THEORETICAL CONSIDERATIONS ON EMOTION: SPINOZA, DESCARTES, DARWIN - WHO WIN THE RACE?

Ilan Shalif (1) Isaac Lewin(2)
ABSTRACT

A study was carried out to reconcile between the "Equisitionist" and the "Nativist" approaches in the emotional domain. It was hypothesized that in the verbal domain, the three general and abstract cognitive dimensions of the "Equisitionists" would prevail, while in the nonverbal domain, the main components are the seven specific "basic" emotions of the "Nativists". Resident students (202) assessed the extent to which 104 photographs of facial expressions, and 148 words signifying emotional states were close to their own momentary emotional state. Data were analyzed by a Nonparametric Multidimensional Profile statistic. It was found that the assessments of both verbal and nonverbal items were made along bipolar dimensions that are in accord with a Nativist genetic approach. Results support Darwin's claim that each of the basic emotions is a bipolar entity. Two assumed basic emotions - Happiness and Sorrow were not found to be independent of each other, but are opposing poles of the same emotional continuum.

_____________________________
(1) This article is based on a study (1987) that was carried out by the first author, under the supervision of Professor Isaac Lewin, towards a Ph.D. dissertation entitled: "The Emotions and the Dimensions of discrimination Among Them in Daily life**". On 1.1.1991, the final draft of the thesis was approved by the Senate of Bar-Ilan University.

** A somewhat abbreviated and non-professional translation of the first draft of the thesis (1978) available for down-load (230kb as text at: http://www.etext.org/Psychology/Shalif/emotion1
or as multi-page HTML tagged at: http://members.tripod.com/~alternativ_psy/diserta1.htm , or http://www.shalif.com/psychology/diserta1.htm
or
will be sent via E-Mail to the one who have failed the download and will ask for it by E-mail.

(2) Address all correspondence to Dr. Ilan Shalif - E-mail - mailto:ilan@shalif.com
-----------------------------

There has been a long and heated controversy between ancient, as well as more recent, philosophers about the basic components of the universe, of the human being and of human emotions. This controversy is also prevalent between psychologists.

Some contemporary scientists follow the ideas of Spinoza (1677/1955)and later De Rivera (1977). They espouse the Equisitionist approach to human nature, and emphasize the importance of learning over heredity. They usually put forward parsimonious cognitive models for the various domains of human life - including the emotional domain.

-2-

Spinoza and De Rivera claimed that the following concepts: a) Persistence (or striving) b) Attainment and c) Necessity, are related to the main aspects of emotional life. From these concepts they derived the emotional dimensions of 1) Desire 2) Pleasure and 3) Pain. Even though none of the modern scientists of this trend adhere exactly to the above variables, they still claim that there are some three components with abstract content that are the main essence of emotion.

The bulk of modern studies that assume dimensionality of emotional phenomena were carried out using the Semantic Differential Technique of Osgood (1969) and show nearly the same trinity:

I. Evaluation (or Pleasantness - Unpleasantness), which parallels Spinoza's Attainment (b) and Pleasure (2);

II. Activation (or Attentional Activity), which parallels Spinoza's Persistence (Striving) (a) and Desire (1);

III. Potency (or Intensity, or Control), which to some extent parallel Spinoza's Necessity (c) and Pain (3).

Osgood (1969), and Ekman & Friesen (1982) claim that the above three are the only clear dimensions that have been identified across most studies, and that additional dimensions are still to be found. Russell, who is one of the most parsimonious cognitivists, vacillated between this three dimensional model (Russell & Mehrabian, 1977), and a two dimensional solution of 1) Evaluation and 2) dynamism (Russell, 1980; Russell & Bullock, 1986).

On the other hand, there are those who follow the ideas of Descartes (1647/1967). These scientists espouse the Nativist approach, and emphasize the importance and the richness of the hereditary biological and psychological equipment of the individual. They usually support one version or another of the genetic approach to emotion, originated by Darwin.

Descartes (1649/1967) identified six "simple passions" (of the mind) and listed them as: 1) Wonder; 2) Love; 3) Hatred; 4) Desire; 5) Joy; 6) Sadness. He claimed that "all the others are composed of some of these six or are species of them" (p362).

Modern scientists of this trend claim that the subjective emotional experience and its communication (i.e., both perception and expression) are mainly results of appraisals done by inborn mechanisms of the brain. They usually call these mechanisms "basic emotions".

Some of them, like Plutchik (1980, 1982), Izard (1971, 1977, 1984), Frijda (1986), Ekman (1982) and Ekman & Friesen (1982), put forward lists of seven to ten basic emotions - with a good congruence between these lists and between their lists and that of Descartes. Ekman & Friesen (1982) summarized their comprehensive comparative study with a list of seven basic emotions commonly judged from still photographs of posed facial behavior, namely: 1) surprise 2) interest 3) anger 4) disgust/contempt 5) happiness 6) sadness 7) fear. They added that this list is only of the clearly depicted basic emotions but not an exhaustive one.

The modern scientists following Descartes' ideas are explicitly or implicitly also following Darwin's (1872) first scientific theory of emotion. In his theory, Darwin claimed that there are specific inborn emotions - each of them includes a specific pattern of activation of the facial expression and behavior.

-3-

According to the first principle of Darwin's theory, these inborn basic emotions originated in serviceable associated habits - with relations between them like those between vomiting and the expression of disgust. In the second principle of his theory, he claimed that each of the basic emotions consists of a pair of bipolar antitheses - like the two opposing poles of fear and serenity. According to his third principle, the emotional phenomena can result from direct actions of the nervous system - like jumping up and down without any aim, when excessively excited.

