You are on page 1of 12

The Caprice of O....

On a Proto-Indo-European substantivization type and its excesses in Ancient Greek

ECIEC 37
June 15 — 17, 2018, University of Michigan, Ann Arbor
Stefan Höfler, Harvard University
hoefler@fas.harvard.edu

1 Starting Point
• Schindler 1980:390 lists examples of thematic adjectives and corresponding (i.e. derived) i-stem
abstracts:
o *dhegu̯ h-ró- ‘burning’ (subst. Gk. τέφρη f. ‘ashes’ Il.+)
→ *dhegu̯ hri- f. ‘burning heat’ (Lat. febris, -is f. ‘fever’)
o *n̥ bh-ró- ‘wet’ (subst. Ved. abhrá- n. ‘cloud’ RV+, Gk. ἀφρός m. ‘foam, froth’ Il.+)
→ *n̥ bhri- m. ‘wetness, wet thing’ (Lat. imber, -bris m. ‘rain, storm’ Plaut.+)

• In note 17, Schindler adds:


o “In gr. ὄµβρος ist die Substantivierung durch o-stufige Vṛddhi gekennzeichnet.”
§ Here “Vṛddhi” does not refer to the morphological process of “Vṛddhi derivation”
(*di̯ éu̯ -, *diu̯ - ‘heaven, sky’ → *dei̯ u̯ -ó- ‘belonging to the heaven, sky’ > ‘a god’), but
rather to the descriptive phenomenon of adding a new vowel “vorne im Wort”.

• The process of substantivization could be set up as:


o *n̥ bh-ró- ‘wet’ (Ved. abhrá-, Gk. ἀφρός)
→ *ónbhro- > *ómbhro- > Gk. ὄµβρος m. ‘rain, storm’ (Il.+)
§ The development *-mbhr- > Gk. -µβρ- is not contradicted by anything substantial.1
• Note the Schwebeablaut-like position of the “substantivizing O” (not †nóbhro-!).
• Since Schindler 1980, not very much ink has been spilled over this substantivization process.

• Aim of this talk:


o Assess the status of this word-formation process in PIE, especially in relation to similar
processes (substantivization through accent retraction, e-Vṛddhi)
o Provide examples from different languages / language branches.
o Marvel at some Greek “capricious O’s”.
o Point at some open questions.

1
For a general development of *-mbh- > Gk. -µβ- see Schwyzer 1959³:333; for *-mbh- > Gk. -µβ- “hinter betontem Vokal” see Neri
2016:14. A sequence -µφρ- is only found at a morpheme boundary (e.g. ἔµ-φρων ‘sensible’, etc.), and in the obscure πεµφρηδών
f. ‘a wasp’ (Nic.).
1
2 Substantivization
• For PIE & IE languages, one has to distinguish two basic processes of substantivization of adjectives:
(1) the substantivization via ellipsis, i.e. the omission of a substantive that an adjective was originally
paired with, so that only the adjective remains in a substantivized meaning (cf. for example
Wackernagel 1926:53; Panagl 1989:188f.).
A bottle of red, a bottle of white / It all depends upon your appetite. [= a bottle of white (sc. wine).]
(Billy Joel – Scenes from an Italian Restaurant).

(2) a process that, for the time being, will be named “direct substantivization”, i.e. the adjective is
substantivized without the existence of a prior syntagma of SUBST. + ADJ.; the result can be (a)
2
abstracts or (b) concrete individualizations (cf. Rau 2008:166f. ; Nussbaum 2014:304ff.).
The conjunctiva is the membrane that lines the eyelid and covers the white of the eye.
• In the ellipsis case (cf. also Watkins’s [1995:156] use of the term “transferred epithets”), it is
conceivable that the substantivization process did not have any “morphological effect” (i.e. no
suffixation, no accent shift, no change in ablaut). The gender of the substantivization depends on the
gender of the ellipted word.
o Ved. pr̥ thivī́-, pr̥ thvī́- f. ‘(Mother) E/earth’: fem. adj. *pl̥ th2-u̯ -ih2 ‘broad’ via ellipsis ‘the broad
earth’, in Ved. kṣā́ m … pr̥ thvī́m (RV 10.31.9); YAv. ząm pərəϑβīm (Y.10.4; Yt.13.9).
o Ved. mahiṣá- m. ‘buffalo’ from the masc. adj. mahiṣá- ‘tremendous’ (RV 10.66.10 mahiṣásya
tanyatós ‘of the tremendous thunder’) used in combination with mr̥ gá- m. ‘animal’ (cf.
mahiṣáṃ mr̥ gám ‘the tremendous animal’ RV 8.69.15, etc. for ‘buffalo’).
• NB: a PIE ellipsis is, of course, difficult to demonstrate, other than indirectly through the lack of any
overt substantivization marker.
• In the case of “direct substantivization” (above (2); note that “direct” refers to the fact that there is no
ellipsis involved), we find a whole array of morphological means to mark the substantivization:
o Suffixation: *n̥ bhró- ‘wet’ → *n̥ bhr-i- m. (Lat. imber m. ‘rain, storm’)
o Accent shift: Ved. kr̥ ṣṇá- ‘black’ → kr̥ ́ ṣṇa- m. ‘(black) antilope’
o Change in ablaut: *pum-és- ‘having a beard’ (>> Lat. pūbēs, -eris ‘pubescent;
downy (of plants)’) → *p(e)um-ōs m. (Ved. púmāṁs- m. ‘man’)3
o “Zero” (if not ellipsis): Ved. tamasá- ‘dark’ → tamasá- n. ‘darkness’
• Note that suffixation is an external mechanism, while accent shift and change in ablaut (and
arguably “zero”) are internal mechanisms. We will return to this distiction below.
• Let us have a look at internal substantivization mechanisms in thematic adjectives (accent shift,
e-Vṛddhi, o-Vṛddhi)

