Everything is culture, but culture is not everything

Comments on Oloyede's paper on mental illness and culture

Peter Ventevogel


In zijn commentaar op Oloyede’s paper stelt de auteur dat deze in zijn pleidooi voor culturele sensitiviteit in psychiatrie tot cultureel relativisme vervalt. In zijn betoog bepleit hij drie onderling verbonden noties over cultuur en psychiatrie: dat transculturele validiteit in psychiatrie diverse lagen heeft; dat culturele verandering ook verandering te weeg brengt in ziektebegrippen; en dat cultuur niet de enige factor is waarmee men verschillen tussen groepen verklaart.

This paper is all about culture, though from a critical perspective. Oloyede points to the relatively small degree of cultural sensitivity in psychiatry. ‘Undoubtedly’, he argues, ‘a culturally sensitive psychiatry is needed.’[note 1]The contribution of social science – and cultural anthropology in particular – to the coming of age of transcultural psychiatry is substantial. Founding fathers of psychiatry as Emil Kraepelin (whose nosological work is the conceptual basis of the current DSM classifica- tion), W.H. Rivers (known for his studies in war related psychopathology) and Sigmund Freud used ethnografic data to substantiate their theories. Mid century psychoanalysts as Abraham and Devereux studied the influence of culture on psychopathology. Contemporary transcultural psychiatrists as Kleinman, Fabrega, and Kirmayer, draw heavily upon insights of medical anthropology.
One can hardly disagree. But in the course of his argument, it seems that to him cultural sensitivity amounts to cultural relativism. It gives the paper a scientific naiveté that I cannot endorse. I will elaborate on some issues Oloyede pays little attention to, while to me they seem of vital importance for a good understanding of the complex relationship between culture and mental illness. Firstly I will discuss the transcultural validity of psychiatric disease concepts in cross-cultural research. Secondly, I would like to elaborate on the vanishing boundaries between cultures and the limitations of the concept of ‘culture bound syndromes’. Finally, I would like to stress that focusing exclusively on culture as an explanatory factor for differences in psychiatric morbidity among groups, could lead to scientific neglect of socio-economic and political factors that are at least as important as culture.

Transcultural validity

Oloyede’s paper is prompted by the discomfort of a group of postgraduate students who were recruited as interviewers for a survey utilizing biomedically driven research instruments. The students felt a lack of correspondence between the employed research methodology and the population for which it was used. This is an important issue. Cross cultural research should include a critical examination of the appropriateness of the methodology, examining whether the research instruments provide the data we are looking after. In epidemiology this process is called ‘cross cultural equivalence or validity’. In a classic article Flaherty et al. (1987) describe five aspects of cross cultural equivalence. I feel this description might help to clarify these points.

Content equivalence

Any item in a questionnaire or checklist should refer to a phenomenon that is relevant to the cultural context of the respondents. For example, a questionnaire designed by the World Health Organisation (WHO), that was to be used worldwide asked, in the section on substance abuse, about the use of cocaine. In Afghanistan this is a substance that is not available, while the use of raw opium is prevalent but is not covered in the questionnaire. The SCL-90 (Symptom Checklist) – a self rating questionnaire commonly used to detect psychopathology – includes questions about avoiding going to the cinema or to travelling on public transport. In the whole of eastern Afghanistan there are no cinema’s, and some villages are inaccessible by road, so posing these types of questions about these issues would obviously fail to gain relevant information on the psychopathology of Afghan individuals.

Semantic equivalence

After translation of a question the original meaning, including its associations should be roughly equal to the original question. This is sometimes difficult to achieve. Some cultures lack equivalent words for emotional states such as feeling depressed or anxious. In a Peruvian study on migration the English ‘to have an adventure’ could not be translated literally because the Spanish word aventura was associated in the Peruvian context with sexual encounters. This connotation was absent in the original Anglo American context (Flaherty 1987).

Technical equivalence

This refers to the methods that are used to collect information. There can be differences among groups in their response to ‘closed questions’ (Boniface & Burchell 2000). In some cultures it is considered impolite to say ‘no’, consequently largely increasing the chance that bias may occur. Questionnaires designed as self-rating scales have to be read aloud by an interviewer when the participants in a study are illiterate. In addition, the presence of an interviewer may seriously affect the quality of the data. For example, people might be less inclined to answer positively on sensitive or shameful topics, which they might have found less difficult with a self administered questionnaire.

Criterion equivalence

Criterion equivalence refers to the instrument’s capacity to assess a variable in both cultures under study and to the equality of the interpretation of the results. This is doubtful in items in standardised questionnaires that in the original American context refer to one of the somatoform disorders, but in a Latin American context tap a cultural idiom of distress not necessarily identical with a somatoform disorder.

