Íslenskt mál og almenn málfræði - 01.01.1981, Blaðsíða 90

88
Karen C. Kossuth
þat er hann fekk. (Haralds saga ins hárfagra, Hkr. 1:107)
DEM REL NOM 3 Sg
(those) which he got.
Cataphoric definites do at least appear preceding proper nouns, parti-
cularly non-human ones:
(4) Þorkell hafði með sér sverðit Skofnung . .. (Laxd.:222)
NOM 3 Sg PREP PHR ACC-DEF ACC
Thorkell had along the sword Skofnung . ..
d) By far the majority of definite NPs in Old Icelandic is exophoric,
the category closest to the original deictic function of the article, or
anaphoric, as markers of given (vs. new) information. Both these uses
of the definite article are cohesive in discourse. The more purely narra-
tive cohesion comes of course from anaphoric definites, which tie
subsequent references to the original mention of an NP. They can
appear at a distance of several clauses, if the NP is important enough
for the discourse:
(5) meyna Herdísi dreymði, at kona kœmi at henni;
ACC-DEF ACC 3 Sg CONJ NOM 3 Sg PREP PHRASE
The maiden Herdis dreamed that (a) woman came toward her;
sú var .. . ekki sýndisk henni konan sviplig. (Laxd.:223)
NOM-DEM 3 Sg NEG 3 Sg DAT NOM-DEF PRED-ADJ
She was . . . To her (Herdis), the woman did not seem handsome.
Though the suffixed definite may in principle accompany every com-
mon noun (except perhaps guð ‘God’), it is frequently omitted in Old
Icelandic narrative prose, in an apparently archaic syntax. Some of the
omissions are formulaic—the result of frequent repetition over cen-
turies of the unmarked definite. References to sun, sky and earth are
just such repeated expressions. Further, the omission of definite
marking on cataphoric expressions is probably due to the fact that they
are specified immediately by the phrase or clause they precede. Definite
marking in these cases might have been felt to be redundant, or perhaps
more likely, the earlier definite markings were purely anaphoric, and
only later extended to include the cataphoric expressions, so that the
saga texts are in flux. Many anaphoric definites are still unmarked in
Modem Icelandic, as Smári (1920:45) observes: