EMOTION PROCESSING
in
ALEXITHYMIA
by
Philippa D. Parker
M.A., University of London, 1972

THESIS SUBMITTED IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
in
PSYCHOLOGY

©Philippa D. Parker, 1996
THE UNIVERSITY OF NORTHERN BRITISH COLUMBIA
October 1996
All rights reserved. This work may not be
reproduced in whole or in part, by photocopy or
other means, without the permission of the author.

Abstract

Alexithymia refers to a lack of words for feelings and is considered to be a
personality trait, leaving a person at risk for psychiatric and psychosomatic disorders. A
reliable measure of alexithymia, the Toronto Alexithymia Scale-20 comprises three factors
which reflect the affective, cognitive and behavioural characteristics of the disorder:
difficulty in identifying emotion, difficulty in describing emotion and externally oriented
thinking. The construct is associated with low positive affectivity (PA) and high negative
affectivity (NA).
The goal of these studies was to disentangle NA and alexithymia, and to investigate
the structural or functional impairment in emotion processing. There were 46 male and
100 female volunteer subjects, aged between 17 and 55, who participated in three studies.
Subjects discriminated between facial identity and facial expression given temporal
constraint in Study 1, and bet~een facial expression of emotion stimuli and neutral
expressions given temporal constraint in Study 2. In Study 3 subjects colour named
emotion and arousal words in a Stroop Task.
In Study 1 all subjects were significantly more accurate in the identity condition. In

Study 2 high alexithymics were significantly less accurate in the recognition of emotion
given temporal constraint. Subjects with the greatest difficulty in describing emotions
were the most impaired. There was a significant difference in the recognition of sad,
angry and fearful stimuli. NA appeared to have a significant effect on perceiving
differences in Study 1, but not in recognizing emotions in Study 2. While there were no

ii

between group effects in Study 3, there was a significant difference between conditions,
with response time quickest for the baseline, and slowest for the arousal condition.
These studies provide support for alexithymia as a viable construct. They help to
disentangle NA and alexithymia, with NA having a significant effect in the attention stage
of processing and alexithymia appearing to be a functional deficit, associated with
difficulty in describing emotion, in the recognition stage of emotion processing. There
appear to be overall differences in the accuracy of recognizing sadness, anger and fear.
These studies also demonstrate that time, or temporal constraint, has a significant affect on
emotion processing.

iii

Table of Contents

Abstract

11

Table of Contents ....... ........ ................ ................ ................................ .... ......... .

1v

List of Tables .................... ............. ............. ....... ... .... ............. ....... .. ........... .....

vn

List of Figures .......... .. ... ............ ............ ...... ....... .......... .................. .... ..............

vn1

Acknowledgement ..... .... ......... ........ .... .. ............ ... .... ............... ....... .......... .........

IX

Chapter 1 - Introduction ......... ........... ........................... ................. ............ ... .....

1

Chapter 2 - Literature Review ............. .... .. ........ ... .. ... .... ...... ... ........... ...... ........ ..

5

2.1 Measuring alexithymia .... .... ... ....... ... .. ......... ........................

5

2.2 Construct validity of the TAS-20: a pilot study ....................

7

2.3 The construct of alexithymia: a general review

9

2.4 Psychosomatic illness and alexithymia .... ........ ... ... . ............... 11
2.5 Negative affectivity associated with alexithymia ...... ...... ......

13

2.6 Cognizing and related health issues .. ........... ........ ..... ............. 14
2.7 Recognizing emotional stimuli

.. .... ..... ... .. ..... .............. ... .. ..... 16

2.8 Biological substrate to recognition of emotional stimuli ... ... ... 17
2.9 Emotion and cognition ...... ....... ... .. ... ..... .. .. ..... .... ..... ..... ..... .

19

2.10 Fast tracking of sensory stimuli ... .... ... ... ... ..... ............ .. ... .... .. 20
2.11 Summary and purpose of research . .......... ... .... .... ............. .... 23
2.12 Hypotheses .. ... ... ................. .......................... .............. .. ...... 24
Chapter 3 - Method .. ....... .. ...... ....... .. .... .... .......... ... ... ... ........ .. ... .. .... .. ... .... ....... ... 27
3.1 Subjects ............. ...... .... ........ .. .... .... ................... ............. .... .. 27

iv

3.2 Measures .............................. ....... ....................... .. ............ ... 27
3.3 Materials .... ........ ... ............... ............. ....... .. .. .. ................ ... .. 28
3.4 Materials for Study 1 .......... .............................. .... ....... ......... 30
3.5 Materials for Study 2 .. ................. ........................................ 32
3.6 Materials for Study 3 ... .......... ................... ......................... .. 33
3.7 Procedure .. ....... .. .. ........... .......... .... ... ... .... ...... .................. .. .. 35
3.8 Procedure for Study 1 ........... ... .............. ....... ....................... 36
3.9 Procedure for Study 2 ............................................ .............. 38
3.10 Procedure for Study 3

............................... ......................... 39

3.11 Debriefing ..... ......................... .. ......... .. .............................. .. 40
Chapter 4 - Results ............ .......................................................................... .....

41

4.1 Questionnaires .. ............. .. .................................. .. ................ . 41
4.2 Study 1 .. ..... ................................... ............... ... ................... .. 45
4.3 Study 2 .................................... ..................................... ........ 46
4.4 Study 3 ........ ......... ............. ................................. ............. .. ... 55
Chapter 5 - Discussion ................. ........................... ......................... .................. .. 58
5.1 Study 1 ................. .......... .. ....... .. ........................................... 59
5.2 Study 2 ................... ... .. .. .. ........ .. ............... .. ............... ........ .. . 62
5.3 Study 2 and the TAS-20 factors ........................... .................... 65
5.4 Study 3 .......................................... .... .... .. ...... ... .. ............. ...... 66
5.5 Disentangling negative affectivity and alexithymia ................. 70
5.6 The role oflanguage in alexithymia .. ..................................... 73

v

5.7 Blindsight and alexithymia ... ...... ...... ............... ....... ... ..... ...... .

75

5.8 Implications of Blindsight for health ...... ............ ....... .............

76

5.9 Differences in the recognition of sadness, anger, and fear .....

79

5.10 Theoretical implications of differences between emotions .... .

80

5.11 Alexithymia and the development of emotions ........... ... .. .. ....

82

5.12 Limitations of Studies 1, 2, and 3 .... ......... .. .... ................ ......

84

Chapter 6 - Summary, Conclusions and Directions for Future Research ... .. .. .. ......

86

6.1 Theoretical implications of Study 1 and ideas for future research.. 86
6.2 Theoretical implications of Study 2 and ideas for future research.. 88
6.3 Theoretical implications of Study 3 and ideas for future research.. 90
6.4 Future research for the 'fast tracking' of stimuli ................... .. .... 90
6.5 The importance of time in emotion processing .. .. ... .. .. .. ....... ... ... 92
6.6 Temporal processing and clinical populations ................ ...... ..... 93
6.7 The possible primacy of affect .... ............................ ... ... ... . ..... .. 94
6.8 Developmental implications of alexithymia ....... .. ....... ... .. .. ....... 95
6.9 Disentangling the TAS-20 factors ............. .. .... ..... ..... ............... 96
6.10 Summary .... ........... ... ....... ... ..... .... ..... .... ..... ..... ......... ... ... ..... ... 97
References ............... ...... ................................................................................. ...... 99
Appendix A - The Toronto Alexithymia Scale-20 (T AS) ...... .. ..... .. ..... .. ..... ...... ..... 111
Appendix B- The Positive and Negative Affect Schedule (PANAS) ......... .. .. ......... 114
Appendix C - Consent Form ............. .. .. ............ ... ...... ..... ..... ............... .... .. ... .... .. .. . 116

vi

List of Tables
Chapter 4
Table 4.1

Descriptive statistics for the TAS-20 and PANAS scales

Table 4.2

Intercorrelations among the TAS-20 and PANAS measures

Table 4.3

Means for the phase order and task interaction

Table 4.4

Alexithymia group means for emotion recognition at the one and
three second levels

Table 4.5

Means for the interaction of Factor II groups (Difficulty in Describing
Feelings) with the task of recognizing emotions at the one and the three
second levels

Table 4.6

Means for the TAS-20 population in each of the four Stroop Task
conditions

Table 4.7

Means for the Stroop Task in the Baseline, Neutral, Emotion and
Arousal conditions in each of the four orders of presentation

vii

List of Figures
Chapter 3
Figure 3.1

Pairs of faces for both the identity and the Emotion conditions in Study 1

Chapter 4
Figure 4.1

Factor II differences in means for the recognition of the combined emotions
at one and three second intervals in Study 2

Figure 4.2

Overall differences in means for the separate emotions at one and three
second intervals in Study 2

Figure 4.3

Factor I differences in means for the recognition of sadness, anger and fear
at one and three second intervals in Study 2

Vlll

Acknowledgements

I should like to thank my committee, Dr. Ken Prkachin, Dr. Glenda Prkachin, Dr.
Bruno Zumbo, and Dr. Colleen Haney for their insights and contributions. I especially
acknowledge my thesis supervisor, Dr. Ken Prkachin, who always appeared to be
supportive and interested, if sometimes bemused, by my ideas and endeavours. I should
also like to acknowledge Dr. Glenda Prkachin for being my mentor in the area of
neuropsychology.
I am very grateful to the many people who generously donated their time as subjects
in these studies, and without whom this research would not have been possible.
Finally, I thank my two children, Saranna and Romany, for their patience with having
an 'experimental' mother, and for their help during this thesis process.

ix

1

CHAPTER ONE
Introduction

Alexithymia is a term coined initially by Sifneos in 1972 to refer to a lack of words
for feelings: a (=without), lex (=word) and thymia (=feeling). The construct encompasses
a cluster of affective, cognitive and behavioural characteristics observed in people, often
psychiatric patients with psychosomatic complaints. A common denominator appears to
be difficulty in identifying and describing emotions. This factor may be accompanied by
reduced fantasy involvement, occasional outbursts of emotion, stiffness and paucity in
facial expression, externally oriented and concrete thinking and confusion over
discriminating between bodily sensations and emotions.
Alexithymics have been described as at risk for psychosomatic and psychiatric
disorders (Taylor, 1994). They present a challenge for therapy since they do not respond
well to insight psychotherapy (Taylor, 1984). The aim of the present study was to shed
some light on possible differences in emotion processing between alexithymics and non
alexithymics by focusing on detection, attention, perception and cognition mechanisms.
The Toronto Alexithymia Scale-20 (TAS-20) was used to identify high, moderate
and low alexithymics using a non-clinical student population. The Positive and Negative
Affect Schedule (PANAS) was also used in order to identify subjects who were high and
low in negative affectivity. While the two constructs, negative affectivity (NA) and
alexithymia are associated and tend to overlap, it has been suggested that they are
independent of one another (Lane et al. , 1996). This study attempted to more clearly
identify and disentangle these constructs. A hypothesis was that there is a distinction

2

between the negative affectivity and alexithymia constructs, since it was expected that
some subjects would be both high alexithymic and high NA, and others would be low
alexithymic and high NA.
Much of the literature concerning alexithymia has been on construct validity (Bagby,
Parker, & Taylor, 1994; Bagby, Taylor, & Parker, 1994; Linden, Wen, & Paulhus, 1995;
Taylor, 1994; Parker, Bagby, Taylor, Endler, & Schmitz, 1993). Personality correlates
of alexithymia have also been widely researched (Fukunishi & Rahe, 1995; Kauhanen,
Kaplan, Julkunen, Wilson, & Salonen, 1993; Parker, Taylor, Bagby, & Acklin 1993;
Rubino, 1995). Health related issues constitute a third major area associated with
alexithymia (James & Large, 1991; Jimerson, Wolfe, Franko, Covino, & Sifneos, 1994;
Kauhanen, Kaplan, Cohen, Salonen, & Salonen, 1994; Millard & Kinsler, 1992; Newton
& Contrada, 1994; Sivik, 1993; Wheeler & Broad, 1994). Several studies to date have

investigated physiological reactivity in alexithymia (Martin & Pihl, 1986; Newton &
Contrada, 1994; Wehmer, Brejnak, Lumley, & Stettner, 1995). Others have indicated
that alexithymia is associated with a deficit in accurately interpreting emotional stimuli, but
results are conflicting (Berenbaum & Prince, 1994; Lane et al., 1996; McDonald &
Prkachin, 1990; Parker, Taylor, & Bagby, 1993a; Parker, Taylor, & Bagby, 1993b).
The alexithymia construct is somewhat of a paradox, in that subjects can and do
identify, describe and express emotions; yet by definition the construct implies that they
have no words for feelings. In order to complete the TAS-20, an awareness of emotion
has to be present. The aim of this project was to continue investigating and replicating in
part previous studies of the recognition of emotional stimuli by alexithymics.

Emphasis

3

was placed on understanding detection processes, attention mechanisms and the
integration of sensory information in alexithymics, to see where they may differ from nonalexithymics.
This construct is pivotal to the emotion-cognition debate.

It is argued on the one

hand that emotion and cognition are independent processes, and on the other hand that
emotion is conditional on cognition (Buck, 1993; Dodge, 1991; Lazarus, 1982; Lazarus,
1991; Zajonc, 1980). In order to better understand alexithymia, there is a need to
understand how emotions are processed, with particular attention to the negative emotions
which seem to cause the alexithymic most confusion, and discomfort (Berenbaum &
James, 1994; McDonald & Prkachin, 1990; Taylor, 1994). Based on findings of
personality correlates with the TAS-20 (Bagby, Taylor, & Parker, 1994), alexithymics
appear to be poor at differentiating between components of emotional distress,
occasionally exhibiting rage or sobbing but without being able to elaborate on what they
are feeling (Taylor,1984). Differences in sensory processing may lead to differences in
cognitive appraisal and therefore emphasis will be on understanding alexithymic behaviour
through an understanding of the neurological substrate. The literature review includes
recent studies on attention mechanisms, fear processing systems and multimodal sensory
information processing mechanisms.
Alexithymia is a potential drain on our health care system since alexithymics may be
presenting in considerable numbers in medical clinics with somatoform disorders (Taylor,
1984). There is evidence of early or possibly unnecessary diagnosis of coronary heart
disease in angina patients with alexithymic characteristics (Kauhanen et al., 1994).

4

However, prospective studies investigating the predictive validity of the TAS-20 for
psychosomatic disease and psychiatric disorders are needed. Without a clear
understanding of how alexithymics may differ from the normal population in emotion
processing, it will be difficult, if not impossible to address the need for adequate diagnosis
and ensuing therapy. Work by Hudley and Graham (1993) on attributional bias suggests
that cognitions and emotions work independently and not sequentially, and that social
cognitive behavioural therapy may change cognitions but not emotions. If there is a preappraisal emotion processing difference for alexithymics, therapy will be needed to address
that particular difference.

5

CHAPTER2
Literature Review

Alexithymia is a relatively new construct (Sifneos, 1972). However, there has been
a steadily increasing amount of literature concerning this deficit. Interest in the construct
is probably due to the fact that alexithymia is thought to be associated with many other
disorders. However, the high level of interest also indicates that researchers are not only
taking the construct seriously, but that it appears to be important to the understanding of
certain health related disorders. The review of the literature touches briefly on the
measurement of alexithymia and associated disorders. The main objective, however, is to
narrow the field to the area of emotion processing, and its biological substrate, with
particular emphasis on the 'fast-tracking' of sensory information. Further issues explored
in the literature review include the physiology associated with alexithymia, the cognizing
of emotions and the effect of cognizing on health, 'blindsight', the role of language in
emotional awareness and the effects that possible deficits in emotion processing may have
on psychosomatic illness. The emotion-cognition debate is briefly reviewed since
alexithymia may, as a construct, shed some light on the interaction of cognition and
emotion in information processing. Findings from these studies may have theoretical
implications for the primacy of emotion.

2.1 Measuring Alexithymia
A review of the assessment of alexithymia (Linden, Wen, & Paulhus, 1995) focused
on seven instruments: the Beth-Israel Questionnaire, the Schalling-Sifneos Personality

6

Scale, the Alexithymic Provoked Response Questionnaire, the Minnesota Multiphasic
Personality Inventory Alexithymic Scale, the Archetypal Test, the Toronto Alexithymia
Scale and the Analog Alexithymia Scale. Evidence suggests that the self report measure,
the Toronto Alexithymia Scale-20 (TAS-20), has adequate reliability (Cronbach's alpha=
.81) and test-retest reliability (.77) in both clinical and non-clinical populations (Bagby,
Parker, & Taylor, 1994; Taylor,1994).

Moderate discriminant validity is reported and

high convergent and concurrent validity (Bagby et al.). Bagby et al. reported a strong
and negative correlation (-.55) between the TAS-20 and the Need for Cognition Scale
(Caccioppo, Petty, & Kao, 1984), and the Psychological Mindedness Scale (-.68 ; Conte
et al. , 1990), and a significant correlation between the TAS-20 and the Beth Israel
Hospital Psychosomatic Questionnaire (Sifneos, 1973). Using the NEO Personality
Inventory (Costa & McCrae, 1985), Bagby et al. reported a non-significant correlation of
the TAS-20 with the traits of agreeableness, conscientiousness, excitrnent-seeking and
activity. Replication, however, is needed in diverse cultural groups (Parker, Bagby,
Taylor, Endler, & Schmitz, 1993). Criticisms of this scale include the fact that there has
not been enough clinical validation, there is overlap ofTAS-20 items with negative
affectivity (NA) measures, and the fact that three independent factors have to be
considered for the diagnosis of alexithymia (Linden, Wen, & Paulhus, 1995).
This scale originally comprised four factors and has been shortened and revised as
the TAS-20, (Bagby, Parker & Taylor, 1994), by dropping the daydreaming factor. It
now includes three factors: Difficulty in Identifying Feelings (Factor 1), Difficulty in

7

Describing Feelings (Factor II), and Externally Oriented Thinking (Factor ill) where
Factors I and II correlate, and Factors II and III correlate (Bagby et al., 1994).

2.2 Construct Validity of the TAS-20: A Pilot Study
A pilot study (Parker & Prkachin, 1995) conducted at UNBC evaluated the
construct validity of alexithymia by examining convergence and divergence of the TAS-20
with other stable measures of affect. These measures included the Positive and Negative
Affect Schedule or PANAS (Watson, Clark, & Tellegen, 1988), the Social Support
Measure or SSM (Seeman & Syme, 1987), the Interpersonal Reactivity Index or IRI
(Davis, 1983) measuring empathy and interpersonal reactivity, the State University of
New York Psychosomatic Symptom Checklist or SUNYA (Attanasio, Andrasik,
Blanchard, & Arena, 1984), the Interest and Preference Test or IPT (Zuckerman,
Eysenck, & Eysenck, 1978) measuring sensation seeking, the Improvisational Situations
Test or 1ST (Fletcher & Averill, 1984) measuring fantasy proneness, and the Affect
Intensity Measure or AIM (Larsen, Diener, & Emmons, 1986).
A wide range of factors have been associated with alexithymia. These behavioural
aspects tend to correlate either with the first and second factors, i.e. identification and
expression of emotion, or with the third factor, externally oriented thinking. In the above
mentioned pilot study (Parker & Prkachin, 1995) Factor I correlated significantly with the
SUNYA scale measuring psychosomatic disorders (.4), with the PANAS scale measuring
low positive affect (-.32) , and high negative affect (.47) and with the IRI measuring
personal distress (.28). Factor II correlated significantly with the IRI measuring personal
distress (.23), with the PANAS meauring high negative affect (.23) and with the SSM

8

measuring lack of social support (-.29). Factor Ill was also significantly associated with
the PANAS measuring high negative affect (.31), with the IPT measures of boredom
(.26), and disinhibition (.23), and with the IRI measures of empathy (-.51), lack of
perspective taking (-.33), and reduced fantasy (-.27). This diversity reflects in part the
independent sensory, cognitive and behavioural elements of the construct.
A later factor analytic study (Parker, 1995), based on the data from the 77 UNBC
psychology students' responses in the pilot study, was not consistent with the
interfactorial correlations found by Bagby et al. (1994), and revealed low interfactorial
correlations for the TAS-20 factors . This lack of interfactorial correlation may have been
due to the homogeneity and comparatively small sample size (JJ=77) of the pilot study
(Parker & Prkachin, 1995). It may possibly reflect the distinctiveness of the facets
comprising the alexithymia construct. There appears to be independence of the three
basic components: confusion over identifying emotion, difficulty in describing emotion,
and lack of psychological insight, or externally oriented thinking. These factors may
reflect in part different neurological mechanisms responsible for the whole construct. The
affective component or confusion over identifying emotion may impinge on the cognitive
component or difficulty in verbalizing emotion, which in tum may lead to a behavioural
component, or externally oriented thinking.

