Scando-Slavica 49 (2003), 35-62.
The Names of the Dnieper Rapids in Chapter 9 of
Constantine Porphyrogenitus De administrando imperio
Elsa Melin
The ninth chapter of Constantine Porphyrogenitus De administrando
imperio (DAI) was composed c. 944. For some 300 years several distin�
guished scholars in the fields of Slavonic and Scandinavian languages
have laboured over the interpretation of the names of the rapids in this
chapter. In the 1940s, when two Scandinavian Slavicists, A. Karlgren and
K.-O. Falk, published extensive studies devoted exclusively to the names
of the Dnieper rapids, this discussion was given a new lease. Regarding
earlier literature I refer to Falk 1951, and to the comprehensive commen�
tary and bibliography on DAI published by D. Obolensky (1962). Below,
I shall discuss Falk s contribution, as well as reviews and articles pub�
lished subsequently by scholars such as G. Y. Shevelov, R. Ekblom, A.
Zariba, A. I. Tolka%0ńev, P. Rehder, M. Shapiro, B. Struminski, A. Dany-
lenko and Th. Andersson.
In his strongly critical review of Karlgren s and Falk s works on the
names of the Dnieper rapids Shevelov (1955/1971, 144) explained the
problems in interpreting the names, and why the criterion of synonymity
must be met:
The problem facing the scholar in a study of the names for the seven
falls of the Dnepr [...] consists not merely in deciphering them and
giving their etymologies. The main difficulty lies in the fact, that, as is
well known, Porphyrogenitus regularly gives two names for each
fall one  Slavic , and the other Scandinavian, and next he gives their
etymology or translation into Greek; moreover, he gives this as a
translation of both names. From this it would follow that the Slavic
and Scandinavian names must be synonymous, and their meaning must
correspond to the Greek translation or explanation. However, this cor�
respondence as often as not is absent [...] so far, no one has succeeded
in bringing into harmony the three basic parts of the material in Por�
phyrogenitus.
Taylor & Francis
DOI: 10.1080/00806760310000892 � 2003 Taylor & Francis
TtyotS. Franci Croup
�
8
Downloaded by
[Texas State University, San
Marcos] at 09:25
1
September 2013
36 Elsa Melin
In a note entitled  Postscript 1969 in the reprint edition Shevelov
(1955/1971, 164) declared that his opinion of 1955 remains unchanged
even after the subsequently suggested interpretations. Tolkacev (1962)
lays down that there is semantic correspondence in two names only.
However, it is crucial to point out that a reliable correspondence be�
tween the three available names can be reached only if they have been
correctly rendered into Greek.
The Name of the First Rapids  ���ż�Ą� (Explanation: ź�
�żąźŹ�ąą)
The Emperor Constantine reports a single name  ���ż�Ą�  which
he took to mean ź� �żąźŹ�ąą  do not sleep in  Russian (Old Swedish)
as well as in Slavonic. If we are to imply Slavonic ne (= ź �) we must
borrow the final v of the preceding word, ���ż�żźąś�ź��ż�, a procedure
commonly adopted by scholars.
Falk (1951, 85) suggests a completely different explanation for the
name  ź���ż�Ą� = ustupi =  rocks . Some scholars have recognized a
Norse word in the name thus Karlgren (1947, 107) supandi and Sahl-
gren (1950a, 145) cisupi. The critics o f Falk s suggestion have objected
that his reading requires several emendations (Zariba 1958, 153f), and
that it does not fulfil the criterion of synonymity with the explanation
(Shevelov 1955/1971, 146; Sahlgren 1952,7).
Tolkacev regards the problem as quite unresolved, the name, however,
as Old Swedish. Several scholars have expressed criticism o f the inter�
pretation of the name as an imperative.
The most recent explanation involving negative imperatives was
brought forward by Strummski, who (1996, 149) gives negative impera�
tives for both  old Ukrainian name ne sT>pi and  old Nordic name ne
siirbi =  shouldn t be sleepy . Greek explanation:  do not fall asleep .
I would like to suggest a quite different explanation. In the Codex Su-
prasliensis, 479, 19 (Meyer 1935, 140), we find the Church Slavonic
word neusypaj^i, the definite form in masc. nom. of the present participle
o f ne usypati,  he who does not fall asleep, who is always awake . It oc�
curs in a text read at Easter, attributed to an archbishop of Constantinople.
It is a direct translation o f Greek Ź �żŻź��ż�. Slovar' russkogo jazyka
XI-XVII vv. gives from Lavr. let. 105 the same example as occurs in the
Suprasliens is: neusypajaj: cervb neusypajai. In this participle -ajaj may,
due to a haplological loss, have resulted in -aj, neusypaj, which is the
Scando-Slavica Tomus 49, 2003
8
Downloaded by
[Texas State University, San
Marcos] at 09:25
1
September 2013
The Names o f the Dnieper Rapids 37
same form as the imperative  don t fall asleep . The Greek on may have
been contracted into -�. Similar final contractions are found in Gignac
(1976, 248, 2): Ą �Ź �� Żż � Ą � � � ą ą, �żźŻ�� for �żźŻ�ąą. Sreznevskij
(1893-1903, s.v.) gives the corresponding Old Russian participle: neusy-
pajai monastyrb  źż�� �� � Ź�żąź��� � was a monastery near the
outlet of the Black Sea at Constantinople.
The vowel sequence �� could be monophthongized into �, especially in
older Modem Greek dialects such as Pontic. This phenomenon is treated
by Dieterich (1898, 78f; 279, d) and Oekonomides (1908, 5 no. 9).
The sequence o u reflects [y], as in �ż � � �ł� ą � � (DAI 9/7) and in
Ł�� ż� �ż ��, the seventh rapids. This rendition is common also in Sla�
vonic place-names in Greece.
In later periods these rapids were called Kodak, Kodatskij porog, and
several further variants (see Evarnickij 1890, 35). At one point the Turks
had built a fortress here, according to Struminski (1996, 138) with a
Turkish name *Kiid�g  Watching , lacking a source reference. This mea�
ning is found in the Uigur word kyd�sci < kyd�s + ci  the watchman ,
which according to Radloff (1893-1911, 2:1487) was formed on the verb
stem kyd�s  to guard .
Hrusevskyj (1906, 151 fn. 2) categorizes Uigur in this period as a West
Turkish language, following Munk�csi:  ogur  ugur {Namen der westli�
chen T�rken, Ujguren)".
In C. Brockelmann s 1921 article on the language of the 11th-century
Turks verbal adjectives are treated on p. 34. To hardi =  der Mann ging
was formed the participle bardaci =  der gehende Mann . On this analogy
kyd�sci might have the sense o f  he who watches, is vigilant . I find it
possible that it is this Turkish name that has been translated into Old Rus�
sian. They corresponded in sense as well as in form, as a present participle
masc. nom. The name signified the rapids as well as the fortress.
Translations of Turkish names were common in Byzantine literature,
according to G. Moravczik (1943, 23). On p. 229 he deals with the name

Ź���, a Chazar fortress on the Don. The Turkish word means  white
house , and the Slavonic translation is bela veza,  white tower . L. Nied-
erle (1902-24, 1:4 121) explains  ą �Ą�ż� �Ź���ż� in ch. 9:91 o f DAI
as a direct translation of Turkish Akkerman,  white fortress , in Russian
Belgorod. This fortress stood on the south bank of the Dniester. If my
interpretation is correct there were fortresses, with Turkish names trans�
lated into Slavonic, on all the great Russian rivers, the Don, the Dnieper
and the Dniester. But apparently the Slavonic name of these Dnieper rap�
Scando-Slavica Tomus 49, 2003
8
Downloaded
by [Texas State University, San Marcos] at 09:25
1
September 2013
38 Elsa Meliti
ids and the adjoining fortress was not widely known and used. No Old
Swedish name is mentioned, merely that its significance is the same as
that o f the Slavonic name. Was the translation made for the sole use of the
Emperor?
The Second Rapids  OuXjBopai,  OaTpo[3ouvi7Tpax
(Explanation: to vricriov tou 9payi_iou)
The significance of these names, O uX^opoi,  O oTpoPouvrrrpax  the
rapids with the islets  has been uncontested ever since Thunmann s
interpretation in 1774 of the Scandinavian name Hulmfors. However, in
my opinion the forms have still not been satisfactorily explained.
OuA(3opai is described by Obolensky (1962, 45) as a dative-locative
form. Struminski as well (1996, 139) explains the name as a dative, refer�
ring to Sahlgren. But Sahlgren had changed his view in 1950, concurring
in Torbjomsson s (1911) opinion that -i is a Greek ending. Sahlgren as�
serts (1950a: 143f) that there are no datives whatsoever in the Scandina�
vian names for rapids. This is also my own opinion.
The rendering o f the reduced vowel b as o u in  OaTpo(3ouvnrpdx
has been long discussed. Basing his hypothesis on ^-spellings in Church
Slavonic texts Shevelov (1955/1971, 147) explains ou as a development
o f -b rather than b. In line with all his other explanations of the phonetic
development o f the names o f the rapids Falk regarded this ou as a
Ukrainian ir. On Falk s Ukrainianisms Shevelov states in conclusion
(1955/1971, 161) that features of Ukrainian phonetics do not appear
in these names .
Among the names of the rapids there is yet another name with the ad�
jectival suffix -*bttyj  (3ouAviTrpdx, a name for the fifth rapids. It is
generally recognized that in this name there is no trace of &-. In my view
ou in  OaTpo(3ouviTrpdx is unlikely to be a reflection of b or t>. I would
like to offer an explanation o f ou based simply on the Greek pronuncia�
tion o f the consonant sequence [v] + [n], a combination which does not
exist in Greek (see Meyer 1896, 337). Thumb states (1910, 19):  Vor v ist
P> zu n geworden. In Vasmer 1941 there are three options for expressing
[v] + [n]: 1. by inserting the vowels i, e o r a i between (3 and v, as in
Z a p o p iv a = *Zarovbna (1941, 62); 2. by substituting n v fo r (3v:
C uA iuva (1941, 158) = *SIivbna; 3. by (3v:  Opo^viK (1941, 194) =
*Orechovbriikb.
