«IIII. .. ,. 4.5 a 20.4 i illfighwihauflgrgka ... Hilary?” vhf); Fig)! . . . . : . . .. .. . . . {.(Qi (”NJ Haiku!” .331”!!! £53.!!! . z ‘. . .4 ... . x v }.v‘|ll.- '. .c .Q’.\| . A: .p.‘.l.. . . v ,.. . . ‘ A. ‘ 5.3; 3:... 1 N»: 7.9.: .. p. ..v v . . . v:¢.flv.vlt\slul ...v.)o . . V ....5....2.t§.a. . . . . Qtfivfimfl. ”.3. "a . ‘ A . .. ‘ c-rl.d..a... . . uvx .. .EY. , 1 .:.5..\4....o- . 231 .. 7 .n.......3nJ.le.a.If .u .. u of... If T: 7.5!}..8. 01 A 1‘ . I .‘l-? {frill- Vt fiflfiwnr '31.. .11.. .33: «ytlni.lulfil. all '~\ll‘i.:ata..l. 1...!- |4 ..l¥vllltl‘&xsa‘l IIIEIIBHIIIHH II uI l I 1w I j gIIIIIIIIIgISIIIIIIIIZIIIIIZIICIIIIIIIIHI E Z. LIBRARY Michigan State University This is to certify that the dissertation entitled AN INTERACTION ANALYSIS OF NEGOTIATOR AND HOSTAGE-TAKER IDENTITY-GOAL, RELATIONAL—GOAL, AND LANGUAGE INTENSITY MESSAGE BEHAVIOR WITHIN HOSTAGE NEGOTIATIONS: A DESCRIPTIVE INVESTIGATION OF THREE NEGOTIATIONS presented by Randall Gage Rogan has been accepted towards fulfillment of the requirements for Ph.D. d . Communication egreem {ffiéw/M/L— Major professor July 10, 1990 Date MS U i: an Affirmative Action/Equal Opportunity Institution 0- 12771 PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE“ DATE DUE DATE DUE A ‘AA $3.8 '2 ‘9 6'93 MSU Is An Affirmative ActioNEquel Opponunfly lnditution AR ma! ANALYSIS (E WIATOR AND warm-um Wm, RELATIGIAL-GJAL, MD mm mm W BEIAVIOR WITHIN mm murmurs: A WINE IRVESTIGATIGI (I THREE WATERS By Retrial]. Guam A DISSEM’ATICN Sutnitted to Hichigan State [hiversity in partial fulfillment of the mire-rents for the degree of mam Wt of Cc-Inicatim ABSTRACT Al'INTIIICIIOI’ANALYSIS»OPNIBBOTIAIUBLAIDIHDSTAGBJZAKBR IDIIHIIYHGDAL, RBLAITOIALHGOAL» AlDILAlGUAGBIIlTllSITY MESSAGIIEIHANIORIUITHIB’HDSIAGBLIEGDTIAIIOIS: AI'DISCIIPTIVB INVESTIGATIOIIOE’THREBHIBGUTIAIIOIS By Randa11.Gege'Rognn This investigation defines hostage negotiations as intense communicative conflict interactions. It is argued that extant hostage negotiation literature is devoid of commuication-based behavioral interpretations. A review of communication conflict literature reveals potential applications of communication variables to hostage negotiations. Specifically, this study conceptualizes hostage negotiations as functional goal-based interactions, in which negotiators and hostage-takers devote their communicative energies to satisfying identity goals (face behaviors), and relational goals (integrativeness/distributiveness), and which are marked by patterns of language intensity. Several research propositions and research questions are posited to investigate these goal behaviors. As this investigation is exploratory, descriptive results are presented for each of the questions and propositions. Some general findings indicate that hostage negotiations are primarily integrative interactions involving retroactive face support. Language intensity is Obfuscated by its association with both integrativeness and distributiveness. The implications of these findings for communication research, and hostage negotiations are discussed. This dissertation is dedicated to my parents, Ray and Marian Rogan, whose unconditional love and support made this dream a reality. iii It is an ardous task to delimit the countless individuals who have helped shape and influence my graduate career. Nonetheless, there are several persons whom I wish to acknowledge for their special contributions. To begin with, I wish to thank the members of my graduate committee: Dr. William A. Donohue, Dr. James B. Stiff, Dr. Frank.A. Boster, Dr. Lawrence A. Hesse, and Dr. Frank A. Fear, for their support, guidance, and suggestions throughout my graduate program. In particular, I wish to thank Dr. William Donohue, my advisor and Dissertation Director, for his special guidance and assistance in helping me develop a research program in hostage negotiations. I am also indebted to Dr. James Stiff for his support and encouragement through the dissertation process. I also wish to acknowledge the key assistance of Special Agent Fred Lanceley of the Special Operations and Research Unit of the Federal Bureau of Investigation for providing the hostage negotiation material upon this research is based. I also want to thank Dr. Richard Smith and Dr. Gary Kaufman of the Michigan State Police Behavioral Science Unit for their input concerning the details of hostage negotiation. Special thanks goes to my family for their continued support through the several years of my graduate program. I specifically wish to thank my parents Ray and Marian, my brother Reg, my sister-in-law iv Karen, and my two special nephews, Ian and Leland for their encouragement. I also wish to thank William Berti for his support and friendship during the dissertation process . Finally, I wish to thank Jesus Christ for the strength and courage to conquer the challenges that were laid before me. TABLBIOE'CUITENTS INTRODUCTION ..................................................... 1 An Interaction Analysis of Negotiator and Hostage-Taker Identitthoal, Relational-Goal, and Language Intensity Message Behavior in Hostage Negotiations: An Exploratory Investigation of Three HostageJTakings ...................... 1 } General Description of the Study ............................ 4 ,/ Research Implications ....................................... 5 Organization of the Disseration ............................. 6 CHAPTER ONE: HOSTAGE NEGOTIATION ................................. 8 Characteristics of Hostage Negotiation ...................... 9 Physical and Situational Parameters .................... 9 Emotional Excitation ................................... 13 Psychological Attributes of the Hostage-Taker .......... 17 . Face-Work in Hostage Negotiation .................. 18 i Control and Productive Management of Hostage Negotiation.... 22 4 CHAPTER THO: COMMUNICATION AND HOSTAGE NEGOTIATION ............... 24/ The Function of Communication in Conflict ................... 25 Mechanistic Interpretation of Communication in Conflict ............................................... 26 Psychological Interpretation of Communication in Conflict ............................................... 29 Face-Work in Conflict ............................. 33 Systems-Interpretive Interpretation of Communication in Conflict ............................................ 34 Effect of Communication on Conflict Interactions ....... 40 An Integrated Approach to Studying Communicative Conflict Behavior .................................................... 42 Communication Behavior in Hostage Negotiation ............... 45 Stages of Hostage Negotiation .......................... 48 Implications for Communication in Hostage Negotiation.. 49 Rationale and Scope of the Study ............................ 51 Research.Propositions and Research.Questions ........... 54 Time Phase Model of Hostage Negotiation ........... 54 Identitthoal and Relational-Goal Conflict Behavior .......................................... 55 Language Intensity Conflict Behavior .............. 80 CHAPTER.THREE: RESEARCH METHODS .................................. 65 Transcripts ................................................. 65 General Description of the Selected Negotiations ....... 88 Transcript Number One: NCPO ....................... 88 Transcript Number The: Majors ..................... 69 Transcript Number Three: MPD ...................... 70 0 V1 Phase Descriptions of the Negotiations ................. 71fll Coding Schemes .............................................. 7S,J Identitthoal Coding Scheme ............................ 76 Coding Procedure and Results for Identitthoal Coding Scheme ....................... 77 Relational-Goal Coding Scheme .......................... 78 Coding Procedure and Results for Relational-Goal Coding Scheme ..................... 80 Language Intensity Coding Scheme ....................... 81 Simple and Elaborate versions of the Qualifiers Coding Category ................................... 83 Procedure and Results for Language Intensity Coding Scheme ..................................... 84 Coding Scheme Validation .................................... 87 IdentitybGoal Coding Scheme validation ................. 87 Results of the Identitthoal Coding Scheme Validation Check .................................. 88 Relational-Goal Coding Scheme Validation ............... 93 Results of the Relational-Goal Coding Scheme validation Check .................................. 95 Language Intensity Coding Scheme Validation ............ 98 Results of the Language Intensity Coding Scheme Validation Check .................................. 100 Conclusions From the validation Checks ...................... 102 CHAPTER.FOUR: RESULTS ............................................ 104 Results for the Research Questions .......................... 107 Results for the Research Propositions ....................... 137 CHAPTER FIVE: DISCUSSION ......................................... 179 Interpretation of the Findings .............................. 180 Identitthoal Behavior ................................. 180 Relational-Goal ........................................ 184 Language Intensity ..................................... 189 Interrelationship Between Identity-Goal, Relational- Goal and Language Intensity Behavior ................... 194 Hostage-taker and Negotiator Relational-Goal Interaction ....................................... 194 Negotiator Identitthoal and Hostage—taker Relational-Goal Interaction ....................... 195 Negotiator and Hostage-taker Identitthoal Interaction ....................................... 196 Language Intensity and Relational-Goal Interaction ....................................... 196 Implications of the Results ................................. 199 Limitations of the Study .................................... 203 Limitations of the Transcripts ......................... 203 Limitations of the Research Methodology ................ 205 Recommendations for Future Research ......................... 207 REFERENCES ....................................................... 211 APPENDICES ....................................................... 221 APPENDIX A: AN EXAMPLE TRANSCRIPT ........................... 221 APPENDIX B: PHASE SYNOPSES OF NEGOTIATIONS .................. 222 vii APPENDIX C: IDENTITY-GOAL (DDING SCHEME ..................... 232 APPENDIX D: RELATIONAL—GOAL (DDING SCHEME ................... 233 APPENDIX E: BASIC LANGUAGE INTENSITY CODING SCHEME .......... 236 APPENDIX E1: SIMPLE AND ELABORATE QUALIFIER CODING PmCEDURE ................................................... 237 APPENDIX F: EXAMPLE IDENTITY-GOAL CODING SCHEME VALIDATION INSTRUMENT .................................................. 239 APPENDIX G: EXAMPLE RELATIONAL-GOAL CODING SCHEME VALIDATION INSTRUMBIT ....................................... 241 APPENDIX H: LANGUAGE INTENSITY CDDING SCHEME VALIDATION INSTRUMENT .................................................. 244 TABLES ........................................................... 247 FIGURES .......................................................... 277 Table 1. Table 2. Table 3. Table 4. Table 5. Table 6. Table 7. Table 8. Table 9.. Table 10. Table 11. LISTWOF'TIBLES Self-Directed and Other-Directed Face Behaviors ........ 247 Identitthoal Category Means and Standard Deviations for Identitthoal Coding Scheme validation Measures....248 Grouped Identity-Goal Category Means and Standard Deviations for Grouped Identitthoal Coding Scheme validation Measures ............................. 249 Relational-Goal Category Means and Standard Deviations for Relational-Goal Coding Scheme validation Measures .................................... 250 Integrative and Distributive Relational-Goal Category Means and Standard Deviations for Relational-Goal Coding Scheme Validation Measures ...................... 251 Simple Qualifier and Elaborate Qualifier Language Intensity Coding Scheme Means and Standard Deviations for the Language Intensity Coding Scheme Validation Measure ................................................ 251 Frequency of Negotiator Integrative and Distributive Relational-Goal Category Codes by Transcript for all Phases ................................................. 252 Frequency of Hostage-Taker Integrative and Distributive Relational-Goal Category Codes by Transcript for all Phases .............................. 252 Frequency of Negotiator Relational-Goal Responses Over all Phases for all Transcripts .................... 253 Frequency of Hostage-Taker Relational-Goal Responses Over all Phases of all Transcripts ..................... 254 Frequency of Hostage-Taker Integrative and Distributive Relational-Goal Category Codes for all Transcripts by Phases .................................. 255 ix Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table 26. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. Frequency of Negotiator Integrative and Distributive Relational-Goal Category Codes for all Transcripts by Phase ............................................... 255 Frequency of Negotiator Integrative and Distributive Relational-Goal Category Codes for Transcript Number One: HCPO .............................................. 256 Frequency of Hostage-Taker Integrative and Distributive Relational-Goal Category Codes for Transcript Number Tho: Majors .......................... 256 Frequency of Negotiator Relational-Goal Responses by Phase for Transcript Number One: NCPO .................. 257 Frequency of Hostage-Taker Relational-Goal Responses by Phase for Transcript Number One: "CPO ............... 258 Frequency of Negotiator Integrative and Distributive Relational-Goal Category Codes for Transcript Number Two: Majors ............................................ 259 Frequency of Hostage-Taker Integrative and Distributive Relational-Goal Category Codes for Transcript Number Two: Majors .......................... 259 Frequency of Negotiator Relational-Goal Responses by Phase for Transcript Number Two: Majors ................ 260 Frequency of Hostage-Taker Relational-Goal Responses by Phase for Transcript Number Two: Majors ............. 261 Frequency of Negotiator Integrative and Distributive Relational-Goal Category Codes for Transcript Number Three: MPD ............................................. 262 Frequency of Hostage-Taker Integrative and Distributive Relational-Goal Category Codes for Transcript Number Three: MPD ........................... 262 Frequency of Negotiator Relational-Goal Responses by Phase for Transcript Number Three: MPD ................. 263 Frequency of HostageJTaker Relational-Goal Responses by Phase for Transcript Number Three: MPD .............. 263 Frequency of Negotiator Identitthoal Responses by Transcript for all Phases .............................. 265 Frequency of Hostage-Taker Identitthoal Responses by Transcript for all Phases .............................. 266 X Table 27. Table 28. Table 29. Table 30. Table 31. Table 32. Table 33. Table 34. Table 35. Table 36. Table 37. Table 38. Frequency of Negotiator Identity-Goal Responses by Phase for Transcript Number One: HCPO .................. 267 Frequency of Hostage-Taker Identitthoal Responses by Phase for Transcript Number One: HCPO .................. 268 Frequency of Negotiator Identitthoal Responses by Phase for Transcript Number Two: Majors ................ 269 Frequency of Hostage-Taker Identitthoal Responses by Phase for Transcript Number Two: Majors ................ 270 Frequency of Negotiator Identitthoal Responses by Phase for Transcript Number Three: MPD ................. 271 Frequency of Hostage-Taker Identity-Goal Responses by Phase for Transcript Number Three: MPD ................. 272 Grand Mean Composite Language Intensity Scores for Hostage-Takers and Negotiators for all Three Negotiations ........................................... 272 Grand Mean Composite Language Intensity Scores for Hostage-Takers and Negotiators by Phase for Transcript Number One: "CPO ............................ 273 Grand Mean Composite Language Intensity Scores for Hostage-Takers and Negotiators by Phase for Transcript Number Two: Majors ..................................... 273 Grand Mean Composite Language Intensity Scores for Hostage-Takers and Negotiators by Phase for Transcript Number Three: MPD ...................................... 274 Frequency of Negotiator Identity-Goal Responses by Phase for all Transcripts .............................. 275 Frequency of Hostage-Taker Identitthoal Responses by Phase for all Transcripts .............................. 276 xi LISTWW MILL. Frequency of negotiator and hostage-taker relational- goal behavior over all phases of all transcripts. . . . .277 W Frequency of negotiator relational-goal behavior over all phases of all transcripts. ....................... 278 W Frequency of hostage-taker relational-goal behavior over all phases of all transcripts. .................. 279 W. Frequency of negotiator and hostage-taker relational- goal behavior by phase for all transcripts. .......... 280 W. Frequency of negotiator and hostage—taker relational- goal behavior by phase for "CPO. ..................... 281 W Frequency of negotiator and hostage-taker relational- goal behavior across phases for "CPO. ................ 282 W Frequency of negotiator and hostage-taker relational- goal behavior by phase for Majors. ................... 283 W. Frequency of negotiator and hostage-taker relational- goal behavior across phases for Majors. .............. 284 HEELS... Frequency of negotiator and hostage-taker relational- goal behavior by phase for MPD. ...................... 285 PM Frequency of negotiator and hostage-taker relational- goal behavior across phases for MPD. ................. 286 W. Frequency of negotiator and hostage-taker identity- goal behavior over all phases of all transcripts. . . . .287 W Frequency of negotiator and hostage-taker identity- goal behavior over all phases of "CPO. ............... 288 W Frequency of negotiator and hostage-taker identity- goal behavior over all phases of Majors. ............. 289 W Frequency of negotiator and hostage-taker identity- goal behavior over all phases of MPD. ................ 290 xii Figura_1§. Negotiator and hostage-taker mean language intensity scores by phase for all transcripts. ................. 291 Figure_16.,Negotiator mean language intensity scores by phase for all three transcripts. ........................... 292 Eightn_12..Hostage-taker mean language intensity scores by phase for all three transcripts. ........................... 293 Eignre_18. Negotiator and hostage-taker mean language intensity scores by phase for "CPO. ............................ 294 Eignza_19..Negotiator and hostage-taker mean language intensity scores by phase for Majors. .......................... 295 Figure_20.,Negotiator and hostage-taker mean language intensity scores by phase for MP0. ............................. 296 xiii INTRODUCTION AN INTERACTIINHANALYSISIOFWNEGOTIATORMAND HOSTAGBHTAKER IDENTITTHGOAL, RELATIONALHGOAL,.ANDILANGUAGE INTENSITY MESSAGEIBEHANIORINITNIN’NOSTAGEINEGOTIATIONS: AIOBSCRIPTIVE INVESTIGATION OF'THNEB NEGOTIATIONS Negotiating for the nonviolent resolution of a hostage-taking is a complex and arduous communicative task. Hostage negotiators have responsibility for persuading a hostage-taker to release his/her prisoners and to surrender peacefully. This is not an easy job when one considers that most hostage-takers are mentally and emotionally distraught, and in a state of heightened emotional excitation, prepared to either fight or flee. Because of the life threatening nature of these interactions, and the hostage-takers' desperation to gain a sense of control and certainty over his/her life, face maintenance issues add to the complexity of the interaction. Knowing what to say to reduce the hostage-taker's potentially violent affective state, to support his/her face, and to gain the release of the hostages is paramount to the negotiator's ability to gain control over the incident and facilitate a nonviolent resolution. Consequently, the physical and situational features of hostage-takings, the hostage-taker's heightenedp level of emotional excitation, and the hostage-taker's psychological I and emotional disposition are three critical determinants affecting the 2 negotiation process (Fuselier & van Zandt, 1987; Soskis & van Zandt, 1986; Taylor, 1983). ,7, Law enforcement agencies involved in hostage negotiations, such as the Federal Bureau of Investigation (FBI), encourage negotiators to engage in supportive and nonthreatening communication (e.g., listening, paraphrasing, self-disclosure, face-support) (Lanceley, Ruple, & Moss, 1985; Soskis & van Zandt, 1986; Regan, Donahue & Lyles, 1989). Specific time-phase models are also endorsed as methods to guide and time proper negotiator communicative behavior (Donohue, Kaufman, Smith & Ramesch, 1989). Implicit in these hostage negotiation strategies is the belief that specific communicative behavior on the part of the negotiator can foster open and cooperative interaction between the negotiator and the hostage-taker. More specifically, it is believed that a negotiator's use of properly timed supportive communication can enhance his/her ability to gain control of the negotiation and to elicit cooperative behavior from the hostage-taker by reducing the hostage-taker's level of physiological and emotional excitation, and by addressing his/her concern for face (Fuselier & van Zandt, 1987; Donahue, et al., 1989). The underlying premise for this belief is that communicatively negotiating with a hostage-taker is the most effective strategy for achieving a successful and nonviolent resolution as opposed to tactical assault options (FBI Training Seminar, December, 1987). While training material and literature from.both the FBI and Michigan State Police (MSP) support this proposition, its justification is equivocal and inconclusive. To begin with, theoretical and empirical support for current hostage negotiation techniques is derived 3 primarily from psychotherapeutic counseling practices. Since most counselor-patient interactions are devoid of intense conflict and mixed-motive bargaining, this makes existing negotiation strategies theoretically suspect. A sizable body of conflict behavior research has been generated which offers insight into the effects of different communicative conflict strategies and tactics on reciprocal conflict behavior and outcomes (see Chatman & Sondak, 1989; Johnson, 1973; Putnam & Jones, 1982; Putnam & Poole, 1987). For example, conflict behavior research has been conducted on the effect of promises, concession-making, commitments, threats, face-saving messages, and language intensity on conflict outcomes (Bonoma & Tedeschi, 1974; Fitzpatrick & Winks, 1979; Guyer & Rapoport, 1970; Krauss & Deutsch, 1966; Johnson, 1971, 1973). Results from these investigations have clearly demonstrated that the use of cooperative and integrative conflict tactics typically foster reciprocal integrative behavior (Chatman & Sondak, 1989; Johnson, 1973; Putnam & Jones, 1982; Putnam & Poole, 1987). Additionally, investigations of language intensity in conflict interactions has related lower lexical intensity with integrative behaviors and outcomes (Donahue, in press). According to Nilson & Putnam (1990) individual pursuit of identity and relational goals are two critical dimensions of conflict that impact conflict behavior and outcomes. Pursuit of identity goals, which is typically manifest in face behaviors, can impact conflict negatively, with individuals incurring significant losses and costs to maintain a.positive identity (Brown 1968, 1970, 1977). Relational goal pursuit, which is manifest in the message content of conflict tactics, 4 reflects the willingness of interactants to either cooperate (integrative acts) or compete (distributive acts) (Sillars, Coletti, Parry & Rogers, 1982). Yet, there is an obvious lack of integration of this body of information into current hostage negotiation literature. Given the importance of communication to hostage negotiation, this represents a critical gap in theory and research. And finally, to date, there is a.dearth of systematic scientific research that empirically investigates the relationship between a negotiator's communicative behavior with that of a hostage-taker within naturalistic hostage negotiations. More importantly, no attempts have been made to evaluate the communicative behaviors of negotiators and hostage-takers from an interactive perspective. Consequently, this dissertation proposes an investigation into the interactive process of naturalistic hostage negotiation to discern possible relationships and patterns in the communicative behavior of hostage-takers and negotiators. GENERAL DESCRIPTION OF THE STUDY Due to the lack of extensive empirical research investigating the communicative behavior of negotiators and hostage—takers in naturalistic hostage negotiations, this study is exploratory and descriptive. Specifically, this research investigation involves an interaction analysis of three actual hostage negotiations that were transcribed according to Jefferson's rules for transcription as outlined in Schenkein (1978). Trained research assistants coded the transcribed interactions according to three coding schemes devised specifically for this investigation. The schemes tapped into three 5 facets of hostage-taker/negotiator conflict behavior. The first scheme focused on hostage-taker and negotiator pursuit of identitybgoals as manifest in face behaviors. By drawing on the work of Brown (1968, 1970, 1977) a reconceptualized 2 x 3 typology of active face behaviors is presented as the basis for this coding scheme. Next, the work of Sillars et a1. (1982) provides the basis for the relationalegoal coding category. This second scheme is tri-level with eighteen subcategories split between integrative and distributive acts. A third level of this second scheme consists of backchannels and salutations, denoting structural and conversational features of interaction. Finally, a slightly modified version of Donohue's (in press) language intensity coding scheme was used to assess the level of intensity of interactant message content. The results are analyzed in a fashion to test and answer several research propositions and research questions concerning the communicative conflict behavior of negotiators and hostage-takers. RESEARCH IMPLICATIONS The implications of this study are significant. The uniqueness of this post-hoe analysis of actual hostage negotiation dialogues can shed some new light on the dynamic communicative features of hostage negotiation. A critical analysis of this nature can help both researchers and practitioners alike to better understand the basic elements of communicative interaction between hostage-taker and negotiator. The three coding schemes used are not exhaustive of all forms of conflict communication. Yet, they do provide a general framework from which the structure of negotiations can be studied. This structural framework represents a point from which future analyses 6 can build to investigate other communicative features of hostage negotiation. ORGANIZATION OF THE DISSERTATION This dissertation is divided into five chapters. Chapter One provides a description of the three dominant characteristics of hostage negotiations, emphasizing the physical context, the hostage-taker’s heightened emotional excitation, and the hostage-taker's emotional and mental instability, of which face maintenance issues are an important dimension. A reconceptualized typology for the study of active face behaviors is advanced. The importance of communication to the negotiator's ability to productively manage and control a hostage negotiation is discussed. Chapter Two provides a thorough review of conflict behavior research, with particular emphasis on the communicative aspects of conflict interaction. Special attention is given to the traditional mechanistic, psychological and systems-interpretive analyses of communicative conflict behavior. From this review, a rationale is provided for an integrated analysis of hostage negotiation which incorporates the investigation of identity-goal and relational-goal behavior as manifest in face-work activities and conflict tactics. Language intensity is also identified as a significant part of this integrated approach. This is followed by a review of the role of communicative behavior in hostage negotiation, from which several research propositions and research questions are advanced for investigation. 7 Chapter Three describes the research methodology of the study. This includes a complete description of the transcripts and the transcript coding scheme. Also included in Chapter Three is a complete discussion of the validation results for the coding procedures. Chapter Four presents the results of the data analysis. And, Chapter five offers a discussion and possible explanations for the results, a review of the implications of the results for hostage negotiations and communication-based conflict research, identification of the limitations of the study, and finally some recommendations for future research. CHAPTERIONE MOSTAGB NEGOTIATION Although humans are considered to be gregarious beings, individual self-interests often impede goal achievement for the larger social unit; whether it be a.dyad, group, organization, society or nation. To minimize incompatibility, norms, mores, folkways, and laws are created to help clarify and define behavioral expectations for the individual. Observance and adherence to these guidelines is argued to promote the betterment of both the individual and the group. Unfortunately, group goals do not always reflect the goals of individual members, nor do the goals of individuals within the group consistently mesh. When such goal incompatibility exists, interpersonal conflict is a typical resultant. The belief of many scholars is that conflict is an inevitable facet of human interaction as long as individuals exist within a social environment (Folger & Poole, 1984; Hooker & Wilmot, 1978; Miller & Steinberg, 1975). However, the jury is still out concerning the value of conflict. History provides numerous examples reinforcing the negative interpretation of conflict as a violent and destructive form of interaction. Alternatively, conflict is posited to be a necessary and productive force in promoting positive growth and change for both individuals and interpersonal relationships (Folger & Poole, 1984; Hooker & Wilmot, 1978). While this latter perspective has merit, the idea of conflict still causes many people to think of competitive, confrontational interactions that result in negative 8 9 consequences for either one or both interactants. Hostage-taking is one form of conflict interaction which typically evokes such reactions. CHARACTERISTICS OF HOSTAGE NEGOTIATION Hostage-taking incidents are complex communicative interactions representing serious social phenomena which involve an interplay of extreme physical, emotional, and psychological forces. It is the unique combination of these forces which distinguishes hostage-takings from other forms of conflict and negotiation. Specifically, hostage-takings can be distinguished by three characteristics: 1) the physical and situational context; 2) the heightened level of emotional excitation of the hostage-taker; and 3) the psychological and emotional instability of the hostage-taker. As will be demonstrated, these three features are critical determinants affecting the process of negotiation (Fuselier & Van Zandt, 1987; Soskis & Van Zandt, 1986; Taylor, 1983). Physical And Situational Parameters The physical and situational features of hostage negotiations are two-fold. First, hostage-taking incidents involve the detainment of one or more person(s), held against his/her will, by at least one other person who is using the former as a bargaining resource for achieving his/her own goals. (Fuselier, 1988; SOARU, 1985; Soskis & van Zandt, 1986). This detainment characterizes what Muir (1977) defines as an ”extortionate transaction.“ According to Muir, an extortionate transaction is an interaction in which one party is held .20 captive by a second party in an attempt to extort certain desirable outcomes from a third party. Its success rests in the expressed threat of inflicting harm to the captive party (Muir, 1977). An extortionate transaction involving people as captives can be an extremely effective compliance gaining technique since individual human life generally represents a valued and precious commodity, particularly when it is "innocent" of any wrongedoing or becomes the object of violent physical power. This type of extortionate transaction could be particularly effective in interactions with law enforcement agencies since they are charged with the primary responsibility of protecting human life and preserving social norms. Hostage—takings are coercive relationships in which hostage-takers attempt to achieve compliant behavior from law enforcement officials by threatening to harm the hostages. The hostage-taker's success is dependent upon police endorsement of the hostage as a valuable commodity (Hocker & Wilmot, 1978; Muir, 1977). Interdependency between the hostage-taker and police is established once the police acknowledge the importance of the hostage and take steps to gain his/her release. In this way, the police are also held hostage in the sense that they are controlled by the hostage-taker in their efforts to secure the safety of the hostage. An interesting paradox of this interdependency is that in holding a hostage, the hostage-taker likewise becomes a captive. Briefly, the hostage-taker is dependent upon the police for his/her own continued well-being, a condition which in turn is dependent upon the continued safety of the hostage(s). In a very real sense then, the hostage-taker is a hostage to himself. As a consequence, mutual goal interdependency is created 31 between the hostage-taker and the negotiating law enforcement officials (Muir, 1977). A crucial ingredient of this relationship is the hostage-taker's endorsement of his/her own life as one of value. A second form of extortionate transaction is the barricaded subject situation. These situations typically involve a single individual who has barricaded himself in somewhere as a result of some antecedant criminal or crisis incident (Fuselier, 1986). Because of the antecedant crisis, the barricaded subject is said to respond to his/her environment from a purely spontaneous emotional level (Lanceley, et. al., 1985). Often times, the perpetrator possess some type of weapon (e.g., gun, bomb). Consequently, the FBI considers these incidents to constitute a threat to the emotional, psychological, and physical well-being of the barricaded subject (Lanceley, et. al., 1985). As such, barricaded subject situations are similar to hostage-takings in that the barricaded individual represents a threat to his/her own well-being, and as a result becomes a hostage to himself. Yet, these situations also pose a threat to the physical safety of other individuals in close proximity to the barricaded subject, as well as the actual physical structure in which the perpetrator is barricaded (e.g., plane, building). Further, law enforcement agencies are charged with maintaining social order. Barricaded subject situations represent a violation of the social order, and require police intervention. In this way, an extortionate transaction is created, whereby society, the situation, the physical environment, the police, and the barricaded subject all become hostages. Although no individuals are directly detained by the perpetrator within a specific 12 ,4 ——-’ physical location, a hostage-taking is considered to exist because of the extortionate nature of the situation. Therefore, for the remainder of this dissertation, the term hostage-taker is used to denote the perpetrator of both barricaded subject situations and hostage-takings. This is in keeping with extant hostage negotiation/barricaded subject literature (Fuselier, 1986; Fuselier & van Zandt, 1987; Soskis & van Zandt, 1986). The second physical aspect of hostage-taking events is their fixed geographic location. Hostage-takings typically occur within specific physical parameters and are confined to that location through the direct efforts of the involved law enforcement agencies. For example, hostage-takings involving barricaded gunmen are isolated to an individual building or structure (e.g., bank, store), while hijackings involve the taking of hostages aboard some mode of transportation (e.g., bus, plane, ship). One primary objective of law enforcement officials involved in these situations is to "contain“ the hostage-taker and hostages to a specific physical location (Fuselier, 1988; Fuselier & van Zandt, 1987; Soskis & van Zandt, 1986). Police units (i.e., SWAT teams) act to control the physical environment of the hostage-taker and restrict his/her mobility by establishing inner and outer containment perimeters. The inner perimeter constitutes a restricted zone in which only the hostage-taker, SWAT units and negotiators are present. An outer perimeter represents the safety or buffer zone between the general public and the hostage-taker. Access to the outer perimeter is restricted to media and law enforcement officials. Only those directly involved in the negotiation process itself are allowed access ,13 into the inner perimeter (Fuselier, 1986; Fuselier & van Zandt, 1987; Soskis & Van Zandt, 1986). Just as entry into the zones is limited, so to is the hostage-taker's ability to exit from the inner containment area. In this way, the hostage-taker's mobility is controlled. This physical containment is a necessary requisite to the actual negotiation process and a primary feature which distinguishes hostage-takings from kidnappings and abductions (Fuselier, 1986; FUselier & van Zandt, 1987; Miran & Goldstein, 1979; SOARU, 1987a). More specifically, according to the Michigan State Police (1990), if a person is held captive, but the police do not exercise physical control over the situation, then it is an abduction or kidnapping, but not a hostage-taking; But, if the police have physical containment of the captive(s) and captive-taker, then the situation constitutes a hostage-taking; This holds true for international events as well, such that the "hostages" in Lebanon are technically not hostages, but rather captives or political prisoners. Emotional Excitation The second dominant force present in hostage-takings is emotionality (Lanceley, et. al., 1985; Miller, 1980; Hiron & ‘ Goldstein, 1979). As a consequence of the perceived and potential life threatening nature of hostage-taking incidents, participants experience an heightened level of emotional excitation. This is a normal human reaction to negative external stimuli, which serves the function of preparing the body for either "fight or flight” (Cannon, 1929). This basic psychophysiological "emergency" response is a complex interaction of three organic bodily systems: the 14 I skeletal-muscular, the autonomic nervous system, and the endocrine-hormonal system (Brady, 1970; Lacey & Lacey, 1970; Kety, 1970; Roth, 1982). An individual's skeletal-muscular reaction to a perceived threat is the most directly visible response. Increased bodily muscle strength, flexibility, and agility represent changes in the motor and postural functions which prepare an individual for either a fight or flee response to a perceived threat (Roth, 1982). Yet, less visible modifications in the autonomic nervous system occur to generate the skeletal-muscular changes. Autonomic nervous system regulation and modification is dependent upon two sub-systems: the sympathetic and parasympathetic systems (Brady, 1970; Lacey & Lacey, 1970). The sympathetic nervous system is thought to be the one system which is activated during periods of stressor stimuli and which accounts for the predominant autonomic changes (Roth, 1982). For example, sympathetic stimulation is typically responsible for increased heart rate, increased surface perspiration and evaporative water loss, pupil dilation, and increased blood flow to the muscles away from the skin surface (Brady, 1970; Lacey & Lacey, 1970). It is these autonomic modifications which facilitate the skeletal-muscular responses. The effect of the parasympathetic system during the presence of aversive stimuli is unclear. It is believed that the parasympathetic component of the autonomic nervous system.may actually be suppressed during periods of severe stress, or may serve to counteract the sympathetic alterations (Roth, 1982). As such, the specific £5 interactive effect of the sympathetic and parasympathetic systems on the skeletal-muscular system remains equivocal. Finally, stressor effects on the endocrine system are the least directly visible, yet, are perhaps the most extensive. Diverse bodily chemical changes involving such hormones as androgen, epinephrine, insulin, pepsinogen, and free fatty acids all tend to increase during periods of stress (Brady, 1970; Kety, 1970). These changes tend to place the overall system in a state of chemical imbalance. The effect is an increased susceptibility to inimical viruses when the imbalance is for a prolonged period of time. Like the skeletal-muscular and autonomic systems, the endocrine system is considered to be free from complex cognitive interpretation as a precursor to primary system modification in response to a stressor (Zillman, 1983). In other words, interpretation of a situation as potentially threatening and thinking about an appropriate response is not necessary to initial emotional excitation reactions. However, perception is central to stimuli interpretation when the period of stress is for a prolonged time period, allowing for cognitive assessment of the perceived threat and modification of behavioral response (Leventhal & Hosbach, 1983; Roth, 1982). Integration of the skeletal-muscular, autonomic and endocrine systems is manifest in emotional excitation. Insight into the effect of excitation as it occurs in hostage—takings can be gleaned from Zillman's (1971, 1974, 1979, 1983) work in excitation transfer theory. Specifically, Zillman's transfer theory of excitation postulates a cumulative effect of excitatory reaction to contiguous aversive stimuli resulting in an heightened level of emotional excitation for 16 subsequent stimuli. According to Tillman (1983), "residues of excitation from a.prior stimulation combine additively with excitation in response to subsequent stimulation, with the combined sympathetic activity expected to intensify the emotional experience and emotional behavior during the subsequent stimulation" (p.216). In other words, one's excitatory reaction to an aversive stimuli at time 2 combines additively with excitation residue from time 1 to generate an intensified emotional reaction at time 2. This combined emotional excitation is expected to be greater than would be experienced if either time 1 or time 2 stimuli were experienced independently. Operationalization of this theory is clearly manifest in hostage-takings. For instance, a hostage-taker experiences increased emotional arousal at the time of committing a crime (e.g., robbery, hijacking a plane) above the normal level of excitation necessary for human existence (i.e., pulse, heart rate, brain waves). Initial contact with law enforcement officials, who act to contain the hostage-taker, constitutes a second and contiguous stimuli which serves to generate additional emotional excitation for the hostage-taker. According to Zillman's theory then, the level of excitation following the arrival of the police is cumulatively integrated with the excitation associated with committing the initial criminal act. The consequence is a hostage-taker whose level of emotional excitation is elevated beyond the normal level that would be associated with a single stimulus. The individual is now at a level where he/she is prepared for either a fight or flight response. In addition, the individual's ability to rationally and cognitively process both incoming information and internal drive states is greatly 17 diminished (Miller, 1980; Roth, 1982). It is at this point that the hostage-taker is most likely to engage in irrational, spontaneous, and violent behavior. Thus, the primary task of the negotiator is to reduce the hostage-taker's level of excitation and to foster cognitive processing of alternative behavioral outcome options (Lanceley, et. al., 1985; Miller, 1980; Hiron & Goldstein, 1979). Psychological Attributes of the Hostage-taker The third and final characteristic feature of hostage—takings is that they are typically perpetrated by individuals who are characterized by mental and/or emotional instability (Fuselier, 1986; Lanceley, 1979, Lanceley, et. at., 1985; SOARU, 1987b; Soskis & van Zandt, 1986; Strentz, 1983, 1986). According to the FBI, hostage-takings generally occur due to one of four reasons: 1) a mentally unstable or emotionally distraught person takes hostages as either a show of force or as a plea for help; 2) criminals trapped during the execution of a crime take captives; 3) prison guards are seized as hostages during'a prison uprising; or 4) individuals are seized as an act of political extortion or terrorism (Fuselier, 1986; Lanceley, et. al., 1985; BOARD, 1974; Soskis & Van Zandt, 1986). In a recent analysis of 245 hostage-taking incidents conducted by the FBI, 60% were attributed to hostage-takers who were defined as ‘mentally'unstable or emotionally distraught (Soskis & van Zandt, 1986). A common attribute of individuals experiencing stress is high situational and personal uncertainty and ambiguity (Albrecht & Adelman, 1987; Hishel, 1984). This is a particularly salient feature for the hostage negotiation process. For the negotiator to be 18 fi// effective, he/she must assist the hostage-taker in dealing with both antecedent and current emotional stressors, as well as any perceptual incongruities of social reality. For this reason, most hostage negotiating units include a professional psychologist to assist in the negotiation process. j E-lll'fll lil'l' Because of the intense emotionality of these conflict interactions, and the predominantly unstable psychological and emotional disposition of hostage-takers, face issues represent a critical dimension of hostage negotiation. Add to this the public visibility of hostage—takings, and the salience of face for the hostage-taker is even greater. According to Goffman (1967), face is "the positive social value a person effectively claims for him/herself by the line others assume he has taken during a particular interaction" (p. 5). In other words, face is an individual sense of self as conveyed to others through verbal and nonverbal interactive behavior. Face is central to conflict interaction as it is the manifestation of an individual's identity, pride, and self-esteem (Brown, 1968, 1970, 1977; Brown 8: Levinson, 1978, 1987 ; Donohue 8: Diez, 1985; Folger & Poole, 1984; Goffman, 1955, 1956, 1959; Hooker & Wilmot, 1978). Brown (1977) and others (Folger & Poole, 1984) identify several dimensions to facedwork. These include: 1) face-saving, 2) face-loss, 3) face-restoration, 4) face-attack, and 5) faceegiving. Face-saving denotes an individual's efforts to protect him/herself from future loss of face (Brown, 1977). For example, hedges and qualifiers are verbal expressions frequently used to minimize one's 19 appearance as being weak, vulnerable, or unintelligent. These expressions represent techniques to reduce the potential loss of face through the negative evaluation of one's self—image as a consequence of the inability to maintain a positive interactive impression. Face—loss may also occur as a result of another's efforts to attack one's face. Embarrassment is the most common form of face-loss (Goffman, 1956). Face-restoration and face-saving are frequently considered to be synonymous. Yet, Brown (1977) distinguishes between the two. According to Brown, face-saving is proactive, while face-restoration is retroactive. Hence, face-restoration denotes an individual's efforts to regain a positive self-presentation lost as a consequence of self face-loss or other face-attacks. Disclaimers, retractions, clarification of previous statements, warnings, and positional steadfastness represent verbal forms of face-restoration tactics. Humor, scapegoating, and excuses are examples of face-saving tactics. While face—saving, face-loss, and face-restoration each constitute self-directed processes, face-attack and faceegiving are other-directed. Face-attack involves an individuals' efforts to negate another's positive social image. Accusations, threats, sarcasm, discredits, interruptions, shunning, and namecalling are all verbal forms of face—attack activities. Conversely, faceegiving'denotes one party's attempts to help restore another's lost self-image, lost as a consequence of perceived and actual threats to self. In essence, faceegiving’involves confirmation of another's worth (Cissna & Sieburg, 1981). Face-giving behaviors involve such activities as 20 supportive listening, conciliation, empathy, politeness, the expression of experiential similarity, and confirmation (Brown & Levinson, 1978, 1987; Gibb, 1961; Shimanoff, 1985; Spradlin, Germeroth, & Bhargava, 1989). Although Brown's (1977) typology of face-work behaviors is heuristically valuable, it is a bit incomplete and imbalanced. Briefly, as delineated in Brown's schema, there are three forms of self-directed face behaviors: face-loss, face-saving, and face-restoration; while there are only two other-directed face behaviors: face—attack and face-giving. Of the three self-directed, only face-saving and face-restoration can be argued to constitute active behavioral manifestations. Face—loss, as defined by Brown (1977) and Folger and Poole (1984), denotes one's inability to maintain a positive social image, resulting primarily from face-attacks originating with others. As such, face-loss does not constitute an active face-work behavior. In other words, face—loss is either a consequence of face-attacks, or an antecedent to face-restoration or faceegiving. Consequently, Brown's typology actually consists of only four active face behaviors: face-saving, face-restoration, face-attack, and face-giving. Scrutinization of these four remaining face activities reveals a 2 x 3 matrix of active face—work behaviors, as displayed in Table 1. One dimension distinguishes between self and other-directed activities. The second dimension consists of three face behaviors: attack, defend (saving), and restoration (restoration and giving). Host face-attack behaviors are primarily conceptualized as other-directed (i.e., "You're a fool", "You jerk"). Yet, it is also 21 common that individuals attack their own face (e.g; “I'm no good." "I'm a failure.") in an attempt to either gain socioemotional support, or simply to reinforce their own self-image. Face defend behaviors are typically viewed as proactive and self-directed ("I'm not sure, but..", "I could be wrong about this, but.."). However, individuals do provide proactive support for others in an attempt to protect them from possible face-loss (e.g. “Don't take this personally ...”, "Before we begin, I just want to tell you that you are very special to me.