Etymological Dictionary of Proto-Celtic

Ranko Matasovic

This is the first etymological dictionary of Proto-Celtic to be published after a hundred years, synthesizing the work of several generations of Celtic scholars. It contains a reconstructed lexicon of Proto-Celtic with ca. 1500 entries. The principal lemmata are alphabetically arranged words reconstructed for Proto-Celtic. Each lemma contains the reflexes of the Proto-Celtic words in the individual Celtic languages, the Proto-Indo-European (PIE) roots from which they developed, as well as the cognate forms from other Indo-European languages. The focus is on the development of forms from PIE to Proto-Celtic, but histories of individual words are explained in detail, and each lemma is accompanied by an extensive bibliography. The introduction contains an overview of the phonological developments from PIE to Proto-Celtic, and the volume includes an appendix treating the probable loanwords from unknown non-IE substrates in Proto-Celtic.

See Brill.com for more information on print edition.

Purchase Access
Author
Introduction
The Non-Indo-European Elements in the Celtic Lexicon
Bibliography
Ranko Matasovic, Ph.D. (1995), is Professor of Linguistics at the University of Zagreb. His interests include Celtic and Indo-European linguistics and language typology. He published nine books, including Gender in Indo-European (Winter, Heidelberg 2004).

Etymological Dictionary of Proto-Celtic

INTRODUCTION

1.The organization of the dictionary

This dictionary contains the lexical entries that can be more or less reliably reconstructed for Proto-Celtic. It is intended to contain Proto-Celtic words rather than roots, but in several cases, where the word formation of cognates in the attested Celtic languages differs, a rather speculative choice had to be made in order to decide on the Proto-Celtic form. In some cases the OIr. form was projected to Proto-Celtic, but in many instances the form with most parallels in other IE languages was postulated for Proto-Celtic as well. Whenever the exact Proto-Celtic form is underspecified, for one reason or another, this is clearly stated in the discussion following the lemma.

In this dictionary, a Proto-Celtic form is reconstructed whenever at least one of the following two conditions are met:

 

(1) Cognates are attested in at least two primary branches of Celtic. By primary branches I understand Goidelic (Irish, Scottish, and Manx), British (Welsh, Cornish, and Breton), Continental Celtic (Gaulish and Lepontic), and Celtiberian. Whether British was dialectally closer to Goidelic (the ‘Insular Celtic’ hypothesis) or to Continental Celtic (the ‘P-Celtic’ hypothesis) was considered irrelevant in deciding whether a given word was reconstructible for Proto-Celtic.

 

(2)  Probable cognates of a word, attested in only one branch of Celtic, exist in at least one other IE language.

 

PCelt. words are given as bare stems, e.g. the n-stem *talamon ‘earth’ is adduced rather than the Nom. sg. *talamū. Where ablaut patterns within paradigms of PCelt. nouns can be reconstructed, this was done in the discussion of particular lemmas. If the etymologically related words within Celtic do not agree in word-formation, the simpler form was usually projected to Proto-Celtic. For example, PCelt. *barinā ‘rocky ground’ is reconstructed on the basis of OIr. bairenn; it is assumed that the Brittonic forms (W brennigen, Bret. and Co. brennik) represent derivatives thereof.

The meaning of Proto-Celtic words is often rather difficult to reconstruct. Where meanings of cognates in various Celtic languages do not agree, either all of the attested meanings were projected to Proto-Celtic, or the meaning deemed most basic was reconstructed. Whenever the meaning of a particular PCelt. word remained the same in one or more of the attested languages, the meaning of the attested word was not adduced in the fields containing these reflexes. For example, PCelt. *wiro- ‘man’ has reflexes with identical meanings in OIr. and MW, so the meanings of OIr. fer and MW gwr were not adduced separately. The same principle was followed in adducing the meanings of the PIE forms and their reflexes: since the meaning of PIE *wiHro- ‘man’ was preserved in its reflexes (e.g. Skt. vīrá-, Lith. výras, etc.), it was not adduced in the field containing the attested forms in IE languages. The list of the attested cognates of the Proto-Celtic lemmata is not meant to be exhaustive. For the sake of conciseness, I usually adduced only cognates from two or three IE branches, usually those that are most relevant for the PIE reconstruction, and added a reference to Pokorny’s Dictionary (IEW), Lexikon der indogermanischen Verben (LIV), and/or Encyclopedia of Indo-Eruopean Culture (EIEC), where more detailed lists of cognates can be found.

Reflexes of the reconstructed PCelt. forms were given from all of the attested Celtic languages. However, since most of our knowledge about Gaulish, Celtiberian and Lepontic is derived from onomastic analyses, cognates in these languages are sometimes adduced although they are not established beyond reasonable doubt.

Every investigator of Celtic etymology must make a principled choice: one can argue that, since Celtic is a branch of Indo-European, it is a priori likely that words in Celtic languages have Indo-European etymologies. If one accepts this assumption, then finding any possible cognate in the IE lexicon is preferable to not giving an etymology at all. On the other hand, one could argue that we cannot possibly know the percentage of words that Celtic borrowed from non-Indo-European languages, so that any Celtic etymon may be equally likely to be inherited as it is to be borrowed from some unknown source. If this is the case, then one needs more than possible cognates in other IE languages in order to make an etymology plausible.

Let us take one example: OIr. ail (phonologically [al']) ‘rock, cliff’ is a very short form, consisting of only two segments. It could, in principle, represent a variety of Proto-Celtic forms (*ali-, *fali-, *yali-), and these could go back to an even larger number of possible PIE roots (*h2el-, *h2elH-, *ph2el-, *pelH-, *(s)pel-, *ph2el-, *yel-, *yeh2l-, etc.). It is obvious that, with such a short and isolated form, the possibility of finding chance resemblances in other IE languages is considerable. Many linguists would therefore consider any etymology of such a word hypothetical, and leave open the possibility that it was borrowed from some non-IE language. On the other hand, if one assumes that this word is much more likely to have been inherited than borrowed from some unknown source, then finding a possible set of cognates from the PIE root *pel- ‘rock’ (OHG feliza, etc.) is enough to make a plausible etymology.

I am not sure which of these two methodological principles one should adopt, but I thought it would not be fair to the reader to be too critical with respect to possible, but uncertain Indo-European etymologies of Celtic etymons. To do so would mean to limit oneself to trivial and well-established etymologies, and my feeling is that potential readers of this book do not expect it to contain just the information that, e.g., OIr. athir is related to Lat. pater. Etymological dictionaries are usually not best-sellers, but this does not mean that they have to be boring. This means that many lemmata in this dictionary should be understood as proposals to be evaluated, rather than as a collection of well-established scientific facts.

However, it was my intention to avoid too speculative etymologies, especially those that rely on alleged reflexes of PCelt. words in only one, poorly attested Celtic language. For example, the Gaul. month name ELEMBIU from the Coligny Calendar is usually[1] derived from the PIE word for ‘deer’ (PIE *h1eln-bho- > Gr. élaphos). The Greek month name elaphobolēion, derived from the same PIE word, is often adduced in support of this etymology. However, I did not include it in my lexicon, since the meaning of ELEMBIU is far from being assured, and there are no traces of this word in other Celtic languages (but cf. PCelt. *elantī, a different formation arguably from the same root). The form found in Coligny is actually compatible with many other interpretations, and in order to relate it to PIE **h1eln-bho- one would also have to explain the unexpected reflex of the syllabic nasal in Gaulish (em instead of *am). I have also tried to avoid all ‘last resort’ etymologies, which are often repeated in the handbooks simply because there do not seem to be any better Indo-European etymologies of particular words. A case in question is, e.g., OIr. dúil ‘creation’, which is commonly derived from PIE *dhuh2li-, from the root *dhuh2- ‘smoke’ (Lat. fūmus, etc.). Now, although it is possible to imagine a series of steps in semantic development that would lead from ‘smoke’ to ‘creation’, I find it difficult to believe this etymology: it seems to me that accepting it would be a sign of desperation, rather than the result of a sound consideration of probabilities.

In many similar cases, the fact that some often adduced etymology is not included in the lexicon means that I found it too incredible. On the other hand, I am sure that there are some good Celtic etymologies that were left out simply because I was unaware that they had been proposed.

2. The sources

In compiling the material for this lexicon, I have consulted all of the existing etymological dictionaries of Celtic languages published after 1950. I have not systematically used the older reference works, such as A. Holder's Alt-celtischer Sprachschatz, or W. Stokes' Urkeltischer Sprachschatz, because the material they contain has been well analyzed in later etymological dictionaries. So far the largest collection of Celtic etymologies can be found in Vendryès’ Léxique étymologique de l'irlandais ancien (LEIA); these are generally reliable, but often inconclusive and seldom very imaginative. Unfortunately, LEIA remains unfinished. Mac Bain's etymological dictionary of Scottish Gaelic is completely outdated and unreliable. Etymological notes in Geiriadur Prifysgol Cymru (GPC) are short, but often correct, and they remain the most valuable etymological resource for Welsh. A. Falileyev’s dictionary of Old Welsh is useful mostly for its rich philological documentation. Another valuable etymological source is Xavier Delamarre’s Dictionnaire de la langue gauloise, although it contains the etymologies only of those Celtic words that are attested in Gaulish. Gaulish loanwords in French and other Romance languages can be gathered from the relevant etymological dictionaries (e.g. Gamillscheg and FEW for French), but they have also been the subject of several articles, e.g. Bolelli 1941-2, Corominas 1976, Campanile 1983, Fleuriot 1991). Latin words of Celtic origin have been treated quite exhaustively in a paper by M. L. Porzio Gernia (1981). Words attested in Celtiberian inscriptions have been gathered and subjected to a careful philological and etymological analysis by Dagmar Wodtko in her Wörterbuch der keltiberischen Inschriften (MLH V.1). For Breton, we have two etymological dictionaries. The dictionary by Guyonvarc'h was conceived very ambitiously, but only a few fascicles were published; the new dictionary by A. Deshayes (2003) is reasonably complete and generally reliable, but does not offer detailed Proto-Celtic reconstructions and any IE etymologies. Furthermore, for Old Breton, we have a very careful and exhaustive work by Léon Fleuriot, Dictionnaire des gloses en vieux breton (DGVB). Finally, for Cornish we have only one etymological dictionary by E. Campanile, who analyzed the lexicon of the Old Cornish glosses. I have also made good use of Stefan Schumacher's Die keltischen Primärverben, which contains a lot of detailed etymological analyses of Celtic verbs with an Indo-European pedigree.

Apart from the mentioned sources, I consulted the reference works on Indo-European etymology, most notably Pokorny’s dictionary (IEW), EIEC, and LIV. Unfortunately, Nomina im indogermanischen Lexikon (Wodtko et alii 2008) appeared too late for it to be used systematically in the preparation of this dictionary. I also profited a lot from the etymological databases prepared for the ‘New Pokorny’ project by my Leiden colleagues, especially the Indo-Aryan database by A. Lubotsky, Latin and Italic by M. de Vaan, Hittite by A. Kloekhorst, and Baltic and Slavic by R. Derksen.

3. From Proto-Indo-European to Proto-Celtic

It is assumed here that PIE had the following phonemes:

 

I. Vowels

 

short                                                long

i                               u

        e          o                      ē          ō

             (a)

 

Phonemically, *i and *u were presumably just allophones of the semi-vowels *y and *w. The status of PIE *a is controversial. Following the Leiden school, I believe that PIE had no *a in the original, inherited lexicon (Lubotsky 1989), but this vowel occurs in several words that are probable loanwords from unknown, non-IE sources. In some cases, *a served as an epenthetic vowel separating difficult consonant clusters, e.g. Lat. pateo, < *pt-eh1- (cf. PCelt. *fatamā ‘palm of the hand, talon’).

 

 

II. Consonants

 

stops:                      

 

 labials         *p        (*b)      *bh

       

 

dentals         *t         *d        *dh

 

 

velars           *k        *g        *gh

 

 

palatalized   *        *       *h

velars  

 

labiovelars   *kw      *gw       *gwh

 

fricative:                              *s

 

laryngeals:                          *h1       *h2       *h3

 

resonants:                            *m       *n        *l         *r

 

glides:                                  *y        *w

 

The phonemic status of the difference between pure velars and palatalized velars in PIE is a disputed matter. It is quite probable that the phonological opposition between them was restricted to just a few environments. The syllabic resonants were just allophones of the non-syllabic resonants, occurring in the syllable nucleus. Therefore, they are not distinguished graphically from the non-syllabic resonants in the PIE reconstructions. The exact phonetic realization of PIE stops is a matter of controversy; the traditional ‘voiced stops’ may have been ejectives, perhaps in Early PIE. The phonetic realization of the ‘laryngeals’ is unknown, so they are marked with indexes (*h1, *h2, *h3). Laryngeals may have been lost in some environments already in PIE, or dialectally, not long after the dissolution of the proto-language, e.g. before *y (Pinault’s rule), or after the sequence *oR (de Saussure’s rule). However, the validity of these rules of laryngeal loss, as well as their exact formulation, are controversial.

