Alexithymia and temperament and character model of personality in alcohol-dependent Turkish men


*Cuneyt Evren, MD, Bakirkoy State Hospital for Mental Health and Neurological Disorders, Istanbul, Turkey, Alcohol and Drug Research, Treatment and Training Center (AMATEM), Icadiye Cad. Mentes Sok. Selcuk Apt. 1/17 34674, Kuzguncuk, Uskudar, Istanbul, Turkey. Email:


Aims:  Alexithymia, a personality trait characterized as having problems identifying, describing, and working with one's own feelings, often marked by a lack of understanding of the feelings of others, is only partly described within the context of personality. The aim of the present study was therefore to study the prevalence of alexithymia among male alcohol-dependent inpatients and investigate the relationship between alexithymia and the dimensions of Cloninger's psychobiological model of personality.

Methods:  The Turkish version of the Toronto Alexithymia Scale (TAS-20) and the Turkish version of the Temperament and Character Inventory (TCI) were administered to 111 male alcohol-dependent inpatients.

Results:  TAS-20 scores correlated positively with harm avoidance and self-transcendence and negatively with self-directedness and cooperativeness. Regression analysis identified high harm avoidance and self-transcendence and low self-directedness as independent predictors of alexithymia. Also harm avoidance and self-transcendence predicted alexithymia in a logistic regression model.

Conclusions:  Alexithymia can be explained by specific dimensions within Cloninger's psychobiological model of personality in alcohol-dependent Turkish men.

ALEXITHYMIA IS AN emotional processing disturbance clinically manifested by difficulties in identifying and verbalizing feelings, in elaborating fantasies, and by a tendency to focus on and amplify the somatic sensations accompanying emotional arousal. Alexithymia is thought to be a stable personality trait and a predisposing risk factor for a variety of psychiatric disorders. Salient features of alexithymia are the inability to distinguish one's feelings from the accompanying bodily sensations, the inability to communicate feelings to others, and an externally oriented cognitive style reflecting an absence of inner thoughts and fantasies. These three lower-order concepts reflect separate, yet empirically related facets of the alexithymia construct.1

The term ‘alexithymia,’ was first introduced by Sifneos in 1972. The word was composed from Greek words: a for lack, lexis for word, and thymos for feeling. The Toronto Alexithymia Scale (TAS) was developed in 1985 as a reliable and valid self-report measure of the alexithymia construct, which has been modified twice into the Revised Toronto Alexithymia Scale and the Twenty-Item Toronto Alexithymia Scale (TAS-20). Nevertheless, according to Sifneos, the optimal way to measure alexithymia should be to use both the Andreasen ‘affect rating scale’ and the Beth Israel Questionnaire, in conjunction with the well-validated original Toronto Alexithymia Scale.2

Studies suggest relatively high prevalence rates of alexithymia both in men with high alcoholism risk3 and in men with alcohol use disorders.4,5 This prevalence rate among alcoholic patients has been reported to range from 42% to 79% in several studies.3,5–9 The prevalence rate was not different in Turkish alcohol-dependent men (48–56%).10–12 Studies on this issue suggested that alexithymic features may negatively affect the long-term alcoholism treatment outcomes and that they may also predict alcohol use.13,14

In recent years, several researchers have emphasized the need for linking personality constructs with neurobiological processes that might underlie individual differences in personality.15 Some studies have explored the relationship between alexithymia and personality traits measured with the Neuroticism, Extraversion/Introversion, and Openness to Experience–Personality Inventory (NEO-PI), which is a measure of the five major domains of personality as well as the six facets that define each domain. Bagby et al. found a positive correlation between alexithymia and neuroticism and a negative correlation with extraversion and openness and found no relationship to the conceptually unrelated traits of agreeableness and conscientiousness.16 Other correlational studies also confirmed these relationships.16–20 Wise et al. suggested that alexithymia is a unique personality trait that is not fully explained by the NEO-Five-Factor Inventory (NEO-FFI).18 A recent study found a moderate positive correlation between TAS-20 total score and neuroticism, but weak but significant negative correlations with openness to experience and conscientiousness domains of the NEO-FFI.21 In another study the TAS-20 and its subscales were analyzed to investigate correlations with the NEO Personality Inventory–Revised (NEO-PI-R) in an obese sample of 259 patients. The TAS-20 and its subscale scores were found to be correlated with elevated neuroticism, lower levels of extraversion and openness, in line with the previous research, but also, somewhat unexpectedly, were found to be correlated with lower conscientiousness, and also with lower agreeableness in women.22 People with high neuroticism tend to think unrealistically, to be unable to control their anger, and to be poor at coping with stressors. People with low openness are deficient in imaginative activity, and tend not to seek out new experiences, and therefore may be limited in their opportunities to describe their emotions to others or learn about others' descriptions of emotions. Individuals low in conscientiousness have a tendency to a lack of self-control and consistent effort.23