The TAS-20 may be measuring broadly a

progression of sensory, cognitive and behavioural aspects of the alexithymia construct,
and may reflect the spectrum of biopsychosocial facets of the construct which are open to
disentanglement and investigation.

9

The TAS-20 appears to be an effective measure of alexithymia, although some of
the items in the pilot study cross loaded and had low communalities. While this may have
been due to the small sample size (JJ=77), these items also appear to load differently
depending on cultural differences (Parker, Bagby, Taylor, Endler, & Schmitz, 1993).
While an overview of the literature suggests that many of the studies on alexithymia
concentrate on the first two factors (difficulty in identifying and describing emotion), and
tend to exclude or ignore the third factor (externally oriented thinking) in their analyses
(Berenbaum & James, 1994; Berenbaum & Prince 1994; Prince & Berenbaum, 1993;
Rubino, 1995), the above pilot study revealed that the items loading onto Factor III
remained the most cohesive even in a five factor analysis. Evidently more work is needed
on structuring the TAS-20, although it appears to be the most reliable, current measure of
alexithymia, and the easiest to administer, being a self-report measure (Linden, Wen, &
Paulhus, 1995).

2.3 The Construct of Alexithymia: A General Review
Alexithymia appears to be independent of socio economic status and general
intelligence (Taylor, 1994). Investigators have found that alexithymia has an inverse
relationship to psychological mindedness, verbal IQ, empathy, need for cognition, social
support, positive affect, constructive daydreaming, anger out, hypnotizability, attentional
control, disclosure, phobia, social hedonia, coping, social desireability, ego strength and
controllability over emotions (Fukunishi & Rahe, 1995; Linden, Wen, & Paulhus, 1995;
Prince & Berenbaum, 1993; Rubino, 1995; Taylor, 1994).

10

A positive relationship has been identified for depression, trait anxiety, somatization,
anger in, negative affect, symptom reporting, guilt and fear of failure daydreaming,
substance abuse, eating disorders, physical hedonia, post traumatic stress disorder
(PTSD) , and panic (Linden, Wen, & Paulhus, 1995; Taylor, 1994). While alexithymia is
positively related to depression, there is evidence to suggest that it is a separate construct.
Parker, Bagby, and Taylor (1991) found virtually no overlap in their correlation matrix
comprising items from the Beck Depression Inventory (BDI) and theTAS. Clients
assessed as both alexithyrnic and depressive have been shown to retain their alexithymic
characteristics after treatment for depression (Haviland, Shaw, Cummings, & MacMurray,
1988).
There may be both primary and secondary alexithymia, although differentiating
between the two requires both prospective and retrospective studies. Primary alexithyrnia
may indicate a genetic or inherited trait, such as a neurophysiologically based disposition
to experience anhedonia or somatization (Fukunishi & Rahe, 1995; Taylor, 1984).
Secondary alexithymia is a term used for people who begin to manifest alexithymic traits
later in life, and might be a reaction to say, stress, childhood trauma, or somatic disease
(Fukunishi & Rahe, 1995). Secondary alexithymia has been associated with PTSD and
with environmental factors in the home, such as lack of positive communication
(Berenbaum & James, 1994), and could be confounded with other psychological
problems, such as depression, eating disorders, or poor social coping skills (Jimerson et
al., 1994; Kauhanen et al, 1993; Prince & Berenbaum, 1993). It would seem to be
important to clarify and understand the issues underlying primary alexithyrnia, before

11

considering the possibility of secondary alexithymia. Although the two behaviours may
appear the same, they are not necessarily caused by the same structural or functional
deficit, and they may not require the same type of therapy.
There seems to be evidence to suggest that primary alexithymia is a personality trait
(Fukunishi & Rahe, 1995; Taylor,1994). Based on correlations of the TAS-20 with the
NEO personality Inventory (NEO-PI) as a measure of five basic dimensions of personality
(Bagby, Taylor & Parker,1994), Taylor (1994) believes that alexithymia should be viewed
as a dimension rather than a category, with high alexithymics exhibiting a deficiency of
affect at one end of a personality continuum. The construct appears to be cross-cultural
(Taylor, 1987).

2.4 Psychosomatic Dlness and Alexithymia
It has been suggested that internalizing emotions rather than spontaneously

expressing those emotions, may be associated with psychosomatic illness. A common
element in psychosomatic patients, be they alexithymics, repressors or intemalizers,
appears to be a difficulty in verbalizing emotions, and then expressing them (Buck, 1984).
Repression and alexithymia, though similar, appear to be different constructs (Myers,
1995). Repressors are thought to actively focus attention away from threatening stimuli
and minimize the cognitive impact, exhibiting high arousal and low subjective expression
(Weinberger, 1990). Alexithymics behave differently and are more similar to high anxious
subjects. They report subjective anxiety yet measures of their physiological activity are
lower than repressors (Newton & Contrada, 1994).

Lower levels of physiological

reactivity, higher baseline heart rate and tonic activity have been confirmed in several

12

studies on alexithymia; alexithymics seem to be physiologically distinct from repressors,
low disclosers and internalizers (Wehmer, Brejnak, Lumley, & Stettner, 1995).
Difficulty or confusion over how one is feeling may impinge negatively on health
outcomes through the inability to label and identify an emotion. One of the reasons for
attempting to understand the alexithymia construct better is to help identify those people
who lack the psychological insight necessary for psychotherapy. It has been suggested
that up to thirty percent of patients presenting in medical clinics are displaying
psychosomatic disorders associated with alexithymic characteristics (Taylor, 1984). They
appear to be difficult to diagnose (Kauhanen et al., 1994). Confusion over bodily feelings
may lead a person to misinterpret an emotion such as sadness, and assess it as hunger,
leading to eventual eating disorders (Jimerson et al. , 1994; Wheeler & Broad, 1994).
There seems to be a need for some brief measure to help the practitioner determine
whether these patients are having difficulty in discriminating between physical and
emotional sensations. Administration of such a measure might help in determining the
focus of treatment. It would also seem to be important to consider the causes for and
reasons behind the manifestation of this disorder by pinpointing the neurophysiological
mechanisms linking alexithymia with psychosomatic disorders and symptoms since
alexithyrnics do not appear to benefit from traditional insight oriented therapy (Taylor,
1994).

An understanding of the neurophysiological processes should help in modifying

therapeutic treatment. Taylor (1994) suggests raising alexithymics ' emotional experience
from the sensorimotor to a more sophisticated representational level, whilst allowing the
client to feel more comfortable using emotions intra and interpersonally.

13

2.5 Negative Affectivity Associated with Alexithymia
Alexithymics are apparently predisposed to experience undifferentiated distress, with
emotional impairment occurring within the interface between perception and cognition.
They appear to be implicitly extracting information from the environment realizing that
something is amiss, and then they appear to seek medical help in symptom reporting,
rather than assessing their level of emotional distress. In their confusion they require help
interpreting sensations, apparently being unable to make the implicit-explicit or cognitiveaffective connection, thereby acknowledging little controllability over their emotions
(Newton & Contrada, 1994): somewhat similar to patients experiencing a panic attack.
Low or negative affectivity may impact negatively on an individual' s socialization. High
negative affectivity may imply that a person is attending to danger or is more vigilant:
this vigilance may help or inhibit the individual depending on the situation. High positive
affectivity, on the other hand, may imply that a person is more flexible and therefore better
able to attend more freely to their environment. Fridja (1986) predicts a negative balance
for quality of life without some form of self deception, self-denial and self defense.
Repressors report both low NA and low alexithymic traits, while alexithymics tend to
report high NA (Myers, 1995). While repressors, intemalizers and low disclosers may be
practising avoidance, it seems unlikely that alexithymics are repressing or denying
information (Newton & Contrada, 1994). They appear rather to be naive or confused,
and to be at a loss as to what is happening to themselves emotionally. It seems, therefore,
unlikely that alexithymics are in a position to practise self-deception. Negative emotions
tend to persist under adverse conditions, while positive emotions disappear with

14

continuous satisfaction in hedonic asymmetry (Fridja, 1986). There is some evidence that
positive and negative affectivity may be independent dimensions (Clark & Watson, 1994).
Alexithymics seem to have a reduced ability to discriminate among the negative emotions
(Taylor,1994), and a deficit in the spontaneous expression of negative affect (McDonald
& Prkachin, 1990) may reflect differences between alexithymics and controls in the

processing of negative affectivity. McDonald and Prkachin (1990) suggested that a
deficient threat appraisal mechanism might underlie alexithymics' lack of affective
language.

2.6 Cognizing and Related Health Issues
Cognizing, or affective-cognitive integration may be beneficial to health and mental
wellbeing (Pennebaker & Hoover, 1983). When people cognitively confront and verbalize
emotional or traumatic events they seem to feel and do better (Pennebaker, Hughes, &
O'Heeron, 1987; Spiegel, 1995). It may be that symbolic representation of emotional
experience leads to greater emotional equilibrium. It is possible that some people, such as
alexithymics, fail to deal with an emotion and confuse emotional arousal with somatic
distress. Initial verbalization may intensify the negative experience momentarily but it
may reduce the individual's vulnerability for future distress (Lane & Schwartz, 1987): in
other words, short term pain for long term gain. Similar to PTSD clients, alexithymics
appear to be unable to easily identify, express or label emotions even though they seem
aware of their presence.
The way people deal with threatening stimuli may affect their ability to cognize.
This in tum may indirectly affect their ability for disclosure of emotion. Fear and

15

inhibition have been associated with the prefrontal cortex of the right hemisphere (Kalin,
1993). Lexical decision however, is associated with the left hemisphere with the corpus
callosum required for transfer of higher level information (Hoptman & Davidson, 1994).
If alexithymics are experiencing difficulty in integrating the cognitive lexical aspect of

emotion with the incoming sensory information, whether exteroceptive or interoceptive,
they may be demonstrating by default that verbalization is the brain' s way of by-passing
right frontal inhibition. By focusing attention away from that particular fear circuit or at
least allowing a feedback loop, reassessment of fear may be modified and then integrated
with other sensory information. Labelling a problem may lead to a healthier and more
functional secondary cognitive appraisal. Dispositional negative affect has been associated
with greater right anterior activation (Davidson, 1994). There has been the suggestion
that alexithymia is linked with deficits in interhemispheric transfer (Zeitlin, Lane, O'Leary,
& Schrift, 1989). Cognitive flexibility and the ability to attend to, and switch from, one

or the other hemisphere has been demonstrated in high hypnotizable subjects (Crawford &
Gruzelier, 1992). While there seems to be a positive correlation between verbal IQ and
hypnotizability (James & Large, 1991), alexithymia appears to have an inverse relationship
to hypnotizability (Linden, Wen, & Paulhus, 1995). A reduced ability to disclose
traumatic or emotional information may have ramifications for health (Pennebaker,
Kiecolt-Glaser, & Glaser, 1988; Petrie, Booth, Pennebaker, Davison, & Thomas, 1995).
It seems likely, therefore, that a deficit in verbalizing emotions, whatever the neurological

cause of that deficit, may have an impact on health.

16

2.7 Recognizing Emotional Stimuli
Since the field for exploring the construct of alexithymia is so broad, the focus in
this study will be on possible differences between alexithymics and non-alexithymics in
initial sensory information processing, or the area of perception with emphasis on
detection, attention and multimodal integration. There is evidence to suggest that
alexithymics are less accurate than non alexithymics in the recognition of verbal and
nonverbal emotional stimuli (Lane et al., 1996; Parker, Taylor, & Bagby, 1993a; Parker,
Taylor, & Bagby, 1993b) although Berenbaum and Prince (1994) suggest that
interpretation, rather than accuracy in identifying emotion, should be the focus of future
research.
Differences in attention mechanisms between alexithymics and controls have been
implicated (Parker, Taylor, & Bagby, 1993a). The Stroop task (Stroop, 1935) is used to
study attentional processes. Subjects are asked to identify the colour in which a word is
written, rather than its meaning. When the meaning of the word conflicts with the colour
it is written in, there is interference with the subject's attentional processes. Alexithymics
tend to be slower than controls on the Stroop task when arousal words are used, and
comparable to controls on non-emotion evoking stimuli (Parker et al., 1993a). It is not
known whether this difference is due to mechanisms of attention, imbalance within the
sensory integration mechanisms, or problems with interhemispheric transfer time
differences.

17

2.8 Biological Substrate to Recognition of Emotional Stimuli
Since the object of this study is to investigate emotional processing differences
between alexithymics and controls, it is necessary to briefly describe the relevant biological
mechanisms which may be involved in attention, recognition and perception of emotion.
If there are differences in either salience and or reaction times when responding to

emotional stimuli, attention mechanisms involving the posterior parietal cortex (PPC), the
superior colliculus (SC), the pulvinar and the frontal cortex, are likely to be implicated.
According to Posner and Petersen (1990), attention involves a network of anatomical
functions, three of which are orienting to sensory events, detecting signals for focal or
conscious processing and maintaining a vigilant or alert state. Shifting attention appears
to involve the PPC, SC and posteriolateral thalamus. Lesions in the parietal lobe may
affect the ability to disengage attention, while damage to the SC may affect the ability to
shift attention, and damage to the thalamus may affect the ability to engage attention. A
language attention system may involve the anterior cingulate gyrus, prefrontal cortex and
posterior parietal lobe (Posner & Petersen, 1990).
Multimodal sensory integration is thought to involve the temporal lobe, specifically
the superior temporal sulcus (STS). The temporal lobe and the STS appear to act as a
multimodal sensory convergence area (Kolb & Whishaw, 1990). Lesions here result in
deficits in language, affect and memory. Although it may not be possible to generalize
from animal studies to humans, since phenomena in humans may be different to animals,
studies of macaque monkeys indicate that the STS has cells which respond to faces,
direction of gaze and head posture (Perrett et al., 1990). Walsh and Perrett (1994) have

18

further implicated the STS in selective attention to others ' gaze. In a study of rhesus
monkeys, gaze aversion and selective viewing were measured over the first few weeks of a
monkey's life, by recording visual fixations while the monkeys were presented with faces
looking away and at them. Responsiveness to faces was found to be associated with
extreme emotionality in monkeys in their third week. Emotional responses were
diminished by the seventh week, with implications for the prefrontal cortex (Mendelsohn,
Haith, & Goldman-Rakic, 1982). Likewise, children between seven and twelve months
show fear of strangers and increased activity in the prefrontal cortex. Regulation of fear in
social referencing is modulated by interpreting expressions on another's face, with the
prefrontal cortex important in discriminating between fearful or threatening stimuli (Kalin,
1993). Three fear regulating areas appear to be the amygdala (for fear generation), the
prefrontal cortex (for assessing danger cognitively), and the hypothalamus (for
neurochemical response). For the prefrontal cortex to efficiently assess danger there must
be automaticity in interhemispheric transfer of higher level information; deficits in
interhemispheric transfer have been implicated in alexithymia (Zeitlin et al. , 1989).
Hoptman and Davidson (1994) suggest that for cognitive and high level attentional
functions ' temporal aspects of this interaction have been largely ignored' and that
'subcortical pathways are less efficient than cortical ones ' (p.214). They suggest
temporal analysis for interhemispheric interactions, with particular emphasis on attentional
systems. These present studies investigated temporal aspects of emotion processing in
attentional and recognition systems.

19

2.9 Emotion and Cognition
The construct of alexithymia seems to stand in the centre of the emotion-cognition
debate. How does one define cognition and where is the line separating emotion and
cognition? Which comes first in the path of information processing?
processes?

Are they separate

One problem is that alexithymics are supposed to be people without words

for emotions; however, they obviously have to be aware of a difficulty in emotion
processing in order to be able to complete the TAS-20 measure. Perhaps alexithymia is
similar to the notion of blindsight, where there is a demonstrated implicit knowledge while
explicit knowledge remains hidden to the subject (Weiskrantz, 1986). Alexithymics may
be relying on non-declarative and slower memory pathways for emotional situations as
outlined by Petri and Mishkin (1994). Recent studies on alexithymia indicate a
decoupling of physiological processes and subjective appraisal (Kauhanen et al., 1994;
Martin & Pihl, 1986). While alexithymics may subjectively feel anxious, their
physiological measures do not reflect this heightened arousal. Ratings of subjective
experience and physiological arousal tend to move in opposite directions, with
sympathetic activity remaining stable while subjective responses rise and appear
dissociated from the physiological experience suggesting that some affective processes
may be independent of appraisal and cognition. There may be two levels of processing,
one for cognition and one for emotion, or there may be an imbalance of implicit and
explicit levels of awareness, with less emotional stimuli entering focal awareness, or some
hitch in the parallel feedback loops necessary for integration. It seems possible that
alexithymia is a deficit of affective-cognitive integration due to the sensory stimulus being

20

assessed at slightly different moments by the affective and cognitive processes, leading to
reduced automaticity in interhemispheric transfer and an ensuing lack of informational
integration.
For the purposes of this study, cognition is held to be knowledge by description
(DK). There is reflection, awareness, and knowledge about reality. Knowledge by
acquaintence (AK) would be at the stage of detection: pre-cognitive, pre-appraisal, along
Gibsonian lines of direct perception and extraction of information from the environment
(Buck, 1993). Like agnosia, alexithymia may be a failure to transform AK into DK, a
failure to put words to sensations. The stimulus is processed, but without meaning.
Buck (1984) divided emotion behaviour into three components: autonomic and endocrine
responses, expressive behaviours (as in facial expressions), and subjective experience.
Integration of these three should be essential for emotional well-being. An emotional
dysfunction may be investigated by examining these areas and their biological substrates.
The aim of this study was to clarify where alexithymics differ from normal or healthy
controls in emotion information processing and where the first point of divergence might
lie. The underlying premise of this study was that alexithymics would differ from controls
after the detection stage of information processing, since their visual mechanisms are likely
to be intact, and there is evidence to suggest that they do not differ from normal subjects
in visual recognition of non-emotional stimuli (Parker, Taylor, & Bagby, 1993a).

2.10 'Fast-tracking' of Sensory Stimuli
The higher baseline heart rate and tonic arousal of alexithymics (Wehmer et al.,
1995) would suggest vigilant attention systems and habitual threat appraisal.

Ledoux,

21

Sakaguchi, and Reis (1984) found support for the neurological bases of AK with evidence
of an affective 'fast track' in the initial sensory processing which projects directly to the
amygdala, an area thought to be crucial in emotion processing (Ledoux, 1993). This
affective shortcut, which lacks contextual and identity information, is several synapses
shorter than the neocortical route: sensory input may travel via the medial geniculate body
(MGB) of the auditory system to the amygdala.
Sensory information about a stimulus may travel in two ways in a parallel processing
system. Information about the stimulus features may exit the sensory thalamus via the
extralemniscal areas and travel to the lateral nucleus of the amygdala, while at the same
time parallel information may travel via the lemniscal nuclei or the extralemniscal nuclei to
primary sensory association cortex areas (Ledoux, 1995). Direct projections from the
sensory thalamus to the lateral nucleus of the amygdala have fewer synapses than stimulus
feature information. This 'fast-tracking ' system has temporal advantage, but lacks
detailed information about the stimulus. This 'fast-track' system, therefore, has a
temporal advantage over perceptual, contextual, memory and polymodal features
processed in hippocampal or primary association cortex areas. The 'fast-track' pathways
exit the ascending sensory system prior to object and event representation in the unimodal
or polymodal cortex and hippocampus, allowing the amygdala to respond to simple
features with incomplete stimulus information from the thalamus (Ledoux, 1993): a form
of AK. This system may then take precedence over the more complete object information
from the cortex, at the expense of perceptual precision (Ledoux, 1993), with implications
for classical conditioning and lack of habituation (Ledoux, 1995).

