General Pronunciation

The pronunciation of Gothic is, with a few stark exceptions, fairly straightforward. I’ve written extensively about it on my website at http://ling.everywitchway.net/germanic/east/gothic, but I’ll try to paraphrase here.
Consonants:
Stops:
Unvoiced:
p [p]
t [t]
k [k]
q [kw] or [kʷ]
Voiced:
b [b] except when intervocalic. Becomes /f/ when word-final.
d [d] except when intervocalic. Becomes /þ/ when word-final.
g [g] except when intervocalic or word-final.
Continuants:
Unvoiced:
f [f]
þ [θ]
s [s]
g [x] when word-final or before an unvoiced consonant.
h* [h] or [x] or [ç]… This is one of those stark exceptions. I’ll start a separate post about it.
Voiced:
b [β], later [v] only when intervocalic.
d [ð] only when intervocalic.
z [z]
g [γ] only when intervocalic.
Sonorants:
r [r] or [ɾ] – syllabic when it occurs by itself. Unvoiced after /h/.
l [l] – syllabic when it occurs by itself. Unvoiced after /h/.
Nasal:
m [m] – syllabic when it occurs by itself.
n [n] – syllabic when it occurs by itself. Unvoiced after /h/.
gg [ŋg]
gk [ŋk]
gq [ŋkw] or [ŋkʷ]
Vowels:
a [a]
ā [a:]
aí [ɛ]* another exception…more later on the ai/au situation.
ai [ɛ:]* see ai/au post once i post it.
ē [e:] (/e/ is always long.)
i [i] or [ɪ]
ei [i:]
u [u] or [ʊ]
ū [u:]
ō [o:] (/o/ is always long.)
aú [ɔ]* another exception… more later.
au [ɔ:]* another exception… more later.
Diphthongs:
ái [ai]* see ai/au post once i post it.
áu [au]* see ai/au post once i post it.
iu [iu] or [ɪʊ], possibly later [y:] or [ɪy]. * I might post more about this later, but I don't have a lot to say about it or any evidence for my theories.

This is pretty much exactly how I pronounce gothic too. I suppose you will continue on this explanation later, especially about ai and au.

I have a few things I'd like to ask. First of all concerning b,d and g. As intervocalic they hare ß, ð and ɣ as you wrote.
So for example we have the word daga (dative singular of dags) which would be [daɣa]. But what if the previous word ends in a vowel? Would 'himma daga' be ['him:a'ðaɣa] or ['him:a'daɣa]?

Second, b,d and g do not change to f, þ and h when word final or before -s ending if its precedes by e.g. l. Such as kalds. But would the d here be [d] or maybe voiceless like a [t] affected by the voiceless -s?

There has been arguing about f, e and o. I have read about f beeing pronounced bilabial instead of labiodental, more like a frictive p. And e a little bit more closed so it approaches [i:] and similar with o so it approaches [u:].

Thoughts about this?

Hi Friþureiks,

I would guess (and there's not a lot that I can do aside from that) that the change of stop > continuant would not apply over word boundaries, just because that's not a terribly common thing to do in most languages, at least not as a persistent rule. It does happen in some common word combinations in English, and as a rule in Swedish where one word ends with -r and another begins with s- (e.g. du sitter här [dy: sɪt.ter hær] vs. här sitter du [hæɾ ʃɪt.ter dy:]), but I think it's more likely that this did not occur in Gothic. My evidence? Well, mostly my gut feeling, but you do not see devoicing occurring across word boundaries in the environment where a word beginning with a voiced stop follows an unvoiced stop, which would have been more clear from the orthography available at the time. This sort of "mutation" is common in Celtic languages, but I don't know of any similar sound shift in any Germanic languages. That's not to say, of course, that it didn't happen.

As for the stops not devoicing after /l/, I don't know how the stop would be pronounced exactly, but my gut says that it would not be devoiced in words like kalds. Again, if I had to make a guess, it would be that the voiced stop would been perceived to be grouped with the preceding /l/, rather than with following /s/, which would have been seen as more of a semantic ending (i.e. nominative case), and while this is, of course, all occurring on a subconscious level, I think the Gothic mind would have translated and subsequently pronounced it as kald-s rather than kal-ts.

/f/ = [f] vs. [ɸ]. Hmm. Maybe. This did definitely occur in Old Norse in words like aptr (Got. aftra), but once again, there's no particular evidence for it that I know of, unless you count the b > f shift when intervocalic b > ß (probably later [v]). That's one of those differences, though, that is so subtle that it might well have varied by dialect. And, really, [f] and [v] are easier for a native English/German speaker like me to pronounce. If a Goth ever calls me out on it at sword-point, I will happily change I will also note, though, that in Voyles' Early Germanic Grammar, he spends a great deal of time discussing the ft > ɸt shift in Old Norse, but doesn't once mention anything similar in Gothic.

/e/ and /o/ - sure, why not? Again, probably subtle variations by dialect. I tend to pronounce them a little close to [i:] and [o:], just because I'm concentrating more on distinguishing them from /ai/ and /au/ than from /ei/ and /ū/. Vowels are pretty fluid, though; I'd say the standard would probably have been somewhere between [e:] and [i:] (and [o:] and [u:], respectively).

I'll try to get some time to expound on all the things I said I would expound upon soon, although I'll admit I really don't like this interface much. All my nice formatting disappeared when I posted, and it makes things a little less clear than I'd like. In the meantime, I hope this helps! I'd also welcome any thoughts you have on any of the above, since I'm hardly the world's leading expert.

Cheers,

Jamin

Maybe an initial ð in daga would sound weird to a goth. But isn't d and ð allophones and not considered different phonemes?
Another example of this is spanish, or more exactly chilean spanish where b, d and g becomes continuants initially if previous word ends in a vowel.
Maybe I have been a little too influenced by chilean spanish and sometimes I use "extreme" fricative continuant b,d,g intervocally. Daga can from my mouth sound a little like [daɰa].

I also wonder if g/h are pronounced [x] and there is a palatal version [ç] as allophone which I suppose is used around front vowel. Would other velars also be palatal in the same positions as well?
Even if k and g only in a doubtful situation would be [c] and [ɟ] around front vowels I would say that at least fricative g would be [ʝ].

E.g. so giba [so:'ʝißa] (applying the also doubted idea of fricative in word boundaries.)