“). And finally, face-restoration behaviors represent an individual's attempt to restore either their own face ("Don't treat me like that.", "Where do you get off saying that to me?"), or to provide support for the face of another in a retroactive fashion, after face-loss has already occurred ("Don't be so hard on yourself.", "Don't be so concerned about that."). Table 1 about here Consequently, a reconceptualized 2 x 3 typology involving self and other orientation with the three types of face behaviors: attack, defend, and restore is advanced in this dissertation. The final six types of face behaviors include: face attack-self, face attack-other, face defend-self, face defend-other, face restore-self, and face restore-other. This reconceptualized schema serves to streamline and integrate the active face-work behaviors of Brown's (1977) original typology. Dealing with face concern issues represents a challenging task for hostage negotiators. Because of the potentially explosive behavior /" .:-,__ 22 that a hostage-taker may express in—response to an attacked face, a negotiator must frequently forego his/her own concern for face for that of the hostage-taker's. This type of selfless behavior is not easy, as the negotiator works to gain control of the interaction, and ultimately the cooperation of the hostage-taker. However, dealing with face concerns is not the only demand placed on negotiators. As discussed throughout this chapter,;hostage negotiations are unique communicative interactions in which the physical context of the incident, along with the hostage-taker's psychological disposition and level of emotional excitation affect the negotiator's ability to productively manage and direct the course of the interaction. At the outset of negotiations, the hostage-taker may be said to exert the greatest influence and control over the situation, as he/she forces the involved law enforcement officials to respond to his/her barricaded seizure. Yet, successful and productive management of hostage negotiations is dependent upon the negotiator's ability to gain and exercise control over the interaction (Fuselier, 1986; Miron & Goldstein, 1979).g CONTROL AND PRODUCTIVE MANAGEMENT OF HOSTAGE NEGOTIATION For the negotiator, control denotes a two-fold responsibility. On one level, the negotiator is partly responsible for controlling the physical environment of the hostage-taking, which includes containment of the incident to a single physical location. As noted earlier, this is what distinguishes hostage—takings from abductions. In some cases this requires affecting the environmental conditions of the hostage-taker (e.g., food, water, heat, lighting). On a second, and 23- equally important level, the negotiator is responsible for guiding and directing the behavior of the hostage-taker. The negotiator must convince the hostage-taker that it is in his/her own best interest to release the hostages and to surrender peacefully. Toward this end, the negotiator must work to reduce the hostage-taker's level of emotional excitation and stabilize his/her personal and situational uncertainty (Fuselier, 1986; Lanceley, et. al., 1985; Miron & Goldstein, 1979; Taylor, 1983). It is this second form of control which represents the most significant communicative challenge to the negotiator. As previously discussed, the psychological and emotional disposition of the hostage-taker has implications for the form and function of the communicative strategies utilized by the negotiator to gain control of the interaction. The negotiator must determine which messages will foster trust and rapport, aid in the reduction of emotional excitation, and inhibit violent behavior. The magnitude of this task reflects the critical importance of understanding the communicative features of hostage negotiations which will facilitate cooperative behavior from the hostage-taker, and result in successful, nonviolent conclusions. As such, this dissertation now turns to a review of communication-oriented research of conflict behavior, with emphasis on the interactive nature of communication within hostage negotiation.{ -1 ammo mmmmmmmm The study of conflict behavior has been an area of interest to researchers for many years, resulting in a multitude of research articles and publications (Brown, 1968; Chatman & Sondak, 1989 ; Deutsch, 1958, 1960; Deutsch 8: Krauss, 1962; Donohue, Allen, & Burrell, 1986; Donolme, Diez, & Hamilton, 1984; Donohue, Diez, & Weider-Hatfield, 1984 ; Fitzpatrick & Winks, 1979 ; Loomis, 1959 ; Mcgrath & Julian, 1963; Johnson, 1971, 1973; Putnam & Geist, 1985; Putnam & Wilson, 1987; Putnam & Jones, 1982; Putnam 8: Poole, 1987; Sillars, 1980; Swingle & Santi, 1972; Terhune, 1968 ; Thibaut & Coules, 1952;). The vast majority of these studies have looked primarily at the psychological, economic, social, and contextual factors of conflict interaction as they affect both individual and interpersonal conflict behavior. Early research of conflict behavior typically pitted student subjects against one another in some form of zero-sum or mixed-motive bargaining game (i.e., prisoner's dilemma, Acme trucking game). (Bonoma, Schlenker, Smith 8: Tedeschi, 1970 ; Borah, 1963 ; Deutsch & Krauss, 1962; Druckman & Bonoma, 1976; Gruder, 1965; Hornstein, 1965; Kelley, 1965). Many useful and interesting results were generated from these investigations. Perhaps the most important finding was the noticeable effect that oommicatim has on subjects' ability to successfully deal with, and resolve conflicts. Specifically, conflict research demonstrated that when subjects are provided with an 24 25 opportunity to communicate, the number of unsuccessful interactions is reduced (Deutsch, 1960; Johnson, 1973; Loomis, 1959; Rubin & Brown, 1975; Swingle & Santi, 1972; Terhune, 1968). No doubt this is an important finding; one which has provided the foundation to subsequent investigations. For communication is a basic and central ingredient to all forms of human interaction, including conflict. THE FUNCTION OF COMMUNICATION IN CONFLICT The centrality of communication to conflict is clearly evident in the definitional features of conflict interaction. For example, Hooker and Wilmot (1978) define conflict as: "the expressed struggle between at least two interdependent parties who perceive incompatible goals, scarce rewards, and interference from the other party in achieving their goals" (p. 23). Further, Folger and Poole (1984) define conflict as: "the interaction of interdependent people who perceive incompatible goals and interference from each other in achieving those goals." As evidenced in these two definitions,}three key features are central to the definition of conflict: interaction, interdependence, and expressed goal incompatibility (Putnam & Poole, 1987). Each of these three elements is tightly interwoven with the other two. For it is through interaction that interdependence develops, and through interdependence that interaction is maintained, which may ultimately result in the realization and expression of goal incompatibility. Communication is at the heart of this interaction. According to Putnam and Poole (1987) and Putnam and Jones (1982) communication is the very essence of human interaction. In fact, Putnam and her colleagues (Putnam & Jones, 1982; 26 Putnam and Poole, 1987) argue that communication is the "sine qua non" of all interaction, of which conflict is just one type. Communication is basic to conflict in several key ways. First, communication behavior often creates conflict (Chatman, & Sondak, 1989; Hooker & Wilmot, 1978). It is only through the expression of incompatible goals that parties establish a conflict interaction. The keyword here is "expressed' goal incompatibility. For if goal incompatibility is not expressed, either verbally or nonverbally, then no relational conflict can be said to exist (Folger & Poole, 1984). This holds true for all conflicts, including hostage negotiations. For example, if a hostageetaker fails to communicate with law enforcement officials the fact that he/she holds another person captive, thereby creating an extortionate transaction between him/herself and the police, then no hostage-taking exists (Michigan State Police Training Seminar, March, 1989). Second, communication is critical to conflict in that communication tends to reflect conflict. In other words, verbal and nonverbal behavior manifests latent goal incompatibility. The form of communicative behavior expressed by interactants mirrors the level of intensity and importance of the incompatibility for each interactant. Finally, communication is a central feature of conflict in that communication is the vehicle for either the destructive or productive management of the goal incompatibility (Folger & Poole, 1984; Hocker & Wilmot, 1978). In essence, communication is the key for dealing with } conflict. This tripartite relationship of communication to conflict is what Garfinkel (1972) describes as one of self-reflexivity. According to 27 Garfinkel's perspective, communication is concomitantly a structurally defining element of conflict, as well as a functional feature of the interaction. It is both the object and the means of conflict interaction and the negotiation process. Most of the research investigations which have attempted to assess the effect of communication in conflict have typically evaluated it as a secondary or moderating variable for other features (e.g., social context, psychological motivation, economic rewards and costs) as they affect outcomes. Putnam and Jones (1982), Putnam and Poole (1987), Johnson, (1973) and Chatman and Sondak (1989) provide excellent reviews of conflict behavior research involving communication. A synthesis of their reviews produces three broad categories of perspectives concerning the function of communication in conflict. These are the mechanistic, psychological, and systems-interpretive perspectives. This dissertation now turns to a review of research from each of these orientations. Mechanistic Interpretation of Communication in Conflict The most basic research of communication in conflict has sought to determine what effect the presence or absence of disputant communication has on conflict resolution (Deutsch, 1960; Deutsch & Krauss, 1962; Krauss 8: Deutsch, 1966; Loomis, 1959; Steinfatt, Seibold & Frye, 1974; Swingle & Santi, 1972; Terhune, 1968). Research of this orientation has looked primarily at the effect of different modes of communication, timing of information exchange, and disputants' orientation toward the interaction as they affect negotiation outcomes. 28 Numerous studies have demonstrated that when disputants are allowed to exchange information, more successful negotiation outcomes are realized than when no communication is permitted (Deutsch, 1960; Loomis, 1959; Johnson, 1973; Putnam & Poole, 1987; Shure, Meeker, & Hansford, 1965; Terhune, 1968). Generally, the longer the duration of these interactions, the greater the likelihood of a successful resolution (Johnson, 1974). While access to verbal communication facilitates positive resolutions, the potential for successful outcomes is even greater when interactants are provided with both verbal and nonverbal cues (Johnson, 1973; Putnam & Jones, 1982; Satir, 1964; Wichman, 1970). One explanation for these findings is that the ability to communicate reduces the level of formality associated with negotiation interactions and fosters increased interpersonal supportiveness that is devoid in no-communication interactions (Putnam & Jones, 1982). Alternatively, Johnson (1973) posits that unrestricted communication reduces ambiguity of perceived intentions and motivations, by allowing interactants to share information and explain their positions. According to Johnson, the combination of verbal and nonverbal cues is most effective in reducing ambiguity and facilitating successful negotiation outcomes because of the opportunity for interactants to obtain informational confirmation. While unrestricted communication has been found to have a positive effect on conflict outcomes, the orientation and disposition of the interactants tends to mediate its effect. In several experiments conducted by Krauss and Deutsch (1962) bargainers' orientations were manipulated and defined as either cooperative or competitive. When the interactants were of a cooperative orientation toward one another 29 communication was effective in fostering a negotiated resolution. But, when the disputants were of a competitive orientation, the presence of communication was less effective in facilitating a positive outcome. In fact, unrestricted communication often exacerbated the interaction because of participants' willingness to use threats (Deutsch & Krauss, 1962). Message content, motives, and perceived intent of one's opponent was argued to undermine the simple presence of communication. As such, researchers began to explore the impact of different message content on outcomes in terms of its reflection of intent. This has been the primary focus of psychologically-based conflict behavior investigations. Psychological Interpretation of Communication in Conflict The majority of research studies investigating the role of communication in conflict behavior have been of the psychological orientation (Bonoma & Tedeschi, 1974; Brown, 1968; Krauss & Deutsch, 1986; Fitzpatrick & Winke, 1979; Gahagan & Tedeschi, 1968; Guyer & Rapoport, 1970; Huseman, 1977; Johnson 1971, 1973; Linkskold, Tedeschi, Bonoma, & Schlenker, 1971; Putnam & Poole, 1987; Saine, 1974; Tjosvold, 1974). These investigations have focused primarily on the use and characteristics of different communicative messages within conflict interactions as they affect outcomes. The underlying premise of these studies has been that interactants process and interpret the communicative behavior of an opponent in order to make attributions about his/her cognitive and emotional orientation and motivations. According to Rubin and Brown (1975) conflict is a process of social influence, and accurate behavioral attributions are critical to 30 developing and implementing successful influence strategies. The research reviews of Putnam and Poole (1987), Putnam and Jones (1982), and Johnson (1973), and Chatman & Sondak (1989), along with the actual research investigations themselves, provide valuable insight into some of the majors findings concerning the psychological implications of communication and its affect on conflict behavior. A number of studies have generated findings which reveal the differential effects that various message content has on behavior and outcomes. For example messages which convey a sender's desire to cooperate, expectations for the opponent to cooperate, prescriptions for how to cooperate, along with a willingness and readiness to retaliate for uncooperative behavior typically generate perceptions of trustworthiness and cooperativeness, followed by cooperative behavior (Dorris, 1972; Druckman & Bonoma, 1976; Johnson, 1971, 1973; Loomis, 1959; Shure, Meeker, & Hansford, 1965). In addition, when a rationale is provided for why one should behave cooperatively, cooperative behavior typically results (Dorris, 1972; Johnson, 1973). Furthermore, messages which reduce the ambiguity of the message sender's intentions, his/her expectations, and which focus on the management of a task likewise produce cooperative behavior (Putnam & Jones, 1982; Terhune, 1968). A particularly interesting finding by Johnson (1967, 1971) is that paraphrasing of an opponents position and feelings also fosters cooperative behavior. The general consensus of these studies seems to indicate that positive, open, honest, tasksdirected messages are the most effective in encouraging positive negotiation results (Johnson, 1973; Putnam & Jones, 1982; Putnam & Poole, 1987). 31 Specific behaviors and message types which account for these findings include concession-making, promises, commitments, and threats. Several studies that investigated the effects of communicator concessionemaking, generated results indicating that negotiators who made positive concessions were able to elicit more reciprocal concessions than were their counterparts who made no concessions (Druckman & Bonoma, 1976; Harford & Solomon, 1967; Johnson, 1973; Putnam & Jones, 1982; Putnam & Poole, 1987). Apparently, concession-making is viewed as a willingness to cooperate. Once this willingness is expressed, it is interpreted as a commitment to cooperate, resulting in reciprocal, yet guarded cooperation (Swensson, 1967; Terhune, 1968). The use of threats in conflict has produced mixed results. For the most part, threats are viewed as a form of punishment and consequently elicit primarily negative reactions. According to Guyer and Rapoport (1970) threats tend to intensify the conflict interaction with an increase use of coercion and less successful resolutions. However, threats also foster cooperation and compliance when they are perceived as legitimate forms of retaliation, or as an attempt to save face (Bonoma & Tedeschi, 1974; Brown, 1968; Tjosvold, 1974). In comparison, promises are generally more effective in promoting cooperation than threats (Krauss & Deutsch, 1966). However, since promises are perceived as being linked to conditional rewards for cooperative behavior, their effectiveness is often mediated by the perceived credibility of the message source (Gahagan & Tedeschi, 1968; Putnam & Jones, 1982). 32 Individual orientations towards the interaction, as well as certain dispositional characteristics have been found to influence the effect of message content on conflict behavior and conflict outcomes (Johnson, 1973). Specifically, competitive (distributive) orientations commonly encourage perceptions of deceit, mistrust, and trickery, leading participants to behave less cooperatively, regardless of the specific message strategy (Deutsch & Krauss, 1962; Shure, Meeker & Hansford, 1965; Terhune, 1968) . In contrast, cooperative (integrative) orientations tend to promote more trust and cooperation (Johnson, 1973; Putnam, & Jones, 1982; Shure, Meeker, & Hansford, 1965;). Further, the psychological disposition of conflict interactants also affects behavior and outcomes. According to Bixenstine & Douglas (1967) individuals with a psychopathology tend to mistrust and violate cooperative efforts because of their psychological suspicions, making cooperative agreements even more difficult to realize. This has significant implications for hostage negotiations which consists primarily of individuals who are characterized by mental and emotional instability, and as functioning from a distributive bargaining orientation. Two explanations are offered for why individuals within conflict situations respond to particular message content and strategies as they do. The first is based on Gouldner's (1960) "norm of reciprocity." According to this perspective, individuals living within North American culture are socialized to reciprocate in kind for both positive and negative behavior that is directed toward them. As such, positive behavior is said to beget positive behavior and negative behavior to beget negative behavior. However, North American behavior, which is in 33 part based on Judeo-Christian doctrines of behavior, also encourages individuals to "turn the other cheek" to avoid hostility. Consequently, a strict reciprocity-based conceptualization of conflict behavior fails to adequately account for individuals accepting great personal loss to retaliate against perceived hostile behavior (see Brown, 1968, 1970). Alternatively, the concept of "face—saving” has been advanced as an explanatory factor in conflict behavior (Brown, 1968, 1970, 1977; Deutsch & Krauss, 1962). Willie}: As in all interactions, communication conveys both content and relational information (Watzlawick, Beavin & Jackson, 1967). The same is true for conflict interactions. Disputants exchange information concerning not only their positions and interests, but also the nature of their relationship (e.g., power distribution, affection, confirmation) (Donohue, Lyles & Rogan, 1988; Folger & Poole, 1984, Hooker & Wilmot, 1978). Within conflict interactions, pressures to maintain face are heightened as interactants exchange messages involving both content and relational information. As a consequence, interactants act in a manner to defend their self-image, to avoid appearing weak, foolish, or incapable of defending themselves when confronted by threats or aggressive behavior. Similarly, when an opponent is cooperative, it is important to maintain a positive face by likewise cooperating, rather than being hostile and uncooperative. According to Brown (1968, 1977) concern for face increases significantly when an individual is visible to a salient referent group. In fact, individuals have a tendency to engage in behaviors costly to their own potential rewards to avoid being embarrassed or 34 looking foolish. Like message content, concern for face has important implications for conflict interactions, and hostage negotiations in particular. According to Folger and Poole (1984), face-saving'and face-restoration activities, as a response to real and possible face—loss, constitute a “threat to flexibility in conflict interactions\ by adding an additional dimension to conflict“ (p. 153). Face-saving A and face-restoration efforts frequently result in conflict stalemates, impasses, costly retaliatory activities, issue avoidance, and counterattacks (Brown, 1968; Deutsch & Krauss, 1962; Folger & Poole, 1984; Goffman, 1967; Janis 8: Mann, 1977; Pruitt, 1971). Face-attack activities serve to spawn these consequences. Conversely, face-giving efforts function to reduce interactant hostility, and individual self—doubt by emphasizing mutual acceptance, affirmation, and conciliation (Cissna & Sieburg, 1981). Rather than exacerbate conflict interactions, face-giving activities tend to ameliorate tensions and foster productive conflict management (Folger & Poole, 1984). Understanding the effects of both relational and content dimensions of communicative behavior is critical to increasing the knowledgebase about how and why people behave in conflict interactions. The systems-interpretive orientation of communication in conflict offers some insight into these issues by focusing on the process of conflict interaction. Systems—Interpretive Interpretation of Communication in Conflict A variety of systems-interpretive investigations of conflict interactions have been conducted, focusing specifically on the 35 communicative features of negotiation processes (Donohue,Diez & Stahle, 1982; Donohue, Diez & Weider-Hatfield, 1984; Lewis and Fry, 1977; McGrath & Julian, 1963; Morely & Stephenson, 1977; Putnam & Geist, 1985; Putnam, Wilson, Waltman 8: Turner, 1986; Sillars, 1980; Sillars, Coletti, Parry & Rogers, 1982; Theye & Seiler, 1979; Tutzauer & Roloff, 1988). Characteristically, these studies have been concerned with describing and evaluating the sequential relationships of various communicative behaviors on negotiation outcomes, conflict cycles, phases of conflict development, and general patterns of communicative acts (Chatman & Sondak, 1989; Putnam & Jones, 1982; Putnam & Poole, 1987). The typical methodology involves coding communicative messages into behaviorally functional categories that enable researchers to identify potential regularities. The most basic systems—interpretive investigations of communication processes have generated descriptive analyses of communication behavior within the negotiation process. For example, Sillars and his colleagues (Sillars, 1980, Sillars, Coletti, Parry & Rogers, 1982) have investigated the content dimension of verbal statements made during interpersonal conflicts by categorizing them according to a three-dimensional coding scheme of conflict tactics. Within this typology, the three types of messages include passive-indirect statements, distributive (noncooperative/ individualistic) messages, and integrative (cooperative) statements. According to Sillars (1980) these three message types represent the degree to which a tactic facilitates information exchange between interactants and the degree to which the tactic encourages individual or mutual goal achievement. Passive-Indirect statements denote low 36 information exchange and an unwillingness to deal with the conflict. Distributive tactics denote high information exchange that is verbally competitive and individualistic in orientation. Conversely, integrative statements denote explicit information exchange for the purpose of fostering mutual benefit. In a study of roommate conflict interaction, Sillars (1980) asked participants to report about interpersonal problems they experienced with roommates, and to describe the types of strategies and messages used to deal with the problem. Participants were also asked to describe their level of satisfaction toward their roommate in dealing with the conflict, the length and frequency of conflicts, and the level of resolution of conflicts. Using the three dimensional coding scheme, Sillars (1980) correlated the relationship of each strategy type with the reported information. Results clearly demonstrated that integrative tactics were more closely associated with positive conflict resolution, and a higher level of satisfaction toward partner than either the passive-indirect or distributive tactics. However, in terms of overall frequency of use, integrative strategies were used much less than were passive-indirect or distributive. According to Sillars (1980) this pattern reflects a dramatic underutilization of integrative strategies in conflict interactions. Putnam and her colleagues (Putnam & Geist, 1985; Putnam & Wilson, 1987; Putnam, Wilson, Waltman & Turner, 1986) have looked at the effect of differing argumentation and bargaining strategies on the integrativeness-distributiveness of negotiation outcomes. These researchers have primarily studied teacher-union contract negotiation bargaining sessions to discern argumentation strategies as they evolve 37 during interaction. In these studies, Putnam and her colleagues (Putnam 8: Geist, 1985; Putnam 8: Wilson, 1987; Putnam, Wilson, Waltman & Turner, 1986) hypothesized that certain argumentation tactics were associated with integrative outcomes, while other tactics were related to distributive results. In one particular study, Putnam and Wilson (1987) used a seven-level coding typology, to code argument statements into one of the following categories: exploratory problem solving, initiation, ask for concessions, information giving, charges faults, threats, and commitments. The first four argument codes were hypothesized to encourage integrative outcomes, while the latter three were hypothesized to produce distributive outcomes. Results of this investigation revealed that exploratory problem-solving and asking for concessions were associated with integrative outcomes, while initiations and information giving were not. Threats were associated with distributive outcomes while charges faults and commitments were not. These findings tend to reflect the findings generated in the psychologically-oriented investigations, demonstrating that expressions of cooperative intent, and asking for cooperation typically produce positive outcomes; while, threats convey hostile intentions and generate resistance. Donahue and his colleagues (Donohue, Allen & Burrell, 1988; Donohue & Diez, 1985; Donohue, Diez 8 Hamilton, 1984; Donohue, Diez & Stahle, 1983; Donohue, Lyles & Rogan, 1989) have conducted several investigations of naturalistic interactant behavior within conflict contexts. A large part of this research has evaluated the communicative behavior of mediators during divorce mediations sessions 38 (Donohue, Allen & Burrell, 1988; Donohue, Diez, & Hamilton, 1984; Donohue, Lyles & Rogan, 1989). These studies have produced findings which identify the characteristics of successful mediators. For example, successful mediators were those who focused more on the factual and interest concerns of the disputants, than relational issues (Donohue, Allen & Burrell, 1988; Donohue, Lyles & Rogan, 1989). Successful mediators have also been identified as working to structure the interactions more with formal rules for behavior, and as attempting to reframe the interactants' positions (Donohue, Allen, & Burrell, 1988). Other studies conducted by Donohue and his colleagues have focused more on the interactive behavior of disputants involved in conflict. In one study conducted by Donohue and Diez (1985) these researchers investigated the effect of negotiation context variables (i.e., participant goals and interests, negotiation procedures, relational history, and the topic of interaction) on the use of directive statements. According to Donohue and Diez (1985) directives are illocutionary acts concerned with informational requests. These researchers argued that there are four features to directives: level of politeness, which includes the concept of "face," relational features of the interaction, illocutionary force embedded within the directive, and the contextual focus of the directive. Five separate conflict interactions were analyzed for the form and functions of directives used. One major finding was generated from their investigation. Specifically, face threatening directives were found to be used more frequently when individuals' goals are discrepant, the context is less cooperative, interactants' have strong, negative relational history, 39 and the interaction is of high personal involvement. In other words, aggressive directives are typically associated with participant's attempts to assert themselves in response to the contextual and relational characteristics of the interaction when these features appear to be inimical to the individual. In another study, Donohue (in press) investigated the effect of language intensity on conflict behavior within a mediation context. Language intensity is defined as the ”the quality of language which indicates the degree to which the speaker's attitude toward a concept deviates from neutrality" (Bowers, 1963, p. 345). According to Donohue (in press), language intensity is a common feature of conflict interaction that is associated with distributive behavior. As such, he posited that successful mediations should be characterized by language low in intensity, but high in cooperativeness and flexibility. Much like Sillars (1980), Donohue hypothesized that message intensity affects a person's attribution about the communicator, and consequently the opportunity for successful resolution. To investigate this hypothesis, Donohue coded the message behavior of interactants involved in divorce mediation sessions according to six levels of intensity. These included obscure words, metaphors, qualifiers, profanity, sex metaphors, and death metaphors. Twenty transcripts, consisting of ten successful and ten unsuccessful mediations were used. Employing the five correlation scores for language intensity developed by Bowers (1964), Donohue (in press) computed composite weights of intensity for each category. Analysis of the results indicated that lower language intensity messages were associated with successful mediations. Donohue (in 40 press) posited that intensity may also be a strategy for exercising control over an interaction, since the mediators' level of intensity was consistently higher than either of the interactants. Further, he posited that interactant predisposition to either integrative or distributive strategies, and a‘genuine desire to reach a settlement may have influenced the level of language intensity within the different sessions. Nonetheless, the general conclusion is that more cooperative behavior is associated with lower language intensity. This finding reflects the importance of considering the multiple dimensions of message behavior within conflict interactions. Effect of Communication on Conflict Interactions In summary, research of communication in conflict has generated some fairly consistent results. Mechanistic research has demonstrated that communication is a critical variable in the resolution of conflict (Deutsch 8 Krauss, 1962; Druckman & Bonoma, 1976; Hornstein, 1965; Krauss & Deutsch, 1966; Loomis, 1959; Terhune, 1968). Yet, it has also been demonstrated that not all forms and types of communication encourage integrative outcomes (Deutsch & Krauss, 1962; Krauss & Deutsch, 1966). Rather, how and what an individual says influences the behavior of the opponent and the consequent outcome. In general, "being nice" has been found to foster integrative agreements (Putnam & Jones, 1982; Theye & Seiler, 1979). More specifically, as messages increasingly involve an expression of unambiguous intent, expectations for cooperation, explanations for how cooperative behavior should be fulfilled, and a.willingness and readiness to retaliate for uncooperative behavior, then the more likely 41 cooperative behavior will occur and a successful resolution will be realized (Brown, 1968; Dorris, 1972; Guyer & Rapoport, 1970; Johnson, 1971; Terhune, 1968; Tjosvold, 1974). Further, messages which help to reduce ambiguity of motivations, and which encourage fairness and cooperation will likewise facilitate cooperative behavior (Krauss & Deutsch, 1966; Terhune, 1968). And as Johnson (1967, 1971) has reported, accurate paraphrasing of another's feelings and position (i.e., comforting messages) will also help in reducing uncooperative behavior. As demonstrated from the systems-interpretive research, integrative, problem-solving messages of lower intensity will also encourage cooperation and integrative outcomes (Donohue, et al., 1984; Donohue, et al., 1988; Putnam 8: Wilson, 1987; Sillars, et al., 1982). Taken together, the vast number of research studies of communication in conflict generally support the proposition that the more integrative and cooperative one person is in a conflict, the greater the possibility that the other interactant will respond with cooperative and integrative behavior. Conflict research has also demonstrated that an individual's concern for face can greatly influence the conflict interaction, prompting individuals to incur significant costs and losses to either protect or regain a lost positive face by engaging in highly competitive conflict behavior (Brown, 1968, 1970; Deutsch & Krauss, 1962; Folger & Poole, 1984; Janis & Mann, 1977). Therefore, face-work activities represent a second critical dimension to conflict interaction. However, most conflict behavior research has only evaluated message content in terms of the tactical and strategic function (Donohue & Diez, 1985; Donohue et al., 1988; Sillars et al. 1982). 42 Coding techniques of this kind have occasionally assessed face behaviors, but only on a general level, and just as another type of conflict tactic (i.e., threat, support) (Donohue et al., 1989; Sillars, 1980; Sillars, et al., 1982). The specific assessments of facedwork have primarily involved experimental manipulations of an individual's self-image in the form of embarrassment, to determine the impact of face-loss on subsequent conflict tactics. No research has looked specifically at either the relationship or effect of positive face messages on conflict tactics. Further, research of both face and tactics has failed to adequately tap into the interactive nature of conflict processes. More specifically, past research has typically treated face as a unidimensional, static variable within the conflict interaction, while conflict tactics are treated as dynamic. In actuality, both are dynamic variables which can be investigated as independent, yet interrelated components of conflict behavior. As such, a bilevel analysis of conflict interaction seems warranted. AN INTEGRATED APPROACH TO STUDYING COMMUNICATIVE CONFLICT BEHAVIOR Conflict has been defined as an interactive process in which participants perceive the achievement of their individual goals to be incompatible, or hindered by an opposing party (Folger & Poole, 1984, Hooker & Wilmot, 1978). Consequently, goals represent a central feature of conflict behavior. Yet, the idea that conflict behavior is rational and goal-based requires some clarification. As discussed by Craig (1986) and O’Keefe (1988), goals can be defined as either ”intentional" or "functional." 43 According to Craig, intentional goals denote cognitively recognized desired and states that an individual desires from a particular interaction. For example, wanting food when one is hungry constitutes an intentional goal. Comparatively, functional goals represent behaviors that are implicitly defined as inherent to a particular situation. For example, presenting a positive face (impression management) is intrinsic to all social interactions. But, wanting a positive face does not constitute a cognitively desired and state in the same way that wanting food does. It is always inherently present to the situation, but not a rational goal of all interlocutors in all interactions (Brown & Levinson, 1978). O'Keefe (1988) makes the point that "functional" goals are "socially constituted objectives" that are implicit in predicted human interactions, whereas "intentional" goals denote "individually selected and recognized" objectives (p 82). Therefore, a rational goal-based message design logic, such as those developed by O'Keefe and her colleagues (O'Keefe, 1988; O'Keefe & Delia, 1982; O'Keefe & Shepard, 1987) can be used to describe the functional goals of communicative behavior within a given situation. This descriptive use is the primary purpose of Wilson's and Putnam's (1990) goal-based typology for argumentative behavior. According to Wilson and Putnam (1990) negotiation participants pursue three general types of interaction goals: instrumental, relational, and identity. Instrumental goals are specific individual Objectives which require either modification or the elimination of an opponent's objectives. Specific demands, wants, interests, and needs are all forms of instrumental goals. Relational goals are reflected in 44 an individual‘s efforts to either maintain a symmetrical relationship with an Opponent, or to gain the superior position within a complementary relationship (Watzlawick, Jackson, Beavin, 1967). Distribution of power and control, willingness to cooperate, resistance, and threats are all characteristic features of relational goals. Finally, identity goals denote an individual's desire to create or maintain a positive social image. Face-work activities are the primary indicants of identity goals. Like face-work types, identity goals can be either positive (support, defend), or negative (attack). In this way, identity goals also reflect the nature of conflict interactants' relationship. As such, identity goals are interdependent with relational goals, and both identity and relational goals are reflective of an individual's efforts to effect his/her instrumental goals. The separation of both identity goals (face behaviors) and relational goals (conflict tactics and strategies) from conflict message content can provide valuable insight into the dynamic interactive relationship of conflict participants. More specifically, by assessing message content for both types of goals (identity and relational), knowledge can be acquired about the transactional, interactive communicative behavior of conflict participants, and the interrelationship of their individual communicative behaviors in their pursuit of identity and relational goals. The benefit of such a bilevel analysis would be the disclosure of the unfolding interaction between interactants as they exchange information to structure and create their own unique relationship, juxtaposed to their individual goal pursuit. 45 Increased knowledge of this kind has significant implications for hostage negotiations. As discussed in Chapter One, both face issues and control are central to the hostage negotiation process. Gainrmg insight into the interrelationship of a negotiator's and hostage-taker's identity goal-based and relational goal-based message content, as it relates to the message content of each participant, could aid in understanding the successful, nonviolent resolution of hostage-takings. By understanding the form, function, and interrelationship of various message behaviors, predictions could be made concerning the use of message content within the negotiation process. This requires analysis of hostage negotiation interactions. Yet to date, there is a lack of communication-based research of either simulated or naturalistic hostage negotiations. Consequently, this dissertation proposes an interaction analysis of naturalistic hostage negotiations, and by drawing on the results of other communication—oriented studies of conflict interaction, advances several research propositions concerning the relationship of a negotiator's behavior with that of a hostage-taker. //.\_‘ COMMUNICATION BEHAVIOR IN HOSTAGE NEGOTIATION E The importance of communication to the hostage negotiation process , is emphasized by domestic law enforcement agencies responsible for conducting domestic hostage negotiations. According to the FBI, communicative interaction with a hostage-taker will be effective in resolving 96% of all hostage—takings (FBI Training Seminar, December, 1987). This claim is further supported by the Michigan State Police who contend that they have implemented tactical options in only 4 out 46 of 400 incidents (Michigan State Police Training Seminar, March, 1989). Unequivocally, communication is an integral feature in the nonviolent, successful resolution of hostage-takings. Still, the issue of which communicative techniques are appropriate and effective remains unanswered. From a strict compliance-gaining perspective, a variety of communication strategies and tactics could be used to encourage desirable behavior from the hostage-taker. For example, Harwell and Schmitt (1967) have developed a scheme of sixteen different compliance-gaining tactics. Similarly, Wiseman & Schenk-Hamlin (1981) have assembled a taxonomy of fourteen compliance-gaining strategies. Both frameworks include common. techniques which range from positive reward-based motivations to strong negative sanction strategies. While the ultimate goal of the negotiator is to gain the safe release of the hostage(s) and the surrender of the hostage-taker, hostage negotiation is a slightly more complex process than merely trying to persuade someone to do something. Chapter One highlighted some of the unique features of hostage negotiation that can complicate the negotiation process. Perhaps the most critical complications are the hostage-taker's heightened level of emotional excitation, sense of personal uncertainty, and concern for face (Miller, 1980). An effective negotiator should be able to facilitate excitation reduction, stabilize the hostage-taker's sense of well-being, and not threaten his/her face to minimize potentially violent behavior (Lanceley, et. al., 1985; Miron & Goldstein, 1979). Toward this end, it is important for the negotiator to redefine the conflict for the hostage-taker from a zero-sum situation of only limited and violent options, to a 47 perceived win-win solution of peaceful surrender (Taylor, 1983). It is necessary for the negotiator to reframe the hostage-taker's perception of the interaction from one of competitiveness (distributive) to one of cooperativeness (integrative) (Sillars, 1980; Sillars, Coletti, Parry, & Rogers, 1982). In actuality, the hostage-taker's only options are death or surrender. Nonetheless, the negotiator must help the hostage-taker to process perceived options that negate his/her own death, the death of the hostage(s), and/or a tactical assault by the police, and which encourage a nonviolent surrender. Undoubtedly, the negotiator's attempts to guide the thinking and behavior of the hostage-taker is a complex and arduous task. Because of the predominance of psychoemotional factors, and given the prevalence of mentally and emotionally distraught hostage-takers, hostage negotiators have looked primarily to the psychotherapeutic and counseling literature for guidance (Fuselier, 1986; Lanceley, 1979; SOARU, 1987a; Strentz, 1983, 1986). Active listening, paraphrasing, illocutionary pauses, self-disclosure reciprocity, and open-ended questions are just a few of the psychotherapeutic counseling techniques which have received endorsement from law enforcement agencies engaged in hostage negotiation. These methods are seen as enhancing the negotiator's control over the interaction and for reducing the hostage-taker's level of emotional excitation (Fuselier, 1981; Fuselier & van Zandt, 1987; Lanceley, et. al., 1985; Miron & Goldstein, 1979; SOARU, 1985, 1987a). The timing and placement of these illocutionary acts within the negotiation process is considered to be a critical factor. Consequently, various time phase models of negotiation, such as that employed by the Michigan State Police, have been proposed. 48 Stages of Hostage Negotiation To effect successful, nonviolent resolutions, the Michigan State Police have proposed five functional stages of hostage negotiation to help guide and structure the behavior of negotiators (Donohue, Kaufman, Smith, & Ramesch, 1989). The first stage is the "Pre-Work, Intelligence Gathering" phase. This stage precedes actual communicative interaction between negotiator and hostage-taker, occurring during the first hour or two of police arrival at the scene. During this phase, police units involved in the hostage-taking gather intelligence concerning the physical conditions of the situation, the hostage-taker, and hostages, along with as much psychological S,ae/, 77/~EZ; background information about the hostage-taker as possible. Stage two is the "Introduction and Establish Contact" phase. This phase constitutes the actual beginning of the negotiation process, and occurs at the outset of negotiation interaction. The primary task of the negotiator during this stage is to convey an image of concerned helper, elicit dispositional information about the hostage-taker and hostages, and attempt to stabilize the situation. The third stage is the "Relationship Building” phase, accounting for approximately two to eight hours of negotiation. During this time period, the negotiator strives to create a trusting relational bond with the hostage-taker and to increase his/her understanding the of the hostage-taker's thoughts and feelings. Towards this end, the negotiator conveys support, confirmation, and understanding, as disclosing trivial personally private information. Providing face support is also considered a critical part in building a relationship 49 during this phase. No attempts at problem—solving or problem resolution occur at this time. Problem negotiation takes place during the fourth phase, "Problem Negotiations." During this time period, the task of the negotiator is to help direct the hostage-taker into devising his/her own nonviolent surrender by encouraging cooperative and integrative behavior. The focus of the interaction at this point in time is on needs and interests, and not on specific resolution techniques. Estimated time frame for this phase is the eighth through twelfth hour of interaction. The fifth and final stage is the "Resolution" phase, during which the negotiator works with the hostage-taker to create a specific resolution/surrender strategy. While this final stage is often very difficult to work through, success at this point is dependent upon the negotiator's ability for having established a trusting relationship with the hostage-taker during stages two through four. Implications for Communication in Hostage Negotiation While FBI training material and other reference articles discuss the use and value of various communicative tactics and strategies for use in hostage negotiations, they fail to take into account several important issues. First, hostage negotiations, the actual interaction between negotiator and hostage-taker, tend to be viewed as occurring within a communicative vacuum, independent from other extraneous factors, noises, and other participants. For the most part, law enforcement agencies do acknowledge the potential influence of the media, hostages, and other possible unofficial negotiators as possible complications to the negotiation process. Yet, there is a failure to 50 specify just how these factors relate to, and influence negotiations. Even the phase structured model of the Michigan State Police, while acknowledging the interactive nature of hostage negotiations, looks only at the interaction between negotiator and hostage-taker, excluding other possible interactions between the hostage-taker and other negotiators (e.g., relatives, priests, friends, media). Ideally, law enforcement officials attempt to restrict these other interactions, yet they do occur. As such, these other negotiation interactions do impact the negotiation process between hostage-taker and negotiator. Consequently, these extraneous processes should be identified, acknowledged, and evaluated. Second, interpretation of specific hostage negotiation between a negotiator and hostage-taker appears to be based on a unidirectional conceptualization of communication, where what the negotiator says is perceived to relate to and affect the behavior of the hostage-taker, but not reciprocally. More specifically, there is a lack of acknowledgement of communication as a reciprocal, interactive, transactional, process wherein both negotiator and hostage—taker, in pursuit of their individual instrumental, identity, and relational goals, structure and influence the direction of the interaction through the verbal content of their message behavior (Miller 8 Steinberg, 1975). Both the negotiator and hostage-taker engage in a transactional interactive process of sequential turn-taking to achieve their individual goals. Consequently, hostage negotiation needs to be investigated from an interactive process perspective, where message content is assessed to determine the relationship of message behavior for goal achievement between interactants. 51 Finally, the majority of theoretical and empirical support for negotiation techniques comes from psychotherapy. Compared with hostage negotiations, counselor - client interactions are nonconflictual and noncompetitive. Thus, there emerges a concern about the direct application of counseling strategies to hostage negotiations. Communication-based conflict behavior research of naturalistic hostage negotiations is needed to empirically investigate the relationship of the negotiators' communicative behavior with that of hostage-takers'. An analysis of conflict message behavior can help explain the occurrence and relationship of strategies and tactics within hostage negotiations. This dissertation proposes such an investigation. Although all three of these issues represent significant conceptual and operational shortcomings to extant hostage negotiation literature, this dissertation will attempt to address only the latter two problems. No doubt, all three are critical. Yet, by addressing the issues of reciprocal interaction between negotiator and hostage-taker, and the relationship of their individual message content, insight can be gained to build a theoretic foundation upon ' which future research into the holistic process of hostage negotiation can be conducted. RATIONALE AND SCOPE OF THE STUDY Hostage negotiations have been described as highly charged communicative conflict interactions in which the physical containment of the situation, the hostage-taker's heightened level of emotional excitation, and psychological disposition are critical determinants of 52 the negotiation interaction and situational outcome. Due to their level of excitation and uncertainty, hostage-takers have been defined as being extremely concerned about protecting their face, and their desire to influence the situation. Yet, negotiator control of the negotiation process and the overall interaction has been identified as critical to the nonviolent and successful resolution of hostage-takings. Toward this end, law enforcement agencies involved in hostage negotiations have endorsed the strategy of having negotiators employ various comforting behaviors (i.e., listening, self-disclosure, self-disclosive questioning) (Fuselier, 1981; Fuselier 8 Van Zandt, 1987; Lanceley, et. al., 1985; Miron 8 Goldstein, 1979; SOARU, 1985, 1987a). Further, specific time phases have been identified to guide and structure negotiator behavior (Donohue et al., 1989). The rationale for these techniques derive primarily from psychotherapeutic counseling literature and research. While communication is acknowledged to be critical to hostage negotiations, there is a paucity of integration of communication-based conflict behavior research with hostage negotiations. Yet, communication-based conflict behavior research has generated some potentially significant results for application within hostage negotiations. Specifically, findings have been generated which indicate the effect of different conflict tactics on conflict outcomes. For example, conflict research has clearly demonstrated that an individual's use of cooperative and integrative tactics generally results in reciprocal cooperative and integrative behavior when compared to the use of competitive, distributive tactics (Chatman 8 Sondak, 1989; Johnson, 1973; Putnam 8 Jones, 1982; Putnam 8 Poole. 