Most of Darwin's followers took from his three principle theory of emotion only the first one, which says that the origin of the inborn pattern of each basic emotion is an hereditary serviceable associated habit. Like Darwin, they regard the inborn facial expression of emotion to be the main aspect of emotion. Like him they regard the subjective experience of emotion as the main aspect of emotion observable by oneself. Like Darwin, they regarded cross-cultural studies as the main test for their specific approaches. But, none of them studied the bipolarity inherent - according to Darwin - in each of the basic emotions.

Based on Darwin, these scientists searched for cross-cultural inborn facial expressions of emotion, both to establish the truth of the genetic approach to the theory of emotion as a whole and their own version of it - as well as to decide about specific candidates for their collection of "basic emotions", which is the modern term for Descartes' "basic passion of the mind".

Paradoxically, to date there are studies that support both approaches, thus, in a way, refuting both. Results of thousands of studies, which were mostly carried out in the verbal domain and used various dimensional analyzes, support parsimonious cognitive approaches. These studies seem to confirm the claim that two or three abstract dimensions are the basic components of all specific instances of emotion.

However, results of many studies carried out by scientists of the genetic approach, show that, the basic emotions of their lists have the same specific facial expressions all over the world, and that these basic emotions construct complex emotional experiences. To make things more complicated, results of many studies - in the verbal and the nonverbal domains - seem to support, in a way, both approaches.

In many of the more complex studies - like that of Osgood, Suci & Tennenbaum (1957) - the contents of few of the variables and dimensions discerned in them converge with the above three abstract ones and the contents of others converge with that of the basic emotions of the geneticist approach.

Izard (1971) and Frijda (1970) who reviewed results of studies that contradicted the genetic approach or were not unequivocal, claimed that these results are irrelevant. They argued that multidimensional analyzes might be inappropriate for the study of the basic variables of emotion.

On the other hand, Osgood, Russell and others of the parsimonious cognitive trend claim that the concrete emotional phenomena and the specific basic emotions found are the creation of the abstract cognitive emotional variables together with non-emotional variables (and are not primary emotional variables). They claim that the entire emotionality of each phenomenon can be relegated to the two or three abstract emotional continua of the brain proposed by them.

-4-

Scientists like Leventhal (1979, 1982), Plutchic (1980), Ekman & Friesen (1982), and Smith & Ellsworth (1985) try to reconcile the findings and the theories of the two approaches. The proposed solutions are essentially of two kinds:

a) There are those who offer one kind or another of dualistic approaches to emotion. Like Leventhal and Ekman, they claim that both variables are at work, but they do not integrate them into a united multidimensional model, nor do they try to reconcile the contradicting results.

b) There are those - like Plutchik or Smith, who offer an integrated multidimensional model, supported by empirical findings. Usually, the space of their models has dimensions with a somewhat abstract content. But, in the space of these models, the concrete basic emotions can be located as points or regions.

The theory presented in this study is an attempt to reconcile the apparent contradiction between the two contending schools in a different way. It may be regarded as a refinement of the first approach. We capitalized on the fact that most of the studies supporting the parsimonious cognitive approach were leaning heavily on the verbal domain, and those of the genetic approach were leaning mainly on facial expressions. In that theory it was claimed that the emotional phenomenon is bi-modal - one mode is a verbal and abstract, and one mode is nonverbal and concrete.

THE MAIN HYPOTHESES TO BE EMPIRICALLY TESTED WERE

a) In evaluating the similarity between their own emotional feelings at a certain moment and those depicted in facial expressions, subjects use mainly the basic concrete emotions.

b) In evaluating the similarity between their own emotional feelings at that moment and those which are conveyed in words, subjects use mainly the abstract dimensions of emotion.

***

To assess the appropriateness (feasibility) of our theory or refute it, we designed a study based on two autonomous parts - one nonverbal and one verbal - both fitting the multidimensional unfolding model - like that of Hirschberg (1980), Takane, Young & DeLeeuw (1977) or Shalif, Lerner & Dasberg (1981).

-5-

THE METHODS

SUBJECTS:

202 subjects participated in the study - 101 male and 101 female - resident students from two different universities, diverse backgrounds and of various faculties (psychology not included).

MATERIALS:

a) 105 photographed facial expressions
The photographs included: a\I - the 48 items of the test of Szondi (1947), which are spontaneous facial expressions of mixed emotions (i.e., without external directives or manipulations); a\II - 33 artificial expressions comprising of photographs of two samples of seven basic emotions taken from Ekman & Friesen (1975) and two samples of nine basic emotions taken from Izard (1971, 1977) - 2 to 4 items for each of the basic emotions suggested by these authors(3); a\III - 24 artificial expressions of combinations of basic emotions - also from Ekman & Friesen's book(3).

b) 148 words.
The words were mainly drawn from studies of the emotional domain published by the authors cited in this paper (translated to Hebrew). The list consisted of: b\I - a wide variety of 96 words of emotion(4); b\II - 52 words which were included for control of various response sets. Part of them are words of emotion and another part are names of dimensions of emotion previously found.

PROCEDURE:

Each subject was asked to evaluate to what extent her/his emotional state (i.e., feelings) of the moment fitted the emotion expressed in each item, on a six grade scale(5). In this way, the subjects assessed each of the 105 photographed facial expressions, then each of the 148 words, and then reassessed the photographs.

