DIE SPRACHE - Zeitschrift fur Sprachwissenschaft, 39 (1997) 2

1m Auftrag der Wiener Sprachgesellschaft herausgegeben von Heiner Eichner

unter Mitwirkung von Hans Christian Luschutzky, Robert Nedoma, Oskar
E. Pfeiffer, Klaus T. Schmidt, Chlodwig H. Werba und unter redaktioneller Mitar­
beit von Melanie Malzahn. D. REZENSIONEN
Anschrift: Institut fur Sprachwissenschaft der Universitat Wien, Luegerring 1,
A-lOW Wien, Osterreich Vaclav B I a z e k : Numerals. Comparative-etymological analyses of numeral
systems and their implications.Brno, Masarykova univerzita, 1999, 339 p.,
paperback (ISBN 80-210-2070-9).

Inhalt
A. Aufsatzs
I. Balles, Reduktionsformen in langen Wortformen als Ursprung mor­
phologischer Doppelformen
R. N edoma, Die althochdeutsche Fassung des Ersten Bcseler Rezepts
B. Kurzaufsatzo
D. Schurr, Lydisches IV: Zur Grammatik der Inschrift Nr. 22 . . . ..
D. Stifter, Celtiberian -unei, Luguei . . . . . . . . . . . . . . . . . . ..
The present monograph collects nineteen essays on the numerical systems
of several African and Eurasian languages; most of these essays have already
been published as separate studies in various journals of linguistics. The articles
141 have been gathered in the present book which serves as the author's 'Ha­
168
bilitationsschrift' at Brno university. B.divides his subject into three parts: A. Non­
Indo-European numeral systems (viz. the Saharan numerals, the Nubian numerals,
201 the Egyptian numerals, the Berber numerals, the Kartvelian numerals, the Uralic
213

numerals, and the Altaic numerals, p. 1-140), B. the Indo-European numerals (p.
C. Rezensionsaufsatz 141-324), and C. Patterns of creating numerals (p. 325-337). The Indo-European
Kritische Bemerkungen zu den Editionen der Tocharischen Sprachreste family thus forms the central part of the book, and its numerals are discussed in
(K.T.Schmidt) 224 separate subchapters dealing with 'I', '2', '3', '4', '5', '6', '7', '8', '9', '10', '100' and
D. Rezensionen . . . . . . . . . . . . . . ., .. 239 '1000'.
Register zu Band 39 (H. D. Pohl) . . . . . . . . . . . . . .. 247 After every chapter and subchapter of the book, we find a separate
bibliography of the works cited in that chapter. Since many of the works cited are
mentioned in more than one (sub)chapter, the book could have numbered a few
pages less if all the references had simply been given once, at the beginning or at
the end of the book. The present repetition of references is probably due to the
fact that many of these (sub)chapters have appeared in print before, and the reader
Aile redaktionelle Korrespondenz, Manuskripte und Bucher sind an den Heraus­
cannot help but get the impression that the separate essays have been lumped
geber (Anschrift wie oben) zu richten. Fur unverlangt eingesandte Bucher kann
together without much of a final redaction. This impression is strengthened by
weder eine Besprechung noch Rucksendung garantiert werden.
the lack of any introduction to the subject of numerals. The two-pages introduction
Die Aufnahme von Repliken und personlichen Erklarungen wird prinzipiell abge­
(p. III-IV)which precedes the 'contents' does not discuss the history of scholarship
lehnt; die Mitarbeiter sind ihrerseits zu einer streng sachlichen Formulierung
in this field, nor does it give a satisfactory outline of the premises on which B.
angehalten.

bases his analyses of the numerals. For instance, it appears from his discussion at
Autoren erhalten fur Aufsatzs 25, fur Rezensionen 10, fur kleine Anzeigen 5 Son­
various points that B.adheres to the Nostratic theory, but this is not made explicit
derdrucke.

