Part III: Language

Large file: Patience please

Inner speech and imagined interactions form a perceptual core for the study of intrapersonal communication processes. In one study I conducted of 300 students, all except three understood the concept of talking to oneself either aloud or mentally. This concept is different from thinking, different from seeing oneself a sender and receiver, different from many contexts of communication. But, what is it? Perhaps language directed to oneself is a key to understanding this area of study. These articles are related to language as a fundamental aspect of the study of intrapersonal communication processes.

 The Good Side of Talking to Yourself (Donna R. Vocate)

A Typology of Inner-speech Patterns (Joan E. Aitken)

Imagined Interactions, Recurrent Conflict and Thought About Personal Relationships: A Memory Structure Approach (James M. Honeycutt)

Inner Voices as Alter-ego Mental Constructs: Auditory Verbal Hallucinations (Mark A. Hamilton)

Verbal Hallucinations in Normals: A Research Agenda (Terry R. Barrett)

Verbal Hallucinations In Normals, IV: The Physiological Arousal Hypothesis (Karen Hamlet, Terry R. Barrett, Peggy Sabel Paxton, and Kelly L. Blair)

The Good Side of Talking to Yourself

Donna R. Vocate

About the Author: Donna R. Vocate is on the faculty of the College of Communication, Boston University, 640 Commonwealth Ave., Boston, Massachusetts. Her recent work, Intrapersonal Communication: Different Voices, Different Minds is one of the most important pieces of scholarship in our field.

Abstract: Talking to yourself is an activity generally associated with being a bit daft if not actually insane. Consequently, no matter how rich and prolific our self-talk may be, it is never a source of pride and we may even be embarrassed when its existence is brought to our attention. My purpose in this brief essay is to modify that attitude toward self-talk to a more positive one by emphasizing its role in personal growth. Self-talk, after all, is the means by which humans can constructively influence their own development in the cognitive, social, or behavioral realms.

Perspective:

1. Think of a time when you were asked by a coach or a teacher to use self-talk to enhance your performance. Did it work? Why or why not?

2. Do you use self-talk to adapt to your partner in a conversation? What are some typical comments that would characterize this dialogue with the self?

Self-talk exists because of the phenomenon of inner speech (discussed at length in the work of L. S. Vygotsky (1934, 1986) and his colleagues). Inner speech, in brief, results from the internalization of talk or spoken language and its principles in the child. It undergoes drastic changes in structure as it inheres in the individual so inner speech is much more abbreviated and semantically concentrated than ordinary conversation. Once inner speech has developed in each of us, we are then capable of coding pure thought into intelligible expressions, and as a result of that coding, capable of having a dialogue with ourselves and others. Both processes of inner speech (coding and dialogue) are presented in detail elsewhere (Vocate, 1994) and will not be reviewed here since the focus in this essay is on the effects of self-talk.

A brief operational definition is in order for clarity. I define self-talk as a dialogue with the self existing in two forms: (a) the silent, internal dialogic process of inner speech, and (b) the audible, external dialogue addressed to the self although others may hear it. (Some psychologists label the latter form "private speech" and limit its functioning to the regulation of behavior.) In self-talk, the self is both the source and the object of interaction. Both the silent and the audible forms call forth an interpretive response or feedback, and both derive their existence from the coding or decoding process of inner speech. The distinctive attributes of self-talk given this definition, are: (a) self-awareness, or what Mead (1934) termed reflective consciousness, (b) its dialogical nature, (c) a stimulus, either sign or symbol, originating from the self, and (d) an interpretive, symbolic response or feedback from the self. Self-talk may be intentional or unintentional, silent or vocalized. Neither intentionality nor audibility affects its distinguishing self-awareness nor that the self is eliciting a response from itself as it would from another person on the other levels of human communication. The level of self-awareness may vary qualitatively from one occurrence to another, but some minimal level must exist in order to address the self as an object. This definition excludes vague impressions, sensations, images, fragments of thought, and natural sign behavior responses (Faules & Alexander, 1978), such as blushing from the domain of self-talk, thereby rendering it more accessible for study.

Change is an inevitable consequence of life since no living organism can remain totally constant. Piaget labeled the interactive process whereby living entities use their environment to enhance and benefit themselvesñadaptation. Change occurs because the adaptation process that increases the organization of the organism simultaneously requires adjustments in either the individual being or the items to be ingested, or both. The simplest manifestation of adaptation is physical, such as eating or breathingñbut it also occurs on the mental, social, and behavioral planes for humans where self-talk plays a vital role.

It is the blessing of humans that we can, if we wish, deliberately influence the direction and magnitude of any change tied to symbols in our adaptation. The means by which we accomplish this is self-talk. For example, we are capable of directing our own cognitive change whether we are learning about NAFTA from the newspaper or about intrapersonal communication from a teacher in the classroom. We can tell ourselves to remain open-minded. We can relate the concept or principle to other similar cognitions that we already possess. We can color the information with emotional attitudes that we hold toward the source of information, and then dissuade ourselves from attaching that positive or negative orientation to the new knowledge we just acquired. We can, of course, engage in self-talk to selectively decide which elements and associations concerning new information to retain or to reinforce and which to discard.

A similar process of adaptation via self-talk occurs socially. When we talk to the boss, when we meet our friend's parents, when we discuss social norms with someone from another culture, we must adjust our comments and behaviors to the various circumstances if we wish to avoid social embarrassment. Rod Hart's (Hart, Carlson, & Eadie, 1980) concept of rhetorical sensitivity advocates an adjustment to the context and the other person that would be impossible without the analysis, self-direction, and flexibility made possible by a dialogue with the self.

Using self-talk to guide behavior via self-fulfilling prophecies is a process essential to athletic coaches and to individual performers in any field. We can organize and reframe our perception of the situation and how we will behave to ensure that our behavior will be as flawless as possible. We may, if we wish, utilize self-talk during the process itself to monitor and control our behavior moment to moment. An illustration of this is that during a speech, we may comment to ourselves that an audience member looks puzzled, so we tell ourselves to expand our explanation to include more specific details and vivid examples to alleviate her or his confusion.

Years ago Bill Cosby had a comedy routine which portrayed self-talk very humorously with the line, "So, I said to myself, 'Self'. . ." The use of a title when addressing oneself was funny because it established a separateness that seemed foreign to our personal experience. Yet, we also know that we experience dialogues with the self daily. We may call these dialogues "thinking it over," or "considering our options," or "planning my day," or "trying not to hurt her or his feelings," but the reality is that we are talking to ourselves. Self-talk is shaping our world and our experiences in it. Such self-talk may be used also to transform the self as well as its experiences and perceptions. Mead (1934) and other sociogenetic theorists portrayed the self as a miniature society wherein biological and cultural facets of the personality interact (self-talk with the self as the topic) and produce a new synthesisña new self. Thus, self-talk is so vitally important to us that rather than being embarrassed about doing it, we should work to strengthen and improve its functioning and our use of it. Talking to ourselves is talk that shapes, and can serve to enhance, our performance in all the realms of human functioning.

References

Response from Leonard J. Shedletsky

Professor Vocate's chapter is a call to the significance of self-talk, its power and influence; and this is a needed call, since self-talk is such a "transparent" behavior. But, can we assign to self-talk the functions of analysis and flexibility, for instance? And how do we distinguish self-talk from thinking, categorizing, and inferring?

A Typology of Inner-speech Patterns

Joan E. Aitken

About the Author: Joan E. Aitken (Ed. D., University of Arkansas, 1985) is an associate professor in the Department of Communication Studies, University of Missouri-Kansas City, 5100 Rockhill Road, Kansas City, Missouri 64110-2499, telephone (816) 235-1337. An earlier version of this article was presented at the Commission of Intrapersonal Communication Processes and Social Cognition, 1994 Annual Meeting of the Speech Communication Association, New Orleans, LA.

Abstract: This study employs Q-Methodology to provide insight into the study of intrapersonal communication processes by seeking a typology of patterns of inner speech. An analysis of measures completed by 62 people suggests three person-types: the planner, the playful type, the quiet talker.

Perspective:

1. Do the types seem viable to you?

2. Which types remind you of yourself or people you know?

3. What might constitute inner-speech competence?

People talk to themselves, if not aloud, in their minds. Sometimes that self-talk includes thinking about a previous conversation. Sometimes people might think of things they wish they had said. Sometimes people visualize and hear others in their minds. This process of inner-speech includes communication rehearsal, rehashing conversations, talking to one's self through internal dialog or writing, communication with God, imagined interactions with media personalities, and similar behaviors.

The importance of inner-speech and imagined interactions has been demonstrated by previous researchers in this line of research (Critchley, 1991; Crouse & Crouse, 1988; Ericson, 1990; Johnson, J. R., 1984; Weaver, 1985). Among the many functions of inner-speech, Losee (1986) discussed the role of self-talk in stress and Neumann (1987) found a link between imaging and creativity. Beyond defining how inner-speech is used, some researchers have made recommendations regarding the improvement of self-talk skills. Weaver and his colleagues (1988), for example, offered several suggestions for controlling negative self-talk. Weaver and Cotrell (1986) discussed the role of imaging material pictures in intrapersonal communication.

Research in intrapersonal communication has included many different directions. While on one hand research over recent years has given many insights into intrapersonal processing, there is still much to understand. Some new directions in intrapersonal communication research have included the investigation of processing (Stacks, 1989) and other areas (Honeycutt, Zagacki, Edwards, 1989; Edwards, Honeycutt, and Zagacki, 1989). Because it is a relatively new area of study, as recently as 1987, Honeycutt indicated that imagined interactions are barely understood by those who are currently conducting research.

One area of needed research is the nature of how inner-speech varies between types of intrapersonal communicators. Although various communicator typologies have been developed over the years (e.g., Stephen, 1985), this area needs development regarding the use of inner-speech. While various researchers have worked to define inner-speech and its functions (e.g., Korba, 1990; Streff, 1987; Hitch, Halliday, Sebastian, Schaafstal, & Heffernan, 1991), recent studies of intrapersonal communication processes have examined inner voices (Hamilton, 1994), the role of memory in inner speech (Honeycutt, 1994) and other aspects of how people communicate with themselves. The purpose of this study is to provide insight into the study of inner-speech by seeking a typology of patterns of inner-speech.

Method

Q-Methodology is particularly suited to studying intrapersonal communication processes because it is theoretically based on the importance of communicability. As people talk about a given subject, their concourse of statements defines the nature of that subject. Inner-speech is a topic about which people seldom talk, thus the first step was to develop a concourse of statements for the purpose of designing the measure. Stephenson, (1986) indicated that "Q is based on communication and meaning as reflected in the concourse." By using student essays, interviewing subjects, and acquiring statements from their common language, the author collected a concourse of hundreds of statements that manifested aspects of the nature of inner-speech in our culture. Of course, one may question whether or not intrapersonal communication processes can adequately be translated into verbal constructs. In addition, because of the social taboos about "talking to yourself," few people have articulated their inner speech experiences. Thus, in hopes of translating inner speech experiences into common language, this concourse is particularly important in helping to define inner-speech. Statements for a Q-sort were selected from the concourse.

The result was a 66 item Q-sort. Respondents were given the following response grid: -5 (3 statements), -4 (5 statements), -3 (6 statements), -2 (7 statements), -1 (8 statements), 0 (8 statements), +1 (8 statements), +2 (7 statements), +3 (6 statements), +4 (5 statements), +5 (3 statements). Conditions of instruction were as follows: "You have been given a stack of 66 statements for a Q-sort. The statements are about self-talk; or how we think about communication, practice conversations, and imagine interactions in our minds. There are no right or wrong answers. Follow your first impressions. BE SURE TO MIX UP THE STATEMENTS BEFORE YOU BEGIN. Then, sort the statements into three stacks: agree (pleasure), disagree (unpleasure), and neutral (undecided)." From there the individual re-sorted the statement to fit in the grid.

Subjects. The results of 62 Q-sorts were used in this study, which is consistent with appropriate use of Q-methodology (Casey & Graham, 1988; Stephen, 1985, p. 193).

Q-sorts were given to members and friends of four groups: (a) students enrolled in an upper class interpersonal communication class, (b) a group of employees at a government agency, (c) undergraduate pharmacy school majors, (d) upper class communication majors enrolled in an advanced public speaking course. The groups were not intended to be representative, but to reflect diversity of student and non-student adult populations. These data reflect the results of 62 people, 69% female, 71% white, aged 15-62, high school through graduate school graduates.

Measure. The procedure in developing the measure was as follows. After studying a unit on intrapersonal communication, nearly 300 students enrolled in a basic communication courses at a mid-sized, urban university wrote essays on the nature of their self-talk. Only three students indicated they did not understand the assignment or the concept of self-talk. Statements of opinions or feelings about self-talk were extracted from the essays, providing the basis for a Q concourse. To provide other perspectives for a quality concourse (Casey & Graham, 1988, p. 3) additional statements were collected from interviews with 25 people (non-students). The development of the measure was used as a means of teaching communication students about intrapersonal processes. By examining the nature of these statements, patterns of statement content were categorized, from which a sample was drawn.

Because a measure of inner-speech was needed to determine how different people use inner-speech, the author checked statements for inclusion of elements established by recent researchers in the measure. Thus, the statements were categorized according to the nature of the content. Edwards (1989) raised issues about the differences by gender, memories, frequency and pleasantness of imagined interactions. Another key idea regarding inner-speech is how individuals use such intrapersonal communication as rehearsal for communication interaction (Edwards, 1988; Zagacki, 1992). In addition to mental processing, the Q-sort needed to reflect manifestations of inner-speech through writing. Internal dialog, self-talk, second thoughts, dissonance have been examined through the written forms of speech drafts, letters, journals, and diaries (Jensen, 1983, 1984, 1990; Litowitz & Gundlach, 1987; Shedletsky, 1988), which have provided a valid way to study the intrapersonal process. Because of the importance of romantic partners, family members, and relational significance in imagined interactions (Honeycutt, 1987) statements concerning the target of the inner-speech were included in the measure. Finally, some individuals imagine interaction with media personalities and those interactions may be typical of certain kinds of communicators. Parasocial interaction as it relates to media personality has been well documented (e.g., Allen, 1988; Conway & Rubin, 1991; Perse & Rubin, 1989; Rubin & Perse, 1987; Stamm, 1987), thus, the concept was included in the measure.

As a pilot study, the measure was administered to a group of upper class communication students and their friends and relatives. Students were asked to respond to the clarity of wording of the statements. Consensus itemsñ statements that all respondents perceive the same wayñ were omitted and new statements within the same categories were substituted. Other statements were reworded for clarity based on student suggestions.

A sample of 66 statements was then extracted that represented each category and positive and negative responses to various aspects of self-talk. This informal structure was used in creating the final measure (see use of informal structure, Brenner, 1988, p. 13). Under the general larger category of purpose, the following items were included: practice and planning (5, 26, 31, 45, 51, 53, 62), amusement and escape (25, 37, 42, 52, 56, 58), rehashing (6, 12, 38, 61), controlling emotions of apprehension (1, 8, 9, 35), controlling emotions of anger (2, 4, 22, 30, 63), providing comfort (3, 28, 57, 60), problem solving (26, 29, 40), motivation (7, 44, 46, 50, 59, 64).

