Talk:International Alphabet of Sanskrit Transliteration

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Writing systems (Rated Start-class, Low-importance)
WikiProject icon This article falls within the scope of WikiProject Writing systems, a WikiProject interested in improving the encyclopaedic coverage and content of articles relating to writing systems on Wikipedia. If you would like to help out, you are welcome to drop by the project page and/or leave a query at the project’s talk page.
Start-Class article Start  This article has been rated as Start-Class on the project's quality scale.
 Low  This article has been rated as Low-importance on the project's importance scale.
WikiProject India (Rated Start-class, High-importance)
WikiProject icon This article is within the scope of WikiProject India, which aims to improve Wikipedia's coverage of India-related topics. If you would like to participate, please visit the project page.
Start-Class article Start  This article has been rated as Start-Class on the project's quality scale.
 High  This article has been rated as High-importance on the project's importance scale.
Note icon
This article was last assessed in May 2012.

There's a table of some of the html codes for the IAST characters at Talk:National_Library_at_Calcutta_romanization ; if anyone wishes to check them, add anything, and perhaps add columns to indicate what the equivalents are in other transliterations (and scripts), that would be useful. Imc 21:31, 30 May 2005 (UTC)

It wouldn't hurt to add ळ - l, ड़ - ṛ, and ढ़ - ṛh. And it would be useful to add foreign sounds (in Hindi) and other scripts as well. Anyone who wants to tackle it, omniglot's webmaster uses IAST. Also mention there is no unicode (as far I know of yet) for the ring under ऋ - ṛ, thus confusion with the flapped retroflex. Anyway, I have been such a huge supporter of this system, but I had no idea who developed it or what it was called - only that anybody who's anybody in the scholarly world uses this transliteration transcription! Kudos. Khirad 10:40, 29 September 2005 (UTC)

I think you're wrong - those characters aren't used in Sanskrit, and I don't think those transliterations are part of IAST. ऋ is definitely ṛ rather than r̥ - look in any Sanskrit textbook. What Omniglot is using is a related but separate system, which unfortunately I don't know the name of - maybe it's correct National Library at Calcutta, though the table in the Wikipedia entry uses r̥ for the vocalic... Looking around, maybe it's ISO 15919... The difference in the use of ṛ is annoying, but there you are. Butsuri 23:41, 20 October 2005 (UTC)
That's a very minor nit, but while ळ (retroflex l) is not used for Classical Sanskrit (post-Paṇinian), it is used for Vedic Sanskrit; it even appears in the first verse of ऋग्वेदः Ṛgveda... (of course, the original prononciation of it is debatable). One can also consult Monier-Williams' dictionnary introduction; note there are no mention of ड़ and ढ़ there. AntoineL 10:32, 17 November 2006 (UTC)
[to Khirad] Well, it does hurt, since then you've an ambiguity in the meaning of both ḷ (if you choose it for ळ, as many do) and ṛ; and while it usually obvious for the human reader with the aid of the context, it is a definitive problem for those overly stupid computer algorithms and similar tools, since it breaks the fondamental property of reversibility (lossless). About the ring below r for ISO 15919 ऋ, it is there, as U+0325 (see also; and yes, this use of a combining diacritic is also a problem for those stupid computer algorithms too, albeit to a lesser degree. AntoineL 10:32, 17 November 2006 (UTC)

r̥ is sometimes used, mainly by Indo-Europeanists (since that's how they write the parent phoneme in the parent language), but it's certainly not IAST. dab () 10:47, 2 January 2006 (UTC)

Yes, I'm sorry, since my main knowledge is Hindi not Sanskrit, I forgot that this system under the title of IAST was for Sanskrit exclusively. In my naïvetée I thought that this was an umbrella term, since I was directed here from pages not dealing in Sanskrit content. You are right, I am thinking of the National Library at Calcutta scheme. Thanks! Khiradtalk 14:11, 18 January 2006 (UTC)

Formatting of this page.[edit]

I've got problems here now this page has been reformatted. Using Firefox 1.5 under Linux, I can only see the first column of the table. The rest of the table stretches away about 18 feet to the right, which is not much use as I only have a 17inch monitor. Using Firefox under Windows it displays OK, but I do most of my work here using Linux. Imc 19:31, 1 January 2006 (UTC)

Sorry about that, Imc. I had checked it using Firefox and as you said it worked under Windows. Have changed the formating somwhat (not the ideal one though as it has to have spaces between each phonetic set). Do let me know if its okay now (I don't have Linux to test it). --ΜιĿːtalk 06:47, 2 January 2006 (UTC)
That's fine now. Thanks. Imc 09:29, 2 January 2006 (UTC)

