Organisational Dynamics and Behaviour

Psychological Review 1994, Vol. 101, No. 2,211-221

Copyright 1994 by the American Psychological Association, Inc. 0033-295X/94/$3.00

The Varieties of Emotional Experience: A Meditation on James-Lange Theory

Peter J. Lang

James-Lange theory influenced a century of emotion research. This article traces the theory's ori- gins in philosophical psychology, considers differences in the thinking of James and Lange, and assesses Cannon's early critique and the resulting debate. Research is reviewed evaluating physiolog- ical patterns in emotion, the discordance of reported feelings and visceral reactivity, and the role of generalized arousal. NeoJamesian theories of attribution and appraisal—and alternative views based on dynamic psychology—are critically examined. A conception of emotion is presented, on the basis of developments unknown to James in conditioning theory, information processing, and neurosci- ence. Computational models of mentation are discussed, and implications are drawn for the classical debate on cognition and emotion. In concluding, new paths for emotion research are outlined and homage paid to the inspiration of William James.

From the first sentence of "The Physical Basis of Emotion," William James (1894) is clear that the essay's focus will be the human experience of emotion. Taking into account criticism of his original presentation of these thoughts in Mind (1884), and of his subsequent development of the same theme in the Princi- ples (1890), his Psychological Review Discussion is a closely rea- soned defense of what was already a well-known and controver- sial theory of emotional consciousness. His specific aim in all of these works was to show how certain "feelings" (the direct experience of the "coarser" emotions) are evoked in the con- scious mind. When finally free of the master's pen, however, James-Lange theory quickly escaped these repeated efforts at definitional limitation. Variously interpreted and elaborated by others, it has been alternatively the guiding light—or a lightning rod for criticism—for generations of emotion researchers. This commentary adumbrates that history—considering the theo- ry's origins in philosophic psychology, showing how reactions to James's view influenced the course of emotion studies in this century, and concluding with some thoughts on how James- Lange theory might be usefully considered by psychologists today.

Philosophical Psychology In the last decade of the 19th century, the new science of psy-

chology was quintessentially the study of mental life, and its quiddity was the conscious experience of human beings. "Ex-

This work was supported in part by National Institute of Mental Health Grants MH37757, MH41950, and MH43975, and Grant AG09779 from the National Institute of Aging to Peter J. Lang.

Thanks to Bruce Cuthbert for his comments on an earlier draft and to Margaret Bradley for her critique of later versions and editorial assis- tance.

Correspondence concerning this article should be addressed to Peter J. Lang, Center for Research in Psychophysiology, Box 100165 Health Sciences Center, University of Florida, Gainesville, Florida 32610- 0165.

perience" was the traditional domain of philosophy, and thus, theories offered in the nascent science overlapped markedly with precedent philosophical speculation. Furthermore, al- though these scholarly theories were highly sophisticated and creatively articulated, they shared the postulates of folk psy- chology. That is, the thoughts, images, feelings, wishes, and de- sires of reported mental life were all considered to be substan- tive facts. Thus, for those who were then scientifically inclined, feelings, for example, were either irreducible primitive entities or they were raw data to be explained.

James took the latter view—that feelings required explana- tion—and not surprisingly, the first critic he dealt with in the 1894 restatement of James-Lange theory was one of the great founders of psychology, Wilhelm Wundt. Wundt (1896) held feelings (Gefuhle) to be primary, unanalyzable, and in this sense similar to primitive visual or auditory sensations. Wundt al- lowed that different Gefuhle might be combined or elaborated, that physiological changes might exert a secondary influence, and conceded that the resulting broad Affekt was a complex mental event. He stoutly maintained, however, that initially ex- perienced feelings were the beginning and primitive essence of the emotion.

William James was no less convinced than Wundt of the im- portance of conscious feelings. He proposed, however, that emo- tional feelings were not truly givens. They were actually second- ary phenomena, prompted by the perception of somatic and visceral changes that had been elicited more or less directly by external stimulation. It is noteworthy that neither of these great scholarly antagonists based their views on what we would now consider experimental data (nor, incidentally, did Lange), but instead on thought. Wundt's Gefuhle were elements in a prodi- gious classificatory organization of mental experience founded on the introspective method; James's idea was a singular intu- ition, bolstered by anecdote and driven by a functional, prag- matic (if not yet behavioral) view.

Whatever its epistemological origins, James's idea has had profound effects both on the way psychologists think about emotion and subsequent laboratory studies. Almost every text

211

212 PETER J. LANG

on emotion research published in this century has begun with the James-Lange theory, and a host of related experiments have appeared in the journals over the years. Despite this avid inter- est, no experiment has directly tested its central hypothesis. In- deed, as stated by James, the conjecture that emotional feelings are the conscious leavings of physiological interoception is a sci- entific aporia and not open to direct test. Whereas physiological events can be measured with objective precision, the target of the enterprise—the subjective feeling state—does not yield to the scientist's clock, scale, or ruler.

James was, I think, well aware of the problem. In the 1892 edition of his Principles, he pondered the psychology of his time. He concluded that it had indeed become a natural science but "a psychology particularly fragile, and into which the waters of metaphysical criticism leak at every joint" (p. 467). Perhaps for this reason, James grew away from psychology in the last two decades of his life and rededicated his main efforts to philosoph- ical, psychical, and religious issues (Boring, 1950, p. 511; see also Murphy & Ballou, 1969; Vanden Burgt, 1981). After 1897, James happily left to Miinsterberg the direction of the Harvard laboratory he had founded and never sought experimental proof for his theory of emotion.

James and Lange

The Jamesian inheritance is the legacy of a transitional fig- ure—at once resonant of a dualistic philosophic past and herald of a biobehavioral future. His influence radiates forward from his high position on this great divide. Because of these mixed origins, however, his testament has often been misread, the be- quest spent elsewhere than was intended, and the theory vari- ously declared deceased, buried, disinterred and found alive, only to be buried again. James—were he still a witness to events—would surely marvel at the peregrinations of emotion research in this century.

James's general thoughts about the significance of emotion's physiology were not wholly unique. For example, the Australian psychologist Alexander Sutherland (1898) independently devel- oped a similar conception, from which he drew even broader social and biological implications. James, of course, felt that he and Lange (1885/1922) shared in the theory's discovery and gave Lange equal place in its 1894 restatement.

Viewed from afar, however, Lange's (1885/1922) position seems to have a separate emphasis. The Dane was a physiolo- gist—not a philosopher—and showed a scientist's concern for objectivity and an impatience with the merely mental: "No ob- ject can be dealt with scientifically unless it possesses objective qualities, concerning whose properties investigators agree" (p. 37). He did not share James's interest in explaining conscious feelings in the formal terminology of the time but sought, in- stead, a general explanation of emotional phenomena as mea- surable physiology. Furthermore, whereas James described a varied catalogue of somatic and visceral responses that were prompts to conscious emotional experience, Lange's hypothesis was specific: Emotion was a cardiovascular event. "The truly scientific problem in this field is the determination of the emo- tional reaction of the vasomotor system to various influences" (p. 80).

The Jogic of Lange's approach led him to create new organi-

zational categories, for instance, to "put joy and anger together in one class!"—for which he was chastised by Wundt. James appears to have agreed with Wundt (i.e., that conscious experi- ence is the arbiter of classification) and could only apologize for Lange's "too great stress on the vasomotor." Contemporary scientists might show more sympathy for Lange's approach— having seen 20th century quantum physics overturn the ac- cepted isomorphism of the experienced and the measured— and whatever the limitations of his taxonomy, appreciate a po- tentially data-driven scheme. In any event, Lange surely took Wundt's denigration as an accolade—that his theory was "one of those psychologischen Scheinerklarungen which assumes that science is satisfied when a psychic fact is once for all referred to a physiological ground" (James, 1894, p. 516).

