Somatization is a persistent and puzzling clinical problem.  Patients who seek medical care for somatic symptoms but have little or no identifiable disease are said to be ‘somatizing’. The definition in DSM 1V (American Psychiatric Association, 1994) for somatoform is very specific, and it is often the case that somatization, as it commonly presents to health care professionals, does not meet these stringent criteria.  However, this phenomenon still causes much distress to patients and difficulty to the health service in general. Patients are often sent for many consultations, usually starting in Primary Care, where it is estimated that they may account for 10% to 30% of visits to family doctors (Kellner, 1990).  In many cases they are referred by their General Practitioners to multiple medical and surgical specialists, costing the National Health Service a great deal of money each year (Bass & Murphy, 1990).

In this section:

Definitions of somatization

Lipowski (1968) defined somatization as ‘‘the tendency to experience, conceptualise and / or communicate psychological states or contents as bodily sensations, functional changes or somatic metaphors’’.  This definition puts emphasis on the patient’s interpretation of the symptoms.  Kellner (1991) conceptualises somatization as involving the “occurrence of physical symptoms that are not supported by recognisable or sufficient physical pathology”.  This definition does not rule out the presence of organic pathology altogether, which Lipowski’s definition appears to do, neither does Kellner’s definition make any assumption that psychological or emotional states are at the root of the problem, which Lipowski’s explicitly does. 

 Defining somatization is complex. The term has been an umbrella word for different and overlapping concepts, and related research has often used different definitions of somatization.  De Gucht & Fischler (2002) point out that there are two theoretically distinct concepts of somatization, which map approximately on to the two definitions given above. One posits a causal relation between the experience of psychological distress and the presentation of somatic symptoms, seeing the somatic symptoms as an alternative expression of psychological distress.  The other defines somatization simply as the presence of medically unexplained symptoms, without the assumption that there is a psychiatric disorder or psychological problem also present.

According to De Gucht & Fischler (2002) these two strands have their roots in Freud’s distinction between conversion hysteria and neurasthenia.  Conversion hysteria was defined as being psychic in origin, whereas neurasthenia was defined as somatic in origin.  The difference was that conversion hysteria had its roots in the symbolic expression of infantile sexual conflicts whereas neurasthenia, an actual neurosis, was the consequence of unresolved sexual tension in the present.  This distinction is a useful one in understanding how the somewhat confusing and imprecise conceptualization of somatization has arisen.  Kirmayer & Robbins (1991) classed these concepts as two separate clinical phenomena, termed presenting somatization and functional somatization, respectively. The following theories will be discussed with this distinction in mind, in an attempt to structure and clarify the construct of somatization.  

Theories of somatization

Psychological formulations of the aetiology and maintenance of somatization disorders have included psychosomatic, behavioural, cognitive, social learning and psychoanalytic theories.  A model involving emotional processing is also discussed here.  There are many overlaps between these models, which will be discussed later in this article when the models have been outlined. 

Psychosomatic theory

Psychosomatic theorists (e.g. Lask and Fosson, 1989) argue that somatic symptoms are a means of expressing distressing emotions. For those individuals that find acknowledging and expressing emotions difficult, they may, instead, learn to display their distress through bodily symptoms. This expression may be learnt by the individual’s own experience of illness or through observation of this in the immediate social network. The types of symptoms that are developed are thought to be dependent on the individual’s unique physiological weaknesses.  According to Craig, Drake, Mills & Boardman (1994), individual variations in physiological response may have a part to play in making an individual more vulnerable to somatization; a person whose physiological responses are heightened may notice their own responses more, and therefore be more likely to focus on them, and interpret them as illness

Psychosomatic theory seems to be most closely related to De Gucht & Fishler’s presenting somatization definition. It is making an assumption that psychological or psychiatric distress is present, and is being channelled into psychosomatic symptoms.  Definitions constructed by psychosomatic theorists have embodied this principle eg Bridges & Goldberg (1985) who considered that, to be considered ‘somatisers’ for the purposes of their research, patients must meet the following criteria: they must be seeking medical help for somatic symptoms, and not for psychological or psychosocial distress: they must attribute their symptoms to physical illness: and must report when properly interviewed, symptoms that justify psychiatric diagnoses.  Researchers within this perspective such as Bridges and Goldberg (1985) and Craig, Boardman, Mills, Daly-Jones & Drake (1993) have therefore screened out patients who do not display psychiatric symptoms, and their research has therefore tended to find associations between somatization and psychiatric/psychological distress for this reason. 

