Íslenskt mál og almenn málfræði - 01.01.1983, Síða 121

Learning about -ari
119
speaking children of the same age used -ari (71% versus 49%) might
be accounted for by showing that -er is a more productive instrumental
suffix than -ari is.
The difficulty, of course, comes in assessing the relative productivity
of equivalent derivational processes in the two languages. One
possibility is to equate the frequency of a suffix with its productivity.
I carried out a very preliminary comparison of this type, based on
the relative requencies of -er and -ari as instrumental endings in the
English and Icelandic versions of the same children’s book, Richard
Scarry’s What Do People Do All Dayl (Icelandic: Öll eru þau önnum
kafin í Erilborg). I chose this data base because it contained a large
number (161) of instrument names in each language for an identical
set of referents. I also felt that the vocabulary used in a children’s
book would be more representative of instrument names familiar to
children than a sampling of instrument nouns from the adult languages
more generally. This limited comparison showed that theEnglish -er
instrumental ending is used 3.5 times more often that the Icelandic
-ari.5 It may be, then, that the relative productivity (i.e. greater fre-
quency) of English instrumental -er contributes to its somewhat earlier
acquisition.
A crosslinguistic comparison of the frequency of infinitive-like in-
strument nouns (i'.e. zero-derived forms, at least for English) is less
supportive of relative productivity as an explanatory factor. Because
Icelandic children use infinitive-like nouns much more than American
children (23% versus 0%), productivity would lead us to expect that
the adult languages exhibit a similar difference, with formally identical
noun-verb pairs being more frequent in Icelandic. In fact, the compara-
tive data indicate the opposite, with infinitive-like instrument nouns
occurring nearly three times more often in English than in Icelandic
(2.7:1). Clearly, some other explanation will have to be found for
this discrepancy in the acquisition patterns for the two languages.
While it would be premature to draw any conclusions about the role
of productivity in the acquisition of derivational morphology based
on these preliminary comparative data, they do support the idea that
5 If -or is included as a spelling variant of -er (as in indicator or generator), the
ratio of -er/-or:-ari increases to 4.3 : 1. It seems likely that American children would
perceive -er and -or as being a single morpheme in the spoken language.