CC BY
36
R. Matasovic
LATIN TREE AND PLANT NAMES WITH THE SUFFIX -IX, -EX
Латинские названия деревьев и растений с суффиксом -ix, -ex. В статье трактуются латинские названия деревьев и растений с суффиксом -e/i-x, которые обычно считаются заимствованными из некоторого неизвестного доиндоевропейского субстрата, нпр. rumex, carex, filix, и ilex. Утверждается, что все-таки некоторые из этих слов имеют надежные индоевропейские этимологии: лат. carex 'осока' связывается с праславянским ^авь 'то же' и выводится из и.-е. корня *k'eh2- 'острить'; лат. rumex 'щавель' выводится из и.-е. корня *hirewg'- 'рвать', а лат. ilex 'вид дуба' связывается с праславянским *Пъ 'болото, ил' и гр. ilys 'болото', лот. ils 'черный'. Можно обнаружить, что первоначальная форма суффикса *-ik- правильно отражена в словах carex, rumex, и т. д., с переходом *i > *e /_-CC, а форма суффикса в словах salix, larix, и т. д. является результатом перехода *e > *i/_-ks в словах с корнями структуры CVR-, где V - любой краткий гласный кроме i и u.
Ключевые слова: латинские фитонимы, суффикс -ix, -ex, rumex, carex, filix, ilex. salix, larix.
Latin plant and tree names with the suffix -e/i-x include rumex 'sorrel', carex 'sedge', ilex 'holm-oak', salix 'willow, osier', larix 'larch-tree', vitex 'chaste-tree, Abraham's balm' and filix 'large fern, bracken'. Some of them are usually considered to be of substratum origin (cf. e.g. De Vaan 536 for salix, Balles 2008: 49 for larix, WH I: 678 for ilex), and the etymology of most of these words is at best uncertain and often completely unknown. Without denying that some of them are indeed early loanwords from unknown sources, in this paper we shall try to show that most of them have viable IE etymologies.
Latin carex, caricis 'reed-grass, sedge' has not been etymologized successfully. De Vaan (93) mentions the possibility that it is connected with carrere 'to card (wool)' (cf. also WH I: 167), but the semantics of the connection are far from clear, and, besides, this etymology would involve the additional hypothesis that the original form was *carrex, for which there is no evidence, and that carex is due to the alternation seen in littera vs. litera (the so-called 'littera-rule').
I propose to connect carex with Proto-Slavic *sast < *sast 'sedge' (Croat. sas, Slov. sas, sasje (coll.), Ukr. sas, Cz. dial. sas, Slovak sasina, sdsovina). These Slavic words which have exactly the same meaning as carex are sometimes considered to have been borrowed from Hungarian sds 'sedge' (Snoj 720, Machek 1997: 603), but it is important to note that the Hungarian word itself does not have an etymology (EWU 1309), so deriving the Slavic word from Hungarian is an obscurum per obscurius. At first sight the distribution of the reflexes of *sast in Slavic seems to be in those areas where the Hungarian influence is possible, but at a closer look this impression disappears. In Croatian, the reflexes of *sast are admittedly more common in dialects closer to the Hungarian border, but they are also attested in Cakavian, where ancient Hungarisms are very rare (almost unattested), e.g. Vrgada soasa (Jurisic 1973: 204), Pakostane sasa (ibid.), Gacka sasa. Both the forms sas (m.) and sasa (f.) are attested in some Stokavian dialects very far from the area where Hungarian influence is plausible, e.g. West Herzegovina sasa 'Carex acuta', Crna Gora sasak, saska (Jurisic 1973: 204)1. It is then plausible to conclude that PSl. *sast (perhaps also *sasa, because of Croatian dialectal forms) were inherited in Proto-Slavic, but preserved only in some central dialects of Slavic, including Ukrainian, Slovak, Croatian and some Czech dialects2. It was borrowed into Hungarian from Slavic, like several other Hungarian words for flora and fauna (e.g. repce 'colza, kind of cabbage' < PSl. *repa 'beet', bab 'bean' < PSl. *bobt, bodza 'elder, Sambucus' < PSl. *buzt, vidra 'otter' < PSl. *vydra, etc.).
From the formal side, PSl. *sasb is assimilated from *sast, and the reflexes of this form are preserved in Croatian and Serbian dialects: Skok (III: 383) notes sas, G sg. sasa in Crna Gora (Montenegro), and Sugar (2008: 608) has the attestation sas from Slavonia. The original form *sast is regularly derived from *k'eh2s-yo-, which can be from the same PIE root *k'eh2s- as Lat. carex (with -r-
1 Sugar (2008: 677) adduces also attestations of sas from Dubrovnik and Ston, as well as the original form of this plant name, sas 'a water plant' (Sugar 2008: 608), noting that this form is attested in Otok (Slavonija). The form sasa is attested in Herzegovina and sasika u vodi 'Typha L' is attested in an 18th century source from Dubrovnik. Sugar also adduces the form sasina 'Carex U in another 18th century document from Dubrovnik (this form corresponds to Slovak sasina noted above).
2 I have not been able to confirm the existence of Bulg. sasani 'Miesstroh' adduced by Skok (III: 383). Slovak sachor 'sedge' seems to be secondarily built to *sast with -ch- introduced on the analogy of examples where *s and *x alternate, cf. Croat. grah 'beans': grasak 'peas'.
from *-s- by rhotacism). Both PSl. *sast and Lat. carex can be derived from the PIE root *k'eh2- 'sharp, sharpen' (IEW 541-2), cf. Skr. sisati 'sharpens', Ved. sita- 'sharpened', Lat. catus 'sharp-minded, wise' (< *k'h2-to-), YAv. saeni- pointed' (< *kh2-i-ni-), with the suffix *-s-. LIV and NIIL reconstruct the PIE root as *k'eh3-y- (with *h3 rather than *h2) because of Lat. cos, cotis 'whetstone' and Arm. sowr 'sharp', but this is not decisive, as both can be derived from *k'oh2-t- and *k'oh2-ro- respectively3. The suffix *-y- seen in PGerm. *haino 'whetstone' (OE han, ON hein < *k'h2-ey-n-) may have been abstracted from the present-stem *k'h2-ey- (Kroonen 2013: 210). If Arm. sayr 'blade' is related, it may represent *k'h2-ri- or *k'h2ter (NIIL 412).
The semantic motivation for the etymology proposed here is in the sharp leaves of the plant, cf. the Czech (ostrice) and Bulgarian (ostrica) names of sedge, from the root of PSl. *ostrb 'sharp' (Russ. ostryj, Croat, ostar, etc.). Indeed, English sedge (< OE secg, as well as MLG segge) is also derived from the PIE root *sek- 'to cut' (Lat. secare, OCS sesti, etc.), just like Russ. osoka 'sedge' (with the o-grade of the root *sek-).
Latin rumex 'sorrel' is without an etymology. This word cannot be found in De Vaan's dictionary, and WH and EM do not offer any plausible etymologies either. I believe the key to understanding the origin of rumex is the Lithuanian noun rugstyne 'sorrel', which is derived from rugti 'to become sour' (Smoczynski 523). Indeed, in many languages (including English) the word for 'sorrel' is derived from the root meaning 'sour', cf., e.g. Croat. kiselica (from kiseo 'sour'), Germ. Sauerampfer (sauer 'sour'), etc. Lith. rugti has long u because of Winter's law, just like its Slavic cognate, *rygati 'to vomit, belch' (Croat. rigati, Russ. rygat', etc.). Both verbs can be regularly derived from PIE *h1rewg'- 'to belch' (Gr. ereugomai 'to belch out, vomit', Arm. orcam, Lat. rugio 'to bellow, roar', erugo 'belch' (De Vaan 528)4. Lat. rumex can be the regular reflex of PIE *h1rug-sm-, as *-gs- would have been lost before *m (cf. iumenta < *yowgsmenta, OLat. iouxmenta), cf. also the similar process in luna 'moon' < *lowksna < *lewksneh2 (Av. raoxsna-). It is possible that an original *rugsma or *rugsmen was changed analogically to *rugsmiks > rumex on the analogy with the other nouns denoting
3 Beekes (815) doubts that Gr. konos 'fruit of the pine cone, cone' belongs together with these words and prefers to consider it "of foreign origin".
