Hi Robert,

Here come some morning musings on this topic. I'm not a lexicographer, but I use their products so I guess this is a kind of customer feedback. Thanks for your ideas.

>
>Perhaps some thought could be given to a reverse index to this new work or
>to existing dictionaries - but even that would require checking every single
>entry for accuracy and modern usage (for the English).

My guess is that historically, this has been the usual method, except that the reverse index has been converted to main entries in a separate volume. Both Apte's and MW's English-Sanskrit dictionaries appear to be based on the same material as their corresponding Sanskrit-English works. My bet is that Cone's material would be the best source for an English-Pali dictionary.

If a Pali-English dictionary were digitalized, a simple search function ought to be enough make it function as a halfway decent English-Pali dictionary right off the bat. If the database were structured to optimalize multiple sorts of searches, the same body of dictionary material could be a good lexicon for translations both ways, a synonym dictionary, and a research tool for finding examples of usage. Ultimately this should be on the same DVD-rom as a complete searchable corpus. I'm certain this sort of solution is the wave of the future. Work in this direction has almost certainly been done within better funded areas of linguistics and computer science.

As an addition to the above, English vocabulary items/translations could (in the envisioned tool) be marked with classificatory info along the lines of genus/class. This can be a mulitiple layer hierarchy. For example 'goose' would be marked as being a kind of 'bird'. 'Sword' as a kind of 'weapon'. There are already ambitious databases describing and categorizing entire language-worlds along these lines. With such a structure you could, for example, do a search for all words either meaning 'bird' or which are names of types of birds. You could even search for words connected with birds (eggs, nests, wings, feathers etc) This ought to at the very least, be doable for the ordinary language of narration. Structuring this for doctrinal and theoretical terms would be a very difficult project, however, and any solution would probably always be open to criticism.

As far as printed dictionaries go, I'd be happy if there were an English-Pali dictionary which had the same level of quality as Apte's Students English-Sanskrit dictionary. This functions as a dictionary of synonyms, as an aid for beginners when reading texts together with translations (if you know a word for 'lotus' has to be in that compound somewhere, you look up 'lotus') and as a reference when composing practice texts in Sanskrit. It serves a well-defined purpose within the context of learning to read, and to a certain degree write, classical Sanskrit. A Pali equivalent would be very helpful, even if it were based on a particular relatively homogenous stratum of the lanuage, such as commentarial narrative prose.

Even if such a dictionary were intended to be printed, it would be a good idea to develop it within a database environment which is structured to be infinitely expanded later, both in terms of lexical items, methods of cross-referencing and forms of 'tagging' and sorting items. The theoretical problems have probably already been solved by people working in other fields, and it would primarily be a matter of adapting their work to the special needs of Pali.

If the funding available to Pali studies won't support all this, we'll just have to wait for AI tools to become available to do all this for us. It wouldn't surprise me if computers will be able to crunch a language within the next 50-100 years. Just give it a corpus and some rudimentary grammar, and the program will start asking questions of it's 'teacher'. Some number of days, weeks, months later it will churn out all the dictionaries you could ever want.

best regards,

/Rett