Gothic for Goths: Some More Thoughts on Pronunciation

I got most of it out of my system in that last post, but let me just touch on one more pronunciation issue while i'm at it. Fortunately, this one really isn't quite as controversial as the last; this is really just a clarification, though i promise there is a bit of contention to overcome for those of you just tuning in for the drama.

While everyone agrees that the digraphs ai and au are the most contentious to pronounce in gothic, one of the most confusing letters by far is g.  (Well, giba, really, but i'm not going to confuse the issue by using actual gothic letters in my blog.)  Fortunately, there are a pretty solid and undisputed set of rules governing its pronunciation in various circumstances.  I'll try to enumerate them all here for your reference and reflection, as well as the one contentious bit where i once again have to disagree with my hero Mr. Joseph B. Voyles, whose thoughts about g aren't so much wrong as his thoughts about h.  But that's another story.

• g = [g]
• when word-initial.
• gaggan [gaŋgan] (to go)
• giba [giβa] (gift)
• when adjacent to a voiced obstruent or sonorant (other than g).
• baurgja [bɔrgja] (citizen)
• brigdil [brɪgdɪl] (bridle)
• g = [γ]
• when intervocalic.
• agan [aγan] (to fear)
• igil [ɪγɪl] (hedgehog)
• (maybe?) before a syllabic sonorant.
• hagl [haγl̩] (hail)
• baugms [bauγm̩̩s] (tree)
• g = [x]
• when word-final.
• dag [dax] (day, acc.)
• galiug [galyx or galɪʊx] (lie)
• before an unvoiced obstruent.
• dags [daxs] (day, nom.)
• dulgs [dʊlxs] (debt)
• g = [ŋ]
• before a velar consonant (i.e. g, k, or q).
• gaggan [gaŋgan] (to go)
• drigkan [drɪŋkan] (to drink)
• igqis [ɪŋkwis] (y'all)

I would also throw a clarification in there that "word-initial," in my view anyway, includes after prefixes, so "gagaggan" would be [gagaŋgan], not [gaγaŋgan].

Voyles postulates that initial g may in fact have been continuant ([γ]), and while i find this unlikely, i don't argue that it is a possibility.

The real controversy i promised is Voyles' (not unreasonable) speculation that the appearance of g instead of h for [x] is evidence that all instances of h were pronounced [h] and never [x].  This contrasts with the other stops b (which becomes f in similar circumstances) and d (which becomes þ).  Once again, i attribute this to orthographical conventions rather than phonetics, and hold firm to my belief that h = [h] when word- or word-segment-initial (except perhaps before a sonorant, i.e. hl-, hn-, or hr-), and [x] elsewhere.  Or perhaps we're both sort of correct, and where i have posited that h = [x], it may in fact be something more like [ç].