IE-VerbalInflection.htm
Indo-European
Verbal Inflection
and
Nominal Declension
by
Patrick C. Ryan
11/
26
/2004
T
he premise of this essay is that the pattern of Indo-European verbal inflections
is
analyzable in terms of current linguistic theory.
The assumptions made are the following:
1) There was a stage of Indo-European (or a language ancestral to it), in which the
only
vowels were
e
,
e:
,
a:
, and
o:
; and the apparent vowels,
i
(derived from
y
) and
u
(derived from
w
) in
reduced
-grade
syllables.
2) During this stage, the
only
verbal inflection was an alternation of iambic (
v
-
) accentuation, which conveyed
momentariness
, opposed to a trochaic (
-
v
) accentuation, which conveyed
durativity
(
Lehmann 1974:186
);
a)
This resulted in two-forms for
CVCV
verbal roots:
CV'CV
,
momentary
;
'CVC(Ø)
,
durative
.
3) At this stage, Indo-European (or a language ancestral to it) was ergative (
Beekes 1995:193-5
), containing two core cases:
ergative
and
absolutive
:
a) For pronouns, an ergative series: 1
st
pers.:
*me
('converser'); 2
nd
pers.:
*se
('clansman'); 3
rd
pers.:
*te
('tribesman');
1)) Singular and (dual and) plural were not distinguished at this point;
b) For pronouns, an
absolutive series
: 1
st
pers.:
*a:
('here'); 2
nd
pers.:
*o:
('there'); 3
rd
pers.:
*e:
('there yonder').
4) These developments yielded two
aspectual stems
for the very early Indo-European verb:
a)
Momentary Stem
:
*widhé
, 'place in the ear' = 'learn';
1)) There is, properly,
no
zero-grade but rather only a reduced grade, occasioned by the absence of the primary stress-accent, that
may
only eliminate the vowel if a helping vowel is available [
e.g.
*é
(
:
)-] that precedes the first consonant; or, if the first consonant is followed by a resonant (
l
,
m
,
n
,
r
) or a semi-vowel (
y
,
w
, which are then vocalized).
b)
Durative Stem
:
*wéidh
, 'have placed in the ear' = 'know'.
5) The two aspectual stems could be construed with either of the core cases of the pronoun, which cross-referenced respectively the ergative agent and absolutive patient:
a) When the momentary stem was combined with an ergative cross-reference, the thematic aorist
proto
-form
**widhé-m
(
Beekes 1995:235-7
), '(I) learn of' = '(I) find', was produced. This is directly attested only in combination with
*e
(
:
)
-
.
b) When the momentary stem was combined with an absolutive cross-reference, the perfect form
*widhé-a:
(
Beekes 1995:239-44
), 'am learned of' = 'understand', was produced, the
-a:
of which was subsequently shortened due to the absence of stress-accentuation; and the stress-accent of which was subsequently retracted, producing secondarily stress-accented
ó
in the first syllable:
*wóidh-a
(contrary to
Lehmann 1974:186
, who considers the perfect
+durative
).
c) When the durative stem was combined with an ergative cross-reference, the athematic (root) aorist proto-form (injunctive)
*wéidh-m
(
Beekes 1995:235-7
), 'can know' = 'understand', was produced, many of which were later "thematized" for phonological reasons:
*wéidh(-e)-m
.
d) When the durative stem was combined with an absolutive cross-reference, the middle form
*wéidh-a:
(
Beekes 1995:239-44
), 'know myself' = 'am aware', was produced, the
-a:
of which was subsequently shortened due to the absence of stress-accentuation:
*wéidh-a
.
e) When the durative stem was combined with an ergative cross-reference
and
an absolutive cross-reference, the transitive middle form
*wéidh-m-a:
(
Beekes 1995:239-44
), 'know of / for myself' = 'know', was produced, the
-a:
of which was subsequently shortened:
*wéidh-ma
.
6) The adverb
*é
(
:
)
-
, 'then (
Beekes 1995:226
; derived from 'there yonder')', could be prefixed to
any
of the above described forms, and signified
any
time period
other than
the immediate present.
