LINGUIST List 15.1554

Fri May 14 2004

Review: Historical Linguistics: Levin (2002)

Editor for this issue: Naomi Ogasawara <naomilinguistlist.org>


What follows is a review or discussion note contributed to our Book Discussion Forum. We expect discussions to be informal and interactive; and the author of the book discussed is cordially invited to join in. If you are interested in leading a book discussion, look for books announced on LINGUIST as "available for review." Then contact Sheila Dooley Collberg at collberglinguistlist.org.

Directory

  1. Robert Mailhammer, Semitic and Indo-European II

Message 1: Semitic and Indo-European II

Date: Fri, 14 May 2004 13:39:00 -0400 (EDT)
From: Robert Mailhammer <Robert.Mailhammerweb.de>
Subject: Semitic and Indo-European II

AUTHOR: Levin, Saul
TITLE: Semitic and Indo-European II
SUBTITLE: Comparative morphology, syntax and phonetics
SERIES: Current Issues in Linguistic Theory 226
PUBLISHER: John Benjamins
YEAR: 2002
ANNOUNCED IN: http://linguistlist.org/issues/13/13-2961.html


Robert Mailhammer, University of Munich

Key to typographical symbols used:

?: glottal stop
!: pharyngeal counterpart of glottal stop
6: schwa (central vowel)
s,: emphatic sibilant
t': voiceless dental fricative
b': voiced bilabial fricative
d': voiced dental fricative
g': voiced velar fricative
s': voiceless palato-alveolar fricative
h': voiceless velar fricative
c^: voiceless palato-alveolar affricate
�,`,^: following a grapheme indicate stress,
^ above vowel grapheme denotes length
Akk.: Akkadian
Arbc.: Arabic
Gk.: Greek
Gmc. Germanic
H.: Hebrew
Hitt.: Hittite
L.: Latin
ModE: Modern English
OHG: Old High German
OE: Old English
Skt.: Sanskrit
All examples from languages with non-Latin alphabet are
reproduced in phonemic brackets (/.../) symbolising the
transcriptions in the book. Other examples and sounds
appear in underscores (_..._) symbolising italics.


OVERVIEW

"Semitic and Indo-European II" is Saul Levin's second monograph in a
series of comparative studies of Indo- European and Semitic
languages. The first volume (1995) focuses on lexical etymology and
suggests numerous correspondences between the two language families
going well beyond the proposals made in e.g. Moeller 1911 and
subsequent investigations. Although it is a complementary database
frequently referred to in the present book the first volume is not
essential for comprehension. "Semitic and Indo-European II"
concentrates on structural comparisons between the two language
families. Though Levin does not rule out a common proto-language, this
is not a "Nostratic book" in the strict sense as it presents a
collection of proposed structural parallels on all levels of grammar
attributed to situations of language contact at various times and
under diverse circumstances.

SUMMARY

"Semitic and Indo-European II" is divided into eight chapters starting
with Chapter VI, as it is a continuation of the first volume.

In Chapter VI the formal "structure of roots and of uninflected words"
is examined. Levin posits that lexical roots in Semitic and
Indo-European can be abstracted to purely consonantal
structures. Although this is quite uncontroversial for Semitic
languages, roots in Indo- European are not usually seen that
way. Although Levin (p. 6) mentions that the morphological system of
Indo-European "does not so readily lead to the positioning of
consonantal roots", he uses consonantal roots for Indo- European
throughout the book and illustrates his view with an example from
Modern English: In _sing ^� sang ^� sung_ the vowel is subject to
regular change as it fills an empty slot in the consonantal root which
therefore may be spelt _s-ng_. This is very similar to the consonantal
root in e.g. Arbc. _k-t-b_ 'write', though English has only one
ablaut-slot. That this is an abstraction can be seen from the fact
that actual words in both languages cannot be formed without the help
of vocalic segments. The basic unit, according to Levin (p. 16), is
the biconsonantal structure which sometimes occurs in a triconsonantal
manifestation. Both languages exhibit certain phonological constraints
in the combination of consonants. Moreover, there is also a small
number of roots that cannot be reduced beyond a triconsonantal
structure.