A second larger category was the target of the inner-speech, with statements directed toward the higher self (13, 15, 19, 36, 49), and toward an important person (14, 17, 21, 48).

The third larger category was patterns of inner-speech included the type of self-talk (10, 11, 16, 24, 33, 34, 43, 47, 54, 65) and frequency of self-talk (18, 20, 23, 27, 32, 41, 55, 66). The value of such a design is primarily to make sure essential elements in the process of inner-speech are included in the measure. This final 66 statement Q-sort fell within the parameters of most Q-sorts (Brown, 1987, p. 98; Brown, 1986, p. 59).

Data Collection and Analysis

Respondents sorted statements according to their inter-relationships on an agreeñdisagree continuum. Preliminary data analysis was by the most widely used computer program for Q analysis (Stephen, 1985, p. 204), Van Tubergen's (1975) computer program which uses varimax rotation. By looking at the descending array of z-scores and item descriptions for each type, the author examined how each person-type would ideally arrange statements on an agree-disagree continuum. A synopsis of each type was made based on a study of the arrays and comparisons between types. A report of the data included in Appendix A.

Results

Using a four factor solution, there was low correlation between types, with 36% of the variance accounted for by the first three types. A three factor solution was extracted using varimax (orthogonal) rotation. The solution was optimized after 7 rotations. Factor 3 was 33% negative, so negative items were extracted and formed into an additional type. There were no significant correlations between types other than -0.4 between type 3 and 4.

On the final measure, nine consensus items emerged. These consensus items indicate statements about which the respondents think similarly. They are important in that they suggest a commonalty of pattern among all respondents. These scores fall in the neutral range, but whether they are positive (agree) or negative (disagree) is indicated at the end of each statement.

#30. I yell at someone in my imagination about once a week (+).

#10. Sometimes I have imaginary conversations out loud (+).

#54. In my mind, I often write letters to friends and family. I think of things that I would like to say to others, but know that I never will say (-).

#58. I think because I am shy I have more imagination than most people. I can imagine things to make myself feel better (-).

#15. In my imagination, I have conversations with myself mostly. I call myself by name, but I am speaking to another me (-).

#48. In my imagination I have conversations with people of power (-).

#29. I sometimes talk to someone I don't see any more or who died, who was very special. I picture what that person would have told me to do (-).

#32. I seldom have imaginary conversations because most of the time I say what I feel (-).

41. Having inner conversations make me feel foolish and lose touch with my surroundings

(-).

Discussion

The first area of consideration is how respondents agreed in their use of inner-speech. Apparently expressing anger through inner-speech is a typical use. Respondents were neutral in their attitudes about several items in the Q-sort, including having imaginary conversations out loud and writing mental letters. Respondents seldom perceived themselves as having more imagination because of shyness, speaking to a second-self, using self-talk to speak to people in power or to a deceased person. Apparently respondents agreed that they feel comfortable about frequent inner-speech because the two consensus items with the highest disagreement were: "I seldom have imaginary conversations because most of the time I say what I feel." and "Having inner conversations make me feel foolish and lose touch with my surroundings."

The second area of consideration is the typology of inner-speech patterns. Q-methodology is a combination of quantitative and qualitative approaches. After the results are massaged according the precepts of a particular program of data analysis, the researcher then attempts to make sense of the results according to subjective interpretation of person-types. These descriptions do not represent real people, but a composite picture of people whose similarities make them group together. Some people land in a category because they are similar to that type (positive correlation), while others land in a category because they are dissimilar to that type (negative correlation). The number of people in each categoryñpositive and negativeñis indicated after each description below. Negatives would suggest the reverse of the description provided (changing the meaning of the description for those types). Some people are unlike any type, which has two possible explanations. First, the potentially time-consuming and complicated nature of Q-sorting sometimes results in incorrect responses due to respondent confusion or impatience (particularly with young respondents). Another explanation is that these persons may perceive self-talk in a significantly different way from the other people in the study. Descriptions of their views are not reported here. Below are descriptions of how the 3 basic person-types might describe their use of inner-speech.

The Planner: Using Self-Talk to Manage Interpersonal Relationships.

I often think through how I will handle an upcoming anxiety-producing situation. I try to imagine different problems that may arise and how I'll deal with them. Self talk has been incredibly helpful to enable me to sort out things I want to do and how I want relationships to develop. Self talk is like a practice session for me. I find inner conversations to be most helpful. Talking to myself helps me vent frustration, so when I do confront someone I am more calm and controlled. Sometimes I talk to myself about things I have to do and when they have to be done. In my imagination, I have conversations with people I should have confronted.

I have imaginary conversations in which I replay a conversation the way I would have liked it to have taken place. I am unlikely to use self talk to talk to imaginary, famous, or deceased people, and I probably avoid discussing my self-talk with other people.

I am more likelyñthan the other typesñto write notes or letters as reminders to myself. I am more apt to use my imagination to have conversations with people I should have confronted. I am more likely to talk to myself about things I have to do and when they have to be done. I am more apt to find inner conversations to be helpful. I am less likelyñthan the other typesñto talk to imaginary people or people I don't know well. (This type was comprised of 35 positive and 1 negative loadings.)

The Playful Type: Using Self-Talk Creatively. I think one of the reasons I communicate well in most situations is because I can imagine things in advance. I am good at making sense of things afterwards too. I can picture situations, think about what others might say or do, and adapt to the various responses from other people. Self-talk enables me to manage situations, people, and myself.

Every once in a while I have conversations in my imagination when I am distracted or bored. My self-talk is part of my creativity. I could imagine conversations with a performer I idolize, for example. In my imagination, I have conversations with people without feeling shy or having to be careful about what I say. I might even have an imaginary friend I talk to. I know it's not real, but it's fun. I might have an imaginative conversation with a famous person or even a person who has died. In my imagination, I even have conversations with my higher self. That part of me guides me and provides wisdom and insight when I need it. I feel well about inner conversations and am quite imaginative in my use of self-talk.

I am most likelyñas compared to other typesñto have an imaginary friend to talk to, or to have imaginative conversations with famous people, or to imagine conversations with a performer I idolize. I am also more likely to perceive conversations with my higher self; a part of me that guides me and provides wisdom and insight when I need it. I am more likely than other people to talk about my self-talk. I am less likely than other types to use self-talk to plan or rehash conversations. I am more apt to use self-talk for my own entertainment. (This type was comprised of 6 positive, and 3 negative loadings.)

The Quiet Talker: Using Frequent Self-Talk to Cope with Reality. I use my self-talk to sort out feelings and as a way of coping. In my imagination, I have conversations with my "X" (friend, lover, spouse). I calm myself down to sleep by imagining I'm talking to someone. I am apt to carry on conversations with myself most of the time, during any available moment. I have imaginary conversations with myself when my feelings get hurt, so I can let those frustrations out, without showing my anger or as preparation before I confront the person. I am inclined to "yell at" someone in imaginary conversations. Sometimes I use self-talk so much, I'm confused about whether or not I've actually said things to others. I am, however, quite grounded in the role of reality in such conversations.

I am more likelyñthan the other typesñto record messages to myself on tape and play them back to me. I suspect that I use self-talk more than the other types. I am more likely to imagine telling important people about things that need to be changed. I am more likely to dream of places where I want to be with my friends or loved ones, and imagine us talking and enjoying each other. I am least likelyñas compared to other typesñto blame myself and use self-talk to rehash problems. I am the least likely type to talk to myself aloud. I am least likely to use self-talk as conversation with God. (This type was comprised of 7 positive, and 3 negative loadings.)

These three typesñthe planner, the playful type, and the quiet talkerñreflect general purposes and behaviors of inner-speech. Although the patterns here are not representative of everyone, they indicate that people use inner speech for different purposes, aimed at different targets, and through different pattern types and frequencies. The types raise questions about whether the inner-speech patterns might be comparable to external communication patterns. Comparisons to measures of communication apprehension, personality inventories, cognitive style measures, and measures of interpersonal communication competence, for example, would give insight into the effectiveness of these inner communication patterns. Are high communication apprehensive that way because they lack the skills to effectively and creatively visualize communication interactions?

The patterns resulting in this study also raise questions about the implications of communication competence in inner-speech. Because inner-speech is a topic that receives little discussion in a student's education, what principles of effective inner speech can be determined and taught? Previous publications in this area suggest strategies for more effective inner speech, but with specific patterns of use already in place, it may be wise to recommend strategies that best fit within certain strategies. This study suggests that inner-speech patterns do exist, that certain people rely on certain inner-speech behaviors more than other behaviors, and that different people use inner-speech for different purposes. 


References

Appendix A

Inner-speech Q-sort: Typal Array Z's

Note: Some statements are shortened to conserve space.

Initial four factor solutions for types:1234

Item Descriptions:N'S for each type are:388115

1. In my imagination, I have conversations with my0.7-1.31.1-0.6

lover, especially about thing1s I'm afraid to discuss.

2. When I get upset with someone, I don't call or write,0.4 0.1-0.31.8

I talk it out with myself. Then later, I talk to

them about why I'm upset.

3. I calm myself down to sleep by imagining I'm talking-0.5-0.11.7-1.1

to someone.

4. I seldom "bitch" at someone in my imagination.-0.9-0.4-1.81.4

5. While I'm driving in the car, I sometimes have0.9-1.31.1-0.8

imaginary conversations that I plan to have later.

6. In my imagination, I have conversations with people I1.4-1.11.2-0.7

should have confronted.

7. Sometimes I imagine myself on a job interview,0.40.71.3-1.7

talking to the boss, or being promoted.

8. I imagine getting up the nerve to tell someone1.20.1-0.81.5

something that is very difficult to say.

9. In my imagination, I have conversations with people0.61.71.4-0.2

without feeling shy or having to watch what I say.

10. Sometimes I have imaginary conversations out loud.-0.30.6-0.20.1

11. Sometimes I record messages to myself on tape and-1.9-1.11.3-1.0

play them back to me.

12. In my imagination, I have conversations with my "X"0.1-0.82.10.1

(friend, lover, spouse).

13. I have daily conversations with God.0.4 1.2 0.6 -1.4

14. I imagine conversations with a performer I idolize.-1.8 1.7 0.8 0.6

15. In my imagination, I have conversations with myself-0.8 -0.3 -0.5 -0.1

mostly. I call myself by name, but I am speaking to

another me.

16. I often keep a journal, write about stressful things,-0.7 -0.9 0.2 -0.4

or write down particularly significant feelings or thoughts.

17. I have imaginative conversations with famous people,-2.0 1.4 0.1 0.7

sometimes ones who have died.

18. I really don't have conversations with anyone in my-1.6-2.1 0.0 -0.6

mind.

19. God doesn't speak to me like Oral Roberts, but we do0.7 -0.7 -0.5 0.2

communicate.

20. I think people would think you are "off the wall" if-0.8 0.2 0.1 0.2

you had imaginary conversations.

21. I have imaginary conversations with important people-0.7 -0.8 1.8 0.8

and tell them about things that need to be changed.

22. I have imaginary conversations with myself when my 0.8 1.1 1.4 -0.1

feelings get hurt, so I can let those frustrations out,

without showing my anger or before I confront the person.

23. I carry on conversations with myself most of the-1.0 -0.6 1.5 -1.3

time, during any available moment.

24. The content of my self-talk varies depending on my1.1 1.2 -0.3 0.8

mood.

25. Every once in a while I have conversations in my0.3 1.7 0.1 0.2

imagination when I am distracted or bored.

26. I use imaginary conversations to help me think 1.3 0.5 0.5 -0.2

through problems.

27. I never have conversations in my imagination.-1.8 -2.1 0.2 1.7

28. In my imagination I have conversations with myself.-0.4 -0.3 -0.4 -1.7

After talking aloud, I feel consoled. It's not like I am

actually talking to someone.

29. I sometimes talk to someone I don't see any more or-0.5 -1.1 -0.7 -1.0

who died, who was very special. I picture what that

person would have told me to do.

30. I yell at someone in my imagination about once a-0.2 0.3 0.8 0.7

week.

31. In my imagination I have conversations with people in0.3 -0.8 -1.2 0.9

my future. I practice and rehearse in my mind

conversations and speeches.

32. I seldom have imaginary conversations because most of-0.7 -1.5 -0.8 -0.6

the time I say what I feel.

33. In my mind's eye, I have conversations with myself,1.0 0.8 0.4 -1.0

especially when I do something great or bad.

34. Sometimes I talk to myself in the mirror, at the-0.1 1.0 0.7 -1.5

computer, or places like that.

35. I often think through how I will handle an upcoming1.8 0.6 1.0 0.7

anxiety-producing situation. I try to imagine different

problems that may arise and how I'll deal with them.

36. When I'm sorry about something, or something good0.6 -0.8 -1.5 1.6

happens, or just whenever, I have a conversation with God

37. I have imaginary conversations with someone I care0.5 -1.2 -1.7 1.5

about who is away from me.

38. I have imaginary conversations in which I replay a1.4 0.5 -0.1 0.2

conversation the way I would have liked to have taken place.

39. I seldom daydream.-1.1 -0.3 -0.7 0.4

40. In my mind, I hear advice from someone important to-0.2 0.6 0.8 -0.8

me.

41. Having inner conversations make me feel foolish and-1.1 -1.9 -1.2 -1.8

lose touch with my surroundings.

42. I have an imaginary friend I talk to. I know it's-1.9 1.7 0.8 -0.4

not real, but it's fun.

43. I talk about my imaginary conversations with other-1.2 0.3 -1.0 -1.8

people.

44. Sometimes I talk to myself about things I have to do1.4 0.3 -0.5 -0.3

and when they have to be done.

45. I have imaginary conversations with people I'm trying0.8 -0.2 0.3-0.8

to influence in my direction. I practice what I might say.

46. When I exercise, I imagine results.0.6 0.9 -0.1 0.8

47. Sometimes I talk to someone in my imagination, then I-1.1 0.4 1.3 0.9

become confused about whether or not I actually said it to

them.

48. In my imagination I have conversations with people of-1.1 -1.1 -0.4 -0.3

power.

49. In my imagination I have conversations with my higher-0.1 1.3 -1.1 -0.2

self. That part of me guides me and provides wisdom and

insight when I need it.

50. Sometimes I write letters or notes as reminders to1.1 0.0 -0.5 -1.5

myself.

51. Self talk has been incredibly helpful to sort out 1.5 1.1 0.0 -0.5

things I want to do and how I want relationships to develop.

Self talk is like a practice session for me.

52. I imagine romantically related conversations.0.7 -0.8 0.30.1

53. I think one of the reasons. . . I can imagine things in1.3 1.8 -0.4 -0.2

advance. . . good at making sense of things afterwards too. I

can picture situations Self-talk enables me to manage.

54. In my mind, I often write letters to friends and-0.7 -0.1 -0.0 0.2

family. I think of things that I would like to say to

others, but know that I never will say.

55. I used to think I was the only persons having-0.7 -0.2 -0.0 1.5

conversations in my mind, but I guess most people have them.

56. I dream of places I want to be with my friends or0.9 -1.4 1.2 -0.8

loved ones. I see us talking and enjoying each other.

57. I find myself talking aloud when trying to solve a0.2 0.2 -1.0 1.3

problem. Sometimes I sit and imagine a problem in my mind

until I find a solution. Then relax about the problem.