4 It is possible that this root is also reflected in the Germanic words for (sour) cream, Germ. Rahm, OE ream, ON rjumi, perhaps also Av. raoyna-'butter', Kluge s. v. Rahm.
plants with the suffix *-ik-. A similar derivation can be observed in pumex, pumicis 'pumice' which is commonly derived from the root of the word for 'foam' (PIE *spoHimon / *spHimnos, cf. Lith. spaine, OCS pena, OHG feim)5, which is attested as an a-stem in Latin (spuma, De Vaan 593).
Latin ilex 'holm-oak, Quercus ilex' has, according to De Vaan (298) "no etymology" (cf. also WH I: 678, EM 308). However, I believe that it can be plausibly derived from the root *HeyHl- found in PSl. *ilt 'silt, clay' (Russ. il, G sg. ila, Cz. jil, Bulg. il, etc., cf. Vasmer I: 478, ESSJa VIII: 221f.) , Latv. ils 'very dark, black' (Karulis, s. v. Tls, Derksen 2008: 211), Gr. Tlys 'mud, slime, dregs, impurity ' (Beekes 589)7. The semantic motivation may lie in the dark, blackish bark of the holm-oak, or (less likely) in the muddy areas in which this tree grows (actually, it tends to grow in dry, rather than muddy areas; it is often found on slopes of hills, where muddy terrain is rare). It would not be unusual for a tree to be named after the colour of its bark, cf. PIE *bhrh1g'- 'birch' (Skr. bhurja- 'a kind of birch', Lith. berzas, Russ. bereza, OHG birihha, perhaps also Lat. fraxinus 'ash-tree', De Vaan 240f., IEW 139f.), from the root *bhreh1g'- 'to shine' (Skr. bhrajate, Lith. breksti 'to dawn'), cf. Kazansky 2015.
As noted by De Vaan (536) and other etymological dictionaries (EM 590f., WH II: 469), the Latin word for 'willow' has certain cognates in OIr. sail 'willow' (G sg. sailech), W helyg (< *sal-ik-), and OE sealh (< *salko-). De Vaan's assumption that this word is a loanword from some non-IE source seems hardly warranted. If the preceding etymology is correct, and the word for 'holm-oak' is derived from the adjective that originally meant 'dark, blackish', then it is plausible to derive salix 'willow' from the root *sal- (or *sHl-) meaning 'grey', cf. OHG salo 'dirty grey', OE salu 'id.', Russ. solovoj 'greyish' (< *sal-wo-), OIr. sal, saile 'dirt', salach 'dirty', W halog 'id.' This is of course not a new etymology of salix
5 See Matasovic 2004: 126 for the details of the PIE reconstruction. The semantic connection between pumice and 'foam' lies in the fact that pumice looks like dried foam (De Vaan 483).
PSl. *ilt was probably an original u-stem, as evidenced by the derivatives such as Russ. ilovatyj 'muddy', Croat. ilovaca 'muddy earth'.
7 Cf. also Gr. eily: melan (Hesych.). If this gloss-word is indeed related to the others, it may testify that the original meaning in Greek was 'dark, black', as in Latvian. Macedonian ilaks 'a kind of tree' mentioned by WH I: 678 may be related, if it existed.
(cf. IEW 879)8, but the semantic parallelism with ilex renders it slightly more probable. It is unclear which laryngeal should be posited, but if Gr. helike 'willow' is related, then we would have to reconstruct the root as *seh1l- / *sh1l-, noting that the full-grade *seh1l- appears to be unattested.
