Discourse Analysis of Requirements and Knowledge ... - CiteSeerX
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
Discourse Analysis of Requirements and Knowledge Elicitation Interviews Rosío Alvarez University of Massachusetts [email protected] Abstract This paper sets forth a framework for examining the discourse of interviews that take place during requirements analysis and knowledge elicitation. Research has shown that requirements and knowledge elicitation techniques are very similar if not identical in some instances. Moreover, eliciting user requirements is a critical and difficult activity of both information and knowledge based systems development. In this paper, requirements and knowledge elicitation are discursively mediated social practices. Through discourse analysis, this paper shows that narratives emerge during elicitation interviews. In particular, clients propose a “storytelling frame”, where the talk is inviting and rich, which conflicts with that proposed by analysts. A frame brings with it certain ‘rules’ about turns and topics. The paper portrays a conflict between the preferred frames of the analyst and client which points to some of the problems that characterize elicitation interviews. Explanations of why frames are proposed, resisted and sustained are explored.
1. Introduction Communication between analysts and clients is important in all phases of systems development but more so during requirements analysis. In its simplest form, requirements determination entails eliciting and encoding into the new system the requirements that clients verbalize to analysts. Similar to information systems development, expert systems development relies on what is known as knowledge acquisition. Knowledge requirements are formally defined and transferred from some knowledge source (usually human) to a computer program. It has been argued that requirements analysis and knowledge acquisition are almost identical processes [9]. Both requirements
analysis and knowledge acquisition are communicative acts that involve the elicitation of data and information from clients as domain experts or end users. As such, both methods rely on effectual communication between analysts and clients. It is no surprise that systems analysts and knowledge engineers must be skilled in communicating as well as understanding clients motivations and concerns, as research has shown [9]. Yet, one consistent finding in the literature concerned with systems development is the perennial “communication problem” between analyst and client [5, 12, 29]. While analysts working in both requirements analysis and knowledge acquisition use a variety of tools and techniques to obtain information, the interview continues to be the favored elicitation technique [1, 32]. Independent of the type of interview (e.g. structured or unstructured) this elicitation technique depends on interactional talk. Since the discovered data is, in this sense, partly a function of the talk between a client and an analyst, the study of this talk is central to the understanding of how information is captured and the client-analyst relationship in general. Research shows that interviews are recognized as a potentially rich but difficult method of data elicitation [40]. Researchers have argued that communication between analysts and users is often problematic due to issues such as cognitive limitations and vocabulary differences [1, 9]. Sutton [55] suggests that “[D]ifferent social realities bring forth different languages and users are most of the time inhabiting worlds with very different concerns and objects of interest from each other, let alone from providers” (p. 117). Yet, while much attention has been given to the problematic nature of analyst-client communication and language usage, very little of the research is based on direct observation and analysis of these interactions. This paper argues that requirements and knowledge elicitation interviews are a communicative event where narratives are produced. Further, critical discourse analysis (CDA) is used to examine the
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
1
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
narratives that emerge during interviews. Discourse is defined as a systematically-organized set of statements which give expression to meaning. “A discourse provides a set of possible statements about a given area, and organizes and gives structure to the manner in which a particular topic, object, process is to be talked about” [34, p.7]. During requirements analysis interviews, analysts, in their official roles, propose a “framing” which conflicts with the framing proposed by clients during interviews. Frame is defined as the speaker’s instructions to the listener on how to understand the message [24,56]– a metamessage about the context of the talk [57]. The article portrays a mismatch in the expectations of the analyst and client during interviews which explains some of the frustrations which characterize this type of encounter.
2. Requirements analysis and knowledge elicitation as social practice Requirements analysis and knowledge elicitation are two techniques whose concepts and processes are in some cases identical in others very similar. Requirements analysis and knowledge elicitation are given special attention here because of their critical role in systems development. Research suggests that many system failures can be attributed to a lack of clear and specific information requirements [11, 15]. Furthermore, errors during requirements analysis that are not found until later stages of the implementation process can cost significantly more to fix [4, 39]. Similarly, knowledge elicitation is frequently referred to as being one of the most difficult aspects of expert systems development [38, 59]. In its simplest form requirements analysis seeks to identify the information needed to automate organizational tasks and to support information workers in their tasks. Similarly, knowledge elicitation elicits information from domain experts and interprets this information to produce a model which evolves into the expert system used to support knowledge workers [33]. While a difference in lexicon exists, the process and end result of both methods are indistinguishably similar. Moreover, both methods demand similar skills of analysts, for instance building rapport, ability to communicate, understanding client’s motivations and concerns [25, 28]. Clearly, both methods seem to highlight the importance of viewing social context as a critical component for eliciting information. Several currents in the information systems literature argue for and analyze information systems development as a social practice [c.f. 31, 46, 47]. These researchers attend to the ‘organizational’ issues such as power, conflict and control that critically
impinge upon information systems development. Information systems researchers theorize requirements analysis as a process which is socially mediated [22, 20] where requirements are socially constructed through clients’ stories [2]. Others suggest that requirements analysis is a process fraught with conflict, inconsistency and competing viewpoints where one ‘voice’ is not identifiable [42, 13]. Other research suggests that ‘expertise’ is subjective, subtle and tacit [58]. Therefore, requirements analysis and knowledge elicitation would best be served by developing approaches that examine the social situation where information and knowledge are produced. Unfortunately, most requirements and knowledge elicitation depends on technology-driven models and methods. These approaches focus on technical issues and give little if any attention to social context. Instead, requirements analysis techniques produce a “tidy” representation of the organization that is mobile, stable, and combinable [37], thereby allowing analysts to decontextualize requirements, represent the organization in a durable form and make requirements capable of manipulation at a distance far away from the interruptions and clamor of clients [60]. The approach in this paper contributes to understanding requirements and knowledge elicitation as polyphonic, i.e. as containing multiple voices. This polyphonic linguistic interaction is socially situated, where words bring with them the context where they have lived. In particular, this paper examines the fundamental substance of requirements analysis interviews – talk. This paper examines both the formal structure of language by using methods borrowed from sociolinguistics, such as conversational and frame analysis. The analysis here also locates the individual within their social matrix by subjecting the data to criticism that examines how power, control and identity are produced and negotiated during interviews. In this manner, requirements analysis takes on the form of a discursively mediated and constructed social practice.
3. Analyzing discourse through narratives The framework for the study of requirements analysis in this paper draws on critical discourse analysis [10, 17, 18, 19] and narrative analysis [41, 45]. Critical discourse analysis is grounded in interactional talk and text data which employs sociolinguistic methods [50]. Sociolinguistics assigns special significance to the structure of speech and texts, and provides methods for specifying the linguistic features of different types of discourse units
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
2
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
and the way they are tied together into larger units of meaning. However, unlike other forms of discourse analysis, critical discourse analysis also concerns itself with examining social context along the lines of ideology, power and inequality. Through discourse examination, topics of power inequalities usually along the lines of race, class, gender, sexuality and occupation are exposed. Critical discourse analysis therefore, demystifies what is taken to be ‘common sense’ by defamiliarizing it and signaling its functions and consequences in sustaining the social order. As Fowler [21] suggests, critical discourse analysis goes “beyond the formal structure of Language as an abstract system, toward the practical interaction of language and context” (p.10). In this sense, language is seen as a social practice, as a mode of action that is always socially situated “in a dialectical relationship with other facets of ‘the social’…it is socially shaped, but it is also socially shaping, or constitutive” [19, p.131]. From this viewpoint, discourse is seen as constitutive of social reality in a general sense. Narratives are particular types of discourse that emerge during requirements analysis interviews. Research shows that during interviews people “use the narrative form as a pattern for uniting the events of their lives” [43, p.153]. Through narratives, people organize their temporal experiences and make them meaningful. Other research sees the narrative as a “politically motivated production of a certain way of perceiving the world which privileges certain interests over others” [41, p.114]. They are used to create believable explanations for the teller’s actions. To better understand narratives, theorists have developed definitions describing the structure and function of narratives. Some suggest a somewhat narrow definition where narratives are stories about specific past events [35, 36]. These stories follow a chronological sequence where order of events move in a linear way in time. Reissman [45] extends this definition to include other genres such as habitual narratives, which describe repetitive events with no peak in action, and hypothetical narrative, which describe events that did not happen. Researchers of organization and information systems have begun to employ discourse and narrative analysis in various forms to examine modern businesses. For instance, Wynn and Novick [61] examine the issue of ‘turns’ during cross-functional meetings. They find that what is considered a ‘valid’ contribution is based on discourse style. In particular listeners receive the story versus the professional style of discourse differently. Others such as Sayer and Harvey [49] study implementation discourses. In particular they examine how the use of an electronic mail system is used as a technology of power to
manipulate discourses during a BPR project. And finally, Suchman and Bishop [54] examine discourses of ‘innovation’ that function to conserve rather than change existing institutional orders. In the area of narratives, Brown [6] examines the use of narratives that organizational members produced to create meaning and gain political advantage during an IT implementation. Davidson [14] uses narrative analysis to examine sensemaking and interpretation during an IS development project. And finally, Dube and Robey [16] examine stories as a symbol of organizational culture to generate insights into the collective interpretation of management practices by competing groups during a software development project. In general, narrative and discourse analysis draws our attention to the modern business organization as a site for power struggles. Another assumption that guides this research is that the interview, as a discourse genre, is a critical site in which identity is performed. A fundamental observation of discourse analysis is that speakers’ identities emerge from discourse [7]. Drawing on Brian Street’s [52] metaphor “culture is a verb” Sarangi and Roberts [48] propose that “identity is a verb” suggesting that similar to culture, identity is something that people do. Said differently, this paper adopts a performative view of identity where a speaking subjects’ identity is produced through discourse. As feminist theorist Judith Butler [8] states, gender is a copy for which there is no original, suggesting that identity exists and is produced though its performance. The concept of identity construction is of particular relevance for interviews. Unlike sermons, political speeches or educational discussions, the interview as a discourse genre is one where the individual verbalizes thoughts, intentions, and consciousness. It is through this ritual that self-reflection, self-knowledge and selfexamination are obtained. As we confess we tell a story that constitutes a drama in which we are a leading character, and the meaning of this role is found through recollection. In other words, in narrating the past we define ourselves as the implied subject generated by the narrative. Now consider the requirements analysis interview, this act involves participants detailing actions, problems, hopes and needs in regard to work life. Through the elicitation of ‘requirements’ the subject is constituted as the object of knowledge (revealed through self-avowal) and given meaning. This performance functions to construct the identity of the speaking subject. In this sense, it is through linguistic expression that social identities and relations are produced.
