The Etymology of Irish ‘liäig’ (Leech, Doctor, Physician)

David Stifter

Fig. 1: Physician letting blood from a patient. Attributed to Aldobrandino of Siena, Li Livres dou Santé, in London, British Library MS Sloane 2435 (France, c. 1270)

Internal Irish and Celtic reconstruction

The most common term for a ‘leech, doctor, physician’ in early Irish texts, namely OIr. liäig, liëig, gen. lego, lega, must originally have been disyllabic. To judge from the entry in eDIL and from the impressionistic note by Hamp 1984, there is no decisive metrical evidence for its disyllabic nature. However, the spelling alternation between the stressed -ia- in the nominative and short -e- in those cases where an extra syllable is added, is typical of words with hiatus and is therefore ortho­graphic evidence for it (McCone 1996: 130, 135; 1997).

The first vowel of liäig must historically have been i or e, the second was probably a back vowel, *a or *o (short or long). A back vowel is implied by the fact that it caused the lowered -e- of the syncopated allomorph of the word. Only these combinations are compatible with the alternation –ia- and -e- in stressed position. The hiatus in liäig presupposes a Proto-Celtic con­sonant between the two vowels that was lost in Early Irish. This lost sound could have been Primitive Irish *h < PC *s, *i̯, or PC *φ < PIE *p. These sounds disappeared at different stages in the prehistory of the language, but they all left a trace in the form of a hiatus in Old Irish. In the absence of known cognates in the other Celtic languages, the internal Proto-Celtic recon­struc­tion, based entirely on Irish, results in the rather unwieldy form *l{e/i}{s/φ/i̭}{a/ā/o/ō}gi-.

External comparison

Fortunately, the word has an apparent correspondence in the synonymous Proto-Germanic *lēkja-, which allows us to narrow down some of the options. This is continued in Gothic lekeis ‘doctor’, Old Norse lækir ‘doctor’, Old English lǣce ‘doctor, leech’, Old Frisian lētza ‘doctor’, Old Dutch lake ‘leech’, Old High German lāhhi, lāchi ‘leech’ (Kroonen 2013: 331). It is usually assumed that it is a borrowing from Proto-Celtic or an ancient Celtic language before the oper­ation of Grimm’s Law in Germanic, i.e. predating the middle of the 1^st millennium B.C. How­ever, Kroonen (l.c.) makes an alternative proposal for a Germanic-internal explanation of the word: under the assumption that the original meaning was ‘leech, blood-letter’, it could be understood as a derivative from PG *lekan- ‘to leak, drip’; cf. the formally very similar *lēkjōn- ‘rivulet (vel sim.)’. In that case the similarity with the Irish word would be entirely accidental and the two words would not be related with each other.

If, on the other hand, Proto-Germanic *lēkja- is a Celtic loanword, a preform *le.egi- would be op­ti­mal (Van Sluis et al. 2023: 201, 218). If the two words are connect­ed, their comparison limits the Proto-Celtic recon­struction to *le.agi- or le.ogi-. Since both *s and *i̭ were fully re­tained in Proto-Celtic and in the ancient Celtic languages, and since both of these sounds would have been borrowed into Ger­manic, the only remaining reconstruction is PC *leφagi- or *leφogi- < pre-Celtic *lepagi- or *lepogi- in which the loss of *φ < PIE *p had already occur­red. The *ē (< *e.e) of Proto-Germanic must then be due to phonological adaptation during the borrowing. A borrowing in the opposite direction, i.e. from Proto-Germanic into Celtic, is formally improbable.

Pre-Celtic *lepagi- (less likely *lepogi-) can be analysed as a compound with the second element *‑agi- (or *-ogi-) from the Indo-European roots ¹*h₂eg̑- ‘to drive, impel’ (LIV 255–256), i.e. zero-grade *h₂g̑-i- (thus Irslinger in NIL 267) or full-grade *h₂eg̑-i-, but hardly o-grade *h₂og̑-i-. The homonymous root ²*h₂eg̑- ‘to say’ (LIV 256) is usually not considered. The formation is reminiscent of Sanskrit bhiṣáj- ‘doctor’, Young Avestan bišaziia- ‘to cure, heal’ < Indo-Iranian *bʰiš-aj́-, a compound of *bʰiš- ‘cure, remedy’ + *aj́- <  *h₂eg̑- (Mayrhofer 1996: 264). The Indo-Iranian word could thus either refer to some­one who ‘drives a cure’ or who ‘says, utters words of healing’. The etymology of the first ele­ment *bʰiš- ‘cure, remedy’ is unclear, but it could be connected with Indic *bhāṣ- ‘to speak’ < Indo-Iranian *bʰā-s- ‘to speak’ < PIE *bʰeh₂- ‘to speak, say’ (LIV 69–70), i.e. Iir. *bʰiš- < *bʰh₂-s- ‘to speak to, charm, hex’ (Mayrhofer 1996: 262, 264; NIL 269 n. 2).