The main idea shared by James and Lange was that emotion did not begin with the conscious experience of an affect. They both proposed that bodily and behavioral responses in emotion were prior events. The main difference in the thinking of Lange and James was that whereas James made conscious emotion the central focus of his theory (shifting its function from that of initiator of action to passive perceiver of physiological change), for Lange, emotion was its physiology, and conscious experience was of little importance, if not entirely epiphenomenal (see also Wenger's, 1950, views). These separate ideas have been beaten together historically, producing an omelette that tastes different to different diners. Thus, the James-Lange theory is cited vari- ously both in support of and against physiological reduction- ism, as the precursor of the behavioral analysis of emotion, and as the starting point for cognitive theories of affect.

Emotion Research After James: Physiological Studies

Cannon's Critique

The great emotion debate of this century's first half pitted the many defenders of James-Lange theory and the deceased James [1910] against a very lively physiological researcher, W. B. Can- non. Cannon (1927) held that the visceral changes in emotion were part of the body's general mobilization for action ("fight or flight"), a process that did not vary significantly over affects. Furthermore, he generated a compelling list of reasons why emotions could not be a perception of unique visceral changes: In short, chemically induced visceral responses alone do not produce emotions and the surgical isolation of the viscera does not wholly stop them; in any case, the viscera are too slow, too insensitive, and not sufficiently differentiated in reactivity to ac- count for the latency, dynamic range, and variety of emotional expression.

Cannon gave little thought to differences between the epony- mous theory's two authors. James, of course, was concerned with "feelings" as percepts and never limited their stimulus source to the viscera. Knight Dunlap (1920; an ardent positivist and pre-Watsonian advocate of behavioral psychology) dis- missed what he called James's "dualism" and defended the the- ory's Langeian version. At the Wittenberg symposium in 1927, Dunlap argued that similarity of autonomic patterns in intense affects was to be expected, and further, that Cannon's emphasis on visceral reactivity actually supported the Langeian view (Reymert, 1928, p. 159). Indeed, Lange and Cannon could have

SPECIAL ISSUE: JAMES-LANGE THEORY 213

discussed the issues quite amicably: Both held that emotion was fundamentally a physiological phenomenon and that the orga- nizing structures were subcortical. Lange (1885/1922, p. 74) lo- cated the emotion mediator in the vasomotor centers of the brainstem; Cannon (1928, with new data) placed the critical relay higher up in the diencephalon.

James never really addressed the question of brain pathways in emotion, except to speculate in the Principles that there were probably no specialized centers (James, 1890, chap. 25). Can- non clearly disagreed and opposed this view with his own tha- lamic theory. On the other hand, Cannon never studied experi- mentally the issues of reported conscious experience that con- cerned James. That is to say, he did not, himself, systematically record the introspective reports of emotion in human subjects and assess their relationship to concurrent somatovisceral changes. Similarly to James, he appealed to anecdote and case studies, leaned heavily on Maranon's (1924) study of the effects of sympatheto-mimetic drugs (as many have since), and per- force depended on behavioral evidence to define the parameters of emotion in his animal subjects.

The Psychophysiology of Emotion

Whatever its limitations, the agenda had been set by Cannon, and subsequent researchers worked from his outline of the crit- ical issues. Some (e.g., Bard, 1934) pressed forward on the main path, studying brain structures and circuits mediating emo- tion's behavior and peripheral physiology, using data gathered in the animal laboratory, and leaving the question of conscious- ness to future deliberations. Studies of affect and physiology in human subjects came to focus on two questions: Is arousal un- differentiated, as Cannon said, or can different physiological patterns be discerned for different emotions? Do peripheral physiology and verbal reports of emotion covary in ways that would support the hypothesis of interoception?

Activation theory. Cannon's concept of undifferentiated arousal—the same for fight or flight—was supported by many behaviorally oriented psychologists. Elizabeth Duffy (1934) even proposed that scientists give up the concept of emotion entirely, arguing that the referent phenomena were simply vari- ations in the energy level expended in goal-directed behavior. Auguring Schachter, she suggested that "the subjective qualities of'emotion'" were both a Jamesian "awareness" of the bodily changes that occurred with energy shifts and an "awareness" of the stimulus situation and response set (Duffy, 1941, p. 289). Activation theory was subsequently taken up by Lindsley (1951), bolstered on the one hand by new electroencephalo- graphic (EEG) studies of human subjects, and on the other, by animal research implicating the reticular activating system as a central mediator of activity level (Moruzzi & Magoun, 1949). Both data sets supported the idea of a continuum of activation from sleep to intense affect, with little indication of a difference between intense affective states.

Patterns of emotion. Despite this evidence, the idea of spe- cific emotional patterns has not lacked for supporters. This ad- vocacy is not exclusively Jamesian, however; it also reflects the broader, classical view that there are a finite number of funda- mental affective states (e.g., Jorgensen, 1928/1973; Darwin, 1872). More recent proponents of this conception include Tom-

kins (1962, 1963), Izard (1977), Plutchik (1962), and Ekman (1984), each with their own list of fundamental emotions, key differentiating responses, and view of the physiology-affect re- lationship.

Whatever its interpretation, evidence of physiological differ- entiation appeared apace with improvement in bioelectric re- cording apparatus. Malmo, Shagass, and Davis (1950), Ax (1953), J. Schachter (1957), and Graham, Stern, and Winokur (1958) were among those who found significant pattern differ- ences in emotion; similar findings have been reported more re- cently, for example, by Levenson, Ekman, and Friesen, (1990); Miller et al. (1987); and Roberts and Weerts (1982). In all this research, however, the amount of variance accounting for differences between emotional contexts is admittedly small, whereas the amount of variance accounting for generalized ac- tivation is large. Furthermore, differences are generally reflected in just a few of the measures assessed and these tend to vary over experiments. Although the overall mean values may be signifi- cantly different, only a subset of the subjects actually show the suggested affect discrimination.

John Lacey (1959, 1967) did much to organize how psychol- ogists might think about data of this kind.1 He held that pat- terns of physiological responses were driven by a host of in- teracting factors, many of which are psychologically silent and relevant only to the body's necessary housekeeping. Lacey con- strued psychophysiological differences in emotion to be in- stances of the more general hypothesis of "situational stereo- typy": There is a uniquely patterned, consistent reaction of all subjects to a defined group of specific stressor situations. Lacey also catalogued the difficulties in satisfying the hypothesis's re- quirements: Subjects do not all show "intra-stressorstereotypy" a reliable patterned reaction with repeated presentations of the same stressor. Many show "inter-stressor stereotypy": They re- act with the same general pattern to a variety of affective stimuli. Furthermore, visceral patterns change with the intensity of stimulation and with background levels of tension. Finally, re- activity is affected as much by the specific task context and its requirements as by emotional stimuli. Thus, heart rate, for ex- ample, tends to decrease—even in the presence of distressing affective content—when subjects are attending visually or audi- torily; on the other hand, heart rate accelerates when subjects recall and think about the very same unpleasant emotional ma- terial (Lacey, 1959; more recently, Lang, Bradley, & Cuthbert, 1990; Vrana & Lang, 1990). In summary, there is much evi- dence that (a) somatovisceral events reflect stimulus differenti- ation, (b) there are multiple patterns of arousal, and (c) the psy- chophysiology of these "environmental transactions" (to use Lacey's term) can be meaningfully interpreted. There is also every reason to doubt, however, that these data can be consis- tently organized around the same list of specific emotions.

It is not clear that James would be discomfited by this turn of events. Unlike Lange, he did not offer a specific catalogue of emotions, believing instead that they could be as varied as the "shapes of rocks on a New Hampshire farm." Furthermore, he was comfortable with individual differences: "The bodily vari-

1 Fora contemporary review of these general issues, see Cacioppo and Tassinary(1990).

214 PETER J. LANG

ations are within limits" and "still preserve enough, functional resemblance, to say the very least, in the midst of their diversity to lead us to call them by identical names" (James, 1894, p. 520). "Functional" was the key word here; no platonic idea of universal emotions was intended. As he said of "anger," "there is no definite affection . . . in an 'entitative' sense" (p. 520). That is, like Lacey, he held that an emotion has no reality ab- stracted from the circumstances of its occurrence.

Physiology and feelings. Studies of the association between physiological change and reports of feelings come closer—more so than patterning—to addressing the primary Jamesian query: Is emotion's feeling the apprehension of its visceral and somatic expression? Maranon's (1924) experiment, cited by Cannon, and Hohmann's (1966) more recent clinical investigation are representative of efforts to answer the question in a context of radically altered physiology. Maranon administered epinephe- rine to subjects, eliciting a general activation of the viscera, and asked them how they felt. Hohmann studied patients with di- minished visceral afference, because of spinal cord injury, and posed similar questions. It is not surprising that Maranon's sub- jects did not suddenly hallucinate an emotional episode, but many reported "as if" emotions (e.g., "I feel as if I were anx- ious—hut I'm not"). Hohmann's paraplegic subjects reported that they felt emotions less strongly than formerly. The reduc- tion did not "necessarily" include emotional behaviors, how- ever, and some patients were even "regarded as being extremely aggressive." Results like these please only the committed. How does one discriminate real feelings from "as if"? How does one understand "reduced" feelings in individuals when "they swear, they throw their urinals around, they show overt signs of anger" (Hohmann, 1966, p. 154)? In James's day, few questioned that emotions were all of a piece: a fusion of experienced emotion, expressive behavior, and physiology. Experiments performed more recently, however, have brought this common sense view into question.

Response discordance. In an important series of experi- ments, Mandler and colleagues showed that there is surprisingly little covariation between judgments of anxiety or of visceral arousal and actual visceral change during stress. Correlations between autonomic perception and autonomic activity gener- ally accounted for less than 9% of total experimental variance (Mandler & Kremen, 1958), with many individuals showing systematic reporting biases and either over- or underestimating actual physiological response (Mandler, Mandler, & Uviller, 1958). It is interesting that higher correlations with visceral change (as high as .75) were found for expressive verbal behavior (e.g., spontaneous affective expressions ["That's horrible!"] or interference in normal speech patterns; see Mandler, Mandler, Kremen, & Sholiton, 1961). Thus, the most dramatic discor- dance specifically occurred with reports of feelings. It is the sub- ject's judgment of felt affect (the main concern of James and Wundt) that shows this remarkable independence.

Subsequent research with anxious and fearful patients has yielded similar evidence of emotional response discordance (e.g., Cook, Melamed, Cuthbert, McNeil, & Lang, 1988; Lang, 1964, 1968, 1978; Marks & Huson, 1973). Furthermore, three responding systems (verbal judgments of fear, fear behaviors, and physiological reactivity) show a highly desynchronous pat- tern of change with treatment (Rachman & Hodgson, 1974).

These results undermine the Jamesian hypothesis. How can re- ports of feelings be a perception of the physiology, if the two events are not reliably coincident? The problem does not ap- pear with the average response of samples: In general, mean rat- ings of distress increase with various physiological parameters. The problem lies in the diversity of individual reactions. Al- though there is a rough majority of apparent Jamesians— frowning and sweating, and calling themselves anxious—there is also a palpable minority doing the exact opposite. Figure 1, for example, shows individual correlational results from a re- cent experiment (Lang, Greenwald, Bradley, & Hamm, 1993) in which subjects looked at 21 emotionally evocative pictures (e.g., happy babies, grisly mutilations, erotica, frightening ani- mals). The pictures that prompt the largest average physiologi- cal reactions are those for which the strongest feelings are re- ported, and there are interesting and reliable sex differences. The correlations within individuals, however, show enormous range, and many fail to show any significant covariation either between feelings of displeasure and somatically mediated facial display muscles or between felt arousal and measures of sympa- thetic activation (e.g., skin conductance). When researchers have tackled interoception directly, as a nonaffective perceptual task, they have obtained similar findings: Whereas some sub- jects are able to monitor selected visceral responses with some accuracy, many others are wholly unaware of even their own heart beat (e.g., Katkin, 1985; Schandry, Bestler, & Montoya, 1993).

Cognition and Emotion

Modern Phenomenology

The most popular, current view of emotion owes much to gestalt psychology. Gestalt theory developed independently and somewhat later than James's functionalism: first, as a separate alternative to the Wundtian mental chemistry, and later, as a bastion of subjective reality against the behaviorists. The Ge- staltists retrospectively applauded James's efforts to refute Wundt's structuralism but, at the same time, viewed physiolog- ical interoception as too narrow a definition of emotion (Koffka, 1935, p. 401): Emotions involved the whole phenome- nal field, to which the physiological changes made only a partial contribution.

The gestalt movement subsequently (e.g., through Lewin, 1936) had a strong impact on social psychologists who, at least since McDougal (1910), considered the emotions their mandate for study. In this domain, the ideas that would come to intersect most directly with those of James were promulgated by Fritz Heider. Heider (1944) developed a highly influential theory of phenomenal causality as the nexus of social perception. He pro- posed that "a change in the environment gains its meaning from the source to which it is attributed" (p. 372). This causal inte- gration determines whether the individual responds affectively or not—as decreed by gestalt laws—with the same inevitability that different spatial organizations of three dots prompt percep- tion of either a triangle or a straight line. This approach was very close to earlier philosophical roots. Thus, Heider (p. 365) quoted the 17th century sage Spinoza who, despite the formal language, sounds remarkably modern: "If we remove a distur-

SPECIAL ISSUE: JAMES-LANGE THEORY 215

QJ

8 c 0

CBo> a.

1.0

0.8-

0.6-

0.4-

0.2-

0.0-

-0.2-

-0.4.

-0.6

Reported Unpleasantness & Corrugator ('Frown') EMG

f°Dlmm

aODOOOO°°(l||."

na0 »* D **

QDD m*

»»* •«"*

™

Bo D

* D Females 1 * i Males 1

c

Reported Arousal & Skin Conductance Activity

-,00

OD

Individual subjects, ranked (within gender) by size of correlation

Figure 1. Individual covariations for each of 64 subjects relating their reports of emotional experience (feelings of arousal or displeasure) and physiological reactions (skin conductance reactivity or corrugator "frown" muscle activity). Each point is one subject's zero-order Pearson correlation between affective judg- ments and physiological responses when viewing a set of 21 emotional pictures. Correlations above the upper line (r > .37) are statistically significant; those between the two lines (0 < r < .37) are in the predicted direction but" not significant; and those below the lower line (r < 0) are opposite to those expected. Some of the individual differences are accounted for by sex, with women appearing to be better Jamesians for facial muscle activity, whereas men are more reliably Jamesian for skin conductance (see Lang, Greenwald, Brad- ley, & Hamm, 1993). EMG = electromyogram.

bance of the spirit, or emotion, from the thought of an external cause, and unite it to the other thoughts, then will the love or hatred toward that external cause, and also the vacillations of spirit which arise from these emotions, be destroyed" (Spinoza, Circa 1632-1677/1936, p. 254).

It was S. Schachter (1959; S. Schachter & Singer, 1962) who brought the Spinoza-Heider conception to bear on the James- Lange theory and thereby created what is for many the modern view. Schachter agreed with James's first principle: "The indi- vidual will react emotionally or describe his feelings as emo- tions only to the extent that he experiences a state of physiolog- ical arousal" (S. Schachter & Singer, 1962, p. 382). On the other hand, he also proposed that arousal "differences are at best rather subtle and that the variety of emotion, mood, and feeling states are by no means matched by an equal variety of visceral patterns" (p. 380). In short, Schachter disagreed with James and Lange on the patterning issue; he agreed with Cannon, Duffy, and Lindsley that emotional physiology was essentially an un- differentiated activation.