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
3
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
4. Research methodology
5. Research site
This paper draws on two discursive mechanisms, footing and framing, adapted from the ideas of Erving Goffman, to analyze requirements interviews. Erving Goffman [24] uses the term footing to refer to these interactional stances or positions vis-à-vis one another. Footing defines the relationship between speakers. Speakers have available to them a number of social roles from which the can move in and out of. In one moment a speaker may be a friend concerned about the health and well-being of an employee, in another moment the same speaker may be a competitive peer vying for control of resources. As speakers communicate linguistically, they choose from a wide range of social attributes and roles that are made available to them and which they perceive as best fit to the situation. Footings are relational and shifting. As we speak, we shift footing, move in and out of social roles, all of which assist us in establishing the relationship with others such that the meaning of the utterance is understood. Another useful concept is framing. Framing defines the speakers’ instructions to the listener about what has been said and how to understand the utterance [3, 23, 56]. Through subtle signals like pitch, voice, intonation, and facial expression the contextual information is jointly created. These signals provide a frame or metamessage about the context indirectly [57]. Gumperz [30] refers to these signals as “contextualization cues” which may be prosodic, paralinguistic and non-verbal. These surface messages function to call up shared experiences, they are powerful means of negotiating social identity and legitimating preferred styles of communicating in the predominantly asymmetrical interactions in workplace settings [48]. The person proposing the frame establishes how talk is to be understood in that context [57]. In this sense attempts at re-framing, counterframing or any other out-of-frame activity will likely be resisted. Out-of-frame activity refers to lines of activity that are somehow outside of the main official line [23, p.201]. This behavior is distracting and threatens the context, which in turn may be treated by participants with active disattention so that the main line of activity can continue. The discursive mechanisms presented here integrate Goffman’s views of social interaction as involving particular presentations of self and dynamic shifts between footings and frames. This paper will analyze the self, or subject positions as Foucault and Fairclough would call it, and the manner in which these identities are produced, sustained, or discredited during requirements analysis interviews.
The organization chosen for the study is State University1 a large public research university located in the northeastern United States. The annual budget of State University exceeds $.5 billion. It has enrollments of 24, 000, faculty of 1184 and staff of approximately 3600. At the time of the fieldwork, the university had recently allocated $11M for an information systems development project. The organization was to determine its information requirements. This involved determining the mandatory and desirable requirements of a system, documenting them and then evaluating the various vendor solutions according to these requirements. The new information system was to handle all the major administrative functions of the university including, financial accounting, human resources, and all student services such as housing, enrollment, financial aid and admissions. The legacy systems that were in place at the start of this project, which we shall call the Administrative Information Systems project (AIS), had been in use for over fifteen years. These legacy systems consisted of a combination of custombuilt and vendor applications which had been extensively modified for specific university functions. The student registration system was built in-house using a System 2000 hierarchical database with some VSAM files. Information Associates, a vendor of higher education software, provided applications that handled student financial aid, bursar and financial functions. All three applications had been discontinued and were now updated and maintained in-house. Information Systems Incorporated provided the human resource application, which was also a discontinued application now updated by in-house staff. The operating system consisted of MVS ESA v4.3, CICS v 2.1.2, and OS/VS Cobol 2.4. All applications were located on an IBM 3090. A voice response system provided by EPOS was in use for student access to information and registration. Reporting tools included Data Analyzer and SAS. For this research, the author was hired by the CIO of the university to document the software selection process. This role was disclosed to all organizational members involved in the ERP selection and implementation. The case study was completed and a report was provided to the CIO who read and approved its contents. The entire information systems development project began in January 1996 and was still in implementation at the writing of this paper. The focus of this paper is on the requirements analysis 1
Pseudonym is used.
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
4
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
phase which lasted from August 1996 to December 1996. During this shorter phase, the university contracted a consulting firm to conduct the requirements analysis study and produce all relevant documentation.
6. Findings The data collection was conducted through participant observation of requirements analysis interviews. While my primary role was that of observing and documenting the selection process, my recognized technical expertise by clients and analysts often prompted questions directed at me, which required my participation in the discussion. In all interviews at least two analyst-interviewers were present. The analyst’s strategy was to have one interviewer ask question and the other function as ‘scribe.’ Conversations during interviews were taped. Data was collected at thirty-two meetings in which eighty-two individuals from eight different departments participated. The data set consists of sixty hours of tape recordings, along with field notes where tape recording was not feasible. The emphasis during data analysis was on linguistic structure and interaction, therefore it was important to include intonation, volume, pacing and other qualities of speech to capture the mood and feel of the interview. To achieve a balance between capturing all the detail necessary and providing a readable transcript, transcription conventions were based on the work of Reissman [45] and Gronn [26, 27] (adapted from Stubbs [53]). These conventions ensured that overlaps, exclamations, questions, pauses and emphasis were maintained. Symbols used in the transcribed extracts are: // overlapping talk from the first to the last slash (x) pause of x seconds [ ] explanatory note {} nonlexical utterances italics word emphasized by speaker ! exclamation ? question Speakers are identified by Name (of narrator) and by Int. (for interviewer). Open coding was used to identify frames, and examine their interactional accomplishment, maintenance and resistance. The core features of open coding are: 1) the inductive development of provisional categories; 2) ongoing testing of categories through conceptual analysis and comparison of categories with data that is already
coded; and 3) the altering of existing categories as other ones are created or eliminated [51, p.11-13].
6.1. The storytelling frame Many of the interviewees in this study followed a pattern which I call a storytelling frame. The client attempts to stay well within this frame, while the analyst, for the most part, resists and reframes the interaction. A frame brings with it certain ‘rules’ about what talk should be about (or not be about), who can talk, in what sequence and who controls the activity. In the storytelling frame the speaker tells a story or short narrative. This frame creates a more symmetrical relationship between client and analyst. Moreover, the role of information worker resides in the background while another subjectivity is developed in the foreground. The analyst prefers to resist this frame and sustain a more official role that brings with it an asymmetrical relationship; here the analyst controls the turn-taking as well as the topic of discussion by asking the questions. The following passage portrays a shifting into and out of the storytelling frame. Here two members from the housing office, Shelly and Carol, are asked questions by two interviewers, Int1 (who is female) and Int2 (who is male). They are asked about the process by which students received bills pertaining to damages of the dormitory rooms. The extract illustrates a pattern of shifting frames which produce a certain degree of conflict. 1 Int1 -- so you don't have a 24-hour turnaround? 2 Carole -- the other time when there is a little backlog is check 3 out in December. December is a real problem because they are 4 checking out and our staff wants to go out on Christmas break 5 also. So, {um}, we frequently have to badger to get those 6 invoices 7 Int1 -- so December and May are your tough times when you 8 don't have a 24-hour turnaround. What do you think the 9 processing time is in most cases? 10 Shelley – [Shelley looks a Carole] (5) it could be as much as 11 two weeks 12 Carole – yeah, and maybe in some cases depending upon staff 13 in the area offices even three weeks 14 (3) because the staff and the area offices have many other 15 responsibilities besides the billing, that 16 (3) we could badger them, but they can't do it, and we 17 /understand 18 Int1 -- the example/ you gave us suggested that at check out the 19 inventory form was completed with the student present, is that 20 not uniformly true? 21 Shelley -- that is not uniformly true. 22 You're trying to check out 11,000 people 23 in a week's period of time with 300 RA staff, who are also 24 taking finals and have 25 there are sometimes where students will check out alone and 26 waive their right to appeal, by checking out and then someone 27 will come in after the fact and complete the inventory form and 28 do a billing that's related to that
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
5
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
29 Int1 -- how often does this happen? 30 Shelley – (3) we could certainly get you statistics on how many 31 students do that, I, I, don't know I assume a lot. We could do, 32 not, not, full range statistics but take a pile of /inventories 33 Carole -- see how/ many were checked out 34 Int1 -- is that universally true, that if the student checked out 35 without the inventory being completed he waives his right to 36 appeal 37 Carole -- absolutely in all cases, so if his choice is get out of 38 here I wanna leave or wait around till someone can go with him 39 and /complete and so he 40 Int1 -- ok ok/ 41 Carole – ok, because I have been on the phone with parents 42 Int1 –[begins talking to other interviewer, looking away from 43 Carole and Shelley] 44 Carole -- I've been on the phone with parents, [interviewers 45 look at Carole] explaining that when they've gotten bills and 46 have been very upset, and very irate and indicated to them that 47 your son or daughter made a choice and when we went back and 48 inventoried the room this is what we found. 49 I tell you, as a parent it's not a very comfortable thing to hear 50 and, /and the have to 51 Int1 --they /don't accept it? 52 Carole -- not all the time 53 Shelley -- we're looking at a process to possibly not, we used to 54 bill three dollars if the student didn't leave their mattress cover 55 on their mattress and were trying to put a minimum I think what 56 we're suggesting a $25 minimum per billing… 57 Int1 -- except that the cost of mattress covers has gone way up, 58 I know [Interviewer 1 laughs, Carole and Shelley laugh] 59 Int2 – ok, so this is just one way they can get a notice, then it 60 gets on to BRS and is simply an amount due. 61Carole – [2] yes
I analyze this excerpt by examining turn-taking and moves that each speaker accomplishes through their turn. The excerpt opens with Carole responding to the question posed to her by the interviewer about turnaround time. Carole tells us a story about problems they encounter in December. She takes us back to a time when staff has competing demands to leave for Christmas break. Rather than give a yes or no answer, Carole steps into a more personal frame, the storytelling frame, changing the footing from information provider to caring woman. Through the storytelling frame she appears as a concerned woman who cares about employees desires and demands. The interviewer reframes the interview by returning to the question about processing time (line 7). By asking questions the analyst imposes the obligation of a response on the Carole. This places a constraint on Carole’s next move. More importantly, it places Carole in a follower role where the appropriateness of her response will be measured by the question that preceded. Carole hesitates for several seconds. At this point Shelley looks at Carole and intervenes attempting to shift the footing back to less personal one by answering the question more directly with a quantifiable amount.
