From Proto-Indo-European to Indo-European:

After the Great Voice Shift and the Morphotactic Fusion

Roland A. Pooth*
Abstract. This is a very brief outline of what must have happened to the PIE basic ver-
bal inflectional types (“root” formations) after PIE broke up
Keywords. PIE verb inflection, Linguistic reconstruction
*Ghent University, mail to [email protected], [email protected]

1 PIE basic verbal inflectional types

In Pooth (2016b) we reconstruct six basic verbal inflectional types for Proto-
Indo-European (PIE) encoding different aspectual categories (aspects). We use the
term binyan (pl. binyans), which is borrowed from Classical Hebrew grammar (pl.
binyānî́m). Each PIE binyan was derived from the underlying inflectable base by
means of transfixal vowel melodies (e.g. _é_o_) and “root” and pattern inflection.
The inflectable base was most often a discontinuous consonantal skeleton (e.g.
*st_u-, maybe *s-t_u-), but there were a few continuous a-roots (cf. Pooth
2015[6]). Notice that “root” is not a proper term for this unit in all cases, since
inflectional bases that carried derivational affixes (the prefix s- or a so-called
“root enlargement”) were derivational bases, rather than roots. The six PIE basic
aspect binyans are listed in Table 1.

2SG/3SG 3SG
DIRECT (-t-) INTRANSTIVE
ACTIVE DETRANSITIVE aspectual value
wh wh
I *g ént(i) *g nó(i) depending on the aspect of the verb
II *stḗut(i) *stéuo(i) DURATIVE-IMPERFECTIVE (DUR)
III *uóide(i) STATIVE-HABITUAL (STAT)
IV *uóid(i) TRANSITIONAL (CHANGE OF STATE) (TRIT)
V *suṓp(i) INCHOATIVE-STATIVE (ISTA)
VI *suopé(i) DISTRIBUTIVE-ITERATIVE (DSTR)

Table 1. PIE basic aspectual categories I-VI

The so-called “Narten type” is analyzed as a DURATIVE-IMPERFECTIVE (DUR)

aspect (cf. Kümmel 1998). The word-final suffix -i additionally marked the
PROGRESSIVE aspect, as the present progressive -ing form in English (cf. Pooth
2009b). It must have been used less frequently with terminative and telic verbs
 Roland A. Pooth 2

(e.g. *gwem- ‘to come hither, go thither’ or *dheh1- ‘to put sth. swh. make, do’). It
was perhaps not used with specifically punctual verbs.
We can see that the vowel o of the forms of the right column of Table 1 must
have been a DETRANSITIVIZING (DETRANSITIVE) or “protomiddle” marker. This
vowel was morphotactically independent from the consonantal frame or discon-
tinuous morphological skeleton surrounding it (cf. Pooth 2016b).

2 The PIE antipassive construction

The intransitive detransitive forms of all PIE binyans, e.g. *gwhnó(i), *stéuo(i),
*uóide(i), *uóid(i), *suṓp(i), *suopé(i), were detransitive forms that had to be
used within a syntactically intransitive construction including only one core case,
either the sigmatic agent case (-s) or the absolutive case (cf. Pooth 2015b). How-
ever, these forms could be used within a specific intransitive antipassive construc-
tion including an extended core ALLATIVE-DATIVE or LOCATIVE case, which was
attached to the intransitive core. Even if this construction lacked a specific anti-
passive marker, the pattern clearly corresponds to the typical antipassive patterns
among the languages of the world. Therefore, we simply call it the PIE “antipas-
sive construction”. Notice that alongside its aspectual meaning, forms of binyan II
(*stéuo type) had a modal ABILITATIVE (ABIL) meaning and a modal HORTATIVE
(HORT) meaning, and could also be used in the sense of an a EXPECTATIVE-
PROSPECTIVE (EXPECT) future tense.

