Universal alphabet

« previous post | next post »

Not that I think this is any sort of panacea, but our good friends at BBC have seen fit to ask: "Could a new phonetic alphabet promote world peace?"

Although backers of this supposed universal alphabet claim that "it will make pronunciation easy and foster international understanding", I have doubts that SaypU (Spell As You Pronounce Universal project) constitutes a viable route to world peace.

It is curious that SaypU, which uses 24 letters, dispenses with precisely those three letters that give persons who are unfamiliar with Pinyin (official PRC Romanization for Mandarin) the most trouble:  "c", "q", and "x".  It also adds a reverse "e" for schwa, but, since standard keyboards do not have this symbol, the asterisk (*) may be used to represent it.

The home page of the SaypU project has a conversion tool that you can use to render various traditional spellings into SaypU.  For example, if you enter

English "Let's meet at Leicester Square", the result in SaypU is "Let's miit at Lestɘr skwer."

Turkish "Ben İstanbul'da yaşıyorum" ("I live in Istanbul") gives "Ben İstanbul ' da yashɘyorum."

Vietnamese "Tôi ăn một bát phở, được không?" (lit., "I'll eat a bowl of phở, okay?", i.e., "May I have a bowl of pho?") yields "Toy an mot baat fɘɘ, dɘɘk khong?"

Hindi मेरा नाम बहादुर है ("My name is Bahadur") gives "Mɘeerɘaa nɘaam bɘhɘaadɘur hay."

Japanese ここに 語句 もしくは 文章 を 入力 して 下さい ("Please enter a text") results in "Kokoni go ku moshikuwa bunshoo wo nyuuryoku shite kudasai."

Georgian იოსებ სტალინი იყო საქართველოდან ("Joseph Stalin was from Georgia") yields "ioseb stalini iko sakartvelodan."

After you do the conversion to SaypU, you can push a "play" button and the pronunciations will be read out — quite accurately, it seems.

When I showed this BBC article to my colleagues at Language Log headquarters, the only responses I received are these two:

"The Babel fish, by effectively removing all barriers to communication between different cultures and races, has caused more and bloodier wars than anything else in the history of creation."

–Douglas Adams, The Hitchhiker's Guide to the Galaxy

"Il est faux de prétendre que les peuples et les personnes humaines se foutent sur la gueule parce qu'ils ne se comprennent pas.  Ils se foutent sur la gueule parce qu'ils se comprennent."

– Romain Gary, Pseudo

[Thanks to Mark Liberman, Paul Kay, Deven Patel, Philip Lutgendorf, Katherine Wang, Erika Gilson, Eric Henry, Peter Golden, John Colarusso, and Bill Hannas]



  1. Martin J Ball said,

    February 27, 2013 @ 8:16 am

    So, the Vietnamese alphabet, which usefully tells us something about the tones to be used, gives us a SaypU version where tones are totally ignored? This is a step forward…?

  2. Rowland said,

    February 27, 2013 @ 8:23 am

    Because it's not like there's already any International Phonetic Alphabet or anything…

  3. Jason said,

    February 27, 2013 @ 8:48 am

    I wouldn't be completely cynical about this, despite it's naivity. Depending on how well it works, I could see this replacing the IPA for multilingual dictionaries, if there was enough interest. As a jenral orfografi intended tu replase Inglish or uther orfografi, however, this stands the same chanse of suxess as all the uther futecheristik or uetopeian propozels for reforming Inglish orfografi — that is tu say, alas, vurtchualli nun. Peepel just dont wont tu expend the effort tu lern a nue skript all over aggen, let alone the inkompatibilities with all the mateerial in legasi spelling form.

  4. NW said,

    February 27, 2013 @ 9:39 am

    So, let's see, someone's invented a way of writing down the sounds of a language, using only some English letters and no confusing diacritics, and then spoilt this elegant and intuitive simplicity by a backwards E. This radical scheme – is it pronounced "say poo" or "say pew", as I can't tell from the way it's written?

  5. Gianni said,

    February 27, 2013 @ 9:39 am

    We need it for world peace and the elimination of errors in communication and translation.

  6. Ben said,

    February 27, 2013 @ 9:46 am

    Proper Turkish please!

    "Ben İstanbul'da oturuyorum."

  7. Pedro said,

    February 27, 2013 @ 9:50 am

    This reminds me of Bliss, the writing system invented to end all cross-linguistic misunderstanding.

  8. Simon K said,

    February 27, 2013 @ 10:02 am

    Because of course the major barrier to world peace at the moment is communication errors. That's why you never hear of people who speak the same language going to war with each other, right?