There have been few studies conducted with the Temperament and Character Inventory (TCI). In a study including 254 psychiatric inpatients and outpatients, the TCI dimensions of harm avoidance, low self-directedness, and low reward dependence were found to be independent predictors for alexithymia.24 These personality traits may contribute to the characterization of highly alexithymic patients as they experience poorly differentiated emotional distress because they lack the necessary psychological capacities for modulating emotions.25 In a healthy college student sample, Picardi et al. measured alexithymia with the TAS-20 and reported that alexithymia was correlated with both temperament and character dimensions of personality, indicating that both genetic and environmental factors might contribute to alexithymia development.26 In that study the TAS-20 total and subscale scores were correlated with harm avoidance, low self-directedness, low cooperativeness, and low reward dependence.

Consistent with studies conducted on the general population,27 a recent study suggested that alexithymia, as a personality trait, might be a vulnerability factor for alcohol dependence.28 Alexithymia may be a risk factor for alcoholism and it also may affect treatment results.29 Little is known, however, about how to approach alexithymia in this patient group. Thus, understanding alexithymia and the variables that predict alexithymia among this population may help us in terms of prevention as well as in treatment.

The aim of the present study was to identify the prevalence of alexithymia and the relationship of Cloninger's psychobiological model of personality with alexithymia in a sample of alcohol-dependent Turkish men. The previous studies were conducted in a heterogeneous population (psychiatric inpatients and outpatients) or non-patient population (healthy college student sample). Although TAS-20 and TCI are known to be culturally stable instruments, alexithymia and TCI dimensions that predict alexithymia may differ in more homogeneous populations, such as alcohol-dependent inpatients and also for a population in a developing country with a different religion. We hypothesized that alexithymia would correlate positively with harm avoidance and negatively with self-directedness, cooperativeness, and reward dependence and that alexithymia would be unrelated to novelty seeking, persistence, and self-transcendence subscales of the TCI.



The study was conducted at the Alcohol and Drug Research Treatment and Training Center (AMATEM) at the Bakirkoy Mental Hospital, in Istanbul, Turkey. The study was approved by the Ethics Committee of the AMATEM and all patients gave written informed consent before participation. The sample consisted of 111 male alcohol-dependent inpatients, according to DSM-IV. Clinical interviews with the patients were done after a detoxification period, 4–6 weeks after the last use of alcohol. Eleven patients who were younger than 18 years of age, had mental retardation or cognitive impairment, were unable to read or had comorbid psychotic disorders were excluded.


For psychiatric assessment, the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I)30 was used. Patients were also assessed with a semi-structured sociodemographic data questionnaire. All patients were assessed using a semi-structured sociodemographic form. Alexithymia was assessed with the Turkish version of the 20-item TAS-20.16,31 The TAS-20 is a self-report scale consisting of 20 items, which are rated on a 5-point Likert scale, ranging from 1, strongly disagree, to 5, strongly agree, with total scores ranging from 20 to 100. Items 4, 5, 10, 18, and 19 are negatively keyed. The first factor (F1) in the three-factor model for the TAS-20 consists of seven items (items 1, 3, 6, 7, 9, 13, 14) assessing the ability to identify feelings and to distinguish them from the somatic sensations that accompany emotional arousal. Factor 2 (F2) consists of five items (items 2, 4, 11, 12, 17) assessing the ability to describe feelings to other people. Factor 3 (F3) consists of eight items (items 5, 8, 10, 15, 16, 18, 19, 20) assessing externally oriented thinking. The total scores of the TAS-20 were dichotomized as a score of ≥61, which indicated alexithymia, and a score of <61, which indicated no alexithymia. The Turkish version of the TAS-20 has been translated by Sayar and Kose. The approved form has been validated in a Turkish population study.32,33 In the present study Cronbach's alpha for the TAS-20 was 0.74.

