Skip to main content

How to read Böhtlingk–Roth (PWG)

Böhtlingk & Roth's Sanskrit-Wörterbuch — the Großes Petersburger Wörterbuch, seven volumes, 1855–1875 — is the deepest citation apparatus in Sanskrit lexicography, and the first page of any entry throws its abbreviations at you: R., H., P. 6,1,116, Verz. d. B. H. This page is the dictionary-literacy pack: what the editors say the book is, which sigla you will actually meet (ranked by how often the digital text really uses them), and how to tell a literary citation from a grammarian's rule or a manuscript source.

For the online lookup mechanics (displays, input schemes, entry markup) see the PWG dictionary guide; for the full transcribed front matter with English and Russian translations see OCR'd prefaces.

What the Vorwort declares​

The Foreword to volume 1 (1855) states the design in its own words (English translation):

Our dictionary consists of two elements not entirely resolvable into one another: the material of the word-collections made by learned Indians, and our own collections drawn from Sanskrit literature.

That sentence explains most of what you will see on the page. Indian lexica (Amarakośa, Hemacandra, the Medinī…) are reproduced and checked against real usage, so entries record where the two disagree — the editors insist on keeping even the "residue of Indian lexicography that does not resolve into actual usage" rather than discarding it. The other declared commitment is Veda first: the editors treat complete excerption of Vedic literature as "the right beginning for our work" (Foreword, 2), and read the hymns from the texts themselves — Sāyaṇa and the commentators are demoted to "only one of the aids", not the standard of interpretation (Foreword, 3).

The Foreword also declares the search conventions a reader needs (Foreword, 5):

  • Verbal roots are listed without the vowels ṛ, ṝ, ḷ and without final diphthongs; final -ṛ of nominal stems is written -ar.
  • Denominatives stand under their suffix, not in root form.
  • The accent is marked in a simplified notation (Vedic examples keep the manuscripts' marking), and no verb class is given — it follows from the quoted 3rd sg. present.
  • The work was divided: Roth covered Vedic literature, Suśruta's Āyurveda and plant names; Böhtlingk the rest and the arrangement of the whole.

The sigla you will actually meet​

The front matter legend declares 395 abbreviations; the digital body actually uses 257 of them (65%). Frequency is heavily concentrated — the twenty sigla below account for the bulk of all citations, and the very shortest marks (single capital letters) were assigned to the works cited most. Counts are measured occurrences in the digital text (csl-orig/v02/pwg/pwg.txt), from the site's legend feed pref-legends.json:

#SiglumWork (as the legend declares it)Body citations
1R.Rāmāyaṇa (books 1–2 Schlegel's ed. unless marked)234,443
2H.Hemacandra's Abhidhānacintāmaṇi (synonymic lexicon)197,887
3P.Pāṇini's grammar — a rule, not a text passage95,965
4N.Nalopākhyāna in Böhtlingk's Chrestomathie90,941
5M.Manu's law-book (ed. Loiseleur Deslongchamps)83,397
6MBh.Mahābhārata, Calcutta ed.67,272
7RV.Ṛgveda — cited by maṇḍala, sūkta, ṛc58,146
8Bhag.Bhagavadgītā, ed. Schlegel33,406
9Bhâg. P.Bhāgavatapurāṇa30,744
10AK.Amarakośa28,273
11Kathâs.Kathāsaritsāgara, ed. Brockhaus25,078
12AV.Atharvaveda-saṃhitā, ed. Roth & Whitney21,528
13Śat. Br.Śatapathabrāhmaṇa (Mādhyandina)20,941
14Suśr.Suśruta, Calcutta ed. — vol., page, line20,306
15an.Anekārthasaṃgraha (= H. an.)20,269
16St.Stenzler19,558
17Hariv.Harivaṃśa (vol. 4 of the MBh. ed.)15,689
18Varâh. Bṛh.Varāhamihira's Bṛhajjātaka15,212
19Med.Medinīkośa15,136
20Varâh. Bṛh. S.Varāhamihira's Bṛhatsaṃhitā14,323

Two habits worth forming immediately:

  • A siglum's numbers point into one specific edition. The legend entry tells you which: MBh. means the Calcutta edition; R. without further mark means Schlegel's edition for books 1–2 and Gorresio's for books 3–6; Suśr. is cited by volume, page and line of the Calcutta print. Reading a citation without knowing its edition is how misquotations start.
  • The single letters are earned, not arbitrary — R., H., N., M., P. mark the five workhorses of the whole dictionary. Learn these five and a third of the apparatus reads itself.

