should not confuse a speech act with an utterance. An utterance is initiated and completed by a change of the speaker. A speech act can comprise a sequence of utterances, a single utterance or part of an utterance (Greene, 1986; van Dijk, 1997). Furthermore a speech act should be considered as part of a wider class of acts. Although conversation interaction is often about what people say some acts within the conversation can include non-speech components or may be entirely non-verbal. It is useful to consider the exercise of splitting a transcription into speech acts in terms of variables, categories and cases. The variable would take the form of a question. For example one might ask the question, 'do local people influence local politicians during a consultation?' One would then have to decide the criteria by which speech acts could be categorised, that would allow the question to be answered. So going back to the consultation example one might set the following criteria for a speech act showing a decision-maker being influenced by a local person: The decision-maker pledges to do something not pledged to before that comment, as a means of addressing an issue raised by a local person. The analyst could then ask of everything that is said whether or not it shows 'a decision-maker has been influenced?' Each speech act being a case would be placed into one of two categories: ‘speech acts meeting the criteria’ or ‘speech acts not meeting the criteria'. If speech acts are found that do not readily fit into either category, then the conceptualisation of the field of study would have to be reconsidered. It may be that a different variable and category set, or just a different category set is required. It is often useful to create a ‘deferred cases’ category into which one can put cases that prove difficult to code. In deferring cases one can wait for further examples to crop up. By considering multiple cases of an anomaly one may better understand its nature. Once one has categorised all instances, one can perform further analysis on cases within a particular category. For example one might choose to categorise speech acts which show decisionmakers being influenced, by the type of action the decision-maker claims she will take and/or by when the action would take place. As an aside it is often helpful to start one’s analysis by identifying procedural stages in interaction, although not all interaction may be so ordered. Often interaction starts with some sort of introduction or greeting followed by a period of negotiation or information sharing before proceedings are bought to an end. Take for example committee meetings, where interaction is often characterised by the following stages: • The Chair introduces an item on the agenda and then either speaks or asks someone to speak to it. • The Chair or speaker speaks to the item and then asks other group members if they would like to ask questions or make comments. • Other members put their hands up and wait for permission from the chair before asking questions or making comments. The speaker to the item responds. • The Chair asks the group to agree on a decision and/or thanks the speaker for disclosing information. • The Chair closes the item and moves on to the next one. One can then look at how participants take different types of action at different stages along this procedure. The analyst should be careful not to ignore odd patterns of interactions or actions. A consideration of why it is that participants take odd actions can illuminate one’s understanding of why participants take commonly chosen actions. Research Avenues In this sub-section I suggest a number of features of interaction that a student of influence and power can usefully study. First, one can look at if and how anyone explicitly attempts to influence how others participate and whether they do so successfully. To do this one first has to build a picture of who is attempting to influence participation and how. One should identify speech acts that show participants attempting to order some aspect of the participation of one or many participants. One can then further analyse these cases, by asking these questions of them: • Who gives the order? • Who is the order directed at? • What subject is the person asking others to speak or focus on? • Who is asked to speak (if someone is asked to speak)? • How are participants positioned in relation to the subject being discussed? That is are participants asked to receive information on the subject, make suggestions to the local authority on the subject, make decisions on the subject, or agree with a preformulated suggestion on the subject. The next step is to look at how people within the conversation actually act. The researcher should analyse the transcripts and ask: • Who speaks? • In which part of the conversation does the person speak? • What does the speaker speak about? • Does what the speaker says relate to what other participants have requested he talk about? • In what position does the speaker place him or herself in relation to the subject? In comparing the results of these two analyses one can find out the extent to which participants attempt to order participation and the extent to which they are successful. One can also identify if anyone disregards orders. The analyst can look at when it is that people speak, what they speak (or do not speak) about, and how they speak about an issue as a result of following or disobeying orders. One can also study how participants influence group decision-making. In studying group decision-making one would need to start with a study of how group decisions are made and who makes them. I have found that the making of a group decision usually involves one person announcing that a decision has been made, or saying something which implies that a decision has been made. This is so even when groups vote, for it is usually only one person who interprets and/or announces the result of the vote. Once one has found out how decisions are made, or whom it is that makes those decisions, one can analyse when other people within the meeting influence them to make decisions. To do this one should collate speech acts fulfilling these criteria: The decision-maker refers to a point made by another to create a proposition that s/he successfully gets the group to accept as a group decision. The decision must be a new one and not have been made previously. Speech acts falling into this category may be further categorised into one of the following types: • Unintended influence, where the group decision made is not something asked for by the influencing person. • Instrumental influence, where the group decision made serves as a means of pursuing the subject raised by the influencing person. • Intended influence where the group decision made is the decision specified by the influencing person. During a meeting an individual may make a decision not on behalf of the group that is meeting, but on behalf of another group or organisation to which she belongs. One can similarly study how people