Íslenzk tunga - 01.01.1963, Qupperneq 57

OLFUS
55
umlaut. Since neither of the two stems *glf- and *0lf- is otherwise known, the
umlaut must have taken place in the compound itself, caused by the vowel of the
second member, for in dissyllabic simple words a in the first syllable was auto-
matically replaced by g before an o (later spelling u) in the second syllable.
This change was structurally conditioned. The same is not true of e or g before
an o in the second syllable, even though an n-umlaut of e or q might take place.
This fact further supports the first alternative, ölf- < *alf-. The second member
of the compound was not the masculine noun óss ‘mouth of a river’ as some
scholars have maintained, but "‘ö.s, a neuter synonym of the same origin, cf.
Swedish and East Norwegian os n. ‘moutli of a river or a brook’. At the time of
the vowel-change in the first syllable of the compound the second member had
already been reduced to the status of an unstressed suffix.
Section IV treats the question of the gender of os in the Scandinavian lan-
guages, printarily in Swedish and Norwegian. In Danish it is always of common
gender (masculine at an earlier stage). In the Southeast of Norway it is neuter,
but masculine or conunon gender in other parts of the country. According to
Swedish dictionaries the word appears in both genders, and the original gender
is considered to be masculine as in Old Icelandic óss. By a careful examination
of Swedish dictionaries and place names, however, the author comes to the
conclusion tliat Swedish os actually has been neuter from the 13th century on,
except for two isolated examples of common gender, both later than 1850. It
is therefore plausible that os has been neuter in the East of Norway from time
intmemorial.
In section V the first member of the compound is shown to be the stem of
elfr f. ‘river’. This word, as well as ös n., is first and foremost East Norse.
During the early Middle Ages nsually only tliree rivers in Scandinavia were
called Elfr or had names in -etfr, i. e. Göta alv (most of the time called simplv
Elfr, sometimes Gautelfr), Glomma (Raumelfr) and Klaralven (called Elfr in
Ynglinga saga). ln the tract of land bounded by tliese three rivers the words
elfr and ös n. supposedly existed in ancient times, just as Norw. elv, Swed. alv
and Norw., Swed. os n. are frequent there today. A compound composed of
these words may thus be expected in the sante area. Norw. elveos n. actually
is a common word. But a Swedish alvos n., found in a couple of place names in
Bohusliin and Alvsborgs lan, does not exist in tlie Swedish “rikssprák” and has
never found its way into any printed dictionary. Nevertheless, the appellative
íilvos occurs in Swedish dialects, in Varmland aml Bohuslan. A handful of
examples here published (pp. 41—43) liave been found in the dialect archives
in Uppsalu and Gothenburg and one in the word collection of the dictionary
of the Swedish Academy iu Lund. The compound elfös n. (Norw. elveos n.,
Swed. dial. alvos n.) is, in otlier words, known chiefly from the area bounded
hy the three rivers with names in -eljr. Finally, in section V, the author dis-