7) The adjective formant
*-y
, 'like', could be suffixed to
any
of the above described (secondary) forms to produce the primary series of forms, and signified
progressivity
(contrary to
Lehmann 1974:99-100
, who considers it
+declarative
). This is the same formant found in the Indo-European verbal noun (
Beekes 1995:251-2
); see 11) below.
8) The adjective formant
**s(e)-'
, '**strong (from Proto-Language
SA
)', could be prefixed to
any
of the above described forms, and signified
intensivity
; this is the Indo-European
s-mobile
.
9) The adjective formant
'-*s(e)
, '*for one's self (from Proto-Language
SE
)', could be suffixed to a
durative
stem form and combined with an ergative cross-reference to produce the
s(igmatic)-
aorist
(
Beekes 1995:235-7
), which was properly a middle, and which signified a
beneficial reflexive
. In addition, this formant can be seen in "presents" in
*-s
like Greek
aékso:
, 'I grow'; and in
desideratives
and infinitives in
*-s
.
10) A proper reflexive in
*-'dhi
, '*self (from Proto-Language
T[?]SA
-¿E
, 'body-like' = 'self') is found as a formant in the 2
nd
pers. sing. of the imperative of (usually momentary-stem) intransitive verbs
11) The adverbial particle
*-ye:
, 'yea', which Beekes reconstructs as "
-ieh
1
" (
Beekes 1995:246
) and Lehmann as "
-ye?
", could be suffixed to durative or momentary stems to form the
optative
, which, according to
Lehmann (1974:135-9)
, has a "voluntative" basal meaning. Indo-European
*-ye:
is derived from Proto-Language
¿E
-
?A
, 'say-stative' = 'said'. If this is its original meaning, the phrases interpreted by Lehmann can be seen as more accurately promissory or affirmative, and this explains the optative used as "potential". This also explains the lack of the optative in Hittite, the function of which was fulfilled by the enclitic
-ma
, 'but, however', derived from Indo-European
*sme/e:
.
12) The verbal formant
*-e
(
:
)(
i
), 'go(ing) to', could be suffixed to the durative stem (and therefore, contrary to Beekes
(1995:245)
, is
a
thematic
) to produce a primary series of forms, and signified
subjunctive
, which, according to
Lehmann (1974:130-5)
, has a "obligative" basal meaning. Indo-European
*-e
(
:
)(
i
) is derived from Proto-Language
HHE
-
¿E
), 'go(ing) to', which is seen in Indo-European
*1. ei-
. If this is its original meaning, the phrases interpreted by Lehmann can be seen as more accurately simply predictory, and this explains the subjunctive used as "future"; it portrays an action as imminently to-begin/to-be-begun; and is therefore properly
inceptive
, uncannily parallel with modern English 'is going to . . .'
13) The adverbial particle
*-t/eo
, 'nearby, soon', which is derived from Proto-Language
T[H]O
, 'accompany', could be suffixed to imperative forms,
e.g.
Old Indian
i-tá
, '(soon) go (plural)' [the stress-accent of the termination is the sole mark of the plural].
14) The adverbial particle
*-u
, 'again, ever', could be suffixed to imperative forms,
e.g.
Old Indian
é-t-u
, '(soon) let him (ever) go'.
This interpretation of the data is somewhat similar to that provided by Lehmann (
1974:186
); it is different in characterizing the "
m s t
" endings as
ergative cross-referential
rather than indicating (person +) "imperfective" and "
x the ø
" [for Ryan's
*-a
(
:
),
*-o
(
:
), and
*-e
(
:
); Beekes'
*-H
2
(e)
,
*-tH
2
e/o
, and
*-e/o
)] as
absolutive cross-referential
rather than indicating (person +) "perfective" aspect.
The explanation for the discrepancy between my reconstruction and those of Lehmann and Beekes is that, after a shortening of
*-a
(
:
),
*-o
(
:
), and
*-e
(
:
),
a
was able to maintain its phonemic distinctiveness while
*-o
and
*-e
were
Ablaut
allophonic variations of a phoneme that were semantically undistinctive; accordingly, a new method of distinguishing the 2
nd
and 3
rd
persons was introduced by combining
*-'t(e/)o
, which Beekes terms a
"verbal adjective" (1995:250-51)
formant with the earlier inflection [
*-
(
H
)(
e
/)
o
(
:
)], yielding (almost) Beekes'
*-tH
(2)
e/o
(the subscript
2
, indicating an
a
-coloring "laryngeal" is incorrect). This
*-'t(e/)o
was derived from Proto-Language
T[H]O
, properly an
iterative
, which acquired a resultative nuance when combined with the
durative stem
(Beekes' "root"). That this is very old is suggested strongly by the employment of
-t(j)
in the 2
nd
persons of the Old Perfective in Egyptian.