Scando-SIavica Tomus 49, 2003
8
Downloaded by
[Texas State University, San Marcos] at 09:25
1
September 2013
The Names o f the Dnieper Rapids 39
The last example is taken from Fiorina in Macedonia, a region where
the population is now, as it always has been, predominantly Slav. Fiorina
is the only area in this material where � and � were used in conjunction.
In the same records of names anaptyxis o f o and ou is found in words
which have never contained a reduced Slavonic vowel, e.g. �ż��ś�ą =
gvozd(Vasmer 1941, 37)  forest .
E. Schwyzer gives (1953, 277ff) the following description of epen�
thetic vowels in Greek:  Der �bergangslaut erh�lt teilweise die F�rbung
des Vollvokals der folgenden oder auch der vorhergehenden Silbe, teil�
weise wird er ohne R�cksicht auf die Umgebung durch i oder � wieder�
gegeben. He continues by stressing that this phenomenon is almost al�
ways vernacular and sporadic. Epenthesis o f such anaptyctic vowels is
outlined by Dieterich (1898, 42f). In Modern Greek the most common
vowels to appear between a consonant and a lateral, v or  (in rare cases
p), are i and ou. This development took place primarily in older, eastern
dialects, such as Pontic, but was not realized consistently. Consequently, I
regard ou in Ś � ��ż�ż��ąĄ �Ź � as an anaptyxis between the consonants
[v] and [n], which in Greek could not be pronounced in conjunction.
As regards -Ą�Ź� I believe it to be a Church Slavonic metathesis, just
as e.g. Sachmatov (1916, 80f) and Falk (1951a, 114) assumed. I agree
with Tolkacev in his claim (1962, 57):  napisanie prag~b  bylo vosprin-
jato Konstantinom v vide Ą � ą ł (bez glasnogo -b) ... � moglo otrazat' is-
konnomu g v zaimstvovannych slovach , but in my opinion it was not
written but only spoken.
The Third Rapids  ��ą���Ż (Explanation: ��ż� Ć �ąłźż�)
For these rapids only a single name is extant. This is now unanimously
explained as the present participle of OSw gcella,  roar . The question of
why the name ends in -5pt instead of, as might be expected, in -�ą, has
been brought up once more through the contributions of Strumm ski
(1996) and Danylenko (2001). Several scholars, including Falk, have
regarded -��ą as part of the Greek word [ ą�]��Ż. Vasmer (1971, 873),
however, sees an influence from the Greek word for  cold in the femi�
nine  ł�ą����.
As early as 1911 Pipping (p. 17) understood the word as a plural of the
present participle. His view was supported by Torbj�msson, who pointed
out (1911, 4) that it is not uncommon for rapids consisting o f several
Scando-Slavica Towns 49, 2003
8
y
Downloaded
b
[Texas State University, San
Marcos] at 09:25
1
September 2013
40 Elsa Meliti
ridges to receive a plural name. Torbj�msson also (1911, 8) interpreted
the final -i as a Greek ending, in my opinion a correct interpretation.
Strumiriski (1996, 142) has concurred in Pipping s view of this name
as a plural. On the ending, however, he comments:  The pi-ending proba�
bly reflects the palatal pronunciation o f R . If the date assumed by Nordic
philologists for the disappearance of the palatal [r ] is correct the graph�
eme pi cannot reflect this sound. Noreen (1970, ż265) maintains:  Das
urn. R ist schon vorliterarisch, am fr�hesten nach dentalen und inter�
dentalen Konsonanten, mit altem r zusammengefallen.
On the corresponding Slavonic name o f the rapids, which is not re�
ported by Constantine, Vasmer states (1971, 873) that it does correspond
to Zvonec, reported later. In the oldest record of the names, that by Las-
sotas of 1594 (Kleiber 1959, 97), which contains all names later than
Constantine s DAI, we find the plural form Zwoncze, not as Vasmer
says the singular Zvonec. However, in Danylenko s opinion the records
of Lassotas are not reliable, they are polonized. Instead, Danylenko (2001,
50) sees in the r-form a personification o f the present participle, as in the
personal name Stigander (= Stig) and the animal name ]>rondr. On the
other hand, these words signify animates, usually people Prondr being
used as a male name and outside this category I have not been able to
find any present participle ending in -r in the singular.
These third rapids consisted of four ridges. Every time the water
plunged down a ridge there was the same roaring sound. What could be
more natural than using the plural for these repeated roars.
The Fourth Rapids  ą �ąĆ � �, � �ą ��� (Explanation: � ą� �ą
Ć� ��ż��ą� oi Ą���Ź�żą �Ż� �Ź ą�Ź �ąą �ż � Ć �ąłźż�)
In DAI 9/45-56 the Emperor Constantine describes the struggles of the
Rus' at these rapids. These are the only rapids where the boats are carried
over land six miles, which according to Obolensky (1962, 48) is
equivalent to 8.8 kilometres. At the other rapids the boats are pushed on�
wards by their crews wading barefoot, except in the last rapids, where the
boats pass along the main channel.
ą �ąĆ��
V. Thomsen (1919, 307) analysed the name as consisting of ei (e , a i),
 ever and forr  violent . His explanation was accepted by Falk (1951,
Scando-Slavica Tomus 49, 2003
8
y
Downloaded
b
[Texas State University, San
Marcos] at 09:25
1
September 2013
The Names o f the Dnieper Rapids 41
135) and others. Sahlgren (1930, 146 f) suggests a completely different
interpretation Aidfors, Modern Swedish Edforsen,  the rapids by the
isthmus ; the Northmen had to port their boats across an isthmus. Ekblom
(1951, 171f) rejects in my opinion justly Aidfors on paleographical
grounds, and because a e ^ o p is not a semantic parallel of VEaor|T. Ek�
blom is sceptical also towards Aiforr and, like Obolensky (1962, 47), he
maintains that so far the name has not been given a satisfactory explana�
tion.
With Thomsen (1919) I understand ctsi as OSw ai, ON ei. This is, ac�
cording to Hellquist (1939, 172, sub 1. e-prefix), a frozen inflected form
that could be combined with nouns as well as adjectives. In combination
with a noun it is evidenced in the personal name Eric Emuni, according to
Knudsen, Hornby, and Kristensen (1936-64:2, 248)  the man with the
everlasting memory . The word thus means  continuous as well as  al�
ways .
In 9 op I recognize the Old Swedish noun for. For its interpretation I
refer to the well-known Hogby stone (Brate 1911 no. 81), which was
erected around A.D. 1000. The runes read  Fiall a furi fr0 kn draengr As-
mundr . The dative a furi is the most intensely discussed place-name on
the stone. Pipping (1936, 71) proposed F0ri instead of the previously
suggested Fyri. Sahlgren (1939, 162) proposed a specialised sense for
F0ri  a place, whether in a stream or a lake, so shallow that people can
wade through the water. Hellberg (1950, 45) points out that F0ret ap�
pears in Runic Swedish on the Hogby stone. He concludes that the sense
proposed by Sahlgren is found in place-names from Uppland containing
F0ri. Andersson (1971, 29) declares that Hellberg had proved to full satis�
faction that Sahlgren s hypothesis was correct.
Sahlgren continues (1939, 162):  The word, which evidently had its
origin in Uppland, seems to have spread along the ancient itineraries. It
also appears in place-names, such as Forkarby.
Accordingly, I interpret the name  AE1 9 0 P as  the rapids that are al�
ways shallow . They were so shallow that the boats could not be punted
even when unloaded.
Rygh (1904, s.v.) cites For and F(?r for A.D. 1322, and gives 21 ex�
amples of these roots in Norwegian river-names. Since this meaning of
the word originated in Uppland this sense must have been well known to
Vikings setting out on expeditions. One o f the routes led from Kungs-
hamn in Uppland through Foret,  the shallow lake  formerly much
larger than today to Ostra Aros (Uppsala), see Hafstrom 1949, 67.
Scando-Slavica Tomus 49, 2003
8
Downloaded
by [Texas State University, San Marcos] at 09:25
1
September 2013
42 Elsa Melin
!aor|T
N eaariT derives from ne-j^-sytb  insatiable . But this Church Slavonic
word also means  pelican , and most modern commentators have trusted
Constantine s explanation that  pelicans build their nests on these rocks .
A large number o f hypotheses involving birds have been proposed over
the years. Certainly it is possible to maintain with Kleiber (1959; 1960)
and other scholars that rapids or streams have got their names from
birds. Adding -fors is unnecessary, nor need there be any nests; some�
times the occasional appearance o f large flocks o f birds has been suffi�
cient. But there should be some logical grounds for the naming if this
hypothesis is to be believed. Struminski (1996, 144; 149) interprets
NeaoriT as  seagull , and in order to find a corresponding Norse word he
reads, rather than det9 op, *Maeicp6p a word in which he assumed an n
taken from the particle hev.
M. Shapiro (1982) has dealt exhaustively with *ne-jq-sytb, a word
which in the sense o f  pelican occurs just once in the Church Slavonic
Bible, in Psalms 102. The sole reason, Shapiro says, why all bird names
must be rejected is the strict synonymity between the Old Swedish and the
Slavonic names. His view is also mine. Shapiro interprets this name as
 ever violent, ever rapid . In its original sense o f  insatiable VEaar|T
would refer to the wild and dangerous rapids, which must have taken a
great many lives. But for my own part I cannot see  ever violent, rapid as
a strict parallel of  insatiable .
Since the root of VEaar)T is -syt- it can be associated with terms for
water. According to D al' sytaja voda means  polnaja, v koryte; samaja
vysokaja v beregach, bez razliva ili poema . Moreover, the sense o f the
verb sytet' is exemplified;  o vode pribyvat', polnet', ne razlivajas' esce
za berega . Without the negation the name o f these rapids would signify
 filled with water to the brim, but not flooding the banks . N eaar|T is
likely to mean that the water did not even reach the banks.
I suspect, with Thomsen (1919, 306) and other scholars, most recently
Shapiro (1982, 169), that the original name of the rapids might well have
been what they were called in later records Nenasyt, + suffix, Nenasy-
tecb.
NEaar|T,  insatiable , means that the rapids were never sated with
water. As Obolensky points out (1962, 47):  This is an eminently suitable
name for the largest and most fearsome o f the rapids, the only one which,
even in the spring, is never wholly covered by the water.
Scando-Slavica Tomus 49, 2003
8
y
Downloaded
b
[Texas State University, San
Marcos] at 09:25
1
September 2013
The Names o f the Dnieper Rapids 43
It was quite possible for Greek words to end in p, but final -� is as a
rule simply dropped.  Only in cultivated speech is it retained, as in
�  ą� Ź � ��, though even here its pronunciation appears rather affected
(Jannaris 1897 żż 218, 218b). However, in e.g. Pontic all consonants may
appear in final position (see Jannaris 1897, ż 219b). Besides, there are no
traces of a reduced vowel in the Slavonic name. According to P. Rehder
(1968, 154) the fact that -j%- is not rendered by a nasal proves neither the
existence nor the disappearance o f a Slavonic nasal in this period, merely
that certain consonant sequences, e.g. -��-, were not permitted in Greek.
The Fifth Rapids  �ą�ż�Ć��ż^, � ż � � � Ą � Ź � (Explanation:
�ą� �ą ź �łŹ � � Żź��� Ź Ą ż����)
�ą�ż�Ć��ż�
Most scholars accept Lehrberg s interpretation (1816, 369) of this name
as  the wave-rapids , Old Icelandic bara  wave and fo rs in a grecized
form. Only in Jurgevic do we find a warranted distrust:  ... we find in Old
Icelandic the form bara  wave , but are we therefore justified in attempt�
ing to prove that �ą�ż�Ć��ż� must necessarily be half Icelandic and
half Swedish, a compound of Old Icelandic bara  wave and Modern
Swedish/ors? (Falk 1951, 159)
Thunmann (1774, 387) was the first scholar to read � as [v], Falk
(1951, 163) reads Varuforos, which he understands as  the rapids that
abound in tall cliffs and islets . In Modem Greek the letter � can be pro�
nounced either [v] or [b]. The question is whether, in this case, both op�
tions were open. Vasmer (1941, ż15d) shows that in Slavonic place-
names in Greece they were restricted. He maintains that in early Modem
Greek sandhi variation in certain consonant sequences entailed the pro�
nunciation of the second consonant as a plosive; in e.g. ��� �ą�ąą��� �
must be pronounced [b]. The phrase in our text is ź�� �ą�ż�Ć��ż$. In
this respect Thumb (1910, ż15) puts articles and particles in the same
category. If Vasmer is correct we should read [b].
I read Barouforos, but see no reason to understand the word as  the
wave-rapids . I have not been able to ascertain a single Swedish, Danish
or Norwegian name for lake or stream containing bara  wave , neither in
Hellquist 1903-06, Jonsson 1966 and Kousgard-S0rensen 1968-78 nor in
Rygh 1904 or Hovda 1966. L. Ryman of Sprak- och folkminnesinstitutet
(personal communication) confirms that neither bara  wave , nor the cor�
Sccindo-Slavica Tomus 49, 2003
8
Downloaded by
[Texas State University, San
Marcos] at 09:25
1
September 2013
44 Elsa Meliti
responding Modem Swedish b�ra, is found as a place-name element in
their Swedish etymological files.
I understand (3apouin the sense  free from everything that normally covers the ground . B ar
in this sense is part o f many place-names. A few examples from the
province of Skaraborg (Ortnamnen Skaraborg 1950-81) are Barebacka
(11:1,7); Barm ossen (13:212) B arefjall (1:305), etc. We find it also
among Scottish names: Barra, usually interpreted as St. Barr s Isle, may
really mean  bare isle (Mackenzie 1931, 128). In Mackenzie 1931, 141
we read  Barra: the derivation may be from ON berr, Dan. and Swed. bar
 bare . In the SAOB vol. B p. 284 we find from the year 1605 the ex�
pression  nar sjon var bar o f water that is open, not frozen over.
The vowel ou is likely to be a Greek anaptyxis (see sub Rapids nos. 2
and 6).
The element bar was, in initial position, probably unique in expressing
the sense of  free from , but usually with reference to vegetation only.
What, then, was it that Barforsen was free from? In my opinion all
those obstacles that cluttered the other rapids rocks, sandbanks, islets,
etc. For the crews this was useful information, which could be supplied in
a name.
BouXvriTTp�x
A clear majority of scholars about twenty have regarded the first ele�
ment of the name as an adjective from *vblna  w ave , in recent years
Obolensky (1962,48) and Danylenko (2001,45).
I see several factors that contradict this interpretation. Constantine did
not write the expected form o f the adjective vblntnyj but only
PouXvr)-. This slip has been explained as a haplology bn was left out.
Neither Sreznevskij (1893-1903) nor Dal' (1903) adduces an adjective
vblnbnyj formed on the noun volna with the meaning of  wave- . The only
meaning given by Sreznevskij is  wool . Slovn�k jazyka staroslov%1łnsk�ho,
however, cites an adjective derived from volna  wave . Thus we cannot
know with certainty whether this adjective was used in Russian or
Ukrainian territory during this period. Today, the adjective volnovoj is
used in Russian, while Ukrainian has a different noun for  wave  iilija.
When Obolensky (1962, 49) derives the name o f these rapids from
volna  wave he finds support in Miller. But Miller (1885, 26), quoted by
Falk (1951, 160), believes that the g in Volnigskij, a derivation from
Scando-Slavica Tomus 49, 2003
8
Downloaded
by [Texas State University, San Marcos] at 09:25
1
September 2013
The Names o f the Dnieper Rapids 45
volna, has dropped out o f Constantine s name for the rapids because
Constantine had confused this name with that of the ninth rapids. These
rapids were later called Vilbnyj, which means  free . If there was indeed a
confusion it was due to Constantine himself. However, it is my opinion
that in Constantine s record there was no confusion, nor a slip of the pen.
In Kleiber 1959 (p. 97) we find a table of the names of the rapids, re�
corded by scholars and writers from 950 to 1892. On the order o f Peter
the Great B. Baturin and I. Chruscev carried out geodetic investigations in
the Ukraine for seven years, 1721 to 1728. They drew maps of the country
which were published by the Russian Academy of Sciences. From 1730
onwards we find the names Volnoi and Volnoj for the seventh rapids as
well as for the fifth, i.e. identical names for two rapids. These two explor�
ers are likely to have had time enough to ascertain the correct names for
their maps. It follows that in those days the name of the fifth rapids was
Volbnyj.
Danylenko (2001, 46) recognized the problem of the numerous names
for the fifth rapids which contained Vol'n- and Vil'n-. These occurred in
maps from the 17th century and later, or were mentioned in literature. He
concluded:  This puzzle still resists comprehensive solution.
There is no way that the sense of  the wave-rapids can be supported
by Constantine s explanation  because it forms a large lake . The English
translation of Żź�� into  lake is inadequate. Liddell and Scott (1925-40,
s.v.) give  pool o f standing water, left by the sea or a river , Pape (1914,
s.v.)  stehendes Wasser . Evidently there was here a quiet stretch of water,
contrasting with the main channel of the rapids.
Hence, Lehrberg suggested that Żź��� should be replaced by �Ż���
 whirlpool , a conjecture accepted by Thomsen and others but rejected by
e.g. Malitsky (quoted here from Obolensky 1962, 49). Malitsky pointed
out that in these rapids the fall was not very great, and that there had even
been a small harbour at the end of the rapids. For Żź �� he suggests
 creek .
My conclusion is that the two names mean  the free rapids , i.e. free
from stones, rocks and sandbanks. Already Thunmann (1774, 388) under�
stood it so:  Der slavische Name Wolny = Prach bedeutet einen freien
und ungehinderten Wasserfall . Falk, too, interpreted ł ż ���Ą �Ź � as
 the free rapids , but believed them to be comparatively free from
water. But he changed the order of the rapids, placing the fifth where
others have the seventh. For this he was criticized by Slavicists such as
Ekblom and Shevelov. Moreover, Falk himself (1951, 170) admits that
Scando-Slavica Tomus 49, 2003
8
Downloaded
by [Texas State University, San Marcos] at 09:25
1
September 2013
46 Elsa Melin
Varnforos,  the rapids that abound in rocks , and BouXvniTpdx,  the
rapids free from water , do not constitute a semantic parallel. His inter�
pretation thus contradicts his chief thesis that all the Ukrainian names
were direct translations from Old Norse.
The expression volnaja voda refers also to lakes; according to Dal'
1903 it means  glubina, na kotoroj sudno stoit bezopasno, ne na meli,
dostatocna dlja sudna glubina, pri naibol'sej ubyli, otlive, vygone . In
later periods they were called Sina Luka,  the blue bend , which also goes
well with the sense  blue, open water surface .