58. I think because I am shy I have more imagination than-0.7 0.0 -0.4 -0.4

most people. I can imagine things to make myself feel

better.

59. Sometimes I scold myself or try to motivate myself1.2 -0.6 -1.1 1.4

with self talk.

60. I find inner conversations to be most helpful.1.4 -0.1 0.5 -0.4

Talking to myself helps me vent frustration, so when I do

confront someone I am more calm and controlled.

61. I often blame myself for things that go wrong. I 0.4 1.3 -1.3 1.0

replay things in my mind. I think "if only" I had . . . . I

always find a fault in myself to blame for the problem.

62. I talk to myself so I can figure out how to sound0.6 0.9 -1.4 0.5

like I know what I'm talking about when I speak to other

people.

63. Lots of my "conversations" with myself are actually-0.4 0.3 0.5 -1.5

scenarios in which I do all the talking without concern for

what another person might say back.

64. I anticipate favorable things with mental pictures,0.5 1.1 -1.4 0.6

and I find that they happen with ease. I'm not

disappointed if there are different outcomes than

I pictured.

65. When I'm around people, I have to remember not to-1.1 -0.3 -2.0 1.8

talk to myself out loud. Talking aloud to myself seems

natural, but we're not "supposed" to do it.

66. Imaginary conversations are just that, not real, and-0.6 -1.4 -2.1 1.3

reality can only be faced in reality, not in my mind.

Response from Leonard J. Shedletsky


To a reader not familiar with Q Methodology, the subjective interpretation of person-types is the point in the analysis process that leaves a door open for questioning. But if we accept the researcher's interpretation of person-types, and that seems an acceptable tact for the time being, we have what appears to be a useful handle for pursuing this line of research. Moreover, for one interested in learning more about oneself this typology gives us a starting point.

Imagined Interactions, Recurrent Conflict and Thought About Personal Relationships: A Memory Structure Approach

James M. Honeycutt

About the Author: James M. Honeycutt is the director of the Center for Imagined Interaction Research in the Department of Speech Communication, Louisiana State University, Baton Rouge, LA 70803-3923.

Abstract: Developmental models of communication and relational dynamics have been popular in the social psychological literature for over 20 years (e.g., Altman & Taylor, 1973; Knapp & Vangelisti, 1992). This article reviews criticisms of linear models of relational growth and outlines a memory structure approach to relational dynamics based on imagined interactions. A major source of information about information that creates expectancies for relationship development are imagined interactions. These are internal dialogues with significant others that allows individuals to review relational encounters as well as serving a rehearsal function for future relational encounters. Imagined interactions create habitual scripts for various encounters. A memory structure approach reveals how individuals organize information about relational events.


Perspective:

Think about your partner, spouse, or best friend for the following questions;

1. How often do you imagine talking with that person about problems or topics of interest to you?

2. How often do you replay conversations in your mind with this person? Does this make you think about what was said, could have been said that was not, or what you will say next time you see this person?

3. What are your expectancies for the development of close, intimate relationships? Where did these expectancies come from? Do you think these expectancies will change over time?


People have memories about their personal relationships that are often experienced through covert dialogues or imagined interactions (IIs) with relational partners (Honeycutt, 1989; 1991). We can imagine dialogue with a spouse, parent, sibling, or any significant other that helps us to rehearse anticipated interaction, ruminate about previous encounters, and relieve pent-up feelings that may have resulted in anxiety, conflict or tension. Imagined interactions also keep conflict alive as individuals replay old disagreements with significant others in their minds. Individuals may become angry as they relive old conflicts in their minds. Thus, imagined interactions serve a function of catharsis and self-understanding while keeping relationships and recurrent conflict alive in our thoughts. Our expectancies for relationships emanates from imagined interactions through replaying images from the electronic media as well as conversations with parents, siblings, peers, or novels.

People have memories about their personal relationships that affect current or projected relationships. Our memories about relationships form later scripts or expectancies for appropriate behaviors in relationships. The content of one's expectancies for the rise, demise, and maintenance of personal relationships reflects a relationship memory structure. Relational memory structures act as rules for expected behaviors in relationships (e. g., Be faithful, Be trustworthy, Show affection, Share time together, Self-disclose, Stand up for the partner). Some persons may have complex expectancies for relational behavior while others have very simple expectancies. Memories about relationships may be functional or dysfunctional. Memories may decay or be recalled clearly during critical times. For example, during critical times, we may replay conversations in our minds with loved ones and remember memorable messages that impacted us greatly (Knapp, Stohl, & Reardon, 1981).

Imagined interactions and memories about relationship act as a gyroscope that leads persons through current relational space. Memories about relationships reduce the relationship to a mental model of individual expectancies. Thus, relationships exist in the minds of individuals as well being recognized from an outsider's perspective (Olson, 1977) in terms of observable interaction episodes and encounters between relational "partners." While the study of interpersonal relationships has intrigued individuals since the beginning of time, an examination of the memory structures of relationships brings order to a chaotic and perhaps a "normless" complex social world. Hence, relationships can be studied within the minds of individuals as well as in the communication and observable companionship between two individuals.

The study of imagined interactions in personal relationships also explains why there is often recurrent conflict and themes that characterize the encounters between relational partners. Imagined interactions link a series of encounters together as individuals replay what was previously said (e.g., "Last time I said X with poor results, but next time I see her or him I am going to say Y with alternative results") as well as anticipating what may be said at the next encounter. Conflict may be kept alive by reliving old arguments and imagining the next interaction such that the next encounter may become a self-fulfilling prophecy as the interaction expectancy is enacted. More information on the linkage function of imagined interactions for recurrent conflict will be discussed later in this chapter. A memory structure approach to the study of relationships examines how individuals mentally create their relationships and keep conflict alive through imagined interactions and observation of others. Indeed, it is the mental creations of relationships that guide individuals in interpreting the meaning of behaviors in everyday interactions and relational encounters. For example, self-disclosure may symbolize trust and the escalation of intimacy in a couple. However, the valence of the disclosure may signify the decline of intimacy particularly if the impact of the disclosure is negative (Baxter & Wilmot, 1983; Gottman, 1979). Individuals may imagine self-disclosing episodes in the effort to determine how the other person will react to "negative" disclosures. They may rehearse what will be said and imagine various reactions to the way something is said. After repeated experience, rehearsal is no longer necessary as the individual is able to mindlessly execute "tried and proven" statements (e.g., comforting messages used at a funeral, greeting rituals, farewells). Before discussing the role of imagined interactions in creating relational memory structures, it is necessary to briefly review how personal relationships have been studied and conceptualized in the communication and social psychology literature.

Developmental Models of Relationship Growth

A developmental view of relationships assumes that individuals progress through various phases. Models of relationship development range from simple three stage models such as Levinger's (1974) old awareness model to more complex multi-stage models involving constant negotiation and recalibration (Altman, Vinsel, & Brown, 1981; Baxter, 1985; Knapp & Vangelisti, 1992; Taylor & Altman, 1987). Relationships are constantly moving organisms rather than static entities. Relationships, like organisms can be growing as well as appear to be in some type of steady state or decaying. Having achieved a certain level of intimacy, the individuals may have to recalibrate themselves as a function of cultural influences that may be communicated through television images of relationships (e.g., consider soap-opera images of relationships as combustible materials) influence of family, friends, and associates. Having achieved a level of intimacy, individuals can call on their expectancies of what should happen in close relationships in order to interpret the state and quality of the ongoing relationship experiences.

The idea that individuals have various levels of expectancies for the development of relationships is hardly provoking. Yet, the ability to articulate expectancies is a function of individual differences such as gender and previous success in relationships (Honeycutt, 1993). It is not known how helpful relational expectancies are in interpreting behavior within relationships. Further, do individuals really think about relationships going through stages that parallel the models that social psychologists have developed over the past 40 years? For example, family researchers have used the notion of the "family life cycle" for the past 40 years even though the cycle historically has assumed the existence of an intact marriage. Yet, other family forms such as remarried individuals, single-parents, or older couples having additional children who are younger than their grandchildren have not been captured well by the stages of the family life cycle. In essence, these older couples are in multiple phases of the cycle.

If relationships develop through a series of phases, how do the individuals decide on boundary points between phases? There is the question of the representational validity of developmental stages. Poole and Folger (1981) have discussed representational validity of researcher-imposed classifications of subjects' behaviors in terms of the psychological meaning of categories for the subjects that the researchers have studied. Whereas they have discussed this in terms of behavioral coding schemes, the validity notion applies to falsification of developmental stage notions. For example, relationship partner's reports of what they were doing in an interaction with their partner may be incongruent with a researcher imposed interpretation of the meaning of the behavioral acts (cf., Poole & Folger, 1981). The former represents the so-called insider's views of relationship dynamics while the latter has been referred to as the outsider perspective (Olson, 1977).

Duck, Pond, and Leatham (1991) have advised relational researchers to partition the insider's view by differentiating reports about relationships that are common to all observers from reports that result from one or the other partner's relatively constant ways of perceiving relationships.[l] If it can be demonstrated that individuals have relatively common expectations about behaviors that characterize the rise and demise of close relationships, then we are beginning to specify the insider's (respondents') views of relationships. A memory structure approach can provide us with information from an insider's perspective on agreement about relationship characteristics common to a variety of observers.

Aside from specifying components of insiders' views of relationships, a memory structure approach to relationship development provides a way of analyzing thought processes about relationships. Developmental stage theorists have been criticized for the arbitrariness of so-called stage movement. Why do some individuals move rapidly or slowly through different stages? A memory structure approach assumes that expectancies would be a factor in determining the rate and direction of movement. This has indirectly been tested through a series of experiments in which persons sort through a random set of escalating and de-escalating relationship behaviors (Honeycutt, Cantrill, & Greene, 1989). Sorting time is correlated with previous relationship history such that those who have been in more relationships take less time in sorting random actions.

Other critics of developmental models have discussed the so-called linearity problem in which it is assumed that individuals sequentially progress through stages in a systematic fashion. Some stages may be skipped while others are never experienced (Lee, 1984). Conversely, if developmental theorists allow for random movement, then the process is chaotic and we are left with the old adage of "it all depends on the particular relationships." All relationships would be idiosyncratic without similar progressions of behaviors across relationships. There would be individual quid-pro-quo contracts that are negotiated in each relationship. Yet, memory structures for relationships provide the initial information by which relationship rules are negotiated.

The problem of behavioral falsification, stage parsimony, direction of movement, and rate of movement through relationship phases seemed to signal the demise of developmental models. A saturation point also seems to have been reached with the diversity of models (e.g., Altman & Taylor, 1973; Murstein, 1986; 1987; Knapp, 1984; Baxter, 1985; Duck, 1982; Levinger, 1974; Lee, 1984; Coleman, 1977). The theoretical explanation for movement through stages seemed to be exhausted. The explanations centered on satisfaction of interpersonal needs or assessing the cost-benefits of relationships. Aside from social exchange principles seeking to explain relationship stability on the basis of maximizing profit given available comparison alternatives (Hinde, 1979), the reduction of uncertainty as a drive mechanism was examined (Berger & Roloff, 1982). Stability can be enhanced through the reduction of uncertainty in order to sustain predictable and coordinated interactions. Planalp and Honeycutt (1985) found that in long-term relationships, some events increase uncertainty which may result in the relationship ending as well as becoming more intimate.

The exchange perspective has essentially depicted an image of individual rationality in which individuals evaluate the rewards and costs of a relationship compared to possible alternatives (e. g., being single, being in a different relationship that is seen as more rewarding or less costly). This perspective has been criticized for its neglect of unintended consequences and the emphasis on hedonistic drives (Lannamann, 1991). Relational partners were depicted as conscious decision-makers who could list material, symbolic, and affectional rewards or costs. On the other hand, Berger and Kellermann (1994) discuss how the assumed rationality of individuals in a variety of communication models has gone unquestioned without considering the habitual nature of much of our communication behaviors. Langer (1989) notes how individuals use habitual, overlearned schemata as guides to information processing.

IIs and Mindlessness/Mindfulness

The argument can be made that progression through relationships is the result of behaviors that are routinely enacted such that the person is on "automatic pilot." Once, the pilot light is disrupted, the individual may be more mindful. Some scholars have taken the position that mindful processing precedes mindlessness so that through repeated experiences, new behaviors are assimilated within existing categories (Nisbett & Ross, 1980). In this regard, someone with more dating experience might see increased disclosure as indicative of a particular state of intimacy compared to someone with little experience (Honeycutt & Cantrill, 1991).

How does mindfulness lead to the formation of categories which may lead to subsequent cognitive commitments when a similar scenario is next encountered with mindlessness being reflected in subsequent encounters? One way this can be done in the social arena is by having IIs in which relevant procedural records can be activated. For example, having a retroactive II after a given encounter has occurred puts one in a mindful state while replaying the encounter. Further, the individual may think of contingency plans while imaging what another individual might say in the future. Thus, an II that is concurrently retro and proactive can be used to plan for contingencies. The critical point is that IIs are somewhat mindful, particularly when used for rehearsing messages (Honeycutt, 1991).

The memory structure approach assumes that individuals have particular expectancies some of which are more mindful than others in terms of what should happen in the progression of a relationship that can be used as an anchor prototype for categorizing the type of relationship observed between others as well as one's own relationships (Honeycutt, in press). Relationship expectancies may be reinforced as observed behavior is assimilated into existing categories (e.g., self-disclosure). Occasionally, accommodation takes place in which the expectancies are modified to account for new observed behavior (Planalp, 1985). For example, a "faithful partner" loses trustworthiness after being discovered for infidelity. Regardless if assimilation or accommodation takes place, the person may play over in their mind, images of conversations with their relational partner. These imagined interactions may relive previous encounters and link these to anticipated conversations. The imagined interactions may serve a function of keeping relationships intact, rehearsing for the ending of relationships as well as the initiation of relationships such as planning to ask someone for a date (Berger & Bell, 1988). The function of imagined interaction in creating relationships as well as keeping conflict alive in personal relationships is discussed in the next section.

Imagined Interactions, Conflict, and the Creation of Relationships

Imagine the following encounter: You have had an argument with your best friend and are thinking about the argument while driving on the highway. You go over in your mind what was said and what you will say the next time you encounter your friend. This type of intrapersonal communication reflects what has previously been referred to as a linked imagined interaction (Honeycutt, 1991) in that outside of real encounters, individuals replay prior conversations while preparing for anticipated interactions. Indeed, this is one way by which themes of conversations in relationships can easily be discovered; analysis of II topics of relational partners.

Previous studies of II partners reveal that many of the IIs occur with romantic partners, friends, family members, individuals in authority, work associates, ex-relational partners, and prospective partners (Honeycutt et al., 1989b; 1989-90). However, recurrence of IIs with individuals is one way that relational themes are identified. Some individuals have IIs with a variety of others while others have recurring IIs with the same individuals over selected topics.

Additional functions of IIs include catharsis, self-understanding, and psychological relationship maintenance. IIs can function to create catharsis for the individual by relieving tension and reducing uncertainty about another's actions. An individual may feel better after reviewing an encounter because she or he is able to gain a different perspective on the encounter. Elsewhere my colleagues and I have distinguished between the omniscient and direct perspectives while reliving encounters (Honeycutt et al., 1989). The omniscient perspective is where the individual sees her or him self interacting with someone else in a particular context. It is similar to watching yourself on videotape. The direct perspective is where attention is concentrated on the other much as in everyday conversation. A retroactive, omniscient II may result in a different interpretation of the meaning behind messages in prior encounters. In this way, self-understanding can be enhanced through clarifying thoughts. Caughey (1984) has discussed how inner dialogue may help individuals to retain a sense of values and purpose.

