Verb-forming suffixes

For the type Ν-ate some marginal counterexamples are attested in the OED, (affectionate (obsolete), compassionate, divisionate), thereby contradicting Fabb's claim. What is more important, though, is the fact that the productive rule of Ν-ate formation is semantically restricted to chemical substances, whose names, similar to Latinate animal names mentioned above, usually do not involve suffixes. Hence no suffix combinations with -ate can arise. Fabb's example originate and a few others are not productive formations but must be individually listed. In other words, Fabb's observation follows from independently needed mechanisms, namely the listedness of unproductive formations and the semantic restriction imposed on productively formed derivatives.¹

The domain of -ify derivatives (denominal and deadjectival alike) is restricted by the fact that bases must have either ultimate primary stress or must end in [ɪ]. Given that only very few nominal or adjectival suffixes in English have primary stress (-ee, -eer, -esque, and -ese), or end in [ɪ] (-ly and -y), possible suffix combinations with -ify boil down to -ee-ify, -eer-ify, -esque-ify, -ese-ify, and -ly-ify,-y-ify. All other combinations (e.g. with nominal -er, adjectival -al, -ous) are a priori excluded. No examples exist of combinations with primarily stressed suffixes, but it is unclear whether the non-existence of such verbs is really indicative of any structural constraint on -ify, or rather the consequence of pragmatic factors or of some prosodic mechanism affecting primarily stressed monosyllabic suffixes. Thus, it seems that these suffixes are subject to their own peculiar restrictions, perhaps guided by phonological principles.²² Further research is called for to explain the behavior of the three primarily stressed suffixes.

The denominal suffix -ize seems to attach quite often, and naturally, to suffixed nouns of various types. Consider for example computerize, christianize, preacherize, protestantize (if one assumes that the stem is a noun and not an adjective). It is interesting to note that, according to Fabb, deadjectival -ize belongs to a different group of suffixes, i.e. the 'problematic' group 4, which contains suffixes that attach to several preceding suffixes. This has the theoretical consequence that one must assume the existence of two distinct homophonous -ize suffixes, one denominal, the other deadjectival. We will see, however, that both in terms of meaning and in terms of phonology, denominal and deadjectival can be unified and that therefore the homophony account must be rejected. Under the assumption of a unitary suffix -ize it does not come as a surprise that Fabb's generalization concerning denominal -ize turns out to be wrong.

In sum, our discussion of verbal suffixes has shown, that many of the observations by Fabb do not hold up against a larger data base and that existing restrictions on the combinability of verbal suffixes with other suffixes are primarily the result of prosodic constraints.

To conclude our review of Fabb's group 1 suffixes we can state that many of Fabb's empirical claims are wrong and that the proposed base-driven selectional restrictions, semantic and phonological constraints are conceptually and empirically superior to Fabb's selectional restrictions.

¹ In Plag (1996:787) it was still claimed that the whole domain of denominal -ate was lexically governed. Since then, a more thorough investigation into the nature of -ate derivatives has convinced me that this is not the case.

² Derivatives in -esque and -ese seem to take only -ness, and -ee is only attested with following -ism. Both -ness and -ism seem to be phonologically peculiar in the sense that they allow certain structures that are otherwise uncommon with suffixes. For example, -ism may follow words ending in schwa without truncation of schwa (as in Indianaism, example from Goldsmith (1990:261), but cf. propagandism, where schwa is deleted), and -ness seems to tolerate deletion in coordination (cf. happy- and sadness). It is not clear how these facts can be explained, although making reference to notions like the phonological word seems promising (see, for example, Wiese's remarks on a similar case, German diminutive -chen (1996a:69-70).

مواضيع ذات صلة

Non-uniform preservation of secondary stress

No vowel deletion: trochaic vowel-final bases

Vowel deletion: dactylic bases

Preservation of main stress

Stress assignment in English nouns: Pater (1995) OT model

Phonological constraints on -ize derivatives: A new account

Summary: Hiatus, truncation and stress

Truncation and stress

Haplology and stress

Towards well-formedness conditions on -ize derivatives Hiatus and stress

Previous analyses

Optimality Theory: An introduction