Carole then returns to the conversation, staying in the more official frame by providing direct quantifiable information about timing, but after a pause she shifts the footing again. Carole reframes the interaction and now the professional role of information worker remains largely in the background. Instead a relationship of caring woman figures more prominently. Carol returns to the storytelling frame. She describes overworked staff that “have many other responsibilities” (line 14) and inscribes herself in this narrative as an understanding woman who could badger them but does not. Through this brief narrative she persuades us into believing that she is an understanding person. Through short narratives about daily work practices, Carole’s identity as a caring woman begins to take form. The analyst abruptly switches to a factual request for information. She now searches for uniformity (line 20), quantification (line 29) and universals (line 34) in client responses. To the first request Shelley responds with a direct answer, but then switches back to the storytelling frame by narrativizing the daily habitual difficulties of managing workers who have lives with other demands (“RA staff who are also taking finals”). When the storytelling frame is interrupted by the analysts’ request to quantify her response, Shelly stalls, then attempts to comply with the request by offering ‘statistics’. Shelley is compelled to participate and reluctant to resist the framing proposed by the analyst for fear of being evaluated as incompetent information worker. She reluctantly accepts the role of respondent, by offering her rendition of statistics, and lets go of other possible footings so as to be favorably viewed by the analyst. Carole, on the other hand, resists the frame proposed by the analyst more often by counterframing the encounter with stories. In line 40 Carole’s hypothetical narrative about what students might do at check out time is interrupted by the analyst, with “ok, ok”. At this point the interviewers begin to talk amongst themselves, ignoring Carole who has launched into another personal narrative about her interaction with parents. The non-verbal cues by the interviewers, who turn away, signal to Carole that they have transitioned their attention away from her and to their documentation. At this point the pace and pitch of Carole’s talk increases (line 44). She repeats the previous statement, regains control of the interview and reframes the interaction, back to a storytelling one. She tells a story about parents who are surprised when billed for damages to dormitory rooms done by their children. The story peaks by Carole identifying herself as a parent (line 49). This confession contributes to creating a more nuanced identity of Carole, now as a caring mother.
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
6
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
The passage ends with a change in footing initiated by the female interviewer (Int1). Throughout the interaction she has resisted the storytelling frame, by counter framing it whenever it was in place. However, at line 57 the woman interviewer responds with a bit of personal information about herself as a consumer. The prosodic emphasis on “I know” signals to the other women that interviewer has direct experience with or information about purchasing the item of concern, a mattress cover. She shares an understanding with the other women as a consumer, perhaps as a mother as well. In performing this move, the analyst steps out of her official role and Carole and Shelley momentarily leave their roles as information providers to assume that of women consumers conversing about the costs of keeping a home. This move creates an affiliation among the women as they laugh together. The interview is suspended. However, the fleeting moment of intimacy among the women fades as the male interviewer regains control of the interview and reframes the interaction. With an abrupt shift in footing the male interviewer changes the context. He reduces the complexity of concerns and dilemmas of work practices as narrativized by Carole and Shelley to “simply an amount due on the bill”.
7. Discussion and conclusion In examining the discursive construction of narratives during requirements analysis interviews, we find that shifts in framing and changes in footing create an ongoing tension between the analyst and client. Throughout the interview we observe a struggle in reframing which can be attributed to an underlying conflict between the way the analysts view the encounter and the clients view it. For the analyst the encounter is a ‘professional’ encounter, for the client it is about personal work experiences and dilemmas. Thus, both Carole and Shelley propose framings in the direction of a social encounter through storytelling. In this frame the clients’ identity beyond information worker takes form; as caring mother and co-worker. The analysts resist the storytelling frame by shifting out of the frame. This continual shift and apparent conflict exemplifies a basic difference between clients and analysts. While both are engaged in and contribute to producing an interview, clients and analysts have very different understandings of what each one’s role is. The clients approach the encounter from a more intimate angle, attempting to be more personal, funny, friendly, or even dramatic, the analyst comes at it from a very professional stance looking for brevity and directness. In reframing,
Carole and Shelley’s communicative styles change to providing more information and more involvement. By accomplishing a (re)framing Carole is able to signal the importance of larger social world issues. However, the analysts shift out of the storytelling frame and in so doing, confine Carole to a limiting and unidimensional role of information worker and no one else. The persistence of clients to remain in the storytelling frame does not seem to alter the analysts reframing of the interview. Why do clients succumb to this reframing, albeit with some resistance? Perhaps clients may at times see the storytelling frame as less ‘professional’ and fear being seen as incompetent information workers if they remain. As an example, Shelley responds to a request with the concept of statistics perhaps because she believes it will make her appear more professional, a more scientific subject. Secondly, reframing is a difficult task to achieve. It is easier to stay than to resist or react, which takes additional effort, skill and power. When the analysts reframe the interview, interrupting the storytelling frame, power relations are skewed in the direction of the analyst who possesses institutional legitimacy. The analyst is poised to control the interview, ask the questions and determine what gets talked about. In this paper, narratives produced in the discourse of requirements analysis are examined to illuminate issues of identity, conflict, power and communication ‘gaps’ during the interview interaction. This approach expands our understanding of how professional identities are constituted and relations of power are fashioned out of talk during requirements analysis and knowledge elicitation interviews. The paper has shown that clients propose a storytelling which is continually resisted by the analysts. The storytelling frame allows the clients to engage in narratives that position the client within their cultural and social environment. The narrative response provides a space where clients may embed themselves within the story and thereby provide a more nuanced image of their identity, concerns and dilemmas. We are able to see a tug of war between roles assumed and modes of talk during the interview. The client prefers the storytelling frame which frames the interview as a social encounter. Here the talk is friendly, inviting and very rich. However, the analyst opts to reframe, relegating the client to a professional identity and succinct mode of talk. Yet, the storytelling frame would inevitably bring about a more involved and animated person willing to provide information that better describes the ‘motivations and concerns’ of the worker as well as share ‘subjective’ interpretations of what they believe constitutes expertise.
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
7
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
8. References [1] Agarwal, R. and Tanniru, M.R. 1990. “Knowledge Acquisition Using Structured Interviewing: An Empirical Investigation,” Journal of Management Information Systems, Vol. 7, No.1, pp.123-140. [2] Alvarez, R. and Urla, J. (forthcoming). “Tell Me a Good Story: Using Narrative Analysis to Examine Information Requirements’ Interviews During an ERP Implementation,” DATABASE for the Advances of Information Systems. [3] Bateson, G. 1972. Steps to an Ecology of the Mind. New York: Ballantine. [4] Boehm, B. W. 1981. Software Engineering Economics, Englewood Cliffs, NJ: Prentice Hall. [5] Bostrom, R.P. 1989. “Successful Application of Communication Techniques to Improve the Systems Development Process,” Information and Management, Vol. 16, pp. 279-295.
[14] Davidson, E.J. 1997. “Examining Project History Narratives: An Analytic Approach,” in A.S. Lee, J. Liebenau, and J.I. DeGross, (Eds.) Proceedings of the IFIP TC8 WG8.2 International Conference on Information Systems and Qualitative Research, Philadelphia PA, pp. 123148. [15] Davis, G. B. 1982. “Strategies for Information Requirements Determinations,” IBM Systems Journal, Vol. 21, No. 1, pp. 4-30. [16] Dube, L. and Robey, D. 1999. “Software Stories: Three Cultural Perspectives on the Organizational Practices of Software Development,” Accounting, Management and Information Technology, Vol. 9, pp. 223-259. [17] Fairclough, N. 1989. Language and Power. London: Longman. [18] Fairclough, N. 1992. Discourse and Social Change. Cambridge, MA: Polity Press. [19] Fairclough, N. 1995. Critical Discourse Analysis: The Critical Study of Language. London; New York: Longman.
