Verner's Law says that voiceless fricatives, when immediately following an unstressed syllable in the same word, underwent voicing. The Germanic prefix 'ge-' as in German 'genug' or English 'enough' is generally derived from PIE *kom-. As this prefix is always word-initial, and there are plenty of examples of word-initial voiceless fricatives that stay as voiceless fricatives, how is this derivation possible?
It seems irregular to me as well.
I found this PDF that mentions this exact matter on page 43, in the abstract for Germanic *ga(-) revisited. Some thoughts on etymology, phonology and inherited word formations by Christiane Gante:
There are quite a lot of papers on [the issue of the etymology and phonology of the ge- prefix]. But it is still somewhat ambiguous whether Germanic *ga(-) can be smoothly linked to the semantically equivalent Latin co(m/n)-/cum, because of the initial consonant's shape.
Here is a downloadable PDF I found of the notes that go with the presentation: Germanic ga revisited. I read through it, and disappointingly, all it says is:
Common opinion is, that indeed PGmc. *ga- equals Lat. cum, co(m/n)-, OIr. co n, com- etc. and thus can be traced back to PIE *kom. But there is still no consensus on why, in this morpheme, PIE *k is represented by PGmc. *g in (almost) all cases.
However, it does state that this has been extensively discussed and has some citations to these previous papers, so I'll look through those next.
I have found several sources that treat this etymology as speculative; for example:
I take the ge- prefix to have been the default SCS prefix in Proto-Germanic. It presents certain problems that the other prefixes do not. It's etymology is unclear
–Lexical Template Morphology: Change of State and the Verbal Prefixes in German, by B. Roger Maylor, from 2002
I also found the following exchange on the talk page of the Wiktionary article for the Proto-Germanic equivalent ga-:
How does the ‘g‘ in ga- derive from *kom / *ḱóm? I thought Verner's law only applied to consonants following an unstressed syllable, this is not the case for ga-. 220.127.116.11 16:40, 5 November 2014 (UTC)
Preverbs seem to have been treated strangely in Pre- and Proto-Germanic, being unstressed and not wholly bound to the verb; for example, clitics could intervene between preverbs and verbs. I'm not really certain about the specifics, but you can see the same development in *bi(-), originally *h₁pi, from PIE *h₁epi. Anglom (talk) 23:34, 17 October 2015 (UTC)
Also from Wiktionary, I found a claim that Gothic at least allowed stacking of pre-verbs in such a way as to place them word-internally (and this seems to be backed up by other sources such as Phrasal Verbs: The English Verb-Particle Construction and its History by Stefan Thim):
As in other Indo-European languages, a verb in Proto-Germanic could have a preverb attached to it, modifying its meaning (cf. e.g. *fra-werþaną "to perish", derived from *werþaną "to become"). In Proto-Germanic, the preverb was still a clitic that could be separated from the verb (as also in Gothic, as shown by the behavior of second-position clitics, e.g. diz-uh-þan-sat "and then he seized", with clitics uh "and" and þan "then" interpolated into dis-sat "he seized") rather than a bound morpheme that is permanently attached to the verb (as in all other Germanic languages). At least in Gothic, preverbs could also be stacked one on top of the other (similar to Sanskrit, different from Latin), e.g. ga-ga-waírþjan "to reconcile".
–the article "Proto-Germanic grammar" on Wiktionary
I'll update this answer once I find more discussion in scholarly journals.
Morphological levelling is the force working against Verner's law in this situation.
Note also that many forms with the ge- prefix may be created after Verner's law was active; there is a modern trend to date the activity of Verner's law even before Grimm's law was active.
This is a really old problem and a lot of research has been written on this.
There are several ways to tackle it:
Reject PIE*kom > PGm *ga. However, as far as I can see, communis opinio is the opposite (e.g. Lehmann 1986), cf. the following suggested cognates: Proto-Italic *kom (e.g. Latin perfectivizing co- and preposition cum), Proto-Celtic *kom, Slavic ** sъ-, and Hittite =kkan (encl. locatival sentence particle) ‘?’ Kloekhorst 2007 writes it was Sturtevant 1927 who first suggested this etymology. The reconstructed PIE form is * ḱom.
Reinterpret the conditioning factors of Verner's Law.
The low tone (in the initial pretonic position): Ivanov 1999 (and earlier). He suggests that "the Indo-European particle/adverbial element *kom could be used both as an enclitic or as a proclitic."
It seems no one is willing to abandon the ban on the Anlaut position though.
A couple of very important notes.
First, gone is the time of the notorious Neogrammarian motto "sound laws have no exceptions." See Hoptman 2002 and Tsimmerling 2007 for more exceptions to Verner's law; Bernhardsson 2001 for a possible explanation.
Second, we don't really know for sure when Verner's law started to function (in Proto-Germanic? Before or after Grimm's law?) and how long it functioned. Also, there have been several attempts to merge Grimm's and Verner's laws into one.
And, finally, the most importing thing to remember is that
"Any "sound law" is NOT simply an algebraic comparison of a protolanguage with a descendant, but a real sound change (or, in cases like GL, possibly a sequence of sound changes) that occurred in a real speaking population in real time" (Donald Ringe, p.c.)
Those inlaut (and possibly Auslaut) fricatives didn't get all voiced in all words overnight nor did stress in Proto-Germanic shift within a week. Verner's Law didn't stop functioning on the winter solstice. These processes took a long time, they overlapped, not all words got affected (due to various reasons) etc.
This is exactly what we can see with a lot of Gothic preverbs (including ga-) that have not been fully grammaticalized, could have been followed by other enclitics etc. See Arkadiev 2015: 209-210, Buscko 2008, Sizova 1978 and Sizova 2007 for further details. This is a natural, living language in its development.