Additionally, Levin (p. 8) proposes a pattern for cognate roots where
one of the two consonants is weak, i.e. a laryngeal or guttural
segment: Accordingly, Semitic usually retains this element, whereas it
is lost in Indo- European languages. The only exception is Germanic
for which a syllable-initial, pre-vocalic glottal stop can be inferred
from vocalic alliteration in Germanic verse. As a result, Levin
suggests that some Indo-European roots consisting of a CV- or VC-
sequence are originally biconsonantal with the missing element being a
weak consonant. One of the many examples is Gmc. *ala- 'to nourish'
which is interpreted by Levin as a cognate of H. /!ale/ - 'go up',
thus positing a biconsonantal root for Pre-Germanic. On the basis of
comparative data Levin argues against the inclusion of vowels in basic
verbal roots, as consonantal elements tend to be more stable.
Moreover, Levin (p. 31) posits that Indo-European CVCV- roots are
likely to be borrowings from Semitic representing simplifications of
consonantal clusters. This is because Indo-European is more tolerant
than Semitic concerning consonantal onset clusters. The last part of
this chapter provides additional data from Afro- Asiatic languages,
which, according to Levin (p. 50ff), further strengthens the case for
biconsonantal roots as basic units.

Chapter VII ("Stative Inflections") is devoted to "the development of
subsidiary morphemes to express a stative relation between certain
basic morphemes" (p.58). The first two sections deal with _e_
expressing a stative function as part of a pronoun or as part of a
verbal root. Levin (p. 59) proposes that stative morphemes form a
"fairly extensive sub-system" in both languages. In pronouns _e_ shows
up in the second person, e.g _te_, which is found in this function
chiefly as a prefix in Akkadian, Hebrew and Arabic, or as a separate,
preposed word in Latin and Greek. An example for this with assumed
cognate verbs is L. _n�n t� pudet_ 'you (sg.) are not ashamed'
vs. H. /lo? t'eb'o�ws'iy/ 'you (fem. sg.) will not be
ashamed'. Generally, however, in Indo-European the relevant pronouns
are used in a much broader sense than in Semitic and the stative use
with an impersonal verb is restricted to Latin, Greek and maybe Old
English. The distribution of _e_ as part of roots is similar to that
of pronouns: There are several proposed correspondences between
Semitic and Greek/Latin but only few include Sanskrit, which,
according to Levin, is mainly due to phonological reasons as Sanskrit
has a simplified vowel system compared to the Indo-European parent
language. Although most posited cognate pairs involve triconsonantal
roots, there is also an example for a biconsonantal root, the noun
(or, in Hebrew a verb as well) 'seat', H. /s'e�b'et'/
vs. Gk. /he�dos/. Additionally, Levin (p. 92- 101) assumes some
stative function expressed by _o_ or _a_ between the first and second
consonant of a root exemplified by the Hebrew static verb /kos'6ro�h'/
'it (fem.) is proper/ok' vs. the Greek (Attic) adjective /kathar�/
'pure/clean' (fem.). Moreover, some cases, particularly from
Indo-European languages are adduced from which Levin construes a
stative or passive meaning for _u_.