[6] Brown, A. 1998. “Narrative, Politics and Legitimacy in an IT Implementation,” Journal of Management Studies, Vol. 35, No. 1, pp. 35-59.
[20] Flynn, D. J. and Jazi M. D. 1998. “Constructing user requirements: A social process for social context,” Information Systems Journal, Vol. 8, No. 1, pp. 53-83.
[7] Bucholtz, M. 1999. “Transgression and Progress in Language and Gender Studies.” In Bucholtz, M., Liang, A.C. and Sutton, L. A. Reinventing Identities: The gendered Self in Discourse. New York: Oxford University Press.
[21] Fowler, R. 1996. “On Critical Linguistics.” In CaldasCoulthard, C. R. and Coulthard, M. 1996. Texts and Practices: Readings in Critical Discourse Analysis. London: Routledge. pp. 3-14.
[8] Butler, J. P. 1992. Contingent Foundations: Feminism and the Question of "Postmodernism", in Butler, J.P. and Scott, J.W. Feminists Theorize the Political, New York: Routledge.
[22] Galliers, R.D. and J. A. Swan. 2000. “There's More to Information Systems Development than Structured Approaches: Information: Requirements Analysis as a Socially Mediated Process,” Requirements Engineering Journal Vol. 5 No. 2, 74-82. [23] Goffman, E. 1974. Frame Analysis. New York: Harper and Row.
[9] Byrd, T. A., Cossick, K.L., Zmud, R.W. 1992. “A Synthesis of Research on Requirements Analysis and Knowledge Acquisition Techniques,” MIS Quarterly, Vol. 16, No.1, pp. 117-138. [10] Caldas-Coulthard, C. R. and Coulthard, M. 1996. Texts and Practices: Readings in Critical Discourse Analysis. London: Routledge. [11] Cooper, R. B. and Swanson, E. B. 1979. “Management Information Requirements Assessment: The State of the Art,” Database, Vol. 10, pp. 5-16. [12] Cronon, T. P. and Means, T. L. 1979. “System Development: An Empirical Study of User Communication,” Database, Vol. 10, pp. 5-16. [13] Darke, P. and Shanks, G. 1996. “Stakeholder Viewpoints in Requirements Definition: A Framework for Understanding Viewpoint Development,” Requirements Engineering Journal, 1, pp. 88-105.
[24] Goffman, E. 1981. Forms of Talk. Philadelphia, PA: University of Pennsylvania Press. [25] Graham, I. And Jones, P.L. 1988. Expert Systems: Knowledge, Uncertainty and Decision, Suffolk, England: St. Edmundsbury Press. [26] Gronn, P. C. 1983. “Talk as Work: The Accomplishment of School Administration,” Administrative Science Quarterly, Vol. 28, pp. 1-21. [27] Gronn, P.C. 1985. “Committee Talk: Negotiating ‘Personnel Development’ at a Training College,” Journal of Management Studies, Vol. 22, No. 3, pp. 245-268. [28] Guinan, P.J. 1988. Patterns of Excellence for IS Professionals: An analysis of Communication Behaviour, Washington, DC: ICIT Press.
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
8
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
[29] Guinan, P.J. and Bostron, R.P. 1986. “Development of Computer-Based Information System: A Communication Framework,” Database, Vol. 17, No. 3, pp. 3-16.
[43] Polkinghorne, D. E. 1988. Narrative Knowing and the Human Sciences. Albany, NY: State University of New York Press.
[30] Gumperz, J. J. 1982. Discourse Strategies. Cambridge: Cambridge University Press.
[44] Reissman, C. K. 1990. Divorce Talk: Women and Mean Make Sense of Personal Relationships, New Brunswick, NJ: Rutgers University Press.
[31] Hirschheim, R. and Newman, M. 1991. “Symbolism and Information Systems Development: Myth, Metaphor, and Magic,” Information Systems Research, Vol. 2, No. 1, pp. 29-62. [32] Holtzblatt, K. and Beyer, H.R. 1995. “Requirements Gathering: The Human Factor,” Communications of ACM, Vol. 38, No. 5, pp. 30-32. [33] Kidd, A. 1987, Knowledge Acquisition for Expert Systems, New York: Plenum Press. [34] Kress, G. 1985. Linguistic Processes in Sociocultural Practice, Victoria: Deakin University Press. [35] Labov, W. 1972. “The Transformation of Experience in Narrative Syntax,” in W. Labov (Ed.) Language in the Inner City: Studies in the Black English Vernacular. Philadelphia: University of Pennsylvania Press. [36] Labov, W. 1982. “Speech Actions and Reactions in Personal Narrative,” in D. Tannen (Ed.), Analyzing Discourse: Text and Talk. Washington, DC: Georgetown University Press. [37] Latour, B. 1990. “Drawing Things Together,” in M. Lynch and S. Woolgar (eds.) Representation in Scientific Practice, Cambridge MA: MIT Press, pp. 19-68. [38] Littman, D.C. 1987. “Modeling Human Expertise in Knowledge Engineering: Some Preliminary Observations. International Journal of Man Machine Studies, Vol. 26, No. 1, pp. 81-92. [39] Marakas, G.M. and Elam, J. J. 1998. “Semantic Structuring in Analyst Acquisition and Representation of Facts in Requirements Analysis,” Information Systems Research, Vol. 9, No.1, pp.37-63. [40] Moody, J. W., Blanton, J. E. and Cheney, P.H. 1998. “A Theoretically Grounded Approach to Assist Memory Recall During Information Requirements Determination,” Journal of Management Information Systems, Vol. 15, No. 1, pp.7998. [41] Mumby, D. K. 1987. “The Political Functions of Narrative in Organizations,” Communication Monographs, Vol. 54, pp. 113-127. [42] Nuseibeh, B. 1996. “Conflicting Requirements: When the Customer is not Always Right,” Requirements Engineering Journal, 1, pp. 70-71.
[45] Reissman, C. K. 1993. Narrative Analysis. Newbury Park: Sage. [46] Robey, D. and Markus, M.L. 1984. “Rituals in Information Systems Design,” MIS Quarterly, Vol. 8. pp. 515. [47] Robey, D. and Newman, M. 1996. “Sequential Patterns of Information Systems Development: An Application of a Social Process Model,” ACM Transactions on Information Systems, Vol. 14, No. 1, pp. 30-63. [48] Sarangi, S. and Roberts, C. 1999. “The Dynamics of Interactional and Institutional Orders in Work-Related Settings. In Sarangi, S. and Roberts, C. Talk, Work and Institutional Order: Discourse in Medical, Mediation and Management Settings. Berlin: Mouton De Gruyter, pp. 1-60. [49] Sayer, K. and Harvey, L. 1997. “Empowerment in Business Process Reengineering: An Ethnographic Study of Implementation Discourses. In Kumar, K. and DeGross, J. (eds) Proceedings of the eighteenth ICIS. Atlanta, GA, pp. 427-440. [50] Schiffrin, D. 1994. Approaches to Discourse. Oxford: Blackwell. [51] Strauss, A. 1987. Qualitative Analysis for Social Scientists, New York: Cambridge University Press. [52] Street, B. V. 1993. “Culture is a Verb.” In D. Graddol, L. Thompson and M. Byram (eds.) Language and Culture, Clevedon: BAAL and Multilingual Matters, pp. 23-43. [53] Stubbs, M. 1983. Discourse Analysis, Oxford: Basil Blackwell. [54] Suchman, L. and Bishop, L. 2000 “Problematizing ‘Innovation’ as a Critical Project,” Technology Analysis and Strategic Management, Vol. 12, No. 3, pp 327-33. [55] Sutton, D. C. 2000. “Linguistic Problems with Requirements and Knowledge Elicitation,” Requirements Engineering, Vol. 5, pp. 114-124. [56] Tannen, D. and Wallat, C. 1987. “Interactive Frames and Knowledge Schemas in Interaction: Examples from a Medical Examination/Review,” Social Psychological Quarterly, Vol. 50, No. 2, 205-216. [57] Tannen, D. 1986. That’s not what I meant! New York: William Morrow.
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
9
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
[58] West, D. 1992. “Knowledge Elicitation as an Inquiring System: Toward a ‘Subjective’ Knowledge Elicitation Methodology,” Journal of Information Systems, Vol. 2, pp. 31-44. [59] West, D. 1995. “The Appreciative Inquiry Method: A Systemic Approach to Information Systems Requirements Analysis.” In Stowell, F.A. (ed) Information Systems Provision: The Contribution of Soft Systems Methodology, McGrawy_Hill, Maidenhead, UK, pp. 140-158. [60] Westrup, C. 1999. “Knowledge, Legitimacy and Progress? Requirements as Inscriptions in Information Systems Development,” Information Systems Journal, Vol. 9, pp. 35-54. [61] Wynn, E. and Novick, D. 1996. “Relevance Conventions and Problem Boundaries in a Work Redesign Team,” Information Technology & People, Vol. .9, No. 2, pp 61-80.