  9. Ellen K. said,

    February 27, 2013 @ 10:08 am

    It curiously seems to use a rhotic British standard. "top" is "top", indicating a British rounded vowel. Thus one would expect that RP is that standard. Yet the spellings are rhotic. And there's no acknowledgement of the existence of varying accents in English. Seems to me it should properly have a statement about what accent it's transcriptions are based on.

  10. GeorgeW said,

    February 27, 2013 @ 10:10 am

    If only Democrats and Republicans in Washington had a common orthography what a wonderful world it would be.

  11. J. Marshall Unger said,

    February 27, 2013 @ 11:05 am

    I once heard Mary Beckman, arguably today's leading experimental phonetician, answer a student's question, "What does IPA stand for?" with the brilliantly understated lesson "International Phonemic Alphabet." That is, from her standpoint, the IPA is just an attempt to come up with an inventory of typographic symbols broad enough to let you phonemicize any language without resort to digraphs, and mnemonic enough to prevent it being too hard to use. SaypU seems to be just an inferior move toward the same end.

    The quotations from Gary and Adams (the latter, one of all-time my favorites) highlight another point: language is, strictly speaking, speech, not writing. The irony is that people believe that language is merely a code that can be excised from the proximate and longer-term cultural contexts in which it is used, a notion promoted by the fetishization of written language as the "real thing" popular among prescriptivists.

  12. Joshua said,

    February 27, 2013 @ 11:08 am

    What, Tengwar isn't good enough any more?

  13. chris said,

    February 27, 2013 @ 11:30 am

    And there's no acknowledgement of the existence of varying accents in English.

    Isn't this always the bane of phonetic spelling? When two speakers of the same language pronounce the same word differently, do you spell it inconsistently or allow the phonetic spelling to break for one of them? Then, over time, the pronunciation changes and the spelling stays the same — or you can't read old books.

    I have a hard time believing something based on the Latin alphabet can possibly have enough vowels even for non-tonal languages, too.

    Also, even if it works, how much benefit can you really derive from being able to pronounce something you still don't understand?

    P.S. Not sure if the fact that the abbreviation can be pronounced "say poo" is a subtle hint that the whole thing is a joke, or was just overlooked.

  14. J.W. Brewer said,

    February 27, 2013 @ 11:32 am

    The three letters omitted from the usual set of 26 are those that can most obviously be scrapped in a phonetic/phonemic orthography for English; I assume their peculiar status in pinyin is a coincidence (unless it's a result of the same thing – that their seeming redundancy in certain other languages led the designers of pinyin to think they could be more or less randomly reassigned in ways that Anglophones would find non-inuitive).

  15. John Lawler said,

    February 27, 2013 @ 11:43 am

    "It is by universal misunderstanding that all agree. For if, by ill luck, people understood each other, they would never agree." – Charles Baudelaire

  16. Dan Lufkin said,

    February 27, 2013 @ 11:52 am

    I was about to ask whether anyone knew the fate of G.B. Shaw's elegant phonetic alphabet, but I Googled around a little and can report that it lives on, as one might expect, as a hobby in the UK. (Is there anything that is not a hobby in the UK?)

    I found Shavian easy to learn to read and once had the edition of Androcles and the Lion that the Shaw estate commissioned. Ellen K's remark about the necessity of having a standard accent is cogent. Shaw was severely non-rhotic and the transcription of the play was sometimes hard to decode. I think that I'll take up Esperanto written in Shavian as my world peace project. Contributions cheerfully accepted.

  17. Robert Furber said,

    February 27, 2013 @ 12:03 pm

    Skwer for square seems to represent an accent with the merry-Mary merger. But it could also represent a Lancashire or a Liverpudlian pronunciation (though the former would probably be written skwər). It seems unlikely, therefore, that the transcription was based on RP.

    Curiously, the page for pronunciations of the letters advises the Irish/Cockney /heitſ/ pronunciation for the name of h.

  18. Victor Mair said,

    February 27, 2013 @ 12:16 pm

    From Peter Golden:

    This could be fun, but the possibilities for confusion are endless. Moreover, since it is "say it as you pronounce it," what about dialect pronunciations? I had a great aunt who grew up in the west, on the frontier (Colorado mining towns, among other places) and one of her favorite expressions to dismiss someone was "he ain't worth the powder to blow him to hell." "Ain't" came out something like "haint," "hell" was "heyal" – "worth" approached "warth." When she learned that I was going to grad school at Columbia's Russian Institute, she cautioned me to be wary "of them Commonists." My son's North Carolinian in-laws pronounce "where" as "wahr" etc. In addition, "y'all" has variant pronunciations and tonal inflections, each of which add different nuances. A Brooklyn accent would be a SaypU dream.

    It is a Pandora's box.

