Íslenskt mál og almenn málfræði - 01.01.1983, Síða 112

110
Randa Mulford
shuffled and spread out face down. The child was allowed to select
a card, which the researcher then picked up, without showing the child
the picture. The adult questioned the child about each picture as
above and rewarded him with the picture when he gave either a noun
or verb response, whichever the ,,game“ required. As much as poss-
ible, children were encouraged to give something other than sup-
pletive answers, i.e. not to name known instruments or agents such
as hnífur ‘knife’ for tæki sem er að skera ‘a machine that is cutting’
or pabbi ‘Daddy’ for maður sem er að baða einhvern ‘a man who
is bathing someone’. When they gave this kind of response, they were
prompted to give another answer, with varying degrees of success.
Only a few of the younger children insisted on actually seeing the
pictures before they would answer the experimenter’s questions.
A second researcher wrote down all responses. Responses were also
recorded on a portable cassette recorder as a backup. If a child gave
several answers for picture, all answers were noted, although only
first responses are considered in the quantitative analyses that follow.
When the child had successfully ,,won“ the first set of twenty pic-
tures, either agents or instruments, that set was put aside and the
game continued in the same way with the second set of twenty. With
some of the younger children, it was occasionally necessary to do the
two halves of the experiment on separate days, but most children com-
pleted the task in one 15-20 minute session.
2. Results
2.0
In this study, we asked children to demonstrate their knowledge
of the derivational suffix -ari in two ways: (i) by showing that they
could extract the appropriate verb from an unfamiliar -ari noun and
(ii) by suffixing -ari to a verb stem to derive a new agent or instrument
noun. Their success in the first task, extracting the verb, is shown
in Table II. Even the youngest children performed very well, with
more than 75% correct responses. Generally speaking, performance
is slightly, though not significantly, better on agents than on instru-
ments.
Table III shows the extent to which children used the -ari suffix
to form new agent and instrument nouns. Children at all ages were