(1) *h2nér(-s) *stéuo(i) *diéu-m

man-(AGENT) praise:DUR:DTR:ITR\3SG sky-ALLATIVE
(a) DUR ‘man is praising (at) god’
(b) ABIL ‘man can (is able to) praise (at) god’
(c) EXPECT ‘man will (is expected to) praise (at) god’
(d) HORT ‘man shall (it is suggested/demanded that he) praise (at) god’

The intransitive detransitive forms of all PIE binyans were polysemous for the
causer and agent role in the sense that they could be used both semantically in-
transitively and transitively, both with a non-agentive causer or agentive causer
role, and without. Depending on the underlying verb meaning, they could be used
like “agentless” passives, but the addition of an oblique agent was ungrammatical,
as in many languages of the world that exhibit this pattern. With agentive transi-
tive meaning and concomitant agent case marking of the agentive causer they
could be used within the antipassive construction, as in (3).

(2) *diéu *stóue

sky.god(.ABS) praise:STAT:DTR:ITR\3SG
STAT ‘sky-god is always praised (at)’
 From Proto-Indo-European to Indo-European: After the Great Voice Shift and the Morphotactic Fusion 3

(3) *h2nér(-s) *stóue *diéu-m

man-(AGENT) praise:STATIVE:DTR:ITR\3SG sky.god-ALLATIVE
STAT ‘man always praises (at) sky-god’

3 The Great Voice Shift and the Morphotactic Fusion

PIE syntax changed dramatically due to the GREAT VOICE SHIFT and the shift
from PROGRESSIVE TO PRESENT IMPERFECTIVE. Our starting point for the change of
the PIE templatic root and pattern morphological system is based on a few cate-
gorical shifts, typically occurring among the languages of the world (cf. Bybee,
Perkins Pagliuca 1994), see (4). We can see that shifts 2a, 2b, and 2c are three
sub-developments of the emergence of a broad and more polysemous
IMPERFECTIVE (IPFV) aspect category and general tensedness, at least of present
imperfective forms.

(4a) 1 ANTIPASSIVE CONSTRUCTION TO NEOTRANSITIVE NOM-ACC

(4b) 2a PROGRESSIVE TO PRESENT IMPERFECTIVE (CUMULATIVE EXPONENCE)

2b DURATIVE-IMPERFECTIVE TO IMPERFECTIVE
2c STATIVE-HABITUAL TO IMPERFECTIVE

Our starting point for the fundamental change of the original PIE voice system
is the reanalysis of the antipassive construction to neotransitive, including reanal-
ysis of the PIE sigmatic agent case as new nominative (NOM) and the original PIE
allative-dative case (ALL) as new accusative (ACC):

(5) *h2nér(-s) *stóue(i) *diéu-m

man-(AGENT) praise:STAT:DTR:ITR\3SG sky.god-ALL
→ NOM praise:3SG.IPFV.IND.ACT sky.god-ACC

This shift had a very destructive effect on the transfixal marking system in that
the vowel o could not be analzyed as a discontinuous detransitive marker any
longer. The original transfixal marking system thus immediately collapsed. The
PIE direct (-t-) vs. inverse (-s-) marking system (cf. Pooth 2016b; Jacques & An-
tonov 2014) was also destroyed by the introduction of this new transitive con-
struction. The formal effect of shifts 1 and 2b and 2c was merger of the originally
detransitive stative-habitual inflectional type III *uóide with the original active
forms of the durative-imperfective inflectional type II, see (6).

