Alexithymia in Relation to the Use of Qualifiers, and the Quality of Emotional Tone

Alexithymia in Relation to the Use of Qualifiers, and the Quality of Emotional Tone

by Kim McMillan

Abstract

In the present study an investigation was made of the association between Alexithymia scores and the use of qualifiers as well as the quality of emotional tone in reports of an intense emotional experience. Levels of Alexithymia were measured by the Toronto Alexithymia Scale. Written descriptions of a personal experience of anger were collected from Brandon University students recruited from Introductory Psychology classes. Analysis of the written reports involved the direct counting of the number of qualifiers used and measurement of emotional tone by the application of a computer program called LOGOS. Alexithymia scores were not related to the number of qualifiers used nor were they related to the quality of emotional tone in these anger reports.

Emotions play an important part in our lives. Much of what we do and how we do it is influenced by our emotions and the conditions that generate them. Emotions are generated and controlled by the personal implications for well-being conveyed by relationships with the environment (typically social) and comprehended through an appraisal process that, whether short-circuited or not draws heavily on evolved intelligence and knowledge (Lazarus, 1991). As Lazarus says, we cannot understand people unless we understand their emotions. Imagine if you could not even understand your own emotions. Many of us have felt confused at one time or another as to what we were feeling. What if that was a regular occurrence, as it is with people who are alexithymic?

The term alexithymia was coined by Sifneos in 1972 to designate a cluster of cognitive and affective characteristics (Taylor, 1994). The most striking characteristic was a difficulty in identifying and describing feelings verbally. Alexithymic individuals have difficulty distinguishing between affect states and show an impaired ability to recognize both verbal and non-verbal emotional stimuli. The concept of alexithymia was derived from clinical observations of patients with classical psychosomatic diseases. However, development of the Toronto Alexithymia Scale (Taylor, Parker, and Bagby 1992, 1993, 1994) has revealed that alexithymia appears to be a personality trait that is normally distributed in the general population. Current prevalence has been estimated at 10% of the general population (Louth, Hare, and Linden, 1998). There are five major content areas that are consistently described: these are (1) difficulty describing feelings; (2) difficulty distinguishing between feelings and bodily sensations that accompany emotional arousal; (3) lack of introspection; (4) social conformity; and (5) impoverished fantasy life and poor dream recall. Focus of the present study is on difficulty identifying, describing, and communicating feelings.

Several studies have examined the relations between alexithymia and basic dimensions of personality characteristics. Bonnanno and Singer (1990) suggest that alexithymia may be similar to or an aspect of the repressive-defensive coping style, in which there is little tendency to experience emotional distress despite increased autonomic nervous system arousal. However, recent studies have shown that alexithymic individuals are prone to high levels of negative emotion and somatic distress, which would not be expected if alexithymia is an effective coping response. Taylor (1994) showed alexithymia to be associated positively with negative emotionality, negatively with positive emotionality, and negatively with the openness to experience dimension of personality, which encompasses personality characteristics that play an important role in affect regulation. Taylor discusses the three components to emotion response systems in humans (1) neurophysiological (2) motor- and behavioral-expressive domain and (3) the cognitive-experiential domain. All three are involved in the regulation of emotion so that one domain affects the others. The salient features of the alexithymia construct reflect deficits in the cognitive-experiential domain. Alexithymics are unable to identify their own feelings and therefore communicate to others very poorly.

Louth, Hare and Linden (1998) studied affective and behavioral similarities between psychopathy and alexithymia in a sample of female offenders, on the premise that alexithymic individuals have poor communication of their emotions this could lead to violent outbursts. They found that alexithymia, but not psychopathy was negatively related to measures of affective speech content. Both psychopathy and alexithymia were associated with a history of violence. Although detailed speech analyses were not conducted in this study and only affective words were singled out for attention, they found that alexithymics tended to use an excessive number of qualifiers such as “probably, maybe, and I imagine.” I found this observation very interesting and examined the problem further in the present study, predicting that the higher a person scores on the alexithymia scale (i.e. the more alexithymic tendencies they show), the more qualifiers he or she would use.

The purpose of the present study was to investigate in a sample of university psychology students, the association between alexithymia scores and the use of qualifiers as well as the quality of emotional tone reported by these students.

