Íslenskt mál og almenn málfræði - 01.01.1983, Blaðsíða 122

120
Randa Mulford
detailed crosslingustic comparisons may be a valuable way of identi-
fying general principles influencing language development.
4.3 General principles governing the acquisition of word formation
Once possible methodological and language-specific factors which
may contribute to the patterns of results reported for American and
Icelandic children have been identified, is there anything left to ex-
plain? Are there similarities or differences in the results for the two
languages which suggest that general principles may be governing the
fundamental course of acquisition? Clark and Hecht described the
overall pattern of development for production of novel agent and in-
strument nouns in English as follows:
. . . the children fell into three groups: (1) the youngest children
made only inconsistent use of -er and relied on simple com-
pounds, especially for agents, and on familiar words [suppletives],
especially for instruments; (2) slightly older children began to
make consistent use of -er, but only for one of its meanings,
usually the agentive one; and (3) the oldest children produced
-er consistently for both agents and instruments. (p. 1)
This general pattern is an accurate description of the results for Ice-
landic as well, except that younger Icelandic children favored repeated
infinitives over suppletives or compounds, especially for instruments.
To confirm that the three levels of development noted for English
were also found in Icelandic, I did a consistency analysis (Table VIII).
In it I looked at how consistent each individual child’s answers were
for each meaning (agent or instrument) which the -ari suffix can have.
The criterion for a consistent response pattern was considered to be
at least eight of ten responses of the same type (cf. Clark and Hecht
1982). The only totally inconsistent children were in the youngest
group (n=2). The number of children who used a consistent response
pattern for only one meaning, or who used consistently different pat-
terns for the two meanings, declined with age, while the number of
children who used -ari consistently for both instruments and agents
increased. The mean age of children who never used -ari consistently
was 3;6:8 (n=9), consistently with one meaning 4;5:12 (n = 16), and
consistently with both meanings 4;9:27 (n=23).