Khmer (Austro‑Asiatic, Mon‑Khmer)

Some deverbal nouns (with an overt derivational suffix) are used as verbs.

1 Background

In common with other Southeast Asian languages, Khmer gives few overt morphological cues to word class membership, and roots may well be ambiguous, functioning equally well in different categories. Nevertheless, different word classes can be identified, both on syntactic grounds, by the fact that some roots roots are unambiguous, restricted to a single category: numeral, pronoun, noun, preposition, verb/adjective, conjunction, adverb and interjection (Haiman and Ourn 2003: 507-8). Furthermore, there is some word-class changing derivational morphology, which provides an overt morphological indication of word class membership. It is with such words that mismatches may occur.

2 Word-class deponency in Khmer

Khmer has a number of overt means to derive nouns from verbs. Most of these involve compounding:

There is also a deverbal nominalizing infix, <Vm(n)>, which is not entirely productive, but is nonetheless common.

These nouns typically denote concrete physical entitites (as in the examples above) or are action nominalizations (which are syntactically nominal).

However, examples can be found where these nominalized forms are used as verbs; note that adjectives (or rather, their translation equivalent) are construed as verbs in Khmer. (Note that this behaviour is not displayed by the other deverbal nominalizing devices, i.e. those derived by compounding.)Haiman and Ourn (2003: 518-20)) refer to this phenomenon as 'syntactic backsliding'. In every case, the original underived verbal root is an acceptable alternative:

The infix -Vm(n)- also has some other functions, e.g. causativization (slap 'die' → s<am>lap 'kill'), adjectival modification (cah 'old' → c<amn>ah 'oldish') and, under restricted conditions which Haiman and Ourn are unable to characterize, passivization (s’ap 'hate' → s<am>’ap 'hated'). In order to account for the various functions of -Vm(n)-, Haiman and Ourn make the following proposal:

It may be that both the semantic polyvalence and the occasional meaninglessness of the nominalizing infix -Vm(n)- are a reflection of its diachronically original semantic function, which, we suggest, was nil. (p. 522)

That is, they suggest that the two forms (with and without the infix -Vm(n)-) were originally stylistic variants of each other, e.g. a full form and an allegro form, and that a functional contrast was ascribed to the formal contrast only after the fact. What the infixed forms share is that they add some meaning to a verbal root. On this interpretation, the instances of 'syntactic backsliding' illustrated above represent an archaism.

References

Haiman, John and Noeurng Ourn. 2003. Nouns, verbs and syntactic backsliding in Khmer. Studies in Linguistics 27/3. 505-528.