A Testimony of an Ex-Mormon

Jack Kettler

Copyright 1998 Jack Kettler
Mormons like to give testimonies. Unlike Mormon testimonies, I will seek to show that the testimony of my conversion to Christianity is based upon the teaching of Scripture. There are many that claim to have had spiritual experiences. It is seen under close scrutiny that these experiences have actually affected how certain individuals interpret Scripture. Scriptures are often reinterpreted in light of the experience. This approach is fraught with dangerous pitfalls. II Corinthians 11:13-15 warns us of false workers who transform themselves into ministers of Christ. In Matthew 7:22,23 we find that there are those who have even worked miracles, but in the end Christ says, I never knew you. Even miracle workers may be enemies of Christ. We are to be on guard against false ministers and apostles. All organizations, whether secular or religious, offer testimonials. How are testimonials evaluated? Are there false testimonies? If Christ's Spirit is really working in someone's life, what will be the nature of the testimony? Many people do not grasp the potential for self-deception. Not only may we deceive ourselves, the apostle John tells us: Beloved, believe not every spirit, but try the spirits whether they are of God. I John 4:1. False spirits may deceive us. The apostle tells us the nature of a true testimony in his gospel: "Howbeit when he, the Spirit of truth, is come.he shall not speak of himself.He shall glorify me" John 16:13,14. A true testimony must glorify Jesus Christ. Glory given to another makes for an invalid testimony. It is readily seen that Mormon testimonies concerning Joseph Smith (Mormon founder) do not qualify as being inspired by the Holy Spirit. Why? Mormon testimonies consistently exalt Joseph Smith or the supposed truthfulness of the Mormon Church. In l Corinthians 15:1-4 the apostle Paul tells us the nature or content of a true testimony: Moreover, brethren, I declare unto you the gospel which I preached unto you, which also ye have received, and wherein ye stand. By which also ye are saved, if ye keep in memory what I preached unto you, unless ye have believed in vain. For I delivered unto you first of all that which I also received, how that Christ died for our sins according to the scriptures; And that he was buried, and that he rose again the third day according to the scriptures. In this testimony I refer to several passages of Scripture that are relevant to Christ's resurrection, the work of the Holy Spirit in exalting Christ, man's condition, and 1
how men are saved. Do Mormons refer to these passages and the teachings contained therein when sharing their testimony? When Mormons give testimony to their faith in Christ, it is always supplemented by their larger testimony that Joseph Smith is a prophet and the Mormon Church is the only true church. The apex of Mormon testimonies goes beyond Christ to another, namely Joseph Smith. My testimony can go no further than scripture. I confess by the grace of God the truth of what the apostle Paul teaches in l Corinthians 15:1-4. In another place he declares the following concerning man's condition: As it is written, There is none righteous, no, not one.that every mouth may be stopped, and all the world may become guilty before God Romans 3:10,19. This was my condition. Paul goes on to say: For the wages of sin is death, but the gift of God is eternal life through Jesus Christ our Lord Romans 6:23. I had earned the wages of death. God in his mercy gave me the gift of eternal life. The only thing that I earned and deserved was death. Eternal life came as a gift. One thing I am certain of and that is this: there was and is absolutely nothing in me that caused God to give me this gift. Jesus Christ gets all the glory and praise. I attempt to do as the writer of Hebrews sets forth: Looking unto Jesus the author and finisher of our faith Hebrews 12:2. I look to Jesus by giving him the glory. God gave me the gift of faith. I am saved by grace and even my faith is a gift. Ephesians 2:8 says: and that not of yourselves. What is not of yourselves? Faith! Did I choose Christ and exercise faith? Yes, but why? Who gets the glory? Christ? Or me? Why did I choose to believe? Ephesians 1: 4,5 supplies us the answer. According as he hath chosen us in him before the foundation of the world, that we should be holy and without blame before him in love: Having predestinated us unto the adoption of children by Jesus Christ to himself, according to the good pleasure of his will. Was this salvation in my hands to choose or reject? If this were the case, then could I not glory in and of myself? How can that be so? Because I would have done something others had not done. The following verse tells us that the predestination is: according to the good pleasure of his will. So then it is not of him that willeth, nor of him that runneth, but of God that showeth mercy Romans 9:16. The doctrine of predestination more than any other teaching of scripture takes salvation out of man's hands and places it in God's control. Men do not like God's control. The cause of God's choosing or predestination is found in Him. If we insist that we played a part in God's choice, then human merit is brought into the picture. Salvation then becomes synergistic rather than monergistic. Biblical salvation is monergistic. Christ alone, by his complete and finished work saved me. Within a synergistic scheme, salvation becomes a cooperative effort. My work takes away from the work of Christ. How? I made a contribution. I played a part in my salvation. If I was not willing, then God could not save me. A synergistic scheme of salvation not only steals Christ's glory, it limits God's power. God can only do what I allow him to do within this type of system. Again, I confess by the grace of God that: Not by works of righteousness which we have done, but according to his mercy he saved us, by the washing of regeneration, and renewing of the Holy Ghost Titus 3:5.

This is the close of my testimony. To God only wise, be glory through Jesus Christ for ever. Amen. Romans 16:27. heirs according to the promise" (Gal. 3:28,29).
Mr. Kettler is an elder in the Orthodox Presbyterian Church and Chairman of the Waco Committee with Citizens for the Constitution. Mr. Kettler is also a member of the John Birch Society.

