The article deals with the problems of the study of caste in social anthropology. Caste, which has continuously been the essential feature of the Indian society, is also the main issue of discussion between the two groups of social anthropologists: those who traditionally (from the birth of sociology) believe that the caste is a special case of a very strict social stratification and can exist in different societies (of which India is a case par exellence) and the revolutionary view of Louis Dumont who contended that it is a uniquely Indian phenomenon. After a careful analysis of both the results of field studies and the main Hindu sacred literature, Louis Dumont stated that the term of caste could be used only for studies of Indian society since no other so called caste case structure outside India (to be precise, outside Hinduism as the ideological source) has the essential features of Indian caste system: perfect hierarchy which is based on the opposition of purity and pollution and on the separation between status and power. For these conclusions L. Dumont was and continues to be quite fiercely criticized, called Indologist, Indophile and the supporter of caste oppression. The arguments against the view of Dumont by Gerald Berreman, one of the constructive critics, show that Dumont has in reality overlooked some of the dynamic features of the caste system - the eternal struggle for a change by the oppressed castes, and that he ignored the disagreement of the lower castes with the brahmanic view of the caste. By emphasizing the variety of caste systems in India itself, Berreman proposes not to limit the use of the term of caste to India, but to rather expand the universal notion of caste as that of a case of strict social stratification plus the sociocultural interrelation. The two opposing views can be seen both as representing the static and dynamic aspects of the caste system, as well as the views ’from above’ (the idealized point of view of religio-cultural authority) and ’from below’ (the point of view of the oppressed castes which are not allowed to disagree by their religio-cultural authority).lt is suggested that the prospects of the study of a caste lie somewhere in between the two paths and that social anthropology, being the best equiped method to research caste, is still very much a domain of Dumont whose revolutionary ideas provoked the most fertile critique