The first assessment of the photographs (and the only one for the words) was made by means of structured scaling. Each item was rated separately, on a fixed scale of six grades of congruence: one being "differ very much"; two - "different"; three - "slightly more different than similar"; four - "slightly more similar than different"; five - "similar"; six - "very similar" (to be called hence: S-TASK). The photographs were numbered and were presented one after the other in the same order for all subjects. The 48 Szondi Test items were presented first, and according to their original order.

Then, were presented the 33 items of material a\II and the 24 items of material a\III - according to their order in appendix 1. Subjects scored the photographs on a page containing numerical-graphical scales of six steps for each item. The full six verbal-numerical scale was printed both at the top and the bottom of that page.
____________________
(3) In appendix 1 appear the exact references for these photographs.
(4) In appendix 2 appear the English translation of the 148 Hebrew words used in this study.
(5) To evade the problematic and biasing use of collections of rating scales of previous studies, the "one scale of common range" approach was chosen for both studies. This is possible when using Guttman's (1957, 1968) "mapping sentence" approach to the combined theory building and design of experiment.
In this paradigm the subject ratings can be done along a common range for all items - if the content of a domain can be summed up clearly in one phrase. ------------------------

-6-

Then the subject assessed the 148 words which were printed alphabetically on one page - with a six grade graphic-numerical scale near each. Here too, the full six verbal-numerical scale was printed both at the top and the bottom of the page.

Afterwards, the subject assessed the 105 photographs again. This time by means of forced choice, to control for bias and ceiling effects (to be called hence - Q-SORT). In that task, the photographs were divided into groups of 7 to 12 items - which included one sample of each emotion wherever possible.

The judgments were expressed here by arranging the items within each group in six grades of the same scale of the S-TASK, with the added instruction of "relatively to the other items of the group". For groups of seven and nine items a seventh grade of 3.5 was added; for groups of eight nine and ten items there were two places for the intermediate grades; for the groups of 12, each grade had two places.

RESULTS

The finding of the dimensions of the studied domain(s) is based here on the matrix(es) of correlation between the items. The interpretation of the content of those dimensions can be objectively interpreted in two ways: a) according to the location of the items on the non-parametric multidimensional analysis projection maps - wherever the items are clearly interpretable. b) using a multidimensional unfolding model like that of Hirschberg (1980), Takane et al. (1977), or Shalif et al. (1981). In this kind of analysis the interpretation of the content of a dimension can be made according to the correlations between subjects' dimension score (of that dimension) and their scores for items inside and outside the multidimensional analyzes. To test the hypotheses of the study, the following steps were taken: Step 1) We analyzed the dimensionality of the scores of the 48 mixed emotional photographs (material a\I) of the S-TASK (which had a reasonable spread). It was done by the S.S.A.-I of the Guttman Lingoes series of non-parametric multidimensional analysis technique (Lingoes, 1973) - all solutions from two dimensions to ten (the maximum possible) were examined.

Step 2) For each subject, the dimensional scores were computed for each dimension of the 10 dimensional analysis of the S.S.A.-I. The subject's dimension score for each of the 10 dimensions was the sum of 48 products - those of his raw score for each of the 48 items multiplied by the coordinate of that item on that dimension(6). --------------------------
(6) The mathematical expression is: "the multiplication of a row of 48 numbers by a matrix of ten columns by 48 rows".
--------------------------

Step 3) For each subject, two group scores were computed for each of the nine groups of photographs of basic emotions (material a\II) - one for the S-TASK and one for the Q-SORT.

Step 4) The dimensionality of the 96 emotional words (material b\II) was analyzed using the same procedure as in step 1.

-7-

Step 5) For each subject, the dimensional scores were computed for the ten dimensional analyzes of the 96 words as done in Step 2.

Step 6) a comprehensive matrix of Spear-man non-metric correlations among the variables was computed. It included the dimensional scores of step two and five (10+10), the group scores computed in step three (9+9), the raw scores of the words (148) and those of the photographs of the two procedures (105+105). This was done within each group of scores, as well as with all the scores of the other groups - resulting in a matrix of 396 by 396.

FOR THE FIRST HYPOTHESIS

The correlations between each of the 9 groups of photographs of basic emotions in the two tasks are brought in the following table 1. In it, only the sixth dimension was not correlated significantly with any of the group scores of basic emotions.

Table No. 1: The correlations between the dimensions' scores of the 48 Szondi
items (step 2) and the groups' scores of basic emotions (step 3) - S for the
S-TASK and Q for the Q-SORT, with level of significance.