Using IAST in articles[edit]

cut from article:

Usage of correct IAST transliteration of Sanskrit terms in the article can enhance the linguistic value of the article. Along with IPA, it can also help clarify any apbhraṃs (distortion) that may have got introduced in the pure format of the term. Wikipedia has various template supports for IAST. For the correct rendering of IAST glyphs, include the term within {{IAST|-term-}} format. After using IAST in the article, consider including {{IASTText}} template in the article, which would add it to the appropriate category.

this may be useful on Wikiproject Hinduism or someplace, but it doesn't belong on the article. dab () 13:51, 18 January 2006 (UTC)

I have been editing for a short while and have already seen a lot of inconsistency in the way IAST is used. Personally I feel that it is essential that we encourage the use of IAST. There is no other way to get any academic respect unless we do. When I add IAST to articles, sometimes the IAST is quicly reverted to simple English, while other times it stays. Is there some way to build more agreement about these issues? I have tried commenting about this on other pages but have failed to create a groundswell. So I post here, in the temple of IAST, as perhaps my last refuge. What to do? Buddhipriya 01:30, 3 March 2007 (UTC)
Why IAST is used instead of IPA worldwide? IAST weirdly assigns IPA [j] for "y" instead of "j", IPA [ɟ] for "j" instead of "ʝ", in manner of English spelling, but not in manner of spelling of most European languages. In this and other similar matters IAST should follow IPA, and should be reformed, because IAST spelling in these matters looks weird for most Europeans due of its influence by English spelling. (talk) 17:56, 10 January 2008 (UTC)
One problem with the {{IAST|-term-}} format is that, at least in Firefox on the compputer I use, the font changes. So śāntiḥ and śāntiḥ look different—the first is in the same font as the rest of the text (Tahoma?), and the second in some serif font. Weird. This makes articles using the tag look bad. What could be the cause of this? I am tempted to remove the tag for this reason! Devadaru (talk) 03:13, 11 January 2008 (UTC)


Is the value of ś definitely /ʃ/ and not /ɕ/? I'm quite certain that the sound represented by the first letter of śiva does not exist in English. --Grammatical error 18:48, 30 April 2006 (UTC)

Inclusion of the character ळ[edit]

In sanskrit and in many other south-indian languages, the character ळ is widely used. It is not found in the IAST in wikipedia. Where is it? Aanand Pranav Sharma 07:56, 29 August 2006 (UTC)

It's there in ISO 15919, but I'm not sure if it's there in IAST. The ISO standard is (largely) a superset of IAST, by the way. Ambarish 08:22, 29 August 2006 (UTC)
it's also in the National Library at Kolkata romanization, but not in IAST afaik. dab () 13:48, 17 November 2006 (UTC)

IAST vs. ISO 15919[edit]

Is there a reason that IAST is used over ISO 15919 on Wikipedia? The previous discussion item is related to absence of a transliteration for Devanagari ळ in IAST. ISO 15919 this letter is transliterated as l-underdot. Also, in recent years, scholars have begun to represent the short and long forms of vocalic r and l with underring instead of underdot, reserving the latter diacritic, more appropriately, for retroflex and similar letters. Sarayuparin 00:11, 19 November 2006 (UTC)

No there isn't actually. I was attempting to formulate proposals for ISO 15919 but I haven't had time to finish it (see Wikipedia:Naming conventions (Indic)). You're more than welcome to assist to get it off the ground! Sukh | ਸੁਖ | Talk 11:34, 19 November 2006 (UTC)
There is nothing to discourage you from using ISO 15919, especially when transliterating languages other than Sanskrit. When transliterating Sanskrit, the difference of IAST vs. ISO 15919 is really splitting hairs: to people who know Sanskrit, it won't matter whether they see ṛ or r̥ because the meaning is obvious either way, and to people who don't, it own't matter either, because they'll just see an r with some diacritic speck. Using ṛ is more convenient because it doesn't need a combining diacritic in Unicode, and I suggest we stick with {{IAST}} for Sanskrit. In Sanskrit, ळ will only ever come up when discussing specific Vedic Sanskrit citations, otherwise, you'll just use its allophone ढ. In such cases, it may make sense to use ISO 15919. dab () 14:13, 19 November 2006 (UTC)

Sanskrit e and o, as also a before t.[edit]

I totally disagree with the current opinion that Sanskrit पे (IAST e) and पो (IAST o) are pronounced as long /eː/ and /oː/. The Old Prussian language shows, that theese characters have to be pronounciated and written as ei [ei] and ou [ou]. For example OP. deivs 'god', Skr. devaḥ, must be deivaḥ, OP. dātvei 'to give', Skr. dātave, must be dātavei, Latvian govs 'of cow', Skr. gos, must be gous etc.
The Skr. ai really comes from long diphthong āi, au from āu, so along theese changes the short diphthong ai turns into ei and au - into ou. e: and o: are later developments from ei and ou and can't be introduced in Classic Sanskrit!
And now abaut a before t. Indo-Iranian languages have the feature, that tautosyllabic n or m (ṃ) falls out in some cases before t and in the end of the word. And in order to preserve ethymology, such a(n)/a(m) have to be written in IAST by using character ą [an]. So the Skr. śata '100' must be written as śątą (<*čamtam 'hand', cf. Latvian cimds 'glove', čamdīt 'to fumble, to paw'), not śata!, nava as navą (<*navam 'new'), dataḥ 'tooth' as dątaḥ (<*adhantas 'eating') etc. Cognates Proto-Baltic čimtan/cimtan '100' and dantas 'thooth'. Roberts7 01:41, 17 March 2007 (UTC)

This article has nothing whatsoever to do with the Prussian language. dab (𒁳) 16:15, 19 March 2007 (UTC)
That was only an example pointing to Sanskrit real pronounciation of पे. The main story is this. Ok, if we even accept पे and पो beeing pronounced as [ē] and [ō], why the appropriate IAST signs are e and o and not ē and ō?! If [ā] is ā, [ī] is [ī], [ū] is ū, why [ē] is e and [ō] is o? This totally dissarange the whole transliteration system and throws into confusion foreigners which don't know the pronounciation rules. Historically e and o come from Latin language, where they were pronounciated as [ē] and [ō], but Latin e and o have nothing common with Sanskrit पे and पो. The fact that [ē] and [ō] does not have their short counterparts doesn't pass any critisism. The system must be uniform. I think that IAST has to change पे and पो transliteration to ē and ō, if we take as standard nowadays pronounciation. Roberts7 19:07, 20 March 2007 (UTC)
it's just a convention. Yes, ē and ō have been used (ISO 15919), but IAST didn't adopt them (probably because ē sort of implies existence of an e, which would be undefined). IAST is IAST, it is intended for people who know the phonological context (and not as a teaching aid to explain said context) and it happens to be the most widespread standard. dab (𒁳) 21:19, 20 March 2007 (UTC)

More on use of IAST[edit]

The category for articles containing IAST just disappeared because apparently people did not think it was of value. [1] I am wondering if this is a good time to review principles for use of the IAST tag. I have been trying to develop some consensus around use of IAST [[2]]. I have gone over that thread and have tried to refactor it to get a draft of what may become the nucleus of a statement on IAST that we may be able to agree on.

Questions about the use of the IAST tag as opposed to the Unicode tag, or use of the Template:IAST versus other encoding schemes, suggest a lack of clarity about what the IAST tag is for. In my view the purpose of the IAST tag is to define which of the different romanization systems for Devanagari is being used, not to define which of the glyph computer implementations is being used. IAST is a set of glyphs that has implementations in both Unicode and ASCII.

If we can get a clear summary that many editiors can agree on, perhaps we could share our points of agreement with others to see if there can be more general consensus reached. Building agreement slowly around even a few points would help the Hinduism Project on this issue. Please give me your reactions to this restatement of ideas:

1. Basic assumptions

  • IAST is the academic standard for the romanization of Sanskrit.
  • In at least some cases there is the potential for significant information loss if IAST is not used, and the nature of the loss in meaning may not be obvious.
  • Currently there is no clear official standard for the use of IAST within the Hinduism Project, leading to inconsistent use of IAST across multiple articles.
  • Since Wikipedia is intended for use by English readers, clarity of communication with English readers is of great importance.
  • IAST may be harder to read for a general audience.
  • While use of IAST may reduce readability for some readers, failure to use IAST may reduce credibility of articles among other readers, but the impact of this loss of credibility is not clear.
  • For some editors, use of IAST is a matter of acdemic integrity. Good faith should be assumed when practice varies between editors, e.g.:
  1. when a term is first used anywhere within an article (not just in the lead) if the editor feels that use of IAST is important to preservation of meaning.
  2. when an editor feels that a particular term may lead to confusion with another Sanskrit term unless IAST is used consistently.
  • Editors who do not know IAST or prefer not to use it are not required to add it in order to create new content. Later editors may adjust IAST usage at a later time in such cases.