For the evaluation of temperament and character traits, the Turkish version of the TCI34 was used. The TCI is a 240-item self-administered questionnaire that measures the four temperament dimensions (novelty seeking, harm avoidance, reward dependence, and persistence) and the three character dimensions (self-directedness, cooperativeness, and self-transcendence). The Turkish version of the TCI has been validated in a Turkish sample.35

Data analysis

Statistical analyses were done using SPSS for Windows, version 12.0 (SPSS, Chicago, IL, USA) statistical program. Patients with a TAS-20 score ≥61 were compared with non-alexithymic patients with a score <61. For categorical data, χ2 tests were used. An unpaired Student's t-test was used for group comparisons of continuous variables. Pearson's correlation analysis, linear and logistic regression were performed for assessment of predictor variables.


The mean value in the present sample for the TAS-20 total score was 57.6 ± 11.2, whereas for difficulties in identifying feelings, the score was 21.7 ± 6.3, for difficulties in describing feelings, the score was 15.0 ± 4.3, and for externally oriented thinking the score was 21.0 ± 4.6.

In the total sample, 43.2% of patients (n = 48) had a score ≥61 on the TAS-20, and were therefore considered alexithymic. There were no statistical differences between groups in terms of current age, marital, and educational status. The unemployment rate of patients was higher in the alexithymic group (39.6%) than the non-alexithymic group (19.0%; Table 1).

Table 1.  Sociodemographic variables
n = 63%n = 48%
Age (mean ± SD) (years)42.0 ± 9.240.6 ± 8.7t = 1.830.41
Marital status    5.5920.06
Divorced, widowed, Separated1219.01735.4   
Employment status    8.530.037
Without employment1219.01939.6   
With employment2946.01939.6   
Part time1117.524.2   
Retired, Student1117.5816.7   
Education status    6.530.088
Middle school (6–8th grades)1117.51020.8   
High school-2 (9–11th grades)1930.21020.8   

Educational achievement was lower in the alexithymic group (mean, 8.5 ± 3.4 years) than in the non-alexithymic group (mean, 10.3 ± 3.8 years). Alexithymic subjects were significantly less educated than the non-alexithymic subjects (t = 2.7, P = 0.009). The age at onset of alcohol use was lower in the alexithymic (mean, 17.2 ± 4.0 years) than in the non-alexithymic group (mean, 18.8 ± 5.3 years), although the difference did not reach significance (t = 1.7, P = 0.094). The age at which alcohol-related problems started was 24.5 ± 6.8 years in the alexithymic group, and 27.4 ± 7.1 years in the non-alexithymic group. The difference between groups was statistically significant (t = 2.14, P = 0.035). The mean duration of alcohol use disorder was 15.3 ± 8.0 years, which did not differ significantly between the alexithymic group (mean, 16.1 ± 8.3 years) and the non-alexithymic group (mean, 14.7 ± 7.8 years).

A positive correlation was found between the TAS-20 score and the temperament dimension of harm avoidance and the character dimension of self-transcendence. There was also a negative correlation between the TAS-20 score and the character dimensions of self-directedness and cooperativeness. Similar correlations were found between the first factor of the TAS-20 (difficulty in identifying feelings) and the temperament and character dimensions of the TCI (Table 2).

Table 2.  Correlations between TAS-20 and TCI scores
  • *

    P < 0.05;

  • **

    P < 0.01;

  • ***

    < 0.001.

  • C, cooperativeness; DDF, difficulties in describing feelings; DIF, difficulties in identifying feelings; EOT, externally oriented thinking; HA, harm avoidance; NS, novelty seeking; P, persistence; RD, reward dependence; S, self-directedness; ST, self-transcendence; TAS-20, 20-item Toronto Alexithymia Scale; TCI, Temperament and Character Inventory.

DDF 0.150.84***0.070.37***−0.22*−0.22*−0.38***−0.23*0.17
EOT  0.49***−0.070.02−0.030.08−0.13−0.160.11
TAS-20   0.110.41***−0.18−0.17−0.44***−0.31**0.29**
NS    0.14−0.03−0.31**−0.16−0.31**0.13
HA     −0.16−0.44***−0.59***−0.44***0.13
RD      0.26**0.24*0.36***0.01
P       0.24*0.25**0.26**
S        0.60***−0.35***
C         −0.12

Temperament dimensions of harm avoidance were higher and persistence was lower in the alexithymic group than in the non-alexithymic group. Character dimensions of self-directedness and cooperativeness were lower and self-transcendence was higher in the alexithymic group than the non-alexithymic group (Table 3).