Hinton et al. (2000) found in their study on panic disorder among Cambodian refugees in the USA, that the leading question on panic attacks in the SCID (a standardized diagnostic interview using DSM criteria) does not connect clearly to the experiences of the Cambodian patients. The question: “Have you ever had a panic attack, when you suddenly felt frightened or anxious or suddenly developed a lot of physical symptoms?” was found to be an inadequate ‘panic probe’ in their study. Often a Cambodian patient answered “no”, but upon further inquiry it was discovered that they did suffer panic attacks. For the Cambodian refugee the question did not clearly state the most culturally relevant symptoms of a panic attack. In the revised instrument the researcher asked the patient: “Do you sometimes experience orthostatic dizziness?” When the patient confirmed this, then additional questions were asked about other symptoms related to panic attacks.

Allow me to provide another example: some DSM items mean to refer to an abnormality but in some cultural situations refer to normal behaviour. For example, ‘avoidance of leaving the house unaccompanied’ might in the Dutch context be associated with agoraphobia, while in the context of Pasthun culture it does not if the respondent is female. It refers to normal behaviour. Instead, when a woman answers the question affirmatively this would demonstrate a sign of seriously deviant behaviour.

Problems with criterion validity have compelled the WHO to use different cut-off points for the Symptom Rating Questionnaire. This instrument is designed to detect possible psychopathology among primary care attenders in various cultures. The cut off point is the minimum score to be counted as a ‘case’. The cut off point varied from 4 in Sudan, to 11 in Colombia, with the Philippines, India, and Brazil somewhere between these cultural extremes. For each population a culturally appropriate cut off point has to be defined (Kortmann 1986; cf Kortmann 1990).

Conceptual equivalence

Conceptual equivalence is the crux of the matter: Is an illness construct designed in a specific cultural context valid in another cultural context? In their search for cultural idioms anthropologists seem to be pre-programmed to answer this question with ‘no’.

Oloyede describes the problems of the Xhosa interviewers with a question about headaches. He outlines how this difficulty was regarded by the project leader as a problem with semantic equivalence, while the students seem to have had doubts about the conceptual validity and suspected that this was an example of what Kleinman (1987) has called the ‘category fallacy’. This term is used to describe what happens when diagnostic criteria are used in a culture where these symptoms do not constitute an illness entity. It can be illustrated by the hypothetical example of using the constituting criteria of the Indian folk diagnostic term dhat in a population of New York males (Obeyesekere 1985). The symptoms of dhat are a whitish discoloration of the urine, nocturnal discharge of semen, feelings of weakness, and anxiety. A self administered questionnaire with these symptoms would undoubtedly trace a number of persons who fulfil the diagnostic criteria of dhat but are they indeed dhat patients? Of course they are not, because the concept of dhat is invalid in downtown Manhattan, and the symptoms do not constitute a syndrome, instead are a meaningless constellation.

The analogy with cross cultural psychiatry is tempting: “How can we be sure that the condition afflicting a Yoruba tribesman in Nigeria is the same as the disorder afflicting a lawyer in New York or a fisherman in Nova Scotia? Furthermore who is to say what actually constitutes the bedrock of depression: emotions or mixture of emotional and bodily complaints with no clear organic cause?” (Kleinman & Cohen 1997: 76). It seems that we can come to a conclusion now, and sweep away all studies using DSM criteria in a non-western context. The only things we need are ‘thick descriptions’ of culturally unique syndromes: each culture its own illnesses and its own classification system.

Or is it possible that people suffer from conditions their culture has no concepts for? Immediately after the lines I just cited Kleinman and Cohen present an impressive list of DSM disorders and claim that these occur on a global scale: “Such questions aside, several maladies, including organic mental disorders, substance abuse, depression, manic depression (bipolar disorder), various anxiety disorders and schizophrenia, are almost certainly global” (Kleinman & Cohen 1997: 76).[note 2]Kleinman and Cohen assume that many of the diseases in the domain of psychiatry occur worldwide, in various different contexts. This does not necessarily mean that the expression of the diseases is always and everywhere identical. It is clear that there is cultural variability in the phenomenology of psychiatric diseases. Most transcultural psychiatrists agree that when the etiology of a disease is more biological the cultural variation in symptom expression is less. The conditions most clearly caused by biological factors, such as delirium or dementia, look the more or less the same in various different contexts. On the other extreme of the spectrum we find illnesses such as dissociative trance disorders and personality disorders, where the complex interaction between culture and psychopathology is highly visible. For these disorders understanding of the cultural context is essential for any understanding of the condition; cultural factors play a key role in the etiology, in the precipitation of symptoms, in the surrounding circumstances, and the interpretation of the behaviour (Alarcon 1996). Illnesses such as depression and schizophrenia are somewhere in the middle of the spectrum. But to avoid any misunderstanding: Also in severe illnesses culture plays a role. Also in illnesses that seem to be ‘purely cultural’ biological factors play a role. In personality disorders – so visibly interwoven with the cultural context – hereditary and other biological factors contributing to the genesis of these disorders have been identified. In schizophrenia, with its clear disturbances in the cerebral neurotransmitters, cultural factors do play a role, as Fabrega (1989b) has convincingly described.
In my opinion ‘cultural uniqueness’ of illnesses is neither a principle nor a prerequisite. It is a hypothesis that could be tested. In emphasizing the uniqueness of indigenous cultural syndromes one can easily overlook similarities with syndromes in other cultures, or syndromes as described in psychiatry. A brief illustration of this point: Rasmussen (1992) presents an in-depth account of the condition of tamazai among the Tuareg. Persons who suffer from tamazai tend to remain alone in their tents, and avoid contact with others, and appear withdrawn. The Tuareg think the syndrome is provoked by situations of sudden change, such as the sudden death of a loved one, or the disruption of personal routine and social support. As Leff (1994) comments, it is compelling to compare the syndrome with the psychiatric concept of ‘depression’. Rasmussen does not investigate the similarities and differences with depression, and this is a missed opportunity.

