Hostname: page-component-89b8bd64d-mmrw7 Total loading time: 0 Render date: 2026-05-07T15:00:50.424Z Has data issue: false hasContentIssue false
  • English
  • Français

Mutation in Gaulish

Published online by Cambridge University Press:  22 April 2026

Louis H. Gray
Get access Rights & Permissions [Opens in a new window]

Abstract

[Despite the paucity of material, and notwithstanding the conflicting evidence, it would appear that Gaulish possessed a system of mutation like that of the other Celtic languages; and that this system resembled the Brythonic rather than the Goidelic. Certain analogous phenomena occur in French (and Engadine); but whether these are Gaulish survivals is a question, rather, for Romance scholars.]

Information

Type
Research Article
Information
Language , Volume 20 , Issue 4-Part1: October-December , December 1944 , pp. 223 - 230
Copyright
Copyright © 1944 Linguistic Society of America

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

Article purchase

Temporarily unavailable

References

1 See e.g. R. Thurneysen, Handbuch des Altirischen 68–81, 139–45, 286–8 (Heidelberg, 1909); J. Vendryes, Grammaire du vieil-irlandais 84–6 (Paris, 1908); H. Pedersen, Aspirationen i Irsk (Leipzig, 1897) and KZ 35.315–444 (1899). For Manx, see J. J. Kneen, Grammar of the Manx Language 42–3 (Oxford, 1931).

2 J. Strachan, Introduction to Early Welsh 8–20 (Manchester, 1909); J. Baudiš, Grammar of Early Welsh 1.85, 87, 88, 89, 97, 98, 99, 115 (Oxford, 1924); J. Morris Jones, Welsh Grammar, Historical and Comparative, 161–7 (Oxford, 1913); H. d'Arbois de Jubainville, Études grammaticales sur les langues celtiques 1.61*–82* (Paris, 1881); J. Loth, Vocabulaire vieuxbreton 8–9, 11–20 (Paris, 1884) and Chrestomathie bretonne 75–7 (Paris, 1890); J. Le Gonidec, Grammaire celto-bretonne, nouv. éd., 13–33 (Paris, 1838); L. Le Clerc, Grammaire bretonne du dialecte de Tréguier2 15–28 (St. Brieuc, 1911); H. Jenner, Handbook of the Cornish Language 68–72, 175–7 (London, 1904). For Gaulish phonology in general, R. A. Fowkes, Lang. 16.285–99 (1940). For Celtic mutation as a whole, Pedersen, Vergleichende Grammatik der keltischen Sprachen 1.242, 413–7, 427–76 (Göttingen, 1909).

3 J. MacNeill, Notes on ... Irish Ogham Inscriptions, Proceedings of the Royal Irish Academy, Section C 27.337–8, 340–2; Pedersen, Aspir. 68, and, on Celtic generally, Gram. 1.427–8; Strachan 10, 11; E. Norris, Ancient Cornish Drama 2.225–6 (Oxford, 1859).

4 Loth, Vocabulaire 8–15.

5 Pedersen, Gram. 1.436. The Ogham inscriptions are chiefly of the 5th and 6th centuries (MacNeill 331, 332).

6 G. Dottin, La Langue gauloise 100–1, 124–5, 307 (Paris, 1920).

7 L. Weisgerber, Die Sprache der Festlandkelten, Deutsches archäologisches Institut: Römisch-germanische Kommission 20.217 (1931).

8 Pedersen, Gram. 1.436–7, 533.

9 A. Meillet, review of Dottin, BSLP 22.90–1 (1921).

10 Fowkes 291; L. H. Gray, AJP 63.443 (1942).

11 Revue archéologique 5.13.100 (1921). The same statements hold good for mutation of initial consonants (ibid. 114).

12 Cf. OPers. kāra ‘army, people’, Gk. κοίρανος 'lord,' Goth. harjis, Lith. kãras 'army” etc. (Walde-Pokorny, Etymologisches Wörterbuch der indogermanischen Sprachen 1.462 [Berlin and Leipzig, 1927–32]; Pedersen, Gram. 1.362). For survivals of the alternation c/g in Gaulish loan-words in Romance, J. Jud, Archiv. Roman. 6.192 (1922). Intervocalic c is sometimes written g in VL inscriptions, e.g. fegit (E. Diehl, Vulgärlateinische Inschriften no. 982 [Bonn, 1910]).