Here are the principal Celtic sound changes ordered into an approximate relative chronology:[2]

 

 

A) Dialectal IE changes:

 

1. *h1e > *e, *h2e > *a, *h3e > *o

 

2. *eh1 > *ē, *eh2 > *ā, *eh3 > *ō

 

3. *CHC > *CaC, cf. PIE *ph2tēr > PCelt. *fatīr (OIr. athir).

 

4. *CstopHCstop > CstopCstop in non-initial syllables, cf. PIE *dhugh2tēr ‘daughter’ > PCelt. *duxtīr (Gaul. duxtir). This development is somewhat uncertain in the light of Celtib. tuateros ‘daughter’ (Gen. sg.).

 

5. TT > *-ss-, cf. PIE *krd-tu- > PCelt. *krissu- ‘belt’; the same development is found in Italic and in Germanic.

 

6. *CRHC > *CRaHC (> *CRāC), cf. PIE *plh1no- ‘full’ > PCelt. *flāno- (OIr. lán), PIE *rHno- ‘grain’ > PCelt. *grāno- (OIr. grán). Laryngeals were probably preserved after *Ra until the operation of Dybo’s law (A7), and then lost, with the compensatory lengthening of *aH > *ā. The change *CRHC > *CRāC occurred in Italic as well.

 

7. *VHC > VC in pretonic syllables (Dybo’s law, cf. Dybo 1961): PIE *wiHró- ‘man’ > PCelt. *wiro- (OIr. fer). In all non-problematic examples of Dybo’s law the laryngeal was lost after *i, *u, or *a which is the result of the development of syllabic resonants before laryngeals (A6).[3] It is assumed here that the laryngeals had already been lost after *e and *o, which were lengthened (A2). Dybo’s law was posterior to the change of CRHC > CRāC (A6) because of the development of *sfraxto- ‘eloquent’, *frati- ‘fern’, and *klad-o- ‘dig’. Something like Dybo’s law also operated in Italic,[4] and, in some form, probably in Germanic as well (cf. Lat. uir, OE wer < *wiHró-; maybe the vowel shortening (or laryngeal loss) was restricted to the position before resonants in Italic and Germanic). I assume that the operation of Dybo’s law in Celtic was general (i. e. unrestricted by phonetic environment).[5] The apparent exceptions to the operation of Dybo’s law in Celtic are best treated as analogical re-introductions of vowel length from the forms of the root where the length was preserved regularly. Of course, since the position of the accent in PIE cannot be established for many PIE etymons of PCelt. words, the operation of Dybo’s law can often be just assumed, but not strictly proved.

 

8. #RHC- > RaC (cf. Beekes 1988). Although this change is not universally accepted, it is found in the development of the following etyma: *latyo- ‘day’, *natu- ‘poem’, *mati- ‘good’, *mak-o- ‘increase’, *mad-yo- ‘break’, *laxsaro- ‘shine’ (PIE *r could not occur word-initially, so here R = m, n, and l). The same change occurred in Italic and Balto-Slavic, and probably in other European IE branches. The development of *#yHC- and *wHC- is uncertain, but cf. the lemmata *yalo- ‘clearing’ and *waxto- ‘bad’ for the possibility that *H > *a in this position.

 

9. Merger of PIE palatalized velars and pure velars, cf. PIE *dem  ‘ten’ > PCelt. *dekam (OIr. deich). This development is shared by all Centum branches of Indo-European.

 

 

B) Early PCelt. changes:

 

1. *gw > *b, cf. PIE *gwow- > PCelt. *bow- > ‘cow’ (OIr. ).

 

2. Deaspiration of aspirated stops, cf. PIE *bher-o- > PCelt. *ber-o- ‘carry’ (OIr. berid). This sound change was obviously posterior to *gw > *b (B1), because PCelt. *gwh > *gw, cf. PIE *gwher- > PCelt. *gwer-o- ‘heat up’ (MIr. geirid).

 

3. CLCstop > CLiCstop (where L = r, l), cf. PIE *ḱrd- > PCelt. *krid-yo- ‘heart’ (OIr. cride). It is probable that the same development occurred before PIE *m (cf. *kwrmi- ‘worm’ > PCelt. *kwrimi-, *h1lmo- ‘elm’ > PCelt. *limo- (s. v. *lēmo-, *limo-). This change was anterior to the general change of CRC > CaRC (B5) which was otherwise unrestricted by phonetic environment.

 

4. *eRa > *aRa (Joseph’s rule, cf. Joseph 1982), cf. PIE *terh1tro- ‘auger’ > PCelt. *taratro- (OIr. tarathar,W taradr). PIE *e did not become *a before *Rā (cf. PIE *gwenh2 >> *gweneh2 > *gwenā > PCelt. *benā, OIr. ben), and the vowel *e was restored analogically before *Ra in many instances, e.g. in the reduplicated syllables in the perfect (PCelt. *me-mad- > OIr. memaid ‘broke’, 3sg. perf. of maidid ‘breaks’ < PCelt. *mad-yo-). This change preceded the decomposition of syllabic nasals (B5) because of the development of PCelt. *elan(t)ī ‘doe, hind’ < *h1eln(t)ih2 (rather than *alan(t)ī), but after the vocalization of laryngeals between consonants, because of the development of *taratro- < *terh1tro- above.

 

5. CRC > CaRC, cf. PIE *dnt- > PCelt. *danto- ‘tooth’ (OIr. dét, W dant), PIE *mrwo- > PCelt. *marwo- ‘dead’ (OIr. marb, W marw), PIE *bhrso- > PCelt. *barso- > *barro- ‘point, top’ (OIr. barr). Note that syllabic liquids had already developed to *ri, *li before stops and *m (B3).

 

6. Loss of laryngeals in non-syllabic position. This change is later than the development of syllabic resonants (B5), because of, e.g., PIE *lHeto- > *kalHeto- > PCelt. *kaleto- ‘hard’, PIE *wlHo- > PCelt. *walo- ‘ruler, chief’, PIE *smh2eli- > PCelt. *samali- ‘similitude’, PIE *snHi > PCelt. *sani- ‘without’.

                                                                           

7. *p...kw > *kw...*kw, cf. PIE *penkwe > PCelt. *kwenkwe ‘five’ (OIr. cóic, MW pymp). This change predated the development of  *kw  > *x before stops (C1) if PCelt. *kwerxt- ‘bush’ (W perth) is from PIE *perkw- ‘oak’. It is assumed here that the similar assimilation in Italic (cf. Lat. quercus) is a parallel development (for arguments see below).

 

8. *ē > *ī, cf. PIE *Hrē- ‘king’ > PCelt. *rīg- (OIr. ). This change must predate PCelt. *p > *f > Ø (C4), because of PIE *h1epirom > PCelt. *efirom (> *eyrom > *ērom, OIr. íar ‘after’, not **ír).

 

9. *ō > *ū in final syllables, cf. PIE *ḱwōn ‘dog’ > PCelt. *kwū(n) (OIr. cú, W ci).

 

11. *V:RC > *VRC (Osthoff-type shortening before resonants in closed syllables), cf., e.g., PIE *h2weh1nto- ‘wind’ > *wēnto- > *wīnto- > PCelt. *winto-, perhaps also PIE *sih2m-do- > *sīndo- > PCelt. *sindo- ‘that’. This change was obviously posterior to *ē > *ī (B7).

 

 

C) Late PCelt. changes:

1. *C1C2 > *xC2 (where C2 stands for any stop and *s), cf. PIE *septm ‘seven’ > PCelt. *sextam (OIr. secht). This change is posterior to TT > *ss, and also to CRCstop > CRiCstop (B3) because of PIE *prptu- > PCelt. *frixtu- ‘form’, *mrgwto- > PCelt. *mrixto- ‘variegated’, *h2mlto- > PCelt. *mlixto- ‘milk’.

 

2. *pL > *bL (where L stands for any liquid), cf. PIE *pi-prh3-se- > PCelt. *pibrase- ‘will bestow, will give’ > *fibrase- (OIr. ebraid), PIE *dwey-plo- ‘double’ > PCelt. *dwēblo- (OIr. díabul).

 

3. *pN > *wN (where N is any nasal), presumably only after back vowels, cf. PIE *supno- > PCelt. *suwno- > *sowno- ‘sleep’ (OIr. súan, W hun).

 

4. *p > *f, cf. PIE *ph2tēr > PCelt. *fatīr ‘father’ (OIr. athir).

 

5. *ō > ā, cf. PIE *deh3no- > *dōno- > PCelt. *dāno- ‘gift’ (OIr. dán); this change is obviously later than the change of *ō > ū in final syllables (B8). Clear examples of this change in Celtiberian are lacking, but there are no counter-examples.

 

6. *ey > *ē, cf. PIE *(H)reyd- > PCelt. *rēd-o- ‘ride’ (OIr. réidid). This change was obviously later than *ē > *ī (B7). There is some uncertainty whether this change also occurred in Celtiberian.

 

7. *ew > *ow, cf. PIE *newyo- > PCelt. *nowyo- ‘new’ (OIr. núae, W newydd).

 

8. *uw > *ow/_C, cf. PIE *supno- > PCelt. *suwno- > *sowno- ‘sleep’ (OIr. súan). This change probably did not apply before *-i- in the next syllable, because of *dru-wid- > OIr. druí ‘druid’ (rather than **droí). It is unclear whether this change applied in Proto-Celtic, or just in Goidelic and Brittonic (data from Gaulish and Celtiberian are lacking).

 

 

D) Some other probable PCelt. changes:

 

1. The liquid assimilations *rp > *rf > *rr (PIE *serp- > PCelt. *serrā) and *lp > *lf > *ll (PIE *kulp- > PCelt. *kul(f)o-), *-rs- > *-rr- (PIE *bhrso- > PCelt. *barro-), *rst > *rt (PIE *trstu- ‘thirst’ > OIr. tart). All of the attested languages show the results of assimilations, so it is simpler to project those changes to Proto-Celtic. It is possible, however, that at least some of the assimilations were parallel innovations of individual languages after the break-up of Proto-Celtic.

 

2. The assimilation of *mw > *ww, cf. *kom-wīro- > PCelt. *kowwīro- ‘true’ (W cywir).

 

3. The lengthening of the vowel before the cluster *xsL, cf. *toḱ-slo- ‘axe’ > *tōxslo- > PCelt. *tāxslo- (OIr. tál); it is possible that PCelt. *x was lost and that the preceding vowel was subject to compensatory lengthening. However, *x is conventionally retained in the PCelt. reconstuctions because the regularity of this change is uncertain; cf. the lemmata *dīro-, *kīsrā, *muxto‑, *sego- and *skāxslo- for possible instances of this change, but also *tullo- for a possible counter-example.

 

4. The ‘liquid metathesis’ (*ar, *al > *ra, *la) between labials and a cluster of two dentals: PIE *mlsto- > PCelt. *mlasto- ‘taste’, PIE *gwrsto- > PCelt. *brasso- ‘great’, PIE *h1wrsto- > PCelt. *wrasto- ‘shower’, PIE *wlsno- > *walsno- > PCelt. *wlanno- ‘blood red’, etc. Like the preceding one, this change is proposed here for the first time, so it will probably raise some controversy. Another possibility is to assume an analogical zero-grade CCLaC of the roots with the full grade CLeCC (after the regular pattern with the zero-grade CaLCC  and the full grade CeLCC).[6]

 

5. The loss of laryngeals after *ey (and *oy?) before consonants (*VyHC > *VyC), cf. PCelt. *bēto- ‘food’ (W bwyd) < PIE *gweyh3to-, PCelt. *dēno- ‘fast’ (OIr. dían) < *deyHno-, PCelt. *fētu- ‘(grass-)land, territory’ (OIr. íath) < PIE *peyHtu-, PCelt. *kwēno- ‘long’ (OIr. cían) < *kweyHno-, PCelt. *rēno- ‘large expanse of water’ (OIr. rían) < PIE *(H)reyHno-, PCelt. *wēro- ‘crooked’ (OIr. fíar, W gŵyr) < PIE *weyh1ro-, etc. There are only two apparent counter-examples in this dictionary: W rhaeadr ‘torrent’ and OIr. ríathor (with disyllabic ía), do not represent PCelt. *reyatro-, but rather *riyatro-; likewise, OIr. disyllabic bíad ‘food’ can be derived from *biyato- or, less probably, *biwato- (rather than *beyato- < *gweyh3to-). There are no truly reliable examples for the loss of laryngeals after *oy, but cf. PCelt. *koylo- ‘thin’ (which can be from PIE *koyHlo-) and *oyno- ‘one’ (which some linguists would derive from *oyHno- < *h3eyHno-).

