Dear Gunnar,

You wrote:

> Strictly speaking, yes; but when using Indian scripts, there is an additional problem.
>
> I don't remember anything about Sinhalese script, but I think the system is similar to Devanagari; there, to write "eva.m", you write two signs for the syllables "e" and "va", and then you just put a dot above the second syllable to indicate the niggahita/anusvaara.
>
> To write "evam", on the other hand, you have to write three signs for the syllables "e", "va" and "ma", and then put a hook (viraama) at the third syllable to indicate that the final "a" should be mute.
>
> So in many manuscripts, the dot for .m has been used instead of the more complicated sign for an m not followed by a vowel, to save space and work.
>
> There's no use to do so when using the Latin alphabet, of course.

In my opinion, this is not a problem of different scripts. Rather it is a case of Sandhi.

First, let me discuss the nature of Sandhi for the sake of beginners. Sandhi is a phenomenon that we can find in every language. Take the English sentence, for example, "It is an aeroplane". A native fast-speaking person will tend to pronounce it as "Itisanaeroplane", a continuous sequence of sounds. It is sandhi. However sandhi will not appear in writing because white space will compulsorily separate the words. Another example: the prefix "in" in the words "intolerable", "impossible" and "inconplete" are spelled the same but pronounced differently.

In the case of Pali and Sanskrit texts, words are supposed to be written as they are pronounced. For instance, the acc. singular form "eta.m" of eta(d) will appear differently in different phonetic contexts, e. g., "eta.m sabba.m, eta~n hi, etam atta.m". And the verbal stem "dadaa" will appear differently in different phonetic contextts---"dadaa" when followed by single consonants (e.g., "dadaati"), "dada" when followed by double cons. (e.g., "dadattha") and "dad" when followed by vowels (e.g., "dadanti").

This is why sandhi is much more important for Pali and Sanskrit than in modern languages like English.

Now, you will notice that in the Pali examples I have given, I have assumed "basic" forms for words. "eta.m" is the basic form from which "eta~n" and "etam" are derived. "dadaa" is the basic form from which "dada" and "dad" are derived. Such assumptions are made by both classic and modern Pali grammars so that the language can be learnt and understood easily. However, this is not linguistically accurate. "There is no such a thing as THE form of a word---there are various alternating forms, all of which are equally representative of the word in their particular environments". (W. Allen. Sandhi: The Theoretical, Phonetic and Historical Bases of Word-Junction in Sanskrit. The Hagues: Mouton & Co. 1962, p. 19) E.g., "eva.m" and "evam" are legitimate words. To derive "evam" from "eva.m", or vice versa, is only a grammatical convention.

Therefore it would be an editorial mistake to, say, replace all instances of "eva.m" with "evam". To justify such an act by means of the difference in scripts would be another mistake. For scripts have never been part of the languages Pali or Sanskrit. All classical grammars have defined these languages as systems of sound (saddasattha). The mistake of taking scripts into account in textual criticism has resulted in regrettable mistakes---an editor of a romanized Pali text (I cannot recall the title) clearly says in his introduction that he has made all final nasals consistent as ".m"!

with metta

Ven. Pandita