One of the challenges of cross cultural psychiatry is to bridge the gap between anthropological accounts of cultural expressions of suffering, and the rapidly expanding body of psychiatric knowledge. We need a culturally informed epidemiology in which social scientist, psychiatrists and epidemiologist collaborate in research projects (Weiss 2001; De Jong & Komproe 2002).[note 3]The marriage between anthropology and epidemiology could be fruitful, but the partners might need some relation therapy. Some anthropologists tend to neglect the objective world, or would wish it away, while most epidemiologists would dispense with subjectivity entirely if they could (Hahn 1995: 101).
Before application in a cross cultural setting psychological tests or psychiatric rating scales need to be contextualized and adapted through qualitative research methods (Van Ommeren et al. 1999). A good example of fruitful collaboration of cultural anthropologists and psychiatric epidemiologists is the work on ataques de nervios in Puerto Rico and the relationship of this ‘idiom of distress’ with psychiatric morbidity. Ataques de nervios are characterized by episodes in which a variety of symptoms can occur such as trembling, palpitations, a sense of heat rising to the head, and numbness (Guarnaccia et al. 1989a). Ataques de nervios are not a circumscript set of symptoms, but are in essence a culturally sanctioned explanation for emotional experiences that Hispanic individuals otherwise explain, or cannot, or do not want to control (Salmán et al. 1998: 241). It is not simply a culturally shaped version of a specific psychiatric disorder, though it is associated with a higher psychiatric morbidity. In a large study in Puerto Rico (Guarnaccia et al. 1993)subjects with ataque were four times more likely to meet criteria for a psychiatric diagnosis, particularly affective and anxiety disorders. There is a salient connection with panic disorder, but also with posttraumatic stress disorder and other anxiety disorders (Lewis-Fernandez et al. 2002).

Changing cultural boundaries

An important limitation of Oloyede’s paper and relativistic cross-cultural psychiatry in general, is that it fails to recognize that cultures are dynamic. The past decades have shown dramatic cultural changes. Cultures no longer live in isolation from one another. Cultures are integrating values. Can we still speak about ‘Ndembu diagnosis of sickness’ as Turner could? Can we talk about ‘Yoruba belief’, or about the ‘two basic etiologies in Hausa culture’? Or are we trying to collect butterflies that have since long lost their natural niche?

I feel somewhat uncomfortable with how authors reiterate examples of ‘culture bound syndromes’, as amok, that might have little to do with everyday reality of clinical practice. Syndromes exclusively confined to specific cultures are rarities, if they exist at all. The classical examples of culture bound syndromes have either become extinct in the contemporary world (who ever sees a patient with piblotoq nowadays?), or moves beyond their original cultural borders and evaporate in a globalizing world.

Latah or latah-like syndromes are found in many parts of the world, from Siberia to Japan to Thailand to the Philippines. Another example is koro, the syndrome of the shrinking penis, originally described in certain Asian cultures; but koro-like syndromes are now found all over Asia, while increasing numbers of the syndrome are described among other groups including white Caucasian males (Chowdury 1996). Even anorexia nervosa, hailed as the example of a syndrome fundamentally connected to the Euro-American world is found not to be limited to Western societies (Lee 1996).

Culture bound syndromes are not purely ‘cultural’. They lend themselves to universal causal principles. Why wouldn’t they? The notion that culture bound syndromes cannot be physiologically explained because of their assumed specificity to cultural boundaries is untenable. Are persons suffering from a culture bound syndrome supposed to not have a body? Why is there this fear of acknowledging the existence of universal psychobiological mechanisms underlying culturally patterned behaviour? As Fabrega writes: “the discovery of a root lesion or biological marker for a psychiatric disorder would in no way deter the efforts of contemporary cultural psychiatrists from explicating the influence of cultural factors in affecting what a disorder looks like and how it is brought about or played out” (1989: 423).