In this view, emotional specificity—and variety—is created by cognitive factors; that is, a state of arousal prompts an "eval- uative need" (Festinger, 1954), an effort by the person to ac-

count for the perceived activation. The need is resolved "in terms of his knowledge of the immediate situation" (S. Schachter & Singer, 1962, p. 381). Thus, undifferentiated acti- vation in the presence of a beautiful woman would lead a man to "decide that he was wildly in love" (p. 381); given the same physiology, James's attacking bear would prompt fear. As in James's (1894) analysis, Schachter gave little consideration to how physiological arousal comes to be there in the first place. (Similarly to Maranon, Schachter induced sympathetic activa- tion chemically; external "emotional" stimuli were not the pri- mary variables.)

Attribution and Appraisal

Mandler (1975,1984) accepted much of the attribution doc- trine proposed by Schachter but also tried to address the pri- mary stimulus issue. In his elaboration of the theory, he sug- gested that activation in emotion resulted from any interrup- tion or disruption in a goal-directed behavioral sequence. This concept was put forward earlier by Lewin (1935,1940) and also by the more biologically oriented Hebb (1949), who described

216 PETER J. LANG

emotion as the consequence of discontinuity in a phase se- quence.

Other cognitive theorists, however, held that the emotional quality of a stimulus was determined by a prior mental analysis. They viewed the role Schachter assigned to evaluative judg- ment—attribution of an ongoing arousal—as too narrow and too late. Lazarus (1982; Lazarus, Averill, & Opton, 1970), for example, proposed that the whole phenomenal field (to use the gestalt term) was subject to evaluative scrutiny and interpreta- tion as part of the affective process. Elaborating a term coined by Arnold (1960), he argued that all emotion began with an appraisal of stimulus significance. Furthermore, appraisal oc- curred in stages: initially, to determine if an input is indeed emotionally evocative; and then a reappraisal, to assess any late arriving returns (perhaps including ongoing physiological events) and, also, to make a response selection.

These ideas have been highly appealing to clinicians, whose mills turn on patients' introspective reports, that is, patients' interpretations of their physiology and behavior, and the mean- ing they attribute to life events. Aaron Beck, for example, has developed a whole school of psychotherapy based on this view (Beck, Emery, & Greenberg, 1985). Thus, he suggested that anxiety and panic might first arise from a misattribution of physiological arousal (e.g., as indicating a heart attack). Think- ing about this distressing "cognition" is, however, itself a prompt to further arousal and more anxious behavior, and so the causal chain becomes an invidious cognitive/physiological loop. In this circumstance, the therapist's task is to break the loop—to replace the misguided "cognitions" with a more real- istic stimulus assessment, as Spinoza had earlier suggested.

It is difficult to know how turn-of-the-century thinkers would view (could they now return) these contemporary exercises in phenomenological analysis. Lange would probably have little patience with attribution. Given a short briefing on modern ter- minology, however, the more philosophical James could have disputed quite capably in this arena. After all, he had already said that the "primitive arousers" of the emotional reflex "take their place, as experience goes on, as elements in total 'situa- tions'" (James, 1894, p. 518). Thus, he was on record as think- ing that other "ideas" modify emotional reactions. Further- more, he dealt effectively in his Psychological Review article with Mr. Iron's 19th century attribution theory (remarkably similar to Beck's), that the "bodily symptoms" in fear were ac- tually the fear object, because "of their unknown or indefinite evil consequences" (p. 522). But would James be pleased to see that these Spinozian issues are still au courant, or would the scientist in James have been disappointed?

A Post-Modern View of the Jamesian Legacy

William James did not experience this century's great sea changes in psychology: Behaviorism's bold, mindless surge was less than a nascent swell in the last decade of James's life. James met Sigmund Freud in 1909 at Clark University (during the famed psychiatrist's only visit to America), and with his usual receptivity to new conceptions, he later wrote, "I hope that Freud and his pupils will push their ideas to the utmost limits, so that we can learn what they are" (Rosenzweig, 1992, p. 174). / think James would have been astonished, however, to see how

successful they would be—how psychoanalytic theory would, for a time, carry all along in its dark current. James, of course, had no inkling at all that the modest flow of experiments, by Cannon, Bard, and other pioneers, would become the flood of contemporary neuroscience. Finally, nothing at the century's start suggested that a fresh wind would send scudding the mind's lowering Cartesian clouds: James could not know that a whole new concept of mentation would emerge from Turing's and von Neumann's mid-century visions of the computer (e.g., see Goldstine, 1972)—that the brain's mind could be seen as an information processor, needing no homunculoid pilot, ap- praiser, or perceiver.

Motivation and Emotion

These developments are of great scope and, in their aggregate, require a basic reformulation of the issues in emotion that James addressed. First, whatever excesses have been promul- gated in the name of behaviorism, its central, positivist lesson is incontrovertible: The raw data of the psychological enterprise can only be the observable reactions of organisms. Reports of emotional experience can be observed and analyzed, but the reports themselves, however honestly professed, are not them- selves scientific observations. By this I mean, of course, that they are private and not confirmable.

James's theory could be seen as an effort to render experience measurable or, at least, to create a meaningful psychophysics of emotion. When we ask subjects how angry or afraid they are, and to rate that feeling on a scale from 1 to 9, it is somewhat like asking them how loud a sound is or how bright a color is. In the latter two cases, there are external, physical referents for intensity and luminosity, and we can plot subjects' estimates against these referents, discovering lawful relationships between sensory knowledge and known external stimuli. If James had been correct, and feelings were perceptions of emotion's physi- ology, then an analogous psychophysics—interoceptive, rather than exteroceptive—could have been established.

It now seems clear, however, that emotional feelings are not these simple physiological perceptions: (a) Reports of feelings and visceral arousal are often noncoincident; (b) visceral moni- toring is generally poor; and even leaving aside the question of conscious perception, (c) efforts to show that emotional lan- guage—or behavior—depends primarily on feedback from the viscera (or the face, as some have suggested; e.g., see Laird, 1984) have not produced compelling data. It appears, rather, that emotion's language, expression, and physiology can be shaped by a variety of environmental events and that changing contextual demands, sex, and temperamental factors contribute to the formation of unique individual response patterns.

Despite this diversity, a synergy of language, behavioral, and physiological responses (and a reduction in individual differ- ences) generally characterizes strong emotions—a fact James appears to have appreciated, as he chose to limit his theory to the "coarser" affects. This increased covariation with intensity implies an underlying organizing factor. The conditioning liter- ature of the 20th century suggests that this factor is biological and motivational. As Konorski (1967; see also Mackintosh, 1983) and other Pavlovians have argued, affects are engaged, on the one hand, by appetitive states, prompting approach, con-

SPECIAL ISSUE: JAMES-LANGE THEORY 217

summatory, nurturant, positive behaviors; and on the other by aversive states, prompting defensive, protective, negatively va- lent responses. Both motivational states vary in intensity of ac- tivation, modulating the energy expenditure of the organism. As a vast literature confirms, related incentives and reinforcers steer behavior, controlling what is learned and remembered.

