Wednesday, November 23, 2022

SearchResearch Challenge (11/23/22): How to read other scripts and languages?

 I'm never one to turn down an SRS opportunity... 

As you can imagine, in the back pocket of my mind I've got 4 or 5 SearchResearch Challenges queued up and ready to go, questing ever onward bringing you the tidbits of knowledge we all need in our searching lives. 

But every so often, a chance comment tilts the universe in a slightly different direction by pointing out an opportunity for a new SRS topic.  

In a comment this week, Mathlady wrote that "I also encountered the problem of needing to translate signage in an unfamiliar language using an alphabet unknown to me. So I started wondering how to use Google Translate with a different alphabet..."  

Excellent question. 

This is a great chance to talk about how to do exactly that--read in writing systems (what the pros call "orthographies") that you don't recognize.  In some cases (such as Arabic), I can't even recognize the letter boundaries.  

Thanks, Mathlady, for the chance to pose a couple of interesting and potentially tricky Challenges: 

1.  For each of these images below: What is the language?  What is the translation of the text?  





As always, be sure to tell us how you did the translation.  (Did you find a friend? Or did you use an app to do this for you?)  And, as Mathlady points out, how did you enter the text to do your search???  

I'll tell you how I did this next week--but in the meantime, have a great American Thanksgiving!    روز شکرگزاری  or  ਧੰਨਵਾਦੀ   or  Þakkargjörð... or... fill in your own orthography here! 

Search on! 


  1. This could hardly have come at a better time for me. Early next year I will be visiting Hebrew and Arabic speaking countries. I’ve been to that neck of the woods before and have conceded defeat at deciphering their words. I have relied on the kindness of English/French/Spanish speaking strangers. Maybe I can do better.

    Even if I pick up a few words and phrases in those languages, I cannot match them to the written word.

    On a related note, I once worked in an Arabic-speaking country and my North American commuting colleagues and I played a sort of game whereby we would call out the numbers on license plates because the numbers used in that country were not what we in the US call “Arabic” numerals. I think I once saw those other numerals on a clock face in Istanbul, where they longer use Arabic script.

    Heard on the Stanford campus earlier this week:

    “What are you thankful for?”

    “The James Webb telescope.”

    1. Oh - a misleading typo. The last phrase of the third paragraph should read, "where they no longer use Arabic script".

  2. Hi! I took a screenshot, then chopped it down to just the relevant text and saved to Google photos. Then in Photos I used lens and clicked the translate option. This gave me the translation but not the original language. So I clicked on open in Google Translate and the language translated from was then clear.

  3. I normally use the camera in the Google translate app. While it was translating the text, it wasn't showing me the language.

    I was on my iPad.
    I saved each image.
    Launched Google Mobile Search app.
    Tapped the camera icon.
    Selected the image. Translation was working, but again it wasn't always giving me the language.
    Tapped Text.
    Tapped the Select all button under "Test found in image."
    Tapped the bottom Translate button (there are two on the screen at this point.)
    A. Khmer - Learn English Introduction
    B. Hebrew - Significant discount For soldiers in uniform
    C. Amharic - Up to date style (double checked this with an Ethiopian friend. She can only read Up to date in the image.)
    D. Arabic - Deep Excavation (this one also showed a matching road sign on the results page in English and Arabic.)

    Persian, Punjabi and Icelandic detected ;-)

    Typing this up on my computer, just found each image is saved with language name. :-)
    Feliz Acción de Gracias.
    Ich bin dankbar für die Herausforderungen bei der Suche. (via

    1. Yeah... saving the files with the language names probably wasn't the cleverest move. I just forgot to give them arbitrary names. Ah well..

  4. just for grins - from the first image: Bonjour Hej el Hello Здравейте Hola হ্যাল েI Kaabo ສະບາຍດ ີ Përshëndetj Dia duit Ciao こんにちは npwBer Alo Tere Helo नमस्ते
    ជំរាបសរ Saluton_sveiki
    …all can be input in gTranslator as text now…

    found a tool — an Image to Text Converter – OCR (Optical Character Recognition)… roundabout way to arrive…
    know there is an easier Google translate method, especially with the phone, camera, google lens, trans app etc…
    look forward to learning how it should be done… seems it all (the app) is still evolving…
    A. Learn English Introduction - Khmer
    B. Significant discount for soldiers in uniform - Hebrew
    C. Up to date style - Amharic
    D. Deep excavations - Arabic
    from a desktop, then to GooTranslate
    un feliz día de acción de gracias gringo para todos — el año que viene polenta frita con chorizo y brócolini asado con chile y lima
    the way of the Goo…
    helpful bits in the video, but a couple years old…

    it's ALL Icelandic to me… or Punjabi or Persian or
    the Greek variety Ευχαριστία
    sometimes reading the language doesn't help…

  5. Super easy to ID the languages. Used 'save as' and found somebody had conveniently labelled them all. Not you of course. Now of to translations.