13 A. Holder, Alt-Celtischer Sprachschatz 2.1630 (Leipzig, 1896 sqq.); cf. Dottin 65.

14 Holder 1.618–9; cf. MIr. brocc, Breton broc'h ‘badger’, IE *bherek̑e- ‘shine’ (Walde-Pokorny 2.169).

15 Holder 2.1191; for the etymology of catu- see Walde-Pokorny 1.339.

16 Holder 2.712; 1.1331; 3.218–9; for the etymology (cf. OIr. nemed ‘sacellum‘) see Walde-Pokorny 2.330–1.

17 Holder 1.841; cf. also 2.1687.

18 Cf. Skt. kīrtí- ‘fame’ < *qәrәti- etc., IE *qare- ‘praise’ (Walde-Pokorny 1.353), or as Fowkes suggested to me (28 Oct. 1938) : Ionic and Epic κάρτος, Attic κράτος ‘strength, vigor’, Hom. κρατύς ‘strong, mighty’, Goth. hardus, OHG hart, Eng. hard etc., IE *qare- ‘hard, strong’ (Walde-Pokorny 1.354).

19 Dottin 63, note 1, suggests that this may be due to a devoicing of g [which is scarcely probable] or to a façon d'écrire analogous to the OIr. use of gg or c for [g] (cf. Vendryes 25–30). For the interchange cf. also Holder 1.650, 1504; 3.1013. In VL inscriptions c is sometimes miswritten for g, e.g. quinquacinta, coiuci, acente, lece, dicitos, oricinis, vicinti (Diehl nos. 84, 133, 475, 830, 851, 1202, 1345).

20 Cf. Gaulish brogae ‘ager’ (scholion on Juvenal), MIr. mruig, bruig, MnIr. bru(i)gh ‘region’, Lat. margo ‘edge’ (Pedersen, Gram. 1.97).

21 Cf. Gaulish orge ‘occide’, OIr. orgim ‘I kill’, OBret. orgiat ‘caesor’ (miswritten 'caesar'), Armen. harkanem 'I strike', IE *perege- (Walde-Pokorny 2.43).

22 Cf. Skt. sáhas- ‘might’, Gk. ἔχω ‘I have’, ὀχυρός 'firm, strong,' Goth. sigis 'victory', IE *seĝhe- (Walde-Pokorny 2.481–2).

23 Cf. OBret. guerg ‘efficax’; probably cognate with Gk. ἔργον ‘work’, OHG werk, Eng. work etc. (Walde-Pokorny 1.290–1; less probably Pedersen, Gram. 1.105; cf. Weisgerber 212–3).

24 Cf. OIr. traig (gen. traiged) ‘pes’, MWelsh plur. traet, MBret. plur. treit, Corn. plur. treys : Gk. τρέχω ‘I run’, etc.; but with o-vocalism in MnIr. troigh (dat. troigh[idh]), MWelsh troet, MBret. troat, OCorn. truit : Gk. τροχός ‘wheel’, etc. (Pedersen, Gram. 1.39, 98, 522, 533; 2.201; Baudiš 93; H. Zimmer, KZ 32.232–4 [1893]); Holder 2.1901; less probably Walde-Pokorny 2.752–3, 874–5).

25 With -magos cf. OIr. mag, MnIr. magh ‘plain’; cf. Walde-Pokorny 2.258 (otherwise Pedersen, Gram. 1.96); Holder 2.384–5; with rīgo- cf. Skt. rā́j-, Lat. reg- ‘king’, etc. (Walde-Pokorny 2.362–5).