Here I would like to recall Kleiber s description (1959, 76f) o f the
changes the Dnieper has undergone over the centuries. Kleiber believes
that the river must have been considerably deeper when Constantine
wrote about its rapids a thousand years ago, before the extensive defores�
tation of the region:  Der Unterschied im Wasserstande von damals und
jetzt muss nicht in Zentimetem, sondem in Metem gemessen werden.
In BouAvr|TTpdx [o] is rendered by ou. I interpret ou as a Greek ren�
dering o f Slavonic [o] in an unstressed syllable, as first pointed out by
Tolkacev (1962, 53). (See the section  What Greek Dialect Is Repre�
sented in the Names of the Rapids? on page 57.)
The Sixth Rapids  Aeccvti  |3EpouTŁr| (Explanation: (3p a a |a a
vspou)
A eavT i
Obolensky (1962, 49) comments on 9/61-62:  The name is quite clear, it
corresponds to the Old Swedish participle le(i)andi, from lea  to laugh
[...], which, as Falk has pointed out, was an onomatopoeia. B. Ravndal
(1938, 92) writes:  Number 6 vaunts the name o f Leanti or  Laughing ,
corresponding to the Minnehaha of America s Indians. Thomsen as well
(1919, 310) had referred to this Indian name, but it does not support the
sense o f  laughing . N. Holmer, a scholar eminently versed in North
American Indian languages, has declared that the word quite simply
means  water-fall (personal communication to Falk, see Falk 1951, 179).
As early as 1774 Thunmann (p. 389) had proposed  laughing as the
sense of le(i)andi. Miller (1885, 27) mentioned this interpretation, but
suggested another possibility, the participle  hlyjandi,  sogrevajuscij no
znacenie malo podchodit k (3pdana vepou . Danylenko (2001, 52) takes
issue with the sense of  laughing :  But this interpretation is not remotely
Scando-Slavica Tomus 49, 2003
8
Downloaded
by [Texas State University, San Marcos] at 09:25
1
September 2013
The Names o f the Dnieper Rapids 47
similar to the explanation of �spour^ri as vertigine actum. Struminski
(1996, 147) develops Miller s suggestion of  warming , proceeding from
the intensive ya-verb *hl0ja  to boil hot very much , while Danylenko
(2001, 54) associates this verb with the IE root *kel-, Latin calere.
Shevelov (1955/1971, 151) takes issue with Falk s interpretations of
the two names:
The correspondence in meaning between the names for the falls is in
itself questionable. In order to make things fit, Falk ([1951,] 186f) first
has to declare that vbreti designated a sound, and then has to ascribe to
its Scandinavian correspondence the meaning not only of laughter in
general, but also of the most diverse sounds. This all, however, is con�
firmed neither by East Slavic nor by Swedish linguistic facts.
Falk himself pointed (1951, 186) at a crucial argument against the expla�
nation  smiling, laughing :  It should be mentioned that he [Hellquist,
EM] does not adduce a single name for a lake formed on the verb  le , and
that the senses of  smile and  laugh are not given. And further on in fn.
2:  A fresh examination of place-names formed on Le- might provide
new, divergent results.
I choose to base my argument on the Swedish lake names Stora Le and
Lel�ngen, as well as on place-names that contain the element le-. I will
exemplify by these place-names from the province of V�rmland (Ortnam-
nen V�rmland 1922-62): 11:60 Letj�rnet, 111:71 Lesj�n, Letj�men,
111:51 Lesj�n, Lesj�fors�lven. In the province of Skaraborg (Lundahl
1950-81) we find 16:12 Lefors, in V�sternorrland (Ortnamnen V�ster-
norrland 1955) 4:6 Lean, ibid. 65 Lesj�n. The first part of the compound
is frequently regarded as uncertain. No comprehensive survey of these
names seems to have been made.
For the explanation of Le Hellquist (1939, s.v.) refers to Noreen (1909)
and to Kock (1899). ON hier, from the Edda, with the senses of  lake
and  seagod , is treated by both o f these Scandinavian philologists.
Noreen regards hier as fully identical with the lake name Le. He derives it
from the IE root *hlewa, Latin cluo,  besp�le, reinige , and refers to the
Old Irish river-name Cluad. In Pokorny 1959 the corresponding root is
*kleu-, *kl�-, *klu- (p. 607). Noreen does not mention Kock s interpreta�
tion of hier. Kock associates Hier < *HIewaR,  loud , with examples that
we find sub *kel-  rufen, schreien (Pokorny 1959, 548), and makes no
attempt to explain the lake name.
Scando-Slavica Tomus 49, 2003
8
Downloaded by
[Texas State University, San
Marcos] at 09:25
1
September 2013
48 Elsa Melin
In Denmark Kousg�rd-S0rensen cites s.v. Le:  Lehulta sj�, Leesi�,
Lebaek (1968-78:4, 291f). For the Danish names he surmises an uncom�
pounded lake name Le, and links them with the  etymologically obscure
Le and Lel�ngen. He points out that nobody has attempted to explain le -
lae in the names discussed; the East ON correspondent of We- ought to be
lae. Thus he rejects the derivations proposed by Nor�en and Kock. Possi�
bly, he continues, Le might derive from *Hlawian, an adjective meaning
 lukewarm .
Evidently there is no consensus among these three philologists on the
derivation and signification of Le. Hence I would like to propose a hy�
pothesis of my own. E. Ekwall (1928, 250) comments on the name Lent:
 The name may be derived from Welsh lliant,  a torrent, flood, stream .
Lliant looks like a participial formation from the root lei-, on which see
Leen. Sub this name he says (p. 248):  Leen may be identical with Leon
{Lion). [...] The earliest example is Lionhina 1040. [...] OE Lion [...] is a
derivative of the root lei-  to flow; water . [...] The base may be *liian- or
*liion- from *leian-, *leion-. Ekwall refers to Walde 1930-34:2, 392,
corresponding to Pokornż 1959, 664.4:  l%1łi-, giessen, fliessen, tr�pfeln,
Idg. *leio-, regnen, str�men. The senses noted here, which are relevant
for Ekwall s river-names, are  Flut, Meer, �berschwemmung, Meeres�
gott, Lear, str�mt. Lent is categorized by Ekwall as an originally Celtic
name with the sense of  river, stream . The stress is, according to Ekwall
(1928, lxiii), before 1100 likely to have fallen on the final syllable, Lli�nt.
Thus the stress in A eó c v t i may quite well have fallen on the a.
The commentary on Stora Le in Keyser 1846-95 sub A.D. 1268 reads:
er Lee mykit vatn,  Lee means a lot of water. Was the original sense still
current at this time, and is this the best etymology given for the word?
Thus, the name A e ó c v t i may be understood as  the violently flowing
rapids ,  the torrential rapids . The name is not likely to be onomatopoeic.
Ekwall thus regards lliant as a present participle. H. Krahe, on the
other hand, maintains (1951/52, 159) that there were no nY
-participles in
Celtic.  W�ren participiale �J-Flussnamen alt, so k�nnten sich gerade in
ihnen auch f�r das Keltische Reste verbaler �r-Bildungen erhalten haben.
[...] Das Keltische besitzt �berhaupt keine participial gebildeten Fluss�
namen. Although Krahe repeatedly quotes from Ekwall 1928 he seems to
have overlooked Ekwall s explanation. Nicolaisen (1957, 219) deals with
the chronology of these Celtic names:  These river-names must have been
around 500 B.C. Pokomy, writes Nicolaisen, is to be credited with indi�
cating the possibility of pre-Celtic but still Indo-European name elements
Scando-Slavica Tomus 49, 2003
8
Downloaded by
[Texas State University, San
Marcos] at 09:25
1
September 2013
The Names o f the Dnieper Rapids 49
in the British Isles. Krahe (1964, 78) indicates the geographical distribu�
tion of these the oldest of our names:  diese voreinzelsprachliche Hydro-
nomie umfasst nicht die Gesamtheit der indogermanischen Sprachen die
zugeh�rigen Gew�ssernamen sind von Skandinavien bis Unteritalien, von
Osteuropa, namentlich den baltischen L�ndern, bis zu den Britischen In�
seln und den K�sten des Atlantischen Ozeans verbreitet...  .
According to Pokorny (1959, 664.4) the root lei is found in several
European languages. He suggests the possibility that the name Lietuva
(Lithuania), and also Latvian LeTtis  Lithuanians , can be derived from
this root. OtrQbski (p. 117) provides yet another explanation, but from the
same root, for the much discussed name Lithuania. He regards the older
form Leituv� > Letuv� as a modification of *Leit�s, according to him a
region on the river *Leita or *Leite, forms he had himself reconstructed
on the root *lei-. OtrQbski (1911, 188) gives the present-day Lithuanian
rivername Leita, which he describes as  ein bedeutender Gewinn f�r die
Flussnamenforschung .
We have seen that there is evidence o f the root lei- in river-names in
the British Isles, Scandinavia, and the Baltic countries. This corresponds
to Krahe s view of the geographical distribution of the oldest river-names,
which according to him had been coined as early as 1500 B.C. Among
these he does not, however, mention names formed on the root lei-.
In Sweden these ancient Indo-European names are well known and
well covered by scholars, e.g. Boren by Hellquist (1903-06, 66), from IE
*bha  glitter .
BepouT^ri
Most scholars have seen Bepoux^ri as the present participle of vbreti  to
boil . In recent years Kleiber (1959; 1960), Shevelov (1955/1971) and
Danylenko (2001) have objected to the term  present participle . Dany-
lenko (2001, 55) calls it  an archaic denominative . The change of the
vowel from %to q is treated by Leskien (1922) in ż138. Forms such as
gorQste  gorqsta occur already in Old Church Slavonic texts. Vasmer
(1971, 872) deals with Russian �f-participles with the corresponding -ja-
w-vowels, such as Gremjaca, Gremucij in hydronymics. I will keep the
term  substantival present participle , a direct parallel to OSw Leanti.
