Volume 20, Number 2
Fall 1999
C ARCHEOLOGY
CONTENTS of this issue
GP. Rice Dental Variation and Human History 1
(Reviw of Scot and Turner)
0. Ba-Yoser And You Shall Eat th Plants ofthe Feld 4
(Review of Damania, etal, eds)
LE. Rasaessex Language and Prehistory 8
AND F. CavOTO (Review of Greenberg)
MJ. O'BRIEN Commentary: The Futur of Eeolutionary Archaeology 1“
AND RL. LyMan
RETROSPECTIVE REVIEW: FROM THE ARCHIVES
P.Beuwoop (Cloniing te Limits 23
(Review of Irwin; Bahn and Flenley; Lee)
A Bern Middle Pleistocene Behavior in the Dordogne 23
(Review of Rigaud, ed)
P.A. CouNvaLx ‘The Goduin Archive 3
(Review of Mahaney, ed)
DF. Dincavze “Monk's Caes and Shrt Memories 35
(Review of Neudorfe)
RIL For Reconstracing Historie Subisence 38
(Review of Reitz and Scarry)
DH. Keusy Writing at Teotibuacon a
(Review of Langley, Berri, ed.)
SJ. O1ses Feaunal Frauds and Other Ponned Hoaxes 45
(Review f Feder)
DA.Ror ‘The Neopolacltie 48
(Review of Gamble)
PH, West Archaolgy inthe Press 52
(Review of Prpuler and Soi-Popular Article’
Tie ARcusrovocxcat RECORD 58
Isnex To Vous 19-20 628 The REVIEW OF ARCHAEOLOGY
Linguistics
Language and Prehistory
By Jens Elmegird Rasmussen and
Fabrice Cavoto
niversity of Copenhagen)
{Indo-European and its Closest Relatives. Toe Eurastatic Lan-
_Buage Family. vel | 2000) Joseph H. Greenberg. Stanford Uni
Yersty Press, Sunford. x+ 317 Pp
It will take up too much space to describe the creden-
tials of Joseph Greenberg who has done pioneering clas-
sificatory work on practically all of the world’s lan-
‘guages. Since no one can be an expert on all the many
languages and language groups treated by Greenberg—
although in the case of many he definitely is—Green-
berg’s ceuvre is perhaps more than anything a grand
scale test of the validity of the state of the art of all the
‘many fields and subfields on which his theories and con-
clusions are based, Far from all relevant languages are
equipped with ideal tools of study in the shape of clear
and adequate grammars, etymological dictionaries and
ceasy-tocconsult treatises of their phonological and gram-
‘matical history. A wellstudied linguistic subfield ought
to be ready for the comparative handling to which
Greenberg subjects them all; some are fully prepared for
such an undertaking, others are currently just getting
there, while some will perhaps not be "Greenberg-ipe”
for quite some time. Born in 1915, Greenberg cannot
wait for Utopia to materialize, but has wisely chosen to
act now. One can only be grateful that he has taken up
the challenge on a basis which he would undoubtedly
himself have wished to be better in many parts.
‘A large-scale comparison comprising the Eurasiatic seg
ment of “Nostratic” is by definition stil 2 pioneering
work. The study of Nostratic, long since out ofits infan:
cy, may be said to be now in its teens, strugeling with
the question of which of two promising identities it will
have when it finally grows up. Future will show whether
etymologies deriving Indo-Buropean *D, *d, "g (with the
variants “g and “g*) from Nostratic plain ‘p, "t, *® (ihus
MlSvity€ and his followers) or those departing from
slotalized ‘p, ‘t, “%’ (he position defended by
Bomhard) will carry the day, or perhaps some reconcilia
ton of the two schools will prove possible. Since there
are numerous lexemes that do not involve the controver-
sial parts of the phonological system, one will have to
acknowledge that enough within the Nostratic relation-
ship looks so sound thar it would be foolish not to follow
the lead.
‘The methods used by Greenberg in his big compar-
isons have been criticized on several occasions. The
reviewers can only say, So what? Were Rask’s, Grimm's
and Bopp’s methods any better when they did the
‘groundbreaking work that founded what has since been
hallowed as the model for all comparative linguistics to
follow? In many respects Greenberg's new book resem-
bles Bopp's classic Conjugationssystem of 1816 which
‘to many is simply the birth certificate of Indo-European
studies. Like Greenberg now, Bopp was not at all con-
‘cemed with sophisticated phonetic laws, but contented,
himself with superficial similarity between the forms of |
the languages he compared, And since the naked eye can
see the underlying identity of the Ist, 2nd and 3rd plural
Vol. 20, Number 2
endings of Persian barim, ‘i, and ‘ake’ and those of
Gothic sokjam, soketp, sokjand ‘seek’ quoted by Bopp
1816:117 who does not even bother to mention the man-
ifestly supporting endings of Greek and Latin, the Indo-
European felationship was never in doubt. Likewise one
may quote Finnish ipl -mme, 2pl tte along, with the
basic personal markers “m, ts comesponding to Lat.
‘me, tse as defite proof that the comparison of Uralic
‘with Indo-European, regardless of the details of phonetic
Correspondence rules, #8 based on more than. wishful
thinking. The same markers tum up as Eskimo-Aleut “my,
"cand the intervening Chukotian group also has /a/
‘Ad for 1s and 2nd person. Is precisely the aim of Green
berg’ new book to add the morphological suppor to the
etymologies already suggested. As with Indo European
this work. if successful, will be decisive, since, while per
haps any specific lexematie unit can be borrowed from
ane language to another, obvious correspondences in the
material used. for purely. grammatical purposes are
extremely unlikely to reflec anything other than straight
forward relatedness, be. the situation that all the lan
guages concemed are in effect simple descendants of @
Single ancestor. I is the firm impression of the reviewers
thatthe only defensible conclusion i positive one
‘One of the most tenacious controversies pertaining £0
the idea of a Eurasatic subgroup within the Nostatic
relationship concems the position, indeed the very exis
tence, of Alsi. Here the testimony of comparative mor
phology is surprisingly clear. In view of the existence of
ist person marker /m/~ /b/ and a 2nd person //
(whence in places /9,//, /8) in Turkic and Mongolian
‘which bring to mind TE “m, °s, pl “me, “te, combined
‘with, eg. Turkic interogatve /k/ for persons and. /m/
Tor things which are hard to separate from Hungarian Bf
‘who’, mf "what, there can hardly be any doubt that
these languages do indeed belong to Nostratic. And ifthe
‘m/o alteration inthe Ist person marker covers Turkic,
Mongolian and Tungusic there ean be no serious doubt
that they form a special branch within Nostractie. In
GGreenbers’s words, “the chance of an iregular alterna
tion bé/men arising independently three times 1s infin
tesimal” (14). Note that tis will be no Tess true if the
interchange is found to be regular after al
‘Geenberg's new book is thus net so much concerned
with Indo-European which often plays only a minor par,
1s with its “Clovest Relatives", the basic topic being the
Unity and delimitation of Buraslati. This is taken to con-
Sist of Eiruscan, Indo-European, Urlic and. Yukaghir,
Altaic with Korean and Japanese, Ainu, Gilyak, Chukot-
fan, and Eskimo Aleut. The attention given to Etruscan is
expectedly smal, andthe author hestates ast ts satus
as either 2 separate branch of Eurasatic or a8 “a third
‘member of Indo-European, alongside of Anatolian and
Indo-European proper” (23). Thus, the basic message of
the frst chapter, “The Historical Background” (1-2), is
that an Altae node is real, while the other branches are
‘separate units of the large Eurasiatic subgroup of Nostrat-
iclof which the non-Eurasiatic members are Kartvlian,
‘Afvoasatic and ‘ElamoDravidian. Some hesitation is
expressed conceming the position of Karwvelian which,
judging by the taxonomic chart of Bomhard 1996:22,
Seems to be more closely related to EA than the other
Outside branches, so that tis in effect a matter of defint
tion whether one wants to include itor not. The specific
arguments are left forthe following chapters, the relative
iy short “Some Aspects ofthe Comparative Phonology of
Eurasatic™ (24460), and the majestic and central Chapter
3 "Grammatical Evidence for’ Burssiatic™ (61-240), AnFall 1999
¢laborate appendix treats Ainu vowel alterations (241-
76), and a conclusion offers a survey "Classification of
Eurasatic Languages" (27982), 0 which are added refer
‘ences and indices.