IIs may function to maintain relationships (Honeycutt, in press). Knapp (1984) originally posited increased use of covert dialogue as relationships began to come apart. Partners in these kinds of situations are marking time until the relationship ends or they may remain in this state of stagnation. During this stage, overt dialogue is at a standstill because partners feel they 'know' how an anticipated interaction will go because of previous repetitive interaction experiences going nowhere. Since they 'know' how it will go, it is not necessary to say anything. However, we have found that IIs occur very early in relationships. In addition, positive emotions may be attributed to the excitement that accompanies relational initiation and growth (Honeycutt et al;., 1989b). For example, individuals may imagine pleasant activities with relational partners, such as engaging in small talk, planning dates, and discussing shared interests. Caughey (1984) also discusses cases in which individuals fantasize that they are involved with a celebrity and live in their own private world. There are parasocial relationships in which individuals who have never actually had a relationship with the celebrity imagined talking to television characters portrayed by the celebrities (Rubin & Perse, 1987).

A major way that IIs function to maintain relationships is by thinking about previous encounters and reexperiencing the emotional involvement associated with such encounters. Research by Klos and Singer (1981) reveals that induced IIs may elicit different kinds of emotions in thinking about parental conflict (cf., review Honeycutt, 1991). They studied the determinants of adolescents' ongoing thought following simulated parental confrontations. They examined the effect of fulfilled versus unresolved situations with parents, mutual nonconflictual parental interaction versus mutual, conflictual interaction, and a simulated interaction in which the parents' attitude was coercive or collaborative. It was proposed that exposure through a simulated interaction to these conditions would affect later recurrence of simulation-relevant thoughts (e.g., thinking about the parent).

Individuals were in one of six conditions: (a) collaborative decision-making with parent, resolved, (b) collaborative decision-making with parent, unresolved, (c) collaborative confrontation with parent, resolved, (d) collaborative confrontation with parent, unresolved, (e) coercive confrontation with parent, resolved, and (f) coercive confrontation with parent, unresolved. Subjects engaged in simulated interactions with one of their parents while a research assistant read a predeveloped parental script appropriate to each situation. The experimenter asked subjects to think about the last visit of three days or longer with the same-sex parent.

Coercive confrontation was operationalized as the parent's trying to win the argument while not listening to the subject's viewpoint. Collaborative confrontation was operationalized as the parent's expressing her or his own view while trying to understand the subject's view and being open to a mutually acceptable compromise. Collaborative decision-making was operationalized as the parent and subject working together to find a solution to an interpersonal problem they shared that was external to their relationship. Resolution (nonresolution) of problems was operationalized by having subjects reach a solution (leave the dialogue unresolved) at the end of three imagined interactions. After these IIs, subjects were taken to another room and over a period of 20 minutes, thought samples were elicited by sounding a buzzer at 20 random-intervals in which subjects reported what they were thinking, feeling, or imagining in their minds. Thoughts were coded as simulation-related if they were direct references or associations to the simulation conditions. Affect was measured before and after the simulations using Likert scales (1 = absence of emotion, 5 = strong emotion) reflecting interest, anger, distress, joy, disgust, and contempt. Stress with parents was measured using items assessing the extent of interpersonal conflict and satisfaction of needs such as acceptance, recognition, and support.

Results revealed that after the simulations were over, emotions of anger were higher in the coercive conditions rather than the collaborative conditions. The thought-sample intervals revealed that once exposed to a simulated parental conflict, students with a history of parental stress demonstrated as much as 50% of their later thoughts being concerned with the simulations. Klos and Singer (1981) surmised that the reawakening of unpleasant past experiences is enough to sustain arousal and recurrent thought even if the conflict is resolved. It could be that the thoughts of adolescents who have a history of parental stress may be very unpleasant because many environmental cues (e.g., television plots, film) trigger recurrent conflict thoughts. The outcome (resolved, unresolved) is contingent upon the particular evoked records and its reconstitution in the scenario.

It is interesting that the idea of repetitive, stimulus-relevant cognitions may be related to linked IIs in which multiple functions may be served. In the case of long-standing conflict between parents and children, the conflict may be kept alive and maintained in the absence of the other by having retro and proactive IIs. Even though a retroactive II is experienced, it may be immediately linked with a proactive II (e.g., "Last time, I bit my lip. Next time, I see her or him, I am going to say exactly how I feel."). Given that IIs tend to occur with significant others, it may be that many of them are linked and occur between encounters reviewing and previewing conversations.

The linkage function of IIs explains recurring conflict in personal relationships. Indeed, many marital therapists lament how counseling and intervention may not result in long-term benefits in getting married couples to communicate constructively. Often there is a regression to dysfunctional patterns of communication after intervention has ended (Floyd, 1988). Conflict may be kept alive through retro and proactive IIs that link a series of interactions. There may be the experiencing of negative emotions as one "replays" the encounter (Zagacki et al., 1992). The linkage function of IIs for recurrent conflict helps explain why classroom instruction and counseling sessions on rational models for conflict resolution often fail as individuals regress to previous old ways for resolving conflict (e.g., I win, you lose mentality. "I'm right, you are wrong." "I'm OK, You're not OK." "It is my way or else. There is no need for further discussion. You are wrong."). Old interaction scripts are mindlessly called up from long-term human memory. Thus, conflict episodes may pick up where they last left off despite a period of physical separation. In the meantime, the conflict has been kept alive in the human mind through imaging conversations using the retro and proactive (rehearsal) functions of IIs.

Aside from linking conflict episodes, IIs also can psychologically maintain relationships by concentrating thought on relational scenes and partners. For example, individuals may have a number of IIs with partners who are living in distant communities. Long distance relationships can be maintained through imagining interaction with the absent partner. In addition, internal thoughts including IIs may be associated with memories from childhood and recent experiences which are intertwined in a complicated network of images and expectancies (Kihlstrom & Cantor, 1984; Singer, 1987).

When assuming a functional approach to analyzing IIs, there is an implicit belief that there may be therapeutic benefits at some level whether it be increased self-awareness, tension-relief, or feeling pleasant thoughts about the II. Support for this assumption is available in a study by Schultz (1978) investigating the use of imagery to alleviate depression. Schultz studied four imagery conditions across 60 depressed, male psychiatric patients. One condition labeled the "aggressive imagery procedure" had depressed males recalling someone saying something which angered the self. In the "socially gratifying imagery" condition, the individual was instructed to recall someone saying something which was very pleasing. The "positive imagery" condition had the patient recall a place she or he used to visit in order to relax. Finally, the "free imagery" condition had patients reporting all images, thoughts, fantasies, and ideas which occurred to her or him without trying to direct the thoughts.

Ratings of depressive feelings taken 10 minutes after the imagery induction revealed that the first two conditions produced lower levels of depression than the less socially oriented conditions. It could be argued that the aggressive and socially gratifying inductions forced the patient to have retroactive IIs that were negative and pleasant, respectively. Schultz also reports that in comparing the socially gratifying and positive imagery conditions, depression was lowered after the socially gratifying induction. Whatever the case, IIs may elicit different kinds of emotional affect and through transference linger on for a period of time.

IIs and Emotional Affect

Markus and Nurius (1986) discuss the idea of "possible selves" in which individuals may think about previous and future representations of self and how this relates to the current self. For example, "in the past, I was a student but now I am an accountant. In the future, I could be head of the accounting division." Markus and Nurius believe that this kind of thinking can be viewed as cognitive representation of goals, aspirations, fears, motives, and threats. Thus, self-conceptions may be current selves, past selves that no longer characterize us but could be relevant again, or they may be future selves that have not been realized but that are desired or feared. Markus and Nurius (1986) indicate that the idea of possible selves "is broadly construed as an effort to tie self-cognition to self-feelings or affect" (p. 958). They wonder how motives, goals, and values are communicated within the self. IIs play a role in this communication as procedural records for particular situations are called up and reconstituted.

Klinger (1987) notes how we are more likely to react with emotion when we encounter cues related to the pursuit of some goal. There is also evidence that imaginary experiences have effects on emotion similar to "real" experiences and that images are initiated through motivational as well as emotional processes like "real" behavior (Klinger, 1981). If this is the case, then IIs that serve a rehearsal function should be related to measures of emotional affect insofar as it can be argued that rehearsal is goal-directed and makes us attentive to cue that may produce incongruity between desires and expectancies. Because IIs can be conceived as a planning activity containing corresponding intentions, IIs should be associated with emotional affect.

We have found that pleasant emotions were associated with having more IIs that are similar to actual encounters (Honeycutt et al., 1989-90). This supports Mandler's (1984) ideas that emotions are linked to transference in terms of the incongruity between desires, expectancies, and the information provided in a specific situation. If expectancies are confirmed, we may experience joy or be unaware of emotion. If the situation is one of moderate incongruity, we may feel interest or curiosity. If the situation is of gross incongruity, we may feel surprise or be afraid. We may feel anger or distress if the incongruity persists (Singer, 1987). Positive emotions are more likely to be aroused when incongruity is quickly reduced (Singer, 1984). Humor involves a period of moderate incongruity followed by a resolution which matches our expectations of a "punch line" or "explanation" resulting in smiling or joy.

In some thought-sampling research, the data has revealed that thoughts were concerned with the subject's present life concerns 67% of the time, with past and future concerns 24% of the time. Thoughts about the past tended to involve reviewing previous events and using critical evaluation (Klinger, 1987). Thoughts about the future tended to be in the form of rehearsals and setting up alternative scenarios for anticipated events. According to Singer (1987), this "helps us to maximize the lessons we draw from past experiences and to plan better ways of dealing with upcoming contingencies" (p. 8).

Support for Klinger's (1987) notion of emotion being concerned with current concerns is reflected in some of our findings. For example, we have had coders evaluate the level of emotional intensity (1 = low intensity, 2 = medium, 3 = strong) as well as having respondents to the SII report their own level of emotional feeling after having recalled the most recent II (1 = positive, 2 = neutral, 3 = mixed feelings, 4 = negative) and the level of communication satisfaction they felt with the II (Zagacki et al., 1992).

Strong intensity was associated with low communication satisfaction while medium intensity was associated with high communication satisfaction. Strong intensity was also associated with negative and mixed feelings while medium intensity was associated with positive emotions. In addition, we found that highly intense IIs occurred with romantic partners and family members while less intense IIs occurred with work partners. High intensity also was associated with II topics dealing with relational conflict while medium intensity was associated with topics related to school, work, and small talk (Zagacki et al., 1992).

IIs and Relational Expectancies

Imagined interactions help create relational expectancies and thereby contribute to our memories about relationships. Expectancies for relationships can be envisioned in the form of knowledge or memory structures for relationships. Memory structures for relationships help us make sense of behaviors that we observe in others' relationships, our own relationships, and provide a sense of trajectory about where relationships are going or remain cast in some type such as a long-term friendship or platonic, casual friendship. The question then arises as to the relationship between memory structures for relationships and other cognitive structures as well as the prototypical actions and behaviors that are believed to characterize intimate relationships.

Expectancies for behaviors comprise schemata for relationships. Relationship schemata allow individuals to categorize bits of seemingly-related information into a structure that allows them to recognize future behaviors as indicative of close or distant relationships (cf., Planalp, 1985). We have structures in memory that allow us to interpret behaviors, make sense of seemingly unrelated behaviors and hierarchically organize experiences. Memory structure theory assumes that the habitual nature of behaviors is the result of following the script for a particular goal. More specifically, information processing about events in relationships would be a function of the content of relationship expectancies. The expectancies may be well-developed or relatively unarticulated. The idea that individuals have expectancies derived from previous experiences directs attention toward intrapersonal thought processes. Hence, developmental phases may exist in the minds of individuals. This says nothing about behavioral support for parsimonious categorizations of phases. The developmental models may serve as a heuristic anchor for processing relationship behaviors. The stage concept may prescribe prescriptions for so-called appropriate behaviors.

Knapp and Vangelisti (1992) suggest that the idea of developmental stages is descriptive rather than prescriptive. Stages of intimacy and decay are not seen as good or bad. The content of memory structures for relationships may resemble researcher-imposed stage classifications. The study of behavioral referents for developmental stages has been questioned for the omission of individuals' beliefs for relationship trajectories. Honeycutt and his associates (1989, 1990; 1993) argue that a number of research programs that focus attention on co-temporal behaviors have excluded the analysis of individuals' beliefs and expectancies that ultimately guide ongoing interactions. A memory structure approach allows researchers to examine the anchor that as guides for interpreting relationship behaviors. The memory structures may be more or less developed and vary as a function of gender, experiences, and even current relationship quality (cf., Duck, Leatham, & Pond, 1991).

Duck and Miell (1986) indicate that persons have knowledge about relationships even in the absence of an actual relationship partner. Relationships can be viewed as changing mental and behavioral creations of persons who are bonded through communication. Duck indicates that the daily talk between relational partners is the observable relationship. Yet, persons have trajectories for relationships in the form of expectancies for relationships.

In order to explain how individuals interpret a series of co-occurring behaviors in a situation, Schank and Abelson (1977) discuss the notion of scripts.

General expectancies for relationships consist of underlying memory structures that represent relational activities, behaviors, or events. Memory structures are composed of a variety of related scenes. Memory structures represent scenes that are not contextually bound. For example, dating could contain scenes of recalled first dates (e.g., underlying scripts for greeting, small talk at a

Figure 1

Hierarchical Relationship Between Meta-Memory Structures, Memory structures, Scenes, and Scripts for Memory About Events in Escalating Relationships

Meta-memory

Structure (Escalating relationship)  

     

Memory

Structures

(Small

talk)		 Dating Arguing

 

 

(Talking

 

 

Initial date

 

 

Talk about political issues

Scenes

on phone   at a social gathering

with friend)

Meeting a

stranger

 

Recall of last

date

 

Disagree on division of family household duties

Scripts

(Hello, Hello, Get in Ask partner opinion on some
 

Talk about

weather,

activities,

"nice to talk

with you, See

you later, Bye."

car, Arrive at destination, Buy food, Talk, Return home issue (social welfare), Assess strength of their opinion, Tell partner your beliefs, try to persuade him/her on weaknesses in their position and why your position is "right"
 

Hello, Nice

to meet you,

Where are you

from? how do

you like it

here?

Hello, Pickup date Arrive at movie, go to cafe, Talk, Return home Tell spouse it's their turn to clean up the kitchen. Discuss inequity of cleaning duties, Indicate that both must share in duties in order to be fair or equal

 

particular location), double-dating, informal dating, and so on. The underlying scripts can be accessed in terms of retroactive IIs from previous experiences and encounters. Figure 1 presents a memory structure model based on imagined dialogue in categorizing events in memory.