Chapter VIII ("Inflections of Active Verbs and Verbal Nouns")
investigates the diverse correspondences between Semitic and
Indo-European in the inflections of active verbs and verbal nouns. The
first section, dealing with athematic forms (i.e. where the ending
directly follows the root), finds only few possible cognate
structures, one of them involving the verbal roots for 'to come' in
e.g. H. /bo�?-/ and Gk. (non-Ionic) /ba^/. By contrast, it is the
thematic formation (i.e. where a variable vowel _e_ or _o_ follows the
root) which, in Levin's opinion, displays the "most fruitful
correspondences" (p. 110), particularly where this construction is in
statu nascendi. A recurrent example for this proposition is the
comparison of the Indo-European thematic imperative, e.g. Gk. /phe�re/
'carry', to H. /p6re�h'/ 'bear fruit', the neatest match in Levin's
view. Since the thematic suffix seems to have a wider distribution and
more variants in Indo-European, Levin (p.125) tentatively concludes
that the thematic suffix entered Hebrew and Akkadian from an
Indo-European source. Recalling the correspondences of the stative
element in the second person endings (_te_), the next section
investigates second person suffixes in general, mostly involving the
sequence _-tV-_. Although Levin acknowledges that _t_ as part of
second person morphemes also occurs in numerous unrelated language
families he maintains that the correspondences put forward are due to
language contact.

One example for proposed cognates involving second person endings is
the comparison of the second person plural endings in the perfect
tense, e.g. Skt. _c^ac^ar-tha_ 'you (sg.) have wandered'
vs. H. /kora�t-to/ 'you (masc. sg.) have cut' with an aspirate
consonant in the Hebrew ending, as Levin points out.

Moreover, according to Levin, the distinction of gender in second
person endings in Semitic sometimes corresponds to a temporal
differentiation in Indo-European: Arbc. /!alim- ti/ 'you (fem sg.)
know' and /!alim-ta/ 'you (masc. sg.) know' vs. Hitt. _s'ak-ti_ 'you
(sg.) know' (present) and _s'ak-ta_ 'you (sg.) knew' (preterit).

The section on third person endings advances cognates involving
_-at(i)_ and _-Vn_ sequences, e.g. Skt. _a�j�v- at_ 'lived' (3rd
sg. imperfect) vs. Arbc. /Hayy-at/ 'lived' (3rd sg. fem. perfect'. An
analysis of the perfect formation in both language families shows that
the Indo- European kind of reduplication is generally not found in
Semitic. However, Levin, posits a perfect prefix for Indo- European,
reconstructable from what has been seen as anomalous reduplication in
Sanskrit in e.g. _jabh�ra_ 'has brought/borne' and a Germanic prefix
_gV_ in e.g. OHG _giboran_, corresponding to the Hebrew perfect prefix
/yo/_.

This chapter is concluded by an investigation of agent nouns
exhibiting the vowel _o_ and generally feminine action nouns which
contains more lexical than structural etymology. According to Levin
(p.193), the match between pre-classical Greek /(w)oikodo�moi/ and
earliest Biblical Hebrew /bo�neh'b'o��yit'/, both 'house-builders'
illustrate the shared cultural experiences between both language
communities and points to close linguistic contact.

The subject of Chapter IX ("Case-endings and Other Suffixes of Nouns
and Adjectives") is shared case endings and other nominal or
adjectival suffixes. In this, thematic endings denoting the accusative
singular are viewed as exhibiting the neatest congruence:
Arbc. /t'awr- an/ vs. Gk. /tau^r-on/, together with data from other
Semitic and Indo-European languages (e.g. L _taurum_ and
Akkad. _s'�rnam_) point to an ending featuring the sequence "back- or
central vowel + nasal sound" (p.195) which is widespread in Semitic
and Indo-European. However, Levin (p.198) notes that it is the vowel
bearing the distinctive function in the former and the consonant in
the latter language family. From the fact that the system of cases and
declension types in Indo-European is richer and more diverse, Levin
(p. 199) concludes that the distinction of cases spread into Semitic
from Indo- European where it was subsequently adapted to the native
morphonological system.