Participles in Slavonic river-names are treated by Krahe (1951/52). To
be sure, he notes, they are far fewer than in the Germanic languages, but
they can still be established with confidence. On p. 154 he adduces
Scando-Slavica Tomus 49, 2003
8
Downloaded
by [Texas State University, San Marcos] at 09:25
1
September 2013
50 Elsa Melin
Veroutzi in DAI, used as a noun. Subsequently (p. 160f), Krahe points out:
 Es zeigt sich jedenfalls, dass wir mit mehreren Schichten participialer nt-
Flussnamen zu rechnen haben: einige (nicht sehr h�ufige) erweisen sich
auf Grund ihrer formalen Verh�ltnisse als Sch�pfungen vorhistorischer
Perioden, andere  und zwar die �berwiegende Mehrzahl  scheinen
j�ngeren Zeiten ihre Entstehung zu verdanken. Krahe also indicates
(p. 161)  die M�glichkeit ni-Flussnamen auf verbalen Grundlage in ver�
schiedenen Stadien der Sprachentwicklung immer neu zu erschaffen .
The discussion on this name has also concerned a possible rendering
by e o f the reduced vowel b in weak position. Falk (1951, 190) initially
explained e as a Greek svarabhakti vowel, serving to facilitate the pro�
nunciation o f the first part o f the word, as Skok (1928, 230) had done
earlier concerning the e in the fortress name �spouAAia. In his review of
Falk s work Zariba showed (1958, 167) that the sequence �p- occurs in
an initial position in Greek words; thus there was no need for a Greek
svarabhakti vowel. Initial vr- is uncommon in Greek. In ż245a Meyer
states that  vr- ist im Laufe der Zeit durch das allgemeine Schwinden des
v au f'p reduciert worden . In ż 179 he deals with a later development of
vr- from mr-, e.g. � paxug. Such developments are confirmed by etymo�
logical dictionaries. According to Greek grammarians the development of
svarabhakti, or anaptyctic, vowels is not clearcut or predictable. M ost
Greek grammars deal with this phenomenon. Initial br- has developed an
a in e.g. � a p d y x ia (Meyer 1896, ż94a; more examples in the following
żż). Curtius (1879, 729) writes:  Vor p erzeugt sich solches e im aeol.
UETEppos = UETpios; � Eppauos = TTpiauos. On papyri Gignac (1976,
311) writes:  Development o f a vowel between two consonants occurs
more frequently than elsewhere in Greek, probably because o f the col�
loquial nature o f the language o f the papyri ... (see the section on the
second rapids, page 39). Since the phenomenon was not realized consis�
tently, but was sporadic and limited to vernacular speech, �sp- could oc�
cur alongside �p-, and Zariba s objections are thus irrelevant.
In Vasmer 1941 we find u, i, and e as anaptyctic vowels where there
had never been a reduced Slavonic b. Vasmer (1941, 246) cites instances
involving p  C u piajjov = ZipsA, *ZepeAi = *zre/o-and A is
dealt with on p. 139: Z uA i�aiva, ZiAi�Eva = *Slivbna. Evidently there
was a long tradition of anaptyctic vowels developing in Greek, even from
the Classical period, and in my view we have the same anaptyctic vowel
in �spouT^n as in �spouAAia.
Scando-Slavicci Tomus 49, 2003
8
y
Downloaded
b
[Texas State University, San
Marcos] at 09:25
1
September 2013
The Names o f the Dnieper Rapids 51
The literal sense o f ��Ź�źą ���ż�,  the boiling o f water , can be un�
derstood in three ways: 1. The heating o f water. 2. The seething and bub�
bling of water. 3. The movement of water when boiling.
I regard the first interpretation as nonsensical for the rapids, and con�
sequently also Miller s (1885), Struminski s (1996) and Danylenko s
(2001) derivation o f the name o f these rapids from the root *hlyja. No. 2
would be onomatopoeic, but I fail to hear the similarity. To establish the
signification of the verbal noun ��Ź�źą I turn to the corresponding verb,
��Ź � � � , according to Liddell and Scott (1925-40, s.v.)  to shake vio�
lently (of the sea) . In Pape (1914, s.v.) there is a participle of this verb
referring to the sea, signifying  in heftige Bewegung gesetzt . This evi�
dence should be sufficient proof that ��Ź�źą primarily refers to the vio�
lent movement o f the rapids.
The Seventh Rapids  Ł��ż��ż�� � ąĄ��ś� (Explanation:
źą���� Ć �ąłź��)
Ł��ż��ż��
While the signification of the name,  current, small rapids , was recog�
nized by Thomsen (1919, 312f), its form constitutes a problem that has
not been decisively solved.
Falk cites (1951, 207-10) a large number of instances o f struk, mainly
from northern Sweden. There is also one instance o f stryk, in the sense of
 current, small rapids . He regards the name as the dative plural serving as
a nominative. To this explanation Sahlgren objected several times in
Namn och Bygd, especially in 1951, 159. Ekblom (1951, 154) and Obol�
ensky (1962, 56) regard the form as a dative plural following a prepo�
sition, at Strukum. Sahlgren, on the other hand, believes the nominative to
be strukn, a word which, according to Ekblom (1951, 154) and Karsten
(1933, 39ff), is nowhere to be found. Th. Andersson discussed this type of
place-name in an exhaustive article (1980); he established that the forma�
tion o f place-names with e.g. the suffix -n is not likely to have been
prevalent in the Viking Age. In rare cases words or suffixations may be
reconstructed for chronological reasons. Such a reconstruction is the
Dnieper rapids name strukn,  passage, current, small rapids  all accord�
ing to Andersson. This name has thus not been found, only reconstructed,
as Andersson himself admits (p. 19).
Scando-Siavica Tomus 49, 2003
8
Downloaded by
[Texas State University, San
Marcos] at 09:25
1
September 2013
52 Elsa Meliti
It is my intention to prove that what we have in this name is neither
strukn nor a dative plural o f struk.
In his book (1960) on plurals in the earlier Scandinavian naming sys�
tem Lars Hellberg demonstrated that singular names for water-courses
frequently form the basis for plural names for settlements along these
water-courses. The plural Nora stands on a singular nor,  narrow straits
(1960, 73). Am, a dative plural, stands on a single river (a) (1960,72). In
the same way those place-names in the dative plural that Falk cites are
more likely to represent  a number of houses on a river than  a number of
rivers .
For obvious reasons we are not concerned with the names o f the set�
tlements along the Dnieper. I believe that Ł��ż��ż�� contains the nomi�
nativ singular stryk and the Greek ending -ouv. The root might vary in
this word strak, streke, struk, or stryk. To be sure, Hellquist (1939, s.v.)
regards stryk as a fairly modem formation, but other philologists express
the opposite opinion. Torp (1919) relates stryk, defined as  strong current
in a water-course , to Old Norwegian/Icelandic strykr. Lars Hulden
(1960, 16) believes stroke, from Osterbotten, to be an extension of Old
West Norse stryk,  main channel .
In Swedish dialects, and outside Sweden as well, stryk, masc. and
neutr., has a wider distribution than struk. From the archives of the Dia�
lect Institute at Uppsala (Sprak- och folkminnesinstitutet; personal com�
munication from L. Ryman) it appears that stryk and its variants are
known in the provinces of Norrbotten, Vasterbotten, Dalama, Varmland
and Orebro, in Osterbotten in Finland and in the Nucko dialect in Estonia.
Through these provinces led itineraries for levies, and along these the
Vikings may have acquired stryk as a word for small rapids. SAOB as
well (s.v.) cites stryk n. for  the part of the main channel where the current
is strongest. However, as pointed out by Falk and Sahlgren, no instances
of stryk, struk have been found in such central parts of Sweden as Up-
pland and Sodermanland.
The neutral ending -ouv is a variant of -ov, which is extremely com�
mon in Greek place-names. The same change o f o into ou occurs as in
� ż �  � ąĄ � Ź � , in this case after a guttural k in post-tonic position. In
Vasmer (1941, 90) we find such variants as
Ź��ż� and
Ź��ż�, and
final v could be pronounced or left out in different dialects.
It is obvious that the interpretation  źą���� Ć�ąłź�� goes very well
with Stryk, having the same sense and the same number.
Scando-Slavicci Tomus 49, 2003
8
y
Downloaded
b
[Texas State University, San
Marcos] at 09:25
1
September 2013
The Names o f the Dnieper Rapids 53
Falk (1951) justifies his interpretation o fS T p o u K o u v on p. 214:
 From the Slavonic correspondent, Nastre~
je, it is evident that the Norse
name refers not merely to these rapids but also to the surroundings.
Nastre~
je is formed on streib  struka , and its meaning is  the area around
the rapids . Thus, the Slavonic name induced Falk to add  the area
around to the sense of the Norse name as well. The discussion will there�
fore be continued sub the Slavonic name of the rapids.
N a T rp e C j�  N a a T p E � ii
A number of unsuccessful attempts at explaining this name have been
made, see Falk 1951, 192-206. Falk suggested o t for i t , since these let�
ters were easily confused. With Obolensky (1962, 51) I find his conjec�
ture sound. Synonymity is achieved, the root *strbzb having the sense of
 the rapid current of a river . Moreover, Sreznevskij (1893-1903, s.v.
Strbibn-b) has found Stre~
ent for A.D. 1393. In the Dnieper basin, in the
lower reaches of the tributary Desna the Ukrainian form Strilenb is ad�
duced twice, and we also find StY
i~
na (Korepanova 1969, 47). Obolensky,
however, comments in summarizing (1962,51):  the exact relationship
between NaoTpe�ri and stY
e~
� may remain controversial . According to
Falk (1951, 221) the formation of *Nastre~
je is quite straightforward; it is
a common type of word in the Slavonic languages, like other formations
involving prepositions, e.g. Pomorbje. Among other references Falk cites
the work by Karas (1955) on place-names formed on prepositions. On p. 7
Karas gives the example Nawsie, from na wsi.