 

 

E) Some doubtful changes:

 

1. *(C)RHCdentalC > *(C)RaCdentalC, cf. PIE (?) *prh3-sneh2 ‘gift, share’ > PCelt. *frasnā (OIr. rann); however, PCelt. *frasnā, just like Lat. pars, can be derived from a root without laryngeal (generalized after the present stem *pr-neh3- / *pr-nh3- (> PCelt. *far-na-), and explained by ‘liquid metathesis’. Alternatively, the short *a in the Celtic reflexes of PIE roots of the form *CeRHC can always be the result of Dybo’s law, and it is, of course, more economical to assume fewer sound changes.

 

2. #RCvoicedC > #RaCvoicedC (Schrijver’s rule, cf. Lat. magnus < *mnos); however, all of the alleged examples involve the root *meh2- ‘great’ (cf. PCelt. *maglo-, *magyo-, and *magos). None of those etymologies is beyond reproach.

 

3. *h2rCstop- > *arCstop- and *h3rCstop- > *arCstop- (but *h1rCstop- > *riCstop-, cf. *rig-o- ‘go’ < *h1rgh-o-).[7] However, examples of this change are few in number and quite controversial. PCelt. *orgyā ‘testicle’ can be derived from *h1orhi- rather than *h1rhi- or *h3rhi- assumed by some linguists; PCelt. *arto- ‘bear’ is from PIE *h2rto-, but it probably went through the stage *h2rþk’o- (and *#rþC- > *#arþC- may be assumed just like *#rsC > *#arsC). PCelt. *arganto- ‘silver’ is a problem, since the reconstruction *h2rnto- seems somewhat more probable than *h2ernto-. However, the word for ‘silver’ may have had an ablauting paradigm in PIE (Gr. árgyros can be both from *h2er- and *h2r-, and Skt. árjuna- ‘shining’ is clearly from *h2er-).

4. The Problem of Italo-Celtic

Although Italic shares a number of sound changes with Celtic, I remain unconvinced of the ‘Italo-Celtic hypothesis’. Very few phonological and morphological changes are actually exclusive Italo-Celtic isoglosses, and, more importantly, one cannot really establish a relative chronology of those isoglosses, as one can in the case of, e.g., Balto-Slavic. However, there is little doubt that Italic and Celtic developed from a group of closely related Western Indo-European dialects. For a recent discussion of the Italo-Celtic hypothesis see Kortlandt 2007: 151-157.

The following phonological innovations of Italic and Celtic are shared:

 

1. The development of PIE syllabic resonants followed by laryngeals, PIE *CRHC > *CrāC, cf. PIE *rHno- ‘grain’ > Lat. grānum, PCelt. *grāno-. Note, however, that in PCelt. the development was actually from *CRHC to *CRaHC, and then to *CRāC with loss of the laryngeal and compensatory lengthening of *a (see above, changes A6-A7 in the relative chronology). It is uncertain whether the same two-step development occurred in Italic.

 

2. The assimilation *p...kw > *kw...kw. However, this change appears to be late in Celtic. It failed to occur in OIr. deac, deëc ‘10’, which is often derived from *dwey-penkw-om ‘two-fives’, and when this compound was formed (in Proto-Celtic) *p was not in the beginning of the word, and so it regularly changed to *f > Ø rather than assimilating with *kw (see Watkins 1966: 145, but also the lemma *dekan below for problems with this etymology). In any case, such an assimilation is phonetically trivial (cf. the reverse assimilation in PIE *penkwe ‘5’ > Go. fimf).

 

3. The shortening of vowels in pretonic position (Dybo’s law mentioned above); however, this change may not be limited to Italic and Celtic, because it appears to affect Germanic as well, at least in some examples, cf. OE wer ‘man’ < *wiHró- (Skt. vīrá-, Lith. výras, Lat. uir, OIr. fer).

 

Morphological Italo-Celtic isoglosses are not more convincing:

 

1. The genitive ending *-ī is neither pan-Celtic (it is lacking in Celtiberian) nor pan-Italic (it is lacking in Sabellic), and it is not exclusively Italo-Celtic (it occurs in Messapic and probably in Tocharian). Actually, it is an old petrified adjectival form (see Matasović 2004) and, as such, does not represent a common innovation in Italic and Celtic.

 

2. The generalization of the *so- stem of the PIE demonstrative pronoun *so-/*to- is a parallel development, since there are clear traces that PCelt. still had the pronominal stem *to- (see PCelt. *tod). Moreover, it is unclear whether Celtiberian shared the generalization of the *so- stem.

 

3. The introduction of the Gen. ending *-strom in the 1st and 2nd person plural pronouns is not very significant, since the stems of the 2nd person plural pronouns are different in Celtic and Italic, and the forms that actually have this ending are attested only in OIr. (nathar) and Latin (nostrum), so we cannot be sure if they were ever pan-Celtic and pan-Italic.

 

4. The spread of the Dat./Abl. pl. ending *-bhos is uncertain, since in Gaulish we only have -bo, and Irish generalized the ending *-bhis from the Instrumental.

 

5. The superlative ending *-smmo- is indeed a shared exclusive isogloss, but in itself it is not enough to prove the existence of an Italo-Celtic protolanguage.

5. The Sub-classification of Celtic

The exact genetic subclassification of the Celtic languages is still an unsettled matter. Two approaches dominate the current discussions:

 

(1)           The traditional view, defended, among others, by Schmidt (1977), Koch (1992), and de Bernardo Stempel (2006) who classify Brittonic together with Gaulish (and Lepontic, which is probably just an early offshoot of Gaulish) into Gallo-Brittonic, while the Goidelic languages remain as a separate branch of Celtic (see Fig. 1.1. below).

 

(2)           An alternative theory, defended by e.g. McCone (1996) and supported by Schrijver (1995) and Schumacher (2004), who see Brittonic and Goidelic as a separate Insular Celtic branch, while Gaulish and Lepontic are viewed as the Continental Celtic branch. Celtiberian, as is becoming increasingly clear, is almost certainly an independent branch on the Celtic genealogical tree, one that became separated from the others very early (see Fig. 1.2. below):

 

 

Fig. 1.1:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 


Fig. 1.2:

 

 

 

 

 

 

 

 

 

 


The problem of sub-classification within Celtic is relevant to this work only inasmuch as it affects the reliability of Proto-Celtic reconstructions. If there was an Insular Celtic branch within Celtic, then forms reconstructed on the basis of just Old Irish and Middle Welsh, the two best attested early Celtic idioms, need not go back to common Proto-Celtic, but may instead represent Proto-Insular Celtic. Likewise, if we assume the existence of a Gallo-Brittonic branch, then we should be careful in attributing reconstructions based on evidence from the Brittonic languages and Gaulish to Proto-Celtic.

As I have argued elsewhere (Matasović 2008), I tend to view Insular Celtic more as an areal than as a genetic grouping. This does not, however, imply that I believe in Gallo-Brittonic as a valid genetic grouping, either: in the matters of genetic sub-classification within Celtic I think it is wise to remain agnostic, until more is known about Gaulish, Lepontic, and Celtiberian. In any case, since the argument about Insular Celtic vs. Gallo-Brittonic tends to revolve more around the morphological than around the phonological isoglosses, it is unlikely that the eventual resolution of the debate will substantially affect the Proto-Celtic reconstructions proposed here.

6. On the reconstruction of Proto-Celtic

There are several unresolved issues in the reconstruction of the Proto-Celtic phonological system. I have generally tried to follow the consensus opinion, where there is any, but in some cases difficult choices had to be made. It is assumed here that Proto-Celtic had the following phonemes:

 

 

I. Vowels

 

a) short                                            b) long

 

i                               u                      ī                                   ū

        e          o                                              ē

                               

             a                                                           ā

 

 

c) diphthongs

 

        oy                 ow

 

ay    aw                                            āy        āw

 

I take the monophthongization of PIE *ey > *ē to be a Proto-Celtic change, although it is not absolutely certain that this change occurred in the prehistory of Celtiberian (cf. MLH V.1: XVII). I also believe that the change *ew > *ow is Proto-Celtic, and that instances of alleged eu in Gaulish (e.g. in Neviodunum, a toponym in Slovenia) are just spelling variants of a diphthong that did not exist in Latin at the time of the adaptation of the Roman alphabet to Gaulish (cf. McCone 1996).

 

 

II. Consonants:

 

a) stops:

 

                    b

 

t                   d                     

 

k                  g

 

kw                     gw

 

I assume that *kw merged with *kw in Proto-Celtic, so I reconstruct PCelt. *ekwo- ‘horse’ (OIr. ech, etc.) from PIE *h1ewo- (Lat. equus etc.). Apparent exceptions, such as W ci ‘dog’ < PIE *wōn can be explained by assuming early delabialization of *kw in certain environments (e.g. before PCelt. *ū as in the preceding example: *wōn > PCelt. *kwūn > *kūn > W ci). Similarly, the reflexes of *gw(h) and *g(h)w are indistinguishable in Celtic, cf. PCelt. *tangwāt- ‘tongue’ < *dnhw- (OIr. tengae, W tafod).

 

b) fricatives:

 

f      s           [x] (an allophone of *k before stops and *s]

       [z] (an allophone of *s before voiced consonants)

 

I do not assume that there was a PCelt. phoneme *ts (from PIE clusters with two dentals, and/or from PIE *-st-). I believe that PIE *st was preserved in PCelt. (as it is in Celtiberian), and that PIE *TT yielded *ss already in PCelt. (see Schrijver 1995). The fricative *f is the regular reflex of PIE *p. It may have been a bilabial voiceless fricative [φ] phonetically, rather than a labiodental fricative [f]. I also assume that the assimilations of *rs > *rr and *ls > *ll are Proto-Celtic (see McCone 1996); however, I adopt the ‘etymological’ spelling for the clusters *-sr-, *-sl-, *-sn-, *-sm-, and *ly, as if they were intact in Proto-Celtic reconstructions, although they could have changed to *-rr-, *-ll-, *-nn- and *-mm-, respectively, already in PCelt.[8]

 

c) resonants:

 

m     n          l           r

 

I assume that the change of PIE syllabic *m, *n > *am, *an is pan-Celtic. The fronting of *am, *an > *em, *en in Goidelic is a later development that occurred only in some environments (see McCone 1996 for details). I also assume that word-final *-m was preserved, as it is in Celtiberian, and occasionally in Gaulish.

 

d) semivowels:

 

y      w

7.The Celtic languages

For the purpose of this lexicon we adopt the following periodization of the attested Celtic languages.

 

1. Lepontic (attested from the 7th, or early 6th century B.C. until ca. the 1st century B.C.). In all likelihood, Lepontic is just an early offshoot of Gaulish. The evidence for Lepontic as a separate branch of Celtic heavily relies on the archaeological data, especially on the early individualization of the (Lepontic) Golasseca Culture (see Uhlich 1999: 285-293).

 

2. Gaulish, attested onomastically since the 6th century B.C., but with a sizeable corpus of inscriptions only from the 3rd century B.C. (inscriptions in Greek alphabet). Inscriptions in the Roman alphabet are attested later, chiefly after the Roman conquest of Gaul (2nd half of the 1st century B.C.). It is unclear when Gaulish died out, but it was probably spoken until the 6th or 7th century A.D., at least in some isolated pockets in Gaul. Although Gaulish is attested for at least a millennium, no attempt has been made to distinguish between early and late Gaulish in this lexicon. However, the source of Gaulish words (except for names) is always indicated.