(6) */h2nér-s/ *stóue ~ *stḗut */diéu-m/

NOM praise:3SG.IPFV.IND.ACT sky.god-ACC
 Roland A. Pooth 4

We reconstruct PIE 2sg and 3sg sigmatic (INVERSE) forms (e.g. *stḗus) (Pooth
2016b), which merged with the (DIRECT) 3sg forms in -t- and the 3sg forms of the
*uóide type due to shift 1. Alongside the shift from progressive aspect to present
imperfective, this merger can be considered a further step in the collapse of the
original PIE aspect system. The result of this merger was a “mixed paradigm”
*stḗut ~ *stḗus(t) ~ *stóue, as reconstructed by Jasanoff (2003). But recall that
Jasanoff’s reconstruction of a “mixed paradigm” has remained obscure on the
functional level. However, regarding the formal level of his model, we can follow
him, but claim that it is essentially post-PIE.
The type *uóide(i) was thus reanalyzed as belonging to the (neo-) active voice
by the time the original PIE detransitive voice-marking system collapsed. The
original protomiddle (presumably distributive-iterative) inflectional type *suopé(i)
became (neo-) active, too.

(7) *stóue(i) STAT → (neo-) active IMPERFECTIVE

*stoué(i) DSTR → (neo-) active (DSTR-) IMPERFECTIVE

The second shift that had a destructive effect on the original system was the
shift from PROGRESSIVE aspect to PRESENT IMPERFECTIVE tense and aspect, intro-
ducing present-tensedness and cumulative exponence to these forms. It was ac-
companied by reanalysis of the original transfixal marking system of the final part
of the word form and its surrounding consonantal segmental string to a fusional
portmanteau ending (e.g. -t- and -i → -ti). Temporal and aspectual CUMULATIVE
(PRESENT + IMPERFECTIVE) EXPONENCE was thus introduced to the word final
segmental strings of active 3sg forms ending in -ti, -ei, and corresponding middle
endings -oi, -toi, etc. The emergence of linear inflectional endings is thus part of
the phenomenon that we term the Great Morphotactic Fusion here. It involves a
process of morphotactic fusion of one of the original transfixal vowels with the
consonantal string surrounding it (e.g. -t_o_i → -toi). After the shift from progres-
sive to present imperfective, present tensed imperfective endings (-ti and -ei) were
distinguished from general imperfective endings and non-progressive or neutral
endings (these are the so-called “secondary endings”). The two categories B
(IMPERFECTIVE) and C (NON-IMPERFECTIVE) were still tense-neutral (NON-TENSED)
and thus used for past time and future time reference.
After the fusion, leveling of the zero grade in imperfective plural forms (*stéu-
→ *stu-´, cf. Vedic stuvánti) took place, whereas the respective ablaut grade of
the singular forms was generalized in non-imperfective plural forms (*dhh1-´ →
*dhéh1-) to increasingly distinguish categories A and B from category C. The stem
allomorphs of the plural forms are given in brackets in Table 2. Notice that *su̯ ṓpi
must still have belonged to both categories A and B after the voice shift due to its
original PIE polysemous non-tensed meaning ‘is sleeping / falling asleep now’
 From Proto-Indo-European to Indo-European: After the Great Voice Shift and the Morphotactic Fusion 5

(present → A) and ‘was sleeping / falling asleep’ or ‘is always (generally, habitu-
ally) sleeping / falling asleep (→ B). PIE progressive forms of the transitional
aspect (binyan IV) must have been used as present and anterior imperfectives, e.g.
*g̑ ónh1i ‘is in the resulting state of having generated now; is (being) generated
now; is in the (present) state of just (recently) having been generated, has just
been generated’, much like IE perfects (e.g. Vedic ones). In anterior function, they
were narrowed to aorist (perfective) stems later and thus preceded the Vedic “pas-
sive aorist” (jáni, ájani).

