The strict R^ishi-sharing graph between the 10 RV maNDalas (click on figure to enlarge)
The blue lines represent jamadagni who is only a co-author of a single sUktaM in maNDala 3.
-Note the dense clustering between 8,9,10,1
-Note the lack of clustering of family books 2,3,4,6,7
-Note tendency of family books 5 and 8 to overlap with each other and with 1,8,9 in particular.
The R^igvedic family books can be divided into two types: 2,3,4,6,7 which have few R^iShis and the majority of sUkta-s are compositions of a single dominant R^ishi. The coherence of these books in terms of the language and refrains, do support the contention that they were indeed composed by in large part by single dominant R^iShis. Book 5 and 8, in contrast, are more composite pieces with several R^iShis belong to a single clan (atri and kaNva respectively) and potentially spanning a large time range. They also incorporate more material from R^iShis of other clans than do the former type of family book.
This seems to suggest two different themes in family book collection: 1) The first one (type-I family books) where a single R^iShi composes a large number of hymns to various deva-s and various ritual contexts. For example, viShvamitra has compositions in the context of various rituals like raising the yupa-stambha (3.8), agni-praNayana(3.27), sAmidheni (3.27), AprI (3.4) and mAdhyandina savana (3.32). On the other hand vasiShTha has hymns to every devatA of the vedic pantheon and likewise encompassing various ritual contexts. This would mean that these maNDala-s were generated by a single author, who was more or less recreating new recitations for a wide swath of the archaic shrauta ritual. 2) In the second theme (type-II family books) seen in 5, 8 we do not see a single author composing a wide range or sukta-s and a given author might cover only one rite. Likewise the author might merely cover one or a few deities in his(her) sUkta-s or produce collective vaishvadeva hymns. For example, vasushruta Atreya composed the AprI for the atri-s (5.5), dyumna-vishvacharShaNi atreya composed their agni praNyana hymn (5.23),and the woman R^iShikA vishvavArA atreyI composed their sAmidheni (5.28). A very prominent R^iShi of this clan the early astronomer shyAvAshva Atreya composed 12 sUkta-s but of them 10 are to the maruts and 2 to savitA. Only atri bhauma the founding father of this clan cover a large number of devatA-s. The atri-s also incorporate sUkta-s of different a~Ngirasa-s, kashyapa-s, a vasiShTha and a possible vaishvamitra. The kANva-s, likewise, incorporate other a~Ngirasa-s clans, bhArgava-s, atreya-s, kashyapa-s and mana maitravaruNi. What this implies is that in the case of book 5 and 8 the tendency was to collate compositions over time by different authors rather than create a whole new set for much of the ritual sequence. They also were not averse to collecting material from gotra-s other than their own and incorporating it into the family book.
maNDala 9 stands out as the hub in the above graph with connections to every other book. However, it is most connected to maNDala-s 8, 1, 10 and 5 and least connected to the type I family books mentioned above. In part while this reflects the fact that all these books have a large number of R^iShi-s, there is also high inter-connectivity between most of these books which are highly connected to 9. This suggests that all these books shared a certain process in their creation and editorial history. This is best illustrated in the case of maNDala 1 and 9 which not only share several authors but also exhibit synteny and similar spatial clustering of authors. The first 4 collections in both maNDala 1 and 9 are : madhuchChandas, medhAtithi, shunaHshepa and hiraNyastUpa. Down the line kaNva, praskaNva, nodhas, parAshara, kutsa, kashyapa, ambarIsha/vArshagirA are clustered together in both 1 and 9. While a similar synteny is not seen with the other books their authors are certainly shared with 9.
In contrast, the main authors of the type I family books do not tend to show any presence, or a very limited presence (mostly in the composite sUkta-s attributed to multiple R^iShis) in book 9, despite their copious single author compositions in their own book. For example, vAmadeva gautama, the main author of book 4, is absent from book 9, with only the ikShvAku emperor trasadasyu from book 4 contributing to book 9. Similarly, gR^itsamada contributes to single fragment of 3 R^ik-s within a much large composite sUktaM (9.86), and likewise vishvAmitra’s and bharadvAja’s only contributions are single tR^icha-s in 9.67 . vasiShTha, who is the most prolific vedic R^iShi too is very limited in his contributions to book 9. It is only the descendants of the type-I family books who contribute to book 9 — thus, we see later members vasiShTha, vishvAmitra, bharadvAja clans here. This distinction also extends to some early authors seen in book 1 : e.g. vasiShTha’s brother agastya and the pUru king paruchChepa daivodAsi– It is only their descendants, rather than themselves who are in book 9. Thus the type I family books and their equivalent upa-maNDala-s of book 1 appear to represent a tradition who hardly composed any specialized pavamAna soma hymns, but only their descendants appear to have taken up this practice.
