N.B. Links to references and abbreviations are for use with frames.
Abbreviations: BEN=Bengali script, DEV=Devanagari
In DEV and BEN, various schemes have been used to represent Perso-Arabic characters. These schemes involve a various numbers of modified Indic characters. In Hindi, the five dotted characters nuqta-ka, -kha, -ga, -ja, -pha are well-known. These may be termed the 'basic five'. BEN nuqta-ja was authorized by the University of Calcutta in 1936 for use when /z/ is needed in any loan-word.
Current Hindi uses at most the basic five modified DEV characters, but there are also more elaborate schemes. The following chart is based on the scheme of John Borthwick Gilchrist , [1810/1825] and various other works; and the dictionaries of Platts , Forbes , and Shakespear .
Underdots may also be found in other positions. In order to represent all Perso-Arabic signs, two underdots may appear below DEV aa-maatraa, i. These elaborations, and tanwin, probably do not need to be covered by the proposed international standard.
- The dot is placed under a DEV full vowel.
- Dotted DEV va is used by Shakespear in certain circumstances; otherwise ordinary DEV va is used.
- Arabic tanwin.
Gilchrist's Grammar of the Hindoostanee Language (1796) was the earliest printed work in DEV (Naik  Vol.1, p.246).
Gilchrist himself (, p.7) confirms that his forms corresponding to nuqta-ja in the adjacent column are 'all the varieties of [nuqta-ja] in Nagree'. His other unusual forms in Fig.1 are variants of the modern characters shown there. In the
second edition (, p.cxlvii) of his 1804 work, Gilchrist says,
The chief deviations from the naguree of my former publications will appear in the subjoined prospectus, and as they have been adopted by many of the native writers, who follow my plan, their insertion here may do some good, and no possible harm
[Here follows some undotted DEV characters of no direct interest to us, and a few dotted characters agreeing with those in Fig.1 (Gilchrist columns) except that two dots under sa are horizontal under the character, and the variants of ja for 'z', 'zh' are new ones.]
The 100 stories, etc., of the Hindoostanee Story Teller which appear in this edition are given in Roman transliteration, DEV, and Perso-Arabic script. In the DEV section we find the dotted characters fully in use.
Forbes ( Part I, p.vi) makes the following remarks on representing Perso-Arabic letters in DEV:
The plan adopted in this case is to represent the letters in question by such Nagari letters as approximate nearest to them in sound, which in printed books are generally distinguished with a dot underneath, - thus
[here follow the first 14 characters in Fig.1, Dictionary columns]
In some printed books, for instance in Dr. Gilchrist's "Hindee Story-
teller" an attempt has been made to form distinct Devanagari letters for the various forms of the Persian and Arabic z, which it will be observed, are all represented by nuqta-ja; but in reality the object is not worth the labour. ...
This gives us the transition from Gilchrist's complete system to the simpler one of Fig.1 (Dictionary columns).
In his Grammar , Forbes continues the same remarks with:
In the first place, the Hindús, who alone use the Devanágarí character, are sparing in the use of Persian and Arabic words, to one or other of which the various forms of the letter z belong; and, secondly, such words as they have in the course of centuries adapted have become naturalized, or, if the critic will have it, corrupted, so as to suit the elements of the Nágarí: thus [P-A. házirí] is written and sounded [DEV] hájirí. In a new edition, in the Devanágarí character, of the 'Adventures of .Hátim ..Tá,f,' [. and .. indicating underdots] which
we have lately received from India, almost all dots and double letters are discarded, as a useless incumbrance.
[Earlier editions presumably had dots.]
This corresponds to the current usage of minimising the use of even the basic five. It is interesting that Fallon's Dictionary  uses only the basic five.
At the end of this Grammar is a collection of anecdotes in both DEV and Perso-Arabic script, the first seven of which use the full system of modified DEV characters.
An Urdu poet of my acquaintance has used the less usual DEV letters, especially nuqta-sa, nuqta-ha, with the meanings given in Fig.1.