The etymology of larix 'larch-tree' is unknown (De Vaan 328, EM 342, WH I: 765). This word has been claimed to be borrowed from some Central European substratum, perhaps Celtic, but this claim represents no more than obscurum per obscurius, In the light of the etymologies for ilex and salix presented here, it is tempting to see a colour term in the root from which larix is derived. We might start from PIE *g'helh3- 'yellow, green' (Skr. hiri- 'yellow', hari-'green', Lith. zelvas 'greenish', OCS zelent 'green', Gr. khloros 'pale green, greenish yellow', Lat. holus, -eris 'greens, vegetables', perhaps also Lat. helvus 'yellow, dun', IEW 429 f.). Larch is indeed a very green tree, and its (male) flowers are yellow, but there are several formal problems facing this connection. First, initial *gl-(from *g(')hl-) should have been preserved in Latin (cf. gleba 'lump of earth' vs. Lith. glebti 'to embrace, clasp'), but it has been noted that *g- is sometimes lost before laterals, e.g. in ravus 'yellowish gray' vs. OHG grao 'gray' and gravastellus 'grey head' < *g' reh2wo- or *g' rh1wo- (Meiser 1998: 103). Meiser thinks that the loss of *g- before laterals is expected in Sabellic, and several Latin colour terms were indeed borrowed from Sabellic (e.g. rufus 'red-haired', perhaps also helvus 'yellow, dun'). Moreover, if Larva 'evil spirit, demon' (De Vaan 328) is derived from *g'hlh3-swo- (the semantic motivation being in the putrid, greenish colour of dead bodies), then this would represent another instance of the loss of original *g'h before *l in Latin9.
Another problem is the short -a- in Lat. larix, as we would expect *Clh3C- to yield Lat. *ClaC-, with long -a- (e.g. lana 'wool' < PIE *wlh2neh2, cf. Lith. vilna, Croat. vuna, Goth. wulla, etc.). However, the short vowel is attested in glaber 'smooth, bald' < *ghlh2dh-ro- (OCS gladtkt 'smooth', Lith. glodus 'smooth', OHG
8 It is unclear whether Skr. sara-, sala- 'grey' (mentioned in IEW 879) belong here as well (these words are not in EWA), and the same holds for OE sol 'dirty, dark'. If they do, they would have to be derived from PIE *seHl-, with the full-grade of the root.
9 Other examples may be Lat. landica 'clitoris', if it developed from *glandica (cf. Lat. glans 'acorn' and 'glans (penis)', cf. WH I: 758, and Lat. lac, lactis 'milk', which is probably from *glakt (cf. Gr. gala, galaktos 'milk').
glat, De Vaan 263, WH I: 603, EM 275, IEW 431f.). Schrijver (1991: 188) thinks that short -a- is the result of a special vocalization rule *CRHCC- > *CRaCC-, but there are too few examples for this to be ascertained. Hence, larix can be connected with the root *g'helh3- only if the admittedly speculative assumption of Sabellic intermediary is made. The proto-form may have been *glariks < *g'hlh3r-ik-, with *-ik- added by analogy after the other tree-names with this suffix. The starting point of the derivation would have been the adjective *g'hlh3-ro- parallel to Gr. khloros10.
Lat. vitex 'chaste-tree, Abraham's balm' is derived from the root *weh1i- 'to plait, weave, wind' (Lat. vieo, Skr. vyäyati 'wind, wrap', Croat. viti 'twist, wind', De Vaan 677, IEW 1120ff.), from which we also have Lat. vitis 'grape-vine'. The semantic motivation of this etymology lies in the uses of twigs of chaste-tree in basketry. The suffix may have spread to this word by analogy with salix, since chaste-tree and willow are quite similar.
Finally, filix 'large fern, bracken' (var. felix Paul. ex Fest.) is commonly derived from the root for 'henbane' *bhel- (W bele, Russ. belenä, OE beolone, belene, cf. De Vaan 220, IEW 120). The form filix is probably assimilated from earlier felix. The semantic connection is in the resemblance of the stems of henbane and the leaves of fern, as well as in the medicinal use of both plants.
