The nasal infix is a reconstructed nasal consonant or syllable *⟨n(é)⟩ that was inserted (infixed) into the stem or root of a word in the Proto-Indo-European language. It has reflexes in several ancient and modern Indo-European languages. It is one of the affixes that mark the present tense.
In the Proto-Indo-European language (PIE), the nasal infix *⟨n(é)⟩ is one of several means to form the athematic present tense. It is inserted immediately before the last consonant of the zero-grade root.
The infix appeared as *⟨né⟩ in the forms where a full-grade stem would be expected, and as *⟨n⟩ in forms where zero-grade would be expected. For example, the PIE root *weik- "to win" would yield a nasal-infixed present stem *wi⟨né⟩k- ~ *wi⟨n⟩k-.
Since the linguistic ancestor of PIE is not known, there can only be speculations about the origins of the nasal infix. It has been suggested that it arose from a suffix (also related to *-neH- and *-neu-) which underwent metathesis.
Besides the nasal infix, PIE employs a number of affixes to mark the present: *-u-, *-neu-, *-neH-, *-sḱe-, *-de-, and others. All in all, PIE has at least 18 ways to form the present tense. For many verbs, several of these presents can be reconstructed simultaneously. For example, Scottish Gaelic loisg "to burn" goes back to *l̥h₂p-sḱé-, a sḱe-present of the root *leh₂p- which is also the source of Ancient Greek λάμπειν (lámpein) "to shine" via its nasal present *l̥h₂⟨n⟩p-.
It is not clear why there were so many different types of present forms with no or little discernible differences in meaning. The authors of the Lexikon der indogermanischen Verben proposed that they were derived from a number of prior grammatical aspects with distinct (but lost) meanings.
In Latin, Ancient Greek and other daughter languages, the *n was assimilated to m before labial consonants (b, p), and to ŋ, spelled n in Latin and γ in Ancient Greek, before velar consonants (g, k, qu). Latin rūpit "has broken" / rumpit "breaks", from *rup- / *ru⟨n⟩p-, is an example of the first case.
The phenomenon of nasal-infixing as inherited from Proto-Indo-European is found in Sanskrit with the greatest morphological transparency, and is taken as a guide to examining the feature in kindred languages.
Three of the ten classes identified by traditional Sanskrit grammarians have nasal infix of some kind, which are higher-grade and accent-bearing in the strong forms, and reduced-grade in the weak forms. The behaviour[a] of the class-7 root √yuj-[b] class-5 √śru-[c] and class-9 krī-[d] can be seen thus:
While these were seen as 3 separate classes by the ancient Sanskrit grammarians, Ferdinand Saussure demonstrated, as part of his landmark work in postulating the Laryngeal theory, that these were slightly different manifestations of the same nasal infix.
Greek has some verbs that show a nasal infix in the present as opposed to other forms of the verb:
English and the other Germanic languages show only vestiges of the nasal infix. The only certain remaining example is English stand, with the past tense stood lacking the n. However, it can still be seen in some pairs of Latin loanwords:
†The Latin reflexes of the PIE aorist came to be used as the perfect.
‡It is uncertain whether *sleh₂gʷ- had a nasal infix already in PIE, since Greek λαμβάνω is only attested after Homer.
In J. R. R. Tolkien's constructed language Quenya, the nasal infix forms the past tense of verbs ending in any consonant besides -m, -n, or -r. Thus, cen- "to see" has the past tense cen-në, but mat- "to eat" has not *mat-në but the metathesised ma⟨n⟩t-ë.