    That great aunt (on my mother's side) was a real character and told wonderful stories about her days in Arizona, among other places, before it was a state. She ended up in Columbus, Ohio, after detours in Chicago and Oklahoma. Her daughters (my mother's cousins) all spoke with Oklahoma accents. She occasionally "came East," as she phrased it, to see her kinfolk (my mother's family) that had ended up in NYC. To my great disappointment as a child much taken with Westerns, she said that she had never witnessed a shootout, Wyatt Earp style.

  19. Lazar said,

    February 27, 2013 @ 12:21 pm

    The vowels "a e i o u ǝ", combined with "y" and "w", seem workable enough for English with its glidey free vowels, but I don't see how they could accommodate the the vowel inventory of French. /ɔ/ and /o/; /ɛ/ and /e/; /œ/ and /ø/ and /y/; you'd either have to obscure a bunch of phonemes or make up a new, convoluted set of spelling rules for them (giving us a third standard to learn in addition to the traditional orthography and the IPA).

    @Ellen K., chris: Well, if I were making a reformed orthography for English, I would aim for something diaphonemic (i.e. using an inventory of historical phonemes triangulated from General American, RP, and perhaps other varieties) – thus it would be rhotic, LOT-PALM and LOT-THOUGHT distinguishing, lacking the TRAP-BATH split, lacking the LOT-CLOTH split, and so on. That seems like the only alternative to a total balkanization of English spelling.

    @J. Marshall Unger: That matter of the "International Phonemic Alphabet" is the basis for Luciano Canepari's criticism of the IPA – he points out that it's poorly suited to narrow phonetic transcription, and he's proposed a voluminous replacement for it. I don't agree with all his observations and prescriptions, but his work has prodded me to think about phonetics on a more detailed level, and I do find myself increasingly ambivalent about the official IPA as a phonetic script. Why are there symbols for [x] and [χ] but not for dental and alveolar [t], or for apical, laminal and non-sibilant [s]? Why is there a symbol for [ǝ] but not for mid [e], mid [o] or central [a]? Anyway…

  20. Breffni said,

    February 27, 2013 @ 1:13 pm

    Dan Lufkin: 'Shaw was severely non-rhotic' – on the contrary, judging from these videos…


    …I'd say that Shaw had a basically rhotic accent, typical of the Anglo-Irish Dublin speech of his day, with arguable non-rhotic intrusions from time to time, like 'sir' and 'there' at around 00:45 in the second one, which contrasts with 'year' a few seconds earlier.

  21. Levantine said,

    February 27, 2013 @ 1:20 pm

    Ben, 'İstanbul'da yaşıyorum' is perfectly good Turkish, even if the sentence would sound more idiomatic with 'oturuyorum'.

  22. peter said,

    February 27, 2013 @ 1:26 pm

    J. Marshall Unger said (February 27, 2013 @ 11:05 am)

    "The irony is that people believe that language is merely a code that can be excised from the proximate and longer-term cultural contexts in which it is used, a notion promoted by the fetishization of written language as the "real thing" popular among prescriptivists."

    Fetishizers of the written over the spoken have waged a long campaign in the West. One of their early successes was the switch from oral to written examinations in the mathematics tripos at Cambridge University in the 1760s, on the basis that written exams were allegedly fairer. Russian and East European mathematicians still manage with oral examinations.

  23. Ellen K. said,

    February 27, 2013 @ 1:33 pm

    Chris, no, lack of acknowledgement of the existence of different accents is not always a bane of phonetic spelling attempts. No reason it should be.

    Now, the existence of different accents will always be an issue; a bane if you prefer. But acknowledging that it's an issue is simple and easy enough.

  24. X said,

    February 27, 2013 @ 1:43 pm

    It seems like the primary use of such a thing would be for tourist guides, where one knows the user will have no chance of correctly pronouncing anything, but the given phonetic spelling might get them to blunder close enough for a native speaker to figure out what is intended: "may hɘvɘrkraft iz ful ɘv iilz"

  25. J.W. Brewer said,

    February 27, 2013 @ 1:48 pm

    Wouldn't a hypothetical purely-phonetic orthography (especially if used cross-linguistically) tend to help defetishize the written word by making it easier for ordinary folk to conceptualize of writing as merely a transcript of speech? The gizmo that turns SaypU texts into an understandable robot voice would seem to be particularly helpful in that regard. Me, I think many modern linguists/anthropologists tend to overshoot in reaction and instead fetishize the spoken word in ways which may tend to obscure how language use actually functions in highly-literate modern societies – including the fairly recent phenomenon of being able to non-verbally interact with others in e.g. blog comment threads.

  26. Faldone said,

    February 27, 2013 @ 2:04 pm

    @NW: is it pronounced "say poo" or "say pew", as I can't tell from the way it's written?

    According to the orthography of SaypU it should be pronounced /saɪpʊ/.

  27. Antariksh Bothale said,

    February 27, 2013 @ 2:23 pm

    I wrote about the SaypU project on my blog. Frankly, I find LL's treatment too charitable. Even as romanticized universal alphabets go, the design of SaypU seems particularly ignorant.

    "A simplified universal alphabet would end not only misunderstanding. It would help foster peace around the world, he believes."

    This seems to come from a person who’s never come out of their English bubble, which is rather weird because the inventor has been described as a Syrian banker, so I’d assume he’d at least be familiar with Arabic. I dunno how SaypU is supposed to represent Arabic sounds, and exactly how it would help even if I could bypass the Arabic script if I didn’t know Arabic in the first place.

    Nor do I understand how merging all the tonal and many of the consonantal differences of Mandarin would end misunderstanding.

    “If people pronounce and speak in the same way it makes people feel closer to one another. I do think the world with a single alphabet would be a more peaceful place.”

    Dialects don’t exist because of ambiguity in spelling. English spelling is fairly standardized (even the US/UK differences are puny), yet you’d find hundreds of English dialects. This is NOT about spellings systems. It’s about language variation, which happens independent of orthographic representation.

    I don’t think the inventor of SaypU even understands what it means to have a Universal Alphabet. You need a writing system that can (fairly) unambiguously represent sounds from all the languages of the world. The International Phonetic Alphabet (IPA) is the most widely used standard for this purpose. It has over 100 characters and over 50 diacritics.
    I don’t know how SeypU is expected to represent any non-English sound with its 24-character set.

    Feel free to read the whole post (click on my name above) if you wish.

  28. Peter S. said,

    February 27, 2013 @ 2:56 pm

    Considering that carry is "kari" and marry is "meri", I don't think this enterprise is very well thought out.

  29. Cy said,

    February 27, 2013 @ 3:03 pm

    This is just another ridiculous attempt – I mean, the problems with learners and tourists pronouncing non-native languages is that they get wrong the sounds that are completely different and "foreign" from their language, and they get glaringly wrong the things that are CLOSE to sounds in their own language. As many above alluded to, this is somehow broader than the broadest IPA transcription.

    To say nothing of tonal problems, which are a real thing. You can't just skip them, even though in context the silly mistakes listed in humor columns never really occur. Of course people understand the context. But if the pronunciations guided by this system are as far off in the consonants and vowels, then there's no chance for speakers to somehow auger the other segmental levels, like stress and tone. And the cognitive barrier for learners after the inevitable modifications is just – well, try reading the latin system for Hmong without knowing the language natively first.

    I can say that the translation for Vietnamese doesn't even approach either of the two primary dialects, and has deeper problems than the English transcription mentioned by many commenters. I tried a few phrases and it's a simple Caesar cipher the program applies to some list of graphs – no narrow transcription, no non-english phones, no phonotactics, no liaison. This is dumb programming. Probably statistically, someone reading this transliteration and someone reading the original orthography would both have the same amount of errors, ignoring the tone problem.

  30. Cy said,

    February 27, 2013 @ 3:15 pm

    As Antariksh Bothale notes above, LL is being quite generous to this thing. But I'd like to point out the generosity shown also to the BBC website. It is truly insane how poorly researched and written and conceived the linked story is.

    Since I'm way down here and nobody reads that many comments, I'll just illustrate a few problems that make me think they don't have journalism schools in the UK!

    You probably would have gotten cheese instead of pho.
    "We will bury you" was still pretty grim in its "true" translation.
    Syria is not torn asunder by shibboleths. I don't think our banker understands politics or humans very well.
    I know nothing of Finnish but I'm going to say it is not 100% phonetic, unless every Finn sounds exactly alike and children aren't taught to read.

    Do all British publications treat language the way that Americans treat religion?

  31. E said,

    February 27, 2013 @ 3:55 pm

    It might be a wonderful system, but I'd still be unable to read that name as anything but "say P.U.".

  32. Vireya said,

    February 27, 2013 @ 4:44 pm

    The Hindi is just plain silly. It doesn't represent how the words are pronounced at all. They have just transcribed the "sound" of each vowel and consonant, ignoring how they work together. So for मेरा, (mera, my), they have Mɘeerɘaa, as if there were four syllables; Mɘ ee rɘ aa. This is wrong because the word has only two syllables, "me" and "ra". If a vowel sound is written as part of the consonant, it replaces the inherent unwritten ɘ.

  33. Rodger C said,

    February 27, 2013 @ 8:55 pm

    Is it too late to adopt 'Phags-pa?

