Íslenskt mál og almenn málfræði - 01.01.1981, Qupperneq 99

Unmarked Definite NPs and Referential Cohesion
97
nouns accompanied by a demonstrative adjective hinn, sá, þessi, such
as sá skjgldr ‘that shield’ or þessi tíðendi ‘these tidings’ (examples from
Haraldssona saga, Hkr. 111:318). The unmarked nouns which seemed
to be anaphoric were also counted as definite. The result of the count,
given in Table III is a distribution not significantly different from that
found by Givón (1974) for English.
Table III. The Distribution of Definite vs Indefinite
according to Sentence Part
DEF INDEF % DEF % INDEF
subjects 810 122 87% 13%
direct objects 216 172 55% 45%
datives 133 18 88% 12%
genitives 153 32 82% 18%
prep. phrases 464 165 75% 26%
The distribution of definite and indefinite noun phrases here confirms
Old Icelandic participation in the topicality hierarchy in which subjects
tend to be the topic, objects the assertion.
5.2 Word Order
Old Icelandic further uses word order (Kossuth 1978a) to indicate
discourse function: Verb initial order and ok ‘and’ with verb initial
order indicate no change in topic. This ongoing word order never begins
paragraphs, and usually has a pronoun topic (Kossuth 1978b). Thus
when an unmarked noun appears in a verb initial sentence, it is prob-
ably the topic. If it is also the subject, there is no doubt whatsoever of
its topicality. With some variation (Kossuth 1980a), the word order
pattern is: First sentence: topic introduced, SV word order. Second
sentence: Topic pronominalized, still SV word order. Third sentence:
both pronominalization and VS word order show that the topic is
established. In case other NPs, which might cloud pronoun reference,
intervene, the unmarked noun is used and it may appear with either
verb initial or subject initial word order. The word order patterns given
above in the table for the passage on King Sigurd provide an example
of how word order interacts with grammatical function and definite
marking to indicate topic in discourse.
íslcnskt mál III 7