Prefixation in English and Catalan - Departament de Filologia ...

syllables), and strengthenV/lengthenV (strongA and longA end in a nasal). The focus here,
though, is on A-based Vs derived by the suffix –en (cf. 55).
Although these en-suffixed forms have the same Θ-percolations as enrich (cf.
51), now the zero-suffix is replaced by –en. This third type of deadjectival Vs provides
further evidence to say that the en-prefix does not have any features, or if it does, they
are probably the same as those of the base, and due to the Rel. RHR, the features of the
base get percolated, obscuring those of the prefix.
The present panorama predicts the existence of deadjectival Vs constituted by
both prefixation and visible suffixation. As seen earlier, the source of the [+c] role is a
suffix which can be full or empty of phonological content (widenV, sweetenV, brightenV
vs. richV, nobleV, clearV), the former subject to some phonological constraints. If that is
correct, and the constraints on the suffix’s attachment are satisfied, the prefix en- should
be able to attach to bases with both types of suffixes, deriving prefixed deadjectival Vs,
suffixed and non-suffixed. Considering historical data (Marchand (1969)), one observes
that this prediction is borne out. There was one period where forms like enwiden,
ensweeten and embrighten existed alongside of forms like enrich, ennoble, and enclear.
However, of all these prefixed and suffixed forms, embolden is the only existing word
nowadays. The productivity of the different types of deadjectival Vs will be discussed
in section 4.3.
Regarding the disparate behaviour of English A-based Vs with respect to the
transitive-unaccusative alternation, we have seen that the first group (e.g. enrich,
enlarge) does not allow the V to have an unaccusative variant, an observation unnoticed
until now as far as I know (cf. 48-50b), while the third group (e.g. sweeten, redden)
does (cf. 56-57b). As for the second group, there is no uniform behaviour (53-54b). Vs
like clean cannot have the two variants, whereas Vs like thin can. HK’s (1993, 1998,
2002) syntactic theory of argument structure cannot explain this. According to their
theory, all deadjectival Vs, i.e. Vs incorporating As should participate in the transitiveunaccusative
alternation. By looking at the numbers of Vs in each group that allow the
alternation, one sees that their claim is generally true. Deadjectival Vs admitting both an
unaccusative and transitive variants exceed those Vs that do not. However, there is still
a group of Vs which need some explanation. As observed by Kiparsky (1997), the real
generalization behind the transitivity alternations does not have to do with the category
(A) which gets incorporated into the V, but with the notion of agentivity. In Kiparsky’s
words, ‘the availability of the causative alternation depends on the nature of the Agent’s
40