Institut de Lingüística Aplicada

Third Mediterranean Meeting on Morphology (MMM3)


Carmen Kelling (University of Konstanz)
The Role of Agentivity for Suffix Selection

The process of deriving nouns from verbal stems is a fairly regular process crosslinguistically and one which has been argued to involve argument structure processes (e.g., Grimshaw 1990, Booij 2000). In particular, there has been much discussion of the licensing of syntactic arguments of the deverbal noun and the effect of the nominalizing suffix on the argument structure of the base verb (e.g., Booij & van Haaften 1988, Laczkó 1999, Markantonatou 1995). However, there has been rather less work on the more subtle issue of which verbal stems are compatible with which nominalizing suffixes and why.

Dowty's 1991 Proto-Role proposal replacing discrete thematic roles (Agent, Patient, Theme, etc). by a set of verbal entailments allows to determine different degrees of agentivity. This is not only fruitful for the explanation of the mapping between syntax and semantics (cf. Zaenen 1993), but also to account for the selection of suffixes.

This talk focusses on the selection of the French nomilizing suffixes -age and -(e)ment. Several authors have proposed suffix selection accounts claiming that -age attaches to transitive verbs and -(e)ment is added to intransitive, reflexive or passivized verbs (Dubois 1962, Trésor 1971f, Lüdtke 1978). I argue that an analysis in terms of syntactic subcategorization is not sufficient for an account of these nominalizations. Rather, the lexical argument structure properties of the verbal bases must be analyzed more precisely in terms of Dowty's (1991) notion of Proto-Roles. It can be shown that the number of proto-agent entailments (a: volition, b: sentience, c: cause, d: movement, e: existence) is important for the choice of -age and -(e)ment if both can in principle attach to one and the same verbal stem: -age is added if more proto-agent entailments are present, -(e)ment is attached if there are fewer, cf. (1) and (2):

  1. Max étire le métal. / 'Max stretches the metal.'
    verb agent-entailments of SUBJ suffix derivation
    étirertr 'stretch' a, b, c, d, e -age étirage 'stretching'
  2. Max s'étire en baillant. / 'Max stretches himself, yawning.'
    verb agent-entailments of SUBJ suffix derivation
    s'étirerrefl a, b, e -(e)ment étirement 'stretching'

In (3) I give an analysis for furetage which is an exception for the syntactic approach. The argument realized as subject is a proto-agent and therefore -age is chosen although fureter is intransitive.

  1. Max a fureté. / 'Max has searched.'
    verb agent-entailments of SUBJ suffix derivation
    fureterintr a, b, c, d, e -age furetage 'searching'

My approach thus presents an alternative analysis to the traditional perspective in that the basic (syntactic) valency of the verb is not solely responsible for the selection of the suffixes. Instead, the degree of agentivity of the subject of the base verb drives the combinatory possibilities of verbal stem and nominalizing suffix. This perspective allows an analysis which makes appropriate predictions for all cases including those that are exceptions for the traditional syntactic account.