A Strange and Sublime HASANT

by Michael S. Kaplan, published on 2007/12/05 19:59 -08:00, original URI: http://blogs.msdn.com/michkap/archive/2007/12/05/6674157.aspx

Now as I said in Learning to spell in Bengali (when one doesn't know the language) -- admitting I said many different things in that post -- I wanted to get a copy of that book by Amit Chaudhuri.

It arrived yesterday (actually a collection of three of his novels entitled Freedom Song).

I have no idea why it seemed so important. I really don't.

But it did.

I jumped immediately into the first story (I had to read A Strange and Sublime Address first, you knew I did too, I'd wager).

I am blown away by beautiful language and sharp imagery. Of places, and people, and food, and more.

And a world that I admit I know nothing about.

Yet I find myself learning.

Because I am there, in Calcutta.

Visting when Sandeep, who is only ten, is.

Could I be ten, and there? It seems unfathomable.

By tonight I have finished the novel.

By the end of the book I find that I am so completely lost in the language that I worry how I will get back.

...He looked down again at the paper box, and gazed at the word "sandesh" printed on its side in neat, stylist Bengali letters.

The letters, curving, undulating, never still, curving into a kinetic life of their own, reminded him of Calcutta, of buying and selling, of people on the pavements, of office-goers in the mornings, and homecomings in the evenings, of children reading books, of arguments and dissensions in tea-shops, of an unexpected richness in myriad rooms, all festivities of colour and light. He wanted to return to the city where all things curved and arched and danced like those letters; it exhausted him to lie in this room with these other still figures He longed to come back to life.

I suddenly realize how I will get back -- I will be distracted, by the script.

(in between the above two sections is the word sandesh in the Bengali script)

I find myself thankful that the script does not appear too often in the book or I would never be able to finish it as fast!

The word sandesh in Bengali seems to be pretty well known and findable on the web as follows:


Which is to say, U+09b8 U+09a8 U+09cd U+09a6 U+09c7 U+09b6 or SA NA HASANT DA E SHA.

And I find myself wondering why that it would not end with a HASANT to suppress the vowel at the end, making it


and when I look at the script in the book, I am distracted by the small slant on the bottom of that letter which could mean that HASANT is there after all.

But Idoubt it -- no separation, barely a slant. I am reaching here, big time.

I try a Google Fight to see which string is more common and seem to hit a bug in Google Fight since the application won't finish (see it here).

Though when I search for them individually the reason for the problem is clearer -- without the HASANT gets 60 hits, including the Bengali Wikipedia article here, and with the HASANT gets 0 hits (after this article goes live it will probably find everybody who picked up the full article, but it is zero right now).

Now I am lost again, my foundation into the language began with the script, and that definitely came from Unicode. But clearly the VIRAMA model for pure consonants is not followed universally here.

Maybe when the word is said in Bengali the a is still there.

Just ever so slightly.

I suddenly want to look at my Daniels and Bright again and see what it says, but it probably won't help me here, and the book is in the office and I am home.

I wonder for a moment -- the office is not far away; I could just head over.

There is so little like getting lost in that tome, in the richness of the description of the world's writing systems.

Instead I decide to stay home and re-read A Strange and Sublime Address again.

Because I did find somewhere else to get lost.

I'll go back to Calcutta.

And the office will wait for tomorrow....

This post brought to you by  (U+09cd, aka BENGALI SIGN VIRAMA, aka HASANT)

# Philip Newton on Thursday, December 06, 2007 4:16 AM:

I'm told that the inherent vowel is not pronounced in syllable-final position in Hindi -- so that someone named after King Ashoka would be called "Ashok", for example.

I wouldn't be surprised if a similar process had happened in Bengali, in which case you wouldn't need a halant because you'd know that syllable-final inherent A is silent anyway.

# John Cowan on Thursday, December 06, 2007 8:32 AM:

Well!  Next you'll be surprised that we write "give" and "love" instead of "giv" and "luv" in English.

Bengali shares with most of the other modern Indo-Aryan languages a pervasive loss of final vowels.  That being so, it's customary not to bother with a final virama, and the reader just has to know when there's still a final vowel (usually in a borrowing, particularly a borrowing from Sanskrit, where final vowels were still much in evidence) and when there is not.

The explanation of "give" and other words with "-ve" is that until about the 18th century "u" and "v" were basically glyphic variants, and the rule was to pronounce a consonant if a vowel followed, and a vowel otherwise.  So final silent "-e" (another case of loss of final vowels) was always retained after "v" whether the preceding vowel was long or short.

As for the "o" in "love", that's from Middle English, where it was hard to read handwritten sequences like "un", "um", and "uu" ("uv") because they were just runs of ups and downs (minims), so "u" was often changed to "o", as in "son" (but not "sun"), "come", "love".  This didn't affect the pronunciation at all, of course.

