Peust, Carsten: Egyptian phonology: an introduction to the phonology of a dead language (Göttingen, 1999)

Our results differ from Roquet's (1973b) in several points. Different decisions were
sometimes made in drawing the borderline between a significant and an accidental
underrepresentation concerning a specific pair of consonants. This problem cannot be
simply overcome by mathematically determining the statistical level of significance,
since the major problem is posed by the uncertainty of many of the input data (deciding
on how to transcribe a word, or where to assume morphological boundaries, etc.). Note
also that we examined token frequency whereas Roquet probably considered type
frequency. Last but not least, our corpus was different from his. The most salient
differences between Roquet's and our results are the following:

• Roquet does not list any incompatibilities involving w since he left this consonant
out of consideration (cf. Roquet 1973: 112).

• We allow for the combination of ' + s which Roquet (1973b: 108, 110) considers
totally incompatible.2^1

• Numerous of Roquet's "relative incompatibilities" are not covered by what we term
"strong restriction".

• Some of our "strong restrictions" do not even appear among Roquet's "relative
incompatibilities" (b + h, b +1, h + d, h + t, h + h, h + z, h + t).

4.11.4 Roots with complete or incomplete reduplication

Since two consonants that share the same place of articulation do not normally co-occur
in an Egyptian root, we should expect that two identical consonants do not co-occur
either. However this is not true. One of the reasons for this is the fact that a root can be
reduplicated. There are many reduplicated verbs in Egyptian such as b'b' "to bubble",
ptpt "to trample", nmnm "to move", gmgm "to smash", snsn "to associate".
There are also several Egyptian triconsonantal roots in which the first and the third
consonant are identical. A possible explanation would be that these are incomplete
reduplications (cf. Greenberg 1950: 181). Examples:

• "' "to ejaculate" (note that <S = /d/, <i> = /r/)
V "to stand"

• boh "to be speedy"

• grg "to establish"

It can be remarked that the second consonant in words of the type ABA is preferably a
sonorant. These words were perhaps originally reduplications of the type ABAB, and

231 There is a common root s'h "dignity" which is well attested from the Old Kingdom
on. Other words containing s and ' are sm' "pole" (from the Old Kingdom on), s'b "to
castrate" (from the Middle Kingdom on), sm' "vegetable fibre (Meeks 1977/79: II,
326; from the Middle Kingdom on) and sh'.t "rabbit" (from the Middle Kingdom on);
for the latter three, a reading with z would likewise be possible, but the combination
+ z is excluded according to our as well as to Roquet's minds.

197