in the Introduction. Another point in question are the typological observations

Gedruckt mit Unterstutzung des Bundesministeriums fur Wissenschaft und Ver­
which we can make when comparing the different ways in which the world's
kehr, Wien.
languages form their numerals. Throughout the book, B. mainly relies on two
© Wiener Sprachgesellschaft, Wien 1997
Die Zeitschrift und aile in ihr enthaltenen Beitrage und Abbildungen sind ur­
heberrechtlich geschutzt. Jede Verwertung auBerhalb der engen Grenzen des
Urheberrechtsgesetzes bedarf der Zustimmung der Wiener Sprachgesellschaft.
Das gilt insbesondere fur Vervielfaltigungen jeder Art, Ubersetzungen, Mikro­
verfilmungen und fur die Einspeicherung in elektronische Systeme.
Gesamtherstellung: AZ Druck und Datentechnik GmbH, Kempten.
Gedruckt auf alterungsbestandigem Papier
Printed in Germany
ISSN 0376-401X
(well-attested) ways to form numerals, viz. the derivation from body parts, and
the derivation from other numerals by means of simple arithmetic operations.
Nobody will deny that these are the two most important origins of numerals
cross-linguistically, but B. becomes explicit about them only in his very final
chapter (see below); especially for students of Indo-European languages, where
numerals have a notoriously unclear etymology, this would have been a welcome
theme for an introductory chapter.
B.expresses the following two targets of his study (p. III-IV):"(a) to apply
the methods of comparative-historical linguistics to the etymological analysis of
Die Sprache 39/2 (1997), 239-245
 240 Rezensionen
numeral systems in selected language families; (b) to formulate some general rules
of creation of numerals in confrontation with the 'transparent' numeral systems."
What is behind this second aim only becomes clear when we read the final chapter
calIed 'Patterns of creating numerals', which provides a short overview of different
counting systems which exist in the world's languages. B. concludes this chapter
with the observation (p. 336) that "the primary semantic motivation of numerals
was based on body part names." This implies the counting of 1,2, 3, etc. by means
of fingers etc., as can be observed in different languages all over the world. B.
proceeds with the assumption that "the higher numerals originate by means of
elementary arithmetic operations combining them." This is another well-attested
formation type of numerals, as in Latin duodevtginti'lb', undevrginti'19', vtgintt
'20', literally 'two-from-twenty', 'one-from-twenty', 'twenty'.
It is obviously B.'s main aim to show that the Indo-European numerals
have also been formed by the two processes mentioned, viz. the use of body part
names to start counting, and application of arithmetics to derive other numerals
from the former. For the numerals 4-6 and 8-10 (7' is regarded as a loan from
Semitic), B. offers the following etymologies: '4' *k~etyor'set of fingers' or *k~et-l).
'span', derived from *k~et- 'to stretch'; '5' *penk~e 'keeps (hand in fist ?)' from a"
verbal root *penk~- 'to handle' or 'to gather'; '6' * KSl).eks (K indicating an
unspecifiable velar) < *g"s-l).eks from * g"es-'hand' + *l).eks- 'grow' = 'overgrowing
hand'; '8' *h2okto-h j 'set of points' (= dual of 'fingers or knuckles of hand'); '9'
*hjnel) adv., orig. ace. 'in lack'; '10' *dekIp & *dektadv., orig. ace. 'in the end' &
participle 'reaching'. We can see that not only body parts and arithmetics are
involved.
In itself, the idea that the PIE numerals, which lack an etymology, can or
must be explained from general numerical formation patterns is nothing new.
Yet the lack of agreement which has hitherto existed on this subject is not
accidental: it is simply very difficult to find an internal IE etymology for those
numerals, and the suggested connections with non-Indo-European languages alI
suffer from phonological deficiencies. In order to fit the IE numerals into their
supposed etymological motivation, it seems to me that B. stretches the evidence
too far in more than one case. Although he has diligently gathered the many
previous analyses of IE numerals from the literature, and follows the authority of
previous scholars rather than coming up with many new interpretations, B.'sown
conclusions are often open to great doubts. As for '5', a verbal root *penk~- is
simply unattested in Indo-European. Of course, "penkie '5' could theoreticalIy be
a remnant of an otherwise lost root *penk~-, but we simply cannot prove it. The
suggested connection with Hittite psnku- 'total; community' is impaired by
expected development of PIE <enk- to Hit. -ink- (Melchert 1994: 139), which is
absent from psnku-; for this reason, the connection of the latter with Skt. belui­
'much, many' must be retained. In the case of '6' (p. 242), the alleged compound
*g"s-l).eks does not conform to any known PIE composition type. As Viredaz (1997)
has recently shown, the PIE word for '6' was *sl).eks without an initial velar, so
that all reconstructions which depart from *Ks- must be given up. PIE '9' should
be reconstructed with final *-n as *hjnelW, which makes its reconstruction as the
acc.sg. of a supposed noun *hjenu- impossible.
Die Sprache 39/2 (I997), 239-245
Rezensionen 241
For '8' (p. 269),B.relies on Henning's interpretation (1942: 235)of Av. a~ti-as 'the
breadth of four fingers' (which he derived from PIE *h2ok- 'pointed') and his
subsequent derivation of PIE *h2oktoh j as the dual of the PIE preform of Av. a~ti­
(Henning 1948). Yet the meaning 'breadth of four fingers' is not secured for Av.
esti-. The form only occurs in the compound esti.msseb- 'as large as an a~tf, which
the Pahlavr translation of NIrangestan 66 and VIdevdad 13.30 does not specify
any further, and in uz-ssti- 'breadth of eight fingers' (Frahang-i otm 734). The
latter form appears in an enumeration of length measures: F 731ptIoam '16 fingers'
(as per Henning 1942: 236, footnote 2), F 732 vitesti '12 fingers', F 733 disti '10
fingers', F 734 uzesti '8 fingers'. Since vitssti and disti represent regular verbal
abstracts in *-ti- from the respective PIEroots "tens- 'to stretch' and *deik- 'to show',
uz-esti- may well contain a derivative *h2{1k-ti- to the root *h;zIIek- 'to reach',
corresponding to Skt. a$p- f. 'reaching'. The semantic motivation is obvious: vitssti­
is the space between a stretched out thumb and little finger, disti- are the ten
fingers which we use for indicating, and we can accordingly translate a~ti- as
'stretching', which is the position of (twice) the four fingers of a hand without the
thumb. Av. uz-ssti- can be interpreted in two ways: either with uz- 'out, up' as
'stretching upwards' (d. KIingenschmitt 1968: 239 "die Handspitzen nach oben
habend"), or with uz- 'beyond' as 'more than an a~tl (d. Henning 1942:235"super­
a~ti'), as in us-aya- 'very bad' to aya- 'bad'. The latter interpretation would save
Hennings translation of a~ti- as 'the breadth of four fingers', which he posits on
comparison of the Avestan measures with the ancient Greek measures. Within
the Avestan system of length measures, the connection of a~ti- with Skt. a$p- seems
to me far more plausible than the unspecified connection with the PIE root for
'pointed' which Henning suggests. 'The breadth of four fingers' is then Simply
'the stretching of the four fingers of a hand'. These considerations remove alI
indications that the root *h2ok- 'pointed' supplied a word for 'four', of which
Henning 1948 and B. suggest that it provided the basis for the derivation of
*h2oktoh j '8'.
Leaving aside the criticism on details, B.'s book shows the merit of
bringing together an enormous amount of evidence on numerals, their relations
and developments in many languages. These data, rather than suggesting
unwarranted reconstructions, inspire one to ponder about the different possibilities
that languages have with numerals, and how the different numeral systems came
about. It must be regretted that B.does not elaborate more on this theme, which
apparently was his main interest (p. 336 "The creation of numerals confirms more
than any other human activity that man is a measure of himself.") B. has some
very interesting remarks about the distribution of the vigesimal counting system
in Indo-European languages, which automatically leads one to the question why
different languages have different systems such as binary, ternary, quaternary,
quinary, decimal, or vigesimal.
English is not B.'s native language, and this may account for the some­
times strange syntax and choice of words in B.'s text. UsualIy these errors are not
disturbing, but one item should be mentioned. When B. speaks about a 'most
probable cognate' form of a word, he quite consistently writes 'most hopeful'.
This is probably no more than an innocent error in the translation from Czech to
Die Spraehe 39/2 (I997), 239-245
 242 Rezensionen