influence this type of decision-making. This method of studying influence has three limitations. Firstly, you will recall that an occasion of influence depends on the decision made being a new one. The researcher's ignorance of previous decisions may cause him to mistake 'apparent influence' for genuine influence. It is thus necessary to be rigorous and genuinely committed to gathering information about what decisions have already been taken. Secondly it is possible that influence could have occurred in situations that are not accessible to the researcher: i.e. participants make decisions outside the interaction based on what was said during it. Thirdly, there may be a considerable time lag in the operation of influence. Therefore any findings on influence should be considered revisable in the light of new evidence. One can also look at how those who set the agenda of meetings respond to people who bring up none-agenda issues. Speech acts would be firstly categorised into those acts that do and those that do not show agenda setters recognising that a none-agenda issue had been raised. Acts that did show agenda setters recognising none-agenda issues could then be categorised according to what agenda setters used their recognition of the issue to do. Agenda setters might condemn the individual who raised the issue for bringing it up, diminish the importance of the issue, agree to take the issue on board, pledge to pass it on to another group or individual or state that the issue is or will be addressed in the future. Stage 2: The Consequences of Interaction In the second stage one considers the affects of what participants decide on different groups and individuals. It is up to the researcher to decide which groups or individuals should be considered. For example an academic specialising in older people might survey the affects on pensioners of the outcomes of the Cabinets deliberations on the budget. This type of research can sometimes constitute a research project in itself. However in the three-stage approach one can relate how the affects of decisions made in interaction are influenced and intended by certain individuals. At the same time one can explore how different parties influence decision-making and how the decisions they influence affect the lives of groups and individuals beyond the immediate environment of the interaction. It may be that the decisions reached affect others in ways unintended by any of those involved in the interaction. Stage 3: Intentions and Perceptions A complete analysis of influence and power requires more than just the empirical analysis of action and interaction achieved in stage 1. It also requires a theorisation on the objectives and perceptions underpinning participants' actions. This allows the researcher to provide answers to a range of interesting questions about accounts of interaction: • Why do people explicitly order participants to behave in a certain manner? • Why do people announce or imply that the group has made decisions? • Why do some people follow orders of participation in some instances and not in others? • In settings where there is an absence of explicit ordering, why do people act and react to one another in the manner that they do? • Why are people influenced by others to make certain types of decision? In considering how participants think during the interaction under observation it is helpful to consider why they took the decision to participate in the first place. This alerts one to some of the objectives they orient to during the interaction. As part of this exercise one should also seek to find out who was invited to take part in the interaction and who did the inviting. This will give an idea of who was planning to use the event as the opportunity to advance an agenda, and who was attending to respond to it. Nevertheless one should remember that intentions and perceptions are also developed during the interaction. The analyst should pay attention to how a participant’s objectives and actions change as his or her perception of the environment changes; and to how the objectives and actions of certain individuals are contingent on the behaviour of ‘powerful’ individuals. Methods A consideration of how people think requires more than just armchair speculation. A number of methods can be usefully employed: • Interviews with participants. • Participant observation. • Analysis of the transcripts (which in the third stage would involve looking for instances of people making explicit reference to their objectives and perceptions). • A review of biographical and historical data on how participants have acted in the past. • Sociological theories on social interaction. Interviews are useful because they allow interviewees to maintain their anonymity and encourage them to be more frank about their objectives and perceptions. Nevertheless, even during an interview an interviewee's self-portrayal will be coloured by a desire to present him or herself in a certain light (ten Have, 1999). Neither are interviewees always likely to have access to the unspoken intentions that underlie their more automatic, habitual or less considered actions (Wegner & Vallacher, 1977). The dynamic nature of one's thoughts, that is the way in which one's perceptions and objectives change during interaction cannot be captured in their entirety by a retrospective account. In conclusion, an interview on intentions and perceptions will not supply all the relevant information and there may be occasions where what the interviewee says is not true. However, despite its limitations the interview should be regarded as a means of explaining some of what individuals intend when they act. Participant observation is a useful way of understanding how others perceive one another. The use of participation is premised on the notion that the social world involves subjective meanings and experiences constructed by participants in social interaction (Robson, 1993). Through participation the researcher accesses these meanings through seeking similar objectives to participants in the same setting. That is the researcher lifts his, ‘idiosyncratic experiences to the level of the consensual and shared meaning’ (Denzin, 1970, p.12). Having provided an account of how participants’ objectives and perceptions underpin their actions, the analyst can take an additional analytical step. That is to theorise on how the objectives and perceptions bought to the setting by participants are borne from interaction in other social situations and are therefore related to participants’ membership of certain social groups and categories. Here's One I Prepared Earlier In this section I present an analysis of how local people, and local authority bureaucrats and local politicians interact within consultative forums. The aim is to show how one can combine findings from the three analytic stages in one holistic presentation. Consensus Building and The Management of Dissent In recent years the government has required local authorities to fulfil a number of requirements in strategic