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
10
Discourse Analysis of Requirements and Knowledge Elicitation Interviews Rosío Alvarez University of Massachusetts [email protected] Abstract This paper sets forth a framework for examining the discourse of interviews that take place during requirements analysis and knowledge elicitation. Research has shown that requirements and knowledge elicitation techniques are very similar if not identical in some instances. Moreover, eliciting user requirements is a critical and difficult activity of both information and knowledge based systems development. In this paper, requirements and knowledge elicitation are discursively mediated social practices. Through discourse analysis, this paper shows that narratives emerge during elicitation interviews. In particular, clients propose a “storytelling frame”, where the talk is inviting and rich, which conflicts with that proposed by analysts. A frame brings with it certain ‘rules’ about turns and topics. The paper portrays a conflict between the preferred frames of the analyst and client which points to some of the problems that characterize elicitation interviews. Explanations of why frames are proposed, resisted and sustained are explored.
1. Introduction Communication between analysts and clients is important in all phases of systems development but more so during requirements analysis. In its simplest form, requirements determination entails eliciting and encoding into the new system the requirements that clients verbalize to analysts. Similar to information systems development, expert systems development relies on what is known as knowledge acquisition. Knowledge requirements are formally defined and transferred from some knowledge source (usually human) to a computer program. It has been argued that requirements analysis and knowledge acquisition are almost identical processes [9]. Both requirements
analysis and knowledge acquisition are communicative acts that involve the elicitation of data and information from clients as domain experts or end users. As such, both methods rely on effectual communication between analysts and clients. It is no surprise that systems analysts and knowledge engineers must be skilled in communicating as well as understanding clients motivations and concerns, as research has shown [9]. Yet, one consistent finding in the literature concerned with systems development is the perennial “communication problem” between analyst and client [5, 12, 29]. While analysts working in both requirements analysis and knowledge acquisition use a variety of tools and techniques to obtain information, the interview continues to be the favored elicitation technique [1, 32]. Independent of the type of interview (e.g. structured or unstructured) this elicitation technique depends on interactional talk. Since the discovered data is, in this sense, partly a function of the talk between a client and an analyst, the study of this talk is central to the understanding of how information is captured and the client-analyst relationship in general. Research shows that interviews are recognized as a potentially rich but difficult method of data elicitation [40]. Researchers have argued that communication between analysts and users is often problematic due to issues such as cognitive limitations and vocabulary differences [1, 9]. Sutton [55] suggests that “[D]ifferent social realities bring forth different languages and users are most of the time inhabiting worlds with very different concerns and objects of interest from each other, let alone from providers” (p. 117). Yet, while much attention has been given to the problematic nature of analyst-client communication and language usage, very little of the research is based on direct observation and analysis of these interactions. This paper argues that requirements and knowledge elicitation interviews are a communicative event where narratives are produced. Further, critical discourse analysis (CDA) is used to examine the
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
1
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
narratives that emerge during interviews. Discourse is defined as a systematically-organized set of statements which give expression to meaning. “A discourse provides a set of possible statements about a given area, and organizes and gives structure to the manner in which a particular topic, object, process is to be talked about” [34, p.7]. During requirements analysis interviews, analysts, in their official roles, propose a “framing” which conflicts with the framing proposed by clients during interviews. Frame is defined as the speaker’s instructions to the listener on how to understand the message [24,56]– a metamessage about the context of the talk [57]. The article portrays a mismatch in the expectations of the analyst and client during interviews which explains some of the frustrations which characterize this type of encounter.
2. Requirements analysis and knowledge elicitation as social practice Requirements analysis and knowledge elicitation are two techniques whose concepts and processes are in some cases identical in others very similar. Requirements analysis and knowledge elicitation are given special attention here because of their critical role in systems development. Research suggests that many system failures can be attributed to a lack of clear and specific information requirements [11, 15]. Furthermore, errors during requirements analysis that are not found until later stages of the implementation process can cost significantly more to fix [4, 39]. Similarly, knowledge elicitation is frequently referred to as being one of the most difficult aspects of expert systems development [38, 59]. In its simplest form requirements analysis seeks to identify the information needed to automate organizational tasks and to support information workers in their tasks. Similarly, knowledge elicitation elicits information from domain experts and interprets this information to produce a model which evolves into the expert system used to support knowledge workers [33]. While a difference in lexicon exists, the process and end result of both methods are indistinguishably similar. Moreover, both methods demand similar skills of analysts, for instance building rapport, ability to communicate, understanding client’s motivations and concerns [25, 28]. Clearly, both methods seem to highlight the importance of viewing social context as a critical component for eliciting information. Several currents in the information systems literature argue for and analyze information systems development as a social practice [c.f. 31, 46, 47]. These researchers attend to the ‘organizational’ issues such as power, conflict and control that critically
impinge upon information systems development. Information systems researchers theorize requirements analysis as a process which is socially mediated [22, 20] where requirements are socially constructed through clients’ stories [2]. Others suggest that requirements analysis is a process fraught with conflict, inconsistency and competing viewpoints where one ‘voice’ is not identifiable [42, 13]. Other research suggests that ‘expertise’ is subjective, subtle and tacit [58]. Therefore, requirements analysis and knowledge elicitation would best be served by developing approaches that examine the social situation where information and knowledge are produced. Unfortunately, most requirements and knowledge elicitation depends on technology-driven models and methods. These approaches focus on technical issues and give little if any attention to social context. Instead, requirements analysis techniques produce a “tidy” representation of the organization that is mobile, stable, and combinable [37], thereby allowing analysts to decontextualize requirements, represent the organization in a durable form and make requirements capable of manipulation at a distance far away from the interruptions and clamor of clients [60]. The approach in this paper contributes to understanding requirements and knowledge elicitation as polyphonic, i.e. as containing multiple voices. This polyphonic linguistic interaction is socially situated, where words bring with them the context where they have lived. In particular, this paper examines the fundamental substance of requirements analysis interviews – talk. This paper examines both the formal structure of language by using methods borrowed from sociolinguistics, such as conversational and frame analysis. The analysis here also locates the individual within their social matrix by subjecting the data to criticism that examines how power, control and identity are produced and negotiated during interviews. In this manner, requirements analysis takes on the form of a discursively mediated and constructed social practice.
3. Analyzing discourse through narratives The framework for the study of requirements analysis in this paper draws on critical discourse analysis [10, 17, 18, 19] and narrative analysis [41, 45]. Critical discourse analysis is grounded in interactional talk and text data which employs sociolinguistic methods [50]. Sociolinguistics assigns special significance to the structure of speech and texts, and provides methods for specifying the linguistic features of different types of discourse units
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
2
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
and the way they are tied together into larger units of meaning. However, unlike other forms of discourse analysis, critical discourse analysis also concerns itself with examining social context along the lines of ideology, power and inequality. Through discourse examination, topics of power inequalities usually along the lines of race, class, gender, sexuality and occupation are exposed. Critical discourse analysis therefore, demystifies what is taken to be ‘common sense’ by defamiliarizing it and signaling its functions and consequences in sustaining the social order. As Fowler [21] suggests, critical discourse analysis goes “beyond the formal structure of Language as an abstract system, toward the practical interaction of language and context” (p.10). In this sense, language is seen as a social practice, as a mode of action that is always socially situated “in a dialectical relationship with other facets of ‘the social’…it is socially shaped, but it is also socially shaping, or constitutive” [19, p.131]. From this viewpoint, discourse is seen as constitutive of social reality in a general sense. Narratives are particular types of discourse that emerge during requirements analysis interviews. Research shows that during interviews people “use the narrative form as a pattern for uniting the events of their lives” [43, p.153]. Through narratives, people organize their temporal experiences and make them meaningful. Other research sees the narrative as a “politically motivated production of a certain way of perceiving the world which privileges certain interests over others” [41, p.114]. They are used to create believable explanations for the teller’s actions. To better understand narratives, theorists have developed definitions describing the structure and function of narratives. Some suggest a somewhat narrow definition where narratives are stories about specific past events [35, 36]. These stories follow a chronological sequence where order of events move in a linear way in time. Reissman [45] extends this definition to include other genres such as habitual narratives, which describe repetitive events with no peak in action, and hypothetical narrative, which describe events that did not happen. Researchers of organization and information systems have begun to employ discourse and narrative analysis in various forms to examine modern businesses. For instance, Wynn and Novick [61] examine the issue of ‘turns’ during cross-functional meetings. They find that what is considered a ‘valid’ contribution is based on discourse style. In particular listeners receive the story versus the professional style of discourse differently. Others such as Sayer and Harvey [49] study implementation discourses. In particular they examine how the use of an electronic mail system is used as a technology of power to
manipulate discourses during a BPR project. And finally, Suchman and Bishop [54] examine discourses of ‘innovation’ that function to conserve rather than change existing institutional orders. In the area of narratives, Brown [6] examines the use of narratives that organizational members produced to create meaning and gain political advantage during an IT implementation. Davidson [14] uses narrative analysis to examine sensemaking and interpretation during an IS development project. And finally, Dube and Robey [16] examine stories as a symbol of organizational culture to generate insights into the collective interpretation of management practices by competing groups during a software development project. In general, narrative and discourse analysis draws our attention to the modern business organization as a site for power struggles. Another assumption that guides this research is that the interview, as a discourse genre, is a critical site in which identity is performed. A fundamental observation of discourse analysis is that speakers’ identities emerge from discourse [7]. Drawing on Brian Street’s [52] metaphor “culture is a verb” Sarangi and Roberts [48] propose that “identity is a verb” suggesting that similar to culture, identity is something that people do. Said differently, this paper adopts a performative view of identity where a speaking subjects’ identity is produced through discourse. As feminist theorist Judith Butler [8] states, gender is a copy for which there is no original, suggesting that identity exists and is produced though its performance. The concept of identity construction is of particular relevance for interviews. Unlike sermons, political speeches or educational discussions, the interview as a discourse genre is one where the individual verbalizes thoughts, intentions, and consciousness. It is through this ritual that self-reflection, self-knowledge and selfexamination are obtained. As we confess we tell a story that constitutes a drama in which we are a leading character, and the meaning of this role is found through recollection. In other words, in narrating the past we define ourselves as the implied subject generated by the narrative. Now consider the requirements analysis interview, this act involves participants detailing actions, problems, hopes and needs in regard to work life. Through the elicitation of ‘requirements’ the subject is constituted as the object of knowledge (revealed through self-avowal) and given meaning. This performance functions to construct the identity of the speaking subject. In this sense, it is through linguistic expression that social identities and relations are produced.