  34. Matt said,

    February 27, 2013 @ 11:29 pm

    I read this article as a sort of… what's the word for something that would be satire, if it weren't actually meant seriously? — one of those, on how ridiculous and out-of-touch privilege can make people. I mean, when a normal person realizes that they actually can't do something they want to, they either resign themselves to that fact or start the long, hard process of learning how to do it. They don't propose that the entire world dumb down its approach to that activity so thoroughly that even their own amateur, inept attempts at it become technical successes.

    On the other hand, the unintentionally insulting sidebar headline "Why does anyone learn Esperanto?" made me smile, so overall I think the link was worth it.

  35. Ran Ari-Gur said,

    February 28, 2013 @ 1:30 am

    @Matt: Re: "what's the word for something that would be satire, if it weren't actually meant seriously?": The phrases "unintentional satire", "accidental parody", and so on, have all had some uptake. Alternatively, I guess we could call it a "reverse Poeism", after Poe's law.

  36. maidhc said,

    February 28, 2013 @ 4:02 am

    Pedro mentioned Bliss further up, but Bliss's idea was quite different. He tried to create a graphical writing system based entirely on meaning, that everyone could read in their own language.

    Bliss was stuck in Shanghai during WWII and he was impressed with how Cantonese and Mandarin speakers could read the same text.

    Bliss's ideas have worked quite well in limited spheres such as traffic signs or the signs that you see in airports directing you to ATMs or luggage pickup (not his actual symbolics though). Whether the system could be expanded to the extent that you could print a newspaper, for example, is another question.

    It's not hard to find examples of people speaking the same language fighting nasty wars, so I don't think this is the key to world peace.

  37. Filip S. said,

    February 28, 2013 @ 7:20 am

    As has already been noted, even if you ignore the ideology behind the project, the transcription scheme itself is pretty clumsy for some languages – even if you ignore the tones.

    For Polish, in particular, it seems to completely ignore some distinctions, like that between palatal and postalveolar affricates before other consonants ('ćma' and 'czma' both end up as 'tschma'), or that between stop+fricative consonant clusters and genuine affricates (these are distinctive in most dialects!).

    Why get rid of "c" when it's such a good candidate for palato-alveolar affricates? The transcription scheme seems to be particularly deep in the "English bubble", looking slick when shortening English words, but at the same time introducing 'kh' and 'tsch' into Spanish and Indonesian. And what did Hindi do to deserve monstrosities such as "Mɘeerɘaa"?

  38. Pete said,

    February 28, 2013 @ 8:33 am

    Has anyone read the BBC article. It's incredible – almost every sentence is gobbledegook!

    It begins: You are in Vietnam and want a bowl of soup. You ask a local where you can get "pho". After momentary confusion you are handed a book. It's the curse of phonetics. Pho was correct. But you failed to emphasise the vowel and so articulated in Vietnamese "copy" (of a book).

    If "pho" means "book" then it wasn't correct! And what does "emphasise the vowel" mean? Presumably it means "use the correct tone" but tones are completely ignored by SaypU, so it wouldn't have helped anyway.

    The next paragraph reads English has more pitfalls than most other languages. "Don't desert me here in the desert" is a classic example of the heteronym, words spelt the same but pronounced differently. Bill Bryson remarked in his book Mother Tongue that there were nine separate pronunciations of hegemony.

    The claim that English has more pitfalls than other languages is a classic of pop linguistics. But the competing pronunciations of "hegemony" are not the same thing as "desert" (verb) vs "desert" (noun) at all.

    I could go on – there are as many errors as there are sentences.

    I particularly love the fact that the expert linguist the article quotes is Bill Bryson!

  39. Pete said,

    February 28, 2013 @ 8:41 am

    @Cy: I've read your comment! You're completely right – it's a disgracefully shabby article. I've pulled the first two paragraphs apart in my comment above too.

    Also the new alphabet itself is stupid. It doesn't even represent English properly. For example it has no symbol for ʃ (it uses "sh"); it uses the same symbol "j" for ʒ and ; it has no symbols for θ or ð and uses the same digraph for both.

  40. Ted said,

    February 28, 2013 @ 10:26 am

    @maidhc: Whether the system could be expanded to the extent that you could print a newspaper, for example, is another question.

    You can judge for yourself after you read
    this post from Victor Mair

  41. Gianni said,

    February 28, 2013 @ 10:29 am

    Most of the reply here are from a linguistic background. However, the application of such kind of alphabets will facilitate communication, which is beyond linguistics. Any claim that these alphabets lack accuracy ignores the fact that in everyday communication, a mutually understandable form of writing different languages is needed so that errors and misunderstanding can be reduced.