Psychophysiology, 38 ~2001!, 275291. Cambridge University Press. Printed in the USA. Copyright 2001 Society for Psychophysiological Research
Constraints for emotion specificity in fear and anger: The context counts
GERHARD STEMMLER, MARCUS HELDMANN, CORNELIA A. PAULS, and THOMAS SCHERER
University of Marburg, Germany
Abstract We investigated psychophysiological responses to fear and anger inductions during real-life and imagination. Female participants ~N 158! were assigned to a fear-treatment, fear-control, anger-treatment, or anger-control group. Context ~real-life, imagination! was varied in two sessions of fixed order. Eleven self-report and 29 somatovisceral variables were registered. Results showed that ~a! except during anger imagination, control groups were emotionless; ~b! in control groups, contexts prompted diverging somatovisceral responses, but similar emotion self-reports; except during fear imagination, the emotion inductions ~c! were successful and ~d! produced specific emotion reports; ~e! during real-life, somatovisceral fear and anger responses exhibited a marked cardiovascular defense reflex; ~f ! in addition, real-life fear showed an adrenaline-like specific response pattern, whereas real-life anger showed specific forehead temperature and EMG extensor increases, accompanied by an elevated DBP during imagination. A Component Model of Somatovisceral Response Organization is proposed. Descriptors: Emotion, Specificity, Defense reflex, Fear, Anger, Context
Somatovisceral activation constitutes the interface between brain and behavior. The study of somatovisceral responses is therefore an important field of psychological inquiry because, beyond mere reflexlike and homeostatic reactions, such responses may inform about the psychobiological processes leading to anticipated or actual behavior. However, diverse influences may simultaneously impinge on the somatomotor and the autonomic nervous systems, threatening the internal validity of studies by allowing alternative ways of interpreting empirical data. Research about the physiological specificity of emotions is a pertinent example. For instance, the finding of physiological differences between fear and anger inductions ~Ax, 1953! could have been misrepresented as evidence for emotion specificity simply because the inductions differed both physically and in their behavioral demands ~Stemmler, 1984!. Similarly, Stemmlers ~1989! report that emotional autonomic patterning was specific to the context of emotion induction was criticized by Ekman ~1994! for measuring physiology a considerable period after the induction was over, suggesting the emotional processes had already subsided.1 Boiten ~1996! argued that the directed facial action task used by
We are indebted to Wm. Peter Flannery for commenting on an earlier draft of the manuscript. This research was conducted with the help of a grant by the Deutsche Forschungsgemeinschaft, grant # Ste 40506. Address reprint requests to: Prof. Dr. Gerhard Stemmler, PhilippsUniversity, Fachbereich Psychologie, Gutenbergstr. 18, D-35032 Marburg, Germany. E-mail: stemmler@mailer.uni-marburg.de. 1 The argument will not be considered here in any detail. But the physiological profiles during real-life inductions of fear and anger were nevertheless found to differ.
Ekman, Levenson, and collaborators ~Ekman, Levenson, & Friesen, 1983; Levenson, Carstensen, Friesen, & Ekman, 1991; Levenson, Ekman, & Friesen, 1990; Levenson, Ekman, Heider, & Friesen, 1992! exerted its autonomic effects primarily through effortrelated changes in respiration rather than through the activation of emotion circuits in the brain. Levenson and Ekman ~submitted! in turn argue that Boiten has misrepresented their results. In sum, many concerns prevail about the internal validity of studies in this field, and opinions about the empirical validity of physiological emotion specificity are strongly diverging. Conceptual objections against the view that somatovisceral responses could be specific for emotions ~James, 1884!, or at least for some ~Ekman, 1992!, came in two waves. During the first wave of objections it was argued that autonomic activity is largely undifferentiated and unitary, eliminating any basis for specific bodily emotion signatures ~Cannon, 1929; Duffy, 1962; Mandler, 1975; Schachter & Singer, 1962!. The second wave of objections is more recent: Instead of reflecting the arousal of a particular emotion, somatovisceral processes would indicate the demands of an action disposition, action tendency, or action proper ~Davidson, 1993, 1994; Frijda, 1986; Lang, Bradley, & Cuthbert, 1990!. Moreover, emotions could motivate actions that, across emotions, were likely to overlap. The first objection against somatovisceral emotion specificity, that is, that autonomic activity is largely undifferentiated, is no longer tenable. There is ample evidence from both physiology ~Hilton, 1975; Hilton & Redfern, 1987; Jnig, 1988; Jnig & McLachlan, 1992! and psychophysiology ~Berntson, Cacioppo, & Quigley, 1991; Fahrenberg, 1987; Lacey, 1967; Stemmler, 1992a; Stemmler & Fahrenberg, 1989! that somatovisceral processes are 275

