Language-identification questions are off-topic on linguistics.stackexchange.com, or I would have posted this over there. Since it's also a puzzle, I figure puzzling.stackexchange might appreciate it.

The following is an excerpt from (an English translation of) Karel Čapek's 1936 Czech sci-fi satire The War with the Newts. Mister Povondra collects newspaper clippings about the newts from all over the world — "in the Cyrillic alphabet, Greek, Hebrew, Arabic, Chinese script, Bengali, Tamil, Javanese, Burmese or Taalik ... printed in the Malabar, Tibetan or Coptic scripts ..." And then there's one which the author decides to print untranslated, as an image inserted into the text.


I am 95% confident that this text is in a fictional script invented by Čapek for humorous effect. But maybe not! And maybe it's got some cryptic element to it — for example, if it were highly distorted mirror-writing in the Czech language, I doubt I'd personally be able to tell. So I thought I'd throw it out there for others to analyze.

It is also quite possible that Čapek explained the mystery somewhere (possibly in Czech), which would turn this into a [knowledge] question.

  • 1
    $\begingroup$ Further on in the book, there's another note in Japanese kanji. Translation requested for that one also, even though it's much less "puzzling." $\endgroup$ Jul 9, 2019 at 3:40

2 Answers 2


I was notified from my Wikipedia talk page about this puzzle, and I tried to identify it, but after halfway there's more and more unrecognized characters so I stopped there.



The partially-reconstructed text:

Reconstructed image

Edit: here's the unicode:


First, the handwriting(?) style is not a common one used in Java. Plus, if it's handwritten by the author, it also depend on how familiar the author with the script and the writing rules of Javanese. Because of that the reconstruction could have several possibilities. This is just one of them.

Second, to transliterate back to Latin is another thing. Javanese script is an abugida (syllables-based) and it doesn't have space break. I can tell you what each character transliterate to, but without knowing which language, I couldn't tell you what the words are. It's not Javanese, and I don't know Czech. So I will transliterate what I have so far, each syllables, for you to consider. The only visible breaks are the full stop marks.

First sentence: ya sa ra ma pu ha(=or 'a') hi(=or 'i') wo p ti ra sa shmar.

Second sentence: wa pa ra ma yu pi.

Third sentence: (first character is impossible combination) pa p ma ha dha ra ma ha wa lu pa ling ma wa par ysa rte ma ra pu pa gha ha ...

The rest are quite illegible


This appears to be Javanese script. Here's some sample text from Omniglot so you can see the similarities:

enter image description here

The first few characters look like ga, na, ra, ma, and maybe a final _w consonant? (The "ε" shape of the ma is pretty distinctive, but the rest is much harder to identify.)

The handwriting quirks make it difficult to transcribe without knowing the language, though (and it actually could be a related script, since there are many in this family: Balinese also looks similar).Translation will also be difficult since the script has no word boundaries.

In any case, the answer to the implicit question here, "is this a puzzle made by Čapek", is "no". You'll likely have more luck with the actual transcription and translation (which are not related to puzzles) elsewhere.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.