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
3
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
4. Research methodology
5. Research site
This paper draws on two discursive mechanisms, footing and framing, adapted from the ideas of Erving Goffman, to analyze requirements interviews. Erving Goffman [24] uses the term footing to refer to these interactional stances or positions vis-à-vis one another. Footing defines the relationship between speakers. Speakers have available to them a number of social roles from which the can move in and out of. In one moment a speaker may be a friend concerned about the health and well-being of an employee, in another moment the same speaker may be a competitive peer vying for control of resources. As speakers communicate linguistically, they choose from a wide range of social attributes and roles that are made available to them and which they perceive as best fit to the situation. Footings are relational and shifting. As we speak, we shift footing, move in and out of social roles, all of which assist us in establishing the relationship with others such that the meaning of the utterance is understood. Another useful concept is framing. Framing defines the speakers’ instructions to the listener about what has been said and how to understand the utterance [3, 23, 56]. Through subtle signals like pitch, voice, intonation, and facial expression the contextual information is jointly created. These signals provide a frame or metamessage about the context indirectly [57]. Gumperz [30] refers to these signals as “contextualization cues” which may be prosodic, paralinguistic and non-verbal. These surface messages function to call up shared experiences, they are powerful means of negotiating social identity and legitimating preferred styles of communicating in the predominantly asymmetrical interactions in workplace settings [48]. The person proposing the frame establishes how talk is to be understood in that context [57]. In this sense attempts at re-framing, counterframing or any other out-of-frame activity will likely be resisted. Out-of-frame activity refers to lines of activity that are somehow outside of the main official line [23, p.201]. This behavior is distracting and threatens the context, which in turn may be treated by participants with active disattention so that the main line of activity can continue. The discursive mechanisms presented here integrate Goffman’s views of social interaction as involving particular presentations of self and dynamic shifts between footings and frames. This paper will analyze the self, or subject positions as Foucault and Fairclough would call it, and the manner in which these identities are produced, sustained, or discredited during requirements analysis interviews.
The organization chosen for the study is State University1 a large public research university located in the northeastern United States. The annual budget of State University exceeds $.5 billion. It has enrollments of 24, 000, faculty of 1184 and staff of approximately 3600. At the time of the fieldwork, the university had recently allocated $11M for an information systems development project. The organization was to determine its information requirements. This involved determining the mandatory and desirable requirements of a system, documenting them and then evaluating the various vendor solutions according to these requirements. The new information system was to handle all the major administrative functions of the university including, financial accounting, human resources, and all student services such as housing, enrollment, financial aid and admissions. The legacy systems that were in place at the start of this project, which we shall call the Administrative Information Systems project (AIS), had been in use for over fifteen years. These legacy systems consisted of a combination of custombuilt and vendor applications which had been extensively modified for specific university functions. The student registration system was built in-house using a System 2000 hierarchical database with some VSAM files. Information Associates, a vendor of higher education software, provided applications that handled student financial aid, bursar and financial functions. All three applications had been discontinued and were now updated and maintained in-house. Information Systems Incorporated provided the human resource application, which was also a discontinued application now updated by in-house staff. The operating system consisted of MVS ESA v4.3, CICS v 2.1.2, and OS/VS Cobol 2.4. All applications were located on an IBM 3090. A voice response system provided by EPOS was in use for student access to information and registration. Reporting tools included Data Analyzer and SAS. For this research, the author was hired by the CIO of the university to document the software selection process. This role was disclosed to all organizational members involved in the ERP selection and implementation. The case study was completed and a report was provided to the CIO who read and approved its contents. The entire information systems development project began in January 1996 and was still in implementation at the writing of this paper. The focus of this paper is on the requirements analysis 1
Pseudonym is used.
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
4
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
phase which lasted from August 1996 to December 1996. During this shorter phase, the university contracted a consulting firm to conduct the requirements analysis study and produce all relevant documentation.
6. Findings The data collection was conducted through participant observation of requirements analysis interviews. While my primary role was that of observing and documenting the selection process, my recognized technical expertise by clients and analysts often prompted questions directed at me, which required my participation in the discussion. In all interviews at least two analyst-interviewers were present. The analyst’s strategy was to have one interviewer ask question and the other function as ‘scribe.’ Conversations during interviews were taped. Data was collected at thirty-two meetings in which eighty-two individuals from eight different departments participated. The data set consists of sixty hours of tape recordings, along with field notes where tape recording was not feasible. The emphasis during data analysis was on linguistic structure and interaction, therefore it was important to include intonation, volume, pacing and other qualities of speech to capture the mood and feel of the interview. To achieve a balance between capturing all the detail necessary and providing a readable transcript, transcription conventions were based on the work of Reissman [45] and Gronn [26, 27] (adapted from Stubbs [53]). These conventions ensured that overlaps, exclamations, questions, pauses and emphasis were maintained. Symbols used in the transcribed extracts are: // overlapping talk from the first to the last slash (x) pause of x seconds [ ] explanatory note {} nonlexical utterances italics word emphasized by speaker ! exclamation ? question Speakers are identified by Name (of narrator) and by Int. (for interviewer). Open coding was used to identify frames, and examine their interactional accomplishment, maintenance and resistance. The core features of open coding are: 1) the inductive development of provisional categories; 2) ongoing testing of categories through conceptual analysis and comparison of categories with data that is already
coded; and 3) the altering of existing categories as other ones are created or eliminated [51, p.11-13].
6.1. The storytelling frame Many of the interviewees in this study followed a pattern which I call a storytelling frame. The client attempts to stay well within this frame, while the analyst, for the most part, resists and reframes the interaction. A frame brings with it certain ‘rules’ about what talk should be about (or not be about), who can talk, in what sequence and who controls the activity. In the storytelling frame the speaker tells a story or short narrative. This frame creates a more symmetrical relationship between client and analyst. Moreover, the role of information worker resides in the background while another subjectivity is developed in the foreground. The analyst prefers to resist this frame and sustain a more official role that brings with it an asymmetrical relationship; here the analyst controls the turn-taking as well as the topic of discussion by asking the questions. The following passage portrays a shifting into and out of the storytelling frame. Here two members from the housing office, Shelly and Carol, are asked questions by two interviewers, Int1 (who is female) and Int2 (who is male). They are asked about the process by which students received bills pertaining to damages of the dormitory rooms. The extract illustrates a pattern of shifting frames which produce a certain degree of conflict. 1 Int1 -- so you don't have a 24-hour turnaround? 2 Carole -- the other time when there is a little backlog is check 3 out in December. December is a real problem because they are 4 checking out and our staff wants to go out on Christmas break 5 also. So, {um}, we frequently have to badger to get those 6 invoices 7 Int1 -- so December and May are your tough times when you 8 don't have a 24-hour turnaround. What do you think the 9 processing time is in most cases? 10 Shelley – [Shelley looks a Carole] (5) it could be as much as 11 two weeks 12 Carole – yeah, and maybe in some cases depending upon staff 13 in the area offices even three weeks 14 (3) because the staff and the area offices have many other 15 responsibilities besides the billing, that 16 (3) we could badger them, but they can't do it, and we 17 /understand 18 Int1 -- the example/ you gave us suggested that at check out the 19 inventory form was completed with the student present, is that 20 not uniformly true? 21 Shelley -- that is not uniformly true. 22 You're trying to check out 11,000 people 23 in a week's period of time with 300 RA staff, who are also 24 taking finals and have 25 there are sometimes where students will check out alone and 26 waive their right to appeal, by checking out and then someone 27 will come in after the fact and complete the inventory form and 28 do a billing that's related to that
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
5
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
29 Int1 -- how often does this happen? 30 Shelley – (3) we could certainly get you statistics on how many 31 students do that, I, I, don't know I assume a lot. We could do, 32 not, not, full range statistics but take a pile of /inventories 33 Carole -- see how/ many were checked out 34 Int1 -- is that universally true, that if the student checked out 35 without the inventory being completed he waives his right to 36 appeal 37 Carole -- absolutely in all cases, so if his choice is get out of 38 here I wanna leave or wait around till someone can go with him 39 and /complete and so he 40 Int1 -- ok ok/ 41 Carole – ok, because I have been on the phone with parents 42 Int1 –[begins talking to other interviewer, looking away from 43 Carole and Shelley] 44 Carole -- I've been on the phone with parents, [interviewers 45 look at Carole] explaining that when they've gotten bills and 46 have been very upset, and very irate and indicated to them that 47 your son or daughter made a choice and when we went back and 48 inventoried the room this is what we found. 49 I tell you, as a parent it's not a very comfortable thing to hear 50 and, /and the have to 51 Int1 --they /don't accept it? 52 Carole -- not all the time 53 Shelley -- we're looking at a process to possibly not, we used to 54 bill three dollars if the student didn't leave their mattress cover 55 on their mattress and were trying to put a minimum I think what 56 we're suggesting a $25 minimum per billing… 57 Int1 -- except that the cost of mattress covers has gone way up, 58 I know [Interviewer 1 laughs, Carole and Shelley laugh] 59 Int2 – ok, so this is just one way they can get a notice, then it 60 gets on to BRS and is simply an amount due. 61Carole – [2] yes
I analyze this excerpt by examining turn-taking and moves that each speaker accomplishes through their turn. The excerpt opens with Carole responding to the question posed to her by the interviewer about turnaround time. Carole tells us a story about problems they encounter in December. She takes us back to a time when staff has competing demands to leave for Christmas break. Rather than give a yes or no answer, Carole steps into a more personal frame, the storytelling frame, changing the footing from information provider to caring woman. Through the storytelling frame she appears as a concerned woman who cares about employees desires and demands. The interviewer reframes the interview by returning to the question about processing time (line 7). By asking questions the analyst imposes the obligation of a response on the Carole. This places a constraint on Carole’s next move. More importantly, it places Carole in a follower role where the appropriateness of her response will be measured by the question that preceded. Carole hesitates for several seconds. At this point Shelley looks at Carole and intervenes attempting to shift the footing back to less personal one by answering the question more directly with a quantifiable amount.