276 highly patterned and are likely to reflect the activity of multiple activation0inhibition systems. The second objection, that somatovisceral activation during emotions reflects action tendencies but not emotions per se, is a serious challenge to the notion of somatovisceral emotion specificity. This argument was motivated by the low consistency of somatovisceral configurations in studies of emotion ~see reviews by Cacioppo, Klein, Berntson, & Hatfield, 1993; Stemmler, 1998!. In addition, there is ample evidence that not only physical tasks, body postures, and their associated motor behaviors, but also psychological influences ~e.g., attention, mental effort! exert an effect on somatovisceral activation. In the following, the term context will be used to denote this ensemble of nonemotional physical, behavioral, and psychological factors that influence somatovisceral activation during emotion. Beyond the context, emotions might still influence somatovisceral activation. But their effects could well be confounded with context effects, rendering their separate identification difficult. Because contexts vary widely across emotion studies, a low degree of consistency of somatovisceral emotion configurations would be predicted. Theoretically, one of us has discussed context-deviation specificity as a model of emotion specificity ~Stemmler, 1984, 1989, 1992b! whichrather than the model of absolute specificity should be applied when there are contextemotion confounds. Context-deviation specificity views emotion specificity as a conditional concept. An emotional stimulus is supposed to modify a context-bound physiological pattern. Thus, somatovisceral emotion and context effects are usually confounded. Emotion specificity could then be demonstrated only after this confound has been pulled apart. Emotion specificity is found whenever there are systematic and specific deviations of the emotion-plus-context pattern of physiological reactivity from the context-alone pattern. Absolute emotion specificity assumes that, even across different contexts, at least some emotions are accompanied by specific physiological patterns. However, neither contextemotion confounds nor the validity of context-deviation specificity have yet been studied systematically. The present investigation is a first attempt to fill this gap. In particular, we compared a real-life with an imagination context with respect to emotion specificity. In this report, we chose to study the negatively valenced emotions of fear and anger. Reasons for this choice were threefold. First, the database for fearanger contrasts is larger than for any other emotion combination. Second and perhaps more importantly, fear and anger have been induced in different contexts, including real-life and imagination settings, and the directed facial action task. Apart from fear and anger, only happiness has been induced during real-life inductions ~see the studies reviewed by Cacioppo et al., 1993!. Finally, somatovisceral differences between fear and anger, should they exist, could not be attributed to their common negative valence. There are at least 15 studies that have compared fear and anger in two or more somatovisceral responses ~Adsett, Schottstaedt, & Wolf, 1962; Ax, 1953; Boiten, 1996; Chessick, Bassan, & Shattan, 1966; Funkenstein, King, & Drolette, 1954; Levenson et al., 1990, 1991, 1992; Miller et al., 1987; Roberts & Weerts, 1982; Schachter, 1957; Schwartz, Weinberger, & Singer, 1981; Sinha, Lovallo, & Parsons, 1992; Sinha & Parsons, 1996; Stemmler, 1989!. Findings across these studies demonstrated few consistent somatovisceral differences between fear and anger. During anger, diastolic blood pressure, total peripheral resistance, and somatomotor activity increases were larger than during fear ~in five out of eight, three out of four, and two out of six studies, respectively!. During fear, finger temperature decreases and cardiac output increases were

G. Stemmler et al. larger than during anger ~in three out of seven and in two out of four studies, respectively!. The present investigation sought to replicate these consistent findings but also to extend the present knowledge in using a much broader sample of somatovisceral variables than did most of the studies cited above. A broader sample of variables might allow conclusions about the physiological regulation patterns underlying emotion responses that could reach beyond what can be inferred from a few variables alone. All of the studies cited above used a within-subjects control condition against which the emotion effects were compared. This procedure is perfectly well-suited for determining whether there are somatovisceral emotion differences, but the controlemotion comparison is confounded by context differences if they exist. Thus it will be difficult to determine what the context-free emotion signatures might look like. To illustrate, context effects were likely to operate in diastolic blood pressure, even though the small number of studies defies any definitive conclusions. All real-life ~N 5! and all imagination ~N 3! studies found significant diastolic blood pressure increases during anger compared to the control condition. In contrast, during fear, all real-life ~N 4! but none of the imagination ~N 3! studies found significant diastolic blood pressure increases. As a consequence, only two real-life studies reported significantly larger diastolic blood pressure increases during anger than fear, but all of the three imagination studies did. Thus, depending on the emotion induction context there is a real risk for the test of emotion specificity to fail for artifactual reasons. To determine what the context-free emotion signatures might look like, a control condition as similar as possible to the emotion induction period needed to be constructedexcept for the emotion effect per se ~Boiten, 1996; Nyklicek, Thayer, & van Doornen, 1997; Stemmler, 1992b!. In this study, we tried to approach this goal through a between-subjects control design where the control group received exactly the same treatment as the treatment group with the one difference that controls were completely preinformed about the impending emotion inductions. This control procedure should alleviate the emotional meaning of the impending emotion induction but leave the nonemotional context effects almost unchanged. Our expectation of a negligible emotional engagement of the control group is based on the powerful effects appraisals can have under conditions where individuals have to make sense of their environment. Referring, for example, to the cognitivemotivational-relational theory of emotions, Lazarus ~1991! distinguishes among three forms of primary appraisal: goal relevance, goal congruence, and type of ego involvement. If an individual is convinced that the impending threat to his or her self- or social esteem is just fake, the person will not very much care about the forthcoming encounter ~goal relevance!, the threat will not be strongly inconsistent with his or her goals ~goal congruence!, and the individuals self- or social esteem will not be markedly endangered or assaulted, as in fear and anger, respectively. In sum, the aims of the present study were fourfold. First, we tried to determine context effects of two alternative induction methods: real-life and imagination. If present, such effects could blur the effects of emotion on somatovisceral responses. In contrast to somatovisceral responses, we expected only minor influences of the context on self-reports of emotion, because feeling states reflect meaning structures that are largely independent of situational properties. Second, we asked what the emotion responses to fear and anger emotion inductions look like, ifin light of our claim of contextemotion confoundsemotion and context effects are pulled apart by contrasting appropriate treatment and control groups. This second aim also touches upon the question of whether these emo-