2
When speaking of substantivization marked through accent retraction and concomitant insertion of a new full grade (see below)
“It shows up both in cases where the substantivization functions as an adjective abstract ... and further where the substantivization
functions as an “individualization,” in effect referring to someone or something that has whatever quality the adjective denotes ...”
3
Cf. Adams 1985; Höfler 2015:178f. For the semantics cf. Lat. barbātus ‘bearded’ and Romanian bărbat ‘man’.
2
3 Substantivization with accent retraction
Cf. AiGr III:88; Nussbaum 1986:157; Schaffner 2001:328ff.; Rau 2008:165; Steer 2014:397; Höfler 2017.
• Examples:
o Vedic: kr̥ ṣṇá- ‘black’ (RV+) → kr̥ ́ṣṇa- m. ‘(black) antilope’ (RV+), etc.
o Greek: κνηκός ‘pale yellow, tawny’ (Thespis, S., Theocr.,) → κνῆκος f. ‘safflower’ (Myc. ka-
na-ko; Hp., Arist.), etc.
o Germanic: *barza- ‘erect, pointy’ (OHG parremo ‘erectus’; ON barr ‘ardent’) → *barsa- m.
‘bass [the fish]’ (OHG bars, OE bears, Gm. Barsch), etc.
• Accent retraction also appears with *-h2-:
o *pélh2-(o)s n. ‘cover(ing)’ (Gk. °πελας n. ‘skin’; den. verb in Umbr. pelsa- ‘bury’4)
→ *pl̥ h2-s-ó- ‘covering’
́ 2seh2 ‘the covering (one)’ (Hitt. palaḫša- /palḫša-/ c. ‘a garment for
→ *pl̥ h
women; a cloth covering’; PItal. *palasā > Lat. palla f. ‘women’s garment; a
curtain’ Plaut.+).
o *(s)pérH-os n. ‘feather, wing’ (Slov. pero, -esa n. ‘feather’)
→ *(s)pr̥ H-s-ó- ‘having feathers’
→ *(s)pr̥ ́Hseh2 ‘the feathered one’ (Lat. parra, Umbr. parfa- ‘name of a
bird’); cf. Höfler 2017.

4 Substantivization with new accented e-grade (“Substantivierungsvṛddhi”, e-Vṛddhi)


Cf. Schindler apud Mayrhofer, EWAia II:269f.; Eichner 1973:93 note 46; Schaffner 1997:108 note 53 with refs.; Schaffner
2001:235f. note 335, 334f.; Rau 2008:166f.; Steer 2014:397; Steer 2015:33f., 49, 76f., 165; Lipp 2009 II: 416 note 138; Nussbaum
2018:257.
• Examples:
o *bhr̥ hxg̑ -ó- ‘white’ (subst. as Ved. bhūrjá- m. ‘birch’, alb. bredh ‘white fir’5)
→ *bhérhxg̑ -o- ‘the white one’ (Lith. béržas m.; cf. also OHG birka, Gm. Birke, Engl.
birch); cf. Schindler apud Mayrhofer, EWAia II:269f.
o *g̑ n̥ h1-tó- ‘born’ (Lat. (g)nātus, i m. ‘son’, Ved. jātá- m. ‘born; son’, ON kundr ‘son’)
→ *g̑ énh1-to- n. ‘born one’ (OHG kind n., Gm. Kind ‘child’); cf. NIL:152 n. 101.
o *k̑ rn̥ to- ‘horned’ (subst. as MLG runt, Dutch rund ‘bovine’, OE n. pl. hrȳðer ‘cattle’)
→ *k̑ rénto- ‘the horned one’ (OHG (h)rind n., Gmn. Rind, OS hrīth ‘cow’); cf.
Nussbaum 1986:7, 258; Schaffner 2001:203-213.
o PIIr. *Hr̥ -tá- ‘true’ (Ved. r̥ tá- ‘proper, right, true’, subst. as n. ‘truth, order, rule’)
→ *Hár-ta- n. ‘true thing’ (OAv., YAv. aṣ̌a- n. ‘truth, order’); cf. Schaffner 1997:108
note 53. (could also be o-Vṛddhi)

4
Cf. Weiss 2010:305ff.
5
Cf. Schumacher/Matzinger 2013:230.
3
• An explanation for the “new” e-grade that suggests itself is that it is a result / side effect of accent
retraction (cf. Schaffner 1997:108 note 53: “älteste Schicht der Substantivierungen mittels
Kontrastakzent”).6
o It is strange, however, that we find doublets in closely related languages/language branches
(ON kundr ‘son’ vs. OHG kind; MLG runt vs. OHG (h)rind; Ved. r̥ tá- vs. Av. aṣ̌a- n.)

• In any case: e-Vṛddhi also appears in *-h2-stems:


o *tm̥ H-s-ró- ‘dark’ (Lith. tim̃ sras, West-Aukšt. tìmsras ‘dark red’)
→ *témHsreh2 ‘darkness’ (Ved. támisrās f. pl. in RV 2.27.14, Lat. tenebrae f. pl.).
o *h2us(-s)-r-ó- ‘early, matutinal’ (Ved. usrá-; subst. as usrā́ - f. ‘daylight, dawn’)
→ *h2éu̯ sreh2 ‘morning thing’ (Lith. aušrà f. ‘dawn, daybreak’, Gk. αὔρα f. ‘morning
breeze’); cf. Peters 1980:34; Nussbaum 1986:244.

5 “Substantivierung durch o-stufige Vṛddhi” (short: o-Vṛddhi)


Cf. Schindler 1980:39017; Hamp 1982/83:171-177; Nussbaum 1986:8; Rasmussen 1989:178ff.; Neri/Ziegler 2012:27; Neri
2013:197f; Harðarson 2014:47³; Steer 2015:100; Neri 2016:24, 30, 32, 34, 40.
• The substantivization of a thematic adjective is marked by a new accented o-grade at the beginning of
the word.7
o The most straightforward and most easily recognizable examples are derived from adjectives in *-Có- (*-ró-,
*-tó-, *-mó-, *-nó-, *-s-ó-, etc.).
• Some examples (from Rasmussen 1989:186; slightly modified).
o *str̥ h3-nó- ‘spread, strewn, scattered’ (Ved. stīrṇá- ‘id.’) → *stor(h3)neh2 (OCS strana f. ‘side,
land’; Gk. στόρνη ‘belt’ Call. ?).
o *témhx-os ‘darkness’ (ved. támas-) → *tm̥ hx-s-ó- ‘dark’ (Latv. tùmšs, tìmšs) or *temhx-s-ó-
‘id.’ (OS thimm) → *tom(hx)seh2 ‘darkness’ (Lith. tamsà f.).