Behavioural theory

Behavioural theory suggests that illness behaviours can be understood in terms of the functions that they serve, and also by the ways in which behaviours are reinforced by the responses and consequences they receive (Fordyce, 1976).  Individuals may learn to adopt sick-role behaviours after receiving positive reinforcement.  The child’s behaviour may then be inadvertently reinforced, either positively, by care and attention, or negatively, because the child is able to avoid disliked tasks, high expectations or confrontation.  After many repetitions, certain cues in the environment (for example, child is excused from chores due to having a stomach ache) may develop into discriminative stimuli, which can trigger episodes of ‘illness behaviour’.  Illness behaviour is a set of behaviours that a person engages in when they define themselves as ill, and illness behaviour itself is heavily influenced by learning within the family environment and also by modelling from caregivers as discussed below.  (For a discussion of illness behaviour in somatization disorder see Rief & Hiller, 1999.)  These patterns, once established, can continue into adulthood.  Behavioural theory would appear to rest within the second definition, functional somatization, as there appears to be no assumption that psychological distress or psychiatric disorder is present; rather illness behaviour is a result of reinforcement patterns in the environment.

Social learning theory

Social learning theory (Bandura 1973) focuses on the learning that takes place vicariously, through watching others i.e. modelling.  People in a child’s early environment, particularly parental figures, have an impact on this learning.  Parents model the behaviour that it is acceptable to display when feeling ill, which also illustrates and communicates their attitudes towards illness.  For example, there is often a great deal of difference between families as to how soon it would be acceptable to define oneself as ill and express this to others and how much it is appropriate or acceptable to express discomfort and ask for support.  The child learns various behaviours from close family members through modelling, and this in turn impacts on the formation of the child’s beliefs about illness (this is discussed in the section on cognitive behavioural theory).  The social learning perspective does not appear to assume that psychological factors are being expressed through physical distress, and, from this point of view, appears to fit better with the functional somatization definition.

However, it is also possible that methods of dealing with psychological distress would be learned in similar ways, through modelling.  If this is the case, unacceptable emotion may be repressed or rechannelled in line with what is socially or culturally acceptable. Difficult emotions may come to be expressed somatically if this is modelled within the family: for example, if open displays of emotion are avoided, but attention is focused on physical symptoms of illness in family members. This process, influenced by modelling, would have more in common with the presenting somatization perspective, with its assumption that psychological distress lies behind unexplained medical symptoms. Therefore, social learning theory could be consistent with either the presenting somatization or the functional somatization perspective, depending on how it is viewed. 

Cognitive behavioural theory

This theory of somatization focuses on a person’s pattern of attributions and beliefs, which may cause him or her to experience physical sensations in certain ways. (See Rief, Hiller & Margraf, 1998).  Beliefs about illness usually develop within family and culture.  For example, beliefs vary from culture to culture, and within that from family to family, concerning how socially acceptable is it to be ill, how dangerous is it, to what extent one should continue to ‘soldier on’ or otherwise, how much care and sympathy can be expected.  Attitudes to illness are modelled by the behaviour of others, and children also acquire attitudes to illness by the way in which their own illness is responded to by parental figures.  Expectations about illness are also learnt in childhood, for example, how much support and comfort is likely to ensue when one is ill; whether illness reduces expectations on the child and the likelihood of punishment; whether illness affords escape from disliked chores or responsibilities.  This clearly overlaps with behavioural theory, and with social learning theory, as what a child acquires by patterns of reinforcement, and learns through modelling, both influence the beliefs and attitudes they develop.

Attributions can affect the experience of illness; for example physical sensations could be attributed to physical illness or pathology when they are in fact secondary to emotion. These attributions can be modelled directly by caregivers who make these attributions themselves, or can be influenced by beliefs about illness.  Certain attributions such as ‘I might have a heart attack if I am not careful’ can result in focus of attention on the body which increases the likelihood that innocuous symptoms will be noticed.  Any negative beliefs about the consequences of illness (eg incapacity, death) may then be triggered, leading to anxiety.  Anxiety creates further physical sensations that can be further misinterpreted, leading to a cycle of physical responses to anxiety and worry.