NEOACTIVE RESIDUAL MIDDLE aspectual value

wh wh wh wh
A *g én-ti (*g n-´) *g n-ói ~ *g n̥ -tói PRESENT.IPFV
*dhéh1-ti (*dhh1-´) *dhh1-ói ~ *dhh1-tói
*stḗu-ti (*stéu- → *stu-´) *stéu̯ -oi ~ *stéu-toi
*u̯ óid-ei (*u̯ eid- → *u̯ id-´)
*su̯ op-éi
(*su̯ ṓp-i) *su̯ ṓp-i (*su̯ óp-)
*g̑ ónh1-i (*g̑ n̥ h1-´) (+ANTERIOR)
B *stḗu-t (*stéu- → *stu-´) *stéu̯ -o ~ *stéu-to IPFV
*stḗu-s (*stéu- → *stu-´)
*u̯ óid-e (*u̯ eid- → *u̯ id-´)
(*su̯ ṓp-i) *su̯ ṓp-i (*su̯ óp-)
*g̑ ónh1-i (*g̑ n̥ h1-´) (+ANTERIOR)
C *gwhén-t *gwhn-ó ~ *gwhn̥ -tó NON-PROG
*dhéh1-t (*dhh1-´ → *dhéh1-) *dhh1-ó ~ *dhh1-tó (→ NON-IPFV)
*g̑ ónh1
*u̯ id-ó

Table 2. The system after the shifts 1 and 2a-c.

4 Natural Consequences
From a morphotactic perspective, the subsequent developments within the sys-
tem are almost natural consequences of the given evolutionary tendencies. After
the great fusion, the following developments occurred immediately:

4.1 Paradigmatic leveling and blending processes

As a natural consequence of the paradigmatic merger of the (proto-) active
forms of type II and the detransitive forms of types III and VI, paradigmatic level-
 Roland A. Pooth 6

ing and blending processes took place affecting the roots and endings, as illustrat-
ed by Table 3, where emphasis in bold type marks the innovative forms. These
leveling and blending processes must necessarily be dated to a period after the
Great Morphotactic Fusion and the introduction of cumulative exponence. These
processes probably started very closely after the stage when PIE broke up into
dialects. The Great Voice Shift and the Morphotactic Fusion is dated here to
closely after the period where PIE was spoken as a unitary language, presumably
by a smaller number of people, before something happened that made dialects
move apart. We suggest that these processes occurred within a “Vulgar” PIE dia-
lect continuum, comparable with that of Vulgar Latin or so.

NEOACTIVE RESIDUAL MIDDLE aspectual value

A *gwhén-ti *dhugh-ói → *dhugh-óitoi PRS.IPFV
*dhéh1-ti *dhh1-ói → *dhh1-óitoi
*stḗu-ti *stéu̯ -oi
*u̯ óid-ei → *u̯ (ó)id-eiti
*su̯ op-éi → *su̯ op-éiti
*su̯ ṓp-i → *su̯ ṓp-i̯ ei ~ *su̯ ṓp-i̯ eiti *su̯ ṓp-i → *su̯ ṓp-itoi
*g̑ ónh1-i → *g̑ (ó)n-itoi
B *stḗu-t *stéu̯ -o → *stéu̯ -oto IPFV
*stḗu-s → *stḗu-s(t) (the new 3sg suffix
*u̯ óid-e → *u̯ óid-et -r was perhaps
initially introduced
*su̯ ṓp-i → *su̯ ṓp-i̯ et ~ *su̯ ṓp-is(t) *su̯ ṓp-i → *su̯ ṓp-i̯ ot(o) as a mediopassive
*g̑ ónh1-i → *g̑ (ó)nh1-i̯ ot(o) marker here,
*su̯ ṓp-i → *su̯ ṓp-ito see section 4.5)
*g̑ ónh1-i → *g̑ (ó)nh1-ito
C *gwhén-t *gwhn̥ -tó NON-IPFV
*dhéh1-t (*dhh1-´ → *dhéh1-) *dhh1-tó
*g̑ ónh1 → *g̑ ónh1-to
*u̯ idó → *u̯ id-ót(o)

Table 3. Leveling and blending processes (NB that these forms are meant to be types)

4.2 Backformations
As a natural consequence, new (neo-) active forms were increasingly back-
formed from the residual middle forms, as illustrated by Table 4, where emphasis
in bold type marks the neoactive backformations. This development can be con-
sidered a straightforward consequence of the blending processes and another con-
sequence of the Great Voice Shift. The active voice category now lost the original
opponent agentive meaning that most of the protoactive forms once had (excep-
tions were protoactive forms of the */bhéuh2t/ class). Whether the given partici-
 From Proto-Indo-European to Indo-European: After the Great Voice Shift and the Morphotactic Fusion 7

pant involved in the event was agentive or not, was not indicated by the active
versus middle voice marking any longer, and the middle was increasingly nar-
rowed to residual “mediopassive” reflexive, benefactive, O-possessive, and pas-
sive function.