The idea that the variety of emotion might be organized around simpler motivational parameters has also emerged from studies of affective language. Curiously, this approach can be traced back to Wundt (1896) who, on the basis of introspection, proposed that there were three primary dimensions of feeling: Lust (pleasure), Spannung (tension), and Beruhigung (inhibi- tion). Titchener (1908) brought this view to the United States, where he gave primary emphasis to the pleasantness-unpleas- antness dimension. Osgood, Suci, and Tannenbaum (1957) later reexamined the issue empirically. They collected subjects* evaluative responses to a great variety of stimuli, using factor analysis to sort out the covariation between different verbal de- scriptors. Dimensions similar to those of Wundt reappeared, with affective valence (pleasant-unpleasant) accounting for the most of the variance, followed closely by arousal (calm- aroused). More contemporary investigators (e.g., Mehrabian, 1970; Mehrabian & Russell, 1974; Russell, 1980) have repeat- edly reaffirmed that valence and arousal are reliable, factorial dimensions that organize the diversity of evaluative judgments. Furthermore, these same dimensions account best for somatic and autonomic changes that accompany emotional perception and the retrieval of affective memories (Lang, Bradley, & Cuthbert, 1990).

This parallel in concepts and data from widely different liter- atures—semantic analysis and conditioning—is of major sig- nificance. It encourages the hypothesis that emotions are driven by an underlying biology—similar in both human beings and animals—determining not just the intensity (activation level) of physiological mobilization but also the general directional disposition (appetitive approach or aversive withdrawal) of sub- sequent emotional actions. This conception is very different from the neophenomenological view that only activation is bio- logically grounded and that response direction is determined by a rational, attributive, or appraisal process.

Neuroscience on the March

William James concluded his 1894 essay with a definitive statement on the nature of emotion: "The name 'emotion' should be suffered to connote organic excitement as the distinc- tive feature of the state" (p. 529). Thus, for James, activation was no less pertinent a feature of emotion than it later was for Lindsley (1951) or Duffy (1941). It is important to note that James further held that the initiating processes that determine particular types of emotional reactions were also fundamentally physiological. Affective consciousness—the focus of so much theoretical concern over the century—was, for James, clearly a secondary event. Also interesting is that James did not present his ideas on the primary determination of emotions in his 1894 discussion of affective consciousness, nor is it considered in the chapter on emotion in the Principles. The issue is addressed separately in an earlier chapter entitled "Instincts,"

In James's (1890) view, emotions begin with stimuli that

evoke instincts—"reflex type" reactions that are built into or- ganisms: "Instinctive reactions and emotional expressions shade imperceptibly into each other. Every object that excites an instinct excites an emotion as weir (James, 1890, p. 442). Fear, for example, could be evoked automatically in infants by such stimuli as loud "noises," "dark places," or "strange men, and strange animals" (what Seligman, 1971, might call "prepared" stimuli). Furthermore, as instincts, such reactions might "ripen at a certain age and then fade away" (James, 1890, p. 398)—like the imprinting phenomena later described by Lorenz (1952). In a similarly modern way, emotional reflexes were not viewed as invariant: These "reactions will be much modified if the in- stincts combine with experience" (James, 1890, p. 390). All or- ganisms had a "bundle of such reactions"—man more than most—that are shaped and combined in environmental transactions, rather like Gibsonian (1966) "affordances."

Although unable to elucidate its physical basis, James (1890) emphasized that his was a "physiological conception of what an instinct is" (p. 391). Since his time, however, neuroscience has greatly increased understanding of this first level of emotional processing. Furthermore, the newer work does not only expli- cate neural circuits relevant to activation (the path pioneered by Cannon and Bard); in addition, stimulated in part by the early theorizing of Papez (1937), researchers have also shown the significance of specific neural circuits (e.g., the hypotha- lamic and limbic systems; McLean, 1949; Olds & Olds, 1961) and their neurochemistry for the appetitive or aversive direction of emotional reactions.

Recent studies (Davis, 1989; LeDoux, 1990; see also Pank- sepp, Sacks, Crepeau, & Abbott, 1991; Thompson, 1986) have traced, for example, the input/output pathway of uncondi- tioned reflexes to painful shock, and there is now a good neural model of learned fear. The data suggest that aversive stimulation activates a subcortical pathway proceeding from the thalamus to amygdaloid nuclei deep beneath the temporal lobes. From this site there are separate branching circuits (LeDoux, 1990) either through the central pay, mediating somatic emotional responses (e.g., "freezing" in the rat) or through the lateral re- gion of the hypothalamus, activating the viscera (e.g., blood pressure increase). New stimuli, if evoked contingently, subse- quently activate this same circuit. It is interesting that the learned fear response does not appear to depend on a complex analysis of the cue stimulus (appraisal?): That is, the relay is through the primitive thalamus and will proceed in the absence of an intact sensory cortex (e.g., see DiCara, Braun, & Pappas, 1970).

Once established, this neural network can influence other as- sociated behavior. It has been shown in the rat (Davis, 1989), for example, that activation of the central nucleus of the amyg- dala (by either primitive or learned fear cues) subsequently prompts activity in the reticular activating system, intersecting the arc of the startle reflex. Thus, if a startle response is evoked by an entirely new stimulus—but in the context of the condi- tioned fear cue—the amplitude of the startle reaction is auto- matically increased. This is a primitive example of mood-de- pendent enhancement of affectively congruent behavior: That is, an aversive response (the startle reaction) is potentiated in the context of a learned aversive state. Human subjects (e.g., see Lang et al, 1990; Bradley, Greenwald, Petry, & Lang, 1992)

218 PETER J. LANG

show a similar potentiation of the startle reflex to probes ad- ministered while they view unpleasant, fearful events (and a rel- ative startle inhibition in a pleasant context). This same neural circuitry might mediate more complex behavior, for instance, as when, following the Velten (1968) unpleasant mood induc- tion procedure, human participants are more attentive to aver- sive than positive input or more readily recall unpleasant than pleasant memories (Blaney, 1986).

Back to the Future: Mentation and Emotion

William James, like nearly every thinker at the century's turn, held to some variant of the Cartesian theory of mentation. This general view presumes a subsisting knower and doer, holds that conscious experience is self-evident fact, and enshrines ra- tional deliberation as the model of cognition. From this per- spective, knowledge is what we tell ourselves, and mental life is reflected in meaningful narrative; more important, the basic categories and organization of conscious thinking and feeling dictate the parameters of psychological and physiological the- ory. These are still the unquestioned psychological assumptions of the average man, of the law, and of our social and political institutions. We are all, more or less, under the spell of this view—witnesses to, and heroes of, our own lives.

Freudian Interlude

Despite its ubiquity, this grand conception poses serious problems for a science of emotion. Most important, as theory it cannot wholly accommodate the data of human affective behav- ior. This was the main message that Sigmund Freud brought to the Clark University campus in 1909, where James heard his lecture. From observations of his patients, it was clear to Freud that complex behavioral and physiological responses in emo- tion could occur unaccompanied by either conscious intention or awareness. Freud was alerting his psychological colleagues to what we earlier referred to as response discordance. That is, the components of emotion do not always fit together as neatly as the theory of conscious experience supposes, and in psychopa- thology, this discontinuity can be extreme.

Freud's solution to the problem was a prodigous tour de force. He invented a whole new level of experience—the uncon- scious—sometimes populated with not one but several homun- culi, all with intentions, desires, and feelings, and all beyond the reach of normal awareness. This new level of experience could, according to psychoanalysis, account for all the troubling events that have no conscious explanation.

As Skinner (1954) once suggested, the Freudian unconscious is like a giant Ptolemaic epicycle, shoring up a world conception that really needed replacement. Thus, it did not solve the prob- lem of mentation's relationship to emotion and the other co- nundrums of mind and behavior but merely shifted explanation from one set of untestable hypotheses to a parallel set, even fur- ther beyond the pale. Nevertheless, the cultural impact of the Freudian view has been, and continues to be, enormous. It is more difficult to assess its significance for science: On the one hand, psychoanalysis provided a broad rationale for addressing phenomena empirically that might otherwise have been ne- gtected, for example, problems of sexual motivation and social-

personality development; on the other hand, it could provide no real computational framework for the data collected in its name.

