I see, for only one example, the name of the Burmese (Myanmarese?) leader written as Daw Aung San Suu Kyi. But it is almost always pronounced something like "Ahng san soo chee. Presumably, her name is written in her native language, maybe with an entirely different alphabet or script system. Maybe her name in her native language sounds like, Ahng san soo chee. So, why do we write it Kyi? Wouldn’t it be better to write it in English closer to the way it sounds? I assume that the letters we use are some sort of transliteration of the symbols they use. But if they don’t sound like her name the way we write it, why write it that way? How do these translations or changes come about?
A quick Google shows there are at least four different systems to Romanize the Burmese alphabet, some of which emphasize the orthography (transcription), and some of which emphasize the pronunciation (translation).
Just as in English, the way a word is spelled in Burmese may not match its pronunciation. So you have to make a choice as to whether you try to follow the spelling or the pronunciation.
Shouldn’t that be (transliteration)?
Sorry, I screwed that up. It should be “some emphasize the orthography (transliteration), and some emphasize the pronunciation (transcription).”
I see a lot of these kinds of issues when we’re dealing with sounds/pronunciations which simply have no “match” in English.
Well, I screwed it up in the exact same way, so…
Part of the answer to the OP’s queston may be that some languages didn’t historically have alphabets of their own, so the languages were transcribed by sound value into the alphabet of some other language. To get those into English, the transcribed sounds were transliterated into our Latin alphabet. Each step will introduce errors.
Once a leading source uses those values, others in the field will continue to use them rather then re-inventing the wheel. Eventually, the whole world’s calling a city Pusan, and nevermind that the city’s inhabitants refer to it as Busan.
That makes a good deal of sense and explains how there can be many errors introduced into the process. However, that explanation is built around the notion of trying to reproduce the sounds of the original source, which I would say would be the natural and most reasonable approach. And MIGHT explain how the “chee” sound in Burmese might mistakenly be translated and re-translated into “Khi.” It doesn’t, however, explain why we would continue the error. Once it’s discovered, why wouldn’t it be corrected? Why do newspapers continue to call her Ahng san soo khi in print when her name is sounded differently?
CH sounds are hard. They’re normally transliterated as a KH sound, as in loch. But the sound in English church is really a TSH sound, which almost demands simplification (tshurtsh?)
Why wouldn’t it be corrected? Inertia, mostly. There’s not much money in re-designing the wheel, and it’ll take years to catch on. It’ll take something like government intervention for wholesale change.
But you’re helping to correct it now. Raising awareness and (dare I say) fighting ignorance is the first step.
No, they refer to it as Pusan. That initial sound is an unvoiced (and unaspirated) consonant.
As pointed out, there is value to a transliteration system where the spelling in the source language maps to the spelling in Roman letters. There are disadvantages too, for example that English speakers will (quite reasonably) assume that Busan is pronounced with a /b/ sound because it is written with one.
Ah, thanks. Another example for the OP, then–sounded with a /p/, spelled (now) with a B. A quick search suggests that the reason for the difference is a 2000 change in the South Korean government’s official transliteration guide.
I admit I don’t know about Korean phonology, but this illustrates an important point so well I can’t not use it:
Different languages have different phonologies, different collections of sounds speakers use to produce and distinguish between words. Not the sounds they are capable of producing, which is dictated by physiology, but the sounds their dialect uses to make words and distinguish between different words. That second part is important: A sound distinction which is never used to distinguish between two different words in a given language (that is to say, a sound distinction which is not phonemic) is going to be extremely difficult for speakers of that language to hear and reproduce reliably.
Something which is almost an example in English is the difference between the “th” sound in “thick” (/θ/ in IPA) and the “th” sound in “that” (/ð/ in IPA), which is, as I said, almost never used to distinguish words in English. There are two exceptions I can think of: “thigh” and “thy”, and "thistle and “this’ll”, both of which involve words at least one of which is somewhat uncommon.
For an example of a phonetic distinction English never makes, there’s /p/ versus /pʰ/, or the distinction between the “normal” (to an English speaker) unaspirated “p” sound (/p/ in IPA) and the aspirated “p” sound (/pʰ/ in IPA), which is accompanied by a strong puff of air. In Arabic, “p” is always aspirated, and in Korean, the difference between aspirated and unaspirated consonants is phonemic. English speakers genuinely can’t hear the distinction being made here, but anyone can feel it if they say “pill” (aspirated) and “spill” (unaspirated) with their hand right in front of their mouth.
