Phonetic languages?

[linguoboy]

Gah, I can never keep the difference between those two straight in my head.

Part of the problem is that "phonetic" has made its way into popular use whereas "phonemic" never has, so the distinction simply isn't there for most people.

Here's a quick illustration: Take the German word Tag "day". The actual normative pronunciation of this word is ['tʰa:kʰ]. So a truly phonetic spelling of it would be "Thaakh". But there's no need to go into such detail since orthographies are designed to serve fluent speakers and fluent speakers know the phonetic rules of their own languages. German fortis stops (i.e. /p/, /t/, /k/) are always aspirated, so there's no reason to indicate this in the spelling. Whether you write t or th, it will always come out [tʰ]. Aa is one way to represent [a:] in German, but it's seldom used; a rule that stressed vowels are long before single orthographic consonants works just as well most of the time.

And then we come to the final sound. It's spelled g not just because historically it was [g] but because it still is [g] when a vocalic ending is added, e.g. Tage ['tʰa:gə] "days". Standard German doesn't allow [g] in final position, so again it doesn't matter if you write g or k (or gk or kh ); if you've internalised the phonological rules of German it will naturally come out [kʰ]. Of course, this doesn't prevent other languages with the same kind of regressive voicing assimilation (e.g. Turkish) from indicating the change in their spelling, so ultimately it comes down to a choice between phonetic detail and paradigmatic consistency.

Since there are no languages with perfectly phonetic orthographies and precious few (if any) with perfectly phonemic ones either, these aren't the most useful terms available. Most specialists prefer to talk about "orthographic depth" and refer to the relative "deepness" or "shallowness" of a language's orthography. A language like Finnish has a relatively "shallow" orthography since the correspondence between phonemes and letters is very nearly one-to-one whereas English's is "deeper" because the number and complexity of correspondences is much greater. Greater still is the depth in most Southeast Asian scripts, and logographic scripts like Chinese or Ancient Egyptian are the "deepest" of all.

[gyrus]

I thought Czech was supposed to have one of the most phonemic orthographies?

[Jurgen Wullenwever]

I know that Spanish is, Esperanto and Polish too, and that English, French, and German definitely aren't. What other languages are spelled the way they are written?

Aren't all languages spelled the way they are written?

I thought someone must have commented on this little slight, but I didn't find it.

As it seems, no language has a clear, fully phonetic spelling, but many languages are fairly close, even English.

Since there are no languages with perfectly phonetic orthographies and precious few (if any) with perfectly phonemic ones either, these aren't the most useful terms available. Most specialists prefer to talk about "orthographic depth" and refer to the relative "deepness" or "shallowness" of a language's orthography. A language like Finnish has a relatively "shallow" orthography since the correspondence between phonemes and letters is very nearly one-to-one whereas English's is "deeper" because the number and complexity of correspondences is much greater. Greater still is the depth in most Southeast Asian scripts, and logographic scripts like Chinese or Ancient Egyptian are the "deepest" of all.

I wonder if one could say that Chinese writing is unphonetic. I know less than nothing of Chinese, but my impression from what I have read is that the wordsigns are composed with twelve types of lines, and if one regards each of these lines as something corresponding to an alphabetic letter, then the spelling should be rather arbitrary in relation to the phonetic value. This might be a silly idea, but is it a workable point of view, to look at Chinese writing as an alphabetic script with twelve unphonetic letters?

[Mutusen]

I thought Czech was supposed to have one of the most phonemic orthographies?

I'm not an expert in Czech, but (as far as I know) while it is easy to know how to read a word, it is more difficult to know how to spell an unknown word: i/y and í/ý are pronounced the same way (though i can change the sound of the preceding consonant), so are ú and ů (but there are rules and these letters cannot appear anywhere), final consonants are devoiced (so d and t are both prounounced [t] at the end of a word), and mě and mně (different cases of the pronoun I) have the same sound.

[Ludwig Whitby]

I thought Czech was supposed to have one of the most phonemic orthographies?

I'm not an expert in Czech, but (as far as I know) while it is easy to know how to read a word, it is more difficult to know how to spell an unknown word: i/y and í/ý are pronounced the same way (though i can change the sound of the preceding consonant), so are ú and ů (but there are rules and these letters cannot appear anywhere), final consonants are devoiced (so d and t are both prounounced [t] at the end of a word), and mě and mně (different cases of the pronoun I) have the same sound.

Guys! Serbo-Croatian!?

every phoneme is represented by one and only one letter! Every letter is always pronounced the same way.
The only thing is that we don't mark vowels for length, and we don't mark stress.
No final devoicing, no nothing.
There are only a few exceptions, and that is only when the sound change rules are not applied because it would make the word 'unrecognizable' (a load of crap if you ask me)
pod + tip = podtip (if the rules were applied it would be potip, and it's pronounced like that)
post + diplomski = postdiplomski (pronounced pozdiplomski)

[Jurgen Wullenwever]

Guys! Serbo-Croatian!?

every phoneme is represented by one and only one letter! Every letter is always pronounced the same way.
The only thing is that we don't mark vowels for length, and we don't mark stress.
No final devoicing, no nothing.
There are only a few exceptions, and that is only when the sound change rules are not applied because it would make the word 'unrecognizable' (a load of crap if you ask me)
pod + tip = podtip (if the rules were applied it would be potip, and it's pronounced like that)
post + diplomski = postdiplomski (pronounced pozdiplomski)

Are there no dialects with different pronunciations?

[Ludwig Whitby]

Guys! Serbo-Croatian!?

every phoneme is represented by one and only one letter! Every letter is always pronounced the same way.
The only thing is that we don't mark vowels for length, and we don't mark stress.
No final devoicing, no nothing.
There are only a few exceptions, and that is only when the sound change rules are not applied because it would make the word 'unrecognizable' (a load of crap if you ask me)
pod + tip = podtip (if the rules were applied it would be potip, and it's pronounced like that)
post + diplomski = postdiplomski (pronounced pozdiplomski)

Are there no dialects with different pronunciations?

No.
Only the stress is different.

EDIT: wait, There is. The Torlakian dialect in South Serbia (around the city Pirot) pronounces two letters differently
ć (/tɕ/) is pronounced like /c/
and đ (/dʑ/) like /ɟ/
Though I don't see why that matters.

[linguopolitano]

Guys! Serbo-Croatian!?

every phoneme is represented by one and only one letter! Every letter is always pronounced the same way.
The only thing is that we don't mark vowels for length, and we don't mark stress.
No final devoicing, no nothing.
There are only a few exceptions, and that is only when the sound change rules are not applied because it would make the word 'unrecognizable' (a load of crap if you ask me)
pod + tip = podtip (if the rules were applied it would be potip, and it's pronounced like that)
post + diplomski = postdiplomski (pronounced pozdiplomski)

Are there no dialects with different pronunciations?

No.
Only the stress is different.

EDIT: wait, There is. The Torlakian dialect in South Serbia (around the city Pirot) pronounces two letters differently
ć (/tɕ/) is pronounced like /c/
and đ (/dʑ/) like /ɟ/
Though I don't see why that matters.

Exactly!
And there are also some different vowels in some dialects, like /ε/ and some nasal (or nasalized) vowels, which are not present in standard language, but, you are right, that has nothing to do with our alphabet which is phonemic.

[gyrus]

BUT, it doesn't show the pitch accent.

[csjc]

Hebrew seems to be entirely phonemic if vowels and accents are included.

[kibo]

BUT, it doesn't show the pitch accent.

Yes, but in many cases, if not most, the pitch accent and vowel length are not a distinctive feature. I wouldn't describe "phonemic" as an absolute category, it's more like a tendency. And Serbian (Croatian, Bosnian, etc.) orthography is certainly more phonemic than non-phonemic. Rumpetroll has already names some of the instances where there are exceptions to the phonemic principle, where the orthography follows a morphophonemic rule. But that's a huge minority of cases.

[linguoboy]

I wonder if one could say that Chinese writing is unphonetic. I know less than nothing of Chinese, but my impression from what I have read is that the wordsigns are composed with twelve types of lines, and if one regards each of these lines as something corresponding to an alphabetic letter, then the spelling should be rather arbitrary in relation to the phonetic value. This might be a silly idea, but is it a workable point of view, to look at Chinese writing as an alphabetic script with twelve unphonetic letters?

But why should one regard the brushstrokes as corresponding to alphabetic letters? After all, the Roman alphabet is also composed of a limited number of strokes, but these have no relevance outside of calligraphy.

Wikipedia has a reasonably thorough introduction to the Chinese script which delves into the "phonetic" aspects of the script. These consist of rough clues rather than anything remotely resembling a one-to-one correspondence, but one can't say that they are absent entirely.

[TeneReef]

Vietnamese is very phonetic, and it looks beautiful:

[Levike]

Thaaat's not beautiful, more like frightening.

Romanian: Far from being phonetic,
you have to study a lot of grammar to guess how some words should be pronounced.
A very grammar oriented spelling.

Hungarian: Kind of phonetic, maybe just a bit more than Spanish.
But you have to put funny accents all over the place
in order to represent how it's pronounced, so that makes it complicated.
A very few times we write dt and pronounce tt.
And there are two ways to write the sound /j/, namely j and ly.
There are probably others, but none comes to my mind now.

[linguoboy]

Vietnamese is very phonetic

That's debatable. In Hanoi pronunciation (the basis for the current standard), /z/ is represented by both d and r, /s/ by both s and x, c by both tr and ch, etc.

[Thon]

Any language written in IPA can be considered phonetic! (just kidding)

Seriously, I thought Hungarian and Turkish were completely phonetic.

[linguoboy]

Seriously, I thought Hungarian and Turkish were completely phonetic.

Nope; Standard Hungarian merges a and o as [ɒ] and j and ly as [j], exhibits regressive voice assimilation (e.g. [fudbɒlː] for futball), simplifies consonant clusters (e.g. [vɒrtɒm] for varrtam), and so on and so forth. I would put it on a par with German in terms of orthographic depth.

Turkish is nearer the mark, but even it has irregularities. Many of these relate to ğ, which can represent [ʋ], [j], [ː] (i.e. lengthening of the vowel), or nothing at all depending on its position within a word.

[Levike]

Nope; Standard Hungarian merges a and o as [ɒ]

Ahh, no, never. a is /ɒ/ and o is /o/. It takes a lot of alcohol for those to merge.

j and ly as [j]

Yes, but it's archaic so let's keep it.

simplifies consonant clusters (e.g. [vɒrtɒm] for varrtam).

Thanks God for that.

I would put it on a par with German in terms of orthographic depth.

Lemme go full retard, Wut? No way! Nem, nem, nem!

[md0]

Turkish is nearer the mark, but even it has irregularities. Many of these relate to ğ, which can represent [ʋ], [j], [ː] (i.e. lengthening of the vowel), or nothing at all depending on its position within a word.

Also, the fronted vowels aren't as clearly marked as I hoped. I don't remember any examples right now, but there's a number of words which I'd expect to be spelt with â, but they are just a.

It was the first wave of disappointment for my fascination over the Turkish alphabet.

[OldBoring]

I used to think that Archaic Latin should be phonetic, since the Latin alphabet was designed for that language. But... they had <C> and <K> for the same sound. And <C> was also used for the sound /g/.