It is difficult to believe that Witzel is serious in his incredible assertions [about the Sarasvati river]... And when other verses do refer to a riv… - Shrikant G. Talageri

To see some really “inconsistent statements” and “cavalier” establishment of “divisions” of the composer families of the RV, the reader should read Witzel’s 1995 papers, where Witzel shows himself to be completely and (as his present statements show) irretrievably lost at sea (see pp. 446-449 OF MY BOOK): there, at one point, he “wants to limit the clans involved in the composition of the Rgvedic hymns” to only three families, the Vishvamitras, the Atris and the Angirases (in the third of which, he includes all the other Rishis); and, at another point, his broom sweeps all the Rishis in Mandala 8 into two “divisions”, the Kanva and the Angiras. At another, he counts the Vishvamitras in the Bhrgu family, and then goes on (in the absence of even the faintest hint to this effect anywhere in the RV, or even in any subsequent text) to place Vishvamitra at the head of the coalition against Sudas in the Dasharajna battle (not to mention minor(?) slip-ups like treating the Shaunakas as non-Bhargavas, and Ghora as a son or descendant of Kanva)!

A detailed and path-breaking analysis (HOPKINS 1896a) shows large categories of words found in the Late books (1,8,9,10, and often 5), but missing in the Early (6,3,7) and Middle books (4,2) except in a few stray hymns classified by the western academic scholars as Late or interpolated hymns within these books. These include such categories as words pertaining to ploughing or to other paraphernalia of agriculture, words associated with certain occupations and technologies (and even with what could be interpreted as the earliest references to the castes), words where the r is replaced by l (playoga and pulu for prayoga and puru), a very large number of personal names (not having to do with the name types, common to the Rigveda, Avesta and Mitanni records, analyzed by me), various suffixes and prefixes used in the formation of compound words, certain mythical or socio-religious concepts (Sūrya as an Āditya, Indra identified with the Sun, the discus as a weapon of Indra and the three-edged or three-pointed form of this weapon, etc), various grammatical forms (cases of the resolution of the vowel in the genitive plural of ā stems, some transition forms common in later literature, the Epic weakening of the perfect stem, the adverb adas, etc.), particular categories of words (Soma epithets like madacyuta, madintara/madintama, the names of the most prominent meters used in the Rigveda, etc.), certain stylistic peculiarities (the use of reduplicated compounds like mahāmaha, calācala, the use of alliteration, the excessive use of comparatives and superlatives, etc.), and many, many more.

But this, besides being seemingly "possible" (by straining the credulity of even the most credulous and partisan reader to the utmost limit) only in respect of a very few names, would not help in explaining the almost complete absence of Western geographical data in the Early Books. Therefore, Witzel also tries to transfer eastern geographical data to the west,.... or by creating dual entities (eg. an Eastern Haryana-Sarasvatī, as well as a Western Afghan-Sarasvatī, both referred to in the Rigveda, with Witzel being the only person possessing the key to distinguish which Sarasvatī is being referred to in which verse.