It is my intention to demonstrate that nouns formed on na- and a col�
lective noun ending in -bje are not always based on a prepositional phrase.
Miklosich (1927) treats both unprefixed collectives (p. 207) and na-
prefixed place-names. We find e.g. kY
� and n�kY
�  copse (p. 273), dvorje
and n�dvoY
�  farm (p. 241), lŁse and n�les�  forest (p. 304).  Das Suffix
ije bildet Collectiva, he writes, but does not comment on the na-prefixed
names. The formation is Primitive Slavonic. In a minor group o f Russian,
Ukrainian, Polish and Czech appellatives, all containing the prefix na-, I
have found a collective formation with no connotation of  place for (all
from D aF 1903 s.v.): napóTe  obsirnye polja; n�boloe
e  bolotistoe
mes to', nag�r'e  nagornaja polosa, storona. Corresponding words are
found in Ukrainian and Polish.
Only a very few such appellatives have been preserved. A particular
type of formation may however be productive in place-names long after it
Scando-Slavica Tomus 49, 2003
8
Downloaded by
[Texas State University, San
Marcos] at 09:25
1
September 2013
54 Elsa Melin
has ceased to function in appellatives. This is the reason why there are
many more place-names than appellatives of this origin.
The -io- suffix is Indo-European. It was used i.a. for diminutives, e.g.
"ą���ąż�,  manikin . In the Germanic languages it appears as -ia, which
in rare cases had a diminutive signification. Wess�n notes (1943, 2 Iff)
that the suffix had long since lost its productivity even in the Germanic
languages. It was used i.a. to form collectives, but in compounds the con�
notation o f collectivity was often undetectable. These had earlier been
indicated by the prefixes ga- or ge-; thus the German correspondent of
nagorbje is Gebirge. In Scandinavian words a diminutive and collective
connotation can be found in e.g. aspe,  aspen grove, small aspen wood .
In the Slavonic languages I have found a diminutive sense in Czech only,
and only in two words, n�vra
� and the obsolete n�hon, both meaning  pa�
horek ,  hill (PY
�ru%0ńn� slovn�k jazyka %0ńesk�ho). In these words there is no
collective connotation, as is quite clear from the quotation from B.
N%1łmcov� (ibid.)  Vid%1łly z�mek na mal�m n�vra
f ,  They saw a castle on
a small hill . In most Slavonic languages this diminutive function can be
found in verbs and adjectives, and in adverbs with the prefix na-. In ex�
actly the same manner as in n�vra
� it affected � ą ����ś�, as is evident
from the explanation źą���� Ć�ąłź��. The name is well suited for
Lia
nyj, a small rapids with a fall o f merely 0.4 metre (Obolensky 1962,
52).
I have suggested fresh interpretations for the following names:
The name of the first rapids  (�) ���ż�Ą�  I derive from the
Old Church Slavonic/Old Russian present participle neusypaj^i, neusy-
pajaj,  watching . This Slavonic word may be a translation o f the Turkish
name for the fortress and the rapids Kodak < Kydas%0ńi, an Uigur word for
 the watcher ,  the guard .
The names of the fourth rapids  ą �ąĆ ��  � �ą���  I interpret
respectively as  the rapids that are always shallow and  the rapids that are
never sufficiently filled with water .
The names of the fifth rapids  �ą�ż�Ć��ż�  �ż���Ą�Ź� sig�
nify in my interpretation  the rapids that are bare, free  free that is from
sandbanks, islets and other obstacles to the passage of the boats.
The sixth pair of names is � �Ź��ą  ���ż��ś�. I link the present
participle � �Ź ��ą with the IE root *lei-, which occurs also in the Celtic
present participle lliant,  stream . This root also appears in the Lithuanian
river-name L�ita, the origin of the name Lithuania, ���ż��ś� is the pre�
Scando-Slavica Tomus 49, 2003
8
Downloaded
by [Texas State University, San Marcos] at 09:25
1
September 2013
The Names o f the Dnieper Rapids 55
sent participle of * vbreti, literally  the boiling . Since ( 3 p d a n a V E p o u
refers to the violent movement of water I find it reasonable to attribute the
sense of  violently rushing to both of these names. With H. Krahe
(1951/52, 160f) I maintain that Slavonic river-names could be formed on
present participles in several periods and with varying forms.
Except for -Ttpax, Neccotit and  Asupop all names in the nominative
case ending in a consonant have been given Greek endings: -1, -os, -ouv.
In consequence, S t p o u k o u v cannot be a dative plural, nor does it reflect
a reconstructed  *strukn . Regarding the word-formation I have demon�
strated that NaTrpeŁr|  NaaxpE^ri is a collective formation with a
diminutive meaning but with no connotation of  the area around .
The Reduced Vowels
Since the names of the Dnieper rapids were recorded very early they may
be useful for throwing light on the phonetic development of the reduced
vowels. Recently, however, Danylenko (2001,49) has expressed his be�
lief that nothing definite can be said about final reduced vowels in the
names. These vowels are always missing in final position, once in
Neocotit and twice in --rrp&x. In the latter cases the absence of a vowel is
a prerequisite for the Greek development of y into x , as Tolkacev says
(1962, 57). Despite this he claims (ibid.) that at the time in question the
reduced vowels had not yet been dropped. In medial position b is missing
not only in |3ouAvr|TTpdx but also in NccaTpE^fj, and in DAI 9/6 in Te-
A i o u t ^ ccv, where -bb- has fallen. All these losses occur in weak positions.
In strong position we have e in NaaTpE^fj and in T^Epviycbyav (DAI
9/6), in my opinion also an 1 in MiAiviaKccv, from e < b. Their renderings
in PEpouT^r) and  OcrrpoPouviTTpax are incorrect; e and ou are incon�
sistent with other spellings and Slavonic phonetic development. As I have
attempted to explain sub these names they are signals merely of scribal
tradition and Greek phonetic development. Shevelov (1955/1971, 152)
justifies the ou-spelling in  O oT poP ouvnrpax with the large number of
cases of -b for b in adjectives in Church Slavonic texts, but in these texts
ou-spellings are missing completely. He further rejects the analogy e for b
in pEpouT^ri and (3EpouXXia, on the grounds that the latter name had
been fetched from a southern Slav region which had for a long time been
under Byzantine administration and in which the scribal traditions were
Byzantine.  Of course, in the region of the Dnepr falls there were neither
Byzantine administrators nor any Romance population with the old Byz�
Scando-Slavica Tomus 49, 2003
8
Downloaded
by [Texas State University, San Marcos] at 09:25
1
September 2013
56 ElsaMelin
antine cultural tradition (Shevelov 1955/1971, 154). But Constantine
composed his works in Constantinople, subjected to the rules of the
Graeco-Byzantine scribal tradition. Horrocks (1997, 162) describes
Constantine s language in these works:  They range stylistically from a
near academic register, with influences from the language of imperial
administration, to a mere basic variant which owes much to the conven�
tions of the chronographic tradition.
I am quite convinced that the Greeks did not perceive the Slavonic
weak reduced vowels. To be sure, the name corpus is tiny, but the syn�
copes are consistent and the inaccurate substitutes in accordance with
Greek scribal tradition. It seems unlikely that these vowels, which were
not written down by the Greeks, were pronounced by the Slavs.
The Greek Explanation o f the N am es o f Rapids 2, 3 and 6
The explanations  �� ���Żż� �ż� Ć�ąłźż� (the second rapids), ��ż�
Ć �ąłźż� (the third rapids), and ��Ź�źą ���ż� (the sixth rapids), con�
sist of two nouns, of which the second is in the genitive. Lehrberg (1816)
pointed out that the word-order should have been the reverse  Ć � ą ł �
ź�� �ż� ���ąż� (p. 357) and Ć�ąłź��  � � � � (p. 361). No scholar has
made an attempt to explain this word-order. Since it is repeated it is not
likely to have been an accidental mistranslation. I will submit an explana�
tion to the discrimination of experts.
Ossetic, the sole survivor of the Northern Iranian languages, contains
archaic remnants in which headword and attribute have switched places,
in comparison with other Indo-European languages. The ancient name for
Kerc, Pantikapaion, is an Ossetic compound with the sense of  road-fish ,
while we would have expected  fish-road , the road to the fish of the Sea
o f Azov. Don, Dnieper and Dniester are all, according to G. Schramm
(2001, 9f), Ossetic names containing postposed adjectives meaning  wide
and  deep , added to the word for  river .
In the Russian chronicles the Ossetes are called jasy . Other Iranian
peoples mentioned are the Antics and the Alans, who may have mixed
with the Slavs. Vernadsky claims (1943, 159) that  the Antic ruling clan
must have been o f Iranian origin . From the 6th and 7th centuries the
Antics were settled around Kiev. On p. 347 Vernadsky points out that  in
some border districts a mixture o f Slavic and Ossetian may have been
used . Possibly we have in  �� ���Żż� �ż� Ć�ąłźż� a case o f Iranian
word-order, attested in names from Slav territories.
Scando-Slavica Tomus 49, 2003
8
Downloaded
by [Texas State University, San Marcos] at 09:25
1
September 2013
The Names o f the Dnieper Rapids 57
What Greek Dialect Is Represented in the Names o f the Rapids?
In Melin 2003 I pointed out the Pontic Greek features in Teąio�t�ccv,
MiAivioKcxv and BouaeypaS�. In the names of the rapids as well there
are features which can be identified as Pontic, being particularly frequent
in this Greek dialect. These are anaptyctic vowels ( e in BEpo�x�ri, ou in
Bapoucp�po�,  O aTpoPouviTrp�x), and loss o fv in e u [N]Eoaourrfi.
