Skip to main content

How to read the PW (Böhtlingk's shorter dictionary)

Böhtlingk's Sanskrit-Wörterbuch in kürzerer Fassung (PW, seven parts, 1879–1889) is the dictionary most readers of German actually keep at hand: the whole headword coverage of the seven-volume Großes Petersburger Wörterbuch, corrected and extended, at a fraction of the bulk. But "shorter" was achieved by a set of deliberate conventions — an asterisk that flags dictionary-only words, citations that assume you can reach the larger work, accents only where the sources have them. This page reads the front matter so you don't have to learn those conventions by being burned.

For online lookup mechanics and the entry markup see the PW dictionary guide; the complete front matter, with English and Russian translations, is in OCR'd prefaces.

What the Foreword declares​

The single Foreword page of 1879 (English translation) sets out the contract with the reader:

The new dictionary was thus not meant to supplant the older one, but … to improve where something was to be improved, and to add where something was lacking, while, by the omission of all the citations and passages given there, constantly reminding the reader that the larger dictionary must remain the principal source.

The operative conventions, in the Foreword's own order:

  • The asterisk * marks "a word, a meaning, a construction or a gender hitherto adduced only by grammarians or lexicographers" — the single most important reading mark in the book. No asterisk means the form is attested in literature; with it, you are looking at the Indian grammatical tradition's testimony only. (A grammarian using a word as an author, though, counts as literature — and the Bhaṭṭikāvya, as a versified Pāṇini exercise, is denied "the name of an independent literary work".)
  • Accents appear only for words that occur in accented (Vedic) texts. Where only a vocative was attested, the accent is left undetermined and the source work is named instead.
  • Quoted passages are gone by design — a PW entry gives you the judgment and the reference; the evidence lives in PWG. The two books are meant to be used as a pair.
  • The work was collaborative: Roth, Kern, Stenzler and Weber are credited as the four main contributors of corrections and new material, with a dozen further scholars named.
  • Spelling variants in व/ब and श/ष/स are listed once, under the older or better attested spelling — check the other spelling before concluding a word is absent.

The sigla you will actually meet​

The "Abkürzungen der Werke" legend (three front-matter pages) declares 302 abbreviations; the digital body attests 209 of them (69%). As in PWG, the shortest sigla mark the most-used sources. Counts are measured occurrences in the digital text (csl-orig/v02/pw/pw.txt), from the site's legend feed pref-legends.json:

#SiglumWork (as the legend declares it)Body citations
1M.Mānavadharmaśāstra (Manu)68,781
2R.Rāmāyaṇa — kāṇḍa 1–2 Schlegel's ed., 3–6 Gorresio's56,864
3P.Pāṇini13,706
4H.Hemacandra's Abhidhānacintāmaṇi, ed. Böhtlingk & Rieu13,185
5Chr.Böhtlingk's Sanskrit-Chrestomathie, 2nd ed.7,328
6MBh.Mahābhārata — parvan, adhyāya, śloka of the Bombay ed.4,710
7Spr.Indische Sprüche, ed. Böhtlingk, 2nd ed.3,272
8Çâṇḍ.Aphorisms of Śāṇḍilya (Bibl. indica)3,259
9Çâk.Śākuntala, ed. Böhtlingk (unless Pisch./Premaḱ. added)3,190
10Âpast.Āpastamba's Dharmasūtra, ed. Bühler3,078
11AV.Atharvaveda, ed. Roth & Whitney2,084
12HemâdriHemādri's Caturvargacintāmaṇi (Bibl. indica)1,986
13Ind. St.Indische Studien (Weber)1,942
14ṚV.Ṛgveda1,909
15Med.Medinīkośa1,870
16Gal.Galanos' dictionary, in Weber's manuscript copy1,737
17Âpast. Śr.Āpastamba's Śrautasūtra — manuscript (Garbe & Roth)1,718
18Bhâvapr.Bhāvaprakāśa, Calcutta 1875 + manuscript1,697
19Kâd.Kādambarī, Calcutta saṃvat 1919 (Kern)1,658
20VP.Viṣṇupurāṇa, Bombay ed.1,546

Conventions the table already illustrates:

  • A superscript number is an edition switch: VP.² (594 citations, not shown above) is the second edition of Wilson's Viṣṇupurāṇa translation, ed. Hall — a different book from VP..
  • A parenthesised name is a channel, not a work: (Kern), (Roth), (Garbe und Roth) in the legend record who supplied or checked the source.
  • Default edition unless marked: Çâk. cites Böhtlingk's own Śākuntala edition unless Pisch. (Pischel) or Premaḱ. (Premacandra) is appended — the same edition-discipline as PWG's R..
  • Note the recension shift against PWG: the Ṛgveda siglum is spelled ṚV. here but RV. in PWG, and MBh. now points into the Bombay edition where PWG cited the Calcutta print. Sigla do not transfer between the two dictionaries unexamined.

The full legend with every key, expansion and count is rendered on the PW guide's abbreviations section.

The grammatical layer​

The legend itself is almost entirely work sigla (301 of 302; the one grammar-classed key is Pischel, de Gr. pr. — Pischel's De grammaticis prācriticis, 25 citations). The grammatical authorities are cited as works: P. (Pāṇini, 13,706), Pat. zu P. (Patañjali on Pāṇini, 172), Kâty. Śr. and kin for the ritual sūtras. But PW's real metalanguage lives inside the entry, not in the legend:

  • the asterisk described above — the grammarian-only flag;
  • the grammatical tags (<ab>…</ab> in the digital text: gender, priv., case labels) — see the worked sample in the PW guide;
  • the citation-free abridgement itself: where PWG argues with the glossators at length, PW compresses the verdict into marks and word order.

So the literacy skill for PW is less "which grammarian is this?" (usually Pāṇini) and more "is this form literature or lexicography?" — exactly what the asterisk answers.

The rare keys are intentional​

302 declared keys against 209 attested is, as with PWG, a typed and explainable gap, not noise:

  • Manuscript sources carry real weight. Âpast. Śr. — a manuscript, per the legend — has 1,718 body citations; Bhâvapr. combines a Calcutta print with a manuscript; Maitr. S. (1,507) was cited from L. Schröder's edition while it was still being prepared for publication — which is why some of its accents only appear in the addenda. The manuscript catalogues are sources too: Verz. d. Oxf. H. (Aufrecht's Oxford catalogue, 181 citations) and Verz. d. B. H. (Weber's Berlin catalogue, 67).
  • The measured residual is typed. The census leaves 89 legend keys with no body match (pref-only); the feed carries a typed residual class on 26 keys: 16 true_unused (mostly external scholarship cited in the front matter, not in entries), 6 ocr_key (Böhtlingk's special letters ḱ, ǵ in keys like Harshaḱ.), 2 ambiguous, 1 rare, 1 ortho. In the H1560 50-key sample the dominant residual cause was orthography — the legend's German j where the body writes y — a transliteration-era artifact, not a phantom source. Full decomposition: pref-only rollup.

Meeting an unknown siglum in PW, assume first a spelling-variant of a declared key (j/y, ḱ/c), then a front-matter-only reference — and remember the full evidence trail for any entry continues in PWG.

Where these numbers come from​

Frequency counts and classifications trace to src/data/pref-legends.json (PW: 302 legend entries, generated 24-07-2026 from the OCR'd front matter × digital body census). The census: crosscheck rollup (PW: 302 keys, 209 attested, 69%, 89 pref-only). Foreword quotations follow the English translation in PWK/prefaces/ (consolidated English edition).

See also​