    One example is phrasebooks. It must be better for a worldwide traveler, whether for business or for leisure, to learn basic phrases through written materials. A unified written system can facilitate you to cross the borders or fly among different destinations in several days without being bothered by any non-familiar diacritical or irregular marks.

    Admittedly, you cannot utter the accurate words like natives, but this give you a simple way of solving problems. For practical reason, this kind of idea is innovative. The phrasebooks for UN security forces or Peace Corps can be improved based on these ideas without formal linguistic training to the individuals so that international aid can be more smooth. Those asking for accuracy should help refine such projects for win-win results, but not destroy it from the academic perspective.

    From an academic perspective, I am wondering how this alphabet can be improved to represent the click consonants widely attested in Khoisan and Bantu languages. To avoid any kind of linguistic Chauvinism for a universal alphabet, major languages of all over the world should be taken into consideration.

  42. T.Merkel said,

    February 28, 2013 @ 12:42 pm

    Funny that Georgian is included — probably one of the most faithfully matching writing and pronunciation systems out there. (And so typical of a non-Georgian to relegate ყ and ქ to the same sound!) In fact, we wouldn't do so badly if tried to used Georgian as a universal alphabet — we'd run disgracefully short of vowels, to be sure, but it would certainly cover a lot more languages a lot more faithfully than Latin, especially if you also include the Megrelian symbols for shva and elifi.

  43. a George said,

    February 28, 2013 @ 4:55 pm

    @Cy: "Since I'm way down here and nobody reads that many comments, I'll just……"

    Well, I read the comments on this weblog because I want to learn something. It would be pretty stupid to break off at some arbitrary point. So, unless the subject is so esoteric that I cannot comprehend anything from the original post, I read to the end. Sometimes I even revisit, but that is a luxury that you can only afford if you follow few weblogs.

    @Gianni: “One example is phrasebooks. It must be better for a worldwide traveler, whether for business or for leisure, to learn basic phrases through written materials. A unified written system can facilitate you to cross the borders or fly among different destinations in several days without being bothered by any non-familiar diacritical or irregular marks. Admittedly, you cannot utter the accurate words like natives, but this give you a simple way of solving problems. “
    I have historical phrase books with phonetic transcriptions that are very useful for understanding the principles behind their construction. One is “Praktisches Gesprächsbüchlein für den Kriegsdienst. Deutsch-Rumänisch mit der richtigen Aussprache“. This Austrio-Hungarian phrasebook was used by my grandfather in WW1, and there is nothing but phonetic transcription, no indigenous text – in other words,“die richtige Aussprache”. Another is “Deutsch-Norwegischer Soldaten-Sprachführer. Auch in Dänemark zu verwenden“. This was used by German soldiers in Denmark in WW2. This is somewhat better, because there is also the Danish written version.
    But in my view, phrasebooks do not really serve the non-local traveller well in uttering what he/she wants. Rather, they can serve to aid your ear in recognising the utterances of the locals and understanding them. Once you have done that, if you have the capacity for imitation, you will be able to imitate most of the local sounds, and even tonality. That way you may gradually become a better traveller. Clicks is a different matter, that is something you have to learn, like circular breathing for playing wind instruments.

  44. JS said,

    February 28, 2013 @ 7:22 pm

    Many comments above point to phonological contrasts within various languages which this proposal fails to provide representation for…

    But the creators' intent could hardly have been to represent all attested segments or contrasts; as has been noted, the IPA is the (however limited) tool for that purpose.

    Instead, here, the phonological inventories of many languages not-X are mapped to the inventory of a language X, naturally enough resulting in numerous imperfect matches and cases where multiple sounds are mapped to one; e.g., the Vietnamese vowels /ɨ/ () and /~ɘ but frontish and pursed?/ both end up as /ɘ/ (a case exemplifying both of these mapping problems.)

    SO, the real problem (or rather, misrepresentation) here is that X = English, more or less: such a system can have no claim to universality, as a particular phonemic inventory has been implicitly adopted to serve as baseline. CPU could conceivably become a tool with which English speakers might deal with other languages, but not one with which speakers of not-X might sensibly deal with other-not-X, as those two languages will have different sorts of compatibilities. To give a trivial example of this very major problem, take the Japanese vowel generally represented as IPA /ɯ/ and not so distant from Vietnamese /ɨ/ (): the former becomes /u/ in CPU and the latter /ɘ/, so a hypothetical Japanese reader of CPUed Vietnamese will be dramatically short-changed, or if you prefer, "watered-down":

    It's as if, ourselves only using and only able to perceive the colors green, orange and purple, we were to devise glasses capable of filtering the colors of all of the paintings of the world into green, orange and purple… and then we suggested that EVERYBODY use them.