However, Levin supposes that Semitic also influenced certain groups
among the Indo-European family. He (p. 254ff) adduces the Latin
genitive singular, _taur�_, _haed�_ (cf. Arbc. /t'awrin/), also the
proposed source of the suffix _-i-_ to express 'daughter/son of' found
e.g. in Gk. /telam�nios/ 'son of Telamon'. Additionally, a Semitic
origin for the suffix _-isk-_ forming ethnic terms and other
adjectives, e.g. OE _Iudeisc_ or Gk. /ne�ni�skos/ 'hypocoristic of
/ne�ni��s/ 'young man', as well as the feminine suffix _-issa-_ in
Gk. /basili�ssa/, which then spread as far as Modern English (e.g.
_countess_), is advanced: As the source of the former morpheme Levin
(p.273) assumes Hebr. /?iys'/ 'man' and for the latter
Hebr. /?is's'o�h'/ 'woman'.

In addition to that, Levin compares the nominative case endings and
advances a hypothesis on the development of the neuter gender in
Indo-European: There is a certain tendency for some neuter endings in
Indo-European to correspond to feminine endings in Semitic. In
particular, this holds true for words referring to young animals.
Levin concludes that the old differentiation of animate vs. inanimate,
still tangible in Hittite, was changed due to the differentiation of
the animate category in masculine and feminine. This split, according
to Levin, originated in Semitic where it reflected the increased
importance of the female over the male animal to breeders extending
the classification by gender from livestock nouns to nouns in general,
and which subsequently spread into Indo-European languages. This
caused a major upheaval in the grammatical system reflected in
wavering correspondences and gender assignments.

Chapter X ("Syntax") examines the syntactical similarities between
Indo-European and Semitic in conjunction with the morphological
correspondences. This is because, as Levin (p. 283) notes, agreements
on the level of syntax alone are "open to the suspicion that it may be
merely typological ^� i.e. not due to common ancestry or prehistoric
contact". Due to the typological difference in word order, VX in
Semitic and a tendency for XV in the majority of the ancient
Indo-European languages, both language families exhibit clear
divergences such as the serialisation of noun plus adjective which is
compulsory N+Adj. in Semitic but free in Indo-European. Nevertheless,
the Celtic languages largely display the Semitic verb order, a fact
that Levin attributes to contact with Semitic, wondering whether
further parallels can be found. Another suggested contact feature is
the peculiar sequence ART + NOUN + ART + attributive ADJECTIVE which
is shared between Greek and Hebrew. It is normal for Semitic but
limited to Greek on the Indo-European side where the more frequent
pattern ART ADJ NOUN is found in accordance with the typological
predictions of an XV language. Moreover, Semitic and several
Indo-European languages have prepositions. From the fact that Hittite
generally has postpositions Levin infers that this represents the
original state of affairs which is supported by attested postpositions
in the older Indo-European languages like Greek and Sanskrit. From
this Levin (p. 317) concludes that the prepositions of Greek Latin and
the other Indo- European languages evolved under Semitic influence.
The investigation of the rules for agreement in case and gender yields
much the same results for the two language families, although without
possible correspondences.

Chapter XI ("Corresponding Consonants") constitutes an integral part
of this book as it deals with corresponding consonants, hence
providing a basis for the comparison of linguistic material. The
chapter is subdivided into sections on consonants categorised
according to the place of articulation. The various relations are
generally illustrated with several examples from both language
families. Nonetheless, Levin specifically states that "few, if any of
them [the correspondences, R.M.] can be securely traced back to an
ancestral proto-language, hypothetically called Nostratic". As
expected, the rule of thumb is that those proposed cognates displaying
the neatest phonological matches are those with the shortest
prehistory.

