Note: Most itrans notation used in this article and in the database is in kanTrans, or Baraha convention.
Intro
The Rigveda or Rig Veda (Sanskrit: ऋग्वेद, IAST: ṛgveda; kanTrans: RugvEda) is an ancient Indian collection of Vedic Sanskrit sUktas. It is one of the four sacred canonical Hindu texts - SRuti. The RugvEda saMhita is the core text and is a collection of 10 books (maMDalas) with 1,028 hymns (sUktas) in about 10,600 verses (Riks).
The Anukramaṇī (Sanskrit: अनुक्रमणी, IAST: Anukramaṇī or Anukramaṇikā, kanTrans: anukramANi) are the systematic indices of the vEda recording poetic meter, content, and traditions of authorship of each sUkta and in some cases each Rik.
The Beginnings
I was doing some research on the RugvEda for my previous article "Daasharaajna: The First (Bhaaratiya) Itihaasa", and wanted to cast some of the data in my own way. I figured an Anukramani database would be the first thing available, and hunted around for it, but found nothing worthwhile. So, I figured, I'd make one myself. And hence, I set out on a journey of research, discovery, reading, programming, and, of course, humongous amounts of data entry and recasting the data for better representation in Excel. So here goes the story.
Sources
The primary source for all of the Mandala/Sukta/Rik, Ashtaka/Adhyaya/Varga, the RuShi, the dEvata, and the chaMdas for each sUkta, and (where available) for each Rik was the venerable Rigveda Samhita of Asthana Mahavidvan H P Venkata Rao of the erstwhile Mysore State under the patronage of HH Chamarajendra Odeyar Bahaddur. This monumental work comes in 36 volumes running to several tens of thousands of pages. My base reference for building this database was the little table that Asthana Mahavidvan H P Venkata Rao provides at the beginning of each sUkta, that looks like this:
Fortunately, this book in all of the 30 volumes is available at at least three different locations (archive.org, ndl.iitkgp.ac.in and vyasaonline.com) as independent scans (which helped immensely when one scan quality was bad - I simply switched to another source). I scrolled through the literally tens of thousands of pages and screen-grabbed each of these 1028 tables into a separate document, so I could have a ready source and did not have to scroll back and forth.
Apart from this, I also referred to Shrikant Talageri's Rigveda: A Historical Analysis, and somewhat infrequently to the Ralph Griffith translation for insights.
As to my primary reference, the text is in kannaDa, it is also an old print and certainly not conducive to OCR - should a reliable kannaDa OCR be available, which I am not aware of. Therefore, all data that I entered had to be visually read and entered by hand into a spreadsheet. What I realized as I went along was that my primary reference is a print from the 1950s, and has a fair few print mistakes (not a large number, given the extent of the work, but they certainly exist), which meant I had to flip those tens of thousands of pages multiple times, to either read the kannaDa translation of the sAyaNa BAShya, or in some cases, go directly to the sAyaNa BAshya to get the gist of the matter in question.
Despite everything, I found only 10522 overall Riks, compared to Talageri's count of 10552 - I ran short on about 30 Riks in Mandala 1 (1976 vs 2006), and therefore 30 short overall (10552 vs 10522). I cross-checked with the Griffith version and the numbers tally with mine to 10522. So, not quite sure about the missing 30.
Process
Eventually, once I finished entering the raw data into a spreadsheet, my work was not yet done. There were several other steps. Each entry that encompassed multiple riks (for example riks 1-6 had the dEvata, iMdra) had to be expanded to separate individual entries (in the example case, 1 iMdra, 2 iMdra... 6 iMdra and so on). In order to stay consistent on this, I wrote up a Python program to take do this for each of the RuShi, dEvata and chaMdas entries. Then came the additional data for each Rik.
Chronology: Talageri persuasively argues over the chronology of the RugvEda, dividing it into 4 (or actually 5) time-periods: the Early, Middle and Late periods, along with a Final Period and a 'General' period that may have mixed or unclear chronology. I have picked up his (Talageri's) assignment of the chronology directly.
