BMCR 2007.11.05

The Etymologies of Isidore of Seville. With the collaboration of M. Hall

, , The etymologies of Isidore of Seville. Cambridge: Cambridge University Press, 2006. 1 online resource (xii, 475 pages). ISBN 0521837499 $150.00.

Table of Contents

The Etymologiae or Origines of Isidore of Seville, composed in the first half of the seventh century, are what we would call today an encyclopedic dictionary, though one in which the attempt to provide accurate, informative definitions is not the overriding concern, and the study of a word’s origin or etymology, which gives the work its title, is often more important than its definition. The work’s twenty books start from the seven liberal arts (grammar, rhetoric, dialectic, arithmetic, music, geometry, and astronomy, Bks. 1-3), then broaden in scope to include other spheres of human activity (medicine, Bk. 4; law, Bk. 5) and of the natural and mineral world (e. g. the sky, the earth and other geographical topics, Bks. 13 and 14; the animals in Bk. 12; gems, stones and the metals in Bk. 16). Bks. 6-8 are devoted to Christian subjects, spanning from the Scriptures to liturgy and heresy. The work was described by Isidore himself as unfinished (some lemmata, for example, are left without an explanation).

The Etymologies of Isidore of Seville is a co-operative project by an academic team of unofficial Isidore specialists, some of them not even classical scholars in a formal sense. The final result proves this not to have been a problem, and the team has succeeded in producing an accurate, literal translation, which contributes significantly to the field of Isidorian studies. The new translation is based on Lindsay’s 1911 OCT, but takes account of editorial progress made by single-book editions appeared after Lindsay, mainly the volumes devoted to Isidore in the Belles Lettres ALMA series (“Auteurs Latins du Moyen Âge”). The authors make no claim to independence in textual matters, but I have found some helpful and convincing discussions of problems of the Latin text in the footnotes.1 No facing text is provided, but Latin and Greek words are quoted in round parentheses when they are necessary to clarify an etymology, or some other argument impossible to grasp without the original.

The introduction, in addition to providing adequate information about the historical and biographical background of the author, and about Isidore’s other works, (4-10) dwells on what we know about Isidore’s sources, his working methods, and his purposes in producing this encyclopedia (10-24).

A complete picture of Isidore’s use of sources has not yet been drawn, but it is easy to prove that Isidore to a large extent merely transcribed, often verbatim, what he read in Pliny, the Vergil commentary of Servius, Solinus’ paradoxographica, the Latin grammatical writers and lexicographers, and others. My only reservation on this section of the introduction is that the authors are too wary of giving credit to Isidore for independent observation. For one thing, Isidore’s work has preserved a great number of vocabulary items not known from any other source, in part relating to the Latin spoken in the region where he was active, Baetica. This is especially true of Books 15-17 and 19-20, for which Roman encyclopedic handbooks such as Varro’s and Pliny’s had less to offer. In particular, it would be worth mapping out the overlap of the latter books (16-20, dealing with vocabulary from non-intellectual spheres of human activity, such as wool-working, agriculture, masonry, sailing) with the bilingual glossaries collected in the third volume of the Corpus glossariorum Latinorum, which are full of everyday terms. The relationship between Etymologiae and CGL seems closer than that with the more erudite, archaizing lexica of Nonius Marcellus and Verrius-Festus.

Isidore, for example, is our only source for the Latin word for ‘loom’, telaria, if that is indeed the meaning of 19.29.1 (p. 389) ‘Cloth ( tela) is named for the length of its threads (cf. τηλοῦ, “afar”). The word ‘loom’ ( telaria) is derived from it’, tela pro longitudine staminum dicta, cuius diriuatiuum est telaria. Du Cange’s medieval Latin dictionary has the form telarium (post XI century), but in Meyer’s Lübke’s REW the Latin form is absent. Classical sources only have tela. We know now that the Latin word is found in a near-contemporary text, the Vita Sanctae Brigidae, by Cogitosus of Kildare (7th-8th century; PL 72.785D Migne), incidens ligna telaria, in quibus textura telarum operabatur, in pastu ignis, where the Saintly woman cuts to pieces her precious loom to light up a fire and entertain some guests.

Isidore is also the only evidence for talaria as the name of an everyday footwear (in OLD only for the ‘winged sandals’ of Mercury in Verg. Aen. 4.239): 19.33.7 ‘ talares (i. e. talaria) are slippers ( soccus) that seem to be so named because they are so shaped that they come to the ankles ( talus)’, talares calcei socci sunt, qui inde nominati uidentur quod ea figura sint ut contingant talum. The existence of this word has been confirmed by the fourth-century “Celtis glossary” partially published by C. Dionisotti in JRS 72 (1982), 102, l. 55 where we find calligulas = ταλάρια.2

The translation is accurate, the book well produced, and misprints or factual errors very few indeed. The brevity of the following list is a tribute to the accuracy and scholarship of the translators.3

P. 49 (1.18.5) ‘the grave accent can occur with another single accent in a single word, but never with two’ ( grauis accentus cum uno accentu poni potest in dictione una, cum utrisque numquam). The reference, in the footnote, to ‘the rule in Greek that a word may have two accents if it is followed by an enclitic’ seems unhelpful, and I cannot find that rule in Donatus, 4.371-2 Keil. Perhaps it should be made clear that utrisque refers to both types of accents discussed so far, the acute and the circumflex.