  45. JS said,

    February 28, 2013 @ 7:26 pm

    ^ oops, Quốc Ngữ horned u/o gone missing above… (maybe here too)

  46. Matt said,

    February 28, 2013 @ 8:11 pm

    The phrasebooks for UN security forces or Peace Corps can be improved based on these ideas without formal linguistic training to the individuals so that international aid can be more smooth.

    No, they cannot be improved based on these ideas, because these ideas are terrible. That's like saying that global health outcomes could be improved if we gave everyone garden tools and told them to have at it when they saw someone suffering from disease or injury. In the vast majority of cases, shovels and rakes wouldn't be "better than nothing", and in any case there's already something, even for those without formal medical training. There are first-aid kits and AEDS and all sorts of useful tools that non-doctors can use. There are professionals whose entire job is to develop ways for non-doctors to be more useful in emergencies. Garden tools wouldn't even be in the top hundred, the top thousand proposals for addressing this issue.

    I don't know what the phrasebooks for UN security forces or peace corps are like. Maybe they really could be made more user-friendly for people without linguistic training. But if so, there are already dozens, hundreds of approaches to the problem. There's no reason to suppose that yet another amateur reinvention of that wheel will give us better results than those we already have.

  47. Gianni said,

    March 1, 2013 @ 11:21 am


    What I feel uncomfortable to many ideas criticizing these stuff is their simple assertion of its futility. Specialists can help amateurs to develop better tools, in contrast.

    The history of tools itself is evident that tools were improved by all users, not specialists only. I am surprised to see so many opposing ideas for a new tool that has the potential to facilitate multilingual communication, which is unprecedentedly seen in the previously dominant bilingual circumstance.

    Time will give an answer to the development of human writing. As a tool for communication, accuracy and convenience must be achieved in balance. Historically, convenient writings always replace complex ones, given that no writing is absolutely accurate.

    @a George

    I respect the Austrian-Hungarian tradition, but words and phrases have evolved. We need new books in the global times for those global travelers, the number of whom is becoming larger and larger. Global traveling that requires multilingual communication in days and hours, for business and leisure, are no more the privilege of the learned and the elite. Specialists can help facilitate and improve these stuff if they want.

  48. Barbara Phillips Long said,

    March 2, 2013 @ 3:43 pm

    In the Japanese transcription, I ran into another problem:

    "Kokoni go ku moshikuwa bunshoo wo nyuuryoku shite kudasai."

    I was mentally sounding out the words and read "shite" as though the e was silent, because in English it would be. But in this transcription, I assume there are no silent letters, so I made the mistake of reading something that resembled English as though it was English, even though I didn't make that mistake with other words in the transcription.

    The alphabet isn't unfamiliar enough for me to revise my mental default-to rules, apparently, so I switched back and forth between reading letter-by-letter and reading English-style.

  49. David S. said,

    March 2, 2013 @ 7:08 pm

    @Dan Lufkin: Esperanto in Shavian is definitely already a thing.

    @Gianni: This is a re-invention of the wheel. It's not a new tool; we have the IPA and a number of friends, any number of conscripts, and going back further, Latin, Chinese, Arabic and Cyrillic have been adapted to write many languages. Phrasebooks already have ad-hoc phonetic alphabets, which is all this will be until a number of people start using it. Where's the evidence it's better then what we have?

    Specialists get tired of amateurs coming up with "cool new ideas" and acting like the specialists should drop everything to work on something that's been done before many times.

  50. Gianni said,

    March 2, 2013 @ 9:32 pm

    @David S

    If you read another of Prof. Mair's post (http://languagelog.ldc.upenn.edu/nll/?p=4498) about a refugee in Nepal, you will probably agree that one can transcribe the record of her speech to help identify her original language with a universal writing based on modified Latin script. Of course you can use IPA, but it far exceeds the number of the key on the keyboard. At this time, we are doing international aid. Therefore a universal script is better than any ad hoc invention. It would be welcomed if Chinese, Arabic or Cyrillic be used for such situations, but to me a revised universal Latin alphabet is better.

  51. David S. said,

    March 3, 2013 @ 8:58 pm

    IPA does not far exceed the number of keys on the keyboard. I count 105 basic IPA characters; there are many 105 key keyboards. As a practical matter, my keyboard has 47 non-number pad, non-control keys, which let you have 94 characters with a single shift state (since IPA doesn't have upper and lower case). With some compose keys, and an easy way to shift keyboards, it's definitely typeable.