53 1987). Further, language intensity has also been demonstrated as relating to conflict outcomes, with lower language intensity positively related to integrative outcomes (Donohue, in press). Concern for face has also been demonstrated to be a significant factor influencing conflict behavior, with face-loss and face-attacks prompting individuals to incur significant costs, or losses in order to either protect or regain a.positive face (Brown, 1968, 1970). Other conflict research as posited that interactants' concern and pursuit of individual relational and identity goals are reflected in the message content and type of conflict tactics employed (Sillars, et al., 1982; Wilson 8 Putnam, 1990). Specifically, face behaviors reflect identity goals, while integrative and distributive tactics convey relational goal pursuit. In sum, several significant features of conflict behavior which are communication-based have been identified and evaluated. However, to date there is a dearth of research which attempts to investigate hostage negotiation from a communication-based conflict perspective. Furthermore, no attempts have been made to assess naturalistic negotiation behavior of hostage-takers and negotiators from an interactive, relational perspective. Consequently, this dissertation proposes such an investigation. By drawing upon both hostage negotiation and conflict behavior research, this dissertation advances several research propositions and research questions concerning the relationship of hostage-taker and negotiator communicative behavior. Because of the lack of extensive empirical research investigating the communicative behavior of negotiators and hostage-takers in 54 naturalistic hostage negotiation interactions, this study is exploratory and descriptive. Specifically, this investigation involves an interaction analysis of three authentic hostage negotiations. Trained research assistants coded the transcribed interactions for the presence of identityegoal statements (i.e., face-work behaviors), relational-goal statements (i.e., integrative/distributive conflict tactics), and language intensity according to a devised coding scheme. The results are analyzed in an effort to test and answer the following research propositions and research questions. Research Propositions and Research Questions Iima_Ehase_Mndel_n£_HQstaEe_N£EQLiatiQn The Michigan State Police time phase model of hostage negotiation has been cited as a general behavioral framework for hostage negotiators. As outlined earlier in this chapter, the model posits five phases to the negotiation process. Phase One is the "Pre~Work, Intelligence Gathering" stage. No actual negotiation transpires during this period. Police units merely collect background information on the hostage-taker and hostage situation. Phase Two begins the actual interaction. This is the "Introduction and Establish Contact" phase. During this time the negotiator identifies him/herself and attempts to portray him/herself as a concerned helper. Phase Three, "Relationship Building” requires the negotiator to develop a positive relational bond with the hostage-taker. Face support and empathy are considered as critical to this phase. Phase Four “Problem.Negotiations" involves the negotiator working with the hostage—taker to develop solutions to the hostage-taker's problem(s). Emphasis is on the hostage-taker 55 developing his/her own solutions rather than the negotiator telling the hostage-taker what to do. The final stage "Resolution" is the point where the negotiator works with the hostage-taker to develop a specific resolution/surrender strategy. Specific behavioral instructions are now conveyed by the negotiator to the hostage—taker. According to the Michigan State Police, this model represents the desired hostage negotiations procedure for negotiators to follow. However, this model has never been empirically evaluated, nor tested in relationship to negotiator and hostage-taker conflict tactic message behavior. Still, given the behavioral expectations for negotiators during each of the four actual negotiation phases and past behavioral trends of hostage-takers, certain predictions can be made concerning the relationship between hostage-taker and negotiator behavior within each of the time periods. Therefore, research propositions concerning negotiator and hostage-taker expression of identity-goal behavior, relational-goal message content, and language intensity are defined in terms of occurrence across the four negotiation stages. Consequently, analysis of negotiator and hostage-taker behavior is in terms of the research propositions and questions according to the four active phases of negotiation outlined in the MSP time phase model. I! l'l G 1 I B J |° J G J C EJ° | E l . As reviewed earlier in this chapter, a significant amount of research has been devoted to investigating the form and function of conflict tactics. Empirical research from the psychological orientation of communication in conflict has revealed that tactics which convey a desire to cOOperate, expectations for cooperation, and a willingness to retaliate for hostile behavior generally promote an 56 integrative resolution to the conflict (Chatman 8 Sondak, 1989; Johnson, 1973; Putnam 8 Jones, 1982; Putnam 8 Poole. 1987). In terms of relational-goal orientation, these types of tactics reflect integrative behavior (Sillars, et al., 1982; Wilson 8 Putnam, 1990). Conversely, hostile and aggressive conflict tactics typically generate reciprocal behavior (Deutsch 8 Krauss, 1960, 1962; Guyer 8 Rapoport, 1970; Hornstein, 1965;). Again, from a relationalegoal perspective, such tactics reflect distributive conflict behavior. Reciprocity has been argued to be a key to understanding conflict interaction (Gouldner, 1960). Specifically, aggressive and hostile behavior is said to foster competitive behavior, while cooperative and integrative behavior is said to promote noncompetitive behavior. Although reciprocity is inadequate to account for all conflict behavior, it can provide a general theoretical framework for understanding patterns of conflict interaction. Consequently, the following research propositions are advanced concerning the relational goal conflict behavior between hostage-takers and negotiators. RPl. Relationship Between Negotiator Use of Integrative Relational-Goal Conflict Tactics and Hostage-Taker Use of Integrative Relational-Goal Conflict Tactics RPlat The frequency of negotiator use of integrative relationalegoal conflict tactics will remain constant over time. RPlb. The frequency of hostage-taker use of integrative relationalegoal conflict tactics will increase over time. 57 RPlc. The difference in frequency between negotiator use of integrative relational-goal conflict tactics and hostage-taker use of integrative relationalegoal conflict tactics will decrease over time. Because of the unique nature and characteristics of hostage negotiations and hostage-takers as outlined in Chapter One, police negotiators are encouraged to foster and create as integrative a conflict interaction environment as possible (Donohue, et al., in press; Lanceley, et al., 1985; Taylor, 1983). Negotiators strive to present the hostage-taker with the impression that the negotiator is working with and for the hostage-taker, and that the hostage-taker has options other than killing him/herself and the hostages (Miron 8 Goldstein, 1979; Miller, 1980; Taylor, 1983). Consequently, expectations for negotiator behavior in response to hostage-taker behavior deviates from traditional negotiation and bargaining interactions. As such, the following research propositions are proposed concerning negotiator integrative conflict behavior in response to hostage-taker distributive conflict behavior. RP2. Relationship Between Hostage-Taker Use of Distributive Relational-Goal Conflict Tactics and Negotiator Use of Integrative Relational-Goal Conflict Tactics RP2a, The frequency of hostage-taker use of distributive relationalegoal conflict tactics will decrease over time. 58 RP2b. The difference in frequency between hostage-taker use of distributive relational-goal conflict tactics and negotiator use of integrative relational-goal conflict tactics will decrease over time. Concern for one's face has been posited as an alternative explanation and cause for individual conflict behavior (Brown, 1968, 1970, 1977; Deutsch 8 Krauss, 1960). Facedwork behavior research by Brown (1968, 1970) has demonstrated that the loss of face or perceived threat to face can prompt an individual to incur significant costs and losses in his/her own potential rewards to protect and regain that positive face. Specifically, perceived and actual loss of face results in the use of aggressive and hostile conflict tactics (Brown, 1968, 1970). Conversely, research by Johnson (1967, 1971) has revealed that conflict tactics which convey support and understanding (i.e., paraphrasing other's feelings and position) can facilitate cooperative conflict behavior. This link between face issues and conflict behavior is described by Wilson and Putnam (1990) as an individual's simultaneous pursuit of identity and relational goals. The interplay of identity and relational goal behavior has been argued to be a significant feature of hostage negotiations. Emphasis on positive face support messages is stressed for the negotiator as a method to reduce emotional intensity and foster a relational bond with the hostage-taker (Donohue, et al., in press). Therefore, the following research propositions are advanced. 59 RP3. Relationship Between Negotiator Use of Positive Face-Other IdentityeGoal Tactics and Hostage-Taker Use of Integrative Relational-Goal Conflict Tactics RP3a. The frequency of negotiator use of positive face-other identity-goal tactics will remain constant over time. RP3b. The difference in frequency between negotiator use of positive face-other identity-goal tactics and hostage-taker use of integrative relational-goal conflict tactics will decrease over time. RP4. Negotiators will use more positive face~other identity-goal conflict tactics (face defend-other, face restore-other) than positive face-self (face defend-self, face restore-self) and negative face-other identity-goal conflict tactics (face attack-other) in response to hostage-taker use of negative face-self tactics (face attack-self). Again, because of the unique nature of hostage negotiations, and the novelty of this investigation, there exist several questions concerning the relationship between negotiator and hostage-taker identity-goal and relational-goal conflict behavior. Consequently, the following research questions are posed to further investigate this relationship. RQl. ROB. ROS. 60 What is the frequency of negotiator and hostage-taker use of integrative and distributive relationalsgoal conflict tactics over time? . What is the difference in frequency between hostage-taker use of integrative relational-goal conflict tactics and negotiator use of distributive relational-goal conflict tactics over time? What is the frequency of negotiator and hostage—taker use of the different identity-goal conflict tactics over time? . What is the frequency and percentage of response of negotiator use of relational-goal conflict tactics (integrative/distributive) in response to (n + 1) hostage-taker use of negative identity-goal face-other tactics (face attack-other)? What is the frequency and percentage of response of negotiator use of identityegoal conflict tactics in response to (n + 1) to hostage—taker use of negative face-other tactics (face attack-other)? I I I 'l C [1' l E l . Drawing on the work of Bowers and his colleagues (1963) language intensity is defined as "the quality of language which indicates the degree to which the speaker's attitude toward a concept deviates from 61 neutrality" (p. 345). Based on this definition, language intensity is interpreted to be an indicator of a communicator's attitudinal disposition. As such, intensity of language is a means by which interactants express their attitudes and reactions toward each other, as well as serving as a marker for message source attributions (Bradac, Bowers, 8 Courtwright, 1979). Most research of language intensity to date has involved the evaluation of the effect of persuasive messages manipulated for intensity, to assess the impact on attitude change (Bowers, 1963; Bradac, et al., 1979; Burgoon, Jones, 8 Stewart, 1971; Burgoon 8 Miller, 1971; Infante, 1975). Results of these investigations indicate that language intensity affects the inferential processing of information, such that a receiver's attributions about a message source are influenced, thereby affecting attitude change (Bradac,et al., 1979; Burgoon 8 Miller, 1971; Carmichael 8 Cronkhite, 1965;). Further research has demonstrated that cognitive stress and language intensity are inversely related (Bradac, et al. 1979). Specifically, both high and low stress tend to correspond positively to messages characterized by low language intensity, while moderate levels of stress correspond positively to high message intensity (Burgoon 8 Miller, 1971). Operationalization of stress within these studies ranged from.perceived audience hostility to suicidal expressions. Although the manipulations are very different, these findings still have implications for hostage negotiations. In Chapter One, hostage negotiations were characterized as involving an extreme emotional intensity on the part of the hostage-taker. A simple qualitative review of naturalistic hostage 62 negotiations reveals a great deal of stress and intense, emotional language. One possible way to differentiate a hostage-taker's level of stress is to evaluate his/her willingness to live. If a hostage-taker who is in a life threatening interaction wants to live, we could speculate a fairly high level of emotional intensity reflected within his/her message behavior. In comparison, individuals who are intent on committing suicide can be argued to no longer be experiencing high levels of stress as they have already determined their fate. It is when the fate is uncertain, that emotional excitation is the greatest. Hostage negotiations are typically characterized as involving hostage-takers who want to live, yet who are also uncertain about their fate. Consequently, hostage-takers’ level of emotional excitation is extremely high. For this reason, language intensity is considered a critical lexical feature for investigation within hostage negotiation. As also discussed earlier in this chapter, recent work by Donohue (in press) involved the evaluation of message intensity from naturalistic divorce mediation sessions using a coding scheme derived from Bowers' (1964) typology of intensity correlates. Other language intensity research fails to look at naturalistic interactions, and assess level of intensity. As such, Donohue's study represents a unique attempt to investigate language intensity. Results of Donohue's investigation indicate that successful mediations were characterized by low language intensity, while unsuccessful mediations were characterized by relatively higher language intensity. In other words, low language intensity is more positively related with integrative conflict outcomes and high language intensity with distributive outcomes. 63 Drawing primarily on the work of Donohue (in press) the following research propositions and research questions are forwarded concerning the relationship between negotiator and hostage-taker language intensity and conflict message behavior. RPS. Relationship Between Language Intensity and Integrative RP5a RPSb. RP5c. RPSd. RP5c. RPSf. Relational-Goal Conflict Tactics There will be a negative correlation between negotiator mean language intensity and the frequency of hostage—taker use of integrative relational-goal conflict tactics. There will be a positive correlation between hostage-taker mean language intensity and the frequency of negotiator use of integrative relational-goal conflict tactics. There will be a negative correlation between negotiator mean language intensity and the frequency of negotiator use of integrative relational-goal conflict tactics. There will be a positive correlation between negotiator mean language intensity and the frequency of negotiator use of distributive relational-goal conflict tactics. There will be a positive correlation between hostage-taker mean language intensity and the frequency of hostage-taker use of distributive relationalagoal conflict tactics. There will be a.negative correlation between hostage-taker mean language intensity and the frequency of hostage-taker use of integrative relational-goal conflict tactics. 64 Again, because of the unique nature and characteristics of hostage negotiations, along with the paucity of language intensity research within conflict interactions, several research questions are posed to investigate the relationship of language intensity with communicative conflict behavior in hostage negotiation contexts. Specifically, the following questions are asked. R06. What is the relationship between negotiator langUage intensity and hostage-taker use of distributive relational-goal conflict tactics over time? R07. What is the relationship between hostage-taker language intensity and negotiator use of distributive relationalegoal conflict tactics over time? R08. What is the mean composite score for negotiator language intensity in response to (n + 1) hostage-taker use of negative identity-goal tactics (face attack-other)? No doubt, these research propositions and research questions represent only the beginning of possible issues that can be investigated concerning negotiator and hostage-taker communicative behavior. By investigating and seeking answers to these propositions and questions, this investigation can begin to lay the foundation for future research efforts. CHAPTER‘THREE RESEARCH METHODS TRANSCRIPTS Recordings of three hostage negotiations obtained from the FBI Special Operations and Research Unit (SOARU) at the the FBI Academy in Quantico, Virginia. This particular unit is the primary federal hostage negotiation research and training agency in the United States. These three negotiations constitute the data base for this investigation. The interactions were transcribed from original audio recordings provided by law enforcement officials at the scene of each respective hostage-taking incident. Due to the frenetic nature of initial interaction between negotiator and hostage-taker, the transcripts are incomplete in terms of the negotiator's introductory dialogue with the hostage-taker. Further, one of the dialogues ends without official description or explanation of the situational outcome. Nonetheless, the transcribed portions of these interactions provide a substantial record of naturalistic hostage negotiations. Each of the negotiations were conducted over the telephone by means of phone-jack recording devices. This technique results in the elimination of the majority of extraneous background noise. The final typed transcripts were reviewed for accuracy by trained research assistants who listened to the audio recordings in conjunction with reading the transcripts. The structure and format of the transcription followed Jefferson's guidelines for transcription as outlined in Schenkein (1978). An example transcript is attached in Appendix A. 65 66 Since this investigation is exploratory and descriptive, only three transcripts were selected for analysis. The rationale for this decision is two-fold. First, although the presence and potential influencing effect of other interactants (i.e., relatives of the hostage-taker, hostages, and media.persons) was previously identified as a significant facet of hostage negotiations, the specific focus of this dissertation has been narrowed to only those interactions between hostage-takers and police negotiators. This decision was made because the effect and relationship of the third party participants within hostage negotiation is unique and atypical of other third party negotiation involvement. Commonly, "other" nonhostage-taker and non-negotiator participants have a vested interest in the hostage-taker (i.e., spousal, interpersonal relationship). As such, analysis of this participation requires a reconceptualization of third party motivation and behavior specifically for hostage negotiations. Obviously, this represents a critical issue for hostage negotiation, and an area for significant future research. Second, the decision to use only three transcripts is also pragmatic. The three negotiations that were selected are actually part of a larger set of twelve interactions that were obtained from the FBI Special Operations and Research Unit. All twelve negotiations were evaluated for comprehensibility, conversational clarity, and type of negotiation (i.e., criminal, mental/emotional distress). Of the total set of twelve, the three that were selected were the most intelligible. Other transcripts simply were not as comprehensible nor limited to primarily only negotiator and hostage-taker interaction. Given the relative complexity of the three coding schemes, and the lack of 67 conversational clarity within natural hostage negotiations, it seemed prudent to analyze a set of interactions which offer intelligible dialogue. Further, the three selected all involve mentally/emotionally distraught hostage-takers (versus criminal), thereby providing a thread of commonality. Consequently, these three represent the best negotiations for this investigation. With only three transcripts, the question of generalizability becomes an issue. However, these three interactions cumulatively account for 1,814 utterances, a fairly sizable database even though the actual number of negotiation situations is only three. Individually, Transcript Number One contained of 269 utterances, all of which are between a single negotiator and hostage-taker. Transcript Number Two has 494 utterances, of which all but ten utterances are equally divided between one negotiator and one hostage-taker. The other ten utterances belong to a second negotiator, and are interspersed with those of the first negotiator and the hostage-taker. And Transcript Number Three is comprised of 1243 utterances, 85% (1052) of which are between a two negotiators and a single hostage-taker. The first negotiator accounts for 78% (410 of 527 utterances), while the second negotiator accounts for the remaining 22% (117 utterances). Negotiator number two begins interacting with the hostage-taker during the middle of Phase Seven and continues through Phase Eight. Analysis of each individual interaction, as well a comparison of results across the three negotiations can provide valuable insight into the communicative behavior of hostage-takers and negotiators. Research of this sort can offer insight into possible trends, similarities, and dissimilarities both within and across negotiations. Therefore, the 68 purpose of this investigation was to produce exploratory and descriptive information that could facilitate model building, model refinement, and ultimately increase the knowledge base of hostage-taker and negotiator communicative behavior within hostage negotiations. General Description of the Selected Negotiations The three negotiations chosen for investigation are each unique, yet similar. As is the case with all hostage negotiations, the geographic location, interactants, and issue(s) of contention are all different. However, these three cases are similar in that all three hostage-takers are mentally and emotionally distraught. This commonality suggests there are basic similarities in the communicative behavior of both the hostage-takers and police negotiators. To assist in understanding the nature of the three situations the following brief synopses are provided. W This first negotiation interaction involves one hostage-taker and one negotiator. The “hostage-taker" in this situation has barricaded himself inside a radio station. Although there are no other people being held hostage, the hostage—taker is holding both the radio station and himself hostage. The hostage-taker is clearly emotionally and mentally distraught, expressing a overwhelming desire to commit suicide. He argues that he has been the source of drug abuse among his friends, with one incident resulting in the apparent death of his girlfriend. The hostage-taker consistently berates himself for his failures and his responsibility in his girlfriend's death. As a result, he feels that he must take his own life, and that he is no 69 longer needed, nor a contributing member of society. The negotiator attempts to demonstrate to the hostage-taker the value of his life, his worth, and potential contributions to society. This tactic works for a short time as the conversation turns away from the hostage-taker committing suicide. Yet, in the end, the hostage-taker refuses to accept support or comfort, and ultimately kills himself. The total length of this interaction is 269 utterances. Transcrint_Number_Tuoi_Maiors This second interaction is also quite interesting and fortunately more successful in terms of a positive resolution than the first. The situation is primarily between one negotiator and one hostage—taker. However, toward the conclusion of the interaction, a second negotiator is introduced who assists the hostage-taker in surrendering. Yet, the second negotiator accounts for only ten lines of conversation. As such, the majority of the interaction is between the hostage-taker and one negotiator. The basic issue of negotiation in this case is the hostage-taker having barricaded himself inside his garage after shooting and killing his father. The hostage—taker is an ex-soldier who is suffering from terminal brain cancer. He argues that the CIA offered him a deal to kill his father, a supposed Soviet agent. The hostage-taker consistently argues that he has no intentions of harming anyone else, and that he is simply following orders, as any responsible soldier would. Picking up on this characteristic, the negotiator consistently confirms the hostage-taker's loyalty and responsibility for following orders. The negotiator builds a bond with the hostage-taker by highlighting their common military experience. The negotiator proposes 70 a meeting with CIA officials as a.means for the hostage-taker to surrender peacefully and to save face. This suggestion, in fact becomes the final solution and ultimate surrendering details are arranged. Total length of this interaction is 494 utterances. W The third and final transcript also involves a mentally and emotionally distraught hostage-taker. Unfortunately, the transcript begins during the middle of an initial interaction. The situation involves the hostage-taker holding his children hostage because of an argument he had with is wife, in which she accused him of "snitching" on her cousin for selling drugs. As result, he and his family were supposedly threatened. In an attempt to protect himself he purchased some weapons. He apparently believes that his wife is the source of the misunderstanding and wants to talk with her and conduct a news conference in which they clarify the situation. However, efforts to find the wife prove unsuccessful. During the course of the negotiation the hostage-taker threatens to commit suicide and is often incoherent because of crying and displays of emotional distress. A female friend of the family is brought in to intervene to win the release of the children. The negotiator, in cooperation with the female neighbor, work out a.plan with the hostage-taker in which they intend to meet him at his door and escort him to a meeting with his wife at the police station. A second negotiator takes over as the first leaves to meet face-to-face with the hostage-taker. However, the hostage-taker, while seemingly willing to comply, continues to stall his surrender. He raises new demands for meeting with his wife and the media. At this point the negotiation ends and the negotiator notifies 71 his superiors that negotiations are failing. In total, this negotiation consists of 1243 utterances, 1052 of which are between the hostage-taker and negotiators. Interaction with non~police negotiators are clearly marked by new telephone conversations, and account for approximately 15% of the interaction. These utterances were excluded from coding and are not part of the data set. Only the 1052 of negotiator ~ hostage~taker interaction were coded. Phase Descriptions of the Negotiations To assess negotiator and hostage-taker communicative behavior according to the Michigan State Police phase model of hostage negotiations discussed in Chapter Two, each transcript was divided into eight time units of approximately equal length by computing the total number of hostage-taker and negotiator utterances and dividing by eight. Although the Michigan State Police model of negotiation consists of only four active negotiation phases, it is advantageous to differentiate at least eight time periods for analytic purposes. According to Poole (1981) and Jones (1988) time phase analytic investigations should differentiate interaction periods into at least twice as many as are hypothesized by the phase model. This procedure is desirable for two reasons. First, it allows for greater variation in phase structure within interactions by allowing for variation in phase segments across interactions. And second, it negates the possibility of the results fitting the proposed model simply because the results are differentiated according the proposed model. In other words, if a four phase model is hypothesized, and interactions are divided into four phases, then the likelihood of 72 finding results fitting into a four phase model is increased simply because of how the data are divided, grouped and analyzed. Doubling the number of analytic time periods reduces this possibility. in To briefly review, the Michigan State Police Phase Model of negotiation, as discussed in Chapter Two, is divided into five stages (Donohue, Kaufman, Smith, 8 Ramesch, in press). The first stage is the "Pre-Work, Intelligence Gathering” phase. This stage precedes actual communication between negotiator and hostage-taker, accounting for the first hour or two of police arrival at the scene. During this phase, police units involved in the hostage-taking gather intelligence concerning the physical conditions of the situation, the hostage—taker, and hostages, along with as much psychological background information about the hostage-taker as possible. Actual negotiations begins with stage two, which is the "Introduction and Establish Contact" phase. This phase occurs at the outset of negotiation interaction. The primary task of the negotiator during this stage is to convey an image of concerned helper, elicit dispositional information about the hostage-taker and hostages, and attempt to stabilize the situation. This stage corresponds to phases one and two of the eight analysis phases. The third stage is the "Relationship Building" phase, accounting for approximately two to eight hours of negotiation. During this time period, the goal of the negotiator is to increase his/her level of understanding of the hostage-taker's thoughts and feelings by conveying support, confirmation, and understanding, as well as self-disclosing trivial personally private information in an effort to create a trusting relational bond with the hostage-taker. Providing face 73 support is considered a critical part in building a relationship during this phase. No attempts at problem-solving or problem resolution occur at this time. This stage corresponds with analysis phases three and four. Problem negotiation takes place during the fourth phase, "Problem Negotiations." During this time period, the task of the negotiator is to help direct the hostage-taker into devising his/her own nonviolent surrender by encouraging cooperative and integrative behavior. The focus of the interaction at this point in time is on needs and interests, and not on specific resolution techniques. The estimated time frame for this phase is the eighth through twelfth hour of interaction. This fourth stage corresponds to analysis phases five and six. The fifth and final stage is the "Resolution” phase, during which the negotiator works with the hostage-taker to create a specific resolution/surrender strategy. While this final stage is often very difficult to work through, success at this point is dependent upon the negotiator's success at addressing the goals of the first four stages. This final stage corresponds to analysis phases seven and eight. For the sake of brevity, the phase descriptions are not presented here. However, a complete synopsis of each negotiation by the eight phases is presented in Appendix B.{ CODING SCHEMES As discussed in Chapter Two, the focus of this dissertation is the communicative message behavior of hostage-takers and police negotiators as it occurs within naturalistic hostage negotiations. Interest in 74 these dyadic interactions is with the message content of both the negotiators and hostage-takers as it reflects their individual pursuit of identity and relational goals (Wilson 8 Putnam, 1990), as well as their respective level of language intensity. The schemes used for coding and analyzing the communicative interactions are three-fold, tapping into identityegoal based content, relational-goal based content, and language intensity. The categories within each coding scheme are intended to be exhaustive of possible coding options for hostage negotiations (Guetzkow, 1950). The overall unit of analysis for this study is the hostage-takers' and negotiators' uninterrupted speaking turn (utterance). Each unit received a single code for each of the three coding schemes. Where an utterance reflected more than one coding category within a coding scheme, a consensus code was determined. In instances where an utterance failed to fit into a coding category for any of the three schemes, the utterance was awarded a "missing data" code. The underlying logic for the coding schemes utilized in this investigation is based in part on the argument made by Donohue, Diez and Stable (1983) that one of the three primary contextual features which influences interlocutor communicative choices within a negotiation context is the goal framework each individual possesses toward both the interaction and his/her opponent. Research has demonstrated that distributive negotiation dispositions tend to encourage less individual flexibility in communicative interaction with one's opponent, as opposed to more integrative goal structures in which negotiators are less rigid in their demands and more cooperative in style (Donohue 8 Diez, 1985). According to Sillars, Coletti, Parry and 75 Rogers (1982) distributive interactions are characterized by individualistic and competitive behaviors, while integrative discourse involves verbal acts denoting behavioral cooperativeness and a mutualistic problem-solving orientation. Hostage negotiations can be thought of as being predominately distributive in goal structure orientation for both the hostage-taker and negotiator as they both desire goal achievement at the other's expense (a zero-sum profit orientation). Yet, hostage negotiations are marked by the negotiator's attempt to present an integrative orientation by expressing a genuine concern for the hostage—taker's needs and interests in an attempt to facilitate trust and to prompt the latter's surrender. In this way, hostage negotiations represent mixed—motive interactions. As discussed in Chapter Two, Wilson and Putnam (1990) have identified three goal types as being present in most mixed—motive bargaining interactions. These are instrumental, identity, and relational goals. To briefly review, instrumental goals are individually desired objectives which require the opponent's cooperation to realize. Identity goals involve individual attempts at creating and maintaining a desired "face" with the opposing party, And finally, relational goals denote the interactive features of the bargaining situation (e.g., distribution of power, control, cooperativeness, and competitiveness) which develop and evolve within the relationship between the interactants. The coding schemes used in this investigation are comprised of both identityegoals and relational-goals. The rationale for this dichotomy is that a relational-goal interpretation can provide data for the analysis of negotiator and hostage-taker compliance, 76 cooperativeness, and competitiveness; while an identity-goal analysis can tap into "face-work“ behaviors. IdentityeGoal Coding Scheme The primary feature assessed within the identityegoal coding scheme is the type of "face-work" expressed by each of the interactants. Based on the seminal work by Goffman (1959), and as discussed by various researchers (Brown, 1977; Donohue 8 Diez, 1985; Wilson 8 Putnam, 1990) "face—work" is considered to be a critical feature of negotiation interactions. Goffman (1959) identified four types of face-work: face-loss, face-saving, face-attack, and face—giving, Face-loss denotes a reduction or modification in an individual's interactive self-image. Face-saving involves an individual's attempt to maintain or regain lost self-image. Face-attack involves communicative derogation of another's self-worth. And, face-giving denotes an individual's efforts to help restore another's interactive self-image. By redefining face-saving, Brown (1977) and others (Folger 8 Poole, 1984) delineate a fifth type of face work which they label as face-restoration. Briefly, face-saving denotes defensive measures taken by an individual to protect his/her image, while face-restoration involves attempts to regain a lost positive image. In Chapter Two it was argued that Brown's typology of face behavior activities was incomplete. A reconceptualized typology of six face behaviors, representing a 2.x 3 matrix involving a self—other dimension, and three activities (i.e., attack, defend, and restore) was presented in Table 1. For the purpose of this investigation, active 77 face-work identity goals were coded into one of six categories: 1) face attack-other, 2) face attack-self, 3) face defend-self (i.e., proactive face-saving), 4) face defend-other (i.e., proactive face-saving), 5) face restore-self (i.e., retroactive face-restoration), and 6) face restore-other (i.e., retroactive face-giving). A seventh and eighth coding category were used to code "conversational features" (i.e., positive and negative backchannels) respectively. These lexical features do not function to attack or support face, but simply to maintain interaction. The identity—goal coding scheme, along with examples of each communicative "face" act and "conversational feature" are presented in Appendix C. Three research assistants were trained in identification and coding of the eight identityegoal categories. Following several individual practice coding sessions, the coders were administered a sample pretest transcript and asked to code the 150 utterances independently. Overall, there was an 81% agreement among the three coders. This corresponds to a Cohen's kappa value of .76 for the three coders (Cohen, 1960). In comparison, kappa as computed by the Folger, Hewes and Poole method was .79 (Folger, Hewes, 8 Poole, 1984). This latter kappa is determined according to the number of categories within the coding scheme, assuming an equal probability of chance for each category. Cohen's original kappa is computed according the actual distribution of codes among the categories; with the probability of chance based on the actual coding distribution. Cohen's is considered the more stringent method of the two. According to Fleiss, (1981) a Cohen's kappa of .40 to .60 is fair, .60 to .75 is good, and over .75 78 is excellent. Based on these pretest results, it was determined that coding of the transcripts could proceed. All three trained assistants coded the three transcripts independently. However, only the codes of the two coders with the highest percentage of agreement and kappa were selected for analysis. The coding results for Transcript Number 1 were: 75% agreement and Cohen's kappa = .62. For Transcript Number 2 there was a 69% agreement and Cohen's kappa of .57. And for Transcript Number 3 there was a 76% agreement and Cohen's kappa of .68. Once the independent percentage of agreement and kappa were calculated for each transcript, the two coders with the highest score jointly recoded the transcripts for those utterances where they disagreed. This produced a single agreement code for each utterance. Relational-Goal Coding Scheme A tri-level, eighteen item message content—based coding scheme was used to assess relational-goal behavior. This scheme taps into the nature of the interactants' cooperativeness, competitiveness, and behavioral compliance as conveyed within the functional content of their verbal messages. This typology is based in part on the coding scheme of Sillars, Coletti, Parry, 8 Rogers (1982) which assess the level of distributive, integrative, and avoidance behavior of interactants as conveyed within the conflict tactics utilized by each communicator within bargaining contexts. According to Sillars, et. a1, (1982) distributive illocutionary acts denote competitive behavior, while integrative acts denote cooperative behavior. "avoidance" of issues is a third behavioral feature built into the Sillars et al. 79 coding scheme, accounting for interactants' unwillingness to deal with issues of contention, but which is omitted from this coding scheme. Drawing on the three-part category structure outlined by Sillars et a1. (1982) a tri-level, eighteen-item category-set was constructed to assess interactants' relational-goal message content. On the supracategorization level this framework ranges from "integrative acts" (Level I) to "distributive acts" (Level II). The third level is devoted to "structural and conversational features," of interaction including "positive and negative communicative backchannels" (e.g., "uhm," ”uhuh") and "interaction initiations and salutations" (e.g., "hello"). Each of the two relational-goal supracategorization levels (i.e., integrative, 8 distributive) is further differentiated into separate, more specific functional content coding categories. Briefly, Level I — ”integrative acts,“ consists of nine categories, including: 1) "willingness to comply," 2) "questions of emotional disclosure," 3)"questions of fact about individual past experiences," 4) "questions of fact concerning the hostage-taking," 5) "questions of demand," 6) "statements of disclosure," 7) “statements of fact about individual experiences," 8) "statements of fact concerning the hostage—taking," and 9) "statements of support/comfort." Level II - "distributive acts" involved six content coding categories: 1) "expressing personal demands and wants," 2) "topic shifting and topic avoidance, " 3) "avoidance of personal ‘ responsibility," 4) "rejection of other's demands,” 5) "threats against others," and 6) "threats against self“. Level III - "structural and conversational features,‘ consists of three categories: 1) "positive 80 backchannels," 2) "negative backchannels," and 3) "salutations." As a result of pretesting and validation checks, ”topic shifting and avoidance" and "avoidance of personal responsibility" were both dropped from the coding scheme. The former was determined not to be mutually exclusive from the other categories. The latter also failed to be mutually exclusive, and also failed to be identified as a relational behavior by the coders. A copy of the relational-goal coding scheme, along with sample statements for each of the eighteen coding categories are presented in Appendix D. C l' E I I B 1| E B J I' J G J C l' S I Just as with the identity-goal coding scheme, three different research assistants were trained in the identification and coding of the relational-goal code categories. After several practice coding sessions, the assistants were administered a 150 utterance sample pretest transcript. Among the three coders there was an 85% agreement, translating into a Cohen's kappa of .84. The Folger Hewes, and Poole kappa was .85. Again, Cohen's kappa is considered the more stringent of the two. As previously mentioned, "topic shift and avoidance” and "avoidance of personal responsibility” failed to be mutually exclusive from the other categories, as well receive any frequency of codes by the coders. These two categories were subsequently dropped from the coding scheme. Based on these pretest results, coding of the three transcripts was initiated. Similar to the identity-goal scheme, all three assistants coded all three transcripts. Still, only the results of the two coders with the highest percentage of agreement and Cohen's kappa were selected for analysis. For Transcript Number 1 there was a 78% agreement and 81 Cohen's kappa of .71; for Transcript Number 2 there was 57% agreement and a Cohen's kappa of .50; and for Transcript Number 3 agreement was 67% and Cohen's kappa was .61. While the percentage of agreement and kappa for Transcripts Two and Three are low, the kappa values are considered to be fair to good by Fleiss (1981). Once the independent levels of agreement were computed, the two coders with the greatest agreement jointly recoded each transcript for those utterances where they disagreed. This procedure resulted in a single relational-goal code for each utterance. Language Intensity Coding Scheme As discussed in Chapter Two, recent work by Donohue (in press) involved the evaluation of message intensity from naturalistic divorce mediation sessions using a coding scheme derived from Bowers' (1964) typology of intensity correlates. Using 482 terms rated as either low or high in intensity, Bowers' sought to identify possible correlates of message intensity. His investigation revealed five fairly reliable correlates, including: obscure language, metaphors, qualifying adjectives, sex-based metaphors, and death-based metaphors. A sixth correlate - number of syllables - was also postulated to be related to intensity, yet failed to produce a significant correlation upon analysis (.10). Correlations of intensity for the remaining five predictors with the 482 previously rated terms were .59, .83, .89, 1.0, and 1.0 respectively. Donohue employed a slightly modified version of these correlates, including: obscure language, general metaphors, qualifiers, profanity, sex-based metaphors, and death-based metaphors. 82 The addition of “profanity" as a predictor was the only modification Donohue made to Bowers' scheme. In producing a composite measure of language intensity for each utterance from naturalistic interactions, Donohue (in press) weighted the frequency of each predictor within each utterance by Bowers' (1964) correlations of intensity, summed across the six predictor categories, and then divided by the number of words within the utterance. Predictor weightings were as follows: obscure language (.59), general metaphors (.83), qualifiers (.89), and profanity, sex-based metaphors, and death-based metaphors (1.0). This procedure generated an overall intensity measure for each utterance, which was determined by the presence of one or more of these six correlates. Based on this method, intensity values ranged from 0.0 to 1.0. The basic language intensity coding scheme used in this study involved a slight modification to Donohue's intensity measure. Specifically, profanity and sex-based statements were combined into one single category (i.e., sex and profanity statements) rather than coded as two separate categories. This modification was made because of the common occurrence of sex-based messages as profanity in naturalistic discourse, and the sameness of the correlation value for both profanity and sexrbased messages. This change resulted in a five-level typology of language intensity, including: Obscure words, general metaphors, qualifiers, profane and sexual messages, and death-based messages. A copy of this general coding procedure is presented in Appendix E. 83 .um - ... ;.. . - .' '.. .' .; a ; "'- a. .- ; -+- Scrutinization of Bowers' definition of the qualifiers category, as applied in Donohue's coding scheme, revealed a potential weakness when attempting to apply the category to naturalistic discourse. More to the point, Bowers' definition of qualifiers are those words which intensify another word. According to Bowers, "For a term to be classified as qualified, it was required that the qualification be embodied in a separate word. "Greater height,” for example is a qualified term; "higher" is not." (Bowers, 1964, p.352). Bowers' had defined language intensity as "the quality of language which indicates the degree to which a speaker's attitude toward a concept deviates from neutrality." (Bowers, 1963, p. 345). Comparing the qualifiers definition with his overall definition of language intensity reveals an obvious inconsistency between what Bower's defines as intense language and the types of qualifiers that count as intense language, particularly when attempting to code natural discourse. For example, using Bowers' definition of qualifiers to code the two statements: "That is a very nice sweater.‘ and "No, I don't want you to do that." would reveal a lack of sensitivity to intensity in Bowers' definition of what constitutes a qualifier. For the first statement, "very" would receive a code. For the second statement, there would be no qualifiers coded. Yet, the second statement clearly expresses an attitude which deviates from a position of neutrality. Consequently, reliance on Bowers' definition of qualifiers would result in an incomplete measurement of potentially intense language. To deal with this potential methodological weakness, an elaborated definition of the qualifiers category was devised. According to this 84 definition, qualifying adjectives, adverbs, affirmations, negations, auxiliary verbs, and contractions were coded as qualifiers. As such, the statement "No, I don't want you to do that. ” would now have more qualifiers coded and consequently a higher intensity level than would the statement "That's a nice sweater." This Elaborated Qualifiers version of language intensity addresses the lack of sensitivity of the Simple Qualifiers version that was based on Bowers' definition. To evaluate the effectiveness of each scheme when applied to coding of actual discourse, both the Simple Qualifiers and Elaborate Qualifiers versions were used in coding. Instructions for both the Simple and Elaborate Qualifiers coding categories are presented in Appendix E1. Three research assistants were trained to code the four intensity categories of obscure words, general metaphors, profanity/sex, and death statements. These same three coders were also trained to code the qualifiers category according to Bowers' Simple Qualifiers definition. A fourth research assistant was trained to code qualifiers according to the Elaborate version, along with the researcher. As such, there were three coders for the Simple Qualifiers version and four other intensity categories, and two coders for the Elaborate Qualifiers version of language intensity. For data analysis purposes, the Elaborate Qualifiers values were substituted for the Simple Qualifiers, while maintaining the codes for the other four categories. Composite language intensity scores for each unit of analysis (i.e., utterance) were generated using Donohue's computational methodology outlined earlier. Specifically, this procedure involved the coding of words according to the five intensity categories within 85 each utterance for intensity. Each word was weighted by the intensity correlates (i.e., weighting values), summed, and divided by the number of words within each utterance. Where a word was judged as fitting into more than one of the coding categories, coders were instructed to code the unit into the highest intensity category (Donohue, in press). In this way, composite intensity values for each utterance range from 0.0 (low) to 1.0 (high). After several practice training sessions, the coders conducted an independent pretest coding of a sample transcript. Results for all correlate categories, except qualifiers, for all three assistants were: obscure words, 98% agreement, alpha = .81; general metaphors, 95% agreement, alpha = .85; profanity/sex words, 99% agreement, alpha = .95; death words, 99% agreement, alpha 2 .84. Results for the three coders for the Simple Qualifiers category was 95% agreement and alpha 2 .81. Over all five categories, there was a 97% agreement and alpha = .85 for all three coders. For the two coders using the Elaborate Qualifiers version, agreement was 92% with an alpha of.99. With the Elaborate scores substituted for the simple scores there was a 98% agreement and alpha = .89. The results for the two coders with the highest percentage of agreement and alpha are used to evaluate inter-rater reliability for the three transcripts. For Transcript Number One, there was an overall percentage of agreement for obscures, metaphors, profanity/sex, and death of 95% with a mean alpha of .83. For the Simple Qualifiers agreement was 91% and alpha = .88. Across all five coding categories including the Simple Qualifiers, agreement was 94% and alpha = .84. For the Elaborate Qualifiers version agreement was 76% and alpha 2 .99. 