Collections of memory structures constitute what Schank (1982) has referred to as meta-MOPs. A meta-memory structure for the development of personal relationships could contain underlying memory structures for small talk, dating, and brief periods of arguing or disagreement. An escalating meta-memory structure would contain relational actions that are not bounded by particular scenes. Yet, the actions could point to particular scenes (e.g., disclosing about a sibling's chemical dependency in the girlfriend's apartment). Scripts define the sequence of actions in recalled scenes in this hierarchical arrangement of knowledge structures. Scripts represent specific instantiations of more general scenes. Scripts help create plans for accomplishing goals. For example, an individual may have a script for things to talk about on a first date at a restaurant which is part of a dating scene. A person may have a script for disclosing information to someone about a social issue in order to determine compatibility of attitudes.

An ordered array of scenes constitute a memory structure. For example, a memory structure for dating could contain scenes of first dates (and underlying scripts for greeting, small talk at a particular location, etcetera), double-dating, informal dating, and so on. According to Schank (1982), the primary function of a memory structure is processing new information that can be seen as reflecting a particular type of structure that will in turn provide expectations necessary to understand what is being received. For example, a dating memory structure could include a number of dating scenes and would lead the individual to have expectancies of what would happen on "typical" dates.

Honeycutt (1993) provides an example in which a dating memory structure for one person may be accessed by recalling an initial date at a movie. The memory structure may include a movie-date script including actions of waiting in line, ticket purchasing, buying refreshments, locating seats, watching the movie, and leaving the theater. The scene then changes to a restaurant. The theater and restaurant scenes comprise the dating memory structure. The dating memory structure is one of many memory structures that result in the expectancies for the development or deterioration of relationships. The dating memory structure may be part of de-escalating expectancies as an individual imagines spending time with alternative partners (Honeycutt, 1993).

Schank (1982) indicates that memory structures occur at physical, societal, and personal levels. He defines a societal scene in terms of a social relationship between individuals who are "pursuing a goal that the other person is a necessary participant in, at a common time, with a communication link between them. The actions comprising the interaction between the participants defines the scene" (Schank, 1982; p. 96). Physical scenes represent mental or visual images of our surroundings at a specified time. Personal scenes are idiosyncratic and may be thought of in terms of repetitive, private plans.

The division of memory structures into these categories results in the idea that behavior can be processed by asking: What physically happened? Where did it happen? (physical scene) What societal conventions or norms were used? What effect did the behaviors have on the individual's social position? (societal scene) Were the individuals personally effected? What goals were achieved by the behaviors? (personal scene) Thus, a dating memory structure may reflect an encounter at a movie theater (physical scene). The dating memory structure could reflect disclosure at a restaurant so that the persons become more intimate (societal and personal scenes). IIs contain elements of these various scenes as individuals imagine specific settings in which the encounters occur (physical scene) and lines of dialogue (personal scenes).

Pryor and Merluzzi (1985) conducted a series of studies asking persons about the expected sequences of actions for "getting a date" as well as the actions for the "first date." Individuals were initially asked to generate a list of 20 actions that typically occur when a man asks a woman out for a date and the events that would occur on the date. Typical actions were: the male noticing the women, staring between the man and woman, smiling, and finding out about other person from friends. Other actions included manipulating ways to accidentally run into the person, be introduced by a friend, male beginning a conversation, finding similar interests in the conversation, male asking the female for her phone number, and phoning her to ask her. Events for the first date included arriving at the female's domicile, meeting parents or roommates, small talk, going to a movie, buying refreshments, going to get something to eat after the movie, going back home, summarizing the night at the end of the date, kiss, and saying good night.

No differences were found between experienced and novice daters for the frequency of

mentioning the various actions. The only observed difference in terms of expertise was in processing the expectancies in an intuitive "logical" sequential order. A second group of subjects had each action listed on an index card that was presented them in a random deck. Subjects were told to sort the deck into an intuitive logical order. Individuals with more dating experience were able to arrange the randomly organized sequences of dating actions in a shorter time period compared to persons with little dating experience.

The first date study represents a detailed analysis of one of the actions associated with a larger meta-memory structure of escalating relationships. The dating memory structure contains scenes at home, driving the car, being at the movie, going to a restaurant, and returning home. Within these scenes, scripts may be accessed (e.g., script for ordering a meal at the restaurant, script for saying good-bye). The hierarchical structure of these types of knowledge structures (meta-memory structures, memory structures, scenes, scripts) provides a heuristic device that reveals how isolated events such as a single date or an argument can be viewed in broader relational terms within the idea of a meta-memory structure.

Our studies have revealed that the most frequently mentioned escalating action is initial meeting, with 97% of respondents mentioning this action (Honeycutt et al., 1989). The next highest mentioned action was dating (92%). The least mentioned escalating actions were overcome crisis (13%) and talking about future plans as a couple (16%) even though both actions were mentioned by a separate group of subjects as highly typical and necessary. Each of the expectancies can have underlying scenes and scripts.

The prototypical de-escalating memory structures have revealed that the most frequently mentioned de-escalating action was final break-up (58%) followed by avoiding the other (40%). The least mentioned de-escalating actions were talking with friends about relationship problems (10%) and giving excuses for not being able to go out (11 %). These actions were also rated as relatively typical though giving excuses was seen as less necessary (Honeycutt et al., 1992).

We have found that individuals identify six underlying scenes of relational escalation (Honeycutt et al., 1989). Subjects read one of two stories about relationships that started to grow and ended in marriage. The stories contained sentences representing the prototypical actions identified as representative of escalating relationships. The subjects were instructed that, "Some people feel that romantic stories may be divided into several natural parts or stages. We would like you to carefully read the story and decide whether this story may be divided into different parts" (Honeycutt et al., 1989, p. 84).

The first scene reflected the social penetration stages of initiation and experimentation by meeting, engaging in small talk, and phoning the other. The second phase consisted of dating, showing physical affections, sharing time together and seemed to reflect the social penetration stage of intensifying (Knapp & Vangelisti, 1992). Self-disclosure was a separate phase. The fourth phase referred to sexual intercourse. The fifth phase consisted of meeting partner's family members, exchanging jewelry and other gifts, making a verbal commitment, and speaking in the second person by making other-oriented statements reflecting interest in the other's goals. This phase resembles a combination of the intensifying and integrating stages discussed by Knapp and Vangelisti (1992). The final scene was marriage and simply mirrored the social penetration, bonding stage.

Individuals also identified six scenes of relational dissolution (Honeycutt et al, 1992). Subjects read one of two stories about relationships that started to sour based on the prototypical actions identified as representing decaying relationships. The first scene consists of stopping self-disclosure. The second scene represents disagreements and arguments over small things. The third scene represents the avoidance of contact with the partner while the fourth scene reflects reevaluation of the relationship in terms of examining alternative situations. The fifth scene has been referred to as "breaking up" as the relationship is ending and interests in others rises. The final scene simply marks the dissolvement of the relational bond. In addition, the most critical events in the dissolutionment of relationships were stopping the expression of intimate feelings, arguing about little things, fighting and antagonization of the other, talk about breaking up, starting to see others, and the final breakup.

Summary

Individuals have memories about personal relationships derived from imagined interactions that reflect direct and vicarious experience. Memories about relationships can be reflected in terms of expectancies for the rise and demise of relationships. There are a number of sources of information that create memory structures for interpersonal relationships. Of particular interest is the role of imagined interactions; internal covert dialogues that allow individuals to think about relationship partners, scenes, and topics while developing plans for encounters with future relational partners. Research has revealed how conflict may be kept alive in personal relationships through the linkage function of IIs. This occurs when an individual replays a previous conflict while rehearsing for the next anticipated encounter such that the conflict may pick up where it left off at the end of the last encounter.

The memory structures for relationships resemble some of the stages identified in earlier developmental models. Thus, the idea of relational phases or stages exist in the minds of individuals. We remember various relational events through retroactive IIs. The claim can be made that relational stages function as "cognitive maps" by which individuals can chart their location (stage and/or type of relationship) and movement (direction) in their relationships. To use a metaphor, relational memory structures allow individuals to be relational navigators in the jungle of personal relationships. The navigation metaphor is helpful because individuals may recognize that while relationships progress in an infinite number of ways, there is also a prototypical path that they can use as a reference point in order to guide them in current or anticipated relationships (Honeycutt, in press).

Note

1Duck and colleagues (1991) discuss five different ways of partitioning accuracy about relationship symbols. Insiders and outsiders may differ in their perceptions of the meaning of relational symbols and that two insiders can have different perspectives without there being any methodological or empirical problem.

References

Response from Leonard J. Shedletsky

The central concept in Honeycutt's memory structure approach is the memory structure. What is the memory structure? Are memory structures psychologically real? Are they coherent organizations of memories for specific relationships or for generalized relationship beliefs, norms, etc. Is memory structure just another way of saying memory or is there something special about memory structures? Must we choose between memory structures and developmental stages, or can these theories co-exist? What exactly is the link between IIs (Imagined Interactions) and memory structures? What precisely are "scenes"? And, finally, why is navigation the metaphor offered to describe memory structures, as opposed to cognitive maps used to describe relational stages. In short, how does reference to memory structures add to Schank's idea of scripts or to the well established idea of long term memory?

Author's Last Word

Memory structures represent behaviors or events that individuals use to interpret sets of co-occurring actions. When asked to recall the last date, the individual could access a physical scene (where it took place), a social scene (verbal messages "I enjoy being with you"), or a personal scene (A goal was to have a good time and escalate or maintain intimacy with the partner). Memory is hierarchically organized. Hence, memory is dynamic and changing depending upon what scenes are accessed at given time points. The questions posed by Leonard Shedletsky are good for initiating discussion of the interpretation of relational events aside from the questions I posed in the "Perspective" preceding the article. I have answers to his questions that can be found in the articles as well as through deduction. However, I hope that readers will be creative and motivated in class to answer these questions by applying their experiences to the recall of relational events and expectations for the rise or demise of intimate relationships. JMH

Inner Voices as Alter-ego Mental Constructs: Auditory Verbal Hallucinations

Mark A. Hamilton

About the Author: Mark A. Hamilton is on the faculty of the University of Connecticut, Storrs, CT 06269-1085, telephone (203) 486-4569 or (203) 423-7410.


Abstract: Auditory Verbal (AV) hallucinations are unintended verbal sensations that are perceived as a voice. AV hallucination is linked to the invention and reification of an alter-ego mental construct often referred to by those who experience it as an "inner voice." Results of a survey (N = 650) confirmed that normals frequently create alter-ego mental constructs, and that the content of the messages from these constructs reflects the operation of conscience and ego-defense.


Perspective:


1. Do you, like the general population, commonly interpret unintended verbal sensations as an inner voice? If so, do you think they are Auditory-Verbal (AV) hallucinations or mental events confused with auditory hallucinations?

2. Why are many people reluctant to discuss inner voices?

3. Why do people perceive such a variety of sources of inner voices (e.g., God, the Devil, all-powerful individuals, dead relatives, animals, inanimate objects, conscience)?

Auditory-Verbal hallucination is an internally generated but unintended sensation perceived as the utterance of a voice. Normals as well as schizophrenics experience AV hallucinations. Empirical evidence indicates that the propensity to attribute unintended verbal sensations to hallucinated voices is continuously distributed in the population (Chapman & Chapman, 1980; Harrow, Marengo & Ragin, 1985; Jaynes, 1986; Sarbin, 1970; Sedman, 1966a; Smythies, 1956; Strauss, 1969). This conclusion is based on both survey data (Barrett & Etheridge, 1994; Bentall & Slade, 1985; Foulkes & Fleisher, 1975; Mott, Small & Anderson, 1965; Posey & Losch, 1983; Sidgwick, Gurney, Meyers & Podmore, 1894; Young, Bentall, Slade & Dewey, 1986) and case studies (Gordon, 1941; Forrer, 1960; Hamilton, 1985; Medlicott, 1958; Suedfeld & Mocellin, 1987).1 These findings indicate that members of the general population commonly interpret unintended verbal sensations as an inner voice. The messages attributed to an inner voice may be due to true auditory hallucinations, or to mental events confused with true auditory hallucinations, known as pseudo-auditory hallucinations. In either case, the person experiencing the inner voice invents a non-ego source to which the messages can be ascribed. The present study examined normals' propensity to create alter-ego mental constructs, and the types of constructs they might create.

Generating Alter-ego Mental Constructs


Auditory-Verbal (AV) hallucination produces a loss of ego-boundaries that must be reconstructed with reality testing (Gazzaniga, 1985). A lack of reality testing allows alter-ego attributions of hallucinated imagery to flourish. An occasional unintended thought can be attributed to environmental factors such as stress, fatigue or drugs, but persistent unintended voices necessitate the invention of a mental construct (Gazzaniga, 1988, pp. 74-79; Sedman, 1966b).

If the consequences of experiencing an AV hallucination are rewarding for the individual, then this should lower the critical threshold for future hallucination (Slade, 1976). Self-reinforcing hallucinations can generate a positive feedback loop in which lowered threshold and frequency operate synergistically. Mott and colleagues (1965) found that among hallucinators who were normal, 34% had a positive reaction to AV hallucination, 41% had both a positive and negative reaction, 24% had neither a positive nor negative reaction, and none had a purely negative reaction. Thus, nearly three-quarters of their hallucinators claimed to derive some reward from the experience.

By inventing an alter-ego mental construct that is assigned responsibility for the inner voices, the hallucinator may be able to reestablish ego boundaries. If the existence of alter-ego mental constructs helps to reduce the anxiety associated with hallucinating voices, then use of the mental construct will be reinforced, securing its role in the person's mental life. As a consequence, reality testing is further diminished, forming a positive feedback loop that fosters auditory hallucination.

To invent an alter-ego mental construct, the person can look inward, and make an internal attribution of the experience, or look outward, and make an external attribution. An internal attribution requires inventing a mental construct that is the source of the inner voice. An external attribution involves identifying a source of the inner voice existing in the person's environment. Judkins and Slade (1981) found that hallucinating psychotics located the voices they experienced internally (27%), externally (31%), or both internally and externally (42%). The decision to make either internal or external attribution is unrelated to psychosis (Junginger & Frame, 1985), but is probably linked to variables such as religiosity (Chapman & Chapman, 1980), self-esteem, or paranoia.

Religious people who have AV hallucinations tend to confuse them with the voice of God or the Devil (Hamilton, 1985; Jaynes, 1976; Medlicott, 1958; Miller, Johnson & Richmond, 1965; Mott et al., 1965; Sedman, 1966b). Extremely paranoid people who experience AV hallucination tend to interpret it as a message from all-powerful individuals (Sarbin, 1970) or institutions. Less common interpretations of AV hallucination include visitations from dead relatives (Suedfeld & Mocellin, 1987), instructions from animals or insects (Mott et al, 1965), or even dialogue with inanimate objects such as clothes.

Construct Types

A common reaction to AV hallucination is to create a mental construct referred to by those who experience it as an inner voice. Hunter (1984) validated an instrument to assess a person's inner voice experience. This instrument contained three subscales. The first subscale measured the person's propensity to experience an inner voice. The second subscale measured the extent to which the inner voice expressed the concerns of conscience. The third subscale measured the extent to which the inner voice guards the ego against external threat. Thus, the second and third subscales reflect the typical mental activities of the judicial function of conscience and ego-defense.