22

If people are conditioned to avoid emotional stimuli, they are unlikely to attend to

them, and through avoidance fail to learn to cognize such stimuli. Further behaviour is
likely, therefore, to be manifested on the basis of the 'fast-tracked' information, rather
than on the basis of information which has been implicitly identified and given a context.
Alexithymia is associated with panic as opposed to phobia (Parker, Taylor, Bagby, &
Acklin, 1993) which may support the theory that alexithymics tend to 'fast-track'
information at the expense of identity information, since phobias tend to be for specific
stimuli, while panic can be triggered by more generalized situational cues.
One might expect that people who may be processing emotional stimuli in this way
would also have a higher baseline heartrate (HR) and tonic arousal, such as is found in
alexithymia (Wehmer et al, 1995). Higher baseline tonic arousal and HR may affect the
attention systems. Having a vague sensation of danger but without being able to
symbolize, or verbalize the cause might also keep the subjective anticipation of danger at
increased levels, and physiology on constant alert. This vigilence may lead to elevated
cortisol and may damage hippocampal neurons, which are important in motivation,
memory and emotion (Kalin, 1993).
Alexithymics' emotional distress or arousal seems to lack clear differentiation
(Bagby, Taylor, & Parker, 1994), implying that they appear to experience most confusion
when there is a need to discriminate among negative emotions (Taylor,1994). Ledoux
(1995) suggested that individuals whose fear mechanisms are triggered via the subcortical
amygdala route without the hippocampal connection would likely have limited emotional
and psychological insight. He further pointed out that any such emotional memories are

23

likely to continue long after extinction of the emotional behaviour, and that therapy for
such individuals may consist of trying to effect more efficient cortical synaptic links with
the amygdala. Lesions to cortical connections lead to lack of extinction in classical
conditioning. Alexithymics behave as if they have an impaired integration of affective and
cognitive mechanisms. This may have led to lack of extinction or longer habituation for
fearful memories due to overuse of the 'fast-track' sensory information processing, and
possible avoidance of cognitively processing emotions, especially of sad, angry or fearful
stimuli; similar to PTSD clients. Maturational studies suggest that the amygdala matures
earlier than the hippocampus, which implies that early conditioning is non-declarative, and
possibly inaccessible to memory (Ledoux, 1995). Dantzer (1991) points out that if a
defense system is overused during childhood, this pattern may hold for adulthood and
could explain the somatization characteristic of people high on the negative affect scale.
He suggests competition between internal and external stimuli vying for a common
behavioural pathway may lead some people to notice bodily sensations over social clues.

2.11 Summary and Purpose of Research
Alexithymia appears to be an impairment which is associated with many psychiatric
and somatic disorders. It seems, therefore, important to investigate the structural and
functional mechanisms underlying this disorder, so that early assessment and appropriate
therapy may be implemented. Alexithymia is considered to be a personality trait, or a
continuum, which is normally distributed in the population. The TAS-20 seems, at
present, to be the most reliable measure of alexithymia and is easy to administer, being a
self-report questionnaire. Alexithymia appears to be a deficit in emotion processing and

24

those with the greatest difficulty in identifying and describing emotions, are likely to also
experience the greatest difficulty in disclosing their feelings . This deficit may have an
negative effect on their health, and could contribute to psychosomatic disease, psychiatric
disorders, or medical problems, such as eating disorders, panic attacks , or substance abuse
(Taylor, 1994). Since alexithymics appear to be impaired rather than incapable of
integrating cognitive and emotional information, this may be due to impaired fear
processing, selective attention to non-emotional stimuli, reduced intermodal automaticity,
and reliance on non-declarative and slower memory pathways for emotional situations.
The main purpose of this study was to concentrate on the mechanisms responsible
for the deficit in verbalizing and identifying emotions.

Particular emphasis was placed on

the temporal aspects of emotion processing and the implications that any deficit in
temporal processing might have on emotional awareness.

The studies were exploratory

in nature, and their purpose was to re-examine the relationship between alexithymia and
the ability to attend to, recognize, identify, and integrate emotional stimuli. It was
decided to examine selective attention to identity versus emotion stimuli, sensitivity to
facial expressions of emotion versus neutral expressions, and response time to verbal
expressions of emotion or arousal versus baseline or neutral words.

2.12 Hypotheses
Alexithymics appear to have a deficit in emotion regulation specifically in the
integration of affective and cognitive stimuli. They seem to be deficient in the processing
and elaboration of emotional stimuli, rather than in the initial detection of stimuli. While

25

they may be able to detect stimuli normally, they may be less able to shift attention or
integrate affective-cognitive information.
The primary hypothesis was that although alexithymics might be able, given
sufficient time, to recognize and identify emotional stimuli, their ability to do this would be
eliminated or reduced, given temporal constraint. They might be receiving a 'fasttracked' message, followed by more detailed information of the same stimulus. These
two inputs would then arrive slightly out of synchronization in the sensory integration
area, thereby slowing down the individual ' s response to the stimulus.
The second hypothesis was that alexithymics would only respond more slowly to
emotional stimuli, and that their response to non-emotion stimuli would be similar to nonalexithymics, as in the Stroop task (Parker, Taylor, & Bagby, 1993a).

Since alexithymia

is associated with negative affectivity (NA), alexithymic subjects should be more
distracted than non-alexithymic subjects by stimuli evoking states of negative emotional
arousal, and be less able to attend to the task.
A third hypothesis was that the slowest response time would be for negative affect
stimuli, i.e. anger, sadness or fear. This hypothesis was due partly to the premise that they
might have a deficit in threat appraisal, and also to the idea that positive and negative
affectivity are separate dimensions.
The fourth hypothesis was that this deficit might, in part, be due to selective
attention to non-emotional stimuli. There is evidence to suggest that recognition of faces
and processing of emotional facial expression are dissociated (George et al., 1993;

26

Humphreys, Donnelly, & Riddoch, 1993; Magnussen, Sunde, & Dyrnes, 1994; Prkachin
& Prkachin, 1994; Young, 1992).

The fifth hypothesis was that NA and alexithymia are separate constructs, and that a
deficit in the recognition of, or selective attention to, facial expressions of emotion stimuli,
would be due to alexithymia, rather than NA.
The sixth and last hypothesis was that higher levels of alexithyrnia would result in a
reduced ability to respond to emotional stimuli.

27

CHAPTER3
Method

3.1 Subjects

There were 146 volunteers aged between 17 and 55 (M= 26.5). Of the 146
subjects, 79 were recruited from the UNBC Psychology subject pool. Students from the
subject pool received bonus credit for testing. The remaining 67 subjects were recruited
directly on campus. There were 46 males aged between 18 and 55 (M=25.76) and 100
females aged between 17 and 52 (M=26.8). There were 8 lefthanded subjects. Colour
blind subjects were not accepted. There were 11 students (mostly Mandarin speakers) for
whom English was a second language.
3.2 Measures

All subjects were administered two self-report measures: the Toronto Alexithymia
Scale-20 (TAS-20; Bagby, Parker, & Taylor, 1994; see Appendix A), and the Positive
and Negative Affect Schedule (PANAS; Watson, Clark, & Tellegen, 1988; see Appendix
B). The TAS-20 has demonstrated internal consistency, good test-retest reliability,
construct and criterion validity, and a stable and replicable three factor structure (Parker,
Bagby, Taylor, Endler, & Schmitz, 1993). The three factors comprise: Factor I or
Difficulty in Identifying Feelings, Factor II or Difficulty in Describing Feelings and Factor
III or Externally Oriented Thinking. The scale consists of twenty items scored on a
Likert scale with a possible score range from 20 to 100. The PANAS comprises two 10
item mood scales measuring positive and negative mood dimensions. There is a ten item

28

scale measuring negative affectivity (NA) and a ten item scale measuring positive
affectivity (PA). The two scales contain adjectives describing different feelings. The
respondent is asked to describe, using a 5 point Likert scale, the extent of each emotion
experienced, in this case "on the average". The NA scale is internally consistent
(Cronbach's alpha= .87), has high test-retest reliability (.71), with good convergent and
discriminant validity (Watson, Clark, & Tellegen, 1988).

3.3 Materials
The visual stimuli for Studies 1 and 2 were created from Ekman and Friesen's
Pictures of Facial Affect (1976). This was a set of 110 35mm black and white cardboard
mounted slides, numbered and coded. Apart from neutral faces, depictions of only three
emotions were utilized from the collection: sadness, anger and fear. Since Alexithymia is
associated with low PA and high NA (Linden, Wen, & Paulhus, 1995) and difficulty in
distinguishing between anger, sadness, and fear (Bagby, Parker, & Taylor, 1994; Taylor,
1994) the slides used portrayed negative emotions only. The facial expressions of
surprise and disgust were not used, since disgust is hard to differentiate from anger, and
surprise is hard to distinguish from fear (Prkachin & Prkachin, 1994). The emotion slides
were all judged to show the designated emotion by 70% or more of the observers in
Ekman and Friesen's original study (Ekman & Friesen, 1976).

The sadness slides

included numbers 2, 8, 15, 23, 36, 43, 49, 58, 67, 75, 76, 77, 86, 87, 94,102, and 103,
portrayed by 8 females and 5 males. The anger slides included numbers 3, 10, 18, 25,
30, 38, 44, 52, 53, 61, 62, 69, 80, 89, 96, 105, and 106 portrayed by 8 females and 6
males. The fear slides included numbers 9, 16, 17, 24, 37, 50, 51, 59, 60, 68, 78, 79, 88,

29

95, and 104, portrayed by 6 females and 5 males. The neutral slides included numbers 6,
13, 21, 28, 33, 41, 47, 56, 65, 72, 83, 92, 99, and 110, portrayed by 8 females and 6
males. In order to be able to create images from these slides on video or in the printed
form, each slide was scanned using a Sprint Scan Slide Scanner, manufactured by
Polaroid. The scanned slides were converted to computer images and filed in
PaintshopPro under the original Ekman and Friesen coding and numbering. A second
collection was created on file and all numbers or codes visible on the picture were
removed. These pictures were then used to create the visual images for both Study 1 and
Study 2.
On hand in the testing room for all studies were two Sony cassette-corders. A
Sony cassette-corder TCS 580 contained a blank tape (TDK 90) and a Sony cassettecorder TCM 939 contained a pre-recorded tape (Memorex 60). Eight tapes had been
pre-recorded for both Study 1 and Study 2. These tapes were approximately one half
hour in length, and were designed for each of the eight possible orders of stimulus
conditions utilized in the studies. General information about the experiments was given at
the beginning of each tape, followed by detailed instructions for the subjects, practice
sessions for both studies, and the two studies themselves. Audible beeps at one second
intervals were created on the pre-recorded tape by using the pacer mechanism of a Smart
LCD multi-function stopwatch, manufactured by Intertan Inc. (1993).
Record sheets for each study were created so that the experimenter could note the
responses. Numbers on the record sheet tallied with the recorded numbers on the tapes
for each study and each phase.

30

3.4 Materials for Study 1
Study 1 was a repeated measures design consisting of two conditions as the within
subject variable. In one condition subjects discriminated on the basis of facial emotion
and in the other condition subjects discriminated on the basis of facial identity. For both
conditions, pairs of faces were created as one picture, each face set adjacent to the other
(see Figure 3.1 below). Each pair displayed the faces of either a male or a female.
Gender, identity and emotion were counterbalanced. For the Identity condition, two sets
of 50 pairs of faces were created. Each face pair consisted of two faces, portraying the
same emotion. There were 23 male face pairs and 27 female face pairs. Half of the pairs
portrayed two identical faces displaying the same emotion, while the other half portrayed
two different faces portraying the same emotion. For the purposes of the later analysis,
the former were considered "noise" and the latter "signal". For the Emotion condition, an
additional 100 pairs of faces were created. Each face pair consisted of the same face,
represented twice. Half of the pairs showed the same face portraying the same emotion,
while the other half showed the same face portraying different emotions. For the

•

purposes of later analysis, the former were considered "noise" and the latter "signal".
There were 22 male face pairs and 28 female face pairs. Of the "signal" pairs there were
17 face pairs combining sadness and anger, 17 face pairs combining sadness and fear, and
16 face pairs combining anger and fear.

31

Figure 3.1 The. top pair of faces is an example of the pairs of faces shown in Study 1 in

the identity condition, where subjects discriminated on the basis of facial identity. The
bottom pair of faces is an example of the pairs of faces shown in Study 1 in the emotion
condition, where subjects discriminated on the basis of facial emotion.

32

Once the pairs of faces had been arranged as a computer image, these pairs were
printed out onto bond paper. Each face pair image measured Scm by 4cm. The face pairs
were then set in a random order onto a page measuring 21cm by 28cm. There were
fifteen pairs of faces on one page, in five rows, three to a row. Each image was
numbered from 1 to 100 in each condition with the number centered underneath the face
pair. There were six pages with 15 images and one page with ten images for each
condition. Each page of slide images was covered by a protective plastic cover and set in
a loose-leaf binder. Two practice pages of paired slide images, one for the identity
condition and one for the emotion condition were constructed using smaller images
measuring 2.5cm by 2cm. There were 25 practice face pairs for each condition, five to a
row, with five rows.

3.5 Materials for Study 2
Study 2 was a repeated measures design with time as the within subject variable. In
Study 2 subjects viewed 400 slides in total, 200 in each phase or condition. The slides
used for Study 2 were identical to those in Study 1, except that single faces portraying
sadness, anger, fear or neutrality were used. Whilst the recognition rates of individual
slides were judged to show the appropriate emotion by 70% or more of the observers in
Ekman and Friesen's normative sample (1976), the average percentage agreements for the
400 slides used in Study 2 were 87% for sadness, 88% for anger and 87% for fear. These
faces were subsequently printed out, using the same procedure as in Study 1, to measure
4.5cm by 3cm each. Each image was then set on white bond paper measuring 21cm by
28cm, with 16 faces to a page in four rows of four faces, and protected with a plastic

33

cover. Each facial image was numbered from 1 to 200 for each phase. Both phases
were comprised of equivalent sets of images, with 50 neutral faces, 50 sad faces, 50 angry
faces and 50 fearful faces in each condition. The images were arranged randomly, except
that the same identity or emotion was not repeated in succession. For the purposes of
later analysis the neutral faces were considered "noise" while the sad, angry and fearful
faces were considered "signal". Study 2 consisted of two phases, Study 2a and Study 2b.
These phases were essentially the same except the stimuli for Study 2b were set in reverse
order to the stimuli in Study 2a. In both phases the facial expressions of affect were
portrayed by 93 male faces and 107 female faces . Depending on the order to which the
subject had been allocated, the phases were pre-recorded onto audio cassettes in one of
two ways. One phase was recorded at one second intervals using the beeper mechanism
of the Smart LCD stopwatch's pacer function. In other words, there was a one second
interval between the numbers assigned to each facial image. The other phase was prerecorded at three second intervals, so that there was a three second interval between the
numbers assigned to each image, although the beeping remained constant at one second
intervals throughout.
3.6 Materials for Study 3
Study 3 was a partial replication of a previous study by Parker, Taylor, and Bagby
(1993a) and involved the Stroop Task. The original study had three conditions: baseline,
neutral and arousal. The present study consisted of four conditions: baseline, neutral,
emotional and arousal. Four cards were constructed from white poster board measuring
71cm by 56cm. There were 100 words on each card in 20 rows with five words in each

34

row. Each word measured 1.2cm high and was printed in capital letters using Crayola
non-toxic felt tip pens. The colours used were orange, red, black, blue and green. Each
word was arranged randomly and repeated twenty times, four times in each of the five
colours. Neither word nor colour was repeated in succession. The baseline and neutral
cards followed the format used by Parker et al. The baseline card consisted of five
00000' s instead of a meaningful word. The neutral card consisted of the words:

polite, moderate, clever, typical and fair. The emotion card consisted of the words:
anger, surprise, sadness, fear and disgust. The arousal card consisted of the words:
torture, mutilation, cancer, assault and murder.
The arousal words in this study differed from those in the original study by Parker et
al. (1993a). The Committee for this research proposal considered the word "rape" to be
potentially unethical, in that it might evoke an unnecessarily extreme reaction in a subject,
especially if they had been raped. In order to create five equivalent words to the original
group (rape, vagina, penis, Aids , victim), a pilot study was run prior to the testing period,
using UNBC graduate students. These students were presented with 81 words, including
the five original arousal words and the five neutral words.

Arousal was defined as "a

word that evokes in you a state that is different from normal by being distressing or that
brings up unpleasant thoughts or feelings". Adjacent to the listed words was a Likert
scale of one to five, with 5 indicating greatest arousal and 1 the least arousal. The word
"rape" had the highest mean (4.71) and lowest standard deviation (.71). Torture, one of
the words chosen for this study, had a lower mean rating (3.79) and a larger standard
deviation (1.08). The arousal words chosen for this Study had a higher overall mean

35

(3.53) than the original words (2.88). The neutral words evoked the least arousal with an
overall mean of 1.42.

3.7 Procedure
Subjects were tested at UNBC over a five week period. The experimenter was the
same person throughout the testing period. Testing time allotted for each participant was
one hour and subjects were tested on an individual basis. For the three studies there were
eight possible orders of presentation with Study 1 occurring first in each instance. Study
1 was followed by Study 2, and Study 2 was followed by Study 3. Half the subjects filled
out the two questionnaires after completing Study 1, and the other half of the subjects
filled out the questionnaires having completed Study 3. There was no time constraint on
the filling out of questionnaires. Each subject arrived on the hour, e.g. at 9:00 a.m.
Forms for credit were filled out first for 79 of the subjects. The subjects had been told
that they would be participating in a one hour experiment on "information processing".
Subjects were told that there would be two short questionnaires and three experiments
involving visual materials and that these experiments would be timed and fast. They then
read and completed the consent form (see Appendix C) which repeated the above
information and explained confidentiality and their rights as a subject. Gender and age
were noted for each subject at this time. Each subject was assigned a letter and number
code. Subjects then sat in a comfortable chair facing a desk. On the desk surface were
the file folder containing the stimuli for Studies 1 and 2, and the two Sony cassettecorders. There were eight pre-recorded tapes; one for each of the eight consecutive
subject orders, and several blank tapes.

36

3.8 Procedure for Study 1

Study 1 was administered first to all subjects. Half the subjects received the
Identity condition first and half of the subjects received the Emotion condition first, as the
within study order. Study 1a represented the Identity condition, whilst Study 1b
represented the Emotion condition. The experimenter, who was also the researcher,
randomly assigned subjects as they arrived to one of eight orders. Orders 1-4 were
essentially the same as orders 5-8 except that subjects assigned to orders 1-4 filled out the
two questionnaires having completed all the studies, and subjects assigned to orders 5-8
filled out the two questionnaires after completing Study 1. Given the 146 subjects, these
eight orders were rotated 18 times with the final three subjects taking orders 1-3 of the
19th rotation. A record was kept of the order tally and the subject numbers and codes.
The experimenter first turned on the blank tape, explained to the subject that this
tape would pick up their responses, together with the recorded numbers and instructions
from the other tape. They were told that this recording tape would be wiped clean and
reused once all their responses had been correctly entered on the record sheet. Tape
counters were set at zero for each cassette-corder. Next the experimenter recorded the
date, order number and subject number on the recording tape and switched on the prerecorded tape. Subjects were instructed to listen to the pre-recorded tape and follow the
recorded instructions. The initial spoken recorded instructions repeated the information
regarding the two questionnaires and the three studies, and explained about the differing
order of presentations. Subjects were asked to sit comfortably and listen to the numbers
being called out and to answer loudly enough for the second cassette-corder to pick up

37

their voice clearly. Subjects were instructed to say "Yes" if they perceived a difference
between the faces in each face pair and to keep silent if they perceived no difference.
They were told that Study 1 involved speed and not to worry if they missed responses, just
to keep on going and to relax. They were to ask before each testing if there was anything
they did not understand since tests could only be run once. It was explained to the
subjects that they would view 200 slides in Study 1; 100 in phase 1, 100 in phase 2. They
would see two faces at the same time and they should indicate by saying "yes" if the faces
seemed different in any way, and remain silent if they judged the faces to be the same.
The faces could be different in a number of ways but the experimenter would not
specifically indicate the difference. They were to make a judgement and respond within
the given time limit. They were to shift attention to the next slide, only when they heard
the appropriate number. The face pair numbers were read simultaneously with the one
second beeps on the recorded tape. In other words , the numbers were recorded at one
second intervals, apart from those instances where the page needed turning.
Listening to the pre-recorded instructions, subjects were instructed to turn to the
front of the file folder and find the first and second practice pages for studies la and lb.
On both practice pages subjects viewed 25 face pairs. Subjects were told to respond by
saying "yes" if, and only if, the faces seemed different in any way. If they perceived no
difference they were to remain silent. If they felt that they needed more practice, this
option was open to them, although no subjects took advantage of this option. Each face
pair in the study was numbered from 1 to 25 for the two practice pages, and from 1 to 100
for both Study la and lb. At the end of each page of 15 face pairs in Study 1, subjects

38

were asked to "please turn the page". This process continued throughout the first phase,
followed by the second phase. The experimenter was careful to check at the beginning of
each testing that each subject was looking at the correct set of slides, i.e. the phase order
to which they had been assigned, Study la or lb. The experimenter sat behind the
subject, close enough to hear the responses but unable to see which slides were being
viewed. The experimenter recorded by pencil line a "yes" response for each numbered
face pair on the printed record sheets, leaving the silent numbers blank. If the initial
response was unclear to the experimenter or missed, the experimenter listened to the
recording tape later that day to verify the responses. Tapes were wiped clean after this
process.
3.9 Procedure for Study 2
There were four possible within study orders for Study 2.

