Cf. fūh- (constr. Nom. fū-, Acc. fā, Gen. fī) [ibid. 650]
Note a meaning shift
Note the base ʔafu- with suffixes [ibid.]
Marked by unusual developments, still lacking explanation, such as vestiges of 'mimation' in ARM *pūm- (<*pay/w-u-m, with the Nominative case -u ending and the suffixed -m article) and ARB, where vestiges of PSEM three case nominal paradigm with the postpositional -m article ('mimation') are likely preserved, due to some unknown factor, as variant nominal forms.
The related ARB term for 'mouth', fūh- (constr. Nom. fū-, Acc. fā, Gen. fī) [ibid. 650] may be analysed as *fay- with the suffixed possesive 3rd pers. masc. pronoun -hu ('his mouth') again, for some undetermined reason, incorporated into the stem; another possible explanation is that h sometimes serves, both in ARB and other SEM, as a "triconsonantizer", cf. ARB miyāh-, šifāh- (broken plurals from māʔ- 'water' and šaf-at- 'lip'), SEM *ʔilāh- 'God' evidently related to *ʔil- etc.
On morphological problems posed by this root in ARB, cf. [Wright 1 239, 249, 252].
ETH is thought by some authors to be borrowed from CUSH (see [LGz 9] for references), but is rather to be explained from SEM *paw/y- with the ʔa- prefix: cf. the base ʔafu- in GEZ and TGR and the plural forms ʔafaw in GEZ and ʔäfäyǝt in TGR. For SEM origin of the ETH examples also speak vestiges of the case system in GEZ and TGR similar to those in ARB and HBR. Another possible explanation of the Anlaut in ETH *ʔaf(w)- is a contamination with ETH *ʔanf- 'nose', cf. ãf 'mouth' < *ʔanf in part of GUR. As to CUSH forms for 'mouth' [Dolgopolsky 230-31], they are likely related to the present SEM root.
Cf. SEM forms which originally might mean 'mouth, opening': AKK aptu 'window, window opening; opening of the ear' OB on [CAD a2 197], [AHw 61] (that -t is a feminine marker and not part of the root, is seen from the plural apāti), apu 'hole, opening in the ground' NA [CAD a2 201] (neither aptu nor apu is likely to be a borrowing from SUM ab 'window, opening' as is supposed with a question mark in both [CAD] and [AHw]); ARM MND aputa 'opening, mouth (of vessel)' [DM 30] (less likely from PTA < *ptḥ 'to open', as suggested [ibid.]).
Cf. forms with -m as a third radical either related to, or contaminated with, the present root: ARB fʔm 'se remplir la bouche d'herbes (se dit d'un chameau, etc.)' [BK 2 530]; GEZ ʔafʔama 'to put a morsel of food in another's mouth', fāʔm 'morsel', AMH fämma (<*fmʔ <*fʔm ?), GAF (tä)famä 'to take a mouthful' [LGz 154].
Note a rather unusual semantic shift which must have taken place as early as in PSEM, namely 'mouth' > 'edge (of a sword)', in AKK, HBR, SYR and ETH. Cf. the same meaning shift in Copt. tapro (Sah.), tapra (Fay) <tap-r/lV 'tranchant d'épée'
[Fron 44] (*p- 'bocca'); [Holma 21]: AKK, HBR, SYR, GEZ, ARB; [KB 914]: HBR, AKK, UGR, PHO, ARB, GEZ, ARM; [LGz 9]: GEZ, ETH, ESA, ARB, HBR, AKK, ARM, PHO, UGR
