Φωνήεντα Chapter 4

Root-and-pattern PIE?

by C Ryan Moniz

original research· spring 2010 - harvest 2016
updated & published· spring 2022

philology

Φωνήεντα·

1 · 2 · 3 « 4 » 5

The motivation for Indo-Europeanists to explain PIE nominal accent-ablaut using a compositional approach comes in large part from the assumption that paradigms cannot be taken as bearing inherent meaning in Indo-European, but are rather the result of some other process. Indo-Europeanists such as Keydana (2012) find it “unconvincing to derive the reconstructed ablaut patterns from a form of templatic morphology since there is no evidence for this type of morphology in Indo-European.”❦1 Keydana 2012 p 113 Moreover, Sandell (2014) has argued that “templatic morphology is probably too powerful a mechanism to employ to capture the facts about PIE.”❦2 Sandell 2014 p 2 However, there may be reason to believe that PIE was very typologically different from its daughter languages: none of its daughter branches retain the original nominal accent-ablaut in its full form, only one branch retains any of the three laryngeal sounds, and only the oldest attested branch, Anatolian, exhibits a clear animate-inanimate distinction in nouns (whereas the other branches generalized a three-gender system of masculine, feminine, and neuter).

Pooth (2015a, 2015b) boldly proposes a semantically productive templatic system throughout the morphology of Proto-Indo-European. In Pooth’s theory, nouns are not classed into paradigms, but rather the templates act as morphemes unto themselves which convey meaning unto the noun, i.e. they are derivational, not inflectional. The differences between the nouns fitting into the paradigmatic models of Erlangen and Leiden can be explained by positing that only certain forms within the template remained in use after the template lost its productivity, while other forms fell out of use, or survived as related words outside of the noun paradigms proper.❦3 Pooth 2015b p 10, 20 ·❦4 cf. Schindler 1975b Moreover, Pooth believes that the morphological distinctions between nouns mirrored those found in verbs, and a fundamental distinction between agentive words and detransitive words was maintained in both nouns and verbs; detransitivity in nouns was realized either by means of inanimacy, or as a resultative form.❦5 cf. Schinlder 1972 Crucial to Pooth’s reconstruction is the notion that *o is a suprasegmental detransitive morpheme which mapped the feature of [+round] (and perhaps [+back]) onto the vowels in the template in both verbs and nouns. All nouns containing *o are therefore reconstructed to have originally been detransitive.❦6 Pooth 2015b p 2

In addition, nouns can be divided into basic nouns with a “core” template (analogous to strong cases) and an “oblique” template (analogous to weak cases), and relational nouns — i.e. possessed or attributive nouns❦7 Compare the status constructus of Semitic nouns — with a core template based on the oblique (or “broken locative,” see chapter 2) template of the base nouns, and a secondary oblique template.❦ 8 Pooth 2015a p 43 ·❦9 Pooth 2015b p 9 Table 19 below summarizes the derivational templates for PIE nouns:

Table 19 — Nominal template morphology, Pooth (2015a, 2015b)
Agentive Detransitive
R ·S -E R ·S -E
base noun base core é ∅ ∅ ó ∅ ∅
primary oblique ∅ é ∅ é
o
∅		 o
é
ó		 ∅
∅
∅
relational noun relational core
secondary oblique ∅
∅		 ∅
∅ ́		 é
∅		 é
o
∅		 ∅
∅
∅		 o
é
ó

The basic core corresponds to the strong stem of proterodynamic and hysterodynamic paradigms. The relational core stem (e.g. “strong” *ud·én ‘water’), formally identical to the primary oblique or “broken locative” stem (e.g. “weak” endingless LOC *ud·én), corresponds to the weak stem of the proterodynamic paradigm and the strong stem of the hysterodynamic paradigm. The secondary oblique stem corresponds to the hysterodynamic weak stem. The derived stems of the detransitive exhibit three patterns, differentiated by accent placement and by where suprasegmental detransitive marker *o is mapped.❦10 Pooth 2015b p 20 The first two patterns exhibit two surface vowels in the detransitive, wile the third only has one.

Table 20 — Static root accent, detransitive marker is mobile
Agentive Detransitive
base base core *gén·u-❦11 Greek γένυ-, Welsh gen, Gothic kinnu- ‘jaw’ *gón·u-❦14 Greek γόνυ, Sanskrit jā́nu ‘knee
prim. OBL *gn·éu-❦12 Old English cnéow, Old High German kniu ‘knee’ *gén·ou-
relational rel. core
second. OBL *gn·u-é / gn·ú-❦13 Avestan žnu- ‘knee’ *gén·u-o❦15 Hittite gēnuwa- ‘knee’

With these stems, the accent of the detransitive noun is fixed on the root, but the mapping of the detransitive marker progresses with each derived stem. This particular noun is typically reconstructed as either a static root with vowel quality variation — STRONG *gón·u ~ WEAK *gén·u- ‘knee,’ although the reflexes of *gén·u- tend to mean ‘jaw’ rather than ‘knee’ — or as a proterodynamic noun STRONG *gón·u ~ WEAK *gn·éu-, avoiding the reflexes meaning ‘jaw.’ Pooth connects all of these forms under the gloss ‘joint, angle,’ with *gén·u and its derivatives signifying an agentive/animate human joint (‘jaw’ or ‘chin’), and with *gón·u as its detransitive/inanimate counterpart.❦16 Pooth 2015b p 6