3. Celtiberian, attested from the 3rd or early 2nd century B.C. until ca. 1st century A.D. The earliest inscriptions are in Iberian syllabary, but from the 1st century B.C. a considerable number of inscriptions are in Roman alphabet.

 

4. Goidelic, represented by Old Irish and Middle Irish, attested since 4th century A.D. (Ogam inscriptions). We distinguish the following phases of Goidelic:

 

Ogam (4th - 6th centuries)

Old Irish (7th - 9th centuries)

Middle Irish (10th - 11th centuries)

Modern Irish (12th century - present)

 

Usually only the earliest attested form of the word is adduced in the lexicon. This is regularly the Old Irish or the Middle Irish form found in DIL. It should be noted that Old Irish and Middle Irish forms are often not easily distinguished. In principle, all words attested in the glosses and other texts from the Thesaurus Palaeohibernicus are Old Irish, but Old Irish forms can often be found in later manuscripts as well. So, even if a word is only attested in texts, the manuscripts of which were preserved in the Middle Irish period (e.g. in the sagas of the Ulster cycle, or in the Leinster eulogistic poetry), we can often be sure that the same word existed in Old Irish. Therefore, in some cases where I adduced a word as OIr., although it does not occur in the proper OIr. texts, the reader will have to trust my philological judgement, or check the sources for himself.

Scottish Gaelic and Manx forms were not adduced in this lexicon (with a handful of exceptions), since they yield very little additional information about the reconstructed PCelt. words and their origin.

 

5. Brittonic,[9] represented by Welsh, Breton, and Cornish. Dialectal diversity within Brittonic is far greater than within Goidelic, so reflexes of PCelt. words from all three Brittonic languages were adduced, whenever attested. We distinguish the following phases of Brittonic:

 

Old Welsh (7th - 10th centuries)

Middle Welsh (11th - 14th centuries)

Modern Welsh (15th century - present)

 

Old Breton (9th - 11th centuries)

Middle Breton (12th - 16th centuries)

Modern Breton (17th century - present)

Old Cornish (9th -12th centuries)

Cornish (13th - 18th centuries)[10]

 

Middle Welsh, Middle Breton, and Middle Cornish forms are adduced by default. If a word was attested in Old Cornish, Old Breton, or Old Welsh, it is adduced separately in the Cornish, Breton, and Welsh fields, respectively. Modern Welsh forms, as cited in GPC, are adduced only when they are different from MW forms, and often the difference lies only in spelling.

8. Structure of the entries

Each entry in this dictionary consists of several fields. The first field contains the reconstructed Proto-Celtic word and its meaning, as well as the information about the word-class to which it belonged. The following fields contain reflexes of the reconstructed etymon in the primary branches of Celtic, together with some basic grammatical information about them: as a rule the gender of Old Irish and Middle Welsh nouns, the inflectional class of the Old Irish nouns and adjectives, and the attested stems of Old Irish verbs.

In the next field the PIE root of the Proto-Celtic etymon is given, together with a reference to the page in IEW where that root is discussed. After that, there follows a discussion of the proposed etymology and the alternative proposals found in the literature. I have tried to make the derivation of the attested forms from PIE and Proto-Celtic as explicit as possible, without concealing any of the uncertainties or unresolved problems. This field also contains the discussion of other possible cognates of the reconstructed Proto-Celtic etymon. In many cases the etymology proposed here is not the only possible one, but all of the etymologies in this dictionary are meant to be consistent with the Celtic sound laws accepted in this Introduction (see above). I have tried to be maximally clear in making the distinction between mentioning various possible etymological proposals and claiming that a particular etymology is true.

The last field contains the references. The list of references is not meant to be exhaustive, but rather to include the most relevant books and articles, published during the last fifty years, in which etymological discussion of the etyma in question can be found. They are ordered in such a manner that the more general reference works precede publications dedicated to a particular word, or specific problems of phonological development relevant to the lemma in question.

 



[1] See, e.g., McCone 1996: 70, 74.

[2]  Cf. McCone 1996, Isaac 2007, Kortlandt 2007: 117-120 for similar attempts.

[3]  For the alleged loss of laryngeals after PIE *e, *o in pretonic position see the lemmata *siti- and *omo-.

[4] A clear counter-example is the length of Lat. fūmus vs. Skt. dhūmá- < PIE *dhuh2mós. I find none of the explanations of this exception compelling (e.g. the analogy with fūligo "soot", de Vaan 2008). However, the general impression is that Dybo’s law solves more problems than it creates.

[5] Isaac 2007 attempts to limit the operation of Dybo’s depending on the nature of the laryngeal in question, but I do not find his argumentation persuasive.

[6] Note that this explanation will certainly not work for PCelt. *wrasto- (the PIE root is *h1wers-, not **h1wres-).

[7] Joseph 1982: 50-51, McCone 1996: 52, Isaac 2007: 73. It appears that the sequences *#h2L-, *#h3L-, on the one hand, and *#h1L-, on the other, also gave different reflexes in Proto-Tocharian (Hackstein 1998).

[8] All of the attested reflexes have the geminates, cf., e.g., the reflexes of PCelt. *koslo-, *kasninā, *kīsrā, *alyo-, and cf. 1 sg.  pres. of the copula, OIr. am, Gaul. immi. It is quite clear that Proto-Celtic was in the process of developing geminates, because some of the assimilations that produced geminates are certainly early (see above). The reconstructions in this dictionary contain geminated stops, fricatives, and resonants.

[9] This branch of Celtic languages is referred to as "British", "Brythonic", and "Brittonic" in the literature; I tried to be consistent in using "Brittonic".

[10] Cornish is often divided into Middle Cornish (from 12th until the end of the 16th centuries) and Modern Cornish (17th - 18th centuries), cf., e.g., Lewis 1990. I have grouped the words attested from the beginning of the 13th century onwards under a single label, "Cornish".

APPENDIX


The non-Indo-European elements in the Celtic lexicon

Although standards for mentioning possible Indo-European etymologies are rather liberal in this dictionary, there is still a large number of words in the reconstructed Proto-Celtic lexicon that cannot be attributed to any PIE root, and that are, therefore, quite likely to have been borrowed from some non-IE source. In some cases, there are a number of probable cognates in the neighbouring IE dialects (usually Italic and Germanic), but the reconstructed shape of the root distinctly shows non-IE features, which again makes it probable that the Celtic etymon in question was borrowed from some non-IE substratum language, perhaps shared with Italic and/or Germanic.

A number of such words, for which a substratum origin can be assumed, have reflexes only in Brittonic and Goidelic. This can, of course, be the consequence of the poor attestation of Gaulish, Lepontic, and Celtiberian, but in principle we cannot exclude the possibility of substrates shared by Insular Celtic languages, but not by the Continental Celtic.

The following is an alphabetical list of Proto-Celtic forms for which a substrate origin can be assumed; in each case it is indicated whether the etymon in question is attested in Continental Celtic, and whether it has likely cognates in the neighbouring IE dialects (Germanic and Italic).

 

1.      *alten- ‘razor’

2.      *amaro- ‘wailing, crying’

3.      *anderā ‘young woman’ (probably attested in Gaulish)

4.      *bando- ‘peak, top’ (attested in Gaulish; possible cognates in Germanic)

5.      *banwo- ‘young pig, piglet’ (attested in Gaulish)

6.      *baski- ‘bundle’ (probable cognates in Italic)

7.      *birro- ‘short’ (attested in Gaulish)

8.      *blVdV- ‘wolf, large predator’

9.      *bodyo- ‘yellow’ (attested in Gaulish, probable cognates in Italic)

10.  *brano- ‘raven’ (attested in Gaulish)

11.  *bratto-, *brattino- ‘mantle, cloak’

12.  *brokko- ‘badger’ (attested in Gaulish)

13.  *bunno- ‘awl, bittern’

14.  *bussu- ‘lip’ (probably attested in Gaulish)

15.  *butā ‘house, dwelling, hut’

16.  *druko- ‘bad’

17.  *durno- ‘fist’

18.  *esok- ‘salmon’ (attested in Gaulish)

19.  *gweno- ‘smile’

20.  *gulbV-, *gulbīno- ‘beak’ (probably attested in Gaulish)

21.  *gurmo- ‘dun, dark’

22.  *kag-o- ‘get, receive’ (probable cognates in Italic)

23.  *kagyo- ‘pen, enclosure’ (possible cognates in Germanic)

24.  *kalmiyo- ‘skilful, skilled’

25.  *kani- ‘good, nice’

26.  *karbanto- ‘war chariot’ (attested in Gaulish)

27.  *kasninā ‘garlic, leek’ (probably attested in Gaulish)

28.  *katrik- ‘fortification’ (probable cognates in Germanic)

29.  *kayto- ‘wood’ (cognates in Germanic)

30.  *klamo- ‘sick, suffering from leprosy’

31.  *klukā ‘stone, rock’

32.  *knū ‘nut’ (probable cognates in Italic and Celtic)

33.  *koligno- ‘pup, small animal’

34.  *koret- ‘palisade, stone wall’ (possible cognates in Germanic)

35.  *korkkyo- ‘oats’ (probable cognates in Germanic)

36.  *kotto- ‘old’ (probably attested in Gaulish)

37.  *krittā ‘body, frame, shape’

38.  *krok(ke)no- ‘skin’ (probably attested in Gaulish)

39.  *krumbo- ‘round, curved’ (probable cognates in Germanic)

40.  *krundi- ‘round, compact’

41.  *krutto- ‘round object, womb’

42.  *kwezdi- ‘piece, portion’ (attested in Gaulish)

43.  *lēro- ‘diligent’

44.  *liro- ‘sea, ocean’

45.  *lomanā ‘rope, thong’

46.  *lubī/ā ‘herb, plant’ (probable cognates in Germanic)

47.  *lukot- ‘mouse’

48.  *luxtu- ‘content, crowd’

49.  *makinā ‘bellow’ (probable cognates in Germanic and Baltic)

50.  *maylo- ‘bald’ (possible cognates in Germanic)

51.  *mazdyo- ‘stick’ (cognates in Italic and Germanic)

52.  *mesal-kā ‘blackbird’ (cognates in Italic and Germanic)

53.  *menādo- ‘awl’

54.  *metto- ‘decay, blight, shame’

55.  *mokku- ‘pig’

56.  *molto- ‘ram, wether’ (probably attested in Gaulish)

57.  *ninati- ‘nettle’ (probable cognates in Germanic and Baltic)

58.  *nino- ‘ash-tree’ (possibly attested in Gaulish)

59.  *nūsso-, *nowsso- ‘first milk, colostrum’

60.  *rem(r)o- ‘fat, thick’

61.  *rendi- ‘point, peak’

62.  *rowk(k)- / *ruk- ‘tunic, mantle’ (cognates in Germanic and Slavic)

63.  *rūnā ‘secret’ (possible cognates in Germanic)

64.  *sēbro- ‘demon, spectre’

65.  *sfrawo- ‘crow’ (possible cognates in Germanic, Baltic, and Italic)

66.  *sido- ‘elk, stag’

67.  *skamo- ‘light’ (possible cognates in Germanic)

68.  *skublo- ‘bird of prey’ (probably attested in Gaulish)

69.  *slad-yo- ‘hit, slay’

70.  *slattā ‘stalk, staff’ (possible cognates in Germanic)

71.  *swanto- ‘treasure, what is desired’

72.  *smēro- ‘berry’

73.  *subi- ‘strawberry’

74.  *sukko- ‘pig’

75.  *tago- ‘strangle, choke’

76.  *torrV- ‘belly’

77.  *trussko- ‘dirty, leprous’

78.  *trusto- ‘noise, cry’

79.  *wesakko-, *wesākko- ‘grebe, raven’

80.  *wēt(t)ā ‘stream, swamp’

81.  *wimonā ‘sea weed’

82.  *wriggant- ‘vermin’ (possibly attested in Gaulish)

83.  *wroyko- ‘heather’ (possible cognates in Balto-Slavic)

84.  *yoyni- ‘rushes, reed’ (probable cognates in Italic and Germanic)

85.  *yutV- ‘pap, porridge’ (possibly attested in Gaulish)

 

The number of substrate words in Proto-Celtic is actually surprisingly low. Only 85 out of the total 1490 Proto-Celtic words can be ascribed to a non-IE substrate, which is under 6%. This number is probably slightly higher, since several of the IE etymologies proposed in the dictionary might turn out to be false, but even so, it probably does not exceed 10%. Since many of the nouns listed above have probable cognates in other Western Indo-European languages (primarily Italic and Germanic), we might argue that there was no pre-IE substrate exclusive to Celtic, i. e., there was no common substrate in Western Europe from which Celtic, and only Celtic, borrowed words. There are, of course, many words in Welsh and Irish with obscure, presumably non-IE etymology, but it is rather surprising how few of those words go back to Proto-Celtic, or Proto-Insular Celtic (if one believes in that). Again, this may point to the conclusion that there was no single substratum language (or a group of closely related languages) prior to the arrival of the Celts in the British Isles. Judging by the amount of language diversity before the Roman conquests in other parts of Europe, for which we have more data (e.g. for Spain or Italy), this is not so surprising.