NEOACTIVE NEOACTIVE BACKFORMATIONS

A *gwhén-ti *dhugh-óitoi → *dhugh-óiti
*dhéh1-ti *dhh1-óitoi → *dhh1-óiti
*stḗu-ti
*u̯ óid-ei → *u̯ (ó)id-eiti
*su̯ op-éi → *su̯ op-éiti
*su̯ ṓp-i → *su̯ ṓp-i̯ ei ~ *su̯ ṓp-i̯ eiti *su̯ ṓp-i̯ otoi → *su̯ ṓp-i̯ oti//i̯ eti
*g̑ ónh1-i̯ otoi→ *g̑ (ó)nh1-i̯ oti//i̯ eti
*su̯ ṓp-itoi → *su̯ ṓp-iti
*g̑ ónh1-itoi → *g̑ (ó)nh1-iti
B *stḗu-t *stéu̯ -o → *stéu̯ -ot/et
*stḗu-s → *stḗu-st
*u̯ óid-e → *u̯ óid-et
*su̯ ṓp-i → *su̯ ṓp-i̯ et ~ *su̯ ṓp-is(t) *su̯ ṓp-i̯ oto → *su̯ ṓp-i̯ ot/i̯ et
*g̑ ónh1-i̯ oto→ *g̑ (ó)nh1-i̯ ot/i̯ et
*su̯ ṓp-ito → *su̯ ṓp-it
*g̑ ónh1-ito → *g̑ (ó)nh1-it
C *gwhén-t *gwhn̥ -tó
*dhéh1-t (*dhh1-´ → *dhéh1-) *dhh1-tó
*g̑ ónh1 → *g̑ ónh1-t
*u̯ idó → *u̯ id-ó/ét

Table 4. Backfomations of neactive forms (NB that these forms are meant to be types)

As a second consequence, new active forms of category A were backformed

from forms of category B or C. This is illustrated by Table 5, where bold type
indicates the new backformed forms.
The result of the paradigmatic leveling, blending, and backformation processes
was an increased length of the innovative word forms (cf.Table 5). The parts of
the new endings closer to the root were thus prone to be identified as new stem
suffixes on their own, since the final parts of the new long endings were still
standing side by side with short endings (-ti, -toi, -ei) that were identical to them
and inherited from the older system.
 Roland A. Pooth 8

NEOACTIVE NEOACTIVE BACKFORMATIONS

A *gwhén-ti *dhugh-óitoi → *dhugh-óiti
*dhéh1-ti *dhh1-óitoi → *dhh1-óiti
*stḗu-ti
*u̯ óid-ei → *u̯ (ó)id-eiti
*su̯ op-éi → *su̯ op-éiti → B *su̯ op-éi̯ et
*su̯ ṓp-i → *su̯ ṓp-i̯ ei ~ *su̯ ṓp-i̯ eiti *su̯ ṓp-i̯ otoi → *su̯ ṓp-i̯ oti//i̯ eti
→ B *su̯ ṓp-i̯ et *g̑ ónh1-i̯ otoi→ *g̑ (ó)nh1-i̯ oti//i̯ eti
*su̯ ṓp-itoi → *su̯ ṓp-iti
*g̑ ónh1-itoi → *g̑ (ó)nh1-iti
B *stḗu-t *stéu̯ -o → *stéu̯ -ot/et → A *stéu̯ -eti
*stḗu-s → *stḗu-st
*u̯ óid-e → *u̯ óid-et → A *u̯ óid-eti
*su̯ ṓp-i → *su̯ ṓp-i̯ et ~ *su̯ ṓp-is(t) *su̯ ṓp-i̯ oto → *su̯ ṓp-i̯ ot/i̯ et → A *su̯ ṓp-i̯ eti
→ A *su̯ ṓp-i̯ eti *g̑ ónh1-i̯ oto→ *g̑ (ó)nh1-i̯ ot/i̯ et → A *g̑ n̥ h1-i̯ eti