However, there is an intimate connection in many languages between expressed ergatively marked agents and perfectivity/pastness (
Dixon 1998:97-101
) and definiteness.
I conjecture that the pattern presently observable in Burushaski (
Dixon 1998:99-100
) is plausibly similar to the situation we should expect for the earliest Indo-European:
1) For an imperfective transitive statement, an
antipassive
(
Dixon 1998:97-101
) construction was employed:
*k^wón
(1)
ek^
w
ó-m
(2)
éd-e
(
:
)
(S)dog+ABS(olutive) (O)horse+SUPERE(ssive)
(2)
(V)eat+DUR(ative)+ABS(olutive cross-reference)
A dog eats (on a) horse.
2) For a perfective transitive statement, an
ergative
(
Dixon 1998:8ff.
) construction was employed:
*k^wón-s
(3)
(4)
(becomes
k^wó:n
)
ék^
w
e
(4)
edé-t
(S)dog+ERG(ative) [from
*-s(e)
, COM(itative)
(4)
] (O)horse+ABS(olutive)
(V)eat+MOM(entary)+ERG(ative cross-reference)
The dog eats (up) the horse.
3) For an imperfective intransitive statement, an
absolutive
(
Dixon 1998:8ff.
) construction was employed:
*k^wón
(1)
(4)
éd-e
(
:
)
(S)dog+ABS(olutive) (V)eat+DUR(ative)+ABS(olutive cross-reference)
A dog is eaten.
4) For a perfective intransitive statement, an
ergative
(
Dixon 1998:8ff.
) construction was employed:
*k^wón-s
(3)
( becomes
k^wó:n
)
edé-t
(S)dog+ERG(ative) [from
*-s(e)
, COM(itative)]
(V)eat+MOM(entary)+ERG(ative cross-reference)
The dog eats up.
the latest revision of this document can be found at
HTTP://WWW.GEOCITIES.COM/Athens/Forum/2803/IE-VerbalInflection.htm
Patrick C. Ryan * 9115 West 34th Street - Little Rock, AR 72204-4441 * (501)227-9947
PROTO
-LANGUAGE@email.msn.com
NOTES
1.
From earlier
*k^éw(e)
+
-(H
1
e)n
(
Beekes 1995:169-70
), with a shift of the stress-accent to the right on account of the suffix and secondarily stress-accented
-*é-
becoming
-ó-
:
*k^(ø)wón
.
2.
cf.
Gaelic
má
, 'at (from Proto-Language
M[H]A
)'.
3.
At this stage of development, the "genitive" in
*-
(
o
)
s
, from which the later ergative (and nominative) is derived (
Beekes 1995:194
), will still have been properly a comitative (
v.
4.
below), and the earliest genitive will have simply been an adjective formed with
*-y
, '-like (from Proto-Language
¿E
)', that may still be seen in
locative
*-(o)i
, and in the retained genitive
*-i:
(from
V
+
y
) in Latin, Celtic, and Illyrian-Messapian; and, most importantly, in consonant-stem duals (
v.
6.
below).
4.
The earliest
absolutive
form of the normally reconstructed Indo-European
*ekwo-s
('grass-
chew[ing]') will have been
*ék
w
e
(which is most clearly seen in the vocative of the
o
-stems; and in the "accusative (from absolutive)" of personal pronouns:
e.g.
me
,
te
), to which
comitative
*-s
(
e
) (
cf.
Russian
s
[
o
], 'with'; from Proto-Language
S[H]O
) was added, drawing the stress-accent one syllable to the right, and changing secondarily stress-accented
-*é-
to
-ó
(
*ek
w
ó-s
); subsequently, the primary stress-accent was transferred to the first syllable:
*ék
w
o-s
. For
*ék
w
o:s
, 'horses', see
5.
below.