It is not surprising that most of the non-IE words in Celtic are nouns, since nouns are much more often borrowed than verbs, or words belonging to other word classes. It is also understandable that nouns of substrate origin often denote birds, plants, and small animals.

On the formal side, one should note that substratum words in Celtic often have geminates and the vowel *a in the root. Both of these features have been recognized as characteristic of substratum words in other European language groups, especially in Germanic. What is more surprising is the fact that words of non-IE origin in Celtic have the vowel *u in the root much more often than could be attributed to chance (24 out of 85 words, or more than a quarter of the total). Moreover, the donor language(s) seem not to have had a length contrast in their vowel systems. The only long vowels that appear in the roots of non-IE loanwords in Celtic are *ē (which can be from the diphthong *ey) and *ū (in two instances, once alternating with *ow). Finally, labiovelars are extremely rare (they occur in only two words), which probably means that the donor language(s) lacked them. The significance of these findings is yet to be evaluated in the realm of the Celtic linguistics.

References

Adams D. Q. AdamsA dictionary of Tocharian B, Rodopi, Amsterdam and Atlanta 1999.
Adams 1985D. Q. Adams“The Indo-European word for ‘apple’ again”, Indogermanische Forschungen 90/1985: 79-82.
Ahlqvist 1980A. AhlqvistVaria 3. “Old Irish bé, ben ‘woman’”, Ériu 31/1980: 156-163.
André 1985J. André“Noms de plantes gaulois ou prétendus gaulois dans les textes grecs et latins”, Études Celtiques 22/1985: 179-198.
Bammesberger 1974-5A. Bammesberger“La formation de vieil-irlandais céssaid”, Études Celtiques 14/1974-5: 205-206.
Bammesberger 1985A. Bammesberger“Vieil-irlandais téit ‘il va’”, Études Celtiques 22/1985: 202-203.
Bammesberger 1989A. Bammesberger“L’origine de vieil-irlandais bés”, Études Celtiques 26/1989: 69-71.
Bammesberger 1996A. Bammesberger“L’étymologie du vieil-irlandais cerd”, Études Celtiques 32/1996: 139-141.
Bammesberger 1998A. Bammesberger“The etymology of Irish áit ‘place’”, Ériu 49/1998: 41-43.
Beekes 1982R. S. P. Beekes“The PIE Word for ‘Moon, Month’ and the perfecet participle”, JIES 10/1982: 53-64.
Beekes 1987R. S. P. Beekes“The PIE words for ‘name’ and ‘me’”, Die Sprache 33/1987: 1-6.
Beekes 1988R. S. P. Beekes“PIE RHC- in Greek and other languages”, Indogermanische Forschungen 93/1988: 22-45.
Beekes 1992R. S. P. Beekes“‘Widow’”, KZ 105/1992: 171-188.
Beekes 1994R. S. P. Beekes“’Right’, ‘left’, and ‘naked’ in Proto-Indo-European”, Orbis 37: 87-96.
Beekes 1997R. S. P. Beekes“Ancient European loanwords”, KZ 105/1997: 215-236.
Benveniste 1963E. Benveniste“Le nom celtique du fer”, Celtica 3/1956: 279-283.
Benveniste 1969E. BenvenisteLe vocabulaire des institutions indo-européennes, Paris 1969.
Billy 1993P.-H. BillyThesaurus Linguae Gallicae, Hildesheim etc. 1993.
Billy 1995P.-H. BillyAtlas Linguae Gallicae, Hildesheim etc. 1995.
Birkhan 1970H. BirkhanGermanen und Kelten bis zum Ausgang der Römerzeit, Vienna 1970.
Bjorvand & LindemanH. Bjorvand & F. O. LindemanVåre arveord. Etymologisk ordbok Novus forlag, Oslo 2000.
Blažek 2006V. Blažek“Celtic ‘smith’ and his colleagues”, in: Evidence and Counter-Evidence, Festschrift for F. Kortlandt, Rodopi, Amsterdam & Atlanta, I: 35-53.
Bolelli 1941-2T. Bolelli“Le voci di origine gallica del REW di W. Meyer-Lübke, L’Italia dialettale 17/1941: 133-194 and 18/1942: 33-74.
Boutkan 2005D. Boutkan and S. M. SiebingaOld Frisian etymological dictionary, Brill, Leiden 2005.
Buck C. D. BuckA dictionary of selected synonyms in the principal Indo-European Languages, The University of Chicago Press, Chicago 1949.
Campanile 1974E. CampanileProfilo etimologico del cornico antico, Pacini, Pisa 1974.
Campanile 1974aE. Campanile“Un arcaismo morfologico del celtico”, Incontri Linguistici 1/1974: 51-54.
Campanile 1982E. Campanile“Note sur vieil-irlandais bronnaid”, Études Celtiques 19/1982: 151-154.
Campanile 1983E. Campanile“I Galli nella Cisalpina e i dialetti galloitalici”, in: E. Campanile (ed.) Problemi di sostrato nelle lingue indoeuropee, Giardini, Pisa 1983: 27-35.
Campanile 1989E. Campanile“Zur Etymologie von altir. arae und eirr”, ZCP 43: 174-178.
CGL Corpus Glosariorum Latinorum (ed. Löwe and Götz)Teubner, Leipzig 1923.
Charles-Edwards 1972T. M. Charles-Edwards“Some Celtic kinship terms”, Bulletin of the Board of Celtic Studies 24: 105-122.
CIIC Corpus inscriptionum insularum Celticarumed. by R. A. S. Macalister, Dublin 1945-1949.
CIL Corpus Inscriptionum Latinarum. Begründet von Th. Mommsen 1853: Akademie der Wissenschaften, Berlin.
Corominas 1976J. Corominas“Elementos prelatinos en las lenguas romances hispánicas”, in: Jordá, F. et alii (eds.) Actas del I coloquio sobre lenguas y culturas prerromanas de la península ibérica, Ediciones Universidad de Salamanca, Salamanca: 87-164.
Cowgill 1967W. Cowgill“On the fate of *w in Old Irish”, Language 43/1967: 129-138.
Cowgill 1970W. Cowgill“Italic and Celtic superlatives and the dialects of Indo-European”, in: G. Cardona (ed.), Indo-European and Indo-Europeans, University of Pennsylvania Press, Philadelphia, 1970.
Cowgill 1980W. Cowgill“The etymology of Irish guidid and the outcome of *gwh in Celtic”, Lautgeschichte und Etymologie. Akten der VI. Fachtagung der Indogermanischen Gesellschaft, ed. by M. Peters and O. E. Pfeiffer, Wiesbaden 1980: 49-78.
Cowgill 1987W. Cowgill“Introduction”, in: Indogermanische Grammatik, 1, 1. Halbband, Winter, Heidelberg 1987.
de Bernardo Stempel 1987P. de Bernardo StempelDie Vertretung der indogermanischen liquiden und nasalen Sonanten im Keltischen, IBS, Innsbruck 1987.
de Bernardo Stempel 1997P. de Bernardo Stempel“Spuren gemeinkeltischer Kultur im Wortschatz: 1. ‘Tochter’; 2. bri(u)gu; 3. Banba; 4. Vercelli; 5. Banassac; 6. Plumergat”, ZCP 49-50/1997: 92-106.
de Bernardo Stempel 1999P. de Bernardo StempelNominale Wortbildung des älteren Irischen. Stammbildung und Derivation, Tübingen 1999.
de Bernardo Stempel 2000P. de Bernardo Stempel“Ptolemy’s Celtic Italy and Ireland: a linguistic analysis”, in: D. N. Parsons & P. Sims-Williams (eds.), Ptolemy: Towards a linguistic atlas of the earliest Celtic place-names of Europe, Aberystwyth 2000: 83-112.
de Bernardo Stempel 2006P. de Bernardo Stempel“Indogermanisch und keltisch ‘geben’, kontinentalkelt. Gabiae, gabi/gabas, keltib. Gabizeti, altir. Ro-(n)-gab und Zugehoeriges”, KZ 118/2006: 185-200
de Vaan 2002M. de Vaan“The etymology of English to brag and Old Icelandic bragr, in: North-West European Language Evolution 41/2002: 45-58.
de Vaan 2004M. de Vaan“’Narten’ roots from the Avestan point of view”, in: A. Hyllested et alii (eds.) Per Aspera ad Asteriscos. Studia indogermanica in honorem Jens Elmegård Rasmussen sexagenarii Idibus Martiis anno MMIV, IBS, Innsbruck 2004: 591-599.
de Vaan 2008M. de VaanEtymological dictionary of Latin and other Italic languages, Brill, Leiden 2008.
Degavre 1998J. DegavreLexique gaulois, Mémoires de la Société belge d’études celtiques 9, Bruxelles 1998.
Demiraj 1997B. DemirajAlbanische Etymologien, Rodopi, Amsterdam & Atlanta 1997.
Derksen 2008R. DerksenEtymological Dictionary of the Slavic Inherited Lexicon, Brill, Leiden 2008.
Deshayes 2003A. DeshayesDictionnaire étymologique du breton, Douarnenez 2003.
DGVB L. FleuriotDictionnaire des glosses en vieux breton, Klincksieck, Paris 1964.
DIL Dictionary of the Irish LanguageR.I.A., Dublin.
Dottin 1920G. DottinLa langue gauloise, Paris 1920.
Dybo 1961V. A. Dybo“Sokraščenie dolgot v kel’to-italijskix jazykax i ego značenie dlja balto-slavjanskoj i indoevropejskoj akcentologii”, Voprosy slavjanskogo jazykoznanija 5/1961: 9-34.
EIEC J. Mallory & D. Q. Adams (eds.)Encyclopaedia of Indo-European CultureFitzroy Dearborn, Chicago 1997.
Ellis Evans 1967D. Ellis EvansGaulish personal names, Clarendon, Oxford 1967.
Eska 1989J. EskaTowards an interpretation of the Hispano-Celtic inscription of Botorrita, IBS, Innsbruck 1989.
Eska 1990J. Eska“The deictic pronominal *ḱey in Celtic”, Celtica 21/1990: 153-155.
Eska 1994/1995J. Eska“Another look at Lepontic uenia”, Beiträge zur Namenforschung 29-30/1994-1995: 129-134.
Evans 1964D. Simon Evans A grammar of Middle Welsh DIAS, Dublin 1964.
Falileyev A. FalileyevEtymological glossary of Old Welsh, Max Niemeyer, Tübingen 2000.
Falileyev 1997A. Falileyev“Cambro-slavica”, ZCP 49-50/1997: 198-203.
Falileyev and Isaac 2003A. Falileyev and G. Isaac“Leeks and garlic: the Germanic ethnonym Cannenefates, Celtic *kasn-, and Slavic *kesn-”, in: North-West European Language Evolution 42: 3-12.
FEW W. von WartburgFranzözisches etymologisches Wörterbuch, Sbinden, Basel 1928-1966.
Fleuriot 1976-7L. Fleuriot“Le vocabulaire de l’inscription Gauloise de Chamalières”, Études Celtiques 16/1976-7: 173-190.
Fleuriot 1991L. Fleuriot“Celtoromanica in the light of the newly discovered Celtic inscriptions”, ZCP 44/1991: 1-35.
Fowkes 1945R. A. Fowkes“Some Welsh notes”, Language 21, 2/1945: 96-97.
Freeman 2001P. FreemanA comprehensive survey of the language of the ancient Celts in Greco-Roman Asia Minor, The Edwin Mellen Press, Lewiston 2001.
