Many spellings in English look strange but are explained away by saying that the pronunciation of the words have changed but the spellings have just stayed the same. An example is the the word “knight” which at one point was pronounced exactly how it is spelled but is now pronounced “nite”. I have heard that Spanish is remarkably phonetically consistent, but what about the other extreme? Irish had a spelling reform, but Scottish Gaelic didn’t, so I imagine it is a bit divergent. Written Tamil is quite different from spoken Tamil, I’ve heard.
It might be comparing apples to oranges, but Japanese Kanji might be pretty divergent, with Chinese and multiple Japanese readings for one character.
In short, what spoken language is most divergent from the written Standard form?
Written kanji has no connection whatever to the sound. For example, the same character meaning mountain can e pronounced yama or san depending, they tell me, on which is more euphonic. Mt. Fuji, for example, is Fujisan, not as many westerners think, Fujiyama. (The san has nothing to do with the honorific; it is actually the word for mountain in some Chinese language.) Of the languages I am actually somewhat familiar with, I can say that German is almost completely phonetic, as long as you understand the phonology. But initial consonants are typically voiced, no matter how they are spelled (my wife kept hearing Dochter, when the speaker intended Tochter) and finally consonants are normally unvoiced (so Hand is pronounced as though it were spelled Hant, although the plural, Haende, is pronounced as it is spelled and the final e is not silent).
On the other hand, French is nearly as erratic as English. Final letters are, for the most part, not pronounced, although you cannot count on it. One of the two words spelled “fils” is pronounced fees (unvoiced s) and the other roughly is feel. The set phrase “Qu’est-ce c’est que ca” (which can be translated as “Whazzit” consists of seven words or fragments thereof pronounced in five syllables. And the last fragment is also a contraction for cela. There are six words with five spellings all pronounced as ver: ver (worm), vers (verse), vers (towards), verre (glass), vert (green), and vaire (a kind of fur, weasel I think). (And Cinderella’s glass slipper was a fur slipper in the original.)
I’m guessing that the confusion arose not because the initial consonant was voiced, but because it was unaspirated. That happens all the time with English speakers hearing Indian words beginning with unaspirated unvoiced consonants (k, ch, t, p) as voiced consonants (g, j, d, b).
Even French, though is pretty regular with its word-terminal letters–they may not be pronounced in the word, but through elision, if the next word begins with a vowel, it is pronounced. So the normally silent letters still have a very useful purpose, unlike many English words that are just there for “historical” reasons. Although the “passe simple” is a good example of a divergence between spoken and written forms.
Bokmal from what I understand is quite divergent from Norwegian.
In any orthography, there is a tension between two mutually opposed desiderata:
Sound faithfulness: one written form <–> one pronunciation
Meaning faithfulness: one written form <–> one meaning (“morpheme”, technically)
(Disclaimer: while I think most linguists would agree with me, the terms “sound faithfulness” and “meaning faithfulness” as used in this context are made up by me.)
In English, for example, “meaning faithfulness” wins out for the regular plural suffix, which is spelled “s” even though sometimes it is pronounced “z” (e.g., in “dogs”), whereas “sound faithfulness” wins out for the indefinite article, which is sometimes spelt “a” and sometimes “an”.
Japanese Kanji and Chinese characters are both heavily tipped in favor of “meaning faithfulness”, with (at least in Chinese) many characters pronounced exactly the same way but differing in meaning, and (moreso in Japanese) the same character pronounced several different ways but unified under one meaning.
I say all this just to emphasize that although Japanese and Chinese orthography may look “divergent” along one dimension (“sound faithfulness”), the tradeoff is incredible “convergence” along another dimension (“meaning faithfulness”). Spelling is a multidimensional affair.
Were you in Franconia (e.g. Nuremberg) or in (certain parts of) eastern Germany? Most dialects, including the “standard” dialect don’t do that.
That’s a good example that gives an idea of German orthography. It’s not the simple one-to-one mapping between letters and sounds that you would want in a proper phonetic alphabet. There are various context-dependent rules that you have to take into account. Fortunately they are very predictable. For example, a more precise version of this rule would be that all obstruents at the end of syllables are voiceless – even if the spelling appears to suggest something else because it is kept consistent with other forms of the word. When speaking or reading native speakers apply those rules automatically. This goes so far that it is hard to supress when you are speaking a different language. That’s why the above rule is such a vital ingredient of any nice fake German accent. (“Ve haff…”) On the other hand it is a bit harder to explain to a German child why “Hund” (“dog”) and “bunt” (“colorful”) rhyme perfectly but they are spelled in different ways and that you can tell from the inflected forms which way it is.