Subjects allocated to the

first and fifth order, viewed the first phase (Study 2a) first at intervals of one second.
Subjects allocated to the second and sixth order, viewed the first phase (Study 2a) first at
intervals of three seconds. Subjects allocated to the third and seventh order viewed the
second phase (Study 2b) first at intervals of three seconds. Lastly, subjects allocated to
the fourth and eighth order viewed the second phase (Study 2b) first at intervals of one
second.
At the commencement of Study 2, subjects had either just completed Study 1, or
had completed Study 1 followed by completion of the TAS-20 and PANAS
questionnaires. Listening to the pre-recorded instructions, subjects were instructed to
turn to the front of the file folder and find the third and fourth practice pages for Studies

39

2a and 2b. On both practice pages subjects viewed 25 single faces measuring 2.5cm by
1.5cm each. Subjects were told to repond by saying "yes" if, and only if, they perceived
an emotion. If the face appeared to be emotionless, they were to remain silent. Each
face in the study was numbered from 1 to 25 for the practice pages, and from 1 to 200 for
Study 2a and 2b. Presentation of both phases in Study 2 was essentially the same except
that one was presented more quickly (one second intervals) than the other (three second
intervals). Subjects were again instructed to shift attention to the following facial image
only after hearing the appropriate number. If they missed a response they were to relax
but to keep on going. Mter the sixteenth image on each page, they were instructed;
"please tum the page" and the process resumed until all 200 facial images had been viewed
for each phase. The experimenter followed the procedure already outlined under general
procedure.
3.10 Procedure for Study 3

Study 3 followed Study 2. The recording tape was left running, but the prerecorded tape was switched off at this point. Subjects turned round to face the
experimenter. The experimenter, blind to the subjects' TAS-20 and PANAS scores,
balanced the poster board so that it faced the subject. There were four possible orders for
the Stroop Study, and subjects were randomly assigned to one of the four orders. This
presentation order differed from the original study by Parker et al. (1993a), in that they
had presented the Stroop cards in a fixed order. The present Stroop orders, where
!=baseline, 2=neutral, 3=emotional and 4=arousal were: first Stroop order = 1,2,3,4,
second Stroop order= 4,3,2,1, third Stroop order= 3,4,1 ,2, and the fourth Stroop order

40

= 2,1,4,3. Subjects were instructed by the experimenter to name the colour each word
was written in, from left to right, top to bottom. They were asked to ignore the meaning
of the text, and just to say the colour the word was written in, as fast and as accurately as
possible. Although subjects had been screened for colour blindness, each subject was
asked to identify and name out loud the five colours prior to the Stroop Task. To make
sure that the subjects had understood the instructions, they were asked the question: "If
you saw the word 'experience' written in red, what would you say? (Red). Using the
Smart LCD stopwatch, the experimenter timed the subjects' total time taken to colour
name each Stroop card in hundredths of seconds.
After the task was completed subjects were invited to describe how they felt about
the words, and which words, if any, they remembered. These were recorded. It had
been noted by the experimenter that some subjects tended to laugh easily whilst doing the
Stroop Task, some appeared unaffected, and some seemed tense.
3.11 Debriefing
After completion of the Studies and questionnaires, subjects were asked if they had
any questions or concerns. They were asked to avoid discussion of the experiments
because of the nature of Study 1, but general questions about the nature of these studies
could be addressed at this stage on an individual basis. They were also advised that the
Stroop Task could be stressful, and the theory underlying the Stroop Task was explained
to the subjects.

41

CHAPTER4
Results

4.1 Questionnaires
All statistical results were computed using SPSS 6.1 for Windows. After
completion of the data entry, an error rate was calculated for the data set for Studies 1, 2,
3, and the questionnaires. Data for 15 subjects, more than ten percent of the total number
of subjects, were checked and one error found which resulted in a projected error rate of
.000098 for the entire study.
Since alexithymia appears to be a dimensional rather than a categorical construct
(Taylor, 1994), descriptive statistics for the TAS-20 were used to distinguish among low,
moderate and high alexithymia groups. Subjects were identified as low, medium or high
alexithymic on the basis of their TAS-20 scores, using the 33rd and 66th percentiles to
distinguish among the three groups. The 33rd percentile had a value of 42 and the 66th
percentile had a value of 50. The low alexithymia group was made up of 49 subjects with
a mean score of 35 .14, and a standard deviation of 4. 7 5. The medium alexithymia group
was made up of 53 subjects with a mean score of 46.77 and a standard deviation of 2.33.
The high alexithymia group was made up of 44 subjects with a mean score of 57.41 and a
standard deviation of 5.32. The earlier pilot study (Parker & Prkachin, 1995) comprising
77 UNBC students had a minimum score of 31 and a maximum score of 69, with a mean
of 47 and a standard deviation of 9.38.

Thus, the present sample resembled that in the

previous sample and, in fact, appears quite comparable to the university sample of Bagby,

42

Parker, and Taylor (1994). Descriptive statistics for theTAS and PANAS scales are
shown in Table 4.1.

Table 4.1
Descriptive Statistics for the TAS-20 and PANAS Scales
SCORES

minimum

maxunum

mean

S.d.

TAS

24

71

46.08

9.87

DIP

7

29

16.25

5.12

DIDF

5

23

12.34

4.08

EOT

8

28

17.48

4.10

PA

19

48

34.96

5.68

NA

11

41

20.30

5.82

Note. TAS refers to the Toronto Alexithymia Scale-20; DIF refers to Factor I or
Difficulty in Identifying Feelings; DIDF refers to Factor II or Difficulty in Describing
Feelings; EOT refers to Factor III or Externally Oriented Thinking; PA refers to Positive
Affect; NA refers to Negative Affect.

43

The PANAS scores were identified as either negative affectivity (NA) or positive
affectivity (PA). Subjects were identified as either high or low NA or PA on the basis of
a median split in either scale. The median for the NA scale was 19, while the median for
the PAscale was 35. In order to maintain a reasonable balance within cells given the
limited sample size, the median for NA was taken as 18, rather than 19. Seventy subjects
with scores equal to or less than 18 were identified as low NA with a mean score of 15.59,
and a standard deviation of 9.92. Seventy-six subjects with scores equal to or more than
19 were identified as high NA, with a mean score of 24.64 and a standard deviation of
4.71. Seventy-five subjects with scores equal to or less than 35 were identified as low PA
with a mean score of 30.47 and a standard deviation of 3.71. Seventy-one subjects with
scores equal to or more than 36 were identified as high PA with a mean score of 39.70
and a standard deviation of 2.78.
Whilst 32 percent of the entire TAS-20 population were male, 27 percent of the
low alexithymic group were male, 36 percent of the medium alexithymic group were male
and 32 percent of the high alexithymic group were male. Crosstabulation of the NA and
TAS groups revealed that low TAS and low NA shared 38 subjects or 78 percent of the
low TAS group. High TAS and high NA shared 32 subjects, or 73 percent of the high
TAS group. Low TAS and high NA shared 11 subjects or 22 percent of the low TAS
group, while high TAS and low NA shared 12 subjects or 27 percent of the high TAS
group. High TAS and high P A shared 7 subjects or 16 percent of the high TAS group.
The intercorrelations among the .measures are shown in Table 4.2:

44

Table 4.2
Intercorrelations among the TAS-20 and PANAS Measures
DIP

DIDF

EOT

NA

PA

TAS

DIP
DIDF

.5391
p=.OOO

EOT

NA

PA

TAS

.0819

.3511

p=.326

p=.OOO

.5634

.3644

.1745

p=.OOO

p=.OOO

p=.035

-.4498

-.4297

-.3080

-.4053

p=.OOO

p=.OOO

p=.OOO

p=.OOO

.7755

.8390

.6035

.5154

-.5390

p=.OOO

p=.OOO

p=.OOO

p=.OOO

p=.OOO

Note. DIP refers to Factor 1 or Difficulty in Identifying Feelings; DIDF refers to Factor
2 or Difficulty in Describing Feelings; EOT refers to Factor 3 or Externally Oriented
Thinking; NA refers to Negative Affect; PA refers to Positive Mfect; TAS refers to the
Toronto Alexithymia Scale-20.

45

4.2 Study 1
Following the tradition of Pollack and Norman (1964), a non-parametric index of
sensitivity, the average of all possible maximum and minimum areas under the ReceiverOperating-Characteristic curve defined by the subject's performance, or A prime (A'), was
calculated using the following formula: A'= .5 + ((y-x)(1+y-x))/4y((1-x)), where y equals
the probability of hits and x equals the probability of false alarms (Grier, 1971). A hit was
defined as saying a signal was there when it really was there, and a false alarm as saying a
signal was there when it was not. In this study the number of hits was greater than the
number of false alarms in all cases and therefore it was possible to use this formula.
Before running the split plot ANOV As, a 2 x 2 repeated measures ANOVA was run
as a preliminary analysis with gender as the between subject variable and task (identity or
emotion discrimination) as the within subject variable. There was no significant main
effect of gender F(1,144)=.65, n=.421, nor was there a statistically significant gender by
task interaction .EO, 144)=.85, n=.359. Since all questionnaires were administered after
completion of Study 1, their order was not considered relevant to the analysis.
Consequently, gender and order of questionnaires were omitted from further analyses in
Study 1.
The ultimate repeated measures ANOVA was then a 3 (alexithymia groups) x 2 (NA
groups) x 2 (order of phases) by 2 (task: identity versus emotion discrimination), with the
latter factor involving a repeated measure. There was a main effect of task, fl1, 134) =
20.32,...IL= .001, with the mean A' for the identity discrimination (.94) exceeding the mean
A' for the emotion discrimination (.90). There was also a statistically significant phase

46

order by task interaction EO ,134) = 4.71, l2..= .032. Means for the phase order by task
interaction are shown in Table 4.3 below. For both the identity and the emotion
condition, the means were higher in phase order 1. In other words, if the identity
condition was presented first, the means for that task were higher than those presented
after the emotion condition. If the emotion condition was presented first, the means for
that task were higher than those presented after the identity condition.

Table 4.3
Means for the Phase Order and Task Interaction in Study 1
PHASE ORDER 1

PHASE ORDER 2

M

M

IDENTITY

.95

.05

.94

.06

EMOTION

.91

.09

.89

.11

Note. Phase order 1 is the phase in which the task was presented first; Phase order 2 is
the phase in which the task is presented second.

4.3 Study 2
Study 2 was a between-within experimental design, with TAS and NA groups, order
of phases and order of time as the between subject variables, and task (recognition of
emotion) as the repeated measures factor.

A' values were calculated from the

probabilities of hits and false alarms for emotion conditions at the one second and the
three second level.

In a later analysis A' values were also calculated from the

probabilities of hits and false alarms for each separate emotion condition at the one and

. 47

three second level. A hit was defined as saying an emotion was there when it really was,
and a false alarm as saying an emotion was there when it was not.
Before running the split plot ANOVAs a 2 (order of questionnaires) x 2 (gender) x 2
(task) repeated measures ANOVA was run as a preliminary analysis. There were no
significant main effects for gender f(l,142) = .34, 12..= .561, for order of questionnaires
f(l,142) = 1.05, 12..= .308, or for an order of questionnaires by gender interaction f(l,142)

= .25 , 12..= .619. Consequently, gender and order of presentation of the questionnaires
were omitted from further analysis in Study 2. The next repeated measures ANOV A was
a 3 (TAS groups) x 2 (NA groups) x 2 (phase order) x 2 (order of time) x 2 (task) with
the latter factor involving a repeated measure. Tests involving between subjects effects
showed a statistically significant main effect for theTAS groups f(2 ,122) =3.70, 12..=
.028. Means are presented in Table 4.4 below.

Table 4.4
Alexithyrnia Groun Means for Emotion Recognition at the One and Three Second Levels
MEANS

THREE SECONDS

ONE SECOND
M

SD

M

SD

Low Alexithymia

.94

.03

.96

.03

Mod. Alexithymia

.92

.05

.95

.04

High Alexithymia

.91

.04

.96

.03

Note. Low Alexithymia refers to the low alexithymic group; Mod. Alexithyrnia refers to
the moderate alexithymic group; High Alexithymia refers to the high alexithyrnic group

r--

48

A Tukey's HSD post hoc test (Kirk, 1995) revealed a significant difference between
the high and low alexithymic groups in the one second time condition. There was no
significant main effect for the NA groups f(1,122) = .01, lL= .922. Tests involving within
subject effects revealed a statistically significant main effect of task.E(1,122) = 85.60, lL=
.001.

The mean for the three second condition (.96) exceeded the mean (.92) for the one

second condition. There was no significant TAS group by task interaction E(2, 122) =
1.75, lL= .178, nor was there a significant NA group by task interaction .E(1,122) = .28, 12
= .595.
TAS tercile groups for each TAS factor were substituted for theTAS groups for the
next three repeated measures ANOVAs. These ANOV As were 3 (Factor groups) x 2
(NA groups) x 2 (phase order) x 2 (order of time) x 2 (task) with the latter factor
involving a repeated measure. There was no significant main effect for either Factor I
(Difficulty in Identifying Feelings) E(2,122) =.21, 12=.814, or for Factor III (Externally
Oriented Thinking) E(2,122)=.41, lL=.661. There was a statistically significant main
effect for Factor II (Difficulty in Describing Feelings) E(2,122) = 4.07, 12 = .019. Tests
involving within subject effects revealed a statistically significant Factor II by task
interaction£(2, 122) = 4.02, 12 = .020. Figure 4.1 below shows the interaction of Factor
II (Difficulty in Describing Feelings) with the task (emotion recognition at the one and
three second levels).

49

-._,,.... 1.00
c(

-..

• 1-sec
3-sec

c

0
0

eG

..
--..
..•
..

~

0.95

0

)o.

.a

0.90

g

e
e

c

0.85

Low

Medium

High

Alexithymia Group

Figure 4.1 Interaction of Factor II (Difficulty in Describing Feelings) tercile groups with
the task (recognition of emotion) at one and three second intervals in Study 2.

50

A Tukey's HSD post hoc (Kirk, 1995) revealed a significant difference between the
low and high Factor II group means and between the moderate and high Factor II group
means in the one second condition. Group means for the three second condition were not
significantly different from one another. Table 4.5 below shows the means for the
interaction of Factor II and the task.

Table 4.5
Means for the Interaction of Factor 2 Groups (Difficulty in Describing Feelings) with the
Task of Recognizing Emotions at the One and the Three Second Levels.
MEANS

ONE SECOND

THREE SECONDS

M

M

DIDF 1

.94

.04

.96

.03

DIDF2

.92

.04

.96

.03

DIDF3

.91

.05

.96

.03

Note. DIDF 1 refers to low Difficulty in Describing Feelings; DIDF 2 refers to moderate
Difficulty in Describing Feelings; DIDF 3 refers to high Difficulty in Describing Feelings.

To further analyze the sources of these differences, a series of ANOVAs was
conducted to investigate subjects ' abilities to recognize the individual emotions. The
ANOVA designs were 3 (TAS groups) x 2 (NA groups) x 3 (emotions) x 2 (time) with
repeated measures on the latter two factors. Since neither the phase order nor the order
of time was statistically significant in the previous analyses, these between subject
variables were omitted from further analysis. For the between subject effects there was a

51

statistically significant main effect of TAS groups E(2, 140) = 3.54, n = .032. There was
no significant main effect for the NA groups E(1,140) = .01, n = .935, nor was there a
significant TAS by NA group interaction E(2,140) = .37, n = .692. Tests involving the
emotion within subject effects showed a statistically significant main effect of emotion
E(2,280) = 171.22, n = .001, but no statistically significant TAS group by emotion
interaction E( 4,280)=1.10, n = .358, and no significant NA group by emotion interaction
E(2,280)=.53,...n = .589.

A Tukey's HSD Post Hoc test (Kirk, 1995) showed that the

combined mean for sadness (.92) was significantly different from the combined mean for
anger (.95), and that the combined mean for anger was significantly different from
combined mean for fear (.96). Tests involving time within subject effects showed a
statistically significant main effect for time f(1,140) = 101.33, n = .001 with the means for
the 3 second condition (.96) exceeding those of the one second condition (.92). Tests
involving emotion by time within subject effects showed a statistically significant emotion
by time interaction f(2,280) = 27.84, n = .001 (see Figure 4.2 below). Post Hoc tests
showed there to be a significant difference between means for sadness at the one (.89) and
at the three (.94) second intervals.

Neither anger nor fear had statistically significantly

different means in the one and three second conditions, although the means for anger at
the one (.93) and at the three (.96) second intervals were close to being significantly
different.

52

-~

<(

1.00

t:

...
0
0

E
G)

0.95

....0

...
.c
...as
...

~

0.90

()

G)

Cl)

Q

0.85

Sadness

Anger

Fear

Emotion

Figure 4.2 Differences in means for the separate emotions at the one and the three
second intervals for the combined alexithymia groups in Study 2.

53

The individual emotion analyses were repeated using the TAS-20 factors to
discriminate among groups. In the analysis using Factor II (Difficulty in Describing
Feelings) there was a statistically significant Factor II by time interaction, .E( 2,140) =
4.21 ,J2 = .017 (see Figure 4.1). Tests involving Factor I (Difficulty in Identifying
Feelings) showed a statistically significant Factor I by emotion by time interaction
.E(4,280) = 3.66, 12 = .006. A Tukey's HSD Post Hoc test (Kirk, 1995) revealed a
significant difference in means for the recognition of sadness between the high (.88) and
the low (.90) Factor I groups in the one second condition (see Figure 4.3).

54

--

1

1 sec

3 sec

<t

c 0.96
0
;
0

E
Cl)

0 0.92

-.c~

uns 0.88
....
Cl)
Cl)

c

0.84

Fear

~

Low

AngerSadness
•

Medium

Fear
•

AngerSadness
High

Figure 4.3 Factor I (Difficulty in Identifying Emotions) group means for the recognition
of sadness, anger, and fear, at the one and the three second intervals in Study 2.

55

4.4 Study 3
Study 3 was a between-within experimental design with theTAS tercile groups, the
NA median split groups, and the order of presentation of the task as the between subject
variables, and the task (colour naming words) as the within subject repeated measure. A
preliminary repeated measures analysis was a 2 (order of questionnaires) x 2 (gender) x 4
(task). In the tests involving between subject variables there was no statistically
significant main effect of order of questionnaires .E(1,142) = .45, .Q = .502, no statistically
significant main effect of gender F(l,142) = 1.89, 12 = .171, and no statistically significant
order of questionnaires by gender interaction.£(1,142), =.55, .Q = .460. Consequently,
these factors were not considered in the main analysis which was a 3 (TAS groups) x 2
(NA groups) x 4 (order of presentation) x 4 (task) ANOVA, with the latter factor
involving a repeated measure. Tests involving within subject effects showed a statistically
significant main effect for task .E(3,366) = 87 .20, .Q = .001. A Tukey' s HSD post hoc test
showed that the baseline condition means were significantly different from the neutral
condition, the emotion condition, and the arousal condition means. The arousal condition
means were significantly different from the emotion condition means, while the emotion
and neutral conditions were not significantly different from one another (See Table 4.6
below).

56

Table 4.6
Means for the T AS-20 Population in each of the Four Stroop Task Conditions
CONDTIONS

BASELINE

NEUTRAL

EMOTION

AROUSAL

MEANS

68.86

78.58

78.99

80.82

SD

11.92

13.49

14.95

14.71

There was a statistically significant order of presentation by task interaction
.E(9,366) = 6.24, 12 = .001 (see Table 4.7 below).