Frisk H. FriskGriechisches etymologisches Wörterbuch, Winter, Heidelberg 1967.
GamillschegE. GamillschegEtymologisches Wörterbuch der französischen Sprache, Winter, Heidelberg 1969.
GOI R. ThurneysenA grammar of Old Irish, Dublin 1946.
GPC Geiriadur Prifysgol CymruAberystwyth.
Greene 1958D. Greene“Ir. túaimm, stuaim, W. ystum, ywstwyth”, Celtica 4/1958: 44.
Greene 1963D. Greene“Ir. gnás: W gnaws: OHG kunst”, Celtica 6/1963: 62-64.
Greene 1971D. Greene“Varia II”, Ériu 22/1971: 177-180.
Greene 1973D. Greene“The growth of palatalization in Irish”, Transactions of the Philological Society 1973: 127-136.
Greene 1976D. Greene“The diphthongs of Old Irish”, Ériu 27/1976: 26-45.
Greene 1983D. Greene“Cró, crú and similar words”, Celtica 15/1983, 1-9.
Guyonvarc’h 1973-C.-J. Guyonvarc’hDictionnaire étymologique du breton ancien, moyen et moderne, Rennes 1973-
Haarmann 1970H. HaarmannDer lateinische Lehnwortschatz im Kymrischen, Romanisches Seminar der Universität Bonn, Bonn 1970.
Hackstein 1998O. Hackstein“Tocharisch und Westindogermanisch: Strukturell uneinheitliche Laryngalreflexe im Tocharischen (Uridg. *-Uh1C- vs. *-Uh2/3(C)-und *#h1RC- vs. *#h2/3RC-)”, in: W. Meid (ed.), Sprache und Kultur der Indogermanen, IBS, Innsbruck 198: 217-236.
Hamp 1965E. P. Hamp“Evidence in Keltic”, in: W. Winter (ed.), Evidence for Laryngeals, The Hague 1965: 224-235.
Hamp 1965aE. P. Hamp“Old Irish scál, Gothic skohsl”, Celtica 6/1965: 118-119.
Hamp 1971E. P. Hamp“Varia III, 1. The Keltic words for ‘tear’, 2. The ‘bee’ in Irish, Indo-European, and Uralic”, Ériu 22/1971: 181-187.
Hamp 1973E. P. Hamp“Celtic and Indo-European words in *MVL-i, Celtica 10/1973: 151-156.
Hamp 1973aE. P. Hamp“Formations indo-européennes à second élément *(Ho)kw-”, Bulletin de la Société de Linguistique de Paris 68/1973: 77-92.
Hamp 1974E. P. HampVaria, Ériu 25/1974: 253-284.
Hamp 1974-5E. P. Hamp“Bret. GWAZH, GOAH, GOÉH, OIr. féith”, Études Celtiques 14/1974-5: 201-205.
Hamp 1974-5aE. P. Hamp“Varia Etymologica”, Études Celtiques 14/1974-5: 461-477.
Hamp 1975E. P. Hamp“Varia II”, Ériu 26/1975: 168-174.
Hamp 1975aE. P. Hamp“Old Irish ed, id”, ZCP 34/1975: 20-29.
Hamp 1976E. P. Hamp“*gweiHo- ‘live’”, in: A. Morpurgo-Davies and W. Meid (eds.), Festschrift Leonard Palmer, IBS, Innsbruck 1976: 87-91.
Hamp 1978E. P. Hamp“Celtic *dām- and vṛddhi and dāmos”, ZCP 36/1978: 5-12.
Hamp 1978aE. P. Hamp“MW ludd”, Münchener Studien zur Sprachwissenschaft 37/1978: 65-58.
Hamp 1979E. P. Hamp“Indo-European *gwen-Ha”, KZ 93: 1-7.
Hamp 1980E. P. Hamp“IE *kweh2s- ‘cough’”, in: K. Čolakova et alii (eds.), Studia Linguistica in Honorem Vladimir I. Georgiev, Izd. na BAN, Sofia 1980: 130-134.
Hamp 1980aE. P. Hamp“Notulae etymologicae cymricae”, Bulletin of the Board of Celtic Studies 28/1978-80: 213-217.
Hamp 1981E. P. Hamp“Varia III”, Ériu 32: 158-162.
Hamp 1982E. P. Hamp“The Indo-European root *bher in the light of Celtic and Albanian”, ZCP 39/1982: 205-218.
Hamp 1982aE. P. Hamp“Varia”, Études Celtiques 19/1982: 137-142.
Hamp 1982bE. P. Hamp“*-og- in Celtic and notes on BRO”, Études Celtiques 19/1982: 143-149.
Hamp 1983E. P. Hampimbúarach, imbárachi”, Celtica 15/1983: 53-54.
Hamp 1984E. P. Hamp“Varia III”, Ériu 35/1984: 200-203.
Hamp 1984aE. P. Hamp“Varia XV-XVIII”, Études Celtiques 21/1984: 137-140.
Hamp 1984bE. P. Hamp“Varia I. 1. Irish fert, fertae; 2. Irish serb=Welsh herw; 3. Spanish alamo”, ZCP 40/1984: 275- 279.
Hamp 1985E. P. Hamp“Varia IV”, Ériu 36: 181-183.
Hamp 1985aE. P. Hamp“Indo-European ‘Gerste, orge, barley’”, KZ 98/1985: 11-13.
Hamp 1986E. P. Hamp“Varia”, Études Celtiques 23/1986: 47-51.
Hamp 1987E. P. Hamp“*krei- ‘sift’ and *kwreiHa- ‘buy’ and nominalization in Celtic”, Bulletin of the Board of Celtic Studies 34/1987: 117-118.
Hamp 1989E. P. Hamp“Celtic *orko-, *erko-, IE *porko-”, ZCP 43/1989: 194-195
Hamp 1990E. P. Hamp“Two etymologies”, Celtica 21/1990: 173-177.
Hamp 1990aE. P. Hamp“Varia”, Études Celtiques 27/1990: 179-189.
Hamp 1991E. P. Hamp“Varia II: 4. uisce again”, Ériu 42/1991: 193.
Hamp 1992E. P. Hamp“Welsh elfydd and albio-”, ZCP 45/1992: 87-89, ZCP 45/1992.
Hamp 1995E. P. Hamp“Old Irish ARBAR "corn"”, Études Celtiques 31/1995: 89-90.
Hamp 1996E. P. Hamp“OIr. GOR "pious, dutiful": meaning and etymology”, Ériu 47/1996: 193-204.
Hamp 1996aE. P. Hamp“VARIA II”, Ériu 47/1996: 209-211.
Hamp 1997E. P. Hamp“Varia”, Études Celtiques 33/1997: 81-83.
Hamp 1998E. P. Hamp“Two regular milk products”, in: J. Jasanoff et alii (eds.) Mír Curad. Studies in Honor of Calvert Watkins, IBS, Innsbruck 1998: 241-243.
Hamp 2002E. P. Hamp“IE spleen”, JIES 30/2002: 145-150.
Hamp 2003E. P. Hamp“Gaulish ci, -c, Old Irish , Ogam koi”, Celtica 24/2003: 129.
Hemon R. HemonGeriadur Istoriel ar Brezhoneg, Preder, Plomelin 1979-
Hilmarsson 1985J. Hilmarsson“Toch. A kāc, Lat. cutis, OIcel. húð < I.-E. *kuHtís ‘skin’”, KZ 98/1985: 162-163.
Holder 1896-1913A. HolderAlt-celtischer Sprachschatz, I-III, Leipzig 1896-1913.
Huld 2006M. E. Huld“Indo-European ‘hawthorns’, in: K. Jones-Bley et alii (eds.), Proceedings of the Seventeenth Annual UCLA Indo-European Conference, JIES Monograph Series, Washington, DC 2006: 165-177.
Hull 1960/61V. Hull“Old Irish sissidir and its compounds”, ZCP 28/1960-61: 260-265.
IEW J. PokornyIndogermanisches etymologisches Wörterbuch, Francke, Bern 1959.
Irslinger 2002B. IrslingerAbstrakta mit Dentalsuffixen im Altirischen, Winter, Heidelberg 2002.
Isaac 1996G. IsaacThe verb in the Book of Aneirin. Studies in syntax, morphology and etymology, Niemeyer, Tübingen 1996.
Isaac 2002G. Isaac“W byw, byd, hyd”, Studia Celtica 36/2002: 145-147.
Isaac 2004G. Isaac“W sudd ‘Juice’, hufen ‘Cream’”, Studia Celtica 38/3004: 179-180.
Isaac 2007G. IsaacStudies in Celtic sound changes and their chronology, IBS, Innsbruck 2007.
Jackson 1967K. H. JacksonA historical phonology of Breton, DIAS, Dublin 1967.
Jasanoff 1989J. Jasanoff“Old Irish ‘woman’”, Ériu 40: 135-141.
JIES The Journal of Indo-European Studies.
Jordán Cólera 1998C. Jordán CóleraIntrodución al Celtibérico, Monografías de filología griega, Zaragoza 1998.
Joseph 1982L. Joseph“The treatment of *CRH- and the origin of CaRa- in Celtic”, Ériu 33: 32-57.
Joseph 1986L. Joseph“A survival from the Italo-Celtic legal vocabulary”, Ériu 37: 119-125.
Kalygin 2006V. P. KalyginÈtimologičeskij slovar’ kel’tskix teonimov, Nauka, Moscow 2006.
Katz 1998J. T. Katz“Hittite tašku- and the Indo-European word for 'badger'” KZ 111/1998: 61-82.
Kavanagh / Wodtko 2001S. Kavanagh & D. WodtkoA lexicon of the Old Irish glosses in the Würzburg manuscript of the Epistles of Saint Paul, ÖAW, Vienna 2001.
Kloekhorst 2008A.KloekhorstEtymological dictionary of the Hittite inherited lexicon, Brill, Leiden 2008.
Kluge F. KlugeDeutsches etymologisches Wörterbuch, edited by E. Seebold, de Gruyter, Berlin and New York 221989.
Koch 1992J. Koch“‘Gallo-Brittonic’ vs. ‘Insular Celtic’: The inter-relationships of the Celtic languages reconsidered”, Bretagne et pays celtiques - langue, histoire, civilisation. Mélanges offerts a la memoire de Léon Fleuriot, ed. G. Le Menn, Saint-Brieuc/Rennes, 1992: 471-495.
Kortlandt 1981F. Kortlandt“More evidence for Italo-Celtic”, Ériu 32/1981: 1-23.
Kortlandt 1997F. Kortlandt“On the relative chronology of Celtic sound changes”, KZ 110/1997: 248-251.
Kortlandt 2000F. Kortlandt“Three notes on the Old Irish verb”, Études Celtiques 34/1998-2000: 143-146.
Kortlandt 2007F. KortlandtItalo-Celtic origins and prehistoric development of the Irish language, Rodopi, Amsterdam 2007.
Kroonen 2006G. Kroonen“Gemination and Allomorphy in the Protogermanic mn-Stems: bottom and rime”, Amsterdamer Beiträge zur älteren Germanistik 61/2006: 17-25.
KZKuhn’s Zeitschrift für vergleichende Sprachwissenschaft (since the issue no. 100 appears as Historische Sprachforschung).
Lambert 1979P.-Y. Lambert “Gaulois IEVRV: irlandais (ro)-ír ‘discavit’”, ZCP 37/1079: 207-213.
Lambert 1990P.-Y. Lambert“Notes léxicographiques”, Études Celtiques 27/1990: 197-202.
Lambert 1994P.-Y. LambertLa langue gauloise, Paris 1994.
Lambert 1994aP.-Y. Lambert“Gloses en vieux-breton”, Études Celtiques 30/1994: 221-228.
Lambert 1995P.-Y. Lambert“Préverbes gaulois suffixés en -io-: ambio-, ario-, cantio-“, Études Celtiques 31/1995: 115-121.
Lambert 1997P.-Y. Lambert“Gaulois tardif et latin vulgaire”, ZCP 49-50/1997: 3966-413.
Lambert 2000P.-Y. Lambert“Remarks on Gaulish Place-Names in Ptolemy”, in: D. N. Parsons & P. Sims-Williams (eds.), Ptolemy: Towards a linguistic atlas of the earliest Celtic place-names of Europe, Aberystwyth 2000: 159-168.
Lane 1931G. S. Lane“Celtic Notes”, Language 7/1931: 278-283.
Lehmann W. LehmannA Gothic etymological dictionary, Brill, Leiden 1986.
LEIA J. VendryèsLexique étymologique de l’irlandais ancien, Paris 1959-
Lejeune 1971M. LejeuneLepontica, Paris 1971.
Lewis 1989H. LewisDie kymrische Sprache, IBS, Innsbruck 1989.
Lewis 1990H. LewisHandbuch des Mittelkornischen, IBS, Innsbruck 1990.
Lewis and ShortA Latin Dictionary revised, enlarged and in great part rewritten by C. T. Lewis, Oxford 1980.
LHEB K. JacksonLanguage and History in Early Britain, Edinburgh 1953.
Lindeman 1999F. O. Lindeman“Varia V-VI”, Ériu 50/1999: 179ff.