Constraints for emotion specificity tion responses are consistent across real-life and imagination contexts. Third, using a broad sample of somatovisceral variables, we tried to identify the regulatory patterns underlying fear and anger responses. Finally, we wanted to replicate the small and fragile set of findings suggesting a physiological differentiation of fear and anger. Method Participants One hundred and ninety-seven participants were recruited by flyers distributed on a university campus and by local newspaper ads mentioning a study of stress and strain. Inclusion criteria were female gender, German native language, age between 18 and 45 years, right-handedness, and normal weight according to the bodymass index ~23 kg0m 5 kg0m 2 !. Exclusion criteria were studying psychology, current medical treatment, taking medication affecting the circulation, or pregnancy. We chose a female sample because gender differences in self-reported emotions point to a larger female expressiveness of negative emotions ~Manstead, 1992!. Thirty-nine participants declined after the introductory session or were excluded, leaving N 158 participants for statistical data analysis.2 The average age of these participants was 25.3 years ~range years!; mean body-mass index was 21.1 kg0m 2 ~range 1631 kg0m 2 !. Seventy-seven percent of the participants were university or high school students, the remaining 23% were employees from various professions. Participants were paid 90 DM for approximately 6 hr involvement in the study. Experimental Design Participants were assigned to one of four experimental groups formed by the combination of Emotion ~fear, anger! Group ~treatment, control!. Participants were randomly assigned to groups with the restriction of homogeneity of trait anxiety in fear groups and of trait anger in anger groups. As a result, fear groups were virtually identical in STAI trait anxiety ~M 40.0, 40.9; SD 8.0, 9.4; N 38, 41, for treatment and control groups, respectively! and anger groups in STAXI trait anger ~M 22.2, 21.9; SD 4.9, 5.4; N 40, 39, same order as above!. Context ~real-life, imagination! was a within-subjects factor. Real-life inductions were always performed first. Imagination inductions were performed one week later in a second experimental session. A notable feature of this experiment was the use of multiple induction periods, that is, several inductions of the same emotion right after one another. This procedure allowed us to aggregate measurements across emotion induction periods provided homogeneity of response profiles was given. Aggregation increases measurement reliability. The choice of a between-subjects design for the factors Emotion and Group is quite unusual in psychophysiological emotion research, because a within-subjects design has the virtue of controlling sources of variance between individuals that may be large in somatovisceral variables. However, the internal validity of withinsubjects designs is easily threatened ~Cook & Campbell, 1979!. Relevant problems are the interaction of adaptation effects ~fa2 Reasons to decline were the disturbing loud noise ~N 8! and the experience of claustrophobic feelings ~N 1!. N 15 participants gave no explanation for their declining. Reasons to exclude participants were unwillingness to openly answer the items of a personality questionnaire ~N 2!, the association of the Milgram experiments ~N 2!, and strong artifacts in the physiological recordings ~N 11!.

277 tigue, habituation, motivation, etc.! with presentation order and changing interpretations or expectations on the part of the participants. To be sure, presentation order of fear and anger inductions could in principle be balanced across participants, but obviously not the order of treatment and control conditions ~at least not with our control group procedure!. Of equal importance is the consideration that, from the perspective of the experimenter, real-life inductions always carry an element of surprise and dissembled intentions. Thus, it is not unlikely that, in a within-subjects design, the second emotion induction has less of an impact than the first one. For these reasons we chose a between-subjects design. Setting and Apparatus The experimental room ~4 3.4 m! was sound-attenuated and air-conditioned and had a largely nontechnical appearance. Participants sat comfortably in a reclined position. Electrodes were connected to a customized headbox ~NeuroScan!, where signals were preamplified with a gain of 30 ~input impedance of differential inputs 20 MV!. Transducer-based signals were relayed through an input box to Biopac couplers. Right in front of the participants seat was a customized ergometer ~Kettler EX-1!, which could be pedaled while seated. Three LEDs signaled the participant to slow down, speed up, or continue at the same pace. A custom-made hand dynamometer with a built-in indicator scale for pulling strength was mounted at the right side of the seat. Other equipment included two WS-A10E-W Panasonic 160 W loudspeakers, a JVC TK-1281 video camera, and a 20-in. black-white computer monitor ~Miro!. In an adjacent room were placed a 32-channel SynAmps Model 5083 amplifier with 16 bit A0D conversion ~NeuroScan!; a 16channel Biopac system with couplers for skin conductance ~GSR 100!, skin temperature ~SKT 100!, pulse volume ~PPG 100!, respiration belt ~RSP 100!, general purpose couplers ~DA 100!, and with an MP 100 workstation with 16 bit A0D conversion; a KardioDynagraph ~Diefenbach! for impedance cardiography; a bosotron 2 ~Bosch! blood pressure monitor; a Sony minidisk player MDS 501 for the delivery of instructions and loud noise; and other audiovisual equipment. A Macintosh Quadra 950 ~Apple! microcomputer with a NB-D10-24 digital I0O card ~National Instruments! and four serial I00 ports performed experimental control, data recording, data visualization, and data storage under LabView 3.1.1 ~National Instruments!. With the exception of parts of the fear induction and the imagination session, participants were alone in the room. Whenever necessary, experimenter and participant communicated via intercom. Procedure Introductory session. Participants could familiarize themselves with the experimental room, some experimental tasks ~standard tests!, and the emotion self-report form. They also filled out trait anger and anxiety questionnaires needed for group assignments. Finally, participants signed an informed consent that also mentioned that it might be necessary to obtain a blood sample. Experimental Session I. One to three weeks later, the first experimental session was scheduled. It took about 2.5 hr. Two female assistants positioned electrodes and transducers. The ~male! experimenter explained the standard tests and the emotion self-report form. He reminded the participant to sit quietly to help prevent artifacts in the physiological recordings. The standard tests began with a 10-min rest period during which participants should relax but keep their eyes open. The loud noise