5.
As we have seen in
4.
above, the earliest absolutive form of Indo-European
*ekwo-s
('grass-
chew[ing]' = 'horse') will have been
*ék
w
e
, seen in the vocative of the
o
-stems and in consonant stems as
*-ø
; with the comitative formant
*-s
(
e
) (Proto-Language
S[H]O
), a genitive that supplanted the earlier genitive in
*-y
that still appears in the declension of
o-
stems as locative
*-oi
and consonant stems as
*-i
was formed. When the new genitive was assigned an
ergative
(later nominative) function, this produced a situation where the nominative and genitive forms were identical: both
*ék
w
o-s
. Having observed this pattern, it would be natural to suppose that a similar correspondence should obtain in the plural, where we, however, find nominative
ék
w
o:-s
but genitive
ék
w
o:m
. Showing a lengthened
o:
, we also have the instrumental plural in
*-o:is
. Now, it is well-known that the neuter plural is oftenest
-a:
, which I derive from Proto-Language
¿A
, 'many' (it may be that the animate plural was built with
¿A
, '(nuclear) family', which would have had the same phonetic result in Indo-European:
a:
). We should keep firmly in mind that the major difference between animates and inanimates in general is that an inanimate is not likely to appear in a case form functioning as an ergative agent. If this formant was also used for non-neuters such as
*ék
w
e
, the resulting plural absolutive would have been
*ék
w
e-a:
, which, because of the retention of the stress-accent on the first syllable in nominal compounds, would have resulted in
*ék
w
e:
(through
*ék
w
e-a
). Forming a
comitative plural
with
*-s(e)
would have drawn the stress-accent one syllable to the right, yielding
*ek
w
é:s
; and finally moving the stress-accent to the first syllable would have yielded
*ék
w
o:s
, which would later, on the pattern of the singular, have functioned as first an ergative then later as a nominative form. This lengthening of the final vowel to indicate the plural must be very old because, as Moscati (
1969:87
) writes of the endings of the external masculine plural (nominative
-u:
; genitive/accusative
-i:
): "These endings seem to be the result of the lengthening of the corresponding singular morphemes of the nominative (
-u
) and genitive (
-i
)." The
genitive plural
in
*-o:m
(subsequently in most dialects shortened to
*-om
) could, therefore, be simply explained as a plural superessive, corresponding to singular accusative
*-om
(
singular superessive
). After the shortening of
*-o:m
to
*-om
, which came to be regarded generally as an oblique plural stem (probably helped by the frequency of neuter stems in
-m
from a different origin),
*-o:s
was falsely analyzed as consisting of a plural formant
*-s
, and this new "plural" formant was added to the new superessive to form
*ék
w
oms
, which was euphonically assimilated (apical for labial in contact with apical) to
*ék
w
ons
. The earliest
genitive plural
,
*ek
w
é-a:-i
, may be seen in the Greek nominative plural
-oi
after shifting of the stress-accent (
*ék
w
o:i
), and shortening of the vowel, partially to help distinguish it from the
singular dative
in
*-o:i
from
V
+
*-e:i
[Indo-European
*1. ei-
, 'go (to)'; Proto-Language
HHE
-¿E
, '(move up) from-like=progressive' = '(going) to'].
Had the
dative plural
been formed on the earliest absolutive plural (
*ék
w
e:
), the result would have been
*ék
w
e:
+
*-e:i
, or
*ek
w
é:i
, which, like the dative singular, would have naturally developed into
*ék
w
o:i
. As a result, a new method for forming an unambiguous dative plural needed to be found, and the solution selected, in the
o
-stems, was to build a dative plural on the superessive, which had come to be regarded as a general oblique plural stem (
*ék
w
o
(
:
)
m
) by adding the falsely analyzed plural formant
*-s
and a case formative,
*-u
(Proto-Language
FA
, '(a)round'), which is properly a
circumlocative
or topical inflection seen in Indo-European pronominal forms (
e.g.
tu/u:
) , producing
*ék
w
o
(
:
)
m-u-s
in some dialects; in others, a dative plural was built on the proper genitive plural,
*ek
w
é:i
by adding the falsely analyzed plural formant
*-(e)s
and a
subessive
case formative,
*-bhi
(Proto-Language
P[?]FE
-¿E
, 'foot-like = below'), listed in
Pokorny 1959:34-35
under
*ambhi
, 'around, on both sides', and which apparently took the stress-accent:
*ek
w
o:i-bhy-ás
before the stress-accent was regularized on the first syllable (
*ék
w
o
(
:
)
i-bhy-as
).