For the purpose of this study, qualifiers refer to words used to indicate the level of approximation of propositional content or of the speaker's commitment to the proposition.

To assess the subjects sensitivity to their own emotions they were asked to describe an incident in which they were involved that made them feel very angry. Lazarus and Smith (1998) came to the conclusion that retrospective reports represent an important source of evidence. They believed that even though there are problems posed by all methods of studying cognitive-emotional activity, the retrospective reports are valuable, and in their opinion the richest source of data on appraisal and emotion.

Anger was chosen as the emotion to be reported on because past studies have shown anger to be an emotion that most university students can easily recall and describe. Utilizing retrospective reports and a content analysis approach, Anderson and Ng (1994) studied causes of anger and guilt experiences. The analysis of these written reports involved the application of LOGOS.

A computer system named LOGOS was developed by Anderson and McMaster (1989,1990,1993) to use the connotative meaning word scores in the Heise dictionary to index levels of emotional tone of whole passages of text. LOGOS identifies those words in the text under examination that occur in the connotative meaning dictionary and calls them “hits”. The means of these scores on the dimension of evaluation, activity, and potency over a defined block of total words are referred to as emotional tone scores.

It was anticipated that alexithymia scores would be positively related to the number of qualifiers used and negatively related to the quality of emotional tone. The retrospective reports given by the subjects were in the emotion of anger, therefore, the subjects of lower degree of alexithymia were expected to be more sensitive and report more accurately the negative tone of such events. Specifically, it was hypothesized that alexithymia would be positively related to the evaluation dimension (pleasantness), since this is a report on anger the subjects in the low Alexithymia group should display stronger, or more, unpleasantness. Alexithymia was expected to be negatively related with activity (arousal) scores because the high Alexithymics by definition should not show as much arousal as the subjects who score low in Alexithymia. Also, Alexithymia was expected to be negatively related with potency (toughness) scores, again because this is an angry report, the higher group should show less toughness than the lower group. Alexithymia scores are also hypothesized to be negatively correlated with the polarity scores of each report. Polarity scores are a combination of the three primary emotional tone scores (evaluation, activity and potency) reflecting how much they differ from neutrality.

Another view of the data was obtained by dividing subjects into groups of high and low levels of Alexithymia. After a comparison of the scores of male and female subjects had been made. A repeated measures ANOVA was expected to reveal significant differences between the mean value of the high and low Alexithymia groups on the measure collected, with higher use of qualifiers, higher evaluation score, and lower scores in activity, potency and polarity in the higher Alexithymia group.

Method

Subjects:

The study involved Brandon University students recruited from Introductory Psychology classes. Written descriptions of a situation in which they felt angry were one half to a full page in length and were gathered from the students to be analyzed. In return for their participation, the students received one percent of course credit towards their final grade.

Materials:

Paper and pencils were given to the subjects for them to write their retrospective reports of their last experience of anger. The specific instructions that will be given to the subjects are given in Appendix A. Subjects were also given a copy of the Toronto Alexithymia Scale, which were in an envelope, to complete after their retrospective report. To preserve anonymity, rather than use names, all subjects will be given a number, which were recorded on both their retrospective reports and on their Alexithymia scales.

Procedure:

The experiment was conducted with about twenty students at a time. The subjects were given paper and a pencil and instructed to write between one-half and one full page about the last time they felt really angry. The Toronto Alexithymia Scale were also on each subject's table, although, in an envelope so they could not see it while writing their retrospective reports. The subjects were asked to fill out the twenty-item questionnaire as honestly as possible.

After all reports and questionnaires were in from all subjects the number of qualifiers in each retrospective report were counted. Each report was typed into the LOGOS computer analysis to obtain scores on evaluation, activity and potency. Scores from the Toronto Alexithymia Scale were also computed.

Analysis:

All reports were entered into the LOGOS computer program to generate the emotional tone scores for each report. Scores for the Alexithymia scale as well as the number of qualifiers in each text were counted by the researcher. Interobserver reliability for counting the number of qualifiers was obtained by having two people independently count the qualifiers of sixteen reports and calculate the correlation of their scores. The three primary emotional tones are generated and examined through the LOGOS computer program. Polarity scores are derived by squaring each of the primary emotional tone scores, adding them together and dividing by three. This indexes the emotional tone independent of direction.

The distribution of male and female scores were examined to see if the data from the two groups could be combined.