As there is so much overlap between cognitive, behavioural and social learning aspects of somatization, it is not possible to say whether cognitive behavioural theory fits more with functional somatization or with presenting somatization.  The behavioural aspects of cognitive behavioural theory, i.e. that behaviour is learned or reinforced and exists as a result of these processes, may appear to fit more with the functional somatization perspective. However, the cognitive components fit more with the presenting somatization perspective, as cognitive theory postulates that negative beliefs and attributions are expressed in dysfunctional thoughts, which generate negative emotions.  The negative emotions such as anxiety involve somatic sensations, which are then interpreted as illness.

Psychoanalytic theory

This has developed from Freud’s (1894) theory of conversion hysteria. The basic tenet of this theory was that anxiety aroused by unconscious conflict is converted into physical symptoms.  As the conflict is intolerable to the conscious mind, it is not acknowledged.  However, the more ‘socially acceptable’ physical symptoms are a way of expressing distress.  The person is not aware of this process, as it occurs at the unconscious level.  The gain for the person is the reduction in awareness of the conflict and therefore the reduction in anxiety.  

More recent psychoanalytic conceptualizations of somatization (Bucci, 1997) propose that a disconnection occurs, during development, between verbal and non-verbal information systems, and between symbolic and subsymbolic processing forms.  The nonverbal system incorporates representations and processes in all sensory modalities, including motoric and bodily forms.  In terms of this model, somatization involves a dissociation, within emotional schemata, between somatic activation patterns and the symbolic representation of objects.  An object, in this context, would be the mental representation of a relationship with another person.  If symbolic representation is too far removed from the somatic activation component of emotion associated with the relationship, the person will be unable to make sense of the physical aspects of emotion, and will therefore have no language or ability to resolve them.  For example, a person who was emotionally abused as a child may have felt intense fear towards its only carer, but as the child was totally dependent on this person, and had no safe person to turn to, the understanding of fear could not proceed in the normal way.  The child’s emerging schema system could therefore not form reliable links between feelings of fear, ideas about fear and normal behaviours associated with fear (eg avoidance).  In adulthood, interpersonal fear and discomfort is felt at a somatic level only, and is not connected to cognition in a way that makes sense to the individual concerned.  As a consequence, the individual is unable to develop functional strategies to resolve the problems causing the fear, and will be likely to continue to experience the somatic discomfort in a chronic way, possibly leading to the development of symptoms. 

Within this model, hypochondriasis and conversion disorder are seen as slightly different from somatization as described above.  In these cases, the particular affected part of the body (in hysterical conversion) or the process of worrying/seeking help regarding illnesses (in the case of hypochondriasis) is functioning as a symbol, helping to organize the emotional schemata that cannot be expressed through conscious or verbal channels.  This is thought to represent a more organized process, with regard to the connection between emotional schemata and somatic experiences, than is the case for somatization.  If this were the case, it could be hypothesised that hypochondriasis and conversion should be easier to treat psychologically, as the representations would potentially be more accessible and amenable to change. 

Freudian psychoanalytic thinking was the theory that was used by De Gucht & Fischler (2002) to generate the distinction between presenting and functional somatization, therefore by definition, Freudian theory could encompass both aspects of this distinction.  However, the modern psychoanalytic view represented by Bucci (1997) seems to contain an assumption that difficulty with emotional understanding and processing underlies somatization, and therefore this perspective fits best with the presenting somatization model. 

Emotional processing

It has been noticed clinically that some individuals go to great lengths to control emotion (Baker, 1989;1995), which can be an obstacle to treatment.  According to Rachman (1980) all events that produce strong emotion need to be processed in some way.  If human beings were unable to do this, and carried every slightest disturbance of emotional equilibrium for long periods of time or permanently, emotions would take up too much attentional space and life would become unmanageable.  It is therefore important that strategies for processing emotion are used.  

However, some strategies for processing emotion are less helpful than others are.  Problems of processing include failure to register and respond to important events, a block in the ability to experience emotions, or feeling overwhelmed by emotion.   

Emotions are a physical experience as well as a psychological or cognitive experience.  In developmental terms, emotions are regarded as primarily biological events, with subjective feelings as a secondarily developed component (Taylor, 1997).  However, by the time a person reaches adulthood, he or she is usually able to recognize a stimulus that has caused an emotion and see the physiological reaction as secondary.  In people who somatize, the provoking stimulus may not be recognized, and the physiological reaction is seen as primary (Tyrer, 1973).  Emotional feelings are misattributed as physical sensations; the person experiences physical sensations but does not experience the cognitive aspects of emotion.  This is unhelpful because the person is then unable to find helpful ways of dealing with the provoking stimulus, nor is he or she able to process the emotion.  It may be the case that, in people who somatize, unprocessed emotion may be experienced somatically with little awareness that emotion is involved. 

