This response is in reference to Whitney's comment:

/// I would like to revise that suggestion, and instead propose that we might see here a reflex of another set of South-Indian caste-communities referred to in Tamil as the kuṭimakkaḷ, perhaps best rendered 'people of the village'.  The entry s.v. in the Madras Tamil Lexicon reads (with my transliteration and bracketed translations):

kuṭimakkaḷ , n. < id. +. 1. Sub-castes rendering service in a village, being 18 in number, viz.,vaṇṇāṉ [washerman], nāvitaṉ [barber], kuyavaṉ [potter], taṭṭāṉ [goldsmith], kaṉṉāṉ [brazier], kaṟṟaccaṉ [mason], kollaṉ [blacksmith], taccaṉ [carpenter], eṇṇeyvāṇikaṉ [oil merchant], uppuvāṇikaṉ [salt merchant], ilaivāṇikaṉ [betel merchant], paḷḷi [watchman], pūmālaikkāraṉ [garland maker], paṟaiyaṉ [Dalit, pariah], kōvilkuṭiyāṉ [conch-blower], occaṉ [? another Dalit community], valaiyaṉ [fisherman], pāṇaṉ [tailor]. /// 


1. Re: The term "caste."  You can start from the entry in wikipedia: http://en.wikipedia.org/wiki/Caste. 

2. The Portuguese started using the term "casta" when they came to the southern part of Tamilnadu (in the 16-th century) and tried to learn Tamil and describe Tamil in Portuguese for the sake of further missionary activities. "Casta" was the term they used to describe the various groups of people they noticed in the local society. I happened to have the opportunity to work through the Portuguese manuscript (from 1549 A.D.), by Fr. Henrique Henriques, which describes Tamil in Portuguese. As a "Latin grammarian," Henriques had problems categorizing people's local names when doing the declensions. So … when he lists a set of local names for the first declension, he says, "Names for occupations. Such nouns for occupations are also nouns of castes."  For further details, please see our book "The Earliest Missionary Grammar, Harvard University Press, 2013." 

3. Re: Madras Tamil Lexicon entries. Well … there are discrepancies. For example, if anyone seriously wants to study the history of "castes" or any such thing, one should also look at all their entries beginning with the root/stem of the word. Just for fun, try all the words starting with the stem "kuṭi," and you'll see that words such as "kuṭimakaṉ," "kuṭimakkaḷ," and so on don't have contiguous semantics. 

4. The term "cāti/ஜாதி/சாதி" was never restricted to refer to humans in order to indicate high/low status. It just referred to different types of living beings. 

5. There was no indication of "untouchability" in the Tamil society as reflected in early Tamil poems. 

6. Even when Fr. Henriques (16-th century) listed the names/nouns for people in the Tamil land where he lived and served, he never indicated that there was "untouchability" due to the existence of certain people. 

++++++++++++ 

It IS sad that some modern socio-anthro studies have presented local cultures in ways that the locals cannot understand or accept. To me, it all looks like a cookie-cutter analysis for individual academic progress. 

Thanks and regards,
Rajam