# Michael S. Kaplan on Thursday, December 06, 2007 9:30 AM:

How many people in Bangladesh or Bengal would use that Hasant in the middle of the word, though? :-)

Yes, the point of the post was a little tongue in cheek, though the underlying point (that the Virama model is much harder to swallow when there is no visible Virama) is not.....

# Mihai on Thursday, December 06, 2007 12:40 PM:

For those who are interested: Daniels and Bright is now $46 at Amazon (regular price being around $200)

# Goldie (Godhuli?) on Thursday, December 06, 2007 8:15 PM:

The internal hasant is a tricky little bugger. It's used strictly to form conjunct consonants - which are far more common in Bengali than in other Indic languages. Apparently there's an ongoing fight between the modern and traditional linguists, where the modern linguists are trying to eliminate the conjunct consonant altogether in favor of your very own internal hasant. The traditionalists are bound to win this one, and if your poetic tendencies are indication, I'm sure you'll be one to agree.

About the ending hasant, Indic languages in general tend to not pronounce ending vowels and have no use for the hasant. Sadly, I can't find anything specifically about Bengali, though I've got my feelers out there on a few folks with better Bengali know how then I...

# Michael S. Kaplan on Thursday, December 06, 2007 9:09 PM:

I guess that "silent e" at the end of some English words really is looking more and more like a good analogue (thanks for that one, John!) -- though I have to ask, is it in fact a silent vowel? Or is it believed to not be there at all by people who know the language well?

This of course doesn't prove either way how things should be done, but it would help elucidate why things are looked at a certain way (if Unicode is losing anywhere, it is much more in "hearts and minds" than technical merit, so how people feel becomes very important).

# Gé van Gasteren on Tuesday, December 11, 2007 3:33 PM:

The first post by Philip Newton is correct, according to a Bengali friend of mine.

In other words: The writing is not 100% reflecting the pronunciation, and one has to know/learn how to pronounce written texts. The mechanism seems to be simple in Hindi: Don't pronounce the vowel if it can be done easily, else pronounce it.

The exact correspondence between sound and writing which Michael expected (and anyone learning any language would highly appreciate) seems to be there only in Sanskrit.

Completely off-topic here: Dear Michael, I've posted something about typing in Dutch and smart quotes, but I'm not sure if it "took" as I don't find it anywhere on this site. Should I post it again, and if so, how?

# Michael S. Kaplan on Tuesday, December 11, 2007 3:38 PM:

Your post is in the Suggestion Box (and the copy you sent to the Contact link is sitting in my inbox)....

"Expected" might be overstated in this case, but I know what you mean. :-)

# Gé van Gasteren on Friday, December 14, 2007 1:30 PM:

OK, great that it didn't get lost. Yes, I read that before: "Start a new topic in the Suggestion Box," but I'm not sure where to find that...

Sorry to pollute your topics, although I must say that I find the "sweet disorder" on your site really sweet :)

Ah yes, my friend says that Bengali drops the end-vowel less often than Hindi. For example, shubha would be pronounced "shubha" in Bengali, but "shub" in Hindi.

"bh" can't be pronounced without a following vowel, so it becomes like "b" unless another word starting with a vowel follows it.

Just in case you're not familiar with the sounds in these languages: The Hindi "a" is rather colourless and sometimes it's impossible to say whether it is pronounced or not. The English word "pot" has some sound after the "t" which could be called a rudimentary vowel. That's what becomes of the ending "a" in Hindi all the time.

This is about where my knowledge ends. Someone please correct me, but I think that when such a word is followed by another word, the ending "a" comes back to life, so to say, and is either pronounced more fully -- if the second word starts with a consonant -- or else combines with the following vowel to form a "sandhi".

# Michael S. Kaplan on Friday, December 14, 2007 1:40 PM:

Over on the far right side, in a section that won't collapse titled "Administrivia", it is the seond item right after "Comment Policy" and right before "My Editorial Policy". :-)

You did get there the first time (it's how you added the original item there!)....

# Gé van Gasteren on Friday, December 14, 2007 1:43 PM:

I just discovered that the Suggestion Box was hiding in "Administrivia"...

# Michael S. Kaplan on Friday, December 14, 2007 5:02 PM:

I find myself quoting the end note in Amit Chaudhuri's Real Time:

Two stories here are retellings--and quite personal interpretations--of episodes from the Hindi mythologies. An Infatuation is a retelling of an episode from the Ramayana, in which Lord Ram (often spelt Rama) is exiled to the forest for fourteen years because of a curse; he is accompanied by his brother Lakshman and his wife, Sita. Here, a rakkhoshi (the Bengali word for female rakkhosh, or rakshas in Hindi--a powerful demon), Surpanakha, falls in love with him and tries to seduce him. Ram plays along with her and then humiliates her, as the episode shows. She rushes to her brother Ravan, the king of demons, who will avenge her by abducting Ram's wife, Sita--thus setting in motion the main action of the Ramayana. The Wedding is a retelling of the story of Lord Shiv's (often spelt Shiva) wedding.