2. IAST should be specified when:

  • a term is defined in the lead.
  • its etymology is traced.
  • Transliteration should be preserved in any quoted text; i.e. if the original uses IAST, we have to quote correctly.
  • When Sanksrit shlokas/mantras etc are cited, it is ok to use only IAST and leave out standard english transliteration. The reason for this being that it would be too repetitive to provide the (1) standard English transliteration, (2) the IAST transliteration and (3) English translation ... since including the english translation is standard (for English wikipedia), if we need to provide only one transliteration IAST is certainly preferable, since it is richer in information.

3. Except as noted above, IAST should not be routinely used when:

  • A term has a common English equivalent (e.g., Ganesha for gaṇeśa, Krishna for kṛṣṇa).

Buddhipriya 19:16, 25 March 2007 (UTC)

Good points. The guidelines will obviously look something like this. Note that this discussion belongs on Wikipedia:Naming conventions (Indic) / Wikipedia:Manual of Style (Indic-related articles), which somebody should revive and have it linked from WP:MoS (compare Wikipedia:Manual of Style (Arabic)) -- article talkpages are for article content, Wikipedia_talk:-pages are for style and guideline discussions. Yes, IAST should be used when a Sanskrit term is introduced, or when its etymology is being discussed. In other instances, common diacritic-less anglicizations will do. But I wouldn't actively remove IAST from articles where somebody has taken the pains to spell Sanskrit terms in IAST throughout. Also, in more technical topics (Vedic meter, Devi and Vrkis feminines etc.), more background will be expected from the reader, and IAST spelling will be more widespread. Btw, the {{IAST}} template is not just for rendering instructions, but also for article parseability (and tooltip at mouseover). Compare {{PIE}}, {{IPA}}, {{ArabDIN}}. dab (𒁳) 08:39, 26 March 2007 (UTC)

Chandrabindu in IAST?[edit]

I am working on a project that caused me to notice an issue with IAST specifications in handling of chandrabindu. I must convert an electronic source text from the old CSX+ (Classical Sanskrit Extended) encoding to Unicode, and in the course of it noticed situations like. In ASCII CSX+ encoded fonts, the chandrabindu is specificed by a glyph which perhaps would be described in Unicode as (Latin Small Letter M With Dot Below and Chandrabindu Above). Is there an IAST glyph specification for this? An example is the following sample line from the electronic edition of the critical edition of the Mahabharata which includes the word pāpā(?)l, where the missing glyph is shown by a question mark:

1.71.51c ye nādriyante gurum arcanīyaṁ; pāpā(Á=?)l lokāṁs te vrajanty apratiṣṭhān

The resulting romanization must be IAST-compliant, with Unicode encoding (ideally using MS Arial Unicode as the font). I have not hit this issue before and cannot find any IAST specification for handling it. It cannot be encoded as anusvara because the requirement is to have a lossless exchange of encodings. Is this a missing aspect in IAST, and if so, should it be documented in the IAST article? Buddhipriya 05:46, 19 April 2007 (UTC)

Put a candrabindu on 'm.' (From the Report spec in the resources) (talk) —Preceding undated comment added 20:31, 17 May 2009 (UTC).


According to IAST, anusvara is transliterated with an "mdotbelow", ṃ. But "mdotaccent" ṁ is widely used for anusvara. Why the discrepancy? Is this something to be noted in this article? Where are the IAST standards published? (I still have a doubt …) Devadaru (talk) 09:22, 1 January 2008 (UTC)

Cell background colour?[edit]

What, if anything, is the significance of the ivory background colour used in some cells? (Some cells in two of the tables, no cell in one, and both cells one, to be precise.)

The cell background colour was introduced by Miljoshi when adding formatting to the table in [3], but I can't see any explanation in that version nor in the current. The cell formatting has since changed and the cell background colour was originally darker (#FAFAD2) than the current ivory, but the set of cells with background colour has remained the same.

Is this just a random error propagated by copy-paste before first submitted and never questioned by anyone before I came looking? Sounds rather doubtful ... ;-)

But I still can't figure out what its significance is, and I doubt I'm the only one. Could someone in the know update the article to tell us?