One of Levin's results (p. 333) is that the liquid _r_ seems to be the
most stable consonant in the comparisons presented, i.e. across the
language boundary it is the least likely to be represented
differently, which is exemplified by Arbc. /qarnu/ vs. L
_corn�_. Nonetheless, there are several alternations, possibly
occurring within one language, e.g. _r_ vs. _n_ or _r_ vs. _l_, a fact
that introduces significant complications in proposed
correspondences. One example for this is the comparison of the words
denoting 'field' (acc. sg.): Indo-European shows _r_, e.g. L _agrum_,
where the postulated Semitic cognate has _l_,
e.g. Arbc. /Haqlan/. Additionally, the frequent occurrences of
metathesis add further complications to the relations put forward. One
example featuring a sound substitution is Levin's proposed cognate
pair of Arabic _k-t-b_ 'write' (e.g. /yaktubu/ 'he writes', originally
'to sew together') and Gk. /gr�phe/ 'write', OE _ceorfan_ 'to carve'
(p. 347-348). The Indo-European words are explained as a Semitic loan
including a drastic simplification of the _kt_-cluster, which is
disallowed by Indo-European phonology, by the replacement of _t_ with
_r_. The section on plosives contains a vast collection of proposed
correspondences involving numerous phonemes of which only a small
selection can be presented here. One hypothesis involves the
distinction between "satem" and "centum" languages. Levin (p.351)
posits that whenever Semitic shows a fricative or affricate in a place
where a proposed cognate from a centum language has a plosive, the
Indo-European word is likely to be primary. The reason given for this
is that the satem variety developed secondarily from the centum form
in Indo-European, according to Levin, which is deviant to the
traditional Indo-Europeanist view. Levin illustrates this with Hebr.
/!as,oro�h'/ (voiceless, emphatic sibilant) vs. Gk. /agor�/, both
'assembly'. Other than that, plosives sometimes find neater matches as
e.g. in the well-known etymology L corn� vs. Arbc. /qarnu/. In several
places Levin argues for the reconstruction of Indo-European phonology
involving glottalised plosives advanced by Gamkrelidze & Ivanov
(1995). One example is taken from the section on velar and labio-velar
consonants: Gk. /si�gloi/ is presumably borrowed from Hebrew
/s'iqle�y/. Levin suggests that the Hebrew _q_ goes back to a
glottalised plosive and thus would explain the Greek _g_ as a
borrowing at a time when _g_ was still a glottalised [k] as posited by
Gamkrelidze & Ivanov (1995). By contrast, the Arabic _q_, according to
Levin (p.359), did not have a glottalised articulation being the
source of IE _k_ in L. _corn�_ and its widespread Indo-European
cognates. Several etymologies proposed by Levin presuppose a relation
between Indo-European _s_ and Semitic _t/t'_, such as H. /s'eb'et'/
vs. Gk. /he�dos/. Levin (p.423) argues that the fricative _t'_ could
be a bridge between Indo-European and Semitic in this aspect and
concludes that the data does not permit any clear judgement but
supports some "overlapping" (p. 428).

Additionally, in some cases, there are semantic difficulties despite
phonological matches, as in Hebr. /ke�leb/ 'dog' and OE _cealf_
'calf' (OHG kalb), further complicated by OE _hwealp_ (ModE _whelp_)
which leads Levin (p. 371) to the assumption of a reciprocal process
of borrowing (p. 370-372). Furthermore, Levin (p. 383) posits that
Indo-European words containing /b/ may be borrowings from Semitic as
_b_, if it existed at all, was definitely very rare in
Indo-European. The section on sibilants repeats the well-known loan
etymology for the Indo-European _seven_-word from Semitic and stresses
its cultural significance.

Chapter XII ("Vowels and Suprasegmental Sounds") opens with the
observation that the vowel system of Greek is a great deal closer to
Hebrew than to Sanskrit. In the following section Levin argues that
this similarity can only be the result of an intense contact situation
and cannot simply be a taken as a typological coincidence, in
particular when the correspondences described in the preceding
chapters are considered. In the section on ablaut and accent Levin
argues for the connection of the Indo-European /e/-/o/ ablaut to
accent which, according to Levin, is particularly evident in
Greek. Consequently, he posits that /o/ resulted in "intoned" position
in contrast to /e/, which occurred in unstressed environment.
Following this claim is a rich collection of data taken from Semitic,
Greek, Avestan and Sanskrit. Additionally, Levin tentatively suggests
schwa as a possible minimal vowel for Indo-European and concludes this
chapter with a section on metre which mainly finds correspondences
concerning octosyllabic meters and verse structures.