• Further examples of the latter type (substantivizations of possessive adjectives derived from neuter s-
stems) from PIE are:
o *ku̯ él(h1)-os n. ‘turning, going (around)’ (Gk. τέλος n. *‘turning point’; cf. OCS kolo, gen. sg.
kolese n. ‘wheel’)
→ *ku̯ el(h1)-s-ó- (MW pell ‘far, distant’8; subst. as Gk. τέλσον n. ‘headland, i.e. land
where the plough turned’ Il.)
→ *ku̯ ól(h1)so- ‘the turning one’ (Lat. collus m. ‘neck’, PGmc. *halsa- m.
‘id.’); cf. Neri 2013:198.

6
Comparable new e-grades in the wrong place can also be found in secondary s-stems (derived from thematic adjectives); cf.
Nussbaum 1998:526, also Widmer 2004:108f.: *pl̥ h1-nó- ‘full’ (Ved. pūrṇá- etc.) → IIr. *parHnas- (Ved. párīṇas- n. ‘fullness,
abundance’, Av. parǝnah-uuant-); root *√pleh1 (cf. s-aor. Ved. aprās, Gk. ἔπλησα, Arm. elic‛); *n̥ bh-ó- (?) ‘wet’ (Arm. amp/amb
‘cloud’ ?) → *émbh-os n. (Ved. ámbhas- n. ‘water, flood’); root *√nebh (cf. *nébh-os > Ved. nábhas-, Gk. νέφος, etc.). Likewise
in amphikinetic stems (arguably internally derived): *h2éu̯ sōs f. ‘dawn’ (Gk. Hom. ἠώς, Aeol. αὔως; cf. Lat. aurōra); root *√h2u̯ es
(cf. aor. Ved. ví ... āvas ‘became light’; Ved. vástu- ‘dawning’, MW gwawr ‘dawn’, etc.; cf. Widmer 2004:111ff.; Ozoliņš 2015:40-
65.). And in various hypostases of en-locatives (neo-full grade from endingless locative?); cf. Nussbaum 1986:185ff. with addenda.
7
I owe many an insight regarding this substantivization process to Sergio Neri and an illuminating discussion in Jena in Sept. 2014.
8
For the semantics cf. Greek τῆλε ‘at, to a distance’ etc. (Nussbaum 2018:248).
4
o *h2/3u̯ ébh-os n. ‘weave, weaving’ (Gk. ὕφος n. ‘web’ Pherecr.+)
→ *h2/3ubh-s-ó- ‘weaving (adj.)’
→ * h2/3u̯ obhseh2 f. ‘the weaver’ (Lat. uespa f., Lith. vapsà f., Russ. osá f.,
Mid. Pers. vaβz, etc. ‘wasp’); because of the textured nests.

o *h1ér-os n. ‘turning away, around, aside’ (unattested; for the root *√h1er ‘turn’ cf. Alb. për-jer,
príer ‘turn away, aside; slant’; perhaps Ved. r̥ té ‘without, except’ as the loc. sg. of a *h1r̥ -tó-
‘turned away, around, aside’, Ved. ārá- m. or n. ‘distance’ if *h1or-ó- ‘turned-away (side)’)
→ *h1(e)r-s-ó- ‘turned away, around, aside’ (subst. as Gk. *εἰρό- ‘back, rear,
posterior’ in εἰρεϑύρη· ὀρσοϑύρα. ὁ στροφεύς ‘back door’ Hsch. ?; MIr. err ā, f.
‘hinder-part, end, tail’)
→ *h1órso- m. ‘the back, the rear’ (Hitt. arra- c., Gk. Att. ὄρρος m. ‘end of
the os sacrum’, OHG ars m., OE ears, ‘arse’, Arm. oṙ, cf. Gk. ὀρσοθύρη f.
‘back door’ Od.).
→ *h1orséh2 f. ‘the back, the rear’ (Gk. Ion. οὐρή f. ‘tail; rearguard’).9

• Futher examples: the νόστος type (cf. García Ramón 2016):


o *√nes ‘to return home safely’: *n̥ s-tó- ‘returned safely’ (subst. as Ved. ásta-, Av. asta- n.
‘home’) → νόστος m. ‘return home’ (ep. poet. Il.+)
• ... also as “individualizations”:
o *mr̥ -tó- ‘mortal, dead’ (Ved. mr̥ tá-, Av. mǝrǝta-, Gk. βροτός)
→ *mórto- ‘mortal one’ (Gk. µόρτος· ἄνθρωπος Hsch. Latte; Ved. márta- m., GAv.
maṣ̌a- m. ‘mortal’; cf. EWAia II:327)
o For more examples see Appendix 10.

• Further examples: Baltic, Slavic, Germanic (where the o-Vṛddhi type seems to have had a certain
productivity):
o *gélhx-os n. ‘call(ing)’ (cf. *gl̥ hx(s)meh2 > MIr. glám ā, f. ‘satire; outcry, clamor’; cf. Zair
2012:79)
́ xso- m. ‘the caller’ > Lat. gallus m. ‘rooster’, MIr. gall
→ *gl̥ hx-s-ó- ‘calling’ (→ *gl̥ h
‘swan’; cf. Höfler 2017:20)
→ *gol(hx)so- (OCS glasъ m. ‘voice’, Lith. galsas m. ‘echo’, OIcel. kall n.
‘call’).
o For more examples see Appendix 11.

9
An alternative etymology (cf. Frisk II:427; Stüber 2002:31), viz. that the underlying form is *h3or-s-o- ‘having two humps’ (cf.
Gk. ὄρος n. ‘hill, mountain’), does presuppose an original meaning ‘ass’ for the word, which would be difficult to reconcile with
ὀρσοθύρη ‘back door’ on a semantic level, and with εἰρεϑύρη and MIr. err on the phonetic side. On the other hand, a semantic
development ‘turned away’ or ‘back (part)’ > ‘ass’ (qua euphemism) is much more common: cf. butt and buttock and bottom *‘low
part’, backside, derriere (French derrière ‘behind’ and ‘bottom’), rear end, rear, posterior, etc. Cf. also Germ. Hintern, Spanish
trasero ‘back’ and ‘butt’, also Lat. dorsum ‘the back’ (*dē+uorsum ‘turned away’).
5
6 Observations on the three types (accent retraction, e-Vṛddhi, o-Vṛddhi)
• All three types form substantivizations of thematic adjectives.
• All three types are (arguably; see below) barytone.
• All three types are also attested with *-h2- suffix:
o ́ 2seh2 ‘covering’ (Hitt. palaḫša- c., Lat. palla f. ‘women’s garment; a cloth covering’)
*pl̥ h
o *témHsreh2 ‘darkness’ (Ved. támisrās, Lat. tenebrae f. pl. ‘darkness’)
o *h2/3u̯ obhseh2 ‘the weaver’ (Lat. uespa f., Lith. vapsà f., Mid. Pers. vaβz, etc. ‘wasp’)

• The question is: how do we analyze these *-h2- formations on a derivational level?
o Were the corresponding adjectives first substantivized via accent retraction, e-Vṛddhi, o-
Vṛddhi, and subsequently suffixed with *-h2-?
o Or were the adjectives substantivized through suffixation of *-h2-, and accent retraction, e-
Vṛddhi, o-Vṛddhi are side effects of this *-h2- suffixation?