For the record: I’m a native French speaker who went to a French high school. In one of my second-language English classes, our teacher proclaimed: “English spelling is a bitch.” He was actually laughed at, and not for the use of a certain word.
Is it “apparement,” or “apparament,” or “apparamment” or “apparemment”?
“Impressionisme” or “Impressionnisme”?
“Colone”? “Collone”? “Colonne”?
And, this is without getting into the very, very ugly topic of participe passé, and other grammatical complexities.
In English, children participate in spelling bees. In French, adults participate in dictées. What makes French spelling so difficult that adults who train for the event can still all fail to write down a dictation without fault is the very complicated grammar rules, as much as the arbitrary orthography of the words.
It’s not that the “passé simple” isn’t pronounced the way it is written (apart from the fact that the final letters aren’t pronounced, already mentioned as common in French), it’s that the passé simple is almost never pronounced at all because it’s essentially only used in written form.
For example, smo fi’n gwbod (spoken, “I don’t know”) vs. dydw i ddim yn gwybod (more formal, “I don’t know”) vs. nid wyf yn gwybod (literary, “I don’t know”). And endless variations. But all pronounced exactly as written, given the orthographic conventions of the language.
I’d rule out kanji from this thread altogether, since the question was about graphic representation of phonetic sounds, and kanji is logographic, not phonetic. It represents meanings, not sounds. That’s how a given kanji can represent two completely different words, either the kun-yomi which is the native Japanese word, or the or on-yomi which is a Chinese word adapted into Japanese. Phonetics has nothing to do with it in the first place.
For the most phonetically divergent orthography, Tibetan wins hands down. To the uninitiated, the pronunciation of many words has almost no discernible relation to the orthography. Written Tibetan includes large consonant clusters which in pronunciation are resolved into a single phoneme. For example, bsTangyur is pronounced “tanjur” and sprulsku is pronounced “tulku.” The Tibetan name of Tibet is spelled Bod, but prounounced “P[sup]h[/sup]ö’.” The relationship between Tibetan orthography and phonology is not quite as gnarly as it appears at first, since the unpronounced consonants in initial clusters are now indications of the syllabic tone. (A similar development has occurred in Panjabi.)
That’s not entirely true and you have to learn where. For example, nouns do not usually make a liaison with following adjectives that begin with a vowel: "chambres obscures is pronounced with no liaison. But preceding articles and adjectives do usually make liaison with following nouns: “fines herbes” is pronounced as feen sairb. But “fines homards” would be pronounced as feen omard. Then in songs and poetry, many of the silent letters do get pronounced.
I may be throwing stones at my own roof, but in Spanish the ceceo means that many speakers have problems distinguishing between “c” and “s” when trying to write a word they’re unfamiliar with (they still have no problem pronouncing it properly if they haven’t seen it before, though).
Other people have problems with, for example, insisting in treating every written “c” as a “z” sound including cases where it should be “k”, but these are considered mispronunciations; the seseo is very common (more common than the version where we differentiate the “s” and “z” sounds).
Yeah, Kanji is a special case, and not what the OP seems to be asking about. And Japanese does have two syllabaries that are quite phonetic, even if the one used mainly for writing foreign words (katakana) represents those words as they would be pronounced by a Japanese speaker.
I wouldn’t worry about your rocks; I think the Spanish roof is pretty sturdy on this issue. As a native English speaker and beginning-tourist-level Spanish speaker, I have a better chance at correctly pronouncing an unfamiliar written Spanish word than I do with an unfamiliar written English word.
Spanish is renowned worldwide for its excellent match between spelling and pronunciation. It’s one of the most transparent orthographies in the world. (Italian is no slouch in that department either)
As for Tamil, it isn’t that far off. Spoken Tamil diverges from written not because of opaque orthography, but because of diglossia. That means the language is split into two versions, a “high” Tamil used in writing and formal speaking, and a “low” version used in everyday conversation. Arabic is the same way: there is one Standard Arabic language used all over the Arab world, but in conversation everyone speaks a local dialect.
The “high” registers of Tamil and Arabic are like Latin, while the spoken languages are like the Romance languages of French, Italian, Spanish, etc. It would be like if everyone in those countries used Latin for things like writing, making speeches, TV new broadcasts, but spoke the local Romance language in everyday life. Modern Greek used to have diglossia, but the Socialist government dumped the high register in the 1980s and now the spoken language is standard for formal writing.
Tamil writing allows several allophones for each phoneme, which means a number of different pronunciations for each letter. But the orthography is still fairly transparent, because the pronunciation of each letter is predictable from its position in a word. For example, the letter த <t> can be pronounced [t], [d], or [ð] depending on whether it occurs next to an unvoiced consonant, after /n/, or after a vowel, respectively. Tamil gets by with fewer letters that way for a large number of sounds.