English; but unintentionally, its frequent occurrence never tires of reminding us

that it is not the cognate which is hopeful, but only the author.

Michiel d e V a a n

References:

Henning, W.B.
1942: An astronomical chapter of the Bundahishn, Journal of the
Royal Asiatic Society, 229-248.
1948: Okto(u), Transactions of the Philological Society, 69.
Klingenschrnitt, G.
1968: Farhang-i Dim. Edition und Kommentar, unpublished doctoral
dissertation Erlangen-Numberg,
Melchert, C.

1994: Ana tolian historical phonology, Amsterdam - Atlanta.

Viredaz,R.
1997: 'Six' en indo-europeen, Indogermanische Forschungen 102,
112-150.

Werner Sun d e r man n , IranianManichaean Turfan texts in earlypublications

(1904-1934). Photo Edition. Corpus Inscriptionum Iran-icarum, Supplementary
Series Volume III, London, School of Oriental and African Studies 1996. 51 p. +
192 plates. ISBN 0 7286 0259 7. Preis: £ 60.

Hierrnit werden die SchwarzweiBaufnahmen samtlicher bis anno 1934 veroffent­

lichten Manichaica der Berliner Turfansammlung und der Peters burger Krotkov-,
Koxanovski- und von Oldenburg-Sammlungen vorgelegt und zum ersten Mal
seit Salemann' in kontrollierbarer Form gesammelt. Die Auswahl ist sinnvoll und
homogen: Sundermanns Photo Edition enthalt mit wenigen Ausnahmerf mittel­
persische und parthische Fragmente im manichaischen Alphabet. AuBerhalb der
gewahlten Grenzen stehen zwei Gruppen von Fragmenten: (1) die "Londoner
Hymnenrolle", die zwei parthische Hymnen in chinesischer Transkription ent­

1 Manichaeische Studien I, Memoires de l'Acedemie Imperiale des Scien­

ces de St.-Petersbourg, VlIIe serie, VIII:10,Saint-Petersbourg 1908.

2 Fragmente in: 50ghdisch (M14,Ml15, M133,M583,5018120,MIK III 4981a

und f; 54, L73, L74, L75, L77, L117), 50ghdisch und Parthisch (M132a, M259c),
Mittelpersisch und 50ghdisch (M309a),50ghdisch und Uigurisch (5014411), Mittel­
persisch, 50ghdisch und Turkisch (Ml72), Parthisch im soghdischen Alphabet
(5018120), Mittelpersisch in tiirkischer Runenschrift (Mainz 172, 402a und b).

Die Sprache 39/2 (1997),239-245

L.