• To answer this it is worthwhile to look at some general tendencies: while an external mechanism (i.e.
suffixation) can sometimes entail accent shift or change in ablaut ...
o Gk. στραβός ‘squinting’ → Στράβων m.
§ n-Suffix seems to induce accent shift
o *h2ek̑ -ro- ‘pointy, top’ (Gk. ἄκρος) → *h2ó/ék̑ ri- m. (Lat. ocris, Gk. ὄκρις m., ἄκρις f.)
§ i-Suffix seems to induce o/e-ablaut (cf. Schindler 1980:390; Nussbaum 1998:523)
• ... it is not clear that an internal mechanism (i.e. accent shift, change in ablaut) can at the same time
entail suffixation:
o cf. Ved. kr̥ ṣṇá- ‘black’ → kr̥ ́ṣṇa- m. ‘(black) antilope’
§ no overt suffixation!
o cf. Nussbaum 2014:292ff. (esp. 297) for the claim that internally derived collectives/delibatives
(type Gk. ὕδωρ from *u̯ éd-ōr) actually involve two derivational steps: one suffixless internal
process forming a possessive adjective, and one external process (suffixation of *-h2) forming
the collective/delibative (**u̯ éd-or-h2).

• In other words: *-h2- as a substantivizing suffix ought to attach to adjectives only, and not to
substantives.
• It therefore seems unlikely that *-h2- attached to forms already substantivized via internal mechanisms
such as **pl̥ ́h2so- ‘cover’, **témHsro- ‘darkness’, **h2/3u̯ óbhso- ‘weaver’.
• It seems more likely that the mentioned examples that combine internal substantivization mechanisms
and *-h2-suffixation ...
o ́ 2seh2 ‘covering’ (Hitt. palaḫša-, Lat. palla f.)
*pl̥ h
o *témHsreh2 ‘darkness’ (Ved. támisrās, Lat. tenebrae f. pl. ‘darkness’)
o *h2/3u̯ obhseh2 ‘weaver’ (Lat. uespa f., Lith. vapsà f., Mid. Pers. vaβz, etc. ‘wasp’)
• … actually show accent retraction, e-Vṛddhi, o-Vṛddhi as a concomitant effect of *-h2-suffixation.
6
• In other words: we could conclude that this accent/ablaut behavior (accent retraction, e-Vṛddhi, o-
Vṛddhi) is a result or side effect of *-h2-suffixation/substantivization (and not an independent
phenomenon).
• But how exactly we have to interpret the mechanisms involved has to be left open for now.

7 The “Capricious O” in Greek


• In some cases, the position of the o-Vṛddhi vowel is not at all surprising:

o *g̑ heh1-ro- ‘abandoned, empty’ > Gk. χῆρος ‘bereaved, bereft’ (E., Call., A. R.; cf. χήρᾱ f.
‘widow’ Il.+)
→ *g̑ hóh1ro- (Gk. χῶρος, χώρᾱ ‘(empty) space, room; country(side)’ also ‘cavity
(med.)’, Epid. χώρᾱ ‘eye socket’); cf. IEW:419.

• Sometimes, as in the case of ὄµβρος, however, the o appears at the beginning of the word, prothetically,
almost like a prefix:

o *n̥ bh-ró- > *m̥ bh-ró- ‘wet, cloudy’


→ *ómbhro- (Gk. ὄµβρος m. ‘rain, storm’ Il.+); not †n/móbhro-.

o *tl̥ h2-ó- ‘carrying, carried’ (cf. Slav. *tьlo n. ‘ground, foundation; ceiling’, Ved. tulā́ - f.
‘balance, scales’)
→ *ótl(h2)o- m. ‘thing (to be) carried’ (Gk. ὄτλος m. ‘suffering, burden’ A., S.; cf.
ὀτλέω ‘suffer, endure’ Call., A.R., Lyc.); not †tól(h2)o-.
NB: A prefix/preverb *(h2)o- is semantically unattractive here.

• The strange position10 of the o could be explained as replicating a pattern in adjective/substantivization


pairs where the o-Vṛddhi had been introduced after an anlauting laryngeal11:
o *h1(e)r-s-ó- ‘rear, back’ (cf. Gk. *εἰρό- in εἰρεϑύρη· ὀρσοϑύρα; cf. MIr. err f. ‘end, tail’)
→ *h1órso- m. ‘the back, the rear’ (Gk. Att. ὄρρος m. ‘butt’, Gk. ὀρσοθύρη f. ‘back
door’ Od.)

o *h1régu̯ -os n. ‘darkness’ (Gk. ἔρεβος, Ved. rájas-, Goth. riqis)


→ *h1r̥ gu̯ -s-nó- ‘dark’ (unattested; but cf. the antonym *luk-s-nó- ‘bright’, subst. as λύχνος m.
‘lamp’)
→ *h1órgu̯ sneh2 ‘darkness’ (Gk. ὄρφνη f. ‘darkness, night’ Thgn., Pi.+); not

(h1)rógu̯ sneh2 (cf. Nikolaev 2012/2013 [2015]:201 note 72).