Carole then returns to the conversation, staying in the more official frame by providing direct quantifiable information about timing, but after a pause she shifts the footing again. Carole reframes the interaction and now the professional role of information worker remains largely in the background. Instead a relationship of caring woman figures more prominently. Carol returns to the storytelling frame. She describes overworked staff that “have many other responsibilities” (line 14) and inscribes herself in this narrative as an understanding woman who could badger them but does not. Through this brief narrative she persuades us into believing that she is an understanding person. Through short narratives about daily work practices, Carole’s identity as a caring woman begins to take form. The analyst abruptly switches to a factual request for information. She now searches for uniformity (line 20), quantification (line 29) and universals (line 34) in client responses. To the first request Shelley responds with a direct answer, but then switches back to the storytelling frame by narrativizing the daily habitual difficulties of managing workers who have lives with other demands (“RA staff who are also taking finals”). When the storytelling frame is interrupted by the analysts’ request to quantify her response, Shelly stalls, then attempts to comply with the request by offering ‘statistics’. Shelley is compelled to participate and reluctant to resist the framing proposed by the analyst for fear of being evaluated as incompetent information worker. She reluctantly accepts the role of respondent, by offering her rendition of statistics, and lets go of other possible footings so as to be favorably viewed by the analyst. Carole, on the other hand, resists the frame proposed by the analyst more often by counterframing the encounter with stories. In line 40 Carole’s hypothetical narrative about what students might do at check out time is interrupted by the analyst, with “ok, ok”. At this point the interviewers begin to talk amongst themselves, ignoring Carole who has launched into another personal narrative about her interaction with parents. The non-verbal cues by the interviewers, who turn away, signal to Carole that they have transitioned their attention away from her and to their documentation. At this point the pace and pitch of Carole’s talk increases (line 44). She repeats the previous statement, regains control of the interview and reframes the interaction, back to a storytelling one. She tells a story about parents who are surprised when billed for damages to dormitory rooms done by their children. The story peaks by Carole identifying herself as a parent (line 49). This confession contributes to creating a more nuanced identity of Carole, now as a caring mother.
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
6
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
The passage ends with a change in footing initiated by the female interviewer (Int1). Throughout the interaction she has resisted the storytelling frame, by counter framing it whenever it was in place. However, at line 57 the woman interviewer responds with a bit of personal information about herself as a consumer. The prosodic emphasis on “I know” signals to the other women that interviewer has direct experience with or information about purchasing the item of concern, a mattress cover. She shares an understanding with the other women as a consumer, perhaps as a mother as well. In performing this move, the analyst steps out of her official role and Carole and Shelley momentarily leave their roles as information providers to assume that of women consumers conversing about the costs of keeping a home. This move creates an affiliation among the women as they laugh together. The interview is suspended. However, the fleeting moment of intimacy among the women fades as the male interviewer regains control of the interview and reframes the interaction. With an abrupt shift in footing the male interviewer changes the context. He reduces the complexity of concerns and dilemmas of work practices as narrativized by Carole and Shelley to “simply an amount due on the bill”.
7. Discussion and conclusion In examining the discursive construction of narratives during requirements analysis interviews, we find that shifts in framing and changes in footing create an ongoing tension between the analyst and client. Throughout the interview we observe a struggle in reframing which can be attributed to an underlying conflict between the way the analysts view the encounter and the clients view it. For the analyst the encounter is a ‘professional’ encounter, for the client it is about personal work experiences and dilemmas. Thus, both Carole and Shelley propose framings in the direction of a social encounter through storytelling. In this frame the clients’ identity beyond information worker takes form; as caring mother and co-worker. The analysts resist the storytelling frame by shifting out of the frame. This continual shift and apparent conflict exemplifies a basic difference between clients and analysts. While both are engaged in and contribute to producing an interview, clients and analysts have very different understandings of what each one’s role is. The clients approach the encounter from a more intimate angle, attempting to be more personal, funny, friendly, or even dramatic, the analyst comes at it from a very professional stance looking for brevity and directness. In reframing,
Carole and Shelley’s communicative styles change to providing more information and more involvement. By accomplishing a (re)framing Carole is able to signal the importance of larger social world issues. However, the analysts shift out of the storytelling frame and in so doing, confine Carole to a limiting and unidimensional role of information worker and no one else. The persistence of clients to remain in the storytelling frame does not seem to alter the analysts reframing of the interview. Why do clients succumb to this reframing, albeit with some resistance? Perhaps clients may at times see the storytelling frame as less ‘professional’ and fear being seen as incompetent information workers if they remain. As an example, Shelley responds to a request with the concept of statistics perhaps because she believes it will make her appear more professional, a more scientific subject. Secondly, reframing is a difficult task to achieve. It is easier to stay than to resist or react, which takes additional effort, skill and power. When the analysts reframe the interview, interrupting the storytelling frame, power relations are skewed in the direction of the analyst who possesses institutional legitimacy. The analyst is poised to control the interview, ask the questions and determine what gets talked about. In this paper, narratives produced in the discourse of requirements analysis are examined to illuminate issues of identity, conflict, power and communication ‘gaps’ during the interview interaction. This approach expands our understanding of how professional identities are constituted and relations of power are fashioned out of talk during requirements analysis and knowledge elicitation interviews. The paper has shown that clients propose a storytelling which is continually resisted by the analysts. The storytelling frame allows the clients to engage in narratives that position the client within their cultural and social environment. The narrative response provides a space where clients may embed themselves within the story and thereby provide a more nuanced image of their identity, concerns and dilemmas. We are able to see a tug of war between roles assumed and modes of talk during the interview. The client prefers the storytelling frame which frames the interview as a social encounter. Here the talk is friendly, inviting and very rich. However, the analyst opts to reframe, relegating the client to a professional identity and succinct mode of talk. Yet, the storytelling frame would inevitably bring about a more involved and animated person willing to provide information that better describes the ‘motivations and concerns’ of the worker as well as share ‘subjective’ interpretations of what they believe constitutes expertise.
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
7
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
8. References [1] Agarwal, R. and Tanniru, M.R. 1990. “Knowledge Acquisition Using Structured Interviewing: An Empirical Investigation,” Journal of Management Information Systems, Vol. 7, No.1, pp.123-140. [2] Alvarez, R. and Urla, J. (forthcoming). “Tell Me a Good Story: Using Narrative Analysis to Examine Information Requirements’ Interviews During an ERP Implementation,” DATABASE for the Advances of Information Systems. [3] Bateson, G. 1972. Steps to an Ecology of the Mind. New York: Ballantine. [4] Boehm, B. W. 1981. Software Engineering Economics, Englewood Cliffs, NJ: Prentice Hall. [5] Bostrom, R.P. 1989. “Successful Application of Communication Techniques to Improve the Systems Development Process,” Information and Management, Vol. 16, pp. 279-295.
[14] Davidson, E.J. 1997. “Examining Project History Narratives: An Analytic Approach,” in A.S. Lee, J. Liebenau, and J.I. DeGross, (Eds.) Proceedings of the IFIP TC8 WG8.2 International Conference on Information Systems and Qualitative Research, Philadelphia PA, pp. 123148. [15] Davis, G. B. 1982. “Strategies for Information Requirements Determinations,” IBM Systems Journal, Vol. 21, No. 1, pp. 4-30. [16] Dube, L. and Robey, D. 1999. “Software Stories: Three Cultural Perspectives on the Organizational Practices of Software Development,” Accounting, Management and Information Technology, Vol. 9, pp. 223-259. [17] Fairclough, N. 1989. Language and Power. London: Longman. [18] Fairclough, N. 1992. Discourse and Social Change. Cambridge, MA: Polity Press. [19] Fairclough, N. 1995. Critical Discourse Analysis: The Critical Study of Language. London; New York: Longman.