Thanks! — the Sidhekin (talk) 17:48, 7 February 2008 (UTC)

In one word: Diacritics.
It's been over 5 years since your message, but I've only just noticed this myself when editing the article's IPA links.
I first thought those sounds feature some kind of phonetic property common to them all, but it's apparently a typographical issue. They draw attention to letters needing diacritics. There are many of the vowels that do but surprisingly few consonants.
I've added an explanation to the article. —Menchi (talk) 17:08, 18 November 2013 (UTC)

Example Text[edit]

It would be great if this page had a paragrah of example text. —Preceding unsigned comment added by (talk) 07:27, 13 August 2008 (UTC)

Online conversion?[edit]

Is there any Web site where you could punch in Devanagari strings and get the correct IAST out? Jpatokal (talk) 03:49, 2 December 2008 (UTC)

Try this one: mahaabaala (talk) 15:04, 1 October 2011 (UTC)

Pronunciation reference added[edit]

To the layman, this article would be hopeless without a pronunciation guide. I have added a link in the References to an external International Phonetic Alphabet chart with pronunciation guide. This will help alleviate what will otherwise be confusion and dismay on the part of people uninitiated to the IPA. (talk) 17:26, 6 February 2009 (UTC)

Any chance corresponding phonemes could be added into the table itself? It's pretty standard in pages on different scripts I've seen to include a column/space in each chart for them. 2602:306:3AE4:A210:8D61:6451:10EA:7502 (talk) 01:22, 27 July 2015 (UTC)


The candrabindu is missing — how is that one transliterated in IAST? — N-true (talk) 01:33, 2 March 2009 (UTC)

while in general it is true that "due to written Devanāgarī's isomorphic correspondence with spoken Sanskrit, any transliteration of Devanāgarī is automatically a phonetic transcription of Sanskrit, and vice versa", this isomorphism breaks down in the case of the "non-letters" upadhmaniya, jihvamuliya, candrabindu and friends. Afaik, usage of these symbols vary between Vedic schools and manuscript traditions, and we would need to consult expert literature to figure out exactly who uses which to signify what. The simple IAST to Devanagari correspondence breaks down if you really want to go into the gory details of manuscript tradition. If it is deemed desirable to transcribe candrabindu, usually simple m̐ will be used, but I do not know if this was part of the original 1894 definition of IAST. --dab (𒁳) 12:10, 18 May 2009 (UTC)--dab (𒁳) 12:10, 18 May 2009 (UTC)

It was, as were the symbols for upadhmaniya and jihvamuliya. -- Arvind (talk) 10:44, 21 May 2009 (UTC)
Incidentally, in case anyone wants to look up the original specification, whilst copies of the original report are quite hard to find, an English translation of the report was published in JRAS 1895 at pp 879-892. -- Arvind (talk) 00:37, 22 May 2009 (UTC)
A PDF of the commission report can be found at the very bottom of the resource link (talk) 22:32, 24 September 2009 (UTC)


Om is also one letter that is used a lot in Sanskrit literature. Please add it. -- 11:00, 24 Jul 2009 —Preceding unsigned comment added by (talk)

oṃ isn't a letter it's a syllable made up on the vowel o ओ or au औ with anusvāra ं or anunasika ँ. Therefore it has no place in the alphabet. mahaabaala (talk) 15:05, 1 October 2011 (UTC)


The pages states that: owing to confusion of the vowel l sign in Sanskrit (here ḷ) and the need for the same sign for the retroflexive consonant ḷ, which is found in Pāḷi.

In practice there's no confusion because Sanskrit does not have a retroflex consonant ळ; and Pāli does not have a vowel ऌ. mahaabaala (talk) 15:04, 1 October 2011 (UTC)

Confusing conjuncts[edit]

Probably silly question, but: As I understand it, IAST should provide a 1-1-mapping from Devanagari to roman and back. But क्‍ह and ख both map to IAST kha, likewise ब्‍ह and भ to bha. Which point am I missing? --WolfgangRieger (talk) 18:37, 11 March 2013 (UTC)

@WolfgangRieger:, can you think of a Sanskrit word that has these combinations of consonants. From what I know, these should be impossible given the phonotactics of Sanskrit. Uanfala (talk) 22:43, 13 June 2016 (UTC)

IPA values?[edit]

Were there ever ipa values in the main table? From reading the talk page, it seems like there were at some point. Where have they gone? (talk) 22:02, 13 June 2016 (UTC)

They were removed with this edit back in 2013. If you're interested in what the tables looked like with IPA, you might have a look at this old revision. Pinging Kwamikagami in case they might have anything to say on why the IPA was dubious. Uanfala (talk) 22:41, 13 June 2016 (UTC)
I removed them because we don't know in detail how Sanskrit was pronounced and authorities differ. Just as with Latin, Classical Greek, etc. Also, I think the IPA here differed from what we had in the Sanskrit article. Better to avoid content forks by restricting the IPA to Sanskrit so we don't get out of sync again. — kwami (talk) 21:23, 14 June 2016 (UTC)
But Sanskrit had a much stronger tradition of classical phonetic texts than either Latin or Greek. I'd imagine for all segments with the exception of the syllabic consonants it will be relatively easy to give a generally accepted broad IPA value. Uanfala (talk) 19:55, 15 June 2016 (UTC)