Lindeman 1999aF. O. Lindeman“Old Irish inne”, Celtica 23/1999: 155-156.
LIV Lexikon der indogermanischen Verben, ed. by H. Rix et alii, Reichert, Wiesbaden 1998 (22000).
Lockwood 1987W. B. Lockwood“Wortkundliche Parerga: 1. Ir. cunуg "Sturmtaucher, Puffinus"; 2. Kymr. mwyalch, gl. lon "Amsel"; 3. Die keltischen Namen der Ente; 4. Ir. lacha; 5. Ir. tonnуg; 6. Kymr. gwydd, ir. gй(d) "Gans"; 7. Ir. úan "Junges"”, ZCP 38/1981: 179-186.
Lockwood 2005W. B. Lockwood“Old Irish fann: a possible etymology”, Studia Celtica 39/2005: 199-200.
LPH. Lewis & H. PedersenA concise comparative Celtic grammar, Vandenhoeck and Ruprecht, Göttingen 1961.
Lubotsky 1989A. Lubotsky“Against a PIE phoneme *a”, in: New Sound of Proto-Indo-European, ed. by T. Vennemann, de Gruyter, Berlin & New York 1989: 53-66.
Lubotsky 1994A. LubotskyRV. ávidhat. In: Früh-, Mittel- und spätindogermanisch: Akten der IX. Fachtagung der Indogermanischen Gesellschaft vom 5. bis 9. Oktober 1992 in Zürich, eds. G. Dunkel, G. Meyer, S. Scarlata and C. Seidl, Reichert, Wiesbaden 1994: 201-206.
Lubotsky 1994aA. Lubotsky“Av. θβōrəštar and the Indo-European √turḱ”, Die Sprache 36/1994: 94-102.
Lubotsky 2000A. Lubotsky“The Vedic root vṛ- and its present”, in: Indoarisch, Iranisch, und die Indogermanistik, ed. by B. Forssman and R. Plath, Reichert, Wiesbaden 2000: 315-325.
Lubotsky 2006A. Lubotsky“Indo-European ‘heel’”, in: R. Bombi et alii (eds.) Studi linguistici in onore di Roberto Gusmani, Alessandria 2006: 1005-1010.
Luján 2003E. R. Luján“Gaulish personal names: an update”, Études celtiques 35/2003:181-241.
Mac BainA. Mac BainAn etymological dictionary of the Gaelic language, Edinburgh 1911.
Mac Cana 1991P. Mac Cana“Irish maccóem, Welsh makwyf”, Ériu 42: 1991: 27-36.
Mac Éoin 1974G. S. Mac Éoin“The etymology of Ir. coi "cuckoo"”, ZCP 33/1974: 66.
Mallory & Adams 2006J. Mallory & D. Q. AdamsThe Oxford Introduction to Proto-Indo-European and the Proto-Indo-European World, Oxford University Press, Oxford 2006.
Markey 1988T. L. Markey“Eurasian ‘apple’ as arboreal unit and item of culture”, JIES 16/1988: 49-68.
Markey 2003T. L. Markey“Anextlomārus revisited”, KZ 116/2003: 295-301.
Matasović in pressR. Matasović“Sun and Moon in Celtic and Proto-Indo-European”, paper from the Celto-Slavica II conference in Moscow, September 2006. To appear in the proceedings.
Matasović 1995R. Matasović“A re-examination of Winter’s law in Baltic and Slavic”, Lingua Posnaniensis 37/1995: 57-70.
Matasović 1996R. MatasovićA Theory of Textual Reconstruction in Indo-European, Frankfurt a/M 1996.
Matasović 2000R. Matasović“The Sanskrit and PIE words for ‘tear’”, in: Z. Matišić (ed.) Trava od srca (Hrvatske Indije 2), Sekcija za orijentalistiku HFD i Filozofski fakultet, Zagreb 2000: 77-88.
Matasović 2004R. MatasovićGender in Indo-European, Winter, Heidelberg 2004.
Matasović 2008R. Matasović“Insular Celtic as a Language Area”, in: H. L. C. Tristram (ed.) The Celtic Languages in Contact, Potsdam University Press, Potsdam 2008: 93-112.
Matasović 2008aR. MatasovićPoredbenopovijesna gramatika hrvatskoga jezika [A Comparative and Historical Grammar of Croatian], Matica hrvatska, Zagreb 2008.
Mayrhofer M. MayrhoferEtymologisches Wörterbuch des Altindoarischen, Winter, Heidelberg 1992.
McCone 1985K. McCone“Varia I”, Ériu 36/1985: 176.
McCone 1986K. McCone“From Indo-European to Old Irish: Conservation and innovation in the verbal System”, in: Proceedings of the Seventh International Congress of Celtic Studies held at Oxford from 10th to 15th July, 1983, ed. by D. Ellis Evans et alii, Oxford 1986: 222-266.
McCone 1987K. McConeHund, “Wolf und Krieger bei den Indogermanen”, in: W. Meid, ed. Studien zum indogermanischen Wortschatz, IBS, Innsbruck 1987: 101-154.
McCone 1991K. McCone“The PIE stops and syllabic nasals in Celtic”, Studia Celtica Japonica 4, 37-69.
McCone 1991aK. McCone“OIr. -ic ‘reaches’, ithid ‘eats’, rigid ‘stretches, directs, rules’ and the PIE ‘Narten’ present in Celtic” Ériu 42: 1-11.
McCone 1992K. McCone“Varia I”, Ériu 43: 193-197.
McCone 1992aK. McCone“OIr. torc, Av. θβərəsō < PIE *tworḱos ‘(cutter), boar’, Münchener Studien zur Sprachwissenschaft 53/1992: 99-100.
McCone 1992bK. McCone“OIr. aub ‘river’ and amnair ‘maternal uncle’”, Münchener Studien zur Sprachwissenschaft 53/1992: 101-111.
McCone 1993K. McCone“Zisalpinish-Gallisch uenia und lokan”, in: F. Heidermanns et alii (eds.), Sprachen und Schriften des antiken Mittelmeerraums. Festschrift für Jürgen Untermann zum 65. Geburtstag, IBS, Innsbruck: 243-249.
McCone 1994K. McCone“An tSean-Ghaeilge agus a réamhstair”, in: Stair na Gaeilge, ed. by K. McCone, D. McManus et alii, Maynooth: 61-219.
McCone 1994aK. McCone“Zum Ablaut der keltischen r-Stämme”, in: J. E. Rasmussen (ed.), In Honorem Holger Pedersen, Reichert, Wiesbaden 1994: 275-284.
McCone 1995K. McCone“OIr. senchae, senchaid and preliminaries on agent noun formation in Celtic”, Ériu 46/1995: 1-10.
McCone 1996K. McConeTowards a relative chronology of ancient and medieval Celtic sound change, Department of Old and Middle Irish, NUI Maynooth, Maynooth 1996.
McCone 1996aK. McCone“From Indo-European to Old Irish: Conservation and Innovation in the Verbal System”, in: D. Ellis Evans et alii (eds.) Proceedings of the 7th International Congress of Celtic Studies, Oxford 1996: 222-266.
McCone 1998K. McCone“‘King’ and ‘Queen’ in Celtic and Indo-European”, Ériu 49/1998: 1-12.
McCone 1999K. McCone“OIr. erbaid ‘entrusts’, orb ‘heir’ and orbae ‘inheritance’”, in: Studia Celtica et Indogermanica. Festschrift für Wolfgang Meid zum 70. Geburtstag, ed. by P. Anreiter & E. Jerem, Archaeolingua, Budapest 1999: 239-242.
McCone 2005K. McConeA First Old Irish Grammar and Reader, NUI Maynooth, Maynooth 2005.
McCone 2006K. McCone“Greek Keltós and Galátēs, Latin Gallus ‘Gaul’”, Die Sprache 46/2006: 94-111.
McCone 2006aK.McConeThe origins and development of the Insular Celtic verbal complex, The Department of Irish, NUI Maynooth, Maynooth 2006.
McManus 1997D. McManusA guide to Ogam, An Sagart, Maynooth 1997.
Meid 1980W. MeidGallisch oder lateinisch? Soziolinguistische und andere Bemerkungen zu populären gallo-lateinischen Inschriften, IBS, Innsbruck 1980.
Meid 1991W. MeidAspekte der germanischen und keltischen Religion im Zeugnis der Sprache, Innsbruck 1991.
Meid 1993W. MeidDie erste Bottorita-Inschrift, IBS, Innsbruck 1993.
Meid 1994W. MeidGaulish inscriptions, Archaeolingua, Budapest 1994.
Meid 1994aW. MeidCeltiberian inscriptions, Archaeolingua, Budapest 1994.
Meid 1996W. MeidHeilpflanzen und Heilsprüche. Zeugnisse gallischer Sprache bei Marcellus von Bordeaux, IBS, Innsbruck 1996.
Meid 2005W. MeidKeltische Personennamen in Pannonien, Archaeolingua, Budapest 2005.
MLH IVJ. Untermann & D. S. WodtkoMonumenta Linguarum Hispanicarum IV. Die tartessischen, keltiberischen und lusitanischen Inschriften, Wiesbaden 1997.
MLH V.1D. S. WodtkoMonumenta Linguarum Hispanicarum V.1. Wörterbuch der keltiberischen Inschriften, Wiesbaden 2000.
Morris Jones 1913J. Morris JonesA Welsh grammar, historical and comparative, Oxford 1931.
Motta 1992F. Motta“Vues présentes sur le celtique cisalpin”, Études Celtiques 29/1992: 311-318.
Nussbaum 1986A. NussbaumHead and Horn in Indo-European, De Gruyter, Berlin and New York 1986.
Ó Donaill N. Ó DonaillFoclóir Gaeilge-Béarla, Oifig an tSoláthair, Baila Átha Cliath 1977.
Ó Flaithearta 1997M. Ó Flaithearta“Altirisch tess, echtar und die Frage der Konsonantengruppe -Xst- im Keltischen”, ZCP 49-50/1997: 653-663.
O’Brien 1056M. O’Brien“Etymologies and notes”, Celtica 3/1956: 168-184.
Oettinger 2003N. Oettinger“Neuerungen in Lexikon und Wortbildung des Nordwest-Indogermanischen”, in: A. Bammesberger & T. Vennemann (eds.), Languages in Prehistoric Europe, Winter, Heidelberg 2003: 183-193.
OLDOxford Latin Dictionary, Oxford 1980.
Orel 2003V. OrelA handbook of Germanic etymology, Brill, Leiden 2003.
Pedersen H. PedersenVergleichende Grammatik der keltischen Sprachen, Vandenhoeck & Ruprecht, Göttingen 1913.
Pennaod 1986J. Pennaod“La désignation de l’année en celtique”, Études Celtiques 23/1986: 53-56.
Picoche 1992J. PicocheDictionnaire étymologique du français, Le Robert, Paris 1992.
Pokorny J. PokornyIndogermanisches etymologisches Wörterbuch, Francke, Bern 1959.
Porzio GerniaM. L. Porzio Gernia“Gli elementi celtici del latino”, in: E. Campanile (ed.), I Celti d’Italia, Giardini, Pisa 1981: 97-122.
Quin 1960E. G. Quin“Old-Irish ol ‘inquit’”, Celtica 5/1960: 95-102.
Remmer 2002-3U. Remmer“Das idg. Suffix -mon- im Altirischen, Teil 1”, Die Sprache 43/2002-3: 171-211.
Remmer 2004U. Remmer“Das idg. Suffix -mon- im Altirischen”, Teil 2, Die Sprache 44.1/2004: 26-69.
Rieken 1999E. RiekenUntersuchungen zur nominalen Stammbildung des Hethitischen, Harrassowitz, Wiesbaden 1999.
RIGRecueil des inscriptions gauloises, Vol. I: Textes gallo-grecs, Vol. II.1: Textes galloétrusques; Textes gallo-latins sur pierre, Vol. II.2: Textes gallo-latins sur instrumentum, Paris 1988-2002.
Rix 1977H. Rix“Das keltische Verbalsystem auf dem Hintergrund des indo-iranischgriechischen Rekonstruktionsmodells”, in: K. H. Schmidt and R. Ködderitzsch (eds.) Indogermanisch und Keltisch. Kolloquium der Indogermanischen Gesellschaft am 16. und 17. Februar 1976 in Bonn, Wiesbaden 1977: 132-158.