Table 4.7
Means for the Stroop Task in the Baseline, Neutral, Emotion and Arousal Conditions in
each of the Four Orders of Presentation
MEANS

BASELINE
M

NEUTRAL

EMOTION

AROUSAL

SD

M

SD

M

SD

M

SD

ORDER1

68.53

(11.00)

78.75

(11.70)

79.78

(13.22)

84.18

(13.98)

ORDER2

69.47

(12.37)

81.14

(16.13)

80.17

(16.56)

77.43

(14.43)

ORDER3

70.08

(13.58)

79.95

(13.20)

75.95

(12.60)

80.70

(14.03)

ORDER4

67 .37

(10.85)

74.42

(12.02)

79.99

(17.11)

80.98

(16.15)

Nate . Order 1 = baseline, neutral, emotional, arousal. Order 2 = arousal, emotional,
neutral, baseline. Order 3 =emotional, arousal, baseline, neutral. Order 4 = neutral,
baseline, arousal, emotional.

57

Since this was a partial replication of a previous study (Parker, Taylor, & Bagby,
1993a), there was another repeated measures ANOVA using cutoff scores for the low
(scores less than or equal to 51), medium (scores ranging from 52 to 60) and high (scores
greater than 60) alexithymia groups that were used by Parker et al. (1993a). There was
neither a statistically significant main effect of TAS groups_E(2, 125) = .37, .12 = .690, nor
was there a statistically significant alexithymia by task interaction f(6 ,375) = .81, .12 = .566,
in this 3 x 2 x 4 x 4 repeated measures ANOVA.

58

CHAPTERS
Discussion

The underlying purpose of these three studies was to examine the relationship
between alexithymia and the processing of emotional stimuli. It was hypothesized that
alexithymics' ability to detect visual stimuli would be normal, but that their ability to then
attend to or integrate affective-cognitive information might be impaired.
In Study 1 selective attention to identity versus emotion stimuli under temporal

constraint was examined.

It was expected that alexithymics would respond less

accurately than non-alexithymics when perceiving differences in emotion stimuli, but that
alexithymics and non-alexithymics would respond similarly when perceiving differences in
identity stimuli.
In Study 2 recognition of facial expressions of emotion versus neutral expressions

was examined in two timed conditions.

It was expected that alexithymics would respond

less accurately than non-alexithymics to facial expressions of emotion given temporal
constraint, but that they would respond similarly to non-alexithymics in the slower
condition.
In Study 3 the overall response time to verbal expressions of emotion and arousal
versus neutral and baseline words was examined in a Stroop Task.

It was expected that

alexithymics would respond more slowly than non-alexithymics in the emotion and arousal
conditions, but that they would respond similarly to non-alexithymics in the baseline and
neutral conditions.

59

In all three studies it was hypothesized that higher levels of alexithymia would be

associated with less accurate and slower response times to negative emotional stimuli, but
that higher levels of alexithymia would result in no difference in accuracy or response time
to non-emotional stimuli. It was also hypothesized that alexithymia and negative
affectivity, although clearly associated, are demonstrably separate constructs. It was
hoped that the results of these studies would help disentangle the two.

5.1 Study 1
In Study 1 all subjects responded less accurately to emotion stimuli than to identity

stimuli. There were no significant differences among the three alexithymic groups,
although the low alexithymic group responded more accurately than the moderate group,
who in tum responded more accurately than the high alexithymic group. There was no
significant difference between the NA groups, although the low NA group responded
more accurately than the high NA group. The fact that no significant differences were
found among the alexithymic groups may be due to the specific question posed in the
experiment. In Study 1 subjects were in fact required to compare faces and to say
whether the two faces presented were the same or different.

Study 1 therefore measured

the ability to compare faces and perceive a difference, rather than the ability to say
whether an emotion was present or not. Having to compare faces may involve less
complex processing, and different neural systems, than having to discriminate among
emotions, or having to recognize emotions.

Discrimination may imply that some kind of

implicit labelling is required in order to differentiate between emotions: Recognition of
emotions may involve similar processes. Discrimination and recognition of facial

60

expressions of emotion also imply a knowledge of emotion. Comparison of faces in the
emotion condition may involve more processing than comparing faces in the identity
condition, but less processing than for the recognition of emotion.

Being able to

discriminate among emotions may involve more complex processing than comparing faces,
or perceiving a difference between faces. The results show a trend for alexithymic groups
to be similar in the identity condition, but increasingly, though not significantly, less
accurate in the emotion condition. The objective of Study 1 was to see whether
alexithymics attended to identity rather than emotion, by hypothesizing that the high
alexithymic group would be less accurate in the emotion condition. It appears that all
subjects attended more accurately to facial expressions of identity than to facial
expressions of emotion given temporal constraint.
The results of Study 1 seem to imply separate processes for the recognition of
expressions of identity and emotion.

They may imply fewer synapses or less complex

processing for the task of perceiving differences, whatever those differences may be. The
overall difference in accuracy between the two conditions, identity and emotion, suggests
both a qualitative and quantitative difference in the way that we attend to and process the
recognition of facial identity from the way we attend to and process the recognition of
facial expressions of emotion. It suggests that, given temporal constraint, we are better
able to attend to, and process stimulus features of facial identity, and that we process them
more accurately and hence more quickly than we process stimulus features of facial
expressions of emotion.

61

Finding a difference in accuracy whilst perceiving emotion versus identity stimuli is
consistent with the evidence suggesting that facial identity and facial expression are
processed separately (George et al., 1993; Humphreys, Donnelly, & Riddoch, 1993;
Magnussen, Sunde, & Dyrnes, 1994; Prkachin & Prkachin, 1994; Young, 1992).
However, while the findings of Study 1 are consistent with the concept of separate
mechanisms for the processing of facial expression and facial identity, they are not entirely
consistent with prior research suggesting a generalized or shared processing for identity
and expression (Braun, Denault, Cohen, & Rouleau, 1994; Endo, Endo, Kirita, &
Maruyama, 1992).

The findings of Study 1 may lend support for the idea that facial

identity and facial expression of emotion share common neural mechanisms for processing,
but may diverge for higher order processing (George et al., 1993; Weddell, 1989).
The increased accuracy in the identity condition may suggest fewer synapses in
identity processing, since the information appears to be transmitted faster. Several
subjects were unable to respond with any accuracy to the emotion condition. This was
not true for the identity condition. The fact that more subjects had difficulty with the
emotion condition when required to shift from one condition to the other, suggests that
processing of facial expression overlaps with, and perhaps builds upon, the processing of
facial identity. It may suggest that people are pre-disposed to attend to features of
identity. Future research is needed to disentangle these two processes. In Study 1
subjects were required to respond within a one second time interval. It would be
interesting to see whether there is any difference in accuracy between the two conditions if
subjects are asked to respond within a slower time interval. It may be important to see

62

what happens if the exposure duration is manipulated in both directions; i.e. either faster
or slower.
The fact that each condition was perceived more accurately when presented first
may suggest that priming plays a role in attention mechanisms and possibly affects our
ability to shift to, disengage from or neglect stimuli. It may equally well suggest that
fatigue contributes to less accurate responses. The fact that both the identity and emotion
conditions were affected to the same degree by order presentation may lend support for
the speculation that both abilities do share common pathways at some stage in the
processing of information.
In summary, Study 1 was intended to shed some light on the role of alexithymia in

selective attention processing. The hypothesis that alexithymics would attend to identity
stimuli rather than to emotion stimuli was not confirmed.

On the whole, all subjects

appeared to process facial expressions of identity more accurately than they did facial
expressions of emotion. These results suggest that, while processing for facial
expressions of identity and facial expressions of emotion may share common pathways,
there are also differences in the processing required for these two types of stimuli. The
results also imply that alexithymia is not a significant factor in the attention stage of
information processing.

5.2 Study 2
The results of Study 2 were consistent with the expectation that alexithymics would
be less accurate than non-alexithymics in the recognition of emotional stimuli given
temporal constraint, and that higher levels of alexithymia would result in less accurate

63

responses. For the purposes of this study, the term ' recognition' was used to imply an
ability to both detect, and elaborate on, emotional stimuli.

Having to perceive emotional

stimuli within a one second time frame resulted in an overall lower accuracy for all
subjects. However, there was a significant difference between the alexithymic groups.
The least accurate responders were the high alexithymic group, and the most accurate
responders were the low alexithymic group. Responses to emotional stimuli within the
three second time frame were not significantly different for the three alexithymic groups.
This difference in accuracy between the one and three second time frame suggests that the
one second interval was perceived as a constraint by all subjects and placed additional
demands on the emotion processing of the high alexithymic group.
These results suggest that time is an important factor in emotion processing for all
subjects. The fact that high alexithymics recognize the presence of emotional
information similarly to the low and moderate groups given sufficient time, suggests that,
in a normal situation, they are able to detect, attend to and recognize facial expressions of
emotion. This is consistent with the previous suggestions that alexithymics were able to
accurately identify facial expressions of emotion, and pose facial emotions similarly to
non-alexithymics (McDonald & Prkachin, 1990). Alexithymics are able to recognize
emotion, but appear to be impaired when faced with excessive demands, such as temporal
constraint. Since high alexithymics appear to behave similarly to low alexithymics in the
slower condition, the results also suggest that alexithymics have a functional rather than a
structural deficit in emotion recognition since the differences between groups were only
apparent in the temporal constraint condition. This assumption needs further testing.

64

The fact that high alexithymics were less accurate in perceiving emotion given
temporal constraint lends support to the suggestion that, in a situation in which they
experience excessive demands such as temporal constraint, they may process negative
'

emotion stimuli differently to the way in which low or moderate alexithymics process the
same information. The fact that high alexithymics were less accurate in perceiving
emotion within the one second time frame seems to imply that they 'fast track' information
when faced with excessive demands, and that this method of processing sensory or
emotional information leaves them at a disadvantage as to the precise nature of the
stimulus. The term 'fast-tracking ' is somewhat confusing, since it implies that the
processing of information is quicker than normal or multimodal processing. 'Fasttracking' implies that information about a stimulus is processed more quickly, but also
more crudely, given temporal constraint. When a person is faced with excessive
demands, that person may process certain stimuli as if the stimuli were threatening, or as
if the situation demanded that the information should be processed more quickly than
normal. The benefits of 'fast-tracking' are that the person may be able to react
:

appropriately and instantaneously to the stressor. The disadvantage of 'fast-tracking ' is
that the person does not consciously know details about the stimulus. This processing
by-passes, as it were, the contextual and identity features of the stimulus.
It has been suggested that alexithymics may have a deficit in the spontaneous

encoding of negative emotions (McDonald & Prkachin, 1990). Alexithymics are reported
to have higher tonic arousal than normal subjects (Wehmer, Brejnak, Lumley & Stettner,
1995) and this higher baseline state may predispose them to 'fast-tracking' . Their

65

impairment in recognizing the presence of emotion appears to be contingent on the
perception of threat, or distress, or on challenging the perceptual processing systems'
capacities in some other way involving time. In this study, this challenge, or the
perception of threat or distress, was prompted by the condition of temporal constraint.
Again, this seems consistent with the previous finding that alexithymics were only
impaired in spontaneous displays of negative emotion (McDonald & Prkachin, 1990).

5.3 Study 2 and the TAS-20 Factors
Prior to the three studies, the hypotheses had been limited to assuming that the three
factors contributing to the construct of alexithymia; Factor I or Difficulty in Identifying
Feelings, Factor II or Difficulty in Describing Feelings and Factor III or Externally
Oriented Thinking, would be of equal significance.

Previous researchers, however, had

implied that, depending on the area of research under investigation, certain factors
contributing to the construct appear to be more closely aligned (Berenbaum & James,
1994; Berenbaum & Prince, 1994; Prince & Berenbaum, 1993; Rubino, 1995).

Cross

validation of the factor structure suggesting that Factors I and II correlate and Factors II
and III correlate (Bagby, Parker, & Taylor, 1994) seems to be replicated by the
interfactorial correlations in these studies (see Table 4.2). The results from Study 2
suggest that Factor II, or Difficulty in Describing Feelings, plays more of a role in the
ability to recognize emotional stimuli when under time constraint, than the other two
factors ; the greater the difficulty in describing emotions, the greater the impairment in
recognizing emotions given temporal constraint (see Figure 4.1 ).

66

TAS-20 items contributing to Factor II are comprised of questions which indicate
the extent to which the person has words for feelings or is able to articulate those feelings
to others (see Appendix A). The questions measure difficulty in finding the 'right words'
for feelings, and the ability to both describe or reveal feelings to others. Subjects
belonging to the group that had the most difficulty in describing feelings were significantly
less accurate in recognizing emotional stimuli in the one second time frame, but were
similar to the low and moderate groups in the three second time interval (see Figure 4.1).
Neither Factor I, Difficulty in Identifying Feelings, nor Factor III, Externally Oriented
Thinking, appeared to have any significant effect on the ability to recognize emotion
stimuli under time constraint when considering the combined emotions.

It would appear

that, although Factor II or Difficulty in Describing Feelings is moderately correlated with
the other two TAS-20 factors, it is distinctly different from them in its ability to impair
accuracy in recognizing emotions given temporal constraint.

These results are consistent

with recent findings suggesting that Factor II is associated with an impairment in the
recognition of verbal and non-verbal emotional stimuli (Lane et al., 1996). For future
research it might, therefore, be important to investigate the neural mechanisms implicated
in these three factors in order to pinpoint the neural structures which are specific to Factor
II, and which are not shared by the other two factors.

5.4 Study 3
Study 3 was a partial replication of a previous study using the Stroop Task (Parker,
Taylor, & Bagby, 1993a) with the addition of an emotion condition. The original study
had three conditions: baseline, neutral and arousal. In this study an emotion condition

67

was added, thus yielding four conditions: baseline, neutral, emotion and arousal. In spite
of the fact that the subjects in this study appeared to be similarly distributed on the
alexithymia continuum as in the Parker et al. study, their results were not replicated, even
when using the same cut off scores for the high and low alexithymia groups that were used
in their study. This difference may reflect other methodological differences between the
studies. While Parker et al. chose to present the Stroop tasks in a fixed order, the Stroop
tasks in this study were presented randomly and there was a significant order effect, with
each condition having the quickest response time when presented first. Their rationale for
using a fixed order was based on a study of suicide attempters (Williams & Broadbent,
1986) where it was found that using a fixed order or random order made no difference to
the outcome or results. However, the important comparison in that study was between
response times for general negative words as opposed to specific negative words. Only
general negative words were used in this study.
In addition to the difference in order presentation, there was also a difference in

words chosen for the arousal condition. The words selected by Parker et al. (1993a) had
been used previously as high threat stimuli for rape victims, who experienced greater
cognitive interference than the control subjects in this task. The words in this study were
chosen for their general distressing effect as opposed to the specific distressing effect of
the words chosen by Parker et al. This difference in words may also contribute to the
lack of between group differences in colour naming interference. Rape is often associated
with PTSD, which is also associated with secondary effects that resemble alexithymia.
We do not know whether the high scores of the alexithymic group used by Parker et al.

68

were due to primary or secondary alexithymia. If there were subjects in their high
alexithymic group with secondary alexithymia associated with sexual trauma, it would
explain how the specific words used in their study incurred significant group differences.
The word 'rape' had the highest overall arousal effect when tested in the pilot study
administered prior to Study 3. Of the 81 words used in the pilot study, 'rape' clearly had
the highest mean score for all subjects, suggesting that there is something specific about
the word 'rape', which is distressing for both male and female subjects. The arousal
effect of the other words used by Parker et al., e.g. 'penis, vagina, Aids and victim', may
have been compounded when seen in conjunction with the word 'rape' in the Parker et al.
study.
Although there was no significant difference between groups, there was a significant
difference in overall response times across the conditions of Study 3 (see Table 4.6). All
subjects were quickest in the baseline condition. The responses for the neutral and
emotion words were not significantly different from each other. The response times for
the arousal words were the slowest for all subjects. There was a trend for the response
times for the alexithymia groups to move in the expected direction; the quickest responses
being for the low alexithymia groups and the slowest responses for the high alexithymia
groups in each condition. Given the possibility that NA may be a contributing factor in
alexithymia, and in attending to a stimulus, the fact that high alexithymics were slowest in
the baseline condition may be due to a high level of NA.

Results from a later regression

analysis on the data suggest that NA may be implicated in the ability to 'disengage' and
'shift' attention from one stimulus to the other. However, it seems somewhat surprising,

69

in the light of this argument, that NA did not have more of an effect in the Stroop Task.
This could be due to the nature of the stimulus: i.e. , words rather than facial expressions.
The fact that there was no significant difference between groups may also be due to the
nature of the Stroop Task. It appears to be a somewhat clumsy and crude method of
measuring subtle clinical differences, and may be better used in, say, cardiovascular
studies, in order to create stress or tension for susceptible subjects.
In the course of administering Study 3 it became apparent to the experimenter that

the words which each subject found the most distressing did appear to cause the most
interference. For instance, if the subject had particularly noticed the word 'anger' then
the emotion condition tended to be the slowest for that subject. However, those words
which

~e e

to affect individual subjects were specific to each subject and could have

belonged to the neutral, emotional and arousal conditions, although most did seem to
belong to the arousal condition. This variability in specifically distressing words might
explain the variability in response times which may have contributed to washing out any
between group effect or interaction.

However, specifically chosen arousal or emotion

words for specific clinical populations might well incur greatest interference in the colour
naming Stroop task, as seen in the Parker et al. study (1993a). It appears that emotion
words as such are are not distressing for alexithymics. Perhaps the words have little
affective significance due to the fact that emotion words may, in the case of alexithymia,
have been assigned via cognitive, rather than emotion processing systems, as in the case of
' reentrant processing' (Posner, 1994).

I
i

70

I

.!
It may be of significance to the understanding of alexithymia as a construct to
consider that arousal, rather than emotion, appears to be a significant factor. Although
significant group differences were not found in this study, the trend for high alexithymics

to be slower in responding than low or moderate alexithymics was more pronounced in the
i
I

arousal condition than in the emotion condition. Arousal was defined as 'most
distressing' in Study 3. The implication from these studies, and from previous research
(Bagby, Taylor, & Parker, 1994), is that alexithymics experience undifferentiated 'distress'
as opposed to a specific emotion. It may, therefore, be more appropriate to think in terms
of the 'distressing' rather than the 'threatening' or 'stressful' effects of presenting the
stimuli with a temporal constraint condition. This would tie in with the results of Study 2,
where time, in the form of temporal constraint, appeared to cause the most interference.
If the temporal constraint condition acted as a means of 'heightening arousal' in the sense

of being distressing, this could be the situational factor which has the most impact on
alexithymics. In other words, being aroused or distressed may cause the most
interference in the processing of emotional information for alexithymics in particular, and
for everyone to a lesser extent. It may also tie in with the fact that alexithymics are
described as having a higher baseline heart rate and tonic activity (Wehmer et al. , 1995).

5.5 Disentangling Negative Affectivity and Alexithymia
In spite of the substantial correlation between negative affectivity and alexithymia,

negative affectivity did not appear to affect the accuracy of response to emotional stimuli
in either time frame in Study 2. The fact that alexithymics were differentially affected by
temporal constraint in Study 2 but not in Study 1 lends support to the argument that the

I

i
II

71

deficit in emotion processing occurs after a stimulus has been attended to, and is specific
to perceptual mechanisms of emotion recognition, rather than attention mechanisms.

In

other words, alexithymics were not impaired in perceiving a difference in the emotion
condition in Study 1. In Study 2, they were, however, impaired in emotion recognition,
which is a later stage of information processing.
In these studies, alexithymia and negative affectivity do appear to be separate

constructs with different underlying neural mechanisms. The differences between the two
constructs have distinct behavioural implications. Results from a later regression analysis
of these data in which the dependent variables were the respective sensitivity measures of
each study and the explanatory variables were phase order, alexithymia, and both positive
and negative affectivity, mostly confirmed the original ANOVA analyses. An interesting
addition to the results was a significant effect of negative affectivity in Study 1 for
accuracy in perceiving facial differences in the emotion condition. The results from Study
1 suggest that, rather than measuring recognition of emotion versus identity, this study
was measuring the ability to compare and perceive differences between facial features, or
attend to facial features. The results from Study 1 appear to suggest that emotion
processing is more

~

e

than identity processing.