The
instrumental
singular can be seen to be
*-e:
(Beekes:
*-oH
1
), listed in
Pokorny 1959:34-35
under
*1. e/e:, o/o:
, 'away', which I attribute to Proto-Language
HE
, '(move across) from' = '(come) from'. When added to the singular absolutive
*ék
w
e
, this would have yielded
*ek
w
é-e:
, which, when resolved to
*ek
w
é:
, would have resulted in the same form as the resolved absolutive plural (
*ék
w
o:
). The form was retained for the singular instrumental (and probably the earliest ergative) but this necessitated substitution of the genitive plural (
*ék
w
o
(
:
)
i
) for the now undistinctive
instrumental (ergative) plural
:
*ek
w
é-a:-e:
becomes
*ek
w
é:
then
*ék
w
o:
. Obviously, this necessitated a new formation for the instrumental plural, which was constructed on the proper genitive plural,
*ek
w
é:i
by adding the comitative formant
*-s(e)
:
*ek
w
é:i-s
becoming
*ék
w
o:i-s
.
locative plural
**ek
w
é-a:-a:
, when resolved (becomes
**ek
w
é:
then
**ék
w
o:
) required a new form which was constructed on the proper genitive plural,
*ek
w
é:i
by adding the falsely analyzed plural formant
*-s
and the circumlocative formant
-u
discussed above.
The earliest
locative
in
*-a:
(Beekes: instrumental
*-oH
1
), listed in
Pokorny 1959:34-35
under
*1. e/e:, o/o:
, 'nearby', which I attribute to Proto-Language
?A
, 'presence' = 'at'. When added to the singular absolutive
*ék
w
e
, this would have yielded
*ek
w
é-a:
, which, when resolved to
*ek
w
é:
, would have resulted in the same form as the resolved absolutive plural (
*ék
w
o:
). The form was retained for the singular instrumental (and probably the earliest ergative) but this necessitated substitution of the earliest genitive singular
*ék
w
oi
) for the now undistinctive locative singular
after
the comitative singular took its place.
Therefore, the earliest Indo-European nominal declension may be presented in the following table as follows:
SINGULAR
PLURAL
Absolutive (Vocative):
**ék
w
e
(
*ék
w
om
as accusative)
Absolutive (Vocative):
**ék
w
e-a:
(
*ék
w
ons
; neuter *ék
w
a:)
Ergative:
the earliest ergative was probably identical with the instrumental (
**ek
w
é-e:
)
(
*ék
w
os
)
Ergative:
the earliest ergative was probably identical with the instrumental (
**ek
w
é-a:-e:
)
(
*ék
w
o:s
)
Genitive:
**ek
w
é-i
(
*ék
w
os
)
Genitive:
**ek
w
é-a:-i
(
*ék
w
o:m
)
Dative:
**ek
w
é-e:i
(
*ék
w
o:i
)
Dative:
**ek
w
é-a:-e:i
(
*ék
w
omus
)
Superessive:
**ek
w
é-m
(
*ék
w
om
)
Superessive:
**ek
w
é-a:-m
(
*ék
w
o:m
)
Locative:
**ek
w
é-a:
(
*ék
w
oi
)
Locative:
**ek
w
é-a:-a:
(
*ék
w
oisu
)
Circumlocative (Topical):
**ek
w
é-u
Circumlocative (Topical):
**ek
w
é-a:-u
Subessive:
**ek
w
é-bhi
Subessive:
**ek
w
é-a:-bhi
Comitative:
**ek
w
é-s
(
**ék
w
os
)
Comitative:
**ek
w
é-a:-s
(
*ék
w
o:s
)
Instrumental:
**ek
w
é-e:
(
*ék
w
o:
)
Instrumental:
**ek
w
é-a:-e:
(
*ék
w
o:i-s
)
Ablative:
**ek
w
é-de:
(7)
(
*ék
w
o:d
)
Ablative:
**ek
w
é-a:-de:
(
*
various
)
6.