Table 21 — Accent is mobile, detransitive marker static on the root
Agentive Detransitive
base base core *dér·u-❦17 Welsh derw ‘oaks’ *dór·u-❦20 Greek δόρυ, Hittite tāru, Sanskrit dā́ru ‘wood
prim. OBL *dr·éu-❦18 Gothic triu, Old English tréow ‘tree’ *dor·éu-
relational rel. core
second. OBL *dr·u-é / dr·ú-❦19 Greek δρυ-, Lithuanian drū́ta-, Old English trum ‘firm *dor·u-é❦21 Greek δουρός ‘wood (Epic GEN),’ Latin dūr- ‘hard’

With these nouns, the accent is mobile, whereas the detransitive marker remains mapped onto the stem vowel (though Greek δροόν might provide evidence of another detransitive paradigm exhibiting mobile accent and mobile detransitive marking, i.e. prim. OBL / rel. core *dr·óu-, following the paradigm seen below in table 22).

Table 22 — Accent and detransitive marker are mobile
Agentive Detransitive
base base core *uéd·r- *uód·r-❦24 Hittite wātar ‘water’
prim. OBL *ud·én-❦22 Hittite uiten-, Sanskrit udán ‘water’ *ud·ón-
relational rel. core
second. OBL *ud·n-é❦23 Sanskrit udná- ‘water’ *ud·n-ó

Both the accent and the detransitive marker mapping are completely mobile in these nouns.

Pooth also provides an explanation for the vowel quality allomorphy resulting from the change from neuter singular to collective (discussed in chapter 2) by means of a plurative (collective/abstractive/generic/augmentative) marker in the form of an additional *e added to the noun. In the case of nouns like *pék·u ‘domestic animal,’ this simply means that the root vowel becomes long *ē (*pḗk·u, underlyingly */péeku/) in the basic core. However, in situations with mobile accent, the plurative marker *e follows the accent, and when the mobile-accent noun is detransitive, the mapping of *e displaces the mapping of the detransitive marker, e.g. *uód·r ‘water’ → **uóed·r → *uéd·or.❦25 Pooth 2015b p 13

Table 23 — Plurative mapping
‘livestock’
Agentive		 ‘water’
Detransitive
base base core *pḗk·u- *uéd·or-❦28 Gothic wato-, Old English wæter
prim. OBL *pek·éu-❦26 Gothic *faihau-, Old Norse fjá- ‘livestock’ *uod·én-❦29 Gothic watins, Old Norse vatn
relational rel. core
second. OBL *pek·u-é❦27 Sanskrit páśva- ‘livestock’ *uod·n-é

The etymological relationship between the (“hysterokinetic”) relational noun *ph₂·tér ~ ph₂·tr-é ‘[X]’s father’ and the adjectival Greek -πάτωρ ‘having an [X] father’ can be explained by means of an abstractive use of the plurative: *ph₂·tér → **póeh₂·tr- → *péh₂·tor-.❦30 Pooth 2015b p 21

Pooth’s templatic morphology offers a unique picture of PIE as a completely root-and-pattern language, with the paradigms behaving as derivational transfixes upon nouns. These transfixes explain the variation observed in PIE accent-ablaut, and take into account the suspicious formal and semantic relationship between the endingless locative and core relational (“hysterokinetic”) stems. Instead of attempting to explain the differences in vowel quality with a unifying theory of accent and ablaut (where accent predicts ablaut), Pooth provides a semantic explanation for *o-grade, reconstructing a fundamental agentive-detransitive voice opposition (as part of Pooth’s reconstruction of PIE as a language with fluid-S active-stative alignment❦31 Pooth 2015a p 49 ). However, semantic explanations and accentual (phonological) explanations for *e~o allomorphy may not be entirely mutually exclusive.

Φωνήεντα·

1 · 2 · 3 « 4 » 5

references

❦1· Keydana 2012 p 113
❦2· Sandell 2014 p 2
❦3· Pooth 2015b p 10, 20
❦4· cf. Schindler 1975b
❦5· cf. Schinlder 1972
❦6· Pooth 2015b p 2
❦7· Compare the status constructus of Semitic nouns
❦8· Pooth 2015a p 43
❦9· " 2015b p 9
❦10· " p 20
❦11· Greek γένυ-, Welsh gen, Gothic kinnu- ‘jaw’
❦12· Old English cnéow, Old High German kniu ‘knee’
❦13· Avestan žnu- ‘knee’
❦14· Greek γόνυ, Sanskrit jā́nu ‘knee’
❦15· Hittite gēnuwa- ‘knee’
❦16· Pooth 2015b p 6
❦17· Welsh derw ‘oaks’
❦18· Gothic triu, Old English tréow ‘tree’
❦19· Greek δρυ-, Lithuanian drū́ta-, Old English trum ‘firm’
❦20· Greek δόρυ, Hittite tāru, Sanskrit dā́ru ‘wood’
❦21· Greek δουρός ‘wood (Epic GEN),’ Latin dūr- ‘hard’
❦22· Hittite uiten-, Sanskrit udán ‘water’
❦23· Sanskrit udná- ‘water’
❦24· Hittite wātar ‘water’
❦25· Pooth 2015b p 13
❦26· Gothic *faihau-, Old Norse fjá- ‘livestock’
❦27· Sanskrit páśva- ‘livestock’
❦28· Gothic wato-, Old English wæter
❦29· Gothic watins, Old Norse vatn
❦30· Pooth 2015b p 21
❦31· " 2015a p 49

bibliography

philology