In the light of these studies and of previous research suggesting separate neural

mechanisms in processing emotion and identity recognition at higher levels, it seems to
make sense to consider that negative affectivity is significant in the attention stage of
emotion information processing, while alexithymia is significant at the perception and
recognition stage of emotion processing. Higher levels of negative affectivity may result

72

in slower shifting to stimuli, and slower disengaging from, or neglecting of stimuli. In
other words, negative affectivity may affect our ability to shift attention from one stimulus
to another, i.e. back and forth. Negative affectivity may therefore have a diffuse and
generalized impact on information processing, and on emotion processing in particular,
since this impact occurs earlier in the processing of sensory information. The demands of
Study 1 were that the subject had to scan two faces within one second and compare the
faces. In other words, the subjects had to be able to shift their attention from one
stimulus to the other under temporal constraint. This study therefore measured ability to
perceive facial differences and neglect, disengage, shift, and engage attention. However,
there is nothing in the results to suggest that negative affectivity affects our ability to
identify, recognize, or discriminate among emotions. On the other hand, while
alexithymia may not be a crucial factor at the attention stage of emotion processing, it
appears to be a significant factor in the perception and cognition stage of emotion
processmg.
Thus, the results of Study 1 appear to indicate that while alexithymia is not
implicated in shifting attention from one stimulus to the other, negative affectivity may be.
This is an important distinction, since it implies that NA may have a more generalized
effect and that the effects of alexithymia may be specific to emotion recognition. It also
appears that for alexithymia to exist, there may have to be high levels of negative
affectivity, but that conversely, high levels of negative affectivity do not imply alexithymia.
It would be interesting to compare subjects who are high in negative affectivity and low in

alexithymia with subjects high in negative affectivity and high in alexithymia. The

73

descriptive statistics from this subject pool indicated that very few subjects tended to be
low in negative affectivity and high in alexithymia, and only seven subjects were high in
positive affectivity and high in alexithymia. These numbers again lend support to the
speculation that negative affectivity may be a precursor to alexithymia, and may affect our
flexibility in shifting attention, while positive affectivity may be a protective factor for
alexithymia and may aid flexibility in attention mechanisms.
Negative affectivity appears to have an impact on our ability to attend to an
emotional stimulus, be it shifting, neglecting, engaging or disengaging attention.
Alexithymia does not appear to be implicated in our ability to attend to a stimulus. It
does, however, appear to be implicated in our ability to recognize emotions and
discriminate among them.

5.6 The Role of Language in Alexithymia
Alexithymia implies 'no words for feelings', yet it has been unclear from the
literature as to the exact nature of the problem underlying the construct. Although there
are differences between alexithymics and non-alexithymics in emotion processing,
alexithymics are able to detect, attend to, perceive, recognize, label and express emotion.
They appear to be somewhat impaired emotionally, yet their impairment is subtle
(McDonald & Prkachin, 1990; Parker, Taylor, & Bagby, 1993b). The construct has been
clouded by a vague idea that something is amiss; however, it has been difficult to pinpoint
the problem, investigate the underlying neural mechanisms, or understand the ensuing
behavioural ramifications. Results from Study 2 would indicate that this ability to have
words for and describe an emotion might play a pivotal role for emotion recognition in the

74

construct of alexithymia, if the situation demands that subjects respond quickly and under
pressure.
The concept underlying Factor II or Difficulty in describing Feelings, that of
labelling emotion or having words for emotions, is reminiscent of the work by Lane and
Schwartz (1987). They put forward the idea that emotional awareness, descriptive
accuracy and ability to label emotions may be protective factors in psychopathology. The
suggestion was made that there are five levels of emotional awareness, ranging from the
crudest to the most sophisticated form. The crudest form, the 'sensorimotor reflexive'
would allow a person the subjective awareness of bodily sensations. The second level,
described as 'sensorimotor enactive', allows for action and arousal. The third level,
'preoperational', implies that a person has a limited experience of emotional extremes.
The fourth level, or 'concrete operational', implies that a person can differentiate among
emotions, and experiences blending of emotion. The fifth and most sophisticated level,
'formal operational' , allows a person to describe complex and differentiated qualities and
intensities of subjective emotion. If, as is proposed in these studies, alexithymics 'fasttrack' emotional information in certain situations, this would imply that their emotion
processing is somewhat crude. Alexithymics are also described as confusing bodily
sensations with subjective feelings. The crude 'fast-track' processing and emphasis on

!

bodily sensations both suggest the 'sensorimotor reflexive' or lowest level of emotional
awareness as described by Lane and Schwartz (1987). However, this suggestion of
cruder processing in alexithymia applies only to emotion stimuli.

75

5. 7 Blindsight and Alexithymia
The conclusions that can be drawn from the results of Study 2 are also reminiscent
of the concept of blindsight (Weiskrantz, 1986). Blindsight is implicated in situations
where information may be implicitly processed, yet be outside conscious awareness, and
therefore not available for verbal articulation. Blindsight implies that a person may not be
consciously aware of a stimulus. However, there is an implicit knowledge of the stimulus,
but this knowledge is not available for conscious processing. In other words, the person
may not be able to describe or identify the stimulus to themselves or others, but they may
have an inkling that something is awry, yet be unable to know the cause of their
discomfort.
Blindsight is not an easy concept to understand, although it is a fascinating
phenomenon.

It requires a certain level of interest in, and knowledge of, neural

mechanisms. It also requires an acceptance of the concept of parallel processing of
information on both the implicit and explicit levels of awareness.

Blindsight may be

experienced in a number of ways, depending on the neurological damage or functional
impairment. Of relevance to alexithymia is the notion that there is a language attention
system, in which the lateral frontal lobe, the inferor parietal lobule (IPL) and the anterior
cingulate appear to be implicated (Posner & Petersen, 1990). The right anterior cingulate
is also implicated in higher levels of emotion processing (George et al.,1993). The
cingulate cortex is connected to the IPL: this latter area appears to be an "integrative
centre, synthesizing information from exteroceptive sensory systems, limbic and
viscerosensory areas, and from arousal systems" (p. 354) and acts to mediate conscious

76

feelings (Ledoux, 1986). Any evaluation of incoming sensory events which do not reach
the IPL remains implicit or unconscious. In this event emotional expression is automatic
or behavioural and without conscious awareness. A second attempt to process this
already crudely processed information may be made via 'reentrant processing' ; it is
thought that "higher level associations are made by fibers that reenter the brain areas
which processed the initial input" (Posner, 1994, p.7402). However, Ledoux (1986)
suggests that the subsequent evaluation of the stimulus event is made by cognitive, rather
than emotional processing systems, and may be wrong, since it is based on imprecise or
false assumptions or inferences. The underlying principle is the same as in the concept of
'fast-tracking', except that the latter refers to amygdala processing and is associated with
the emotion of fear (Ledoux, 1994).

5.8 Implications of Blindsight for Health
We may assume from the present studies that alexithymics do process emotional
information, but that this processing may not reach conscious awareness in certain
circumstances which are perceived as threatening, distressing, or uncomfortable.
Emotions, therefore, may not be fully processed or cognized by alexithymics. In studies
investigating the relationship between emotion disclosure and health it has been suggested
that the ability to write about trauma and describe the feelings or experiences connected
with the trauma, may act as a protective factor in health (Pennebaker & Hoover, 1983;
Pennebaker, Hughes, & O'Heeron, 1988; Pennebaker, Kiecolt-Glaser, & Glaser, 1988;
Petrie, Booth, Pennebaker, & Davison, & Thomas, 1995). In the light of these studies,
alexithymics then, would be at a disadvantage, since much of their emotion processing

77

does not appear to be conscious, and therefore may not be available for disclosure.
Alexithymics, therefore, may be more at risk for certain health related problems. The
findings of Study 2 imply that alexithymics may be processing certain emotional stimuli in
a crude form given temporal constraint, contributing perhaps to the lack of differentiation
of distress suggested by Bagby, Taylor, and Parker (1994). Low and moderate
alexithymics, for whom the details of the stimulus are understood, may process these
stimuli in a less crude form when faced with excessive demands. For high alexithymics,
this crude processing may be a chronic condition given that many expressions of negative
emotion are fleeting, and the processing of these fleeting emotional stimuli may therefore
place excessive demands on high alexithymics. This is consistent with the suggestion that
alexithymics have a deficit in the spontaneous encoding of negative emotions (McDonald
& Prkachin, 1990).

Alexithymics may habitually process sadness, anger and fear imprecisely. How
then, are they to deal with the fact that often they may be sensing unease or fear, but that
they do not have an explanation for this sensation, nor are they able to discriminate among
the feelings? Whilst they may know on the one hand that something is amiss, they may,
on the other hand, be unable to put the feeling into words, communicate it precisely to
others, or understand what is really happening to them. This feeling of knowing without
knowing is an experience which most people will have had now and again. However,
when it is a chronic condition and occurs habitually in given situations, it may become a
problem with ensuing ramifications. In other words, while most people experience the
inability to put feelings into words occasionally, for high alexithymics it may be a chronic

78

condition. A chronic inability to put feelings into words may lead to somatic problems
(Paez, Basabe, Valdoseda, Velasco, & Iraurgi, 1995). Perhaps this crudely processed and
incomplete information is subsequently processed via 'reentrant processing'.

The

confusion over understanding the original sensation or experience may lead to a
subsequent psychosomatic explanation. In other words, a low level of emotional
awareness and crude processing of the meaning of the stimulus could be subsequently, but
falsely, interpreted as a somatic sensation.
Difficulty in describing emotions seems to be a pivotal factor in the impairment of
emotion recognition in alexithymia. Therefore, the key to understanding the reason for
alexithymics being potentially at risk for psychosomatic and psychiatric disorders may be
found by investigating the neural mechanisms underlying alexithymics' difficulty in
describing feelings. The area of focus should be the interaction of neural mechanisms
responsible for both the 'fast-tracking' of information and the language attention system,
with its putative areas: the lateral frontal lobe, the IPL, and the anterior cingulate. It
seems plausible to consider that the inability to discriminate amongst negative stimuli
when in a distressing situation leaves a person at risk for misinterpreting the situation.
Conversely, it seems plausible to consider that the more accurately we label and
discriminate amongst negative stimuli, the better able we are to accurately interpret the
situation, and the feelings attached to the situation. A chronic inability to discriminate
among negative sensory stimuli could have repercussions in many areas, which is perhaps
the reason for the association of alexithymia with many other psychopathological
disorders. It also seems plausible to consider that levels of emotional awareness and

79

emotional verbal acuity are bound together with the concept of consciousness, in the guise
of AK and DK.

Both the concept of levels of emotional awareness and the concept of

blindsight appear relevant to the construct of alexithymia. It seems reasonable to
speculate that an inability in verbalizing emotions and a form of blindsight in this area may
leave a person at risk for health and well-being (Pennebaker, Hughes , & O' Heeron, 1987).

5.9 Differences in the Recognition of Sadness, Anger and Fear
Study 2 was limited to assessing whether alexithymics differed from nonalexithymics in their ability to perceive whether an emotion was present or not. In other
words, the subjects in this study were not asked to differentiate between, or accurately
label, the separate emotions.

However, an unexpected finding in Study 2 was the

significant difference in the ability to accurately recognize the presence of individual
emotions; sadness, anger and fear.

This finding implies an inherent and implicit ability

to internally attend to, perceive, recognize and discriminate among the three separate
emotions. Given the theories surrounding the language attention system, the concept of
blindsight and of 'fast-tracking' or crudely processing stimulus information, this finding
may imply an inherent ability to implicitly label these emotions, or it may indicate an
explicit impairment in labelling emotions. The overall recognition of sadness was
significantly less accurate than that of anger, and the recognition of anger was significantly
less accurate than that of fear. Only the recognition of sadness appeared to be
significantly impaired in the one second time frame, when compared to the three second
time frame, although there was a trend towards increasing accuracy in the recognition of
anger, and then of fear.

80

Alexithymics are described as having outbursts of anger and sobbing, but not as
having moments of intense fear. Confusion over anger and sadness, therefore, seems
relevant to alexithymia. The high alexithymic group was significantly less accurate in
their recognition of emotions than the low and moderate group. Subjects who had the
most difficulty in describing feelings were the most impaired in their overall recognition of
sad, angry and fearful stimuli in the one second time interval, whilst their responses were
comparable to the low and moderate groups in the slower condition (see Figure 4.1).
Subjects who had the most difficulty in identifying feelings, or who belonged to the high
Factor I group, were significantly more impaired as a group in their ability to accurately
recognize sadness in the one second, as opposed to the three second, time interval (see
Figure 4.3). Factor I, or Difficulty in Identifying Feelings, is comprised of questions
which measure confusion over feelings and the ability to discriminate between physical and
emotional sensations (see Appendix A).

5.10 Theoretical Implications of Differences between Emotions
What are some of the more general implications that can be inferred from finding
differences in the recognition of individual emotions? The data from this study suggest
that fear is the most accurately processed of the three negative emotions presented in this
study in both temporal conditions, and that this accuracy is uniform for all subjects and all
alexithymic groups. The greatest impairment in accurately processing an emotion was in
the facial expression of sadness, and this impairment was evident in all groups. It was
most evident for those with the most difficulty in describing feelings, and most affected by
the condition of temporal constraint for those with the greatest difficulty in identifying

81

feelings. One possible explanation for the difference in accurate recognition of the three
emotions may lie in their specific physiognomy (Ekman, 1977). Facial expressions of
anger and fear as shown in Ekman and Friesen's Pictures of Facial Mfect (1976) tend to
involve larger numbers of physiognomic changes than facial expressions of sadness.
Facial expressions of fear involve the raising of the inner and the outer brow, the lowering
of the brow, the raising of the upper eyelid, the tightening of the lower eyelid and the
horizontal stretching of the lips. Facial expressions of anger involve the tightening of the
lower eyelid, the raising of the upper eyelid, chin raising and lip tightening. Facial
expressions of sadness involve the raising of the inner and outer brow and lip corner
depressing. This quantitative element may contribute to differences in recognition
accuracy for the three emotions.
From an evolutionary standpoint one would expect there to be greater accuracy in
recognizing fear, since it is often externally and contextually driven, and therefore crucial
to survival, whilst sadness tends to be perceived internally and less contextually. Sadness
may be less important from an evolutionary perspective, although it may play an
important role in our health and well-being.

Psychiatric and psychosomatic problems

may arise from a generalized lack of differentiation of negative emotions, or they may arise
from an impairment in the perception of sadness rather than from an impairment in the
perception of fear. Does this then imply that 'fast-tracking', in the case of alexithymics,
has more to do with distress over the inability to recognize sadness? Fear, after all, was
the most accurately recognized of the emotions in these studies, and therefore the least
likely, according to the present analysis, to be 'fast-tracked ' or crudely processed. 'Fast-

82

tracking' may simply be a phenomenon which occurs in moments of distress, and when
excessive demands are placed on a person. It may not be specific to 'fear' processing, but
may apply to any situation deemed distressing for alexithymics.
One interesting inference is that alexithymia may be largely due to an impairment in
the recognition of sadness. Sadness in itself may well be an emotion which has been
overlooked by researchers.

Anger and fear seem more socially acceptable emotions,

while sadness is shunned and often treated with contempt. For instance, it has been
unacceptable, in our culture, for men, or even boys, to cry. Alexithymia is associated
with many psychological and psychiatric disorders, and if sadness, or an inability to
accurately recognize sadness is also associated with alexithymia (see Figure 4.3), this
finding suggests that more attention should be paid to the inter and intrapersonal
experience of sadness, and perhaps even to the necessity of this emotion, and its
importance in the area of health.

5.11 Alexithymia and the Development of Emotions
Another possible explanation for differences in the recognition of emotions may be
found in the literature investigating the development of emotions. According to Lewis
(1995) the expression of sadness may be seen on an infant's face at three months. Facial
expressions of anger usually appear at four to six months, while facial expressions of fear
are not expressed until the infant is seven to eight months old. The initial facial
expressions of fear seen on an infant's face coincide with the timing of increased activity in
the prefrontal cortex: an area of the brain associated with discriminating among
threatening stimuli in humans (Kalin, 1993). Somewhere between seven and twelve

83

months of age children begin to gauge their level of fear by interpreting expressions seen
on a parent's face: a form of 'social referencing' and 'self-regulation'.

Since

alexithymics appear to be able to recognize facial expressions of fear as accurately as nonalexithymics, the impairment seems to be due to a lack of accuracy in recognizing sadness,
and to a lesser extent, anger, when in a demanding situation. This suggests that the
impairment or deficit for primary alexithymia may originate in the first half year of an
infant's life, and leave them at risk for an inability to differentiate quickly among anger,
sadness and fear.
Since alexithymics' recognition of these facial expressions of emotion appears
normal when there is no temporal constraint, the deficit is unlikely to be structural. It
seems more likely that alexithymia is due to a functional impairment: an impairment in the
sense that the infants are interpreting a non-threatening stimulus as an unknown and
therefore distressing stimulus and then overcompensating by 'fast-tracking' the
information. This may happen at a time when the hippocampus has yet to fully mature
and the infant's memory is more dependent on the amygdala. There is evidence to
suggest that the amygdala matures at an earlier stage of development than the
hippocampus (Ledoux, 1995). Any memories relying on the amygdala circuit may well be
inaccessible to conscious processing. The hippocampus appears to provide contextual
information regarding a stimulus, while the amygdala may assign its 'emotional'
significance, irrespective of the method of processing and the details of the stimulus. It
seems likely, then, that there is a developmental window in which an interaction of late
hippocampal activity, high anxiety, and a traumatic event associated with perceived

84

negative affectivity may occur and leave a person at risk for alexithyrnia. It also seems
likely that high anxious infants are predisposed towards alexithymia. However, not all
high anxious infants are at risk, nor will they develop alexithymia. Only prospective
studies could confirm this theory and disentangle the various categories of anxiety and
how they may, or may not, interact with alexithymia.
5.12 Limitations of Studies 1, 2, and 3
The results of these studies should be generalizable to a normal adult population
with a similar cultural background. Although the subjects in these studies belonged to a
student population, the University of Northern British Columbia tends to have a higher
enrollment of mature students, with general workforce experience, than other universities.
This fact is reflected in the higher mean age (M=26.5) of the subjects than is typical in
university samples. The results are only generalizable within the context of these studies.
Alexithyrnia seems to have developmental significance. However, these results are
not generalizable to children, and if children were to be measured there would have to be
changes to the methodology. It seems likely that children may be slower in processing
emotional information. The TAS-20 seems somewhat adult in its orientation and may be
unsuitable for children. These studies addressed the concept of primary alexithymia. No
personal histories were taken from the subjects and it is unknown as to whether the high
scores could be attributable to secondary alexithymia. Secondary alexithyrnia may be due
to delayed development resulting from some traumatic event in a person's childhood.
It was an arbitrary decision to use the tercile split for the TAS-20 scores and a

median split for the PANAS scores. It could be argued that the PANAS scores should be

85

treated similarly to the TAS-20 scores. One potential limitation of this study was the fact
that the TAS-20 and PANAS scales are self-report measures, and that the subjects were
completing these scales without complete anonymity.

It may be wise to use an additional

measure, such as the Levels of Emotional Awareness Scale (Lane, Quinlan, Schwartz,
Walker, & Zeitlin, 1990) in order to assess emotional insight.
The method of presenting the stimuli in Study 1 and 2 in a file folder was efficient,
if somewhat primitive. A more hi-tech approach: using video and editing equipment,
with a computer based printout for the responses, might result in different or more
pronounced results. A more precise presentation of the stimuli on video for 1:30th of a
second exposure, followed by a one or three second interval for the response might result
in greater between group differences. In the present studies the responses were manually
recorded by the experimenter. The vast majority of responses were obviously within the
time frame and were clearly responses to the numbered stimuli. Occasionally the
experimenter had to make a subjective judgement call for late responders . This
subjectivity could be eliminated if the subjects responded by pressing a button connected
to a computer. There are positive and negative aspects to both the high and low
technological approaches. For an exploratory study, the methodology used in these
experiments seems adequate. It should be borne in mind that it may not be possible to
generalize from the laboratory results, using photographs, to the real world.

86

CHAPTER6
Summary, Conclusions, and Directions for Future Research

The experiments undertaken in Study 1 and Study 2 were exploratory in nature.
The hypotheses originated from three different perspectives. One perspective was
behavioural; simply realizing from observation that people do seem to vary in their ability
to verbalize how they are feeling. Another perspective was neurological; considering
brain structures and functions, and how small, seemingly insignificant variations can have
important repercussions and interaction effects. The final perspective came from drawing
the behavioural and neurological perspectives together to find a construct which seemed
to fit the observed behaviour and hypothesizing about the neural mechanisms which might
be implicated in the construct.