86 Substituting the Elaborate for the Simple Qualifiers produces an overall level of agreement of 91% and mean alpha = .86. For Transcript Number Two, there was a.mean 98% agreement between two coders and a mean alpha of .83 for all but the qualifiers category. For the Simple Qualifiers agreement was 97% and alpha = .91. With the Simple Qualifiers included, the mean agreement for all five categories was 98% and alpha = .64. For the Elaborate Qualifiers, agreement was 87% and alpha = .99. Substituting the Elaborate for the Simple qualifiers produced a mean agreement of 96% and alpha = .86. For Transcript Number Three there was an overall agreement between two coders on obscures, metaphors, profanity/sex, and death of 99% with alpha = .91. Simple Qualifiers were characterized by a 97% agreement and alpha = .76. Including Simple Qualifiers, the mean agreement was 98% and alpha = .91. There was an 88% agreement on Elaborate Qualifiers with alpha = .99. Combining Elaborate qualifiers with the other four categories overall agreement was 97% and alpha = .94. These results indicate a high level of agreement between coders across the four categories of obscure words, metaphors, profanity/sex, and death; and for a Simple language intensity coding version based on the Simple Qualifiers coding procedure. Likewise, coder agreement is high when the Elaborate Qualifiers coding procedure is used instead of the Simple Qualifiers, representing an Elaborate language intensity coding scheme. The two coders with the highest level of agreement and alpha for the obscure, general metaphor, profanity/sex, death, Simple Qualifier, and Elaborate Qualifier jointly recoded each transcript where they disagreed to create a single code for each utterance. These recodes constitute the actual data set for this analysis. 87 CODING SCHEME VALIDATION Each of the two goal-based coding schemes were evaluated to test the accuracy of the nominal coding categories for individual message content. Similarly, the language intensity coding scheme was evaluated to validate the accuracy of the composite intensity scores generated for each utterance. Validation measures were made using both the Simple qualifiers and Elaborate Qualifiers coding technique. In all, a total of 561 naive undergraduate students enrolled in undergraduate communication courses at a large midwestern university served as the research participants. However, the actual number of participants involved in each validation was less than this and of unequal size. Participants received extra course credit for their voluntary involvement. Identity-Goal Coding Scheme Validation To assess the validity of the identity-goal coding scheme, an aggregate of 243 undergraduate students enrolled in undergraduate communication courses were randomly assigned one of six evaluation measures. The actual number of subjects for each measure ranged from 38 to 45. Each of the six measures represented one of the six active face behavior coding categories of the overall coding scheme (i.e., self-hostile = face attack-self, other-hostile face attack-other, self-support face defend-self, other-support face defend-other, self-defend = face restore-self, other-defend 2 face restore-other). Each instrument consisted of 30 statements randomly selected from all three transcripts. The statements which were used had been coded into 88 one of the six active face categories by the trained coders. There were five sample statements for each of the six categories. Participants were asked to rate each of the 30 statements independently according to one of the six five-point Likert scales. The participants rated the statements in terms of how self-hostile, other-hostile, self—supportive, other-supportive, self—defend, or other-defend they perceived the statements to be. Each group of participants were read an introductory statement informing them of the nature of the study. In an effort to provide them with a basic contextual framework for interpreting the statements, the introduction also stated that the statements had been taken from hostage negotiations. The overall validation effort constituted a 6 way repeated measures design. For each the six instruments, the design was a one-way repeated measures. An example of the identity-goal validation instruments is presented in Appendix F. Mean scores and standard deviations were computed for each of the six groups of statements within each measurement instrument. Table 2 presents these values for all six categories for all six instruments. As can be seen from Table 2 there was a predominate linear pattern for the means scores across the six instruments. In general, the means for each category reflect a fairly strong match between the participants' perception of the statements and how they were coded. In fact, most of the means are significantly different at the .05 level (i.e., p,< .05) from the mean for the identity-goal category being evaluated within each individual instrument. However, there were some nonsignificant differences. Specifically, the mean ratings for the 89 self-hostile and other-hOstile statements in the self-hostile instrument; the self-support, other-support, self-defend, and other defend in the self-support instrument; the other-hostile and self-defend statements in the self-defend instrument; and the other-support and other-defend in the other-defend instrument were not significantly different at the .05 level. Table 2 about here As was expected, a six-way repeated measures analysis of variance for all six of the validation instruments indicated a significant main effect for instrument by statement category (F (25,237) = 35.30 p < .001, n3 = .62). To explore this effect in greater detail, individual one-way repeated measures analysis of variance were computed for each of the six instruments. For the self-hostile instrument, there was a significant main effect for the statement categories, accounting for 44% of the variance (E (5,185) = 43.96, p < .001, mg .44,). Test of a linear model accounted for 41% of the variance (2 .65, n3 = .41). There was a significant main effect for the statement categories within the other-hostile instrument, accounting for 50% of the variance (F (5,220) = 57.84, p < .001,n£ = .50,). The effect was linear, with the model accounting for 44% of the variance (r = .66, n3 = .44,). There was a small main effect for the identity-goal categories in the self-support instrument (E.(5,185) = 10.45, p,< .001, n3 = .20,). A linear model accounted for 16% of the variance (1 = .40, n3 = .16). The main effect was much larger in the other-support instrument 90 (F (5,215) = 90.26, p < .001, n2 = .62,), with a linear model accounting for all of the variance (2 = .79, n3 = .62). For the self-defend instrument, 34% of the variance was accounted for by the statement categories (R (5,195) = 27.35, p < .001, n3 = .34,). Like the others, a linear model provided the best fit (r = .54, n3 2.30). Though the effect was the smallest in the other-defend instrument (F (5,185) = 3.45, p_< .05, n3 = .06,), a linear model provided the best fit (I: = .24, n.2, = .08). Conceptual arguments can be made to explain the lack of significant difference between the sets of means and the low effect sizes for the self-hostile, self-support, self-defend, other-defend instruments. Though the mean for other-hostile statements within the self-hostile instrument is less than that of the self—hostile set of statements they are not significantly different (p < .75). While participants were instructed to rate the statements in terms of how self-hostile they perceived them to be, participants may have been unable to adequately differentiate between self-directed and other-directed statements. In fact, there appears to be a significant "hostility" statement effect. Further, the statements used in the instruments are randomly sampled from the transcripts and as such, may not be the strongest representations of their respective categories. And perhaps most critically, the statements are naturalistic statements taken out of context. Being out of context may have made it more difficult for the participants to adequately evaluate them. The lack of significant difference between the self-support, <3ther-support, self-defend, and other-defend statements seems to indicate that participants were unable to adequately differentiate 91 between self and other directed statements, as well as support and defend. Yet, there is a clear distinction in ratings between the support, and defend statements and hostile statements. Again, perceptions may have been subsumed by a "hostility" effect. Lack of representativeness among the statements may likewise be a problem as well as the lack of a sufficient contextual framework. Regarding the means for other-hostile and self-defend statements within the self-defend instrument, and other-support and other-defend in the other-defend instrument, their respective values are not significantly different (p < .81, and p < .06). Conceptually, statements which may be used to defend one's face may also be perceived as hostile towards another. Likewise, statements used to support may also provide retroactive defense. Consequently, participants may not have been able to clearly distinguish between self-defend and other hostile as well other-support and other-defend in the other-defend instrument. As with the self-hostile and self-support instruments, the statements are out of context. And finally, the statements used may not have been the best examples for the respective categories. To test for a possible hostile - support - defend effect a three-way repeated measures analysis of variance was computed across the three general categories of statements (i.e., hostile, support, defend). Results indicate a significant instrument by statement category effect (2 (4,240) = 97.81, p < .001, n3 = .47,). Table 3 presents the means for each of the three statement categories within each of the three instrument categories. Individual one-way repeated measures ANOVAS were computed for each of the six instruments across the three general statement categories. 92 As can be seen in Table 3 there is a significant difference between cell means for hostile, support, and defend for the self-hostile instrument. There is likewise a significant main effect for the statement categories within the self-hostile instrument (E (2, 74) = 114.36, p < .001, n3 = .53,) which is linear (r = .71, n3 = .50). There is a less significant effect for hostile statements within the other-hostile instrument (E (2,88) = 31.06, p < .001, mi = 28,) which is also best fit with a linear model (1 = .49, n3 = .24). This is reflected in the nonsignificant difference between the defend and hostile statement categories within the other-hostile instrument. Table 3 about here Within the self-support instrument, there is a nonsignificant difference between the support and defend statement category means. Still, there is a slight linear main effect for the statement categories (F (2,74) = 25.01, p < .001, n3 = .32; r = .52, n3 = .27). For the other-support instrument, all the statement category means are significantly different at the .05 level. As such, there is significant main effect for the statement categories (F (2, 86) = 153.73, p,< .001, mi = .69,). It is also linear (I = .80, n3 = .63). As reflected in Table 3 there is no significant difference between the hostile and defend statement categories within the self-defend instrument. Consequently, there is a very small main effect (F (2,78) ‘ = 18.99, p < .001, n3 = .19). Yet it is linear (r = .41, n2 = .17). Finally, there is no significant difference between the means for hostile and support statements, nor between support and defend; though 93 there is between hostile and defend. As such, the statement category main effect is very slight (F (2,74) = 4.55, p,< .05, n3 = .06). In sum, the results tend to reflect a.matching of perceptions by participants with that of the coded statements. Results in Table 2, along with Table 3, and the 3 x.3 ANOVA of general statement categories by general instrument categories, indicate fairly strong effects for each of the three general categories (hostile, support, and defend). Although the effect is less distinct between hostile and defend in the defend instruments. Still, more detailed analyses reflect a rather erratic effect. As posited earlier, several conceptual explanations can be offered to account for the lack of conclusiveness. Perhaps the two most important are: 1) the statements used may not be strong representations of the respective categories, and 2) the statements are taken out of context. Given these potential limitations in methodology, tentative support has been gained for this particular coding scheme. Relational-Goal Coding Scheme Validation The validation check of the relational-goal coding scheme was conducted in much the same way as was the identity-goal coding scheme validation. In total, 198 undergraduate students enrolled in undergraduate communication courses were randomly assigned one of four validation measures. The actual number of participants for each instrument ranged from 44 for the competitive instrument, to 50 for the cooperative measure, to 52 for both the individualistic and mutualistic measures. The four validation measures each represented behaviors defined by Sillars and colleagues (1982) as indicative of either 94 integrative or distributive conflict tactics. More specifically, verbally competitive and individualistic behaviors are characteristic of distributive tactics, while verbally cooperative and mutualistic are integrative. Each of the four instruments was comprised of fifty-two statements that were randomly sampled from all three of the transcripts, based on the nominal coding value assigned to them by the trained coders as being representative of each of thirteen of the fifteen active relational-goal categories. As indicated previously in this chapter, coding category 11 - "topic shift", was dropped from the coding scheme because it was not mutually exclusive from the other categories. In a similar fashion, coding category 12 - "avoidance of personal responsibility" - was excluded from the validation check due to the infrequency of its occurrence within the three coded transcripts, and it's lack of mutual exclusivity. This resulted in only 13 of the 15 actual active relational-goal categories being validated. Four randomly sampled statements were provided for each of the final 13 categories. Sample statements for each coding category were randomly distributed among the fifty-two to minimize any possible ordering effect. Participants were instructed to evaluate each of the fifty—two statements independently according to a 5-point Likert scale, which corresponded to each of the individual instruments. In other words, participants evaluated the statements in terms of either how cooperative, competitive, individualistic, or mutualistic they perceived them to be. In this way, each instrument constituted an independent one-way repeated measure. To provide participants with a 95 contextual frame of reference for evaluating the statements, they were read an introductory statement describing the sample statements as originating from hostage negotiations. An example of the relational-goal validation instruments is presented in Appendix G. Results_of_the_Ealational:Goal_Coding_Scheme_Yalidation_Check Means and standard deviations of ratings for each of the thirteen coding categories for each of the four validation instruments are presented in Table 4. It is evident from Table 4 that there is a significant difference in participant ratings of the categories defined as integrative as compared to those defined as distributive. Other than for a couple of exceptions, the mean ratings for all of the statement categories within the integrative dimension are significantly different from those within the distributive at the .05 level for all four measures. The two exceptions are the "self-disclosure" and "threat against self” categories within the cooperative, competitive, and individualistic instruments; and the "statement of fact about person" and "threat against self" within the mutualistic measure. These two sets of means are not statistically significantly different. Table 4 about here A four-way repeated measures analysis of variance across the four instruments by the 13 categories produced a significant interaction effect for statement coding category by validation instrument (R (36,194) = 83.98, p < .001, n3 = .74). This was the expected effect. Individual one-way repeated measures analysis of variance were computed for each of the four instruments across all 13 coding 96 categories. For the cooperative instrument there was a significant main effect for the coding categories which accounted for 65% of the variance (E.(12,588) = 123.20, p,< .001, n3 = .65). A repeated measures analysis of variance produced similar results for the competitive measure, indicating that coding categories account for 50% of the total variance (F (12,516) = 55.30, p < .001, n3 = .50). The percent of variance accounted for by the coding categories within the individualistic measure was only 32% (F (12,612) = 29.99, p < .001, n3 = .32). And within the mutualistic measure, the coding categories account for 63% of the total variance (F (12,612) = 112.31, p < .001, n3 = .63). These findings indicate a fairly strong validation between the statement coding categories and participants perceptions of the statements. Additional analyses were conducted in which the 13 statement coding categories were grouped into their apriori integrative or distributive dimensions. As can be seen from Table 5, there is a clear distinction in mean scores between the two dimensions for all four validation instruments. In fact, all mean scores are significantly different at .001 level. A four-way repeated measures analysis of variance for these aggregate categories revealed a significant interaction effect for the coding categories by the validation instruments (F (3,194) = 65.20, p < .001, n3 2 .59). Further, individual repeated measures ANOVAS for each instrument produced positive results in favor of the coding scheme. Within the cooperative instrument, 66% of the variance was accounted for by the coding categories (F (1,49) = 260.24, p,< .001, mg = .66). Sixty-eight percent was accounted within the competitive instrument 97 (E (1,43) = 187.68, p < .001, n3 = .68). Within the individualistic instrument, only 45% of the variance was attributable to the coding categories (E (1,51) = 85.27, p < .001, n3 = .45). And within the mutualistic measure, 64% of the variance was accounted for by the coding categories (F (1,51) = 181.65, p < .001, n2 = .64). Again, fairly strong support has been generated for the validity of the relational-goal coding scheme. Table 5 about here Conceptual arguments are available to explain the lack of significant difference between the "self-disclosure" and "threat against self" categories for the cooperative, competitive, and individualistic measures; and the "statement of fact about person" and the "threat against self" for the mutualistic instrument. To begin with, statements of disclosure were defined as emotionally laden and emotionally-based statements. Such is also the case for statements which convey a self threat. For example, to say "I'm going to kill myself" clearly involves self-disclosure as well as a selfedirected threat. Consequently, finer discrimination of such statements as cooperative, competitive, and individualistic would be difficult to realize. The same argument can be made for "statements of fact about person“ which were defined as self-directed, and "threats against self." Again, both are self-directed and individually focused. Finally, it may have proven difficult for participants to clearly differentiate statements as cooperative, competitive, individualistic or mutualistic. 98 Nonetheless, the results clearly offer validation support for the relational-goal coding scheme. Participants views are fairly consistent with the prediction that the integrative statement categories are seen as cooperative and mutualistic, while the distributive statement categories are seen as competitive and individualistic. Language Intensity Coding Scheme Validation The purpose for validating the language intensity coding scheme was twofold: to assess the accuracy of the coding scheme, and to determine whether the Simple Qualifier of Elaborate Qualifier coding procedure was more precise in detecting intensity. Towards this end, 120 undergraduate students enrolled in undergraduate communication courses were randomly asked to evaluate a set of 50 statements according to a single intensity validation measure. The 50 statements themselves were randomly sampled from the three transcripts. Each statement had been coded by the language intensity coders and possessed a language intensity score ranging from 0.00 to 1.00. To effectuate analysis, statements were divided into five levels according to their individual intensity score. Level 1: "No Intensity" included statements with scores of only .00. Level 2: "Slight Intensity" denoted statements with scores ranging from .01 to .25. Level 3: "Moderate Intensity" included statements with scores of .26 to .50. Level 4: "High Intensity" denoted statements with scores of .51 to .75. And level 5: "Extremely High Intensity" statements had scores of .76 to 1.0. In many cases, statements had an intensity score for the two coding procedures which was noticeably different. For example, 99 the statement "No, you don't want to do that" had a score of .00 for the Simple Intensity procedure and a score of .36 for the Elaborate Intensity coding procedure. Statements were randomly sampled to fit into one of the five levels. However, statements could fit into more than one level, depending upon which coding scheme score was used. Attempts were made to sample 10 statements for each level for each of the two coding procedures. However, because the Simple Qualifiers scheme was not as sensitive as the Elaborate Qualifier, scores were distinctly lower for the Simple Intensity procedure (grand mean = .02, with 85% coded as .00), while in contrast they were noticeably higher for the Elaborate Intensity scheme (grand mean = .23, with 34% coded as .00). As reflected in these values, codes for the Simple coding scheme were positively skewed while codes for the Elaborate scheme approached more of a normal distribution. As a result, the sample statements for each of the two schemes were not equally distributed across all five intensity levels. Within the Simple scheme, "No Intensity“ had 25 statements, "Slight Intensity" had 11 statements, "Moderate Intensity” had 10 statements, "High Intensity had 1 statement, and "Extremely High Intensity" had 4 statements. For the Elaborate scheme, the distribution was a bit more equal, with 6 statements in "No Intensity,” 7 statements in "Slight Intensity," 15 statements in "Moderate Intensity," 11 statements in "High Intensity," and 11 statements in "Extremely High Intensity." Participants were instructed to rate each of the 50 statements in terms of how intense they perceived them to be. Intensity was defined as "how much the statement expresses a position (opinion) which deviates from neutral." Participants rated the statements along a 100 five-point Likert scale which ranged from ”No Intensity," "Slight Intensity," "Moderate Intensity," "High Intensity" to "Extremely High Intensity." These five evaluation levels corresponded to the five intensity score levels. In this way, the design was a one—way repeated measure. Before evaluating the statements, participants were read an introductory statement explaining the purpose and nature of the study. The statement also informed them that the statements were extracted from hostage negotiations as a means of providing them with a general contextual frame of reference for assessing the statements. A copy of the language intensity validation measure is included in Appendix H. {-a . .- ;.u :s‘ . -. ‘ s.‘.- .‘u’ ; '.; 'm. .- A simple correlation analysis between the mean score rating assigned to each of the 50 individual statements by the evaluation participants and the composite intensity scores derived from each of the two language intensity coding procedures characterized the Elaborate Intensity coding scheme as being more strongly correlated with raters perceptions (r = .38), than the Simple Intensity coding Procedure (1 = .31). After correcting for attenuation, the Elaborate ceding correlation increased to .43 (r = .43) while the Simple coding correlation increased to .36 (r =.36). Both were significant at the -001 level. While these scores are relatively low, both reflect a InOder-ate relationship between the coding schemes and raters' Perceptions. However, the Elaborate scheme is the stronger of the two. To investigate the effect pattern of both schemes further, means were computed for each of the intensity level categories. Because of the unequal distribution of sample statements across the five intensity levels, "High Intensity" and "Extremely High Intensity" 101 were collapsed into a single intensity level - " High Intensity" — for both the Simple and Elaborate coding schemes. Mean scores were computed for each of the four remaining intensity levels, based on the validation ratings. Table 6 presents the means and standard deviations for each intensity level for both the Simple and Elaborate coding schemes. Table 6 about here The values for the Simple scheme indicate a significant difference between the "No Intensity" and "Slight Intensity" categories, (9 < .001), as well as between the "Slight Intensity" and "Moderate Intensity” (p < .001). Yet, the mean for the "Moderate " level is less than the "Slight" category. There is no significant difference between the "Slight Intensity" and "High Intensity" levels (9 < .554), while "Moderate " and "High" are significantly different (p < .005). In contrast, mean scores for the Elaborate coding scheme reflect a more linear regression, with "No Intensity" less than "Slight Intensity" and Significantly different (p < .001); "Slight Intensity" less than "Moderate Intensity" and significantly different (p < .001); while "Naderate" and "High" are not significantly different at the .01 level (a < .036). Further investigation of effect strength for each scheme was aSsessed by means of one-way repeated measures analysis of variance. Results for the Simple Qualifier coding scheme revealed a simificant, Vet: Small effect, with the intensity categories accounting for 15% of the total variance (E (3.35?) = 64.22, p < .001, 113 = .15). 102 Decomposing for sums of squares with the values of -3, 1, l, 1 indicated a linear model accounting for 14% of the variance, with a moderate effect size (r = .37, n3 = .14). In comparison, results for the Elaborate Qualifier scheme revealed a fairly strong and significant effect, with the intensity categories accounting for 46% of the total variance (E (3, 357) = 374.73, p < .001, n3 2 .46,). Decomposing for sums of squares indicated a strong linear model accounting for 44% of the variance, with a strong effect (2 = .66, n3 = .44,). These results are quite impressive considering the fact that participants essentially evaluated the statements out of context and without the aid of paralinguistic cues. The validation results for both language intensity coding procedures offer the strongest support for use of the Elaborate Qualifier coding procedure rather than the Simple Qualifier coding method. CONCLUSIONS FROM THE VALIDATION CHECKS Support has been gained to validate each of the three independent coding schemes for this investigation. Support is the weakest for the identity-goal coding scheme. However, conceptual arguments have been advanced which offer explanation for the lack of clear support. Further, the relatively strong and consistent kappas between coders for the three transcripts indicate a fairly reliable measure. Regarding the relational-goal coding scheme, the findings offer fairly strong support in validating the integrative/distributive distinction between the numerous categories. The findings are especially favorable considering the number of coding categories within the scheme. Again, conceptual arguments made previously offer support 103 for the findings, as do the fairly strong and consistent kappa scores across the three transcripts. Finally, validation results for the language intensity coding schemes reflect strong support in favor of the Elaborate Qualifier coding procedure versus the Simple Qualifier coding procedure. Given the findings generated, the preferred method of analysis for this investigation is the Elaborate Qualifiers procedure. In sum, moderate support has been gained for all three coding schemes. While the validation results are not irrefutable, analysis of the data can proceed. Chapter Four presents the results of the data analysis by addressing each of the research propositions and research questions advanced in Chapter Two. However, cautious interpretation of the findings is warranted, given the exploratory nature of this investigation and the equivocality of the coding methodologies. mm REILTS The results of this investigation could be presented in a couple of different ways. One approach would be to divide the chapter into three sections, corresponding to each of the three transcripts, and to discuss the findings for each research question and research proposition in terms of each transcript. This approach would work well as a case study -- allowing for a full description of communication behavior within each transcript to be presented. However, a presentation of this type would make the questions and propositions subordinate to the transcripts, since it would emphasize differences and similarities between transcripts. An alternative approach would be to present the results for each transcript in response to the individual research questions and research propositions. This latter procedure would facilitate discussion of the specific behaviors posited in the questions and propositions as they occur within the individual interactions, as well as a comparison of the behaviors across the three negotiations. In this way, the individual transcripts could be used to explicate potential patterns and differences in the data. Since this investigation explores specific communicative behaviors within hostage negotiations, and attempts to discern potential behavioral patterns across transcripts, it makes more sense to review the findings according to the second organizational approach. 104 105 This chapter is divided into two major sections. The first section is a presentation of the results for the individual research questions presented in Chapter Two. These research questions are explained before the research propositions because the questions are more general in focus. For presentation purposes, the discussion of results for individual questions have been combined in those cases where they addressed the same goal behavior or combination of goal behaviors. The second major section is a description of the results for the individual research propositions. As with the presentation of results for the research questions, thematic propositions with common foci have been combined for the ease of discussion. Since this investigation is exploratory and descriptive, only basic statistical analyses were performed; these include frequencies, frequency distributions, percentages, mean composite language intensity scores, and some correlations. Specifically, frequencies and means were computed for hostage-taker and negotiator use of identityegoal and relational-goal coding category behaviors for the eight time phases of each negotiation, as well as across negotiations. Likewise, negotiator and hostage-taker mean language intensity scores were computed for each of the eight phases of each interaction, and also across the three negotiations. Correlations were computed between mean language intensity scores for each phase with the relational-goal means for each phase at both the individual and aggregate levels. The results are presented on two levels. One level is a discussion of the results for each of the three individual transcripts. This transcript-level analysis is important because it 106 offers insight into the actual communicative behavior of individual hostage-takers and negotiators. This facilitates discernment of individual differences between transcripts according to the idiosyncratic characteristics of each negotiation. The second level of presentation is at the aggregate level, across the three negotiations. While transcript by transcript analysis offers meaningful insight into individual negotiations, the aggregate level analysis is valuable because it provides a cumulative accounting and averaging of negotiator and hostage-taker behavior that can be used to build a communication data base. This data base then can be used to develop and test models of hostage negotiation using other transcripts. It is acknowledged that this data set consists of only three transcripts. Yet, since no communication data base currently exists for hostage negotiations, this investigation represents an initial contribution toward building that data base. Further, while hostage-takings differ according to the type of hostage-taker (i.e., criminal, mental), these three interactions involve hostage-takers who are mentally and emotionally distraught; thereby denoting a theme of commonality. As such, a data base specifically for hostage-takings involving mentally and emotionally distressed hostage-takers can be initiated. 107 RESULTS FOR THE RESEARCH QUESTIONS RQl. What is the frequency of negotiator and hostage-taker use of integrative and distributive relational-goal conflict tactics over time? RQZ. What is the difference in frequency between hostage-taker use of integrative relational—goal conflict tactics and negotiator use of distributive relational-goal conflict tactics over time? Frequencies of hostage-taker and negotiator use of integrative and distributive relational-goal conflict tactics were computed for each the three independent interactions by phase, and across phases; as well as across all phases of all negotiations (i.e., aggregately). The findings for relational—goal tactics are presented at two levels. On the macrolevel, tactics are grouped into the integrative, distributive, or Other category. At the microlevel the individual 15 tactics are delineated. The results are presented aggregately and then for each interaction. Wits There were a total of 1,814 utterances across all three negotiations, 905 of which are attributed to the hostage-takers and 908 to the negotiators. On the macro category level, Table 7 reveals that integrative tactics account for 82.2% of the negotiators' behavior (n = 747), with distributive accounting for only 5% (n = 45). The "other" category (i.e., positive and negative backchannels, and salutations) represents 12.8% (n = 117) of negotiators' utterances. 108 In comparison, Table 8 indicates that distributive tactics account for 10.7% (n = 97) of hostage-takers' behavior. Integrative tactics represent 68.5% (n = 620) of all utterances, with ”other" message behaviors accounting for 20.8% (n = 188). Figure 1 offers a graphic comparison of the relational-goal behaviors between negotiators and hostage-takers across all three interactions. Tables 7 8 8, Figure 1 about here Dissecting the negotiators' relational tactic use into the 15 individual categories reveals a fairly wide distribution among the various integrative strategies (FigUre 2). As reflected in Table 9, "statements of disclosure" were the most frequent (20.6%, n = 187). "questions of fact about person" accounted for 18.4% (n = 167) of all utterances, followed by "statements of fact about person" (12.6%, n = 114), "positive backchannels (12.4%, n = 113), and "questions of disclosure" (10.8%, n = 98). Interestingly, expressions of "willingness to comply" accounted for only 9.3% (n = 98). "statements of wants and demands" accounted for only 4.4% (n = 39) of all relational tactics, but 87% of the distributive-only tactics. "rejection of other's demands" occurred only 6 times (0.7%) across all three negotiations. For the most part, these results reflect a fairly neutral, yet, supportive relational interaction pattern on the part of negotiators. Table 9, Figure 2 about here 109 The distribution among the 15 individual relational-goal tactics for hostage-takers is as equally diverse as it is for negotiators (Figure 3). However, only three tactics account for 67% of all codes (Table 10). Specifically, ”statements of disclosure" account for 29.4% (n = 266) of all behaviors, with "positive backchannels" representing 19.2% (n 174), and "statements of fact about person" constituting 18.5% (n 167). "statements of wants and demands are the fourth most frequent, accounting for 8.5% (n = 77). All other tactics individually account for a single digit percentage. Not very surprisingly, all of the distributive tactic codes are represented. Of these, ”statements of wants and demands" account for 8.5% (n = 77), followed by ”threats against self" (1.2%, n = 11), with "threats against other" and "reject other's demands" each accounting for 0.4% (n = 4). In comparison to negotiators, hostage-takers are also predominately integrative. Yet, hostage-takers do engage in distributive behaviors twice as often. Table 10, figure 3 about here Transcriptaflumber_0nei_ECEO There were a total of 269 coded utterances in the WCPO negotiation. Of these, 135 were made by the hostage—taker and 134 by the negotiator. It is visible from Figure 5 that both the hostage-taker and negotiator engaged primarily in integrative behavior throughout the interaction. The negotiator's behavior was consistently integrative, while the hostage-taker did engage in some distributive message behaviors. As reflected in Table 13, practically 110 all (99.3%, n = 133) of the negotiator's relational-goal tactics were integrative, denoting consistency across all eight time phases. The only deviation occurred in Phase Three in which the negotiator was assigned one "positive backchannel.” This denotes a complete absence of distributive tactics. The hostage-taker's relational-goal behavior was likewise primarily integrative, accounting for 88.9% (n = 120) of the overall coded utterances (Table 14). Distributive tactics represent only 8.1% (n = 11) of the total interaction, with "other" tactics accounting for just 3.0% (n = 4). Tables 13 8 14, Figure 5 about here Phase by phase analysis indicates that distributive tactics were used by the hostage-taker in all but Phases Three, Four, and Five. The greatest frequency of distributive tactics occurred in Phase Seven (n = 4, 23.5%), Phase Two (n = 3, 17.6%), and Phase Six (n = 2, 11.8%). It is important to note that distributive tactics were recorded for the hostage-taker in the last three phases, with a doubling in frequency from Phase Six to Phase Seven. This trend reflects the hostage-taker's resolution to commit suicide, with the final utterance being the culmination of the act itself. As reflected in Table 15, the negotiator utilized practically every integrative tactic listed. However, there were four predominant tactics involved, including "statements of disclosure" (n = 38, 28.4%), "questions of fact about the person" (n = 33,24.6%), "questions of disclosure" (n = 30,22.4%), "statements of support and comfort" (n = 13, 9.7%). Both "statements of disclosure" and 111 "statements of support and comfort" occur with regular consistently throughout the eight phases, except in Phases Three and Four, where the negotiator is employing primarily "questions of fact about person" (Phase Three) and "questions of disclosure (Phase Four). These trends are consistent with the MSP phase model of negotiation in which the negotiator attempts to increase his understanding of the hostage-taker's thoughts and feelings during Phase Two (analysis Phases Three and Four). These tactics also reflect an attempt on the part of the negotiator to refocus the hostage-taker's attention away from committing suicide to positively resolving the situation (Fuselier, 1986). Unfortunately, his strategy fails in the end. Table 15 about here As reflected in Figure 6 and Table 16, the hostage—taker is primarily expressing "statements of Disclosure" (61.5%,n = 83) across all eight phases. This denotes the expression of emotional, thought and feeling-based information. Neutral “statements of fact" account for 20% (n = 27) of the tactics used across the eight phases. However, "threats against self" are the third most frequent tactic (n = 7, 5.2%). These statements occur primarily in the beginning of the interaction as the hostage-taker is conveying his intent to the negotiator (Phases One and Two). During Phases Three, Four and Five, the hostage-taker engages primarily in “statements of fact" and ”statements of disclosure,‘ reflecting a shift of attention away from the issue of suicide. However, "statements of wants and demands," "threats against self," and "statements of disclosure" dominate phases 112 Six, Seven, and Eight as the hostage-taker refocuses his attention to wanting his dog delivered to him and killing himself. The transcript ends with the hostage-taker stating "I'm going right..“ as the phone drops to the floor. Table 16, Figure 6 about here I . I H l I . H . Similar to the WCPO negotiation, both the negotiator and hostage-taker in this second interaction engaged primarily in integrative behavior (Figure 7). According to Table 17, 169 (68.1%) of the negotiator's 248 utterances were coded as integrative, while only 7.3% (n = 18) were coded as distributive. The "other" category accounts for almost one-fourth (24.6%, n = 61) of the 248 utterances. A phase by phase analysis indicates that integrative tactics were their greatest during the fourth and fifth phases (87.1%, n = 27; 83.9%, n = 26), and the lowest during the last phase (53.1% (n = 17). For the remaining phases (One, Two, Three, and Six), the frequency ranged from a low of eighteen (60%) in Phase Three to a high of 71% (n = 22) in Phase Seven. “Other" statements (positive and negative backchannels, salutations) were most frequent in Phases One, Two, and Three (32.3%, n = 10; 35.5%, n = 11; 36.7%, n = 11). Distributive tactics occurred with a frequency of one (3.3%) in both Phases Two and Three. The greatest percentage of overall distributive tactics were coded in Phases Six, Seven, and Eight (16.1%, n = 5; 9.7%4, n = 3; 25%, n = 8). 113 Tables 17 8 18, Figure 7 about here The hostage-taker's communication is also mostly integrative, accounting for 79.9% (n = 195) of his total interaction. As depicted in Table 18, integrative tactics accounted for a low of 64.5% (n = 20) of the utterances in Phase 7 to 87.1% (n = 27) in Phase Five, only 10.2% (n = 25) of the hostage-taker's 244 utterances were coded as distributive, with 9.8% (n = 24) coded into the "other" category. Phase Three consists of most of the distributive codes (22.6%, n = 7), followed by Phase Six (19.3%, n = 6). Phases Five, Seven, and Eight also involve distributive codes, ranging from only 6.5% and 7.1%, n = 2) in Phases Five and Eight respectively, to 12.9% (n = 4) in Phase Seven. As reflected in Figure 8 the negotiator employed each of the integrative tactics at least four times throughout the course of the interaction, while using only "statements of demands and wants” (5.6%, n = 14) and "rejection of other's demands" (1.6%, n = 4) as distributive tactics. According to the results in Table 19, the most frequent integrative tactic used by the negotiator involved “questions of fact about person" (16.9%, n = 42). This is followed by statements expressing a "willingness to cooperate" (15.3%, n = 38), "statements of fact about person“ (11.7%, n = 29), and "statements of disclosure" (10.1%, n = 25). "questions of fact about person" and "willingness to comply” dominate Phases One through Six as the negotiator gathers information from the hostage-taker about his condition, thoughts, and feelings, while consistently maintaining an intent to work with the 114 hostage-taker. ”statements of fact about person" dominate Phase Seven as the negotiator begins to explain his plans and activities for securing the surrender of the hostage—taker. Finally, "statements of wants and demands" dominate Phase Eight as the negotiator tells the hostage-taker what to do to surrender. Table 19, Figure 8 about here Table 20 indicates that the hostage-taker was coded as using all but two of the ten integrative tactics, including "questions of fact about hostage-taking" and ”statements of support and comfort”. The two tactics coded most frequently were "statements of fact about person” (33.2%, n = 81) and ”statements of disclosure" (23%, n = 56). "statements of wants and demands" were the third most frequent (9.4%, n = 23). Both "statements of fact about person" and "statements of disclosure" dominate across all eight phases, with "statements of demands" fairly high in Phases Three and Six. Table 20 about here This pattern tends to make sense given the nature of the interaction. This particular negotiation involves a mentally distressed hostage-taker who apparently is suffering from brain cancer. The majority of the interaction centers around the hostage-taker explaining a supposed deal he made with the CIA to kill a Soviet spy (his father) in exchange for the opportunity to overdose. In Phase Three, the hostage-taker is expressing requests for a way to 115 end his pain, while in Phase Six, he is telling the negotiator that he wants to go to the hospital and he wants the negotiator to work with him to help end his pain. The interaction ends with the hostage-taker surrendering and being taken to the hospital. WW There were a total of 1,053 utterances in this interaction, 528 were made by the hostage-taker and 527 by the negotiator. As with the two other interactions, integrative tactics were the dominant relational—goal coded for both the hostage-taker and negotiator. Figure 9 depicts the distribution of relational tactics across all eight phases for both interactants. Clearly, the negotiator employed almost completely integrative tactics. As shown in Table 21, 84.4% (n = 445) of the 528 utterances were coded as integrative. In comparison, only 5.1% (n = 27) were coded as distributive, and 10.4% (n = 55) as "other." While distributive tactics are recorded for all eight phases, they are the greatest during the last four phases (Phases Five, Six, Seven and Eight). However, they never exceed more than 13.8% (n = 9) of total codes within a single phase (Phase Seven). Integrative tactics never fall below 75.8% (n = 50) of all codes for a single phase (Phase 4). In this case, 22.7% (n = 15) of the 66 codes in Phase Four were coded as "other." Tables 21 & 22, Figure 9 about here 118 Overall eight phases, the hostage-taker's use of integrative tactics account for only 58% (n = 305) (Table 22). Of the 527 utterances made by the hostage—taker, 11.8% (n = 81) were coded as distributive, with 30.4% (n = 180) coded as "other." Integrative tactics are the greatest in Phase Two (77.3%, n = 51) and the lowest in Phase 7 (45.5%, n = 30). They are also quite low in Phase Eight, accounting for only 48.4% (n = 31). Distributive tactics are the greatest in Phase One (27.7%, n = 18), followed by Phase Eight (25%, n = 18). "Other" tactics account for 18.7% (n = 11) in Phase Two, and 51.5% (n = 34) in Phase Eight. This increase in “other" tactic use involving "positive backchannels” reflects the hostage-taker's passive agreement to the surrender plan delineated by the negotiator. However, this is counteracted by the increase in distributive tactic use during the eighth phase, in which the hostage-taker backs down on the plan. Figure 10 indicates that the negotiator used each of the integrative tactics at least seven times throughout the entire interaction. Of these, "statements of disclosure" were the most frequently used, accounting for 23.3% (n = 123) utterances across the eight phases (Table 23). The second most frequent tactic coded was "questions of fact about person" (17.8%, n = 94), followed by "statements of fact about person" (15.2%, n = 80), and "questions of disclosure" (12.1%, n = 840). Within the interaction, the negotiator is primarily soliciting thoughts, feelings, and information from the hostage-taker, along with providing information about himself and his feelings_about the hostage-taker's situation. All but one of the "other” codes are "positive backchannels” accounting for 9.7% (n = 51) 117 of all utterances. Apparently, the negotiator is using this message behavior as a means to maintain positive interaction with the hostage-taker. This appears to be consistent with MSP and FBI training literature about the need to establish trust and rapport with the hostage-taker (Fuselier, 1988). Table 23, Figure 10 about here Regarding distributive tactics, "statements of wants and demands" are the most frequent, accounting for 4.7% (n = 25) of the total 527 utterances. The negotiator's use of these tactics tends to be the greatest during the last four phases, with Phase Seven characterized by nine "statements of demands" codes. During this phase of the interaction, the negotiator is trying to tell the hostage-taker to come outside and meet him with the hostage-taker's female neighbor. This is done under the pretense of gaining the safety of the children (the hostages) and to escort the hostage-taker to the police station to conduct a news interview with his wife (the hostage-taker's primary demand). During Phases Five and Six, the negotiator attempts to bolster the hostage-taker by telling him not to consider suicide as a way out, and not to harm the children. In comparison, the hostage-taker also used most of the integrative tactics, except for "statements of support and comfort" (Figure 10). Of these, "statements of disclosure" were the dominant tactic, accounting for 24.3% (n = 128) of the 528 utterances (Table 24). For the most part, the hostage-taker was engaging in catharsis, trying to explain his plight to the negotiator and to gain the 118 negotiator's compassion. “Statements of fact about person" were the second most common tactic, accounting for 11% (n = 58) of the utterances. Interestingly, the hostage-taker also used ”questions of disclosure" (8.9%, n = 47) and "questions of fact about person" (8.1%, n = 32), apparently to solicit biographical information from the negotiator, as well as the negotiator's feelings about his situation. Of the ”other" tactics, all but 4% were "positive backchannels." In fact, of all the relational-goal tactics coded, 29.3% (n = 154) were "positive backchannels." During Phase Seven, where it appears that the hostage-taker is going to surrender, 49% (n = 32) of the hostage-taker's utterances are “positive backchannels." This behavior may reflect the hostage-taker's lack of commitment to the surrender plan. Table 24 about here Regarding the hostage-taker's distributive tactic use, 85% (n = 52) are coded as "statements of wants and demands.” These are the greatest during Phases One and Right, accounting for 22.7% (n = 15) within each phase. During Phase One, the hostage-taker has initiated interaction with the negotiator, and is clearly conveying his demands. "Statements of demands" continue to be expressed throughout the interaction, but at a lower frequency than in Phase One. However, during Phase Eight, the hostage-taker again begins expressing "statements of demands" (n = 15, 22.7%) as he reneges on the apparent surrender plan, and renews his demands to conduct an interview with 119 his wife. The interaction ends with the hostage-taker screaming that he wants his wife. R03. What is the frequency of negotiator and hostage-taker use of the different identity-goal conflict tactics over time? Frequencies of negotiator and hostage—taker use of identity-goal behaviors were computed for each the three independent interactions by phase, and across phases; as well as across all phases of all negotiations (i.e., aggregately). The findings for are presented at two levels. First on the aggregate level for all phases of all interactions. Second, the results are presented aggregately across phases of each interaction and then aggregately discussed in terms of possible trends as they occur by phase for each interaction. Wits As Figure 11 illustrates, the two primary identity-goal categories represented are the "restore-self" and ”restore-other." According to Table 25, 8.7% (n = 79) of the 909 negotiator utterances were coded as devoid of any identity-goal behaviors. This leaves 830 coded utterances. Of these, ”restore-other" accounts for 87.8% (n = 581), followed by "positive backchannels (18.4%, n = 138), and "restore-self" (13.1%, n = 109). Theoretically, this is as it should be (Fuselier, 1988). There are no codes for either the "other-hostile" or ”self-hostile" categories. The "defend-self" and "defend-other" categories are represented, but cumulatively account for only 2.2% (n = 18) of the 830 coded utterances. In both Majors 120 and MP0 negotiations, "positive backchannels" exceed "restore-self," but occur less frequently in WCPO. Tables 25 & 28, Figure 11 about here For the hostage-takers, 10.1% (n = 91) of their 905 total utterances were coded as "no identity goal" (Table 28). In comparison to the negotiators, "restore-self" accounts for 84.7% (n = 528) of the 905 coded utterances, followed by "positive backchannels (24.2% (n = 195) and "restore-other" (8.9%, n = 58). "Defend-self" accounts for 2.2% (n = 18), followed by "self-hostile" (.8%, n = 5), and "defend-other" (.4%, n = 4). The only code not represented is the "other-hostile" category. Cumulatively, "positive backchannels" account for a greater frequency than all other identity-goal codes in all but Majors. In this particular interaction, "restore—other" exceeds "positive backchannels" by only five codings. Again, these aggregate results tend to reflect the theoretical orientation that hostage negotiations are essentially a hostage-taker defending and restoring his position, and the negotiator acting to offer defense and restoration of the hostage—taker’s face. Traascrint_numher_0nei_HCEO For this first interaction, the negotiator accounted for 134 of the total 289 utterances. As illustrated in Figure 12, "restore-other" was the most frequent category coded. In Table 27 it can be seen that of the 134 recorded codes for the negotiator, "restore-other" represents 84.9% (n = 87), or 70% of the 128 active identityegoal codes, minus the "no identity goal" codes. 