The full legend, with every declared abbreviation, its expansion and its measured count, is rendered on the PWG guide's abbreviations section.

Grammarians' rules are not text citations​

The legend feed classes the 395 keys as 393 work sigla, 1 grammar siglum, 1 meta siglum — but the learner-relevant split is functional. A cluster of sigla cite grammatical authority, and their references read differently:

SiglumAuthorityBody citationsHow to read the reference
P.Pāṇini's Aṣṭādhyāyī95,965P. 6,1,116 = book, chapter, sūtra — a rule
Kâty.Kātyāyana (vārttikas on Pāṇini)8,789usually via the Pāṇini apparatus
gaṇa.the gaṇapāṭha in Böhtlingk's Pāṇini ed.8,419membership in a listed word-class
Siddh. K.Siddhāntakaumudī1,407the later grammatical compendium
Pat.Patañjali's Mahābhāṣya (via Böhtlingk's excerpts)1,152the great commentary on Pāṇini
Sch. / Schol."Scholiast" — the commentator on whatever text is at hand—a meta marker, not one fixed work

When an entry cites P. 6,1,116. it is telling you this form is licensed by that rule — a statement about grammar, not an attestation in literature. The distinction matters because the Vorwort's critical stance (usage beats the glossators) applies to word meanings; for word formation the grammarians remain the cited authority.

The rare and manuscript sigla are intentional​

395 declared abbreviations against 257 used is not sloppiness — the gap is typed and explainable, and the legend was never meant to be a checklist of the body:

  • The legend grew with the work. Each later volume opens with its own front matter, including "abbreviations newly added" lists (volume 3's list is where the Buddhist manuscript sources first appear — Foreword, 3-2). Seven of the 27 transcribed front-matter pages are abbreviation legend.
  • Manuscript sources are first-class citizens. From volume 3, A. Schiefner supplied readings from Buddhist manuscripts in the Paris library: SADDH. P. (Saddharmapuṇḍarīka, 171 body citations), UPAG. AV. (Upaguptāvadāna, 15), DAŚABH. (Daśabhūmīśvara, 12), KAṬHINĀV. (3), down to KĀPIŚĀV. and KRIYĀSAM. with a single citation each. A siglum with one hit is a deliberate record of a source consulted, not an OCR artifact. Kâty. Śr. — cited 8,694 times — is likewise marked Hdschr. (manuscript) in the legend, and Weber's catalogue of the Berlin manuscripts (Verz. d. B. H.) alone accounts for 4,677 citations.
  • The measured residual is typed. The census leaves 133 legend keys with no body match (pref-only); the feed carries a typed residual class on 52 keys: 22 true_unused (mostly long bibliographic titles of European scholarship), 12 rare (manuscript and occasional works), 9 ocr_key (transliteration specials like ḱ, Ǵ), 5 spacing (dual keys such as Vop. oder Vopad. — the body always uses the short form, so the double key itself scores zero), 3 ambiguous, 1 ortho (German j for y). The full decomposition: pref-only rollup.

So when you meet an unfamiliar siglum, the order of hypotheses is: a later volume's addition → a manuscript source → a European reference work — and only then a transcription problem.

Where these numbers come from​

Frequency counts and classifications on this page trace to the site's legend feed src/data/pref-legends.json (PWG: 395 legend entries, generated 24-07-2026 from the OCR'd front matter × digital body census). The census itself: crosscheck rollup (PWG: 395 keys, 257 attested, 65%, 133 pref-only). Vorwort quotations follow the English translation in PWG/prefaces/ (consolidated English edition).

See also​