Figure 1. Continued.

of preparing and giving a speech. Participants were asked to sit quietly. Induction Period 1 began with the instruction that participants were to give a 5-min speech on the topic arguments pro and con the European Union. The speech should be prepared during the next 5 min. Participants should do their best. It would be necessary to record their speech. The speech would be evaluated for verbal intelligence and compared to a norm sample. Actually, the speech preparation lasted 11 min, because it was interspersed with three additional fear induction periods. Induction Period 2 was designed to induce fear of physical harm. It was announced that a blood sample would be drawn. A female assistant in a white coat brought in a metal kidney dish of syringes and placed it in the participants view. Induction Period 3 was designed to induce fear of evaluation. Introduced as a doctoral student, a male assistant,
280 Experimental Session II. About one week later, participants returned to the lab. Standard tests were identical to the first session, but emotion inductions now used emotional imagery. Participants performed two distinct, consecutive imagination tasks: Reliving a personal emotional episode and recalling the real-life emotion induction. Only the latter recall will be used here, because it ~a! leaves imagination content largely constant across participants, ~b! automatically includes treatment versus control group information, and ~c! best equates the stimulus and background information across real-life and imagination contexts. Fear and anger imagination followed the same scheme. After a 4-min rest period with ensuing emotion self-report, the reliving and then the recall task began, each introduced by a 1-min prestimulus period. Participants recalled the induction periods one by one ~four periods for fear, three periods for anger groups!. The male assistant supported the recollections encouraging especially the recall of bodily sensations during that period ~focus on response propositions; Miller et al., 1987!. He then asked the participant to imagine that situation as vividly as possible and to leave her eyes open. With a button press he started a 1-min data recording period. Then he continued with the next induction period. Following the last recall, participants completed the emotion selfreport for the most intensively recalled period and then scored the vividness of her imagery. Finally, participants were unhooked and led to another room for a postexperimental interview ~see Figure 1, Panel d, for the exact timing of the imagination task!. Variables Self-report of emotion. Participants performed an 11-point intensity rating on six unipolar ~0 not applicable, 10 completely applicable! and five bipolar ~505! scales tagged by one to four descriptive adjectives. The scales were selected to capture ~a! expected emotional feelings such as shame ~embarrassed0ridiculed0 ashamed0foolish!, fear ~frightened0timid0afraid0scared!, sadness ~sad0depressed0miserable0dejected!, happiness ~happy0gay0cheerful0delighted!, and anger ~angry0annoyed0mad0sore!; ~b! the bodily sensation of a pounding heart; ~c! feelings of arousal and valence such as tense versus relaxed ~nervous0restless0tense0wound up versus calm0relaxed0placid0at ease!, active versus tired ~energetic0 active0animated0lively versus tired0fatigued0sluggish0exhausted!, positive versus negative ~positive0pleasant versus negative0 unpleasant!; ~d! cognitive states such as alert versus confused ~alert0 attentive0receptive0lucid versus confused0baffled0perplexed!; and ~e! motivational states such as interested versus bored ~curious0 interested0motivated versus bored0indifferent0dull!. Self-reports of emotion were always collected right after physiological registration periods. Because we could not be sure that the emotions induced would last longer than 1 min, participants were not asked to rate their momentary affective state. Instead, participants were asked to recall and rate the emotional state that prevailed right before the registration period, that is, at the end of the preceding condition. Ratings were obtained during standard tests for the end of the rest period, for the loud noise and the handgrip tasks, and for the final minute of the exercise task; during real-life emotion inductions, for the end of the rest period, and separately for each of the three ~anger! or four ~fear! induction periods; during the imagination of emotion, for the end of the rest period and for the most intense of the recall periods. Imagery vividness. Vividness of imagination was assessed on a 7-point scale ranging from perfectly clear and as vivid as the actual experience ~1! to not clear or vivid, but recognizable ~4!