121 "Restore-self" is the second most frequent, accounting for 25.8% (n = 33) of the 128 codes. "Defend—other and "positive backchannels" are the only other two codes recorded, counting for 5.5% (n = 7) and 0.8% (n = 1) of the 128 active codes. Perhaps the most interesting of these results is that "restore-other" is the single most frequent code throughout Phases One to Five. However, in Phase Six, "restore-self" accounts for 43.8% (n=7), while "restore—other represents 50% (n = 8). During Phases Seven and Fight, "restore-self" becomes the most frequent code recorded. It is during this time that the negotiator is defending his beliefs about the hostage-taker's worth and need to live. Tables 27 & 28, figure 12 about here For the hostage-taker, "restore-self" is by far, the single most dominant behavior recorded, accounting for 82.2% (n = 111) of the total 135 utterances. After dropping the "no identity-goal" codes (7.4%, n = 10), "restore-self" accounts for 88.8% of the 125 codes. As reflected in Table 28 and Figure 12, other active codes include "self-hostile" 3.2% (n = 4), "defend-self" (1.8%, n = 2), "restore-other" (1.8%, n = 2), "positive backchannels" (4%, n = 5), and "negative backchannels" (1.8%, n = 2). Across the eight phases, "restore-self" is consistently the predominant behavior. Interestingly, the hostage-taker restores the negotiator's face in Phase Three. During this part of the interaction, the hostage-taker offers retroactive support to the negotiator for past activities which he discloses as an attempt to build a relationship with the 122 hostage-taker. "Self-hostile" behavior is present in Phases One, Two, Four, and Fight as the hostage-taker continually expresses his desire to commit suicide. This focus subsides in Phases Five through Seven as the negotiator has the hostage-taker talking about his past. Unfortunately, the hostage-taker's final behavior is to commit suicide. I . I H l I . H . Figure 13 illustrates the distribution of identity-goal tactics recorded for both the negotiator and hostage-taker. Consistent with Transcript Number One, the negotiator used primarily "restore-other” behaviors while the hostage-taker used "restore-self." Negotiator behavior accounted for 248 of the 492 utterances coded, with the hostage-taker accounting for the remaining 244. Table 29 shows that of the 248 negotiator codes, only 9.3% (n = 23) were coded as "no identity-goal" codes leaving 225 active identity-goal behaviors. As already mentioned, "restore-other" tactics are the most frequent, accounting for 48% (n = 108) of the 225 recorded codes. The second most frequent are the "positive backchannels" accounting for 28.9% (n = 85) of the total codes, followed by "restore-Self" behavior which represents 19.8% (n = 44). Proactive ”defend-other" and "defend-self" codes were recorded, equally 1.8% (n = 4) and .8% (n = 2) respectively, along with 2 (.8%) "negative backchannels.“ There were no "self-hostile" of "other-hostile" behaviors recorded. Tables 29 & 30, Figure 13 about here 123 Retroactive "restore-other“ and "positive backchannels” dominate negotiator behavior from Phase One to Three (Table 29). Beginning at Phase Four, the negotiator begins to engage in more "restore-self" behaviors, shifting back and fourth between "restore-self" and "positive backchannels.“ Throughout the interaction, "restore-other" tactics remain consistently high. According to Table 30, only 8.8% (n = 21) of the hostage-taker's identity-goal behavior was considered uncodable into one of the eight categories, leaving 223 active behavior codes. Of these, 70.4% (n = 157) were coded as retroactive "restore-self” behaviors. Interestingly, 13.9% (n = 31) of the hostage-taker's utterances were coded as "restore—other" in support of the negotiator. While "restore-self" dominates across the eight phases, "restore-other" codes are likewise recorded across seven of the eight phases, with the majority of them occurring in the last three phases. This trend reflects the hostage-taker’s willingness to work with the negotiator and to support the negotiator's requests as the interaction progresses. "Positive backchannels" are the third most frequent behavior, accounting for 11.7% (n = 28) of the total codes. Proactive "defend-self" (2.7%, n = 8) codes are recorded in four of the first five phases, but cease in the last three phases as the hostage-taker begins to work with the negotiator. Likewise, only one "self-hostile" code is recorded (.5%, n = 1) in Phase One, with one (.5%) "negative backchannel recorded in Phase Three. Again, similar to the first transcript, there were no "other-hostile" behaviors recorded, as well as no "defend-other" codes. 124 WWW As clearly illustrated in the results of the two previous transcripts, Figure 14 likewise illustrates that "restore-other" dominated the negotiator's identity-goal behavior, while ”restore-self" dominated hostage-taker behavior. More precisely, according to Tables 31 and 32, 89.4% (n = 388) of the negotiator's codes were identified as "restore-other," while 49.4% (n = 280) of the hostage-taker's were "restore—self." 0f the 527 utterances recorded for the negotiator, 9.5% (n = 50) were coded as "no identity-goal” codes; of the hostage-taker's 528 utterances 11.4% (n = 80) were also coded according to this category. Tables 31 & 32, Figure 14 about here — In terms of active identity-goal behaviors (n = 477), the negotiator used "restore-other" tactics 78.7% of the total interaction. ”Positive backchannels" accounted for 14.7% (n = 70) of the overall coding, followed by ”restore-self" behaviors (8.7%, n = 32). "Defend-self,“ " defend—other," and "negative backchannels" did occur, but only as a small percentage of the total (.8%, n = 3; .4%, n = 2; .8%, n = 4) respectively; The most telling trend across the eight phases is the relative decrease in the number of "positive backchannels" beginning in Phase Six, and the increase in "restore-self" codes starting in Phase Five. Within in the negotiation procedure itself, Phase Five is the point where the negotiator begins to talk the hostage-taker into meeting him outside his apartment and accompanying the neighbor to the police station. 125 Throughout this process, the negotiator is retroactively defending himself and his behaviors to the hostage-taker. 0f the 488 active identity-goal codes recorded for the hostage-taker, 55.8% are the retroactive “restore-self" codes. The second most frequent behavior recorded is the use of "positive backchannels," accounting for 35.1% (n = 184) of the 488 active behaviors. The remaining ten percent of identityegoal behavior is accounted for by "restore-other" tactics (4.5%, n = 21), ”defend-self" (2.4%, n = 11), "negative backchannels (1.3%,n = 8), and "defend-other" (.9%, n = 4). There are no "other—hostile" nor ”self—hostile" codes recorded. Apparently, the hostage~taker is feeling threatened and perceives the need to justify his actions, while simultaneously trying to maintain the rapport he has with the negotiator. What is interesting to note in the trends across phases is that "defend-self” is recorded ten times in Phases One through Four, but only once in Phases Five through Eight. Coinciding with this trend is the decrease in the number of "restore-other" codes across the eight phases to the point where there are none in Phase Eight. Further, “negative backchannels" reflect an increase in Phase Seven and Fight. In terms of actual dialogue during these phases, the hostage-taker is expressing his concern about his situation and the potential solutions to the negotiator in Phases One through Six. During Phase Seven, the negotiator has the apparent agreement of the hostage-taker to surrender. However, he is seemingly only expressing passive verbal agreement (i.e., "positive backchannels“). Phase Eight denotes the point at which the hostage-taker reneges on the surrender deal, and 128 renews his demands while expressing greater uncertainty about his plight. RQ4. What is the frequency and percentage of response of negotiator use of relational-goal conflict tactics (integrative/distributive) in response to (n + 1) hostage-taker use of negative identity-goal face-other tactics (face attack-other)? R05. What is the frequency and percentage of response of negotiator use of identity-goal tactics in response to (n + 1) to hostage-taker use of negative identity-goal face-other tactics (face attack-other)? As revealed in the discussion of results for Research Question Number Three, there were no "other-hostile" codes recorded for hostage-taker identity-goal behavior in any of the three negotiations. Consequently, no analysis can be conducted for these two research questions. RQB. What is the relationship between negotiator language intensity and hostage-taker use of distributive relational-goal conflict tactics over time? Frequencies of hostage-taker use of distributive relational-goal tactics were computed for each the three independent interactions by phase, and across phases; as well as across all phases of all 127 negotiations (i.e., aggregately). Mean composite language intensity scores were computed for negotiator language intensity by phase across the three interactions, as well, as by phase for each of the negotiations. Correlations were computed between the mean language intensity scores and the frequency of hostage—taker distributive relational-goal behaviors by phase. The results are presented aggregately and then for each interaction for each of the research propositions. Wits As revealed in the discussion of results for Research Question One, only 10.7% (n=97) of the 905 total utterances attributed to the hostage-takers across the three interactions were distributive. As Table 8 shows, MPD accounted for 82.8% (n = 81) of the distributive codes, followed by Majors (25.8%,n = 25), and WCPO (11.3%, n = 11). Table 11 reveals a more precise distribution of hostage-taker use of relational-goal statements based on a cumulative frequency of utterances coded into the three categories for all three negotiations. It is clear that hostage-taker use of distributive statements reflect a decrease from Phase One through Phase Four. At Phase Five, the frequency begins to increase again until it is at the same level as Phase One. This trend reflects the difficulty of the fifth stage of the MSP model - the "Resolution" stage. While the negotiator may have been successful in encouraging hostage-taker cooperation through Stage Four, the hostage-taker begins to resist the negotiator's attempts to gain his surrender in Stage Five. 128 Tables 8 & 11 about here Figure 15 illustrates the mean composite language intensity scores for all negotiators and hostage-takers from the three interactions at each phase. The pattern of negotiator intensity is bimodal, indicating a gradual increase in intensity from Phase One through Three, a sudden decrease in Phase Four, a gradual increase again through Phases Five and Six, with a decrease through Phases Seven and Fight. The grand mean composite score for all three transcripts across phases is .172 (Table 33). Table 31, Figure 15 abOut here Correlation analysis of hostage-taker distributive phase frequencies with the negotiator mean phase language intensity scores indicate a weak negative relationship (1: -.18). Figure 18 illustrates the mean intensity scores at each phase for negotiators in each of the three negotiations, while Figure 17 illustrates hostage-taker language intensity behavior for all three interactions. As can be seen, there are no clear language intensity patterns for either the negotiator or hostage-taker on the aggregate level. Negotiation by negotiation analysis can offer more insight into possible situation-specific patterns. Figures 18 & 17 about here 129 WWW As already discussed, the majority of hostage-taker relational-goal behavior across all eight phases of WCPO is integrative. More specifically, only 8.1% (n = 11) of all 135 utterances were coded as distributive (Table 14). The majority of these (83.8%, n = 7) occur within the last three phases of interaction. Phases Three, Four, and Five are devoid of any distributive tactic codes. Tables 14 & 34, Figure 18 about here In comparison, negotiator mean language intensity reflects a steady decrease from Phase One through Phase Four, followed by a fairly consistent increase from Phase Five through Phase Fight (Figure 18, Table 34). The grand mean language intensity score for negotiator behavior across all eight phases was .141. This reflects a low level of language intensity compared to the overall language intensity grand mean. Correlation analysis of hostage-taker distributive behavior for the eight phases with the negotiator's mean language intensity for each phase reveals a strong correlation (r,= .48). In other words, as negotiator language intensity increases there is also an increase in distributive message behavior of the hostage-taker. However, it is uncertain as to the direction of causal relationship, if there exists one. Yet, it is clear that a relationship between the two variables does exist. 130 I . | H l I . H . According to Table 18, 10.2% (n = 25) of the hostage-taker's 244 utterances were coded as distributive reflecting the predominance of hostage-taker use of integrative strategies . The pattern appears to be bimodal, with the majority, (58%, n = 14) occurring during the last four phases, but primarily in Phase Six. As reflected in Table 35 and Figure 19, the negotiator's mean language intensity increase dramatically from Phase One to Phase Three. This is followed by a slight drop in intensity in Phase Four, with an increase in Phase Five. Intensity then begins to decrease from Phase Six through Phase Eight. Average intensity over all eight phases was .195. The fluctuation in intensity can be understood by the hostage-taker's sporadic monologue about his illness. Table 35, Figure 19 about here Analysis of hostage-taker distributive tactic usage by phase in relationship with negotiator mean language intensity for each phase revealed a strong correlation (r = .59). As with transcript Number One, there is a significant relationship between negotiator language intensity and hostage-taker distributive behavior. This relationship denotes that as the hostage-taker becomes distributive, his language intensity increases. WW Table 22 and Figure 9 reflect the hostage-taker's overall reliance on integrative relational-goal tactics across the eight phases of interaction. Again, only a small percentage of the 131 hostage-taker's total behavior is distributive in nature (11.8%, n = 81). A significant portion of these occur in Phase One (29.5%, n = 18), with another 28.2% (n = 18) occurring in Phase Eight. These two phases account for the majority of the distributive codes (55.7%, n = 34). Unlike the first two transcripts, distributive codes are recorded for all eight phases of this negotiation. Table 22, Figure 9 about here As reflected in Figure 20 and Table 38, the negotiator's language intensity is primarily one of gradual decrease, but which begins with a slight increase in intensity occurring in Phases One through Three. This is followed by a predominant decrease from Phase Four through Phase Eight. Another slight increase occurs in Phase Six. The peaks are noted as points at which the negotiator offers positive support to the hostage-taker. Table 38, Figure 20 about here Correlation analysis of the frequency of the hostage-taker's distributive behavior by phase with negotiator mean phase language intensity revealed a strong negative relationship between the two variables (2,: -.58). As with the other two transcripts, there is a strong relationship. Yet, this one reflects an inverse relationship. Apparently, as the hostage-taker becomes more distributive, negotiator language intensity decreases in an effort to regain trust and cooperation. 132 R07. What is the relationship between hostage-taker language intensity and negotiator use of distributive relational- goal conflict tactics over time? Mean composite language intensity scores were computed for hostage-taker language intensity by phase across the three interactions, as well, as by phase for each of the negotiations. Correlations were computed between the mean language intensity scores and the frequency of hostage-taker distributive relational~goal behaviors by phase. The results are presented aggregately and then for each interaction for each of the research propositions. Wits Negotiators across all three transcripts engaged in distributive message behavior only 5% (n = 45) of the 909 turns at talk (Table 7). Only two of the three interactions had distributive codes recorded, Transcript Number Two - Majors, and Transcript Number Three - MPD. Majors accounted for 40% (n = 18) while MPD accounted for 80% (n = 27). As can be seen from Figure 4, negotiator distributive behavior is almost nonexistent. In Table 12 a trend reflecting a continual increase from Phase One through Phase Eight marks negotiator use of distributive tactics across the three negotiations. In fact, 73.3% (n = 33) of the 45 recorded codes occur within the last three phases, with 53.3% (n = 24) occurring during Phases Seven and Eight. This pattern reflects the MSP model Stage Five, in which the negotiator issues directives for surrender. 133 Tables 7 & 12, Figure 4 Hostage-taker language intensity across the three negotiations and eight phases averaged .285 (Table 33). As can be seen from Figure 15, the pattern across the three transcripts and eight phases is rather erratic, and negatively correlated with the negotiator language intensity behavior (3 = -.43). The correlation between hostage-taker language intensity and negotiator distributive relational-goal behavior for all three interactions was fairly strong (2 = .41). As such, there is a strong relationship between the intensity of the hostage-takers' message behavior and the relational-goal orientation of the negotiators' message behavior. However, as illustrated in Figure 17, the pattern of hostage—taker language intensity across transcripts and phases is a bit equivocal. Transcript by transcript analysis can offer insight into possible situational specific trends. Table 33, Figure 15 about here WM As already indicated there were no distributive codes recorded for the negotiator in Transcript Number One. The only deviation from integrative behavior recorded for the negotiator occurred in Phase 3 where there is "Other" code recorded (Table 13). Tables 13 & 34, Figure 18 134 The mean hostage-taker language intensity for WCPO across the eight phases was .325 (Table 34). In relation to the negotiator's language intensity, there appears to be a rough mirroring with distributive tactic use (Figure 18). In fact, the correlation is slightly positive (r = .17). Because of the absence of distributive relational-goal codes for the negotiator in this negotiation, no correlational analysis could be computed. In Figure 7 it is visible that distributive negotiator behavior is practically nonexistent. According to Table 17, only 7.3% (n = 18) of the negotiator's 248 utterances were coded as distributive. Of these, 88.9% (n = 18) occur during the last three phases. The other 11.1% are due to single tactic recordings in Phases Two and Three. Tables 17 & 19, Figure 7 about here Overall mean hostage—taker language intensity was .219. It is plotted along with negotiator language intensity scores for each mean in Figure 19. There is a fairly strong negative correlation between the two (1 = -.42). As can be seen, hostage-taker language intensity is erratic across phases, first increasing (Phase One to Two), then decreasing (Phase Two to Three), increasing (Phase Three to Four), decreasing (Phase Four to Five), and then primarily increasing fairly consistently during Phases Six, Seven and Fight. This pattern tends to make sense in relationship to the content pattern of the transcript. As the hostage-taker is suffering from terminal brain cancer, his behavior fluctuates dramatically from phase 135 to phase throughout the entire interaction. He spends much of the interaction explaining his plight to the negotiator, with the negotiator responding with solace and comfort. During the final Phases (Six, Seven, and Fight) the negotiator is working with the hostage-taker to achieve surrender. As described earlier, the hostage-taker becomes excited as he makes concessions. The correlation between negotiator distributive behavior and hostage-taker language intensity is small but, positive (r =.31). As such, there is a positive relationship between what the negotiator says and the intensity of the hostage-taker's message behavior. Perhaps the negotiator attempts to calm the hostage-taker down by expressing his own statements with intensity. However, it is not clear exactly what the nature of the relationship is. W Of the 909 negotiator utterances coded for this third interaction, only 5.1% (n = 27) were recorded as distributive (Table 21). While these tactics are present in all eight phases, their frequency is quite low throughout, with only Phase Seven denoting the greatest frequency (Figure 9). In fact, Phase Seven accounts for exactly 33.3% (n = 9) of the 27 codes. As with the previous negotiator distributive behavior in the Majors transcript, as well as hostage-taker distributive behavior in all three transcripts, the majority of distributive codes are recorded for the last half of the negotiation (Phases Five through Eight). More precisely, 81.5% (n = 22) of the distributive behaviors recorded are accounted for by Phases Five through Eight (Table 21). 138 The hostage-taker's language intensity behavior is presented in Figure 20, along with the negotiator's language intensity. The overall mean intensity across the eight phases is .314, denoting a fairly high level of intensity in comparison to the other transcripts. The intensity pattern is again erratic. The most notable feature is the significant increase in intensity in Phase Seven. During this phase of the interaction, the negotiator is attempting to direct the hostage-taker in the surrender/meeting procedure. The hostage-taker's message behavior during this phase consists primarily of “positive backchannels,” denoting an apparent agreement on the part of the hostage-taker with the negotiator's surrender plan. However, intensity drops in Phase Eight as the hostage-taker backs down on the plan. The relationship between negotiator and hostage-taker language intensity is negative (1 = -.22). The relationship between the hostage-taker's language intensity and negotiator distributive behavior is quite strong (2 = .72). Again, as reflected in Transcript Number Two, the relationship between the frequency of negotiator distributive behavior within each phase, and the hostage-taker's mean language intensity for each phase are positively related. Yet, it remains unclear as to the precise nature of this relationship (i.e., does one cause the other?). 137 R08. What is the mean composite score for negotiator language intensity in response to (n + 1) hostage-taker use of negative identity—goal tactics (face attack-other)? As revealed in the discussion of results for Research Question Number Two, there were no "other-hostile" codes recorded for hostage-taker identity-goal behavior in any of the three negotiations. Consequently, no analysis could be conducted for this research question. RESULTS FOR THE RESEARCH PROPOSITIONS RPl. Relationship Between Negotiator Use of Integrative Relational-Goal Conflict Tactics and Hostage-Taker Use of Integrative Relational-Goal Conflict Tactics RPla. The frequency of negotiator use of integrative relational-goal conflict tactics will remain constant over time. RPlb. The frequency of hostage-taker use of integrative relational-goal conflict tactics will increase over time. RPlc. The difference in frequency between negotiator use of integrative relational-goal conflict tactics and hostage-taker use of integrative relational-goal conflict tactics will decrease over time. 138 Frequencies of hostage-taker and negotiator use of integrative relational-goal conflict tactics were computed for each the three independent interactions by phase, and across phases; as well as across all phases of all negotiations (i.e., aggregately). The results are presented aggregately and then for each interaction for each of the preceding research propositions concerning negotiator and hostage-taker integrative relational-goal behavior. We As presented in response to Research Questions One and Two, negotiator use of integrative relational—goal conflict tactics accounts for 82.2% (n = 747) of the 909 negotiator utterances, while 88.5% (n = 820) of the hostage-takers' 905 utterances were coded as integrative. As shown in Figure 4, the integrative behavior of negotiators across all three interactions is characterized by slight "ripples" in the frequency across the eight phases of negotiation. As can be seen in Table 12, the exact percentage of integrative tactics does fluctuate from Phase One to Phase Eight. Beginning in Phase One, integrative behavior accounts for 88.8% of all relational-goal behavior. Yet, as seen in Phases Two through Eight, there is a gradual decrease in the overall percentage of relational tactics attributed to integrative behavior, dropping to 80.7% (n = 92) of all utterances in Phases Seven and Eight. By comparison, there is a gradual increase in the frequency and percentage of distributive tactics, beginning primarily in Phase Five. However, including the "other" tactics, which consists primarily of "positive backchannels” (94.2%, n = 113 of 120 codes), to the integrative code frequencies across the eight phases would significantly increase the frequency and 139 percentage of negotiator integrative behavior. This would reveal a more dramatic decrease in integrative behavior from Phase One to Phase Eight, marked by an increase in distributive tactic use. This is consistent with the MSP phase model in which the negotiator issues more demands to achieve the hostage-taker's surrender. Tables 11 & 12, Figure 4 about here Hostage-taker integrative relational-goal behavior is also fairly high across the eight phases (Figure 4). Closer analysis of the phase means in Table 11 reveals a constant fluctuation in hostage-taker integrative behavior across phases. In Phase One, 85.2% (n = 73) of 112 utterances were recorded as integrative. The frequency and percentage increases in Phase Two (79.8%, n = 91), but drops again in Phase Three. The decline continues to a low of 53.4% (n = 83) in Phase Seven, followed by a slight increase in Phase Eight (84.8%, n = 70). In comparison, distributive tactics show a decrease from Phase One to Phase Four, followed by a gradual, and continual increase in frequency from Phase Five to Phase Eight. However, if “other” tactics, which again are primarily "positive backchannels" (95%, n = 174 of 184 codes) are added to integrative tactics of each of the eight phases, then hostage-taker integrative behavior reflects a pattern of increase from Phase One to Phase Five, followed by a.decrease Phase Six to Eight. Tables 11 and 12 reveal a fairly consistent decrease from Phase One to Phase Four in both the frequency and percentage of difference 140 between negotiator and hostage-taker integrative behavior from Phase One to Phase Eight. However, there appears to be gradual increase in discrepancy from Phase Five to Phase Fight, with the greatest difference occurring in Phase Seven. Again, the MSP model predicts a potential increase in cooperation in Stages Two and Three, followed by possible hostility as the interactants work to resolve the problem. Insight into the potential situation-specific variables affecting negotiator and hostage—taker integrative behavior can be gleaned from the individual transcripts. WWW Earlier in this chapter, the results of negotiator relational-goal behavior for WCPO were presented. In that discussion, it was revealed that the negotiator engaged in no distributive behavior throughout the entirety of this transcript. This is clearly illustrated in Table 13, where it can be seen that 133 (99.3%) of the negotiator's 134 utterances were coded as integrative. It is also evident that the negotiator's integrative behavior was constant across the eight phases of negotiation. The only deviation in this relational-goal behavior was the single occurrence of an ”other" code in Phase Three. Tables 13 & 14 about here The integrative tactic use of the hostage—taker likewise dominated relational-goal behavior for this particular interaction. According to Table 14, 120 (88.9%) of the 135 utterances coded were coded as integrative. By phase, the lowest percentage of usage 141 occurred in Phase Seven, when integrative tactics accounted for only 78.5% (n = 13) of the within phase utterances. Looking at all eight phases reveals an apparent increase in integrative tactic use from Phase One through Phase Four. Beginning at Phase Five there is a steady drop in the frequency and percentage of use through Phase Seven. Interestingly, integrative tactics demonstrate an increase in Phase Eight. Yet, the hostage-taker shoots himself as the final act of the interaction. In terms of specific integrative tactics used, the majority involved "statements of disclosure" (81.5%, n = 83) and "statements of fact about person" (20%, n = 27) (Table 18). By comparison, the hostage-taker's distributive behavior in Phases Six, Seven and Eight accounted for 83.8% (n = 7) of the total distributive behavior throughout the interaction. Phase Seven alone accounted for 38.4% (n = 4). It is at the end of the interaction that the hostage-taker openly expresses his frustration and intent to commit suicide. As shown in Figure 5, negotiator integrative behavior remains constant over the eight phases. In comparison, the hostage-taker's behavior fluctuates across phases. The pattern is not a clear upward trend throughout, but rather one of relative consistency across phases. The difference in percentage between hostage-taker and negotiator use of integrative tactics does decrease from Phase One to Phase Four, but increases again from.Phase Five through Seven, with a slight decrease in discrepancy occurring in Phase Eight. Again, early Phases are marked by answering the negotiator's questions, while the latter phases reflect an increased steadfastness to die. 142 FigUre 5 about here Of the negotiator's 248 utterances coded for this interaction, 88.1% (n = 189) were recorded as integrative. As shown in Table 17, a sizable percentage (24.8%, n = 81) were coded as "other" tactics, with the smallest percentage (7.3%, n = 18) coded as distributive. Table 17 reveals a decrease in the negotiator's integrativeness from Phase One to Phase Three. However, this is followed by a large increase in integrativeness in Phases Four and Five, with a predominant decrease in integrative behavior in Phase Six through Phase Eight. The majority of the ”other" behaviors were "positive backchannels" (91.8%, n = 22) (Table 19), which could be defined as noncommittal integrative behavior. In this way, integrative behavior is substantiated across the eight phases. Yet, of particular interest to the analysis is the noticeable increase in negotiator distributive behavior in Phases Six, Seven, and Eight. As such, while negotiator integrative behavior does remain fairly constant across phases, there is a slight decrease in the last three phases, countered by an increase in distributive behavior. Again, this indicates that negotiators become more directive in the last phases to facilitate hostage-taker surrender. Tables 17, 18, & 19 about here 143 The hostage-taker's integrative behavior accounted for 79.9% (n = 195) of the total 244 utterances coded. According to Table 18, distributive tactics accounted for only 10.2% (n = 25) with "other" tactics accounted for 9.8% (n = 24). Hostage-taker integrative behavior reflects a general pattern of consistency from Phase One to Phase Four, with a slight decrease occurring in Phase Three. During this phase, distributive tactics increase markedly. Integrative behavior hits a high point in Phase Five, followed by a sizable decrease in Phases Six and Seven. Again, distributive and "other" tactics increase in frequency and percentage during these two phases. All "other” tactics coded during Phase Seven are coded as ”positive backchannels” (Table 20). Finally, integrative behavior increases in Phase Eight, accounting for 85.7% (n = 24) of total message behavior. Overall, hostage-taker integrative behavior maintains a general pattern of constancy across the eight phases, with moderate increases and decreases occurring throughout the interaction. Table 20, Figure 7 about here As can be seen from Figure 7, negotiator integrative behavior is actually lower than hostage—taker integrative behavior throughout most of the interaction. However, negotiator integrativeness does exceed hostage-taker integrativeness in Phases Four and Seven. Comparing Phase One percentage of integrative behavior with Phase Eight integrative behavior reveals an actual increase in difference between hostage-taker and negotiator behavior (Phase One difference = 15.8; Phase Eight difference = 32.8). As such, while hostage-taker 144 integrativeness remains fairly constant across phases, negotiator integrative behavior actually decreases. Wm For this third transcript, negotiator integrative relational-goal behavior accounts for 84.4% (n = 445) of the 527 total utterances coded. In comparison, distributive behavior accounts for only 5.1% (n = 27), while "other" tactics account for 10.4% (n=55) (Table 21). Negotiator integrative behavior reflects a continual up and down pattern from Phase One through Phase Eight. In other words, there is no consistent trend in which integrative behavior increases, decreases, or remains constant across the eight phases. As such, the frequency and percentage of negotiator integrative behavior does not follow a preset pattern, but rather reflects more of a phase/situation-based response. Yet, it is of interest to note that in those phases (Phases Two and Four) where negotiator integrativeness decreases, ”other" tactic codes increase. "positive backchannels" is the solitary code recorded during these phases (Table 30). So, although integrativeness decreases, actual positive behavior still continues, reflecting a fairly constant pattern of integrative behavior across the eight phases. It is also interesting to note that negotiator distributive behavior reflects a substantial increase in the last four phases, accounting for 81.5% of all distributive tactics recorded. Tables 21 & 30 about here 145 By comparison, hostage-taker integrative behavior likewise illustrates a pattern of continual adjustment, and not a clear linear increase or decrease in integrativeness across the eight phases. As delineated in Table 22, hostage-taker integrative behavior accounts for 58% (n = 305) of the 528 total utterances coded. Distributive accounts for 11.6% (n=81) while "other“ tactics account for 30.4% (n = 180). Integrativeness is at its highest during Phase Two, accounting for 77% (n = 51) of the 88 utterances, and drops to its low during Phase Seven (45.5%, n = 30). But as mentioned, "other" tactics account for a large portion of hostage-taker behavior in this transcript. According to Table 24 all but 3% of these are "positive backchannels.” By combining the frequency and percentage of integrative and "other" tactic codes recorded within each phase, there is aclear pattern of highly consistent integrativeness across the eight phases. However, distributive behavior accounts for a large percentage during both the first and eighth phases, denoting a decrease in integrativeness in Phase Eight. Tables 22 & 24, Figure 9 about here As can be seen from Figure 9, negotiator integrativeness is consistently high in frequency, but fluctuates noticeably from phase to phase. By comparison, hostage-taker integrative behavior likewise fluctuates across phases. Yet, the difference in frequency between the two tends to increase slightly as hostage-taker integrativeness increases from Phase One to Phase Eight, while negotiator 148 integrativeness actually decreases from Phase One to Phase Eight. However, the increase in difference is negligible. RP2. Relationship Between Hostage-Taker Use of Distributive Relational-Goal Conflict Tactics and Negotiator Use of Integrative Relational-Goal Conflict Tactics RP2a. The frequency of hostage-taker use of distributive relational-goal conflict tactics will decrease over time. RP2b. The difference in frequency between hostage-taker use of distributive relational-goal conflict tactics and negotiator use of integrative relational-goal conflict tactics will decrease over time. Frequencies of hostage-taker use of integrative relational-goal conflict tactics and negotiator distributive relational-goal tactics were computed for each the three independent interactions by phase, and across phases; as well as across all phases of all negotiations (i.e., aggregately). The results are presented aggregately and then for each interaction for each of the preceding research propositions concerning negotiator and hostage—taker relational-goal behavior. Aggregate_Besults Overall frequencies of hostage-taker distributive relational-goal behavior and negotiator integrative behavior have already been presented in detail in response to Research Questions One and Two. As discussed, hostage-taker distributive behavior accounts for only 10.7% (n = 97) of the 905 total utterances attributed to the hostage-takers 147 across the three negotiations (Table 7). In comparison, negotiator integrative behavior accounts for 82.2% (n = 747) of the 909 negotiator-made utterances. FigUre 1 illustrates the difference in frequency between these two behaviors for all hostage-takers and negotiators across all three interactions. As can bee seen, the discrepancy in overall frequency is quite large. Tables 7 & 11, Figure 1 about here By phase, the aggregate frequency of hostage-taker distributive behavior denotes a predominant curvilinear trend. As shown in Table 11, the hostage-taker distributiveness reflects a marked decrease in frequency from Phase One through Phase Four, falling from 17% (n = 19) of 112 utterances in Phase One to 5.3% (n = 8) of 113 utterances in Phase Four. However, beginning in Phase Five, hostage-taker distributive behavior increases slightly to 8.2% (n = 7), followed by a jump to 12.9% (n = 15) in Phase Six. It falls slightly in Phase Seven to 8.5% (n = 10), but increases dramatically to 17.8% (n = 19) in Phase Eight. Contrary to Research Proposition 2a, hostage-taker distributive relational-goal behavior does not reflect a downward trend from Phase One to Phase Eight, but rather, a curvilinear pattern marked by an initial decrease in frequency from Phase One to Four, followed by an increase in frequency in the last four phases. As discussed in response to Research Proposition 1, negotiator integrative relational-goal behavior is characterized by a pattern of moderate decrease in frequency and percentage of integrative codes from Phase One to Phase Eight. As can be seen from Table 12, the 148 trend is marked by a "rippling“ of the overall frequencies across the eight phases. As mentioned in the presentation of results for Research Proposition Number 1, inclusion of the "other" tactics, which consists primarily of ”positive backchannels" (94.2%, n = 113 of 120 codes), to each of the eight phases would significantly increase the frequency and percentage of negotiator integrative behavior. However, the general pattern of decrease in frequency would still be prevalent in Phases Six, Seven, and Eight, as negotiator distributive behavior increases. Table 12, Figure 4 about here A comparison of hostage-taker distributive tactics and negotiator integrative tactics is presented in Figure 4. As can be seen, hostage-taker distributive behavior does decrease from Phase One to Phase Four. But, it does begin to increase in Phase Five to a level slightly higher than Phase One and Eight. Negotiator integrative tactics alone are marked by a pattern of relative constancy across phases. Including the ”other” tactic frequencies identified in Table 12 to the negotiator integrative tactics makes a decrease beginning in Phase Six through Eight more distinguishable. Consequently, partial support is provided for Research Proposition 2b on the aggregate level. Support is partial in the sense that the discrepancy between hostage-taker distributive relational-goal behavior and negotiator integrative behavior does increase from Phase One to Four, but decreases again from Phase Five to Fight. 149 WWW Negotiator integrative behavior is highly constant across the eight phases of this interaction, accounting for 99.3% (n = 133) of the 134 codes (Table 13). As discussed earlier, there were no distributive codes. The one "other" tactic code was a "positive backchannel." Alone, the integrative codes denote constancy across phases. Including the one "positive backchannel" simply makes the pattern completely consistent for all eight phases. Tables 13 & 14 about here Hostage-taker distributive behavior within this interaction accounts for 8.1% (n = 11) and is bimodal (Table 14). High points of distributive tactic use occur in Phases Two and Seven. Of the four "other" codes recorded, 50% (n = 2) are "positive backchannels." There is only one "negative backchannel" which occurs in Phase Three. Analysis of hostage-taker behavior in terms of Research Proposition 2a provides partial support for the proposition. Hostage—taker distributive behavior does reflect an overall decrease in frequency from Phase One though Five, even though there is a surge in Phase Three. However, hostage-taker distributiveness does increase at the end of the interaction, with Phase Seven characterized by 23.5% (n = 4) distributive codes. Distributiveness does decrease in Phase Eight, but, the final act of distributive behavior is suicide. Based on the frequency distributions and percentages presented in Tables 13 and 14, partial support can be identified for Research Proposition 2b. As illustrated in Figure 5, the difference between 150 hostage-taker distributive behavior and negotiator integrative behavior does increase during the first five phases of interaction. However, the discrepancy decreases in Phases Six and Seven, as well as Eight, if the hostage—taker's suicide is included. Figure 5 about here I . I H l I . H . As shown in Table 17 and Figure 7 negotiator integrative relational-goal behavior accounts for 88.1% (n = 89) of the 248 total utterances coded for this interaction. A review of the integrative tactics alone reveals a pattern of general decrease in frequency throughout the interaction, marked by three points of high integrativeness. Phase One begins relatively high and is followed by two phases of consistent decrease. Phase Four involves a jump in frequency to a point exceeding Phase One, followed by two phases of decrease. This pattern continues, with Phase Seven marked by an increase in frequency, followed by a decrease in Phase Eight. Combining the "positive backchannels” coded as ”other" tactics to integrative tactics for each phase, reveals a pattern of fairly consistent integrativeness across Phases One to Five. Beginning at Phase Six, integrative behavior decreases and distributive behavior increases. Tables 17 & 18, Figure 7 about here 151 Hostage-taker distributive behavior is marked bimodal. As can be seen from the frequencies presented in Table 18, hostage-taker distributiveness is present in Phases Two and Three, with the high point in total distributiveness occurring in Phase Three. Distributiveness is also present in the last four phases, with Phase Six being the second highest frequency phase. Hostage-taker distributive behavior does reflect a decrease in Phases Seven and Eight, with Phase Eight less than Phase Seven. Regarding Research Proposition 2a, hostage-taker distributive behavior can not be said to decrease linearly from Phase One to Phase Eight for this transcript. Rather, distributiveness increases in the latter stages of the interaction, with certain phase-specific increases. This behavior can be explained by the idiosyncratic characteristics of the hostage-taker's topic of discourse during the individual phases, as was discussed earlier. Comparing hostage-taker distributiveness and negotiator integrativeness for this interaction reveals an interesting trend. As illustrated in Figure 7, as hostage-taker distributive behavior increases, negotiator integrative behavior typically decreases, with negotiator distributive behavior increasing. The exception to this pattern occurs in Phases Seven and Eight, where as hostage distributiveness increases (Phase Seven) negotiator integrativeness increases, and where hostage-taker distributiveness decreases (Phase Eight), negotiator integrativeness decreases, with a corresponding increase in negotiator distributiveness. But, for the most part, in response to Research Proposition 2b, the difference in frequency between hostage-taker distributive behavior and negotiator integrative 152 behavior actually increases. Basically, the negotiator seems to be responding to distributive behavior with distributive behavior. WEED According to Table 22, hostage—taker distributive relational-goal behavior accounts for 11.8% (n = 81) of the hostage-taker's 528 total utterances. Comparatively, integrativeness, accounts for 58% (n = 305), and "other" tactics for 30.4% (n=180). Of the 180 "other" codes 98.3% (n = 154) are "positive backchannels." As can be seen from Table 22, hostage-taker distributive behavior is characterized by a substantial decrease in frequency from Phase One to Three. This is followed by a pattern of distributive behavior maintenance at a fairly low level through Phases Four, Five, and Six; but not a complete nonoccurrence. The hostage-taker's distributiveness drops to a low of only 3% (n = 2) in Phase Seven. This is the phase in which the hostage-taker is apparently agreeing to cooperate with the negotiator's plan for surrender, as noted by the high percentage of "positive backchannels." A sudden increase in distributiveness occurs in Phase Eight, as the hostage-taker rejects the negotiator's surrender plan and issues new demands. Table 22 about here In response to Research Proposition Number 2a then, hostage-taker distributive behavior can not be said to decrease linearly from Phase One to Phase Eight. Rather, a.decrease does occur in the early phases of interaction, but is marked by an increase during the midpoint phases, followed by a sudden increase in the eighth phase. 153 Negotiator integrative behavior, independent of "positive backchannels,“ is fairly consistent across the eight phases (Table 21). Including "positive backchannels" with integrative tactics strengthens this general pattern. However, there are some slight undulations in the frequency distribution. At Phases Five and Seven, negotiator distributive behavior increases while integrative behavior decreases. Interestingly, during those phases when hostage-taker distributive behavior is relatively high (Phases One, Four, Six and Eight) negotiator integrative behavior (including "positive backchannels") accounts for 89 4% or more of his total behavior. »Comparing hostage-taker distributive and negotiator integrative behavior as illustrated in Figure 9, provides partial support to Research Proposition 2b. It is partial support in that the difference in frequency between hostage-taker distributive behavior and negotiator integrative behavior is low during those phases of high hostage-taker distributive behavior. However, the difference does not reflect a linear pattern from Phase One to Phase Eight. RP3. Relationship Between Negotiator Use of Positive Face-Other IdentityeGoal Tactics and Hostage4Taker Use of Integrative Relational-Goal Conflict Tactics RP3a. The frequency of negotiator use of positive face-other identity-goal tactics will remain constant over time. 154 RP3b. The difference in frequency between negotiator use of positive face-other identity-goal tactics and hostage—taker use of integrative relational-goal conflict tactics will decrease over time. Frequencies of hostage-taker use of identity-goal conflict tactics and negotiator integrative relational-goal tactics were computed for each the three independent interactions by phase, and across phases; as well as across all phases of all negotiations (i.e., aggregately). The results are presented aggregately and then for each interaction for each of the preceding research propositions concerning negotiator identity-goal behavior and hostage-taker relational—goal behavior. WW 'As discussed in the presentation of results to Research Question Number Three, negotiator identity-goal behavior was predominated by the use of retroactive face restore-other tactics (81.7%n, n = 581), followed by retroactive face restore-self (12%, n = 109) (Table 25). "positive backchannels" accounted for 15% (n = 138) of the negotiators' 909 utterances. Proactive face defend-other, proactive face defend-self, and "negative backchannels" respectively accounted for 1.4% (n = 13), .8% (n = 5), and .7% (n = 8) of the total codes. There were no recordings of negotiator use of face attack-self or face attack-other. Combined frequency and percentage for use of positive face-other behaviors account for 83% of negotiator total identity-goal behavior. All positive face behaviors, including face defend-other, 155 defend-self, face restore-other and restore-self, and "positive backchannels" account for 90.8% (n = 824) of the negotiators' overall identity-goal behavior. The frequency distribution of both negotiator and hostage-taker specific identity-goal behaviors across all phases of all three transcripts is presented in Figure 11. Tables 25 & 37, Figure 11 about here According to Table 37, negotiator use of all positive Face tactics remains constant across the eight phases. Specifically regarding positive face-other tactics, Table 37 does depict fluctuations across phases. The pattern for negotiator positive face- other tactics is bimodal, with increases occurring in Phases Three and Five. However, these shifts are primarily between face restore-other, face restore-self, and “positive backchannels." Given these findings, support is provided for Research Proposition 3a. Comparing Table 37 with Table 11 fails to reveal any significant pattern or trend between negotiator identity-goal behavior and hostage-taker integrative relational-goal behavior. The discrepancy in frequency between the two goal behaviors fails to decrease in a linear pattern from.Phase One to Eight. Likewise, there is no pattern visible between these behaviors by phase. However, the actual raw frequency discrepancy and percentage of integrative and positive face- other tactics per phase does reflect a decrease from Phase One to Phase Eight. Still, there does not appear to be any direct relationship between negotiator identityegoal behavior involving positive face-other tactics and hostage-taker integrative tactics. 