*su̯ ṓp-ito → *su̯ ṓp-it

*g̑ ónh1-ito → *g̑ (ó)nh1-it
C *gwhén-t *gwhn̥ -tó
*dhéh1-t (*dhh1-´ → *dhéh1-) *dhh1-tó
*g̑ ónh1 → *g̑ ónh1-t
*u̯ idó → *u̯ id-ó/ét → A *u̯ id-éti

Table 5. Additional backformations (NB that these forms are meant to be types)

4.3 The emergence of new stem suffixes I

The emergence of new stem suffixes out of the part of the innovative long end-
ings closer to the root (cf. Table 6) is dated here to a post-PIE period with increas-
ing divergence. The reanalysis must have happened when IE languages had al-
ready split up into more separated dialect clusters or branches, which, however,
presumably were still in close areal contact.
Some of the stem suffixes, e.g. the sigmatic stem suffix -s-, cannot be recon-
structed for Proto-Anatolian (only the ending -s(t)), but most of them are contin-
ued in this archaic branch. Therefore, this third development happened in a period
with much more dialectal variation, but it must still have happened closely before
or within the period when the dialectal group developing to Proto-Anatolian split
up from the residual “Vulgar” PIE continuum. Notice that we are not followers of
the Indo-Hittite hypothesis. Regarding the comparative data, we find that it is
much more plausible infering that speakers of (pre-) Proto-Anatolian split from an
already strongly divergent IE dialect continuum rather than taking the rest of the
IE languages as sprung from a common “Proto-Indo-Greek” proto-language. All
 From Proto-Indo-European to Indo-European: After the Great Voice Shift and the Morphotactic Fusion 9

archaic features of Anatolian languages can be explained as residual archaic fea-

tures inherited from PIE, whereas there are no significant shared innovations of
the rest of the IE family.

NEOACTIVE NEOACTIVE
wh ti
A *g én- *dhugh-óitoi → *dhugh-ói-ti
*dhéh1-ti *dhh1-óitoi → *dhh1-ói-ti
*stḗu-ti
*u̯ óid-ei → *u̯ (ó)id-ei-ti
*su̯ op-éi → *su̯ op-éiti → B *su̯ op-éi̯ e-t
*su̯ ṓp-i → *su̯ ṓp-i̯-ei ~ *su̯ ṓp-iei-ti *su̯ ṓp-i̯ otoi → *su̯ ṓp-i̯ o/e-ti
→ B *su̯ ṓp-i̯ e-t *g̑ ónh1-i̯ otoi→ *g̑ (ó)nh1-i̯ o/e-ti
*su̯ ṓp-itoi → *su̯ ṓp-i-ti
*g̑ ónh1-itoi → *g̑ (ó)nh1-i-ti
B *stḗu-t *stéu̯ -o → *stéu̯ -ot/et → A *stéu̯ -o/e-ti
*stḗu-s(t) → *stḗu-s-t
*u̯ óid-e → *u̯ óid-e-t → A *u̯ óid-e-ti
*su̯ ṓp-i → *su̯ ṓp-i̯ e-t ~ *su̯ ṓp-is-t *su̯ ṓp-i̯ oto → *su̯ ṓp-i̯ ot/i̯ et → A *su̯ ṓp-i̯ e/o-ti
→ A *su̯ ṓp-i̯ e-ti *g̑ ónh1-i̯ oto→ *g̑ (ó)nh1-i̯ ot/i̯ et → A *g̑ n̥ h1-i̯ é/ó-ti
*su̯ ṓp-ito → *su̯ ṓp-i-t
*g̑ ónh1-ito → *g̑ (ó)nh1-i-t
C *gwhén-t *gwhn̥ -tó
*dhéh1-t (*dhh1-´ → *dhéh1-) *dhh1-tó
*g̑ ónh1 → *g̑ ónh1-t → *g̑ ón-H-t
*u̯ idó → *u̯ id-ó/é-t