Skinner's solution to the problem of mind was radical sur- gery. He proposed to simply cut out mentation (and much of physiology) and content ourselves with the study of behavior, unadorned. Although this austere path has its adherents, many have found it difficult to fit and hold together the complexities of behavior with absolutely no mental mortar. Fortunately, a dramatic third alternative has appeared, which both comes to grips with mentation and yet promises scientific rigor: In the way that clockwork technology, new in the 18th century, pro- vided a model of how biological organisms might be understood as mechanisms (La Mettrie, 1748/1953), so the 20th century computer now provides scientists with a new mechanistic model of mind (e.g., see Gardner, 1985;Gunderson, 1971).

Processing Emotion Information

In the contemporary cognitive view, the brain is an organ for storing and processing information. Mentation is conceived as analogous to the operations of a computing machine. Reports of thoughts, memories, and feelings are outputs, reflecting com- putational events in the brain associated with the acquisition of new information and the reorganization of the memory store. The new analytic problem is to define information structures and processing programs that link stimulus input with physiol- ogy, language, and behavior. Taking this view, an emotion theory could be constructed in the classical way, guided by hypotheses from introspection (e.g., as Newell & Simon, 1963, studied problem solving); however, it is also possible to create an artifi- cial intelligence that mediates input and output, unencumbered by traditional (but perhaps misleading) preconceptions about consciousness. Furthermore, the theory can be constrained by our knowledge of neural processes, building a model of mental operations that is both a cognitive and a neural network (Grossberg, 1991).

Emotional memory. In recent years, psychologists have be- gun to consider emotion as information processing (e.g., Bower, 1981; Lang, 1979; Mandler, 1984). This approach is still at an early stage of development, without broad agreement on any specific theory; however, important implications can already be drawn for the issues addressed by James and his successors. The most relevant work concerns the organization and content of memory: Ignoring the debate about structural specifics (e.g., se- rial or parallel; local or distributed; symbolic level or neural only), it is generally assumed that information about events or behaviors has some unit representation and that there is con- nectivity among representations. Thus, a memory of an emo- tional episode can be seen as an information network that in- cludes units representing emotional stimuli, somatic or visceral responses, and related semantic (interpretive) knowledge. The memory is activated by input that matches some of its repre- sentations. Because of the implicit connectivity, the other rep- resentations in the structure are also automatically engaged, and as the circuit is associative, any of the units might initiate or subsequently contribute to this process (Lang, 1985).

The biological foundation of such an emotion network could be a parallel neural net, or more probably, a great many paral-

SPECIAL ISSUE: JAMES-LANGE THEORY 219

lei, interlinked neural subnets. In any event, units are likely to be multiply connected and are potentially parts of different overlapping knowledge structures. Furthermore, such nets are not thought of as static. In the connectionist approach champi- oned by McClelland and Rumelhart (1986), for example, the weighting of individual units changes with activation as the sys- tem stabilizes, and the same net may even have more than one stable state (see the Feldman & Ballard, 1982, model of the Necker cube). It recalls James's notion that similar stimulus sit- uations can engage very different "instincts": "Nature implants contrary impulses to act on many classes of things, and leaves it to slight alterations in the conditions of the individual case to decide which impulse shall carry the day" (James, 1890, p. 392).

Emotional memories (particularly James's "coarse" emo- tions) would differ in important ways from perceptual and de- clarative information nets, and most other episodic memories, in that they are action dispositions (Frijda, 1986). Furthermore, they are about strongly motivated behavior: freezing, fleeing, attacking, sexual approach. Thus, an emotion network would include connections to subcortical motivation circuits, mediat- ing response programs for visceral mobilization, and approach or defense (Lang, 1993). These connections to "hard-wired" re- flex systems lend a network coherence and may account for their persistence, for example, as in simple phobia or the idio- syncratic patterns of sexual arousal. As LeDoux (1990) argued from the animal conditioning data, emotional memories are to some extent indelible.

Cognition or emotion? James and Wundt debated the la- tency of experienced feelings: Do they precede or follow emo- tion's physiological expression? More recently, Lazarus (1982) and Zajonc (1980) disputed a similar question: Do emotional reactions precede or follow their phenomenological "ap- praisal"? As the century nears its end, information-processing models should be at last carrying us past these issues. Only if the definition of cognition is arbitrarily restricted to rational conscious thought can it be placed in opposition to emotion— and even then, only weakly. Emotional expression is itself clearly an elaborate cognitive processing task in which data are integrated from many sources in the brain (often in millisec- onds), occurring, in the main, outside awareness. Contempo- rary analysis encourages the view that the conscious narrative flow of evaluations, appraisals, historical interpretations, and affective judgments—the reported story of the emotion—is of- ten an afterword, significant as a memorial record but only pe- ripherally related to the cognitive mechanics of an ongoing emotion.

The Physical Basis of Emotion: A Postscript

Since the time of James and Lange, our knowledge of the physical basis of emotion has greatly increased in depth and complexity. Powerful new tools are available for neurochemical analysis and brain imaging (e.g., Crease, 1993). Furthermore, advancing general understanding now requires that we consider interactions between three vast and very different literatures: First, there is the objective study of behavior as organized by the principles of conditioning. From Konorski to Rescorla, there is much here that deals effectively with basic psychophysiological

issues in emotion—the acquisition of affects, their motivation, the relationship between affect and context—which needs to be considered by all emotion researchers. Second, cognitive scien- tists need to get themselves deeper into the emotion business, and emotion researchers need to become more attuned to con- temporary cognitive studies. Emotion theory based on intro- spection and phenomenology has had a limited yield; it is time to experiment with the less isomorphic (but potentially power- ful) mental models of artificial intelligence. (Although, if phi- losophers [e.g., Dennett, 1991] succeed in dressing "conscious- ness" in more fashionable clothes, we might yet lend some ob- jective substance to that quirky Jamesian stream.) Third, we naively need to temper behavioral and cognitive views of affect with what we know about the neurophysiology and neurochem- ical foundations of emotional behavior and stick close to our animal models. This helps to limit inference and to keep our cognitive theories away from empty boxology. I do not propose that we naively accept psychologischen Scheinerklarungen (wholly satisfied with physiological explanations), and it may yet be too early for a true, integrative cognitive neuroscience. However, cognitive and neural scientists must remain alert to parallel developments in each other's field, as both have the power to productively guide—and to mislead (e.g., as when neu- rologists are sent searching for the anatomical locus of anxious or depressed "feelings").

Looking back at James-Lange theory from a centennial per- spective: It was the right idea for its time. It provided a bridge, addressing issues from the older philosophical psychology, fa- miliar since Spinoza, Descartes, and Kant, but conceived in the modern way that requires a grounding of psychological con- cepts in objective measurement. In uniting these views, the the- ory provided a focus for researchers of varied disciplines and orientations, leading to the acquisition of important new knowledge and productive efforts to reformulate the research problem. Finally, the theory is instructive for contemporary meditation because it is one of the few that successfully chal- lenged the folk psychology of its time. James's openness, inge- nuity, and vision—and Lange's realism—should inspire no less now than they did at the century's start.

References Arnold, M. R. (1960). Emotion and personality. New York: Columbia

University Press. Ax, A. F. (1953). The physiological differentiation between fear and an-

ger in humans. Psychosomatic Medicine, 15, 433-442. Bard, P. (1934). The neuro-humoral basis of emotional reactions. In C.

Murchinson (Ed.), Handbook of general experimental psychology (pp. 264-311). Worcester, MA: Clark University Press.

Beck, A. X, Emery, G., & Greenberg, R. L. (1985). Anxiety disorders and phobias. New "York: Basic Books.

Blaney, P. H. (1986). Affect and memory: A review. Psychological Bul- letin, 99, 229-246.

Boring, E. G. (1950). A history of experimental psychology. New York: Appleton-Century-Crofts.