10
Another possible example is *g̑ h(e)s-lo- ‘thousand’ (cf. Ved. sahásra-, Gk. χίλιοι, Lat. mīlle) → *óg̑ hslo- m. ‘a (group of)
thousand’ (Gk. ὄχλος m. ‘crowd, mass, multitude; trouble’, pl. ‘the masses’ Pi., Ion., Att.). This is standardly reconstructed as
*u̯ óg̑ hslo- (root *√u̯ eg̑ h ‘move, convey’), which accounts better for the meaning of the denom. verbs (µετ-, ἀν-)ὀχλίζω ‘move by a
lever, heave up’ and ὀχλέω ‘move, roll’ (both in Hom.), but not for the meaning of ὄχλος itself and the lack of digamma traces in
Homer. Two separate lexemes?
11
Cf. Nikolaev 2012/2013[2015]:201 note 72 who reconstructs *(s)hxn̥ bh-ró- (Gk. ἀφρός) and *hxómbhro- (Gk. ὄµβρος). A possible
case is Gk. ὄγκος ‘bulk, mass; weight’ (= Ved. áṁśa- m., Av. ąsa- m. ‘portion’) if from *√h1nek̑ ‘take, receive’ or *√h2nek̑ ‘reach’
(cf. Hitt. nakkī- ‘important’).
7
• This analysis makes the following account of the antonymous ‘moon’-word plausible:
o *léu̯ k-os n. ‘light’ (Ved. rókas-, Av. raocah-, OP raočah-)
→ *luk-s-nó- ‘bright’ (subst. as λύχνος m. ‘lamp’)
or → *leu̯ k-s-nó- ‘bright’ (YA. raoxšna-, OHG liehsen ‘bright’)
→ *lou̯ ksneh2 f. (YAv. raoxšnā- ‘light’, Lat. lūna (Praenest. LOSNA), MIr. lúan, OCS
luna ‘moon’, OPr. f. pl. lauxnos ‘stars’).
NB: While some languages could continue an e-grade formation, OCS luna and OPr. lauxnos
presuppose an o-grade.

• Another example with anlauting laryngeal:


o *h2ek̑ -s-tó- ‘provided with a point, edge; pointy, sharp’ (Latv. aksts ‘nimble’, m. ‘point (of a
feather, pen, sting); cf. *aksti- > Lith. akštìs f. ‘skewer’, russ. ost’ f. ‘awn’, poln. ość f.
‘fishbone’; *akstu- > Lat. astus, -ūs m. ‘craft, cunning’ as *‘sharpness’ ?)
→ *h2ók̑ sto- m. ‘point, top’ (Gk. ὄχθος m. ‘eminence, bank, hill’ Ion. since h.Ap.)
→ *h2ók̑ steh2 f. (Gk. ὄχθη f. ‘any height or rising ground, natural or artificial, bank,
dyke by the side of a river’ ep. poet. Il.+)
NB: For the semantics cf. ἀκτή f. ‘promontory, rocky coast, rough shore, edge’ (Il.).

• The assumption of a “capricious O” could also explain the following formations with a structure R(z) +
“-όνος”, if we assume that the o was introduced between the root and the adjective-forming suffix of a
*R(z)-nó- adjective:
o χρόνος m. ‘time’ (Il.+), κλόνος m. ‘excitement, throng, turmoil (of battle); agitation’ (ep. poet.,
Il.+), θρόνος m. ‘throne, seat’ (Myc. to-no, to-ro-no˚; Il.+), Κρόνος m. ‘Cronus (the Titan)’,
φθόνος m. ‘envy, jealousy’ (Pi., Ion. Att.).12

• Note that in the cases where the root may have had a root-final laryngeal (κλόνος, θρόνος, Κρόνος; see
note 11), the o-Vṛddhi (at least descriptively) seems to have deleted any laryngeal effect.
o *k̑ l̥ h1-nó- ‘agitated’ → †k̑ l̥ h1óno- (†καλόνος), †k̑ lóh1no- (†κλῶνος), †k̑ ól(h1)no- (†κόλλος or

κοῦλος)
§ but rather → *k̑ l(h1)óno- > Gk. κλόνος ‘agitation’
• This seems odd and one might rather want to doubt the etymology, the formal analysis of these words
or the seṭ status of the root, than to accept this strange effect.

• There is, however, another plausible (νόστος type) example:


o *mr̥ hx-tó- ‘congealed, clotted’ (Ved. mūrtá- ‘coagulated’ AV+)
→ *mr(hx)óto- m. ‘curdled stuff’ (Gk. βρότος m. ‘coagulated blood, gore’ Hom.)

12
Gk. χρόνος m. ‘time’ if from *√g̑ her ‘to seize’ or *√g̑ her ‘short’ (*‘a short while’; cf. OIr. gar ‘short’)?; κλόνος m. ‘agitation’ if
from *√k̑ elh1 ‘to spur on’ (κέλοµαι; a seṭ root according to LIV2 Addenda (referring to O. Hackstein and I. Seržant, but not to Pinault
1984:120f.)?; θρόνος m. ‘throne, seat’ if from *√dherh2 (cf. θρᾶνος ‘bench’ and θρῆνυς ‘footstool’, Myc. nom. pl. ta-ra-nu-we; cf.
de Lamberterie 2004 who argues for original *dhór(h2)no- (Myc. to-no, Gk. *θόρνος), whence θρόνος (Myc. to-ro-no˚) in analogy
to θρᾶνος/ θρῆνυς. But also this *dhór(h2)no- could be interpreted as an o-grade substantivization of the adj. *dhr̥ h2nó- > *thrānó-
underlying θρᾶνος/θρῆνυς.); Κρόνος m. ‘Cronus (the Titan)’ if ‘the cutter/cut’ from *√(s)ker(hx) ‘to cut (off)’ (cf. Skt. kīrṇa-
‘injured’ Lex., PSlav. *kъrnъ ‘mutilated’)?; φθόνος m. ‘envy, jealousy’ (Pi., Ion. Att.) if from *√gu̯ hedh ‘to yearn for’ (cf. ποθέω)?
8
8 Conclusion and Open Questions
• Presented possible and (hopefully) plausible examples of the o-Vṛddhi type from a variety of
languages, that taken together confirm the PIE status of this substantivization process.
• Whatever its ultimate origin, I hope to have shown that the o-Vṛddhi type behaves in a similar way as
the e-Vṛddhi type and the accent-retraction type and that they are possibly interconnected.
• The o-Vṛddhi type has experienced some productivity in (the prehistory of) certain language branches.
• Using formally “marked” substantivization mechanisms can be explained by the speakers’ (occasional,
not universal) drive to make the substantivization as dissimilar to the underlying adjective as possible
(without, however, losing the ultimate connection).
• The “Caprice of O”, i.e. the strange position of the o-Vṛddhi vowel (type ὄµβρος and κλόνος), could
be due to this drive, but could also be rooted in reanalysis and analogy of/to established pairs of adjective
and substantivization.
o Open questions:
• If barytone accent was original in all o-Vṛddhi formations, how to account for oxytone examples?
§ cf. Gk. θῆµα n. ‘tomb’ and θωµός m. ‘heap’ (qua *dhoh1-m(n)-o-), φλέγµα n. ‘flame’
and φλογµός m. ‘id.’, etc. (cf. also Nussbaum 2018:242 note 43).
§ cf. *gu̯ hr̥ m-s-ó- ‘hot’ (subst. as *gu̯ hr̥ mseh2 > PCelt. *gu̯ rimsā > OIr. grís ā, f., MW
gwres ‘heat’) → *gu̯ hromsó- m. ‘heat’ (Ved. ghraṁsá- m. ‘id.’ RV, AV, Kauś).
o Analogy to other oxytone R(o)-formations, or phonological conditioning (cf. Rasmussen
1989:197-216)?
• What is the relationship/difference between the accent-retraction, the e-Vṛddhi, and the o-Vṛddhi type?
o Chronology, pure analogy, intraparadigmatic ablaut?
• And why is there sometimes variation in the position of the e-Vṛddhi and the o-Vṛddhi vowel (another
“Caprice of O”) in einzelsprachlich formations that seem to be clearly cognate?