 Emotion  |___________________________Dimension________________________________
----------|  1      2      3      4      5      6      7      8      9      10
Happiness |0.63!| 0.13 | -.05 | 0.10 | 0.07 | -.07 | 0.16+| 0.03 | 0.12 | -.03
         Q|0.56!| -.05 | 0.04 | 0.10 | -.03 | -.07 | 0.16+| 0.02 | 0.17*| 0.01
Surprise S|0.25!| 0.09 | -.11 | 0.06 | 0.17*| -.03 | 0.06 | -.16+| -.02 | 0.02
         Q|0.31!| -.02 | 0.05 | 0.05 | 0.01 | -.08 | 0.07 | -.10 | 0.11 | 0.05
Interest S|0.16*| 0.24!| -.07 | 0.01 | 0.14+| -.02 | 0.02 | 0.07 | 0.07 | 0.04
         Q|0.07 | -.02 | 0.08 | -.01 | -.05 | -.01 | 0.01 | -.04 | 0.04 | 0.20*
Fear     S|-.22!| -.05 | -.30!| 0.03 | 0.11 | -.00 | 0.10 | -.08 | -.16+| -.05
         Q|-.15+| -.11 | -.19!| -.15+| -.04 | 0.13 | 0.04 | -.10 | -.07 | -.14+
Anger    S|-.27!| -.01 | -.25!| 0.13 | 0.11 | -.03 | 0.01 | -.11 | -.16*| -.04
         Q|-.23!| 0.03 | -.11 | 0.22!| 0.07 | -.09 | 0.01 | -.00 | -.01 | 0.02
Disgust  S|-.14+| -.00 | -.29!| 0.04 | 0.12 | 0.06 | 0.05 | -.07 | -.15+| -.02
         Q|0.03 | -.11 | -.08 | 0.04 | 0.05 | -.02 | -.03 | 0.03 | -.09 | 0.10
Sadness  S|-.35!| -.02 | -.24!| -.01 | 0.13 | 0.11 | -.01 | -.06 | -.13 | -.07
         Q|-.42!| -.04 | -.04 | -.06 | 0.01 | 0.06 | -.08 | 0.07 | -.13 | 0.02
Contempt S|-.16+| 0.22!| -.12 | -.00 | 0.07 | -.03 | -.00 | -.06 | -.06 | -.10
         Q|-.06 | 0.17*| 0.11 | -.16*| -.13 | 0.01 | -.03 | 0.04 | 0.05 | -.08
Shame    S|-.27!| 0.25!| 0.02 | -.12 | 0.04 | 0.03 | -.15+| 0.04 | -.05 | 0.01
         Q|-.33!| 0.11 | 0.14 | -.04 | -.10 | 0.08 | -.17*| 0.11 | -.05 | -.02
------------------------------------------------------------------------------
(7) Plus (+) means .02.05 (one tail).

-8-

For each task, the difference between the ten correlations' profile(8) of each of the 9 emotions and the profiles of each of the other 8 emotions was computed. _______________
(8) The profiles are the S rows and the Q rows shown in Table No.1.
---------------

The significance of the differences was estimated by the Hoteling test for the differences between correlated correlations (as suggested by Guilford, 1965). It shows that among the 72 differences (36 for each task) only 5 failed to reach significance - each for a different pair of emotions. Thus leaving five of the 36 pairs of emotions with a significant difference on one task only.

Thus - relying on the nonverbal data alone - results supported the first hypothesis of the study, which predicted that "the assessment of subjects of the extent to which facial expressions of emotion are like their own, would be done mainly along dimensions that have the same content of basic emotions".

However, due to ceiling effects for part of the unpleasant basic emotions in the S-TASK and for the more pleasant ones in the Q-SORT procedure; and due to insufficient content validity of the artificial facial expressions of basic emotions - the results are not as clear cut as we could have wished.

To deepen the understanding of the results and to clarify the picture, a painstaking item-by-item check of scores distribution was done. With the same aim, a thorough check of the 396 by 396 matrix of correlations was carried out. These steps enabled the unequivocal interpretation of all 10 dimensions and resulted in a clearer picture. The contents of the 10 dimensions - as out-putted by the computer program and not rotated by hand afterwards are as follows: I. Happiness versus Sadness; II. Interest in others + leniency versus drowsiness; III. Fear; IV. Anger (or firmness versus embarrassment); V. Surprise (stupefaction); VI. Troubled - guilt-worry-frustration, versus relief; VII. Shame; VIII. Seems to have the content of contempt; IX. Love versus hate; X. Disgust.

FOR THE SECOND HYPOTHESIS

Following are the main results relating to the correlations among the 10 dimensions of discrimination between the 96 words of emotion, and those between the dimensions and all the other 386 variables of the study. The 10 dimension inter-correlation matrix shows that only one of the 45 correlations - that between the first and the second dimensions - is significant: r=-.20, p<.004 (two-tail). This may be the result of uneven sampling of the emotional domain due to our unintentional bias, due to the bias of the language itself, due to uneven distribution of the subjects emotional experience, or due to the relativity (or incompleteness) of the independence between the various basic emotion.

The interpretation of the content of each dimension is made according to the convergence of information that is taken from two main sources. One is the content of the words that are on the edges - the two margins of each dimension of the S.S.A.-I analysis, which is the base of the classical way of interpretation. The other is the detailed inspection of the matrix of correlations of the 10 dimensions with the other 386 scores - the unique contribution of the paradigm of the unfolding model.

-9-

The first dimension: contentment versus distress.
The words that are on the edges (near the margins of this dimension) are: calmness, at ease, and contentment on one side, versus misery and distress on the other side. The words that have the highest correlations with it are: depression -.67; distress -.69; sadness -.67; at-ease +.69; calmness +.74; contentment +.69; satisfied +.72; and happiness +.67 - p<.001 for all of them. The dimension's highest correlations with subgroups of facial expressions of basic emotions are with happiness +.57 and -.40 for sadness.

The second dimension: alertness (vigilance) versus serenity.
On its edges are alertness, tension and curiosity versus serenity, disrespect and relief. Its highest correlations are: alertness +.47; complexity +.46 disappointment +.42; stubbornness +.40; serenity -.38; disregard -.27; happiness -.33; relief -.26; satisfied -.30; and tranquility -.29 - all are of p<.001. The dimension's significant correlations are with the subgroups of happiness and shame.