The three scales formed a Guttman simplex, such that scoring high on the guard subscale necessitated scoring high on the conscience and propensity subscales. Scoring high on the conscience subscale necessitated scoring high on the propensity subscale. This meant that the majority of those who experienced an inner voice received messages related to conscience, and most who received messages related to conscience also received messages related to ego-defense. Hence, there were two dimensions to dialogue with alter-ego mental constructs, one rooted in the conscience, and one rooted in defense of the ego against external threat.

Hunter (1984) used the product rule to test for a hierarchical relationship among the three subscales. If the three subscales form a Guttman simplex, then the correlation between scores on the propensity and guard subscales (r [339] = .72) should be the product of the correlation between the propensity and conscience subscales (r [339] = .89), and the conscience and guard subscales (r [339] = .85). The obtained correlation of .72 is within sampling error of the predicted correlation of .76. The Guttman simplex pattern suggested a four-group typology: those without mental constructs, those with guard and conscience constructs, those with a conscience construct, and those with an unidentified construct.

A low score on all three subscales indicates the absence of alter-ego mental constructs. A high score on all three subscales indicates the creation of guard and conscience constructs. A high score on the propensity and conscience subscales, accompanied by a low score on the guard subscale, indicates the creation of only a conscience construct. A high score on the propensity subscale, accompanied by low scores on the conscience and guard subscales indicates the creation of an unidentified construct. Of the 339 respondents surveyed by Hunter (1984), 57% would count as having conscience and guard constructs, 7% as having only a conscience construct, 7% as having an unidentified construct, and 29% as not having an alter-ego mental construct. Thus, 71% of the Hunter sample were categorized as having an inner voice construct. This figure is consistent with the results of Salzman (1983), which classified 65% of respondents as having an inner voice construct. The present study explored the four-group typology.

Method

Participants were 410 female and 240 male undergraduates in basic social science courses at two major Eastern universities. They read a short description of experiencing an inner voice, then filled out a questionnaire designed to place them in one of the four groups of the Hunter (1984) typology. Following administration of the survey, participants were debriefed.

Materials

The materials used in the study were adapted from Hunter (1984). Before filling out the questionnaire containing the inner voice items, participants read a two page description of types of conscious thought. The two page text appears in the Appendix. This text is a shortened version of that used in the Hunter (1984) study. Its format is similar to that used by Posey and Losch (1983), and Barrett and Etheridge (1992; 1994). The text describes experiencing an inner voice, and contrasts it with conscious thought without an inner voice. It was descriptive and anecdotal, but nonevaluative.

The survey instrument was based on the one used by Hunter (1984).2 Items were presented in 5-point Likert format. Item scores could range from strongly agree (coded +2) to strongly disagree (coded -2). The propensity, conscience and guard subscales of the Hunter instrument had coefficient a reliabilities of .89, .88, and .76, respectively.3

Results

Item quality was assessed using the confirmatory factor analysis program CFA (Hamilton & Hunter, 1992). Three criteria were used to evaluate items: item content, internal consistency (do items measuring the same underlying construct relate to one another in the same way), and parallelism (do items measuring the same underlying construct relate to outside variables in the same way). As expected, the propensity, conscience and guard subscales emerged from the confirmatory factor analysis. The 6-item propensity scale had an alpha reliability of .84. The 14-item conscience scale had an alpha reliability of .92. The 9-item guard scale had an alpha reliability of .80. One of the items designed to load on the conscience subscale (I think it is important to obey rules, even if I will not be rewarded for following them) was shown to lack internal and external consistency, and was consequently discarded.

Scores on the propensity subscale were uncorrelated with being female (r [650] = .04),4 but were correlated with scores on the conscience subscale (r = .86, p < .001) and guard subscale (r = .61, p < .001). Scores on the conscience subscale were correlated with scores on the guard subscale (r = .78, p < .001). Based on the product rule, the correlation between the propensity and guard subscales should be .67. The predicted correlation of .67 was not significantly different from the obtained correlation of .61 (Chi-square 1 = 1.66, p = .20). Thus, the correlations among the three subscales fit the hierarchical model implying a Guttman simplex.

The correlational finding provided tentative support for the four-group typology. The items that emerged from the confirmatory factor analysis were unit weighted and summed to generate scale scores. These scale scores were then used to test the four-group typology. Mean scores on the propensity subscale (M = .22, SD = .92), conscience subscale (M = .12, SD = .77), and guard subscale (M = -.18, SD = .73) reflected the hierarchical relationship among the three scales. Respondents who obtained a positive score on a scale were classified as high on the scale; a negative score or zero was classified as low on the scale. Of the 650 respondents, 176 (27%) were categorized as possessing conscience and guard constructs, 148 (23%) were categorized as possessing only a conscience construct, 37 (6%) were categorized as possessing an unidentified construct, and 274 (42%) were categorized as possessing no alter-ego mental constructs. Only 15 (2%) of those scoring high on the propensity scale defied the proposed typology by scoring high on the guard subscale scale but low on the conscience subscale. Thus, the four-group typology fit the data. Among those claiming to have a mental construct, there was a strong correlation between possessing a conscience construct and possessing a guard construct (r [376] = .48, p < .001).

Discussion

The study found that 58% of normals report possessing an alter-ego mental construct. This rate is comparable to the 65% found by Salzman (1983) and the 71% found by Hunter (1984). The results of the present study support the four-group typology. All but 15 out of 650 respondents were successfully categorized into the four groups proposed by Hunter (1984). The number of unidentified constructs was quite small in the current study (6%) and in the study by Hunter (7%). The only substantial difference between the results of the current study and that of Hunter (1984) was that the percentage of respondents claiming to have a conscience construct but not a guard construct was greater in the present study. This difference could be due to variance in the sample. In the final sections, the nature of the conscience and guard constructs is explored.

The Conscience Construct

The conscience represents the internalization of social rules. For those who have their conscience channeled through an inner voice, appropriate social behavior is praised and inappropriate social behavior is criticized. Guilt for these individuals consists of a period of chastisement from an inner voice. By its nature, the conscience emphasizes group norms over individual needs. Hence, a conscience construct is prosocial, and should be perceived to have a positive influence on behavior. Normals experience inner voices which advise them (Medlicott, 1958), criticize their bad behavior (Hamilton, 1985; Linn, 1977) and praise their good behavior (Hamilton, 1985). Their inner voice may give the pros and cons of engaging in a behavior (Linn, 1977), or attempt to regulate their emotion if they become too upset (Hamilton, 1985). The clinical literature contains examples of patients who experience an inner voice that gives them advice (Gould, 1950), tells them they have been bad, or comments on everything they do (Sedman, 1966b). Sometimes these patients even label the voice their "conscience" (Sedman, 1966b), although other times they associate the voice with that of God or the Devil (Hamilton, 1985; Hill, 1936; Medlicott, 1958; Miller et al, 1965; Sedman, 1966b).

The Guard Construct

The guard construct defends the ego from external threat. An aspect of the guard construct unexplored by the current study is defense of the ego against internal threat. The clinical literature is filled with examples of patients who experience an inner voice threatening them (Linn, 1977; Mott et al, 1965), insulting them (Allen, Halperin & Friend, 1985; Gould, 1949; Hill, 1936; Linn, 1977; Sedman, 1966b; Slade, 1972), or encouraging them to engage in behavior they find immoral (Gould, 1949; Linn, 1977; Sarbin, 1970), such as sex acts (Gould, 1950) or suicide (Allen et al, 1985). Similar messages are received by normals who have created an inner voice. For instance, Gordon (1941) cites normals who heard a voice making offensive remarks about their character.

Notes

1Those who argue that only schizophrenics hallucinate voices can still be found (Allen, 1986; Brett & Starker, 1977; Linn, 1977; Rund, 1986), though they are becoming increasingly rare. Studies that report verbal hallucination in normals as rare have likely underestimated its frequency, given that Western culture has stigmatized auditory hallucination (Foulkes & Fleisher, 1975; Smythies, 1956; Sarbin, 1970).

2Need for social approval appears to have little effect on responses to the propensity subscale (Hunter, 1984), or reports of AV hallucination (Barret & Etheridge, 1992).

3A prestudy (N = 427) employing the 20 item Hunter survey instrument showed that the reliability of the three subscales could be improved by adding items, particularly to the 4-item guard subscale which had a reliability of only .68.

4Data from four studies allowed an estimate of the relationship between gender and AV hallucination. The effect size between being female and hallucination was calculated for the four studies: Mott et al.(1965, r [50] = .37); Launay and Slade (1981, r [200] = .03); Barrett (1991, r [601] = .10); and Chapman and Chapman (1980, r [66] = -.19). Meta-analysis (Hunter & Schmidt, 1990) was used to determine the average weighted effect size, and the standard deviation (corrected for sampling error) across studies. There was a significant amount of variance among the four studies (CHI-SQUARE3 = 9.88, p < .05), with an average correlation of r (917) = .08, and a corrected standard deviation of .08. It would appear the Chapman and Chapman study is an outlier. If that study is eliminated, the average correlation between being female and AV hallucination is homogeneous (CHI-SQUARE2 = 4.72, p > .05). The average correlation was r (851) = .10 (p < .01), with a corrected standard deviation of .04.

References

Appendix

Text Describing Types of Conscious Thought

A Discovery About Conscious Thought

Several years ago, a well-respected psychologist made an intriguing discovery. As he was talking to a female friend of his, he jokingly referred to something bad she might do. His friend replied, "Oh, I couldn't do that, my conscience would torture me for days." When he heard the word "torture," the psychologist became curious. He asked her, "What do you mean by torture? What goes on in your head when your conscience tortures you?" What she said astounded the psychologist (although many people would not be surprised). She said, "My conscience would just call me awful names for hours on end." That statement floored the psychologist, because what he understood her to say was that she heard a little voice in her head that talked to her. After a long conversation, this turned out to be true; her conscience was a distinct voice inside her head that would spew out a stream of insults such as "filthy, worthless, dirty, ungrateful."

As the conversation continued, the psychologist made a confession to his friend that stunned her as much as her confession had surprised him. He told her that he had never heard any kind of voice in his head. (Many people will not be surprised to hear that). To the psychologist, the word "conscience" had never been any more than a metaphor for "thinking about the ethical consequences of an act." By "thinking," he meant a silent process of internal thought generation. Only when he was playing a role or practicing a lecture did he engage in the subvocal speech that sounded to him like his own voice speaking (which, of course, it was). His friend asked him, "Do you mean that your head is just blank with nothing going on?" The psychologist replied that was often true, though silent thought is still thought.

In the years since his conversation with his friend, the psychologist checked through everyone that he knew well enough to ask such a personal question. First, he told his wife what his friend had said. His wife responded "Of course, everybody has a conscience." The psychologist replied, "But you mean a little voice in your head that is always talking to you?" She said, "It doesn't always talk, only when it wants to tell me what to do or what to say or what not to do or what to be afraid of. . ." She said her voice would only shut up when she was reading or doing a lot of heavy thinking (for example, math problems or serious introspection).

Many of the psychologist's friends had "little voices" that talk to them in their heads, and many of his friends did not. Some of those who have "little voices" have voices of the opposite sex, some have voices of the same sex, and some say that they can't really tell the voice's gender. Moreover, some people have more than one little voice. They have a little "devil" as well as a "conscience." That is, some people hear one little voice that eggs them on to try new and "scary" or "bad" things that might be fun or might bring pleasure, and another voice which makes them feel guilty for trying such behaviors.

Types of Conscious Thought

It has been generally believed that all conscious thought is the same. What the psychologist discovered seems to indicate that there is at least one aspect of conscious thought which varies. What follows are descriptions of two types of thought.

Often, conscious thought is experienced as a clearly verbal process, that is, one in which thoughts are put into words and experienced as thought they were vocalized. Several distinct lines of thought may occur at the same time, producing different views, positions or attitudes regarding any subject under consideration. This type of thinking can be easily described since it has already been put into words, but difficulty often arises over the fact that there are actually several thoughts running on parallel tracks simultaneously, sometimes taking the form of an inner dialogue or voice. The conclusions drawn on the one track do not necessarily prevent contradictory conclusions being drawn at the same time on other tracks. "Conscience" within this type of thought typically takes the form of a "voice" speaking or giving advice from within one's own mind. For many people, feeling guilty is perceived as a period of internal criticism from such a voice. It is common for a voice or thought track to be uncomfortable to the individual, and to spontaneously take on a kind of independent existence.

Conscious thought may also be experienced as a clearly nonverbal process, that is, one in which thoughts are silently juggled, merged, redefined and reordered, out of which are produced conclusions. Thinking of this type is not readily subject to description, as it is a flow of concepts and relations which is not generally perceived as something which can be translated into verbal messages. "Conscience" within this type of thought refers to an abstract set of ethical and moral beliefs, and guilt is perceived as a general feeling of distress. For people who constantly experience nonverbalized thought, there may be periods where there is no particular line of thought, especially conscious subvocalized thoughts, going through their minds.

After reading these descriptions, you may have come to feel yourself associated to some extent with one type of thinking or the other. It is possible that you engage in only one type of these two types of thought or both. Neither kind of conscious thought is "better" or "worse" than the other, they are simply different. Nor is one form of thought "abnormal."

IT IS UNDERSTOOD THAT EXAMINING ONE'S MENTAL LIFE IS AT THE SAME TIME EXCITING AND SENSITIVE. WITH THIS IN MIND, WE ASK YOU TO HELP US BY FILLING OUT THE QUESTIONNAIRE AS OPENLY AND HONESTLY AS POSSIBLE. YOUR RESPONSES WILL, OF COURSE, REMAIN CONFIDENTIAL.

IF A QUESTION MAKES YOU UNCOMFORTABLE, OR YOU WOULD RATHER NOT ANSWER IT, FEEL FREE TO LEAVE IT BLANK.

Response from Leonard J. Shedletsky

It is difficult to know just how to describe one's own thought processes, voices, silent thought, nonverbal processes, or some combination of these. Given the choices between "little voices" and "nonverbalized thought" could the subject be biased to reconstruct the experience one way or the other? How do we get at the psychological reality, as opposed to the report, of thought?

 

Verbal Hallucinations In Normals:

A Research Agenda

Terry R. Barrett

About the Author: Terry R. Barrett is a professor in the Psychology Department at Murray State University. The preparation of this essay was supported in part by grant 2-12833 to the author from the Committee on Studies and Institutional Support, Murray State University. Reprints may be requested from the author at the Psychology Department, Murray State University, Murray, KY 42071.


Abstract: The present paper argues that verbal hallucinations are a relatively frequent occurrence for a substantial minority of the general population. A number of research questions are proposed to enhance our understanding of this unique conscious experience.


Perspective:

1. The idea that individuals free of mental illness experience verbal hallucinations is not commonly accepted by mental health professionals. What sort of research project (s) do you think might provide the evidence necessary to convince these professionals that verbal hallucinations are a normal part of conscious experience for some people.

2. You have just met a professional researcher who says she has the resources to pursue only one of the projects suggested in the paper. Decide which project you think she should pursue and then develop an argument that might convince her to proceed in that direction.

A recent edition of the New Statesman & Society magazine (Linehan, 1993) reported a conference held in Manchester, England by the Hearing Voices Network. This organization promotes the idea that verbal hallucinations (hearing voices outside the head when no one actually said anything) should be considered a normal aspect of conscious experience rather than as an indication of mental illness. Although this effort flies in the face of the historically accepted understanding of hallucinations (e. g., Asaad, 1990; Sarbin & Juhasz, 1967; Slade & Bentall, 1988), recent research is beginning to support their claim.