o *sth2uró-/*stuh2ró- ‘large, stout’ (Ved. sthūrá- ‘id.’, OSwed. stūr ‘big’)


§ → *(s)téu̯ h2ro- m. ‘the stout one’ (YAv. staora- ‘large domestic animal’, PGmc.
*steura- and *þeura- ‘bull’, Ved. sthávira- ‘firm, large’); cf. Rau 2008:167 n. 33.
§ → *(s)téh2uro- m. ‘the stout one’ (Gk. ταῦρος m., Lat. taurus ‘bull’, Lit. taũras, OCS
turъ ‘aurochs’).

o *√u̯ i̯ eh1 ‘to wind’ (Ved. aor. ávyat, Lat. uiēre)


*u̯ ih1-nó- ‘winding’ (*u̯ ih1-(m)n-ó-?; cf. Lat. uīmen n. ‘withe’; Olsen 1999:439f., 840)
→ *u̯ ih1no- n. > Lat. uīnum n., Umbr. vinu ‘wine’
→ *u̯ ói̯ (h1)no- or *u̯ óh1ino- m. > Gk. (ϝ)οῖνος m. ‘wine’
→ *u̯ i̯ óh1no- or *u̯ ih1ono- m. > Hitt. wiyana- c. ‘wine’
→ *u̯ ói̯ (h1)neh2 or *u̯ óh1ineh2 f. > Gk. οἴνη f. ‘vine’ (Hes., etc.)

• But these questions will need to be addressed on a different occasion.