planning and service delivery. Often, a part of each requirement is to consult local people and other organisations over how a centrally defined plan should be implemented in their locality. Consequently local authorities often invite the public and representatives of other organisations to one off or a series of consultative meetings. The local authority plays a major role in 'managing' these meetings. It resources and hosts them, provides the minutes and agendas for them, appoints the chair and often speaks to agenda items. Members of the public, having been invited, attend these meetings for a variety of reasons: meeting other people, visiting Council buildings, seeking information or trying to influence policy-making processes. In accepting an invitation to participate, an individual usually assumes his or her objectives may be achieved through communicating with members of the group on the group’s terms. Hence, no matter his or her initial reasons for participating, each individual attempts to achieve the sub-objective of being accepted as a group member. Each ensures that his or her communication, participation and conversing are aligned or oriented to what is perceived to be the expectations of the group often expressed through the orders of the chair. This involves feeding in views and opinions at points where they are asked for; and on the subjects upon which it is appropriate to speak (most usually the agenda items). Local authorities often experience a tension between the need to respond to the participating public’s issues and the need to work towards both central and local political predefined agendas. Workers of the local authority resolve this tension by requesting participants to focus their attentions and contribute to discussions on the agency’s preformulated proposals. To the extent that participants wishing to maintain their status within the meeting say anything, then they speak on issues related to the local authority’s agenda. In seeking to gain a consensus workers suggest or ask of the group that they accept the proposals. Where participants speak following such a suggestion it is usually to voice their acceptance. This is not to deny the ingenuity of participants when they pledge their support or accept suggestions. Some may have joined the partnership because they wanted to see the agency's plans implemented. The point is that in such situations local people tend to focus on the issues of concern to decision-makers, rather than the other way round. Despite its management skills, the local authority is unlikely to stop a determined dissenting voice. Dissent is taken to mean the putting forward of an alternative agenda to that of the local authority or an attack on the local authority and its agenda. Participants show dissent when they think it unlikely that they will be able to get the group to tackle the issue of their desire through following the orders of participation. Where dissent arises, the local authority, directed by the objective of engendering interest in and support for its agenda, attempts to 'manage the dissent'. This means ending the discussion of the nonealigned agenda. Wherever possible this is done so not to destroy the notions that the participating public’s views are important and capable of influencing decision-making. That is, it is done to maintain the motivation of the public to continue their participation – a central government requirement. There are good reasons for expecting lead agencies to be successful in managing dissent effectively; or rather there are good reasons for expecting local people to discontinue their dissent when given a hint from the local authority. Firstly, in order for a local person to effectively assert an alternative agenda within the meeting he would have to gain the support of other group members. This would be difficult to achieve in a consultative meeting given that other members would be attending to influence the agenda of the local authority, some might be members of organisations funded by the local authority and others would be working for the local authority. Secondly, the dissenting participant may also wish to influence the local authority and its agenda. Continual dissent is likely to harm relationships with local authority representatives and lessen the chances for influence. In effectively managing dissent, the local authority is able to produce an environment in which there is little dissent. Given the lack of visible dissent, participants are likely to perceive the chances of gathering support around an alternative agenda or defeating the existing one to be low. Consequently, a participant wishing to forward an alternative agenda, even if willing to gamble his group standing through sustained dissent, is more likely to depart in silence and cease attendance. Rather than creating citizen centred government the management of dissent creates a group of government centred citizens. Remaining participants tend to be focused on or in support of the government's agenda. Participants with no prior interest in the local authority’s agenda undergo a shift, a change in what they identify as being a matter of concern (Jenkins, 1996). 'Matters assume an importance in the minds of individual citizens which they would have not possessed but for elite intervention' (Parry et al. , 1992, p.8). The realignment of participants' commitments to the government's agenda may take their attentions and energies away from other commitments (Forbes & Sashidharan, 1997). In expecting to be able to influence the government on a given subject, but in actual fact focussing on the government’s agenda, participants may find participation a stumbling block rather than a catalyst to realising their goals (Bewley & Glendinning, 1994). One can conclude that the management of dissent serves to keep in abeyance the creation of coalitions around alternative agendas. Conclusion In this article I presented a method for studying influence and power in conversation interaction. The 'speech act method' was described as a method for categorising conversation into actions that influences participants objectives, perceptions and actions. Influence was argued to be a process whereby an individual changes his or her intentions and actions on the basis of the likely reactions of others. Power was understood to be the capacity to effect action. The model took three stages. The first stage required that the researcher create an account of interaction among participants. The second required that the researcher consider the consequences of these interactions for individuals and groups. The third required that the researcher theorise upon the objectives and perceptions that form the motivation for participants' actions. This article also examined a commonly used method for studying conversation interaction: conversation analysis. Conversation analysis was shown to eschew the study of participants' objectives and perceptions of each other which it was argued hindered the study of power and interaction in conversation interaction.