4 EEG variables and voice parameters were also obtained, but they are not analyzed here.
Constraints for emotion specificity Skin temperatures ~in degrees Celcius! were measured at the volar surface of the left little fingers distal phalanx and at the forehead with a TSD 102a fast response thermistor ~BIOPAC!. Sensitivity was 1.39 8C0V ~with a midpoint at 32.2 8C!; the frequency range was DC0.15 Hz; sampling rate was 1 Hz. Respiratory excursions were recorded with a thoracic strain gauge transducer ~TSD101, BIOPAC!. Amplification was 50, filters were set to DC and 10 Hz; sampling rate was 100 Hz. The BIO25 program determined respiration rate ~in min 1 ! as the frequency with peak amplitude in the power spectrum of 10-s data windows ~with Hanning window!. Electrodermal activity was recorded with a constant voltage of 0.5 V at the volar surfaces of the proximal phalanxes of the index and ring fingers on the left hand. Ag0AgCl electrodes ~TSD 103, BIOPAC! had a surface of 0.38 cm 2 ; they were filled with a 0.05 molar sodium chloride Unibase emulsion. Sensitivity was 100 mS0V with a frequency range of DC10 Hz and a sampling rate of 100 Hz. Phasic responses greater than 0.078 mS ~minimal slope of 0.007 mS0s, maximal half recovery time of 10 s! counted as skin conductance responses ~SCR!. BIO25 parameterization yielded SCR-amplitude ~in microsiemens! and the number of SCRs per minute. Skin conductance level ~SCL; in microsiemens per square centimeter! was separated from the raw signal with a 0.15-Hz low-pass filter and sampled with 1 Hz. Electromyograms ~EMG; in microvolts! were obtained from the m. extensor digitorum, m. zygomaticus major, and m. corrugator supercilii of the left body side through Ag0AgCl surface electrodes ~arm 8 mm, face 4 mm sensor diameter, In Vivo Metric!. Amplification was 1,000 with filters set to 10 Hz and 1 kHz. After rectification, the signals were low-pass filtered at 40 Hz ~24 dB0 octave! and sampled at 100 Hz. The vertical electrooculogram was derived from 4 mm Ag0 AgCl electrodes ~In Vivo Metric! placed on the vertical midline of the left eye right above the eyebrow and about 0.8 cm below the lower lid. Amplification was 200, frequency range was DC 30 Hz, and sampling rate was 100 Hz. The BIO25 program determined the number of eyeblinks per minute ~Eyeblink-No.!. Periods of physiological data recording were always 1 min long. During standard tests, nine periods were obtained: 1st, 5th, and 9th minute of the rest period as well as prestimulus and task values of loud noise, handgrip, and exercise. During real-life emotion inductions, six anger or seven fear periods were recorded: 1st and 4th minute of the rest period, prestimulus, and each of the induction periods. During the recall of Session I induction periods, eight anger or nine fear periods were collected: 1st and 4th minute of the rest period, prestimulus, and reliving, prestimulus, and recall of each of the induction periods. Thus, there were 34 data registration periods for the fear and 32 periods for the anger groups. For the current report, only a subset of periods will be analyzed ~see Figure 1!. Postexperimental interview. The standardized postexperimental interview comprised 30 questions at Session I and 13 questions at Session II, designed to further elucidate cognitive, emotional, and motivational outcomes of the experimental treatments. Personality questionnaires. Participants completed six questionnaires, three of which are relevant here: Betts Questionnaire Upon Mental Imagery ~Sheehan, 1967!, the State-Trait Anxiety Inventory ~Laux, Glanzmann, Schaffner, & Spielberger, 1981!, and the State-Trait Anger-Expression Inventory ~Schwenkmezger, Hodapp, & Spielberger, 1993!.

5 Error sums of squares in Emotion Group ANOVAs performed on difference scores during real-life emotion inductions were indeed markedly reduced by ANCOVAs with one to five covariates to 72%, 61%, 56%, 52%, and 48%. Numbers are averages across 29 physiological variables. 6 Specificity analyses were also performed using simple difference scores between response levels during emotion inductions and the 5th minute of the introductory rest period. Of the 29 t tests reported in the column Specific response rl in Table 1, 23 had the same sign and significance status ~6 negative sign and significant, 15 nonsignificant, 2 positive sign and significant!. ANCOVA showed five significant variables and difference score analysis one significant variable that was nonsignificant in the other analysis. Of the 29 t tests reported in the column Specific response im in
282 test the null hypothesis of no conditional mean differences among groups ~Huitema, 1980!, because ~a! participants had been randomly assigned to experimental groups and ~b! the covariates ~standard test periods! were assessed before any differential treatment had begun. Statistical data analysis. Statistical data analysis was performed in three steps. First, homogeneity of multiple emotion induction periods of real-life fear and anger inductions was checked to determine which periods could be averaged for subsequent analyses. Second, ANCOVAs on averaged induction periods were performed for each variable. Third, these univariate tests were supplemented by multivariate tests across all variables to guard against an inflation of the alpha error in multiple univariate tests. Homogeneity of response profiles during real-life emotion induction periods was determined by a test of profile parallelism ~i.e., the interaction of variables induction periods, see Morrison, 1976!. This test was performed on response levels of variables standardized across all observations registered. Dependent variables in subsequent statistical analyses were the means across homogeneous induction periods. Forming means also allowed calculation of internal consistencies ~Cronbachs alpha! of the dependent variables. To guarantee maximum comparability, means of physiological data during the induction periods of the imagination context were obtained across the same periods as during the reallife context. Self-reports of emotion during imagination could not be averaged because only one rating was performed. The overall statistical model used for hypothesis testing was a 2 ~Group! 2 ~Emotion! 2 ~Context! ANCOVA followed up by a priori specified contrasts ~see below!. Using PROC MIXED of SAS0STAT ~SAS Institute Inc., 1997!, the error variance-covariance matrix was specified to be completely general ~TYPEUN option! and to allow for heterogeneous error variances within each Group Emotion combination ~GROUP option!. These provisions tailored the statistical tests of contrasts to their appropriate error terms. Ten contrasts were specified. They were devised to test for ~a! emotion task effects ~differences between the tasks per se, i.e., without an emotional tuning; comparison of fear and anger induction control groups, separately within the real-life and the imagination context!, ~b! context effects ~comparison of control groups across contexts, separately for the fear and the anger inductions!, ~c! emotion responses ~the effects emotions had over and above task effects per se; comparison of treatment and control groups, separately within real-life fear, real-life anger, imagination fear, and imagination anger conditions!, and ~d! specific emotion responses ~the tetrad difference of the responses to the fear minus the anger emotion inductions in the treatment minus control group means, separately within each context!. An analysis of response profiles during emotion inductions supplemented the univariate analyses and also served to control for the accumulation of the alpha error present in multiple univariate, non a priori tests ~see below!. Two parameters of response profiles were analyzed, profile level and profile pattern. Profile level is the mean across profile variables. Levels were analyzed just like single variables; they depict the general response amplitude. Profile patTable 1, 27 had the same sign and significance status ~1 positive and 1 negative sign and significant, 25 nonsignificant!. In each analysis, one t test was significant but nonsignificant in the other analysis. Overall, ANCOVA displayed slightly higher statistical power but a highly similar picture in the pattern of results as seen with difference score analysis.