9
9 References
ADAMS, Douglas Q. (1985), ‘Sanskrit púmān, Latin pūbēs, and Related Words’, Die Sprache 31,1, 1–16.
AiGr: Jacob Wackernagel (‒ A. Debrunner ‒ L. Renou ‒ R. Hauschild) (1896-1954), Altindische Grammatik, 3 Bde.,
Göttingen: Vandenhoeck & Ruprecht.
DE LAMBERTERIE, Charles (2004), ‘Sella, subsellium, meretrix: sonantes-voyelles et ‘effet Saussure’ en grec ancien’, in:
Indo-European Perspectives, Studies in Honour of Anna Morpurgo Davies, hg. von J.H.W. Penney, Oxford:
University Press, 236-253.
EICHNER, Heiner (1973), ‘Die Etymologie von heth. mehur’, MSS 31, 53–107.
EWAia: Manfred Mayrhofer (1992‒2001), Etymologisches Wörterbuch des Altindoarischen, 3 Bde., Heidelberg: Winter.
FRISK: Frisk, Hjalmar (1960-70), Griechisches Etymologisches Wörterbuch, 2 Bde., Heidelberg: Winter.
GARCÍA RAMÓN, José Luis (2016), ‘Les substantifs du type νόστος, φόρτος, χόρτος en grec et leur préhistoire’, in: Nouveaux
acquis sur la formation des noms en grec ancien, Actes du Colloque international, Université de Rouen, ERIAC, 17-
18 octobre 2013, hg. von A. Blanc und D. Petit, Leuven, Paris: Peeters, 203-224.
HAMP, Eric P. (1982/83), ‘Indo-European substantives in *-mó- and *-mā́ -’, ZfvS 96, 171-177.
HARÐARSON, Jón Axel (2014), ‘Zur Entwicklung der neutralen s-Stämme im Germanischen’, in: FS Oettinger, 46–63.
HÖFLER, Stefan (2015b), ‘Ist der Wettergott ein Himmelsgott? Indogermanische Götternamen und ihr Beitrag zur internen
Derivation’, in: 1. Grazer Symposium zur indogermanischen Altertumskunde, “Der Mensch im Spannungsfeld
zwischen Ritual und Magie”, hg. von Christian Zinko und Michaela Zinko, Graz: Leykam, 148‒189.
––––––– (2017), ‘Observations on the palma rule’, Pallas – Revue d’études antiques 103, 15-23.
IEW: Julius Pokorny (1989²), Indogermanisches Etymologisches Wörterbuch, I. Band. Bern: Francke.
LIPP, Reiner (2009), Die indogermanischen und einzelsprachlichen Palatale im Indoiranischen, 2 Bde., Heidelberg: Winter.
LIV² Addenda: Martin Kümmel, Addenda und Corrigenda zu LIV², abrufbar unter
http://www.martinkuemmel.de/liv2add.html; letzte Änderung 03.02.2015 20:25.
NERI, Sergio (2013), ‘Zum urindogermanischen Wort für ‚Hand‘’, in: Multi Nominis Grammaticus, Studies in Classical and
Indo-European linguistics in honor of Alan J. Nussbaum on the occasion of his sixty-fifth birthday, hg. von A. I.
Cooper et al., Ann Arbor, New York: Beech Stave Press, 185-205.
––––––– (2016), Review of KROONEN, Guus Etymological Dictionary of Proto-Germanic, Leiden: Brill 2013, Kratylos 61,
1-51.
NERI/ZIEGLER 2012: Sergio Neri und Sabine Ziegler, »Horde Nöss«, Etymologische Studien zu den Thüringer Dialekten [=
Münchner Forschungen zur historischen Sprachwissenschaft (MFhS), Bd. 13], Bremen: Hempen.
NIKOLAEV, Alexander (2012/2013[2015]), ‘Homeric ἀάατος: Etymology and Poetics’, Die Sprache 50,2, 182-239.
NIL: Nomina im Indogermanischen Lexikon, hg. von Dagmar S. Wodtko, Britta Irslinger, Carolin Schneider, Heidelberg:
Winter 2008.
NUSSBAUM, Alan J. (1986), Head and Horn in Indo-European, Berlin: de Gruyter.
––––––– (1998), ‘Severe Problems’, in: FS Watkins, 521-38.
––––––– (2014), ‘Feminine, Abstract, Collective, Neuter Plural: Some Remarks on each (Expanded Handout)’, in: Studies
on the Collective and Feminine in Indo-European from a Diachronic and Typological Perspective, hg. von S. Neri
u. R. Schuhmann, Leiden: Brill, 273-306.
––––––– (2014b), ‘Greek τέκµαρ ‘sign’ and τέκµωρ ‘sign’: Why both?’, in: Akten Erlangen 2011, 215-60.
––––––– (2018), ‘Agentive and other Derivatives of “τόµος-type” Nouns’, in: Claire Le Feuvre, Daniel Petit & Georges-
Jean Pinault (eds.), Verbal Adjectives and Participles in Indo-European Languages, Bremen: Hempen, 233-266.
OLSEN, Birgit Anette (1999), The Noun in Biblical Armenian, Origin and Word-Formation – with special emphasis on the
Indo-European heritage, Berlin: Mouton de Gruyter.
OZOLIŅŠ, Kaspars (2015), Revisiting Proto-Indo-European Schwebeablaut, Dissertation: UCLA.
PANAGL, Oswald (1989), ‘Bahuvrīhi, Ellipse, Synekdoche’, in: FS Meid, 185-192.
PETERS, Martin (1980), Untersuchungen zur Vertretung der indogermanischen Laryngale im Griechischen, Wien.
––––––– (1997[2002]), ‘Indogermanische Chronik 35, Teil III’, Die Sprache 39/3, 94–129.
PINAULT, Georges-Jean (1984), ‘Benveniste et le tokharien’, in: E. Benveniste aujourd’hui, Actes du Colloque international
du C.N.R.S. (Tours, 28-30 septembre 1983), ed. by G. Serbat, J. Taillardat, Paris, 1984, vol. II, 109-124.
PROBERT, Philomen (2006), Ancient Greek Accentuation: Synchronic Patterns, Frequency Effects, and Prehistory, Oxford:
University Press.
RASMUSSEN, Jens Elmegård (1989), Studien zur Morphophonemik der indogermanischen Grundsprache, Innsbruck: IBS.
RAU, Jeremy (2008), ‘YAv. haosrauuah- and dǝ̄ uš.srauuah-’, HS 120, 159-168.
SCHAFFNER, Stefan (1997), ‘Der altfriesische Rechtsterminus bas(a)feng’, MSS 57, 97-114.
––––––– (2001), Das Vernersche Gesetz und der innerparadigmatische grammatische Wechsel des Urgermanischen im
Nominalbereich, Innsbruck: IBS.
SCHINDLER, Jochem (1980), ‘Zur Herkunft der altindischen cvi-Bildungen’, in: Lautgeschichte und Etymologie, Akten der
VI. Fachtagung der Indogermanischen Gesellschaft, Wien, 24.―29. September 1978, hg. von M. Mayrhofer, M.
Peters, O. E. Pfeiffer, Wiesbaden: Reichert, 386-93.
SCHUMACHER/MATZINGER (2013): Stefan Schumacher und Joachim Matzinger, Die Verben des Altalbanischen,
Belegwörterbuch, Vorgeschichte und Etymologie, Unter Mitarbeit von Anna-Maria Adaktylos, Wiesbaden:
Harrassowitz.
10
SCHWYZER, Eduard (1959³), Griechische Grammatik, auf der Grundlage von Karl Brugmanns Griechischer Grammatik, 1.
Allgemeiner Teil, Lautlehre, Wortbildung, Flexion, München: Beck.
SKARDŽIUS, Pranas (1943/1996), Lietuvių kalbos žodžių daryba, Rinktiniai raštai 1, Vilnius: Mokslo ir enciklopedijų
leidykla.
STEER, Thomas (2014a), ‘Von der Hysterokinese zur Amphikinese: Akzentgebundener Ablaut bei der Substantivierung
athematischer Adjektive’, in: Akten Erlangen 2011, 397–412.
––––––– (2015), Amphikinese und Amphigenese, Morphologische und phonologische Untersuchungen zur Genese
amphikinetischer Sekundärbildungen und zur internen Derivation im Indogermanischen, Wiesbaden: Reichert.
STÜBER, Karin (2002), Die primären s-Stämme des Indogermanischen, Wiesbaden: Reichert.
VINE, Brent (1998), Aeolic ὄρπετον and Deverbative *-etó- in Greek and Indo-European [= Innsbrucker Beiträge zur
Sprachwissenschaft, Vorträge und Kleinere Schriften 71], Innsbruck: IBS.
––––––– (2016), ‘Latin crassus, grossus, classis’, IF 121, 131-158.
WACKERNAGEL, Jacob (1920-1924), Vorlesungen über Syntax mit besonderer Berücksichtigung von Griechisch, Lateinisch
und Deutsch, 2 vols., Basel.
WATKINS, Calvert (1995), How to Kill a Dragon: Aspects of Indo-European Poetics, Oxford: University Press.
WEISS, Michael (2010), Language and Ritual in Sabellic Italy, The Ritual Complex of the Third and Fourth Tabulae
Iguvinae, Leiden: Brill.
WIDMER, Paul (2004), Das Korn des weiten Feldes, Interne Derivation, Derivationskette und Flexionsklassenhierarchie:
Aspekte der nominalen Wortbildung im Urindogermanischen, Innsbruck: IBS.
ZAIR, Nicholas (2012), The Reflexes of the Proto-Indo-European Laryngeals in Celtic, Leiden: Brill.

Appendix:

10 The νόστος type


Skardžius 1943:321; Probert 2006:174ff.; García Ramón 2016; EWAia I:150; Vine 1998:68; Schwyzer1959³:501; Steer 2015:100.