The third dimension: pride versus embarrassment.
On its edges are pride, firmness, and haughtiness versus fatigue, embarrassment - pride +.46; superiority +.42; courage +.43; firmness +.53; embarrassment -.31; fatigue -.34; anxiety -.27; shame -.25; and weakness -.27; all are of p<.001. The dimension's correlations with subgroups of basic emotions are with sadness -.20, p<.02; shame -.14, p<.05 and anger +.19, p<.02.

The fourth dimension: indifference versus surprise and excitement.
On its edges are surprise, adoration, curiosity, shame and panic versus Indifference, boredom and routine. Indifference +.42; sleepy +.16; boredom +.24; routine +.16; surprise -.59; excitement -.46; gratitude -.48; love -.43; panic -.44; shame -.43; Its highest correlations with subgroups of photographs of emotions are with surprise and fear both of -.22 p<.002.

The fifth dimension: the dimension of pity.
The words on its edges are pity, compassion, leniency versus sympathy, curiosity, surprise and helpless. The highest correlations are with pity -.67; compassion -.53; leniency -.56; tenderness -.41; all are of p<.001. Positively correlated with it are only confusion +.15 p<.038 and love +.17 p<.018. The dimension is also correlated with the interest subgroup: -.18, p<.02 and with mixtures of anger + disgust and anger + contempt - both of +.21, p<.004.

The sixth dimension: suspiciousness (and astonishment) versus longing.
On its edges are: scepticism, astonishment, suspiciousness, curiosity, caution versus longing, compassion, leniency. The highest correlations are with suspiciousness +.34; astonishment +.35; scepticism +.24 versus longing -.48; compassion -.29; courage -.28 and yearning -.27.

-10-

The seventh dimension: responsibility and involvement versus laxity.
On its edges are: routine versus indifference, longing and scepticism. Its highest correlations are with routine +.40; worry +.25 animosity -.27; attraction -.30; delight -.30; mirth -.24; indifference -.29; longing -.31. It is also significantly correlated with the subgroup of surprise: -.15, p<.05.

The eighth dimension: patience and loneliness versus longing and love.
The marginal words on its edges are loneliness, patience and involvement on the one side, versus hope, longing, excitement, desire and delighted which are on the other side. Its highest correlations are with patience +.31 and leniency +.33 (both p<.001). Negatively correlated with it are happiness, craving, excitement, weakness, desire, alarm, slumber, fatigue, longing, and hope - all with p<.002 (the lovers' cluster).

The ninth dimension: guilt and depression versus contempt and disregard.
On its edges are: caution, fatigue and disregard versus loneliness, contempt and scepticism. Correlations of the significance level of p<.002 are with caution, disregard, embitterment and fatigue versus contempt, loneliness and scepticism. These of .002>p<.02 are with guilt, despair, haughtiness, posing, pain, anger, confidence, blur, restlessness, servility, sleepy and slumber versus boredom and involvement. A possible common denominator for the words that are in the direction of caution and disregard is the depression syndrome - guilt, despair and pain, and their results - restlessness and fatigue.

The tenth dimension: hostility versus conciliation.
The marginal words on its edges are: hate, pain and fatigue versus conciliated and serenity. Correlations with the conciliation direction that reach the significance level of p<.002 are of alarm, alertness, adoration, conciliated, contentment, desire, disappointment, longing, repulsion, right(ness), serenity. Correlations with the direction of the hostility having the significance of .002

None of the correlations of the above ten dimensions with control words - the content of which was the name of cognitive dimensions - was higher than those with relevant words of emotion. The correlations of those control words with the above ten dimensions were even lower than the correlations of these dimensions with the relevant photographs of facial expressions of the basic emotions.

And so, results contradict our second hypothesis and consequently refute our reconciliation theory.

The following matrix of correlation in Table no 2. sums up the results of the two sub-studies - the nonverbal and the verbal one. It clearly demonstrates their being of the same domain. Part of the dimensions of the original computer output were rotated for ease of representation. The matrix of 100 correlations between the two sets of 10 dimensions contains in it 10 with p<.01 and among them 6 are of p<.001. There is a clear convergence of 6 of the 10 dimensions of facial expression and 8 of these of the words of emotion. These results deepen the refutation of the second hypothesis of this study.

