[Note: For conceptual clarity, there is no marked Urdu/Devanagari script on this page, so the script bar at the bottom of the page will not cause any changes.]

As you look at the verses, you’ll notice some non-standard spellings. The biggest reason is that the basic transliteration system was designed to reflect Urdu spelling and pronunciation. Since we couldn't consistently generate modern standard Hindi spellings from it, we originally decided to offer direct reproduction of the Urdu letters. Thus you might see, for example, ‘raastah’ rather than ‘raastaa’. At the price of not seeing what you’re used to, you’d gain extra information (usable for metrical analysis and other purposes) about the letters that Ghalib actually wrote. This approach can, for many words, bring you as close as possible to a direct experience of the ghazals as Ghalib composed them. In other words, it can result in a transliteration—a scholarly study tool—like the roman (=English) script transliterations that are also provided.

Originally, only a few very basic exceptions to this policy were made, so that extremely common words like yih and vuh and nah were made to look as they do in modern standard Hindi. But now Sean Pue has been working to come closer to modern standard Hindi spellings for many more words. This is a work in progress.

Above all, consider the vexed case of conjuncts. It is helpful to the transliteration-user to see that a word is pronounced faryaadii or ta;hriir with the far and ta;h as initial long syllables. And of course, it looks fine in Urdu script too. But in Devanagari, it turned out to generate absurdly many odd-looking conjuncts. To avoid this problem, we disabled the conjunct-forms in Devanagari, so that there were no conjuncts, which was bad but still the lesser of two evils. Thus originally we showed no conjuncts at all (which is the way things normally look in Urdu), as the least bad of the available choices. But in this case too, Sean is still working on improvements, so conjunct display is now more of a mixed bag.

Another intractable problem was the Urdu letter ((ain, which is pronounced as a kind of wild-card vowel and simply can’t be consistently rendered into Devanagari by any automated means. However, Sean tells me he has fixed this.

In Urdu, certain Persian words like ;xvush (happy), ;xvushii (happiness), ;xvud (oneself), and ;xvaahish (longing) all appear with a full vaa))o in them, so they may seem to have a long uu sound; but in fact they don't. Such spellings have been retained in the Roman transliterations, but removed from the Devanagari.

The (very rare) letter zhe has no standard Devanagari representation; it has been made to appear as jh with an underdot.

Forms like miraa and tiraa (for meraa and teraa ) reflect the shortening of vowels for metrical purposes.

Traditional Urdu usage frequently has vuh and yih even in plural situations, where modern standard Hindi would have ve and ye .

There are also some minor technical glitches generated by different browsers. Sean has made our system strictly according to the Unicode standard, so it can be expected that these glitches will gradually be fixed. For more on technical issues, see Sean’s ‘more information’ page, which will be enhanced over time.

So please accept my apologies for any remaining oddities in the Devanagari. I’m hoping that in some cases the advantage of receiving accurate, reliable information about what Ghalib actually wrote will more than compensate for the unusual-looking spellings. Count on Sean to keep an eye on the situation and try to figure out the best available options over time. For the present, just remember that in some cases you may be seeing a direct transliteration of the Urdu, rather than an attempt at modern standard Hindi spelling.