This brings me to another aspect of cultural change: the developments in psychiatry as a science. Oloyedes thinks that psychiatry is now at last moving away from the dark era of biological determinism towards the sacred of holism. I wish I could agree. It is misleading to depict psychiatric science as an originally biologically oriented enterprise culminating in the DSM-III, while now at last, the dawn of liberation looms because of the inclusion of some social and cultural factors in the DSM-IV.[note 4]In this perspective I would like to correct the image Oloyede draws of the state of the art in psychiatry. Firstly, I do not know who Oloyede told about the “limited success obtained from treatment using psychotropic medication” but it is far from the truth. Secondly, clinical psychiatry, since its inception, is well aware of social, cultural and economic factors influencing onset and outcome of mental disorders. The pendulum in scientific interests is going toward the other direction, pointing at biological and genetic factors, but I still assume that most psychiatrists will not agree with Andreasen that psychiatry merely seeks to identify the biological factors that cause mental illness.
There have always been schools in psychiatry emphasizing the psychological and social roots of mental problems. Think of Freudian psychoanalysis, or the ‘common sense psychiatry’ of Adolf Meyer in the US and the rise of social and community psychiatry in the last few decades. I am afraid that the ‘anthropological understanding of mental illness’ is not at all regarded as ‘new’ by psychiatrists but on the contrary, is easily set aside as old fashioned and irrelevant. Mainstream psychiatry profoundly distrusts anthropological knowledge because it is discredited by cultural relativistic claims pushed to the extremes. To cite Fabrega again: “From the legitimate claim of possible differences in personality, social behaviour, illness pictures, and other parameters of illness, because of cultural differences, the derived illegitimate claim was that western psychiatric illnesses are but products of social labels and cultural conventions, and hence are psychobiologically fictive” (1989: 418).

… But culture is not everything

In the past decade transcultural psychiatrists have reached consensus that the universalist and cultural relativist approaches have to be integrated (cf. De Jong 1992; Patel 2001). Studies comparing local concepts of illness with international psychiatric classifications can provide valuable insights in the relative contribution of culture to mental illness. The relationship of major depressive disorder and cultural idioms of distress is studied in Bengal (Chowdury et al. 2001), Rwanda (Bolton 2001), and Zimbabwe (Patel 1998). The study in Zimbabwe found a clear association between the Shona concept of kufungisisa (‘thinking too much’) and high scores on a depression scale. The results also indicated that multiple somatic complaints such as headache and fatigue are the most common presentations of depression. When specifically asked many patients freely spoke of cognitive and emotional symptoms. The study also found a strong association between depression and certain forms of anxiety (Patel et al. 2001). Cross cultural studies such as this suggest a clear lack of distinction between depression and anxiety in primary care. This poignantly demonstrates the need to reconsider certain boundaries between diagnostic categories in the DSM. Incorporation of local illness categories in research instruments can thus effectively challenge the psychiatric classification.

The new culturally informed epidemiology does not only challenge psychiatric classification. It also challenges the position of ‘culture’ as the primary factor in explaining the differences in psychopathology among culturally different groups. Socioeconomic and political factors might be far more important. Studies in Brazil, Zimbabwe, India, and Chile found a consistent and significant relation between impoverishment and depression (Patel 2001). Other factors that were found cross-culturally associated with depression are gender-related life events, education, and economic deprivation.

Recent studies in four post-conflict settings in low income countries identified risk factors for the development of posttraumatic stress disorder. These risk factors were similar to those found in western studies: conflict related trauma, torture, violent death of a family member, separation from family, alcohol abuse of parents, youth domestic stress, daily hassles. Interestingly the pattern of these risk factors differed per country (De Jong et al. 2001) and in this way findings of cross cultural studies point at the importance of contextual differences in the study of psychopathology.


This discussion has emphasized three interrelated notions: transcultural validity is a multilayered concept; cultural changes influence illness conceptions; and culture is only one of the factors that may explain differences amongst groups. Culture is pervasive. There is culture in everything but culture is not everything. Naive cultural relativism is potentially dangerous. In its extreme form it could lead to therapeutic nihilism (‘every treatment has the right to exist’) and a romanticized picture of ‘indigenous treatment systems’ resulting in a disregard of patients suffering from conditions treatable by modern medicine.


Peter Ventevogel (1967) is a cultural anthropologist and a psychiatrist working for the Dutch NGO HealthNet International in Jalalabad, Afghanistan. He is coordinator of a new project to integrate mental health in the basic health care structures of the rural areas in Eastern Afghanistan (email: pventevogel@cs.com).

The author thanks Marlene Barber and Marianne Vysma for editing this text.