-11-

Table No. 2: The matrix of correlations between the dimensions' scores of
the words and those of the facial expressions with two-tail significance.
-----------------------------------------------------------------------------
             |  1  |  2   |3$(9)|   4 |  5   |  6  |  7$ |  8$ |  9  | 10
* Facial     |Hppi |Inter |  F  |  A  |  Su  |Troub|  S  |  Co |Love |Disgust
  * express  |ness |   est|  e  |  n  |  rp  |  led|  h  |  nt |     |   -
   - *  ions |Sad- |  in  |  a  |  g  |  ri  |Light|  a  |  em |     |Satisf
Words  *     | ness|others|  r  |  er |  se  | ness|  me |  pt | Hate| fction
-------------|-----|------|-----|-----|------|-----|-----|-----|-----|-------
1 Contentment| .60 | .25  | -.12| .14 |-.04  |-.08 |-.10 | .02 | .13 | -.01
  - Distress |.001 |.001  | .094|.044 |**(10)| **  | **  | **  | .065|  **
-------------|-----|------|-----|-----|------|-----|-----|-----|-----|-------
2 Alertness  |-.28 | .10  | -.13| .08 | .23  | .17 | .05 | .00 | .12 | -.05
 - Serenity  |.001 | **   | .060|  ** | .001 | .018| **  | **  | .097| **
-------------|-----|------|-----|-----|------|-----|-----|-----|-----|-------
3 pride - Emb| .01 | .06  | .01 | .29 | .05  | .01 |-.19 | .09 | -.03| -.04
   arrassment| **  | **   | **  |.001 | **   | **  |.004 | **  | **  | **
-------------|-----|------|-----|-----|------|-----|-----|-----|-----|-------
4$ Excitement|-.06 |-.07  | -.09| .00 | .00  |-.07 | .05 |-.10 | -.04|  .07
-Indifference| **  | **   | **  |  ** | **   | **  | **  | **  | **  | **
-------------|-----|------|-----|-----|------|-----|-----|-----|-----|-------
5$ Pity      | .04 | .15  |  .01| .03 | .05  | .06 |-.03 |-.02 | -.10|  .10
             | **  |.018  | **  |  ** | **   | **  | **  | **  |  ** | **
-------------|-----|------|-----|-----|------|-----|-----|-----|-----|-------
6$ Longing - | .12 |-.11  |  .06| .02 |-.16  | .24 | .08 | .01 |  .04|  .03
   Scepticism|.093 | **   | **  |  ** |.022  |.001 | **  | **  |  ** | **
-------------|-----|------|-----|-----|------|-----|-----|-----|-----|-------
7$ Nonrespon |-.07 | .02  | -.02| .05 | .14  |-.08 | .03 |-.05 |  .04|  .14
  -  sibility| **  | **   | **  |  ** |.050  | **  | **  | **  |   **| .055
-------------|-----|------|-----|-----|------|-----|-----|-----|-----|-------
8 Patience   |-.05 | .09  |  .01|-.19 | .00  |-.01 | .05 |-.03 | -.04|  .04
      - Love | **  |  **  | **  |.006 | **   | **  | **  | **  |  ** | **
-------------|-----|------|-----|-----|------|-----|-----|-----|-----|-------
9 Guilt  -   | .05 |-.05  | -.02|-.05 | .04  | .18 |-.06 | .04 | -.01| -.08
     Contempt| **  |  **  | **  | **  | **   |.006 | **  | **  |  ** | **
-------------|-----|------|-----|-----|------|-----|-----|-----|-----|-------
10 conciliati| .09 | .11  |  .08|-.08 | .03  |-.09 |-.03 |-.02 | .05 | -.10
 on-Hostility| **  |  **  |  ** | **  | **   | **  | **  | **  |  ** | **
-----------------------------------------------------------------------------
(9)  $ means that the initial dimension was rotated 1800 for presentation
     clarity.
(10) ** means that the significance level is of p>.10 two-tail.

In the above matrix there is one puzzling correlation. Though the correlation of -.13 between the alertness direction of the second dimension of the words, and the fear direction of the third dimension of the facial expression missed being significant (p>.06) it still blots the clear picture a bit. As it converges with a smaller positive correlation between the fear direction and that of happiness of the nonverbal sub-study: correlation of -.05 with the unrotated third dimension of the S-TASK of table No.1, it might be the result of specific factors of the season and the subjects - the end of the year examinations.

-12-

DISCUSSION

Results supported the first hypothesis of this study while refuting the second one. Thus was refuted our theoretical effort to reconcile the contradictions between the "parsimonious" cognitive approach of Osgood and Russell (as representatives of the Equisitionists), and their abstract dimensions of emotion, with the genetic theory of emotions (of the Nativists) and its basic emotions which have concrete content.

The support of our first hypothesis together with the refutation of the second hypothesis may be seen as a considerable support for Descartes' philosophical arguments about emotional phenomena, as well as for Darwin's version of the genetic approach to emotion.

Results thus refute the claims of Spinoza and that of the abstract cognitive approach to emotion who claim that their cognitive dimensions are dimensions of emotional experience. It might be that their dimensions are the dimensions of abstract cognitive conceptualization of emotion or emotionally related processes and phenomena. Still, if one inspects the content of the first three concrete dimensions of the words of emotion, one can see that the three abstract dimensions of the parsimonious-cognitive approach are an abstraction of them.

It is still too early to decide which was the main reason for that abstraction in previous studies. It could be the result of the design of the studies - where according to Hirschberg (1980), subjects can use emotional or non-emotional dimensions of discrimination on the same items - according to the wording of the instructions. It could also be the result of subjective interpretation of the content of the dimensions by the scientist - who could not rely in the past on rich inter-correlation matrixes including both dimensions and other variables.

Our results also show that the dimensional approach to the emotional domain does not compete with that of the basic emotion - as suspected by Ekman & Friesen (1982); nor that there is a contradiction between the two - as suspected by Izard (1971). Our results clearly show that the two converge.

Furthermore, our results also show that previous studies of the basic emotions were lacking. For instance, nearly all the categorical research of the basic emotions claimed that happiness and sadness - called also by some: sorrow or distress - are two different and independent basic emotions.

However, both the sub-study of the verbal domain and the sub-study of the nonverbal domain have found them (independently) to be two opposing poles of the same emotional dimension - the first and most loaded one in both.
In addition, though not so clearly, many of the other dimensions of emotional experience and expression were found in this study to have a bipolar content, both in the verbal and the nonverbal sub-studies.