158 Perhaps a trend will be revealed in a review of the individual transcripts. WM Of the negotiator's 134 utterances coded in this transcript, 84.9% (n = 87) were coded as face restore-other tactics. This was followed by face restore-self (24.8%, n = 33), face defend-other (5.2%, n = 7), and "positive backchannels" (.7%, n = 1). As indicated in the previous section, there were no negative face behaviors recorded for the negotiators in any of the three transcripts. Taken together, face restore-other and face defend-other account for 70.1% (n = 94) of all recorded Identity codes. Table 38 reveals an interesting pattern for positive face-other tactics across the eight phases. There is a generally high and consistent frequency of positive face-other tactics (involving primarily face restore-other from Phase One through Phase Five. However, beginning at Phase Six there is a marked decrease through Phase Eight. This decrease in positive face-other tactics is countered by a noticeable increase in face restore-self behaviors. Tables 38 & 14 about here Table 14 indicates that hostage-taker integrative tactics account for 88.9% (n = 120) of all behaviors. As previously discussed, hostage-taker integrative behavior is bimodal, with relatively high frequencies of distributive behavior occurring in Phases Two and Seven. Comparing Tables 38 and 14 reveals a pattern which indicates that as hostage-taker behavior becomes less integrative (i.e., more 157 distributive), that negotiator Identity tactics switch from positive face-other tactics to positive face-self. Specifically, as hostage-taker integrativeness drops, negotiator identity-goal behavior becomes more retroactive face restore-self. In this way, support is provided for Research Proposition 3b. Yet, the frequency discrepancy between negotiator positive face-other tactics and hostage-taker integrative behaviors does not decrease in a linear pattern from Phase One to Eight. I . | H l I . H . Of the negotiator's 248 utterances, 43.5% (n = 108) were coded as face restore-other, followed by "positive backchannels (43 5%, n = 85), face restore-self (17.7%, n = 44), face defend-other (1.8%, n = 4) and face defend-self and "negative backchannels" (.8%, n = 2) (Table 29). The combination of face defend and restore other tactics account for 45.1% (n = 112). Across the eight phases, positive face-other tactics reflect a pattern of gradual increase from Phase One to Phase Five. Interestingly, face restore-self tactics likewise increase from Phase One to Four. During Phase Five, negotiator identity—goal behavior consists primarily of face restore-other and "positive backchannels." In Phase Six, there is a substantial decrease in the frequency of face restore-other, countered by a noticeable increase in face restore-self tactics. This single episode is followed by a pattern of increase in face restore-other tactics in Phases Seven and Eight, accompanied by a decrease in face restore-self behaviors. In response to Research Proposition 3a, positive negotiator identity-goal behavior does not remain constant across 158 phases in this interaction. The pattern is mostly one of increase from Phase One to Phase Eight. Tables 29 & 18 about here In review, hostage-taker integrative tactic use in this interaction accounts for 9.9% (n = 195) of the total codes (Table 18). The pattern over the eight phases is basically bimodal. Hostage-taker integrativeness decrease noticeably in Phases Three and Seven, with companion increases in distributiveness. In comparing negotiator other-directed positive identity-goal tactics and hostage-taker integrativeness, no visible pattern emerges on a phase by phase basis. However, in looking at face defend-self tactics, there seems to be a slight increase in frequency of use by the negotiator as the hostage-taker's distributive behavior increases (Phases Three and Six). This response is most noticeable in Phase Six. Regarding research Proposition 3b, these findings offer weak support for the proposition, limited primarily by the continual increase in the negotiator's positive face-other identity-goal tactics. W As illustrated in Table 31 89.4% (n = 388) of the negotiator's 527 coded utterances were recorded as face restore-other. Of second greatest frequency were "positive backchannels" accounting for 13.3% (n = 70), followed by face restore-self (8.1%, n = 32), "negative backchannels (.8%, n 4), face defend-self (.8%, n = 3), and face defend-other (.4%, n 2). Again, there were no face attack codes recorded. Combined positive other tactics account for 89.4% 159 (n = 388), increasing the face restore-other frequency and percentage nominally. Across the eight phases, negotiator positive face-other tactics are, for the most part, relatively constant. The only single phase in which they drop significantly is Phase Two. During this phase, the negotiator employs a fairly high level of "positive backchannels." Other than for this one phase, the frequencies displayed in Table 31 reflect a relatively stable pattern across the eight phases. In reference to Research Proposition 3a then, these findings tend to offer support for the proposition. Tables 31 & 22 about here Of the hostage-taker's 528 codes, 58% (n = 305) are as integrative tactics. As discussed earlier, hostage-taker integrative relational-goal behavior is characterized by a fairly uneven pattern (Table 22). However, when the "positive backchannel" codes are combined with the integrative tactic frequencies, as argued for previously, the frequency distribution pattern tends to flatten out across the eight phases. Still, there are two phases in which hostage-taker integrative behavior is markedly lower - Phases One and Eight. By comparison, negotiator positive face-other tactics during these two phases are less than the greatest within-phase frequency. What is interesting is that in both Phase One and Eight, negotiator use of face restore-self tactics are higher in frequency than in the other six phases. As such, the negotiator is essentially retroactively restoring his own face when the hostage-taker engages in distributive behavior. 180 Regarding Research Proposition 3b, the relationship between negotiator positive face-other tactics and hostage—taker integrative only tactics does not merge in viewing their frequencies from Phase One to Phase Eight. However, if the "positive backchannels" are included in the integrative phase frequency and percentile, then there is a slight decrease in discrepancy between negotiator Identity behavior and hostage-taker integrative between Phase One and Phase Eight. Yet, the decrease in difference is not a linear one, but rather involves numerous undulations across the eight phases. RP4. Negotiators will use more positive face—other identity- goal conflict tactics (face defend-other, face restore-other) than positive face-self (face defend-self, face restore-self) and negative face-other (face attack—other) identity-goal conflict tactics in response to hostage-taker use of negative face-self tactics (face attack—self). Frequencies and percentages of response were computed for negotiator use of identity-goal conflict tactics in response to (n + 1) hostage-taker use of negative face-self tactics (n) (face attack-Self) for each the three independent interactions by phase, and across phases; as well as across all phases of all negotiations (i.e., aggregately). Due to the low frequency of hostage-taker face attack-self codes the results are presented aggregately. Still, this discussion of results looks at individual negotiator identity-goal 181 behavior within the negotiations in which hostage-taker face attack-self utterances were coded. W As already indicated, only .8% (n = 5) of the hostage-takers' 905 total utterances were coded as face attack-self. As reflected in Table 8, four of these codes were recorded in the WCPO transcript and one in Majors. Within WCPO, the negotiator's identity-goal response to the hostage-takers face attack-self was consistently retroactive face restore-other (100%, n = 4). The one hostage-taker face attack-self utterance recorded in Majors was followed by a "positive backchannel." (i.e., "Yeah"). For WCPO, the following statements are examples of the self-attacking statements made by the hostage-taker, followed by the negotiator's response. HT: "I’ve messed things up." NEG: ”Again, I'm trying to point out that you're a strong person..you're an extremely strong person and you've got a lotta room to move.“ HT: “I've fowled it up." NEG: "Nothing, there's nothing that can't be corrected, nothing." Table 8 about here Given these sparse results it is impossible to generalize at all. Given the difficulty of the identity-goal coding scheme, it may be the case that coders simply failed to adequately detect face attack-self 182 statements. Still, the consistency of the negotiator's four responses in WCPO are in line with expectations for what negotiator follow-up behavior should be. RP5. Relationship Between Language Intensity and Integrative Relational-Goal Conflict Tactics Frequencies of hostage-taker and negotiator use of integrative relational-goal tactics were computed for each the three independent interactions by phase, as well as by phase across the three interactions (i.e., aggregately). Mean composite language intensity scores were computed for hostage—taker and negotiator language intensity by phase for each of the three negotiations independently, as well as by phase for all three negotiations. Correlations were computed between the mean language intensity scores and the frequency of hostage-taker and negotiator integrative relational-goal behavior by phase. The results are presented aggregately and then for each interaction for each of the research propositions. RP5a. There will be a negative correlation between negotiator mean language intensity and the frequency of hostage-taker use of integrative relational-goal conflict tactics. W112: Hostage-taker integrative relational-goal behavior has already been described in detail in previous sections of this Chapter. In review, 88.5% (n = 820) of all hostage-takers' relational-goal 183 behavior was coded as integrative (Table 8). By phase, hostage-taker integrativeness is fairly erratic (Table 11). In comparison, negotiator language intensity is relatively low across the eight phases of all interactions (x = .172). Negotiator language intensity is bimodal, with peaks occurring in Phases Three and Six (Figure 15, Table 33). Correlation analysis of these two variables revealed a slight positive correlation (r = .15). Although this value is small, it is contradictory to the direction posited in the research proposition. Tables 8, 11 & 33, Figure 15 about here Adding the frequency of the hostage—taker's use of "positive backchannels" for each phase makes the frequency distribution for hostage-takers' integrative tactic use relatively constant across phases. The two exceptions are the low frequencies of integrative behavior in Phases One and Fight. With this modification to the integrative phase frequencies, the correlation between hostage-taker integrative behavior and negotiator language intensity is even stronger (r = .35). Consequently, the prediction posited in Research Proposition 5a fails to be supported by the data on the aggregate level for either method of computing hostage-taker integrative phase frequencies. W Within the WCPO transcript, hostage-taker integrative behavior was shown to be basically curvilinear; reflecting an increase in frequency in Phases Four and Five, with a concluding increase in 184 frequency in Phase Eight (FigUre 5, Table 14). Negotiator language intensity was likewise predominantly curvilinear, but in the opposite direction to hostage-taker integrativeness. As reflected in Figure 18, negotiator mean language intensity drops to a low point in Phase Four (.057), and then increases, to a high of .251 in Phase Eight. The correlation of these two variables is -.81. Adding the "positive backchannels" to the hostage-taker's integrative behavior does little to affect the frequency distribution across the eight phases. However, it does serve to attenuate the correlation (r = -.78). The data in this transcript clearly support the research proposition. Table 12, Figures 5 & 12 about here I . l H l I . H . Hostage-taker integrative relational—goal behavior was described earlier as marked by continual fluctuation. As can be seen from Figure 7, the frequency of hostage-taker integrativeness decreases in Phases Three, Six, and Seven. Of the 244 utterances coded, it accounted for 79.9% (n = 195). Negotiator language intensity was bimodal, reflecting high mean scores in Phases Three and Six (Figure 19, Table 35). The mean across the eight phases was .195; higher than in WCPO. The correlation between hostage-taker integrativeness and negotiator language intensity was -.02. As such, the data fail to adequately support the proposition. Table 35, Figures 7 & 19 about here 185 However, adding the hostage-taker's use of ”positive backchannels" to the integrative behavior frequency score for each phase modifies the frequency distribution markedly, with overall integrativeness dropping in Phases Three and Six. In comparison, these decreases mirror the increases in negotiator language intensity. Pearson's correlation for the integrative tactics and language intensity mean after adding "positive backchannels" changes dramatically to r = -.32. With this modification, support is gained for the research proposition. W As can be seen from Figure 9, hostage-taker integrative behavior increases in Phase Two, but is marked by a continual decrease through Phase Five, followed by a slight increase in Phase Six, and a leveling off in Phases Seven and Eight. As indicated in Table 22, distributive behavior is the greatest in Phases One and Eight, with a period of consistency in Phases Four, Five, and Six. Negotiator language intensity maintains a fairly low level, with no sudden increases in any phases (x = .180). For the most part, the language intensity distribution is bimodal, with the highest point occurring in Phase Three. Pearson's correlation between the hostage-taker's integrative behavior and negotiator language intensity was r = .85. This reflects a clear negation of the research proposition. Table 22, Figure 9 about here By adding the hostage-taker's use of "positive backchannels" to the integrative frequency for each phase produces a distribution that 188 is curvilinear, with low frequencies occurring in Phase One and Phase Eight. However, correlating this combined integrative behavior with negotiator language intensity still fails to generate support for the proposition. The resulting correlation is less (r = .83) than the first. However, it is still positive and in contradiction to the posited direction. Of the three interactions, this one seems to be attenuating support for the relationship between hostage-taker integrative behavior and negotiator language intensity found in WCPO and Majors. RP5b. There will be a positive correlation between hostage—taker mean language intensity and the frequency of negotiator use of integrative relational-goal conflict tactics. Wits Cumulatively, negotiator integrative tactic use accounted for 82.2% (n = 747) of all recorded relational-goal behaviors (Table 8). As can be seen from Figure 23, negotiator integrative behavior was fairly consistently high across the eight phases for all three interactions. There were three low points, Phase One, Phase Seven, and Phase Eight. Negotiator language intensity was relatively high across the eight phases, with a.mean of .285, compared to the negotiator's .172. As shown in Figure 17, negotiator language intensity was most noticeably marked by a peak in Phase Seven, with smaller peaks in Phases Two and Four. Correlating these two variables reveals a fairly negative relationship (r = -.31). As such, this negates the posited proposition. 187 Table 8, Figures 15 & 23 about here However, by adding the negotiator's "positive backchannels" behavior to the integrative frequency causes the negotiator's integrative behavior to become curvilinear, with low frequencies in Phases One and Eight. Correlation of the revised combined integrative behavior with hostage-taker language intensity reveals a small, negative relationship (1 = -.08). Still, at the cumulative level, the data fail to support the proposition. Izansczint_flumbsr_0ne;_HCEQ As revealed previously in this Chapter, negotiator integrative behavior accounted for all but .7% (n = 1) of all relational-goal behavior (Table 13). This one nonintegrative code was a "positive backchannel." Hostage-taker language intensity in WCPO had a mean score of .325 overall the eight phases (Table 34). The pattern was marked by a dramatic decrease in Phase Three from Phases One and Two, followed by increase in Phase Four, decrease in Phases Five and Six, increase in Phase Seven, and a leveling off in Phase Eight (Figure 18). The correlation between negotiator integrative behavior and hostage-taker language intensity was slightly positive (1 = .22). By adding the single "positive backchannel" to the negotiator's integrative behavior, this increases the frequency distribution for integrative to 100%. No correlation can be produced for this relationship because of the lack of variance in negotiator integrative behavior. Still, these data support the research proposition. 188 Tables 13 & 34, Figure 18 about here Negotiator integrative behavior was not nearly as high in WCPO, but still accounted for the majority of relational-goal behavior (88.1%, n = 189). As reflected in Figure 7, integrativeness fluctuated across the eight phases, with the greatest frequency occurring in Phase Four, and the lowest in Phase Eight. Hostage-taker language intensity likewise varied across the eight phases, averaging .219. However, it was characterized by an up and down pattern through Phase Five, after which it appears to level off during Phases Six, Seven and Eight (Figure 19). The correlation for hostage-taker language intensity and negotiator integrative behavior was small, but positive (r = .10). Figures 7 & 19 about here However, when the negotiator's "positive backchannel“ codes are added to the frequencies of integrative behavior, the correlation reverses completely to r = -.10. The negotiator's distribution of combined integrativeness now reflects consistency across Phases One through Five, with slight decreases in Phases Six and Eight. Nonetheless, while weak support is found for the proposition from integrative-only codes, it is negated when the combined integrative . codes are used in the correlation. 189 Transczin£_Nnmb££_Ih:eei_MED In this third interaction, negotiator integrative behavior accounted for 84.4% (n = 445) of all relational-goal codes (Table 21). Negotiator use of integrative tactics across the eight phases was highly constant, with only one major drop in frequency and percentage occurring in Phase Four (Figure 8). In comparison, hostage-taker language intensity was relatively high, averaging .314 across the eight phases (Table 38). The language intensity pattern was marked by peaks at Phase Three and Phase Seven, with Phase Seven being the most dramatic (Figure 22). Pearson's correlation reflects a small, negative relationship between these two variables (r = -.7). However, when the negotiator's "positive backchannel" code frequencies are added to the integrative code frequencies, the correlation increases dramatically, but negatively, to r = -.55. Consequently, the data fail to support the predicted proposition. The only transcript which does lend support is WCPO. Tables 21 & 38 Figures 9 & 22 about here RP5c. There will be a negative correlation between negotiator mean language intensity and the frequency of negotiator use of integrative relational-goal conflict tactics. 170 As the results for each of the behaviors has already been discussed in detail, only the specific results for this proposition are presented. W The relationship between all negotiators' integrative behavior and language intensity scores was found to be fairly negative (2 = -.38), thereby offering support for the research proposition. .After adding the negotiators' "positive backchannel" behavior with the integrative frequencies, the relationship was found to be even more inverse (r = -.50). In both cases, overall results lend support to the proposition. Izanscnini_flumbez_flnei_HCEQ For this individual transcript, the data offered clear support for the proposition (1 = -.44). With the ”positive backchannels" added to the integrative code frequencies, support was found to be even greater (1 = —.78). As such, this transcript supports the proposition. I . | H l I . H . Unlike transcript number one, the data from this transcript failed to support the research proposition. Specifically, the correlation between negotiator language intensity and integrative behavior was found to be positive (r = .23). The relationship did change ever so slightly after the "positive backchannels" were added to the integrative tactics. However, the change was insignificant (r = .22). Therefore, support is not gained for the proposition in this interaction. 171 W The results for this interaction are rather interesting. Using just the integrative code frequencies, the correlation between negotiator integrative behavior and language intensity is negative (1 = -.39). However, after combining "positive backchannels" with integrative codes, the correlations changes dramatically to r = .25. Therefore, support is gained from the integrative-only frequencies, but lost when combined with the "positive backchannels." This may be due to the high frequency of negotiator use of "positive backchannels" throughout the interaction, which consequently increases the correlation. RP5d. There will be a positive correlation between negotiator mean language intensity and the frequency of negotiator use of distributive relational-goal conflict tactics. Wits Across the three interactions, negotiator us of distributive relational-goal conflict tactics accounted for only 5% (n = 45) of all relational codes (Table 7). As illustrated in Figure 10 and delineated in Table 12, the greatest frequency of negotiator use of distributive tactics occurred during the last three Phases (Six, Seven, and Eight). Negotiator language intensity has been described as quite low, averaging .141 across the three transcripts and all eight phases (Table 34, Figure 15). Across all interactions, the relationship between negotiator distributive behavior and language intensity was found to be moderately strong and positive (r =.31). 172 Therefore, support for this fourth research proposition has been gained at the aggregate level. Tables 7, 12, & 34 Figures 11 & 15 about here Transcript Number QDE' HCEQ It is not possible to calculate the correlation between negotiator language intensity and distributive tactic use for this transcript as there were no recorded instances of negotiator distributive behavior. I . I H l I , H . As reflected in FigUre 7 and Table 17, negotiator use of distributive relational conflict tactics was almost nonexistent, accounting for only 7.3% (n = 18) of all relational-goal codes recorded. Again, negotiator language intensity was fairly low, averaging .195 across the eight phases. However, as described earlier, it was characterized by two peaks of intensity (Figure 21, Table 35). The correlation between negotiator distributive behavior and language intensity was found to moderately negative (r = -.21). As such, support was not found for the proposition. Tables 17 & 35 Figures 7 & 21 about here 173 Transcript Number Threa' MED For this third transcript, negotiator distributive tactic use was again quite small, accounting for only 5.1% (n = 27) of all relational—goal codes recorded. As illustrated in Figure 9 the distributive behaviors that were used were implemented primarily in the last four phases of interaction, with Phase Seven accounting for the greatest frequency (n = 9, 13.8%). Negotiator language intensity was slightly lower in this interaction than in Majors, averaging .180 across the eight phases. The relationship between negotiator language intensity and distributive tactic use was a fairly strong and negative one (r = -.51). As such, support is not found for the proposition. Figure 9 about here RP5e. There will be a positive correlation between hostage-taker mean language intensity and the frequency of hostage-taker use of distributive relational-goal conflict tactics. AEEIEEEL2_BesnlLs Across all three interactions, hostage-taker use of distributive relational-goal conflict tactics accounted for 10.7% (n = 97) of all relational codes (Table 20). Hostage-taker us of distributive tactics was the greatest at the two ends of the interactions (Phase One and Eight) (Table 11, Figure 4). Hostage-taker language intensity was predominantly high across the eight phases of all three interactions, averaging .285 (Table 33, Figure 15). Intensity reached the highest 174 level during Phase Seven (Figure 15). In correlating hostage-taker distributive relational-goal behavior with the mean language intensity of each phase, it was revealed that the relationship is negligible (r = -.04). As such, the overall results fail to offer support for this proposition. Tables 11, 20, & 33 Figures 4 & 15 about here WW For this first transcript, hostage-taker distributive behavior accounted for 8.1% (n = 11) of all relational coded recorded (Table 14). As illustrated in Figure 5, distributive tactic use is relatively high in Phases Two and Seven, with the greatest percentage occurring in Phase Seven. As already discussed, hostage-taker language intensity is high across the eight phases, averaging .3251 (Table 34). As shown in Figure 18, intensity begins high, but drops in Phase Three, increases in Phase Four, followed by another decrease in Phases Five and Six, with a subsequent increase in Phases Seven and Eight. The correlation between hostage-taker language intensity phase mean and distributive tactic use was found to be positive (1 = .30). Although the strength of the relationship is not very strong, the data do support the proposition. Tables 14 & 34 Figure 5 & 18 about here 175 I . | H l I . H . Hostage-taker distributive relational-goal behavior accounted for 10.2% (n = 25) of the total relational codes (Table 18). These were fairly well distributed across the interaction, with Phase Three accounting for the greatest percentage (22.8%, n = 7), followed by Phase Six (19.3%, n = 8). Yet, the majority occurred during the last four phases (Figure 7). Hostage-taker language intensity averaged .219 across the eight phases, marked by a pattern of fluctuation (Table 35, Figure 21). The correlation between hostage-taker language intensity and distributive tactic use was moderately negative- (r = -.37). Support for the proposition was not found in the data from this interaction. Tables 18 & 35 Figures 7 & 19 about here WW Hostage-taker distributive relational-goal conflict tactic use was the greatest in this third transcript, accounting for 11.8% (n = 81) of the relational—goal codes recorded (Table 22). As reflected in Figure 9 the greatest frequency and percentage of distributive tactic use occurred in Phases One and Eight. Still, distributive behavior was present in all eight phases. Hostage-taker language intensity during this interaction was the highest of all three interactions, averaging .314 across the eight phases (Table 38). The correlation produced reflects a moderately strong negative relationship between language intensity and distributive tactic use (1 = -.33). Again, 178 support for the proposition was not found. As illustrated in Figure 20, intensity increases dramatically during Phase Seven. However, distributive tactic use is at its lowest point in this phase, further indicating that hostage-taker language intensity and and distributive behavior are not positively related in this interaction. Still, the correlation between hostage-taker language intensity and integrative tactic use is r = -.54 reflecting an even stronger negative relationship than that between hostage-taker distributive and language intensity. Consequently, it becomes apparent that potential flaws exist in the measure of langUage intensity, relational-goal behavior, and integrative goal behavior, or even all three. Tables 20 & 38 Figures 9 & 20 about here RP5f. There will be a negative correlation between hostage-taker mean language intensity and the frequency of hostage—taker use of integrative relational-goal conflict tactics. Hostage-taker integrative relational-goal behavior has been described at length both in earlier subsections and this Chapter. Therefore, it will not be elaborated on in this subsection. Rather, discussion focuses strictly on the data as the relate to the research proposition. 177 W Across the three interactions, the data supported the research proposition that hostage-taker language intensity and integrative behavior would be negatively correlated. Analysis of the data revealed a strong negative correlation (r = -.47). However, this changes dramatically when "positive backchannels" are added to the integrative tactics. Specifically, the finding becomes 1: .08, reflecting a very slight, but positive relationship. Using the integrative-only codes generates support for the proposition, but the combined integrative codes fail to lend support. W The data from this transcript generated support for the proposition. In the integrative code only analysis, the correlation was found to be a fairly strong, and negative one (1 = -.45). Adding the "positive backchannels" with the integrative codes simply strengthened the relationship, 1 =—.58. Again, the proposition found support in this transcript. Unlike Transcript Number One, the data from this second transcript failed to generate support for the proposition. Between the integrative-only codes and the language intensity scores, the correlation was r = .12. After combining the integrative with the "positive backchannels, this value increase to r =.18. As such, the findings negated the prediction posited in the research proposition. W Using the frequencies for the integrative-only codes produces support for the proposition (2 = - 54). But, when the "positive 178 backchannels" are added, the size and direction of the relationship changes dramatically (r =.22). Again, as mentioned in earlier discussion, the hostage-taker's use of "positive backchannels significantly affects assessment of interactive outcomes and relationships among the variables. As illustrated in these results, there are some interesting patterns of behavior and interaction occurring within these three transcripts. Support has been gained for some of the research propositions, but many were proven to be inaccurate predictions, with the findings violating theoretical assumptions outlined in Chapters One and Two. The next chapter offers possible explanations for these findings, along with an assessment of the present research investigation, followed by suggestions for future research. CHAPTERIFIVE DISCUSSION The purpose of this dissertation was to provide exploratory and descriptive insight into the identityegoal, relational-goal, and language intensity behavior of police negotiators and hostage-takers as they occur within actual hostage negotiations. This investigation also sought to answer a number of specific research questions, and to "test" several research propositions derived from communication-based conflict research. As reflected in Chapter Four, the findings of this investigation are rather extensive and somewhat equivocal. Unfortunately, no strikingly clear patterns of behavior jumped out of the data. This may be due to several factors, including the individual differences of the three negotiations, methodological weaknesses of the specific coding procedures, and the attempt to analyze hostage negotiations using a goal-based message design logic. The purpose of this final chapter is four-fold. First, it is necessary to highlight the key findings for each of the three coding schemes. Second, I wish to to discuss some of the implications of these findings pertaining to hostage negotiations in general and communication behavior within conflict. Third, I hope to assess the various strengths and limitations of the methodological procedures. And finally, I will offer recommendations for future research. The chapter begins with the interpretation of the findings. 179 180 INTERPRETATION OF THE FINDINGS This section is divided into four parts. Part One is a discussion of the results for both negotiator and hostage-taker identity-goal behavior. Part Two is a discussion of relational-goal behavior. Part Three reviews the language intensity findings. And Part Four is a synthesis of the relationship between negotiator and hostage-taker behavior according to the three coding schemes. Identity-Goal Behavior In Chapter One, face-work activities were posited to be a critical dimension of hostage negotiations. Communication-based conflict literature, which is replete with testimonies to the importance of face-work in conflict, was cited as the rationale for this proposition (Brown, 1988, 1970, 1977; Brown & Levinson, 1978, 1987; Donohue & Diez, 1985, Folger & Poole, 1984; Goffman, 1955, 1958, 1959; Hooker & Wilmot, 1978). Further, Wilson's & Putnam's (1990) goal-based model for studying argumentation clearly delineates identity-goal behavior (i.e., face-work) as one of three basic interaction goals. Given this theoretical and empirical support, this investigation sought to assess the form and function of negotiator and hostage-taker identity-goal behavior both within and across the three negotiations. As discussed in chapter Four, the majority of both negotiator and hostage-taker identityegoal behavior across the three negotiations was positive. It is no surprise that 83% of all negotiator behavior is focused on either restoring or defending the face of the hostage—taker. This was identified in Chapter One as one of the primary responsibilities of the negotiator as s/he works to reduce the 181 hostage-taker's situational and emotional uncertainty, and to build relational rapport (Fuselier, 1988; Lanceley, et. al., 1985; Miron & Goldstein, 1979). These results were consistent across all situations, demonstrating a commonality among the negotiators. Negotiator use of "positive backchannels" was also a predominant behavior, particularly in Majors and MPD. Within these two interactions, negotiators apparently used backchannels as a method for offering positive feedback to the hostage-taker as the hostage-taker vented his frustrations and concerns. However, WCPO was marked by a noticeable paucity of "positive backchannels" (.7%, n = 1). The primary distinction between WCPO and the other two negotiations was that the hostage-taker in WCPO clearly expressed the intent to commit suicide from the outset of the interaction. This distinction is clearly marked by the fact that 75% of all the hostage-takers' threats against self occurred in WCPO. For the negotiator, this meant that he had to actively dissuade the hostage-taker from his plans and not simply allow him to vent (Fuselier, 1988). In Majors and MPD, both hostage-takers were severely stressed, but neither seriously expressed a suicidal intent. Thus, negotiators could use "positive backchannels“ as a means to allow venting, gather dispositional information about the hostage-taker, and offer implicit support. Retroactive "restore-self” was common among negotiators as they attempted to justify their position and behaviors to the hostage—takers. Interestingly, most negotiators used more "restore-self" behaviors in the last three phases (Six, Seven, Eight) of their respective negotiations than the earlier phases. In relating this to the MSP phase model (Donohue, et. al., 1989), these phases 182 comprise the "Resolution" stage. During this time, as the hostage-takers debate with the negotiators over possible surrender details, negotiators apparently are forced into a position of having to justify their stance and reiterate their concern for the hostage-takers. Hence, the greater proportion of "restore-self" behaviors. As expected, negotiators did not engage in either "attack-self" or "attack-other" behaviors. This finding was consistent across all three negotiations. In keeping with their responsibility of offering support and direction, it would have been alarming to discover negotiators involved in hostile behaviors, particularly towards the hostage-takers. Denial of their desire to express hostility is a critical part of the negotiators' task. Yet, derogation of their own face can be an effective means for supporting the face of the hostage-taker. Hostage-taker identity-goal behavior was also mostly positive. Yet, as expected, this behavior was comprised mainly of retroactive ”restore-self" codes. Apparently, hostage-takers communicate in a fashion to consistently retroactively defend themselves, their positions, and their behaviors. An interesting finding across all identity-goal behaviors is the frequency of hostage-taker utilization of "positive backchannels." This pattern is most noticeable in MPD and Majors. As discussed in Chapter Four, most of the hostage-takers' use of "positive backchannels" occurs within the last four phases of interaction. In terms of the MSP phase model, these phases are equivalent to Stages Four and Five —- "Problem Negotiation" and 183 "Resolution." According to the phase model, these two latter stages denote the period in which the negotiator is trying to work out a resolution strategy with the hostage-taker. Hostage-takers may be responding with "positive backchannels" to the negotiators' proposals to convey passive agreement, or even uncertainty and hesitancy. The pattern of hostage-taker "positive backchannel" use in MPD does lend some support to this interpretation. This communicative behavior may be a marker for negotiators that complete resolution has not been achieved. Another particularly interesting finding was the frequency with which hostage-takers retroactively restored the face of the negotiators. This occurred with the greatest frequency in Majors. This may denote the hostage-takers' agreement with the negotiators' proposal, as well as a willingness to cooperate with the negotiator. This can be seen most clearly in the phase description of Majors presented in Appendix B, in which the hostage-taker expresses a trust and belief in the negotiator. This is also reflected in MPD during Phases Three and Four as the hostage-taker commends the negotiator as being an "A-l cop." Regarding hostile face behaviors, it was rather surprising to discover that none of the hostage-takers in any of the interactions engaged in the use of "attack-other" behaviors. Because of the hostage-takers' hypothesized heightened emotionality, and concern for face (Lanceley, et. al., 1985; Miller, 1980), this was an unexpected discovery. In actuality, such behaviors may have occurred, but were simply not coded as "attack-other" behaviors due to potential limitations in the coding procedure (i.e., unit of analysis too large 184 to detect single hostile expressions within the utterance). These possible limitations will be elaborated upon in a later section of this chapter. Overall, the results reflect a pattern of mostly positive and nonhostile hostage-takers, interacting with positive and supportive negotiators. Still, the integrativeness and competitiveness of individual behaviors constitutes a second critical interaction goal (Wilson & Putnam, 1990). Relational-Goal Behavior In Chapter Two, support was given for the theoretical position that communication conveys both content and relational information (Watzlawic, Beavin, & Jackson, 1987). Wilson's and Putnam's (1990) message design logic for argumentative behavior incorporated this theoretical stance by delineating relational-goal behavior as a second interaction goal. According to Wilson and Putnam, relational-goal behavior is in part concerned with the relative cooperativeness and competitiveness of conflict interactants. A review of communication-based conflict literature from the psychological and system-interpretive perspectives clearly highlighted the importance of positive, honest, and cooperative behavior in promoting positive resolutions to conflict (Bonoma & Tedeschi, 1988; Brown, 1988; Donohue, et. al., 1982; Donahue, et. al., 1984; Donohue, et. al., 1988; Krauss & Deutsch, 1988; Fitzpatrick & Winke, 1979; Huseman, 1977; Johnson, 1971, 1973; Putnam & Poole, 1987; Sillars, 1980; Tjosvold, 1974). Given the importance credited to relational-goal behavior in conflict, this study sought to investigate the form and pattern of negotiator and hostage-taker relational-goal behavior as it occurs in 185 naturalistic hostage negotiations. Several research questions and research propositions were advanced toward this end. As already revealed in Chapter Four, the results were somewhat equivocal, offering tentative support to some of the propositions while negating others. This section offers an interpretation of the findings to those questions and propositions. As already discussed in Chapter Four, the relational-goal behavior of both negotiators and hostage-takers was predominantly integrative for all three negotiations. For negotiators, 82% of their behaviors were coded as integrative. This is consistent with FBI and MSP training material, which emphasizes the need for negotiators to be cooperative and supportive, while also working to gain control of the situation (Fuselier, 1988; Lanceley, et. al., 1985; Miron & Goldstein, 1979; Taylor, 1983). In keeping with this strategy, all three negotiators engaged primarily in integrative behaviors. Unlike Majors and MPD, WCPO reflected a completely integrative approach across all eight phases. This can be explained in part by the hostage-taker's psychological disposition and commitment to kill himself. In response to this situation, the negotiator was forced into using strictly integrative behaviors to encourage the hostage-taker's surrender. In Majors and MPD, neither hostage-takers were committed to killing themselves, thereby altering the complexion of the negotiators' interaction pattern. More specifically, negotiators in Majors and MP0 did engage in distributive behavior during the course of the negotiation. However, it was limited primarily to the last four phases of interaction. 188 This behavior pattern makes sense given what the negotiators' conversational focus should be as described in the MSP phase model. Specifically, these last four phases correspond to the last two stages of negotiation, which are ”Problem Negotiation“ and "Resolution." During these stages, it is appropriate for the negotiator to begin issuing directives and instructions to the hostage-taker as they work through a plan for surrender. Consequently, these findings partially negate Research Proposition number 1a, which posited constant negotiator use of integrative behaviors across entire interactions. Negotiators did engage in a fairly high number of "positive backchannels." Again, WCPO was devoid of these as the negotiator was forced into dissuading the hostage-taker from suicide. "positive backchannels" were common in both Majors and MP0. This specific behavior offers the negotiator the opportunity to provide positive feedback, and to let the hostage-taker vent. As demonstrated in Chapter Four, the combination of "positive backchannels" with integrative-only demonstrates that negotiators' did engage in a fairly consistently high level of integrative behavior across the eight phases of the three negotiations. In this way, partial support was generated for Research Proposition 1a. But, while integrative behavior was relatively constant throughout the interactions, distributive behavior increased, thereby offsetting the integrative behavior. In further keeping with the MSP model of negotiation, negotiators employed statements which offered personal feelings and thoughts about the hostage—taker, as well as neutral fact-based and feeling-based questions about the hostage-taker. As reflected in the individual transcript analyses, these behaviors were used primarily in the early 187 stages of negotiation, which correspond to the "Introduction and Establish Contact,” and "Relationship Building" stages of the MSP negotiation model. For the most part, negotiator distributive tactics involved "statements of wants and demands" which were marked by a noticeable increase in the last four phases. Again, this marks the time when the negotiator is working through a surrender plan with the hostage-taker. Majors offers the clearest picture of this as the hostage-taker ultimately surrenders. MPD also reflects this pattern until Phase Eight, during which the hostage-taker reneges on the plan. The results for the hostage-takers' relational-goal behavior are surprising in the sense that hostage-takers' level of distributive behavior was so relatively low. In fact, only 10.7% of the hostage-takers' behavior were coded as distributive. For WCPO and MPD there seems to be a fairly noticeable trend across the eight phases, marked by a decrease in distributive behavior and an increase in integrative from Phase One through Five, followed by an increase in distributive and decrease in integrative in Phases Six through Eight. Relating this to the MSP model helps make sense out of this curvilinear trend. According to the model, the early stages (Phase One through Five) are the periods in which the negotiator is working to build a relationship with the hostage-taker; working to reduce his/her emotional excitation and hostility. However, the latter stages involve "Problem Negotiation” and "Resolution," during which time the hostage-taker returns to a distributive disposition in negotiating surrender, or in these two situations, reemphasizing personal positions and demands. In comparison, Majors is characterized by an erratic pattern of fluctuations, with an ultimate decrease in distributiveness. 188 In this way, partial, and limited support was provided for Research Proposition 2a. Because of the curvilinear distributive behavior pattern for hostage-taker relational-goal tactics, tentative support was also generated for Proposition 1b. This proposition posited a continual increase in hostage-taker integrative behavior across the eight phases. In combination with the ”positive backchannel" codes, integrative behavior was marked by a highly constant, slightly increasing trend across the eight phases of all transcripts. Yet, as already indicated, hostage-taker behavior was also characterized by an increase in distributive behavior in Phases Seven and Eight. Majors was the one exception to this pattern. In terms of specific relational-goal behaviors, hostage-takers disclosed their feelings, asked the negotiator person specific questions, offered fact-based personal information, and "positive backchannels." Statements of personal wants and demands constituted the predominant distributive behavior. Overall, hostage-takers were primarily cooperative and integrative. However, an obvious flaw in the relational-goal coding scheme is the fact that there are twice as many integrative tactics as there are distributive. Consequently, it seems reasonable that more behaviors were coded as integrative than distributive. Further, the unit of analysis, as mentioned previously, may have been too large for coders to effectively identify distributive behaviors. While distributive behaviors account for only 10.7% of all hostage-taker behaviors, it is still twice as many as negotiator distributiveness. Further, as reflected in both WCPO and MPD, a 189 hostage-taker does not need to engage in more distributive tactics than integrative to refuse to cooperate with the negotiator. In fact, it only requires one distributive statement, as in WCPO to end the interaction in a distributive fashion. Still, the critical discovery is the curvilinear trend of hostage-taker distributive behavior to first decrease and then increase. This offers verification of Fuselier's (1988) proposition that such a pattern is common among most mentally and emotionally distraught hostage-takers. By realizing that this pattern does exist, perhaps negotiators could take steps to counteract the potential increase in hostage-taker distributiveness in the latter stages of negotiation as they work toward resolution. Language Intensity Emotional excitation was identified as one of the three characteristic features of hostage negotiations as defined in Chapter One (Fuselier & van Zandt, 1987; Soskis & Van Zandt, 1988; Taylor, 1983). Zillman's (1971, 1979, 1983) transference theory of excitation was used to highlight the additive effect of emotionality for hostage-takers as they experience a hostage-taking'from an emotional arousal perspective. Bowers' (1983, 1984) work in language intensity posits that certain verbal indicators constitute correlates of emotional intensity as conveyed through language. Drawing primarily on the work of Donohue (in press), language intensity was shown to be associated with conflict behavior. Specifically, interactions characterized by high language intensity scores for mediators were less likely to result in integrative solutions than were those in which the mediator had a low language intensity score. Consequently, language 190 intensity was considered a critical variable for measurement within hostage negotiations. As discussed in Chapter Three, a Simple—Qualifiers and an Elaborate-Qualifiers coding scheme were initially used to measure language intensity. Validation results favored the Elaborate-Qualifiers, indicating that it accounted for 48% of the variance (E (3/357) = 374.73, p <.001, n3 = .48) compared to the Simple-Qualifiers scheme (F (3/357) = 84.22, p <.001, n3 = .15). Therefore, the Elaborate-Qualifiers scheme was used for measurement of hostage-taker and negotiator language intensity. As illustrated in Figure 15 and Table 33, and discussed in Chapter Four, negotiator language intensity averaged .172 across the eight phases of all three interactions, while hostage-taker language intensity averaged .285. The difference in mean scores indicates that hostage-taker language intensity is substantially higher than negotiator language intensity. This lends support to the argument that hostage-taker emotional excitation is heightened during interaction (Fuselier & Van Zandt, 1987; Soskis & Van Zandt, 1988; Taylor, 1983). Pearson's correlation between hostage-taker and negotiator language intensity was fairly strong and negative (1 = -.43). This reflects moderate support for the conclusion that as hostage-taker language intensity increases, negotiator intensity decreases, and vice versa. This conclusion must be qualified however. Based on the methods used for this investigation, it is not possible to scientifically conclude that a causal relationship exists between hostage-taker and negotiator language intensity. It is only possible to conclude that a negative relationship does exist. 