The correlations between each of the report measures and the Alexithymia scores of the subjects were calculated.

After determining the appropriate way of dividing the subjects into a high and low group according to their Alexithymia score, differences between the means of those groups were examined by repeated measures of ANOVA.

Results

Most of the University students did not find it difficult to follow the instructions given. All but one Toronto Alexithymia Scale could be used. The one mentioned was discarded because the student did not complete the back of the questionnaire. All students could remember a time when they were angry and were able to write sufficient amounts of text to be examined for our study.

A calculation of inter-rater reliability was needed for the judgment of qualifiers. The determination of inter-rater reliability rested upon the agreement of judgments about the presence of qualifiers within the retrospective reports. The raters demonstrated very similar evaluations of the reports achieving high inter-rater reliability. As shown in Table 1 the correlation between the two raters was 0.95, which was statistically significant at the p < 0.01 level, with N = 16.

Table 1
Correlation Between Two Judges Counting the Number of Qualifiers.

** Correlation is significant at the p < 0.01 level, two-tailed, N = 16.

Before continuing any further it was necessary to look at the differences between males and females to assess whether or not the data for both genders could be combined. Repeated measures ANOVA showed that the only statistically significant difference was in the number of qualifiers used, as shown in Table 2. Males were found to use more qualifiers than females, although they did not use more words in total. These findings made it possible to combine the genders and proceed to the analysis of alexithymia scores and expressive measures.

Table 2
Differences Between Word Counts and Emotional Tone Scores of Males and Females.

**Significant at p < 0.01 level, two-tailed, N = 32.

The correlations between Alexithymia scores and the seven other variables involved in this study are shown in Table 3. The correlation between hits and total words was found to be statistically significant and therefore no further adjustment for length of reports was needed. There was not a significant correlation between the number of total words and qualifiers eliminating the question of whether more qualifiers are used simply because the topic takes longer to explain. Although it was hypothesized that alexithymia scores would have significant correlations with the measures of emotional tone, this was not supported.

Table 3
Correlations Between Alexithymia, Word Counts and Emotional Tone Scores.

** Correlation is significant at the p < 0.01 level (two-tailed), two-tailed, N = 32.

Because of the centrality of Alexithymia to the theorizing about anger reports, the data was also examined by dividing subjects into high and low levels of Alexithymia based on the standard cut-off score of 51 for the Toronto Alexithymia Scale. As shown in Table 4, by using this division there were nearly equal numbers of subjects in each group.

Table 4
Alexithymia Score of Subjects Below and Above the Standard Cut-off of 51.

Means were computed to find the differences between the high and low Alexithymic groups on the seven variables shown in Table 5. The high Alexithymic group showed a lower number of total words, lower amount of hits, lower number of qualifiers, lower evaluation, higher activity, higher potency, and higher polarity. Although repeated measures ANOVA showed that none of these differences were not statistically significant, the higher potency level was in agreement with prediction.

Table 5
Means of the High and Low Alexithymia Groups on Word Counts and Emotional Tone Scores.

			Word Counts		Emotional Tone Scores
Alexithymia Group	N	Total Words	Hits	Qualifiers	Evaluation	Activity	Potency	Polarity
Low	17	216	34	3.12	-0.34	-0.24	-0.04	2.66
High	15	193	28	2.80	-0.39	-0.10	-0.07	2.67
Difference (Low-High)	32	23	6	0.32	-0.04	0.14	-0.03	-0.01

No differences were statistically significant at p < 0.05, two-tailed.

When the four highest Alexithymia scores and the six lowest scores were separated from the rest of the scores and analyzed very different results were obtained, as shown in Table 6, compared to the previous high and low groups that were used. The means for these extreme scores show the extreme high Alexithymic group to be higher in total words, higher in hits, higher in qualifiers, higher in evaluation, higher in activity, lower in potency and lower in polarity than the extreme low group. The differences in the number of qualifiers, and the scores of evaluation, potency and polarity in this sample are all consistent with what was hypothesized. Also, the differences in the total words, hits, qualifiers, evaluation, activity, and polarity are all statistically significant.

Table 6
Means of the Extreme High and Low Groups on Word Counts and Emotional Tone Scores.