‘The phonological section is the weak part of the
fortress Its main concems are alternations between stop
and nasal and within the vowel system.
“The nasal/non-nasl alternation of Eskimo-Aleut is ten
tatively compared with similar phenomena in Ural and
‘Chukotian. Thus, for Uric, the 2nd person marker is
in Ziryene and Ob-Ugric in some parts of the verbal sy=
tem, but-d/t in others, and also Samoyed (Nganasan and
Nenets) has 2sg -n in wordfinal position (alternating
with medial -r-, conceivably from *+-). This is highly remi-
Ascent of the Eskimo-Aleut nasallzation of word final
stops (Alaskan WEsk. ann ‘man’ etg. nut, Alaskan
Esk. aqum, erg. anutim), while the governing rule for
CChulkchi isthe reverse, 70? ‘thee’: jena “in thee’. Emo-
neously, the author interprets the original Eskimo situs.
tion as one of nasalization in close juncture, which is
indeed found in many dialects, but only as a shallow rule
of sandhi (as the author describes at great length). It is
plain that this is no firm bass for genetie conclusions.
The section about TE wocalism is toally unacceptable
to the reviewers. In the author's view altemations of
height (/e, u/o are typical indications of earlier vowel
harmony. Since such alternations can be cited from some
of the other branches of EA, viz. Chukchi, Korean and
Gilyak, and allegedly tic in well with the living vowel
harmony of Unilic and Alaie, the author claims his
broader perspective confirmed by the existence in indi
vidual IE languages of some odd instances of vowel vara
tion which he projects back beyond the IE protolan-
guage all the way to a common Nostra vowel
Sltemation, In each instance, a more “local” explanation
is belitled as ad hoc or even reflecting fear to face the
facts. To the reviewers, such weak spots in the phono-
logical histories of the IE languages are still superior 10
statements that blatantly violate wellestablished rules as
the author's suggestions mostly do, Infact, iis often not
hhard to explain the troublesome forms as due to quite
natural changes ofa more recent date, Thus, for OCS 1,
Lith. #3 vs. Lat. ex (34f), it may be noted that Balto-Slavic
“iis the antonym of *in (Lith. 9 and so may have taken
‘over its vowel by a commonplace analogy, much as Slav.
‘ve ‘in’ took that of “ud (> OCS vy ‘out’). Gk. W§ ‘night’,
a8 opposed to Lat. nos etc. quite regular from "dks
by “Cowgil's rule” (seen also in dv “nail” from
*byndge*s). Gk. dal, loviarvs. Cova is a simple sponta.
ncous assimilation, while OCS wera ‘yesterday":veter
‘evening’ and the altemate form Sutyre vs. Eetyre four
are mere allegro reductions. The same may g0 for Gk
‘mioupes for normal réooapes ‘four’, but since the Gk.
reflex of “schwa secundum” is /i/ (xitvnut ‘I expand’
from ‘pj-nd-bemD, one may derive all variants from an
old paradigm “kérwores, loc. “ke ursi (he later with
‘ru:> Gk. -6). True, the theory of schwa secundum is
‘now generally abandoned” (36), but by mistake, for
Schindler 1977 irefutably shows it to be the regular
appearance of zero in the environment «TTR Ge,
‘between two inital plosive consonants followed by a
nonsyllabie sonorant) An example like Sk. ras :Avest
sarab “head” is pif, since both represent Indole
“ras from older *kpHos, a (repaired) zero-grade variant
of Gk. xépag, this being normal ablaut e/zero with no TE
1V iavolved, The same goes for Avest.-cina 'at all which
is identical with Skt. cana, not an old alternant of it. San-
skrt sima vs. sama is s0 Isolated that an adhoc explana:
The REVIEW OF ARCHAEOLOGY 9
tion is called for: analogy with , cf *Dynérz > "bnér=, bY PIE Simpl
Fed to “bynér, Gk. dvp "man’; 3) on its way t0 220,
Unstressed “e passed through an olike stage, wherefore
lengthening of unstressed “e- yields °<, cf. “suésor >
Lat soror ‘sister; (4) the acrostatic (Narien”)ablaut &/6
is regular from underlyingly long root vowel, to which
‘one need add only rule of initial accent PIE stéy/stéy-
of Ved. stdu‘i, mid, stdve from stu- ‘invoke’, prelE
‘siéus/stéue, the later alternant developed via “seu! >10 The REVIEW OF ARCHAEOLOGY.