All of the words discussed in this article are said to contain the same suffix, -ex, -icis, which sometimes occurs as -ix, -icis (Leumann 1977: 244, Balles 2008: 49f., Ernout 1946: 140-148). The nominative in -ex is far more common than the one in -ix11. Besides tree and plant names, this suffix is used to form names of parts of trees and plants (frutex 'shrub, shoot', caudex 'trunk of a tree', cortex 'bark, rind'), names of small animals and insects (culex 'gnat, midge', cimex 'bug, bed-bug', pülex 'flea'12, mürex 'purple-fish, shellfish yielding purple dye', natrix 'adder', ibex 'a kind of goat, chamois', laurex 'young rabbit cut out from the mother', fulix (=
10 The final stress of Gr. khloros points to another possible explanation of the short -a- in larix, namely Dybo's law, the shortening of pretonic long vowels in Italic and Celtic (as in Lat. vir 'man', OIr. fer < *wiHros, cf. Skr. viras). However, this law is highly controversial (cf. Matasovic 2012 and the references found there).
11 The ratio is approximately 8 words in -ix vs. more than 26 in -ex (including loanwords).
12 The adjective pedicosus 'full of lice' from pedis 'louse' perhaps points to the existence of yet another noun denoting an insect with the suffix *-ik-, *pediks.
fulica) 'coot, a water fowl, heron (?)', dentex 'a fish, Dentex dentex') and other objects, including a few body-parts (pollex 'finger', podex 'the fundament, anus', pantex 'bowels', varix 'dilated vein', fornix 'arch, vault'13, vertex, vortex 'whirlpool; top of the head, summit', apex 'a kind of pointed mitre, top of a helmet', imbrex 'semi-cilindrical tile', famex 'bruise, contusion', pumex 'pumice', silex 'pebble-stone, flint, lava'). Exceptional are rupex 'a rough, uncultivated man, a boor' and senex 'old man' (G sg. senis)14 which denote persons15. Besides, some nouns with this suffix are Greek loanwords, e.g. sorex 'shrew-mouse' from Gr. hyraks (or from the same source as the Greek word), calix 'chalice' from Gr. kyliks, paelex, pellex 'concubine' from Gr. pallaks, pallake, latex 'liquid, fluid' from Gr. lataks 'drop of wine' and (h)irpex 'a large rake' is considered to be either from Sabellic (hirpus 'wolf') or from Greek (harpaks 'thieving'). Apparently, the semantic core of nouns built with this suffix may have been made of diminutives ('parts of plants, small plants, small animals and insects'), whence its use spread to other domains. The Greek loanwords (and possibly loanwords from other languages) simply adjusted to the existing pattern. One has also to count with the possibility that there were originally two suffixes - one diminutive, and another found in plant and tree names - which were for some reason conflated in Latin. The diminutive *-ik- can be compared to Slavic diminutives in *-tc-, which were, however, thematicized (e.g. PSl. *ajtce 'egg' > Russ. jajco, PSl. *ertcb 'young goat' > Croat. jarac, cf. Matasovic 2014: 142ff.).
The original form of the suffix found in the words discussed in this article is subject to some controversy, as we find both -ix (salix, larix) and -ex (frutex, rumex, Tlex). However, if all of these words contain the same suffix (rather than two different suffixes conflated in Latin), then PCelt. *sal-ik- 'willow' (OIr. sail, G sg. sailech, W helyg, Matasovic 2009: 319f.) and the parallelism between Lat. natrix and PCelt. *natriks (OIr. nathir, G sg. nathrach, W neidr, Matasovic 2009: 286) prove that the original form must have been
13 Later also 'brothel', from the root offornus, furnus 'furnace' andfornax, G sg. fornacis 'kiln, oven' (the semantic connection is in the vault-shaped roof of the furnace).
14 Plautus (Cist. 372) has the pl. senices.
15 The meaning 'a boor' is probably of metaphorical origin; rupex is presumably derived from the same root as rupes 'a rock', so the original meaning may have been similar to the meaning of pumex and silex (for the semantic connection cf. Eng. blockhead).