Families: Much of this is clear from the name (to get the appropriate RuShi name, in many cases, I had to delve into the sAyana BAShya), however, here, while much of the data agrees with Talageri's, I found a handful of assigned families not backed up in the anukramaNi BAShya. Now, I'm unaware of how Shrikant Talageri has assigned these families, but I only assigned them where the data from H P Venkata Rao's book was solid. In cases where it wasn't, I left the family column alone - blank/unknown. Thus this column will show some variance with Talageri's numbers. This variance, however, will be very small - < 2% overall, in my expectation. The rest of the data agrees with Talageri's.
References to Composer: This is straight from Talageri's data. No changes and not entirely verified with the original.
References to BArata/pUru/tRuShi rAja/Leader/RuShi: Directly from Talageri's data, with minor additions based on a basic grep from the Griffith text (easier to grep English text, however, the data is not 100% reliable. Should be close enough for statistical purposes, though)
References to Early/Middle/Late period personalities: Straight from Talageri's data
References to vamSa: Initial seed from Talageri's book, with additional basic scan of the Griffith text. Should be okay for statistical purposes
References to Locations and Rivers: Straight from Talageri's book
References to Flora and Fauna: Seeded from Talageri's book, with additional references from a basic scan of the Griffith text.
Special sUktams: This was a collection of various sources, plus within my primary reference - wherever I could find them.
All of these extras were add-ons to the original database, and were added on one by one with the help of another Python script. Should there be further additions of a similar vein, these, too can easily be tacked on to the database.
Now with all these completed, there are in some cases, references to RuShis, dEvatas or chaMdas on the same Rik. In order to clearly delineate this, I had to enter these multiple references with a " + " separator, split the columns based on this specific separator in Excel Power Query, Unpivot these split columns, then finally replace the duplicate names (with .1, .2 extensions) with the same name for homogenous data.
In Conclusion
There are multiple sources of errors, despite multiple and laborious cross-checks - print errors from the H P Venkata Rao books, transcription errors from the original kannaDa text image to Excel, issues (if any) with Talageri's data that has been borrowed, and any that crept in during the scans of the Griffith text.
With all this done, this database will be freely available to anyone (link). If you find issues in the database, do let me know through comments, hopefully with a reference I can quickly cross-check (include images of a book if you are looking at one, or a weblink), so I can fix the main database, and anyone else retrieving this data will have those corrections. On the whole, though, the number of errors should be relatively small (1-3% in my estimate, worst case, given multiple levels of cross-checks), so statistically they shouldn't affect much.
The MainDB tab is the expanded version of the original, with 10522 entries in the table. The UnPivDB is the un-pivoted database for RuShi/dEvata/chaMdas with a attribute/value pair. This is to aid the enumeration of the these quantities when named jointly - ex. when a Rik is dedicated to both iMdra and agni, it can be counted both with iMdra and with agni.
Why
Why did I do this? What can be done with this? Well, the anukramaNi database it its own thing, hopefully an easy reckoner for someone looking for this kind of information. I know I was.
Let's see a simple case:
A quick count of the dEvatas to whom Riks are dedicated lays bare the general misconception/incorrect/misleading statement that there are more dedicated to agni than to iMdra.
Furthermore, the claim is modified that the agni references are more in the earlier Mandalas as compared to the later ones. A quick Mandala-by-Mandala comparison shows this, too, is mostly untrue. Except for Mandala 9 (marginal) and Mandala 5, dedications 3:2 in iMdra's favor. If indeed agni is anywhere ahead, it is in the late period, in Mandala 5 and the late period sUktams of Mandala 1, but that is more than made up for in Mandala 8 (of the same time-period). Interestingly, Mandala 9 is dedicated almost completely to pAvamAna sOma, to the exclusion of every other dEvata, so iMdra does not feature at all in this late-period Mandala, and agni, only in 8 Riks out of ~1100.
Even further, when divided by time-period, in the Early, Middle, Late and General periods, the ~3:2 dedication in favor of iMdra is consistent, laying bare the general misinformation peddled since it is nigh impossible for anyone to actually cross-check.
With the main database, we find, for example, a total of 2739 Riks are dedicated to iMdra exclusively. However, delving deeper, we find that there are 3144 dedications to iMdra altogether - exclusive and joint dedications combined - this can be gleaned from the unpivoted database. Likewise, the combined number is 2061 for agni, still holding to the same 3:2 ratio as before.
Not to mention, fancy charts like this can be created at a click - for those who prefer pictorial representation:
Extensive statistical analyses of any kind may be conducted with the database.
No comments:
Post a Comment