Increased activity of the sympathetic nervous system, increased tension in voluntary muscles and hyperventilation can all generate somatic symptoms (Sharpe & Bass, 1992: Kellner, 1985; 1987).  In this way, many symptoms can be experienced, for example, pain; gastrointestinal problems; chest and breathing difficulties.  Other symptoms could develop as a result of repeated arousal or tension associated with experiencing the somatic component of emotions, without the ability to recognize and deal with the emotion.  These symptoms are then incorrectly labelled as physical disorders and the emotional situation is not addressed.  The emotional feeling remains, which leads to continuing experience of somatic symptoms.

A construct called ‘alexithymia’ has been used to describe individuals who appear unable to make links between the physical and psychological aspects of emotion, and can therefore seem unable to express, process, emotion in a helpful way, and in fact, may often deny feeling emotion (Sifneos, 1973; Taylor, Bagby & Parker, 1991). Research has shown many links between ‘alexithymia’ and somatization (Cohen, Auld & Brooker,1994; Deary, Scott & Wilson, 1997; Bach & Bach, 1995) and also pain disorder (Cox, Kuch, Parker, Shulman & Evans, 1994).

Alexithymics, or those who do not process emotion well, may suffer a variety of difficulties connected to managing emotion such as impulsive or addictive behaviour, and it is not always the case that the difficulties are expressed through somatization (Taylor et al 1991.)  It may be that difficulty processing emotions, including alexithymia, are a necessary but not sufficient condition for the emergence of somatization, and that experiences and environmental factors, including family and cultural factors, have a part in determining how these difficulties will be managed. The process by which this may occur in somatization will be returned to later in this paper.

It can be seen when discussing somatization from the emotional processing  perspective that there are elements of both the presenting and functional conceptualizations of somatization.  This perspective does contain the assumption that it is a misprocessing of emotion that leads to the somatization, and to the extent that it sees emotional distress as the cause, it is similar to the presenting somatization model.  However, there is no assumption that repressed emotional distress is responsible, but rather that emotion is experienced in a somatic way only, and that conscious or psychological emotion does not in fact arise, but is rechannelled at an earlier stage.  It could be argued, therefore, that emotion as it is usually understood does not arise, and the symptoms are the manifestation of tension in the present.  This is closer to the concept of functional somatization.  

Overlap between models

There is much overlap and many areas of agreement between the theories of somatization cited here. Emotional processing models e.g. Taylor et al (1991) propose that somatization occurs due to a misperception and mislabelling of the bodily symptoms that accompany emotion, which arise as a result of problems in identifying and linking somatic sensations to cognitive and environmental events.  The modern psychoanalytic perspective, of which Bucci’s (1997) model is an example, also seems to be proposing a similar process with regard to a lack of linkeage between physical and cognitive aspects of emotion, due to a failure of normal emotional development. There are also overlaps between both of these perspectives and the cognitive perspective, in the sense that the link between thoughts and feelings is involved.  In addition, attributions regarding physical experiences can lead to mislabelling of physical sensations as a physical illness, representing an overlap between cognitive theory and emotional processing theory.

There are also overlaps between cognitive, behavioural and social learning theories.  Cognitive theory proposes that somatization could develop through the acquisition of dysfunctional illness beliefs that are developed in the person’s early learning environment, and cognitive theorists would probably agree that modelling, and various types of reinforcements that can occur during both the childhood and adult environment, are important in the formation of such beliefs.

How might somatization develop?

The most helpful approach would seem to be a creation of a synthesis of the above perspectives.  It could be that somatization develops as follows:

An individual’s development of schemata to organize helpful links between thoughts, feelings and somatic sensations is interrupted for some reason early in life. (Psychoanalytic perspective).  The individual is more likely to attribute bodily sensations to physical causes rather than emotional causes (Emotional processing perspective).  Bodily sensations are thus misattributed or mislabelled.  This is particularly likely to occur if emotional causes for the sensations are ignored by caregivers and if physical illness is reinforced by the child receiving care or attention (Behavioural theory). The beliefs of family and culture regarding illness may influence the way this is conveyed to a child through modelling and reinforcement, and the child’s own beliefs and attributions will be shaped by this, in conjunction with his or her own experinces. The childhood environment is therefore likely to be crucial (Craig et al, 1993) in affecting whether a person with poorly connected schemata will develop a pattern of somatization disorders, or develop other types of emotional difficulty.