*-ik-. It is generally assumed that PIE and Proto-Italic *e, *a yielded *e in the final closed syllable (Meiser 1998: 71), cf. *tibi-kan > tTbicen 'tibia-player', *haru-speks > haruspex, *arti-faks > artifex 'artist', *avi-kaps > auceps 'bird-catcher', but that final *i is preserved in this position. Meiser (1998: 71) adduces fornix 'vault' < *forn-ik-s (a derivative of fornus 'oven') as an example of this development. The final vowel of iudex 'judge' < *yows-diks is considered to be analogical after the type artifex, haruspex, etc., and pedes (pedess in Plautus, Aul. 528) < *ped-it-s 'pedestrian' and eques < *ekw-it-s 'horseman' must be analogical to mTles, G sg. mTlitis, of unknown etymology16. The reason behind this conclusion is that, if either -iCC or -eCC has to be analogical, the model existed only for -eCC, but not for -iCC, since the merger of *-eCC and *-aCC to *-eCC is undisputed.
If the vowel *-i- was indeed regularly preserved before final *-CC, then the final vowel must be original in salix, larix, and filix, and analogical in vTtex, rumex and Tlex, but it is unclear why the analogy spared the words that preserved the N sg. ending -ix. They do not share any formal or semantic properties not already found in the group with the N sg. ending -ex: besides the plant/tree names salix, larix, filix we find -ix in the animal names natrix 'adder', fulix 'coot', as well as in fornix 'arch, vault', and in the loanword calix. It can be observed that -ix is common in nouns with the canonical root structure CVR- (salix, larix, filix (< felix), fulix, calix, varix, where R = r, l), but natrix and fornix are exceptions, and -ex is found in culex 'gnat, midge' and silex 'flint', which also have the root structure CVR-. Although the exact history of the development of the suffix *-ik- in Latin will probably continue to elude us, the facts presented here make possible two alternative conclusions. The conservative approach would be to claim that final -ix was changed analogically to -ex in Latin except in nouns with the root structure CVR-, where V is any short vowel other than the high vowels i, u17.
16 De Vaan (379) believes that miles is originally a compound of milia 'thousands' and -it- found in ped-it-, equ-it- (from the root of iter 'way'). Miles would then be 'who goes with/by the thousand', but I do not find this really convincing.
17 The vowel -i- in natrix may have been preserved on the analogy with the agent-noun feminine suffix -trix (e.g. amatrix, feminine to amator 'lover'); it is derived from the root of no, nare 'swim', hence 'adder' is etymologically 'a swimmer'. Besides, Leumann (1977: 244) claims that the shortness of -i- in natrix is not established with certainty. Finally, fulix may have preserved -ix under the influence of the variant fulica (attested since Afranius; fulix is attested only since Cicero). Even so, the ending -ix in
The other possibility, which I would prefer, is to claim that *-i-before final *-CC became *-e-, thus merging with the reflexes of * eCC and *-aCC. In that case the forms of iüdex, pedes, eques are regular rather than analogical, and the final short *-e- in carex, ilex and rumex is expected. Subsequently, *-eCs became -iCs in nouns with the canonical structure CVR-, where V is any short vowel other than i and u18. This is the reason why we regularly find short *-i- in the last syllables of larix, salix, calix, etc.
References
Balles, Irene. 2008: "Nominale Wortbildung des Lateinischen", in: Rosemarie Lühr (ed.) Nominale Wortbildung des Indogermanischen in Grundzügen. Band 1: Latein, Altgriechisch, bearbeitet von Irene Balles. Hamburg: Verlag Dr. Kovac. Beekes, Robert S. P. 2011: Etymological Dictionary of Greek, Leiden: Brill.
Derksen, Rick. 2008: Etymological Dictionary of the Slavic Inherited
Lexicon, Leiden: Brill. De Vaan, Michiel. 2008: Etymological Dictionary of Latin and the other
Italic Languages, Leiden: Brill. EM = Ernout, A. & Meillet, A. 1967: Dictionnaire etymologique de la
langue latine, Paris. Ernout, A. 1946. Philologica, Paris.
ESSJa = N. S. Trubacev, (ed.) Etimologiceskij slovar' slavjanskix jazykov, Moscow.
EWA = M. Mayrhofer. 1992-2001: Etymologisches Wörterbuch des
altindoarischen, Winter: Heidelberg. EWU = L. Benkö (ed.). 1993-1997. Etymologisches Wörterbuch des
Ungarischen, Budapest. IEW = J. Pokorny. 1959. Indogermanisches etymologisches Wörterbuch, Bern.