Figure 3. Adjusted means and standard errors of experimental groups ~left graph! and of selected contrasts ~right graph! for somatovisceral variables exhibiting a specific response in Table 1. f-rl fear induction real-life. a-rl anger induction real-life. f-im fear induction imagination. a-im anger induction imagination. er-f emotion response ~treatmentcontrol! contrast fear. er-a emotion response contrast anger. sr specific response ~er-fer-a!. For abbreviations of variables, see Table 1.
Did the Emotion Inductions Work? Emotion self-reports suggest successful emotion inductions, as seen both in the negatively valenced emotion responses ~i.e., treatment control group differences! and in the specific increases in reported
fear and anger during fear and anger inductions, respectively ~see a summary of results in Table 2, panels 1 and 2!. This conclusion applies to both the real-life and the imagination context. Somatovisceral responses, however, draw a different picture. Whereas the

Figure 3. Continued.

real-life fear and anger inductions elicited strong and patterned physiological responses, during imagination, only the anger induction did so, although somewhat dampened ~Table 2, panels 3 and 4!. Fear imagination evoked practically no physiological emotion response. Does this result warrant the conclusion that the fear recall failed to induce a fear response? If, as we propose, the answer is yes, why were self-reports of fear elevated and specific during fear recall?
One explanation would posit the operation of defensive interpretations of the real-life induction, when, one week later, the participants were asked to recall the real-life episodes. Because participants were debriefed at the end of Session I, fear treatment subjects might not have been able to imagine their real-life fear induction without interfering, fear-reducing thoughts that the threat was not real after all. This explanation could account for the lack of somatovisceral arousal; it would, however, be at variance with
288 Table 2. Summary of Emotion Responses and Specific Responses for Self-Report and Somatovisceral Variables within Contexts
Emotion response fear Emotion response anger 1. Self-Report of Emotion, Context Real-Life F Shame, Fear, Sadness, Anger, Heartbeat, Tense, Profile level f Happiness, Active, Positive, Alert, Interest Profiles nonparallel F Shame, Fear, Sadness, Anger, Heartbeat, Tense, Profile level f Happiness, Positive, Alert, Interest Profiles nonparallel 2. Self-Report of Emotion, Context Imagination F Shame, Fear, Sadness, Anger, Heartbeat, Tense, Profile level f Active, Positive, Alert F Shame, Fear, Sadness, Anger, Heartbeat, Tense, Profile level f Positive Profiles nonparallel 3. Physiological Variables, Context Real-Life F P-wave amplitude, relative Q-T time, Ejection speed, Heather index, CO, SBP, DBP, SCR-No., SCR-amplitude, SCL, Respiration rate, EMG zygomaticus, Profile level f Heart period, HP-variability, T-wave amplitude, ST-segment, Pe-Qs time, LVET, PEP, R-Z time, SV, TPR, PVA, PTT Profiles nonparallel F Heather index, CO, SBP, DBP, SCR-No., SCL, TMP-forehead, EMG extensor, Profile level f Heart period, T-wave amplitude, ST-segment, Pe-Qs time, LVET, PEP, R-Z time, SV, PVA, PTT, EMG corrugator Profiles nonparallel fF Fear aF Anger Profiles nonparallel fF Fear aF Anger Profiles nonparallel

289 autonomic responses. A well-documented mechanism, the defense reflex ~Hilton, 1982!, is of particular interest here. In the cat and in the rat, it can be elicited by L-glutamate microinjections into the midbrain periaqueductal gray matter ~Bandler, 1982; Bandler & Keay, 1996; Hilton & Redfern, 1987!. The defense reflex prepares an organism to cope with an emergency and specifically to perform the extreme muscular exertion of flight or attack. This is well exemplified by the pattern of cardiovascular response which is characteristic of the alerting stage of these reactions ~Hilton, 1982, p. 159!. The cardiovascular manifestations of the defense reflex include vasodilatation in skeletal muscle; venoconstriction; vasoconstriction in the splanchnic area, kidneys, and skin; increases in heart rate, in the contractile force of the heart, and in cardiac output; finally and most important for the present discussion, increases in both SBP and DBP, which together with the simultaneous tachycardia point to a profound inhibition of the baroreceptor reflex. The chronotropic, inotropic, and dromotropic response pattern, the marked vasoconstriction in the skin, and the pressor response in both SBP and DBP indeed suggest the elicitation of a defense reflex in both fear and anger real-life inductions. During anger imagination, signs for a defense reflex were weaker; cardiac output and left ventricular contractility increases were missing in particular. In sum, both the fear and the anger real-life inductions provoked an alerting response, which had a pattern that is very similar to the description of the defense reflex. During imagination, this alerting response was largely absent. Specificity in Somatovisceral Responses Nonparallel fear and anger response patterns suggested that at least part of the fear and anger responses did not overlap. The specific fear response ~Table 2! was an even more marked chronotropic and inotropic activation, which however lacked an even larger pressor response. Thus, the specific fear response was not just a stronger defense reflex, but resembled an adrenaline-like response pattern ~Sibolboro Mezzacappa, Kelsey, & Katkin, 1997; Wenger et al., 1960!. Funkenstein had advanced the adrenaline hypothesis of fear years ago ~Funkenstein, 1955; see also Wagner, 1989!. This result found during real-life closely parallels the fear response during imagination reported by Sinha et al. ~1992!, where CO, heart rate, and SBP increases as well as TPR, PEP, LVET, and DBP decreases were observed. The specific fear response in the present study also included more numerous SCR responses and larger EMG corrugator activity. The specific anger response comprised increased skin temperature at the forehead and larger EMG extensor activity during real-life, paralleling the results of Stemmler ~1989!. During imagination, these responses were also present ~see Figure 3!. Perhaps due to the slight anger response in anger controls, they missed significance when compared to the fear response. A hot face and the clenched fist are strong expressive signs of anger, identified here in miniature by their somatovisceral manifestations. The alleged specific DBP increase during anger in the present and in other real-life studies was probably masked by the marked defense reflex common to both fear and anger; but it did show up during imagination. A Component Model of Somatovisceral Response Organization During Fear and Anger This article started out from the notion of emotioncontext confounds suggesting ~a! that emotion and context would each exert their unique effects on somatovisceral responses and ~b! that the control groups would capture most of the context effects. How-