    1. Dagnabit... that was me. Forgot to de-identify the languages. Ah well. File names are frequently useful!

    2. clever jon - earmark of a real sleuth - interesting that it saved as 'cambodian' instead of 'khmer'

      was Leonardo Bigollo Pisano a turkey?

  6. $ minutes later: Cambodian/kymer used Google Lens works great: ' Learn English introduction'

    Hebrew Google Lens shows 'A significant discount For soldiers in uniform'

    Amharic: same process "up to date style"

    Arabic; likewise "deep excavations"

    This was indeed fun....

  7. Happy Thanksgiving Day, Dr. Russell and everyone celebrating in the world.

    I did this:

    Downloaded the images and in Google Photos went one by one
    Using Google Lens.
    Translate chip
    Got the answers.

    The first three Google says only, detect language and got the translation.

    In the last one, the language is Khmer. This language is spoken in Cambodia.

    1. Not related to the Challenge but interesting. Today I learned about Lampreys and lamprey pie. Curious that Queen Elizabeth had one in her Coronation and one in her last Jubilee. More curious for me was that lampreys for this last pie came from the United States and not from UK as tradition because apparently there are no more lampreys available.

      Any of you have tasted lampreys or the pie?

    2. no Lpie for me, thanks -

      taste like chicken... I mean squid...

    3. Queen Cersei and Lord Tyrion dined on lamprey pie as they discussed plans for the Battle of the Blackwater.

    4. way too eel-ly/early for this… now if it was a light lavender macaron with jelly…
      the feast
      Hank 1
      carnes murenarum
      fresh water jelly?

    5. I remembered Dr. Russell article about unintended consequences and the invasion of species with asian carp, mussels and rabbits.

      Searched [flying asian carp vs lampreys invasion]

      From 2003: Canada and the United States spend about $15-million a year to control the sea lamprey. And they ate the trouts killing the industry. Lampreys eat a lot

  8. I did a lot of brainstorming while baking, all for nought.

    I asked [Is there an app to recognize languages?] and was pointed to the Google Translate mobile app (which I have never used before). Using my camera, I got

    A. Khmer: “Learn English Introduction”
    B. Hebrew: “Significant discount for Soldiers in Uniform”
    C. ?: “Up to Date Style”
    D. Arabic:”Deep excavations”

    For some reason, I could not discern the languages in C and D, but would guess that D is a form of Arabic. Translating “Deep excavations” into Arabic gave a result that looked like the one in the challenge. I will work on C. Then on to pedestrian signs and rest rooms.

    Happy Fibonacci Day everyone.

  9. Identifying the languages was easy (even discounting that the file names gave it away - I wish they'd been named something different). Using the images for all except C quickly gave the likely languages. Then using my phone and google translate I thought I could translate the 4 terms. However I always back translate and for example, putting "Learn English. Introduction" into Google translate gave me slightly different characters for the word Introduction (assuming it's the 2nd one for sign A).

    Sign B wouldn't apply to me as I'm not a soldier in uniform but I guess I could negotiate to get that significant discount.

    Sign C was quickly identified - but sadly putting my up-to-date style back into Google translate also gave something different. I did try and type this one with the keyboard but either the language was wrong (and so the filename was too) or the keyboard characters were slightly different. Maybe I "am-harry-ing" it too much!

    Sign D was easy - no need for any deep excavating to find this one (although there are several languages that use the same or similar scripts - Arabic, Farsi, Urdu, Uigher and several others (listed at This means that for many languages, it's not just identifying the script. It's also identifying the language using that script. In some cases that can be very difficult when there are a lot of similarities. (Farsi looks like Arabic but has a few differences e.g. the dots on the letters - that can give clues sometimes).

    I did like identifying scripts A and C as they were least familiar. There are lots more that are unusual - Georgian for a European language nobody recognises. There's also Burmese and Lao that look similar (also to the first one).