Jurisic, Blaz. 1973: Rjecnik govora otoka Vrgade. II. dio: rjecnik, Zagreb: JAZU.
Karulis, K. 1992: Latviesu etimologijas värdnica. Riga: Avots. Kazansky, N. N. 2015: 'Nazvanie berezy v indoevropejskoj perspektive', in: Bereza. Sbornik statej, Moscow: Nauka, 5-15.
fornix is unexpected (-ex would appear to be regular after -or- followed by another resonant, cf. pollex 'finger'). It may be that -i- in fornix was preserved because of the influence of fornicatus 'arched, vaulted' (e.g. in the common expression via fornicata), where -i- is expected in the open internal syllable. Alternatively, it may be an allegro-form of the substantivized adjective *fornicus 'arch-shaped' (this is assumed by WH I: 534).
18 If they are valid, the explanations of the exceptions to the rule, natrix, fulix and fornix remain unchanged, whichever of the two alternative conclusions we accept.
Kluge = Fr. Kluge. 1999: Etymologisches Wörterbuch der deutschen
Sprache, Berlin: De Gruyter. Kroonen, Guus. 2013: Etymological Dictionary of Pro to-Germanic, Leiden: Brill.
Leumann, Manu. 1977: Lateinische Grammatik. Laut- und Formenlehre,
Syntax und Stilistik, München: Beck. LIV = H. Rix et alii. 2003: Lexikon der indogermanischen Verben,
Wiesbaden: Reichert. Machek, Vaclav. 1997: Etymologicky slovnik jazyka ceskeho, Prague:
Nakladatelstvi Lidove Noviny. Matasovic, Ranko. 2004: Gender in Indo-European, Heidelberg: Winter. Matasovic, Ranko. 2009: Etymological Dictionary of Proto-Celtic, Leiden: Brill.
Matasovic, Ranko. 2012: "Dybo's law in Celtic", Zeitschrift für celtische
Philologie 59: 129-141. Matasovic, Ranko. 2014: Slavic Nominal Word-Formation. Proto-Indo-European Origins and Historical Development, Heidelberg: Winter. Meiser, Gerhard. 1998: Historische Laut- und Formenlehre der
lateinischen Sprache, Darmstadt: Wiss. Buchgesellschaft. NIIL = D. S. Wodtko et al. 2008: Nomina im indogermanischen Lexikon,
Heidelberg: Winter. Schrijver, Peter. 1991: The reflexes of the PIE laryngeals in Latin,
Amsterdam & Atlanta: Rodopi. Skok = Skok, Petar. 1974. Etimologijski rjecnik hrvatskoga ili srpskoga
jezika, Zagreb: JAZU. Smoczynski = Smoczynski, Wojciech. 2007: Slownik etymologiczny jqzyka
litewskiego, Vilnius: Vilniaus Universitetas. Snoj = Snoj, Marko 2003: Slovenski etimoloski slovar, Ljubljana: Modrijan.
Sugar, Ivan. 2008: Hrvatski biljni imenoslov, Zagreb: Matica hrvatska. Vasmer, Max. 1955: Russisches etymologisches Wörterbuch, Heidelberg: Winter.
WH = A. Walde & J. B. Hofmann. 1938: Lateinisches etymologisches Wörterbuch, Heidelberg: Winter.
R. Matasovic. Latin tree and plant names with the suffix-ix, -ex
Latin plant and tree names with the suffix -e/i-x include rumex 'sorrel', carex 'sedge', ilex 'holm-oak', salix 'willow, osier', larix 'larch-tree', vitex 'chaste-tree, Abraham's balm' and filix 'large fern, bracken'. Some of them are usually considered to be of substratum origin (cf. e.g. De Vaan 536 for salix, Balles 2008: 49 for larix, WH I: 678 for ilex), and the etymology of most of these words is at best uncertain and often completely unknown. Without denying that some of them are indeed early loanwords from unknown sources, in this paper we shall try to show that most of them have viable IE etymologies.
Keywords: Latin plant and tree names, suffix -ix, -ex, rumex, carex, filix, ilex. salix, larix.