Likewise, the first part of the Celtic word *lepagi- is unclear. It could contain the root *lep- ‘to peel off’ (LIV 412), but the semantic motivation remains obscure. However, Po­korny (IEW 677) operates with a separate root lep- ‘to chat, talk, etc.’, which has not been adopted in LIV, but from which he derives OIr. līaig (sic) via the preform “*lēpagi- ‘Besprecher’” (i.e. ‘char­mer’), i.e. semantically very similar to the Indo-Iranian word. Such a preform, which would appear as *līφagi- > *lī.agi- in Proto-Celtic and in ancient Celtic languages, is not com­patible with the assumed Germanic loanword, but oper­ating with full-grade *leφagi- < *lep-h₂g̑-i- in­stead of lengthened grade *lēpagi- (i.e. lēp-h₂g̑-i- in modern notation) would produce the desired result.

Irslinger’s explanation of liäig as *h₂lei̭(H)-h₂g̑i- from the root *h₂lei̭(H)- ‘to smear’ (LIV 277), i.e. ‘he who drives/does smearing’, is not compatible with the form of the word in the assumed borrowing into Germanic.

Conclusion

While no explanation emerges as immediately self-evident, it is overall likely that an agen­tive compound of the Indo-European root *h₂eg̑- ‘to drive’ lies at the heart of OIr. liäig, which has semantic and formal parallels in other languages. Formally the most promising can­didate is a variant of Pokorny’s suggestion, namely *leφagi- < *lep-h₂g̑-i- in the sense of ‘char­mer, some­one who speaks unto a disease’, however a question mark remains over the status of the postul­ated root *lep-. It is hoped that further investigation into the historical phonology and morphology of OIr. liäig and Proto-Germanic *lēkja– will shed more light on their origin and original meaning.

Further reading:

• Hamp, Eric P. (1984), ‘Varia III. 1. liaig ‘Physician’, Ériu 35: 200
• IEW = Julius Pokorny (1959), Indogermanisches Etymologisches Wörterbuch (Tübingen–Basel: Francke)
• Kroonen, Guus (2013), Etymological Dictionary of Proto-Germanic [= Leiden Indo-European Etymological Dictionaries 11] (Leiden–Boston: Brill)
• LIV = Martin Kümmel und Helmut Rix (2001), LIV. Lexikon der indogermanischen Verben. Die Wurzeln und ihre Primärstammbildungen. Zweite, erweiterte und verbesserte Auflage (Wies­baden: Reichert)
• Mayrhofer, Manfred (1996), Etymologisches Wörterbuch des Altindoarischen. II. Band (Heidelberg: Universitätsverlag Winter)
• McCone, Kim (1996), Towards a Relative Chronology of Ancient and Medieval Celtic Sound Changes [= Maynooth Studies in Celtic Linguistics 1] (Maynooth: Department of Old and Middle Irish, St. Patrick’s College, Maynooth)
• McCone, Kim (1997), ‘Léic úait inna biada milsi (Wb. 6c7) ‘Put from You the Sweet Foods’, in Sound Law and Analogy. Papers in Honor of Robert S. P. Beekes on the Occasion of his 60^th Birthday, ed. by Alexander Lubotsky (Amsterdam: Rodopi), pp. 171–5
• NIL = Dagmar S. Wodtko, Britta Irslinger, Carolin Schneider (2008), Nomina im Indogermanischen Lexikon (Heidelberg: Universitätsverlag Winter)
• van Sluis, Paulus, Anders Richardt Jørgensen and Guus Kroonen (2023), ‘European Prehistory between Celtic and Germanic: The Celto-Germanic Isoglosses Revisited’, in The Indo-European Puzzle Revisited. Integrating Archaeology, Genetics, and Linguistics, ed. by Kristian Kris­tiansen, Guus Kroonen and Eske Willerslev (Cambridge: Cambridge University Press), pp. 193–243