Conversion, such as:

  • permit (verb): I permit you to do so
  • permit (noun): Take this permit

Can be considered to be a morphological (i.e. lexical) process. But there are arguments for it being a semantic process, rather than a morphological one. For example, in the first example sentence permit may indeed fit the syntactic category of being a noun, but is occupying the syntactic slot of a verb, so it takes on the property of a verb.

My problem is that I can't really think of a good argument in favour the syntactic argument and can't find too much clear literature on the topic.

  • I don't think /permit/ is a good example of conversion as there is a phonological contrast between the two words spelt that way (stress location). Conversion usually refers to situations where a word from one part of speech category is derived from one in another, with no evident morphological process, and they remain homophonous. Commented Mar 21, 2012 at 4:54
  • @Gaston You're right, there is a difference in the primary stress location in the noun-verb pair, but this is still conversion. Such movements in stress is one of the methods used to determine the directionality of conversion processes (i.e. whether it is a case of noun > verb or verb > noun) Commented Mar 21, 2012 at 10:06
  • 1
    I'm used to conversion as another term for zero-derivation, are you using the term more broadly? Commented Mar 21, 2012 at 13:00
  • Ah I see. Well I'm talking about conversion not from a zero-derivation perspective Commented Mar 21, 2012 at 13:15
  • 1
    Conversion has been understood differently in linguistic morphology. The Marchandeans (followers of Hans Marchand) have adopted the zero derivation approach. It is called so because Marchand argued that there is a zero morpheme in what is usually known as conversion (to cheat=> cheat0). It seems this approach is predominant now. However, there are many others who haven't accepted it. e.g. Laurie Bauer in his textbook prefers to use the term "conversion" and not zero derivation, for reasons stated on pp. 37-38. Quirk et al. 1985 understand conversion rather broadly, including cases like yours.
    – Alex B.
    Commented Mar 21, 2012 at 23:27

2 Answers 2


I personally prefer the definition of conversion as "the phenomenon in which one base-word (stem) may function as a member of different lexical categories" (Don, Trommelen, and Zonneveld 2000: 943, cf. Mel'čuk 1973, Plungian 2011: 149). In other words, a stem that underwent a functional transposition A=>B starts to behave syntactically as other members of class B:

Class A (a lexical verb)

He permitted me to do so.

He permits me to do so.

He will permit me to do so. etc.

Class B (a count noun)

I need a permit.

How many permits did you have?

Take this permit.

On the other hand, in languages with poor morphology where word classes might not have distinctive markers we find another phenomenon, category indeterminacy (or multifunctionality). For example, in Sranan Tongo (Dutch Guyana), Voorhoeve 1979 reports that the word hebi can mean "heavy; to be heavy; weight." However, Don, Trommelen, and Zonneveld 2000 argue that this phenomenon is different from conversion. In the case of conversion, there is always one "original" stem (class A in the example above), to which you can trace the transposed stem:

OED: permit (noun), from permit (verb)

And the original stem is usually more frequently used than the transposed one:

permit (verb), rank: 2121, frequency: 16764; permit (noun), no data (not even included) http://www.wordfrequency.info/files/entriesWithoutCollocates.txt

  • Thanks @Alex B, this is more the kind of argument I am familiar with, having not done much work with syntax. But the original stem is not always the more common one. There is a good discussion in Plag (2003) on the directionality of conversions in English Commented Mar 24, 2012 at 17:19
  • @DangerFourpence, I never said "always", I said "usually". Plag argues the same, notice his "in general" or "the vast majority of cases".
    – Alex B.
    Commented Mar 24, 2012 at 23:52
  • sorry, I didn't mean to give the impression that you were implying it was always the case Commented Mar 25, 2012 at 13:57

You would like to find an analysis where, in the first sentence, permit "fits the syntactic category of being a noun, but is occupying the syntactic slot of a verb." Commenters are for good reason thinking you are talking about zero-derivation. To take it from a neutral perspective, we can assume that the analysis you are interested in is diagrammed roughly as follows.

enter image description here

A word which is normally a noun is behaving as a verb. We can find more dramatic examples of this kind of behavior outside of English. Consider this pair of Tagalog sentences:

Nagtrabajo ang lalaki
worked     NOM man
"The man worked."

Lalaki ang nagtrabajo
man    NOM worked
"The one who worked is a man."

The first sentence is a basic intransitive clause, where the subject comes after the verb, and ang marks the subject as such. In the second one, things got flipped. What seems like it should be a verb is getting marked with ang, and lalaki, which seems like it should be a noun, is showing up in the verb's usual position. It is no longer clear what is a noun and what is a verb. The problem is that "noun" and "verb" are being taken to mean two different things at once. A noun is something that refers, and something that generally refers to an object. A verb is something that predicates, and generally refers to an action. The approach Van Valin (2008) takes in analyzing the Tagalog sentences is to do away with categories like NP and VP in favor of RP ("referential phrase") and Pred ("predicate"), so that only one of the typical functions of nouns and verbs is implied in the higher levels of the structure. A simplified version is like this:

enter image description here

For the English example, it will be harder to pull this kind of analysis off with the example you have chosen, because the semantics aren't very regular. The Tagalog example works because any noun can predicate and the meaning is "to be a N." It is hard to think of other nouns that could be analyzed like your permit example in English, where the noun predicates and takes a subject, an object and an infinitival complement. On the other hand, you might be able to analyze in this way a kind of aversive construction where just about any noun acts as a transitive verb, e.g.,

  • Thomas pantsed Henry.
  • Thomas knifed Henry.
  • Thomas baseballed Henry.
  • Thomas netflixed Henry.

Meaning "Thomas did something which negatively affected Henry, involving N."

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.