P. 51 (1.21.14) diple περὶ στίχον‘with a point’ — not ‘next to a line’?

P. 126 (5.30.11) consuetudo may be ‘prevailing practice’, but one should be aware that this is also a standard term for ‘current linguistic usage’ — a very interesting remark, because Isidore is discussing the persistent, and eventually prevailing, habit of using pagan names for the days of the week.

P. 183 (8.10.3) ethnici ex Graeco in Latinum interpretantur gentiles, ‘the Latin word gentiles is translated as ethnici in Greek’: Isidore sometimes suggests that Greek has borrowed a word from Latin, but here the direction of the translation is from Greek to Latin.

P. 388 toral should be torale.

P. 384 χειροδύτη should be χειροδύτην).

P. 393 gallicula is not ‘a small Gallic shoe’, but a corruption of caligula, from caliga, although clearly influenced by a popular etymology from ‘Gallic’.

On a more general level, I am puzzled by the frequent rendering of Latini with ‘Latin speakers’, whereas Graeci is mostly simply ‘Greeks’ (e. g. p. 39 (1.3.4) Graeci uero uiginti quattuor (sc. elementa). Latini enim inter utramque linguam progredientes uiginti tria elementa habent, ‘the Greeks use twenty-four (sc. characters). Latin speakers, falling between these two languages, have twenty-three characters). Perhaps ‘Latins’ sounds strange in English, but the distinction seems un-Isidorian. By Isidore’s time, the term had acquired the wider meaning of ‘Westerners’, as opposed to the Byzantines.

In a continuous reading of the Etymologies, which an accessible, attractive translation such as this one may encourage, individual readers are in for all kinds of little surprises and discoveries about the ancient world, and it is impossible, though perhaps unfair to the authors, not to wish that more notes had been provided. To give just one example, one reads in a section on time reckoning that the Romans calculated the day from midnight (p. 126, 5.30.4). Those who are used to associating hora nona with three p.m. will be surprised, and it would have been useful to be reminded that the midnight-system coexisted with the unofficial dawn-system, in which hours were counted from sunrise.

In a few cases the translation is unclear without a note.

P. 49 (1.18.3): the phrase ‘ unde (“whence”) is grave here (i. e. its pitch lowers as we move from the first to the second syllable)’ is incomprehensible without comment. The point is, I think, that some grammarians argued that the some relative adverbs were not accented (= grauari; cf. Prisc. 3.83.21 Keil hoc quoque, quomodo omnia infinita, id est ‘quo, ubi, unde’, [‘qua’] quando relatiuum est, grauatur, aliter suum accentum seruat).

P. 53 (1.27.19, on orthography) ‘ pene, which is a conjunction, [is spelled] with e’: paene, besides being spelled with ae, is unanimously an adverb: perhaps read penes, unless this is an oversight on the part of Isidore.


1. See e. g. p. 390 n. 22, where the MS reading cicadas, a word for a type of gold hairpin, is defended against Lindsay’s and Rodri/guez-Pantoja’s cycladas.

2. Isidore is also the only source tracing a brief history of Latin (9.1.6); he calls contemporary Latin the ‘mixed’ language, lingua mixta. Some examples of words for which Isidore draws on contemporary evolved or ‘vulgar’ Latin are: 20.9.4 mozicia, for a type of vessel whose name is said to be derived from modicus, ‘small’, with an evolved pronunciation typical of the Italians; 15.12.2 capanna for ‘hut’; 15.15.5 actum [i. e. the furrow-measure] prouinciae Baeticae rustici acnuam uocant. The stadard work on the vulgarisms in Isidore’s Etymologiae is J. Sofer, Lateinisches und Romanisches aus den Etymologiae des Isidorus von Sevilla (Göttingen, 1930).

3. The belatedness with which I have produced this review has at least enabled me to take advantage of the excellent assessment of the same book by Ana-Isabel Magallón Garcia, published in BMCR’s electronic twin, TMR, 07.05.30. Magallón Garcia, who is an expert Isidore scholar, has pointed out some omissions in the bibliography, mainly of Spanish publications, and of everything published after 2003. It may be useful to mention the appearance of yet another (Italian) translation, by Angelo Valastro Canale (Torino: UTET, 2004), which has a facing Latin text, and, at Euro 55.40 for two (paperback) volumes, comes at a very convenient price.