Thus, our study gives substantial support to Darwin's (1872) second principle of emotion which claims that the inborn emotions are bipolar. The bipolar findings are also congruent with findings of modern neurological studies of the Amigdala of the Limbic system of the brain. Clearly demonstrating this are findings about the bipolarity function of the Amigdala in the creation of the emotional experience of the basic emotions - as showed by Fonberg (1986) and Panksep (1986).

Our results also support Hirschberg's (1980) claims that emotional experience is too complex to be studied in the prevailing paradigm. She suggested the multidimensional unfolding model as the more appropriate paradigm for this domain and the one that would clarify many unresolved questions about the emotional domain.

********
************

-13-

APPENDIX No. 1: THE SOURCE OF THE 57 ITEMS OF MATERIALS: a\II AND a\III

The item numbers, their source and their placement on the plates:

Materials: a\I - The 48 items of the Szondi test were arranged according to the original arrangement in six groups - two rows on a plate.

Materials: a\II and a\III - The 57 items were arranged in two to three rows on the plate with three to four in each row, according to the number of items.

Plates VII - X are of the 33 items of basic emotions, XI - XII are of the 24 artificial mixtures of emotion taken from Ekman & Friesen (1975).

The list of abbreviations


An Anger     Ha  Happiness Cn Control      D Down   No. Number
Co Contempt  In  Interest  Sc Scepticism   U Up      p  Page
Dg Disgust   Sa  Sadness   Qu Questioning  L Left   71  Izard(1971)
Dt Distress  Sh  Shame     Ne Neutral      M Middle 77  Izard (1977)
Fe Fear      Su  Surprise  Mx Mixed        R Right  75  Ekman & Friesen (1975)

            The arrangement of the photographs on the plates
                   
      VII(11)            VIII                 IX                X
1 Ha 75 p112 L   | 8 In 77  p85 No.5 |17 Ha 75 p112 R   |22 In 71 p329 R
2 Su 75 p42  L   | 9 Ha 71 p236 No.2 |18 Su 75 p45  L   |23 Ha 71 p328 M
3 Fe 75 p181 R U |10 Su 71 p236 No.3 |19 Fe 75 p62  R   |26 Su 71 p328 L
4 An 75 p185 L D |11 Dt 71 p236 No.4 |20 An 75 p42  R   |27 Sa 75 p127 R(12)
5 Dg 75 p30  R   |12 An 71 p237 No.6 |21 Dg 75 p30  L   |28 An 77 p88  No.9
6 Sa 75 p193 L D |13 Sh 71 p237 No.7 |22 Sa 75 p127 L   |29 Sh 71 p329 L
7 Co 75 p183 L D |14 Fe 71 p330 L    |23 Co 75 p25  R U |30 Dt 75 p122 R
                 |15 Dg 71 p328 R    |                  |31 Fe 77 p91  No.1
                 |16 Ha 71 p237 No.9 |                  |32 Dg 71 p237 No.5
                                                        |33 Co 71 p330 M
              XI                                    XII
1 Ne    p38 L  | 7 An+Fe p96  R || 13 Su+Qu p177 R D | 19 Mx     p86  R D
2 Su+Fe p59 R D| 8 An+Cn p97  L || 14 Ne    p51  L   | 20 Ha+Su  p197 L U
3 Su+Dg p73 D  | 9 Ha+Su p108 L || 15 Co+An p185 R U | 21 Ha+Su  p197 R U
4 Sc+D  p74 L  |10 Ha+An p110 L || 16 Co+Dg p72  R D | 22 Ha+Co  p109 D
5 Mx    p86 L U|11 Sa+Fe p122 L || 17 Sc+Dg p74  R   | 23 Sa+Dg  p125 L
6 An+Dg p93 D  |12 Sa+An p123 D || 18 Fe+Dg p75  D   | 24 Sa+Ha  p126 L

_____________________________
(11) The Roman numerals indicate the plates. The Arabic numerals indicate the items in their respective group - a\II and a\III
(12) This item is of Distress according to Izard (1977). But he took it from Ekman & Friesen (1975) who made it represent Sadness.
---------------------------------

-14-

APPENDIX No.2: The list of 148 words of materials b and their subgroups

"C" Words that have the letter "C" to their left, are words of emotion that were included for control and are not part of the basic list of 96.
"*" Words that have the asterisk "*" to their left, are names of dimensions and variables that were found in previous studies and are not emotions.


    activity          delighted         haughtiness     * self-restraint
    adoration       * dependence        helplessness    * restraint
    alarm             depression        hope              right(ness)
    alertness         desire            humiliation     * rigor
    anger             despair           indifference      routine
    animosity       * Dimness           inferiority       sadness
    annoyance         disappointment  * initiative        satisfaction
    anxiety           disgust         * intensity         satisfied
  * appetite        * disqualificate    interest          scepticism
  * approval        C disquiet          involvement       scorn
  * artificiality     disregard         joy               serenity
    astonishment      disrespect        leniency          servility
    at ease           distress          loneliness        shame
    attraction        droopy            longing         * sharpness
  * balanced          embarrassment   C love            C shyness
    belonging         embitterment    * meditative      * simplicity
  C bitterness        enjoyment       C mirth           * sincerity
  * blur            C envy              misery          * honesty
    boldness        * exaggeration    C mockery         * sleepy
    boredom           excitement      * naturalness     * slumber
    calmness          fatigue           nervousness       sorrow
    caution           fear              pain            * stability
  * clearness         firmness          panic             stubbornness
    compassion        fondness        * passiveness       suffering
  * complexity      C friendly          patience        C superiority
  * complexity        frustration       pity              surprise
  C concern         C gaiety            pleasure          suspiciousness
    conciliated     C generosity      * posing            sympathy
    confidence        gloom             pride             tenderness
  * confusion         grateful          quiscence         tension
    contempt        C greediness        regret            tolerance
    contentment       grievance         relaxed           tranquillity
    courage           guilt             relief          * unstable
  C craving           bliss             repulsion       * vigilance
  * criticism         happiness       C respect         * weakness
    curiosity       * haste           C restlessness      worry
    decisiveness      hate              restrained        yearning

-15-

REFERENCES

Darwin, C. (1872) The Expression of Emotions in Man and Animals. London: John Murray (1965. Chicago: Univ. of Chicago press).

