Cebuano, one of the languages spoken in the Phillipines, has infixation. This is a morphological process in which a bound morpheme attaches inside a root. More common morphological processes are prefixation and suffixation.
The following are five nouns from Cebuano:
sibwano (a Cebuano)
ilokano (an Ilokano)
tagalog (a Tagalog person)
inglis (an Englishman)
bisaja (a Visayan)
The language names are the following:
sinibwano (the Cebuano language)
inilokano (the Ilocano language)
tinagalog (the Tagalog language)
ininglis (the English language)
binisaja (the Visayan language)
An analysis of the words sinibwano, tinagalog and binisaja reveals that the affix -in- has been inserted between the onset and the rest of the root. However, with ilokano the affix is in front of the root. In the case of ininglis, it is not clear if the affix -in- is the first or second syllable of the word.
To solve the data it is important to find a simple solution. It is not so simple to argue that the affix -in- is sometimes a prefix and sometimes an infix. Rather, a rule can be created which inserts the affix -in- after the onset of the root. The onset of sibwano is /s/. With insertion of the affix after the onset /s/, the result is /s/ + /in/ + /ibwano/ which derives sinibwano. The same pattern applies to tagalog and bisaja. The three words begin with consonants.
However, the words ilokano and inglis begin with vowels. Here the onset of the first syllable is 0. Therefore, the infix -in- is inserted after a 0 onset. This is also known as an empty onset. Here is the derivation for ilokano: 0 + /in/ + /ilokano/ becomes inilokano.
The word ininglis is derived by the same process. To indicate the position of the affix in the word, we can write /in/(Af) to distinguish the affix from the part of the root. Here is the derivation for ininglis: 0 + /in/(Af) + inglis becomes ininglis.
Cebuano is a language with infixation. The infix -in- attaches to the inside of ethnic group names to derive language names. It may appear that the affix is sometimes a prefix because it attaches before vowel-initial roots. However, the rule which states that -in- attaches after onsets solves this problem. In roots which have no initial consonant, the infix -in- is attached after a 0 onset. This makes it possible to express the process as infixation in all cases.