So. Some languages don’t distinguish between “p” /p/ and “b” /b/, some languages (Japanese, most famously) don’t distinguish between “r” /r/ and “l” /l/, and some weirdo languages don’t distinguish between /p/ and /pʰ/; now, how do you write words in a target language’s writing system when the target language doesn’t distinguish between a key pair of sounds which make those words distinct to native speakers? It is a puzzle, and the more “scientific” transliterations make generous use of punctuation (apostrophes, for example) and diacritical marks to annotate the target text so as to not simply lose information.
Allophones can give any system the dry heaves. Take the [tt] of butter, for instance–most English dialects will use some version of the sound /d/ or a /t/+/d/, but it’s really just the way a doubled /t/ sound is resolved in that particular instance. God knows how it gets transcribed or transliterated into Swahili or whatever.
I’ve sometimes taken to simply throwing in an H after asperated consonants, except in the case of /p/, because I don’t want it to be read as a PH (an /f/ sound). Not long ago, I was transliterating a few passages from Kartvelian languages, and I entirely lost track of what I was doing once I got into the sounds that someone else had previously transcribed as C. I still haven’t figured out whether those are /k/-ish sounds or /s/-ish sounds.
Regardless of what is differentiated by the home language, the transcribed sound should be based on the norms of the target language. If a /b/ and a /p/ are undifferentiated in Korean, the individual words should still be transcribed by how they sound to an English (in this case) speaker.
My Busan/Pusan example was a poor choice. It was just the first thing I thought of and was based on a faulty assumption. I should have researched that better.
Some other examples have been given, but it should be mentioned that Roman letters don’t even correspond to the same sounds in other European languages. In Spanish “g,” “j,” and “ll” have different sounds than in English, while “b” and “v” are barely distinguished if at all. If, for example, a non-Roman alphabet was first Romanized by French scholars the result might be quite different than if it were Romanized by English ones.
Just wait until you’re getting into Icelandic and have to distinguish between the initial consonant in hland (=urine; pronounced with a voiceless* l*) and* land* (=land, with voiced l)!
Whoever developed the commonly used romanization from Burmese used y to soften certain consonants. So the Myanmar currency, spelled kyat, is pronounced chat.
How would you transliterate the Thai word for ‘brake’ which sounds almost exactly like English ‘brake’? The answer seems obvious if you know your target is English-speaking, but many Europeans would assume an Asian word written as ‘brake’ is pronounced like English ‘bra-kay.’
I’d write the word phonetically as /bre:k/, but the Royal Thai General System of Transcription doesn’t allow ‘:’ (let alone symbols like ư, ǣ, ɤ, ɛ, etc.) so the ‘official’ Thai transliteration is brek.
I recall a sad incident based on an American’s inability to render a Thai name in a phonetic form he could remember.
Of course there are examples of where corrections have been made. Beijing and Mumbai come to mind.
Actually,
- even if the language had its own alphabet, and even if the alphabet happened to be the Latin alphabet, there many have been transcription. My grandfather, his mother and his sister all had their lastname spelled differently: it’s a German lastname and the people at the Civil Registry never bothered to ask “how do you spell this?”, they just wrote it as it sounded to them.
- those transcribed sounds might already be “in our Latin alphabet”, but transcribed by Italians, Spaniards, Portuguese… and then suffer additional deformation when being read by someone unfamiliar with the original transcription method. This isn’t a peculiarity of English or anything like that: the same problem exists when a transcription to English is read by someone unfamiliar with it (cif. “Sri Lanka” being pronounced “Esrri Lanka”, with an additional E and a rolled R and with both As equally clear and open, by Spanish TV newscasters).
Good points.
I’m not sure what you mean by “doubled /t/.” Either way it’s only one phoneme; whatever dialect, it’s one sound–the British /t/, for example, or the North American alveolar tap /ɾ/, which is a single sound, rather than two sounds (/t/ + /d/). The fact that it’s written with two letters doesn’t change this. Or is there some dialect where it’s articulated as two that I’m not aware of?