Introduction and Usage

The data presented here form a cumulative supplemental dictionary of Sanskrit, where the results from lexicographical works postdating the smaller petrograde dictionary of Otto Böhtlingk (pw) and the addenda by Richard Schmidt (Schmidt) are collected in the form of brief quotations. These were glossaries added to editions of texts as well as specialized dictionaries. Glossaries that were obviously not up to standard were excluded. The sources quoted are abbreviated, but never modified or harmonized. Testimonia in these quotations are not updated (for instance, by referring to more recent editions) and only one per meaning was given. The quotations were also not translated and not all testimonia could be checked in detail. Obvious errors and misprints were silently corrected. The NWS gives access to a large body of lexicographical material, but does not completely supersede the glossaries that are its source.

The NWS is also neither a new unified dictionary nor a complete thesaurus of quotations, nothing of the sort could be attempted in the three years of its funding. However, we believe that it may greatly facilitate future efforts in these directions, for it accomplishes a radical expansion of our knowledge of Sanskrit lexicography: The NWS contains 107553 entries—the pw has ca. 150000, Schmidt ca. 12000—which are as a rule new lemmata, new meanings, or new testimonia not found in pw and Schmidt, or indeed Monier-Williams, which very rarely goes beyond the pw.

The task of providing a unified search for all dictionaries and glossaries was tricky and although the search can be used without any theoretical knowledge, some background is helpful in order to make most of the portal, which combines the search in the NWS, pw and Schmidt.

  • The lemmas are as far as possible given in the form known from the pw, since the NWS is technically a cumulative supplement to the pw. Users accustomed to the dictionary of Monier-Williams will have to make the necessary adjustments.
  • Nouns are given in stem form, in case of multiple stem forms the pw was the guide (for instance, śreyaṃs; bhaviṣyant; gomant; ātman; rājan; acikitvaṃs; pratyañc). Past passive participles (e.g. gata) are given under the respective roots, other participles form separate lemmata.
  • Compounds, special spellings, and feminine stems with a significantly changed meaning are treated as separate entries. Please note that the pw lists feminine stems fairly consistently under the masculine form withouth regard for relative frequency, for instance sarasvatī is listed under sarasvant.
  • cvī formations are treated in combination with their respective auxiliary verbs (e.g. praguṇīkṛ).

Each entry in the NWS can have the following additional pieces of information:

structure of a NWS dictionary entry

Some of these items, like synonyms, sigla, testimonia and the like, are active, that is, clicking on them opens a new window, but this feature is self-explanatory.

For a more detailed assessment of the search results, the following remarks about the construction of the portal and its three components, the NWS (left), pw (middle) and Schmidt (right) may be of help. First of all, the object of the project funded for three years by the DFG was a cumulative search of a large number of „Nachträge“ (addenda), which resulted in the NWS (left). The left window will display only information not contained in pw or Schmidt. This means that, for instance, the dictionaries of Edgerton and Grassmann are not contained in the NWS as a complete set of data, merely what is missing in the pw is given. The idea behind the portal was to allow the user to review all three sources of lexical information at a glance. The data of the pw and of Schmidt were given to our project for this purpose by Thomas Malten, Köln. Basically they are identical with the data in the background of the Cologne Dictionary Project, but there are obvious modifications to the presentation on screen, and perhaps more importantly, we have started to apply our own corrections to the data, which will thus fork into a new śākhā.

Unfortunately this history of the portal needs to be kept in mind for using the search tools. Only the addenda in the NWS were fully tagged by the present project, whereas the data from the Cologne project could not be reworked in this way. For this reason, all specifications of search only apply to the left window. The same limitation affects the information popping up during lemma entry: the suggestions are taken from a list of NWS lemmata, not of pw lemmata! It is also worth noting that the popups are not static, but change with further inputs. If one tries to search for vyavahāra, entering the first six letters produces the correct lemma. If one continues entering letters, the system produces further lemmata, for instance compounds as vyavahārapāda etc., which are generated from NWS entries only. The difference is made clear when one actually starts the search, which yields many more hits.

The most problematic area when harmonising the divergent sources was the idiosyncratic form of roots and derivations given by Böhtlingk, who—as is well known—uses kar for kṛ, hotar for hotṛ and so forth. Considerable effort that has gone into not letting the user notice the underlying problems, but sometimes, especially when one suspects that the portal has not given expected results, one should try using the fuzzy search or wildcards. This is, by the way, also a good shortcut for searching feminine forms.

One practical example may clarify the issue: searching vallarī yields only results in the NWS, since the pw lists vallarī under vallari. The problem is solved by using the extended search methods, but in different ways, which can be best understood by searching vallara, vallari, and vallarī, which give differing results. In order to broaden one’s search, we can use the following methods.

vallar? — this allows one arbitrary letter, so the results include vallara, vallari and vallarī.

vallar* —this allows any number of following arbitrary letters and thus in addition catches compounds.

Another possibility is to enable fuzzy search, which allows a few arbitrary letter changes. In the case of vallarī this leads to very diverse results, from gallakī to khallara.


Please mail feedback and corrections to