26 For the second component cf. treide, above (10), and note 24; also Holder 2.1910; Meillet 90–1.

27 W. Meyer-Lübke, Das Katalanische 162 (Heidelberg, 1925); cf. MIr. brí (acc. brig n-) ‘hill’, Welsh bry ‘high, above’, Skt. bṛhánt- ‘high’, Goth. baírgs, OHG berg ‘hill’, etc. (Walde-Pokorny 2.172–4; Pedersen, Gram. 1.100; Weisgerber 185, 195).

28 Cf. OIr. bith ‘world’, Skt. jīvá-, Lat. vīta ‘life’, etc. (Walde-Pokorny 1.668–70). Change of b to v is very frequent in VL inscriptions. For Brythonic [vr] < br, see Pedersen, Gram. 1.116–7; and for IE *dhubhro-, PC *dubro-, see Walde-Pokorny 1.840; Holder 1.1362; 3.224.

29 MIr. cob ‘victory’, OIcel. happ ‘luck’, OCS kobŭ ‘fortune, guardian spirit’ (Walde-Pokorny 1.457–8; cf. Holder 1.1054r-5; 3.1245–6). For syncopation in Celtic compounds, see Pedersen, Gram. 2.2–3.

30 Cf. Lat. fascia ‘band, bandage’, etc., IE *bhasqo- (Walde-Pokorny 2.135–6).

31 With dubno- cf. OIr. domun, MnIr. domhan ‘world’, OIr domain, MnIr. doimhin, Welsh dwfn ‘deep’ < *dhubni-, Goth. diups, OHG tief, Eng. deep, Lith. dubùs, etc., IE *dheube-(Walde-Pokorny 1.847–8; 2.190; Pedersen, Gram. 1.35); for Italic bn, Stolz-Schmalz, Lateinische Grammatik5 155 (by Leumann-Hofmann; Munich, 1928); C. D. Buck, Grammar of Oscan and Umbrian, rev. ed., 79 (Boston, [1928]); for Brythonic fn [vn] < bn, Pedersen, Gram. 1.117.

32 Walde-Pokorny 2.588; Pedersen, Gram. 1.118; Walde-Hofmann, Lateinisches etymologisches Wörterbuch, 1.171 (Heidelberg, 1928); Meyer-Lübke, Romanisches etymologisches Wörterbuch3 §1710 (Heidelberg, 1935); Jud 194–6. In VL inscriptions p is occasionally mis-written for b, as sipi, depetis (Diehl nos. 406, 1537).

33 Cf. Skt. áśman- ‘stone’, Gk. ἄκμων ‘anvil’, Lith. akmuõ, OCS kamy ‘stone’, etc. For the Brythonic mutation fn [vn] < mn, as in Welsh safn ‘jaw’ : Gk. στόματος (gen.) < *στομṇτος, Avestan staman- ‘mouth’, see Pedersen, Gram. 1.78, 167 (cf. Walde-Pokorny 2.648); and for rf [rv] < rm, as in Welsh cwrf ‘beer’ (above [17]), and note 35), see Pedersen 1.168.

34 It is uncertain whether the base of the word is *gu̯here- ‘hot’ or *bhere- ‘boil’, but in either case it contains the formative -mo- (cf. Walde-Pokorny 1.687–9, 2.158; Pedersen, Gram. 1.168; Holder 1.491, 493; 3.912, 914). For Romance survivals of the alternation Bormo/Borvo, see Jud 204.

35 Cf. MIr. coirm, cuirm, Welsh cwrw(f), cwrf, cwyrf, OCorn. coref ‘beer’, Lat. cremor ‘thick juice or broth’, Skt. karamb(h)á- ‘gruel’, kulmāṣa- ‘sour gruel’; from Lat. cervisia come Fr. cervoise, Sp. cerveza ‘beer’, etc. (Walde-Pokorny 1.419; Pedersen, Gram. 1.168; Walde-Hofmann 1.207, 287–8; Holder 1.995–8, 1202–3; Meyer-Lübke §1830).