‘stéy., the “inidal accent rule” operating vacuous in
other words because unstressed short vowels have
already been lost, forms lke gen. "p2,trds ‘father's, 3p
‘prondntet “il and even da "eugbYoe Tour being
in fact accented on thei fist full vowel); (6) additional
lengthening of @ produced 3 (nom. “pédz > ‘pods
“foot, probably via timoric /e:/ pronounced in two sy)
lables {€-] one of which was unaccented and so became
{o}, whereupon (€.0] was contracted to a long o-vowel;
stems in clusters shortened the lengthened vowel of
the nome. ef. "ndkets ‘night’ vs. (pbds. An account of
these basi rules is given in Rasmussen 1978/99,
‘Special rules apply to certain varieties of IE /o/: (7) the
“thematic vowel" (Le. a vowel in stemfinal position)
alternates ¢/, the form being /o/ before a voiced seg-
‘ment, /e/ elsewhere, cf. verbal endings in “om, "4s, “2,
ont, ot pronouns in “od, "of, and the telling common
form of ipv. and voc. “e; the only exception is the
nom.sg. in "0s which may be accounted for by assuming
earlier voiced *z for the lengthening nom, marker as
‘opposed to voiceless" of the non-lengthening 2s8 mark.
ef of verbs: (6) the d/zer0 of some reduplicated verbal
Categories (eg. the perfect) must be simple disimilation,
‘ievedeor- being ultimately based on *krerdeer; () an
“infxal 0 must be assumed for causatives, the wordtype
toga and derivatives like Gk. répy7, the forms coming
‘out regular only in case an extra element "-Q- (a conso-
‘ant that changes to o after the working of the ablaut
rules) is inserted into the root segment, cf. Rasmussen
1992 for arguments.
Of this only the thematic vowel and the infixal-o- have
any chance of turing up in extemal evidence. The 10
‘may be ultimately identical, but that i for fuure analysis
to decide.
‘As claimed by IitSvityé the overwhelming frequency
of underlying /e/ inthe TE vocaliam must be due to merg
fer of a more varied older picture preserved in Ural and
Alaic. However, a yet unpublished study by Cavoto
shows that Urllé and Altaie do not demand seven origi
nal vowels, but share a system of four basic vowel
phonemes, viz. a, d, w/o, /, and that the opposition «
2 dissolves when the forms are projected back into the
Nostra provolanguage for which a threefold system
‘ah, u/o, Ve sufices, all variation being ascribable to the
‘Adjacent consonants
‘The central chapter 3 presents 72 grammatical items
as evidence for EA. The trouble with grammatical mark-
ts and pronominal stems is that they are very short and
often of somewhat vague functional identity, this making
the danger of fortuitous resemblances very great. Thus,
some of the entries have litle oF no persuasive power,
‘while others make up small sets that seem hard to dis
miss,
‘Unable to comment on all, we call attention to a num-
ber of specific items which may be of general interest
(he numbers are those of Greenberg's sections)
‘V2. Fistperson mvs. of Uralic with much the same
Aisesbution as TE “mvs. “bp, ie. with m as a more active
or engativelike marker as opposed to an inactive velar,
the difference being restricted to the singular. Compars-
by, Eskimo-Aleut has 19g * opposed 0 a labial in the
dual and plural, athough the facts given are somewhat