Table 3.  TCI scores according to the presence of alexithymia
 Non-alexithymic (n = 63)Alexithymic (n = 48)tP
Mean ± SDMean ± SD
  1. TCI, Temperament and Character Inventory.

Novelty Seeking17.9 ± 4.919.3 ± 3.8−1.60.1
Harm avoidance16.8 ± 5.822.0 ± 4.9−5.1<0.001
Reward dependency14.7 ± 3.014.2 ± 2.70.760.45
Persistence5.5 ± 1.34.6 ±
Self-directedness26.4 ± 6.021.6 ± 5.44.4<0.001
Cooperativeness30.0 ± 5.027.2 ±
Self-transcendence18.3 ± 5.421.3 ± 6.7−3.10.003

Taking the TAS-20 total score as a dependent variable in multiple linear regression and including seven TCI dimensions into the model as independent variables, the results indicated a significant predictive power of harm avoidance, self-transcendence and low self-directedness for the total alexithymia scores (Table 4). Harm avoidance was also found to be a predictor of the ‘difficulties in identifying feelings’ and ‘difficulties in describing feelings’ subscales of the TAS-20, whereas self-transcendence was a predictor of the ‘difficulties in identifying feelings’ and self-directedness was a predictor of the ‘difficulties in describing feelings’ subscales (Table 4). Taking alexithymia as the dependent variable in forward stepwise multiple logistic regression and including dimensions of the TCI in the model as independent variables, the results indicated a significant predictive power of harm avoidance and self-transcendence for the presence of alexithymia (Table 5).

Table 4.  Stepwise linear regression: TAS-20 and subscale scores as dependent variables, and temperament and the character dimensions of TCI as independent variables in the first step
 Difficulties in identifying feelingsDifficulties in describing feelingsTAS-20 Total score
  1. Results are presented as standardized regression coefficients β, *P < 0.05; **P < 0.01; ***P < 0.001.

  2. There was no appropriate model for ‘Externally oriented thinking’ subscale of the TAS-20.

  3. TAS-20, 20-item Toronto Alexithymia Scale; TCI, Temperament and Character Inventory.

anovaF = 21.3, d.f. = 2, 108F = 11.5, d.f. = 2, 108F = 12.1, d.f. = 3, 107
P < 0.001P < 0.001P < 0.001
Adjusted R20.270.160.23
Harm avoidance0.43***0.22*0.25*
Self-directedness −0.25*−0.23*
Self-transcendence0.27** 0.18*
Table 5.  Determinants of alexithymia in forward (Wald) logistic regression model
 BSEWaldd.f.Pexp(B)95.0%CI for exp(B)
  1. CI, confidence interval.

Harm avoidance0.1780.04317.0921<0.0011.1951.0981.300


The prevalence of alexithymia in the present study sample was found to be 45.1%, which is consistent with the prevalence reported in Western populations3,36 and in samples of Turkish alcoholic outpatients.10,12 The present findings also replicated results of Uzun et al. in a larger sample of Turkish alcoholic subjects.12 A high alexithymia rate suggests a strong connection between alexithymia and alcoholism. Studies have shown that alexithymia predicted poor outcome in alcoholic inpatients.9,36 In alexithymic subjects, alcohol use might alleviate stressful situations and facilitate verbal and emotional contacts, but eventually lead to the development of alcoholism.

The second aim of the present study was to identify the relationship between alexithymia and Cloninger's temperament and character model of personality among male Turkish alcoholic inpatients. To our knowledge the present study is the first to investigate this relationship in a sample of alcohol-dependent subjects. The present results supported the findings of Grabe et al., who studied a more heterogeneous sample of psychiatric inpatients and outpatients,24 and the present results were consistent with the general view that alexithymia is not represented by one single personality dimension or subscale but is best explained by a mixture of different traits within Cloninger's temperament and character model of personality.