In 1976, Jaynes proposed the startling hypothesis that verbal hallucinations were a regular part of human experience no more than 5000 years ago. Because of a change in the organization and function of consciousness, however, the ability to experience verbal hallucinations was lost. Motivated by Jaynes' book, Posey and Losch (1983) assessed 375 college students and discovered that 71 percent reported the occurrence of at least one verbal hallucination. Barrett and Etheridge (1992) confirmed these findings and suggested that 25 to 30 percent of college students may experience verbal hallucinations as frequently as once a week.

An important area for future research is the development of a taxonomy of verbal hallucinations. Currently, we have little more than anecdotal evidence about the characteristics of normal verbal hallucinations (e. g., affective tone, semantic content, and type of voices). The situation is not much better with respect to a description of verbal hallucinations associated with psychopathology (however, see Aggernaes, 1972; Goodwin, Alderson, & Rosenthal, 1971; Junginger & Frame, 1985; Junginger & Rauscher, 1987; and Mott, Small, & Anderson, 1965). We also need intergenerational and cross-cultural studies to examine the universality of verbal hallucinations in normal individuals.

We are just beginning to explore the individual differences related to the experience of verbal hallucinations (e. g., see the paper in this volume by Miner and colleagues for an extensive exploration of one set of individual difference variables). Barrett and Etheridge (1994) demonstrated that verbal hallucinators exhibit heightened negative affect and feelings of rejection and incompetence in social situations compared to non-hallucinators. In addition, Etheridge (1990) suggested that verbal hallucinators, as a group, may be relatively high on the dimension of fantasy proneness (Wilson & Barber, 1983). These findings indicate that work comparing verbal hallucinators and non-hallucinators on a variety of standard personality dimensions might prove to be quite productive.

Differences in developmental histories may also be an important cue to unraveling the mysteries of verbal hallucinations. Etheridge (1990) found that verbal hallucinators tended to describe their childhood experiences in more negative terms than did non-hallucinators. However, Bone (1992) was not able to replicate these findings. Clearly more work is needed to determine if a particular set of childhood experiences are related to verbal hallucinations during adulthood.

Classic views of hallucinations often associated hallucinations with vivid imagery. In support of this position, Barrett (1993) found that verbal hallucinators rated their imagery as more vivid than did non-hallucinators. These preliminary findings suggest that a systematic comparison of verbal hallucinators and non-hallucinators in a variety of imagery tasks (e. g., mental rotation and mental paper folding tasks) might yield important insights.

Data obtained from the various research projects proposed above should suggest important hypotheses regarding the cognitive mechanisms that underlie verbal hallucinations. At present, however, there seems to be two preliminary hypotheses that deserve attention. One is the possibility that imagery processes are somehow related to verbal hallucinations. Exactly what those imagery processes are and how they might function to produce verbal hallucinations, however, has yet to be determined. The second hypothesis was suggested by Slade and Bentall (1988) and elaborated by Bentall (1990). According to this hypothesis , hallucinations represent the failure of a cognitive mechanism monitoring the contents of consciousness for the purpose of discriminating between internally generated events (e.g., images, thoughts) and perceptions (our representations of external events). Verbal hallucinations, on this hypothesis , are internally generated events mistakenly judged by this mechanism to be perceptions. Bentall, Baker, and Havers (1991) have produced some evidence in support of this position.

The Hearing Voices Network seems to be on the right track in claiming that verbal hallucinations, in and of themselves, do not indicate psychological illness. Although this conclusion has not yet been embraced by the general mental health community, there are signs of increasing awareness of this fact (Slade & Bentall, 1988). Indeed, one recent author (Watkins, 1986) thinks we make a mistake taking as a goal of therapy, the elimination of verbal hallucinations. Instead, she makes the intriguing argument that emotional health and creativity will be greatly facilitated by encouraging the development of intentional dialogues with our "imaginal others."

References

Response from Leonard J. Shedletsky


With 71% of college students reporting the occurrence of verbal hallucinations and 25 to 30% of college students experiencing verbal hallucinations as frequently as once a week, we need to know more about verbal hallucinations. I do not know from the article just what the individual experiences. Does she or he believe that someone spoke?

Verbal Hallucinations In Normals, IV: The Physiological Arousal Hypothesis

Karen Hamlet, Terry R. Barrett, Peggy Sabel Paxton, and Kelly L. Blair

About the Authors: Karen Hamlet is a graduate student and Terry R. Barrett is a professor, both in the Psychology Department at Murray State University. Peggy Sabel Paxton is a member of the West Kentucky Regional Mental Health/Mental Retardation Board, Inc. Kelly L. Blair is on the faculty in the Psychology Department at the Whitaker Regional Rehabilitation Center.

Abstract: In this article, the authors report three experiments designed to test the hypothesis that individuals in the general population who report frequent verbal hallucinations have higher levels of physiological arousal than non-hallucinators. Arousal was measured under conditions of relaxation, under conditions of heightened arousal, and in a startle paradigm. In none of these situations did hallucinators have higher levels of arousal than non-hallucinators. The authors conclude that there is, as yet, no convincing data in the literature on hallucinations linking the concept of heightened physiological arousal to the production of verbal hallucinations. See the reference list for related articles in this series of studies by these and other authors.

Perspective:

1. The idea being tested in this series of studies, that heightened physiological arousal and verbal hallucinations are related, comes, in part, from data linking increased levels of stress to the production of verbal hallucinations. Do you think it is reasonable to equate the concept of stress with the concept of physiological arousal? Why or why not?

2. According to the authors, it is not unreasonable to accept the findings presented in this article because of a recent understanding that has emerged about the notion of arousal. Describe that new understanding of arousal and how it changes our ideas about the relationship of hallucinations to physiological arousal.

Hallucinations reported by individuals in the general population have been the focus of occasional study for more than a century (e. g., Bentall & Slade, 1985; McKellar, 1968; Parish, 1914; Sidgwick, 1894; West, 1948). The common finding in these studies is that some individuals in the general population report hallucinatory experiences that do not seem to be associated with psychopathology, organic disorders, sleep, or drug use. The types of hallucinations reported include auditory hallucinations, visual hallucinations, hallucinations of touch, and hallucinations involving more than one sensory system.

In 1983, Posey and Losch surveyed 375 college students about the occurrence of a specific type of auditory hallucination, called verbal hallucinations. These hallucinations involve hearing a voice outside the head say something when no one is actually present. They found that fully 71% of their subjects reported having had at least one such hallucinatory experience. They selected 20 of those reporting frequent verbal hallucinations to complete Hathaway and McKinley's 1967 Minnesota Multiphasic Personality Inventory (MMPI) and to participate in an interview. These authors reported that, for all but one of these individuals, the MMPI profiles did not indicate any sign of psychopathology. In addition, the interview information corroborated the survey results and further suggested that these hallucinations were, indeed, experienced as voices coming from outside the head when no one was around.

Barrett and Etheridge (1992) extended and elaborated these results. They surveyed nearly 600 college students using a modified version of the Posey and Losch instrument. These authors found nearly identical reporting rates to those obtained by Posey and Losch (1983). In addition, they found that for a relatively large minority of those reporting verbal hallucinations, the frequency of such experiences was quite high. For example, 37.2% of the 586 individuals indicated they had experienced hearing their own voice aloud, outside their head, when no one had actually said anything. Furthermore, 47% of those who reported such an event indicated the experience occurred at least once a week.

One hypothesis that was tested and rejected was that subjects who reported such events were doing so because of experimental demand characteristics. This hypothesis suggests that subjects were reporting the occurrence of verbal hallucinations, not because they had experienced them, but because they wished to "help" or be seen in a positive light by the experimenter. To test this hypothesis , individuals who reported frequent verbal hallucinations and individuals who reported never having had such an experience were compared on five measures of this tendency. No significant differences were found, indicating that subjects, for the most part, were probably not lying about the occurrence of verbal hallucinations.

Another hypothesis considered by these authors was that reports of verbal hallucinations might be related to the presence of various forms of active, or incipient, psychopathology. Although this was considered to be improbable since all of their subjects were full time students, verbal hallucinations have often been taken to indicate signs of psychopathology. To test this hypothesis , they compared verbal hallucinators to non-hallucinators using the MMPI and a symptom check list (called the SCL-90R, Derogatis, 1983) regularly used by mental health centers to screen for indications of

psychopathology. Neither of these instruments was able to differentiate between individuals reporting frequent verbal hallucinations and those reporting no verbal hallucinations.

Barrett and Etheridge (1992) concluded that reports of verbal hallucinations by individuals in the general population could not be accounted for by the presence of emotional problems nor were these reports due to experimental demand characteristics. They went on to suggest that the reports appeared to be veridical accounts of conscious experiences and that a relatively large minority of college students, do indeed, experience frequent verbal hallucinations.

Slade and Bentall (1988), after a careful review of the literature on hallucinations, proposed that there are a number of factors that underlie the occurrence of hallucinations. Of particular relevance to this article is the idea that stress-induced physiological arousal is an important component in predicting the occurrence of hallucinations. They argued that "there is ample evidence that stress-induced arousal increases the probability that a person will hallucinate" (pg. 221). Other authors have agreed that physiological arousal is an important mechanism of hallucinations (e. g., West, 1975; Slade, 1972, 1973; Siegel & Jarvik, 1975; Hartmann, 1975).

There are two categories of evidence often cited in support of the physiological arousal hypothesis . One category is the apparent relationship between acute life events and hallucinations. Examples of particularly stressful life events that seem related to hallucinations include the loss of a loved one (Grimby, 1993; Rees, 1971), life-threatening situations (Comer, Madow, & Dixon, 1967; Lilly, 1956), sustained military operations (Belenky, 1979), and being held hostage (Siegel, 1984). The other category of evidence consists of three studies that attempted to directly measure arousal in hallucinators and non-hallucinators (Cooklin, Sturgeon, & Leff, 1983; Slade & Bentall, 1988; Toone, Cooke, & Lader; 1981).

The first category of reports suffer from several major shortcomings as evidence for the relationship between physiological arousal and hallucinations. First, there never has been (as far as we can tell) any attempt to compare the frequency of hallucinations during or after a particularly stressful event to the frequency of hallucinations that occur in the absence of such stressful events. We now know that many individuals in the general population experience quite frequent verbal hallucinations. It might be that hallucinations in these stressful situations are not more frequent than in non-stressful situations. Rather, these hallucinations may just receive more attention.

Second, these types of stressful events are often related to other factors associated with hallucinations such as sensory deprivation (Zuckerman & Cohen, 1964) and loss of sleep (Berger & Oswald, 1962). For example, loss of a loved one and sustained military operations can often lead to much loss of sleep. In addition, the life-threatening accidents that have been associated with hallucinations are all ones that involve much sensory deprivation (e.g., being trapped in an underground mine and being lost at sea). Finally, being held hostage is a situation in which both sensory deprivation and loss of sleep are common.

Third, it is not at all clear what the exact relationship is between the stressful life event and physiological arousal in these various situations. Generally, one might imagine increased physiological arousal as a result of the event. However, at least with respect to hallucinations associated with the death of a loved one, Rees (1971) indicated that the hallucinations were common for 10 years following the death. Rather than increased physiological arousal, one might just as plausibly hypothesize that these hallucinations are related to decreased physiological arousal due to prolonged depression triggered by the loss. In general, then, studies such as the ones mentioned above, while important in their own right, do not tell us much about the correlation between physiological arousal and hallucinations. More importantly, they tell us nothing about the hypothesized causal relationship between arousal and hallucinations.

The second category of evidence consists of studies that have attempted to assess the role of arousal in hallucination using direct physiological measures of arousal. As far as we can tell, there have only been three such studies. Toone, Cooke, and Lader (1981) compared schizophrenic patients with a history of hallucinations to matched controls on measures of skin conductance (a standard measure of physiological arousal). They found that the hallucinating schizophrenics exhibited higher skin conductance, indicating higher physiological arousal, than the matched controls. These authors claimed that this data was supportive of the notion that high arousal levels caused hallucinations. The problem with this data, however, is that the hallucinating schizophrenics were chosen because they were experiencing constant, on-going hallucinations. Consequently, it is quite possible that the increased arousal was not a cause, but rather was a consequence, of the hallucinations that occurred during the physiological recording session.

The other two studies in this category actually provide evidence more consistent with this latter hypothesis than the original hypothesis . Cooklin, Sturgeon, and Leff (1983) found that increases in skin conductance were reliably associated with the onset of verbal hallucinations in a group of hallucinating schizophrenics. Similarly, Slade and Bentall (1988) reported that in an unpublished study, one patient demonstrated increased skin conductance during periods of verbal hallucinatory activity.

Although the physiological arousal hypothesis of hallucinations seems to have been largely accepted by the scientific community, there does not seem to be any compelling data in support of the hypothesis . The purpose of this chapter is to report the results of three experiments designed to explicitly test the physiological arousal hypothesis of verbal hallucinations. Each of these experiments compared individuals in the general population who reported experiencing frequent verbal hallucinations with matched control subjects who had never experienced verbal hallucinations (we call the former group hallucinators and the latter group non-hallucinators). The hypothesis was that hallucinators would have higher levels of physiological arousal than non-hallucinators.

Experiment 1

In this experiment, two direct measures of physiological arousal were taken. These included a measure of muscle tension and a measure of skin temperature. Both of these have been taken to be indicators of physiological arousal with higher levels of muscle tension and lower levels of skin temperature indicating higher arousal (Duffy, 1962). These measures were taken during three phases of the experiment: baseline recordings immediately after the subject had been connected to the equipment; recordings during a muscle-relaxation task; and recordings following a series of reaction time tasks. The muscle-relaxation task was used to obtain measures of arousal during relatively low levels of stimulation while the reaction-time tasks were used to obtain measures of arousal immediately following a period when the subject was alert and vigilant. We also used a self-report measure of arousal. Thayer's (1967) Activation-Deactivation Adjective Checklist is designed to measure a person's overall phenomenological sense of physiological arousal. Because this instrument measures the sense of general bodily activation, Thayer argued that it might be a better measure of arousal than any direct physiological measurement.

Method

Subjects. The 32 subjects were introductory psychology students who received extra credit for their participation. There were 16 hallucinators and 16 non-hallucinators. The subjects were chosen from a large pool of individuals who had completed the Verbal Hallucination Questionnaire (Barrett & Etheridge, 1992). Hallucinators were randomly selected from the group of individuals who scored in the top 25 percent on that questionnaire. Non-hallucinators fell in the bottom 25% of scores on the same test. The experimenter was blind to group membership throughout the testing session. Subjects in the two groups were matched on age, sex, and composite ACT scores.

Materials and Procedures. Muscle tension and skin temperature were recorded using equipment produced by Autogenic Systems, Inc. Muscle tension readings were taken from the frontalis muscle of the forehead and skin temperature was taken from the middle finger of each subject's non-dominant hand. Each person first completed Thayer's Adjective Checklist (1967). Next, muscle tension and skin temperature were recorded during a two-minute baseline period. This was followed by an 18-minute progressive-muscle-relaxation training period. During the muscle-relaxation training, muscle tension and skin temperature were recorded during a two-minute period four minutes into relaxation training, 10 minutes into relaxation training, and 16 minutes into relaxation training. At the end of the muscle-relaxation training period, subjects again filled out Thayer's Adjective Checklist.