The final chapter (XIII, "Epilogue: Echoes of Prehistoric Life")
recapitulates some points considered particular important by Levin,
such as the significance of livestock and the influence the two
language families exerted on each other in this semantic field, and
presents some additional thoughts. The book concludes with several
indices for the reader's convenience.

DISCUSSION

In this book Levin raises numerous controversial points, only a small
selection of which will be discussed here for reasons of available
space. From the existence of postpositions in Hittite, Levin concludes
that this represents the original state of affairs and that the
Indo-European languages received their prepositions through contact
with Semitic languages. However, the non- occurrence of prepositions
in the Indo-European parent language is common knowledge and included
in the standard handbooks, such as Tichy (2000:39), Meier-Br�gger
(2000:258). Moreover, it is also to be expected according to the
typological predictions for an XV language (see e.g. Lehmann 1974:15,
Vennemann 2003a:333-336). Both points are not mentioned by
Levin. Additionally the database, comprising only Sanskrit, Greek,
Hittite, Latin and Germanic on the Indo-European side, seems
insufficient to warrant his proposal. It would be interesting to see
how languages developed for which the contact with Semitic is more
difficult to posit. As a matter of fact, the assumption that Germanic
developed its prepositions through contact with Semitic requires an
additional explanation of this proposed contact situation, which is
not given by Levin. As he (p. 317) notes the evolution of prepositions
must have occurred comparatively late, as postpositions are attested
at an early stage in most languages reviewed by Levin and prepositions
only spread later on. Sanskrit, for example, has regular postpostions.
It seems particularly odd then that Hittite, despite extremely close
contact with Akkadian (p. 315 and Watkins 2001), kept its
postpositions whereas other Indo-European languages should have
acquired their prepositions in historical times for which the contact
situation was not as close or may have never existed at all. The
occurrence of prepositions in Celtic is only one of the numerous
syntactical features this language shares with Semitic languages,
alongside the position of the verb, all of which have been known for
some decades and which have been attributed to contact phenomena
(cf. Pokorny 1927-30, Gensler 1993, Vennemann 2003a). It is difficult
to understand why Levin mentions nothing of this, thus missing an
opportunity for the investigation of structural correspondences.

Another problematic position is the notion (chapter VI) that roots in
Indo-European are consonantal, just like in Semitic. Although similar
abstractions have been posited by some Indo-Europeanists (see
e.g. Pooth 2003) this is doubtful because in an Indo-European root the
vowel carries lexical information. This can be inferred from the fact
that some Indo-European roots show the basic vowel _a_ rather than the
more frequent _e_ (see e.g. Kurylowicz 1956:187, Jasanoff 2003:3). By
contrast, consonantal roots presuppose that the vowel is irrelevant
for the lexical meaning of words based on such roots. Thus, the vowel
is a distinctive property and does not alternate according to
morphological function as in Semitic. Therefore, vowel gradation in
Indo-European does not have the same mophological significance as in
Semitic where it expresses grammatical categories directly (see
e.g. Lipinski 1997:358). In Indo-European vowel gradation is a
concomitant of particular types of stem formation, where it generally
is a recessive and redundant property (see e.g. Kurylowicz 1961:13,
Meier-Br�gger 2000:135f, Tichy 2000:37). Additionally, it is often
tied to word stress (e.g. in the case of zero grade), a fact that
Levin himself notes in his section on ablaut. Nevertheless, it is
noteworthy that Levin's support for consonantal roots comes from
Modern English, a Germanic language, whose verbal stem formation
exhibits major typological divergences from that of the Indo-European
language family due to the systematisation and functionalisation of
ablaut in the Germanic strong verbs (see e.g. Prokosch 1939:120,
Szemer�nyi 1980:86). As a matter of fact, the structural
correspondences as well as the striking and unique typological
proximity between Germanic and Semitic concerning ablaut are neither
analysed nor mentioned by Levin although they have been pointed out
before (e.g. Kortlandt 1992:102) and language contact has been
suspected as the reason behind it (e.g. Stedje 1987, Vennemann 1998).