An examination of the remaining authors of book 9 fills in the remain parts of the puzzle. The single most prolific clan in maNDala 9 are the kAshyapa-s with 34 sUkta-s. They have a solid colinear block of 20 sUkta-s, which includes the AprI sUktaM of their gotra. This suggests that at its heart maNDala 9 began not as a soma maNDala but as the family book of the kAshyapa-s. Their AprI is unique in having pavamAna soma in place of one or both forms of agni- i.e. tanu napAt or narAshaMsa. This suggests ( as also proposed by Shrikant Talageri) that kAshyapa-s had a unique sacrificial tradition based primarily on soma rite rather than the fire rite. It thus appears that they were one of the main popularizers of the pavamAna rite during the evolution of the classical shrauta system.
The remaining prominent authors of the maNDala 9 are of course the atri-s overlapping with book 5 and kANva-s overlapping with book 1 and 8 and a group of “new” bhR^igu-a~Ngiras who do not have a notable presence elsewhere in the R^ig with the exception of 8 and 10. These bhR^igu-a~Ngiras contribute about 30 sUkta-s, distinct from those of the atri-s and kANva-s The atri-s and kANva-s are clearly specifically interacting clans. Both of them were ritualists for the Iranian rulers of the rushama clan (8.3, 8.4, 8.51 and 8.30). The atri-s also mention kaNva-s as being hotar-s in sacrifices with them (5.41.4). Hindu tradition also records the love-hate relationship between the bhR^igu-s and atri-s and the shrauta ritual tradition records a chatur-rAtra rite where an atri initiated the bhR^igu aurva in to a soma ritual. Thus, these clans indeed seem to belong to a linked system. However, a point to note about the “new” bhR^igu-a~Ngiras in book 9 is that, like the atri-s and kAshyapa-s, they preserve several sUkta-s of their ancient ancestors of a much older layer like the great R^iShi ushanA kAvya, who is remembered as one of the early ritualists and his father kavI bhArgava. This suggests that the bhR^igu-s and some of these new a~Ngiras were also ancient soma ritualists. But, unlike the kashyapa-s they also had the fire ritual which is at the heart of the main shrauta tradition.
Several models could explain these observations but we believe the following seems plausible. There were two traditions in the early shrauta period: 1) A compositional mode in which each main director of rites composed a new set of hymns, which he used during his tenure for all his major shrauta actions and recited these with his assistants. 2) The editorial mode in which the director of rites collected hymns from his ancestors and memorized them rather than composing fresh collections for everything. He might supplement these merely with pavamAna litanies or a few new ones to show his standing as a vipra. The former spawned the type-I family books and the latter the type-II family books. We suspect that zarathuShTra represented an Iranian equivalent of the type-I activity.
The type-II authors were also linked with formalization of specialized pavamAna soma rite which became the basis for the soma ritual of the classical soma yAga. It is likely they composed their soma-sUkta-s as lyrics on which sAman-s were sung which became the basis of the collections of the bahiSh pavamAna and Arbhava pavamAna songs. At their core these pavamAna litanies have the ancient soma hymns of ushanA kAvya, kavi bhArgava, early kashyapa-s and manu. They are padded with more recent litanies of the descendants of vasiShTha and vishvAmitra along with those of late a~Ngira clans like the kapi-s and vaikhAnasa-s. This shows that the pavamAna songs of the 3-fold soma pressings in the classical shrauta ritual arose from a process of amalgamation around an ancient core with inclusion of new material to more or less account for all major brahmin families. This amalgamation of ancient R^ishi-s of certain clans and descendants of the type-I family book R^iShi-s suggests that an atri-kAshyapa-bhR^igu-a~Ngira alliance introduced these to the latter at a later point. It was only after the amalgamation did the extant maNDala 9 and 1 emerge and their descendants then put together maNDala 10 by means of their new compositions amalgamated again with other assorted old material. A further round of this was again seen in the creation of the khila of the RV. In doing this they were merely following the precedence set by the atri-s and kaNva-s in composing their family books. It spatial and temporal terms, I suspect the pavamAna transition actually represent a second influx of new Indo-Aryan settlers from the more north-western zones of modern Afghanistan. This was prior to the rise of the kuru-pa~nchAla confederation but probably coeval with the rise of the pUru-s (perhaps under bharata daushyanti).
In course of my vedic education I gradually reached these conclusions. I wanted to put this down for long, but then in the middle there was Talageri’s work and that of some white Indologists. Both of these while seeing some of these elements had their own problems. I was mentally distracted by the need to rebutt their theories, but then I thought it was just not worth it. Anyone who studies the veda, and does not dismiss Hindu tradition as meaningless, can see that what I outline above is the essence of the whole thing, and is about as much as we can draw in a cursory analysis. Of course there are numerous specific points concerning specific aspects of the shrauta ritual and to some extant the gR^ihya rituals that we might cover if we feel so inclined in the reminder of our life. All this we fear would become extinct in our clan with me, who is the last somapAyin in my current family.