After the relaxation training, subjects were given a series of 60 trials on a simple reaction-time task. A computer was used to present the signals and to record the responses. Subjects were asked to respond as fast as they could after a stimulus was presented. Immediately following the simple reaction-time task, muscle tension and skin temperature were measured during a two-minute period. Then subjects completed 60 trials in a choice reaction-time task. Here subjects had to respond differentially to one of two stimuli. At the end of this second reaction time task, muscle tension and skin temperature were again recorded during a two-minute period. Finally, subjects completed Thayer's Adjective Checklist for a third time

Results and Discussion

For muscle tension and skin temperature, there were six separate two-minute recording periods: (1) baseline, (2) after four minutes of relaxation, (3) after 10 minutes of relaxation, (4) after 16 minutes of relaxation, (5) following the simple reaction-time task, and (6) following the choice reaction-time task.

For each measurement period, a two-group analysis of variance was calculated comparing hallucinators to non-hallucinators on muscle tension scores and on skin temperature scores. None of these analyses produced a significant effect. Similar analyses indicated no differences between hallucinators and non-hallucinators on any of the three administrations of Thayer's Adjective Checklist.

The results of this experiment are clear. Under the conditions imposed by the various experimental tasks, verbal hallucinators were not more physiologically aroused than non-hallucinators. In addition, verbal hallucinators did not rate themselves as more aroused than non-hallucinators on three separate occasions during the experiment. These data are inconsistent with the physiological arousal hypothesis of verbal hallucinations.

There are, however, two arguments for further research. First, there is considerable agreement that while muscle tension and skin temperature both measure arousal, skin conductance is a more sensitive measure of physiological arousal (Duffy, 1972; Porges, Ackles, & Truax, 1983). Consequently, it may be that measures of skin conductance would be more sensitive to the hypothesized arousal differences between verbal hallucinators and non-hallucinators. Therefore, in Experiment 2 we compared hallucinators and non-hallucinators on measures of muscle tension and skin conductance.

The second consideration has to do with the concept of arousal itself. Fahrenberg, Walschburger, Foerster, Myrteck, and Muller (1983) argued that arousal should be conceptualized as consisting of two different kinds of processes. One of the processes is state arousal which is related to the environmental stimulation a person receives. The second process is trait arousal which is understood to be a relatively stable and consistent pattern of activation over various stimulus situations.

It is possible that the general experimental situation used in Experiment 1 was one that tapped state arousal rather than trait arousal. The subjects were constantly exposed to environmental stimulation over the course of the whole relaxation period. They were required to listen to a progressive-muscle-relaxation tape and engage in muscle-relaxation training. In addition, the experimenter was in the room during the course of the entire experiment in order to monitor and record the physiological measurements. From this perspective, one might conclude that hallucinators and non-hallucinators do not differ on state arousal. It may be, however, that the environmental stimulation provided during the entire course of the experiment was enough to mask potential group differences in trait arousal. Consequently, in Experiment 2 we measured arousal under conditions of minimal stimulation to look more closely at potential differences between verbal hallucinators and non-hallucinators on trait arousal.

Experiment 2

In this experiment, the physiological measures of skin conductance and muscle tension were taken during a 20-minute period of minimal stimulation. In addition, we gathered further self-report data related to arousal. Subjects were asked to complete the State-Trait Anxiety Inventory designed by Spielberger, Gorsuch, and Lushene (1970). This instrument measures state anxiety (a transitory emotional state of apprehension and tension) and, more importantly for the purposes of this research, trait anxiety (a relatively stable characteristic of anxiety proneness). Subjects were also asked to complete the Perceived Stress Scale developed by Cohen, Kamarck, and Mermelstein (1983). This measures the degree to which individuals appraise situations in their lives as stressful.

Method

Subjects. The 28 subjects (14 hallucinators and 14 non-hallucinators) tested in this experiment were selected using the same procedure as described in Experiment 1. As in Experiment 1, the experimenter was blind to group membership and the two groups were matched on age, sex, and ACT composite scores.

Materials and Procedures. Muscle tension and skin conductance measures were obtained using the Self Regulation Systems ORION Base Unit connected to a computer. Upon arrival to the laboratory, each subject was asked to complete the State/Trait Anxiety Inventory (Spielberger et al., 1970) followed by the Perceived Stress Scale (Cohen et al., 1983).

After completing the self-report instruments, the subject was connected to the equipment. Muscle tension sensors were attached to the frontalis muscle of the forehead and skin conductance electrodes were attached to the tips of two fingers, the middle finger and the ring finger of the non-dominant hand.

After the sensors were attached to the subject, the lights were dimmed and the subject was instructed to close her eyes and breathe normally, concentrating with each exhalation upon the number 1. It was suggested that the subject think of a large block figure of the number 1. When the subject indicated that she understood the instructions, she was asked to close her eyes and to attempt relaxation by means of the procedure described. Automatic recording of two physiological measures, muscle tension and skin conductance, was then started and continued for 20 minutes with summary statistics obtained for each one-minute period of monitoring. The experimenter left the room once the 20-minute period started and did not return until the end of the recording session.

Results and Discussion

For both muscle tension and skin conductance, a 2 x 20 analysis of variance was computed where the grouping variable was Groups (hallucinators versus non-hallucinators) and the repeated-measures independent variable was Trials (1-20). These analyses produced only one significant effect, the main effect of Trials on the skin conductance measure [F(19, 494) = 5.44, p < .0001]. Skin conductance means decreased over the first ten minutes of testing and then tended to rise again, especially in the last five minutes of testing. Separate two-group analyses of variance performed on the State-Trait Anxiety Inventory and the Perceived Stress Scale did not produce any significant effects.

The results of this study are clear. Under conditions of minimal stimulation, individuals who report frequent verbal hallucinations were not more physiologically aroused as measured by muscle tension and skin conductance than persons who never experienced verbal hallucinations. This study replicated the muscle tension results of Experiment 1, as well as indicating no differences in skin conductance. Furthermore, none of the self-report measures indicated any differences. As such, the physiological arousal hypothesis of verbal hallucinations was, once again, disconfirmed by the data.

One problem with both experiment 1 and 2 is that they depend on the historically accepted notion that physiological arousal can be reliably assessed by a variety of measures during periods of relative relaxation. Neiss (1988) argued forcibly that the available data is simply inconsistent with such a view of arousal. This suggests that the way we have been measuring physiological arousal may be flawed.

In fact, there is a different way of conceptualizing arousal that makes use of the notion of reactive potential. In this view, arousal is operationalized in terms of a startle response rather than as an ongoing physiological process that can be tapped during relaxation. There is some evidence suggesting that this perspective might be a productive one. For example, posttraumatic stress disorder is a delayed reaction to extreme stress that is characterized by, among other things, heightened arousal as defined by an exaggerated startle response (APA, 1987).

Mueser and Butler (1987) studied veterans with posttraumatic stress disorder. They compared veterans who experienced verbal hallucinations with those who did not have hallucinations on demographic, military, postmilitary, and symptom variables. They found that the hallucinators tended to have more intense symptoms of posttraumatic stress disorder than non-hallucinators, suggesting that hallucinations might be related to exaggerated startle responses.

Recently, Lang, Bradley, and Cuthbert (1992) proposed a model that relates arousal and affective state to the level of an individual's startle response. This model suggests that high arousal, alone, is not enough to produce exaggerated startle responses. In fact according to these authors, the combination of high arousal and high positive affect actually inhibits the startle response. On the other hand, the combination of high arousal and high negative affect produce exaggerated startle responses. Barrett and Etheridge (1994) have demonstrated that verbal hallucinators in the general population show clear signs of heightened levels of negative affect compared to non-hallucinators. Given these considerations, the physiological arousal hypothesis of verbal hallucinations predicts that hallucinators will have a greater startle response than non-hallucinators.

Experiment 3

In experiment 3, we compared hallucinators and non-hallucinators in a situation designed to produce a reliable startle response. Subjects viewed a videotape containing film clips of startling and non-startling scenes. Arousal levels were determined by the measurement of skin conductance.

Method

Subjects. The subjects were 32 introductory psychology students (16 hallucinators and 16 non-hallucinators) who were selected using the same procedures as in Experiments 1 and 2. As in the first two experiments, subjects in these two groups were matched on sex and age. However in this experiment, subjects were not matched on ACT composite scores. Once again, the experimenter was blind to the subject's group membership during testing.

Materials and Procedures. Skin conductance measures were again obtained using the Self Regulation Systems ORION Base Unit connected to a computer.

Twelve three-minute film clips (two from each of six films) were selected from a large collection of obscure horror films. These clips had been previously rated so that there was one startling film clip and one non-startling film clip for each of the six movies represented. The startling scene occurred at approximately 70 seconds into each of the startling film clips. Previous ratings also indicated no differential startle potential in the first 60 seconds of the startling and non-startling film clips.

Two videotapes were created. Each videotape contained six of the 12 three-minute film clips, startling scenes from three of the films and non-startling scenes from the other three films. The order of placement of the film clips was random on both videotapes with the restriction that three clips of the same type could not occur together. Each tape was divided into two sections to allow a rest period. In addition, after every clip there was a 90 second section of blank tape to allow the subject's skin conductance to return to baseline.

The subject was seated comfortably in front of a television screen and connected to the physiological recording equipment. The video tape was started and skin conductance recording began. After viewing the first three clips, the subject rested for approximately two minutes and then viewed the remaining three clips.

Results and Discussion

Initial analyses indicated no differences between hallucinators and non-hallucinators in mean skin conductance during the three-minute baseline period prior to the first clip. Additional analyses were performed on mean skin conductance scores for the 90 second blank periods following each of the film clips. These analyses also indicated no differences between hallucinators and non-hallucinators.

Figure 1 shows mean skin conductance for each 30-second period in the startling and in the non-startling sessions separately for hallucinators and non-hallucinators. A 2 x 2 x 6 factorial analysis of variance was calculated. The grouping variable was Groups (hallucinators versus non-hallucinators). The two repeated-measures independent variables were Type of Clip (startling versus non-startling) and Period (with six levels where the levels of this variable represented the mean skin conductance collected during each 30 second portion of the film clip). There was a significant Type of Clip by Period interaction [F(5, 150) = 17.48, p < .00001]. As shown in Figure 1 for both hallucinators and for non-hallucinators, skin conductance scores showed a gradual decline in the non-startling condition across the six 30-second periods but showed a precipitous rise during period three in the startling condition. This, of course, is the 30-second period in which the startling scene was presented. This interaction indicates that we were quite successful in producing situations that created a strong startle response in the subjects. However, this analysis did not produce any other effects. As can be seen in Figure 1, non-hallucinators showed exactly the same magnitude of startle response during the startling clip as the hallucinators.

The data from this experiment are clear. There was no indication, whatsoever, that hallucinators


Figure 1

Mean Skin Conductance

Mean skin conductance for each of six 30-second periods as a function of type of film clip shown separately for hallucinators and non-hallucinators.

 

 

 

showed an exaggerated startle response compared to non-hallucinators. We are left with the inescapable conclusion that verbal hallucinators in the general population simply do not differ from matched non-hallucinators in terms of physiological arousal.

Summary and Conclusions

Three experiments were conducted in an attempt to produce support for the physiological arousal hypothesis of verbal hallucinations reported by individuals in the general population. In Experiment 1, we measured muscle tension and skin temperature during a baseline period, during a muscle-relaxation task, and immediately following two types of reaction-time tasks. In addition, subjects filled out Thayer's Adjective Checklist, a self-report measure of arousal, three times during the experiment. There were no significant differences between hallucinators and non-hallucinators on any of these measures.

In Experiment 2, we recorded muscle tension and skin conductance during a 20-minute period of minimal stimulation to get a clearer picture of hypothesized differences in trait arousal. Subjects also filled out two self report measures, the State-Trait Anxiety Inventory and the Perceived Stress Scale. As in Experiment 1, there were no differences between hallucinators and non-hallucinators on any of these measures.

In Experiment 3, we compared hallucinators to non-hallucinators in a startle paradigm. Although we were successful in measuring a clear startle reaction, that reaction did not differ in hallucinators and non-hallucinators.

Overall, then, this series of experiments has provided no evidence for the physiological arousal hypothesis of verbal hallucinations reported by individuals in the general population. Slade and Bentall (1988) and Bentall (1990) have argued that hallucinations reported by schizophrenics are most likely produced by the same set of mechanisms as those reported by college students. To the extent this is true, the experiments reported here also suggest that physiological arousal is probably not tied to the verbal hallucinations often associated with schizophrenia. Based on our review of the available literature and the present set of experiments, we do not believe there is any convincing evidence suggesting that increased arousal causes verbal hallucinations.

Our conclusions, in a very real way, have been anticipated by progress made in understanding the complex relationship between cognition, emotion, and physiology. Neiss (1988) has pointed out that the concept of arousal was introduced into the psychological vocabulary in the first half of this century because it seemed to be unidimensional and to be the physiological equivalent of the concepts of fear and anxiety. Much evidence (reviewed by Neiss, 1988) now contradicts these initial assumptions. Various physiological systems, all thought to index arousal, show little correlation with one another in any given situation. In addition, these systems seem to operate in a relatively independent way. Finally, these systems are not related in any unidimensional way to affective states. These findings have led theorists in the area of physiological brain activity, biochemistry, and emotional states to reject the concept of a global activation concept as an explanatory device. It is not surprising, then, that the concept of physiological arousal has not turned out to be helpful in our understanding of verbal hallucinations.

These conclusions should not be taken to mean, however, that we advocate the abandonment of a search for the physiological correlates of verbal hallucinations. We certainly do not. Instead, we believe that verbal hallucinations are most profitably understood to be the product of a unique psychobiological state of consciousness characterized by a particular configuration of cognitive processes, feeling states, and physiological activation. Consequently, we predict the most useful research strategy will be a multifaceted one exploring each of these interdependent dimensions within a contextual framework.

Notes

Experiment 1 was conducted as part of Kelly L. Blair's master's thesis in Psychology at Murray State University. Experiment 2 was conducted as part of Peggy Sabel Paxton's master's thesis in Psychology at Murray State University. Experiment 3 was conducted as part of Karen Hamlet's senior honors thesis in Psychology at Murray State University. Experiment 3 was supported by grants 2-12869 and 2-12756 to the second author from the Committee on Institutional Studies and Research, Murray State University.

References

Response from Leonard J. Shedletsky

An impressive program of research not withstanding, one cannot help but ask such questions as these:

1. What sorts of things did the voices say?

2. How did the individual react to the voice?

3. In what context did the hallucination occur?

4. What sense did the individual make of the experience?

Perhaps it is unfair to ask such broad sweeping questions of a highly focused research. Nevertheless, these questions do nag at the reader, and there may even be a clue buried in their answers about the phenomenon itself.

Homepages: Communication Resources (Aitken) | Communicating (Berko) | Intimate Communication | Intrapersonal | Public Speaker/Public Listener (Berko) | "Speech Communication Teacher" (NCA)aker/Public Listener (Berko) | "Speech Communication Teacher" (NCA)