  10. language lingo - does lens do binary?
    "01001000 01100001 01110000 01110000 01111001 00100000 01010100 01101000 01100001 01101110 01101011 01110011 01100111 01101001 01110110 01101001 01101110 01100111 00100000 00010100 00100000 01110011 01101111 01101101 01100101 01110111 01101000 01100101 01110010 01100101 00100000 01101001 01101110 00100000 01110100 01101000 01100101 00100000 01100101 01110100 01101000 01100101 01110010 00100110"

  11. Okay, folks, it looks like everyone except me got this in a snap and I figured it out with the hint of an app, though my first brainstorm was: “Would Google Images work somehow?” This exercise was extremely useful – as soon as I set foot in another country I will be taking photos of every sign I see.

    Question: Will this camera function work in all scenarios, or might we sometimes need to enter the text ourselves, admittedly slow but oddly fun and satisfying? This was in some sense an intellectual exercise though it was rooted in a real situation in which I found myself.

    Out in that real world I suspect we will not have file names to clue us in about languages.

    remmij – Thanks for the Google Translate video. I think it will be useful. A picture of the mathematician Cahit Arf and his famous Arf function appear on the back of the Turkish ten lira note.

  12. This comment has been removed by the author.

  13. a couple other bits via Kim Komando regarding gLens:
    (fwiw - she indicates this is an android only app, no iPhone...)

    1. I saw a demo of this in 2019; special purpose but useful. Not only can it read text in many languages to its wearer but it has facial and object recognition.

  14. a wooly quandary - it keeps coming back as 'sheepResheep'… not sure of the language…
    another sign to translate…
    ardi resheep jarraitzen du… or something like that
    back to Kafka — Franz might be pleased

  15. Replies
    1. I thought about including Basque as one of the language examples as it's fairly unrecognizable to many people. But I stuck with non-Latin character-based scripts.

    2. Basque is hard to recognise - although Europeans (especially those who have been to North Spain) will have seen it. Finnish is also hard. My favourite though is this one - and even though the translation is on the sign, identifying the language is harder.

    3. Or this one - spoken by over 40m people (and apparently 27th most speakers. ಹುಟ್ಟುಹಬ್ಬದ ಶುಭಾಶಯಗಳು

      And to confuse there's ווינטשן איר אַ זייער גליקלעך ניטל

    4. I got "Peaceful way" in Georgian for the road sign, though neither that nor "Happy journey" seems to match the script.

    5. It is Georgian. In Google Translate, "Way Peaceful" gives გზა მშვიდობიანი - so the right word order but the last two characters are wrong. "Peaceful way" gives მშვიდობიანი გზა so wrong order (and again two wrong characters). Using the keyboard to edit the Georgian gives "The way of peace" which shows how even if you recognise a language, you may not get all the idiomatic and grammatical bits using the automatic translators. However the greeting "Happy Journey" and "The way of peace" are close enough to be understood - neither have a verb so both mean "May your journey be peaceful".

      I always do a back-translate switching between Bing and Google's translators to try and overcome these differences. Even in English there can be variations. In the UK we'd still commonly say "Merry Christmas" or if we want to be politically correct and reflect other religions, we'd say "Seasons Greetings". Nobody says "Happy Holidays" unless influenced by Americans. And at the moment there's the Football World Cup. It's not the soccer world cup as we don't call it soccer.

      I'd also never go out wearing only pants - unless I wanted to get really odd looks. I always wear a pair of trousers over my pants (technically "underpants").

      Language differences can be funny - and are often odd. My favourite - I do interviewing for work and was asking about a particular manager of a company selling Scotch Whisky. This manager had a reputation for sailing close to the wind. My interviewee commented that he'd got the job after finishing his porridge. My response - understanding the British idiom - was to ask what he'd been inside for as porridge is British slang for jail (or is that gaol). I'd love to see what an automatic translator would have made of that conversation.)

      Another example from the past that impacted Google searching. Google used to include synonyms when you added in the ~ operator (now deprecated). If you searched for the metal ~Aluminium you'd get things to do with Aluminium and also a lot to do with the US spelling of Aluminum (not, as far as I know used anywhere else). However a search for ~Aluminum also gave results for Al Pacino, Al Jolson and if I remember, Al Quaeda. I quickly realised that Google was making the chemical symbol "Al" a synonym for Aluminum - as that was the US term, and was not doing so for Aluminium. Hence anything with Al in the name was included.

