A MetaFilter post introduced me to the writer Saadat Hasan Manto, who at the time of Partition left a successful screenwriting career in Bombay for a miserable existence in Lahore, where he wrote brilliant stories and drank himself to death in early 1955. (There’s an affecting account of his life here.) He’s been compared to Lawrence and Gorky for the unblinking honesty of his writing, which frequently got him in trouble with the government; his most famous story is Toba Tek Singh, which is available in both Urdu and English here (there’s even a set of notes on the language). You won’t soon forget the lunatic’s nonsense mantra (اوپڑ دی گڑ گڑ دی اینکس دی بے دھیانا دی منگ دی دال آف دی لالٹین, uupa;R dii gu;R gu;R dii enaks dii be dhyaanaa dii mu;Ng dii daal aaf dii laal;Ten).


  1. Your Urdu text seems to be jumbled up – it reads ‘di be dhyaana di moong di daal of di lalten upar di gur gur di enaks’.
    Speaking of nonsense, Babu Gopinath is another Manto story with memorable phrases like ‘continuitily’ and ‘anty ki panty po’

  2. The Urdu is correct in the post itself, it’s just that MovableType doesn’t know how to wrap it between lines, so you have to start reading at the end of the first line and continue from the right in the second line. Sorry about that.
    Thanks for the recommendation; I’m looking forward to exploring this writer!

  3. If that isn’t right, what are the rules / native expectations for line-breaking of counterflow text? I was involved with a system that tried to do it, but that was ages ago.
    When the reader encounters a transition, they must move to the end of the counterflow piece and begin reading in the opposite direction. This includes the case where a LTR quotation is embedded in a RTL quotation that is itself embedded in a LTR document. But I did not think that this movement ever went outside a line. It sounds like perhaps sometimes it does.
    The layout algorithm that I knew was roughly the following.

    • Divide the text into words using the storage (logical / reading) order.
    • Measure the words. This process can take ligatures into account, and so respect the ultimate rendering direction. There is also some trickiness about reversible glyphs, like parentheses, if the reverse doesn’t have the same kerning.
    • Break the words into lines. Here one can also include hyphenation and widow and orphan elimination. The breaks are inserted into the logical flow, which still isn’t directional.
    • Take each line. Lay it down by working from the deepest embedding up to the shallowest odd embedding, reversing the pieces each time. What this means is that the doubly-nested LTR text in the above example will be reversed twice, and so stay LTR, but it will occur in the proper place within its RTL parent.

    Thus the margin at which the reader starts is determined by the shallowest embedding within any given line. So, if the Urdu were three lines long, the second line would have started naturally at the right. But here, the second line has outer English and so starts at the left margin. There the reader immediately encounters a transition and must scan up to the comma and read back to the margin.

  4. Greg Goulding says

    I managed to turn up a copy in both Nagari and Nastaliq of another of his stories, “Khol Do” here (or here for Nastaliq); furthermore, the fact that the Devanagari version was copied from a source text gives me hope of finding such an edition for myself… or, of course, I could just finally learn how to read the script…
    “Khol Do” for those who’ve only read the translation, is the really mean one that ends with the daughter hearing a doctor say “open the window” and pulling down her pants, at which point her father starts shouting, “my daughter is alive!”
    I also think It might be worthwhile to try to turn up a copy of Ghalib, the film he wrote. Has anyone seen this?
    Oh, oh, more for those who can read hindi: the BBC has this collection of articles on Manto that seems quite good, though I haven’t read much of it…
    It’s a strange point to be on the threshold of being able to read whatever one likes in a language, but very frustrating too…

  5. Besides the line break problem, I read the quote with lots of boxes on-screen, and when copied to Word most of it appeared in Times New Roman, but the h of dhyana didn’t connect and it as well as the retroflexes were in Tahoma. Moreover, the retroflexes didn’t connect correctly either.
    Not the first time I had that experience.
    So, I again had to make it all into that ugly Tahoma to have all the letters of ‘lantern’ (lalTen) and others connect the way they should. But line wrapping still is a pain. Any of you who know of a nice-looking Urdu Unicode font that would work in similar cases? My favourites from e.g. Global Office don’t do Unicode, and the GO people say they don’t intend to upgrade or provide conversions. A pity; their phonetic keyboards are grrrreat.
    On those occasions, I wish that Word had that trusty ole WordPerfect feature, the Code View. In that way, we would be able to leave the LtoR/RtoL codes or whatever in their proper places, and move only the text inside of them.

  6. If you’re looking for a Devanagari edition of Manto’s stories, there are several available here in India. I have the one published by Kitab Ghar – found it online here.
    Manto was a writer for hire in the Bombay film world in the 40s. The Ghalib film you’re referring to is Mirza Ghalib – Manto wrote the story, not the screenplay and dialogues. The film was made and released after he left India for Pakistan.

  7. Does anyone know where I can find an online copy/ version of Toba Tek Singh in the Devnagari script? It would be of great help to me. Thanks

Speak Your Mind