191 Further, results for the language intensity scores failed to generate findings similar to Donohue's (in press), in which he noted a decrease in intensity over time. Both the hostage-takers' and negotiators' language intensity scores were characterized by continual fluctuations at the aggregate level. Patterns for the aggregate scores were marked by extreme ambiguity. However, individual trends can be deciphered from the individual interactions. For example, in WCPO, negotiator language intensity was characterized by continual decrease from Phase One through Four. However, it begins a fairly steady increase until reaching its highest level in Phase Eight. The decrease in intensity is related to the negotiator's focus on statements and questions of fact and feeling toward the hostage-taker. During the latter phases, negotiator behavior has already been described as attempting to persuade the hostage-taker to not kill himself. This becomes quite clear in Phase Eight in which the negotiator makes his final attempt to dissuade the hostage-taker. In comparison, the WCPO hostage-taker's language intensity score demonstrates a pattern of decrease, increase, decrease, and levelling off. This intensity behavior can in part be explained by the MSP phase model stage. The early stages involve a calming'down of intensity as the hostage-taker and negotiator get to know each other. This is followed by an increase as the negotiator begins to discuss the problem. A decrease occurs as the discussion shifts to neutral fact and feeling based statements by the hostage-taker. Yet, in the last two phases, hostage-taker intensity increases and levels-off as the hostage-taker reaffirms his desire to commit suicide. The correlation 192 between hostage-taker and negotiator language intenstiy was positive but small (x = .17). Both negotiator and hostage-taker language intensity in Majors are characterized by extreme fluctuation across the phases. Hostage-taker intensity behavior can be described according to his psychological disposition. The hostage-taker claims to have a brain tumor, which no doubt influences his cognitive processing abilities. As discussed in the review of relational-goal behavior, the hostage-taker's behavior is quite sporadic, marked by expressions of cooperativeness and individualistic expressions about the problem. Hence, the hostage-taker's language intensity reflects his wildly fluctuating state of mind. The negotiator's language intensity is most notably distinguished by its increase from Phase One to Three and its decrease in the last three phases. Phases One through Three denote the period during which the negotiator makes initial contact with the hostage-taker and attempts to establish rapport with him. Specifically during Phase Three, the negotiator struggles to keep the hostage-taker focused on working out a surrender plan. During Phases Six, Seven, and Eight, a surrender plan is arranged, but the negotiator works to restrain the hostage-taker from acting hastily. According to the MSP model, these last three phases mark the surrender and "Resolution" stage of negotiation. Interestingly, negotiator intensity decreases during this final stage. Overall eight phases, the correlation indicated a fairly strong negative relationship (2 = -.42) between hostage-taker and negotiator language intensity. 193 This means that hostage-taker intensity increases as negotiator intensity decreases. Again, no causal explanations can be offered. In MPD, hostage-taker and negotiator language intensity are again negatively correlated (r = -.22). Negotiator intensity reflects a continual decrease over time beginning in Phase Four. This is in contradiction to WCPO and Majors in which negotiators' intensity scores fluctuate with hostage-taker intensity. Hostage-taker intensity was consistently higher across the eight phases, reflecting the hostage-taker's heightened emotionality. Within MPD hostage-taker language intensity does not seem to follow any clear pattern. Although, in Phase Seven, intensity is at its highest point, as are the number of "positive backchannels." This offers insight into the language intensity coding scheme as being influenced by "positive backchannels." This indicates that language intensity is sensitive to both integrative and distributive language behavior. Because of this no clear accounting of language intensity can be offered for MPD, nor can unequivocal insight be gained into the relationship between hostage-taker language intensity and negotiator language intensity. In sum, the findings for language intensity behaviors are very confusing, reflecting no clear pattern across interactions. Even within negotiations, the patterns are perplexing. As such, it is not possible to offer integrative interpretations beyond individual interactions. Although the Elaborate-Qualifier scheme accounted for a greater percentage of the variance, methodological shortcomings place limitations on the interpretability of language intensity behaviors. 194 Interrelationship Between Identity-Goal, Relational-Goal, and Language Intensity Behavior Several research propositions were advanced in Chapter Two concerning the relationship between the three types of negotiator and hostage-taker behavior. In Chapter Four it was revealed that the data offered only partial support to some of the propositions, while totally disconfirming others. Given these findings, and the limitations of this investigation, it is not possible to offer definitive conclusions regarding the interactive nature of negotiator and hostage-taker behavior patterns. Further, because of the nature of naturalistic hostage negotiations (i.e., third party involvement, hostage involvement, other extraneous influencing factors) it is not possible to make causal inferences concerning the effect of one individual's behavior on the behavior of the other interactant. Nonetheless, it is worthwhile to evaluate these relationships toward the end of developing theoretical models of hostage-taker and negotiator interaction patterns. Hostage:taker_and_NegQtiator_Belational:§oal_lnteraction Only partial support was generated for the research propositions which posited a decrease in frequency between negotiator and hostage-taker integrative tactics use, and distributive and integrative behaviors. As was demonstrated in Tables 11 and 12, the cumulative difference between negotiator integrative behavior and hostage-taker integrative behavior began to decrease through Phase Four, but then increase from Phase Five through Eight. Consequently, as hostage-taker integrative behavior increases, distributive decreases, denoting an increase in frequency discrepancy between hostage-taker distributive 195 and negotiator integrative behaviors. This pattern holds true across the three negotiations. A potential cause for this trend could be the negotiators increased use of distributive tactics in the laSt four phases. Hostage-taker integrative behavior does increase from Phase One to Four, yet it is countered by a slight decrease in Phases Five through Eight, and a slight increase in distributive in Phases Five through Eight. The MSP model implicitly predicts this trend as a result of the negotiator's relationship building efforts during Stages Two and Three, followed by potential hostility and defensiveness during Stages Four and Five as the negotiator attempts to focus the conversation toward problem negotiation and resolution. Again, it is not possible to determine whether the hostage-takers' behavior causes an increase in negotiator distributive behavior, or if negotiator distributiveness sparks an increase in hostage-taker distributiveness. (-¢. '- . .-. ' - .. ;.. .. ;--- :.- :- ; '... a. . - : '.. As was illustrated in Table 25, cumulative negotiator identity-goal behavior was dominated by positive face-other tactics across the eight phases of interaction. This was verified in Table 37, which showed the negotiator's consistent use of positive face-other tactics throughout the interactions. It was already indicated that hostage-taker integrative behavior does increase from Phase One to Phase Eight. Consequently, tentative support was offered for the proposition that negotiator positive face-other and hostage—taker integrative tactics move closer in frequency across the phases of interaction. This pattern was weak, but still detectable for all three negotiations. Yet, as indicated in Chapter Four, this trend is 198 not a linear one, but rather one characterized by numerous undulations. Still, it is not possible to claim that negotiator behavior necessarily causes hostage-taker behavior. H I. | I H | -| I II I'l G 1 I | I. As discussed in Chapter Two, interactants' concern for face can have a significant impact on conflict interaction and resolution. Because of the characteristically unstable psychological disposition of most hostage-takers, the issue of face represents a critical dimension to hostage negotiations (Fuselier, 1988; SOARU, 1987b; Strentz, 1983, 1988). To reduce emotional excitation, foster rapport, and facilitate cooperation, police negotiators must continually offer positive face support to hostage-takers (Fuselier, 1988; Lanceley, et. al., 1985; Miller, 1980; Miron & Goldstein, 1979). Consequently, it was posited that negotiator's would employ more positive face-other tactics than any other face behaviors in response to hostage—taker face attack-self statements. As noted earlier, hostage-takers engaged in negative face behaviors only five times. Four of these codes were recorded in WCPO only, reflecting the hostage-taker's suicidal manifestations. Analysis of negotiator follow-up statements revealed that positive face-other tactics (i.e., face restore-other) were the dominant response, accounting for 75% (n = 4) of all negotiator responses. The other response was a "positive backchannel." Therefore, positive, but limited support was provided for the proposition. Several other research propositions were advanced regarding the relationship between language intensity and relational-goal behaviors. 197 Unfortunately, the data failed to offer any clear or consistent associations between language intensity and relational-goal behavior on the aggregate level in support of these propositions. Nonetheless, some interesting interaction patterns did emerge at the individual level. To begin with, negotiator language intensity is negatively correlated with hostage—taker integrative behavior. This is most noticeable for WCPO (r = -.81) and Majors (r = -.02). The results are even more dramatic when the "positive backchannels" are added to the hostage-takers' integrative tactic frequency (WCPO: ; = -.78; Majors: z = -.32). However, MPD is characterized by positive correlations (I = .85; r = .83). The differences between transcript correlations is most notably due to the idiosyncratic characteristics of each interaction. In WCPO the negotiator has a low intensity score because of his efforts to dissuade the hostage-taker from committing suicide. But, in both Majors and MPD, the negotiator issues statements of demands and directives to the hostage—taker, working toward surrender. These results would lead one to believe that negotiator language intensity might be positively correlated with negotiator distributive messages content. But, the correlation between negotiator language intensity and distributive tactic use was negative for both Majors (r = -.21) and MPD (r = -.51), and zero for WCPO. Further, the correlation between negotiator language intensity and negotiator integrative behavior was also negative for WCPO (r = -.44) and MPD (r = —.39), but positive for Majors (r = .23). However, this changes when "positive backchannels” are added with integrative codes. Even 198 still, WCPO was the only interaction to maintain a negative correlation and support for the proposition, while Majors and MPD were both positive (WCPO: z = -.78; Majors: r = .22; MPD: ; = .25). Consequently, it is not possible to definitively link an increase in language intensity with integrative-only or distributive-only tactic use by negotiators. The only consistent finding was the negative correlation between hostage-taker integrative-only behavior and hostage-taker language intensity for all three transcripts. When "positive backchannels" are added, the results vary, with only WCPO reflecting a strong negative correlation (r = -.58), while Majors and MPD reflect positive correlations (Majors: r = .18; MPD: r = .22). In comparison, the correlations between hostage-taker distributive behaviors and language intensity for Majors and MPD were moderately negative (Majors: r = -.37; MPD: r = -.33) while WCPO was positive (1 = .30). Regarding the correlation between hostage-taker language intensity and negotiator integrative behaviors, again, only WCPO reflected a consistently positive correlation (r =.22). As similarly reflected in these findings, it is not possible to claim consistent relationships between hostage-taker language intensity and hostage-taker relational—goal behavior, nor hostage-taker language intensity and negotiator relational-goal behavior at the aggregate level. Still, there are some possible explanations for these findings. WCPO is the only interaction which produced findings which correspond to the predictions of the research propositions. WCPO is critically unique from Majors and MPD in that the negotiator employed only integrative tactics throughout the entire interaction. As such, there 199 is no means for testing language intensity in relation to distributive tactic use. Yet, WCPO represents a negotiation in which suicide is the issue. Majors and MP0 do not deal with suicide, but rather the hostage-takers' problems and how to resolve them nonviolently. In this way, it becomes evident that the situational demands of the interaction influence the negotiation strategies of the negotiator. Another potential explanation for the findings lies within the . definition of language intensity. For the purpose of this dissertation language intensity was defined as "the quality of language which indicates the degree to which a speaker's attitude toward a concept deviates from neutrality," which is Bowers' definition (1983, p.345). Intensity was measured according to Donohue's (in press) coding scheme, with an elaboration to the qualifiers coding category. This modification allowed for the coding of both affirmative and negative adjectives and adverbs. Consequently, language intensity is a measure of both positive and negative attitudes towards a concept. As a result, the strength and pattern of separate correlations between language intensity and a positive and negative attitude are obfuscated. This represents one potential limitation of this study, which will be elaborated upon in the next section. IMPLICATIONS OF THE RESULTS The results of this investigation have application and implications for hostage negotiations and communication-based studies of conflict behavior. Regarding hostage negotiations, several summative interpretations can be derived from the data. First, the results reveal that these three negotiations involve primarily 200 integrative relational-goal behaviors on the part of both the hostage-taker and negotiator. This is comes as no real surprise for the negotiator, as cooperative communication is emphasized in negotiation training literature (Fuselier, 1988; Fuselier & van Zandt, 1987; Lanceley, et. al., 1985; Soskis & Van Zandt, 1988). But, it is rather surprising for the hostage-taker. The expectation is that the hostage-taker would be more competitive in style. Apparently, hostage negotiations are really struggles of cooperation, with both parties striving to reach an agreement to resolve the situation. While they still constitute ”extortionate transactions" (Muir, 1977), interdependency demands cooperative behavior. Still, hostage negotiations only require one distributive act to cause a violent conclusion, as in WCPO. It was also interesting to discover that all three interactions were devoid of any face attack-other codes. The expectation would be that such behavior should occur at least occasionally because of the hostage-takers heightened level of emotional excitation. Yet, these negotiations were characterized by predominately face restore-self behaviors on the part of the hostage-taker, and face restore-other on the part of the negotiator. The fact that these interactions all involved individuals who were mentally and emotional distraught, rather than criminals thwarted during a crime, or prisoners revolting, may account for this self-justifying and other-supportive behavior (Fuselier, 1988). Language intensity behaviors were more equivocal, making interpretation of patterns difficult. Nonetheless, negotiator language intensity was consistently lower through the interactions than the 201 hostage-takers’. Yet, negotiator intensity did increase at various times, primarily during those phases in which the negotiator was trying to convince the hostage-taker of his sincere concern for the hostage-taker's well-being, and while giving instructions for surrender. Hostage-taker language intensity fluctuated across phases without any clear rhythm. With the language intensity coding procedure employed, both positive and negative verbal attitudinal expressions were recorded, making interpretation of intensity as an indicator of possible aggressiveness extremely difficult. Consequently, it is important to note that language intensity can reflect both positive and negative emotional states, and may not be undesirable. Overall, these results indicate that negotiations transpire in relatively close approximation of the MSP phase model (Donohue, et. al., 1989) and the suggested strategies for dealing with mentally unstable individuals (Fuselier, 1988). Specifically, negotiators begin by establishing rapport and trust with the hostage-taker. Only after the hostage-taker begins to calm down and focus on neutral issues does the negotiator work towards resolving the problem. This is reflected primarily in the increase in the negotiators' distributive relational-goal behavior during the last phases of negotiation. The other key finding is the tendency for hostage-takers to use "positive backchannels." It was abundantly clear in MP0 that the hostage-taker was not completely committed to the surrender plan, as he engaged in primarily "positive backchannels" up to the time of his refusal to cooperate. While this may be a key to detecting incomplete cooperation, further research of other situations needs to be conducted. 202 There are implications from this investigation for the study of communication in conflict at both the methodological and theoretical levels. At the methodological level, two goal—based coding schemes have been developed and validated for use in naturalistic conflict interactions. While the relational-goal based scheme is focused almost primarily toward hostage negotiations, the identity-goal scheme has diverse applications. It represents an initial attempt to discern face behaviors from naturalistic discourse. As such, it does require modifications for improvement. An improved coding procedure to calculate language intensity was also generated as part of this investigation. The results reflect the fact that language intensity is not merely associated with distributive messages, but with integrative as well. However, the results also reveal the difficulty in interpreting the language intensity behaviors because of the fluctuating patterns. While this coding plan represents an improvement over previous methods (Bowers, 1983, 1984; Donohue, in press) it requires much improvement. Theoretically, this investigation offers additional support to the mechanistic-based perspective (Deutsch, 1980; Deutsch & Krauss, 1982; Loomis, 1959; Terhune, 1988) that communication is a critical ingredient in the resolution of conflict. It also demonstrates how tenuous communication is in such interactions, and that it only takes one competitive behavior for the entire process to end tragically. Further, this study indicates that positive communication behaviors do tend to pervade conflict interactions (Chatman & Sondak, 1989; Johnson, 1973; Putnam & Geist, 1985: Putnam & Jones, 1982). Asking research participants what conflict strategies they use (e.g., Sillars, 203 1980) may have a reverse social desirability effect in the sense that participants simply say they use distributive tactics to present an impression of self-assertiveness to the researcher, when in actuality they employ mostly integrative tactics. However, this investigation cannot verify this claim. Comparative experimental and naturalistic research needs to be conducted to test this idea. As the psychological perspective of communication in conflict indicates, the disposition of interactants has a profound effect on the course of the conflict (Bonoma & Tedeschi, 1988; Brown, 1988; Krauss & Deutsch, 1988; Fitzpatrick & Winke, 1979; Huseman, 1977; Johnson, 1971, 1973; Putnam & Poole, 1987; Tjosvold, 1974). This investigation adds to that literature the case of hostage negotiations. Perhaps in no other conflict situation is disposition so critical. As revealed in the discussion of WCPO, the hostage-taker's unfaltering desire to kill himself precludes the conclusion. LIMITATIONS OF THE STUDY Throughout this dissertation, a variety of factors have been identified and discussed in terms of their potentially limiting effect in both the measurement and interpretation of the data. These issues can grouped into two general categories: 1) limitations associated with the database, and 2) limitations of the methodology. Limitations of each of these two areas will be discussed. Limitations of the Transcripts Regarding the transcripts, it was argued in Chapter Three that the use of these three transcripts provided ample data from which to 204 conduct an exploratory investigation. The number of utterances coded, and the subsequent number of codes totaled 1,814; representing a fairly sizable data set. However, the fact remains that there are only three transcripts. As such, it is impossible to generalize beyond the individual transcripts. Further, the situation of each transcript is unique, particularly in WCPO. The psychological disposition of the hostage-taker varies, as does the motive for the hostage-taking. As reviewed in Chapter Two, the psychological disposition of conflict interactants does influence their behavior (Bixenstine & Douglas, 1987; Deutsch & Krauss, 1982; Johnson, 1973; Putnam & Jones, 1982). These individual differences need to be taken into account in hostage negotiations and investigated more closely. The results have indicated fairly strong support for the MSP phase model of negotiation as evidenced primarily in the behavior of the negotiator across the eight phases of the three negotiations. However, all but Majors is incomplete in actual negotiation between hostage-taker and negotiator. As such, it is not possible to unequivocally discern the negotiation structure of the police negotiators. In addition, MPD ends without indication of final outcome, making predictions of behavior patterns and outcomes impossible. Complete recordings could offer more definitive insight into possible trends in relation to the MSP phase model. Numerous other limiting factors existing within the data derive primarily from extraneous variables that could influence both hostage-taker and negotiator behavior, but which go unaccounted for. For example, third party participants, such as in MPD no doubt influence how the hostage-taker interacts as well as the negotiator. 205 Yet, third parties need not be limited to only those who talk on the phone. For the hostage-taker they can include the police SWAT team, the hostages, the media, and other hostage-takers. Third party influence is potentially greater for the negotiator, including such persons as the hostages, the media, the police SWAT team leader, other police officials, the hostages and hostage-takers' relatives, and even other members of the negotiating team. Consequently, naturalistic conflict interactions such as these are replete with troublesome limitations. However, these factors should not negate efforts to investigate these interactions. Limitations of the Research Methodology Throughout Chapter Three, potential weakness of the research methods were identified in relation to the specific coding procedure. Validation of the three coding schemes did reflect moderate support for each procedure. However, certain limitations do pose serious problems for the methods. First, with the unit of analysis being the utterance, much of the finer aspects of an individual's turn at talk are suppressed by the more dominant theme of the statement. This means that the actual frequency of identity-goal or relational-goal behaviors may not be completely accurate in relation to actual discourse. Second, the validation results for both the identity-goal and relational-goal coding schemes are suspect because of potential weaknesses associated with the sample statements used in the validation procedures. Specifically, the sample statements were taken out of context creating problems for validation participant interpretation. Because the statements were randomly sampled according to the coder 208 ratings they may not have been the best representations of the specific coding categories. Yet, the validation results do offer moderate support for both coding schemes. The language intensity coding procedure is also suspect. The operational definition of what constitutes intense language is not completely clear. Though the overall average reliability among coders was high for all five categories, (alpha = .89) the individual categorical definitions require clarification. For example, modification of the qualifiers category resulted in a noticeable improvement between language intensity coding results when validated, with the Elaborate—Qualifiers procedure accounting for three times as much variance as the Simple-Qualifiers (n3 = .48). Further clarification could be offered for the other intensity categories. But, more importantly, the language intensity coding scheme is devoid of any paralinguistic cues as metrics for assessing intensity. Effort needs to be directed toward improving this measure by evaluating the ratings of content-only statements with full verbal and nonverbal ratings to check for potential indicators. Additional insight could be gained be asking subjects to rate the intensity of statements, and then asking for their rationale. In combination, these procedures could aid in the production of an accurate language intensity coding scheme for conflict interactions . A final limitation of this investigation was its lack of interactive analysis. Communication in all contexts is interactive (Miller & Steinberg, 1978). This study clearly demonstrates the need to analyze naturalistic interactions from a lag sequential perspective. 207 As such, other investigations of communication must also strive to assess behavior interactively. RECOMMENDATIONS FOR FUTURE RESEARCH Since this investigation was exploratory, the directions for future research are numerous. But, rather than attempt to delineate all possibilities, only a few key directions are discussed. There are two primary areas in which research can proceed. First, additional experimental investigations can be conducted in an attempt to better understand the behaviors being measured by the three coding schemes. In this vein, effort should be made to improve the individual coding categories. For example, using sample statements from the transcripts and having raters agree on how to code them could help in clarifying the types of statements that should be coded into the individual categories. Of particular interest is the language intensity coding scheme. Attention should be given to fine-tuning the conceptual and operational definitions of each of the five intensity categories, especially "qualifiers." Having subjects rate the actual statements from an audio recording and comparing the intensity rating with the intensity score may reveal possible similarities or dissimilarities. Asking raters to code the statements and then asking them their reasons for the assigned ratings could offer additional insight into what makes a statement intense. This would be an improvement over traditional language intensity research in which researchers had subjects evaluate experimenter fabricated statements (Bowers, 1983; Bradac, et. al., 1979; Burgoon, et. al., 1971; Burgoon & Miller, 1971; Infante, 1975). 208 In any case, language intensity represents a potentially promising area for future research. An additional suggestion for future methodological research involves using the same transcripts and procedures, but with smaller units of analysis than the utterance. As indicated in an earlier section, potential behavioral patterns within long utterances may have been suppressed by more dominant conversational behaviors. For example, research could focus on individual thought units. Further, using the same data set, lag sequential analysis could be conducted to discern specific interactive behavior patterns. This would offer greater insight into possible causal relationships, the results of which could be used in developing specific theoretical models of hostage negotiation. Another possibility involves dividing the transcripts into the eight time phases based on content rather than number of utterances. In other words, the phases could be established by having raters identify those points in the transcripts where they perceive a major shift in negotiator conversationl focus. This might enable a more sound assessment of the MSP phase model. Finally, although simulated negotiations fail to provide complete external validity, they can offer complete recordings of entire negotiations. This would allow for more accurate communicative assessments across negotiations over time. In terms of other potential theoretical issues, this research could differentiate between the hostage—taking type (i.e., criminal, mental) and the hostage-taker's disposition to investigate potential behavioral patterns within specific negotiation types. Research by Donohue, Rogan, Ramesh and Borchgrevink (1990) indicates unique 209 differences between hostage-taker behaviors based on the type of hostage-taking. The effect of individual disposition on conflict behavior has been well documented in the communication-based conflict literature (Bixenstine & Douglas, 1987; Deutsch & Krauss, 1982; Johnson, 1973; Putnam & Jones, 1982), as well as in the FBI training literature (Fuselier, 1988; Lanceley, et. al., 1985; SOARU, 1987b). Future investigations should attempt to study the behavior patterns both within and across different hostage-taking situations. Additional efforts could be directed toward understanding the effects of third party interactants on both the hostage-taker and the negotiator. Hostage negotiations do not occur within a vacuum. Numerous extraneous variables influence the entire process. Perhaps the most easily accessible would be third party interaction with hostage-takers. As indicated earlier in this dissertation, several other negotiations are available to the researcher, many of which contain significant other interaction with the hostage-taker. Consequently, this represents a potentially viable area for additional exploratory research. Finally, simply continuing to investigate other naturalistic hostage negotiations according to the current methodology, with improvements to the coding procedures, would add to the data archive of hostage-taker and negotiator communicative behavior. Through this continuing research, potential trends may reveal themselves as the number of cases increase. In any case, research into the communication of hostage negotiations represents an highly pragmatic and important area for investigation. Continuing research can only help increase our 210 understanding of this intense conflict behavior, so that one day violent conclusions can become an outcome of the past. LIST OFW Albrecht, T. L. & Adelman, M. B. (1987), Communicating 39958] supper: New York: Sage Publications. Bakenan. R &Gottman J M (1988) WM W New York: Cambridge University Press. Bixensteine, V. E. & Douglas, J. (1987). Effect of psychopathology on group consensus and cooperative choice in a six-person game. W3? -.37 Bonoma, T. V, Schlenker, B. R. Smith, R. B, & Tedeschi, J. T. (1970). Source prestige and target reactions to stress. EsmhonometriLSciences.._191 111-113 Bonoma, T. V. & Tedeschi, J. T. (1974). The relative efficacies of escalation and deescalation for compliance-gaining in two-party conflicts. MW 212- 218 Borah, L. A. (1983). The effects of threat in bargaining: Critical and experimental analysis W 37—44. Bowers, J. W. (1983). Language intensity, social introversion, and attitude change. Spee5:;h_M5:nqgraphshhfigL 345-352. Bowers, J. W. (1984). Some correlates of language intensity. Quarterly W 415-420. Bradac, J. C, Bowers, J. W. & Courtwright, J. A. (1979). Three language variables in communication research: Intensity, immediacy, and diversity W 257- 289. Brady, J. V. (1970). Endocrine and autonomic correlates of emotional behavior In P Black (Ed ) EmcsiologicaLcorrelatestmotion. New York: Academic Press. Brown, B. R. (1988). The effects of need to maintain face in interpersonal bargaining W W 107-122 Brown, B. R. (1970). Face—saving following experimentally induced embarrassment. 255— 271. 211 Brown, B. R. (1977). Face- -saving and face- restoration in negotiation. In D Druckman (Ed ). Negotiationsiocialmmhologiaal perspectives. (pp. 275- -299). Beverly Hills, CA: Sage Publications. Brown B. R. & Levinson, S. (1978). Universals in language usage: Politeness phenomena. In E. N. Goody (Ed. ). Questions_and (pp. 58- -289). Cambridge: Cambridge University Press. Brown. B R &Levinson. S (1987) Walsh language_usage1 Cambridge: Cambridge University Press. Burgoon, M. & Miller, G. R. (1971). Prior attitude and language intensity as predictors of message style and attitude change following counterattitudinal advocacy. Journal_g£_£ezsonality MW 240- 253 Burgoon, M, Jones, S. B. & Stewart, D. (1971). Toward a message- -centered theory of persuasion: Three empirical investigations of language intensity. Human_Communicat1on ResearaLll 240- 258 Carmichael, C. W. & Cronkhite, G. L. (1985). Frustration and language intensity. SeaacLMonograahssazl 107-111. Cannon, W. B., (1929). Bodily changes in pain, hunger, fear and rage. New York: Appleton—Century. Chatman, J. A., & Sondak, H. (1989). Integrating communication and negotiation research. Paper presented at the Conference on Research on Negotiations in Organizations. March 31—April 2, 1989. Cissna, K. N. & Sieburg, E. (1981). Patterns of interaction; confirmation and disconfirmation. In C. .Wilder-Mott & J. H. Weakland, (Eds), W (on 253- -282) New York: Praeger Cohen, J. A. (1980). A coeffecient of agreement for nominal scales. EducationaWhWLMaammnentszfli 37-46. Craig, R. T. (1988). Goals in discourse. In D. G. Ellis & W. A. Donohue (Eds. ), ' processes1 (pp. 257-274). Hillsdale, NJ: Erlbaum. Deutsch, M. (1958). Trust and suspicion. Journl_o£_§onflict W 265- 279 Deutsch, M. (1980). The effect of motivational orientation upon trust and suspicion. W 123-139. 213 Deutsch, M. & Krauss, R. M. (1982). Studies of interpersonal bargaining W 52- 78 Donohue. W A (in preSS) Communicatimmjaritaljisputsandjimm Med1a119n1 New Brunswick, NJ: Laurence Earlbaum. Donohue, W. A, Allen, M. & Burrell, N. (1988). Mediator communicative competence. Communieatinn_McnnEraphs1_5§1 102-119. Donohue, W. A., & Diez, M. E., (1985). Directive use in negotiation interaction. Ccmmunicathnmmsfizi 305-318. Donohue, W. A, Diez, M. E. & Hamilton, M. (1984). An expanded model of communication rule use in negotiation. Humen_£emmunieatien 8252mm. 403-425 Donohue, W. A, Diez, M. E, & Stahle, R. (1983). New directions in negotiation research. In R. Bostrom (Ed. ), Cgmmunieatien Xearhegk121 (pp. 249- -279). Beverly Hills, CA. Sage. Donohue, W. A, Diez, M. E, & Weider-Hatfield, D. (1984). Skills for successful bargainer: A valence theory of competent mediation. In R N Bostrom (Ed ) Competencsimmmunicathn (pp. 219- -258). Beverly Hills, CA Sage. Donohue, W. A., Kaufman, G., Smith, R., & Ramesch, C. (1989). Crisis bargaining: A framework for understanding intense conflict. Paper presented at the International Association for Conflict Management conference, Athens, Georgia. (June). Donohue, W. A, Lyles, J, & Rogan, R. G. (1988). Issue development in divorce mediation. Mediatien_QuarLerly1_241 19- 28. Donohue, W. A., Rogan, R. G., Ramesh, C., & Borchgrevink, C. (1990). The role of relational development in hostage negotiation. Paper presented at the Central States Communcation Association, Detroit, Michigan. (March). Dorris, J. W. (1972). Reactions to unconditional cooperation: A field study emphasizing variables neglected in laboratory research. W121 387- 397 Druckman, D. & Bonoma, T. V. (1976). Determinants of bargaining behavior in a bilateral monopoly situation II: Opponent' s concession rate and similarity. Behazigral_591eneee1_211 252-262. FBI Training Seminar. (1987). Michigan State Police Headquarters, Lansing, Michigan. December. Fitzpatrick, M. A. & Winke, J. (1979). Ybu always hurt the one you love: Strategies and tactics in interpersonal conflict. W311 214 Fleiss, J. L. (19881). WW5. New York: Wiley. Folger, J. P., Hewes, D. E. & Poole, M. S. (1984). Coding social interaction. In B. Dervin & M. J. Voigt (Eds. ), Ezogress_1n communicatimciencanoLA. Norwood NJ: ABLEX Publishing Co. Folger. J P 8: Poole. M S . (1984). W commnnication_perspectiye. Glenview, IL: Scott, Foresman & Company. Fuselier, G. D. (1988). A practical overview of hostage negotiations. WWW 1-11. Fuselier, G. D. & Van Zandt, C. R. (1987). A practical overview of hostage negotiations. Unpublished manuscript, SOARU, FBI Academy, Quantico, Virginia. Gahagan, J. P. & Tedeschi, J. T. (1988). Strategy and the credibility of promises in the prisoner 's dilemma game. Journal_of_Confiligt 85551111125142. 224- 234 Garfinke1,H. (1972). Remarks on ethnomethodology. In J. Gumperz & D. Hymes (Eds ), WWW communication (pp. 301- -324). New York: Holt, Rhinehart and Winston. Gibb, J. (1981). Defensive communication. Journal_Q£_Cgmmuniga;igng_2+ 141—148. Goffman, E. (1955). On face-work. Esychiatzx1_l&1 213-231. Goffman, E. (1956). Embarrassment and social organization. Amer1gan W 264- 271. Goffman. E. (1959). WWW. Garden City, NY: Anchor Books, Doubleday and Company, Inc. Goffman. E (1967) Wm M. Garden City, NY: Anchor Books, Doubleday and Company, Inc. Gouldner, A. (1960). The norm of reciprocity: A preliminary, statement. WA 161-178 Gruder, C. L. (1971L Relationships with opponent and partner in mixed-motive bargaining WM 403-416. Guetzkow, H. (1950L Unitizing and categorizing problems in coding qualitative data ClinicaLEthlogLA. 47- 58 215 Guyer, M. & Rapoport, A. A. (1970L Threat in a two-person game. Wmll- -25. Harford, T. & Solomon, L. (1967). "Reformed sinner" and' 'lapsed saint" strategies in the prisoner 's dilemma game. Jnnrnal_n£_Cnn£l1nt Resolutinn.41.104 -.109 Hocker, J. L. & Hilmot, W. R., (1978). Internersnnal_nnn£11ct (2nd ed.). Dubuque, IA: Wm. C. Brown Publishers. Hornstein, H. A. (1985L The effects of different magnitudes of threat upon interpersonal bargaining WW EsxnhnlngL_i.282- -293 Huseman, R. C. (1977L Interpersonal conflict in modern organization. In R. C. Huseman, C. M. Logue, & D. L. Freshley (Eds. ), Bend1ngs inlntnrnersnnaLandflrganizatinnaannnuninatinnwrded) Boston, MA: Hollbrook Press. Infante, D. A. (1975L Effects of opinionated language on communicator image and conferring resistancre to persuasion. Hestern_$peenh Communicatinmja. 112-119 Janis, I. & Mann, L. (1977). Denisinn_naking. New York; Free Press. Johnson, D. W. ,(1967L The use of role reversal in intergroup competition. MW 135-141. Johnson, D. W. (1971L Effects of warmth of interaction, accuracy of understanding, the proposal of compromises on the listener s behavior MW 207- 216 Johnson, D. W. (1973L Communication in conflict situations: A critical review of the research. lnternntinnnlJErnnp_Ien5119n5. 31 48- 65. Johnson, D. W. (1974). Communication and the inducement of cooperative behavior in conflicts: A critical review Sneegh_ngnngranhs1_411 64-78. Jones, T. S. (1988L Phase structures in aggreement and no-aggreemnet mediation. Cnmmn1natinn.Besnarnh.45. 470-495 Kelley, H. H. (1965). Experimental studies of threats in interpersonal negotiations. W 79-105 Katy, S. S. (1970L Neurochemical aspects of emotional behavior. In P. Black (Ed ) Winn. New York: Academic Press. 216 Krauss, R. H. & Deutsch, M. (1966). Communication in interpersonal bargaining WWW 572-577. Lacey, J. I. & Lacey, B. C. (1970). Some autonomic-central nervous system interrelationships. In P. Black, (Ed.). Ehyshlnginal nnzrelates_n£_emntinn1 New York: Academic Press. Lanceley, F. J. (1979). The antisocial personality as a hostage-taker. Unpublished manuscript manuscript, SOARU, FBI Academy, Quantico, Virginia. Lanceley, F. J., Ruple, S. W. & Moss, C. G. (1985). Crisis and suicide intervention. Unpublished manuscript, SOARU, FBI Academy, Quantico, Virginia. Leventhal, B., & Hosbach, P. A. , (1983L The perceptual—motor theory of emotion. In J. T. Cacioppo & R. E. Petty, (Eds. ), Saninl nsxnhnlngL_A_5nurnebnnk. (pp 353- -388) New York: The Guilford Press. Lewis, S. A. & Fry, W. R. (1977) Effects of visual access and orientation on the discovery of integrative bargaining alternatives W 75- 92. Lindskold, S. Tedeschi, J. T. Bonoma, T. V. & Schlenker, B. R. (1971L Reward power and bilateral communication in conflict resolution WW3. 415-418 Loomis, J. L. (1959L Communication, the development of trust, and cooperative behavior. Human_BnlaLinns._lZ. 305- 315. Harwell, G. & Schmitt, D. R. (1967). Dimensions of compliance-gaining behavior: An empirical analysis. Saninmetzy1_afl1 350-384. McGrath, J. E. & Julian, J. W. (1983L Interaction process and task outcomes in experimentally created negotiation groups. lgurnal_nf EmhnlnginaLStudieLJA. 117-138 Michigan State Police (1990). Surviving hostage-taking and abductions. H' 1' SI I E J' N Michigan State Police Training Seminar. (1989). March. Hiller. A. H. (1980). Wanna. Boulder. CO: Westview Press. Hiller, G. R. & Steinberg, M. (1975). anntnrnersnnalmmnunmatinn Chicago: Science Research Associates. 217 Hiron, M. S. & Goldstein, A. P. (1979). Easinge1 New York: Pergamon Press. Hishel, H. H. (1984L Perceived uncertainty and stress in illness. WM, 163— 171. Morley. I & Stephenson. G (1977) MW Barg1n1ng1 London: George Allen and Unwin Ltd. Muir. 14. R., (1977). Wham. Chicago: Chicago University Press. 0' Keefe, B. J. (1988L The logic of message design: Individual differences in reasoning about communication. Communicatinn W 80-103. O'Keefe, B. J, Delia, J. G. (1982L Impression formation and message production. In M. Roloff & C. Berger (Eds. ), Sng1nl_nogn1119n_nnd nnmmuninatinn1 (pp. 33- -72). Beverly Hills, CA: Sage. O'Keefe, B. J. & Shepard, G. J. (1987L The pursuit of multiple objectives in face- to-face persuasive interactions: Effects of construct differentiation on message organization. Cnmmnninatinn MW 396-419 Poole, H. S. (1981). Decision development in small groups: A comparison of two models W45. 1— 20 Pruitt, D. (1971L Indirect communication and the search for agreement in negotiation. MW 205- 239 Putnam, L. L. & Geist, P. (1985L Argument in bargaining: An analysis of the reasoning process W 5Q1 225- 245. Putnam, L. L. & Jones, T. S. (1982L The role of communication in bargaining HumanfinmmuninatimResnarth. 262- 280 Putnam, L. L. & Poole, M. S. (1987L Conflict and negotiation. In F. Jablin, L. Putnam, K. Roberts, & L. Porter (EdsL Handhnnk_n£ WW (pp 549-599) Sage: Newbury Park Putnam, L. L. Wilson, S. R, Waltman, M. S. & Thrner, D. (1986L The evolution of case arguments in teachers' bargaining. Jnnznnl_nf 1W3. 63-81 218 Putnam, L. L. & Wilson, S. (1987). Argumentation and bargaining strategies as discriminators of integrative and distributive outcomes. Paper presented at International Association for Conflict Management conference. George Mason University, Fairfax, VA. (JuneL Rogan, R. G., Donohue, W. A., & Lyles, J. (1989). Empathic perspective-taking as a strategy for gaining and exercising control in hostage negotiation situations. Paper presented at the International Association of Conflict Management conference. Athens, Georgia, (June). Roth, W. T. , (1982). The meaning of stress. In F. M. Ochberg & D. A. Soskis (Eds ). Yintim5_nf_Inrrnrism. (pp 37- -57) Boulder. CO: Westview Press. Rubin. J & Brown B (1975) The_Snnial_E51nhnlngx_nf_Bargaining_and Nagot1a11on1 New York: Academic Press. Saine, T. A. (1974). Perceiving communication conflict. Spaooh Monogranh5._41. 49-58. Satir, V. Con1o1n1_Eam1ly_Ihorany1 Palo Alto: Science and Behavior Books. Schenkein. J (Ed ) (1978) Studie5_in_1he_nrganizatinn_nf nonzersatinnal_intnrantinn. New York: Academic Press Shimanoff, S. B. (1987L Types of emotional disclosures and request compliance between Spouses Cnmmnninatinn_Mnnngranh5._§g. 85- 100 Shure, G. H. Meeker, R. J. & Hansford, E. A. , (1965). The effectiveness of pacifist strategies in bargaining’games. Journal nf_Cnnflint_Basnlutinn._9. 106- 117 Sillars, A. L. , (1980L Attributions and communication in roommate conflicts Cnmmuninatinn_nnnngranh5._41 180- 200. Sillars, A. L., Coletti, S. B., Parry, D., & Rogers, M. A. (1982). Coding verbal conflict tactics: Nonverbal and perceptual correlates of the' 'avoidance—distributive-integrayive" distinction. Human_Cnmmuninatinn_Besearnh._B 83- 95. SOARU. (1974). The political assassin. Unpublished manuscript, FBI Academy. Quantico, Virginia. SOARU. (1985). Hostage negotiation training seminar material. FBI Academy, Quantico, Virginia. SOARU. (1987a). Negotiation/tactical concepts for on-scene commanders. Unpublished manuscript, FBI Academy, Quantico, Virginia. 219 SOARU. (1987b). Abnormal psychology for hostage negotiations. Unpublished manuscript, FBI Academy, Quantico Virginia. Soskis, D. A. & Van Zandt, C. R. (1986). Hostage negotiation: Law enforcements most effective nonlethal weapon. Ibo_EBI_Managomont 911mm 1-8. Spradlin, A. L., Germeroth, D., & Bhargava, C. (1989). Face-work in conflict: A review of the literature. Paper presented at the International Association of Conflict Managment conference. Athens, Georgia, (June). Steinfatt, T. M., Seibold, D. R., & Frye, J. K. (1974). Communication in game simulated conflicts: Two experiments. Sooooh_Monogzaohs1 411 24-35 Strentz, T. (1983). The inadequate personality as a hostage-taker. Unpublished manuscript, FBI Academy, Quantico, Virginia. Strentz, T. (1988). Negotiating with the hostage-taker exhibiting paranoid schizophrenic symptoms Womd WM 12- 18 Swensson, R. E. (1967). Cooperation in prisoner' s dilemma game: I. Effects of asymmetric payoff information and explicit oommunication.BehanoraL£ciemoe.._121 314- 322. Swingle, P. G. & Santi, A. (1972). Communication in nonzero-sum games.Journal_of_Earsoualim_and_Soo1aLEsxoholom._231 54-83 Taylor, R. W. (1983). Hostage and crisis negotiation procedures: Assessing police liability. Iziali_lfli 64-69. Terhune, K. W. (1988). Motives, situation, and interpersonal conflict within prisoner s dilemma. Wail Esxohologyo Monograph Supplement. Theye, L. D. & Seiler, H. J. (1979). Interaction analysis in collective bargaining: An alternative approach to the prediction of negotiated outcomes. In D. Nimmo (Ed. ), Communioation Ioarhook1_31 New Brunswick, NJ: Transcation-International Communication Association. Thibaut, J. W. & Coules, J. (1952). The role of communication in the reduction of interpersonal hostility. Journal_of_Abnormal_and WM 770-777 Tjosvold, D (1974). Threat as low-power person s strategy in bargaining: Social face and tangible outcomes. Intoznational MW 494- 510 220 Tutzauer, F. & Roloff, M. E. (1988). Communication processes leading to integrative agreements: Three paths to joint benefits. W451 360- 380. Watzlawick, P., Beavin, J. H. & Jackson, D. D. (1967). Ezagmatios_of l . I' . E I I E . | I' 1 II DaLthQEiesi_and_EaIade251 New York: U. W. Norton & Company. Wichman, H. (1970). Effects of isolation and communication on cooperation in a two-person game lournaLoLEersonalitLand W 114- 120. Wilson, S. R. & Putnam, L. L. (1990). Interaction goals in negotiation In J Anderson (Ed ).Qommn1oation_1earbook._131 Sage: Newbury Park, CA. Wiseman, R. L. & Schenck-Hamlin,W ., (1981). A multidimensional scaling validation of an inductively—derived set of compliance-gaining strategies Communioationjonographsiifir 251— 270. Zillman,D ., (1971). Excitation transfer in communication-mediated aggressive behavior JoumalmtfimmentaLSooialjsmholom 21 419—434. ' Zillman, D, & Bryant, J. , (1974). Effect of residual excitation on the emotional response to provocation and delayed aggressive behavior. W 782— 791. Zillman, D., (1979). Hostility_and_aggrossion. Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers. Zillman, D., (1983). Transfer of excitation in emotional behavior. In J.T. Cacioppo & R. E. Petty (Eds.), Sooial_osxohology1_A souroobook (pp. 215-240). New York: Guilford Press. WICES 0001 0002 0003 0004 0005 0008 0007 0008 0009 0010 0011 0012 0013 0014 0015 HTOI: PNBO: HTOl: PNBO: HTOl: PNBO: HTOl: PNSO: HTOl: PNSO: HTOl: PNBO: HTOl: PNBO: HTOI: APPENDIXflA AN EZAMELE'TRANSCRIPT HELLO HI JIM? YES. DALE METHOUSE AGAIN. HOW'S IT GOIN? S'ALRIGHT OKAY, WE GOT A CALL OUT FOR VOGEL...UM...HE'S IN THE FIELD SOMEPLACE, WE'RE TRYING TO GET A HOLD OF HIM...UM...WE CHECKED ON MCRACKEN...HE'S IN THE CITY HERE BUT WE'RE NOT SURE EXACTLY WHERE RIGHT NOW BUT I EXPECT TO HEAR BACK...WELL...PROBABLY WITHIN THE NEXT HALF HOUR OR SO EXACTLY WHERE HE IS. THAT'S A LONG TIME. WELL, IT'S NOT THAT LONG. (PAUSE) WHERE DO WE STAND NOW? WHAT DO YA...WHERE DO WE GO FROM HERE? I HAVE...I HAVE TO GO OUT...I HAVE TO TERMINATE MYSELF. I INDUCE NO OTHER WAY OUT. WELL...AS I SAID BEFORE I THINK YOU'RE SELLING YOURSELF SHORT ON THAT. NOBODY SELLS THEMSELF SHORT IF THEY HAVE TO PICK BETWEEN THE TWO. BUT YOU'RE MAKING ASSUMPTIONS THAT YOU ONLY HAVE A FEW CHOICES AND I DON'T THINK THAT'S NECESSARILY THE CASE. WELL IT ISN'T FAIR. I HAVE TO JOIN MELANIE. WELL AS I SAID BEFORE...I'M STILL TRYING TO CHECK ON ON THAT SITUATION TOO. NO SHE'S GONE I HAVE TO JOIN HER. 221 APPENDIXCB PHASE SYNOPSIS OF'REGOTIATIONS Phase Description of Transcript Number One: WCPO Ehaso_Qno. The tape for this interaction begins shortly after the initial contact has been made between negotiator and hostage-taker. The negotiation begins with the negotiator introducing himself, and explaining how he is trying to locate another police officer for the hostage-taker. Throughout this first phase, the hostage-taker expresses his strong desire to commit suicide because of past mistakes, and his apparent involvement in the drug overdose of his girlfriend. The negotiator attempts to minimize the hostage-taker's need to kill himself by arguing how much other people need him. The hostage-taker denies his ability to help anyone, especially from jail. Ehasoolwoo Phase two involves the negotiator asking the hostage-taker fact-based questions regarding his past group membership. The hostage-taker is reluctant to disclose any information, but continues to focus on his desire to commit suicide because of his girlfriend's death. The negotiator offers support by minimizing the finality of the hostage-taker's problems. The hostage-taker calms down a bit and begins to answer the negotiator's questions. The hostage-taker admits to bad drug side effects influencing his behavior. Picking up on this, the negotiator outlines possible contributions that the hostage-taker could make to others by 222 223 drawing on his past drug experience. The negotiator errs by disqualifying all of the hostage—taker’s negative behavior to the side effects. This causes the hostage-taker to become hostile and defensive. EhasoJThroo. The negotiator attempts to refocus the hostage-taker's attention away from the drugs by asking fact-based questions about his possible involvement in revolutionary groups. The hostage-taker begins to open-up to the negotiator by answering his questions. At one point the hostage-taker jokes about the negotiator's inability to locate the other cop. The negotiator refocuses the discussion toward the hostage—taker's past martial arts experience by asking neutral questions. The hostage-taker cooperates by responding freely. EhaS§_Eou21 The conversation continues to be primarily integrative throughout this phase, with the negotiator still focusing his questions on the hostage-taker's martial arts and exercise activities. The hostage-taker responds openly and without hesitation. Interaction is relaxed and cooperative. Ehaso_Eiye1 During the early part of this phase, the interaction continues to be cooperative, friendly, and open with the interactants discussing the hostage-taker's physical exercise. However, half way through the phase, the negotiator switches topics and attempts to draw the hostage-taker back into the present by asking for an explanation of how the hostage-taker got to the present situation. The hostage-taker responds freely, explaining that his revolutionary group's plans simply went awry; due mostly to his own shortcomings. 