Neurobiological approaches to human disease ~pp. 143173!. Toronto: Huber. Jnig, W., & McLachlan, E. M. ~1992!. Characteristics of function-specific pathways in the sympathetic nervous system. Trends in Neurosciences, 15, 475 481. Lacey, J. I. ~1967!. Somatic response patterning and stress: Some revisions of activation theory. In M. H. Appley & R. Trumbull ~Eds.!, Psychological stress ~pp. 1437!. New York: Appleton-Century-Crofts. Lang, P. J. ~1968!. Fear reduction and fear behavior: Problems in treating a construct. In J. M. Shilen ~Ed.!, Research in psychotherapy ~Vol. 3, pp. 90102!. Washington, DC: American Psychological Association. Lang, P. J., Bradley, M. M., & Cuthbert, B. N. ~1990!. Emotion, attention, and the startle reflex. Psychological Review, 97, 377395. Lang, P. J., Bradley, M. M., & Cuthbert, B. N. ~1997!. Motivated attention: Affect, activation, and action. In P. J. Lang, R. F. Simons, & M. T. Balaban ~Eds.!, Attention and orienting: Sensory and motivational processes ~pp. 97135!. Mahwah, NJ: Lawrence Erlbaum. Laux, L., Glanzmann, P., Schaffner, P., & Spielberger, C. D. ~1981!. Das State-Trait-Angstinventar. Weinheim: Beltz. Lazarus, R. S. ~1991!. Emotion and adaptation. New York: Oxford University Press. Levenson, R. W., Carstensen, L. L., Friesen, W. V., & Ekman, P. ~1991!. Emotion, physiology, and expression in old age. Psychology and Aging, 6, 2835. Levenson, R. W., & Ekman, P. ~submitted!. Difficulty does not account for emotion-specific heart rate changes in the directed facial action task. Levenson, R. W., Ekman, P., & Friesen, W. V. ~1990!. Voluntary facial action generates emotion-specific autonomic nervous system activity. Psychophysiology, 27, 363384. Levenson, R. W., Ekman, P., Heider, K., & Friesen, W. V. ~1992!. Emotion and autonomic nervous system activity in the Minangkabau of West Sumatra. Journal of Personality and Social Psychology, 62, 972988. Mandler, G. ~1975!. The search for emotion. In L. Levi ~Ed.!, Emotions Their parameters and measurement ~pp. 115!. New York: Raven. Manstead, A. S. R. ~1992!. Gender differences in emotion. In A. Gale & M. W. Eysenck ~Eds.!, Handbook of individual differences: Biological perspectives ~pp. 355387!. Chichester, England: Wiley. Marwitz, M., & Stemmler, G. ~1998!. On the status of individual response specificity. Psychophysiology, 35, 115. Mezzacappa, S. E., Kelsey, R. M., & Katkin, E. S. ~1997!. The effects of epinephrine administration on cardiovascular function. Psychophysiology, 34, S64. Miller, G. A., & Kozak, M. J. ~1993!. Three-systems assessment and the construct of emotion. In N. Birbaumer & A. hman ~Eds.!, The structure of emotion: Psychophysiological, cognitive and clinical aspects ~pp. 31 47!. Toronto: Hogrefe & Huber. Miller, G. A., Levin, D. N., Kozak, M. J., Cook, E. W., McLean, A., & Lang, P. J. ~1987!. Individual differences in imagery and the psychophysiology of emotion. Cognition and Emotion, 1~4!, 367390. Morrison, D. F. ~1976!. Multivariate statistical methods. New York: McGraw-Hill. Nyklicek, I., Thayer, J. F., & van Doornen, L. J. P. ~1997!. Cardiorespiratory differentiation of musically-induced emotions. Journal of Psychophysiology, 11, 304321. Roberts, R. J., & Weerts, T. C. ~1982!. Cardiovascular responding during anger and fear imagery. Psychological Reports, 50, 219230.

~Received September 14, 1999; Accepted June 18, 2000!