Russell 1990P. RussellCeltic word-formation: the velar suffixes, DIAS, Dublin 1990.
Schindler 1969J. Schindler“Die idg. Wörter ‘Vogel’ und ‘Ei’”, Die Sprache 15/1969: 144-167.
Schindler 1975J. Schindler“Armenian erkn, Greek odýnē, Irish idu”, Indo-European Studies II, ed. by C. Watkins, Harvard University Press, Cambridge 1975: 252-274.
Schindler 1977J. Schindler“A thorny problem”, Die Sprache 23/1977: 25-35.
Schmidt 1957K. H. SchmidtDie Komposition in gallischen Personennamen, Max Niemeyer, Tübingen 1957.
Schmidt 1967K. H. Schmidt“Keltisches Wortgut im Lateinischen”, Glotta 44/1967: 151-174.
Schmidt 1980-82K. H. Schmidt“The Gaulish inscription of Chamalières”, Bulletin of the Board of Celtic Studies 29/1980-82: 256-268.
Schmidt 1987K. H. Schmidt“Handwerk und Handwerker im Keltischen und Germanischen”, in: W. Meid (ed.), Studien zum indogermanischen Wortschatz, IBS, Innsbruck 1987: 265-286.
Schmitt 1997Ch. Schmitt“Keltisches im heutigen Französisch”, ZCP 49-50/1997: 814-829.
Schrijver 1990P. Schrijver“Latin festīnāre, Welsh brys”, Münchener Studien zur Sprachwissenschaft 51/1990: 243-247.
Schrijver 1991P. SchrijverThe Development of the PIE Laryngeals in Latin, Rodopi, Amsterdam & Atlanta 1991.
Schrijver 1993P. Schrijver“Varia IV. OIr. dëec, dëac”, Ériu 44/1993: 181-184.
Schrijver 1994P. Schrijver“The Celtic adverbs for ‘against’ and ‘with’ and the early apocope of *-i”, Ériu 45/1994: 151-189.
Schrijver 1995P. SchrijverStudies in British Celtic historical phonology, Rodopi, Amsterdam & Atlanta 1995.
Schrijver 1996P. Schrijver“OIr. gor ‘pious, dutiful’: meaning and etymology”, Ériu 47/1996: 193-204.
Schrijver 1997P. SchrijverStudies in the history of Celtic pronouns and particles, Maynooth 1997.
Schrijver 1998P. Schrijver“The British word for ‘fox’ and its Indo-European origins”, JIES 26/1998: 421-434.
Schrijver 1999P Schrijver“Vowel rounding by Primitive Irish labiovelars”, Ériu 50/1999: 133-137.
Schrijver 2003P. Schrijver“The etymology of Welsh chwith and the semantics and etymology of PIE *k(w)sweibh-“, in: P. Russell (ed.), Yr Hen Iaith. Studies in early Welsh, Aberystwyth 2003: 1-23.
Schrijver 2004P. Schrijver“Indo-European *smer- in Greek and Celtic”, in: J. Penney (ed.) Indo-European perspectives: Studies in honour of Anna Morpurgo Davies, Oxford University Press, Oxford 2004: 292-299.
Schumacher 1995S. Schumacher“Old Irish *tucaid, tocad and Middle Welsh tynghaf tynghet re-examined”, Ériu 46/1995: 49-57.
Schumacher 2000S. SchumacherThe Historical Morphology of the Welsh Verbal Noun, Maynooth 2000.
Schumacher 2004S. SchumacherDie keltischen Primärverben, IBS, Innsbruck 2004
Sharpe 1979R. Sharpe“Hiberno-Latin laicus, Irish láech and the devil’s men”, Ériu 30/1979: 75-92.
Sims-Williams 1981P. Sims-Williams“The development of the Indo-European voiced labiovelars in Celtic”, Bulletin of the Board of Celtic Studies 29/1981: 201-229.
Sims-Williams 1990P. Sims-Williams“Dating the transition to Neo-Brittonic: Phonology and history, 400-600”, in: A. Bammesberger & A. Wollmann (eds.) Britain 400-600: Language and History, Heidelberg 1990: 217-261.
Sims-Williams 1999P. Sims-Williams“Old Irish feda (gen. fedot): a ‘puzzling’ form in the Cambrai Homily and its implications for the apocope of /i/”, in: P. Anreiter & E. Jerem (eds.), Studia Celtica et Indogermanica: Festschrift für Wolfgang Meid zum 70. Geburtstag, Archaeolingua, Budapest 1999: 471-474.
Sims-Williams 2003P. Sims-WilliamsThe Celtic inscriptions of Britain: Phonology and chronology, Blackwell, Oxford 2003.
Sims-Williams 2007P. Sims-WilliamsAncient Celtic place-names in Europe and Asia Minor, Blackwell, Oxford 2007.
Solinas 1995P. Solinas“Il celtico in Italia”, Studi etruschi 60/1995: 311-408.
Stalmaszczyk & Witczak 1995P. Stalmaszczyk & K. T. Witczak“Celto-Slavic language connections: New evidence for Celtic lexical influence upon Proto-Slavic”, Linguistica Baltica 4/1995: 225-232.
Stifter 1998D. Stifter“Study in red”, Die Sprache 40/1998: 202-223.
Stokes 1894W. StokesUrkeltischer Sprachschatz, Göttingen 1894.
Stüber 1998K. StüberThe historical morphology of N-stems in Celtic, The Department of Old Irish, NUI Maynooth, 1998.
Tichy 1993Tichy, E."Kollektiva, Genus femininum und relative Chronologie im Indogermanischen," KZ 106/1993: 1-19.
Tovar 1972-3A. Tovar“Kollektiva auf -r im Keltischen”, Études Celtiques 13/1972-3: 411-428.
Uhlich 1989J. UhlichDOV(A) and lenited -b- in Ogam”, Ériu 40/1989: 129-134.
Uhlich 1993J. Uhlich“Die Reflexe der keltischen Suffixfarianten *-yo- vs. *-iyo- im Altirischen”, in: M. Rockel and S. Zimmer (eds.), Akten des ersten Symposiums deutschsprachiger Keltologen, Narr, Tübingen 1993: 353-370.
Uhlich 1993aDie Morphologie der komponierten Personennamen des Altirischen, Witterschlick, Bonn 1993.
Uhlich 1995J. Uhlich“Altirisch domun ‘Welt, Erde’ und domain ‘tief’”, KZ 108/1995: 278-289.
Uhlich 1999J. Uhlich“Zur sprachlichen Einordnung des Lepontischen”, in: S. Zimmer, R. Ködderitzsch and A. Wigger (eds.), Akten des zweiten deutschen Keltologensymposiums, Max Niemeyer, Tübingen 1999: 277-304.
Untermann 1975J. UntermannMonumenta Linguarum Hispanicarum I, 1-2, Wiesbaden 1975.
Untermann 1989J. Untermannarganto- ‘Silber’ im Keltiberischen”, in: K. Heller et alii (eds.), Indogermanica Europaea: Festschrift für Wolfgang Meid, Institut Für Sprachwissenschaft der Universität Graz, Graz 1989: 431-450.
Villar 1997F. Villar“The Celtiberian language”, Zeitschrift für celtische Philologie 49/50: 898-949.
Villar 2004F. Villar“The Celtic language of the Iberian peninsula”, in: Ph. Baldi and P. U. Dini, Studies in Baltic and Indo-European linguistics in honor of W. R. Schmalstieg, Benjamins, Amsterdam 2004: 243-273.
Vine 1999B. Vine“Greek rhíza ‘root’ and “Schwa Secundum”, UCLA Indo-European Studies 1/1999: 5-29.
Wagner 1967H. Wagner“Zu altir. marnid ‘verraten’”, ZCP 30/1967: 1-6.
Wagner 1971H. WagnerStudies in the Origins of the Celts and of Early Celtic Civilisation, Niemeyer, Tübingen 1971.
Walde-HoffmannA. Walde and J. B. HoffmannLateinisches etymologisches Wörterbuch, Winter, Heidelberg 1934-1950.
Watkins 1959C. Watkins“The etymology of Old Irish ind-aim”, Language 35/1959: 18-20.
Watkins 1962C. WatkinsThe Indo-European origins of the Celtic verb I. The sigmatic aorist, Dublin 1962.
Watkins 1965C. Watkins“Indo-European metrics and archaic Irish verse”, Celtica 6/1965: 194-249.
Watkins 1966C. Watkins“Italo-Celtic revisited”, in: Ancient Indo-European Dialects, ed. by J. Puhvel and E. C. Polomé, UCLA Press, Los Angeles 1966.
Watkins 1970C. Watkins“The etymology of Irish dúan”, Celtica 11/1970: 270-277.
Watkins 1976C. Watkins“Varia I”, Ériu 27/1976: 119-122.
Watkins 1978C. Watkins“Varia III”, Ériu 29/1978: 155-165.
Watkins 1979C. Watkins“Old-Irish saithe, Welsh haid: etymology and metaphor”, Études Celtiques 16/1979: 191-194.
Watkins 1987C. WatkinsTéora ferba fíra”, in: G.Cardona & N. H. Zide (eds.), Festschrift for Henry Hoenigswald. On the occasion of his Seventieth Birthday, Niemeyer, Tübingen 1987: 400-407.
Watkins 1990C. Watkins“A Celtic-Latin-Hittite etymology”, in: T. Abusch et alii (eds.) Lingering over words. Studies in ancient Near Eastern literature in honor of William N. Moran, Scholars Press, Atlanta 1990: 451-453.
Watkins 1993C. Watkins“Another thorny problem”, Linguistica 23, 243-248.
Watkins 1995C. WatkinsHow to kill a dragon in Proto-Indo-European, Oxford University Press, Oxford 1995.
Watkins 2005C. Watkins“The Old Irish Word for ‘Fleshfork’”, in: Heroic Poets and Poetic Heroes in Celtic Tradition, A Festschrift for P. K. Ford, ed. by J. F. Nagy and L. E. Jones, CSANA-Yearbook 3-4, 2005: 377-378.
Weiss 2006M. Weiss“Latin orbis and its cognates”, KZ 119/2006: 250-272.
Widmer 1997P. Widmer“Zwei keltische t-Stämme”, KZ 110/1997: 122-127.
Widmer 2001P. Widmer“Air. tál ‘Zimmermannsaxt”, in: H. Eichner et alii (eds.) Fremd und Eigen. Untersuchungen zu Grammatik und Wortschatz des Uralischen und Indogermanischen in Memoriam Hartmut Katz, Praesens, Vienna 2001: 293-302.
Widmer 2004P. WidmerDas Korn des Weiten Feldes. Interne Derivation, Derivationskette und Flexionklassenhierarchie: Aspekte der nominalen Wortbildung im Urindogermanischen, IBS, Innsbruck 2004.
Willi 2002A. Willi“Old Irish (h)uisse ‘just, fitting’”, Ériu 52/2002: 235-240.
Wodtko et alii 2008D. Wodtko, B. Irslinger, and C. SchneiderNomina im indogermanischen Lexikon, Winter, Heidelberg 2008.
ZCPZeitschrift für Celtische Philologie.
Ziegler 1994S. ZieglerDie Sprache der altirischen Ogam-Inschriften, Vandenhoeck and Ruprecht, Göttingen 1994.
Zimmer 1987S. Zimmer“Three Welsh etymologies: gellyg ‘pears’, ebol ‘colt’, buddelw ‘cowpost’”, Cambridge Medieval Celtic Studies 14/1987: 61-63.
Zimmer 1991“Some Welsh etymologies”, in: L. Isebaert (ed.), Studia Etymologica Indoeuropaea A. J. Van Windekens dicata, Peeters, Louvain 1991: 313-317.
Zimmer 1994S. Zimmer“Zum britischen s-“, Journal of Celtic Linguistics 3, 149-164.
Zimmer 1995S. Zimmer“Indogermanisch *hsu- und *dus- im Kymrischen”, ZCP 47/1995: 176-200.
Zimmer 2000S. ZimmerStudies in Welsh word-formation, DIAS, Dublin 2000.