The Sanskrit term jati literally means ‘birth’ and comes to mean a ‘specific group’ or ‘species’. Its variations in the different modern languages of India refer to the everyday social groupings we would normally understand as castes, including those such as are shown on the map of Bisru in the case study. However, it is important to recognize that the idea of jati can operate on a number of levels, from a rather wide grouping across vast regions of India, to a locally-based group of much closer kinship. A significant marker of jati identity is provided by marriage patterns.  Jatis are generally endogamous groups, although this pattern can change according to the ‘level’ of jati operating.  Some broader groupings allow for a form of endogamy which may increase the status of one (sub-) jati in relation to another.  The Patels of Gujarat provide a good example (Pocock, 1972; see also Tambs-Lyche, 1980).  These points indicate that the idea of jati is relatively fluid; the relationship of groups in any particular region or locality can change, according to a number of factors.

One key factor is economic change.  Jatis which are gaining in economic strength may shift their relations with other groups in a region as their power grows.  This is one way, for example, that we can interpret debates over the status of non-Brahmin castes in nineteenth century Maharashtra (see O’Hanlon, 1985).  British rule led to a period of rapid change as the regional economy was more fully integrated into the global networks of the Empire.  This change provided the impetus for social changes through which existing relations of power were challenged.  In particular, some low caste Mahratha-Kunbi cultivators looked to challenge the dominance of local Brahmin castes by projecting their history as a Kshatriya history (in particular, this was the strategy of the activist Jotirao Phule (1826-1890) – see O’Hanlon, 1985).  This points us towards a critical issue in understanding the operation of caste in India.  Frequently, the relationship between jatis in a locality or region is established by reference to varnaVarna in this sense operates as a kind of theoretical framework for the articulation of relations between jatis, and can provide the rationale for any changes in these relations.  Generally, the processes of change associated with this reference to varna are known as sanskritisation.  The famous Indian anthropologist M.N. Srinivas used this term in his book Religion and Society amongst the Coorgs of South India.  He wrote:

The caste system is far from a rigid system in which the position of each component caste is fixed for all time. Movement has always been possible, and especially so in the middle regions of the hierarchy. A low caste was able, in a generation or two, to rise to a higher position in the hierarchy by adopting vegetarianism and teetotalism, and by Sanskritizing its ritual and pantheon. In short, it took over, as far as possible, the customs, rites, and beliefs of the Brahmins, and the adoption of the Brahminic way of life by a low caste seems to have been frequent, though theoretically forbidden (1952: 30; see also Srinivas, 1956).

One key way of making such a change is by demonstrating a genealogy which links one to a higher position in the varna hierarchy, as in our example of Phule above.  Other ways include changing eating or occupational practices. The latter are significant because jati is often associated with norms of ritual purity and pollution.  These norms usually relate to what the anthropologist Ursula Sharma refers to as ‘forms of bodily communication’ (1999: 36).    The kind of things that groups eat, and the kind of work that they do, can provide markers of their relative ritual purity.  Vegetarianism and the consumption of dairy products are generally perceived as ritually pure, meat-eating of various kinds and the drinking of alcohol as relatively impure.  In terms of occupation, working with dead or waste matter such as tanning, sweeping or cutting hair is often perceived as impure.   This appears very similar to the views in Dumont’s Homo Hierarchicus which we criticised above, but it is important to note that purity and pollution in the sense mentioned in this section provide a kind of language for negotiating the relative position of jatis in specific localities, rather than the universal governing principle of a ‘system’. Particularly in modern urban environments, one can easily see how these markers of status related to jati may be dissipated.  As a result, we might expect that this identification by jati  would become less significant in such circumstances.  Many writers have demonstrated, however, that jati status operates in a particularly intense form in urban environments, as it provides networks of kinship which can be a major support to survival and progress.  The language of purity and pollution may be retained in these situations, even where the practice of ritual restrictions has fallen away.