Inadequate in showing this: Thus, the Esk. ergative Tse
sma is quoted as representing Ist person “sm, while the
‘Aleut form -min shows that the person marker is really
‘de, whence word-inal "9, so that Esk. -ma isin fact from
older “mena, i. erpative marker *m-+ Isgy <*k + an
‘optional panicle not present in the Aleut form. Note also
Vol. 20, Number 2
the difference in the 1p! possessive, inergative “put,
‘which is somehow from *-B-5 with a labial (< *m12) of 1st
person + plural “8, hardened by general rule to stop “in
‘wordsfinal position, contrasted with the ergative counter-
part “mita from ergative *m- + 1st pers. “B+ plural “t<
£8, again with word-inal treatment before the particle 2
With fewer preceding consonants there is wordinternal
treatment in the Isg inerg. possessive in Esk. -ba Aleut
without the particle -1). The Esk-Al. Auslautgeseze
Changing underiying stops into nasals, and spirants into
stops, are all but given in Swadesh 1952:168 and have
been discussed at some length in Rasmussen 1979 (curt
cna they ae noe mentioned Inthe survey of Eaimo
rules given in the introduction to Fortescue
AD 198s, Altai sso as Ist person my, retained word
internally as in Turk, -miz ‘our, and changed into
‘when wordinitial (Turk. béz “we'), except when the
‘word contains a second nasal, as “min ‘T (Uzbek men,
Mod.Turk. analogical bin). We suggest this formulation as
superior to the traditional explanation of b > m by assim-
lation which the author rightly dismisses.
45. Second person t seems inescapable, but the alterna-
tive $ looks more than anything like an original cond
tioned variant. Thus, in IE, the variation 23g *s : 2pl “te
and 2sg pronoun ‘tu, *te can be paralelied, ¢8, by
“némos ‘reverence’ : *németo-s ‘holy’ and other cases of
an interchange *@t/"s collected in Rasmussen. 1994,
Likewise for Altai: if Mong. is derived from "1 (72),
‘one wonders why si of other Altaic languages cannot be
4 further development of the same. Surely Eskimo-Aleut
2pl poss. ci must contain a plural marker also; note that a
derivation from *4-5 would make it parallel with 2du Esk.
‘ak, Al. -Otx from "47, the palatal character of Esk. *6
(reflected as (6), [2], (6] and () in individual dialects)
makes the development to {¢) with an auxiliary Fvowel
fully comprehensible, There thus seems to be no real ev
dence for the author's Nostratic 2nd person *s
7. A pronoun base ge or ga is suggested to account for
the IE pronoun ‘T'and the stem nuclei of pronouns like
Chukchi 787/70" or Hung. en-gem/téged “Wthou' and
Eskimo -miedn ‘L.. thee’. While the rest may look prom-
‘sing, an IE “eg(@2)om is supported only by Indoranian
(Skt. abd), which has an extension ~am on many other
pronouns also (Skt. fuim ‘thou’, aysdim “he’), the com.
bined evidence leads rather to TE *(é@. The stem of
the Esk-Al, personal pronouns is found to consist of a
dental + velar cluster, perhaps /ty/, ef. for details Ras-
rmussen 1987/99. This makes things far too complicated
for a simple equation of ‘ef Jom with Kamassian
Samoyed fgd-mt ‘Lam’
89. A third-person pronoun /e and a demonstrative
stem a/e are distinguished on quite shady grounds. If //
‘of IE enclitc “im, “id is a regular weakening of the “the-
matic vowel” “e/-, the eform of the orthotone gen.
