Arabic is well-known in phonological circles for the diversity and complexity of its broken plurals (jam` at-taksiir جمع التكسير) - that is, plurals formed at least partly by internal modification of the word. The commonest type for four-consonant roots is mainly characterised by an -aa- after the second consonant. For example:
daftar-un > dafaatir-u "notebook"
kawkab-un > kawaakib-u "planet"
Certain rather regular complexities emerge when a long vowel is present in the singular; depending on position, it is either treated as an extra consonant or affects the length of an output vowel:
xaatam-un > xawaatim-u "ring"
risaal-at-un > rasaa'il-u "letter"
qaanuun-un > qawaaniin-u "law"
If the stem has more than four consonants and takes this plural types, the later ones get dropped off the edge, so to speak:
`ankaabuut-un > `anaakib-u "spider"
Now this has been the subject of some interesting work, notably in autosegmental phonology, where such phenomena have been taken as a strong argument for separating consonants and vowels into separate tiers. For Arabic, the plural morphology itself - in this case, the skeleton -a-aa-i[i]-, but there are many others - never seems to involve infixing a true consonant; diminutives in -u-ay-i- can be explained away by treating -y- as a semivowel. But that changes if we look beyond Arabic...
Jibbali/Sheri is a Semitic language spoken on the southern coast of Oman, and (despite its location) is neither descended from nor mutually intelligible with Arabic. Among other changes, it no longer has distinctive vowel length. Its commonest equivalent of the Arabic plural form described above involves the insertion of -ab-/-ɛb- instead of -aa-:
dəftɔr > defɛbtər "notebook"
kənsed > kenabsəd "shoulder"
mɛrkɛb > mirɛbkəb "boat"
muṣħar > muṣɛbħar "branding iron"
although it does have a more Arabic-like form with -o-/-ɔ- in some (mainly feminine) cases:
maħfer > moħofur "basket"
ħalḳũ-t > ħɔloḳum "Adam's apple"
Note that the -ab- plural is productive enough to apply to Arabic borrowings like dəftɔr. I would love to know how this form emerged; as far as I know, no other Semitic language has a b-infix plural.
Ratcliffe, Robert R. The 'Broken' Plural Problem in Arabic and Comparative Semitic. John Benjamins: Amsterdam 1998.
An anachronistic paired word in Chuvash
3 hours ago