Bower, G. H. (1981). Mood and memory. American Psychologist, 36, 129-148.

Bradley, M. M., Greenwald, M. K., Petry, M. C., & Lang, P. J. (1992). Remembering pictures: Pleasure and arousal in memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 18, 379-390.

220 PETER J. LANG

Cacioppo, J. X, & Tassinary, L. G. (1990). Inferring psychological sig- nificance from physiological signals. American Psychologist, 45, 16- 28,

Cannon, W. B. (1927). The James-Lange theory of emotions: A critical examination and an alternative theory. American Journal of Psychol- ogy, 39, 106-124.

Cannon, W. B. (1928). Bodily changes in pain, hunger, fear and rage (2nd ed.). New York: Appleton-Century-Crofts.

Cook, E. W., Melamed, B. G., Cuthbert, B. N., McNeil, D. W., & Lang, P. J. (1988). Emotional imagery and the differential diagnosis of anx- iety. Journal of Consulting and Clinical Psychology, 56, 734-740.

Crease, R. P. (1993). Biomedicine in the age of imaging. Science, 261, 554-561.

Darwin, C. (1872). The expression of the emotions in man and animals. Chicago: University of Chicago Press.

Davis, M. (1989). Neural systems involved in fear-potentiated startle. Annals of the New York Academy of'Sciences, 563, 165-183.

Dennett, D. (1991). Consciousness explained, Boston: Little, Brown. DiCara, L., Braun, J. J., & Pappas, B. (1970). Classical conditioning

and instrumental learning of cardiac and gastrointestinal responses following removal of neocortex in the rat. Journal of Comparative andPhysiologial Psychology, 73, 208-216.

Duffy, E. (1934). Emotion: An example of the need for reoriention in psychology. Psychological Review, 41, 184-198.

Duffy, E. (1941). An explanation of "emotional" phenomena without the use of the concept "emotion." Journal of General Psychology, 25, 283-293.

Dunlap, K. (1920). Mysticism, Freudianism and scientific psychology. New York: Appleton-Century-Crofts.

Ekman, P. (1984). Expression and the nature of emotion. In K. S. Scherer & P. Ekman (Eds.), Approaches to emotion (pp. 319-343). Hillsdale, NJ: Erlbaum.

Feldman, J., & Ballard, D. (1982). Connectionist models and their prop- erties. Cognitive Science, 6, 205-254.

Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 117-140.

Frijda, N. H. (1986). The emotions. Cambridge, England: Cambridge University Press.

Gardner, H. (1985). The mind's new science. New \fork: Basic Books. Gibson, J. J. (1966). The senses considered as perceptual systems. Bos-

ton: Houghton Mifflin. Goldstine, H. H. (1972). The computer from Pascal to von Neumann.

Princeton, NJ: Princeton University Press. Graham, D. T., Stern, J. A., & Winokur, G. (1958). Experimental in-

vestigation of the specificity of attitude hypothesis in psychosomatic disease. Psychosomatic Medicine, 20, 446-457.

Grossberg, S. (1991). Neural network models of conditioning and action. Hillsdale, NJ: Erlbaum.

Gunderson, K. (1971). Mentality and machines. New York: Doubleday. Hebb, D. (1949). The organization of behavior. New York: Wiley. Heider, F. (1944). Social perception and phenomenal causality. Psycho-

logical Review, 51, 358-374. Hohmann, G. W. (1966). Some effects of spinal cord lesions on experi-

enced emotional feelings. Psychophysiology, 3, 143-156. Izard, C. E. (1977). Human emotions. New 'York: Plenum. James, W.( 1884). What is emotion? Mind, 19, 188-205. James, W. (1890). The principles of'psychology (Vols. 1 &2). New "fork:

Holt. James, W. (1894). The physical basis of emotion. Psychological Review,

7,516-529. Jorgensen, C. (1973). A theory of the elements in the emotions. In

M. L. Reymert (Ed.), Feelings and emotions: The Wittenberg Sympo- sium (pp. 310-315). New "fork: Arno Press. (Original work published 1928)

Katkin, E. S. (1985). Blood, sweat, and tears: Individual differences in autonomic self-perception. Psychophysiology, 22, 125-137.

Koffka, K. (1935). Principles of gestalt psychology. New \brk: Harcourt, Brace.

Konorski, J. (1967). Integrative activity of the brain: An interdisciplin- ary approach. Chicago: University of Chicago Press.

Lacey, J. I. (1959). Psychophysiological approaches to the evaluation of psychotherapeutic process and outcome. In E. A. Rubinstein & M. B. fsaloff (Eds.), Research in psychotherapy (Vol I, pp. 160-208). Washington, DC: American Psychological Association.

Lacey, J. I. (1967). Somatic response patterning and stress: Some revi- sions of activation theory. In M. H. Appley & R. Turnbull (Eds.), Psychological stress: Issues in research. New "York: Appleton.

Laird, J. D. (1984). The real role of facial response in the experience of emotion: A reply to Tourangeau and Ellsworth, and others. Journal of Personality and Social Psychology, 47, 909-917.

La Mettrie (1953). Man, a machine. Chicago: G. Bussey. (Original work published 1748)

Lang, P. J. (1964). Experimental studies of desensitization psychother- apy. In J. Wolpe (Ed.), The conditioning therapies (pp. 38-53). New York: Holt, Rinehart & Winston.

Lang, P. J. (1968). Fear reduction and fear behavior: Problems in treat- ing a construct. In J. M. Shlien (Ed.), Research in psychotherapy (Vol. 3, pp. 90-103). Washington, DC: American Psychological Associa- tion. Reprinted in I. Florin & W. Tunner (Eds.), Therapie der Angst: Systematische Desensibilisierung, Fortschritte der Klinischen Psy- chologie(l975, Vol. 8). Urban & Schwartzenberg: Munchen-Berlin- Wien.

Lang, P. J. (1978). Anxiety: Toward a psychophysiological definition. In H. S. Akiskal & W. L. Webb (Eds.), Psychiatric diagnosis: Explora- tion of biological predictors (pp. 365-389). New "York: Spectrum.

Lang, P. J. (1979). A bio-informational theory of emotional imagery. Psychophysiology, 16, 495-512.

Lang, P. J. (1985). The cognitive psychophysiology of emotion: Fear and anxiety. In A. H. Tuma & J. D. Maser (Eds.), Anxiety and the anxiety disorders (pp. 131-170). Hillsdale, NJ: Erlbaum. Reprinted 1985 in Psychotherapeutisch Pasport, 3(2), 3-62.

Lang, P. J. (1993). The motivational organization of emotion: Affect- reflex connections. In S. Van Goozen, N. E. Van de Poll, & J. A. Sergeant (Eds.), The emotions: Essays on emotion theory (pp. 61-93). Hillsdale, NJ: Erlbaum.

Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1990). Emotion, atten- tion, and the startle reflex. Psychological Review, 97, 377-395.

Lang, P. J., Greenwald, M. K., Bradley, M. M., & Hamm, A. O. (1993). Looking at pictures: Affective, facial, visceral, and behavioral reac- tions. Psychophysiology, 30, 261-273.

Lange, C. (1922). The emotions (I. A. Haupt, Trans.). Baltimore: Wil- liams & Wilkins. (Original work published 1885)

Lazarus, R. S. (1982). Thoughts on the relations between emotion and cognition. American Psychologist, 37, 1019-1024.

Lazarus, R. S., Averill, J. R., & Opton, E. M. (1970). Towards a cognitive theory of emotion. In M. B. Arnold (Ed.), Feelings and emotions: The Loyola Symposium (pp. 207-232). San Diego, CA: Academic Press.