    If you're stuck in ASCII, you can use X-SAMPA. If you want a transcript of her speech, your alphabet simply won't do. There are a number of sounds that it can't represent at all, and worse, it can't distinguish between any number of pairs of sounds represented distinctly in IPA and thus in X-SAMPA. If you're trying to parse out someone's accent through a text representation, those details are going to be critical.

    You say a universal script is better then any ad hoc invention, but SaypU is more ad hoc then X-SAMPA. You have to earn universality, or occasionally have it dictated from above. SaypU hasn't done that.

    And the best way, in the 21st century? A simple audio recording. No specialized training needed to make, no fine details obscured.

    (Universal-wise, some languages have writing systems finely tuned to them, with features you're obliterating. You can't record Irish's contrast between palatized and velarized consonants; you can't record tones used in Vietnamese and Chinese; you can't record several consonants in Hungarian or Hindi. It's not effectively useful for all or even most of the world's languages.)

  52. Gianni said,

    March 3, 2013 @ 10:45 pm

    @David S

    I would prefer WorldBet than X-SAMPA, though WB is even less convenient for keyboard user than IPA. I don't think an X-SAMPA text is readable to non-specialists. Of course, I am not a specialist of alphabet and have not compared different alphabets for specific languages. I am an amateur in the studies of alphabets, so I value convenience more than accuracy if I have to write something down.

    Your suggestion of audio recording reminds me of this article:


    I agree that languages are better to be recorded and translated orally than via the written media in our age.

  53. Chandra said,

    March 5, 2013 @ 5:22 pm

    Gianni, saying that you don't need a linguistics background to create an alphabet that accurately captures the pronunciation of all the world's languages is like saying you don't need a background in culinary arts to prepare a banquet that will please the palates of a thousand people. Sure, any random person can walk in off the street and give it a try, but that doesn't mean anybody's going to swallow it.

  54. Ray Dillinger said,

    March 5, 2013 @ 11:48 pm

    I once attempted to create a phonemic alphabet for English. I used twelve vowels and twenty-two consonants. It wasn't as good as the system Shaw had already come up with. And I wasn't such a fool as to think it would represent, say, Spanish, German, or French (all of which are relatively closely related to English) without modification. Once you start down that path, you wind up with IPA, or something very much like it.

    I have never understand why anyone would say that having a common alphabet/language/whatever will help prevent wars. Even the briefest and most cursory study of history will refute that ridiculous notion. Many of the bloodiest and most brutal wars have been fought between folk who shared a language. Wars are fought due to genuine conflicts of interest or competition for a scarce resource, and language differences have hardly ever been important.

  55. Victor Mair said,

    March 6, 2013 @ 12:23 pm

    From Jaber George Jabbour, Director of the SaypU Project (he tried to post this several times himself, but it didn't go through):

    Thank you all for the feedback. We would like to stress that SaypU is a collaborative experiment that is still at its infancy. The spelling that is available on the website is not the final product – it is only the first draft. Therefore, we would be grateful for any help you may be able to offer in improving/correcting/voting on the spelling of words on the website.

    Unlike the IPA, the pronunciation using the SaypU alphabet doesn't have to be accurate. The objective is to try to make the pronunciation as understandable and consistent as possible when pronounced by people who are familiar with the Roman alphabet, but who don't know anything about the foreign language. Naturally, the degree of success would vary across languages and words. However, over a very large number of years, we hope that a standardised way of pronouncing words by foreigners could emerge.

    Finally, the answers to some of the points made above could be found here: http://www.saypu.com/aboutus.php

  56. Gianni said,

    March 6, 2013 @ 1:05 pm

    To speak in favor of the amateurs against academia is a joke, and I understand why I received so many criticism. But I will still stick on inaccuracy of a universal writing. It is inaccuracy that accompanies natural writings, not accuracy. If anyone wants to design a universal script with absolute accuracy, he/she can of course do this in the lab. But writing is in society, which is not a lab. I bet this person's design will end up futile in practice.

    In today's world, the promotion of peace is not simply the prevention of warfare but also the elimination of discrimination and poverty. I have said in previous replies about international aid.

    My suggestion for SaypU is to find their targeted users. Such a script cannot be omniglottic, but can encompass a certain number of world population. How wide? Where should it be used? How would they market their product? This is beyond linguistics. Marketing of a designed script is a fantastic topic.

  57. Mark Mandel said,

    June 8, 2013 @ 8:31 pm

    Oh, God, not another one! Yes, of course, another one. I miss some worthwhile posts on ADS-L (the American Dialect Society listserver) because of one crackpot member whose ceaseless and brainless advocacy of his "true spelling" for English was so annoying that I added the person's name to my spam filter; and since many people still reflexively or by default apply "quote all", any thread he's "contributed" to is blocked.

RSS feed for comments on this post

Leave a Comment