*éxjo, dat. "ésmof in ho way separates it from the second
‘of the two items. As for the vocalism, there appears to be
widespread crosslinguistic agreement on front vowels
for near deixis as opposed to darker vowel in stems of
more distant deixis, cf. Hung. ez/az ‘thisthat’, it/ott
‘here/there’. There is a very interesting, and possibly sig-
nificant, resemblance between the Ainu detransitivizing
prefix f and the Eskimo suffix quoted as + (its main var
and) of same function; the Eskimo morpheme is found to
be ultimately “+ in Rasmussen 1979:71 (thus also, per-
hhaps independently, Fortescue & al. 1994:396), which
offers no good support for its being in origin an agglutt
nated object pronoun,
10. A demonstrative ku of unclear deixis (distant inFall 1999
Gilyak, near in Altaic cum Japanese) is taken to underlie
Hitt, kin, the acc.sg. of Ra ‘this': however, also apa
‘that’ has acc. apiin, the wform being in all probability
simply’ a reduced form of the vowel influenced by the
‘old “m, cf, esp., the anaphoric stems ani- and uni. built
‘on two different forms of the acc.sg. of the endlitic-a
NNo such rounding occurred in Luwian 2a, za-n which
proves an IE palatal ike Lith, 3s (whose /i/ is from
enelitic use); it should be noted that there is no Hiero-
alyphic Lowian variant®., this being simply an antiquated
reading for 2a. The Eskimo 3rd person suffixed pro-
nouns quoted for the “participial mode” in-lu-m, tut,
-lu-y6t are those of East Eskimo; the full picture indicates,
PEsk. “yu, “Ak, “kA (thus sil in SW Alaska) with the
special consonant gradation of original word-final clus-
ters (Rasmussen 1979:33); it is not easy to define the orig
inal identity of the elements of the series *-KX, “KX-7
*-KX-6 thus pointed to (metathesis of the pronoun stem
/e¢P VER), but it is no evidence for an old vowel alter-
‘ation u/i
11, A demonstrative “t¥ seems inescapable and repre-
sents a classical item in distant comparison. The nom sg.
masc. "50, fem. “sab of the IE stem *te-/'to-is traditional
ly assigned to’a different stem; however, given the IE
alternation s/t, itis not inconceivable that *s0 is by dis
similation from older *s0-s, and this in turn by assimila
tion from regular “fos (thus Rasmussen 1994). Forms
(with different vowels) are quoted from the entire Eurasi-
atic area, Eskimo offering its only prefix, the anaphoric
ta- (possibly really *taé) attached to deictic local expres
sions, The reported “distance distinction” tu/to (fa) vs.
tite (neat) again has a sound:-symbolic fee! to it.
12. The alleged demonstrative s, dispensable as such
for IE, may be properly a reflexive. Its development into
4 3rd person pronoun in Uralic (Finn. han) is paralleled
at least by the Greek enclitic anaphoric & (ws. orthotone
reflexive €) and, for the possessive, by German sein and,
French son Gilso Hitt. 3/3 ‘his, also the Latin
accusativus cum infinitivo and the Eskimo use of the
reflexive of the recurrent 3rd person open avenues for
such a development. The Eskimo reflexive possessive in
‘i acts on a preceding stem lke a single consonant and
so is underivingly c which with word-fial nasalization
gave “71 and later “mi (Aleut -m); this Is proved by the
‘dual *20b (Aleut O¢x) from intervocalic *¢*y Rasmussen
1979:370. The full array k/m, fc vs. TE *b,/m.t, 8 const
tutes incontrovertible proof of the Nostratic genetic rela-
tionship in the classical Neogrammarian sense of the
terms
1416. Dual in a velar occurs in many places, often
accompanied by a plural marked by a dental. The com-
bined weight is considerable, indeed in itself enough to
prove the relationship to any unbiased observer. Thus,
Uralic pl ris combined with a du. in -7in Vogul, and in
‘other velars in Ostyak and Samoyed, and the Ist and 2nd
pl. of the verb end in “mek, “ek. This is remarkably
Close to the picture of Chukchi which also has pl. with
nouns and verbal endings -mdk, -tak. It is shockingly
close, abo, to TE “me, “te which disagree with the plural
marking of nouns in *s or *4 (both from *-8), but fully
shyme with the dual type ‘pdytér-e ‘two fathers’. It would
seem that in the verbal endings, which were properly
“poy. “yey, Bm, “54-5 (ie. predicative agent noun
with adequate number marking followed by enclitic pro-
nouns with their own number markings, cf. Rasmussen
19872:100 [1999:267f), the redundant second number
smarking was standardized using the form of the dual,
This looks very much like a common innovation of TE,
The REVIEW OF ARCHAEOLOGY i
Unalic and Chukotian. The author's presentation is some-
‘what compromised by his suggestion that the Armenian
pi. & (from IE *s) continues a dual sk. Mistaken 1s2so
the assessment of the Sirenik Eskimo plural marker
which is simply the word-inal development of “4; with
regular vowel weakening a form like ‘pana- ‘spears’
yields ‘pand} > Sir. pangf this is no ~4 plural” (104) on a
par with Lat. bi, bae, baec. A palatal shade of the plural
‘marker is further indicated by the Uralic 4/4 alternation
Ginn. talo-, inessive talo-tssa ‘in the houses’), even if
the niles are not congruent with those for IE *s/.