    6. driving in Georgia -not sure i'd have the phone out using translate…
      "Renting a car in Tbilisi are now becoming more comfortable, easier and more affordable… sure"
      "We reached back in Tiblisi at 7:30 pm on 20 December checked in Konka Hotel. We went to refuel the vehicle, traffic in city was rough with rare road signs.
      We called our agent to hardback the vehicle before given time, he reach our hotel in no time check the car condition returned the security deposit and took handover signatures.
      Driving through the military highway through the Georgia was amazing experience with some astonishing sceneries and people there are clam and gorgeous. Language was the problem for us, majority of the people don’t understand English atoll. Anyway we concluded our trip as per our planned routes without much problems"

    7. This comment has been removed by the author.

    8. Idioms and slang must be the most difficult part of translation. I’ve lived in the US all my life but lately have tried to get to the Irish and British Isles every year and I watch a lot of British TV. I’m constantly learning new words and phrases. I made my first trip to England when I was 21 and came back telling my mother stories about the “funny” things the English said (like “mind the gap”). Her reply was, “Honey, it’s their language”.

  16. Once again, search sherpa Daniel has managed to chart a course for his clowder of Felis catus to
    find their own paths up the side of the Babel-onian tower to shared translational nirvana.
    Now that the tool is found, what will the tool do?… the search begins anew.
    מִגְדַּל בָּבֶל‎,
    herding… search, living the dream
    language death
    ancient writing systems
    giddy up (or is it kitty up?)… and the cat veers off:
    image search

    1. It's been too long since I re-watched "Living the Dream" cat herding commercial for EDS. That's brilliant. (And it definitely reminds me of my job.)

  17. I must say that I feel more than a bit foolish at asking and pursuing a question that was so trivial to the SRSers. Of course there is an app for that! I should have considered that first, especially since that app was already on my phone.

    1. Rest assured, if it was a question for you, it was also a question for many others. Not to worry--this gave us a great opportunity to talk about different writing systems and translation!

    2. not foolish or trivial at all... in fact, a solid launch foundation to search from - plus few things are static
      and it is worthwhile to check for the current state-of-the-art (especially in the image realm) - it almost makes me wish I had a phone. There are no foolish questions, only foolish non-searches... where would math be without

    3. The question itself wasn’t trivial, but the answer seemed obvious to those who responded. The original question was important to me because translating the obscure sign I had seen taught me something about the place where I found myself. I also have upcoming travels which will be made easier, and likely enriched, by my new useful/fun tool/toy.

      I have learned some interesting things about these translating tools which I will share later. In the meantime, I have studied mathematical problem solving for years but have been studying myself as a non-mathematical problem solver. I recently saw this which struck a chord (Oz Johnson, NYT, 25 November 2022):

      “There is a Sufi story I love about the wise fool, Mullah Nasreddin. It goes like this: Darkness had fallen, and Nasreddin had lost his keys. He knelt by a lamppost, searching. A friend joined him, and after a long while, asked, ‘Where exactly did you lose the keys?’ ‘In my house,’ Nasreddin said. The friend said, ‘What? In your house? Why are we looking here?’ To which Nasreddin replied, “’There is more light here.’”

    4. I'm in the dim... from wiki …'s_ring

    5. suitable for lens translation , text & location? handy while traveling - deciphering.
      near Gdańsk

  18. numerical translation…
    what is XLII*
    "42 and the asterisk symbol – Forty-two is the ASCII code for the symbol * also known as the asterisk. This symbol is often thought to translate to anything or everything. In this instance, 42 = everything, the meaning of life."
    'That's it. That's all there is.'
    Narrator: There is a theory which states that if ever anyone discovers exactly what the Universe is for and why it is here, it will instantly disappear and be replaced by something even more bizarre and inexplicable. There is another theory mentioned, which states that this has already happened.
    …hello Google…

  19. something even more bizarre and inexplicable:


    1. Of course mathematical notation is inexplicable to the uninitiated. (Why did I get a spelling error on “unexplicable?”) For example, in the expressions f(x + 2) and a(x + 2) the sets of parentheses have completely different meanings. I seem to remember reading that Richard Feynman devised some alternate mathematical symbols which made more sense to him.

      This made me remember a time in elementary school when our teacher told us to write down a sentence she was going to say. She said, “There are three ways to spell the word –“ and then she said to/too/two (take your pick) and none of us could figure out which word she meant. Yet in everyday conversations we resolve homonyms all the time, which is a form of translation.