We found no relationship between alexithymia and personality disorders among Turkish alcohol-dependent subjects in a previous study, in which personality disorder criteria according to the DSM-III-R were used.11 This finding supported the idea that alexithymia is a construct associated with distinct personality dimensions.16–18,37 Several studies have shown that alexithymia is multidimensional rather than unidimensional, consisting of emotional and cognitive components.29,38 The emotional component is characterized by an incapacity to describe and identify feelings, and the cognitive component is characterized by a lack of daydreaming and the presence of externally oriented thinking. In the present study, participants' TCI scores correlated with the incapacity to describe and identify feelings. This suggests that difficulty in identifying feelings is the most important factor that determines the relationship with the temperament and character dimensions of personality in alcohol-dependent subjects. The regression analysis identified the TCI dimensions of high harm avoidance and self-transcendence and low self-directedness as independent predictors for TAS-20 total scores. This supports the conclusions of Grabe et al. and Picardi et al. that both genetic (temperament) and environmental (character) dimensions of personality contribute to the alexithymia construct.24,26 Alexithymic deficits in emotion regulation are related to the reduced ability to control, regulate, and adapt behavior according to chosen goals and values that characterize individuals with low self-directedness. Alexithymic deficits in emotion regulation are also plausibly related to cautiousness, fearfulness, pessimism, and shyness in individuals who score high in harm avoidance.

The alexithymic group had higher harm avoidance and self-transcendence scores and lower persistence, self-directedness and cooperativeness scores than the non-alexithymic group. Nevertheless, not all of these variables were correlated with or were predictors of alexithymia. The first hypothesis that alexithymia was positively correlated with harm avoidance and negatively with self-directedness was confirmed. Harm avoidance is a hereditary tendency to inhibit or cease behavior, and can include pessimistic worry in anticipation of future problems, passive avoidant behaviors such as fear of uncertainty and shyness of strangers, and fatigue. Character dimensions self-directedness include acceptance of responsibility for one's own choices, identification of individually valued goals and purposes, development of skills and confidence in solving problems, and self-acceptance.34 Instead of low reward dependence found in the study by Grabe et al.,24 we found self-transcendence as an independent predictor for alexithymia in the present sample. Character dimension of self-transcendence is described as self-forgetfulness, transpersonal identification, and spiritual acceptance.34 This difference could be attributed to sampling differences, different treatment settings (the present sample consisted of inpatients), or cultural differences.

Freyberger divided alexithymia into two types: a personality trait (primary) and a state reaction (secondary).39 Primary alexithymia could constitute a clinical trait that would characterize subjects who are prone to psychosomatic or psychiatric disorders. Secondary alexithymia could constitute a clinical state reflecting stress or adaptation during severe psychosomatic or psychiatric disorders. This proposal has been supported by recent studies.29,40–43 Haviland et al. found that alexithymic characteristics of newly abstinent alcoholic subjects decreased as their treatment progressed, and hence concluded that alexithymia was a state reaction in alcohol-dependent patients.29 Nevertheless the latest study evaluating the stability of alexithymia in alcohol-dependent patients suggested that alexithymia is a stable personality trait rather than a state-dependent phenomenon.28

The present study had several limitations. First, all measures used in the present study were self-reported. Alexithymic subjects are perhaps inherently unable to evaluate themselves correctly because of their difficulties in cognitive processing of emotions. We conducted the study after the detoxification phase to eliminate this potential confound. Although the latest study confirmed that alexithymia is a stable personality trait in alcohol-dependent subjects,28 an important limitation was that we did not included anxiety, depressive and somatic symptoms as independent variables in regression models. Another limitation of the present study was that all the patients were male. The study group was restricted to a treatment-seeking population, and it was not possible to generalize the present findings to non-treatment groups. Finally, we did not use any scale to measure severity of alcohol use.

The rate of alexithymia is high among adult alcohol-dependent Turkish men and the present results suggest a strong relationship between alexithymia and alcoholism in this population. Specific therapeutic interventions to improve the ability to identify and describe subjective feelings and to establish conscious relationships between some overwhelming emotions and the occurrence of episodes of alcohol craving and relapse might be clinically relevant in alexithymic alcohol-dependent subjects.28 Alexithymia is explained by specific dimensions within Cloninger's psychobiological model of personality in alcohol-dependent Turkish men, but further studies are needed to understand the associated or underlying mechanism of alexithymia in population samples other than North American or British samples.


The authors would like to thank Ebru Ilicali and Jennifer G. Schnellmann, PhD, ELS, for their assistance with editing the manuscript.