**Significant at 0.01 level, two-tailed.
*Significant at 0.05 level, two-tailed.

Discussion

The purpose of this study was to investigate the association between Alexithymia scores of University students and the use of qualifiers, as well as the quality of emotional tone in reports by these students of an intense emotional situation. It was anticipated that Alexithymia scores would be positively related to the number of qualifiers used. Resting on the premise of Alexithymics difficulty in communicating it was expected that they would use more words approximating what they were feeling rather than using words that confidently commit themselves to a specific feeling. In other words, the higher the Alexithmia score the more qualifiers would be used.

The retrospective reports were given in the emotion of anger, therefore, the subjects of a high degree of Alexithymia were expected to be less sensitive and report less accurately the negative tone of such events. Specifically, it was hypothesized that Alexithymia would be positively related to the evaluation (pleasantness) dimension, and negatively related to activity (arousal), potency (toughness), and polarity (a combination of the three showing the quality of emotional tone).

Gender was examined as a variable in an early stage of analysis of the data to determine whether any differences found later might be due to differences between males and females and not just the Alexithymia score. The only difference found between the two genders was in the number of qualifiers used. Males were found to use more qualifiers than females. This was an interesting finding because males are commonly reputed to be less in touch with their feelings and therefore more tentative when reporting them. However, none of the other differences between males and females were found to be significant. Acccordingly, it was possible to further examine the data on the basis of Alexithymia scores of males and females combined.

Alexithymia scores and the variables of total words, hits, qualifiers, evaluation, activity, potency, and polarity were correlated but the results did not support the hypotheses. There were no statistically significant correlations between any of the variables and Alexithymia scores. It may be noted that there was no correlation between the number of total words and the number of qualifiers, eliminating the interpretation that more qualifiers were used when the topic took longer to explain.

Support was not obtained for hypothesized differences between the mean values of the high and low Alexithymia groups on the measures collected except for one. Although none of the differences found were significant, the higher Alexithymia group did show a lower mean potency score which had been predicted. This finding also agrees with Louth, Hare and Linden (1998) who found Alexithymia to be negatively related to affective speech content. The potency score is related to the amount of toughness shown in a person's text, since the text involved in this study has to do with angry emotions, and according to Louth, Hare and Linden the higher Alexithymics should show less affective speech content, it was therefore thought that the higher Alexithymic group would not express as much toughness.

Louth, Hare and Linden also found Alexithymia to be positively related to the number of qualifiers, and this relation was also hypothesized in the present study, although the results did not support this.

The higher Alexithymia group showed lower mean scores on the evaluation dimension which was not consistent with what was hypothesized in this study but this finding does agree with what Taylor (1994) has previously found. Taylor showed Alexithymia to be positively associated with negative emotionality. The evaluation dimension is related to the degree of pleasantness a person shows in their writing, therefore in this study it was predicted that high scores of Alexithymia would be consistent with higher pleasantness because of the anger reports that were used. It was thought that high alexithymics would not show much emotion and would not get as angry and portray those angry, negative feelings but as Taylor has stated and shown, high alexithymics do tend to express negative emotions.

Higher activity was found in the higher Alexithymic group which not only disagreed with what was hypothesized in this study but also disagrees with both the actual definition of what Alexithymia is and with what Bonanno and Singer (1990) found. The activity dimension has to do with the amount of arousal displayed in the person's text. A defining characteristic of Alexithymia is a difficulty distinguishing feelings accompanying arousal, therefore it was hypothesized that arousal would not be shown in the higher Alexithymic group's text because regardless of the level of arousal by definition they still would not clearly relay their feelings. This hypothesis also rested on the suggestion from Bonanno and Singer that Alexithymia was similar to or an aspect of repressive defensive coping style in which there is little tendency to express emotional distress despite increased arousal.

One reason some of the findings were not found to be significant or consistent with the hypotheses in the present study could be the nature of the population from which the subjects were drawn. The subjects came from Introductory Psychology classes which may not contain a large number of persons who score high in Alexithymia. Although the subjects could be divided on the basis of the standard cut-off value their scores were not widely distributed. In other studies, for instance with female offenders as in Louth, Hare and Linden (1998), extreme scores may have been uncovered and therefore significant differences between the two groups were revealed. It may be that differences in the levels of Alexithymia between people do not result in large differences in detection and expression of emotion unless they are extreme. In other words, perhaps only the very high and very low levels of Alexithymia show the defining characteristics and persons scoring at moderate levels do not. Another possible reason for the lack of correlations with measures of expression could be that Alexithymia is only slightly related to the emotional tone of a person's report when a specific emotional situation is described.

