Íslenskt mál og almenn málfræði - 01.01.1985, Blaðsíða 61

Language Shift in an Icelandic Child
59
One which contains Icelandic, the other which contains code-switched
data. That is, those sentences which involve using of two codes, English
and Icelandic, within the sentence boundaries are treated separately.
This is done to acknowledge that monolingual rules are not appropriate
in the analysis of sentences and phrases where two codes are used. The
present report considers only the Icelandic data although it will refer to
the results of the analysis of the code-switched data.
The error analysis is based on comparison with an Icelandic sentence
formed according to Icelandic rules of morphology. This is a conveni-
ent method for the present purpose as the concern is about what happ-
ens to the complexity of the Icelandic morphology under conditions of
limited exposure.
One word can include three types of errors, case, gender and number
errors. This is because one inflection (ending) carries these grammatical
distinctions. When all three are in error the same word can be re-
grouped three times.
3. Results
There were 74 words in 1982 which had an inappropriate inflection
(ending) according to Icelandic rules of morphology. 53 of these are
considered in the analysis. The differences are ambiguous cases and re-
peated occurrence of the same word with the same type of error. There
were 94 words in 1984 which had an inappropriate inflection. In the
data analysis 78 of these are considered. The error analysis is concerned
with error type. In 1984 as in 1982 most of the errors were case errors,
32 or 58.2% in 1982 and 48 or 61.5% in 1984. The table below shows
the number of errors per set of data and grammatical categories. The
first number is the amount, the second the percentage of the errors.
Table 1.
case gender number case case case gender total
gender number number number
gender
Febr. 1982 32 7 5 9 1 0 1 55
58.2 12.8 9.1 16.4 1.8 1.8
April 1984 48 12 4 5 7 1 1 78
61.5 15.4 5.1 6.4 8.9 1.3 1.3