Another problematic point is Levin's position of an Indo- European
perfect prefix is to a considerable extent based on the existence of
the prefix *_ga-_ in the Germanic past participle. This assumption is
clearly untenable as the prefix *_gi/ga-_ did not go with the past
participle in Germanic times but was only used in the attested
daughter languages with varied regularity (e.g. de Boor & Winsniewski
1985:121, Lehnert 1990:104). Additionally, the perfect tense in
Indo-European cannot simply be compared to the perfective aspect of
Semitic which clearly has an aoristic quality. These three points are
just some examples for the major problems of the present volume: In a
large number of cases the data is insufficient to support the often
far-reaching conclusions drawn by Levin. Moreover, several
propositions are either disputed or falsified outright once the actual
data is examined more closely, e.g. the position of an Indo-European
perfect prefix. This also holds true for many of the proposed
correspondences on the phonological/phonetic level, such as the
relation of IE _s_ vs. Semitic _t_, or the position of a Proto-Semitic
glottalised plosive _q_ (disputed in Lipinski 1997) discussed in
chapter XI. Moreover, sometimes the argumentation is circular: The
match of the vowel systems of Hebrew and Sanskrit are taken as
indications of contact because of all the other correspondences
proposed. Numerous suggested contact features are so unspecific that
they hardly support Levin's positions: _i_ in various combinations
occurs as a genitive marker in Indo-European, the occurrences in Latin
and Greek need not be the result of Semitic influence.

Furthermore, it is a considerable problem that Levin does not present
a historical background hypothesis providing a basis for his
proposals. This is all the more surprising as such a theory, linking
Germanic and Celtic to Semitic, has been around for years (see
e.g. collection in Vennemann 2003b). Additionally, several etymologies
or other points suggested by Levin have been proposed elsewhere which,
however, frequently remains unmentioned: The semantic problem of the
_calf/dog_ word (chapter XI) is addressed in Vennemann 1995, the
_seven_-word has recently also found acceptance in the
Indo-Europeanist literature as a Semitic loanword (see e.g. Rasmussen
1995), and the inference of a pre-vocalic glottal stop in the Older
Germanic languages from vocalic alliteration is included in nearly
every handbook on Old English (e.g. Lehnert 1990:35).

EVALUATION

"Semitic and Indo-European II" is a vast collection of data from
Semitic as well as Indo-European languages covering many aspects of
grammar. Nevertheless, this book suffers from numerous severe
problems. Firstly, the material presented is often incomplete, as
significant results from other studies or pertinent data are omitted.
Secondly, no comprehensive framework or theory is outlined to account
for the proposed contact phenomena, and existing approaches are not
considered. Thirdly, the conclusions frequently are ad hoc and are not
supported by detailed analyses. The most substantial results may be
several individual etymologies and some correspondences on the
syntactic level which, however, are not investigated with the
necessary consistency and thoroughness. As a result, large parts of
"Semitic and Indo-European II" are quite speculative, neither
substantiated nor sufficiently researched with the necessary
depth. One does not have to go as far as R. Schmitt (1996) in his
review of the first volume (Kratylos 41, 201-205), who says "Die Zeit
f�r die Lektuere dieses Buchs ist vertane Zeit" (Reading this book is
a waste of time). Nonetheless, "Semitic and Indo- European II", just
as the first volume, would have been of much more benefit to the
reader, if it were shorter, concentrating on fewer, but more
substantial cases and incorporated the relevant literature.