224 The negotiator responds with support and retroactive face saving messages (e.g., " There are people all around the city who are looking up to you."). 'Ehaso_51x1 The negotiator continues to express opinion-based statements in support of the hostage-taker. Yet, the hostage-taker negates the negotiator's comments as "bull" by arguing for his point from a philosophical revolutionary point of view. The phase ends with the hostage-taker asking for his dog with the negotiator stressing the need for the hostage-taker to surrender Ehaso_$oyon1 The hostage-taker reaffirms his desire for his dog to "go with him.“ The negotiator continues to offer support, while positing that no one would be willing to bring him his dog as long as he's armed. The hostage-taker argues for being trusted due to his consistent behavior in not harming anyone. The hostage-taker refuses the negotiator's efforts to get him to surrender by condemning his own behavior as cause for his death. Ehaso_E1gh11 The negotiator becomes more assertive in trying to get hostage-taker to not be so hard on himself. The hostage-taker believes that the negotiator is coming him. The negotiator expresses his sincerity, emphasizing the need for the hostage-taker to trust him and surrender. The hostage-taker argues that he can trust no one but himself and closes the interaction by killing himself (i.e., "I'm going right..."), as the phone drops to the floor. 225 Phase Description of Transcript Number Two: Majors Ehaso_Qno1 The interaction begins with the negotiator and hostage-taker introducing themselves to each other. The hostage-taker starts right out telling'the negotiator to let him explain how he got into his situation. The hostage-taker begins disclosing that he has a brain tumor, which was diagnosed at a military hospital. The hostage-taker argues that the CIA offered him a deal for a drug overdose to alleviate his pain and suffering in exchange for killing a Soviet spy. He explains that the spy was posing as his father. The negotiator simply responds with simple backchannels and occasional fact-based questions. Phaso_Iwo1 The hostage-taker continues to elaborate about his deal with the CIA. He also begins arguing that he has facts that the Soviets, with the cooperation of his mother, killed his father some forty years previous. The negotiator continues his technique of asking questions for clarification, and mostly allowing the hostage—taker to vent his concerns by using simple neutral backchannels ("ck”). Toward the end of the phase the negotiator suggests that the hostage—taker meet with the CIA, and that he can work out such a.meeting for him. The hostage-taker takes hold of this proposal by expressing a willingness to surrender in order to meet with the CIA. Ehaso_Ihroo1, The hostage-taker begins to make suggestions for how he could surrender, issuing a demand that the police do not use handcuffs. The negotiator responds to the hostage-taker's proposal by discussing possible surrender details. However, in the process of 226 working out the details, the hostage-taker lapses back into an explanation of his brain tumor, and how the CIA offered him a deal. The hostage-taker is extremely talkative and open throughout, with only momentary exclamations of demands. The negotiator attempts to refocus the hostage-taker's attention toward the surrender plan, while offering succor and comfort. Ehase_Eon21 The hostage-taker expresses sincere willingness to work with the negotiator to resolve his surrender plan. However, the hostage-taker again drifts off into discussing his current physical state; that he is sitting in his own excrement, covered with gas, and almost out of cigarettes. The negotiator conveys concern that he not smoke, but it becomes evident that the hostage-taker was simply toying with the negotiator. The hostage-taker continues to elaborate about his tumor, his deal with the CIA, and his past military experience. The negotiator maintains an approach of listening and offering support. Ehase_21y21 Throughout this phase, the negotiator struggles to get the hostage-taker to refocus his attention on the proposed surrender plan. However, the hostage—taker continually shifts topics, focusing first on the death of his father, his tumor, and his involvement with the local police in drug busts. The negotiator spends most of his time simply asking questions and offering support. Phase_51x1 The hostage-taker offers the negotiator information to verify his story. The negotiator finally refocuses the hostage-taker's attention toward the surrender plan. The hostage—taker again proposes a plan, identifying concessions he is willing to make 227 to satisfy the negotiator. He asks for handcuffs and the negotiator's guidelines. He is extremely cooperative. The negotiator identifies the procedure, promising to take the hostage-taker to the hospital. Ehaso_$oyon1 The hostage-taker takes unsolicited moves to discard some of his weapons. The negotiator reigns him in a bit, instructing him to wait for authorization before making any moves. The hostage-taker complies and the negotiator makes arrangements for how handcuffs will be delivered to the hostage-taker and how the hostage-taker will surrender his weapons. Phaso_E1ght1 The negotiator spends most of his time instructing the hostage-taker how to surrender so that nothing bad happens. The hostage-taker is extremely cooperative throughout this phases. He answers all of the negotiator's questions and follows all directives without reservation. In fact, the negotiator has to exert control over the hostage-taker's enthusiasm. The interaction ends with the hostage-taker surrendering peacefully. Phase Description of Transcript Number Three: MPD Phase_Qn21 Throughout this first phase, the hostage-taker demands to speak with his wife and a local television news reporter. The negotiator agrees to contact the news reporter and to look for the hostage-taker's wife. The hostage-taker is panicky about his apartment being assaulted by the police. He threatens that "something will happen" if anyone attempts to get in. The negotiator maintains a position of cooperation and nonagressiveness. The hostage-taker then begins to question the negotiator about who he is and his position. 228 The negotiator responds openly. The hostage-taker then begins to explain that his wife left him because he cooperated with the police in a.drug bust. The negotiator offers support by primarily listening. The hostage-taker continues to de extremely defensive and even hostile toward the negotiator. While the negotiator begins the interaction as cooperative and accommodating, he now tells the hostage-taker that he must cooperate if he wants the police to help. The hostage-taker responds by crying and begging for help to find his wife. The negotiator promises to try and find her, and asks the hostage-taker questions about her toward this end. Ehase_IH91 The hostage-taker elaborates on his involvement in a drug bust. He explains that when his wife find out about it she left him. He posits that he was threatened by the drug dealers, and that the police whom he assisted failed to protect him. The negotiator expresses sympathy and understanding, and asks the hostage—taker what he thinks should be to resolve the situation and to satisfy his wants and needs. the hostage—taker gets defensive and aggressive, expressing anger and frustration toward his wife and the police. The negotiator becomes assertive, asking the hostage-taker specific questions and requesting his cooperation. The hostage-taker starts crying and whining again that he is "already dead" and that he has no way out. The negotiator becomes directive, explaining to the hostage-taker that he does have positive options. Ehase_Ihreel During this phase, the hostage-taker thinks he's going to be left alone and that he'll be killed by drug dealers. He begins crying that he is scared. The negotiator offers support by 229 saying that they are "wrapped in together" in trying to solve this problem. The bond of rapport begins to build as the hostage-taker asks the negotiator person-specific questions (who he is, what he does). Yet, the hostage-taker feels hopeless and offers to commit suicide to solve the crisis. The negotiator negates this idea and suggests that the hostage-taker begin his life again in another city. The hostage-taker continues to cry about his plight and the negotiator continues to offer support. However, the negotiator's support is more directive in nature than passive. Ehaso_Eon;1 The hostage-taker and negotiator discuss issues openly now. There is an air of trust and rapport. The hostage-taker continues to request to speak with his wife. The police are unsuccessful in their efforts and the negotiator must convey this information. the hostage-taker remains anxious, threatening to kill himself if he hears any noises. The negotiator continues to solicit information to help locate the wife. The hostage-taker expresses a keen interest in the negotiator's plan for him to start life over in a new city, and asks for how to do it. The negotiator stresses his willingness to help and cooperate, but that hostage-taker must also cooperate. The hostage—taker praises the negotiator as "A—l.“ The hostage-taker begins to cry again about his plight. The negotiator offers suggestions about how the hostage-taker could surrender. A neighbor woman friend calls the hostage-taker and offers to take the hostage-taker's children (i.e., the hostages). Ehaso_Eiyo1 The hostage-taker becomes jumpy hearing noises of people being evacuated, and threatens that "something will happen 230 unless the noise stops." The negotiator asks hostage-taker for additional information that might help them locate his wife. The effort has been unsuccessful up to this point. 'The hostage-taker makes a sudden gesture to step out onto his balcony to get rid of his gun. The negotiator instructs the hostage-taker to wait for him to prepare things. In the interim, the hostage~taker's female neighbor meets with the negotiator who allows her to talk with the hostage-taker. After talking with her the hostage-taker expresses a willingness to cooperate and to meet with the negotiator and neighbor. However, the hostage-taker still requests to see his wife. The negotiator states that they have not found her yet, and refocuses the discussion toward how the hostage-taker will surrender. Ehaso_$1x1 The negotiator proposes the plan for the hostage-taker's surrender and release of the hostages. He stresses the trust and honesty of their relationship to encourage the hostage-taker's compliance. The negotiator emphasizes that the plan is the hostage-taker's only good Option. The neighbor talks with the hostage-taker again, persuading him to cooperate with the police. The negotiator renews interaction by outlining the plan for surrender. The hostage—taker passively agrees to the plan, but expresses hesitancy. He argues that he must prepare for it, and demands to see his wife again. The negotiator argues that while she has not been found, she will be waiting for him at the precinct. Ehaso_$oyon1 In this phase, the hostage-taker begins to fall apart again. He starts crying and begging for help. He agrees to the surrender plan. The negotiator stresses that the hostage-taker is a 231 good guy, and that he is trying to help. The negotiator continues to lay out the surrender plans and the hostage-taker again agrees. However, he stalls for time to prepare. The first negotiator leaves the phone during'the middle of this interaction to meet the hostage-taker. A second negotiator renews the phone dialogue. This second negotiator stresses that the hostage-taker get rid of his gun and that he surrender. The hostage-taker balks. Ehase_Eigh11 the negotiator keeps trying to persuade the hostage-taker to surrender, emphasizing police willingness to cooperate and help him. The hostage-taker insists that he have prove and that he needs to talk with reporters and his wife before surrendering. The negotiator indicates that his wife has not been located, but that she will meet him at the precinct. The hostage-taker begins to balk more, refusing to meet the first negotiator and neighbor as he had originally indicated that he would. The negotiator works at attempting to persuade him to surrender, but the hostage-taker refuses. He panics after realizing that the police are outside his door waiting for him. He becomes hysterical, yelling for the police to leave him alone and for them to find his wife. The tape ends with the hostage-taker yelling for his wife. APPENDIXCC IDENTITTLGOAL.CODING SCHEME 0. No Identity goal in utterance "hello" ”good-bye" Face-Attack 1. Face Attack—Other "you're a liar" "you're just jerking me around" "you're no good" 2. Face Attack-Self "I'm no good" "I just keep screwing up my life” Face-Defend (proactive face-saving) 3. Face Defend-Self "I know that I'm doing the right thing" "This is my only option; I have no other choice“ 4. Face Defend-Other "Let me just say, I think you're doing a great job" "Don't take this the wrong way, but" "You're not a bad person" "Yes you're right, that is important.” Face-Restore (retroactive face-support) 5. Face Restore-Self "I'm not as dumb as you think" "Don't think I don't know what you're doing, I'm no idiot" 6. Face Restore-Other "Don't be so hard on yourself" "That's not a big deal, we can take take of it” Conversational Features 7. Positive Backchannels ..Y811.. "Uhrhuh" Ilokayll 8. Negative Backchannels IINahll "unh—uh" 232 H» APPENDIXID RELATIONALAGOAL CODING SCHEME . No Issue goal expressed in utterance (Blank uttterance; crying) . Integrative . Willingness to integrate/comply with other's demands/wants “I'm trying to work with you, but you need to work with me." "I'll let the woman go if you get me some beer and cigarettes." "Okay, I'll give you what you want." "I'm letting the woman go." . Questions of disclosure: self-based feeling & thinking (excluding suicidal statments) ”How ya doing?" "Why did you stop going?" "How did you feel about that?” . Questions of fact about person: self-based "How long did you live there?" "Do you know ?" "How many people were in your group?" . Questions of fact about hostage-taking situation "How are the hostages?" "How many people are in there with you?" ”Where's my brother?" . Statements of disclosure: self-based feeling and thinking ”I'm scared." "I don't think I can take much more." “I refuse to put up with his/her shit." . Statements of fact about personzself-based “I lived there for four years." "I just lost my job." "I had to leave because of the economy." . Statements about hostage-taking situation "The hostages are doin' fine.” "You've got fifteen cars around the joint." . Statements of support and comfort "Ybu're not a bad prson." "You're doin' a good job." "You're not that bad off." 233 234 9. Question of personal wants and needs (demands) "So where do we go from here?" "So what is it you want?" "What can I do for you?" B. Distributive 10. Expressing personal demands and wants 11. 12. 13. 14. 15. "I want to talk to my cousin." "I want to get out of here." "I want a car and $500,000." Topic shift*** Changing topic of discussion from that in previous utterance "I don't want to talk about that right now" "I told you that that's a mute subject." Avoidance of personal responsiblity*** "It's on your head if the hostages die." "You're the one who controls this thing, not me." "It's not my fault." Reject other's demands and proposals "1 am not going to do that" "I can't give you a car" Threat to take action against others "If you don't give me what I want I'll kill a hostage." "I blow the plane at 1:30” Threats against self “I have to kill myself." "I have to pay for my mistakes.” C. Structural and Conversational Features 18. 17. Positive Backchannels IIOkayIO IOUMUhOO "Alright" IIUhUhII IOUMOI Negative Backchannels "N811" "Unh—uh" 235 18. Initiations and salutations "Hello" IIHiOI ***Note. Relational-Goal coding category #11 was dropped from the final coding scheme because it was not mutual exclusive from the other coding categories. Coding category #12 failed to receive any codes during the transcript coding procedure. APPENDIXCE BASIC’LANGUAGE INTENSITY’CODING SCHEME 1. Obscure words: Instances in speech which involve the use of "uncommon" words or phrases. This includes forgein words, polysyablic words, and rarely used words. For example, "lascivious" rather than lustful; “coup d'etat" rather than a rebellion. 2. General metaphors: This category involves words or phrases in which the denotative meaning of a term is used in a manner other than it is conventionally associated. For example: "I feel like I'm sinking," and "My life is a roller coaster." 3. Qualifying adjectives or adverbs: This includes words which are intended to modify other words by accenting the meaning of the qulaified word. For example " an incredibly attractive women" rather than " an attractive women." This category includes both Simple Qualifiers and Elaborate Qualifiers described in Apeendix E1. 4. Profanity and sex-based statements: This category includes the use of words to convey profanity. A variety of word types and phrases fit into this category. For example: "bullshit," “piss-off,” and "asshole." 5. Death-based statements: This category includes words or phrases which involve the use of death-focused statements. For example: "I 'm going to waste him," "I have to end it," and "I'll kill her." 236 mm 31 SDI’LE mm W m (IBM m I . SIMPLE QUALIFIERS WING PKXIEDURE Code words or phrases which act as intensifiers to adjectives and adverbs. This includes the use of words or phrases which are used to modify other words by accenting the meaning of the qualified word. For example, "incredibly" would be coded as a qualifier for the statement “an incredibly attractive women.” The word "attractive" would not be coded. Key to coding using this system: look for words ending in "y" or “1y." II . ELABORATE QUALIFIERS CODING PMCEDURE There are several categories to the Elaborate qualifiers scheme, these include: adjectives and adverbs, auxillary verbs, affirmations, negations, contractions, pronouns with contractions, and certain "special" words. A. Adjectives and Adverbs Code all adjectives and adverbs. For examle, code ”beautiful" as one qualifier, not just "incredibly" in the state "incredibly beautiful." B. Auxillary verbs Code such words as: will should done do night could can have an had does would are is C. Auxillary verbs with negative contractions Code such things as: won't shouldn't can't wouldn't couldn't don't haven't aren't hadn't isn't doem't hasn't D. Affirmations and Negations Code the following: yes no yah yeah not nah B. Pronouns with contractions Code the following: I'm They're He'll I've You've That's He's It's There's I'll They'll She's We've 237 238 F. Special words Code the following special words which describe the quality of something or someone: all anything nothing too also whatever nobody everybody somebody anybody now right right now initially first APPENDIXLF IEXAMELE IDENTITYFGOAL1000ING SCHEME VALIDATION INSTRUMENT INSTRUCTIONS; Please rate each of the following statements in terms of how ”you perceive it to be. By 'self—supportive" we mean how much the statement attempts to protect or proactively endorse a person's own position. Circle the one number which most clearly represents your perception. Very Somewhat Neutral Not Not At All Self Self Self Self Supportive Supportive Supportive Supportive 1 2 3 4 5 01. "The problems you've had are due to that, so write them off" 1 2 3 4 5 02. "Ybu're lying to me“ ———————— 1 2 3 4 5 03. "I'd like to hear your side of it" --------- 1 2 3 4 5 04. "Nobody sells themself short if they have to pick between the two" 1 2 3 4 5 05. "I have nothing, I can't even terminate myself" 1 2 3 4 5 06. "Ya 'all up there messin with me" ~~~~~~~~~~ 1 2 3 4 5 07. "I'm not enough to handle more than one thing at a time" 1 2 3 4 5 08. "Ybu cops are all the same" --- 1 2 3 4 5 09. "I'm just trying to keep em from.giving peOple the wrong kinda drugs" 1 2 3 4 5 10. "If you go, who's going to help them?" ----- 1 2 3 4 5 11. "Let me explain what happened" 1 2 3 4 5 12. "I have no alternative; you have lots of alternatives" 1 2 3 4 5 13. "But, you've got a lot of people with a strong feeling for you“ 1 2 3 4 5 239 240 Very Somewhat Neutral Not Not At All Self Self Self Self Supportive Supportive Supportive Supportive 1 2 3 4 5 14. "I don't want to be taken in" 1 2 3 4 5 15. "Ybu're kind of a domestic guy, knowing what to do with a baby" 1 2 3 4 5 16. "They promised me a hot shot if I put him down" 1 2 3 4 5 17. "You're a smart guy not to drink coffee” —-- 1 2 3 4 5 18. "I've messed things up" 1 2 3 4 5 19. "You sound like a pretty reasonable guy" --- 1 2 3 4 5 20. "All I want is my wife" 1 2 3 4 5 21. "Ybu're just jerking me around” 1 2 3 4 5 22. "I'm not that crazy" ------ 1 2 3 4 5 23. "You sound like an A—l cop” ---- 1 2 3 4 5 24. "They don't need me" 1 2 3 4 5 25. "That's bullshit. I wil not accept that" -—- 1 2 3 4 5 28. "Let him come and go to another reality with me" --— 1 2 3 4 5 27. "As I said, I think you're selling yourself short” 1 2 3 4 5 28. "I've fowled it up" 1 2 3 4 5 29. "There ain't nothing that can't be corrected" 1 2 3 4 5 30. "Ycu're just like all the other bums" ------ 1 2 3 4 5 APPENDIXIG' IEXAMPLE RELETIOMALAGOIL.CODING SCHEME VALIDATION’INSTRUMENT INSTRUCTIONS; Please rate each of the following statements in terms of how “oompetitige" you perceive it to be. By competitive, we mean how much the statement reflects a desire to satisfy one 's own interests, as well as an unwillingness to cooperate. Circle the one number which must clearly represents your perception. Not At All Not Neutral Somewhat Very Competitive Competitive Competitive Competitive 1 2 3 4 5 1. "I'm gonna have to pay for that" l 2 3 4 5 2 "Let's set something up to help you out” ----- 1 2 3 4 5 3. "Why did you give up on that?” 1 2 3 4 5 4. "I ain't going to give up my weapons" ------- 1 2 3 4 5 5 "You're a smart guy not to drink coffee" --—- 1 2 3 4 5 6 "How come ya don't drink coffee?” ----------- 1 2 3 4 5 7 “If I hear any noises, I'm telling you, it's gonna happen" 1 2 3 4 5 8. "Let's work it out" 1 2 3 4 5 9. "Leave me alone or it's going down" --------- 1 2 3 4 5 10. ”What do you want us to do?" 1 2 3 4 5 11. "Have you ever been arrested before?" ------ 1 2 3 4 5 12. "Are ya doing okay?" 1 2 3 4 5 13. "Where's my wife?" 1 2 3 4 5 14. "That's bullshit, I can't accept that" ----- 1 2 3 4 5 15. "Where is he now?" 1 2 3 4 5 18. "When did you first find this out?" -------- 1 2 3 4 5 241 242 Not At All Not Neutral Somewhat Very Competitive Competitive Competitive Competitive 1 2 3 4 5 17. "That's only one time, I don't think you can say that's consistency" 2 3 4 5 18. "Okay, what is it you want?" 2 3 4 5 19. "She's gone, I have to join her" 2 3 4 5 20. "I want you to leave me alone" 2 3 4 5 21. "Ybu held a pretty important position" ---—- 2 3 4 5 22. "Yah, I worked out there" 2 3 4 5 23. "How's it going?" 2 3 4 5 24. "I'm scared to death" 2 3 4 5 25. "I have to go, I have to terminate myself" 2 3 4 5 26. “I have brain tumor cancer" - 2 3 4 5 27. "Is there anybody else up there with ya?" -- 2 3 4 5 28. "You're wife's not at that address” -------- 2 3 4 5 29. "I've been into it since the 50's“ --------- 2 3 4 5 30. "We can help ya, but you're gonna hafta work with us" 2 3 4 5 31. "We're still tryin' to locate him" --------- 2 3 4 5 32. "I'm gonna hurt you all" 2 3 4 5 33. "I don't trust doctors 'cause they always want to do things to ya" 2 3 4 4 34. "The clothes are in the trunk" 2 3 4 5 35. "What kind of weapons do you have?” -------- 2 3 4 5 36. "Do you want your mom to come get the baby?" 2 3 4 5 37. "Let's talk about how to do it so we don't have any mistakes" 2 3 4 5 38. "Well, you're really not going to be in hot water" 2 3 4 5 243 Not At All Not Neutral Somewhat Very Competitive Competitive Competitive Competitive 1 2 3 4 5 39. "I don't drink coffee" -- - 1 2 3 4 5 40. "How are the children doing?" 1 2 3 4 5 41. "I want my dog in here with me" 1 2 3 4 5 42. "I don't think you want to do that" -------- 1 2 3 4 5 43. "I know it's rough" 1 2 3 4 5 44. "Make 'em leave or I'll blow their brains out" 1 2 3 4 5 45. "I wanna talk to my wife, now” 1 2 3 4 5 48. "When were you last in San Francisco” ------ 1 2 3 4 5 47. "Well, I think we have a little problem with that" - l 2 3 4 5 48. ”I think I would trust him" ——— 1 2 3 4 5 49. "How would you like to do this?" ----------- 1 2 3 4 5 50. "The kids are doin' okay" 1 2 3 4 5 51. "I'm going right ..... ” - ----- 1 2 3 4 5 52. "Get those news vultures outta here” ------- 1 2 3 4 5 APPENDIXIH LANGUAGE INTENSITY’CODING SCHEME VALIDATION INSTRUMENT INSTRUCTIONS; Please rate each of the following statements in terms of how 1ntonsn you perceive it to be. By "intense," we mean how much the statement expresses a position (i.e. opinion) which deviates from neutral. Circle the number for the category which most clearly represents your perception. No Slight Moderate High Extremely High Intensity Intensity Intensity Intensity Intensitx 0 1 2 3 4 01. "But look how strong he is even though he's.." 0 1 2 3 4 02. "I'm going right.(now)..” 0 1 2 3 4 03. "Nobody sells themself short if they have to pick between the two" 0 1 2 3 4 04. "Nothing, there's nothing that can't be corrected, nothing” ------------------------- 0 1 2 3 4 05. "Bad, very bad" -- 0 1 2 3 4 06. "Exactly” 0 1 2 3 4 07. "Well, as I said before, I think your selling yourself short on that." 0 1 2 3 4 08. "No. He's in good shape." 0 1 2 3 4 09. "To kill C. J. Moran." 0 1 2 3 4 10. "Hello." 0 1 2 3 4 11. ”And she's been walking the streets for 28 years with that son-of-aebitch and he's been . calling himself my father." 0 1 2 3 4 12. "It's very simple." 0 1 2 3 4 13. "Hang on, hang on." 0 1 2 3 4 14. "Let me explain what happened." 0 1 2 3 4 244 245 No Slmght Moderate High Extremely High 0 1 2 3 4 15. "Don't give me that bull, fighter for the cause. They think I'm a mad man" 0 1 2 3 4 16. "No, not at all, our plan was much different than this.” 0 1 2 3 4 17. "One hotshot." 0 1 2 3 4 18. "There isn't much time left for me." --------- 0 1 2 3 4 19. "Now I'm not no brave man or nuthin like that.” 0 1 2 3 4 20. "Actually, actually, actually the tear gas cleared up my sinuses." 0 1 2 3 4 21. "But we would like for you to get rid of your revolver." 0 1 2 3 4 22. ”Dead.“ —-- 0 1 2 3 4 23. "And then they promised me a hotshot, enough poison to down a bull elephant if I put down, I'm talking to you man.” --- 0 1 2 3 4 24. "I'll be, I'll hang on here." — -- 0 1 2 3 4 25. "very politely, he can see me right in the mirror." 0 1 2 3 4 26. "I'm almost ready." -- 0 1 2 3 4 27. ”I told you, he thinks he's God." ------------ 0 1 2 3 4 28. "Shit.“ 0 1 2 3 4 29. "No sir, I just told you my job is to have everbody come out of this building safe, without anybody getting hurt." 0 1 2 3 4 30. "You all shot me down out there." 0 1 2 3 4 31. "0k." 0 1 2 3 4 32. "No, you're not going to be in trouble." ----- 0 1 2 3 4 33. "YOu know who Bill Jones is, the news reporter" 0 1 2 3 4 34. "They already have her." 0 1 2 3 4 246 No Slight Moderate High Extremely High 0 1 2 3 4 35. "Yah. That's right, yah." 0 1 2 3 4 36 . "No, you don't want to do that." 0 1 2 3 4 37. "Cause she's the one who called." 0 1 2 3 4 38. "I don't want..." 0 1 2 3 4 39. "Alright." 0 1 2 3 4 40. "No, I don't wanna hear that." 0 1 2 3 4 41. "It's gonna happen." ------- - 0 1 2 3 4 42. "0k, and he's in narcotics." 0 1 2 3 4 43. "No, no, no, no." -- — 0 1 2 3 4 44. "0k, go ahead." 0 1 2 3 4 45. "Yah, have him call me, hear." 0 1 2 3 4 46. "I don't have no win." 0 1 2 3 4 47. "Switched completely to boxing." ------------- 0 1 2 3 4 48. "What I'm saying is I don't have any more information for you." —- 0 1 2 3 4 49. "I've fowled it up." 0 1 2 3 4 50. "No, no you're not, no sir." — 0 1 2 3 4 TABLES Table 1 247 Win: Message Focus Behavior Type Other-Directed Self—Directed Attack: "You're a fool" "I'm no good" "You're a jerk" "I'm a failure" Defend: “I'm not sure, "Don't take this (proactive) but...." personally.." "I could be wrong "Before we begin I about this,..." want to tell you how special you are to me." Restore: "Don't treat me like "Don't be so hard on (retroactive) that." yourself." "Where do you get off saying that to me?" "Don't be so concerned about that." 248 Table 2 II |.| G 1 C | H I SI 1 1 D . I' E II |.| G 1 C 1° 3 l y 1'! l' H Identity-Goal Coding Categories Other Self Self Other ' Self Other Measures Hostile Hostile Support Support Defend Defend Other 1.85 3.46 2.96 3.62 2.54 2.98 Hostile (.55) (.75) (.49) (.67) (.49) (.59) (N=38) Self 2.07 2.13 3.01 3.61 2.53 3.30 Hostile (.75) (.85) (.75) (.46) (.59) (.70) (N=45) Self 3.12 3.65 2.60 2.80 2.77 2.73 Support (.73) (.93) (.65) (.77) (.60) (.59) (N=38) Other 4.19 3.78 3.02 2.14 3.37 2.44 Support (.70) (.69) (.44) (.51) (.49) (.48) (N=44) Self 2.29 2.86 2.55 3.57 2.32 2.97 Defend (.73) (.89) (.44) (.59) (.46) (.53) (N=40) Other 3.13 3.21 2.97 2.80 3.06 2.61 Defend (1.13) (.85) (.59) (.73) (.71) (.65) (N=38) Note. A low mean score denotes a high level of agreement between the coding category and the individual validation measure. Standard deviations are presented in parentheses. 249 Table 3 a or... a" .0: ; ‘f. "-. :0. :ga.qo 0‘ - a. o G I II I'I G J C l' S 1 K 1.! I' H Grouped Identity-Goal Coding Categories Grouped Measures Hostile Support Defend Hostile 2.40 3.30 2.83 (N=83) (.55) (.42) (.49) Support 3.71 2.63 2.83 (N=82) (.64) (.36) (.36) Defend 2.86 2.97 2.74 (N278) (.75) (.39) (.47) Note. A low mean score denotes a high level of agreement between the coding category and the individual validation measure. Standard deviations are presented in parentheses. 250 Table 4 0 I I Eaiaf1onaI:goaI_EaaegozéLMeansflagdiSfandafid_Deyiatinns_fnr Validation Measures Relational-Goal Coding Categories Cooperative Competitive Individual Mutual (N=50) (N=44) (N=52) (N=52) Willingness to 4.50 2.61 2.92 4.59 Comply .(.37) (.80) (.86) (.37) Question of 3.30 2.65 2.85 3.25 Disclosure (.39) (.61) (.62) (.56) Question Fact 3.17 2.90 2.77 3.03 About Person (.31) (.51) (.71) (.51) Question Fact 3.19 2.87 3.02 3.13 About Hostage-Taking (.42) (.66) (.62) (.53) Statement of 2.87 3.18 3.62 2.75 Disclosure (.41) (.49) (.63) (.47) Statement Fact 3.01 2.94 3.47 2.67 About Person (.40) (.66) (.76) (.67) Statement Fact 3.47 2.60 2.58 3.23 About Hostate—Taking (.43) (.60) (.60) (.46) Support/Comfort 3.30 2.69 2.96 3.22 (.36) (.53) (.66) (.56) Question of Personal 3.88 2.60 2.90 3.96 Wants/Demands (.44) (.67) (.75) (.45) Statement of Wants 2.20 3.98 4.02 2.31 and Demands (.73) (.77) (.76) (.76) Reject Other's 2.16 4.18 3.86 2.14 Demands (.64) (.52) (.70) (.57) Threat Against 1.82 4.45 3.94 1.76 Other (.85) (.63) (.90) (.81) Threat Against Self 2.80 3.36 3.84 2.45 (.58) (.63) (.73) (.47) Note. A low mean score denotes a high level of agreement between the coding category and the individual validation measure. Standard deviations are presented in parentheses. 251 Table 5 a ‘2: . :. ;'o 0. .a . r :' ; 0|: ‘0. : ’__o u‘:' :.o ;..:‘O 0‘ .. a. '0 {r ; 0.; ‘0; 0.0.: "u‘ z a. 0. Measures Validation Measures Grouped Relational-Goal Coding Categories Cooperative Competitive Individual Mutual (N=50) (N=44) (N=52) (N=52) Integrative 3.40 2.78 3.01 3.31 (.20) (.37) (.35) .32 Distributive 2.25 3.99 3.91 2.16 (.55) (.46) (.62) (.54) NoLo. A low mean score denotes a high level of agreement between the coding category and the individual validation measure. tandard deviations are presented in parentheses. Table 6 my ‘ 91: . ° :00 :00 : - 91. ° ' :02 :" I ’0 . .00: I‘u‘ v‘30 :00 :n-u .‘ . m o 0‘ :0'4 1:: c ‘0 ' 0100'. SchemelalidatiomMeasure Validation Measure Language Intensity Levels Language Intensity No Slmght Moderate High Coding Scheme Intensity Intensity Intensity Intensity Intensity Simple 1.67 2.27 2.10 2.25 Qualifier (.51) (.52) (.49) (.67) Elaborate .94 1.90 2.13 2.08 Qualifier (.61) (.53) (.48) (.48) Note. A low mean score denotes a high level of agreement between the coding category and the individual validation measure. Standard deviations are presented in parentheses. Table 7 :3 -. a \--o .- o u r- : o ; :go 0. o ‘ ‘r a. ‘0. W25 TRANSCRIPTS Relational-Goal 'Coding Categories WCPO MAJORS MPD TOTAL Integrative 133 169 445 747 Distributive O 18 27 45 Other 1 61 55 117 Table 8 TRANSCRIPTS Relational-Goal --- ------------------ Coding Categories WCPO MAJORS MPD TOTAL Integrative --— 120 _ 195 305 620 ----- Distributive 11 25 61 97 Other 4 24 160 188 253 Table 9 a0 ‘0 ‘\'-o '. o i- - 00: o: {"1000 ‘ ‘1‘ 4 W5 Relational—Goal Cumulative Coding Categories Frequency . Percentage No Relational Goal (0) 27 3.0 Willingness to Comply (1) 84 9.3 Question of Disclosure (2) 98 10.8 Question Fact About 167 18.4 Person (3) Question Fact About 12 1.3 Hostage-Taking (4) Statement of Disclosure (5) 187 20.6 Statement Fact About Person (6) 114 12.6 Statement Fact About Hostate-Taking (7) 18 2.0 Support/Comfort (8) 21 2.3 Question of Personal Wants/Demands (9) 14 1.5 Statement of Wants and Demands (10) 40 4.4 Reject Other's Demands (13) 6 0.7 Threat Against Other (14) 0 0.0 'Threat Against Self (15) O 0.0 Positive Backchannel (16) 113 12.4 Negative Backchannel (17) 2 0.2 6 Salutation (18) 5 0. Table 10 :3 :3 a .- :-‘- z.- 0': :‘WDIO ‘ .A" I‘ ‘ foLAllJ‘ransoripts Relational-Goal Cumulative Coding Categories Frequency . Percentage No Relational Goal (0) 39 4.3 Willingness to Comply (1) 13 1.4 Question of Disclosure (2) 57 6.3 Question Fact About 43 4.8 Person (3) Question Fact About 12 1.3 Hostage-Taking (4) Statement of Disclosure (5) 266 29.4 Statement Fact About Person (6) 167 18.5 Statement Fact About Hostate-Taking (7) 21 2.3 Support/Comfort (8) 0 0.0 Question of Personal Wants/Demands (9) 7 0.8 Statement of Wants and Demands (10) 77 8.5 Reject Other's Demands (13) 4 0.4 Threat Against Other (14) 4 0.4 Threat Against Self (15) 11 1.2 Positive Backchannel (16) 174 19.2 Negative Backchannel (17) 7 0.8 3 0.3 Salutation (18) 255 Table 11 _ . WWII. JSJCI C! E III 'IlEl RELATIONAL-GOAL CATEGORIES Phases Integrative Distributive Other Total 1 73 19 20 112 2 91 11 12 114 3 83 10 21 114 4 85 6 22 113 5 76 7 30 113 6 79 15 22 116 7 63 10 45 117 8 70 19 19 108 Table 12 RELATIONAL-GOAL CATEGORIES Phases Integrative Distributive Other Total 1 99 1 14 114 ——_ 2 87 3 24 114 3 93 2 18 113 4 94 2 26 112 5 96 5 13 114 6 94 9 7 110 7 92 12 10 114 8 92 12 10 114 Table 13 :3 -. I \“I .- I o ‘- : O - ;.I I. .0 . - K: ; I0: ‘0: W RELATIONAL-GOAL CATEGORIES Phases Integrative Distributive Other Total 1 17 0 0 17 2 l7 0 0 17 3 16 0 1 17 4 17 0 0 17 5 17 0 0 17 6 17 0 0 17 7 17 0 0 17 8 15 O 0 17 Table 14 RELATIONAL-GOAL CATEGORIES Phases Integrative Distributive Other Total 1 14 1 2 17 2 14 3 0 17 3 16 0 1 17 4 17 0 0 17 5 16 0 1 17 6 15 2 0 17 7 13 4 0 17 8 15 1 O 17 257 Table 15 E Ell I'I BJI' 1318 12] E W PHASES Relational-Goal Coding Categories 1 2 3 4 5 6 7 8 No Relational Goal (0) 0 1 0 0 O 1 1 0 Willingness to Comply (1) 0 0 0 0 0 0 2 4 Question of Disclosure (2) 3 4 2 10 4 3 4 0 Question Fact About 2 4 12 5 7 2 1 0 Person (3) Question Fact About 0 0 0 0 0 0 0 0 Hostage-Taking (4) Statement of Disclosure (5) 4 6 0 1 4 8 6 9 Statement Fact About Person (6) 2 0 2 1 0 0 0 1 Statement Fact About Hostate-Taking (7) 1 0 0 0 0 0 0 0 Support/Comfort (8) 4 2 O 0 2 1 3 1 Question of Personal Wants/Demands ( 9) 1 0 0 0 0 2 0 0 Statement of Wants and Demands (10) 0 0 0 0 O 0 0 0 Reject Other's Demands (13) 0 0 0 0 0 0 0 0 Threat Against Other (14) 0 0 0 0 0 0 O O Threat Against Self (15) 0 0 0 0 0 0 0 0 Positive Backchannel (16) 0 0 1 0 0 0 0 0 Negative Backchannel (17) 0 0 0 0 0 0 0 0 Salutation (18) 0 0 0 O 0 0 0 0 258 Table 16 :3 -. I' .I :-:- ;.- :- ; 0.: .I; :z‘iOo. - I 03 I Warm PHASES Relational-Goal Coding Categories 1 2 3 4 5 6 7 8 No Relational Goal (0) 1 1 0 O 1 1 1 2 Willingness to Comply (1) 0 0 0 0 O 0 O 0 Question of Disclosure (2) 0 0 0 O 0 0 0 1 Question Fact About 0 0 0 0 0 0 O 0 Person (3) Question Fact About 0 O 2 0 0 0 O 0 Hostage-Taking (4) Statement of Disclosure (5) 13 11 3 11 9 13 11 12 Statement Fact About Person (6) 0 2 11 6 6 1 1 0 Statement Fact About Hostate-Taking (7) 0 0 0 0 0 O 0 0 Support/Comfort (8) O 0 0 0 0 0 O 0 Question of Personal Wants/Demands (9) 0 0 0 O 0 O 0 0 Statement of Wants and Demands (10) O 0 0 O O 1 1 1 Reject Other's Demands (13) O 1 0 O O 0 0 O Threat Against Other (14) 0 O 0 0 0 O 0 0 Threat Against Self (15) 1 2 0 0 0 1 3 0 Positive Backchannel (16) 1 O 0 0 1 O 0 0 Negative Backchannel (17) O O 1 0 0 0 0 0 Salutation (18) 1 0 0 O 0 0 0 O 259 RELATIONAL—GOAL CATEGORIES Phases Integrative Distributive Other Total 1 21 0 10 31 2 19 1 11 31 3 18 1 11 30 4 27 0 4 31 5 26 0 5 31 6 19 5 7 31 7 22 3 6 31 8 17 8 7 32 Table 18 RELATIONAL-GOAL CATEGORIES Phases Integrative Distributive Other Total 1 25 0 5 30 2 26 4 1 31 3 23 7 1 31 4 26 0 5 31 5 27 2 2 31 6 24 6 1 31 7 20 4 7 31 8 24 2 2 30 Table 19 Wow Relational-Goal Coding Categories 1 2 3 4 5 6 7 8 No Relational Goal (0) 0 1 0 0 1 3 1 3 Willingness to Comply (1) 2 9 8 6 3 8 2 0 Question of Disclosure (2) 1 1 0 1 O O 0 1 Question Fact About 10 1 4 10 6 4 4 3 Person (3) Question Fact About 2 0 2 0 O O 0 1 Hostage—Taking (4) Statement of Disclosure (5) O 4 1 5 9 1 2 3 Statement Fact About Person (6) 4 2 1 3 5 2 9 3 Statement Fact About Hostate-Taking (7) 1 O 0 O 0 1 3 3 Support/Comfort (8) 1 O 1 1 O 0 0 0 Question of Personal Wants/Demands (9) 0 l 1 1 2 O 1 0 Statement of Wants and Demands (10) O O O 0 0 3 3 8 Reject Other's Demands (13) 0 1 1 0 O 2 0 O Threat Against Other (14) 0 0 O 0 O O O O Threat Against Self (15) O 0 0 0 O 0 O 0 Positive Backchannel (16) 8 11 10 4 5 7 6 7 Negative Backchannel (17) 0 O 1 O O 0 0 O Salutation (18) 2 0 0 0 0 0 0 0 261 Table 20 31' “O 0. Q. :"- :.‘ {r ; 0.: 1.- {-flII. ‘ I I W PHASES Relational-Goal Coding Categories 1 2 3 4 5 6 7 8 No Relational Goal (0) 1 0 1 1 1 2 3 3 Willingness to Comply (1) 2 3 1 0 0 2 0 1 Question of Disclosure (2) O 2 1 3 2 1 0 1 Question Fact About 1 0 2 O 1 1 1 3 Person (3) Question Fact About 0 O 0 0 0 0 O O Hostage-Taking (4) ' Statement of Disclosure (5) 3 9 7 9 11 6 8 3 Statement Fact About Person (6) 16 8 9 12 12 10 4 10 Statement Fact About Hostate-Taking (7) 3 4 2 0 0 0 3 3 Support/Comfort (8) O 0 0 0 0 0 0 0 Question of Personal Wants/Demands (9) 0 O O 1 O 2 1 1 Statement of Wants and Demands (10) 0 2 7 O 2 10 4 2 Reject Other's Demands (13) O 1 0 O O O 0 0 Threat Against Other (14) O 0 O 0 0 0 O 0 Threat Against Self (15) O 1 O 0 0 0 O 0 Positive Backchannel (16) 5 1 1 4 2 1 7 1 Negative Backchannel (17) 0 0 O 1 0 0 0 O Salutation (18) O O 0 0 O O O 1 262 Table 21 u I I I I I . I 3.0 ’0“ I \“I z I a -- : - ;.I I I - ‘- ; 0.: CaLeEQIE_CQd2s_fQz_IzanSQIiELEIhreei_MED RELATIONAL-GOAL CATEGORIES Phases Integrative Distributive Other Total 1 61 1 4 66 2 51 2 13 66 3 59 1 6 66 4 50 1 15 66 5 53 5 9 67 6 58 4 2 64 7 53 9 4 66 8 60 4 2 66 Table 22 RELATIONAL-GOAL CATEGORIES Phases Integrative Distributive Other Total 1 34 18 13 65 2 51 4 11 66 3 44 3 19 66 4 42 6 17 65 5 33 5 28 66 6 4O 7 21 68 7 30 2 34 66 8 31 16 17 64 263 Table 23 E E H I' I B ] l' J G 1 B l E] E W PHASES Relational-Goal Coding Categories 1 2 3 4 5 6 7 8 No Relational Goal (0) 3 3 2 3 1 O 1 2 Willingness to Comply (1) 15 5 6 6 2 2 1 6 Question of Disclosure (2) 2 7 8 9 11 12 4 11 Question Fact About 18 10 14 11 11 8 9 13 Person (3) Question Fact About 3 0 0 1 1 1 O 1 Hostage-Taking (4) Statement of Disclosure (5) 8 16 20 6 13 22 25 13 Statement Fact About Person (6) 12 7 7 11 13 9 11 10 Statement Fact About Hostate-Taking (7) O O 0 3 1 2 0 3 Support/Comfort (8) 0 1 1 0 O O 2 1 Question of Personal Wants/Demands (9) 0 2 1 0 0 2 0 0 Statement of Wants and Demands (10) 1 2 O 1 5 4 9 3 Reject Other's Demands (13) 0 0 1 O O 0 O 1 Threat Against Other (14) O O 0 O 0 0 O 0 Threat Against Self (15) O 0 0 0 0 O O 0 Positive Backchannel (16) 4 12 5 15 7 2 4 2 Negative Backchannel (17) 0 O O 0 1 0 0 O Salutation (18) 0 1 1 0 1 0 0 0 264 Table 24 :3 z. I .I ;-‘- ;." {z ; II; a; :3‘1... ‘ I I W PHASES Relational-Goal Coding Categories 1 2 3 4 5 6 7 8 No Relational Goal (0) 0 8 1 1 2 5 3 1 Willingness to Comply (1) O 0 0 O 0 0 1 0 Question of Disclosure (2) 6 7 6 5 7 7 3 6 Question Fact About 5 3 3 10 2 2 4 3 Person (3) Question Fact About 0 0 O 3 1 4 O 2 Hostage-Taking (4) Statement of Disclosure (5) 13 19 25 17 11 19 11 13 Statement Fact About Person (6) 7 13 9 6 10 3 8 2 Statement Fact About Hostate-Taking (7) 2 0 0 O 0 0 O 4 Support/Comfort (8) 0 O O O 0 0 0 0 Question of Personal Wants/Demands (9) 1 1 0 0 0 0 0 0 Statement of Wants and Demands (10) 15 3 1 5 5 6 2 15 Reject Other's Demands (13) O 1 1 O O 0 O 0 Threat Against Other (14) 3 0 O O O 0 O 1 Threat Against Self (15) 0 O 1 1 O 1 0 0 Positive Backchannel (16) 13 11 19 17 28 21 32 13 Negative Backchannel (17) 0 O 0 0 0 0 2 3 Salutation (18) O 0 0 O O O 0 1 265 Table 25 . . . ;, -. . .--. ; . .-. p; z-a... - . .. . for_all_2hases TRANSCRIPTS Relational-Goal - Coding Categories WCPO MAJORS MPD No Identity Code 6 - 23 50 ----- Other—Hostile 0 0 O Self-Hostile 0 0 0 Defend-Self 0 2 3 Defend-Other 7 4 2 Restore-Self 33 44 32 Restore—Other 87 108 366 Positive Backchannels 1 65 70 Negative Backchannels 0 2 4 Table 26 . ;, -. . .. ;--— ;.- .-. ..; a-q... ; . foLallihases TRANSCRIPTS Identity-Goal Coding Categories WCPO MAJORS MPD No Identity Code 10 21 60 Other-Hostile 0 O 0 Self-Hostile 4 1 O Defend-Self 1 6 11 Defend-Other 0 0 4 Restore-Self 111 157 260 Restore-Other 2 31 21 Positive Backchannels 5 26 164 Negative Backchannels 2 2 6 267 Table 27 E Ell |.| III'IGJB 121 E W PHASES Identity-Goal --- Coding Categories 1 2 3 4 5 6 7 8 No Identity Code (0) 2 2 0 0 O 1 1 0*— Attack-Other ( 1) 0 0 0 O O 0 0 0 Attack-Self (2) 0 0 0 O 0 0 O O Defend-Self (3) O 0 0 O 0 0 0 O Defend-Other (4) 1 3 1 0 1 1 0 0 Support-Self (5) 1 3 1 O 2 7 10 9 Support-Other (6) 13 9 14 17 14 8 6 6 Positive Backchannel (7) O 0 1 0 O 0 O O Negative Backchannel (8) 0 0 0 0 O O O 0 268 Table 28 =, -.. . .. ;.-_.;,- . . “ z-a... ' PHASES Identity—Goal -- Coding Categories 1 2 3 4 5 6 7 8 No Identity Code (0) 3 2 O_-_ O 1 1 1 2 Attack-Other (1) O 0 O O O 0 0 0 Attack-Self (2) 1 1 O 1 O 0 O 1 Defend-Self (3) 0 1 0 0 0 O O 1 Defend—Other (4) O O O O 0 O 0 O Support-Self (5) 12 13 14 12 15 16 16 13 Support-Other (6) 0 O 2 O 0 0 0 0 Positive Backchannel (7) 1 0 0 3 1 O 0 0 Negative Backchannel (8) O 1 1 0 O 0 0 0 269 Table 29 E EN |.| III'IGJB IE! E W PHASES Identity-Goal —-- -- Coding Categories 1 2 3 4 5 6 7 8 No Identity Code (0) - 8--- 3 -0 4 1 3--—-l----3—- Attack-Other (1) 0 0 0 O O 0 0 O Attack-Self (2) O O O 0 0 0 0 O Defend-Self (3) 0 1 0 O 0 0 O 1 Defend-Other (4) 1 1 O 2 0 0 0 0 Support-Self (5) 4 5 6 7 1 9 8 4 Support-Other (6) 9 9 12 13 20 12 16 17 Positive Backchannel (7) 9 12 11 4 9 7 6 7 O 270 Table 30 W W PHASES Identity-Goal Coding Categories 1 2 3 4 5 6 7 8 No Identity Code (0) 3 0 2 3 1 2 3 7 Attack-Other (1) 0 0 0 0 O O 0 0 Attack—Self (2) 1 O 0 0 0 0 0 0 Defend-Self (3) 2 0 1 1 2 0 0 O Defend-Other (4) 0 O O O 0 0 O O Support-Self (5) 18 28 23 16 25 20 14 13 Support-Other (6) 1 2 4 6 0 6 6 6 Positive Backchannel (7) 5 1 1 4 3 3 8 1 Negative Backchannel (8) 0 0 0 1 0 0 O 1 271 Table 31 E E H |° I II l'l G ] B l E] E W PHASES Identity-Goal -------------- Coding Categories 1 2 3 4 5 6 7 8 No Identity Code (0) 8 6 3 4 7 5 8 9 Attack-Other (1) O O O 0 O 0 0 0 Attack-Self (2) 0 O O 0 0 0 0 0 Defend-Self (3) 0 1 1 1 0 0 0 0 Def end—Other (4) O 1 O 0 0 0 1 O Support-Self (5) 7 2 1 1 3 4 6 8 Support-Other (6) 42 38 54 42 48 52 43 47 Positive Backchannel (7) 9 16 6 18 8 3 8 2 Negative Backchannel (8) 0 2 1 0 1 O O 0 272 Table 32 PHASES Identity-Goal Coding Categories 1 2 3 4 5 6 7 8 No Identity Code (0) 5---10 4 7 5 13 8 8 - Attack-Other (1) O 0 0 O O 0 O 0 Attack-Self (2) O 0 0 O O 0 0 O Defend-Self (3) 6 1 1 2 0 0 1 0 Defend-Other (4) 1 0 0 l 2 0 0 0 Support-Self (5) 36 37 35 33 27 33 19 4O Support-Other (6) 3 3 5 3 4 2 1 0 Positive Backchannel (7) 14 15 21 18 28 20 35 13 Negative Backchannel (8) 0 0 0 1 0 0 2 3 Table 33 :0. 0‘3. 01“.. ‘ 0'34 Phases Hostage-Takers Negotiators 1 .282 .143 2 .294 .171 3 .252 .196 4 .297 .158 5 .264 .168 6 .245 .200 7 .340 .177 8 .302 .165 Table 34 . . .-;. ..m.. - ..q ;,_ . -. . . - . ': ---- ;.- ;.. Phases Hostage-Takers Negotiators ---- 1 .401 .167 2 .400 .132 3 .238 .074 4 .339 .057 5 .306 .085 6 .249 .189 7 .334 .174 8 .335 .251 Table 35 . . V‘s. .ou-o - :oq :e- u -. . . - . .: ::-— :. .. Phases Hostage-Takers Negotiators -- 1 ---------- -- .205 ----------- .092 —--— 2 .238 .160 3 .179 .282 4 .257 .209 5 .188 .260 6 .226 .239 7 .219 .190 8 .238 .127 274 Table 36 H C .I I I | .I S E H l 'I l I H I. | l E] E I . I H l I . H . Phases Hostage-Takers Negotiators 1 .260 .171 2 .243 .221 3 .345 .228 4 .296 .208 5 .299 .159 6 .261 .172 7 .467 .167 8 .337 .117 275 Table 37 E El! I'I III'IGJE IE] E alLTransoripts PHASES Identity-Goal ----- Coding Categories 1 2 3 4 5 6 7 8 No Identity Code (0) 18 11 3 8 8 9 10 12 Attack-Other ( 1) 0 0 0 O O 0 O O Attack-Self (2) 0 O 0 0 0 O 0 0 Defend-Self (3) 0 2 1 0 0 O 0 1 Defend-Other (4) 2 5 1 2 1 l 1 0 Restore-Self (5) 12 10 8 8 6 20 24 21 Restore—Other (6) 64 61 80 72 82 72 65 70 Positive Backchannel (7) 18 28 18 22 17 10 14 9 Negative Backchannel (8) O 2 2 1 1 O O 0 Table 38 . =9 -. °. .. ;-;- ;.- .-. ..; :-q... - . -.. - . PHASES Identity-Goal Coding Categories 1 2 3 4 5 6 7 8 No Identity Code (0) --- 11 12 6 10 7 16 12 17 AttacksOther (1) 0 0 0 0 0 0 O 0 Attack-Self (2) 2 1 0 1 0 O O 1 Defend—Self (3) 8 2 2 3 2 0 1 1 Def end-Other (4) 1 0 0 1 2 0 0 O Restore-Self (5) 66 78 72 61 67 69 39 66 Restore-Other (6) 4 5 9 9 4 8 7 6 Positive Backchannel (7) 20 16 19 25 32 23 43 14 Negative Backchannel (8) 0 1 2 2 O O 2 4 FIGURES 277 328540 odomlquofiqdm «ES 55 oz. 0 10°F [ecu loom woos [com [com a: I -2: ed; // m”, can .muamcomcmcp ppm do mommso ppm 1o>o Lov>ezon peomu_ecowpe_oc cmxmpuommumoz use couomaomoz mo accosoocm.fi mesmem ADNBODBHA F1 ure 2. Frequency of negotiator relational-goal behavior over all phases of all transcripts. 278 l 10 13 14 15 16 17 18 RELATIONAL-GOAL CODES l 'IA 'l/I/l/II/l/Il Ill/I Ill/III Ill/ll IIIIIIIll/IIIIIIIIIIIIIIIIIII[III/III 'I/l/l/lll/l/l/I/ll/l/III/Ill]Ill/IIIIIIIIIIIIII/lll/l/l/l/ll ll/l IIIIIIIIIII/IIIIIIIIIII/IIIIIIIIIIIIIIIIIIIIIII/IIIIII '//Ill/l//////////////////////// l/l/l/l/l/l/II/l/l/l/l/l/l/ ’ll/l/I/l 200 l I l l l O O O O C! co (.0 ‘f N 180- 160“ 140- 120‘ 100‘ AONBOOBEH 23456789 279 of all transcripts. Fi ure 3. Frequency of hostage-taker relational-goal behavior over all phases __9____ Q 7////////////////////////////////////. ’/////////////////////////////////// ’/////////////////////////////////////////////////////// RELATIONAL-GOAL CATEGORIES 2 3 4 5 6 7 8 910131415161718 W l 1 O 300 50- 0 l O m 250 r 200 " 100 " AONEODBMd TOT l-goal behav iona ge-taker relat Fi ure 4. Frequency of negotiator and hosta by phase for all transcripts. 280 HT DIST HT OTHER NEG OTHER s'A'O'A'I'I'I -“““‘ :4 *P’WH‘E‘H 411-7: ’lillfi‘fll‘d‘l 'IIIIIIIIIIIIIIIIIII[IIIIIIIIIIIIIIIIIIIIIIII Fa'a'a'a'a'a ‘m‘ “ *fljuq . r ) “‘1' V . r~mmnr-131 'tII411IrirItIIIIIJ/Ilmalmfml [III/IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII: I O. O O! I‘““ . waummuamm ///v:zi/-IIWJI1J III/IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII ‘V.I‘V‘V‘V‘I.I 7‘ .fl“ W1 fidfiiffl fifi’f'fffff??ff¢'"’ IIIIIIIIIIIIIIIIIIIIIIIIIIIIIJ'IIIIIIIIIIIIIIIII”g '-VQ's'c'.'c‘a'o'a'a'c'a'o'avA'5'. .‘ l..I "1:: , \ , J . ,. ~~wwvt| w .--. 1~ 7.1.- 1 £3 I I flI'J'I IIJ'J‘U' IIIIIIIIIIIIIIIIIIIIIIII’llIIIIIIIIIIIIIIIIIIIIIII'IIII I .‘I'I‘OI'I‘II I'll. Q! '1 : , iph'uwfr.) ..' Illlliwlmfiilil IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII uvvmrrvvauu-rrmra j 1“ I if!!! m IIIIIIIIIllllllllllllllllllllllIIIIIIIIIIIllllllllllllllllo ”131-4,44.— ‘Ifilfl’ldvil'l. IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII’IIIII ADNEODBUJ he to U) 04 '- PHASES ‘IOY‘ ge-taker relational-goal behav F1 are 5. Frequency of negotiator and hosta across phases for WCPO. 281 NEG DIST NEG OTHER HT INT HT DIST HT OTHER 74 IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII l“‘l“fl“‘\\“ IIIIIIIIIIIIII’llIIIIIIIIIIIIIIIII!IIIIIIIIIIIIIIIIIIIIIIIII IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII n“““i m ’IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII/IIIIIIIII. ’IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII ./£) IIIIIIIIIIII[III/IIIIII/IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII I“““““ IIIIIIIIIIIIIIIIIII/IIIIIIIIIIIIIII/IIIIIIIII/IIIIIIIIIIIIIII I'A'A'A'A'A‘n '1 |‘\‘ IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 18 I I I I I I I T I (b V N O on {b V N O ‘— ‘— Q- ADNHI'IDEHJ 5 4 TIME PHASES 282 mph—mvmwvwmworm m N. m LII. memomupqu 440mlgdzocfimm If) v I 'l/II II/l ‘- 11 O m (Ill/Ill]: l/l/l/l/l/l/l/l/l/l/l/ll '0 '///////////////////// N 'll/Il/I/III/l/l/l/l/lIII/ll fi.o -oF low I C V é: In AONBRDBHJ Low>mzma pmomu .oauz so» mommga mmogum pmcowumpmg mempnmmmumo; new Loumwuomm: we zocosamgm aMIMMmmHM Fi ure 7. Frequency of negotiator and hostage-taker relationaI-goal behavior by phase for Majors. 283 NEG OTHER HT OTHER .u“““‘uu“ ‘ ' x1 ILRIIJIJ-IJIJIJ IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 'A'A'A'A t.“ _ - I...;*.L:_ ; . /rJI/JfII/. 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII: o'c'.”d'g'§'c'§'t'a‘ ///I//I/: rllIlIlIIIIIIlIIIIIIIII(IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII. [.0 .“““\“““\‘ .-: ..-~;4:::r , ‘1:J.J..;..-x1,¢714 27‘1“; LL44 //ll/l/I/l/l/I/=/I////i’/ 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII _ IO. t“““‘ AL, .aazd',. , ‘.._~;r\u,tL,..L .‘sg""»l e . ,JJJJJ-V fill/IIIIJIIJIIIIIIII/l 'IIIIIIIIIIIII/IIIIIIIIIIIIIIIIIIIIIIIIII 'A'A’A'A'A"'A'A'A"'n /IIIIIIJII—ftl‘I‘l'I/rlkIII/(l, IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII I I r 7 I m N N ‘— ADNEI'IDBHd M N F RELATIONAL-GOAL CATEGORIES Fi ure 8. Frequency of negotiator and hostage-taker relationaI-goal behavior across phases for Majors. 284 ll/l/l/l/l/l/l/l/l/ll/l/l/l/II/I/////////. 'Il: 'Illl/l/I/t VIII; ’/1 III/ll Ill/III/IIII/III/I/II III/II/IIIIIIIII/l I/ll 'Il/l/l/I/lII/I/ll/l/l/IIII/I/I III III/IIIlI/IIIIIII/IIIIIII/I 'I/I/ll C3 0) I O on I 0 PM 1 O (b I I TI I I OO O O O O M N '- u) ‘t AONSHOBHJ 2 3 4 5 6 7 8 910131415161718 I D RELATIONAL-GOAL CATEGORIES I-goai behavior 1 ona iator and hostage-taker reIat Fi ure 9. Frequency of negot by phase for MPD. 285 NEG DIST NEG OTHER HT OTHER Ivocvcvcvvvcooov km\“\““‘\\‘ '5'a'rn'c'a'c'o'rrra'woman's" ' i _ _ “‘ ' I IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII I— A'A'A'A'A'd'fl't'n'n'a'n'a'n'n"'t't'aV‘K'A'a't'c‘ ‘ I“ *7, :3- m I. A 1% $98!}? a 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII. {'A'O‘.'o'b'b'A'A'A'O'Q'O‘Ivt'O‘ A. , It ‘gn IIII/JIWIIIIJ. 'IllIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII. Ol...tlt.l0ttlllt( cw. IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII‘ P rt¢oottottj “ , -2;sz34-1 ’ I I ‘1'"! ‘I 1‘] $11 3'4. IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII .‘A'A'A'A'A'A'A'fi’A'A'A 'I .n‘\‘§\‘ttwttttt\ ‘ * ,yn IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIll/IIIIIIIIIIIIIIIII l 7 l T l l I h to In 1" In N 1- In 1' N '- RELATIONAL-GOAL CATEGORIES Fi ure 10. Frequency of negotiator and hostage-taker reIationaI-goaI behavior across phases for MPD. 286 III/III/ll // // ’I/l IIll/lIIIII/IIIIIIII/I/III/II/I/ 'I/l/l/Il/I/II/I/I/I/Illl/l/I/Il/IlIII/ll/I/I/l/II 'l/ ’I/l/l/l/l/l/l/l/l/Ill/IIII/I/IIIIIIII 'IIIll/lllllllllllllllllll. ///////////////// Ill/ll T F I W I l l l C O O 0 C3 D c o o m ‘ N c on to V N v- v- v- '- ADNEOOEHJ 2 3 4 5 6 T 8 910131415161718 1 O RELATIONAL-GOAL CATEGORIES 287 mu_~_owuko Low>mcmn pmomuxuvpcmum gmxmuummoumo; use Lopmwpomm: mo xucmscmgu .HH mgamwm ADNBI'IDBHd 288 $568.30 ficm..»._._.~2uo_ m N. m m ¢ m N F _ P m i 9 fl // fl fl M 18 fl fl / fl M 13 / fl fl // fl M r8 fl fl // w / '8 W E I I. 2: 82 g 2: .oauz $o mwmmsn Fpm gm>o 3:23 33-3352 Luggage. 2:. 3:582 3 35:8: .3 AONSOOBUd 289 $555546 4497:.»25 m h m n v m N w o — — ,/ // /; / ./l ‘ I; , O M m m w / / / / / II I/ II / / / / / M M M o / / / / / / / / / w m é fl fl / W {cm fl fl fl 2: / I W IONF fog; om? .mgonmz mo mmmmca ppm gm>o Lom>mcmn Fwomuxupucmum memuummmumo; ucm gopmmuomm: we aucmzcmgu .mfi mgamvm AQNBHOBEJ 290 memouupqu Aqowlttae v mm N v _ p I '///////////. " ’/////////////////////////////////////////////////////////// m loop 1on9 c’: C N ADNBODBHA I O In N Icon Ionm nxuv .oaz $o mommga Fpm gm>o Lom>mgmn Foomuauwucwcw gmxmuummmumo; vcm goumwuommc mo Aucmaomgu .¢H mgamwm 291 muqum _ - a. vto 1m