36 Pedersen, Gram. 1.117, 165, 533; Holder 1.879–82.

37 Cf. Skt. srávati, Gk. ῥἐω ‘flow’, etc. (Dottin 99; Pedersen, Gram. 1.82, 537; Holder 1.1500–1; the assumption of a pre-form *sprutu- [Walde-Pokorny 2.703] seems unnecessary). For survivals of Φρουδις in Gallo-Romance dialects, see Meyer-Lübke ZRPh. 20.531–2 (1896) and Et. Wb. §3545.

38 Dottin 66, 99; Pedersen, Gram. 1.69; Holder 2.1372, 1457–8; Meyer-Lübke §7789. In Brythonic, IE initial prevocalic s regularly > h (probably as a result of generalized sandhi-mutation; Pedersen, Gram. 1.71).

39 Pedersen, Gram. 1.82, 96–7, 117, 168; Baudiš 87, 89, 93, 115; cf. also Vendryes, La Position linguistique du celtique 6, Proceedings of the British Academy 23 (1937) : ‘Dans l'ensemble, la phonétique du celtique continental s'accorde avec celle du brittonique. Quelques traits seulement l'en distinguent, comme le traitement du groupe initial wr-, représenté par br- en celtique continental alors qu'en brittonique il passe à gr- (par l'intermédiaire de gwr-).‘ Cf. also Pedersen, Gram. 1.59–60; Baudiš 150; E. Windisch in G. Gröber, Gdr. d. rom. Phil. 12.398 (Strasbourg, 1904–6).

40 M. A. Pei, Italian Language 13 (New York, 1941); W. J. Entwistle, Spanish Language 40 (London, [1936]); Holder 2.1094–6; Conway-Whatmough, Prae-Italic Dialects of Italy 1.87, 90 (Cambridge, Mass., 1933); Meyer-Lübke §2648; Dottin 64–5, 77. On the whole problem, see E. Wechssler, Giebt es Lautgesetzte? 100–11 (Halle, 1900); Meyer-Lübke, Einführung in das Studium der romanischen Sprachen3 227–36 (Heidelberg, 1920); Windisch 396–404; M. K. Pope, From Latin to Modern French 6, 136–40 (Manchester, 1934); Vendryes, Celtique et roman, Rev. de Ling. Romane 1.262–7 (1925).

41 MnFr. lac, as my friend and colleague Henri F. Muller points out, implies a Lat. *laccus, influenced by laccus < Gk. λάκκος ‘pond, cistern, pit’ (cf. A. Grauer, Les Consonnes géminées en latin 23, 61 [Paris, 1925]). A further factor was, perhaps, the presence of OFr. lai, lay ‘song, poem’, as well as of lai (< Lat. laicus < Gk. λαικός) ‘nonclerical’ (suggested to me by Mrs. Charles O. Fiertz, 19 March 1944).

42 Cf. F. Haag, Vergleichung des Prakrit mit den romanischen Sprachen (Berlin, 1869); for the general principles involved, E. Sievers, Grundzüge der Phonetik5 283–6 (Leipzig, 1901); M. Grammont, Traité de phonétique 200–2 (Paris, 1932; especially 200: ‘Si la consonne est sourde, les voyelles qui l'entourent tendent à lui donner la sonorité: d'une voyelle à l'autre les cordes vocales tendent à ne pas cesser de vibrer, phénomène de moindre action. Si la consonne est déjà sonore, les voyelles, qui ont toujours plus d'aperture que n'importe quelle consonne, tendent à augmenter l'aperture de la consonne; c'est encore de la moindre action‘).

43 Cf. M. Dillon, Germanic and Celtic, JEGP 42.492–8 (1943); Italic and Celtic, AJP 65.124–34 (1944).