The shift of o to ou in BouXvr)TTpax and Sxpo�Kouv is usually ex�
plained as a Northern Greek phonetic development. This weakening,
however, occurred in Pontic as well, although it was here more variable
(see Horrocks 1997, 313). In Pontic, loss of the vowel took place in the
syllable immediately after the stress. Thus, I find it unlikely that bn has
fallen before the stress in (3ouĄvr|TTpąx. In my view the number of Pon�
tic vernacular forms in this chapter are the best witnesses to the Greek
dialect spoken by the informants. Most Greek cities on the northern coast
were founded by colonists from Asia Minor. One of these early Pontic
cities was Olbia, which controlled the commerce between Kiev and Byz�
antium. Later on Varangians settled in Olesie, which was to assume the
function that Olbia used to have (Vernadsky 1943, 56ff). Even today
Pontic lives on, e.g. in a region south of Rostov on the Don.
What Language Produced the Original Names?
An issue long discussed is whether the original names were Slavonic or
Norse. Falk (1951, 39f) believed the Slavonic names to be direct transla�
tions of the Old Swedish names. Tolkacev (1962, 60) is the most recent
scholar to hold the opposite view. In my opinion the frequent appearance
in this region of the name Vol'nyj for rivers and rapids is evidence of the
Slavonic origin of the names. That Barfors was the origin of all these
names is simply out of the question.
Who Was the Informant?
It is quite unlikely that the Emperor Constantine himself communicated or
explained the names. He spoke no Pontic, no Iranian language nor, in all
probability, no Slavonic language. In several contexts it is claimed that he
knew some sort of Slavonic language, since his grandfather was a Slav.
This allegation is firmly rejected by Bury (1906, 542, fn. 2). Constantine s
Scando-Slavicci Tomus 49, 2003
8
Downloaded
by [Texas State University, San Marcos] at 09:25
1
September 2013
58 Elsa Melin
father, Basileios, is described by Horrocks (1997, 137) as  an illiterate
peasant of Armenian descent .
Following the same linguistic argument it is unlikely that a Varangian
would have been able to report the full range o f the names. Several schol�
ars have believed that a southern Slav a Bulgar, a Serb, or a Macedo�
nian acting as an interpreter would have communicated the Slavonic
names, with their thrice repeated metathetic forms. This view as well
seems improbable to me. A fairly long residence in a dialect-speaking
milieu is needed for someone to start using dialect forms. Nor is it likely
that a Pontic Greek would have reported all these names he had no rea�
son to be familiar with the Slavonic bible. The most probably informant
would be a Slav, an Alan or an Ossete in a border territory where Old
Russian, Pontic Greek or an Iranian language were spoken. This infor�
mant must have been familiar with the Church Slavonic Psalter, the
source of NeaariT in the sense of  pelican in the name of the fifth rapids,
and also of the Church Slavonic metathetic forms.
Objections have been raised to the presence of Church Slavonic forms,
most recently by Tolkacev (1962), who claims that at this date Church
Slavonic had not yet gained a foothold among the eastern Slavs. But his
view is not supported by the information provided by Vernadsky
(1943, 369) and Horrocks (1997, 139). Shapiro claims (1982, 169):  Any
person literate in Slavonic in the tenth century had necessarily to be fa�
miliar with the Bible, the Psalter in particular. The most ancient Chris�
tian tradition prevailed among the Pontic Greeks. Among the Slavs Chris�
tianity had been spreading from A.D. 850.  Between the mid-ninth and
mid-eleventh centuries Byzantine culture, spearheaded by Christian mis�
sions that brought with them religion, law, art and literature, was spread
as far as the Baltic (Horrocks 1997, 138).
References
A n d ersson , Th. 1972.  H �gb ystenen s runin skrift . S tu dier i n ordisk filo lo g i
58:1 7 -4 8 .
----------. 1980.  O rd eller suffixbildn in g? T ill fr�gan om rekonstruktion av ord ur
ortnamn . Norna-rapporter 18, 9-40. Uppsala.
B rate , E rik . 1911. � sterg � tla n d s ru n in skrifter (= S v e rig es runinskrifter 2).
Stockholm : K ungl. Vitterhets historie och antikvitets akadem ien.
Brockelm ann, C. 1921.  M ahm ud al K asgh arl �ber die Sprachen und die St�m m e
der T�rken im 11. Jahrh. K �r� si Csom a-Archivum 1 ,2 6 -4 0 . Budapest.
B u ry, J. 1906.  The T reatise D e a d m in istra n d o im p erio  . B yzantinische Z eit�
schrift 15, 5 1 7 -7 7 . Leipzig.
Scando-SIavica Tomus 49, 2003
8
Downloaded
by [Texas State University, San Marcos] at 09:25
1
September 2013
The Names o f the Dnieper Rapids 59
Constantine Porphyrogenitus. 1949-62. D e adm inistrando im perio. V ol. 1. Greek
text edited by G y. M oravcsik . E n glish translation b y R . J. H . Jen kins. B u d a �
pest.
--------- . 1962. D e a dm in istran do im p erio . V ol. 2: C om m entary by F . Dvornik,
R . J. H . Jen k in s, B . L ew is, G y. M o ravcsik , D. O bolen sky, S. Runcim an.
London.
Curtius, G. 1879. G riindziige d e r griechischen E tym ologie. 5. A uflage. Leipzig.
D a l', VI. 1903. Tolkovyj slo va r' zivogo velikorusskogo ja zyk a, 1-4. M oscow .
D anylenko, A. 2001.  The N am es o f the D nieper R apids in Constantine Porphy�
rogenitus R evisited . D ie W elt d er Slaven 46, 4 3-6 2.
D ieterich, K . 1898. U ntersuchungen zur G eschichte der griechischen Sprache (=
Byzantinisches Archiv 1). Leipzig.
E k b lom , R . 1951.  D ie N am en der siebenten D neprstrom sch nelle . S pr�k-
ve ten sk a p lig a s� llsk a p ets i U ppsala f� rh a n d lin g a r 1 9 4 9 -1 9 5 1 . U p p sala ,
W iesbaden.
Ekvvall, E. 1928. English R iver-N am es. Oxford.
E vam ick ij, D. I. 1890. V ol'nosti zaporozskich kozakov. St. Petersburg.
F alk , K .-O . 1951. D n ep rfo rsa rn a s nam n i k ejsa r K on sta n tin VII P orfyro-
gennetos  D e adm inistrando im perio. Lund.
G ign ac, Fr.Th. 1976. A G ram m ar o f the G reek P a p yri o f the Rom an an d B yzanti�
ne P eriods. Vol. 1. Phonology. M ilano.
H afstr�m , G . 1949. Ledung och m arklandsindelning. U ppsala.
H ellb erg, L . 1950. Inbyggarnam n p � -karlar i svenska ortnam n. U p p sala and
Copenhagen.
--------- . 1960. Pluralform i �ldre n ordiskt ortnamnsskick. U ppsala.
H ellquist, E. 1903-06. Studier o ver de svenska sj�nam nen, d era s h�rledning och
historia. ( - Svenska landsm �l 20). Stockholm .
--------- . 1939. Svensk etym ologisk ordbok2. Lund.
H orrocks, G . 1997. G reek: A H isto ry o f the L anguage an d its Speakers. London
and N ew York.
H ovda, P. 1966. N orske elvenamn. O slo - Bergen.
H rusevskyj, M . 1906. G eschichte des ukrainischen Volkes 1. Leipzig.
Hulden, L . 1960.  Ortnamnen i M un sala . S tudier i nordisk filo lo g i 4 1:4 , 1 -160.
U ppsala.
Jannaris, A. N. 1897. An H istorical G reek G ram m ar. London.
Jon sso n , H. 1966. N ordiska o rd f � r vatten sam lin g (= Lu n d astu dier i nordisk
spr�kvetenskap 16). Lund.
K a ra s, M . 1955. N azw y m iejsco w e typu  P odgo ra , Z a la s w j^ zyku polskim i
innych jqzykach stowianskich. W roclaw.
K arlgren , A . 1947. D n eprfossern es N ordisk-Slaviske N am ne. Copenhagen. Fest-
skrift udgivet a f K 0benhavns U niversitet i anledning a f U niversitetets a ars�
fest.
Scando-Slavica Tomus 49, 2003
8
y
Downloaded
b
[Texas State University, San
Marcos] at 09:25
1
September 2013
60 Elsa Melin
K arsten , T . E . 1933.  N o rd isk a p articip ia p resen tis so m �lv- och forsn am n  .
N amn och B ygd 21, 3 9-44.
K eyser, R . (ed.) 1 846-95. N orges gam le L ove in dtil 13 8 7 1-5. Christiania: G r0 n-
dahl.
K leiber, B . 1959.  Zu den slavischen Nam en der Dneprschnellen . Z eitsch rift f� r
slavisch e P hilologie 28, 7 5 -9 8 .
----------. 1960.  N ochm als zu den slavischen N am en der D neprschnellen . Z e it�
schrift ftir slavisch e P hilologie 29, 125-27.
K nudsen, G ., R . H ornby, and M . K ristensen. 1936-64. D anm arks gam le Person-
tiavne 1-2. Copenhagen: G. E . C . G ads Forlag.
K o c k , A . 1899.  E ty m ologisch -m yth ologisch e U ntersuchungen . In d o g erm a �
nische Forschungen 10, 9 0 -1 1 1 . Strassburg.
K orep an ova, A . P. 1969. Slov o tv or%0ńi typy hidronim iv basejn u n y~
n 'oi D esni.
K iev: N aukova dumka.