Descartes, R. (1649) Treatise on the passions of the soul. In the Philosophical Works Of Descartes: E.S. Halden & G.R.T. Ross (Trans.) Cambridge: Cambridge University Press, 1967.

Ekman, P., Friesen, W.V. (1975) Unmasking the Face. Engelwood Cliffs N.J.: Prentice-Hall.

Ekman, P., (Ed.), (1982) Emotions in the Human Face. London: Cambridge University Press.

Ekman, P., & Friesen, W.V., (1982) What emotion categories or dimensions can observers Judge from facial behaviour? In P. Ekman (Ed.), Emotions in the Human Face. London: Cambridge University Press.

Fonberg, E. (1986) Amygdala, emotions, motivation, and depressive states. In R. Plutchik & H. Kellerman (Eds.), Emotion - Theory Research and Experience. New York: Academic Press, Vol. 3.

Frijda, N.H. (1970) Emotions and recognition of emotion. In M.B. Arnold (Ed.). Feeling and Emotions: The Loyola Symposium. New York: Academic Press. Cambridge: Cambridge University Press.

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

Guilford, J. P. (1965) Fundamental Statistics in Psychology and Education. New York: McGraw-Hill, (Fourth edition).

Guttman, L. (1957) Introduction to facet analysis. In Proceedings of the Fifteenth International Congress of Psychology, Brussels 1957. Amsterdam: North Holland Publishing Co.

Guttman L.A. (1968) A generalized nonmetric technique for finding the smallest coordinate space for configuration of points. Psychometrica, 33, 469.

Hirschberg, N. (1980) Individual differences in social judgment: a multivariate approach. In M. Fishbein (Ed.), Progress in Social Psychology. Hilsdale, N.J.: Lawrence and Erlbaum Associates.

Izard, C.E. (1971) The Face of Emotion. New York: Meredith.

Izard, C. E. (1977) Human Emotions. New York: Plenum Press.

Izard C.E. (1984) Emotion-cognition relationships and human development. In C.E. Izard, J. Kagan & R.B. Zajonc (Eds.). Emotion, Cognition and Behavior. Cambridge: Cambridge University Press.

-16-

Leventhal, H.A. (1979) Perceptual-motor procession model of emotion. In P. Plinter, K.R. Blakstein and I.M. Spigel (Eds.), Perception of Emotion in Self and Others. New York: Plenum Press.

Leventhal, H. (1982) A perceptual motor theory of emotion. Social Science Information, 21, 819-845.

Lingoes, J. C. (1973) The Guttman-Lingoes Nonmetric Program series. Michigan: Mathesis Press.

Osgood, C.E., Suci, G.J., and Tennenbaum, P.H. (1957) The measurement of Meaning. Urbana: University of Illinois Press.

Osgood, C.E. (1969) Introduction. In J.G. Snider & C.E. Osgood (Eds.), Semantic Differential Technique. Chicago: Aldine.

Pankssep, J. (1986) The anatomy of emotions. In R. Plutchik and H. Kellerman (Eds.), Emotion - Theory Research and Experience. New York: Academic Press, Vol. 3.

Plutchik, R. (1980) Emotion a Psycho-Evolutionary Synthesis. New York: Harper.

Plutchik, R. (1982) A psycho-evolutionary theory of emotions. Social Science Information, 21, 529-553.

Russell, J.A. & Mehrabian, A. (1977) Evidence for a three-factor theory of emotions. Journal of Research in Personality, 11, 273-294.

Russell, J.A. (1980) A circumflex model of affect. Journal of Personality and Social Psychology, 39, 1161-1178.

Russell, J.A. & Bullock, M. (1986) On the dimensions preschoolers use to interpret facial expressions of emotion. Developmental Psychology, 22, 97-102.

Shalif, I., Lerner, Y. and Dasberg, H. (1981) A symptom profile analysis of anti-psychotic drug treatment: non-parametric multidimensional technique. Psychiatry Research, 4, 1-12.

Smith, C.A. & Ellsworth, P.C. (1985) Patterns of cognitive appraisal in emotion. Journal of Personality and Social Psychology, 48, 813-838.

Szondi, L. (1947) Experimental Diagnosis of Drives. Bern: Hans Huber.(Translated by G. Hull, New York: Grune and Stratton, 1953.)

Takane, Y., Young, F. W. & DeLeeuw, j. (1977) Nonmetric individual differences, multidimensional scaling. Psychometrica, 142, 7-67.

To get back ON-LINE with the Opening page
The other home-pages:
http://members.tripod.com/~alternativ_psy
To the TOP

For suggestions and complains click on: ilan@shalif.com

(=====To the disertation and the theory

To the opening page ---->