6.1 Theoretical Implications of Study 1 and Ideas for Future Research
In Study 1 it was expected that a difference would be found between alexithymics
and non-alexithymics in their ability to attend to identity versus emotion stimuli.

While

there were no significant group differences, there was a significant difference in comparing
facial expressions of emotion and facial identity. This result implies there is a difference
in information processing between emotion stimuli and identity stimuli: this needs to be
further investigated. Future research might also consider investigating identity stimuli
versus separate emotion stimuli. While Study 1 appears to provide evidence for an
overall difference in accuracy while attending to emotion versus identity stimuli, in light of
the results of Study 2 it may be important to test this difference in accuracy while

87

attending to either sadness, or fear or anger versus identity stimuli. In other words, whilst
there appears to be an overall tendency to respond more accurately given temporal
constraint to identity versus the combined emotions of sadness, anger and fear, this
difference in accuracy may not hold when identity stimuli are compared with facial
expressions of fear stimuli. Facial expressions of fear were the most accurately recognized
of the three negative emotions presented in Study 2. We may attend more accurately to
facial expressions of fear than to those of sadness, given temporal constraint, and the
difference in accuracy between fear and identity stimuli may no longer be significant.
Several subjects raised the possibility that gender may also be a factor in emotion
processing. Is it easier for men to process male facial expressions of emotion and women
female expressions of emotion?

A question that was raised by both subjects and the

experimenter concerned the neutrality of neutral facial expressions. How neutral is
neutral? The results indicate that neutral expressions were recognized as such, but the
accuracy in recognizing these 'noise' images may have been influenced by context. How
much does context, or the surrounding images, affect the processing of each individual
image, and in particular, on neutral faces? It has been suggested that context plays a role
in the recognition of emotions but that the relationship between emotion and context is
complex (Fernandez-Dols, Sierra, & Ruiz-Belda, 1993), and that ratings of neutral faces
vary according to the situational cues presented (Carrera-Levillain & Fernandez-Dols,
1994). Neutral expressions are, for the most part, assessed as neutral. However, they do
contain affective meaning, and this meaning appears to have the most influence when the
facial expression is seen in a clear, but unusual context. In uncommon situations, it seems

88

that neutral faces can have as much influence in understanding the situation as prototypical
expressions of basic emotions. It seems unlikely that Study 1 and 2 provided an
uncommon context, but this possibility could be explored in future research. These
studies only addressed a limited context, and future research is needed to disentangle the
contextual implications.

6.2 Theoretical Implications of Study 2 and Ideas for Future Research
In Study 2 it was expected that alexithymics would be more impaired in processing
emotion stimuli when excessive demands, in the form of temporal constraint, were placed
on them. This hypothesis was confirmed, and it was found that the impairment appears to
be associated with difficulty in describing feelings. Future research may investigate the
neural mechanisms underlying language attention systems, in particular the anterior
cingulate, the IPL and the lateral frontal lobe, and how stressors may impact on these
mechanisms. Future research may further examine the differences in emotion processing
of sadness, anger and fear.
Alexithymia is conceptually associated with language. Results from Study 2
indicate that accuracy in recognizing emotion is significantly related to our ability to
describe emotions. In previous research the suggestion has been put forward that the
right anterior cingulate and bilateral inferior frontal gyri are implicated in emotion
recognition in higher order processing (George et al., 1993). The anterior cingulate is
also implicated in the language attention system (Posner & Petersen, 1990). The Inferior
Parietal Lobule, with its connections to the frontal gyri and the anterior cingulate, is
suggested to be the area of "mediation for conscious feelings" (Ledoux, 1986, p.354).

89

Both the neurological evidence and the results from Study 2 appear to lend support to the
argument that alexithymia is implicated in higher order processing, and it appears to be
linked to language, or the ability to put feelings into words. On the other hand, there is
nothing to suggest that negative affectivity is directly linked with this stage of processing.
The hypothesis that higher levels of alexithymia as measured by the TAS-20 would
result in less accurate processing of emotion stimuli was confirmed.

Most importantly,

the results of these studies helped to disentangle the alexithymia and negative affectivity
constructs and provide evidence that alexithymia not only is a viable construct, but also
has particular behavioural ramifications which are not necessarily shared by negative
affectivity. Future researchers may wish to further disentangle NA and alexithymia.
What is it specifically in NA that potentially triggers alexithymia? The PANAS (see
Appendix B) measures various feeling states associated with NA. It is possible that a
particular state, i.e. nervousness or distress, might contribute to alexithymia.

Since

alexithymics do appear to be 'fast tracking' information when under constraint, resulting
in cruder processing, one investigative approach might be to measure cortisol levels, since
cortisol may contribute to the 'vigilant' and alert state suspected in alexithymia. It would
be interesting to find out, in future research, what alexithymics do actually experience,
both subjectively and physiologically, when they view negative emotion stimuli, and
whether their experience is different for verbal and non-verbal stimuli.
In these studies the emphasis was placed on comparing negative affectivity with
alexithymia. Future research may compare alexithymia and positive affectivity, with the
hypothesis that positive affectivity may be a protective factor in emotion processing, and

90

also a protective factor in health and well-being. High levels of positive affectivity may
allow for normal or quicker shifting to emotion stimuli. It has been suggested that
alexithymia is associated with low levels of positive family communication (Berenbaum &
James, 1994).
Another area of future research may be to replicate these studies but with auditory,
rather than visual stimuli. The inference in these studies is that alexithymia is associated
with a functional impairment in the processing of visual emotional stimuli. It is possible
that auditory emotional stimuli processing may have similar impairments. The concept of
'fast-tracking' has been implicated in the auditory system. It is not known how it
functions in the visual system.

It would be interesting to investigate how tone, context

and facial expressions of emotion interact.

6.3 Theoretical Implications of Study 3 and Ideas for Future Research
An observation made by the experimenter during the administration of Study 3, was
that some subjects tended to laugh easily, some appeared unaffected, and some seemed
tense whilst doing the Stroop Task. Future research might consider administering a
humour questionnaire, such as the Situational Humour Response Questionnaire (Martin &
Lefcourt, 1984) in order to shed light on the effect of humour on tension, and indirectly,
on health issues such as cardiovascular disease.

6.4 Future Research for the 'Fast-tracking' of Stimuli
'Fast-tracking' has been described as a response to threatening stimuli, and has been
associated with the processing of fear (Ledoux, 1993). Although it is tempting to assume
that the temporal constraint condition of Study 2 is perceived as threatening by the high

91

alexithymic group, this is an assumption. Perhaps future research could provide evidence
supporting that assumption. Or, conversely, future research might pinpoint the exact
nature of 'fast-tracking' and its situational cues. This study seems to lend support for the
condition of temporal constraint as a situational cue for 'fast-tracking'. In an experiment
similar to Study 2, facial expressions of emotion could be presented in rapid succession in
one condition, and slower in the other condition. Alexithymics could be asked to indicate
whether an emotion is present or not, whether it is similar to the previous exposure, and
whether they liked the expression more or less than the previous exposure. If
alexithymics cannot discriminate between exposures, but they indicate a dislike for the
facial expressions of sadness, anger and fear, this would indicate that they are recognizing
the threat level but not the emotion itself.
Similarly, in a second experiment, alexithymics could be exposed to a rapid
succession of neutral, sad, angry and fearful faces, and asked to state whether an emotion
is present or not, with the expectation that their responses would be less accurate than
non-alexithymics for the sad, angry and fearful faces. Subsequently, they could be
exposed to a rapid succession of facial expressions of emotion on the same faces and
asked to state which faces they recognize. Following the argument put forward by
Zajonc (1980) that memory recognition is intensified by affect, and is better for objects
liked or disliked, it would be expected that alexithymics would have a higher recognition
rate for the angry, sad and fearful faces than the non-alexithymics.

Zajonc (1980) also

argued that people tend to form preferences for things they cannot identify. This
proposed experiment could provide evidence suggesting that, while alexithymics are less

92

accurate in consciously discriminating among emotions, confusing fearful, angry and sad
faces, they do implicitly feel those emotions, and experience them as threatening given
temporal constraint.
This idea that the emotion is somehow subjectively experienced when objectively
processed is reminiscent of 'knowledge by acquaintance' (AK), or the concept of qualia
(Buck, 1993). It also implies that both interoceptive and exteroceptive processing of an
emotion share common neural mechanisms. Emotional experiences and the perception of
expressions of others may be processed at the same time, in the same place or in the same
systems. In other words, if a person witnesses a fearful expression, they also experience
that emotion, albeit implicitly, within themselves, suggesting that the internal experience
and the external recognition are processed simultaneously and together. AK does not
imply any knowledge about the stimulus or 'knowledge by description' (DK). DK
requires an understanding of the experience, the ability to label the experience, and to put
it into words (Buck, 1993). It seems likely, and these proposed experiments might lend
support for the hypothesis, that alexithymics experience AK, but fail to experience DK in
the area of emotion recognition.

6.5 The hnportance of Time in Emotion Processing
Another important aspect of Study 1 and Study 2 lies in the methodology itself.
Presenting subjects with fairly straightforward stimuli and manipulating the temporal
condition is a novel approach to the investigation of emotion processing.

The results

from using this methodology do seem to suggest that time is an important factor in the
processing of emotions, and the manipulation of time may shed light on impairments in

93

neural processing. In hindsight it may seem obvious that temporal constraints will lead to
overall differences in the perception of emotion, and also to group differences in emotion
processing. The idea for manipulating time came from an interest in Ledoux's work on
'fast-tracking' threatening information and also from an interest in the behavioural
problems affecting autistic children.
Autism appears to be an acute disorder due to the lack of integration, and
potentially lack of temporal coordination, of sensory stimuli. Although most clinicians
would see the two disorders, autism and alexithymia, as categorically different, there
appears to be a certain commonality between them. One commonality is that there seems
to be a need for sufficient processing time. Put another way, autists and alexithyrnics do
not appear to tolerate temporal constraint on sensory processing in some situations.
While the impairment in autism is likely found in the attention stage of information
processing (Courchesne et al., 1994; Ross, Radant, Young, & Bommer, 1994;
Wainwright-Sharp & Bryson, 1993), and an impairment in an earlier information
processing stage might imply a more acute and diffuse disorder, alexithyrnia appears to be
a specific problem associated with higher order processing in the recognition of emotions
at a later stage of sensory information processing.

6.6 Temporal Processing and Clinical Populations
Alexithymics as such cannot be differentiated from non-alexithyrnics by their
outward appearance, or by their behaviour for the most part. It is only when excessive
demands or certain situational constraints are in place that difficulties in emotion
processing become obvious. Alexithymics then appear to become more distressed and to

94

experience greater interference with sensory processing.

It seems likely that many

psychiatric and psychological disorders share common deficits in processing sensory
information when presented with some form of temporal constraint, and this method of
experimentation could be used with other clinical populations. The subjects in these
studies were all part of a normal population and the effects seen withfu these studies are
among people who differ largely within a normal range. However, the ramifications of
this impairment are far reaching and may extend beyond the known associated disorders to
other groups of people or clinical categories, i.e. the prison population, sexual and
physical abusers. It may be important to compare alexithymics with similar, but different,
groups. Internalizers, inhibitors, and repressors could be compared with alexithymic
groups using the methodology from Study 2.

6. 7 The Possible Primacy of Affect
It seems likely that affect, rather than cognition, may play a primary, if implicit role,
in behavioural responses of certain clinical populations with specific disorders, given
temporal or other situational constraints. The challenge will be to provide therapy for
disorders where primacy of emotion is suspected in situations of temporal constraint. It
will require teaching people to translate AK into DK in order to halt the 'fast-tracking' of
distressing sensory information. The challenge will be to enable these people to confront
the distressing stimuli, and, in a sense, to relax, in order to learn to slow down their
processing. Then they may be able to catch the details of the stimulus in the first round of
processing, and be able to know and understand it, and be in a position to describe
whatever it happens to be. Although it is tempting to view primary and secondary

95

alexithymia, such as PTSD, as one and the same disorder, this may be an erroneous
assumption. Future research is needed to investigate, and potentially disentangle, these
two disorders. While PTSD clients appear to behave similarly to alexithymics, they may,
however, have more in common with repressors, and therefore warrant a different mode
of therapy. In order to devise therapeutic treatment for alexithymics, there seems to be a
need to pinpoint the interactions between the neural mechanisms responsible for language
attention systems, the processing of facial expressions of emotion and 'fast-tracking'
mechanisms. It may be useful to consider habituation and extinction rates for
alexithymics with the expectation that for certain stimuli, extinction may be eliminated.
Facial expressions of emotion other than sadness, anger and fear may be examined using
this methodology in the context of alexithymia.

6.8 Developmental Implications of Alexithymia
The evidence from these studies only provides support for the hypothesis that
alexithymics are less accurate in recognizing emotion given temporal constraint. Future
research could look at the ability of alexithymics to label or describe emotions both with
and without temporal constraint. It would be expected that alexithymics have the most
difficulty in labelling emotions given temporal constraint, and high alexithymics may
experience greater difficulty than low alexithymics in labelling emotions in a slower
condition. It would be interesting to find out whether they can learn to correctly label
emotions, or whether this ability, similar to other linguistic abilities, is lost once the critical
learning phase is passed. If this were the case, if the ability to label emotions is lost after
a certain age, it would become important to assess children early for alexithymic traits,

96
with a view to helping them discern differences between the emotions. A scale designed
to measure alexithymic traits in children would be needed for early assessment.
A closer look at the association between psychosomatic illness, difficulty in
describing feelings, sadness and alexithymia may be warranted. Being able to recognize
an emotion, label it, and simultaneously experience it, may be the key to understanding the
association between alexithymia and psychosomatic disorders.

6.9 Disentangling the TAS-20 Factors
However, we cannot assume that difficulty in describing feelings is the sole reason
for emotional impairment in alexithymia, or the key factor in understanding the link
between alexithymia and other problems. In the pilot study conducted at UNBC (Parker
& Prkachin, 1995), several measures were correlated with the TAS-20. One of those

measures, the Psychosomatic Symptom Checklist or SUNYA had the highest correlation
with Factor I or difficulty in identifying feelings . This correlation suggests that the ability
to identify feelings is a significant factor in psychosomatic disorders and alexithymia.
Factor II, or Difficulty in Describing Feelings, appeared to correlate most highly with
measures of negative affectivity, personal distress and lack of social support. Although it
might be tempting to narrow the measure of alexithymia by eliminating extraneous
questions or factors from the TAS-20, we do not as yet know how these factors impinge
on one another, and interrelate.

It would appear from the literature (Linden, Wen, &

Paulhus, 1995), and from a factor analysis conducted at UNBC in conjunction with the
pilot study (Parker, 1995), that some revision of the TAS-20 scale might be warranted.

97

The most significant difference in recognizing sadness in the one and three second
intervals was found in the group experiencing the most difficulty in identifying feelings .
Perhaps Factor I helps to pinpoint the key problem of this disorder: an inability to
accurately identify sadness, and to a lesser extent, anger. Factor II, or Difficulty in
describing feelings may pinpoint the importance of being able to verbalize emotions and
communicate feelings to others. Investigation of Factor III, or Externally Oriented
Thinking, may shed light on associated behaviours and may act as a marker for people at
risk for psychiatric and psychosomatic disorders associated with alexithymia. What is
fascinating about this construct is the fact that it is possible to investigate it at each
factorial level. These studies unwittingly focussed on Factor II, or Difficulty in
Describing Feelings, and shed some light on Factor I, or Difficulty in Identifying Feelings.

6.10 Summary
It seems to have been beneficial to run the three studies in conjunction with one
another, since each study focused on a specific stage or type of processing.

This allowed

the experimenter the opportunity to disentangle some of these processes and their
interaction with alexithymia and negative affectivity. Study 1 focussed on attention
processing and the ability to attend to, neglect, disengage from and engage attention.
Study 2 focussed on the recognition of emotion processing. Study 3 shed some light on
the role of arousal in alexithymia. It seems that negative affectivity may be important in
attention processing, while alexithymia appears to be a deficit in emotion recognition. It
seems likely that heightened arousal and distress may be associated with alexithymia.

98

Negative affectivity may interact with, and contribute to alexithymia but high NA may not
necessarily result in alexithymia.
The object of these studies was to investigate the construct of alexithymia in the
context of emotion processing. One important hypothesis has been largely confirmed,
and due to the novelty of the methodology and the fact that the research was primarily
exploratory in nature, the results of these studies have raised an enormous amount of
expected and unexpected questions which may be addressed in future research. One
unexpected finding was the difference in recognition for the separate emotions.
The data from this study lend support for the construct of alexithymia, which
appears to be a continuum or personality trait, with people at the high end experiencing
the greatest difficulty in discriminating among various negative emotions given temporal
constraint. Alexithymia does appear to be a separate construct from negative affectivity.
Alexithymics do appear to be 'fast-tracking' or crudely processing emotional information
given temporal constraint.
Although most behavioural differences were found in the group most impaired in
describing feelings, the demands of these studies were too limited to negate the validity of
the other two TAS-20 factors in emotion processing. As already suggested in the
introduction, these factors may complement one another and also reflect different stages
of information processing. Clearly alexithymia is an important construct: important in
shedding light on emotion processing as a whole, on consciousness and awareness, and on
the interaction of affect, cognition and behaviour.

99
References

Attanasio, V., Andrasik, F., Blanchard, E.B., & Arena, J.G. (1984). Psychometric
properties of the SUNYA revision of the Psychosomatic Symptom Checklist. Journal of
Behavioural Medicine, 1, 247-257.
Bagby, R.M., Parker, J.D.A. , & Taylor, G.J. (1994). The Twenty-item Toronto
Alexithymia Scale - 1. 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 twenty-item Toronto
Alexithymia Scale-II. Convergent, discriminant, and concurrent validity. Journal of
Psychosomatic Research, 38, 33-40.
Berenbaum, H. , & James, T. (1994). Correlates and retrospectively reported
antecedents of alexithymia. Psychosomatic Medicine, 56, 353-359.
Berenbaum, H., & Prince, J.D. (1994). Alexithymia and the interpretation of
emotion-relevant information. Cognition and Emotion, .8., 231-244.
Braun, C.M.J., Denault, C., Cohen, H., & Rouleau, I. (1994). Discrimination of
facial identity and facial affect by temporal and frontal lobectomy patients. Brain and
Cognition, 24, 198-212.
Buck, R. (1984). The communication of emotion. New York: Guilford Press
Buck, R. (1993). What is this thing called subjective experience? Reflections on
the neuropsychology of qualia. Neuropsychology, 1, 490-499.

100

Caccioppo, J.T., Petty, R.E., & Kao, C.F. (1984). The efficient assessment of
need for cognition. Journal of Personality Assessment, 48, 306-307.
Carrera-Levillain, P., & Fernandez-Dols, J-M. (1994). Neutral faces in context:
their emotional meaning and their function. Journal of Nonverbal Behaviour, 18., 281-299
Clark, L.A, & Watson, D. (1994). Distinguishing functional from dysfunctional
affective responses. In Paul Ekman & Richard J. Davidson (Eds.), The Nature of
Emotion: Fundamental Questions (pp. 131 -136). New York: Oxford University Press.
Conte, H.R. , Plutchik, R. , Jung, B.B., Picard, S., Karasu, T.B., & Lotterman, A.
(1990). Psychological mindedness as a predictor of psychotherapy outcome: A
preliminary report. Comprehensive Psychiatry, 31 , 426-431.
Costa, P.T., & McCrae, R.R. (1985). The NEO Personality Inventory Manual.
Odessa: Psychological Assessment Resources.
Courchesne, E., Townsend, J., Akshoomoff, N.A., Saitoh, 0. , Yeung-Courchesne,
R. , Lincoln, A.J., James, H.E., Haas, R.H. , Schreibman, L., & Lau, L. (1994).
Impairment in shifting attention in autistic and cerebellar patients. Behavioural
Neuroscience, 108, 848-865
Crawford, H.J., & Gruzelier, J.H. (1992). A midstream view of the
neuropsychophysiology of hypnosis: recent research and future directions. In Erika
Fromm & Michael R. Nash (Eds.), Contemporary Hypnosis Research (pp.227-266).
New York: The Guilford Press
Dantzer, R. (1991). Stress and disease: a psychobiological perspective. Annals
of Behavioural Medicine, 13, 205-21.