Beekes' reconstruction for the nominative neuter dual is "
*-h
1
(
Beekes 1995:194
)", which I would emend to simply
*-y
, the earliest genitive formant of IE. A quick look at his table of reconstructions will reveal that, in a number of daughter languages,
*-y
is also appropriate for masculine-feminine prototypes (
e.g.
Greek nominative dual
-ei
for
u
-stems, which is impossible to derive from Beekes' postulated
*-h
1
.
I am not sure if a consensus can be reached on the earliest Indo-European term for 'pair' but perhaps
*dwi-n-
might be provisionally accepted. The Indo-European dual is a relatively late formation; and I suggest that it developed from a phrase of the form:
*dwiné ek
w
éi
, 'pair of horse', from which, after the demise of the regular genitive in
*-y
,
*dwin
could be deleted since the form was distinctive (
ek
w
éi
), 'horses (dual)'.
A second method of forming the dual accounts for the major part of the other dual formations. As I mentioned above, the earliest Indo-European absolutive plural had
*-a:
:
*ék
w
a:
, 'horse-family' = 'horses'. When this plural formant became obsolete except for neuters, it became available as a dual for masculine-feminine nouns (but not for neuters since neuters continued to employ
*-a:
as a plural; consequently, had been most conveniently adopting
*-y
). Where we should see this most clearly is in the masculine-feminine consonantal stems; and, in fact, Sanskrit
-a:
and Avestan
-a
confirm the analysis. However, it can be seen clearly in the
o
-stems in Sanskrit and Avestan also:
-a:
and
-a
, the same formants we find in the
neuter
nominative
plural
. It is, to me, inexplicable that Beekes should reconstruct the latter as deriving from
*h
2
and the former from
*h
1
.
7.
The ablative singular is like the dative singular in being recognized by two effects: a lengthening of the stem-vowel and an additional consonant. Pokorny (
1959:182-83
) identified the ablative ending as a metathesis of
*de:
, 'from', listed incorrectly under
*de
, demonstrative stem. I consider
*de:
to be the Indo-European reflex of Proto-Language
T[?]A
-HE
, 'hand-(move across) from' = '(come) from'. In its earliest form, the ablative singular was
**ek
w
é-de:
; for reasons that may be purely euphonic, this was metathesized to
**ek
w
é-e:d
, which resolved to
**ek
w
é:d
, and finally, with shift of the stress-accent to the root syllable:
**ék
w
o:d
. The earliest ablative plural will have been
**ek
w
é-a:-de:
. This was first resolved to
**ek
w
é:-de:
; second, to
**ek
w
é:d
with metathesis. It was, therefore, identical in form to the ablative singular, and could not be distinctive. For reasons that are unclear, various
dative plural
formations of diverse origins were substituted for the undistinctive resolved ablative plural. A new ablative plural might have been formed with the
new plural formant
*-s
, yielding
**-de:s
if
de:-
had still been employed as a verb (meaning probably 'go from the possession of';
cf.
Egyptian
dj
, 'give' = 'cause to go from [the possession of]') but apparently
de:
had been lost as a verb in Indo-European, and was restricted to being an adverbial element. There were, undoubtedly, other cases in Proto-European that were lost so completely that they cannot be justifiably projected back to its early date: Greek
-de
, 'towards', probably represents Proto-Language
T[?]E
-HHE
, 'heel-'(move up) from' = '(going) to' = 'hurry after', seen expanded by
-¿E
, durative, in
*2. dey6-
(for
**deHey-
); Greek
-thi
, 'at', probably represents Proto-Language
T[?]SA
-HHE
, 'side'-'(move up) from' = '(going) to' = 'cause to go to the side of' = 'put beside', seen expanded by
-¿E
, durative, in
*2. dhe:(y)-
(
i.e.
**dheHey-
), 'set, stand, lay'. The elusive
circumlocative
can apparently be seen in Greek
-ou
, 'at' (from 'around'). Even
the earliest locative in
*-a:
can be glimpsed in Greek words like
háma
, 'at the same time'.