Note what he does here for both Ram-->Rama and Shiv-->Shiva, but not Sit-->Sita. Which kind of drives home the issue and one of the challenges that is clearly by multiple reports much more common in Bengali than Hindi....

# Gé van Gasteren on Saturday, December 15, 2007 1:14 PM:

Have a look at the character set for Devanagari, or Bengali: There are two versions of each vowel, one single, one double. (They're usually called "short" and "long".)

The vowel that gets dropped is always a single "a", never a double one -- it just has to much "colour" to be phased out that easily. And as feminine names often end in a double "a", that one is not dropped.

So the problem boils down to the question why single a is written the same as double a.

The cause is probably twofold:

1. The established standard for Sanskrit transliteration in the West, IAST, has always been difficult to type. Even in Unicode, some of the necessary characters are in Latin Extended Additional -- from the name you can already see that they were not high on the priority list.

2. Indians don't bother to write the names accurately in English texts anyway, because they know it can't be done really: Those who know Hindi will pronounce them correctly whatever the writing, and those who don't know Hindi will not pronounce them correctly, again whatever the writing.

To make it complete, here's the IAST spelling of the names in Sanskrit:

Rāma, Śiva, Sītā.

Also noteworthy here (see how the author sometimes writes the ending a, sometimes doesn't): Śūrpaṇakhā, Lakshmana, Rāmāyaṇa/Rāmāyaṇam

In my post, I see little squares for the "n + dot below". Hopefully the other specials are coming through as displayed. Maybe I'll add some ITrans spellings, just in case:

raama, shiva, siitaa, shuurpaNakhaa, lakShmana, raamaayaNa

# Michael S. Kaplan on Saturday, December 15, 2007 4:36 PM:

Even if people speak the language they don't always get it right -- e.g. Goldie and her family, and the long-running issue I talked about with the spelling of Godhuli (the issue there being principally an i vs ii one).

# Gé van Gasteren on Sunday, December 16, 2007 8:14 AM:

All kinds of loose remarks:

Great blog, that one about Godhuli!

She may have been thinking that, as it´s a girl´s name, it should end with an ii, which is typical of words of feminine gender. But I´ll check it with my friend from Bengal...

According to my little knowledge of Sanskrit, it´s Poornimaa,

or पूर्णिमा.

If it doesn´t show properly here, look here:


The battle between ca/cha and cha/chha transliteration has been fierce and unending. Result: when you see "cha" in a ´stray´word, it may stand for the aspirated character, or the unaspirated one.

Now I´m sure that either telepathy or synchronicity exists: Right when you were writing or speaking your tip about the Suggestion Box, I was finding it. I mean, I had not seen your tip when I posted my comment.

# Gé van Gasteren on Monday, December 17, 2007 7:56 AM:

Just one last (?) change to the spelling of Goldie's name:

My friend insists that both the u and the i should be short.

He added the information that godhuli or dusk is commonly considered an auspicious time for marriage in Bengal. Interesting to have such a name!

# Michael S. Kaplan on Monday, December 17, 2007 9:02 AM:

Well, your friend would have gotten on better with them before they determined that the name was spelled differently in the song/poem, as the post explains. :-)

# Gé van Gasteren on Saturday, December 22, 2007 8:10 AM:

My friend deeply apologizes, and Goldie goes around the world:

When I told my Bengali friend about my divulging his insights on the World Wide Web, he got a bit uneasy and asked his cousin (by email or phone) to double-check this "u" in the Tagore songbook that he has.

Result: In the song, it's a double u after all, not a single/short one...

He still wants to check if both versions are possible, but I'll have to help him get Bengali to work on his PC.

Google gives 5040 hits for godhuli with the double u, and 456 with the single u, so he may have a point there.

(Quite good that number of hits, I thought, incidentally. And have a nice Christmas time.)

# Michael S. Kaplan on Saturday, December 22, 2007 11:12 AM:

Well, now you know how surprised they were when they realized what they thought was the name was basically wrong. :-)

The funny part was that they assumed a great of assonance with the last name (Chaudhuri, a very common Bengali name), so you can see what they thought were the vowels for many years....

Enjoy the holidays!

referenced by

2010/05/12 Avoiding a Telugu-tubbies ref (out of respect for the 3rd largest language in India)

2008/01/03 Throwing a BRIC [with Diwali written on it] at Outlook, aka Attn. Outlook: There *is* an 'I' in BRIC

2007/12/06 The subliminal message: 'My search is more testosterone-laced than yours!'

go to newer or older post, or back to index or month or day