23. The case for an *absolutive k” is particularly weak.
‘The term is taken from Eskimo where an inflection
‘arnag ‘woman’, exgative *aknam, pl. “aknat gives the
impression of a suffixed -q in the non-ergative Cabsolu-
tive") form. However, the gemination caused on single
‘consonants, as *alug, *allum, “allut ‘footsole’, demands
aan underlying stem ending in a uvular consonant, i.e.
“arnak, *alus- (> PEsk. *q with normal word final treat-
ment), also before possessive suffixes, 8.
Chaplino aenaysi, aluy-st ‘your (pl.) woman, oar blade’
1s opposed to pana-zi ‘your spear’. Due to simplification
Of geminates in WEsk. and Aleut, the uvular has come to
bbe widely perceived as a semantically empty suffix of
‘erratic occurrence. That there are noun-forming velars in
Uralic, Turkic, Chukotian and Gilyak is not so impressive
if the common function is as vague as that: And even if
Finn, sade’ ‘rain’ (vb. sata), Altai taradk ‘a comb’ (vb.
tara), Kamchadal Cirsug thief Gir ‘steal, Gilyak
‘asgavk ‘younger brother’ (asga- ‘be younger’), or some
Of them, really do contain the same suffix, one woriders
how it has come to be perceived by Greenberg as a petr-
fied article.
724, The accusative in -m is one of the most striking
resemblances between IE and Uralic. One will not like to
dismiss it, and equally would one like to find a compara:
‘dum for the Esk-Al. ergative marker which is also °-mr
It may be suggested that the two m-cases are genetically
‘identical, viz. an original genitive. Ifthe verbal stem is an
‘old agent noun, its object is only expected to be in the
‘genitive (ike Eng. ber lover meaning “the one who loves
her), The reinterpretation of the genitival complement
Of the agent noun as a direct object would then be a
‘most interesting innovation common to TE and Uralic,
40, The curious resemblance between the synonymous
‘Turkic 4f and Eskimo 413 ‘having ~, pointed out as coinci-
dental by Menovitikov 1962:103, is expanded to include
the possessive adjective of Lydian, eg. (mtr) manelid
‘that of Mane", and the Common Anatolian pronominal
genitive in 1 as Hitt. ammiél ‘of me’. However, there can
be little doubt that Lydian + is simply based on the Iat-
ter, just like the Luwian poss. adj. in ~aSS4- is based on
the gen. case-ending seen in Hit. (IE os). The origi
nal locus of the genitive is probably precisely the 15g
Pronoun: The IE possessive pronouns were derived by
*vddhi" Gle., vowel insertion + thematicization) from
the stems of the personal pronouns, eg. *tué ‘thee’ =>
“teués ‘thy’; with ‘me’, however, the original form “mué
‘was regularly simplified to *mé’so that, in this case, a
“vpddhi" vowel had to be prefixed, the result being TE
tems ‘my’; endingless forms of the possessive adjec-
tives were used as genitive case-forms of the pronouns
Gwith initial accent, i.e. probably substantivized), eg.
*téye ‘of thee’; the corresponding 15g *éme apparently
took over the initial °m- from other case-forms like *mé,
this giving *méme of which the actually reconstructible
PIE ‘méne (Avestan mana, OCS mene) must be a dissim-
lated form. Hitt. ammél will be regular from *emele, an12 The REVIEW OF ARCHAEOLOGY
‘obvious dissimilation of ‘emene, itself a blend of ‘emés
and ‘méne. This, then, hardly reflects an old possessive
morpheme, nor does *méne itself appear to contain an
Old genitive -m, even if such a thing may be recoverable
from. eg, Chuvash bi/minu ‘Vy’ and Kamchadal
e/kien “~who/whose’ (n0. 25).