Patterns making a person vulnerable to somatization may develop in childhood, but are likely to persist into adulthood.  The symptoms which develop may focus, at first, on the individual’s unique physiological weaknesses (psychosomatic) but over many years, tension, worry, and focus on the symptoms may lead to the development of a complex syndrome of experience that the patient finds it hard to believe could be linked to emotion, particularly as the person may have never formed a full understanding of how somatic and cognitive experiences of emotion are linked in the first place.

The role of early relationships and attachment

With regard to the developmental failure to form reliable links in schemata, the childhood attachment relationship is likely to be crucial.  It is during the child’s early development, within the primary attachment relationship, that these links are usually formed.  The primary carer assists the child to do this, both by modelling, and through understanding the child’s emotions and acting as a ‘container’ for them (Bion, 1967).  The sensitive caregiver provides feedback of this understanding to the child, both verbally and non verbally, in a manner appropriate to the child’s developmental stage, allowing the child to develop an understanding over a prolonged period of time throughout early childhood.  Eventually the child is able to understand and process his or her own emotions.  An abusive or disrupted attachment relationship, therefore, would leave a child vulnerable to difficulty in this area and would tend to lead to the formation of maladaptive schema patterns (Young, 1990).

However, it may not be necessary for the individual to have been abused or neglected for problems to occur.  If the parent had not formed a good understanding of links between the physical sensations of emotion and the cognitive/psychological aspects, he or she would be unable to help the child to achieve this.  There are a number of ways in which a parent or primary carer could influence a child in this respect.  Two examples are given below.   

If a parent finds it difficult to distinguish between the physiological and psychological aspects of emotion, he or she would be unable to help the child with this distinction.  Suppose the child has a stomach ache triggered by a situation at school.  A parent with difficulty understanding the above distinction may fail to recognize this in the child, and treat the problem as though it were primarily physical in nature. If experiences such as these are continuously repeated, the child may then fail to gain the understanding that some somatic sensations arise as a result of psychological factors, and continue to attribute them to physical causes. This could reinforce this expression, and lead to the formation of beliefs, which will continue to influence the child’s perception of illness and its causes.

Another example of the effect of parenting style on the development of emotional processing could be as follows: a parent may feel embarrassed by the expression of emotion, and would therefore not respond favourably to emotional displays in the child.  The child would be then less likely to express emotion in a social context and would have less opportunities to learn to understand it as a psychological experience, as well as possibly acquiring a belief that displays of emotion are unacceptable and do not elicit care.  If, in addition, the parent attended and gave care when the child displayed physical symptoms, this would reinforce the likelihood that the child would experience only the physical sensations of emotion.

Some experimental support for the effects of both attachment and environment on the development of patterns of somatization comes from Craig, Boardman, Mills, Daly-Jones & Drake  (1993).  These researchers found evidence that people who somatized in adulthood were more likely to have experienced lack of care as a child.  In addition to lack of care, the participants in the study tended to have experienced either prolonged illness themselves, or the chronic illness of a caregiver.  These are environmental experiences which could cause development of dysfunctional illness beliefs or unhelpful patterns of reinforcement as described above.   


This article has attempted to speculate on likely pathways to the development of somatization disorders by creating a synthesis from existing psychological models of somatization disorder.  This has been undertaken in an attempt to promote further discussion and a better understanding of the possible aetiology of this disorder and hopefully, to further the development of more appropriate therapeutic approaches to these puzzling, and often seemingly intractable problems.

If it is the case, as proposed here, that emotional development is central to initial vulnerability to somatization problems, it may be that theories of attachment and emotional development would be helpful in order to understand at what stage in the developmental sequence of emotion difficulties occur, particularly difficulties which may lead to the disruption of normal connections between emotions and physical sensations.  This could well link with Lane & Schwartz’s theory (1987) of the development of awareness of emotion.  An understanding of the stage at which optimum development became blocked may point to ways for helping the client, where possible, to remove the blocks or repair the disconnected schemata. 

The approach presented here also highlights the importance of attributions and beliefs about illness, as well as reinforcements for behaviour associated with illness, and supports the use of cognitive behavioural techniques, which address beliefs and behaviour.  These techniques are already widely used for these types of problems, often with success.  However, it may be that when these techniques appear to be failing, a better understanding of the impact of emotional schemata, and strategies to help the client establish links between the cognitive and somatic aspects of their emotional experience, would be the key to moving the therapy forward. 

“A sound mind is a sound body”

Juvenal, Satires