101

Davidson, R.J. (1994). Honouring biology in the study of affective style. In Paul
Ekman and Richard J. Davidson (Eds.), The Nature of Emotion: Fundamental Questions
(pp. 321-328). New York: Oxford University Press.
Davis, M.H. (1983). Measuring individual differences in empathy: evidence for a
multidimensional approach. Journal of Personality and Social Psychology, 44, 113-126.
Dodge, K.A. (1991). Emotion and social information processing. In J. Garber
and K. Dodge (Eds. ), The Development of Emotion Regulation and Dysregulation
(pp.159-181). NewYork: Cambridge University Press.
Ekman, P. (Ed.). (1977). Emotion in the Human Face (2nd ed.) New York:
Cambridge University Press.
Ekman, P., & Friesen, W. (1976) Pictures of Facial Affect. Palo Alto, California:
Consulting Psychologists Press.
Endo, N. , Endo, M., Kirita, T., & Maruyama, K. (1992). The effects of
expression on face recognition. Tohoku Psychologica Folia, 2.1, 37-44.
Fernandez-Dols, J-M., Sierra, B., & Ruiz-Belda, M.A. (1993). On the clarity of
expressive and contextual information in the recognition of emotions: a methodological
critique. European Journal of Social Psychology, 23, 195-202.
Fletcher, K.E., & Averill, J.R. (1984). A scale for the measurement of roleplaying ability. Journal of Research in Personality, 18., 131-149.
Fridja, N.H. (1986). The Emotions. New York: Cambridge University Press

102

Fukunishi, 1., & Rahe, R.H. (1995). Alexithymia and coping with stress in healthy
persons: Alexithymia as a personality trait is associated with low social support and poor
responses to stress. Psychological Reports, 76, 1299-1304.
George, M.S., Ketter, T.A., Gill, D.S, Haxby, J.V., Ungerleider, L.G., Herscovitch,
P., & Post, R.M. (1993). Brain regions involved in recognizing facial emotion or
identity: an Oxygen-15 PET study. The Journal of Neuropsychiatry and Clinical
Neurosciences, .5., 384-394.
Grier, J.B. (1971). Nonparametric indexes for sensitivity and bias: Computing
formulas. Pyschological Bulletin, 75, 424-429.
Haviland, M.G., Shaw, D.O., Cummings, M.A. , & MacMurray J.P. (1988a).
Alexithymia: subscales and relationship to depression. Psychotherapy and
Psychosomatics, 50, 164-170.
Hoptman, M.J., & Davidson, R.J.

(1994). How and why do the two cerebral

hemispheres interact? Psychological Bulletin, 116, 195-219.
Hudley, C., & Graham, S. (1993). An attributional intervention to reduce peerdirected aggression among Mrican-American boys. Child Development, 64, 124-138.
Humphreys, G. W., Donnelly, N., & Riddoch, M. J. (1993) Expression is
computed separately from facial identity, and it is computed separately for moving and
static faces: Neuropsychological evidence. Neuropsychologia, 31 , 173-181.

103

James, F.R., & Large, Robert G. (1991). Quantifying alexithymia and
hypnotizability in pain clinic attenders and normal controls. In M.R. Bond, J.E. Charlton
& C.J. Woolf (Eds.), Proceedings of the Vlth World Congress on Pain. (pp.247-250).

Elsevier Science.
Jimerson, D.C., Wolfe, B.E., Franko, D.C., Covino, N.A., & Sifneos, P.E. (1994).
Alexithymia ratings in bulimia nervosa: clinical correlates. Psychosomatic Medicine, 56,
90-93.
Kalin, N.H. (1993). The neurobiology of fear. Scientific American, 268, 94-101.
Kauhanen J., Kaplan, G.A., Cohen, R.D., Salonen, R., & Salonen, J.T. (1994).
Alexithymia may influence the diagnosis of coronary heart disease. Psychosomatic
Medicine, 56, 237-244.
Kauhanen, J., Kaplan, G.A., Julkunen, J., Wilson, T.W., & Salonen, J.T. (1993).
Social factors in alexithymia. Comprehensive Psychiatry, 34, 330-335.
Kirk, R.E. (1995). Experimental Design. Procedures for the Behavioural
Sciences. New York: Brooks & Cole.
Kolb, B., & Whishaw, I. Q. (1990). Fundamentals of Human Neuropsychology.
New York: Freeman.
Lane, R.D., Quinlan, D.M., Schwartz, G.E., Walker, P.A., & Zeitlin, S.B. (1990).
The levels of emotional awareness scale: a cognitive-developmental measure of emotion.
Journal of Personality Assessment, 55, 124-134.

104

Lane, R.D., & Schwartz, G.E. (1987) Levels of emotional awareness: a
cognitive-developmental theroy and its application to psychopathology. American
Journal of Psychiatry, 144, 133-143.
Lane, R.D., Sechrest, L., Reidel, R., Weldon, V., Kaszniak, A., & Schwartz, G.E.
(1996). Impaired verbal and nonverbal emotion recognition in alexithymia.
Psychosomatic Medicine, 58, 203-210.
Larsen, R.J., Diener, E., & Emmons, R.A. (1986). Affect intensity and reactions
to daily life events. Journal of Personality and Social Psychology, 51, 803-814.
Lazarus, R.S . (1982). Thoughts on the relations between emotion and cognition.
AmericanPsychologist, 37, 1019-1024.
Lazarus, R.S. (1991). Progress on a cognitive-motivational-relational theory of
emotion. American Psychologist, 46, 819-834.
Ledoux, J, Sakaguchi, A., & Reis, D.J. (1984). Subcortical afferent projections of
the medulla geniculate nucleus mediate emotional responses conditioned to acoustic
stimuli. Neuroscience, .1:, 683-698.
Ledoux, J.E. (1986). The neurobiology of emotion. In J.E. Ledoux & W. Hirst
(Eds.), Mind and Brain: Dialogues in Cognitive Neuroscience (pp. 301-355).
Cambridge: Cambridge University Press.
Ledoux, J.E. (1993). Emotional networks in the brain. In M. Lewis & J.M.
Haviland (Eds.), Handbook of Emotion (pp. 109-118). New York: The Guilford
Press.

105

Ledoux, J.E. (1995). Emotion: clues from the brain. Annual Review
Psychology, 46, 209-235.
Lewis, M. (1995). Self-conscious emotions. American Scientist, 83, 68-78.
Linden, W., Wen, F., & Paulhus, D.L. (1995). Measuring alexithymia: reliability,
validity and prevalence. Advances in Personality Assessment, 10, 51-95.
Magnussen, S., Sunde, B., & Dyrnes, S. (1994). Patterns of perceptual
asymmetry in processing facial expression. Cortex, 30, 215-229.
Martin, R.A. , & Lefcourt, H.M. (1984). Situational Humour Response
Questionnaire: Quantitative measure of sense of humour. Journal of Personality and
Social Psychology, 47, 145-155.
Martin, J.B. , & Pihl, R.O. (1986). Influence of alexithymic characteristics on
physiological and subjective stress responses in normal individuals. Psychotherapy and
Psychosomatics, 45, 66-77.
McDonald, P.W. , & Prkachin, K.M. (1990). The expression and perception of
facial emotion in alexithymia: a pilot study. Psychosomatic Medicine, 52, 199-210.
Mendelsohn, M.J., Haith, M.M., & Goldman-Rakic, P.S. (1982) Face scanning
and responsiveness to social cues in infant rhesus monkeys. Developmental Psychology,

18., 222-228.
Millard, R.W. , & Kinsler, B.L. (1992). Evaluation of constricted affect in chronic
pain: an attempt using the Toronto Alexythymia Scale. Pain, 50, 287-292.
Myers, L.B. (1995). Alexithymia and repression: the role of defensiveness and
trait anxiety. Personality and Individual Differences, 19, 489-492.

106

Newton, T.L., & Contrada, R.J. (1994). Alexithymia and repression: contrasting
emotion focused coping styles. Psychosomatic Medicine, 56, 457-462.
Paez, D., Basabe, N., Valdoseda, M., Velasco, C., & lraurgi, I. (1995).
Confrontation: Inhibition, alexithymia, and health. In J. W. Pennebaker (Ed.), Emotion,
Disclosure, and Health (pp.195-222). Washington, DC: American Psychological
Association.
Parker, J.D.A., Bagby, R.M., & Taylor, G.J. (1991). Alexithymia and depression:
distinct or overlapping constructs? Comprehensive Psychiatry, 32, 387-394.
Parker, J.D.A. , Bagby, R.M., Taylor, G.J. , Endler, N.S., & Schmitz, P. (1993).
Factorial validity of the 20-item Toronto Alexithymia Scale. European Journal of
Personality, 1, 221-232.
Parker, J. , Taylor G., & Bagby, M. (1993a). Alexithymia and the processing of
emotional stimuli: an experimental study. New Trends in Experimental and Clinical
Psychiatry, 2., 9-14.
Parker, J.D.A., Taylor, G.J. , & Bagby, R.M. (1993b). Alexithymia and the
recognition of facial expressions of emotion. Psychotherapy and Psychosomatics, 59,
197-202.
Parker, J.D.A., Taylor, G.J. , Bagby, R.M. , & Acklin, M. (1993). Alexithymia in
panic disorder and simple phobia: a comparative study. American Journal of Psychiatry,
150, 1105-1107.
Parker, P.D. (1995). Factor analysis of the Toronto Alexithymia Scale-20.
Unpublished manuscript, University of Northern British Columbia at Prince George.

107

Parker, P.D., & Prkachin K.M. (1995). Pilot study examining the construct
validity of the Toronto Alexithymia Scale-20. Unpublished raw data.
Pennebaker, J.W., & Hoover, C.W. (1983). Inhibition and cognition: toward an
understanding of trauma and disease.

In R.J. Davidson, G.E. Schwartz & D. Shapiro

(Eds.), Consciousness and Self-Regulation (pp. 107-136). New York: Plenum Press.
Pennebaker, J.W. , Hughes, C. , & O'Heeron, R.C . (1987). The psychophysiology
of confession: linking inhibitory and psychosomatic process. Journal of Personality and
Social Psychology, 52, 781-793.
Pennebaker, J.W., Kiecolt-Glaser, J.K., & Glaser, R. (1988). Disclosure of
traumas and immune function: health implications for psychotherapy. Journal of
Consulting and Clinical Psychology, 56, 239-245.
Perrett, D.I. , Harries, M.H. , Mislin, A.J. , Hietanen, J.K. , Benson, P.J., Bevan, R.,
Thomas, S., Oram, M.W., Ortega, J. , & Brierley, K. (1990). Social signals analyzed at
the single cell level: someone is looking at me, something touched me, something moved!
International Journal of Comparative Psychology, _4, 25-55.
Petri, H.L., & Mishkin, M.

(1994) Behaviourism, cognitivism and the

neuropsychology of memory. American Scientist, 82, 30-37.
Petrie, K.J., Booth, R.J., Pennebaker, J.W., Davison, K.P., & Thomas, M.G.
(1995). Disclosure of trauma and immune response to a Hepatitis B vaccination
program. Journal of Consulting and Clinical Psychology, 63, 787-792.
Pollack, I., & Norman, D.A.A. (1964). A non-parametric analysis of recognition
experiments. Psychonomic Science, 1. 125-126.

108

Posner, M.I. (1994). Attention: the mechanisms of consciousness. Proceedings
of the National Academy of Sciences of the United States of America, 91 , 7398-7403.
Posner, M.l., & Petersen, S.E. (1990). The attention system of the human brain.
Annual Review Neuroscience, 13, 25-42.
Prince, J.D., & Berenbaum, H. (1993). Alexithymia and hedonic capacity.
Journal of Research in Personality, 27, 15-22.
Prkachin, G.C., & Prkachin, K.M. (1994). Adaptation to facial expression of
emotion. Cognition and Emotion, .8., 55-64.
Ross, R.G., Radant, A.D., Young, D.A., & Hommer, D.W. (1994). Saccadic eye
movements in normal children from 8 to 15 years of age: a developmental study of
visuospatial attention. Journal of Autism and Developmental Disorders, 24, 413-431
Rubino, LA. (1995). Comment upon some personality correlates of alexithymia.
Psychological Reports, 76, 544-546.
Seeman, T.E., & Syme, S.L. (1987). Social networks and coronary heart disease:
a comparison of the structure and function of social relations as predictors of disease.
Psychosomatic Medicine, 49,341-354.
Sifneos, P.E. (1972). Short term psychotherapy and emotional crisis.
Cambridge: Harvard University Press.
Sifneos, P.E. (1973). The prevalence of 'alexithymic' characteristics in
psychosomatic patients. Psychotherapy and Psychosomatics, 22, 255-262.
Sivik, T. (1993). Alexithymia and hypersensitivity to touch and palpation.
Integrative Physiological and Behavioural Science, 28, 130-136.

109

Spiegel, D. (1995). Commentary. Journal of Psychosocial Oncology, TI.,_l15.
Stroop, J.R. (1935). Studies of interference in serial verbal reactions. Journal of
Experimental Psychology, 18., 643-661.
Taylor G.J. (1984). Alexithymia: concept, measurement and implications for
treatment. The American Journal of Psychiatry, 141 , 725-732.
Taylor, G.J. (1987). Alexithymia: culture and class relationships. A symposium.
Transcultural Psychiatric research review, 24, 85-95.
Taylor, G.J. (1994). The alexithymia construct: conceptualization, validation and
relationship with basic dimensions of personality. New Trends in Experimental and
Clinical Psychiatry, 10, 61-7 4.
Wainwright-Sharp, J.A., & Bryson, S.E. (1993). Visual orienting deficits in highfunctioning people with autism. Journal of Autism and Developmental Disorders, 23, 113
Walsh, V. , & Perrett, D.I. (1994). Visual attention in the occipitotemporal
processing stream of the macaque. Cognitive Neuropsychology, 2., 243-263.
Watson, D., Clark, L.A., & Tellegen, A

(1988). Development and validation of

brief measures of positive and negative affect: The PANAS scales. Journal of
Personality and Social Psychology, 54, 1063-1070.
Weddell, R.A. (1989). Recognition memory for emotional facial exressions in
patients with focal cerebral lesions. Brain and Cognition, 11, 1-17.

110

Wehmer, F., Brejnak, M.A., Lumley, M., & Stettner, L. (1995). Alexithymia and
physiological reactivity to emotion-provoking visual scenes. The Journal of Nervous and
Mental Disease, 183, 351-357.
Weinberger, D.A. (1990). The construct validity of the repressive coping style.
In J.L. Singer (Ed.), Repression and Dissociation: Implications for Personality Theory,
Psychopathology and Health (pp.337-386). Chicago: University of Chicago Press.
Weiskrantz, L. (1986). Blindsight: a case study and implications. Oxford:
Oxford University Press.
Wheeler, K. , & Broad, R.D. (1994). Alexithymia and overeating. Perspectives in
Psychiatric Care, 30, 7-10.
Williams, J.M.G., & Broadbent, K. (1986). Distraction by emotional stimuli: Use
of a Stroop task with suicide attempters. British Journal of Clinical Psychology, 25, 101110.
Young, A.W. (1992). Face recognition impairments.

Philosophical Transactions

of the Royal Society of London B, 335, 47-54.
Zajonc, R.B. (1980). Feeling and thinking: Preferences need no inferences.
American Psychologist, 35, 151-175.
Zeitlin, S.B., Lane, R.D. , O'Leary, D.S., & Schrift, M.J. (1989). Interhemispheric
transfer deficit and alexithymia. American Journal of Psychiatry, 146, 1434-1439.
Zuckerman, M., Eysenck, S.B.G., & Eysenck, H.J. (1978). Sensation seeking in
England and America: crosscultural, age and sex comparisons. Journal of Consulting
and Clinical Psychology, 46, 139-149.

111

Appendix A

The Toronto Alexithymia Scale-20 (TAS-20)

112

TAS- 20
Using the scale provided as a guide, indicate how much you agree or disagree with each of the
following statements by circling the corresponding number. Give only one answer for each statement.
Circle 1 if you STRONGLY DISAGREE
Circle 2 if you MODERATELY DISAGREE
Circle 3 if you NEITHER DISAGREE NOR AGREE
Circle 4 if you MODERATELY AGREE
Circle 5 if you STRONGLY AGREE
Neither
Disagree
Nor Agree

Moderately
Agree

1--2

3

4

5

for my feelings.

1--2

3

4

5

I have physical sensations that even
doctors don't understand.

1--2

3

4

5

4. I am able to describe my feelings easily.

1--2

3

4

5

5. I prefer to analyze problems rather than
just describe them.

1--2

3

4

5

6. When I am upset, I don' t know if I am sad,
frightened, or angry .

1--2

3

4

5

7. I am often puzzled by sensations in my
body.

1--2

3

4

5

8. I prefer to just let things happen rather
than to understand why they turned out
that way.

1--2

3

4

5

9. I have feelings that I can't quite identify.

1--2

3

4

5

1--2

3

4

5

1--2

3

4

5

1--2

3

4

5

Strongly
Disagree

1. I am often confused about what emotion
I am feeling.

2. It is difficult for me to find the right words
3.

10. Being in touch with your emotions is
essential.

11. I find it hard to describe how I feel about
people.

12. People tell me to describe my feelings
more.

Moderately
Disagree

Strongly
Agree

113

Neither
Strongly
Disagree

Moderately
Disagree

Disagree
Nor Agree

Moderately
Agree

Strongly
Agree

13. I don't know what's going on inside me.

1--2

3

4

5

14. I often don't know why I'm angry.

1--2

3

4

5

15. I prefer talking to people about their daily
activities rather than their feelings.

1--2

3

4

5

16. I prefer to watch "light" entertainment
shows rather than psychological dramas.

1--2

3

4

5

17. It is difficult for me to reveal my
innermost feelings, even to my closest
friends.

1--2

3

4

5

18. I can feel close to someone, even in
moments of silence.

1--2

3

4

5

19. I find examination of my feelings useful in
1--2
solving personal problems.

3

4

5

20. Looking for hidden meanings in movies or
1--2
plays distracts from their enjoyment.

3

4

5

114

Appendix B

The Positive and Negative Affect Schedule (PANAS)

115

THE PANAS
This scale consists of a number of words that describe different feelings and emotions. Read each
item, and then mark the appropriate answer in the space next to that word. Indicate to what extent
you generally feel this way, that is, how you feel on the average. Use the following structure to
record your answers.
1

very slightly or
not at all

2

a little

3

moderately

__interested
- -distressed
__excited
_ _upset
__ strong
_guilty
- -scared
- -hostile
- -enthusiastic
_ _proud

quite a bit

4

_ _irritable
- -alert
- -ashamed
__inspired
_ _nervous
- -determined
- -attentive
_jittery
- -active
- -afraid

5

extremely

116

Appendix C

Consent Form

117

Individual Characteristics Research Project
Information and Consent Form
Investigators: Philippa Parker and Ken Prkachin
Psychology Program, UNBC
We are conducting a study which will last approximately one hour. As a subject,
you will be asked to read and sign this consent form, complete two questionnaires, and
participate in three short experiments. In the experiments you will be shown visual
material and you will be asked to indicate your response orally or by pressing a button.
The order of presentation of the studies and completion of the questionnaires will vary for
each student.
The two questionnaires ask you for information about individual characteristics.
Psychologists often use such questionnaires to try to gain a better understanding of why
people think, feel and act as they do. This present project is a follow-up to a pilot project
conducted earlier this year. The object is to discover differences in responses to various
situations.
The two questionnaires vary slightly in how you are asked to respond. Please read
the instructions for both questionnaires carefully and then answer the questions to the best
of your ability. You may find it difficult to answer some questions. Please try, to the
extent that it is possible, to answer all items. However, if there are some that you truly
cannot answer, or would prefer not to answer, you may leave them blank. The two
questionnaires should take ten to fifteen minutes to complete.
Please be assured that all the information you provide will be kept in the strictest
confidence. We do need to ask you some general information, such as your age, gender
and student ID. However, once the tests are administered and the data entered, the actual
questionnaires will be destroyed. The data will be stored anonymously, and each subject
will have been allocated a unique number. Please be assured that your participation is
voluntary, and that you are free to withdraw at any time.
Please indicate your consent to participate in this study by signing below.
I,
, have read the foregoing information. I understand that my
participation in this research project is voluntary and that I am free to withdraw at any
time. I also understand that the information I provide will be treated in strict confidence.
Based on these understandings, I hereby give my consent to participate.
Signature

Date

Student#

Gender