• It seems reasonable to assume that the type νόστος originates in o-Vṛddhi substantivizations of zero-grade
*-tó- verbal adjectives.
o *√nes ‘to return home safely’: *n̥ s-tó- ‘returned safely’ (subst. as Ved. ásta-, Av. asta- n. ‘home’)
→ νόστος m. ‘return home’ (ep. poet. Il.+)
o *√bher ‘to carry (away), bring’: *bhr̥ -tó- ‘carried (away)’ (Ved. ˚bhr̥ ta-, Av. ˚bǝrǝta-)
→ φόρτος m. ‘load, freight; burden’ (Od.+)
o *√pleu̯ ‘to (over)flow, to float’: *plu-tó- (Ved. plutá- ‘flooded, floating’ AV+, Gk. πλυτός ‘washed’)
→ πλοῦτος m. ‘wealth’13 (Il.+)
o *√pek̑ ‘to pluck’: *pek̑ -tó- (cf. denom. Gk. πεκτέω ‘shear, clip’)
→ πόκτος m. ‘fleece’ (Lyr. Adesp. 73, Hdn.)
Similarly: οἶτος ‘fate, doom’ (PGmc. *aiþa- ‘oath’); κοῖτος ‘bed; sleep’; χόρτος ‘enclosed place, farmyard’ (Lat.
hortus, Osc. húrz).
• García Ramón (2016) assumes analogical influence from τόµος action nouns (on which now Nussbaum
2018), which seems possible for the mentioned forms that all seem to go back to (concretized) verbal
abstracts.

13
For the semantics cf. Skt. pluta- ‘lengthened (of vowels)’, pluti- f. ‘lengthening’ (Pāṇ., Sū+) < *‘overflowing’. Or is the
underlying adjective a *pl̥ h1-u-tó- ‘provided with fullness’ (cf. *pelh1-u-u̯ ent- ‘id.’ as the basis of the PN Πελυεσσιος; cf. Peters
1997[2002]:103f.) > *pluh1tó- ‘rich, abundant’ whence *plóu̯ (h1)to- m. ‘wealth’? Alternatively, one could assume that the laryngeal
had already been lost in the basis *pól(h1)u- : *pél(h1)u- ‘fullness’, resulting in a possessive adjective *plu-tó- ‘provided with
fullness’, the ultimate basis of both Gk. πλοῦτος and Skt. pluta-, pluti-.
11
• Since, however, there are also νόστος type nouns that demand an interpretation as a concrete individualization
of the underlying adjective, it is perhaps easier to regard the νόστος type as a whole as a (morphological)
substantivization process.
o *mr̥ -tó- ‘mortal, dead’ (Ved. mr̥ tá-, Av. mǝrǝta-, Gk. βροτός)
→ *mórto- ‘mortal one’ (Gk. µόρτος· ἄνθρωπος Hsch. Latte; Ved. márta- m., GAv. maṣ̌ a-
m. ‘mortal’; cf. EWAia II:327)
• Accordingly, concrete nouns like πόκτος m. ‘fleece’ can be interpreted as ‘plucked thing’, χόρτος as
‘enclosed thing’, etc. (cf. the criticism in Probert 2006:179 against their interpretation as “verbal abstracts”).
• In a similar manner, one could account for the ὄγµος type (Gk. ὄγµος m. ‘furrow, swathe’ ep. poet. Il.+, Ved.
ájma- m. ‘course, path’ RV < *h2óg̑ -m(n)o-); cf. Schwyzer 1959³:492; Hamp 1982/83:171-177; Nussbaum
2014b:234 and 2018:242(citing oxytone examples).

11 Substantivizing o-Vṛddhi in Baltic, Slavic, Germanic


• The process of substantivizing o-Vṛddhi seems to have experienced some productivity in Baltic, Slavic,
Germanic, as can be seen from the following examples (again, substantivizations of possessive adjectives
based on neuter s-stems):
o *kréi̯ p-os n. ‘turn(ing)’ (cf. Lith. kreĩpti ‘turn’)
→ *krip-s-ó- ‘turning, turned’ (MW crych ‘wrinkled, curly’, Lat. crispus ‘curly’)
→ *kroi̯ pso- m. ‘a turn’ (OCS krěsъ m. ‘summer solstice’, Serb. krȉjes ‘St John’s
fire’).
o *pék̑ -os n. ‘wool, fleece’ (Gk.. πέκος n. ‘id.’)
→ *pek̑ -s-ó- ‘having wool’ (Lat. pexus ‘woolly’14)
→ *pok̑ so- ‘a shock of wool/hair’ (PGmc. *fahsa- n. in OHG fahs ‘headhair’,
OE feax ‘id.’, OIcel. fax ‘mane’).
h
o *u̯ énd -os n. ‘(a) hair’ (cf. OIr. find o, n. ‘Haar’)
→ *u̯ n̥ dh-s-ó- ‘having hair’ (subst. as OIr. fés ?, f.? ‘hair (coll.), pubic hair’)
→ *u̯ ondhseh2 f. ‘hair (coll.)’ (OPr. wanso f. ‘first hairs on the face’)
→ *u̯ ondhso- m. ‘hair’ (RCS ǫsъ m. ‘moustache, beard’, Cz. vous ‘beard hair’;
mostly pl. ‘beard’); cf. IEW:1148; Rasmussen 1989:202f. note 29.
o *sék-os n. ‘cut(ting)’ (Lat. secus n. ‘sex’ Plaut.+)
→ *sək-s-ó- ‘cutting, cut’ (Lat. saxum n. ‘stone, rock’ < *‘hewn piece of rock’)
→ *sokso- n. ‘a cutting thing’ (PGmc. *sahsa- n. ‘knife’ in OHG sahs ‘Messer’,
OE seax ‘dagger, knife’).
o *h3élh1-os n. ‘exhaustion’ (cf. Gk. aor. ὄλεσ(σ)α ‘destroyed, lost’; cf. Ved. alasá- ‘tired’)
→ *h3l̥ h1-s-ó- ‘exhausted’ (Lat. lassus ‘tired, exhausted’15; Lith. *ìlsas ‘id.’ as the basis of
il̃sti ‘become tired’, ilsė́ ti ‘to rest’ ?)
→ *h3ol(h1)seh2 f. ‘exhaustion’ (Lith. alsà f. ‘fatigue’)

14
Colum. 11.3.26: At Cappadocia [sc. lactuca] quae pallido et pexo densoque folio uiret “But the Cappadocian lettuce which grows
with a pale, woolly, thick leaf” (Loeb).
15
For the phonological side cf. Vine 2016.
12

You might also like