K ou sg�rd -S0rensen, J. 1968-1978. D anske s0- og �navne 1-8. Copenhagen.
K r�he, H. 1951/52.  A lteurop�ische Flussnam en  . B eitr�g e zu r N am enforschung
3, 153-70, 225 -4 3 .
--------- . 1964. U nsere �lteste Flitssnamen. W iesbaden.
Lehrberg, A , C . 1816. U ntersuchungen zu r E rl�uterung d e r �lteren G eschichte
R u ssla n d s 5, B esch reibu n g d e s unteren D n eprs und sein e W asserf�lle. St.
Petersburg.
L esk ien , A . 1922. H andbuch d e r altbu lgarisch en (altkirchen slavischen) S p ra �
che6. H eidelberg: C arl W inters U niversit�tsbuchhandlung.
Liddell, H. G. and R . Scott. 1925-40. A G reek-English Lexicon. Oxford.
M ackenzie, W. C . 1931. Scottish Place-N am es. London.
M elin, E lsa. 2003.   Sam b at�s and C ity N am es in Ch. 9 o f Constantine Porphy-
rogenitus D e adm inistrando im perio , D ie W elt d e r Slaw en 48, 187-92.
M eyer, G . 1896. G riechische G ram m atik. Leipzig.
M ey er, K . H. 1935. A ltk irch en sla visch -g riech isch es W �rterbuch d es C odex
Suprasliensis. G l�ckstadt and Ham burg.
M iklosich , F. 1927. D ie B ildung d e r slavischen Personen - und O rtsnam en: drei
A b h a n d lu n g e n (= Sam m lu n g slavisch er Lehr- und H andb�cher. R eihe 3,
T exte und U ntersuchungen 5). H eidelberg. (Q uotations are from D ie s la �
vischen O rtsnamen aus A pellativen.)
M iller, V . 1885. N azvanija dneprovskich p o ro g o v u Konstantina B agrjanorodno-
g o . (= D revnosti. Trudy M o sk . A rch. O b a
%0ńestv a 1. S u pp lem en t 1 9 -3 1 ).
M oscow .
M oravcsik, Gy. 1 9 4 1 ,1 9 4 3 . B yzantinotiircica 1 ,2 . Budapest.
N ico laise n , W . 1957.  D ie alteurop�eisch en G ew �ssern am en der britischen
H auptinsel . B eitr�ge zur Nam enforschung 8 ,2 0 9 -6 8 .
N iederle, L . 1902-1924. Slovansk� staro~
itn osti 1 Praha.
4.
N or�en, A . 1909.  Ein Paar altnordische Seenam en . In dogerm anisch e F orsch �
ungen 26, 2 2 2 -2 4 .
Scando-SIavica Tomus 49, 2003
8
Downloaded
by [Texas State University, San Marcos] at 09:25
1
September 2013
The Names o f the Dnieper Rapids 61
--------- . 1970. A ltnordische G ram m atik 1. 5. A uflage. T�bingen.
O bolensky, D. 1962. See Constantine Porphyrogenitiis, 1962.
O ekonom ides, D. E. 1908. Lautlehre des Pontischen. Leipzig.
Ortnamnen Skaraborg. 1950-81. Ortnam nen i Skaraborgs l�n 1-18. U ppsala.
Ortnamnen V�rm land. 1922-62. O rtnamnen i V�rm lands l�n. U ppsala.
O rtnamnen V �sternorrland. 1955. O rtnam nen i V �sternorrlan ds l�n 1. U p p sala
and Copenhagen.
Otr�bski, H. 1911.  Lietuvą . B eitr�ge zur N amenforschung 9, 116-18, 188.
Pape, N. 1914. G riechisch-deutsches H andw�rterbuch. Braunschw eig.
Pipping, H. 1911.  D e skan dinaviska D neprnam nen . S tu dier i nordisk fd o lo g i
2:5, 1-31. Helsinki.
--------- . 1936.  Fvn. Fżri, fsv. F 0re - och nsv. F�ret , N amn och B ygd 2 4 ,6 6 - 7 5 .
Pokornż, J. 1959. Indogerm anisches etym ologisch es W �rterbuch 1. Berne - M u�
nich.
P Y
�ru %0ńn � slovnik ja zyk a %0ńesk�ho. 1935-. Prague.
R ad lo ff, V . V . 1 8 9 3 -1 9 1 1 . V ersuch ein es W �rterbu ch es d e r Tiirkdialekte. St
Petersburg.
Ravndal, G. B . 1938. Stories o f the E ast-Vikings. M inneapolis.
R ehder, P. 1968.   N e a a r |T  1@>9B^3) � Z um Problem der Entnasalisierung
der slavischen N asalvokale in de A dm inistrando Im perio . D ie W elt d er S la �
ven 13, 137-55.
R ygh, O . 1904. N orske Elvenam ne. K ristiania.
Sachm atov, A. A. 1916. Vvedenie v kurs istorii russkogo ja zy k a , %0ń. 1. Petrograd.
Sah lgren , J. 1930.  V ikin gar i �sterled. N �g o t om D n eprfo rsarn as sv en sk a
namn . Namn och B ygd 18, 135-48.
----------. 1939.  N ordiska ortnamn i spr�klig och sak lig belysning. 13. F�lebro,
Fardhem och Fau r�s . Namn och B ygd 2 7 ,1 4 5 - 6 3 .
--------- . 1950a.  D neprforsam as svenska namn . Namn och B yg d 38, 138-146.
----------. 1950b.  M era om D neprforsarnas sven ska nam n . Nam n och B ygd 38,
159-69.
---------- . 1951.  D neprforsarna. G en m �le tili gen m �le . N am n och B yg d 39,
153-60.
----------. 1952.  L e s nom s su�dois des torrents du D nŁpre chez Constantin Por-
phyrog�nŁte . O nom ata 1 ,4 -7 .
S A O B . 1898-. Svenska A kadem iens ordbok. Lund.
Schram m , G . 2001.  Ein erstarrtes K onzept der Flussnam enphilologie: A lteuro�
pa . Namn och B ygd 89, 5 -1 8 .
Schw yzer, E. 1953. G riechische G ram m atik 1. 2. A uflage. M�nchen.
Shapiro, M . 1982.  Slavon ic *n ej�sy tb  pelican  : The Perpetuation o f a Septua-
gintal Solecism  . The Slavonic a n d E ast E uropean R eview 60, 161-71.
Shevelov, G .Y . 1955/1971.  On the Slav ic N am es for the F alls o f the D niepr in
D e a dm in istran do im perio o f Constantine Porphyrogenitus , in his T ea sers
Scando-Slavica Tomits 49, 2003
8
y
Downloaded
b
[Texas State University, San
Marcos] at 09:25
1
September 2013
62 Elsa Melin
an d A ppeasers. E ssays a n d Studies on Themes o f Slavic P h ilo log y (= Forum
Slavicum 32), 143-64. (O riginally published in W ord 11, 1955).
Sk ok , P. 1928.  O rtsnam enstudien zu D e adm in istran do im perio des K aisers
C onstantin Porph yrogennetos . Z e itsc h rift f � r O rtsn a m en sfo rsch u n g 4 ,
213
44.
S lova r' russkogo ja zyk a X I-X V II vv. 1975-. M oscow .
Slovn�k ja zyk a staroslov%1łnsk�ho. 1966. Prague.
Srezn evsk ij, I. I. 1 8 9 3 -1 9 0 3 [1 958]. M a te ria ly d lja slo va rja drevn eru ssk ogo
ja z y k a 1-3. St. Petersburg: T ip o grafija im peratorskoj A kadem ij nauk. R e �
print: M oscow : G os. izd-vo inostrannych i nacional'nych slovarej,
Strum iH
ski, B . 1996. L inguistic In terrelations in E arly R us'. Edm onton and T o �
ronto.
Thom sen, W. 1919.  D et russiske riges grundlaeggelse ved N ordbom e . Sam lede
A fhandlinger, F0rste Bind, 231
414. Copenhagen and K ristiania.
Thum b, A. 1910. Handbuch d er neugriechischen Volkssprache. Strassburg.
T o lk a%0ńev, A. I. 1962.  O nazvanii dneprovskich porogov v so%0ńinenii K onstantina
B agrjan o ro d n ogo D e a dm in istran do im perio". Isto ri%0ńesk a ja gram m atika i
leksikologija russkogo ja zyk a , 2 9-7 5. M oskva.
T orb j�m sson , T. 1911.  Den sjunde D neprforsen . S tudier i N ordisk F ilologi 2:6,
1- 10.
Torp, A . 1919. N ynorsk etym ologisk ordbok. Kristiania.
V asm er, M . 1941. D ie Slaven in G riech en lan d (= A bhandlungen der Preus-
sischen A kadem ie der W issenschaften, Phil.-Hist. K lasse N r 12). Berlin.
---------- . 1971.  Zu den N am en der Strom schnellen des D nepr bei Konstantin
Porphyrogenitus . Schriften zu r slavischen A ltertum skunde und Nam enkunde
2, 8 72-7 4. Berlin.
V ernadsky, G . 1943. A H istory o f Russia. V ol. 1. A ncient Russia. N ew Haven.
W ess�n, E.. 1943. Svensk spr�khistoria. V ol. 2. O rdbildning. Stockholm .
Z a rib a, A. 1958.  K .-O . Falk, D neprforsam as namn i k ejsar Konstantin V II Por-
fy ro g e n n e to s D e a d m in istra n d o im p erio , Lu nd 1951 . O n o m a s tic a 4,
153-69.
Translated by B. Ellenberger
Elsa Melin
Anslagsv�gen 1
S-226 47 Lund Sweden
Scando-Slavica Tomus 49, 2003
8
Downloaded
by [Texas State University, San Marcos] at 09:25
1
September 2013