Aye. Stops cannot really be sustained. That is quite clear. I don't think I said anything different. But silence can be sustained. These stops are two-parters. First you close the air flow, then you release it. Normally you do the second part practically instantly after the first, but if you like, you can remain silent for a while - very short while, if you're speaking fast. As this pause is in the middle of a 'phoneme', I don't think it's that silly to speak of a long consonant. From start to finish it can easily take as long as any other sustained phoneme. ...Might be a little bit figurative, s'true.
I agree. In linguistics, you sometimes talk about words having "timing slots" - a sound usually takes up one slot, and a long sound takes up two. Languages with doubled stops do exactly what Qvaak described - the airflow is stopped for longer.
(In general we don't think of English as having this kind of doubling, but there's the occasional exception - compare the "k" in "booking" to the one in "bookkeeper". We don't pronounce two separate "k" sounds in "bookkeeper" - but most speakers hold the closure in "k" for noticeably longer.
Now on the other hand, even though the release part is usually the part with the clearest sound, the closing the air flow part does sound quite different in different stop consonants and may well serve as a phoneme at it's own right. So yeah, to speak of doubled consonant is pretty accurate too. If you pause long enough to differentiate the sounds, you get two consonants at the prize of one stop of the air flow. I think in word combinations like at times and top priority even english speakers often do this two for one trick - I do for sure.
I have found some helpful pictures to illustrate the airflow and silence Qvaak describes, from Jeremy O'Brien's paper "Perception and English t-Glottalization":
http://people.ucsc.edu/~jpobrien/papers/obrien_tglot.pdf.
These show a speaker pronouncing the nonsense words "apa", "ata", and "aka", the three English voiceless stops. The blank spot in the middle is the silence that results from closing off the airflow. (You can spot a few little lines at the left-hand edge of the silence - those are pulses from the vocal folds, which carry over the voicing from the previous vowel slightly into the closure. If these were voiced stops - "aba", "ada", "aga" - the voicing would continue through the closure, as you can kinda see here:
http://www.springerimages.com/Images/Biomedicine/1-10.1007_s00221-005-0008-z-1)
The sound waves for the silent part are indistiguishable for all three stops. The reason we can tell them apart is because of the sounds at the beginning and end, when the airflow is being closed and opened.
In the picture, you'll notice some dark bands in the vowel sounds on either side of the silence. The dark areas show the frequencies where there's the most acoustic energy. You'll notice that the patterns of those dark bands are different for different sounds. Those patterns are what our brain uses to convert a constantly changing stream of sound into what we perceive as individual sounds.
I don't have any spectrograms of double consonants handy, and I am at home with a terrible cold so I'm not going to make my own spectrograms. But if we looked at a language with doubled consonants, we'd see that the period of silence is extended for longer, and the dark bands at the beginning and end sound the same. To my ear, this is what Dothraki is doing.
I wouldn't characterize the closure as "a phoneme in its own right", though. Let me explain what I mean.
In English, we sometimes consider sounds the same even though we pronounce them differently. Take the "p" sound in the words "pit" and "spit". English speakers consider it the same sound, and usually can't hear any difference without training. But it's easy to show they're pronounced differently - for example, if you hold a piece of paper in front of your mouth, you'll notice that it moves a lot more on "pit" than "spit".
That's because the "p" in "pit" is aspirated - it's pronounced with a burst of air. The two p's are produced differently - a linguist would say they're different
phones, but for English speakers, they're forms of the same
phoneme.
A phoneme is a mental category of sounds, so you can't test whether something is a separate phoneme by looking at whether you can find differences in the sound at its beginning and its end. The real question is whether speakers consider them different.
As for the IPA notation:
/hrak.kar/ is the simplest possible solution and very defendable. But 1) people not used to doubled consonants will voice two fully executed stops and sound very wrong; 2) If Dothraki also uses fully doubled stops (for example with some prefixes), we'd need to find some different way to notate them.
/hrak:ar/ doesn't seem too popular choice. I saw it used somewhere outside in the internet (with finnish, not with Dothraki of course), but there too it was presented as misleading. Syllabes change in the middle and the notation argues against that.
/hraʔ.kar/ seems like an odd solution. Isn't the ʔ-symbol used unconventionally? As far as I read it marks a stop deep in the throat, a voiceless glottal plosive. You might manage to emulate the way the word sounds quite closely, but if you followed the instruction, you'd do it very wrong and rather awkwardly.
I think it's fine for the Dothraki community to choose whatever's useful here - there are often several possible IPA transcriptions for a word, depending on how much detail you're including. (In the English word "hamster", for example, many people produce a "p" sound in the middle - "hampster". A broad transcription would be /hæm.str̩/, which notes the individual phonemes but not the effect of the rule that inserts the "p". A narrower transcription would be [hæ̃mp.str̩], which shows the insertion of the [p] and the nasalization of the /æ/.) (BTW, the other window with my IPA shows that as a syllabic r, but the little syllabic marker is disappearing when I paste the transcription into the Post Reply form.)
If I were writing a paper on Dothraki, I'd probably go with /hrak.kar/. If it later turned out that some stops had double releases, I'd use the "unreleased" diacritic from the IPA for the double stops that are simply held longer. (I've tried to cut and paste the symbol into this form, but it's not showing up - so it's probably not all that useful for the forum.)
As for the last - Qvaak, you're right, the ʔ is a glottal stop. It's made by closing the vocal folds together and cutting off airflow. (English has it in the word "uh-oh", and languages like Hawaiian and Arabic make much greater use of it - for them, it's just another consonant.)
With voiceless stops (/p/, /t/, /k/, and /q/) it might not sound all that strange if people try to say doubled stops with a glottal stop. To produce a voiceless stop, you have to do two things - close the airflow, and stop the vibration of the vocal folds. Many English speakers do this by closing the vocal folds together, producing what are called "glottalized" stops. With voiced stops, though, the glottal stop will also stop the continuous voicing that you normally get throughout the stop. So that will sound weird.
