Might help for some of the remaining:
Post #4. I’m pretty sure 21 is Urdu. I don’t read Urdu, but it uses the Arabic alphabet, and there’s a word in the right position in the script to be the name of the language which would be pronounced 'uRDU - Urdu.
Aside: in 21, they seem to get the direction of the writing correct and use the correct letter forms, unlike 8, although, again, I don’t know Urdu, so maybe it’s actually just as messed up.
16 is Farsi (the language most Iranians speak). Again, they’ve reversed the word order, and transcribed it letter by letter in reverse, with the wrong letter forms. I don’t read Farsi, but it uses a close variant of the Arabic alphabet, and allowing for reverse transcription, there’s a word in roughly the correct position in the script, that would be pronounced “FARSI - Farsi”.
So what happens when you call that number? Do you get a menu like “If you speak English, type in 1” and if you missed it, you have to listen to all 63 options?
41 is Kurdish. Again, allowing for reverse transcription, there’s a word in roughly the right position that would be “KURDI - Kurdi”.
53 is using Arabic lettering or a close variant, is in the right direction, and seems to using the correct letter forms, but also seems to be using a different font than the other Arabic alphabet entries, and also smeared or something (maybe the font being used, or scanning artifacts, or a combination). I think I can make out the word LGhah, “language”, in roughly the right spot in the script, but other than that, I’m lost.
Based on ကညီ ကျိာ် being the third and part of the fourth word and the appearance of certain tone markers, I’ll guess that 57 is a Karenic language, probably S’gaw.
17 is supposed to be Hindi. They’ve just screwed up the mechanics of the script. I see what’s supposed to be the word “Hindi” but it’s written wrong.
18 looks like Armenian
I’m really having trouble understanding how they managed to mangle some of these translations they way they did. How on Earth are they getting the right(ish) words, spelled with the right letters, but using the wrong letter forms, in the wrong direction? I mean, if you just copy and paste from Google Translate, you’ll probably get an awkward translation that might be hard for a native speaker to parse, but the letters would be at least facing the right way.
And these guys are supposedly offering service in all these languages. As I understand it, these kinds of services usually rely on on-call freelancers, not professional interpreters on staff, but still, you’d think they could have offered a few bucks to one of the on-call freelancers to write the message correctly.
This also makes me think. What if the Rosetta Stone, for example, had been written this way? Are some “mystery scripts” like Linear A and Linear B so hard to decipher because they’re actually incompetent “translations” with the symbols reversed and in the wrong form? (I think that’s actually been seriously proposed by some linguists, and here we have a real world example of it).
The ones using Arabic script are even worse. Arabic letters have different forms, initial, medial, and terminal, depending on their position in a written word. There’s no real equivalent of this in English, but they’re pretty much just using the terminal forms, which are also used when writing a letter by itself. More or less,
H S I L G N E E B D L U O W S I H T.
Have you ever tried to use these alphabets on a computer? Everything has to be set up perfectly in terms of the software settings, and copying and pasting can screw everything up. It’s happened to me a lot.
For example, if you’re writing in the Bengali script or the Devanagari script or related Indic scripts, some of the vowel indicators go before the consonant, even though they indicate the vowel that follows the consonant. In the mind of the person writing or reading Bengali, one isn’t writing or reading the vowel first. One is creating or seeing a complete syllabic (“abugida”) character that is made up of a consonant portion and a vowel portion.
But computer software and keyboards treat these things as individual characters, so the software has to know that when you type, say T-E, that the characters have to be arranged as E-T. Any disparity in the settings when you copy and paste that can screw everything up, and if the user has no idea how to read the language, E won’t notice it.
Yeah, but that doesn’t really match what’s going on in the Arabic scripts.
It’s like they’re seeing,
English
But typing,
H S I L G N E
And it’s really not even that close. It’s not just a matter of leaving the caps lock on. A copy/paste isn’t going to create those errors. It’s like the translator is somehow seeing and typing the individual letters that make up the words without seeing the word itself.
The way the software has to be written for non-Western scripts is to take the key sequences, and re-arrange them, and then make substitutions for variants, such as differences in initial-medial-final forms, or combined characters.
But the underlying code that’s being read is still that original sequence as typed. If that sequence is taken from one software environment to another software environment that hasn’t been “taught” to handle the script correctly, I can very easily see how “English” might wind up as “H S I L G N E.” That’s one of the easiest errors to understand.
The “receiving” software environment doesn’t know to take a sequence that in Western languages is assumed to be characters moving from left to right and reverse it. And it doesn’t know to substitute different letter forms in context.
I don’t know if it’s precisely relevant, but when I lived in Egypt and did English editing for a lot of native Arabic speakers, my MS Word program would get “infected” (not in the literal computer science meaning, since there were no viruses involved, just glitches) with all kinds of left-to-right v right-to-left screwups.
For example, I’d receive a document written in English by someone who did a lot of their work in Arabic. I’d edit the document and the process would go fine.
But then I’d create a new blank document, and everything would type “backwards.” Another thing is that the delete key would delete the text to the right of the cursor instead of to the left.
I eventually figured out some oddball work-arounds which I no longer recall, except that one of the steps was to set the text to right justify so that it would in practice left justify.
It was very strange.
Ok, I think I understand what you’re saying, and that does make some sense. The original might have originally been typed in correctly, but then when it was re-formatted for the printed document, the re-formatting also accidentally broke up the words and reversed them. And no one did a QC on the final product, or at least not a competent one.
I think @Ascenray is right. I have seen that effect.
There are several phases involved in rendering Arabic text: first, the line has to be translated from logical order to the correct visual order, in this case right-to-left. Next, the correct basic form for each character must be selected. Third, they should be joined up into ligatures, etc. If the text engine crashes at Step 1, that is where you get effects like reversed text consisting of isolated forms.
I said it looks like some kind of Syriac font, but it’s too blurry to see exactly what’s going on there, and anyway I don’t presume to know much Aramaic.
Um, I don’t think you did, actually. In post 21, you mentioned Syriac, but didn’t say which one that was. I don’t know any Aramaic, so I’ll take your word for it that it’s a Syriac font. So, probably Chaldean or Assyrian Neo-Aramaic?
Just because this seems somewhat relevant to this topic, I just want to note that if one is using the @ so that it will notify a user, one must spell the user name correctly.