43. A passive participle in , wellknown from IE “to-
(alternating with *n0), is found also in Uralic and, most
clearly, Eskimo. The last distinguishes an active ptc, in
ugh Bug (stem “tuk-P-dus) anda passive pic. in
tagr-baq (stem “4an-/-bak- Greenlandic neri-oq
ing’, nerfsaq ‘eaten’. In Rasmussen ms. arguments are
given for identification of the active form with the TE
actionnoun type in plain “+ (Skt. isubbjt- “carrying
arrows’), and of the 3rd sg, possessive of the passive
form with the TE passive pte. in “to, “birt. ‘Carried’
‘meaning properly ‘belonging to the carrying one’. Ifthe
active form in “tis retained as the 3sg finite verb,
‘object in ‘im may indeed be an old genitive (ef. ad 24.
44, Participle nt: The verbal derivative used in the
‘Vogul present periphrasis,e 8. miritant-em ‘lam going’
highly reminiscent of the IE active participle, prs.
“bléronts ‘carrying’. However, itis hard to see what is
‘gained by analyzing the TE 3rd plural finite present in
“ni as a neuter plural ofthe participle, just because this
is -nti in Sanskrit; as the author notes, this is mostly
equated with Gk. dépovta and derived from IE *n-d
‘True, the development of schwa to Skt. /i/ has been
attacked by Burrow 1979, but if the arguments had been
any good (and we are not aware of any), wouldn't the
author have quoted one?
3. A conative sk is attempted for the IE inchoative
present suffix *she/o- and a FennoUgric companion
posited as “3, the latter being functionally most a fre
{quentative, but also conative. It should of course not be
Suppressed that the IE suffix i also mostly iterative, but
that shade appears to adhere originally to the accompa:
nying reduplication (Gk. StB4oxw ‘teach, ie, begin t0
‘cach again and again, teach litle by tle; but reduplica-
tion may disappear (as it has in the Hittite iterative sk-
verbs). Given the general shortness of suffixes, the
‘Palatality of the FU sibilant and the IE velar attracts atten-
tion and makes this equation look better than most.
(6062. Interrogatives have kj, m. In IE and Uralic, the
‘J forms have become relative (EE “{6s ‘he who’, a most
remarkable common innovation duly pointed out by the
author 225),
‘Adding the existence of a fourth interrogative marker,
viz. m (n0. 64), there is some possibility that the entire
system of interrogatives has very deep roots combining
it with that of other pronouns: Is it a pure coincidence
that the Ist, 2nd and 3rd person markers are m/k, t/n,
3/¥, and the demonstrative stem nuclei may also be
posited as m/k, i/n, s/V on a scale of falling proximity,
and even the interrogatives exhibit m/le and n? Could it
be that the persons were simply marked by morphemes
indicating the degree of closeness to the speaker? And,
finally, is it possible that the interrogatives are simply
ddemonstratives used in questions, perhaps on an ade-
uate tone, of the type “(and) this one?”, meaning “who
she?” —7 What is the chance that the three sets would
just happen to display, in full or in part, a variation
employing the same alternants, albeit not by common.
rules? [FC]
‘The final main chapter on Ainu vowel alternations is
beyond the scope of this review. We only note that the
author finds support for a “northern Asiatic connection”
in the facts he has adduced, esp. the morphemes of dual
Vol. 20, Number 2
k( and plural 1() which are presuasively characterized.
as decisive and diagnostic.
‘A concluding chart summarizes the results as giving
eight main subfamilies of Eurasiatic: Etruscan, 1E, Uralic-
Yukaghir, Altaic, KoreanJapanese-Ainu, Gilyak, Chukot
jan, and Eskimo-Aleut. Miller 1996 supplies very impres-
sive additional arguments for the inclusion of Korean and
Japanese into Altaic under a strict Neogrammarian con-
‘cept of genetic relationship. Thus. itis highly possible,
indeed expected, that future research will simplify the
taxonomic picture of this distant relationship which,