Table 6. Reanalysis of endings (NB that these forms are meant to be types)

3PLURAL 2/3COLLECTIVE
INTRANSITIVE INTRANSTIVE
DETRANSITIVE DETRANSITIVE IE 3pl
wh wh
I *g nró *g náh2 *gwhnáh2ont(o)
II *stéuro *stéuah2 *stéu̯ ah2ont(o)
III *ueidór *uóidah2 *u̯ óidah2ont(o)
IV *uidór *uidáh2 *u̯ idáh2ont(o)
V *suópr *suópah2 *su̯ ópah2ont(o)
VI *suopér *suopáh2 *su̯ opáh2ont(o)

Table 7. From PIE collective-plural forms to new 3pl forms

 Roland A. Pooth 10

4.4 The emergence of new stem suffixes II

As an appendix to the preceding section 4.3, we suggest that PIE 2/3 collec-
tive-plural intransitive forms were extended either by the (proto-) middle and neo-
active 3pl ending *-ont or the newly created and highly productive middle ending
*-onto. In line with the aforementioned sections, these forms developed into IE
stems in *-ā- (see Table 7) (cf. Pooth 2016b).

4.5 The emergence of new mediopassive endings

For the emergence of the -r(i)- endings from prior 3pl endings cf. Pooth (2014:
177) with references. Past tense reference and durative-imperfective aspect pre-
sumably was encoded with middle inflection in the post-PIE period, thus func-
tionally continuing or copying the preceding PIE use of detransitive binyan II
forms in this function. The -r- endings maybe developed to specifically opponent
mediopassive (reflexive, benefactive, O-possessive, passive) endings in opposi-
tion to the old middle endings within category B (IMPERFECTIVE). But we will not
expand on this here.

5 Outcome
We finally arrive at a “Vulgar” PIE or post-unitary PIE model which is formal-
ly almost identical with the traditional “Graeco-Aryan” model, as found in all
present handbooks of Indo-European Linguistics. The novelty of our model is on
the functional and analytical level. We can even provide an internal chronology of
the morphosyntactic changes that must have taken place. We do not want to be
victims of the traditional, but misleading and anachronistic preconceptions, which
are simplistically re-projecting IE morphotaxis, including ablaut grades, stem suf-
fixes, and endings to an artificial pseudo-proto-language that they call “PIE”. The
Templatic Model and its implications are promising and have great advantages.
They can make predictions resulting in a post-PIE model that is closer to the Pro-
to-Anatolian verb system, and they additionally provide a highly plausible sys-
tematic explanation for all the given IE vowel alternations (ablaut grades) without
any preconceptions and sound laws that are invented out of the blue.

References
Bybee, Joan L., Perkins, Revere and Pagliuca, William
1994 The Evolution of Grammar: Tense, aspect and modality in the languages of
the world. Chicago University Press.
 From Proto-Indo-European to Indo-European: After the Great Voice Shift and the Morphotactic Fusion 11