LeDoux, J. E. (1990). Information flow from sensation to emotion plas- ticity in the neural computation of stimulus value. In M. Gabriel & J. Moore (Eds.), Learning and computational neuroscience: Founda- tions of adaptive networks (pp. 3-52). Cambridge, MA: Bradford Books/MIT Press.

Levenson, R. W., Ekman, P., & Friesen, W. V. (1990). Voluntary facial action generates emotion-specific autonomic nervous system activity. Psychophysiology, 27, 363-384.

Lewin, K. (1935). A dynamic theory of personality. New York: McGraw- Hill.

SPECIAL ISSUE: JAMES-LANGE THEORY 221

Lewin, K. (1936). Principles of topological psychology. New York: McGraw-Hill.

Lewin, K. (1940). Formalization and progress in psychology. University of Iowa Studies in Child Welfare, 16(3), 9-42.

Lindsley, D. B. (1951). Emotion. In S. S. Stevens (Ed,), Handbook of experimental psychology (pp. 473-516). New \brk: Wiley.

Lorenz, K. (1952). King Solomon's ring: New light on animal ways. New York: Crowell.

Mackintosh, N. J. (1983). Conditioning and associative learning. New York: Oxford University Press.

Malmo, R. B., Shagass, €., & Davis, F. H. (1950). Symptom specificity and bodily reactions during psychiatric interview. Psychosomatic Medicine, 12, 362-376.

Mandler, G. (1975). Mind and emotion. New \ork: Wiley. Mandler, G. (1984). Mind and body: Psychology of emotion and stress.

New York: Norton. Mandler, G., & Kremen, I. (1958). Autonomic feedback: A correlational

study. Journal of Personality, 26, 388-399. Mandler, G., Mandler, J. M., Kremen, I., & Sholiton, R. D. (1961). The

response threat: Relations among verbal and physiological indices. Psychological Monographs, Whole No. 75.

Mandler, G., Mandler, J. M., & Uviller, E. T. (1958). Autonomic feed- back: The perception of autonomic activity. Journal of Abnormal and Social Psychology, 56, 367-373.

Maranon, G. (1924). Contribution a Fetude de 1'action emotive de 1'adrenaline [Contribution to the study of the emotive action of ad- renalin]. Revue Francaise d'Endocrinologie, 2, 301-325.

Marks, I. M., & Huson, J. (1973). Physiological aspects of neutral and phobic imagery. British Journal of Psychiatry, 122, 567-572.

McClelland, J., & Rumelhart, D. (Eds.). (1986). Parallel distributed processing: Explorations in the microstructure of cognition: Psycho- logical and biological models (Vol. 2). Cambridge, MA: MIT Press.

McDougal, W. (1910). An introduction to social psychology. London: Methuen.

McLean, P. D. (1949). Psychosomatic disease and the "visceral brain." Recent developments bearing on the Papez theory of emotion. Psy- chosomatic Medicine, 11, 29-44.

Mehrabian, A. (1970). A semantic space for nonverbal behavior. Jour- nal of Consulting and Clinical Psychology, 35, 248-257.

Mehrabian, A., & Russell, J. A. (1974). An approach to environmental psychology. Cambridge, MA: MIT Press.

Miller, G. A., Levin, D. N., Kozak, M. J., Cook, E. W., McLean, A., & Lang, P. J. (1987). Individual differences in imagery and the psycho- physiology of emotion. Cognition and Emotion, 1, 367-390.

Moruzzi, G., & Magoun, H. W. (1949). Brain stem reticular formation and activation of the EEG. EEG Clinical Neurophysiology, 1, 455- 473.

Murphy, G., & Ballou, R. O. (1969). William James on psychical re- search. New York: Viking Press.

Newell, A., & Simon, H. (1963). Computers in psychology. In R. Luce, R. Bush, & E. Galanter (Eds.), Handbook of mathematical psychology (pp. 361-428). New York: Wiley.

Olds, J., & Olds, M. E. (1961). Interference and learning in paleo-corti- cal systems. In J. F. Delafresnaye (Ed.), Brain mechanisms and learn- ing^. 153-187). Oxford, England: Blackwell.

Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. (1957). Themeasure- ment of meaning. Urbana: University of Illinois Press.

Panksepp, J., Sacks, D. S., Crepeau, L. J., & Abbott, B. B. (1991). The psycho- and neurobiology of fear systems in the brain. In M. R.

Denny (Ed.), Fear, avoidance, and phobias: A fundamental analysis (pp. 7-59). Hillsdale, NJ: Erlbaum.

Papez, J. W. (1937). A proposed mechanism of emotion. Archives of Neurology and Psychiatry, 38, 725-743.

Plutchik, R. (1962). The emotions: Facts, theories and a new model. New \brk: Random House.

Rachman, S., & Hodgson, R. J. (1974). Synchrony and desynchrony in fear and avoidance. Behavior Research and Therapy, 12, 311 -318.

Reyniert, M. L. (1928). Feelings and emotions: The Wittenberg Sympo- sium. Worcester, MA: Clark University Press.

Roberts, R. J., & Weerts, T. C. (1982). Cardiovascular responding dur- ing anger and fear imagery. Psychological Reports, 50, 219-230.

Rosenzweig, S. (1992). Freud, Jung, and Hall the king-maker: The his- toric expedition to America. St. Louis, MO: Rana House Press.

Russell, J. A. (1980). A circumplex model of affect. Journal of Person- ality and Social Psychology, 39, 1161-1178.

Schachter, J. (1957). Pain, fear, and anger in hypertensives and normo- tensives: A psychophysiological study. Psychosomatic Medicine, 19, 17-29.

Schachter, S. (1959). The psychology of affiliation: Stanford, CA: Stan- ford University Press.

Schachter, S., & Singer, J. E. (1962). Cognitive, social, and physiological determinants of emotional state. Psychological Review, 69, 379-399.

Schandry, R., Bestler, M., & Montoya, P. (1993). On the relation be- tween cardiodynamics and heartbeat perception. Psychophysiology, 30, 467-474.

Seligman, M. E. P. (1971). Phobias and preparedness. Behavior Ther- apy, 2, 307-320.

Skinner, B. F. (1954). Critique of psychoanalytic concepts and theories. Scientific Monthly, November, 300-305.

Spinoza, B. (1936). Philosophy of Benedict de Spinoza (R. H. M. Elwes, Trans.). New \brk: Todor Publishing. (Original work published Circa 1632-1677)

Sutherland, A. (1898). The origin and growth of the moral instinct (Vols. 1 & 2). New \brk: Longman, Green.

Thompson, R. F. (1986). The neurobiology of learning and memory. Science, 233,941-947.

Titchener, E. B. (1908). Lectures on elementary psychology of feeling and attention. New \brk: Macmillan.

Tomkins, S. S. (1962). Affect, imagery, and consciousness: The positive affects (Vol. 1). New York: Springer.

Tomkins, S. S. (1963). Affect, imagery, and consciousness: The negative affects (Vol. 2). New York: Springer.

Vanden Burgt, R. J. (1981). The religious philosophy of William James. Chicago: Nelson-Hall.

Velten, E. (1968). A laboratory task for induction of mood states. Be- haviour Research and Therapy, 6,473-482.

Vrana, S. R., & Lang, P. J. (1990). Fear imagery and the startle-probe reflex. Journal of Abnormal Psychology, 99, 189-197.

Wenger, M. A. ,(1950). Emotion as visceral action: An extension of Lange's theory. In M. L. Reymert (Ed.), Feelings and emotions: The Mooseheart Symposium (pp. 3-10). New York: Wiley.

Wundt, W. (1896). Grundriss der Psychologie [Outlines of psychology]. Leipzig, Germany: Entgelmanri.

Zajonc, R. B. (1980). Feeling and thinking: Preferences need no infer- ences. American Psychologist, 35, 151-175.

Received September 1,1993 Accepted September 1, 1993 "