LC Tables  for Hindi include nuqta-sa, nuqta-ha, and the as yet not understood nuqta-gha, nuqta-.ta [but note that BEN nuqta-gha is a 'variant' of BEN nuqta-ga in Fig.2, and Gilchrist's [1810/1825] nuqta-.dha is almost, but not quite, identical to nuqta-.ta!].
We deduce that a stable system of modified DEV developed from Gilchrist's schemes, and that this has given rise to a body of texts.
In BEN, five different schemes have been used by various scholars in India and Bangladesh:
Three of the schemes have no claim to being standard:
Furthermore, these three schemes are not widely used.
- Scheme 2, 1st row, uses cha as the base character because in some Bengali dialects this is pronounced [sa]. It also uses ya for Ze (5th row; see names in Fig.5), whereas nuqta-ja is standard for /z/, and ya has a very different pronunciation.
- Scheme 4, 4th row, uses ya for Dhaal, which again has a very different pronunciation.
- Scheme 6 is cumbersome, and also uses ya for Ze (5th row).
Schemes 3, 5, which are regarded as alternatives, are widely used by grammarians, lexicographers, and scholars, e.g. Suniti Kumar Chatterji and Jnanendramohan Das. (They are not used in the popular media or general books.) A few available fonts include nuqta-ka, -kha, -ja, -pha. A survey of examples of extended BEN usage has been prepared by Abu Jar M Akkas.
With characters used for some other languages included for convenience, the general scheme of extended Bengali script is:
For the BEN schemes there is evidently a body of texts. The schemes vary, although a stable scheme (3, 5) may be developing. However, there is no need to think of transliterating all the different modified characters in separate ways. Because of the underlying Perso-Arabic script, the favoured schemes and all the others may be dealt with in a simple, unified way, shown in the next section.
Accomodating various schemes
Fig.4 shows how the various schemes may be accomodated. The suggested transliterations apply to all the Indic characters in a given row (which therefore count as 'equivalent orthography' when, and only when, they represent Perso-Arabic characters), and are in general accord with current or draft international standards for Perso-Arabic script. Characters which occur without any modification are omitted from Fig.4.
7 - Gilchrist (1804/1825)_____
8 - later DEV______________
9 - Latin transliteration of Indic
The DEV character nuqta-jha is found in Parivardhita Devanaagarii .
Fig.5 simplifies Fig.4 to show its underlying pattern. The first column shows the Perso-Arabic character corresponding to an Indic character used in some particular scheme. The second column enables one to deduce what the transliteration of the associated Indic consonant should be. An Indic consonant being 'unique' means that it is used for only one Perso-Arabic consonant. A transliteration being 'dominant' means that this transliteration is to be used when an Indian character is used for more than one Perso-Arabic character. Only two consonants have been found to be dominant. (We are still only thinking of modified Indic characters, as in Fig.4.)
(Names given are for convenience of reference only)
The procedure for transliterating a set of Indic characters representing Perso-Arabic characters is then:
For reverse transliteration, one uses the Indic characters of the original Perso-Arabic to Indic scheme. In this way a large part of many schemes may be accomodated.
- Unmodified Indic characters have their usual transliteration.
- Each modified Indic character representing only one Perso-Arabic character is transliterated as shown in Fig.5.
- Any modified Indic character representing more than one Perso-Arabic character is given the dominant transliteration appropriate to those characters (cf. Fig.5).
Electronic transmission and storage requiring 7-bit character set
The new Latin characters with diacritics may be included as follows (with the usual abbreviations):
s_macr-b -> _s z_dot-a -> ;z s_dia-b -> ^s or ~s or ,s
z_macr-b -> _z z_dot-b -> .z h_dia-b -> ^h ~h ,h
t_dia-b -> ^t ~t ,t
opening inverted comma [before vowel] -> .
I gratefully acknowledge the help kindly given me by Abu Jar M Akkas and other correspondents.
Up to Transliteration List
Copyright (C) Anthony P. Stone 1999. This material may be freely used, provided the author is acknowledged
Last updated: 10 June 2002