Jacques, Guillaume & Anton Antonov

2014 Direct/Inverse Systems. Language and Linguistics Compass 8, issue 7, 301-
318.
Jasanoff, Jay H.
2003 Hittite and the Indo-European Verb. Oxford University Press.
Kümmel, Martin J.
1998 Wurzelpräsens neben Wurzelaorist im Indogermanischen. Historische
Sprachforschungen 111, 191-208.
Pooth, Roland A.
2009 Der urindogermanische Progressiv. In: Protolanguage and Prehistory. Akten
der XII. Fachtagung der Indogermanischen Gesellschaft ... in Krakau. Ed.
by R. Lühr & S. Ziegler. Wiesbaden, 381-406.
2014b Die Diathesen Aktiv vs. Medium und die Verbsemantik im Vedischen der
R̥ gveda-Saṃhitā. Proefschrift. Defended 2014-10-23. Universiteit Leiden,
Leiden University Repository, https://openaccess.leidenuniv.nl
2015b Proto-Indo-European Nominal Morphology. Part 1. The Noun. Language
Arts 1, issue 2015-12-23, ms. 2015-12-23.
2015[6] Gr. ἔρι-, ved. 2r̥ und arí-, die uridg. Wurzel *h1er- ~ *h1ar- und uridg. *h1er-
~ *h1ar- ‚Widder, Schafsbock‘, Historische Sprachforschung 128, 95-122
2016b Proto-Indo-European Verb Morphology. Part 1. Inflection. Language Arts 2,
issue 2016 AUG 4, ms. 2016 AUG 4.

Publications
Cf. https://leidenuniv.academia.edu/RolandPooth
Pooth, R.A. 2000: Stativ vs. Medium im Vedischen und Avestischen. Historische Sprachforschung 113, 88-
116.
Pooth, R.A. 2001: Studien zur frühurindogermanischen Morphologie I. ‘Stativ’, ‘Medium’ und ‘Perfekt’.
Historische Sprachforschung 114, 220-258.
Pooth, R.A. 2004a: Ablaut und autosegmentale Morphologie: Theorie der urindogermanischen Wurzelflexi-
on. In: Indogermanistik - Germanistik - Linguistik. Ed by. M. Kozianka, R. Lühr & S. Zeilfelder. Hamburg,
401-471.
Pooth, R.A. 2004b: Zur Genese der späturidg. thematischen Konjugation aus frühuridg. Medialformen. Indo-
germanische Forschungen 109, 31-60.
Pooth, R.A. 2009b: Proto-Indo-European Ablaut and Root Inflection. In: Internal Reconstruction in Indo-
European. Ed. by J. E. Rasmussen & T. Olander. Copenhagen, 229-254.
Pooth, R.A. 2011: Die 2. und 3. Person Dual und das Medium. In: Indogermanistik und Linguistik im Dialog.
Akten der XIII. Fachtagung der Indogermanischen Gesellschaft ... in Salzburg. Ed. by T. Krisch & T. Lind-
ner. Wiesbaden, 473-83.
Pooth, R.A. 2012: Zum Aufkommen transitiver Verben in frühen Vedischen am Beispiel 1r̥ . In: The Indo-
European Verb. Proceedings of the Conference of the Society for Indo-European Studies, Los Angeles 13-
16 September 2010. Ed. by E. Craig Melchert. Wiesbaden, 267-84.
Pooth, R.A. 2014a (ms.): Ein Problem der Methode der komparativen Rekonstruktion von Morphemen, Mor-
phemgrenzen und morphosyntaktischen Kategorien, ms. version 2014-08-07.
 Roland A. Pooth 12

Pooth, R.A. 2014c (ms.): Voice, Transitivity Direction, Case, and Alignment in Proto-Indo-European. IV.
The Proto-Indo-European Case System, version 2014-11-17.
Pooth, R.A. 2014d (ms.): More evidence for Proto-Indo-European transfixes: Two types of ‘lengthened
grades’, version 2014-12-05.
Pooth, R.A. 2015a (ms.): A typological overview of Proto-Indo-European, version 2015-07-23.
Pooth, R.A. 2016+ (ms.): Is the “tēzzi principle” a plausible inference. Paper given at: The precursors of
Proto-Indo-European: the Indo-Hittite and Indo-Uralic hypotheses. Workshop at the Leiden University
Centre for Linguistics, 9-11 July 2015.
Cf. Language Arts 3, 4, and 5, which are also available there.

Supplement to Language Arts 2 | cf. Pooth 2016b