  20. I am trying to use these translation tools in real contexts.

    Adventures in translation – the ambiguous email:

    I received an email from a hotel I have booked. It was first in English and then in Hebrew. The English version read:
    “Dear guest Please note, Parking is limited, it is up to available space, we can't guarantee it. first come first served.”

    I put my mobile app up to the Hebrew message on my screen and got
    “Dear sir, the parking spaces at the hotel are limited, therefore it is not possible to guarantee or reserve parking spaces in advance”

    I tried it several more times and got the same message but beginning with “Dear thanks” and “Dear guest”. Sometimes the word “winner” was appended.

    Since I had written text, I put it into Google Translate and got
    "Dear guest, the parking spaces at the hotel are limited, so it is not possible to guarantee or reserve parking spaces on a first-come, first-served basis"

    This last one doesn’t really make sense to me, but, interestingly, uses the phrase “first come, first served”, as did the English message but not the mobile translation of the Hebrew. According to the OED, that phrase dates back to ca. 1542.

  21. Adventures in translation – The cryptic visa

    About a decade ago I worked for a government agency in an Arabic speaking country and was sent a full page visa. Some of it was in both English and Arabic such as my name, profession, and relevant dates. Other information was in Arabic only, and this I have been able to translate for the first time thanks to the Google Translate mobile app. Some of this information is correct such as my state and country of birth, nationality, sponsor’s name, and that I was allowed to earn an income during my visit.

    Much of the rest is incomprehensible to me. There are thirteen lines of Arabic script down the right side of the page. Using the mobile translator on them I have gotten such words and phrases as “aloe vera history”, “the Achilles heel”, “dog leash data”, “the teeth”, and “place of affliction”. The oddest part is that these change as I move my phone or on subsequent reads. I’ve gotten “mental demise”, “residents of the siege”, you get the idea. I even got one about dog fighting and another about breeding camels. Occasionally I get something that might be correct like “Release date” or “Place of birth”. I imagine the “rogue” words are more artifacts of the translator than reflections of what is actually there.

    Are these inconsistencies because the paper is old, or the lighting differs from time to time, or issues related to translating script from right to left into left to right? Maybe the virtual translator is capricious. Having read passages of The Metamorphosis from different translators, I wonder if the virtual translators are subject to issues such as interpretation and making inferences from context, as human translators are.
    I really should study this more systematically.

  22. odd - now I'm only getting Kazakh links

    entered with voice, worked fairly well - the queen gets demoted to a princess, the tea disappears and lumps become humps...
    still camel themed
    "Ханшайым айтқандай, бір-екі өркеш"

  23. used lens, example - resulting SERP
    how to:

  24. I had an unexpected linguistic experience with Google. I rely on Siri to turn my phone’s torch on and off but I wanted to learn how to do it manually. So I asked [Where is the torch on iPhone?] and was directed to the “Control Centre”, etc. I wondered why the British spelling was used. After mulling it for a while, I asked [Where is the flashlight on iPhone?] and was directed to the “Control Center”. Interesting.

    Of course then I had to try [Où est la lampe de poche sur l'iphone?] with predictable results. But somehow distinguishing between the two English language words surprised me more. Both queries were in English and differed only by one word. On reflection, it makes sense but was unexpected nonetheless. I wonder how it will handle American regionalisms.

  25. the more you notice, the more curious things may become… SORI won't speak to me… (isn't SIRI an  thing? not Goo - I have no face id…)
    wiki AERV
    another grinder from SIRI - she said she had the munchies
    in nuevo England… hella good
    there's almost always a list
    a (real) torch alt - not to be confused with a flame bro
    also may be on your "dog and bone"
    …could math be a virus 2?
    a block buster
    "the creative act is not performed by the artist alone…" Marcel Duchamp
    a generator haiku:
    An flamebeau flashlight same...
    A flamebeau searchinto the tone,
    curious! resesarch again.

    *Si el botón de la linterna del ícono de la linterna no está visible cuando abre el Centro de control, puede cambiar su configuración para agregarlo. Vaya a Configuración > Centro de control, luego toque Torchlight en la lista de controles disponibles.

  26. Regionalisms are likely to cause problems for any voice recognition technology - at leat for a while, as this comedy sketch illustrates:

    But more seriously, MacWorld tried out Siri with a few UK regional accents to mixed results (although I think the Scottish Elevator / Lift makes the point slightly better).

  27. And here is another video showing accent problems - this time French speaking English.