Employing the OED on CD: Practical problems

Results

Of all affixes searched for (eN-, be-, -en, -ize, -ate, -ify, and ∅), only be was totally unproductive in the relevant period,¹ whereas all other processes yielded between two and 488 neologisms.

In order to assess and compare the productivity of the processes under discussion, the number of types as attested in the OED have been summarized in table 1:

The figures in table 1 allow the following generalizations. Of all overt verbalizing affixes, -ize is by far the most productive. Novel formations in -ate are much less frequent, followed by those in -ify. The affixes eN- and -en are attested only in very few isolated coinages and can therefore be considered unproductive. The same conclusion holds for the prefix be-, which is not attested at all in formations of the twentieth century (except in the parasynthetic formation of ornative participles, see note 10 above).

These results are roughly in line with the existing literature, which can be summarized as in Marchand's statement that the "only derivative morphemes Present-day English has for denominal verbs are -ate, -ify, -ize." (1969:364). A more detailed picture is given by Gussmann (1987), who does not only hold -ize to be the only truly productive affix, but also states a sharp contrast between -ate and -ify on the one hand, and -en, eN and be-on the other. Of the latter three affixes he says that they are "completely unproductive" (p.96, see also Marchand 1969:164, 148), which is corroborated by our findings. According to Bauer (1983:223) -en "is only marginally productive, if at all", which seems too conservative a statement, since two new formations in 86 years make an affix clearly belong to the category 'not at all productive'. The two suffixes -ify and -ate are set apart from the unproductive ones by Gussmann not only because of "the greater number of derivatives [...] but primarily because non-accidental redundancies can be stated here." (1987:96-97). According to his analysis, -ify and -ate appear to be semi-productive, which is in line with the results in table 1 Bauer comes to a similar conclusion when he states that -ize is the most productive overt affix, followed by -ify (1983:222-223). Unfortunately, he does not say anything about -ate, but Marchand (1969:258) and Plank (1981:214) point out that a certain type of denominal -ate is productive.

The high number of converted verbs in Table 1 is not surprising at all, since it has long been observed that in Modern English conversion is the most productive verb-deriving process (e.g. Jespersen 1942:112f, Bauer 1983:226-227). We will see that the observed differences in the productivity between -ify, -ate, -ize and ∅ correspond nicely with the structural restrictions imposed on these processes, which lends further support to the validity of the OED-based count. Notice, however, that the quantitative difference between -ize derivatives and converted verbs is perhaps not as big as one might have expected, given the general unrestrictedness of conversion. A possible explanation for this result could lie in the greater saliency of overt morphological marking.

Overall, the observed quantitative differences between the processes under discussion corroborate existing accounts of their productivity. It should not be overlooked, though, that the above-cited claims by other linguists are founded to a large extent on dictionary-based investigations themselves. Therefore, it could be argued that my own count is a mere replication of earlier studies. However, to my knowledge, no explicit comparative dictionary-based account of the productivity of these processes has ever been proposed before.²

Having established a reasonable dictionary-based analysis, we may now move on to a text-based account of the productivity of the processes under discussion.

¹ Note again that parasynthetic formations were eliminated from the data, which excludes productive formations where be- occurs in combination with the or native denominal suffix -ed. This parasynthetic be- is described by the OED as follows (prefix be- 7):

"Forming participial adjectives, which unite the preceding senses, esp. 6 and 2, in the notion of 'covered or furnished with,' usually in a conspicuous, ostentatious, unnecessary, or overdone way. In modern use (e.g. with Carlyle) the force of the be- is often merely rhetorical, expressing depreciation, ridicule, or raillery, on the part of the speaker, towards the appendage or ornamentation in question; cf. booted and bebooted, gartered begartered, wigged bewigged. Some of these words have no form without be-, and closely approach the verbs in 5, e.g. bedaughtered, bepilgrimed 'overrun with pilgrims.' This is now the most frequent use of be-, and the formations of this kind are endless". See Marchand (1969:148) for a similar statement.

² Mahn (1971) provides a comprehensive count of one type of derivatives, those in -ify, but does not include counts of its rival affixes. Based on earlier editions of the OED, data from other dictionaries, and on his own collection of words, Mahn arrives at 29 putative 20th century neologisms, of which only 16 he found documented with their first attestation, whereas the others are assumed to be neologisms on the grounds that they are not listed in historical dictionaries (1971:186). At least three of his putative neologisms are already attested earlier (e.g. bronzify 1855, complexify 1839, tourify 1820, see OED), so that his list boils down to 26 as against my 23.