Conversion vs. -ize, -ify and -ate

We may now turn to the discussion of conversion in relation to the overt suffixes. Semantically, conversion is the most general case in that the meanings of the derivatives with overt suffixes are a subset of the possible meanings of converted verbs. This means that all of the bases attested with the overt suffixes could, in principle, have undergone conversion instead of overt affixation. What made the speakers choose the overt affixes instead? It seems that one reason for this choice lies in the more specific meaning these suffixes express in comparison to the completely indeterminate meaning of conversion.¹ As was pointed out above, the interpretation of converted items relies on the linguistic and extra-linguistic context to an even greater extent than the interpretation of, say, -ize derivatives. Thus, from the view of perception, overtly affixed forms are better than converted items. Furthermore, conversion does not apply to certain kinds of derived adjectives nor to derived nouns, which makes these classes of base words exclusively susceptible to overt suffixation. A look at the neologisms in appendix 1 corroborates this, since, at least with -ize, a high pro portion of base words are morphologically complex.²

In fact, of the 488 converted verbs only 79 (17 deadjectival and 62 denominal ones) actually express meanings that are also associated with the overt suffixes. If we compare this figure with the number of -ize derivatives, we see that it is only about one fourth of the number of -ize neologisms in the OED. This difference can be interpreted in such a way that conversion is certainly not the most productive process in important semantic domains.

Incidentally, not all of the base forms of these 79 converted verbs could have been overtly suffixed, since there are other restrictions at work. One restriction holding for all overt suffixes which I have not yet discussed, is that compounds cannot be suffixed by -ize, -ify and -ate. However, compounds may readily be turned into verbs by conversion. Of the compound-based converted verbs listed in appendix 1, the following express meanings that are also found with overtly suffixed verbs: cobweb, cold-cream, highlight, mothball, pothole, rustproof, scapegoat, streamline, waymark. A parallel restriction seems to hold for base forms involving the prefixed elements multi- (multiplex), cross- (cross-reference) and super- (supercoil), which do not undergo overt suffixation either. The reason for this impossibility is probably not morphological but phonological. English compounds, as well as the forms with the said prefixes, seem to exhibit a stress pattern that is incompatible with the prosodic restrictions imposed on the derivatives involving overt suffixes.³ Taking these restrictions into consideration the overlapping domain is further curtailed.

With regard to the remaining group of truly competing formations a number of overtly suffixed forms are indeed attested (some of them 20th century forms). Consider the forms in (1), where the converted verbs are listed in the left column, their suffixed rivals in the right column:

(1) 

The majority of the derivatives on the left are synonymous to those on the right, which shows again that token-blocking is not to be expected with low frequency items.

The patterning of the data involving overtly suffixed forms as against converted verbs lead to the same conclusions. The semantic and phonological properties of the individual processes curtail the possible overlap of rival domains to a large extent. The number of actually competing forms is therefore much smaller than previously conceived, and in the truly rival domain all affixes are applicable.

¹ This effect may also be responsible for the slight preponderance of derivatives with the more specific -ate in (Rival morphological processes -ize vs. -ate (3) versus (Rival morphological processes -ize vs. -ate(4) earlier.

² The high frequency of complex bases with -ize is not only due to its semantics but also due to the toleration of stress lapses. Many suffixed adjectives, for example, are dactyls or end in a dactyl (e.g .federal).

³ Thus, there seem to be no mechanisms available to accomodate compound stress to the stress patterns overtly affixed verbs must conform to. The details of this phenomenon still need to be worked out.

Note that derivatives on the basis of words featuring photo- as their first element (photoisomerize, photosensitize, photosynthesize) are not compounds. This is evidenced by their stress pattern, which runs counter to the compound stress rule: the base words are pronounced with primary stress on the second element. In other words, photo- behaves like a combining form (Bauer 1983:213), and not like a noun. This analysis is in accordance with the meaning conveyed by photo- in the above derivatives: 'having to do with light', not 'having to do with photographs'.