[6] Brown, A. 1998. “Narrative, Politics and Legitimacy in an IT Implementation,” Journal of Management Studies, Vol. 35, No. 1, pp. 35-59.
[20] Flynn, D. J. and Jazi M. D. 1998. “Constructing user requirements: A social process for social context,” Information Systems Journal, Vol. 8, No. 1, pp. 53-83.
[7] Bucholtz, M. 1999. “Transgression and Progress in Language and Gender Studies.” In Bucholtz, M., Liang, A.C. and Sutton, L. A. Reinventing Identities: The gendered Self in Discourse. New York: Oxford University Press.
[21] Fowler, R. 1996. “On Critical Linguistics.” In CaldasCoulthard, C. R. and Coulthard, M. 1996. Texts and Practices: Readings in Critical Discourse Analysis. London: Routledge. pp. 3-14.
[8] Butler, J. P. 1992. Contingent Foundations: Feminism and the Question of "Postmodernism", in Butler, J.P. and Scott, J.W. Feminists Theorize the Political, New York: Routledge.
[22] Galliers, R.D. and J. A. Swan. 2000. “There's More to Information Systems Development than Structured Approaches: Information: Requirements Analysis as a Socially Mediated Process,” Requirements Engineering Journal Vol. 5 No. 2, 74-82. [23] Goffman, E. 1974. Frame Analysis. New York: Harper and Row.
[9] Byrd, T. A., Cossick, K.L., Zmud, R.W. 1992. “A Synthesis of Research on Requirements Analysis and Knowledge Acquisition Techniques,” MIS Quarterly, Vol. 16, No.1, pp. 117-138. [10] Caldas-Coulthard, C. R. and Coulthard, M. 1996. Texts and Practices: Readings in Critical Discourse Analysis. London: Routledge. [11] Cooper, R. B. and Swanson, E. B. 1979. “Management Information Requirements Assessment: The State of the Art,” Database, Vol. 10, pp. 5-16. [12] Cronon, T. P. and Means, T. L. 1979. “System Development: An Empirical Study of User Communication,” Database, Vol. 10, pp. 5-16. [13] Darke, P. and Shanks, G. 1996. “Stakeholder Viewpoints in Requirements Definition: A Framework for Understanding Viewpoint Development,” Requirements Engineering Journal, 1, pp. 88-105.
[24] Goffman, E. 1981. Forms of Talk. Philadelphia, PA: University of Pennsylvania Press. [25] Graham, I. And Jones, P.L. 1988. Expert Systems: Knowledge, Uncertainty and Decision, Suffolk, England: St. Edmundsbury Press. [26] Gronn, P. C. 1983. “Talk as Work: The Accomplishment of School Administration,” Administrative Science Quarterly, Vol. 28, pp. 1-21. [27] Gronn, P.C. 1985. “Committee Talk: Negotiating ‘Personnel Development’ at a Training College,” Journal of Management Studies, Vol. 22, No. 3, pp. 245-268. [28] Guinan, P.J. 1988. Patterns of Excellence for IS Professionals: An analysis of Communication Behaviour, Washington, DC: ICIT Press.
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
8
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
[29] Guinan, P.J. and Bostron, R.P. 1986. “Development of Computer-Based Information System: A Communication Framework,” Database, Vol. 17, No. 3, pp. 3-16.
[43] Polkinghorne, D. E. 1988. Narrative Knowing and the Human Sciences. Albany, NY: State University of New York Press.
[30] Gumperz, J. J. 1982. Discourse Strategies. Cambridge: Cambridge University Press.
[44] Reissman, C. K. 1990. Divorce Talk: Women and Mean Make Sense of Personal Relationships, New Brunswick, NJ: Rutgers University Press.
[31] Hirschheim, R. and Newman, M. 1991. “Symbolism and Information Systems Development: Myth, Metaphor, and Magic,” Information Systems Research, Vol. 2, No. 1, pp. 29-62. [32] Holtzblatt, K. and Beyer, H.R. 1995. “Requirements Gathering: The Human Factor,” Communications of ACM, Vol. 38, No. 5, pp. 30-32. [33] Kidd, A. 1987, Knowledge Acquisition for Expert Systems, New York: Plenum Press. [34] Kress, G. 1985. Linguistic Processes in Sociocultural Practice, Victoria: Deakin University Press. [35] Labov, W. 1972. “The Transformation of Experience in Narrative Syntax,” in W. Labov (Ed.) Language in the Inner City: Studies in the Black English Vernacular. Philadelphia: University of Pennsylvania Press. [36] Labov, W. 1982. “Speech Actions and Reactions in Personal Narrative,” in D. Tannen (Ed.), Analyzing Discourse: Text and Talk. Washington, DC: Georgetown University Press. [37] Latour, B. 1990. “Drawing Things Together,” in M. Lynch and S. Woolgar (eds.) Representation in Scientific Practice, Cambridge MA: MIT Press, pp. 19-68. [38] Littman, D.C. 1987. “Modeling Human Expertise in Knowledge Engineering: Some Preliminary Observations. International Journal of Man Machine Studies, Vol. 26, No. 1, pp. 81-92. [39] Marakas, G.M. and Elam, J. J. 1998. “Semantic Structuring in Analyst Acquisition and Representation of Facts in Requirements Analysis,” Information Systems Research, Vol. 9, No.1, pp.37-63. [40] Moody, J. W., Blanton, J. E. and Cheney, P.H. 1998. “A Theoretically Grounded Approach to Assist Memory Recall During Information Requirements Determination,” Journal of Management Information Systems, Vol. 15, No. 1, pp.7998. [41] Mumby, D. K. 1987. “The Political Functions of Narrative in Organizations,” Communication Monographs, Vol. 54, pp. 113-127. [42] Nuseibeh, B. 1996. “Conflicting Requirements: When the Customer is not Always Right,” Requirements Engineering Journal, 1, pp. 70-71.
[45] Reissman, C. K. 1993. Narrative Analysis. Newbury Park: Sage. [46] Robey, D. and Markus, M.L. 1984. “Rituals in Information Systems Design,” MIS Quarterly, Vol. 8. pp. 515. [47] Robey, D. and Newman, M. 1996. “Sequential Patterns of Information Systems Development: An Application of a Social Process Model,” ACM Transactions on Information Systems, Vol. 14, No. 1, pp. 30-63. [48] Sarangi, S. and Roberts, C. 1999. “The Dynamics of Interactional and Institutional Orders in Work-Related Settings. In Sarangi, S. and Roberts, C. Talk, Work and Institutional Order: Discourse in Medical, Mediation and Management Settings. Berlin: Mouton De Gruyter, pp. 1-60. [49] Sayer, K. and Harvey, L. 1997. “Empowerment in Business Process Reengineering: An Ethnographic Study of Implementation Discourses. In Kumar, K. and DeGross, J. (eds) Proceedings of the eighteenth ICIS. Atlanta, GA, pp. 427-440. [50] Schiffrin, D. 1994. Approaches to Discourse. Oxford: Blackwell. [51] Strauss, A. 1987. Qualitative Analysis for Social Scientists, New York: Cambridge University Press. [52] Street, B. V. 1993. “Culture is a Verb.” In D. Graddol, L. Thompson and M. Byram (eds.) Language and Culture, Clevedon: BAAL and Multilingual Matters, pp. 23-43. [53] Stubbs, M. 1983. Discourse Analysis, Oxford: Basil Blackwell. [54] Suchman, L. and Bishop, L. 2000 “Problematizing ‘Innovation’ as a Critical Project,” Technology Analysis and Strategic Management, Vol. 12, No. 3, pp 327-33. [55] Sutton, D. C. 2000. “Linguistic Problems with Requirements and Knowledge Elicitation,” Requirements Engineering, Vol. 5, pp. 114-124. [56] Tannen, D. and Wallat, C. 1987. “Interactive Frames and Knowledge Schemas in Interaction: Examples from a Medical Examination/Review,” Social Psychological Quarterly, Vol. 50, No. 2, 205-216. [57] Tannen, D. 1986. That’s not what I meant! New York: William Morrow.
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
9
Proceedings of the 35th Hawaii International Conference on System Sciences - 2002
[58] West, D. 1992. “Knowledge Elicitation as an Inquiring System: Toward a ‘Subjective’ Knowledge Elicitation Methodology,” Journal of Information Systems, Vol. 2, pp. 31-44. [59] West, D. 1995. “The Appreciative Inquiry Method: A Systemic Approach to Information Systems Requirements Analysis.” In Stowell, F.A. (ed) Information Systems Provision: The Contribution of Soft Systems Methodology, McGrawy_Hill, Maidenhead, UK, pp. 140-158. [60] Westrup, C. 1999. “Knowledge, Legitimacy and Progress? Requirements as Inscriptions in Information Systems Development,” Information Systems Journal, Vol. 9, pp. 35-54. [61] Wynn, E. and Novick, D. 1996. “Relevance Conventions and Problem Boundaries in a Work Redesign Team,” Information Technology & People, Vol. .9, No. 2, pp 61-80.
0-7695-1435-9/02 $17.00 (c) 2002 IEEE
Proceedings of the 35th Annual Hawaii International Conference on System Sciences (HICSS-35’02) 0-7695-1435-9/02 $17.00 © 2002 IEEE
10