The results obtained by analyzing the extreme scores in this sample seem to support the previous argument in regards to differences between high and low groups only being found when using extreme scores. The sample of extreme scores used here though, were small.

In future studies similar to this it may be useful to have the subjects report larger amounts of text because as shown in this study there is a positive association between the total amount of words used and the number of hits detected by the LOGOS computer program for analysis. If there were larger amounts of text, therefore more hits for analysis, perhaps significant findings would be extracted.

In light of what has been previously noted with regard to extreme scores future studies of Alexithymia may require different types of populations. The differences that others have reported in expressiveness relative to Alexithymia may only be characteristic of persons with extreme scores.

The findings from this study encourage further investigation into the difference between males and females in the usage of qualifiers. It may be that males are more tentative in their reports of emotional tone, but are not systematically higher in Alexithymia.

References

Anderson, C.W. & McMaster, G.E. (1982). Computer assisted modeling of affective tone in written documents. Computers and the Humanities, 16, 1-9.

Anderson, C.W. & McMaster, G.E. (1989). Quantification of rewriting by the Brothers Grimm: A comparison of successive versions of three tales. Computers and the Humanities, 23, 341-346.

Anderson, C.W. & McMaster, G.E. (1992). Pantextual indices of emotional tone. Reader Response to Literature, 36-55.

Anderson, C.W. & McMaster, G.E. (1993). Emotional tone in Peter Rabbit before and after simplification. Empirical Studies of the Arts, 11(2), 177-185.

Anderson, C.W. & McMaster, G.E. (1997). Modeling pessimistic rumination with computer detected connotative meaning levels. Poster presentation, American Psychological Association Convention, Chicago, IL, August.

Anderson, C.W. & Manstead, A.S.R. (1993). The emotional tone of descriptions of situations provoking emotions. Unpublished Manuscript, Brandon University 0.

Bagby, R.M., Parker, J.D.A., Taylor, G.J., Endler, N.S., & Schmitz, P. (1993). Factorial validity of the 20-item toronto alexithymia scale. European Journal of Personality, 7, 221-232.

Bagby, R.M., Taylor, G.J. & Parker, J.D.A. (1994). The 20-item toronto alexithymia scale-I. Item selection and cross validation of the factor structure. Journal of Psychosomatic Research, 38, 23-32.

Bagby, R.M., Taylor, G.J. & Parker, J.D.A. (1994). The 20-item toronto alexithymia scale-II. Convergent, discriminant, and concurrent validity. Journal of Psychosomatic Research, 38, 33-40.

Bagby, R.M., Taylor, G.J. & Parker, J.D.A. (1998). Alexithymia: relationship with ego defense and coping styles. Comprehensive Psychiatry, 39(2), 91-98.

Elder, L. & Anderson, C.W. (1995). Appraisal components and core relational themes associated with anger, pride, and hope. Unpublished Manuscript, Brandon University.

Lazarus, R.S. (1991). Emotion and Adaptation. New York: Oxford University Press.

Louth, S.M., Hare, R.D. & Linden, W. (1998). Psychopathy and alexithymia in female offenders. Canadian Journal of Behavioural Science, 30(2), 91-98.

Ng, D. (1992). Aggression as a consequence of anger and guilt. Unpublished Manuscript, Brandon University.

Pappas, J.S. (1989). Qualifiers in patient-physician discourse: an analysis of interviews from radio call-in programs. ERIC document # 335928.

Taylor, G.J. (1994). The alexithymia construct: conceptualization, validation, and relationship with basic dimensions of personality. New Trends in Experimental and Clinical Psychiatry, 10(2), 61-73.

Appendix A

The following are the instructions that were given to the subjects at the beginning of the experimental session:

“I would like everyone to come up and get a piece of paper and a pencil. All papers are numbered to preserve anonymity, so please do not write your name on them. I would like you to write one-half to a full page describing everything you felt the last time you remember being angry. Once you are done writing, please pull out the questionnaire with the same number from the envelope where you are sitting and fill out all 20 questions as honestly as possible and then hand everything in.”