REFERENCES

Gensler, O. D. (1993), A typological evaluation of
Celtic/Hamito-Semitic syntactic parallels, unpublished PhD thesis,
Univ. of California, repr. University Microfilms International, Ann
Arbor, Michigan 1995]

de Boor, H. & R. Wisniewski (1985), Mittelhochdeutsche Grammatik,
Berlin / New York

Gamkrelidze, T., V., & V. V. Ivanov (1995), Indo-European and the
Indo-Europeans, 2 vols., Berlin / New York

Jasanoff, J. (2003), Hittite and the Indo-European verb, Oxford

Kortlandt, F. (1992), "The Germanic Fifth Class of Strong Verbs",
North Western European Language Evolution 19, 27- 81

Kurylowicz, J. (1956), L'apophonie en Indo-Europ�en, Wrozlaw

Kurylowicz, J. (1961), L'apophonie en S�mitique, Wrozlaw

Lehnert, M. (1990), Altenglisches Elementarbuch, Berlin/New York

Lipinski, E. (1997), Semitic languages: Outline of a comparative
grammar, Leuven

Moeller, H. (1911), Vergleichendes indogermanisch- semitisches
W�rterbuch, G�ttingen

Pokorny, J. (1927, 1929, 1930), Das nicht-indogermanische Substrat im
Irischen, Zeitschrift f�r celtische Philologie, 16, p. 95-144,
231-266, 363-394; 17, p. 373- 388; 18, p. 233-248

Pooth, R. A. (2003), "Proto-Indo-European Ablaut and root Inflection",
paper presented at 16th International Conference on Historical
Linguistics, Copenhagen, 10th- 15th August 2003

Prokosch, E. (1939), A comparative Germanic Grammar, Baltimore

Rasmussen, J. E. (1995), "Betonte Schwundstufe im Indogermanischen",
in: Smoczynski, W. (ed.), Kurylowicz Memorial Volume Part One, Cracow,
93-100

Schmitt, R. (1996), "S. Levin, Semitic and Indo- European", review in
Kratylos 41, 201-205

Stedje, A. (1987), Warum nur im Germanischen: Altes und Neues zum
Ablaut der starken Verben, Studier i modern sprakvetenskap, NS 7,
96-113

Szemer�nyi, O. (19893), Einf�hrung in die vergleichende
Sprachwissenschaft, Darmstadt

Tichy, E. (2000), Indogermanistisches Grundwissen, Bremen

Vennemann, Th. (1995), "Etymologische Beziehungen im Alten Europa",
Der Ginko Baum Germanistisches Jahrbuch f�r Nordeuropa 13, S. 39-115

Vennemann, Th. (1998), Andromeda and the Apples of the Hesperides, in
Jones-Bley, K., Della Volpe, A., Dexter, M. R., Huld, M. E.,
Proceedings of the Ninth Annual UCLA Indo-European Conference, Los
Angeles, May 23, 24, 1997, Journal of Indo-European Monograph Series,
Washington D.C., 1-68

Vennemann, Th. (2003a), Syntax und Sprachkontakt: Mit besonderer
Ber�cksichtigung der indogermanischen Sprachen des Nordwestens, in:
Bammesberger, A. & Th. Vennemann (eds.), Languages in Prehistoric
Europe, Heidelberg, 333- 364

Vennemann, Th. (2003b), Europe Vasconica - Europa Semitica, edited by
P. Noel Aziz Hanna, Trends in Linguistics: Studies and Monographs 138,
Berlin

Watkins, C. (2001), "An Indo-European linguistic area: Ancient
Anatolia. Areal diffusion as challenge to the comparative method?",
in: Aikhenvald, A., Dixon, R. M. W., Areal Diffusion and genetic
inheritance, Oxford

ABOUT THE REVIEWER

Robert Mailhammer is a Ph.D. candidate at the University of Munich
currently researching the morphological and etymological situation of
the Germanic strong verbs.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue