Because gaul is already taken by 狗. These are two often-used words. I
have not looked into historical phonology, but I think the literary
reading giul, and the Mandarin form jiu3, offer some clue to the
presence of a historical medial. Because 狗 is colloquial/mundane,
that's why I keep the simpler spelling for it. But one could try to go
the other way, too...
The same situation happens gail 改 and giail 解. I used to write 解 as
gyail, but now I prefer giail. In this specific case, I have one point
of support from Vietnamese 解 giải.
The presence of medial -y- or -i- makes connection with Mandarin Hanyu
Pinyin. That's the idea. It's not necessary to bust all the homophone
problems, but for often-used terms, like the two cases above, I would
think something must be done.
Personally I don't like the gy combination too much for gyaul. I have
tried gjaul, too. Any other solution? Perhaps geaul? I don't know.
It's something I still have not figured out. Gy- does not seem too
bad, as many people are familiar with gynecology, gymnastics, and the
Japanese word gyoza 餃子.
Anyway, it's best to try to find related cases to support a more
general rule... 狗 九 so far has been a tough problem to crack. I guess
if there are no further data points, leaving them both as gaul may
also be OK...